This content has been downloaded from IOPscience. Please scroll down to see the full text.

Download details:

This content was downloaded by: jnam27

IP Address: 169.235.18.60

This content was downloaded on 25/02/2014 at 01:55

Please note that terms and conditions apply.

Polyaniline/poly(-caprolactone) composite electrospun nanofiber-based gas sensors:

optimization of sensing properties by dopants and doping concentration

View the table of contents for this issue, or go to the journal homepage for more

2014 Nanotechnology 25 115501

(http://iopscience.iop.org/0957-4484/25/11/115501)

Home Search Collections Journals About Contact us My IOPscience

Nanotechnology

Nanotechnology 25 (2014) 115501 (10pp) doi:10.1088/0957-4484/25/11/115501

Polyaniline/poly(ε-caprolactone) compositeelectrospun nanofiber-based gas sensors:optimization of sensing properties bydopants and doping concentrationKaren Low1,4, Nicha Chartuprayoon2,4, Cristina Echeverria3, Changling Li2,Wayne Bosze2, Nosang V Myung2 and Jin Nam1,5

1 Department of Bioengineering, University of California-Riverside, CA 92521, USA2 Department of Chemical and Environmental Engineering, University of California-Riverside,CA 92521, USA3 Department of Civil and Environmental Engineering, University of California-Los Angeles,CA 90095, USA

E-mail: jnam@engr.ucr.edu

Received 25 October 2013, revised 19 December 2013Accepted for publication 14 January 2014Published 24 February 2014

AbstractElectrospinning was utilized to synthesize a polyaniline (PANI)/poly(ε-caprolactone) (PCL)composite in the form of nanofibers to examine its gas sensing performance. Electricalconductivity of the composite nanofibers was tailored by secondary doping with protonic acidsincluding hydrochloride (HCl) or camphorsulfonic acid (HCSA). FT-IR and diffuse reflectanceUV–vis spectroscopy were utilized to examine doping-dependent changes in the chemicalstructure and the protonation state of the nanofibers, respectively. The oxidation and protonationstate of the composite nanofibers were shown to strongly depend on the doping agent andduration, demonstrating a simple way of controlling the electrical conductivity of the composite.PANI/PCL electrospun nanofibers having various electrical conductivities via varying dopants anddoping concentrations, were configured to chemiresistors for sensing various analytes, includingwater vapor, NH3, and NO2. Secondary doping with Cl− and CSA differentially affected sensingbehaviors by having distinctive optimal sensitivities. Biphasic sensitivity with respect to electricalconductivity was observed, demonstrating a facile method to enhance gas sensitivity byoptimizing secondary doping. A balance between Debye length of the nanofibers and overallcharge conduction may play an important role for modulating such an optimal sensitivity.

Keywords: electrospinning, gas sensor, composite nanofibers, polyaniline, poly(ε-caprolactone)

S Online supplementary data available from stacks.iop.org/Nano/25/115501/mmedia

(Some figures may appear in colour only in the online journal)

1. Introduction

The increasing need for sensors in various applications rangingfrom agriculture, automotive industry, medical diagnosis and

4 Authors made equal contribution.5 Address for correspondence: Department of Bioengineering, University ofCalifornia-Riverside, Materials Science and Engineering Building 331, 900University Avenue, Riverside, CA 92521, USA.

public/national security demands the development of sensi-tive, reliable, fast responsive, and inexpensive chemical andbiological sensors [1]. Solid-state sensors have been soughtafter due to their ability to alter their electrical properties(e.g., electrical resistance and capacitance) upon exposureto analytes. Among the variety of sensing materials usedfor solid-state sensors, conducting polymers (CPs) such aspolyaniline (PANI) [2], polypyrrole (PPy) [3], or polythio-

0957-4484/14/115501+10$33.00 1 c© 2014 IOP Publishing Ltd Printed in the UK

Nanotechnology 25 (2014) 115501 K Low et al

phene (PT) [4] have shown a promising potential as gassensing materials because of their excellent physical, chemical,electrical and material properties including tunable electricalbehavior by dopants and doping levels, diverse monomerchemistry, and ease of functionalization [5]. More importantly,their low operating temperature and power consumption areenticing to configure them for various types of sensors such asamperometric and chemiresistive/conductometric devices [5].

CPs have some physical and chemical limitations, pre-senting difficulties in manufacturing [6]. In this regard, vari-ous polymer hosts have been used to improve processability,resulting in a composite that combines the excellent electricalproperties of the CPs with the robust mechanical properties ofthe polymer host [6]. In addition to enhanced processability,introduction of polymer hosts provide an opportunity to tailorthe gas sensing performance by optimizing the ratio of CPsdispersed in the host polymer [7]. The advantages of thesecomposites include tunable analyte selectivity, lower detec-tion limits, fast response time, and improved environmentalstability [7].

One-dimensional (1D) nanostructures, such as nanowires,nanorods, nanobelts, and nanotubes, have emerged as attrac-tive platforms for the fabrication of these gas sensors. Theyexhibit properties that their bulk material counterparts lack,such as a large surface-to-volume ratio for greater interactionbetween surface and analyte, two-dimensional diffusion, andefficient electron transport to generate fast responses and quickrecovery [8–11]. Fabrication strategies, including solution orvapor-phase approaches [12], template-directed methods [13],solvothermal synthesis [14], and self-assembly methods [15]have been employed to produce 1D CP composite nanos-tructures. Although these methods enable tight dimensionalcontrol, they typically require multi-step fabrication resultingin a low manufacturability. In comparison, electrospinningprovides an alternative approach to mass-produce ultra-long1D composite nanofibers in a facile and cost-effective man-ner [16]. The feasibility of tuning the properties of the gassensor by altering electrospinning parameters that control fiberdimensions and composition and by introducing secondaryprotonic dopants that modulate electrical properties, makeselectrospinning an attractive method for the development of ahighly sensitive 1D nanostructured solid-state sensor towardsvarious gases.

In this study, PANI/poly(ε-caprolactone) (PCL) compos-ite nanofiber mats were bulk-synthesized and configured in achemiresistor for gas sensing studies. Because of its insolubil-ity in water, PCL was selected as the insulating polymer host toprovide structural stability in humid air. The composite offersa competent platform to investigate the effects of secondaryprotonic doping to enhance the sensitivity of electrospunPANI/PCL composite nanofibers.

2. Experimental methods

2.1. Synthesis of PANI/PCL composite nanofibers

A 5 wt% solution of PCL (Mw = 65 000, Sigma-Aldrich,St Louis, MO) was dissolved in 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) (Oakwood Products, Inc., Indianapo-lis, IN). PANI (3 mg ml−1)-camphorsulfonic acid (HCSA,

3 mg ml−1) solution was prepared by first adding (1R)-(–)-10-camphorsulfonic acid (Sigma-Aldrich, St Louis, MO) in HFIPfollowed by rigorous stirring at 1200 rpm to fully dissolveHCSA. Then, PANI in the emeraldine base form (Mw =

100 000, Sigma-Aldrich) was dispersed into the solution. Themixed solution was sonicated for 5 min and stirred at 1200 rpmfor 5 min at room temperature. The sonication/stirring processwas repeated four times, and the solution was stirred for 4 hin order to further disperse PANI. Subsequently, the PANI–HCSA solution was filtered with a 0.2µm pore size membranefilter (Whatman, GE Healthcare Life Sciences, NJ) by vacuumfiltration to remove undispersed particles. The filtered PANIwas mixed with the PCL solution at a 4:1 volume ratio.

To synthesize nanofibers, the PANI/PCL solution waselectrospun using a high voltage DC power supply (GlassmanHigh Voltage, NJ) at −19 kV with a flow rate of 0.4 ml h−1

and a 30 cm tip to plane collector distance to produce anapproximately 2µm-thick fiber mat. The thickness of the fibermat was controlled by adjusting the deposition time.

2.2. Secondary doping by hydrochloric acid (HCl) and HCSA

To investigate the effects of dopants and doping concentra-tion on the electrical resistivity and sensing performance ofPANI/PCL composite nanofiber, the fiber mats were separatelyincubated with a solution of 1 M HCl, 1 M HCSA, or 1 MNaOH for various durations ranging from 1 min to 12 h. Afterthe doping process, the solution was aspirated out and thefiber mats were rinsed with DI water, followed by air dryingfor 1 h. The fiber mats were subsequently characterized for itschemical and electrical properties.

2.3. Characterization of PANI solution and PANI/PCLnanofibers

To determine PANI loss during the filtration process, thePANI concentration after filtration was quantified using a UV–vis spectrometer (DU 800, Beckman Coulter, CA). Variousconcentrations of unfiltered PANI solution were prepared,and their absorbance spectra were obtained from the UV–visspectrometer by scanning the solutions at the wavelengthranging from 200 to 1100 nm to generate a calibration curve.The absorption spectrum of the filtered solution was comparedto the calibration curve to quantify PANI content.

Fourier-transform infrared spectroscopy (FT-IR) using theEquinox 55 FT-IR (Bruker, Billerica, MA) was utilized tocharacterize the chemical structures of the as-prepared anddoped PANI/PCL composite nanofibers. The mat was placedon a highly polished NaCl substrate and measured with 32scans at a resolution of 0.5 cm−1. The transmission spectrawere determined in the wavenumber range between 600 and3000 cm−1.

In order to determine doping states, diffuse reflectanceUV–vis spectroscopy was used to measure the absorbance ofthe as-prepared, secondary doped PANI/PCL composite, andpure PCL nanofiber mats using a UV–vis Spectrometer (Evo-lution 300, Thermo Scientific, Waltham, MA). The absorbancespectra were measured in the wavelength range between 190and 1100 nm.

2

Nanotechnology 25 (2014) 115501 K Low et al

Figure 1. A representative image of the gas sensor.

The morphology and fiber diameter of the fiber matswere characterized using scanning electron microscopy (SEM,Nova NanoSEM450, FEI, Hillsboro, Oregon). The compositenanofibers were sputter-coated with platinum–palladium tovisualize under SEM.

2.4. Characterization of electrical properties

To determine the electrical property of nanofibers, a 4 mm×6 mm sized strip of nanofiber mat with a controlled thick-ness of approximately 2 µm was placed onto 500 µm-thickoxidized silicon substrate. Using an aluminum stencil witha gap size of 1 mm× 8 mm, gold microelectrodes with thethickness of approximately 300 nm were deposited onto thesample by sputtering using the EMS 575X sputter (ElectronMicroscopy Science, Hatfield, PA). The electrical character-ization of PANI/PCL fiber mat was obtained from probingsource and drain microelectrodes by sweeping the potential−0.5 to 0.5 V to generate I –V curves using a source-meter(Keithley 2363, Cleveland, OH). The electrical conductivityof PANI/PCL composite nanofibers was determined by nor-malizing it to the thickness of the sample.

2.5. Device fabrication

Figure 1 shows the sensing device as fabricated. The de-vice was connected to the sensing system as previously re-ported [17]. Briefly, the fiber mat integrated with gold micro-electrodes was mounted onto a sample holder and enclosed bya glass chamber with the volume of 3.15 cm3 with a gas inletand outlet ports for gas flow. The chip was then clipped to aKeithly source-meter to establish an electrical connection.

2.6. Gas sensing measurement

The doped PANI/PCL nanofibers were examined againstvarious gases including water (H2O) vapor, ammonia (NH3),and nitrogen dioxide (NO2) at room temperature with a custombuilt gas sensing system with the Alicat mass flow controllers(MFC) operated by LabVIEW as described elsewhere [8].Prior to exposure to various concentrations of analyte, dryair was introduced over the sensor for an hour to establishthe sensor baseline. Afterward, the sensor was subjected tovarious concentrations of analyte, followed by a recoveryperiod with dry air in a step-wise manner. All the analyteswere thoroughly mixed with carrier gas (dry air) beforeintroduced to the sensor. Water-saturated dry air was prepared

by bubbling the carrier gas through water. Relative humiditywas controlled by varying the mixing ratio of dry air andwater-saturated air. The volumetric flow rate was kept constantto 200 standard cubic centimeters per minute (sccm) whilevarious concentrations/saturations were tested.

3. Results and discussion

3.1. Synthesis of PANI/PCL composite nanofibers and theirmaterial characterization

Homogeneous incorporation of PANI into the electrospunblend PANI/PCL nanofibers is essential to prevent the beadformation of aggregated PANI particles along the nanofibers,which may deteriorate electrical and sensing properties [18].Based on our preliminary studies, HFIP was selected as acommon solvent because of its appropriate vapor pressure atroom temperature that enables electrospinning nanofibers lessthan 100 nm in fiber diameter [19]. Additionally, HFIP mayenhance the electrical conductivity of the composite nanofibersby inducing linear extension of PANI [20]. However, theemeraldine base of PANI has low solubility in organic solventsincluding HFIP [21]. Thus, HCSA was used to protonate theemeraldine base of PANI, making it a polyelectrolyte (orwell-dispersion of PANI) in HFIP to enhance the uniformdispersion of PANI [22]. This dispersed PANI solution wasthen vacuum-filtered to remove non-dispersed PANI particles.

Due to possible changes of PANI content during filtration,the filtered PANI solution was quantified by UV–vis spec-troscopy. Various concentrations of unfiltered PANI solutionwere prepared and their UV–vis absorption spectra were ex-amined (figure S1 available at stacks.iop.org/Nano/25/115501/mmedia). The absorption peak at wavelength of 300 nmthat represents π–π∗ transition of the benzenoid ring [23],was utilized to generate a calibration curve for determinationof the final PANI content [24]. The final concentration offiltered PANI solution was estimated to be ∼5.8 mg ml−1,exceeding the initial PANI content of 3 mg ml−1. The increasein PANI content after filtration may result from evaporationof HFIP during the vacuum filtration process. Multiple ex-periments confirmed the content change is consistent amongexperiments. The filtered PANI solution was mixed with PCLsolution and then electrospun as described earlier. Figure 2(a)shows the microstructure of electrospun PANI/PCL compositenanofibers examined by SEM. The nanofibers are randomlyoriented with relatively uniform size having a typical cylindri-cal morphology. A quantitative analysis of the fibers indicatesthe tightly controlled average fiber diameter at 83 ± 16 nm(figure 2(b)). The surface of the nanofibers after doping withHCl or CSA was closely examined for possible changes, butSEM examination revealed no apparent morphological changeafter protonic acid doping (figure 3).

In order to examine the effects of doping on the compos-ite, FT-IR and diffuse reflectance UV–vis spectroscopy wereperformed. The structure and the oxidation level of PANI/PCLcomposite electrospun nanofibers were analyzed after dopingwith protonic acids (i.e., HCl and HCSA) and de-dopingwith a base (i.e., NaOH) (figure 4). Figure 4(a) displays the

3

Nanotechnology 25 (2014) 115501 K Low et al

Figure 2. Morphological characterization of as-electrospun PANI/PCL composite nanofibers. (a) A representative scanning electronmicroscopy (SEM) image and (b) fiber diameter distribution.

Figure 3. Surface morphological characterization of electrospun PANI/PCL composite nanofibers (a) as-prepared, (b) after Cl− doping, and(c) after CSA doping.

Figure 4. Chemical characterization of PANI/PCL fiber mats with various doping agents and time using (a) FT-IR and (b) diffuse reflectanceUV–vis spectroscopy. Insets in (b) show representative color changes of the samples depending on doping agent and time.

FT-IR absorption spectra of de-doped, as-electrospun, Cl− andCSA doped PANI/PCL nanofibers. The FT-IR characteristicabsorption bands at 2949, 2866, and 1730 cm−1 are specific toasymmetric CH2, symmetric CH2 and carbonyl stretching ofPCL, respectively [25]. The other bands at 825, 1161, 1297,1493, and 1586 cm−1 are specific to C–H bending vibrationof benzene ring, vibration mode of quinoid ring, stretchingvibration of C–N, stretching vibration of N-benzenoid ringand stretching vibration of N-quinoid ring of PANI, respec-tively [26]. To determine changes in the levels of protonationby acidic doping and basic de-doping, the ratio of the peakarea intensity between the benzenoid (protonated PANI) and

quinoid (undoped PANI) rings at 1493 and 1586 cm−1 inthe FT-IR spectra, respectively, was quantified [27]. The ratioof as-electrospun PANI/PCL is 0.91:1. This is close to theintrinsic emeraldine salt of PANI which has 1:1 ratio betweenbenzenoid and quinoid rings [28]. De-doping the samplewith NaOH decreased the value to 0.78:1, inferring that thebenzenoid structure was deprotonated to quinoid rings after thetreatment. Conversely, the increasing ratio values to 1.04:1 and1.28:1 by doping with Cl− and CSA for 5 min, respectively,indicate that PANI was further protonated. The smaller valueobtained from Cl− doping indicates slower protonation ratethan that of CSA.

4

Nanotechnology 25 (2014) 115501 K Low et al

In addition to the PANI/PCL structural changes observedby FT-IR spectra, the protonation states of PANI/PCL compos-ite nanofibers upon de/doping was further closely examined bydiffuse reflectance UV–vis spectroscopy. Figure 4(b) shows thediffuse reflectance UV–vis spectra of 1 min de-doped in NaOH,as-electrospun, 1 or 5 min doped with HCl, 1 or 5 min dopedwith HCSA PANI/PCL samples, and pure PCL sample alongwith their representative fiber colors. A color change fromgreen to blue was observed in de-doped PANI/PCL nanofibersin 1 M NaOH, whereas that of protonic acids (i.e., HCl andHCSA) doping turned the green color lighter proportional todoping duration. These color changes were closely related tothe protonation states of the PANI/PCL characterized by theabsorbance spectra in figure 4(b) [23]. The absorbance peakat wavelength of 206 nm is specific to PCL as shown bypure PCL sample [29]. Two absorbance peaks in de-dopedPANI/PCL in NaOH appear at the wavelengths of 325 and605 nm, respectively, denote the benzenoid and quinoid struc-tures which are characteristics of PANI in emeraldine baseform [24]. The larger absorbance peak at wavelength of 605 nmas compared to rest of the samples, indicates more quinoidcontents, similarly observed in the FT-IR spectrum of de-dopedPANI/PCL. As-electrospun PANI/PCL nanofibers exhibited asignificant decrease in the peak intensity at 605 nm, indicatingthe conversion of quinoid rings to its polaron state [24]. A peakat 425 nm that was missing in the de-doped specimen indicatesthe degree of protonation resulting in bipolaron formation [30].Upon doping with protonic acids, either HCl or HCSA, thepeak at 605 nm vanished, and a free carrier tail became visibleat 850 nm [31]. This characteristic is consistent with therelaxation of the PANI chain as its typically coiled structurebecomes more linearly extended by the reduction of π defectsthat cause the compact structure [32]. This uncoiling processinduces the delocalization of electrons in the polaron bandof the PANI chain [24]. Interestingly, CSA doped nanofibersexhibit higher absorbance intensity of a free carrier tail thanthe Cl− doped samples, indicating greater efficiency of doping.The greater intensity also is observed when the nanofibers weredoped for a longer duration. This observation, combined withthe FT-IR spectroscopy data, demonstrates that protonation ofPANI/PCL composite strongly depends on the doping agentas well as doping duration.

3.2. Electrical properties of electrospun PANI/PCL compositenanofibers

After chemical characterization, the electrical properties andsensing performance of PANI/PCL nanofibers were investi-gated. The electrical conductivity of nanofibers was obtainedfrom I –V characteristics generated by sweeping the voltagebetween −0.5 and 0.5 V across source and drain electrodes.Figures 5(a) and (b) show the typical I –V curves of Cl− andCSA doped PANI/PCL nanofibers, respectively, as comparedto that of as-prepared PANI/PCL nanofibers. The I –V curveof as-prepared nanofibers has a smaller slope with non-linearand asymmetric, indicating its insulating electrical propertywith Schottky contact prior to doping [33]. This observationis consistent with the high quinoid content within PANI/PCL

nanofibers observed from its FT-IR and diffuse reflectanceUV–vis spectra. Electrical current increased after doping withCl− and CSA. The electrical characterization also showeda linear I –V relation, a characteristic of Ohmic contact,indicating that the devices exhibit a good electrical contactbetween the nanofibers and the gold microelectrodes [34].

The electrical conductivity of doped PANI/PCL nanofiberswas determined from slopes of I –V curves normalized bythe geometry of the nanofiber mats after various treatments(figure 5(c)). The as-prepared composite nanofibers exhibitedan electrical conductivity of ∼9 × 10−7 S cm−1. Afterdoping in 1 M HCl for 5 min, the electrical conductivityincreased to ∼1× 10−5 S cm−1, and further increased to∼5× 10−4 S cm−1 after 10 min. In contrast, CSA was amore effective dopant as the electrical conductivity became∼8× 10−2 S cm−1 after 5 min doping. This substantiatesdopant-dependent electrical property of PANI, correspond-ing to the observations from FT-IR and diffuse reflectanceUV–vis spectroscopy. Electrical conductivities of PANI/PCLnanofibers doped with both Cl− and CSA were saturatedafter certain thresholds, indicating that the extent of doping isdiffusion controlled [35].

3.3. Gas sensing properties of electrospun PANI/PCLcomposite nanofibers

Various analytes including H2O vapor NH3, and NO2 were ex-amined. Prior to exposure to various concentrations of analyte,dry air was introduced over the sensor for an hour to establishthe sensor baseline. Afterward, the sensor was subjected tovarious concentrations of analyte, followed by a recoveryperiod with dry air in a step-wise manner. The normalizedchange in electrical resistance (1R/Ro) and sensitivity of Cl−

and CSA doped PANI/PCL nanofibers plotted against differentconcentrations of various gas are shown in figures 6 and 7. TheRo and 1R were defined as the initial baseline resistance andchange in resistance upon subjected to analytes, respectively.

Figures 6(a)–(f) show the 1R/Ro responses of Cl−

and CSA doped PANI/PCL nanofibers, respectively, to H2Ovapor, NH3 and NO2. During H2O exposure, the sensorsshow negative 1R/Ro responses as compared to the positive1R/Ro responses seen in NH3 and NO2 sensing. These sens-ing behaviors may result from different sensing mechanismdepending on analyte interactions with PANI/PCL electrospunnanofibers. The decrease in electrical resistance of PANI/PCLnanofibers upon exposure to H2O vapor may be attributed toproton exchange-assisted conduction of electrons (PEACE)mechanism [36]. The protonating agents, including H2O,protonate the imine group and forms polarons, allowing freemovement of charge carriers along the PANI backbone. Onthe other hand, a deprotonating agent such as NH3 interactswith protons on the PANI backbone and yields ammoniumion, NH+4 [2]. This phenomenon reduces free charge carriers,causing an increase in electrical resistance of PANI/PCLnanofibers. Additionally, the sensors show a recovery in1R/Ro responses during pure dry air flow in both H2O vaporand NH3 due to their easy adsorption and desorption processfrom the PANI surface [37]. Although the sensors also display

5

Nanotechnology 25 (2014) 115501 K Low et al

Figure 5. I –V plots of (a) Cl− and (b) CSA doped electospun PANI/PCL composites. (c) Conductivity of Cl− or CSA doped PANI/PCLcomposite sensor is plotted as a function of doping time (some error bars are not visible because they are smaller than the data point markers).

Figure 6. Transient sensing profiles of ((a)–(c)) Cl− and ((d)–(f)) CSA doped electrospun PANI/PCL composite sensors at variousas-fabricated conductivities in response to various concentrations of ((a) and (d)) H2O vapor, ((b) and (e)) NH3, and ((c) and (f)) NO2.

increases in 1R/Ro in response to NO2, the mechanism ofelectrical resistance changes differs from that of NH3 sensing.Unlike NH3, NO2, a reducing gas, oxidizes emeraldine salt byremoving electrons from the aromatic rings [38]. As a result,no recovery is observed due to strong interaction between NO2

and PANI, which makes the desorption rate of NO2 becomemuch slower, instigating a poisoning effect [38].

Figures 7(a)–(f) show the effect of dopant and dopingconcentration on the sensitivity of sensors. The data demon-strate that the extent of gas sensor sensitivity can be tailored by

6

Nanotechnology 25 (2014) 115501 K Low et al

Figure 7. Normalized responses of ((a)–(c)) Cl− and ((d)–(f)) CSA doped electrospun PANI/PCL composite sensors at various as-fabricatedconductivities in response to various concentrations of ((a) and (d)) H2O vapor, ((b) and (e)) NH3, and ((c) and (f)) NO2.

protonic doping that determines the electrical conductivity ofPANI/PCL nanofibers. Interestingly, it was observed that thesensitivity change increased with increasing gas concentration,but at a fixed gas concentration, there was not a linear relation-ship between sensitivity change and the electrical conductivityof sensor. For closer examination, sensitivity changes per ppmwith respect to different electrical conductivity for variousgases were plotted (figure 8). The semi-log plots of the sen-sitivity are shown as a function of the electrical conductivity.The sensitivity was determined from the slope of the sensorresponse (1R/Ro) plotted against the gas concentration in thelinear region (figure 7). Recent studies using PANI have re-ported sensitivity values of 0.46%/per cent saturation for H2Ovapor [39], 4.23% ppm−1 for NH3 [40], and 19.91% ppm−1 forNO2 [39]. In this study, the Cl− doped PANI/PCL nanofibrouscomposite exhibited sensitivities of 1.36%/per cent saturationfor H2O vapor, 13.75 and 70.31% ppm−1 for NH3 and NO2,respectively. Similarly, CSA doped sensors showed sensitiv-ities of 0.76%/per cent saturation for H2O vapor, 21.43 and54.41% ppm−1 for NH3 and NO2, respectively. In general, thePANI/PCL composite sensors exhibit superior sensitivity ascompared to PANI-based sensors [40–49]. However, recoverytimes of the nanofiber sensors are slower than typical thinfilm-based sensors [39, 40]. The greater sensitivities are likelydue to the high surface area-to-volume ratio that nanofiberspossess, allowing more gas molecules to adsorb onto thesurface of the fiber. However, this inevitably requires moretime for the gas molecules to be released from the surface,resulting in increased recovering time.

Interestingly, the device shows biphasic sensitivity, withthe exception of CSA doped PANI/PCL upon exposure to NH3(figure 8). This observation is significant in that it demon-strates a means to optimize the sensitivity of an electro-spun chemiresistive sensor with a simple secondary dopingprocess (dopant and doping duration) without altering thecomposition of materials, which may affect processabilityand mechanical stability. The mechanisms underlying howdoping modulates the optimum sensitivity are still elusive.One possible explanation could be a balance between Debyelength and overall charge conduction. It has been shownthat Debye length, a scale that describes the alteration ofconductance by an adsorbed charged molecule, is one ofthe significant factors affecting sensitivity of chemiresistivesensors [50–52]. In this regard, we have previously reportedthat engineering the dimension of 1D nanostructure whichis comparable to the material’s Debye length, significantlyinfluences sensitivity of polypyrrole-based sensors [53]. Inaddition to the dimensional control, another method of mod-ulating sensitivity via Debye length is by adjusting chargecarrier density. Debye length is inversely related to chargecarrier density, thus electrical conductivity [54]. In this study,we observed that the sensitivity increases when the electricalconductivity decreases (i.e., less doped) above a threshold(optimum electrical conductivity yielding a maximum sen-sitivity). However, below this threshold, sensitivity reduces bylower electrical conductivity due to decrease in overall chargeconduction. The discrepancies in the optimal electrical con-ductivities for the maximum sensitivity depending on dopants(i.e., Cl− and CSA), are probably due to their differences in

7

Nanotechnology 25 (2014) 115501 K Low et al

Figure 8. Sensitivity of Cl− or CSA doped electrospun PANI/PCL composite sensors at various as-fabricated conductivities to(a) H2O vapor, (b) NH3, and (c) NO2.

hydrophilicity affecting responses to H2O vapor and molecularsize influencing molecular structures [55]. However, howthese individual attributes collectively affect overall sensingperformance needs to be further investigated. Nevertheless,we demonstrated that secondary doping significantly impactsthe sensing performance of electrospun composite nanofibers.

4. Conclusion

In this study, we synthesized a sensitive PANI/PCL compositenanofibrous sensor using electrospinning and characterized itsgas sensing performance. The electrical conductivity of this 1Dnanostructured sensor was further tuned by secondary protonicdoping with Cl− and CSA. More significantly, we demon-strated that optimal sensitivities toward H2O vapor, NH3, andNO2 can be achieved by optimizing doping, where dopant- anddoping duration-dependent biphasic sensitivity were observed.

Acknowledgment

We greatly acknowledge the financial support from KIMSAcademy Laboratory Program.

References

[1] Kolmakov A and Moskovits M 2004 Chemical sensing andcatalysis by one-dimensional metal–oxide nanostructuresAnnu. Rev. Mater. Res. 34 151–80

[2] Virji S, Huang J X, Kaner R B and Weiller B H 2004Polyaniline nanofiber gas sensors: examination of responsemechanisms Nano Lett. 4 491–6

[3] Cabala R, Meister V and Potje-Kamloth K 1997 Effect ofcompetitive doping on sensing properties of polypyrroleJ. Chem. Soc. Faraday Trans. 93 131–7

[4] Bai H and Shi G Q 2007 Gas sensors based on conductingpolymers Sensors (Basel) 7 267–307

[5] Janata J and Josowicz M 2003 Conducting polymers inelectronic chemical sensors Nature Mater. 2 19–24

[6] Strumpler R and Glatz-Reichenbach J 1999 Conductingpolymer composites J. Electroceram. 3 329–46

[7] Hatchett D W and Josowicz M 2008 Composites ofintrinsically conducting polymers as sensing nanomaterialsChem. Rev. 108 746–69

[8] Sandra C H, Debangshu C, Wilfred C, Nosang V M andAshok M 2007 Single polypyrrole nanowire ammonia gassensor Electroanalysis 19 2125–30

[9] Hangarter C M, Hernandez S C, He X, Chartuprayoon N,Choa Y H and Myung N V 2011 Tuning the gas sensingperformance of single PEDOT nanowire devices Analyst136 2350–8

[10] Hangarter C M, Bangar M, Mulchandani A and Myung N V2010 Conducting polymer nanowires for chemiresistive andFET-based bio/chemical sensors J. Mater. Chem.20 3131–40

[11] Chartuprayoon N, Hangarter C M, Rheem Y, Jung H andMyung N V 2010 Wafer-scale fabrication of singlepolypyrrole nanoribbon-based ammonia sensor J. Phys.Chem. C 114 11103–8

[12] Wagner R and Ellis W 1964 Vapor–liquid–solid mechanism ofsingle crystal growth Appl. Phys. Lett. 4 89–90

[13] Tran H D, Li D and Kaner R B 2009 One-dimensionalconducting polymer nanostructures: bulk synthesis andapplications Adv. Mater. 21 1487–99

8

Nanotechnology 25 (2014) 115501 K Low et al

[14] Zhang X, Huang D, Chen F, Chen C and Liu Q 2004 The firstone-dimensional manganese polymer containing4,5-dicarboxyimidazole: solvothermal synthesis, crystalstructure and magnetic behavior of [Mn(phen)(Hdcbi)]n(H3dcbi = 4,5-dicarboxyimidazole, phen=1,10-phenanthroline) Inorg. Chem. Commun. 7 662–5

[15] Palmer L C and Stupp S I 2008 Molecular self-assembly intoone-dimensional nanostructures Acc. Chem. Res.41 1674–84

[16] Huang Z M, Zhang Y Z, Kotaki M and Ramakrishna S 2003 Areview on polymer nanofibers by electrospinning and theirapplications in nanocomposites Compos. Sci. Technol.63 2223–53

[17] Heng C S, Miluo Z, Wayne B, Jae-Hong L and Nosang V M2013 Metal nanoparticles and DNA co-functionalizedsingle-walled carbon nanotube gas sensors Nanotechnology24 505502

[18] Dong H, Nyame V, MacDiarmid A G and Jones W E 2004Polyaniline/poly(methyl methacrylate) coaxial fibers: thefabrication and effects of the solution properties on themorphology of electrospun core fibers J. Polym. Sci. B42 3934–42

[19] Fong H, Liu W, Wang C S and Vaia R A 2002 Generation ofelectrospun fibers of nylon 6 and nylon 6-montmorillonitenanocomposite Polymer 43 775–80

[20] Hopkins A R, Rasmussen P G and Basheer R A 1996Characterization of solution and solid state properties ofundoped and doped polyanilines processed fromhexafluoro-2-propanol Macromolecules 29 7838–46

[21] Kulszewicz-Bajer I, Wielgus I, Pron A and Rannou P 1997Protonation of polyaniline in hexafluoro-2-propanol.Spectroscopic investigation Macromolecules 30 7091–5

[22] Jeong S I, Jun I D, Choi M J, Nho Y C, Lee Y M and Shin H2008 Development of electroactive and elastic nanofibersthat contain polyaniline and poly(L-lactide-co-ε-caprolactone) for the control of cell adhesion Macromol.Biosci. 8 627–37

[23] Stejskal J, Kratochvıl P and Radhakrishnan N 1993Polyaniline dispersions 2. UV–vis absorption spectra Synth.Met. 61 225–31

[24] Zheng W, Min Y, Lee S J, MacDiarmid A G,Angelopoulos M, Liao Y H and Epstein A J 1995 Themolecular conformation of non-doped polyaniline and itseffect on the properties of doped polyaniline MRS OnlineProc. Lib. 413 413–535

[25] Elzein T, Nasser-Eddine M, Delaite C, Bistac S and Dumas P2004 FTIR study of polycaprolactone chain organization atinterfaces J. Colloid Interface Sci. 273 381–7

[26] Lijuan Z and Meixiang W 2002 Synthesis and characterizationof self-assembled polyaniline nanotubes doped withD-10-camphorsulfonic acid Nanotechnology 13 750

[27] Yelil Arasi A, Juliet Latha Jeyakumari J, Sundaresan B,Dhanalakshmi V and Anbarasan R 2009 The structuralproperties of poly(aniline)—analysis via FTIR spectroscopySpectrochim. Acta A 74 1229–34

[28] MacDiarmid A G and Epstein A J 1989 Polyanilines: a novelclass of conducting polymers Faraday Discuss. Chem. Soc.88 317–32

[29] Martins-Franchetti S, Campos A, Egerton T and White J 2008Structural and morphological changes in poly(caprolactone)/poly(vinyl chloride) blends caused by UVirradiation J. Mater. Sci. 43 1063–9

[30] Huang W S and MacDiarmid A G 1993 Optical properties ofpolyaniline Polymer 34 1833–45

[31] Min Y, Xia Y, MacDiarmid A G and Epstein A J 1995 Vaporphase secondary doping of polyaniline Synth. Met.69 159–60

[32] MacDiarmid A G and Epstein A J 1995 Secondary doping inpolyaniline Synth. Met. 69 85–92

[33] Chakrabarty R K, Bardhan K K and Basu A 1991 NonlinearI –V characteristics near the percolation threshold Phys.Rev. B 44 6773–9

[34] Ramanathan K, Bangar M A, Yun M, Chen W,Mulchandani A and Myung N V 2004 Individuallyaddressable conducting polymer nanowires array Nano Lett.4 1237–9

[35] Sengupta P P, Barik S and Adhikari B 2006 Polyaniline as agas-sensor material Mater. Manuf. Process. 21 263–70

[36] Ting Z, Syed M, Bongyoung Y, Nosang V M and Marc A D2009 A gas nanosensor unaffected by humidityNanotechnology 20 255501

[37] Nicolas-Debarnot D and Poncin-Epaillard F 2003 Polyanilineas a new sensitive layer for gas sensors Anal. Chim. Acta475 1–15

[38] Agbor N E, Petty M C and Monkman A P 1995 Polyanilinethin films for gas sensing Sensors Actuators B 28 173–9

[39] Azim-Araghi M E, Jafari M J, Barhemat S andKarimi-Kerdabadi E 2011 Gas mixture sensor based onpolyaniline-chloroaluminium phthalocyaninenanocomposite thin films Sensor Lett. 9 1349–55

[40] Tai H, Jiang Y, Xie G, Yu J and Chen X 2007 Fabrication andgas sensitivity of polyaniline–titanium dioxidenanocomposite thin film Sensors Actuators B 125 644–50

[41] McGovern S T, Spinks G M and Wallace G G 2005Micro-humidity sensors based on a processable polyanilineblend Sensors Actuators B 107 657–65

[42] Matsuguchi M, Okamoto A and Sakai Y 2003 Effect ofhumidity on NH3 gas sensitivity of polyaniline blend filmsSensors Actuators B 94 46–52

[43] Gong J, Li Y, Hu Z, Zhou Z and Deng Y 2010 UltrasensitiveNH3 gas sensor from polyaniline nanograin enchased TiO2fibers J. Phys. Chem. C 114 9970–4

[44] Sharma A L, Kumar K and Deep A 2013 Nanostructuredpolyaniline films on silicon for sensitive sensing ofammonia Sensors Actuators A 198 107–12

[45] Chabukswar V V, Pethkar S and Athawale A A 2001 Acrylicacid doped polyaniline as an ammonia sensor SensorsActuators B 77 657–63

[46] Li D, Jiang Y, Wu Z, Chen X and Li Y 2000 Self-assembly ofpolyaniline ultrathin films based on doping-induceddeposition effect and applications for chemical sensorsSensors Actuators B 66 125–7

[47] Yan X B, Han Z J, Yang Y and Tay B K 2007 NO2 gas sensingwith polyaniline nanofibers synthesized by a facileaqueous/organic interfacial polymerization SensorsActuators B 123 107–13

[48] Bishop-Haynes A and Gouma P 2007 Electrospun polyanilinecomposites for NO2 detection Mater. Manuf. Process.22 764–7

[49] Zhang D, Liu Z, Li C, Tang T, Liu X, Han S, Lei B andZhou C 2004 Detection of NO2 down to ppb levels usingindividual and multiple In2O3 nanowire devices Nano Lett.4 1919–24

[50] Kolmakov A, Zhang Y, Cheng G and Moskovits M 2003Detection of CO and O2 using tin oxide nanowire sensorsAdv. Mater. 15 997–1000

[51] Fan Z and Lu J G 2005 Gate refreshable nanowire chemicalsensors Appl. Phys. Lett. 86 123510

9

Nanotechnology 25 (2014) 115501 K Low et al

[52] Li C C, Du Z F, Li L M, Yu H C, Wan Q and Wang T H 2007Surface depletion controlled gas sensing of ZnO nanorodsgrown at room temperature Appl. Phys. Lett. 91 032101

[53] Chartuprayoon N, Rheem Y, Ng J C K, Nam J, Chen W andMyung N V 2013 Polypyrrole nanoribbon basedchemiresistive immunosensors for viral plant pathogendetection Anal. Methods 5 3497–502

[54] Horowitz G, Hajlaoui R, Bourguiga R and Hajlaoui M 1999Theory of the organic field-effect transistor Synth. Met.101 401–4

[55] John M F, Leandro F, Hieu N, James D B, Yann-Fuu K,Jeja S-N and Christine E S 2008 Biocompatibilityimplications of polypyrrole synthesis techniques Biomed.Mater. 3 034124

10