www.MaterialsViews.com
Sensors
Hybrid Films of Graphene and Carbon Nanotubes for High Performance Chemical and Temperature Sensing Applications Tran Thanh Tung,* Cuong Pham-Huu, Izabela Janowska, TaeYoung Kim,* Mickael Castro, and Jean-Francois Feller
A
hybrid composite material of graphene and carbon nanotube (CNT) for high performance chemical and temperature sensors is reported. Integration of 1D and 2D carbon materials into hybrid carbon composites is achieved by coupling graphene and CNT through poly(ionic liquid) (PIL) mediated-hybridization. The resulting CNT/PIL/graphene hybrid materials are explored as active materials in chemical and temperature sensors. For chemical sensing application, the hybrid composite is integrated into a chemo-resistive sensor to detect a general class of volatile organic compounds. Compared with the graphene-only devices, the hybrid film device showed an improved performance with high sensitivity at ppm level, low detection limit, and fast signal response/recovery. To further demonstrate the potential of the hybrid films, a temperature sensor is fabricated. The CNT/PIL/graphene hybrid materials are highly responsive to small temperature gradient with fast response, high sensitivity, and stability, which may offer a new platform for the thermoelectric temperature sensors.
1. Introduction Low-dimensional carbon structures such as 1D carbon nanotubes (CNTs) and 2D graphene holds great promise for potential sensor applications because of their unique Dr. T. T. Tung, Dr. C. Pham-Huu, Dr. I. Janowska Institut de Chimie et Procédés pour l’Energie l’Environnement et la Santé (ICPEES) ECPM, UMR 7515 du CNRS-Université de Strasbourg 25 rue Becquerel, 67087, Strasbourg Cedex 02, France E-mail: [email protected] Prof. T. Kim Department of Bionanotechnology Gachon University 1342 Seongnamdaero, Sujeong-gu, Seongnam-si Gyeonggi-do 461-701, South Korea E-mail: [email protected] Dr. T. T. Tung, Dr. M. Castro, Prof. J.-F. Feller Smart Plastics Group European University of Brittany (UEB), LIMATB-UBS rue de Saint-Maude´ 56321 Lorient, France DOI: 10.1002/smll.201403693 small 2015, DOI: 10.1002/smll.201403693
nanoscale features, large specific surface area, high carrier mobility, and facile chemical functionalization.[1–8] Previous studies show that their electrical resistance (or conductance) modulates upon exposure to chemical species or environmental changes. It has been shown that the large surface area of these carbons allows for the attachment of a large amount of analyte molecules on their surfaces, leading to the high sensitivity and signal-to-noise ratio (SNR) of the sensor devices, and the fast electron transfer between the carbon and analyte enables the direct electrochemical reaction with rapid response time.[9–11] In addition, their excellent mechanical properties allow for the integration into flexible sensing devices.[12,13] In this context, assembling graphene nanosheets into interconnected CNT networks is of special interest for the realistic development of a new sensing platform because the resulting hybrid materials would offer a combined set of properties such as reinforced mechanical strength and straightforward molecular transport through interconnected pathway, leading to improved electronic and mechanical properties.[15,16] Such hybrid materials have been prepared by either noncovalent bonding (e.g., solution mixing of CNT and graphene) or covalent bonding (e.g., CVD growth of CNT on
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
1
full papers www.MaterialsViews.com
graphene) for applications in the field of flexible electronics, energy storage, catalyst, and multifunctional composite.[17–24] Despite the potential advantages offered by the combination of the 1D- and 2D-structured carbon materials, the exploitation of these hybrid composites as sensor platforms suffered from several practical issues such as controlled assembly of graphene and CNTs on the molecular level and lack of efficient functionalization of carbons on sensing performance. Herein, we present a method to produce a carbon hybrid material consisting of CNTs and reduced graphene oxide (rGO). For the efficient hybridization of CNTs and rGO sheets, we used a ionic liquid-based polymer, referred to as poly(ionic liquids) (PILs), as a stabilizer and linker. In the course of the hybridization process, PIL molecules not only improve the dispersion of CNTs and rGO sheets in solvents against uncontrolled aggregation, but also act as a linker to assemble the two carbons into the nanostructured hybrid materials. Coupling of CNTs and rGO by PIL-mediated hybridization yielded the complex of CNT/PIL/rGO, in which each component was assembled on the molecular level through electrostatic and π–π stacking interactions. The CNT/ PIL/rGO hybrid composites were tested for the two different types of sensing applications such as chemical and temperature sensors. For the chemical sensor application, the CNT/ PIL/rGO film was deposited on an interdigitated electrode and integrated into the chemo-resistive sensors. Upon exposure to various volatile organic compound (VOC) vapors, the device showed an adjustable conductance with a rapid and selective detection capability. The hybrid film was further explored as a new type of temperature sensing platform. The free-standing film of CNT/PIL/rGO was capable to detect
temperature by monitoring temperature-activated carrier transport from hot end to cold end. The potential of the CNT/ PIL/rGO hybrid as active material for chemical and temperature sensors is discussed in terms of their performance.
2. Preparation of CNT/PIL/rGO Hybrid Materials The CNT/PIL/rGO hybrid materials were prepared through the following three steps: (i) modification of CNTs by water soluble PIL of poly(1-vinyl-3-ethylimidazolium) bromide (PIL-Br) to form an aqueous dispersion of PIL-modified CNTs (CNT/PIL); (ii) modification of G–O by hydrophilic PIL followed by chemical conversion into rGO/PIL; (iii) mixing of CNT/PIL and rGO/PIL and subsequent phase transfer reaction by exchanging the counter anion, Br− in PIL with NTf2− to form a hydrophobic CNT/PIL/rGO hybrid composite. The resulting CNT/PIL/rGO hybrid materials were characterized by a number of methods including X-ray diffraction (XRD), FTIR, X-ray photoelectric spectroscopy (XPS), and Raman spectroscopy. Figure 1a shows the powder XRD pattern of the CNT/ PIL/rGO hybrid materials. The XRD pattern showed a broad peak in the 10°–30° 2θ region in which multiple diffraction peaks are overlapped. A (002) deconvoluted diffraction peak at 2θ of 20.1° is assigned to rGO platelets with an interlayer spacing of 0.44 nm. The interlayer distance is larger than that of graphite (0.34 nm), attributable to the residual oxygen functional groups of rGO as well as PIL held in the interlayer
Figure 1. a) Schematic illustration of CNT/PIL/rGO hybrid materials; b) XRD pattern; c) C 1s XPS spectra; and d) Raman spectra of CNT/PIL/rGO hybrid materials.
2 www.small-journal.com
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2015, DOI: 10.1002/smll.201403693
www.MaterialsViews.com
Figure 2. a,b) SEM images; c,d) TEM images of CNT/PIL/rGO hybrid materials.
galleries of rGO.[25] Two peaks at 2θ angle of 26.5° and 43° are likely due to the (002) and (004) diffraction of CNTs, indicating that CNT is intermixed with rGO within the hybrid materials. Chemical analysis of CNT/PIL/rGO was conducted by FTIR (Figure S1, Supporting Information) and XPS (Figure S2, Supporting Information). The survey spectrum of XPS clearly indicates the presence of C (62.24 at%), O (12.53 at%), N (8.06 at%), F (12.30 at%), and S (4.51 at%). The C1s XPS spectrum of CNT/PIL/rGO could be deconvoluted into different peaks that correspond to the following functional groups: carbon sp2 (C = C in aromatic ring, 284.5 eV), epoxy/hydroxyls (C–O, 286.2 eV), and carbonyl (C = O, 287.2 eV) and carboxylates (O–C = O, 288.8 eV).[25–27] Two additional peaks corresponding to carbon bound to nitrogen (C–N, 286.2 eV) and fluorine (C–F, 293.0 eV) were due to the presence of PIL. In addition, the nitrogen content and the nature of the different nitrogen species are presented in Figure S1c (Supporting Information). The nitrogen content in CNT/PIL/rGO was found relatively high (≈8 at%) as compared with that for the chemically reduced rGO (0.8 at%). This is due to the presence of PIL in the CNT/PIL/rGO composite, as the PIL molecules is composed of a nitrogen-containing cation (imidazolium) and anion (NTf2−).[28,29] Raman spectrum was recorded with a laser excitation wavelength of 532 nm under ambient condition. For Raman analysis, the samples were prepared by depositing CNT/PIL/rGO suspension onto glass slide substrate. The Raman spectra show a typical D-band (1345 cm−1) and G-band (1579 cm−1) corresponding to structure defects of sp3 domain and the E2g-vibration mode of sp2 carbon, respectively.[26–28] The 2D-band originating from two phonon double resonance small 2015, DOI: 10.1002/smll.201403693
Raman process was also observed at 2683 cm−1 and the peak at 2920 cm−1 is the D+D′ band corresponding to a defect for an elastic scattering event to provide momentum conservation in the Raman process.[29–33] The intensity of 2D peak is relatively small due to the disorder, and as compared with expanded graphite the 2D peak position in PIL-modified rGO shifted considerably to a lower wavenumber by 33 cm−1, from 2720 to 2687 cm−1, suggesting a successful exfoliation to single or few layer graphene sheets in rGO/PIL.[26,27,34] The 2D band of the CNT/PIL/rGO hybrid is different from that of rGO/PIL complex, which could be the super-imposed peaks due to the electronic band structure of CNTs in the composite.[35,36] All Raman peaks of CNT/PIL/rGO hybrid materials were found to be red-shifted as compared to CNT/ PIL and blue-shifted as to rGO/PIL (Figure S3, Supporting Information). This is an indication that CNTs and rGO are assembled into hybrid composite by PIL-mediated process, suggesting that PIL plays an important role in coupling CNT and rGO on the molecular level as well as stabilizing them in the suspension. The morphology of the CNT/PIL/rGO hybrid materials was examined by scanning electronic microscopy (SEM), transmission electron microscopy (TEM) and atomic force microscopy (AFM). Figure 2a,b displays typical SEM images of CNT/PIL/rGO hybrids at different magnification, showing a CNT network covered by overlapped rGO sheets. TEM images in Figure 2c,d showed that CNTs with lengths of several hundred nanometers to micrometers are adhered on the surface of rGO sheets. The higher resolution TEM image (Figure 2d) revealed that individual CNTs are well-dispersed on the rGO sheets. It is noted that the coupling of CNT and
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
3
full papers www.MaterialsViews.com
Figure 3. Tapping mode AFM topographic images and height profiles of CNT/PIL/rGO hybrid films.
rGO would be the result of the strong cation-π interaction between π electron clouds of sp2-hybridized carbons and imidazolium cation of PIL, which in turn allows for the hybridization of CNT and rGO sheets.[37–39] We also used AFM to characterize the topography of hybrid CNT/PIL/rGO film, which were deposited on a Si/SiO2 substrate by spin-coating of the suspension (in acetonitrile, 0.3 mg mL–1). Figure 3 shows a typical tapping mode AFM images and height profile displaying the morphology and thickness of the CNT/PIL/rGO hybrid films. From the height profiles, the thickness of rGO/PIL and CNT/PIL in the composite was found in the range of 1.5−2.5 nm and 10–14 nm, respectively, which is consistent with those measured for PIL-coated rGO and PIL-coated CNTs (Figure S4, Supporting Information). Some of CNTs and rGO sheets were also found to protrude from the composite films, resulting in a rough surface morphology at nanoscale.
3. Chemical Sensing Application of CNT/PIL/ rGO Hybrids As a proof-of-concept experiment, we fabricated a chemoresistive sensor by spray-coating CNT/PIL/rGO hybrid on an interdigitated electrode array (Figure 4a). The relative differential resistance responses (Ar = ΔR/R0)
4 www.small-journal.com
to various analyte vapors of sensor devices was recorded as detailed in the Experimental Section. Prior to flowing analyte vapors, sensor arrays were stabilized to show constant electrical resistance in dry N2 flow. Upon exposure to analyte vapors, their resistance increased sharply, giving a positive vapor response of ΔR/R0. When analyte vapor flow was turned off and N2 gas flow restored, resistance of the sensors went back to its initial resistance, indicating a complete desorption of vapor molecules from the sensing materials and full recovery of the device. Vapor exposures were interlaced between recovery periods to show reproducibility. Figure 4b shows a typical dynamic response (ΔR/R0 vs time) of a CNT/PIL/rGO device for the detection of various analyte vapors, such as methanol, chloroform, tetrahydrofuran (THF), and benzene at saturated concentration. The as-fabricated sensor devices showed strong responses to all analyte vapors with large changes in the resistance of CNT/ PIL/rGO. The device was highly responsive to a 5% diluted NH3 solution with an amplitude of Ar ≈3.6. The CNT/PIL/ rGO device also responded to methanol with an amplitude of Ar ≈2.2. Considering that the previously reported value for the rGO/PIL-based sensor was Ar ≈1.68, the device in this work showed that the sensitivity was increased by 1.4fold.[40] The significantly improved sensitivity of CNT/PIL/ rGO sensors reported here is mainly attributed to the combined effect of hybridization of 1D and 2D carbons and the
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2015, DOI: 10.1002/smll.201403693
www.MaterialsViews.com
Figure 4. Chemo-resistive responses of CNT/PIL/rGO composite-based sensors: a) schematic diagram to describe the structure of sensor device; b) Normalized resistance change versus time upon sequential exposure to different analytes; c) Normalized resistance change versus time with respect to vapor concentration d) Plot of normalized resistance change versus vapor concentration for different analytes.
presence of PIL molecules between the assembled carbons. It is well known that vapor molecules are adsorbed onto binding sites with low energy such as graphitic sp2–bonded carbon and high-energy binding sites such as vacancies, structural defects, and functional groups on the surface of carbon nanomaterials, leading to the change in resistance.[41] The interconnected networks of 1D CNT and 2D rGO in the hybrid composites would provide more binding sites available for molecular adsorption and therefore enhance the sensing response. In addition, PIL located between the CNT and rGO sheets also serve to absorb VOC vapor molecules and induce changes in their physical properties. Upon exposure to analyte vapors, PIL molecules in the carbon network could swell and the chains rearrange in a way to disrupt the conductive carbon channels with an increased resistance. The ranking of the response for sensor toward different analytes is found as followed: Armethanol > Archloroform > Artetrahydrofuran > Arbenzene. The sensing mechanism can be explained by the van der Waals interactions between analyte molecules and the hybrid sensing materials, which differ with interaction parameters such as diffusion coefficient, average number of solvent molecules per cluster, extra clustering and so on. The presented data are consistent with the Langmuir–Henry– Clustering (LHC) electrosorption model as detailed in the previous reports for conductive polymer composite.[42,43] Figure 4c presented a relationship between sensing response and the concentration for four analyte vapors. When exposed to successively increased concentration of analyte vapors in the range of 1–150 ppm, the resistance small 2015, DOI: 10.1002/smll.201403693
change increased in a linear fashion, as shown in Figure 4d. The SNRs of the sensors at 1 ppm concentration of analyte molecules were estimated as 22.8, 34.5, 48.6, and 59.3 for methanol, chloroform, tetrahydrofuran, and benzene, respectively, as shown in Table S1 (Supporting Information). According to the International Union of Pure and Applied Chemistry (IUPAC) definition, the signal is considered to be a true signal, when the SNR ratio equals 3.[44] Therefore, the results in the present work suggested that the detection limit of CNT/PIL/rGO sensor is in the ppm range, and it can be possibly down to sub-ppm levels. The response times of sensor at a 1 ppm concentration were estimated as 20, 26, 30, and 40 s for methanol, chloroform, tetrahydrofuran, and benzene, respectively (see Figure S5, Supporting Information). This fairly fast response of the device is attributed to the swelling of PIL molecules by molecular adsorption and the relevant diffusion dynamics. The swelling of PIL at the interface of CNT and rGO would result in the separation of the two carbon/ carbon nanojuctions, thereby affecting the tunneling current through the composite.[45–47] Moreover, the higher the diffusion coefficient of analyte molocules through the percolated structure (Dc, Table S2, Supporting Information) the quicker the response. As compared with the previously reported rGO-only sensors, the present device showed an improved performance in terms of sensitivity, detection limit, and fast signal response/recovery toward VOCs and is consistent with other recent progress in nanocomposites based on CNTs and graphene for chemical sensors application.[48–52]
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
5
full papers www.MaterialsViews.com
Figure 5. Experiment setup for temperature sensing characterization.
4. Temperature Sensing Application of CNT/ PIL/rGO Hybrids We further explored the potential use of CNT/PIL/rGO hybrid as thermoelectric temperature sensing materials. For the device fabrication (Figure 5), thin film of the hybrid materials was held between the copper bars and heated from either side using two Peltier modules, while the temperature difference (ΔT) across the sample was monitored with two thermocouples attached to each ends. The voltage (ΔV) generated by temperature difference between the hot and cold ends, was measured over time as shown in Figure 6. Figure 6a shows the response of the device to the stepwise temperature changes, in which the output voltage
curve matches with the temperature profile. The Figure 6b showed the relationship between output voltage of the CNT/PIL/rGO device and temperature difference. The output voltage increases linearly with the temperatures, showing a near-room temperature operation of the device. Figure 6c showed a variation of output voltage of the device under cyclic changes of temperature. The device shows a rapid response and recovery to the cyclic temperature changes without any noticeable time lag. Therefore, the device was highly responsive to the temperature, which is ascribed to the thermoelectric properties of CNT/ PIL/rGO hybrid composite. When a temperature gradient is applied, the excess charge carriers migrate to the cold side of the sample, that the build-up of carriers generate
Figure 6. a) Output voltage of CNT-PIL-rGO composite-based sensor upon a step-wise thermal input; b) Plot of output voltage of the sensor versus temperature; c) Output voltage of the sensor upon cyclic thermal input; d) I–V characteristic of the composites measured between the forward bias (+5 V) and reserve bias (–5 V)
6 www.small-journal.com
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2015, DOI: 10.1002/smll.201403693
www.MaterialsViews.com
a voltage across the materials. This voltage generated by the thermoelectric materials is governed by its Seebeck coefficient (the amount of voltage generated across the material at given temperature gradient (ΔV/ΔT).[53,54] The highly sensitive thermoelectric response of CNT/PIL/rGO composite materials is attributed to the synergistic effects of the individual components, in which PIL plays an important role in enhancing thermoelectric properties. The PIL between the CNT and rGO sheets is proposed to induce the energy filtering or phonon scattering, but diminish carrier transport, thereby leading to an increase in Seebeck coefficient.[55–58] To verify the proposed mechanism, the current–voltage (I–V) characteristic of the CNT/PIL/rGO composite was investigated. Figure 6d shows the I–V curve of the device measured in the bias voltage range of –5 to +5 V. While symmetric, the I–V curve of CNT/PIL/rGO composites is not linear, showing the electrical behavior similar to that of the previously reported semiconducting materials. This suggests that PIL between CNTs and rGO separate two carbon nanomaterials, leading to the hopping or tunneling conduction as the dominant charge-transport mechanism. The sensitivity of the device was further investigated by measuring the minimum thermal input required to produce an output voltage. For this purpose, we monitored the output voltage variation of the composite generated under the stepwise temperature change by few degrees. In Figure S6 (Supporting Information), the sensor showed a remarkable voltage change under subtle temperature change of ≈1 °C, while the output voltage curve matches with the temperature profile. Considering the high sensitivity and stability under room temperature operation, the present device may hold the potential as a temperature sensor in a wide range of applications such as public safety, automotive, and green house.
5. Conclusion A hybrid composite material of 1D CNTs and 2D rGO sheet have been prepared using PIL as inter-linker for coupling them through noncovalent interactions. In this work, the hybrid material was explored for dual missions: (i) detection of trace amount of organic vapors and (ii) monitoring environmental temperatures. For chemical sensing application, the CNT/PIL/rGO hybrid material was used as a chemoresistor to detect various organic vapors at ppm level at room temperature. The CNT/PIL/rGO sensor was found to be selective toward a general class of organic vapors and showed an improved performance in terms of sensitivity, detection limit, and fast signal response. We also demonstrated a temperature sensor based on the thermoelectric properties of CNT/PIL/rGO hybrid material. The device was highly responsive to even small temperature gradient with fast response time, demonstrating the potential of the CNT/ PIL/rGO hybrid as a new type of temperature sensing material. The CNT/PIL/rGO composite presented in this work may offer an alternative platform for the chemical and thermoelectric temperature sensors. small 2015, DOI: 10.1002/smll.201403693
6. Experimental Section Preparation of PIL-Mediated Hybridization of CNTs and Graphene: Imidazolium-based poly(ionic liquids) such as poly(1-vinyl-3-ethylimidazolium) bromide (PIL-Br) and poly(1-vinyl3-ethylimidazolium) bis(trifluoromethylsulfonyl)amide (PIL-NTf2) were prepared according to our previous reports.[59,60] Graphite oxide (GO) was prepared by modified Hummers method (as previous reported[34] and oxidized multiwalled carbon nanotube (CNT) was kindly provided by Nanocyl. First, 10 mg of CNTs and 50 mg of PIL-Br were mixed in 100 mL of DI water and sonicated for 1 h to form an aqueous suspension of CNT/PIL-Br. Separately, 100 mg of GO and 200 mg of PIL-Br were mixed in 100 mL of DI water under stirring for 1 h. The GO in the mixture was then reduced with hydrazine monohydrate under reflux at 100 °C for 24 h. Subsequently, the mixture was filtered through a polytetrafluoroethylene (PTFE) membrane, purified by washing and redispersed in aqueous solution under mild sonication for 30 min, which yielded a homogeneous black-colored suspension without any visible precipitation of rGO/PIL-Br. Following that, CNT/PIL-Br and rGO/PIL-Br were mixed, sonicated for 30 min, and further stirred for 3 h to yield a homogenous and stable aqueous suspension of CNT/PIL-Br/rGO. The as-produced CNT/PIL-Br/rGO suspension was subjected to a phase transfer process with lithium bis(trifluoromethylsulfonyl) amide, Li(NTf2), during which hydrophilic Br− in PIL was substituted with hydrophobic NTf2−, leading to the formation of CNT/PIL-NTf2/rGO. The precipitated CNT/PIL/rGO hybrid materials were collected by centrifugation and filtration, and then dispersed in organic solvents as previously reported.[27,37] Fabrication of Chemical and Temperature Sensor Devices: 2 wt% of CNT/PIL/rGO hybrid materials were dispersed in acetonitrile by mild sonication for 30 min. The suspension was then sprayed onto interdigitated electrodes using spray layer-by-layer (sLbL) technique for chemical sensing test. Resistance of transducers was controlled by number of layers to ensure a good reproducibility of dynamic properties of sensors. Prior to the test, vapor sensors were conditioned at 30 °C overnight in controlled atmosphere. The sensing performances of the devices were evaluated by recording the chemo-resistive responses upon exposure to a standard set of volatile organic compounds (VOCs) such as methanol, benzene, chloroform, and terahydrofuran at room temperature. Gas containing analyte vapors were fluxed into the chamber by a mass flow meter (Bonkhorst Instrument) and high purity nitrogen was used to flush the device between alternating exposures to analyte-containing gas flows. The device response has been estimated via the relative differential resistance responses (ΔR/R0) defined by Equation (1) Ar =
ΔR R − R 0 = R0 R0
(1)
where R is the resistance of the sensing materials when exposed to analyse vapors and R0 is the initial resistance of the sensing materials when exposed to a nitrogen flow. For fabrication of temperature sensors, CNT/PIL/rGO suspension was vacuum-filtered through PTFE membrane and dried at 30 °C overnight in a vacuum oven. The as-obtained 10-µm thick
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
7
full papers www.MaterialsViews.com
hybrid film was cut into a rectangular shape (10 mm in width and 42 mm in length). Silver paste was deposited on both ends of the sample for electrical contact and dried at 50 °C for 3 h. The sample was held between the copper bars and heated from either side, while simultaneously measuring the temperature difference, ΔT, across the sample. The setup consisted of the Peltier devices attached to an aluminum heat sink to create a temperature gradient and two thermocouples to measure the temperature gradient across the sample. The voltage, ΔV, generated by heating at one side of the sample was measured while a constant temperature gradient is forced over the sample.[61]
Supporting Information Supporting Information is available from the Wiley Online Library or from the author.
Acknowledgements This work was partly supported by grants of ICPEES, CNRS-Strasbourg and University University of South Brittany (UBS-Lorient). This research was also supported by the R&D Program for Society of the National Research Foundation (NRF) funded by the Ministry of Science, ICT & Future Planning (Grant No. 2013M3C8A3078806). We thank Prof. Kwang S. Suh, Mr. Hoseok Lee at Korea University and Dr. Isabelle Pillin, Mr. Hervé Béllegou, Dr. Cong-Hanh Pham at LIMATB, the University of South Brittany for their helpful discussion and contributions to this work.
[1] J. Kong, N. R. Franklin, C. Zhou, M. G. Chapline, S. Peng, K. Cho, H. Dai, Science 2000, 287, 622. [2] H. J. Dai, Carbon Nanotubes, Springer, Berlin, Germany 2001. [3] J. Li, Y. Lu, Q. Ye, M. Cinke, J. Han, M. Meyyappan, Nano Lett. 2003, 3, 929. [4] A. Modi, N. Koratkar, E. Lass, B. Wei, P. M. Ajayan, Nature 2003, 424, 171. [5] A. A. Balandin, S. Ghosh, W. Bao, I. Calizo, D. Teweldebrhan, F. Miao, C. N. Lau, Nano Lett. 2008, 8, 902. [6] S. Park, R. S. Ruoff, Nat Nanotechnol. 2009, 4, 217. [7] S. Gilje, S. Han, M. S. Wang, K. L. Wang, R. B. Kaner, Nano Lett. 2007, 7, 3394. [8] W. Gao, L. B. Alemary, L. Gi, P. M. Ajayan, Nature Chem. 2009, 1, 403. [9] F. Schedin, A. K. Geim, S. V. Morozov, E. W. Hill, P. Blake, M. I. Katsnelson, K. S. Novoselov, Nat Mater. 2007, 6, 652. [10] S. Rumyantsev, G. Liu, M. S. Shur, R. A. Potyrailo, A. A. Balandin, Nano Lett. 2012, 12, 2294. [11] Y. M. Lin, P. Avouris, Nano Lett. 2008, 8, 2119 [12] M. F. Yu, O. Lourie, M. J. Dyer, K. Moloni, T. F. Kelly, R. S. Ruoff, Science 2000, 287, 637. [13] C. Lee, X. Wei, J. W. Kysar, J. Hone, Science 2008, 321, 385. [14] L. V. Ruitao, C. S. Eduardo, T. Mauricio, ACS Nano 2014, 8, 4061. [15] F. D. Novaes, R. Rurali, P. Ordejón, ACS Nano 2010, 4, 7596. [16] V. Varshney, S. S. Patnaik, A. K. Roy, G. Froudakis, B. L. Farmer, ACS Nano 2010, 4, 1153.
8 www.small-journal.com
[17] X. Lin, P. Liu, Y. Wei, Q. Li, J. Wang, Y. Wu, C. Feng, L. Zhang, S. Fan, K. Jiang, Nat. Commun. 2013, 4, 2920. [18] D. H. Lee, J. E. Kim, T. H. Han, J. W. Hwang, S. K. Jeon, S. Y. Choi, S. H. Hong, W. J. Lee, R. S. Ruoff, S. O. Kim, Adv. Mater. 2010, 22, 1247. [19] Z. Fan, J. Yan, L. Zhi., Q. Zhang, T. Wei, J. Feng, M. Zhang, W. Qian, F. Wei, Adv. Mater. 2010, 22, 3723. [20] Y. Li, W. Zhou, H. L. Wang, L. M. Xie, Y. Y. Liang, F. Wei, J. C. Idrobo, S. J. Pennycook, H. J. Dai, Nat. Nanotechnol. 2012, 7, 394. [21] R. T. Lv, T. X. Cui, M. S. Jun, Q. Zhang, A. Y. Cao, D. S. Su, Z. J. Zhang, S. H. Yoon, J. Miyawaki, I. Mochida, F. Kang, Adv. Funct. Mater. 2011, 21, 999. [22] H. Sun, Z. Xu, C. Gao, Adv. Mater. 2013, 25, 2554. [23] J. H. Huang, J. H. Fang, C. C. Liu, C. W. Chu, ACS Nano 2011, 5, 6262. [24] D. Yu, L. Dai, J. Phys. Chem. Lett. 2010, 1, 467. [25] S. Some, Y. Kim, Y. Yoon, H. Yoo, S. Lee, Y. Park, H. Lee, Scientific Rep. 2013, 3, 1929. [26] T. Y. Kim, S. W. Kwon, S. J. Park, D. H. Yoon, K. S. Suh, W. S. Yang, Adv. Mater. 2011, 23, 2734. [27] T. Y. Kim, H. W. Lee, J. E. Kim, K. S. Suh, ACS Nano 2010, 4, 1612. [28] C. H. Wang, C. H. Wu, J. W. Wu, M. T. Lee, J. K. Chang, M. D. Gerd, C. L. Sun, Analyst 2013, 138, 576. [29] T. P. Fellinger, F. Hasche, P. Strasser, M. Antonietti, J. Am. Chem. Soc. 2012, 134, 4072. [30] Z. Lin, G. H. Waller, Y. Liu, M. Liu, C. P. Wong, Nano Energy 2013, 2, 241. [31] C. Ferrari, D. M. Basko, Nature Nanotechnol. 2013, 8, 235. [32] C. Shih, A. Vijayaraghavan, R. Krishnan, R. Sharma, J. H. Han, M. H. Ham, Nature Nanotechnol. 2011, 6, 439. [33] A. C. Ferrari, J. C. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri, S. Piscanec, D. Jiang, K. S. Novoselov, S. Roth, A. K. Geim, Phys. Rev. Lett. 2006, 97, 187401. [34] T. T. Tung, H. Ba, T. P. Lai, J. M. Nhut, O. Ersen, D. Begin, I. Janowskaa, D. L. Nguyenc, P. Granger, P. H. Cuong, J. Mater. Chem. A. 2014, 2, 11349. [35] X. Li, W. Cai, J. An, S. Kim, J. Nah, D. Yang, R. Piner, A. Velamakanni, I. Jung, E. Tutuc, S. K. Banerjee, L. Colombo, R. S. Ruoff, Science 2009, 324, 1312. [36] Y. F. Li, Y. Z. Liu, Y. G. Yang, M. Z. Wang, Y. F. Wen, Appl. Phys. A 2012, 108, 701. [37] T. Y. Kim, T. T. Tung, T. H. Lee, J. E. Kim, K. S. Suh, Chem. Asian J. 2010, 5, 256. [38] T. Y. Kim, H. W. Lee, M. Stoller, D. R. Dreyer, C. W. Bielawski, R. S. Ruoff, K. S. Suh, ACS Nano. 2011, 5, 436. [39] T. Y. Kim, S. W. Kwon, S. J. Park, D. H. Yoon, K. S. Suh, W. S. Yang, Adv. Mater. 2011, 23, 2734. [40] T. T. Tung, M. Castro, T. Y. Kim, K. S. Suh, J. F. Feller, J. Mater. Chem. 2012, 22, 21754. [41] J. T. Robinson, F. K. Perkins, E. S. Snow, Z. Wei, P. E. Sheehan, Nano Lett. 2008, 8, 3137. [42] T. Gao, M. D. Woodka, B. S. Brunschwig, N. S. Lewis, Chem. Mater. 2006, 18, 5193. [43] J. F. Feller, J. Lu, K. Zhang, B. Kumar, M. Castro, N. Gatt, H. J. Choi, J. Mater. Chem. 2011, 21, 4142. [44] J. Li, Y. Lu, Q. Ye, M. Cinke, J. Han, M. Meyyappan, Nano Lett. 2003, 3, 929. [45] A. Carrillo, I. R. Martin-Dominguez, A. Marquez-Lucero, Sens. Actuators B-Chem. 2006, 113, 477. [46] M. Knite, K. Ozols, G. Sakale, V. Teteris, Sens. Actuators B-Chem. 2007, 126, 209. [47] P. J. W. Handsa, P. J. Laughlin, D. Bloor, Sens. Actuators B 2012, 162, 400. [48] S. Chatterjee, M. Castro, J. F. Feller, J. Mater. Chem. B 2013, 1, 4563.
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2015, DOI: 10.1002/smll.201403693
www.MaterialsViews.com [49] T. T. Tung, M. Castro, I. Pillin, T. Y. Kim, K. S. Suh, J. F. Feller, Carbon 2014, 74, 104. [50] K. A. Mirica, J. M. Azzarelli, J. G. Weis, J. M. Schnorr, T. M. Swager, PNAS 2013, 110, E3265. [51] S. Some, Y. Xu, Y. Kim, Y. Yoon, H. Qin, A. Kulkarni, T. Kim, H. Lee, Scientific Rep. 2012, 3, 2868. [52] S. Konwer, A. K. Guha, S. K. Dolui, J. Mater. Sci. 2013, 48, 1729. [53] G. J. Snyder, E. S. Toberer, Nat. Mater. 2008, 7, 105. [54] A. Hilaal, R. Seeram, Nano Energy 2013, 2, 190. [55] W. Liu, X. Yan, G. Chen, Z. Ren, Nano Energy 2012, 42. [56] C. A. Hewitt, A. B. Kaiser, S. Roth, M. Craps, R. Czerw, D. L. Carroll, Nano Lett. 2012, 12, 1307.
small 2015, DOI: 10.1002/smll.201403693
[57] C. Bounioux, P. Dıaz-Chao, M. C. Quiles, M. S. Martın-Gonz´alez, A. R. Goni, R. Yerushalmi-Rozene, C. Muller, Energy Environ. Sci. 2013, 6, 918. [58] K. Zhang, Y. Zhang, S. Wang, Sci. Rep. 2013, 3, 3448. [59] T. T. Tung, T. Y. Kim, K. S. Suh, Org. Electro 2011, 12, 22. [60] T. T. Tung, T. Y. Kim, J. P. Shim, W. S. Yang, H. K. Kim, K. S. Suh, Org Electro 2011, 12, 2215. [61] Z. Antar, J. F. Feller, H. Noël, P. Glouannec, K. Elleuch, Mater. Lett. 2012, 67, 210.
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Received: December 13, 2014 Revised: February 23, 2015 Published online:
www.small-journal.com
9
Sensors
Hybrid Films of Graphene and Carbon Nanotubes for High Performance Chemical and Temperature Sensing Applications Tran Thanh Tung,* Cuong Pham-Huu, Izabela Janowska, TaeYoung Kim,* Mickael Castro, and Jean-Francois Feller
A
hybrid composite material of graphene and carbon nanotube (CNT) for high performance chemical and temperature sensors is reported. Integration of 1D and 2D carbon materials into hybrid carbon composites is achieved by coupling graphene and CNT through poly(ionic liquid) (PIL) mediated-hybridization. The resulting CNT/PIL/graphene hybrid materials are explored as active materials in chemical and temperature sensors. For chemical sensing application, the hybrid composite is integrated into a chemo-resistive sensor to detect a general class of volatile organic compounds. Compared with the graphene-only devices, the hybrid film device showed an improved performance with high sensitivity at ppm level, low detection limit, and fast signal response/recovery. To further demonstrate the potential of the hybrid films, a temperature sensor is fabricated. The CNT/PIL/graphene hybrid materials are highly responsive to small temperature gradient with fast response, high sensitivity, and stability, which may offer a new platform for the thermoelectric temperature sensors.
1. Introduction Low-dimensional carbon structures such as 1D carbon nanotubes (CNTs) and 2D graphene holds great promise for potential sensor applications because of their unique Dr. T. T. Tung, Dr. C. Pham-Huu, Dr. I. Janowska Institut de Chimie et Procédés pour l’Energie l’Environnement et la Santé (ICPEES) ECPM, UMR 7515 du CNRS-Université de Strasbourg 25 rue Becquerel, 67087, Strasbourg Cedex 02, France E-mail: [email protected] Prof. T. Kim Department of Bionanotechnology Gachon University 1342 Seongnamdaero, Sujeong-gu, Seongnam-si Gyeonggi-do 461-701, South Korea E-mail: [email protected] Dr. T. T. Tung, Dr. M. Castro, Prof. J.-F. Feller Smart Plastics Group European University of Brittany (UEB), LIMATB-UBS rue de Saint-Maude´ 56321 Lorient, France DOI: 10.1002/smll.201403693 small 2015, DOI: 10.1002/smll.201403693
nanoscale features, large specific surface area, high carrier mobility, and facile chemical functionalization.[1–8] Previous studies show that their electrical resistance (or conductance) modulates upon exposure to chemical species or environmental changes. It has been shown that the large surface area of these carbons allows for the attachment of a large amount of analyte molecules on their surfaces, leading to the high sensitivity and signal-to-noise ratio (SNR) of the sensor devices, and the fast electron transfer between the carbon and analyte enables the direct electrochemical reaction with rapid response time.[9–11] In addition, their excellent mechanical properties allow for the integration into flexible sensing devices.[12,13] In this context, assembling graphene nanosheets into interconnected CNT networks is of special interest for the realistic development of a new sensing platform because the resulting hybrid materials would offer a combined set of properties such as reinforced mechanical strength and straightforward molecular transport through interconnected pathway, leading to improved electronic and mechanical properties.[15,16] Such hybrid materials have been prepared by either noncovalent bonding (e.g., solution mixing of CNT and graphene) or covalent bonding (e.g., CVD growth of CNT on
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
1
full papers www.MaterialsViews.com
graphene) for applications in the field of flexible electronics, energy storage, catalyst, and multifunctional composite.[17–24] Despite the potential advantages offered by the combination of the 1D- and 2D-structured carbon materials, the exploitation of these hybrid composites as sensor platforms suffered from several practical issues such as controlled assembly of graphene and CNTs on the molecular level and lack of efficient functionalization of carbons on sensing performance. Herein, we present a method to produce a carbon hybrid material consisting of CNTs and reduced graphene oxide (rGO). For the efficient hybridization of CNTs and rGO sheets, we used a ionic liquid-based polymer, referred to as poly(ionic liquids) (PILs), as a stabilizer and linker. In the course of the hybridization process, PIL molecules not only improve the dispersion of CNTs and rGO sheets in solvents against uncontrolled aggregation, but also act as a linker to assemble the two carbons into the nanostructured hybrid materials. Coupling of CNTs and rGO by PIL-mediated hybridization yielded the complex of CNT/PIL/rGO, in which each component was assembled on the molecular level through electrostatic and π–π stacking interactions. The CNT/ PIL/rGO hybrid composites were tested for the two different types of sensing applications such as chemical and temperature sensors. For the chemical sensor application, the CNT/ PIL/rGO film was deposited on an interdigitated electrode and integrated into the chemo-resistive sensors. Upon exposure to various volatile organic compound (VOC) vapors, the device showed an adjustable conductance with a rapid and selective detection capability. The hybrid film was further explored as a new type of temperature sensing platform. The free-standing film of CNT/PIL/rGO was capable to detect
temperature by monitoring temperature-activated carrier transport from hot end to cold end. The potential of the CNT/ PIL/rGO hybrid as active material for chemical and temperature sensors is discussed in terms of their performance.
2. Preparation of CNT/PIL/rGO Hybrid Materials The CNT/PIL/rGO hybrid materials were prepared through the following three steps: (i) modification of CNTs by water soluble PIL of poly(1-vinyl-3-ethylimidazolium) bromide (PIL-Br) to form an aqueous dispersion of PIL-modified CNTs (CNT/PIL); (ii) modification of G–O by hydrophilic PIL followed by chemical conversion into rGO/PIL; (iii) mixing of CNT/PIL and rGO/PIL and subsequent phase transfer reaction by exchanging the counter anion, Br− in PIL with NTf2− to form a hydrophobic CNT/PIL/rGO hybrid composite. The resulting CNT/PIL/rGO hybrid materials were characterized by a number of methods including X-ray diffraction (XRD), FTIR, X-ray photoelectric spectroscopy (XPS), and Raman spectroscopy. Figure 1a shows the powder XRD pattern of the CNT/ PIL/rGO hybrid materials. The XRD pattern showed a broad peak in the 10°–30° 2θ region in which multiple diffraction peaks are overlapped. A (002) deconvoluted diffraction peak at 2θ of 20.1° is assigned to rGO platelets with an interlayer spacing of 0.44 nm. The interlayer distance is larger than that of graphite (0.34 nm), attributable to the residual oxygen functional groups of rGO as well as PIL held in the interlayer
Figure 1. a) Schematic illustration of CNT/PIL/rGO hybrid materials; b) XRD pattern; c) C 1s XPS spectra; and d) Raman spectra of CNT/PIL/rGO hybrid materials.
2 www.small-journal.com
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2015, DOI: 10.1002/smll.201403693
www.MaterialsViews.com
Figure 2. a,b) SEM images; c,d) TEM images of CNT/PIL/rGO hybrid materials.
galleries of rGO.[25] Two peaks at 2θ angle of 26.5° and 43° are likely due to the (002) and (004) diffraction of CNTs, indicating that CNT is intermixed with rGO within the hybrid materials. Chemical analysis of CNT/PIL/rGO was conducted by FTIR (Figure S1, Supporting Information) and XPS (Figure S2, Supporting Information). The survey spectrum of XPS clearly indicates the presence of C (62.24 at%), O (12.53 at%), N (8.06 at%), F (12.30 at%), and S (4.51 at%). The C1s XPS spectrum of CNT/PIL/rGO could be deconvoluted into different peaks that correspond to the following functional groups: carbon sp2 (C = C in aromatic ring, 284.5 eV), epoxy/hydroxyls (C–O, 286.2 eV), and carbonyl (C = O, 287.2 eV) and carboxylates (O–C = O, 288.8 eV).[25–27] Two additional peaks corresponding to carbon bound to nitrogen (C–N, 286.2 eV) and fluorine (C–F, 293.0 eV) were due to the presence of PIL. In addition, the nitrogen content and the nature of the different nitrogen species are presented in Figure S1c (Supporting Information). The nitrogen content in CNT/PIL/rGO was found relatively high (≈8 at%) as compared with that for the chemically reduced rGO (0.8 at%). This is due to the presence of PIL in the CNT/PIL/rGO composite, as the PIL molecules is composed of a nitrogen-containing cation (imidazolium) and anion (NTf2−).[28,29] Raman spectrum was recorded with a laser excitation wavelength of 532 nm under ambient condition. For Raman analysis, the samples were prepared by depositing CNT/PIL/rGO suspension onto glass slide substrate. The Raman spectra show a typical D-band (1345 cm−1) and G-band (1579 cm−1) corresponding to structure defects of sp3 domain and the E2g-vibration mode of sp2 carbon, respectively.[26–28] The 2D-band originating from two phonon double resonance small 2015, DOI: 10.1002/smll.201403693
Raman process was also observed at 2683 cm−1 and the peak at 2920 cm−1 is the D+D′ band corresponding to a defect for an elastic scattering event to provide momentum conservation in the Raman process.[29–33] The intensity of 2D peak is relatively small due to the disorder, and as compared with expanded graphite the 2D peak position in PIL-modified rGO shifted considerably to a lower wavenumber by 33 cm−1, from 2720 to 2687 cm−1, suggesting a successful exfoliation to single or few layer graphene sheets in rGO/PIL.[26,27,34] The 2D band of the CNT/PIL/rGO hybrid is different from that of rGO/PIL complex, which could be the super-imposed peaks due to the electronic band structure of CNTs in the composite.[35,36] All Raman peaks of CNT/PIL/rGO hybrid materials were found to be red-shifted as compared to CNT/ PIL and blue-shifted as to rGO/PIL (Figure S3, Supporting Information). This is an indication that CNTs and rGO are assembled into hybrid composite by PIL-mediated process, suggesting that PIL plays an important role in coupling CNT and rGO on the molecular level as well as stabilizing them in the suspension. The morphology of the CNT/PIL/rGO hybrid materials was examined by scanning electronic microscopy (SEM), transmission electron microscopy (TEM) and atomic force microscopy (AFM). Figure 2a,b displays typical SEM images of CNT/PIL/rGO hybrids at different magnification, showing a CNT network covered by overlapped rGO sheets. TEM images in Figure 2c,d showed that CNTs with lengths of several hundred nanometers to micrometers are adhered on the surface of rGO sheets. The higher resolution TEM image (Figure 2d) revealed that individual CNTs are well-dispersed on the rGO sheets. It is noted that the coupling of CNT and
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
3
full papers www.MaterialsViews.com
Figure 3. Tapping mode AFM topographic images and height profiles of CNT/PIL/rGO hybrid films.
rGO would be the result of the strong cation-π interaction between π electron clouds of sp2-hybridized carbons and imidazolium cation of PIL, which in turn allows for the hybridization of CNT and rGO sheets.[37–39] We also used AFM to characterize the topography of hybrid CNT/PIL/rGO film, which were deposited on a Si/SiO2 substrate by spin-coating of the suspension (in acetonitrile, 0.3 mg mL–1). Figure 3 shows a typical tapping mode AFM images and height profile displaying the morphology and thickness of the CNT/PIL/rGO hybrid films. From the height profiles, the thickness of rGO/PIL and CNT/PIL in the composite was found in the range of 1.5−2.5 nm and 10–14 nm, respectively, which is consistent with those measured for PIL-coated rGO and PIL-coated CNTs (Figure S4, Supporting Information). Some of CNTs and rGO sheets were also found to protrude from the composite films, resulting in a rough surface morphology at nanoscale.
3. Chemical Sensing Application of CNT/PIL/ rGO Hybrids As a proof-of-concept experiment, we fabricated a chemoresistive sensor by spray-coating CNT/PIL/rGO hybrid on an interdigitated electrode array (Figure 4a). The relative differential resistance responses (Ar = ΔR/R0)
4 www.small-journal.com
to various analyte vapors of sensor devices was recorded as detailed in the Experimental Section. Prior to flowing analyte vapors, sensor arrays were stabilized to show constant electrical resistance in dry N2 flow. Upon exposure to analyte vapors, their resistance increased sharply, giving a positive vapor response of ΔR/R0. When analyte vapor flow was turned off and N2 gas flow restored, resistance of the sensors went back to its initial resistance, indicating a complete desorption of vapor molecules from the sensing materials and full recovery of the device. Vapor exposures were interlaced between recovery periods to show reproducibility. Figure 4b shows a typical dynamic response (ΔR/R0 vs time) of a CNT/PIL/rGO device for the detection of various analyte vapors, such as methanol, chloroform, tetrahydrofuran (THF), and benzene at saturated concentration. The as-fabricated sensor devices showed strong responses to all analyte vapors with large changes in the resistance of CNT/ PIL/rGO. The device was highly responsive to a 5% diluted NH3 solution with an amplitude of Ar ≈3.6. The CNT/PIL/ rGO device also responded to methanol with an amplitude of Ar ≈2.2. Considering that the previously reported value for the rGO/PIL-based sensor was Ar ≈1.68, the device in this work showed that the sensitivity was increased by 1.4fold.[40] The significantly improved sensitivity of CNT/PIL/ rGO sensors reported here is mainly attributed to the combined effect of hybridization of 1D and 2D carbons and the
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2015, DOI: 10.1002/smll.201403693
www.MaterialsViews.com
Figure 4. Chemo-resistive responses of CNT/PIL/rGO composite-based sensors: a) schematic diagram to describe the structure of sensor device; b) Normalized resistance change versus time upon sequential exposure to different analytes; c) Normalized resistance change versus time with respect to vapor concentration d) Plot of normalized resistance change versus vapor concentration for different analytes.
presence of PIL molecules between the assembled carbons. It is well known that vapor molecules are adsorbed onto binding sites with low energy such as graphitic sp2–bonded carbon and high-energy binding sites such as vacancies, structural defects, and functional groups on the surface of carbon nanomaterials, leading to the change in resistance.[41] The interconnected networks of 1D CNT and 2D rGO in the hybrid composites would provide more binding sites available for molecular adsorption and therefore enhance the sensing response. In addition, PIL located between the CNT and rGO sheets also serve to absorb VOC vapor molecules and induce changes in their physical properties. Upon exposure to analyte vapors, PIL molecules in the carbon network could swell and the chains rearrange in a way to disrupt the conductive carbon channels with an increased resistance. The ranking of the response for sensor toward different analytes is found as followed: Armethanol > Archloroform > Artetrahydrofuran > Arbenzene. The sensing mechanism can be explained by the van der Waals interactions between analyte molecules and the hybrid sensing materials, which differ with interaction parameters such as diffusion coefficient, average number of solvent molecules per cluster, extra clustering and so on. The presented data are consistent with the Langmuir–Henry– Clustering (LHC) electrosorption model as detailed in the previous reports for conductive polymer composite.[42,43] Figure 4c presented a relationship between sensing response and the concentration for four analyte vapors. When exposed to successively increased concentration of analyte vapors in the range of 1–150 ppm, the resistance small 2015, DOI: 10.1002/smll.201403693
change increased in a linear fashion, as shown in Figure 4d. The SNRs of the sensors at 1 ppm concentration of analyte molecules were estimated as 22.8, 34.5, 48.6, and 59.3 for methanol, chloroform, tetrahydrofuran, and benzene, respectively, as shown in Table S1 (Supporting Information). According to the International Union of Pure and Applied Chemistry (IUPAC) definition, the signal is considered to be a true signal, when the SNR ratio equals 3.[44] Therefore, the results in the present work suggested that the detection limit of CNT/PIL/rGO sensor is in the ppm range, and it can be possibly down to sub-ppm levels. The response times of sensor at a 1 ppm concentration were estimated as 20, 26, 30, and 40 s for methanol, chloroform, tetrahydrofuran, and benzene, respectively (see Figure S5, Supporting Information). This fairly fast response of the device is attributed to the swelling of PIL molecules by molecular adsorption and the relevant diffusion dynamics. The swelling of PIL at the interface of CNT and rGO would result in the separation of the two carbon/ carbon nanojuctions, thereby affecting the tunneling current through the composite.[45–47] Moreover, the higher the diffusion coefficient of analyte molocules through the percolated structure (Dc, Table S2, Supporting Information) the quicker the response. As compared with the previously reported rGO-only sensors, the present device showed an improved performance in terms of sensitivity, detection limit, and fast signal response/recovery toward VOCs and is consistent with other recent progress in nanocomposites based on CNTs and graphene for chemical sensors application.[48–52]
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
5
full papers www.MaterialsViews.com
Figure 5. Experiment setup for temperature sensing characterization.
4. Temperature Sensing Application of CNT/ PIL/rGO Hybrids We further explored the potential use of CNT/PIL/rGO hybrid as thermoelectric temperature sensing materials. For the device fabrication (Figure 5), thin film of the hybrid materials was held between the copper bars and heated from either side using two Peltier modules, while the temperature difference (ΔT) across the sample was monitored with two thermocouples attached to each ends. The voltage (ΔV) generated by temperature difference between the hot and cold ends, was measured over time as shown in Figure 6. Figure 6a shows the response of the device to the stepwise temperature changes, in which the output voltage
curve matches with the temperature profile. The Figure 6b showed the relationship between output voltage of the CNT/PIL/rGO device and temperature difference. The output voltage increases linearly with the temperatures, showing a near-room temperature operation of the device. Figure 6c showed a variation of output voltage of the device under cyclic changes of temperature. The device shows a rapid response and recovery to the cyclic temperature changes without any noticeable time lag. Therefore, the device was highly responsive to the temperature, which is ascribed to the thermoelectric properties of CNT/ PIL/rGO hybrid composite. When a temperature gradient is applied, the excess charge carriers migrate to the cold side of the sample, that the build-up of carriers generate
Figure 6. a) Output voltage of CNT-PIL-rGO composite-based sensor upon a step-wise thermal input; b) Plot of output voltage of the sensor versus temperature; c) Output voltage of the sensor upon cyclic thermal input; d) I–V characteristic of the composites measured between the forward bias (+5 V) and reserve bias (–5 V)
6 www.small-journal.com
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2015, DOI: 10.1002/smll.201403693
www.MaterialsViews.com
a voltage across the materials. This voltage generated by the thermoelectric materials is governed by its Seebeck coefficient (the amount of voltage generated across the material at given temperature gradient (ΔV/ΔT).[53,54] The highly sensitive thermoelectric response of CNT/PIL/rGO composite materials is attributed to the synergistic effects of the individual components, in which PIL plays an important role in enhancing thermoelectric properties. The PIL between the CNT and rGO sheets is proposed to induce the energy filtering or phonon scattering, but diminish carrier transport, thereby leading to an increase in Seebeck coefficient.[55–58] To verify the proposed mechanism, the current–voltage (I–V) characteristic of the CNT/PIL/rGO composite was investigated. Figure 6d shows the I–V curve of the device measured in the bias voltage range of –5 to +5 V. While symmetric, the I–V curve of CNT/PIL/rGO composites is not linear, showing the electrical behavior similar to that of the previously reported semiconducting materials. This suggests that PIL between CNTs and rGO separate two carbon nanomaterials, leading to the hopping or tunneling conduction as the dominant charge-transport mechanism. The sensitivity of the device was further investigated by measuring the minimum thermal input required to produce an output voltage. For this purpose, we monitored the output voltage variation of the composite generated under the stepwise temperature change by few degrees. In Figure S6 (Supporting Information), the sensor showed a remarkable voltage change under subtle temperature change of ≈1 °C, while the output voltage curve matches with the temperature profile. Considering the high sensitivity and stability under room temperature operation, the present device may hold the potential as a temperature sensor in a wide range of applications such as public safety, automotive, and green house.
5. Conclusion A hybrid composite material of 1D CNTs and 2D rGO sheet have been prepared using PIL as inter-linker for coupling them through noncovalent interactions. In this work, the hybrid material was explored for dual missions: (i) detection of trace amount of organic vapors and (ii) monitoring environmental temperatures. For chemical sensing application, the CNT/PIL/rGO hybrid material was used as a chemoresistor to detect various organic vapors at ppm level at room temperature. The CNT/PIL/rGO sensor was found to be selective toward a general class of organic vapors and showed an improved performance in terms of sensitivity, detection limit, and fast signal response. We also demonstrated a temperature sensor based on the thermoelectric properties of CNT/PIL/rGO hybrid material. The device was highly responsive to even small temperature gradient with fast response time, demonstrating the potential of the CNT/ PIL/rGO hybrid as a new type of temperature sensing material. The CNT/PIL/rGO composite presented in this work may offer an alternative platform for the chemical and thermoelectric temperature sensors. small 2015, DOI: 10.1002/smll.201403693
6. Experimental Section Preparation of PIL-Mediated Hybridization of CNTs and Graphene: Imidazolium-based poly(ionic liquids) such as poly(1-vinyl-3-ethylimidazolium) bromide (PIL-Br) and poly(1-vinyl3-ethylimidazolium) bis(trifluoromethylsulfonyl)amide (PIL-NTf2) were prepared according to our previous reports.[59,60] Graphite oxide (GO) was prepared by modified Hummers method (as previous reported[34] and oxidized multiwalled carbon nanotube (CNT) was kindly provided by Nanocyl. First, 10 mg of CNTs and 50 mg of PIL-Br were mixed in 100 mL of DI water and sonicated for 1 h to form an aqueous suspension of CNT/PIL-Br. Separately, 100 mg of GO and 200 mg of PIL-Br were mixed in 100 mL of DI water under stirring for 1 h. The GO in the mixture was then reduced with hydrazine monohydrate under reflux at 100 °C for 24 h. Subsequently, the mixture was filtered through a polytetrafluoroethylene (PTFE) membrane, purified by washing and redispersed in aqueous solution under mild sonication for 30 min, which yielded a homogeneous black-colored suspension without any visible precipitation of rGO/PIL-Br. Following that, CNT/PIL-Br and rGO/PIL-Br were mixed, sonicated for 30 min, and further stirred for 3 h to yield a homogenous and stable aqueous suspension of CNT/PIL-Br/rGO. The as-produced CNT/PIL-Br/rGO suspension was subjected to a phase transfer process with lithium bis(trifluoromethylsulfonyl) amide, Li(NTf2), during which hydrophilic Br− in PIL was substituted with hydrophobic NTf2−, leading to the formation of CNT/PIL-NTf2/rGO. The precipitated CNT/PIL/rGO hybrid materials were collected by centrifugation and filtration, and then dispersed in organic solvents as previously reported.[27,37] Fabrication of Chemical and Temperature Sensor Devices: 2 wt% of CNT/PIL/rGO hybrid materials were dispersed in acetonitrile by mild sonication for 30 min. The suspension was then sprayed onto interdigitated electrodes using spray layer-by-layer (sLbL) technique for chemical sensing test. Resistance of transducers was controlled by number of layers to ensure a good reproducibility of dynamic properties of sensors. Prior to the test, vapor sensors were conditioned at 30 °C overnight in controlled atmosphere. The sensing performances of the devices were evaluated by recording the chemo-resistive responses upon exposure to a standard set of volatile organic compounds (VOCs) such as methanol, benzene, chloroform, and terahydrofuran at room temperature. Gas containing analyte vapors were fluxed into the chamber by a mass flow meter (Bonkhorst Instrument) and high purity nitrogen was used to flush the device between alternating exposures to analyte-containing gas flows. The device response has been estimated via the relative differential resistance responses (ΔR/R0) defined by Equation (1) Ar =
ΔR R − R 0 = R0 R0
(1)
where R is the resistance of the sensing materials when exposed to analyse vapors and R0 is the initial resistance of the sensing materials when exposed to a nitrogen flow. For fabrication of temperature sensors, CNT/PIL/rGO suspension was vacuum-filtered through PTFE membrane and dried at 30 °C overnight in a vacuum oven. The as-obtained 10-µm thick
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.small-journal.com
7
full papers www.MaterialsViews.com
hybrid film was cut into a rectangular shape (10 mm in width and 42 mm in length). Silver paste was deposited on both ends of the sample for electrical contact and dried at 50 °C for 3 h. The sample was held between the copper bars and heated from either side, while simultaneously measuring the temperature difference, ΔT, across the sample. The setup consisted of the Peltier devices attached to an aluminum heat sink to create a temperature gradient and two thermocouples to measure the temperature gradient across the sample. The voltage, ΔV, generated by heating at one side of the sample was measured while a constant temperature gradient is forced over the sample.[61]
Supporting Information Supporting Information is available from the Wiley Online Library or from the author.
Acknowledgements This work was partly supported by grants of ICPEES, CNRS-Strasbourg and University University of South Brittany (UBS-Lorient). This research was also supported by the R&D Program for Society of the National Research Foundation (NRF) funded by the Ministry of Science, ICT & Future Planning (Grant No. 2013M3C8A3078806). We thank Prof. Kwang S. Suh, Mr. Hoseok Lee at Korea University and Dr. Isabelle Pillin, Mr. Hervé Béllegou, Dr. Cong-Hanh Pham at LIMATB, the University of South Brittany for their helpful discussion and contributions to this work.
[1] J. Kong, N. R. Franklin, C. Zhou, M. G. Chapline, S. Peng, K. Cho, H. Dai, Science 2000, 287, 622. [2] H. J. Dai, Carbon Nanotubes, Springer, Berlin, Germany 2001. [3] J. Li, Y. Lu, Q. Ye, M. Cinke, J. Han, M. Meyyappan, Nano Lett. 2003, 3, 929. [4] A. Modi, N. Koratkar, E. Lass, B. Wei, P. M. Ajayan, Nature 2003, 424, 171. [5] A. A. Balandin, S. Ghosh, W. Bao, I. Calizo, D. Teweldebrhan, F. Miao, C. N. Lau, Nano Lett. 2008, 8, 902. [6] S. Park, R. S. Ruoff, Nat Nanotechnol. 2009, 4, 217. [7] S. Gilje, S. Han, M. S. Wang, K. L. Wang, R. B. Kaner, Nano Lett. 2007, 7, 3394. [8] W. Gao, L. B. Alemary, L. Gi, P. M. Ajayan, Nature Chem. 2009, 1, 403. [9] F. Schedin, A. K. Geim, S. V. Morozov, E. W. Hill, P. Blake, M. I. Katsnelson, K. S. Novoselov, Nat Mater. 2007, 6, 652. [10] S. Rumyantsev, G. Liu, M. S. Shur, R. A. Potyrailo, A. A. Balandin, Nano Lett. 2012, 12, 2294. [11] Y. M. Lin, P. Avouris, Nano Lett. 2008, 8, 2119 [12] M. F. Yu, O. Lourie, M. J. Dyer, K. Moloni, T. F. Kelly, R. S. Ruoff, Science 2000, 287, 637. [13] C. Lee, X. Wei, J. W. Kysar, J. Hone, Science 2008, 321, 385. [14] L. V. Ruitao, C. S. Eduardo, T. Mauricio, ACS Nano 2014, 8, 4061. [15] F. D. Novaes, R. Rurali, P. Ordejón, ACS Nano 2010, 4, 7596. [16] V. Varshney, S. S. Patnaik, A. K. Roy, G. Froudakis, B. L. Farmer, ACS Nano 2010, 4, 1153.
8 www.small-journal.com
[17] X. Lin, P. Liu, Y. Wei, Q. Li, J. Wang, Y. Wu, C. Feng, L. Zhang, S. Fan, K. Jiang, Nat. Commun. 2013, 4, 2920. [18] D. H. Lee, J. E. Kim, T. H. Han, J. W. Hwang, S. K. Jeon, S. Y. Choi, S. H. Hong, W. J. Lee, R. S. Ruoff, S. O. Kim, Adv. Mater. 2010, 22, 1247. [19] Z. Fan, J. Yan, L. Zhi., Q. Zhang, T. Wei, J. Feng, M. Zhang, W. Qian, F. Wei, Adv. Mater. 2010, 22, 3723. [20] Y. Li, W. Zhou, H. L. Wang, L. M. Xie, Y. Y. Liang, F. Wei, J. C. Idrobo, S. J. Pennycook, H. J. Dai, Nat. Nanotechnol. 2012, 7, 394. [21] R. T. Lv, T. X. Cui, M. S. Jun, Q. Zhang, A. Y. Cao, D. S. Su, Z. J. Zhang, S. H. Yoon, J. Miyawaki, I. Mochida, F. Kang, Adv. Funct. Mater. 2011, 21, 999. [22] H. Sun, Z. Xu, C. Gao, Adv. Mater. 2013, 25, 2554. [23] J. H. Huang, J. H. Fang, C. C. Liu, C. W. Chu, ACS Nano 2011, 5, 6262. [24] D. Yu, L. Dai, J. Phys. Chem. Lett. 2010, 1, 467. [25] S. Some, Y. Kim, Y. Yoon, H. Yoo, S. Lee, Y. Park, H. Lee, Scientific Rep. 2013, 3, 1929. [26] T. Y. Kim, S. W. Kwon, S. J. Park, D. H. Yoon, K. S. Suh, W. S. Yang, Adv. Mater. 2011, 23, 2734. [27] T. Y. Kim, H. W. Lee, J. E. Kim, K. S. Suh, ACS Nano 2010, 4, 1612. [28] C. H. Wang, C. H. Wu, J. W. Wu, M. T. Lee, J. K. Chang, M. D. Gerd, C. L. Sun, Analyst 2013, 138, 576. [29] T. P. Fellinger, F. Hasche, P. Strasser, M. Antonietti, J. Am. Chem. Soc. 2012, 134, 4072. [30] Z. Lin, G. H. Waller, Y. Liu, M. Liu, C. P. Wong, Nano Energy 2013, 2, 241. [31] C. Ferrari, D. M. Basko, Nature Nanotechnol. 2013, 8, 235. [32] C. Shih, A. Vijayaraghavan, R. Krishnan, R. Sharma, J. H. Han, M. H. Ham, Nature Nanotechnol. 2011, 6, 439. [33] A. C. Ferrari, J. C. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri, S. Piscanec, D. Jiang, K. S. Novoselov, S. Roth, A. K. Geim, Phys. Rev. Lett. 2006, 97, 187401. [34] T. T. Tung, H. Ba, T. P. Lai, J. M. Nhut, O. Ersen, D. Begin, I. Janowskaa, D. L. Nguyenc, P. Granger, P. H. Cuong, J. Mater. Chem. A. 2014, 2, 11349. [35] X. Li, W. Cai, J. An, S. Kim, J. Nah, D. Yang, R. Piner, A. Velamakanni, I. Jung, E. Tutuc, S. K. Banerjee, L. Colombo, R. S. Ruoff, Science 2009, 324, 1312. [36] Y. F. Li, Y. Z. Liu, Y. G. Yang, M. Z. Wang, Y. F. Wen, Appl. Phys. A 2012, 108, 701. [37] T. Y. Kim, T. T. Tung, T. H. Lee, J. E. Kim, K. S. Suh, Chem. Asian J. 2010, 5, 256. [38] T. Y. Kim, H. W. Lee, M. Stoller, D. R. Dreyer, C. W. Bielawski, R. S. Ruoff, K. S. Suh, ACS Nano. 2011, 5, 436. [39] T. Y. Kim, S. W. Kwon, S. J. Park, D. H. Yoon, K. S. Suh, W. S. Yang, Adv. Mater. 2011, 23, 2734. [40] T. T. Tung, M. Castro, T. Y. Kim, K. S. Suh, J. F. Feller, J. Mater. Chem. 2012, 22, 21754. [41] J. T. Robinson, F. K. Perkins, E. S. Snow, Z. Wei, P. E. Sheehan, Nano Lett. 2008, 8, 3137. [42] T. Gao, M. D. Woodka, B. S. Brunschwig, N. S. Lewis, Chem. Mater. 2006, 18, 5193. [43] J. F. Feller, J. Lu, K. Zhang, B. Kumar, M. Castro, N. Gatt, H. J. Choi, J. Mater. Chem. 2011, 21, 4142. [44] J. Li, Y. Lu, Q. Ye, M. Cinke, J. Han, M. Meyyappan, Nano Lett. 2003, 3, 929. [45] A. Carrillo, I. R. Martin-Dominguez, A. Marquez-Lucero, Sens. Actuators B-Chem. 2006, 113, 477. [46] M. Knite, K. Ozols, G. Sakale, V. Teteris, Sens. Actuators B-Chem. 2007, 126, 209. [47] P. J. W. Handsa, P. J. Laughlin, D. Bloor, Sens. Actuators B 2012, 162, 400. [48] S. Chatterjee, M. Castro, J. F. Feller, J. Mater. Chem. B 2013, 1, 4563.
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
small 2015, DOI: 10.1002/smll.201403693
www.MaterialsViews.com [49] T. T. Tung, M. Castro, I. Pillin, T. Y. Kim, K. S. Suh, J. F. Feller, Carbon 2014, 74, 104. [50] K. A. Mirica, J. M. Azzarelli, J. G. Weis, J. M. Schnorr, T. M. Swager, PNAS 2013, 110, E3265. [51] S. Some, Y. Xu, Y. Kim, Y. Yoon, H. Qin, A. Kulkarni, T. Kim, H. Lee, Scientific Rep. 2012, 3, 2868. [52] S. Konwer, A. K. Guha, S. K. Dolui, J. Mater. Sci. 2013, 48, 1729. [53] G. J. Snyder, E. S. Toberer, Nat. Mater. 2008, 7, 105. [54] A. Hilaal, R. Seeram, Nano Energy 2013, 2, 190. [55] W. Liu, X. Yan, G. Chen, Z. Ren, Nano Energy 2012, 42. [56] C. A. Hewitt, A. B. Kaiser, S. Roth, M. Craps, R. Czerw, D. L. Carroll, Nano Lett. 2012, 12, 1307.
small 2015, DOI: 10.1002/smll.201403693
[57] C. Bounioux, P. Dıaz-Chao, M. C. Quiles, M. S. Martın-Gonz´alez, A. R. Goni, R. Yerushalmi-Rozene, C. Muller, Energy Environ. Sci. 2013, 6, 918. [58] K. Zhang, Y. Zhang, S. Wang, Sci. Rep. 2013, 3, 3448. [59] T. T. Tung, T. Y. Kim, K. S. Suh, Org. Electro 2011, 12, 22. [60] T. T. Tung, T. Y. Kim, J. P. Shim, W. S. Yang, H. K. Kim, K. S. Suh, Org Electro 2011, 12, 2215. [61] Z. Antar, J. F. Feller, H. Noël, P. Glouannec, K. Elleuch, Mater. Lett. 2012, 67, 210.
© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Received: December 13, 2014 Revised: February 23, 2015 Published online:
www.small-journal.com
9
