Home
Search
Collections
Journals
About
Contact us
My IOPscience
Metallic nanoparticles functionalizing carbon nanotube networks for gas sensing applications
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 Nanotechnology 25 055208 (http://iopscience.iop.org/0957-4484/25/5/055208) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 129.174.21.5 This content was downloaded on 17/06/2014 at 12:41
Please note that terms and conditions apply.
Nanotechnology Nanotechnology 25 (2014) 055208 (10pp)
doi:10.1088/0957-4484/25/5/055208
Metallic nanoparticles functionalizing carbon nanotube networks for gas sensing applications Ahmed Abdelhalim, Alaa Abdellah, Giuseppe Scarpa and Paolo Lugli Institute for Nanoelectronics, Technische Universit¨at M¨unchen, Arcisstraße 21, D-80333 Munich, Germany E-mail: [email protected] Received 13 November 2013 Accepted for publication 10 December 2013 Published 9 January 2014 Abstract
We report the fabrication of carbon nanotube (CNT) based gas sensors functionalized with different metallic nanoparticles (NPs) (Au, Pd, Ag) with exceptionally high responses towards four test gases (NH3 , CO2 , CO and ethanol). The CNT networks were fabricated through a low cost spray deposition process while the NPs were deposited by a thermal evaporation process. CNT based gas sensors functionalized with Au with a nominal thickness of 1.0 nm showed superior response towards NH3 , CO and ethanol. The sensors’ normalized responses reached 92%, 22% and 32% with concentrations of 100 ppm, 50 ppm and 100 ppm for NH3 , CO and ethanol respectively. CNT based gas sensors functionalized with Pd with a nominal thickness of 1.5 nm showed the best performance with CO2 . The normalized response reached 3%, 6%, 12% and 17% with concentrations of 500 ppm, 1000 ppm, 2500 ppm and 5000 ppm of CO2 respectively. We also investigated the morphological and optical changes that occur to the NPs upon thermal treatment. Functionalization of CNT films deposited on glass with Au and Ag showed surface plasmon resonance effects that are dependent on the nominal thickness of the functionalization layer. Keywords: carbon nanotubes, nanoparticles, spray deposition S Online supplementary data available from stacks.iop.org/Nano/25/055208/mmedia (Some figures may appear in colour only in the online journal)
1. Introduction
sensing applications. CNTs generally show p-type properties under ambient conditions [11], so by adsorbing oxidizing molecules (which withdraw electrons from the CNT film) the conduction of the film increases. On the other hand, the adsorption of reducing molecules, resulting in electron donation to the CNT film, decreases the conduction of the film on increasing the number of adsorbed molecules [12]. Despite their high sensitivity with respect to several gases, pristine CNT films lack the capability to differentiate between different gas molecules. The lack of selectivity of CNT based sensors can be partially overcome by metal functionalization of the CNT surface [13–15] or by polymer coatings [16]. Several methods have been used in order to deposit metal nanoparticles (NPs) on top of the CNT network,
Random carbon nanotube (CNT) networks have attracted a broad community of researchers in the last decade due to their impressive structural, mechanical and electronic properties [1–3]. The remarkable properties of CNTs are exploited in a wide range of applications in science and engineering [1, 2]. The most common applications of CNT films involve transparent conductive electrodes [3–5], thin film transistors and circuits [6, 7], and mechanical and chemical sensors [8–10]. CNT films display inherently reduced dimensionality and, hence, high surface-to-volume ratio, which, together with outstanding gas adsorption capabilities, renders CNTs ideal candidates for environmental 0957-4484/14/055208+10$33.00
1
c 2014 IOP Publishing Ltd Printed in the UK
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
such as thermal evaporation [17, 18], sputtering [19] and electrochemical deposition [20]. Penza et al employed a sputtering technique to deposit Pt-nanoclusters at tuned loadings of 8, 15 and 30 nm [19]. The sensors were exposed to different gases such as NO2 , NH3 , CO2 and CH4 . CNT sensors with a Pt load of 8 nm showed the highest normalized response in the case of NO2 and CH4 , while CNT sensors with a Pt load of 15 nm showed the best normalized response in the case of CO2 and NH3 . The main drawback is that the operating temperature of the sensors was 120 ◦ C. The same group also reported CNT sensors functionalized with Au nanoclusters [10]. The highest responses obtained were 10%, 16% and 14% for 10 ppm concentration of NO2 , 1000 ppm concentration of NH3 and 10 ppm concentration of H2 S respectively. These results were obtained at an operating temperature of 200 ◦ C. Zanolli et al fabricated CNT based sensors functionalized with Au NPs and oxygen plasma and exposed them to different pollutants such as NO2 , CO and C6 H6 [21]. They showed that Au functionalization results in better sensing abilities towards the detection of specific gases (NO2 and CO), while in the case of C6 H6 Au NPs do not seem to play a crucial role in the sensing process when compared to pristine CNTs functionalized with oxygen plasma. Although the formation of metal NPs on top of the CNT film improves the response of the sensor towards different gases greatly, the interaction between the metal nanoparticles and the CNT film is still not fully understood. Kim et al were among the first to demonstrate the formation of a nano-Schottky (also known as non-traditional Schottky) barrier at the interface between the deposited Al nanoparticle and the CNT film [22], which in turn modulates the molecular adsorption. Moreover, Kauffman et al reported that upon electrochemical growth of Au NPs on top of CNT film, an electron transfer occurs from the CNT to the NP species. The adsorption of CO molecules on the NP surface is accompanied by transfer of electronic density back into the CNTs [8]. They also reported that the magnitude of electron transfer into the CNT valence band scaled with the work function of the functionalization metal [24]. Mubeen et al also reported the formation of a nano-Schottky barrier at the Au NP–CNT interface which is modulated upon the exposure of H2 S [20]. Furthermore, they stated that the sensitivity of the sensor is strongly dependent on the number and the size of the Au NPs but the sensing mechanism is independent of it. Another series of studies concentrated on physical changes induced on a given surface by metal NPs. Serrano et al discussed, for instance, the surface plasmon resonance (SPR) in Au films deposited on glass substrates and annealed in air at different temperatures [25]. They showed that the resonant absorption of extended surface plasmons of the deposited Au film depends on the thickness of the film. Moreover, the morphology of the film and its optical properties are dependent on the film’s initial thickness which can be altered according to the annealing temperature. Sun et al reached the same conclusion by investigating the temperature dependence of morphological evolution and the corresponding SPR property variation of Au films with
nominal thicknesses of 5 nm upon rapid thermal annealing for 180 s at different temperatures ranging from 100 to 700 ◦ C [26]. Gingery et al reported the evaporation of Au NPs on top of multiwall carbon nanotubes (MWCNTs) which were grown by chemical vapor deposition [18]. They stated that the size and the distribution of the NPs are affected only by the nominal thickness deposited, with no influence of the evaporation rate or the substrate temperature during the evaporation process. In this paper, we report functionalized CNT gas sensors with exceptionally high as well as immediate response towards various test gases at room temperature. The sensors are functionalized with Au, Pd and Ag at different load levels. The response of the different sensors to gases such as NH3 , CO, CO2 and ethanol is examined. Testing towards such gases enables the fabricated sensors to cover different areas of application such as environmental monitoring and toxic gas detection, and in the food packaging industry as well. It is shown that metal functionalization of the CNT layer can alter the performance of the sensor towards different test gases, and hence can enable differentiation between different gases. Au functionalization results in exceptional response to NH3 , CO and ethanol, while Pd functionalization improves the sensor’s response to CO2 exposure. The sensor performance depends not only on the functionalizing material but also on the thermal treatment following the functionalization process. Finally, the effect of the thermal treatment on the samples functionalized with Au and Ag in terms of surface morphology and the change occurring in the frequency of the induced surface plasmons (SPs) on the CNT surface is demonstrated. 2. Materials and methods
CNT solutions are sprayed through an air gun onto a heated substrate. The temperature of the substrate is a process parameter that depends on the solvent and the surfactant used to disperse the CNTs. Besides influencing the surface topography, the spraying parameters have a direct impact on the optical and electrical characteristics of the fabricated films. The layers deposited in this work were sprayed by an air atomizing nozzle. These kinds of nozzles, as well as ultrasonic spray nozzles, provide a fine degree of atomization [5] which in turn results in a more uniform and homogeneous film. A commercially available automatic air atomizing spray gun (Krautzberger GmbH, Germany) was used in depositing the CNT films. Further details on the spray setup and the different spray parameters can be found in [27–29]. 2.1. Solution preparation
In order to dissolve CNTs in an aqueous solution, a high molecular weight cellulose derivative, sodium carboxymethyl cellulose (CMC), is used. This kind of surfactant has been reported previously as an excellent agent for dispersing CNTs in water [30]. Further discussion about the effects of different kinds of surfactants on the prepared CNT solution 2
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
can be found in [31]. CMC is dissolved in distilled water to make 0.5 wt% aqueous solution. The solution is stirred overnight in order to uniformly dissolve the surfactant in water. After this, 5 mg of CNTs are added to 10 g of 0.5 wt% CMC–aqueous solution to make 0.05 wt% CMC based aqueous CNT solution. Sonication of the CMC based solution is performed for 20 min using a probe sonicator at 50% power (∼48 W). The solution is centrifuged for 90 min at 15 000 rpm and 80% of the centrifuged solution is taken from top to be used for the deposition. 2.2. Sensor fabrication
Si wafers with 200 nm of thermally grown SiO2 are used as substrates. An interdigitated electrode structure (IDES) consisting of a 5 nm thick Cr layer, which promotes Au adhesion on SiO2 , followed by a 40 nm Au layer is evaporated on top of the SiO2 . 100 µm is chosen to be the spacing formed between the two electrodes of the interdigitated structure. The substrates are exposed to oxygen plasma for further cleaning and then they are immersed in 1 wt% of APTES in isopropanol solution since APTES acts as an adhesion promoter. On top of the IDES a CNT film is spray deposited in order to form the resistive network for the gas sensing. Figure 1 shows a schematic of the sensor structure and an AFM image of the CNT layer deposited on top. Chemical post-deposition treatment is required to remove the CMC-matrix embedding the CNTs, thus converting the network behavior from insulating to conducting. To this purpose, the samples are soaked in 4 M HNO3 solution for >12 h at room temperature for complete removal of the surfactant, rinsed in DI water and then dried. Metal functionalization is performed by thermal evaporation from a tungsten boat. Au with equivalent nominal thicknesses of 1.0, 1.5 and 2.5 nm is evaporated over the entire structure coated by CNTs, while Pd is evaporated with nominal thicknesses of 0.5, 1.0 and 1.5 nm. Note that all metals are evaporated with ˚ s−1 . the minimum evaporation rate which is 0.1 A The complete sensor module is composed of the CNT sensor mounted on a carrier glass together with a Peltier heating element for temperature control and a Pt100 sensor for in situ temperature monitoring. The heating element is located beneath the glass carrier, while the Pt100 sensor is placed next to the CNT sensor on top of the glass carrier. This arrangement allows precise determination of the actual temperature reached by the CNT sensor. Sensor performance is investigated by exposure to different concentrations of the test gas (NH3 , CO2 , CO and ethanol). The desired concentration is achieved by dilution of the test gas with a carrier gas while maintaining the total gas flux constant. Pure nitrogen is used as the carrier gas for NH3 , CO and ethanol while air is used as the carrier gas for CO2 .
Figure 1. (a) Schematic of the sensor architecture showing the
IDES structure. (b) AFM image of a typical low density CNT layer deposited on top of the IDES structure.
stresses built at the interface between the substrate and the coating material [25]. The stress introduced during the evaporation of thin metal films has been studied for a long time [32–36]. Briefly, the stress introduced in the film is composed of thermal stress, which is due to the difference in the thermal expansion coefficients between the substrate and the evaporated metal, and intrinsic stress, which is due to the accumulated flaws which are built on the surface during the deposition process. Which effect plays the dominant role depends on the physical properties of the material and the evaporation conditions. In general, stress between the substrate material and the coating material results in stress distribution at the interface which promotes migration of the film away from the stressed areas upon heating. The relaxation of the film is a non-uniform process and leads to inhomogeneous nucleation on the surface. This results in the formation of metal islands which tend to modify their shapes upon annealing, mostly becoming more rounded in order to decrease their surface energy. Figure 2 shows SEM images of CNT sensing layers functionalized with Au with nominal thicknesses of 0.5 nm (2(a)), 1.0 nm (2(b)), 1.5 nm (2(c)) and 2.5 nm (2(d)). The images show that the formation of the Au NPs is uniform on the substrate. The size of the deposited NPs is dependent on the load. Moreover, the coverage of the NPs on the tubes is dependent on the nominal thickness of the deposited layer. Thermal treatment is an important process step to stabilize and de-dope the CNT network before gas exposure [10]. On the other hand, performing thermal
3. Results and discussion 3.1. Characterization of the functionalized CNT films
The morphology of the metal NPs on the substrates depends on the initial thickness of the film and the integrated 3
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
Figure 2. SEM images of CNT films functionalized with different loads of Au: (a) 0.5 nm, (b) 1.0 nm, (c) 1.5 nm and (d) 2.5 nm.
treatment on functionalized CNT sensors directly affects the NPs evaporated on top of the CNT films, thus affecting the sensor performance. In order to investigate the effect of the thermal treatment process on the functionalized CNT network, CNT thin films utilized in sensor fabrication were spray deposited on glass substrate with the same process steps as described earlier and subsequently functionalized via thermal evaporation. The nominal thicknesses of Au deposition were 0.5 nm, 1.0 nm, 1.5 nm and 2.5 nm, respectively. The samples were thermally treated in air at 100 ◦ C for 1 h. Figures 3(a) and (b) show the morphological changes that occur to the Au NPs at different loads after thermal treatment. Au NPs tend to form isolated islands of large particles instead of relatively small particles distributed on the whole surface after the thermal treatment. By looking at the AFM images we can identify three parameters for the metal islands: (i) height, (ii) size and (iii) relative distance between them. Before annealing only the CNT film is dominating the features of the image since the thickness of the CNT bundle diameter is higher than the nominal thickness of the evaporated Au. After annealing the CNT film almost disappears underneath the formed metal islands. The maximum height of the formed metal islands becomes 128 nm, 133 nm, 101 nm and 87 nm for samples that have nominal thicknesses of 0.5, 1, 1.5, 2.5 nm, respectively. Moreover, the size of the islands and the distance between them decrease with increasing nominal thickness. The migration of the film from the stressed area to more relaxed areas is easier for a thinner metal layer because a
smaller amount of material needs to be displaced to create the metal island. The optical transmittance measurement shows the existence of SPs in the fabricated samples within the visible range. Au samples evaporated on top of the CNT films showed a drop in the intensity of the optical transmission, as shown in figure 4. For Au with a nominal thickness of 2.5 nm, the drop in intensity occurs at around 590 nm. A blue shift of ∼50 nm can be observed upon annealing the samples for 1 h at 100 ◦ C. Qualitatively the same behavior is observed for the other samples. The variation in the optical transmission is dependent on the size of the metal islands which are formed after the thermal treatment and their distribution density as well. This agrees with the AFM analysis which showed the migration of the films from the stressed areas and the formation of metal islands upon thermal treatment. Using the same procedure, Ag has been evaporated on top of CNT films sprayed on glass with nominal thicknesses of 1.0 nm, 1.5 nm and 2.5 nm, respectively. The Ag samples also showed SPR but at a different wavelength from that obtained for Au. The results obtained are shown in figure S1 (supporting information, available at stacks.iop.org/Nano/25/ 055208/mmedia). The effect of the thermal treatment on the Au nanoparticles was also analyzed using a different process. Instead of annealing the samples for 1 h continuously, we divided the annealing process into intervals of 10 min. This process is repeated for six cycles and after each cycle the surfaces changes are monitored via AFM and optical 4
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
Figure 3. AFM images of CNT films deposited on glass and functionalized with different loads of Au (0.5, 1.0. 1.5 and 2.5 nm). (a) As
fabricated. (b) Annealed continuously for 1 h at 100 ◦ C. (c) Annealed for 1 h in intervals of 10 min.
2.5 nm are evaporated on glass substrates can be found in figure S3 (supporting information, available at stacks.iop.org/ Nano/25/055208/mmedia).
transmission measurements. Figure 3(c) shows that after the stepwise annealing no dramatic change occurred to the surface. The morphological features as well as the optical transmission remained basically unchanged (figure 1(a)). The shift in the SPR curve is considerably smaller than that observed when the annealing is carried out continuously for 1 h. A small blue shift of ∼35 nm occurs. It is worth mentioning that this shift occurs after the first 10 min of annealing, while after that the optical transmission curve remains almost the same. The change in the optical transmission curve measured after each cycle can be found in figure S2 (supporting information, available at stacks.iop. org/Nano/25/055208/mmedia). As we will see later, the insensitivity of the functionalized APTES treatment has been shown to improve the adhesion between the glass substrate and the metal layer [1] by strengthening the film-to-substrate bond. The stronger the film-to-substrate bond, the stronger its capability to withstand the forces produced by the integrated stress throughout the film. Thus, when Au NPs evaporated on APTES-treated glass are subjected to the same thermal treatment as when they are evaporated on a CNT film, no significant change occurs to the surface morphology. The different behavior with respect to the Au deposition on the CNT film is most likely attributable to the hydrophobic nature of the CNT film [37, 38], which causes a weak bond to be formed with the Au evaporated film and in turn a higher stress at the interface. Optical transmission measurements for samples where Au NPs with a load of
3.2. Sensor response
When using CNT films for gas sensing, the CNT network density is one of the critical parameters affecting the device performance [39]. The gas sensors presented in this paper utilize CNT films with rather low tube densities. At such densities, the film operates near its percolation threshold, thereby reducing the number of metallic pathways in the network and amplifying the contribution of semiconducting pathways. Metal functionalization of the CNT network causes a slight decrease in the film resistance, which is in turn dependent on the load of the functionalization metal. Increasing the metal functionalization will tend to create a complete film on top of the CNT network, which will quench the sensitivity of the overall system. The change in resistance of CNT gas sensors upon exposure to different gas species has been studied by several authors [29, 39, 40]. It has been attributed to two main reasons: (i) charge transfer between the adsorbed gas molecule and the CNT, and/or (ii) modulation induced by the gas species at the Schottky barrier at the CNT/metal contact (Au electrode) interface. It has been previously shown that metal functionalization (especially Au) enhances the response of CNT gas sensors greatly. This effect was attributed to the catalytic action of the 5
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
Figure 4. Optical transmission measurements for CNT networks functionalized with Au with loads of 0.5 nm (a), 1.0 nm (b), 1.5 nm (c)
and 2.5 nm (d). Each graph shows the measurement for the as fabricated samples (red), annealed continuously for 1 h at 100 ◦ C (blue) and annealed for 1 h at 100 ◦ C in intervals of 10 min (black).
deposited layer which boosts the gas sensitivity of the CNT film [10]. It is well known that the adhesion between the Au particles and the CNT film is weak due to the hydrophobic nature of CNTs [11]. In our case, the post-deposition process in diluted HNO3 can introduce defect sites and surface oxidization to the CNT film [41], thus allowing Au to bind to the CNT film via the carboxylic groups introduced on the surface. The deposition of Au NPs on top of the CNT film causes the formation of a ‘non-traditional Schottky barrier’ at the interface between the semiconducting tubes and the metal NPs [23, 24]. One reason for the enhancement in the sensor response after the Au decoration could be the increase in the sites of CNT/metal Schottky barriers which can be further modulated upon gas exposure. In the case of NH3 , it coordinates with the formed Au NPs [42], hence charge transfer can take place through the formed nanoparticles towards the CNT film. Possibly, the Au NPs help in binding more than one ammonia molecule to the same defect site instead of binding only one in the case of pristine CNT film. The amount of charge transfer decreases with increasing size of the Au NPs because the catalytic activity of Au NPs diminishes considerably when the particle size exceeds a certain limit [43]. Upon annealing, Au NPs tend to migrate and form new nucleation sites in order to reduce the surface energy, as we
discussed earlier. This leads to two main effects: (i) increase in the film conductivity because the formed Au island acts like a parallel metallic path to the CNT film, (ii) partial uncovering of the CNT film which can interact directly with the exposed gas species. Which of the two effects is dominant depends mainly on the nominal thickness of the evaporated Au NPs. Concentrations between 10 and 100 ppm for NH3 and ethanol and between 5 and 50 ppm for CO have been considered. These concentrations are achieved by diluting the test gas with pure N2 as carrier gas. For CO2 , the concentrations tested cover a range between 500 and 5000 ppm and are achieved by diluting the test gas with normal air as carrier gas. Note that in the case of CO2 measurement, the normal air used as carrier gas typically contains 390 ppm of CO2 , which is the reason for the high concentration range covered in our tests. Each cycle consists of an exposure interval followed by a recovery interval. During exposure, the test gas enters the chamber along with the carrier gas at room temperature at a constant flux of 200 ml min−1 for the desired duration. Recovery is then introduced by heating the sensor module to 80 ◦ C and increasing the carrier gas flux to 1000 ml min−1 for 300 s, after which heating is stopped. The high flux is maintained for another 300 s to accelerate cooling of the sensor and purge any residual test gas molecules out of the chamber. Recovery is then completed by a final 300 s interval under sensing conditions to restore the initial resistance. The increase in 6
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
Figure 5. The change in resistance of CNT based sensors
Figure 6. The relation between the concentration and the
functionalized with 1.0 nm of Au when exposed to different concentrations of NH3 before (a) and after annealing (b).
normalized response of CNT based sensors functionalized with Au with loads of 1.0 nm (a) and 2.5 nm (b) when exposed to NH3 . Each graph shows the response before (blue) and after annealing (red).
thermal energy during the heating cycle is necessary to enhance desorption of gas molecules attached to the CNT film and enable full recovery [39]. As mentioned above, heating the sensor for small intervals does not affect the morphology of the metal NPs formed on top of the CNT film. Figure 5 shows the effect of the thermal treatment on the resistance of a CNT sensor functionalized with 1.0 nm nominal thickness of load. Upon thermal treatment, a noticeable increase in the conductance is obtained. A larger effect is obtained for CNT sensors functionalized with 2.5 nm Au nominal thickness (figure S4, supporting information, available at stacks.iop.org/Nano/25/055208/mmedia). Figure 6 shows the effect of the thermal treatment on the response to NH3 of two CNT sensors functionalized with Au of 1.0 nm and 2.5 nm thickness respectively. A considerable enhancement in the sensor response is obtained after the thermal treatment in the case of 1.0 nm Au coverage while the sensor response remains almost unchanged in the case of 2.5 nm of Au. This could be attributed to the different size of the Au nanoparticles obtained after annealing. Further investigations aimed at clarification of this point are currently being performed. The main figure of merit used to evaluate the sensor performance is the normalized response, defined in equation (1) as the relative change in resistance, where Ri and Rf are the initial and final resistance values of the exposure cycle, respectively, Norm. Response =
Rf − Ri ∗ 100 (%). Ri
Figure 7 shows the response to four different test gases obtained for CNT gas sensors functionalized with Au and Pd with different nominal thicknesses. Taking a close look at the graphs, one can immediately notice that Au (1.0 nm) shows superior performance for NH3 , CO and ethanol. On the other hand, Pd (1.0 nm) has the best response to CO2 . CNT sensors functionalized with Au (1.0 nm) reach normalized responses of 92%, 22% and 30% when exposed to 100 ppm of NH3 and ethanol and 50 ppm of CO, respectively. CNT sensors functionalized with Pd (1.0 nm) reach a normalized response of 17% when exposed to 5000 ppm of CO2 , and the normalized response is 2.1% with CO2 concentration as low as 500 ppm. Sensor response is found to have a rather logarithmic dependence on concentration. As expected, higher functionalization load does not result in higher sensor response, since the contribution of semiconducting CNT paths tend to decrease as a complete metallic film is formed. Each metal can achieve a maximum normalized response within a certain range of surface coverage. The optimal range for Au functionalization is between 0.5 and 2.5 nm, while for Pd it is between 0.5 and 1.5 nm. Although Au and Pd have quite similar work functions, which was reported previously to be a dominant effect in the formation of the potential barrier at the interface of the NP–CNT and also the response of the gas sensors [24], they behave differently when they are used for CNT sensor
(1) 7
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
Figure 7. The relation between the concentration and the normalized response of functionalized CNT based sensors when exposed to
NH3 (a), CO (b), CO2 (c) and ethanol (d). Each graph plots the responses of the CNT based sensors when functionalized with Au with loads 1.0, 1.5 and 2.5 nm, and Pd with loads of 0.5, 1.0 and 1.5 nm.
functionalization. This could be attributed to the difference in stress distribution during the evaporation of each metal, as discussed earlier. An AFM image showing the difference in the nanoparticle formation between Au and Pd evaporated on glass with the same load can be found in figure S5 (supporting information, available at stacks.iop.org/Nano/25/ 055208/mmedia). CNT sensors functionalized with Ag with nominal thicknesses of 1.0, 1.5 and 2.5 nm were also exposed to the four test gases. The results can be found in figure S6 (supporting information, available at stacks.iop.org/Nano/25/ 055208/mmedia). The performance of the gas sensors was also investigated under heating conditions. Figure 8 shows the normalized response of a bare CNT sensor and functionalized CNT sensor with 1.0 nm of Au. The exposure of the NH3 gas was performed at two temperatures, 50 ◦ C (figure 8(a)) and 80 ◦ C (figure 8(b)). Taking a close look at the graph, it becomes clear that the performance of the sensor deteriorates when compared to the results obtained in figure 7(a). This could be due to the increase in the gas desorption rate which increases with increase in temperature [39, 44].
high as well as immediate sensor response to four test gases (NH3 , CO2 , CO and ethanol). CNT films were prepared using a reliable and reproducible low cost spray process technique while the metal NPs were deposited by thermal evaporation. The CNT based gas sensor functionalized with 1.0 nm of Au reached values of 25% for a concentration as low as 10 ppm and 92% for a concentration of 100 ppm under NH3 exposure. Taking into consideration the short exposure time, which was 100 s, the values obtained for NH3 can be considered exceptional. We also analyzed the effect of the thermal treatment on the performance of the Au functionalized CNT based gas sensors when exposed to NH3 . In the case of Au with a nominal thickness of 1.0 nm the sensor response can be as high as 120% with 100 ppm concentration of NH3 after thermal treatment. In contrast, for CNT based gas sensors functionalized with 2.5 nm of Au the response is not significantly affected by the thermal treatment. Moreover, we investigated the changes that occur to the morphology of the NPs upon annealing. Functionalization of CNT films deposited on glass with Au and Ag showed surface plasmon resonance effects that are dependent on the nominal thickness of the functionalization layer. Morphological modifications that occur upon annealing lead to changes in the optical properties of the films. The functionalized CNT based gas sensors presented in this work could be further developed and optimized in order to be
4. Conclusion
We reported on the successful implementation of CNT based gas sensors functionalized with Au and Pd with exceptionally 8
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
[4] Chien Y-M, Lefevre F, Shih I and Izquierdo R 2010 A solution processed top emission OLED with transparent carbon nanotube electrodes Nanotechnology 21 134020 [5] Tenent R C, Barnes T M, Bergeson J D, Ferguson A J, To B, Gedvilas L M, Heben M J and Blackburn J L 2009 Ultrasmooth, large-area, high-uniformity, conductive transparent single-walled-carbon-nanotube films for photovoltaics produced by ultrasonic spraying Adv. Mater. 21 3210–6 [6] Ha M, Xia Y, Green A A, Zhang W, Renn M J, Kim C H, Hersam M C and Frisbie C D 2010 Printed, sub-3 V digital circuits on plastic from aqueous carbon nanotube inks ACS Nano 4 4388–95 [7] Wang C, Zhang J and Zhou C 2010 Macroelectronic integrated circuits using high-performance separated carbon nanotube thin-film transistors ACS Nano 4 7123–32 [8] Li J, Lu Y, Ye Q, Cinke M, Han J and Meyyappan M 2003 Carbon nanotube sensors for gas and organic vapor detection Nano Lett. 3 929–33 [9] Bekyarova E, Davis M, Burch T, Itkis M E, Zhao B, Sunshine S and Haddon R C 2004 Chemically functionalized single-walled carbon nanotubes as ammonia sensors † J. Phys. Chem. B 108 19717–20 [10] Penza M, Rossi R, Alvisi M, Cassano G and Serra E 2009 Functional characterization of carbon nanotube networked films functionalized with tuned loading of Au nanoclusters for gas sensing applications Sensors Actuators B 140 176–84 [11] Martel R, Schmidt T, Shea H R, Hertel T and Avouris P 1998 Single- and multi-wall carbon nanotube field-effect transistors Appl. Phys. Lett. 73 2447 [12] Kong J 2000 Nanotube molecular wires as chemical sensors Science 287 622–5 [13] Penza M, Cassano G, Rossi R, Alvisi M, Rizzo A, Signore M A, Dikonimos T, Serra E and Giorgi R 2007 Enhancement of sensitivity in gas chemiresistors based on carbon nanotube surface functionalized with noble metal (Au, Pt) nanoclusters Appl. Phys. Lett. 90 173123 [14] Kong J, Chapline M G and Dai H 2001 Functionalized carbon nanotubes for molecular hydrogen sensors Adv. Mater. 13 1384–6 [15] Zhao Q, Buongiorno Nardelli M, Lu W and Bernholc J 2005 Carbon nanotube–metal cluster composites: a new road to chemical sensors? Nano Lett. 5 847–51 [16] An K H, Jeong S Y, Hwang H R and Lee Y H 2004 Enhanced sensitivity of a gas sensor incorporating single-walled carbon nanotube–polypyrrole nanocomposites Adv. Mater. 16 1005–9 [17] Scarselli M, Camilli L, Castrucci P, Nanni F, Del Gobbo S, Gautron E, Lefrant S and De Crescenzi M 2012 In situ formation of noble metal nanoparticles on multiwalled carbon nanotubes and its implication in metal–nanotube interactions Carbon 50 875–84 [18] Gingery D and B¨uhlmann P 2008 Formation of gold nanoparticles on multiwalled carbon nanotubes by thermal evaporation Carbon 46 1966–72 [19] Penza M, Rossi R, Alvisi M, Suriano D and Serra E 2011 Pt-modified carbon nanotube networked layers for enhanced gas microsensors Thin Solid Films 520 959–65 [20] Mubeen S, Lim J-H, Srirangarajan A, Mulchandani A, Deshusses M A and Myung N V 2011 Gas sensing mechanism of gold nanoparticles decorated single-walled carbon nanotubes Electroanalysis 23 2687–92 [21] Zanolli Z, Leghrib R, Felten A, Pireaux J-J, Llobet E and Charlier J-C 2011 Gas sensing with Au-decorated carbon nanotubes ACS Nano 5 4592–9 [22] Kim B-K et al 2006 The effect of metal cluster coatings on carbon nanotubes Nanotechnology 17 496–500
Figure 8. The relation between the concentration and the
normalized response of CNT based sensors at temperatures of T = 50 ◦ C (a) and T = 80 ◦ C (b) when exposed to NH3 . Each graph shows the response for a bare CNT layer (blue) and a CNT layer functionalized with 1.0 nm of Au (red).
integrated in a sensor array that is highly selective towards different gas molecules. Acknowledgments
The authors would like to acknowledge Professor Jonathn Veinot from the Department of Chemistry, University of Alberta for fruitful discussions concerning the results presented in this work. This work was partially supported by the Graduate Centre of the faculty of Electrical and Information Technology (FGC-EI) at the Technische Universit¨at M¨unchen. References [1] Hu L, Hecht D S and Gruener G 2010 Carbon nanotube thin films fabrication, properties, and applications Chem. Rev. 110 5790–844 [2] Cao Q and Rogers J A 2009 Ultrathin films of single-walled carbon nanotubes for electronics and sensors: a review of fundamental and applied aspects Adv. Mater. 21 29–53 [3] Zhang D, Ryu K, Liu X, Polikarpov E, Ly J, Tompson M E and Zhou C 2006 Transparent, conductive, and flexible carbon nanotube films and their application in organic light-emitting diodes Nano Lett. 6 1880–6 9
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
[33] Presland A E B, Price G L and Trimm D L 1972 Hillock formation by surface films Surf. Sci. 29 424–34 [34] Rahman Khan M S 1987 Hillock and island formation during annealing of gold films Bull. Mater. Sci. 9 55–60 [35] Thornton J A 1989 Stress-related effects in thin films Thin Solid Films 171 5–31 [36] Hwang S-J, Lee J-H, Jeong C-O and Joo Y-C 2007 Effect of film thickness and annealing temperature on Hillock distributions in pure Al films Scr. Mater. 56 17–20 [37] Song Y I, Kim G Y, Choi H K, Jeong H J, Kim K K, Yang C-M, Lim S C, An K H, Jung K T and Lee Y H 2006 Fabrication of carbon nanotube field emitters using a dip-coating method Chem. Vap. Depos. 12 375–9 [38] Su H-C, Chen C-H, Chen Y-C, Yao D-J, Chen H, Chang Y-C and Yew T-R 2010 Improving the adhesion of carbon nanotubes to a substrate using microwave treatment Carbon 48 805–12 [39] Abdellah A, Abdelhalim A, Horn M, Scarpa G and Lugli P 2013 Scalable spray deposition process for high-performance carbon nanotube gas sensors IEEE Trans. Nanotechnol. 12 174–81 [40] Peng N, Zhang Q, Chow C L, Tan O K and Marzari N 2009 Sensing mechanisms for carbon nanotube based NH3 gas detection Nano Lett. 9 1626–30 [41] Lordi V, Yao N and Wei J 2001 Method for supporting platinum on single-walled carbon nanotubes for a selective hydrogenation catalyst Chem. Mater. 13 733–7 [42] Bili´c A, Reimers J R, Hush N S and Hafner J 2002 Adsorption of ammonia on the gold (111) surface J. Chem. Phys. 116 8981 [43] Corma A and Garcia H 2008 Supported gold nanoparticles as catalysts for organic reactions Chem. Soc. Rev. 37 2096–126 [44] Abdellah A, Abdelhalim A, Loghin F, K¨ohler P, Ahmad Z, Scarpa G and Lugli P 2013 Flexible carbon nanotube based gas sensors fabricated by large-scale spray deposition IEEE Sensors J. 13 4014–21
[23] Kauffman D R, Sorescu D C, Schofield D P, Allen B L, Jordan K D and Star A 2010 Understanding the sensor response of metal-decorated carbon nanotubes Nano Lett. 10 958–63 [24] Kauffman D R and Star A 2007 Chemically induced potential barriers at the carbon nanotube–metal nanoparticle interface Nano Lett. 7 1863–8 [25] Serrano A, Rodrıguez de la Fuente O and Garcıa M A 2010 Extended and localized surface plasmons in annealed Au films on glass substrates J. Appl. Phys. 108 074303 [26] Sun H, Yu M, Wang G, Sun X and Lian J 2012 Temperature-dependent morphology evolution and surface plasmon absorption of ultrathin gold island films J. Phys. Chem. C 116 9000–8 [27] Abdellah A, Fabel B, Lugli P and Scarpa G 2010 Spray deposition of organic semiconducting thin-films: towards the fabrication of arbitrary shaped organic electronic devices Org. Electron. 11 1031–8 [28] Abdellah A, Virdi K S, Meier R, D¨oblinger M, M¨uller-Buschbaum P, Scheu C, Lugli P and Scarpa G 2012 Successive spray deposition of P3HT/PCBM organic photoactive layers: material composition and device characteristics Adv. Funct. Mater. 22 4078–86 [29] Abdellah A, Member S, Yaqub A, Ferrari C, Fabel B, Lugli P and Scarpa G 2011 Spray deposition of highly uniform CNT films and their application in gas sensing Proc. IEEE 11th Conf. Nanotechnology pp 1118–23 [30] Takahashi T, Tsunoda K, Yajima H and Ishii T 2004 Dispersion and purification of single-wall carbon nanotubes using carboxymethylcellulose Japan. J. Appl. Phys. 43 3636–9 [31] Abdelhalim A, Abdellah A, Scarpa G and Lugli P 2013 Fabrication of carbon nanotube thin films on flexible substrates by spray deposition and transfer printing Carbon 61 72–9 [32] Sharma S K and Spitz J 1980 Hillock formation, hole growth and agglomeration in thin silver films Thin Solid Films 65 339–50
10
Search
Collections
Journals
About
Contact us
My IOPscience
Metallic nanoparticles functionalizing carbon nanotube networks for gas sensing applications
This content has been downloaded from IOPscience. Please scroll down to see the full text. 2014 Nanotechnology 25 055208 (http://iopscience.iop.org/0957-4484/25/5/055208) View the table of contents for this issue, or go to the journal homepage for more
Download details: IP Address: 129.174.21.5 This content was downloaded on 17/06/2014 at 12:41
Please note that terms and conditions apply.
Nanotechnology Nanotechnology 25 (2014) 055208 (10pp)
doi:10.1088/0957-4484/25/5/055208
Metallic nanoparticles functionalizing carbon nanotube networks for gas sensing applications Ahmed Abdelhalim, Alaa Abdellah, Giuseppe Scarpa and Paolo Lugli Institute for Nanoelectronics, Technische Universit¨at M¨unchen, Arcisstraße 21, D-80333 Munich, Germany E-mail: [email protected] Received 13 November 2013 Accepted for publication 10 December 2013 Published 9 January 2014 Abstract
We report the fabrication of carbon nanotube (CNT) based gas sensors functionalized with different metallic nanoparticles (NPs) (Au, Pd, Ag) with exceptionally high responses towards four test gases (NH3 , CO2 , CO and ethanol). The CNT networks were fabricated through a low cost spray deposition process while the NPs were deposited by a thermal evaporation process. CNT based gas sensors functionalized with Au with a nominal thickness of 1.0 nm showed superior response towards NH3 , CO and ethanol. The sensors’ normalized responses reached 92%, 22% and 32% with concentrations of 100 ppm, 50 ppm and 100 ppm for NH3 , CO and ethanol respectively. CNT based gas sensors functionalized with Pd with a nominal thickness of 1.5 nm showed the best performance with CO2 . The normalized response reached 3%, 6%, 12% and 17% with concentrations of 500 ppm, 1000 ppm, 2500 ppm and 5000 ppm of CO2 respectively. We also investigated the morphological and optical changes that occur to the NPs upon thermal treatment. Functionalization of CNT films deposited on glass with Au and Ag showed surface plasmon resonance effects that are dependent on the nominal thickness of the functionalization layer. Keywords: carbon nanotubes, nanoparticles, spray deposition S Online supplementary data available from stacks.iop.org/Nano/25/055208/mmedia (Some figures may appear in colour only in the online journal)
1. Introduction
sensing applications. CNTs generally show p-type properties under ambient conditions [11], so by adsorbing oxidizing molecules (which withdraw electrons from the CNT film) the conduction of the film increases. On the other hand, the adsorption of reducing molecules, resulting in electron donation to the CNT film, decreases the conduction of the film on increasing the number of adsorbed molecules [12]. Despite their high sensitivity with respect to several gases, pristine CNT films lack the capability to differentiate between different gas molecules. The lack of selectivity of CNT based sensors can be partially overcome by metal functionalization of the CNT surface [13–15] or by polymer coatings [16]. Several methods have been used in order to deposit metal nanoparticles (NPs) on top of the CNT network,
Random carbon nanotube (CNT) networks have attracted a broad community of researchers in the last decade due to their impressive structural, mechanical and electronic properties [1–3]. The remarkable properties of CNTs are exploited in a wide range of applications in science and engineering [1, 2]. The most common applications of CNT films involve transparent conductive electrodes [3–5], thin film transistors and circuits [6, 7], and mechanical and chemical sensors [8–10]. CNT films display inherently reduced dimensionality and, hence, high surface-to-volume ratio, which, together with outstanding gas adsorption capabilities, renders CNTs ideal candidates for environmental 0957-4484/14/055208+10$33.00
1
c 2014 IOP Publishing Ltd Printed in the UK
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
such as thermal evaporation [17, 18], sputtering [19] and electrochemical deposition [20]. Penza et al employed a sputtering technique to deposit Pt-nanoclusters at tuned loadings of 8, 15 and 30 nm [19]. The sensors were exposed to different gases such as NO2 , NH3 , CO2 and CH4 . CNT sensors with a Pt load of 8 nm showed the highest normalized response in the case of NO2 and CH4 , while CNT sensors with a Pt load of 15 nm showed the best normalized response in the case of CO2 and NH3 . The main drawback is that the operating temperature of the sensors was 120 ◦ C. The same group also reported CNT sensors functionalized with Au nanoclusters [10]. The highest responses obtained were 10%, 16% and 14% for 10 ppm concentration of NO2 , 1000 ppm concentration of NH3 and 10 ppm concentration of H2 S respectively. These results were obtained at an operating temperature of 200 ◦ C. Zanolli et al fabricated CNT based sensors functionalized with Au NPs and oxygen plasma and exposed them to different pollutants such as NO2 , CO and C6 H6 [21]. They showed that Au functionalization results in better sensing abilities towards the detection of specific gases (NO2 and CO), while in the case of C6 H6 Au NPs do not seem to play a crucial role in the sensing process when compared to pristine CNTs functionalized with oxygen plasma. Although the formation of metal NPs on top of the CNT film improves the response of the sensor towards different gases greatly, the interaction between the metal nanoparticles and the CNT film is still not fully understood. Kim et al were among the first to demonstrate the formation of a nano-Schottky (also known as non-traditional Schottky) barrier at the interface between the deposited Al nanoparticle and the CNT film [22], which in turn modulates the molecular adsorption. Moreover, Kauffman et al reported that upon electrochemical growth of Au NPs on top of CNT film, an electron transfer occurs from the CNT to the NP species. The adsorption of CO molecules on the NP surface is accompanied by transfer of electronic density back into the CNTs [8]. They also reported that the magnitude of electron transfer into the CNT valence band scaled with the work function of the functionalization metal [24]. Mubeen et al also reported the formation of a nano-Schottky barrier at the Au NP–CNT interface which is modulated upon the exposure of H2 S [20]. Furthermore, they stated that the sensitivity of the sensor is strongly dependent on the number and the size of the Au NPs but the sensing mechanism is independent of it. Another series of studies concentrated on physical changes induced on a given surface by metal NPs. Serrano et al discussed, for instance, the surface plasmon resonance (SPR) in Au films deposited on glass substrates and annealed in air at different temperatures [25]. They showed that the resonant absorption of extended surface plasmons of the deposited Au film depends on the thickness of the film. Moreover, the morphology of the film and its optical properties are dependent on the film’s initial thickness which can be altered according to the annealing temperature. Sun et al reached the same conclusion by investigating the temperature dependence of morphological evolution and the corresponding SPR property variation of Au films with
nominal thicknesses of 5 nm upon rapid thermal annealing for 180 s at different temperatures ranging from 100 to 700 ◦ C [26]. Gingery et al reported the evaporation of Au NPs on top of multiwall carbon nanotubes (MWCNTs) which were grown by chemical vapor deposition [18]. They stated that the size and the distribution of the NPs are affected only by the nominal thickness deposited, with no influence of the evaporation rate or the substrate temperature during the evaporation process. In this paper, we report functionalized CNT gas sensors with exceptionally high as well as immediate response towards various test gases at room temperature. The sensors are functionalized with Au, Pd and Ag at different load levels. The response of the different sensors to gases such as NH3 , CO, CO2 and ethanol is examined. Testing towards such gases enables the fabricated sensors to cover different areas of application such as environmental monitoring and toxic gas detection, and in the food packaging industry as well. It is shown that metal functionalization of the CNT layer can alter the performance of the sensor towards different test gases, and hence can enable differentiation between different gases. Au functionalization results in exceptional response to NH3 , CO and ethanol, while Pd functionalization improves the sensor’s response to CO2 exposure. The sensor performance depends not only on the functionalizing material but also on the thermal treatment following the functionalization process. Finally, the effect of the thermal treatment on the samples functionalized with Au and Ag in terms of surface morphology and the change occurring in the frequency of the induced surface plasmons (SPs) on the CNT surface is demonstrated. 2. Materials and methods
CNT solutions are sprayed through an air gun onto a heated substrate. The temperature of the substrate is a process parameter that depends on the solvent and the surfactant used to disperse the CNTs. Besides influencing the surface topography, the spraying parameters have a direct impact on the optical and electrical characteristics of the fabricated films. The layers deposited in this work were sprayed by an air atomizing nozzle. These kinds of nozzles, as well as ultrasonic spray nozzles, provide a fine degree of atomization [5] which in turn results in a more uniform and homogeneous film. A commercially available automatic air atomizing spray gun (Krautzberger GmbH, Germany) was used in depositing the CNT films. Further details on the spray setup and the different spray parameters can be found in [27–29]. 2.1. Solution preparation
In order to dissolve CNTs in an aqueous solution, a high molecular weight cellulose derivative, sodium carboxymethyl cellulose (CMC), is used. This kind of surfactant has been reported previously as an excellent agent for dispersing CNTs in water [30]. Further discussion about the effects of different kinds of surfactants on the prepared CNT solution 2
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
can be found in [31]. CMC is dissolved in distilled water to make 0.5 wt% aqueous solution. The solution is stirred overnight in order to uniformly dissolve the surfactant in water. After this, 5 mg of CNTs are added to 10 g of 0.5 wt% CMC–aqueous solution to make 0.05 wt% CMC based aqueous CNT solution. Sonication of the CMC based solution is performed for 20 min using a probe sonicator at 50% power (∼48 W). The solution is centrifuged for 90 min at 15 000 rpm and 80% of the centrifuged solution is taken from top to be used for the deposition. 2.2. Sensor fabrication
Si wafers with 200 nm of thermally grown SiO2 are used as substrates. An interdigitated electrode structure (IDES) consisting of a 5 nm thick Cr layer, which promotes Au adhesion on SiO2 , followed by a 40 nm Au layer is evaporated on top of the SiO2 . 100 µm is chosen to be the spacing formed between the two electrodes of the interdigitated structure. The substrates are exposed to oxygen plasma for further cleaning and then they are immersed in 1 wt% of APTES in isopropanol solution since APTES acts as an adhesion promoter. On top of the IDES a CNT film is spray deposited in order to form the resistive network for the gas sensing. Figure 1 shows a schematic of the sensor structure and an AFM image of the CNT layer deposited on top. Chemical post-deposition treatment is required to remove the CMC-matrix embedding the CNTs, thus converting the network behavior from insulating to conducting. To this purpose, the samples are soaked in 4 M HNO3 solution for >12 h at room temperature for complete removal of the surfactant, rinsed in DI water and then dried. Metal functionalization is performed by thermal evaporation from a tungsten boat. Au with equivalent nominal thicknesses of 1.0, 1.5 and 2.5 nm is evaporated over the entire structure coated by CNTs, while Pd is evaporated with nominal thicknesses of 0.5, 1.0 and 1.5 nm. Note that all metals are evaporated with ˚ s−1 . the minimum evaporation rate which is 0.1 A The complete sensor module is composed of the CNT sensor mounted on a carrier glass together with a Peltier heating element for temperature control and a Pt100 sensor for in situ temperature monitoring. The heating element is located beneath the glass carrier, while the Pt100 sensor is placed next to the CNT sensor on top of the glass carrier. This arrangement allows precise determination of the actual temperature reached by the CNT sensor. Sensor performance is investigated by exposure to different concentrations of the test gas (NH3 , CO2 , CO and ethanol). The desired concentration is achieved by dilution of the test gas with a carrier gas while maintaining the total gas flux constant. Pure nitrogen is used as the carrier gas for NH3 , CO and ethanol while air is used as the carrier gas for CO2 .
Figure 1. (a) Schematic of the sensor architecture showing the
IDES structure. (b) AFM image of a typical low density CNT layer deposited on top of the IDES structure.
stresses built at the interface between the substrate and the coating material [25]. The stress introduced during the evaporation of thin metal films has been studied for a long time [32–36]. Briefly, the stress introduced in the film is composed of thermal stress, which is due to the difference in the thermal expansion coefficients between the substrate and the evaporated metal, and intrinsic stress, which is due to the accumulated flaws which are built on the surface during the deposition process. Which effect plays the dominant role depends on the physical properties of the material and the evaporation conditions. In general, stress between the substrate material and the coating material results in stress distribution at the interface which promotes migration of the film away from the stressed areas upon heating. The relaxation of the film is a non-uniform process and leads to inhomogeneous nucleation on the surface. This results in the formation of metal islands which tend to modify their shapes upon annealing, mostly becoming more rounded in order to decrease their surface energy. Figure 2 shows SEM images of CNT sensing layers functionalized with Au with nominal thicknesses of 0.5 nm (2(a)), 1.0 nm (2(b)), 1.5 nm (2(c)) and 2.5 nm (2(d)). The images show that the formation of the Au NPs is uniform on the substrate. The size of the deposited NPs is dependent on the load. Moreover, the coverage of the NPs on the tubes is dependent on the nominal thickness of the deposited layer. Thermal treatment is an important process step to stabilize and de-dope the CNT network before gas exposure [10]. On the other hand, performing thermal
3. Results and discussion 3.1. Characterization of the functionalized CNT films
The morphology of the metal NPs on the substrates depends on the initial thickness of the film and the integrated 3
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
Figure 2. SEM images of CNT films functionalized with different loads of Au: (a) 0.5 nm, (b) 1.0 nm, (c) 1.5 nm and (d) 2.5 nm.
treatment on functionalized CNT sensors directly affects the NPs evaporated on top of the CNT films, thus affecting the sensor performance. In order to investigate the effect of the thermal treatment process on the functionalized CNT network, CNT thin films utilized in sensor fabrication were spray deposited on glass substrate with the same process steps as described earlier and subsequently functionalized via thermal evaporation. The nominal thicknesses of Au deposition were 0.5 nm, 1.0 nm, 1.5 nm and 2.5 nm, respectively. The samples were thermally treated in air at 100 ◦ C for 1 h. Figures 3(a) and (b) show the morphological changes that occur to the Au NPs at different loads after thermal treatment. Au NPs tend to form isolated islands of large particles instead of relatively small particles distributed on the whole surface after the thermal treatment. By looking at the AFM images we can identify three parameters for the metal islands: (i) height, (ii) size and (iii) relative distance between them. Before annealing only the CNT film is dominating the features of the image since the thickness of the CNT bundle diameter is higher than the nominal thickness of the evaporated Au. After annealing the CNT film almost disappears underneath the formed metal islands. The maximum height of the formed metal islands becomes 128 nm, 133 nm, 101 nm and 87 nm for samples that have nominal thicknesses of 0.5, 1, 1.5, 2.5 nm, respectively. Moreover, the size of the islands and the distance between them decrease with increasing nominal thickness. The migration of the film from the stressed area to more relaxed areas is easier for a thinner metal layer because a
smaller amount of material needs to be displaced to create the metal island. The optical transmittance measurement shows the existence of SPs in the fabricated samples within the visible range. Au samples evaporated on top of the CNT films showed a drop in the intensity of the optical transmission, as shown in figure 4. For Au with a nominal thickness of 2.5 nm, the drop in intensity occurs at around 590 nm. A blue shift of ∼50 nm can be observed upon annealing the samples for 1 h at 100 ◦ C. Qualitatively the same behavior is observed for the other samples. The variation in the optical transmission is dependent on the size of the metal islands which are formed after the thermal treatment and their distribution density as well. This agrees with the AFM analysis which showed the migration of the films from the stressed areas and the formation of metal islands upon thermal treatment. Using the same procedure, Ag has been evaporated on top of CNT films sprayed on glass with nominal thicknesses of 1.0 nm, 1.5 nm and 2.5 nm, respectively. The Ag samples also showed SPR but at a different wavelength from that obtained for Au. The results obtained are shown in figure S1 (supporting information, available at stacks.iop.org/Nano/25/ 055208/mmedia). The effect of the thermal treatment on the Au nanoparticles was also analyzed using a different process. Instead of annealing the samples for 1 h continuously, we divided the annealing process into intervals of 10 min. This process is repeated for six cycles and after each cycle the surfaces changes are monitored via AFM and optical 4
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
Figure 3. AFM images of CNT films deposited on glass and functionalized with different loads of Au (0.5, 1.0. 1.5 and 2.5 nm). (a) As
fabricated. (b) Annealed continuously for 1 h at 100 ◦ C. (c) Annealed for 1 h in intervals of 10 min.
2.5 nm are evaporated on glass substrates can be found in figure S3 (supporting information, available at stacks.iop.org/ Nano/25/055208/mmedia).
transmission measurements. Figure 3(c) shows that after the stepwise annealing no dramatic change occurred to the surface. The morphological features as well as the optical transmission remained basically unchanged (figure 1(a)). The shift in the SPR curve is considerably smaller than that observed when the annealing is carried out continuously for 1 h. A small blue shift of ∼35 nm occurs. It is worth mentioning that this shift occurs after the first 10 min of annealing, while after that the optical transmission curve remains almost the same. The change in the optical transmission curve measured after each cycle can be found in figure S2 (supporting information, available at stacks.iop. org/Nano/25/055208/mmedia). As we will see later, the insensitivity of the functionalized APTES treatment has been shown to improve the adhesion between the glass substrate and the metal layer [1] by strengthening the film-to-substrate bond. The stronger the film-to-substrate bond, the stronger its capability to withstand the forces produced by the integrated stress throughout the film. Thus, when Au NPs evaporated on APTES-treated glass are subjected to the same thermal treatment as when they are evaporated on a CNT film, no significant change occurs to the surface morphology. The different behavior with respect to the Au deposition on the CNT film is most likely attributable to the hydrophobic nature of the CNT film [37, 38], which causes a weak bond to be formed with the Au evaporated film and in turn a higher stress at the interface. Optical transmission measurements for samples where Au NPs with a load of
3.2. Sensor response
When using CNT films for gas sensing, the CNT network density is one of the critical parameters affecting the device performance [39]. The gas sensors presented in this paper utilize CNT films with rather low tube densities. At such densities, the film operates near its percolation threshold, thereby reducing the number of metallic pathways in the network and amplifying the contribution of semiconducting pathways. Metal functionalization of the CNT network causes a slight decrease in the film resistance, which is in turn dependent on the load of the functionalization metal. Increasing the metal functionalization will tend to create a complete film on top of the CNT network, which will quench the sensitivity of the overall system. The change in resistance of CNT gas sensors upon exposure to different gas species has been studied by several authors [29, 39, 40]. It has been attributed to two main reasons: (i) charge transfer between the adsorbed gas molecule and the CNT, and/or (ii) modulation induced by the gas species at the Schottky barrier at the CNT/metal contact (Au electrode) interface. It has been previously shown that metal functionalization (especially Au) enhances the response of CNT gas sensors greatly. This effect was attributed to the catalytic action of the 5
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
Figure 4. Optical transmission measurements for CNT networks functionalized with Au with loads of 0.5 nm (a), 1.0 nm (b), 1.5 nm (c)
and 2.5 nm (d). Each graph shows the measurement for the as fabricated samples (red), annealed continuously for 1 h at 100 ◦ C (blue) and annealed for 1 h at 100 ◦ C in intervals of 10 min (black).
deposited layer which boosts the gas sensitivity of the CNT film [10]. It is well known that the adhesion between the Au particles and the CNT film is weak due to the hydrophobic nature of CNTs [11]. In our case, the post-deposition process in diluted HNO3 can introduce defect sites and surface oxidization to the CNT film [41], thus allowing Au to bind to the CNT film via the carboxylic groups introduced on the surface. The deposition of Au NPs on top of the CNT film causes the formation of a ‘non-traditional Schottky barrier’ at the interface between the semiconducting tubes and the metal NPs [23, 24]. One reason for the enhancement in the sensor response after the Au decoration could be the increase in the sites of CNT/metal Schottky barriers which can be further modulated upon gas exposure. In the case of NH3 , it coordinates with the formed Au NPs [42], hence charge transfer can take place through the formed nanoparticles towards the CNT film. Possibly, the Au NPs help in binding more than one ammonia molecule to the same defect site instead of binding only one in the case of pristine CNT film. The amount of charge transfer decreases with increasing size of the Au NPs because the catalytic activity of Au NPs diminishes considerably when the particle size exceeds a certain limit [43]. Upon annealing, Au NPs tend to migrate and form new nucleation sites in order to reduce the surface energy, as we
discussed earlier. This leads to two main effects: (i) increase in the film conductivity because the formed Au island acts like a parallel metallic path to the CNT film, (ii) partial uncovering of the CNT film which can interact directly with the exposed gas species. Which of the two effects is dominant depends mainly on the nominal thickness of the evaporated Au NPs. Concentrations between 10 and 100 ppm for NH3 and ethanol and between 5 and 50 ppm for CO have been considered. These concentrations are achieved by diluting the test gas with pure N2 as carrier gas. For CO2 , the concentrations tested cover a range between 500 and 5000 ppm and are achieved by diluting the test gas with normal air as carrier gas. Note that in the case of CO2 measurement, the normal air used as carrier gas typically contains 390 ppm of CO2 , which is the reason for the high concentration range covered in our tests. Each cycle consists of an exposure interval followed by a recovery interval. During exposure, the test gas enters the chamber along with the carrier gas at room temperature at a constant flux of 200 ml min−1 for the desired duration. Recovery is then introduced by heating the sensor module to 80 ◦ C and increasing the carrier gas flux to 1000 ml min−1 for 300 s, after which heating is stopped. The high flux is maintained for another 300 s to accelerate cooling of the sensor and purge any residual test gas molecules out of the chamber. Recovery is then completed by a final 300 s interval under sensing conditions to restore the initial resistance. The increase in 6
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
Figure 5. The change in resistance of CNT based sensors
Figure 6. The relation between the concentration and the
functionalized with 1.0 nm of Au when exposed to different concentrations of NH3 before (a) and after annealing (b).
normalized response of CNT based sensors functionalized with Au with loads of 1.0 nm (a) and 2.5 nm (b) when exposed to NH3 . Each graph shows the response before (blue) and after annealing (red).
thermal energy during the heating cycle is necessary to enhance desorption of gas molecules attached to the CNT film and enable full recovery [39]. As mentioned above, heating the sensor for small intervals does not affect the morphology of the metal NPs formed on top of the CNT film. Figure 5 shows the effect of the thermal treatment on the resistance of a CNT sensor functionalized with 1.0 nm nominal thickness of load. Upon thermal treatment, a noticeable increase in the conductance is obtained. A larger effect is obtained for CNT sensors functionalized with 2.5 nm Au nominal thickness (figure S4, supporting information, available at stacks.iop.org/Nano/25/055208/mmedia). Figure 6 shows the effect of the thermal treatment on the response to NH3 of two CNT sensors functionalized with Au of 1.0 nm and 2.5 nm thickness respectively. A considerable enhancement in the sensor response is obtained after the thermal treatment in the case of 1.0 nm Au coverage while the sensor response remains almost unchanged in the case of 2.5 nm of Au. This could be attributed to the different size of the Au nanoparticles obtained after annealing. Further investigations aimed at clarification of this point are currently being performed. The main figure of merit used to evaluate the sensor performance is the normalized response, defined in equation (1) as the relative change in resistance, where Ri and Rf are the initial and final resistance values of the exposure cycle, respectively, Norm. Response =
Rf − Ri ∗ 100 (%). Ri
Figure 7 shows the response to four different test gases obtained for CNT gas sensors functionalized with Au and Pd with different nominal thicknesses. Taking a close look at the graphs, one can immediately notice that Au (1.0 nm) shows superior performance for NH3 , CO and ethanol. On the other hand, Pd (1.0 nm) has the best response to CO2 . CNT sensors functionalized with Au (1.0 nm) reach normalized responses of 92%, 22% and 30% when exposed to 100 ppm of NH3 and ethanol and 50 ppm of CO, respectively. CNT sensors functionalized with Pd (1.0 nm) reach a normalized response of 17% when exposed to 5000 ppm of CO2 , and the normalized response is 2.1% with CO2 concentration as low as 500 ppm. Sensor response is found to have a rather logarithmic dependence on concentration. As expected, higher functionalization load does not result in higher sensor response, since the contribution of semiconducting CNT paths tend to decrease as a complete metallic film is formed. Each metal can achieve a maximum normalized response within a certain range of surface coverage. The optimal range for Au functionalization is between 0.5 and 2.5 nm, while for Pd it is between 0.5 and 1.5 nm. Although Au and Pd have quite similar work functions, which was reported previously to be a dominant effect in the formation of the potential barrier at the interface of the NP–CNT and also the response of the gas sensors [24], they behave differently when they are used for CNT sensor
(1) 7
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
Figure 7. The relation between the concentration and the normalized response of functionalized CNT based sensors when exposed to
NH3 (a), CO (b), CO2 (c) and ethanol (d). Each graph plots the responses of the CNT based sensors when functionalized with Au with loads 1.0, 1.5 and 2.5 nm, and Pd with loads of 0.5, 1.0 and 1.5 nm.
functionalization. This could be attributed to the difference in stress distribution during the evaporation of each metal, as discussed earlier. An AFM image showing the difference in the nanoparticle formation between Au and Pd evaporated on glass with the same load can be found in figure S5 (supporting information, available at stacks.iop.org/Nano/25/ 055208/mmedia). CNT sensors functionalized with Ag with nominal thicknesses of 1.0, 1.5 and 2.5 nm were also exposed to the four test gases. The results can be found in figure S6 (supporting information, available at stacks.iop.org/Nano/25/ 055208/mmedia). The performance of the gas sensors was also investigated under heating conditions. Figure 8 shows the normalized response of a bare CNT sensor and functionalized CNT sensor with 1.0 nm of Au. The exposure of the NH3 gas was performed at two temperatures, 50 ◦ C (figure 8(a)) and 80 ◦ C (figure 8(b)). Taking a close look at the graph, it becomes clear that the performance of the sensor deteriorates when compared to the results obtained in figure 7(a). This could be due to the increase in the gas desorption rate which increases with increase in temperature [39, 44].
high as well as immediate sensor response to four test gases (NH3 , CO2 , CO and ethanol). CNT films were prepared using a reliable and reproducible low cost spray process technique while the metal NPs were deposited by thermal evaporation. The CNT based gas sensor functionalized with 1.0 nm of Au reached values of 25% for a concentration as low as 10 ppm and 92% for a concentration of 100 ppm under NH3 exposure. Taking into consideration the short exposure time, which was 100 s, the values obtained for NH3 can be considered exceptional. We also analyzed the effect of the thermal treatment on the performance of the Au functionalized CNT based gas sensors when exposed to NH3 . In the case of Au with a nominal thickness of 1.0 nm the sensor response can be as high as 120% with 100 ppm concentration of NH3 after thermal treatment. In contrast, for CNT based gas sensors functionalized with 2.5 nm of Au the response is not significantly affected by the thermal treatment. Moreover, we investigated the changes that occur to the morphology of the NPs upon annealing. Functionalization of CNT films deposited on glass with Au and Ag showed surface plasmon resonance effects that are dependent on the nominal thickness of the functionalization layer. Morphological modifications that occur upon annealing lead to changes in the optical properties of the films. The functionalized CNT based gas sensors presented in this work could be further developed and optimized in order to be
4. Conclusion
We reported on the successful implementation of CNT based gas sensors functionalized with Au and Pd with exceptionally 8
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
[4] Chien Y-M, Lefevre F, Shih I and Izquierdo R 2010 A solution processed top emission OLED with transparent carbon nanotube electrodes Nanotechnology 21 134020 [5] Tenent R C, Barnes T M, Bergeson J D, Ferguson A J, To B, Gedvilas L M, Heben M J and Blackburn J L 2009 Ultrasmooth, large-area, high-uniformity, conductive transparent single-walled-carbon-nanotube films for photovoltaics produced by ultrasonic spraying Adv. Mater. 21 3210–6 [6] Ha M, Xia Y, Green A A, Zhang W, Renn M J, Kim C H, Hersam M C and Frisbie C D 2010 Printed, sub-3 V digital circuits on plastic from aqueous carbon nanotube inks ACS Nano 4 4388–95 [7] Wang C, Zhang J and Zhou C 2010 Macroelectronic integrated circuits using high-performance separated carbon nanotube thin-film transistors ACS Nano 4 7123–32 [8] Li J, Lu Y, Ye Q, Cinke M, Han J and Meyyappan M 2003 Carbon nanotube sensors for gas and organic vapor detection Nano Lett. 3 929–33 [9] Bekyarova E, Davis M, Burch T, Itkis M E, Zhao B, Sunshine S and Haddon R C 2004 Chemically functionalized single-walled carbon nanotubes as ammonia sensors † J. Phys. Chem. B 108 19717–20 [10] Penza M, Rossi R, Alvisi M, Cassano G and Serra E 2009 Functional characterization of carbon nanotube networked films functionalized with tuned loading of Au nanoclusters for gas sensing applications Sensors Actuators B 140 176–84 [11] Martel R, Schmidt T, Shea H R, Hertel T and Avouris P 1998 Single- and multi-wall carbon nanotube field-effect transistors Appl. Phys. Lett. 73 2447 [12] Kong J 2000 Nanotube molecular wires as chemical sensors Science 287 622–5 [13] Penza M, Cassano G, Rossi R, Alvisi M, Rizzo A, Signore M A, Dikonimos T, Serra E and Giorgi R 2007 Enhancement of sensitivity in gas chemiresistors based on carbon nanotube surface functionalized with noble metal (Au, Pt) nanoclusters Appl. Phys. Lett. 90 173123 [14] Kong J, Chapline M G and Dai H 2001 Functionalized carbon nanotubes for molecular hydrogen sensors Adv. Mater. 13 1384–6 [15] Zhao Q, Buongiorno Nardelli M, Lu W and Bernholc J 2005 Carbon nanotube–metal cluster composites: a new road to chemical sensors? Nano Lett. 5 847–51 [16] An K H, Jeong S Y, Hwang H R and Lee Y H 2004 Enhanced sensitivity of a gas sensor incorporating single-walled carbon nanotube–polypyrrole nanocomposites Adv. Mater. 16 1005–9 [17] Scarselli M, Camilli L, Castrucci P, Nanni F, Del Gobbo S, Gautron E, Lefrant S and De Crescenzi M 2012 In situ formation of noble metal nanoparticles on multiwalled carbon nanotubes and its implication in metal–nanotube interactions Carbon 50 875–84 [18] Gingery D and B¨uhlmann P 2008 Formation of gold nanoparticles on multiwalled carbon nanotubes by thermal evaporation Carbon 46 1966–72 [19] Penza M, Rossi R, Alvisi M, Suriano D and Serra E 2011 Pt-modified carbon nanotube networked layers for enhanced gas microsensors Thin Solid Films 520 959–65 [20] Mubeen S, Lim J-H, Srirangarajan A, Mulchandani A, Deshusses M A and Myung N V 2011 Gas sensing mechanism of gold nanoparticles decorated single-walled carbon nanotubes Electroanalysis 23 2687–92 [21] Zanolli Z, Leghrib R, Felten A, Pireaux J-J, Llobet E and Charlier J-C 2011 Gas sensing with Au-decorated carbon nanotubes ACS Nano 5 4592–9 [22] Kim B-K et al 2006 The effect of metal cluster coatings on carbon nanotubes Nanotechnology 17 496–500
Figure 8. The relation between the concentration and the
normalized response of CNT based sensors at temperatures of T = 50 ◦ C (a) and T = 80 ◦ C (b) when exposed to NH3 . Each graph shows the response for a bare CNT layer (blue) and a CNT layer functionalized with 1.0 nm of Au (red).
integrated in a sensor array that is highly selective towards different gas molecules. Acknowledgments
The authors would like to acknowledge Professor Jonathn Veinot from the Department of Chemistry, University of Alberta for fruitful discussions concerning the results presented in this work. This work was partially supported by the Graduate Centre of the faculty of Electrical and Information Technology (FGC-EI) at the Technische Universit¨at M¨unchen. References [1] Hu L, Hecht D S and Gruener G 2010 Carbon nanotube thin films fabrication, properties, and applications Chem. Rev. 110 5790–844 [2] Cao Q and Rogers J A 2009 Ultrathin films of single-walled carbon nanotubes for electronics and sensors: a review of fundamental and applied aspects Adv. Mater. 21 29–53 [3] Zhang D, Ryu K, Liu X, Polikarpov E, Ly J, Tompson M E and Zhou C 2006 Transparent, conductive, and flexible carbon nanotube films and their application in organic light-emitting diodes Nano Lett. 6 1880–6 9
Nanotechnology 25 (2014) 055208
A Abdelhalim et al
[33] Presland A E B, Price G L and Trimm D L 1972 Hillock formation by surface films Surf. Sci. 29 424–34 [34] Rahman Khan M S 1987 Hillock and island formation during annealing of gold films Bull. Mater. Sci. 9 55–60 [35] Thornton J A 1989 Stress-related effects in thin films Thin Solid Films 171 5–31 [36] Hwang S-J, Lee J-H, Jeong C-O and Joo Y-C 2007 Effect of film thickness and annealing temperature on Hillock distributions in pure Al films Scr. Mater. 56 17–20 [37] Song Y I, Kim G Y, Choi H K, Jeong H J, Kim K K, Yang C-M, Lim S C, An K H, Jung K T and Lee Y H 2006 Fabrication of carbon nanotube field emitters using a dip-coating method Chem. Vap. Depos. 12 375–9 [38] Su H-C, Chen C-H, Chen Y-C, Yao D-J, Chen H, Chang Y-C and Yew T-R 2010 Improving the adhesion of carbon nanotubes to a substrate using microwave treatment Carbon 48 805–12 [39] Abdellah A, Abdelhalim A, Horn M, Scarpa G and Lugli P 2013 Scalable spray deposition process for high-performance carbon nanotube gas sensors IEEE Trans. Nanotechnol. 12 174–81 [40] Peng N, Zhang Q, Chow C L, Tan O K and Marzari N 2009 Sensing mechanisms for carbon nanotube based NH3 gas detection Nano Lett. 9 1626–30 [41] Lordi V, Yao N and Wei J 2001 Method for supporting platinum on single-walled carbon nanotubes for a selective hydrogenation catalyst Chem. Mater. 13 733–7 [42] Bili´c A, Reimers J R, Hush N S and Hafner J 2002 Adsorption of ammonia on the gold (111) surface J. Chem. Phys. 116 8981 [43] Corma A and Garcia H 2008 Supported gold nanoparticles as catalysts for organic reactions Chem. Soc. Rev. 37 2096–126 [44] Abdellah A, Abdelhalim A, Loghin F, K¨ohler P, Ahmad Z, Scarpa G and Lugli P 2013 Flexible carbon nanotube based gas sensors fabricated by large-scale spray deposition IEEE Sensors J. 13 4014–21
[23] Kauffman D R, Sorescu D C, Schofield D P, Allen B L, Jordan K D and Star A 2010 Understanding the sensor response of metal-decorated carbon nanotubes Nano Lett. 10 958–63 [24] Kauffman D R and Star A 2007 Chemically induced potential barriers at the carbon nanotube–metal nanoparticle interface Nano Lett. 7 1863–8 [25] Serrano A, Rodrıguez de la Fuente O and Garcıa M A 2010 Extended and localized surface plasmons in annealed Au films on glass substrates J. Appl. Phys. 108 074303 [26] Sun H, Yu M, Wang G, Sun X and Lian J 2012 Temperature-dependent morphology evolution and surface plasmon absorption of ultrathin gold island films J. Phys. Chem. C 116 9000–8 [27] Abdellah A, Fabel B, Lugli P and Scarpa G 2010 Spray deposition of organic semiconducting thin-films: towards the fabrication of arbitrary shaped organic electronic devices Org. Electron. 11 1031–8 [28] Abdellah A, Virdi K S, Meier R, D¨oblinger M, M¨uller-Buschbaum P, Scheu C, Lugli P and Scarpa G 2012 Successive spray deposition of P3HT/PCBM organic photoactive layers: material composition and device characteristics Adv. Funct. Mater. 22 4078–86 [29] Abdellah A, Member S, Yaqub A, Ferrari C, Fabel B, Lugli P and Scarpa G 2011 Spray deposition of highly uniform CNT films and their application in gas sensing Proc. IEEE 11th Conf. Nanotechnology pp 1118–23 [30] Takahashi T, Tsunoda K, Yajima H and Ishii T 2004 Dispersion and purification of single-wall carbon nanotubes using carboxymethylcellulose Japan. J. Appl. Phys. 43 3636–9 [31] Abdelhalim A, Abdellah A, Scarpa G and Lugli P 2013 Fabrication of carbon nanotube thin films on flexible substrates by spray deposition and transfer printing Carbon 61 72–9 [32] Sharma S K and Spitz J 1980 Hillock formation, hole growth and agglomeration in thin silver films Thin Solid Films 65 339–50
10
