Accepted Manuscript Title: Development of highly sensitive optical sensor from carbon nanotube-alumina nanocomposite free-standing films: CNTs loading dependence sensor performance Analysis Authors: Abid, Poonam Sehrawat, S.S. Islam, Payal Gulati, Mohammad Talib, Prabhash Mishra, Manika Khanuja PII: DOI: Reference:
S0924-4247(17)31168-8 https://doi.org/10.1016/j.sna.2017.10.062 SNA 10428
To appear in:
Sensors and Actuators A
Received date: Revised date: Accepted date:
24-6-2017 6-9-2017 28-10-2017
Please cite this article as: Abid, Poonam Sehrawat, S.S.Islam, Payal Gulati, Mohammad Talib, Prabhash Mishra, Manika Khanuja, Development of highly sensitive optical sensor from carbon nanotube-alumina nanocomposite free-standing films: CNTs loading dependence sensor performance Analysis, Sensors and Actuators: A Physical https://doi.org/10.1016/j.sna.2017.10.062 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
SC RI PT
Development of highly sensitive optical sensor from carbon nanotube-alumina nanocomposite free-standing films: CNTs loading dependence sensor performance Analysis Abid, Poonam Sehrawat, S.S. Islam*, Payal Gulati, Mohammad Talib, Prabhash Mishra, Manika Khanuja
Abid1, Poonam Sehrawat1, S.S. Islam1*, Payal Gulati1, Mohammad Talib1, Prabhash Mishra1, Manika 1
U
Khanuja1
Centre for Nanoscience and Nanotechnology, Jamia Millia Islamia, (A Central University), New Delhi - 110025, India.
A
N
* Corresponding Author Email:
[email protected]
D
A promising facile and economic technique to develop MWCNTs/Alumina composite film. Optical sensors prepared from free-standing films for varying MWCNTs contents. FESEM analysis corroborates homogeneous dispersal of MWCNTs in host matrix. Fast response, high sensitivity andexcellent repeatability in Vis-NIR region. The technique holds potential for commercial scale production.
TE
M
Highlights
A
CC
EP
Abstract— We report a highly sensitive optical sensor based on free-standing thin films derived from multi-walled carbon nanotubes (MWCNTs)-Alumina nanocomposite by Gel-cast technique. The sensing principle involves the change in the resistance/conductance of the fabricated nanocomposite film on interaction with the optical stimulus. The performance of the sensor strongly depends – on loading and dispersion of MWCNTs in Alumina host matrix; wavelength; and power density of the laser beam. The optimized loading of CNTs to achieve maximum sensitivity was 1.5wt%. The sensitivity of the sensor shows linear relationship with power density of the laser beam and found to be highly sensitive in Vis-NIR region. The maximum sensitivity of the sensor is found to be 13.2% at 635nm wavelength, 3.5mW/mm2power density of laser beam and at 1.5wt% MWCNTs loading in Alumina host matrix. At this loading, the response time and recovery times of the sensor are found to be 1.7s and 2.1s respectively. The additional advantage of the present sensor is that it is facile and cost-effective method to fabricate high performance optical sensors.
Manuscript received………….. This work was partially supported by the Department of Science and Technology, India with grant number no. SR/S2/CMP99/2012 dated 12-12-2013. The authors are with Centre for Nanoscience and Nanotechnology, Jamia Millia Islamia, (A central University), New Delhi - 110025, India (Email:
[email protected]).
2
Index Terms- Carbon nanotubes, nanocomposite, free standing thin films, Optical sensor
SC RI PT
I. INTRODUCTION Optical sensors are attracting huge interest due to their wide-range of chemical, biochemical, medical and healthcare, and imaging applications. Recent advancements in the field of nanotechnology have created great potential towards developing low cost, highly sensitive, portable sensors with low power consumption. Nanomaterials, in particular, carbon nanotubes (CNTs) are considered as one of the most favorable materials due to its unique electronic properties and their extremely large surface area to volume ratio [1-5]. Researchers have tried much for the development of optical sensor, to harness the specific properties of CNTs [6-10].
A
CC
EP
TE
D
M
A
N
U
Various reports are available on photoconductive response of individual as well as microscopic films of CNTs [11-14]. For optical sensors, CNTs can be used for wide range of applications from UV-Vis to IR region due to its tunable band gap [15-19]. Photoconductive response of single-walled CNT sheets was reported by Shaoxin Lu and Balaji Pachapakesan. They have reported the photoconductive response of CNTs as a function of position, contact area, light intensity, ambient pressure, and frequency of laser pulse [20]. The results therein contributed significantly for new researchers in this field. Similarly, Yafai Zhang fabricated the multi-walled carbon nanotubes (MWCNTs) based IR detectors by deposition of MWCNTs on the pre-defined copper electrode on SiO2 substrate [21]. This sensor can work significantly only in IR region. At the same time, Antonio Serra et al. performed photoconductive measurements on large number of aligned single-walled carbon nanotubes (SWCNTs) bundles [22]. This alignment of SWCNTs opens a new method in the field of optical sensing. Avouris et al. observed polarization and light intensity dependent photoconductivity in single SWCNT [23]. Marian B. Tzolov et al. demonstrate CNTs-Si hetrojunction arrays and IR photocurrent response [24]. I.A. Levitsky et al. studied photoconductivity of SWCNTs under continuous wave NIR illumination [25]. Individual CNTs based IR transistors were reported by Hongzhi Chen et al. [26]. Ho-Yin Chan et al. have performed a very unique experiment and fabricated single carbon nanotube based transistors by AFM manipulation system and their photoconductive response was studied [27]. In majority of the aforementioned reported results, device design and its reproducibility make a big setback on the tall claim of CNTs for device commercialization. Sensor design by placing a single fiber between two electrodes is not only a difficult task but expensive too. It needs high-end nanoscale manipulation [28-30]. The other disadvantage of such sensors is that they possess less absorption area which ultimately limits the response of the optical sensor [31]. Hence to enhance the photoresponse of the CNT based sensors, we need to focus on developing a network /mesh of CNTs fibers, which provides sufficient absorption area to overcome the abovementioned impediments. Efforts were put to develop CNTs mesh/network structure by spray deposition, making composite with polymers, glues, and epoxies etc. In a nut shell, the achievements are not so outstanding and possess serious limitations [12, 13, 21, 31]. In this paper, to further improve the sensing performance of optical sensor, we report the fabrication and characterization of optical sensor based on MWCNTs/Alumina nanocomposite free standing thin film. The fabrication of the sensor is carried out using gel-cast technique, a well-known technique used for the fabrication of processed film in a form of mesh of well dispersed CNTs [32]. In the host matrix, CNTs loading is a critical issue and our results showed that the sensor performance in terms of quantification of several important sensor parameters is highly sensitive towards this factor. The organization of this paper is as follows: I. The structural morphology of the films was studied by variable loading (wt %) of CNTs embedded in alumina matrix. II. The sensors fabricated from the said films were characterized by collimating the fiber coupled laser beam of varying wavelength and power density on the top surface of individual sensor. III. A comparative study with suitable analysis is made on the impact of CNT loading (in the alumina composite) on sensor parameters. Maximum sensitivity was found to be 13.2% at a CNT loading level of 1.5 wt% in the composite for the 635nm wavelength laser at constant laser power of
3
3.5mW/mm2. Wavelength dependent studies show that the sensors are sensitive in the visible and NIR region. II.
EXPERIMENT
SC RI PT
A. Preparation of MWCNTs-Alumina nanocomposite Sol gel synthesis route was adopted to prepare Alumina sol, where 20.5g aluminum-sec-butoxide (Al(OCH3) C2H5)3 (Alfa Aesar, 97%) was mixed in 150 ml of DI water with continuous stirring and heating using hot plate magnetic stirrer. The temperature was maintained around 90°C for 1h. To avoid the particle aggregation, 0.5cc concentrated HNO3 (60-70%) was added to the stirring solutions. This process results in the hydrolysis of aluminum-sec-butoxide. Then the hydrolyzed solution was refluxed for 20h to obtain alumina sol. Subsequently, 2 wt% Poly-vinyl alcohol (PVA), 1.0% wt Poly (ethylene glycol) and 0.25% wt Benzyl butyl phthalate were blended, as binder and plasticizer, with the stirring solution. Now, by adding different concentrations of MWCNTs (procured from Chengdu Organic Chemicals Co., China; Diameter: < 8nm, Length: 10–30 μm and dispersed in 0.1% SDS solution in DI) viz. 0.6wt%, 1.0wt%, 1.5wt%, 2.0wt%, and 3.0wt% to the stirring solution, we get MWCNTs-Alumina nanocomposite in the form of a paste or slurry [33-35].
A
N
U
B. Development of MWCNTs/Alumina nanocomposite free standing thin films MWCNTs/Alumina nanocomposite films of varying concentration of MWCNTs in alumina matrix were prepared using Tape cast machine shown in Fig. 1. In this technique, the previously prepared MWCNTsAlumina nanocomposite consisting of a suspension of MWCNTs in Alumina sol was
D
M
Casted in a motor driven Mylar belt. When a constant relative movement is established between the blade and the belt, the slurry spreads on the belt to form thin sheet which on drying results in gel layer. Thickness of the film was precisely controlled by motor speed as well as Doctor’s blade integrated with the machine. The length of the film can be controlled by adjusting the volume of the solution used for casting.
A
CC
EP
TE
C. Fabrication of the sensors Samples of 10mm x 10mm dimensions were sliced from the nanocomposite film prepared in previous section. These samples were annealed at 900 °C in inert atmosphere to remove the volatile organic impurities from the films leading to the formation of MWCNTs conducting network/mat. Silver electrodes were deposited on both ends of these annealed samples by screen printing technique. Fig. 2 (a) shows the nanocomposite films before annealing and Fig. 2 (b) shows the fabricated sensor having silver electrodes. For the experimental purposes of this study, samples of five varying concentration of MWCNTs in the alumina matrix were prepared and examined for the concentration dependent photoresponse. Five different samples were fabricated for each concentration in order to confirm the feasibility to go for batch fabrication. The samples were coded as G1=0.6 wt%, G2=1.0 wt%, G3=1.5 wt%, G4=2.0 wt% and G5=3.0 wt %. Further increase in CNT loading results in their agglomeration which in turns results in the formation of ruptures and voids in the fabricated thin film. D. Experimental set up for sensor characterization The schematic of the experimental setup for the characterization of the fabricated sensor is shown in Fig. 3. On illuminating laser beam on the top surface of the sensor, the resistance decreased sharply due to the generation of excess free carriers, i.e. electrons and holes in their respective bands. The photoresponse was measured with Keithley 4200 SCS system. Three lasers (B&W Tek) of wavelength 635nm, 785nm, and 1064nm with tunable power were employed for photo-excitation of the sensors. The distance between laser source and sample was maintained at 10 mm with a spot size of ~6 mm. The laser power was recorded
4
SC RI PT
using a Newport 843 R power meter. Since, the contact resistance between the sensing element (free standing film) and electrodes was negligible in comparison to that of the fabricated film, therefore the major contribution to the electrical response is supposed to be due to MWCNTs-Alumina nanocomposite sensor element. The current versus voltage (I-V) characteristics shows the change in response on laser light excitation. Fig. 4 shows I-V characteristics of sensor G3 under dark as well as illumination with laser wavelength 635nm at power density 1.5mW/mm2. Ohmic behavior of metal/semiconductor (M/S) contact can be attributed to the quantum tunneling of the free carriers across the metal (electrode) and semiconductor (sensing film) interface.
A
N
U
III. RESULTS AND DISCUSSION To study the surface morphology of the MWCNTs/Alumina nanocomposite films, SEM images were taken for all the sensors and shown in Fig. 5 (a-e). From Fig. 5 (a-c), it is clearly visible from the SEM images that there is uniform dispersion of MWCNTs in the alumina matrix upto 1.5wt %, beyond which MWCNTs agglomerate into lumps/bundles as shown by red circles in Fig. 5 (d-e). The agglomeration is due to strong inter-tube van der Waals forces which prevent uniformity in dispersion of the MWCNTs in alumina. To confirm the multi-walled structure of the carbon nanotubes, TEM image of sensor G3 (=1.5 wt% MWCNTs loading) has been given in Fig. 5(f). In this image, the numbers of walls are clearly visible with an outer diameter of ~16 nm.
CC
EP
TE
D
M
Fig. 6 shows UV-Vis sepectra of MWCNTs/Alumina composite was recorded by Agilent carry 100 spectrometer. Spectra shows strong absorption at 250nm which is due to electron transition of π- π* in MWCNTs. The absorption coefficients and sensitivity are directly related. More the absorption of photons more the generation of free e-h pairs and more the sensitivity is- this is how the operational principle of optical sensor works. Here Fig. 6(a) shows the absorbance of all the sensors in which G3 sensor having MWCNT loading 1.5 wt%, shows the highest absorbance. In Fig. 6(b) we calculated the absorption coefficient for all the sensors at two different wavelengths, i.e., 250nm and 635nm. Although the absorption coefficient at 250nm is higher than at 635nm wavelength but we could not perform sensing studies at this wavelength as the UV source for this wavelength is not available in our laboratory. Therefore, we stuck on the sensing studies at 635nm wavelength. As observed in Fig. 6(b), the absorption coefficient is maximum for G3 sensor irrespective of the wavelength. This is in favor of our experimental result.
A
When light is illuminated on the sensor the following processes occur: (i) photons with the suitable energy get absorbed within the sample and generate excess free carriers; (ii) photo-generated carriers drift towards the electrodes on applying suitable bias across the sample; (iii) in this process, resistance of the sample decreases and hence conductivity increases. If R0 be the initial resistance and Rf be the final obtained resistance then ΔR= (|R0-Rf|) is the change in resistance. From the change in resistance, we can measure the sensitivity of the sensor which is defined as the ratio of change in resistance and the applied resistance. It is a dimension-less quantity and can be expressed in terms of percentage. Mathematically, the sensitivity of the sensor can be expressed as [36]: 𝑆𝑒𝑛𝑠𝑖𝑡𝑖𝑣𝑖𝑡𝑦 =
∆𝑅 𝑅0
× 100 (%)
(1)
5
TE
D
M
A
N
U
SC RI PT
To study the photo-induced response characteristics of the fabricated sensors (G1, G2, G3, G4 and G5), we have conducted systematic studies at different wavelengths and powers of the laser beam. In Fig. 7, the variation of sensitivity has been plotted as a function of power density of the laser beam for all the sensors. This variation of sensitivity has been plotted for three different wavelengths of the laser beam viz. 635nm, 785nm and 1064nm in Fig. 7(a), 7(b) and 7(c) respectively. It has been observed from the graphs that the sensitivity is highest for G3 sensor (corresponding to 1.5wt% of MWCNTs) irrespective of the wavelength as well as the power of the laser. In principle, on increasing the concentration of MWCNTs the sensitivity should also increase. The obtained results contradict because of the non-uniform dispersion of MWCNTs at higher loading of MWCNTs in the alumina matrix. At higher loading, MWCNTs tend to agglomerate and form bundles/lumps owing to strong Van der Waals attractive forces among them. Due to this, only the top surface of the sample interacts with light illumination and generates relatively less excess free carriers resulting in reduced photoresponse and hence, the reduced sensitivity. The variation of sensitivity of all the sensors having varying MWCNT loading is plotted in real time as shown in Fig. 8 (a). It is clear from Fig. 8(a) that the sensitivity exhibits a strong and peculiar dependence on the concentration of MWCNTs in the alumina matrix. The sensitivity is found to be 8.8 %, 9.6%, 13.2%, 8.0% and 8.5% for sample G1, G2, G3, G4 and G5 respectively at 635nm wavelength of the laser beam. The similar fashion of sensitivity is achieved for the other two wavelengths i.e. 785nm and 1064nm. Fig. 8(b), shows the error bar graph indicating the sensitivity of all the sensors (G1, G2, G3, G4, and G5). The error/deviation for each sensor has been calculated from average value of the five samples prepared for each concentration. As can be seen from this Figure, this deviation is very small (around ~0.3% for G3 sample). This indicates the feasibility of scaling this fabrication technique on commercial scale. Also, it is noticeable from Fig. 8(b) that the sensitivity, irrespective of laser wavelength, first increases with increase in the concentration of MWCNTs until a certain maximum value is achieved followed by decrease with further increase in concentration of MWCNTs. The optimum loading of MWCNTs is found to be as 1.5 wt% (G3) for all three wavelengths of the laser beam. The reason for this trend can be interpreted with the dispersion quality of MWCNTs in alumina matrix. Poor dispersion causes agglomeration/lump formation preventing laser beam to intercept majority of CNTs within the lump. As a result, less photoinduced carriers are generated and therefore, less response.
A
CC
EP
Fig. 9 depicts the variation of response- and recovery time of the sensors (G1, G2, G3, G4, and G5) having varied MWCNTs loading for 635nm, 785nm and 1064nm wavelength of the laser beam. The response time is defined as the time taken by the photoresponse to reach from 10% to 90% of its final value. As is evident from the Fig. 9 (a) that the response time varies inversely with the concentration of MWCNTs for all the three wavelengths of the laser beam. The minimum response time is obtained for G5 sensor (=3wt % MWCNTs) and these are 1s, 0.9s, and 1.4s for laser wavelength 635nm, 785nm, and 1064nm, respectively. Similarly, the recovery time of the sensor is defined as the time to reach the photoresponse from 90% to 10% of its value. It is clear from Fig. 9(b) that G5 sensor shows fastest recovery time and the values for its
Recovery times are 0.8s, 1s, and 0.8s for laser wavelength 635nm, 785nm, and 1064nm respectively. However, in terms of sensitivity G5 sensor does not show best performance and was found to exhibit relatively less sensitivity than G3 sensor. The selection of any sensor depends on certain important sensing parameters viz. sensitivity, resolution, minimum lower detection limit, and response- and recovery time. In present study, sensor G3 exhibits maximum sensitivity and this can make its – (i) detection limit very low to enable the sensor to respond at very low illuminations; and (ii) improve the resolution. At the same time, its response- and recovery time are of the order of seconds (
