J Mater Sci: Mater Electron

DOI 10.1007/s10854-017-6685-0

Fabrication, characterization and enhanced sensing performance

of graphene-TiO2 gas sensor device
Mahdi Taymouri Amiri1 · Ali Akbar Ashkarran1 

Received: 14 October 2016 / Accepted: 28 February 2017

© Springer Science+Business Media New York 2017

Abstract Graphene-TiO2 nanocomposite layers (G-TiO2- applications in diverse industries, medicine, indoor and
NCLs) were prepared using sol–gel method along with outdoor air qualities and other environmental studies [2–4].
spin coating deposition and applied for detection of ­CO2 Semiconductor gas sensors have been widely used in port-
gas. The physicochemical properties of G-TiO2-NCLs gas able gas detection systems because of their advantages such
sensors were determined by X-Ray diffraction (XRD), as low cost, easy production, compact size and simple elec-
ultra-violet visible spectroscopy (UV–Vis), scanning elec- tronic measurement devices [5, 6]. In this regard, ­TiO2 has
tron microscopy (SEM), transmission electron microscopy been widely used as a gas senor due to its chemical stabil-
(TEM) and Fourier transform infrared-spectroscopy (FT- ity, absence of toxicity and low price [7, 8]. Lots of reports
IR). The gas sensing properties of the prepared G-TiO2- are available concerning pristine and functionalized T ­ iO2
NCLs was successfully demonstrated by detection of ­CO2 gas sensors prepared by various techniques such as sol–gel,
gas at different operating temperatures and gas concentra- sputtering, electron beam evaporation and spray pyrolysis
tions. Both pristine T
­ iO2 and G-TiO2-NCLs sensors showed [4, 6].
a faster response at operating temperature of 200 °C. It was In recent years, great efforts have been made to develop
found that G-TiO2 gas sensors show higher response toward graphene-TiO2 nanocomposite due to the high electron
sensing ­CO2 gas compared with pristine T ­ iO2 sensors at transport on the surface of the graphene and its large sur-
optimum temperature. The enhancement in the gas sens- face area [9–12]. It was previously shown that introducing
ing performance of G-TiO2-NCLs can be ascribed to the graphene into the semiconductor matrix could enhance the
introducing of graphene into the T­ iO2 matrix. The response gas sensing properties due to the presence of large amount
of G-TiO2-NCLs gas sensors was maximum (1.34) when of adsorption sites on the surface of the graphene compared
0.001 g graphene was incorporated into the T ­ iO2 matrix at with individual materials [3, 13]. Li et al., prepared reduced
200 °C as an optimum operating temperature. graphene oxide (rGO) decorated ­ TiO2 microspheres by
hydrothermal method for detection of ammonia [11]. They
have suggested that covered rGO with ­TiO2 could improve
1 Introduction the ­NH3 adsorption leading to the obvious sensitivity and
excellent selectivity to different concentrations of ammo-
Gas sensing applications of nanomaterials received much nia from 5 ppm to 50 ppm at room temperature. They have
attention in both industries and scientific communities found that when rGO bridged T ­ iO2 particles, the resistance
due to environmental concerns [1]. Sensing of harmful of the composite is lowered consequently making the sen-
gases has become more significant because of its extensive sor work at room temperature. On one hand, the acidity of
the surface of ­TiO2 has a preferential adsorption of base
ammonia and may thus increase the ammonia adsorption at
* Ali Akbar Ashkarran the interfacial area of rGO–TiO2 leading to higher response
ashkarran@umz.ac.ir
to ammonia.
1
Department of Physics, Faculty of Basic Sciences, University Similarly, Acharyya and co-workers prepared a highly
of Mazandaran, Babolsar, Iran efficient room temperature gas-sensor device based on the

13
Vol.:(0123456789)
 J Mater Sci: Mater Electron

hybrid structure of ­TiO2 nanotubes (TNTs) and rGO by different concentrations of graphene (details for synthesis
electrochemical anodization technique followed by electro- of graphene was reported in our previous work [15]) solu-
deposition of rGO layer in the form of discrete islands [14]. tion containing 0.0001, 0.0005, 0.001, 0.05 and 0.1 g gra-
They have shown that uniformly distributed rGO clusters phene in ethanol as a graphene source were used. The next
acted as an efficient interconnect among the neighboring steps were the same as synthesis of pristine ­TiO2 NLs.
nanotubes which lead to extremely fast response times and
recovery times due to the high carrier mobility of rGO. 2.2 Preparation of G‑TiO2 gas sensor device
In this study, we have prepared G-TiO2-NCLs by dop-
ing graphene into the ­TiO2 matrix via sol–gel and spin Experimental set up of G-TiO2 gas sensor device is sche-
coating methods. We have also studied the effect of vari- matically depicted in Fig. 1. Comb-like Cu electrodes were
ous graphene concentrations on gas sensing property of the first patterned onto glass substrate by sputtering method.
produced ­TiO2 NCLs at different temperatures. Moreover, The width of overlap electrodes and the gap between two
the amount of optimum graphene content on gas sensing successive electrodes was 1  mm. Then the prepared gra-
property of ­TiO2 NCLs in addition to optimum working phene-TiO2 solutions were deposited on glass substrate
temperature and gas concentration was investigated. having cleaned Cu electrodes using the spin coating tech-
nique at a fixed speed of 6000 rpm. Then the prepared lay-
ers was dried at 100 °C for 1  h and annealed at 450 °C in
2 Experimental details air for another 1 h to obtain graphene-TiO2 sensor device.
Gas sensing measurements were performed in a vacuum
2.1 Synthesis of ­TiO2 NCLs chamber at the presence of different concentrations of ­CO2
gas by monitoring the electrical resistance of the prepared
The sol–gel and spin coating techniques were used to syn- sensors using a computer-controlled multimeter (Agilent
thesize pristine ­TiO2 and G-TiO2-NCLs. For preparation of 34,410  A). The response of the sensor usually defines as
­TiO2 solution 1 mL T ­ iCl4 (Merck, 99%) was slowly added the relative changes in the resistance of the sensor at the
dropwise into 10 mL ethanol (Merck, 99.8%) under vigor- presence and absence of C­ O2 gas according to Eq. (1):
ous stirring at room temperature. A large amount of HCl
gas was exhausted and a transparent yellowish solution
S = Rg ∕Ra (1)
was formed. The final solution was then stirred for another where ­Rg is the resistance at the presence and R
­ a is the
30 min to get a homogeneous solution. Before coatings pro- resistance in the absence of ­CO2 gas.
cess all substrates (25 × 20 × 2 ­mm3 soda lime glasses were
used as substrate) were cleaned with acetone, HCl 0.2  M, 2.3 Characterization
distilled water and finally using an ultrasonic (30 kHz, for
15  min.). The substrates then were spin coated at a fixed Crystalline structure of the samples was determined by
speed of 6000 rpm in air and room temperature using the as XRD diffractometer (X’pert Philips) with wavelength of
prepared ­TiO2 sol. The obtained layers were then dried at Cu ­Kα radiation in 2Ө range from 10° to 80° by 0.04° s­ ec− 1
100 °C for 1 h and annealed at 450 °C in air for another 1 h steps. UV–Vis spectroscopy of the samples was taken by
to obtain pristine T
­ iO2 NLs. For synthesis of G-TiO2-NCLs, a double beam Optizen POP spectrophotometer (Mecasys

Fig. 1  A schematic illustration

of the G-TiO2-NCLs gas sensor
experimental set up

13
J Mater Sci: Mater Electron

Company, Korea) from 200  nm to 1000  nm with a step spectroscopy of the G-TiO2-NCLs was carried out using a
of 1  nm. SEM analysis was carried out using an electron FT-Raman 960 Raman scattering spectrometer equipped
microscope instrument (Philips XL30) at 10 to 20  keV with an Nd-YAG laser source operating at 1064 nm wave-
accelerating energy. TEM images were taken by a Philips length at room temperature.
CM120 instrument at 120  keV accelerating energy by
deposition of nanostructures onto the copper grid at room
temperature. In fact, before deposition of final layers and 3 Results and discussion
fabrication of gas sensor device a few drops of the cor-
responding sample in liquid form deposited on a copper SEM and TEM images of the pristine T ­ iO2 and G-TiO2-
grid and then transferred into the TEM microscope for NCLs containing 0.001 g graphene are displayed on Fig. 2.
TEM image. The FT-IR spectroscopy of the films was per- It was observed that spherical shaped nanoparticles (NPs)
formed by a Thermo Nicolet dispersive spectrometer with are formed on the pristine T
­ iO2 sample (Fig.  2a). This is
a 100  mW laser power and 10  mm spot size. The Raman while for G-TiO2 sample a few nanoplatelets are prepared

Fig. 2  SEM and TEM images of pristine T

­ iO2 (a and c) and G-TiO2-NCLs (b and d), respectively

13
 J Mater Sci: Mater Electron

which are attributed to the incorporation of graphene into (a) 120

the ­TiO2 matrix, as is shown in Fig. 2b. It was reported that
presence of graphene can cause an increase in nanometric
cracks and as a result increase of surface to volume ratio 100
which is a crucial factor in gas sensor devices. More details
C-OH
about the microstructure of the prepared G-TiO2-NCLs are

Transmitance (%)
80 Ti-O-C
found by TEM analysis. TEM images of the pristine ­TiO2
confirmed that almost spherical T ­ iO2 (NPs) with an aver- 60
age size of ~ 35 ± 4 are formed during the synthesis process
(Fig. 2c). On the other hand, TEM image of G-TiO2-NCLs
40 Ti-O-Ti
which is depicted in Fig. 2d verifies that T ­ iO2 (NPs), with
almost narrow size distribution, are dispersed on the sur-
face of graphene sheets and supported by the transparent 20
TiO2
graphene sheets. The presence of graphene as a template G-TiO2-NCLs
leads to a minor agglomeration of the ­TiO2 (NPs) and con- 0
sequently enhances their active sites for sensing of gas mol- 4000 3400 2800 2200 1600 1000 400
ecules [16]. Such a combination structure is highly desired Wavenumber (cm-1)
for gas sensing applications due to the enhanced transporta-
(b)
tion of electrons between ­TiO2 (NPs) and graphene sheets

(101)
which is also reported by Wang and co-workers [17].
FT-IR spectra of pristine ­TiO2 and G-TiO2-NCLs are
depicted in Fig.  3a. The wide absorption bands around
690 and 3420 cm− 1 are due to the formation ­TiO2 anatase

(200)
(004)
phase which is present for both pristine ­TiO2 and G-TiO2-
Intencity (a.u.)

(211)
TiO2
NCLs [18]. Band appeared around 3420 cm− 1 is assigned

(105)

(204)

(220)
(215)
(116)
to the bending and stretching modes of the Ti-OH groups in
both pristine T­ iO2 and G-TiO2-NCLs samples, respectively
[19]. Moreover, the bands between 400 and 700  cm− 1
are attributed to the deformation of Ti-OH and Ti-O-Ti G-TiO2-NCLs
groups. Bands appeared around 1600  cm− 1 is assigned to
Ti-O-C bond vibration due to the presence of graphene in
G-TiO2-NCLs sample [20]. Based on the appeared bend-
ing and stretching modes in FT-IR spectra of G-TiO2-NCLs
sample we may conclude that graphene sheets are grafted 10 20 30 40 50 60 70 80
with ­TiO2 matrix. In fact, for pristine ­TiO2 and the G-TiO2- 2Ө (degree)
NCLs, the broad absorption below 1000 cm− 1 is apparently
ascribed to the combination of Ti–O–Ti and Ti–O–C vibra- Fig.  3  a FT-IR spectra and b XRD patterns of pristine T
­ iO2 and
tion modes resulting from the chemical interaction between G-TiO2-NCLs
­TiO2 and graphene. Furthermore, the peak appeared around
1600 cm− 1 attributed to the skeletal vibration of graphene
which becomes wider for G-TiO2-NCLs sample compared The single crystalline size determined by Debye–Scherrer
with pristine ­TiO2 [21]. In fact, introducing graphene into formula was about 14 nm. Notably, no diffraction peaks of
the ­TiO2 matrix results in its overlapping with the absorp- graphene were observed in XRD pattern of 0.001 g G-TiO2-
tion band at 1633  cm-1 for Ti–O–Ti stretching vibration. NCLs (Fig. 3b). This is due to the fact that the main charac-
That is why the Ti–O–Ti absorption band decreases in teristic peak of graphene is located around 2Ө = 25 ̊ which
FT-IR spectrum of G-TiO2-NCLs. overlaps with (101) peak of ­ TiO2 antase phase. These
To find the crystalline phase and structure of the final results indicate that dopant have no effect on the crystal-
products XRD analysis was performed. Figure  3b shows line phase and structure of ­TiO2 [4]. The average crystal-
the XRD patterns of pristine ­TiO2 and G-TiO2-NCLs sam- lite size, calculated by Scherrer formula, is ~ 12.8  nm for
ples. XRD result for bare ­TiO2 sample demonstrates forma- G-TiO2-NCLs.
tion of crystalline anatase phase with characteristic peaks at Figure  4 shows the Raman spectrum of the G-TiO2-
2Ө = 25.3, 37.7, 47.8, 54, 62.7 and 75.2 degree which is in NCLs. The bands appeared below the 1000  cm− 1 (about
agreement with 21-1272 standard card from JCPDS [22]. 225 cm− 1, 410 cm− 1, and 645 cm− 1) corresponds to Raman

13
J Mater Sci: Mater Electron

(a) 1.2
200 °C
150 °C
250 °C
Ti-O
300 °C

Sensetivity (Rg/Ra)
1.1

G-band
Intensity (a.u.)

2D-band
D-band

0.9
0 1000 2000 3000 4000 5000 6000 7000
0 500 1000 1500 2000 2500 3000 3500
T (s)
Raman shift (cm-1)
(b)
1.19
Fig. 4  Raman analysis of G-TiO2-NCLs containing 0.001 g graphene

1.17
modes for pristine anatase ­TiO2. On the other hand, two Response (Rg/Ra)
bands appeared around 1330 and 1600 cm− 1 in the Raman
spectrum ascribed to the disordered (D) band and gra- 1.15

phitic (G) band, respectively [23]. The D band is associ-

ated with the defects, while the G band is responsible for 1.13
photons corresponding to s­p2 carbon. This means that in
addition to T­ iO2 the main structure of graphene remained
1.11
unchanged in the G-TiO2-NCLs which is also in agreement
with XRD results. Furthermore, appearance of 2D Raman
band of graphene materials (around ~ 2605  cm− 1) reveals 1.09
that some single layers of graphene may form in the final 150 200 250 300

G-TiO2-NCLs [24]. Consequently, according to the Raman Temperature (°C)

and XRD results we may conclude that graphene single lay-
ers probably distributed within the T­ iO2 matrix. Moreover, Fig.  5  a Sensitivity of the pristine T
­ iO2 sensor at different operating
annealing the samples at 450 °C would not remove carbon temperatures in the presence of 10,000 ppm C ­ O2 gas and b trend of
materials from the final G-TiO2-NCLs [25, 26]. sensor response as a function of temperature
Gas sensing performance of pristine T ­ iO2 and G-TiO2-
NCLs was investigated for detection of C ­ O2 gas. Figure 5a transfers electrons to adsorbed oxygen atoms and form
shows the sensitivity of the pristine ­TiO2 sensor vs. time ionic species such as ­O−, ­O2− and O−2 . The adsorption of
at different temperatures ranging from 150 to 300 °C in the ­O− is more reactive and thus makes the material more sen-
presence of 10,000 ppm ­CO2 gas in order to determine the sitive to the presence of reducing gas [8]. Furthermore,
optimum operating temperature. It was previously reported as the temperature increases the leading process becomes
that the response of sensors is highly affected by the operat- the adsorption of ­ O− and consequently the sensitivity
ing temperature [2]. It was found that sensitivity enhances enhances. On the other hand, by more increase in the tem-
by increasing the operating temperature and arrives at an perature the formerly adsorbed ionic oxygen desorbs and
upper limit value at optimum temperature (200 °C) and sensitivity declines [3]. The maximum response was found
declines again by more increase in the operating tempera- to be 1.17 in the optimum operating temperature (see for
ture. Such behavior in G-TiO2-NCLs gas sensor device can example, Fig. 5b). Figure 6a demonstrates sensor sensitiv-
be explained based on gas adsorption/desorption mecha- ity at optimum operating temperature for pristine ­TiO2 and
nism on the surface of the sensor device [7]. In other word, G-TiO2-NCLs with various graphene content (ranging from
when ­TiO2-NCLs sensor device is exposed to oxygen of 0.01, 0.005, 0.001, 0.0005 to 0.0001 g) at the presence of
atmosphere, it can adsorb and interact with oxygen and 10,000  ppm ­CO2 gas. Pristine T ­ iO2 NCLs showed lower

13
 J Mater Sci: Mater Electron

(a) 1.4 (a) 1.4

200°C

1.3 150°C
1.3

250°C
Sensetivity (Rg/Ra)

Sensetivity (Rg/Ra)
1.2
1.2
300°C

1.1
1.1

TiO2+G(0.0005 g )

TiO2+G(0.0001 g )
TiO2+G(0.005 g )

TiO2+G(0.001 g )
TiO2+G(0.01 g )
TiO2

1
1

0.9
0 1000 2000 3000 4000 5000 6000 7000 8000 0.9
T (s) 0 1000 2000 3000 4000 5000 6000 7000
Time (S)
(b) 1.39
(b) 1.39

1.34
1.34
Response (Rg/Ra)

1.29
Response (Rg/Ra)
1.29

1.24
1.24

1.19 1.19

1.14 1.14

1.09 1.09
0 0.0001 0.0005 0.001 0.005 0.01 150 200 250 300

Concentration of graphene (g) Temperature (°C)

Fig.  6  a Sensitivity of the G-TiO2-NCLs sensors with various gra- Fig.  7  a Sensitivity of the G-TiO2-NCLs gas sensor device contain-
phene concentration at 200 °C in the presence of 10,000 ppm ­CO2 gas ing 0.001 g graphene at different operating temperatures in the pres-
and b trend of G-TiO2-NCLs sensor as a function of graphene content ence of 10,000  ppm C ­ O2 gas and b trend of sensor response as a
function of temperature

response in comparison with G-TiO2-NCLs in the presence

of 10,000  ppm C ­ O2 gas and optimum temperature. This which leads to a decrease in sensitivity of G-TiO2 sensor. In
may be due to the fact that graphene prevents the accumu- other word, more accumulation of graphene results again in
lation of ­TiO2 (NPs) in G-TiO2-NCLs. It is seen that sensor a reduction in electrical conductivity of the G-TiO2 NCLs
response improves and reaches the maximum value (~ 1.34) which directly influence the sensor response.
when the concentration of graphene increases up to 0.001 g Therefore we may conclude that there is an optimum
and then declines by more increase of graphene content amount of graphene in T ­ iO2 matrix for the maximum
(Fig.  6b). This could be ascribed to the excessive active improvement in the gas sensing performance. The same
surface area provided by introducing graphene into the process was applied to find the optimum operating tem-
­TiO2 matrix [9]. In fact, introducing graphene into the ­TiO2 perature of G-TiO2-NCLs obtained at optimum amount
matrix prevents the accumulation of T ­ iO2 (NPs) which of graphene in T ­ iO2 matrix. Figure  7a illustrates the sen-
increases the electrical conductivity. In the other hand, by sitivity of the G-TiO2-NCLs gas sensor device containing
more increase in the concentration of graphene, uniform 0.001 g graphene vs. time at different temperatures ranging
distribution between T ­ iO2 matrix and graphene disappears from 150 to 300 °C in the presence of 10,000 ppm C ­ O2 gas.

13
J Mater Sci: Mater Electron

Similarly, it was found that sensitivity improves by increas- notably enhanced the gas sensing performance of T ­ iO2.
ing the operating temperature up to 200 °C and decreases G-TiO2-NCLs gas sensor devices revealed a remarkable
by more increase in the operating temperature as illustrated increase in the sensitivity of the sensor compared with pris-
in Fig. 7b. This is due to the fact that there is a competition tine ­TiO2, which was concluded to be due to the electronic
between slow kinetics at low temperatures and enhanced modification as a result of incorporation of graphene into
desorption at high temperatures [12]. Therefore, it was sug- the ­TiO2 matrix.
gested that the optimum operating temperature for both
pristine ­TiO2 and G-TiO2-NCLs sensor device is 200 °C.
The sensitivity of the G-TiO2-NCLs sensors in the pres- References
ence of different C ­ O2 gas concentrations and optimum
operating temperature is also demonstrated in Fig. 8. Sen- 1. D. R. Miller, S. A. Akbar, P. A. Morris, Sens. Actuators, B 204,
250 (2014).
sor response as a function of C ­ O2 concentration revealed 2. S. M. Mohammad, Z. Hassan, R. A. Talib, N. M. Ahmed, M. A.
an increase with increasing the C ­ O2 concentration from Al-Azawi, N. M. Abd-Alghafour, C. W. Chin, N. H. Al-Hardan,
500 to 15,000 ppm. This shows that the prepared G-TiO2- J. Mater. Sci. - Mater. Electron. 27, 9461 (2016).
NCLs are promising gas sensing devices for detection of 3. A. Esfandiar, S. Ghasemi, A. Irajizad, O. Akhavan, M.R.
Gholami, Int. J. Hydrogen Energy 37, 15423 (2012)
­CO2, which can be applied for monitoring toxic gases in the 4. D. Meng, T. Yamazaki, T. Kikuta, Sens. Actuators, B 190, 838
environment. (2014).
5. Q. Gao, W. Zeng, R. Miao, J. Mater, Sci. - Mater. Electron 27,
9410 (2016)
6. A. Wisitsoraat, A. Tuantranont, E. Comini, G. Sberveglieri, W.
4 Conclusion Wlodarski, Thin Solid Films 517, 2775 (2009)
7. B. Karunagaran, P. Uthirakumar, S. J. Chung, S. Velumani, E. K.
Pristine ­TiO2 and G-TiO2-NCLs gas sensor devices were Suh,. Mater. Charact. 58, 680 (2007).
fabricated through sol–gel process and spin coating deposi- 8. Y. Kwon, H. Kim, S. Lee, I. J. Chin, T. Y. Seong, W. I. Lee, C.
Lee, Sens. Actuators, B 173, 441 (2012).
tion. Chemical composition of the prepared sensor devices, 9. S. Basu, P. Bhattacharyya, Sens. Actuators, B 173, 1 (2012).
determined by FT-IR, showed formation of ­TiO2 in addi- 10. H. D. Jang, S. K. Kim, H. Chang, E. H. Jo, K. M. Roh, J. H.
tion to XRD analysis which verified formation of T ­ iO2 Choi, J. W. Choi, Aerosol Sci. Technol. 49, 538 (2015).
anatase phase. The results of ­CO2 gas sensing measure- 11. X. Li, Y. Zhao, X. Wang, J. Wang, A. M. Gaskov, S. A. Akbar,
Sens. Actuators, B 230, 330 (2016).
ments revealed that not only there is an optimum amount 12. E.H. Song, Y.F. Zhu, Nanosci. Nanotechnol. Lett 5, 198 (2013)
of dopant for the maximum response of the prepared sen- 13. G. Peng, J.E. Ellis, G. Xu, X. Xu, A. Star, ACS Appl. Mater.
sor but also an optimum operating temperature in which Interfaces 8, 7403 (2016)
the response of the sensor device is maximum. It was 14. D. Acharyya, P. Bhattacharyya, IEEE Electron Device Lett. 37,
656 (2016).
suggested that introducing graphene into the ­TiO2 matrix 15. A.A. Ashkarran, B. Mohammadi, Appl. Surf. Sci 342, 112

(2015)
16. X. Huang, Z. Yin, S. Wu, X. Qi, Q. He, Q. Zhang, Q. Yan, F.
1.3 Boey, H. Zhang, Small 7, 1876 (2011)
15000 ppm
17. Q. Wang, X. Guo, L. Cai, Y. Cao, L. Gan, S. Liu, Z. Wang, H.
Zhang, L. Li, Chem. Sci. 2, 1860 (2011).
18. M.A. Abdullah, F.K. Chong, J. Hazard. Mater 176, 451 (2010)
1.2 10000 ppm 19. M. Abdennouri, R. Elmoubarki, A. Elmhammedi, A. Galadi, M.
Baalala, M. Bensitel, A. Boussaoud, Y. El Hafiane, A. Smith, N.
Sensetivity (Rg/Ra)

5000 ppm Barka, Journal of Materials and Environmental Science 4, 953

2500 ppm (2013)
1.1 20. W. Y. Yan, Q. Zhou, X. Chen, X. J. Huang, Y. C. Wu, Sens.
500 ppm Actuators, B 230, 761 (2016).
21. A.J. Wang, W. Yu, Y. Fang, Y. Song, D. Jia, L. Long, M.P.
Cifuentes, M.G. Humphreye, C. Zhang, Carbon 9, 130 (2015)
22. A.A. Ashkarran, J. Cluster Sci 22, 233 (2011)
1
23. 23. J.J. Zhang, Y. H. Wu, J. Y. Mei, G. P. Zheng, T. T. Yan, X. C.
Zheng, P. Liua, X.X, Guan. Photochem. Photobiol. Sci 15, 1012
(2016)
24. O. Akhavan, Carbon 81, 158 (2015)
0.9 25. J. Hong, M. K. Park, E. J. Lee, D. Lee, D. S. Hwang, S. Ryua,
0 1000 2000 3000 4000 5000 6000 Sci Rep. 3 2700 (2013).
Time (S) 26. A.V. Dolbin, M.V. Khlistyuck, V.B. Esel’son, V.G. Gavrilko,
N.A. Vinnikov, R.M. Basnukaeva, Low. Temp. Phys 42, 57
(2016)
Fig. 8  Sensitivity of the prepared G-TiO2-NCLs gas sensor device at
optimum temperature exposed to various ­CO2 concentration

13