Sensors and Actuators B 238 (2017) 1120–1127

Contents lists available at ScienceDirect

Sensors and Actuators B: Chemical journal homepage: www.elsevier.com/locate/snb

Superior enhancement of NO2 gas response using n-p-n transition of carbon nanotubes/SnO2 nanowires heterojunctions Quan Thi Minh Nguyet a,b , Nguyen Van Duy a,∗,1 , Nguyen Thi Phuong a , Nguyen Ngoc Trung b , Chu Manh Hung a , Nguyen Duc Hoa a , Nguyen Van Hieu a,∗,1 a b

International Training Institute for Materials Science (ITIMS), Hanoi University of Science and Technology (HUST), Vietnam School of Engineering Physics (SEP), Hanoi University of Science and Technology (HUST), Vietnam

a r t i c l e

i n f o

Article history: Received 22 April 2016 Received in revised form 23 June 2016 Accepted 26 July 2016 Available online 26 July 2016 Keywords: CNTs/SnO2 NWs n-p-n heterojunction Gas sensors ppb level NO2

a b s t r a c t The enhancement of sensor performances for the detection of highly toxic NO2 gas is one of the most important technological issues in the realization of devices for environmental pollution monitoring. Herein, the n-p-n heterojunctions of carbon nanotubes (CNTs) and SnO2 nanowires (NWs) are designed and fabricated by a facile method to investigate their NO2 gas-sensing performance. The designed devices are easily realized by first growing the SnO2 NWs on Pt electrodes using a thermal chemical vapour deposition method and then spray coating the single-walled carbon nanotubes (SWCNTs) and/or multi-walled carbon nanotubes (MWCNTs) to make a connection between the two electrodes. The MWCNTs/SnO2 NWs device has much higher NO2 responsiveness compared to the SWCNTs/SnO2 NWs and the pristine SnO2 NWs devices. The fabricated devices showed ultrahigh enhancement in gas sensing performance for NO2 , whereas these characteristics do not significantly improve for H2 S gas. At a low working temperature of 100 ◦ C, the fabricated sensors are able to detect the NO2 gas concentration down to 20 ppb. We also clarify the sensing mechanism of the fabricated devices. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Next-generation gas sensors with improved sensitivity, reliability, energy consumption, and response-recovery speed have an excellent potential for applications in different fields including environmental monitoring, industrial control, medicine, disease diagnosis, and so on [1]. The design and synthesis of effective materials and devices for high performance and inexpensive gas sensors are of crucial important for practical applications. Metal oxide semiconductors are well known materials for the resistive type of gas sensors due to their ability to sense many toxic gases such as NO2 , NH3 , CO, H2 S, and so on [2]. Tin oxide is considered as one of the most popular gas-sensing materials and the development of SnO2 -based gas sensors has been the focus of research by numerous research groups [3]. Different morphologies of SnO2 such as nanoparticles, nanowires (NWs), nanorods,

∗ Corresponding authors at: International Training Institute for Materials Science (ITIMS), Hanoi University of Science and Technology (HUST), No. 1, Dai Co Viet Road, Hanoi, Vietnam E-mail addresses: [email protected] (N. Van Duy), [email protected] (N. Van Hieu). 1 Post address: No. 1 Dai Co Viet Str., Hanoi, Vietnam. http://dx.doi.org/10.1016/j.snb.2016.07.143 0925-4005/© 2016 Elsevier B.V. All rights reserved.

and thin films have been studied for gas-sensing applications [4]. One-dimensional nanostructures such as NWs have shown advantages and have attracted a lot of interest due to their small size and high crystallinity, which has led to enhanced sensing properties [5]. The NWs also enabled the fabrication of nanojunction structures, which were recently reported to improve the gas-sensing performance through the variation in barrier height between NW–NW contacts [6]. Recently, NWs network sensors have been reported in terms of sensing mechanism and some functionalization for their sensing improvement [7,8]. Heterojunctions of p-n, n-n, and p-p of metal oxide NWs were developed for gas sensors, but mainly based on the surface functionalization. For instance, Choi, et al. reported on the improvement of response to reducing gases of the n-type SnO2 NWs by surface functionalization with p-type Cr2 O3 nanocrystals [9]. The decoration of p-type hetero-nanocrystals on the surface of p-type NWs was also reported significantly effect on the sensitivity of the sensors depending on the suppression and/or expansion of the hole-accumulation channel at the interface of heterojunction [10]. The n-p structure of the SnO2 @CuO core@shell NWs was reported to enhance the response to reducing gases, while it deteriorated the sensing performance to oxidizing gas of the SnO2 [11]. The nanofibers of reduced graphene oxide and metal oxide were also reported to show significantly high response to NO2 gas

Q.T. Minh Nguyet et al. / Sensors and Actuators B 238 (2017) 1120–1127

1121

Fig. 1. Processes for device fabrication: (A) Thermally oxidized Si substrate equipped with pair Pt electrodes; (B) The electrodes with selective growth of SnO2 NWs; (C) CNTs/SnO2 NWs heterojunction devices. (1) the step of in-situ growth of SnO2 NWs and (2) the step of CNTs deposition.

[12]. A Schottky contact made of semiconducting NWs and metal electrode was also reported to be very sensitive to the gas adsorption/desorption and could be used for a fast response sensor [13]. The fabrication of a NW–metal Schottky contact, however, requires an expensive technique such as the focused ion beam technique and is thus not suitable for low-cost applications and mass production. In addition, the top metal contact can prevent the analytic molecules from diffusing and adsorbing on the surface of buried NWs, thus limit the sensing performance. Meanwhile, owing to their large specific surface area as well as some advanced electronic properties, carbon nanotubes (CNTs) show great potential in gas-sensing applications, especially at low working temperature [14,15]. However, pure CNTs-based sensors have some disadvantages in practical use due to the strong adsorption of analytic gas molecules on their surfaces, which leads to a long response-recovery time [16,17]. The combination of CNTs and metal oxide for gas-sensing applications has been studied recently owing to the complementary properties and heterojunction formation of the two materials [17,18]. The hybrids or composites of CNTs and metal oxides enable the sensors to operate at room temperature [18,19]. The interesting properties and applications of Schottky contacts between CNTs and several metal oxides were also represented elsewhere [20–22]. For instance, Vuong et al. reported the p-n heterojunction of p-type SWCNTs and n-type tungsten oxide thin film for photodetection [21]. Yoon et al. prepared a p-n heterojunction between SWCNTs and SnO2 nanowires for an optoelectronic device [22]. The fabrication approaches reported were rather complicated and unrepeatable. Moreover, the heterojunction in the form of a thin film limited the ability of the gas molecules to diffuse and adsorb on the bottom layer, thus reducing the sensor performance. In this work, we demonstrated the simple fabrication of CNTs (both SWCNTs and MWCNTs) and SnO2 NWs heterojunctions in an n-p-n structure to study their gas sensing properties. SnO2 NWs were grown in situ on Pt electrode using a thermal chemical vapor deposition (CVD) method, and then the CNTs were spray-coated to bridge the two adjacent electrodes. The bridging structure of CNTs on SnO2 NWs allows us to study the variation in electrical properties of their heterojunction to detect gases. We believe that the CNTs/SnO2 NWs heterojunction is the factor that makes the greatest contribution to the devices’ conductance, leading to excellent enhancement of their sensing performances for NO2 gas.

2. Experimental The procedures for sensor fabrication are illustrated in Fig. 1. The fabrication process involved three simple steps: conventional photolithography patterning of Pt electrodes (step 1), CVD growth of SnO2 NWs (step 2), and spray-coating of CNTs (step 3). The comblike electrodes consisting of Cr (10 nm) and Pt (100 nm)/Au (5 nm) were fabricated via conventional lithography followed by metal deposition techniques. The SnO2 NWs were grown directly on electrodes via the thermal CVD method, as reported elsewhere [23]. In a typical synthesis, 0.1 g of tin powder was loaded on an alumina boat and then inserted into the hot zone of a quartz tube furnace. The sensor chips were kept about 2 cm from the alumina boat. The SnO2 NWs were grown by increasing the furnace temperature to 750 ◦ C at a rate of 30 ◦ C/min and maintained for 15 min under a flow of oxygen at 0.5 sccm. After the growth, the furnace was turned off and allowed to cool naturally to room temperature. Two types of commercial CNTs (SWCNTs and MWCNTs) were used for device fabrication. The size parameters were a diameter of 2 nm and length of 100 ␮m for SWCNTs; the corresponding values for MWCNTs were 10 nm and 10 ␮m. To prepare the solution of CNTs for spraying, 10 mg of each type of CNTs was dispersed in 100 ml of isopropanol under high-power ultrasonic vibration for about 5 min. Then 5 ml of CNTs solution was spray-coated onto the top part of the grown SnO2 NWs electrodes through a shadow mask. After spray-coating of CNTs, the sensors were heat-treated at 350 ◦ C in ambient air for 5 h to remove the solvent and stabilize the CNTs/SnO2 NWs contact. The fabricated devices were observed by field emission scanning electron microscopy (FE-SEM, JEOL 7600F) and Raman spectroscopy. The electrical properties and gas-sensing performance for NO2 and H2 S gases were investigated using a Keithley model 6002A instrument and a homemade gas mixing system. Sensors with pure CNTs coated on Pt electrodes and pristine SnO2 NWs grown directly on electrodes were also studied for comparison [24]. In the case of fabrication of the pristine SnO2 NWs, we maintained a growth time of 30 min, and thus the nanowires were long enough to bridge the two Pt electrodes. 3. Results and discussion The morphologies of the CNTs/SnO2 NWs chip, SnO2 NWs, MWCNTs, and SWCNTs were characterized by FE-SEM and the images are shown in Fig. 2(A–D). The FE-SEM images of the

1122

Q.T. Minh Nguyet et al. / Sensors and Actuators B 238 (2017) 1120–1127

Fig. 2. FE-SEM images of the (A) CNTs/SnO2 NWs heterojunction device (inset is an image of pristine SnO2 NWs device); (B) pristine SnO2 NWs, (C) MWCNTs, (D) SWCNTs. Raman spectra of (E) MWCNTs/SnO2 nanowires, and (F) SWCNTs/SnO2 NWs.

CNTs/SnO2 NWs chip indicate that the SnO2 NWs grew selectively on the Pt electrode fingers where the Au catalyst was deposited. No NWs were found growing on the bare SiO2 substrate due to the vapor–liquid–solid growth of the SnO2 , where the Au serves as seed for the nucleation of NWs [25]. The morphology of the SnO2 NWs grown on the electrode can be seen in Fig. 2(B); it involved a thick layer of entangled NWs mat, and thus the Pt electrode could not be seen. This ensured that the sprayed CNTs made contact with the SnO2 nanowires only but not with the Pt electrodes. The distributions of MWCNTs and SWCNTs on sensor chips characterized by FE-SEM images are shown in Figs. 2(C) and (D), respectively. The networked CNTs are observable; they act as the conductor to transport charges between electrodes in sensing measurements. Amorphous carbon in impurities can also be observed in the FESEM images as a byproduct of CNTs synthesis. The sprayed CNTs adhered strongly to the substrate and lay as a network; thus the

contact resistance between CNTs can be ignored during gas-sensing measurements [14]. The CNTs films were sprayed very thin so that the bare silicon substrate can be seen through in order to ensure the diffusion of analytic gaseous molecules down to the bottom layers [26]. Raman spectroscopy was also used to characterize the quality of the fabricated devices. The typical peaks at 1321 and 1576 cm−1 , corresponding to the D-band and G-band of MWCNTs, are clearly seen in Fig. 2(E). The intensity of the G-band is much stronger than that of the D-band, indicating the high quality of the MWCNTs. In addition, the Raman spectrum taken in the region of MWCNTs/SnO2 showed an active Raman mode of SnO2 . As can be seen, for SnO2 NWs, the highest Raman shifted peak of rutile crystal structure is located at around 632 cm−1 , corresponding to the A1g vibration mode. Other typical peaks at 474 and 774 cm−1 are rather low for nanostructured material, thus could not be seen in the spectra [17].

Q.T. Minh Nguyet et al. / Sensors and Actuators B 238 (2017) 1120–1127

(A)

(B)

(C)

(D)

Fig. 3. I-V curves of (A) and NO2 gas response at 200 ◦ C of pristine SWCNTs (B), MWCNTs (C) and SnO2 NWs (D) sensors.

The Raman spectrum of the SWCNTs/SnO2 NWs sample shown in Fig. 2(F) reveals the typical characteristics of SWCNTs. The Raman peak at 169 cm−1 corresponds to the radial breathing mode of SWCNTs with an average diameter of 1.42 nm [27]. However, it is difficult to observe the Raman mode of SnO2 NWs in this spectrum because of the exceptionally high Raman signals from the SWCNTs. The typical Raman spectrum of the pure SnO2 NWs can be seen in our previous report [6]. Besides fabricating heterojunction structures, the individual material sensors were also prepared to compare their electrical and sensing properties. The current-voltage characteristics of those sensors are presented in Fig. 3(A). The conducting behavior of CNTs is explained by the overlapped band region as well as the narrow band gap structures. It is easy to understand the Ohmic contact properties of CNTs and Pt electrodes shown in the linear shape of I-V results. However, previous work has indicated that a Schottky contact is supposed to form between Pt and SnO2 materials [28]. This property is due to the high work function of about 5.6 eV for Pt and lower value of about 4.6 eV for SnO2 . In fact, there is no Schottky junction characteristic between SnO2 NWs and Pt electrodes in the on-chip fabricated devices in the present work. Our results can be explained as following. By growing NWs directly on the electrodes via the thermal CVD method, the tin metal first formed an alloy with the Au catalyst as well as Pt. From this alloy, the SnO2 NWs were grown following the vapor–liquid–solid mechanism during the oxidation process. The transition of materials from Pt to SnO2 NWs is therefore not as abrupt as in other coating methods. The Ohmic contact characteristics of the SnO2 NWs sensor is indicated as a linear I-V curve, as shown in Fig. 3(A). Firstly, we checked the gas-sensing characteristics of the pristine materials to NO2 at 200 ◦ C. The typical gas-sensing responses of the pure SWCNTs, MWCNTs, and SnO2 NWs sensors for 1 ppm of NO2 are shown in Fig. 3(B). Both pristine SWCNTs and MWCNTs sensors exhibited p-type gas-sensing characteristics where the sensors’ resistance decreased upon exposure to NO2 gas [26]. This is

1123

the result of the capture of electrons from the p-type sensing layer by NO2 adsorption and results in an increase in hole concentration. In contrast, the SnO2 NWs sensor showed an increase in resistance upon exposure to the oxidizing gas NO2 as a result of n-type gas-sensing behavior [5]. The sensor responses to 1 ppm of NO2 of the pristine SWCNTs, MWCNTs, and SnO2 NWs sensors are about 1.05, 1.04, and 2.61, respectively. The response to NO2 of the pure SnO2 is relatively higher than those of the pristine SWCNTs and MWCNTs. As for the CNTs/SnO2 NWs heterojunction devices, the I-V curves measured at 150 ◦ C in air and in 250 ppb of NO2 are shown in Fig. 4(A). The linear plots of I-V curves are shown in Fig. S4 (Supplementary materials). Evidently, the I-V curves showed typically non-linear characteristics. This means that the CNTs/SnO2 NWs heterojunctions have formed potential barriers blocking the current flowing through the electrodes. This characteristic could be explained by the formation of Schottky junctions between the CNTs and SnO2 NWs. Generally, MWCNTs were reported to be a metallic material with a work function of about 4.6–4.9 eV [23,24] whereas SWCNTs were reported to be a narrow bandgap p-type semiconductor with a work function of about 5.05 eV [25], which differs from the work function of SnO2 NWs (4.6–4.8 eV). It is also evident that the potential barrier formed at the heterojunction of MWCNTs/SnO2 NWs should be lower than that of the SWCNTs/SnO2 NWs heterojunctions under exposure to air as well as NO2 gas. This also means that the conductivity of the MWCNTs/SnO2 NWs device is better than that of the SWCNTs/SnO2 NWs device, as illustrated in Fig. 4(A). Additionally, the better conductivity of the MWCNTs/SnO2 NWs device was possibly due to the better conductivity of the MWCNTs compared with the SWCNTs. It should be noted that the response of CNTs/SnO2 NWs junctions is determined by the ratio of IAir /IGas . The I-V curve displayed in Fig. 4(A) reveals that the ratio is dependent on the applied voltage. Thus, to obtain the maximum response, the applied voltage should be optimized. The greatest change in the current ratio in air for 250 ppb of NO2 is achieved at about 2 V for MWCNTs/SnO2 NWs and above 15 V for SWCNTs/SnO2 . Based on this result, the NO2 gas-sensing measurements were done with these two applied voltages for the two types of CNTs/SnO2 NWs sensors. The gas-sensing characteristics of the two types of heterojunction CNTs/SnO2 NWs sensors were first evaluated by determining the optimum working temperatures. Fig. 4(B) shows the transient response versus time of those sensors as a function of working temperature under exposure to 250 ppb of NO2 gas. It can be seen that both sensors exhibited a similar response at 100 and 150 ◦ C, and these responses are much higher than those measured at 200 ◦ C. The response of both sensors measured at temperatures lower than 100 ◦ C may be higher. However, the response-recovery speed is too low to be interesting for practical applications. Thus, taking the response-recovery time into account, the optimum working temperature of 150 ◦ C should be selected. The resistance transience versus time of the MWCNTs/SnO2 NWs and SWCNTs/SnO2 devices upon exposure to different concentrations of NO2 (20–250 ppb) and sensing temperature of 150 ◦ C is shown in Fig. 4(C). Both sensors showed an increase in resistance upon exposure to NO2 gas and recovered to the initial values after a certain number of cycles, suggesting a well reversible adsorption between NO2 molecules and sensing elements. The response and recovery times of the MWCNTs/SnO2 NWs sensor to 250 ppb NO2 are about 125 and 65 s, whereas the corresponding values for the SWCNTs/SnO2 sensor are 115 and 75 s. Compared to the response of the pure MWCNTs, SWCNTs, and SnO2 NWs, the heterojunction devices showed superior sensitivity with a much lower detection limit at the parts-per-billion level. In addition, the sensor response increases nearly linearly with increasing NO2 concentration in the measured range. At a given concentration, the MWCNTs/SnO2 sensor exhibits a higher response compared

1124

Q.T. Minh Nguyet et al. / Sensors and Actuators B 238 (2017) 1120–1127

(A)

(B)

(D)

(C)

Fig. 4. (A) I-V curves of heterojunction devices; (B) transient resistance versus time of devices; (C) sensor response as a function of NO2 concentration measured at different concentrations of NO2 at 150 ◦ C.

to the SWCNTs/SnO2 NWs counterpart (Fig. 4(D)). This result is inverted to that of the pristine materials. A summary of the gas response (S) and response (tres. ) time of fabricated sensors is provided in Table 1. The results show optimal working temperatures of about 200 ◦ C for the SnO2 NWs sensor and about 100 ◦ C for the CNTs/SnO2 NWs sensor, at which the sensors showed the highest sensitivity. The optimal working temperature of the CNTs/SnO2 NWs device is lower than that of the pristine SnO2 NWs sensor. This could be due to the more active performance of the heterojunction barrier at low temperature. At higher working temperatures, the increase in thermal energy would degrade the role of potential barrier formed at the heterojunction. On the other hand, the addition of the CNTs

layer on SnO2 NWs makes the response time longer because it takes time for the gas to diffuse through the extra material layer [13–15]. Additionally, the fabricated sensors were also characterized with H2 S as a reduced gas, and the results are shown in Fig. 5. As can be seen in Fig. 5(A), the typical responses of n-type and p-type materials were found for SnO2 NWs and CNTs sensors, respectively. In contrast to the NO2 gas response, the resistance of pristine CNTs increases with the introduction of H2 S gas, while the resistance of the SnO2 NWs sensor decreases. The result is assumed to be due to the electrons provided to the sensing materials by the H2 S molecules. These electrons would increase the major carrier concentration in the case of SnO2 NWs but lessen the number of positive charge carriers in CNTs. The pristine sensors showed very

Table 1 Summary for sensing properties of (A) SnO2 NWs network; and (B) CNTs/SnO2 bridging structures. (A) SnO2 Nanowire Network

NO2 conc. (ppb)

150 ◦ C 1000 2500 5000 1000

200 ◦ C

2 2.6 2.9 3.2

35 40 42 47

250 ◦ C

2.6 4.6 5.5 6.6

18 20 22 26

300 ◦ C

1.5 2.5 3.2 5.4

8 10 17 25

1.4 2.1 2.6 4.1

5 10 17 23

(B) NO2 conc. (ppb)

MWCNTs/SnO2 NWs 100 ◦ C

20 50 100 250

SWCNTs/SnO2 NWs 150 ◦ C

200 ◦ C

100 ◦ C

150 ◦ C

200 ◦ C

S

tres (s)

S

tres (s)

S

tres (s)

S

tres (s)

S

tres (s)

S

tres (s)

2.1 3.0 5.7 17.9

90 110 120 125

1.7 2.8 5.9 16.9

80 105 120 130

1.4 2.4 3.4 5.0

65 70 55 50

2.8 4.7 7.3 13.3

100 110 120 115

1.8 3.0 6.5 10.6

60 80 90 85

– – – 1.9

– – – 60

Q.T. Minh Nguyet et al. / Sensors and Actuators B 238 (2017) 1120–1127

(A)

1125

(B)

Fig. 5. (A) Transient resistance versus time of pristine materials at 250 ◦ C upon exposure to 1 ppm H2 S; (B) gas response to various H2 S concentrations at different temperatures of the CNTs/SnO2 NWs, and SnO2 NWs sensors (B).

low responses to 1 ppm of H2 S of about 1.0076, 1.012, and 1.11 for SWCNTs, MWCNTs, and SnO2 NWs, respectively. The transient resistance versus time of the CNTs/SnO2 NWs and SnO2 NWs sensors upon exposure to various H2 S concentrations (1–10 ppm) at different temperatures was also measured (Figs. S1, S2, and S3 in the Supplementary materials). A summary of the gas responses of CNTs/SnO2 NWs and SnO2 NWs sensors to different H2 S concentrations at varying temperatures can be seen in Fig. 5B. It can be clearly seen that the optimal working temperature for SnO2 NWs sensors is about 300 ◦ C. Meanwhile, the optimal temperature of CNTs/SnO2 NWs is observed at around 200 ◦ C. Compared with the response of the pristine materials, the results indicate a slight improvement in gas response in the case of SWCNTs/SnO2 NWs sensors. All of these gas sensing results could be explained by using the electrical and potential barrier model, as seen in Fig. 6. In this model, we assume that the heterojunction between p-type CNTs and ntype SnO2 NWs would act as a typical p-n diode. In the fabricated structures, the electrical current has to go through two heterojunctions (named D1 and D2), and thus the device has an n-p-n structure. At the same time, SnO2 NWs and CNTs act as electrical resistances; they are in serial contact (as observed in Fig. 6A). The n-type SnO2 has a wide bandgap (Eg = 3.6 eV, at 300 K) [12], which is much larger than that of the CNTs (∼0.1–0.5 eV) [31]. The carrier density of SnO2 (3.6 × 1018 cm−3 ) is lower than that of CNTs film (2.8 × 1022 cm−3 ) [32]. Therefore, the n-p-n heterojunction can be considered composing of two Schottky diodes (D1, and D2) in oppositely arrangement, as depicted in Fig. 6A. Because the CNTs/SnO2 NWs heterojunctions are not ideal diodes, thus the resistances RJ1 and RJ2 are representative for the leakage current due to the imperfection of heterojunctions. Since there is always one of D1 or D2 operates in reversed bias voltage during gas sensing measurements, the measured current is mainly determined by the leakage current (or the values of RJ1 and RJ2 ). Based on our observations, the pristine SnO2 NWs show much better gas sensing performance than CNTs (both SWCNTs and MWCNTs). If no potential barrier forms between CNTs and SnO2 NWs (D1 and D2 act as ohmic contacts), then the gas sensing performance will be determined by the contribution of individual materials. However, our results showed that the sensitivity of the heterojunction devices is much higher than that of the pristine sensors, demonstrating the significant contribution of the heterojunction potential barrier. Under the reverse bias condition, the CNTs/SnO2 NWs interfaces act as an electric current regulator. The interaction between analytic gas molecules and interfacial layers is magnified due to the regulation properties [28]. In Fig. 6B, we illustrate the band diagram of the system in different measur-

Fig. 6. (A) Equivalent circuit and (B) illustration of potential barrier, and trap states between CNTs and SnO2 NWs of fabricated heterojunction in Air, NO2 , and H2 S.

ing ambient. The gas interaction is assumed to change the barrier height as well as interface states density right on the materials surface. Hence, in this work, we consider the two current components including ideal diode and interface trap assisted recombination. To clarify the working principle of the device, we tested the I-V curves of a CNTs/SnO2 NWs diode in air and in different gases (NO2 , and H2 S) at various temperatures, and the data are shown in Fig. S5 and

1126

Q.T. Minh Nguyet et al. / Sensors and Actuators B 238 (2017) 1120–1127

S6 (Supplementary materials). In general, the total electric current (Jtotal ) through a diode could be expressed as follows [33]: Jtotal = JSHR + Jideal



JSHR = qB∗ Nt exp − ∗

 ˚B 2kB T

 ˚   B

Jideal = q C exp −

kB T

exp

( exp

 qV  a 2kB T

 qV  a

kB T



−1

− 1)

(1)

to other strong oxidizing gases has not checked in this work yet. Anyhow, our sensor design is excellent for application in the monitoring of highly toxic NO2 gas regarding to the simple fabrication process, bipolar applied voltage, and high sensitivity.

(2) Acknowledgement (3)

Where JSHR is trap assisted recombination current; Jideal is ideal diode current; Nt is signed for interface trap density; kB is Boltzmann constant; T is absolute temperature; Va is applied voltage; and ˚B is barrier height. As shown in Fig. S5 (Supplementary materials), when measured in air, the leakage (reverse) current is fairly high of about 2 mA at bias voltage of −2 V due to the existence of trap states (eg. oxygen vacancies) at the interface. When NO2 , a strong oxidized gas and very active at low temperature, is introduced, it occupied the interface trap states of NWs (Nt decreased), thus reducing the leakage current (or the CNTs/SnO2 NWs diode became more perfect) because the adsorption of NO2 molecules blocking the movement of free electrons from CNTs to SnO2 , and hole from SnO2 to CNTs. As a result, the reversed current of device decreased (resistance increased) significantly, and produces superior sensitivity to NO2 gas. Another possible explanation is assumed to an additional barrier height, ˚ B , due to the negative NO2 radicals stuck on the interfacial layer. From Eq. (2) and (3), it is obviously that the contribution of JSHR to the total current increases with Nt . Especially on reverse bias voltage region, when Va is negative, JSHR increases faster than Jideal . Hence, at specific Nt value, JSHR would be prominent to Jideal . Meanwhile on forward bias region Jideal would be dominant. Combining to the NO2 gas sensing results in Fig. S5, we argue that the changing in Nt would play a role as the main driving force of sensor response. Differently, the reduced H2 S gas is not effective to directly create the trap states on SnO2 NWs. It indicates the existence via surface oxygen radicals as the well-known gas sensing mechanism for SMO materials. The H2 S molecules react to the surface oxygen radicals to release interface trap states. However, this indirect influence is negligible at low working temperature leading to a small resistance change. Additionally, because of the rather high leakage current in air, thus the heterojunction resistance is comparable to SnO2 NWs and CNTs resistance. The total sensor resistance would involve RNWs and RCNT . As the result, the impact of junction resistance change to the sensor resistance decreases (Fig. S6, Supplementary materials). Both of the above effects make the sensor response to H2 S degraded. 4. Conclusion We have designed and fabricated CNTs/SnO2 NWs heterojunction devices for gas-sensing applications. The designed sensor devices were easily realized by in situ growth of SnO2 NWs on a pair of Pt electrodes, followed by spray coating of CNTs (both SWCNTs and MWCNTs) to make the heterojunctions. The electrical and gas sensing characteristics of the fabricated sensors were tested for enhancement of NO2 sensors. The results reveal that the heterojunction devices have superior sensitivity compared to the pristine counterparts, and they can monitor the concentration in parts per billion of the NO2 gas. The MWCNTs/SnO2 NWs sensor showed a higher response to NO2 compared to the SWCNTs/SnO2 NWs device. When testing with a reduced gas, H2 S, the heterojunction sensors did not provide a significant improvement. In addition, with an increase of working temperature, sensitivity of the CNTs/SnO2 devices decreased due to the distortion of rectifying properties of the CNTs/SnO2 heterojunctions. These results suggest that the fabricated sensors are suitable for monitoring highly toxic NO2 at low temperature. Despite that the response of the sensors

This research is funded by Vietnam National Foundation for Science and Technology Development (NAFOSTED) under grant number 103.02-2014.32. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.snb.2016.07.143. References [1] G. Peng, U. Tisch, O. Adams, M. Hakim, N. Shehada, Y.Y. Broza, S. Billan, R. Abdah-Bortnyak, A. Kuten, H. Haick, Diagnosing lung cancer in exhaled breath using gold nanoparticles, Nat. Nanotechnol. 4 (2009) 669–673, http://dx.doi. org/10.1038/nnano.2009.235. [2] E. Comini, C. Baratto, G. Faglia, M. Ferroni, A. Vomiero, G. Sberveglieri, Quasi-one dimensional metal oxide semiconductors: preparation, characterization and application as chemical sensors, Prog. Mater. Sci. 54 (2009) 1–67, http://dx.doi.org/10.1016/j.pmatsci.2008.06.003. [3] V.V. Sysoev, T. Schneider, J. Goschnick, I. Kiselev, W. Habicht, H. Hahn, E. Strelcov, A. Kolmakov, Percolating SnO2 nanowire network as a stable gas sensor: direct comparison of long-term performance versus SnO2 nanoparticle films, Sens. Actuators B 139 (2009) 699–703, http://dx.doi.org/ 10.1016/j.snb.2009.03.065. [4] G. Korotcenkov, Metal oxides for solid-state gas sensors: what determines our choice? Mater. Sci. Eng. B Solid-State Mater. Adv. Technol. 139 (2007) 1–23, http://dx.doi.org/10.1016/j.mseb.2007.01.044. [5] N.D. Chinh, N.V. Toan, V.V. Quang, N.V. Duy, N.D. Hoa, N.V. Hieu, Comparative NO2 gas-sensing performance of the self-heated individual, multiple and networked SnO2 nanowire sensors fabricated by a simple process, Sens. Actuators B 201 (2014) 7–12, http://dx.doi.org/10.1016/j.snb.2014.04.095. [6] L.V. Thong, N.D. Hoa, D.T.T. Le, D.T. Viet, P.D. Tam, A.T. Le, N. Van Hieu, On-chip fabrication of SnO2 -nanowire gas sensor: the effect of growth time on sensor performance, Sens. Actuators B 146 (2010) 361–367, http://dx.doi.org/10. 1016/j.snb.2010.02.054. [7] J.-H. Kim, S.S. Kim, Realization of ppb-scale toluene-sensing abilities with Pt-functionalized SnO2 –ZnO core–shell nanowires, ACS Appl. Mater. Interfaces 7 (2015) 17199–17208, http://dx.doi.org/10.1021/acsami.5b04066. [8] J.-H. Kim, P. Wu, H.W. Kim, S.S. Kim, Highly selective sensing of CO, C6 H6 , and C7 H8 gases by catalytic functionalization with metal nanoparticles, ACS Appl. Mater. Interfaces 8 (2016) 7173–7183, http://dx.doi.org/10.1021/acsami. 6b01116. [9] S.-W. Choi, A. Katoch, J.-H. Kim, S.S. Kim, Prominent reducing gas-sensing performances of n-SnO2 nanowires by local creation of p–n heterojunctions by functionalization with p –Cr2 O3 nanoparticles, ACS Appl. Mater. Interfaces 6 (2014) 17723–17729, http://dx.doi.org/10.1021/am504164j. [10] S.-W. Choi, A. Katoch, J.-H. Kim, S.S. Kim, Remarkable improvement of gas-sensing abilities in p-type oxide nanowires by local modification of the hole-accumulation layer, ACS Appl. Mater. Interfaces 7 (2015) 647–652, http://dx.doi.org/10.1021/am5068222. [11] J.-H. Kim, A. Katoch, S.-H. Kim, S.S. Kim, Chemiresistive sensing behavior of SnO2 (n)–Cu2 O (p) core–shell nanowires, ACS Appl. Mater. Interfaces 7 (2015) 15351–15358, http://dx.doi.org/10.1021/acsami.5b03224. [12] J.-H. Lee, A. Katoch, S.-W. Choi, J.-H. Kim, H.W. Kim, S.S. Kim, Extraordinary improvement of gas-sensing performances in SnO2 nanofibers due to creation of local p–n heterojunctions by loading reduced graphene oxide nanosheets, ACS Appl. Mater. Interfaces 7 (2015) 3101–3109, http://dx.doi.org/10.1021/ am5071656. [13] Y. Hu, J. Zhou, P.-H. Yeh, Z. Li, T.-Y. Wei, Z.L. Wang, Supersensitive, fast-response nanowire sensors by using Schottky contacts, Adv. Mater. 22 (2010) 3327–3332, http://dx.doi.org/10.1002/adma.201000278. [14] N.D. Hoa, N.V. Quy, Y. Cho, D. Kim, Porous single-wall carbon nanotube films formed by in Situ arc-discharge deposition for gas sensors application, Sens. Actuators B 135 (2009) 656–663, http://dx.doi.org/10.1016/j.snb.2008.10.041. [15] P. Qi, O. Vermesh, M. Grecu, A. Javey, Q. Wang, H. Dai, S. Peng, K.J. Cho, Toward large arrays of multiplex functionalized carbon nanotube sensors for highly sensitive and selective molecular detection, Nano Lett. 3 (2003) 347–351, http://dx.doi.org/10.1021/nl034010k. [16] D.H. Shin, J.S. Lee, J. Jun, S.G. Kim, J. Jang, Detection of hazardous gas using multidemensional porous iron oxide nanorods-decorated carbon nanoparticles, ACS Appl. Mater. Interfaces 7 (2015) 1746–1751, http://dx.doi. org/10.1021/am507314t.

Q.T. Minh Nguyet et al. / Sensors and Actuators B 238 (2017) 1120–1127 [17] N.D. Hoa, N.V. Quy, D. Kim, Nanowire structured SnOx -SWNT composites: high performance sensor for NOx detection, Sens. Actuators B 142 (2009) 253–259, http://dx.doi.org/10.1016/j.snb.2009.07.053. [18] F. Mendoza, D.M. Hernández, V. Makarov, E. Febus, B.R. Weiner, G. Morell, Room temperature gas sensor based on tin dioxide-carbon nanotubes composite films, Sens. Actuators B 190 (2014) 227–233, http://dx.doi.org/10. 1016/j.snb.2013.08.050. [19] S. Mubeen, M. Lai, T. Zhang, J.H. Lim, A. Mulchandani, M.a. Deshusses, N.V. Myung, Hybrid tin oxide-SWNT nanostructures based gas sensor, Electrochim. Acta 92 (2013) 484–490, http://dx.doi.org/10.1016/j.electacta.2013.01.029. [20] F.-L. Meng, Y. Jia, J.-Y. Liu, M.-Q. Li, Y.-F. Sun, J.-H. Liu, X.-J. Huang, Nanocomposites of sub-10 nm SnO2 nanoparticles and MWCNTs for detection of aldrin and DDT, Anal. Methods 2 (2010) 1710, http://dx.doi.org/10.1039/ c0ay00424c. [21] N.M. Vuong, H.N. Hieu, D. Kim, An edge-contacted pn-heterojunction of a p-SWCNT/n-WO3 thin film, J. Mater. Chem. C. 1 (2013) 5153, http://dx.doi. org/10.1039/c3tc30845f. [22] J. Yoon, K.W. Min, J. Kim, G.T. Kim, J.S. Ha, Nanowires Nanotechnol. 23 (2012) 265301, http://dx.doi.org/10.1088/0957-4484/23/26/265301. [23] D.T.T. Le, N. Van Duy, H.M. Tan, D.D. Trung, N.N. Trung, P.T.H. Van, N.D. Hoa, N.V. Hieu, Density-controllable growth of SnO2 nanowire junction-bridging across electrode for low-temperature NO2 gas detection, J. Mater. Sci. 48 (2013) 7253–7259, http://dx.doi.org/10.1007/s10853-013-7545-9. [24] V.V. Quang, N.V. Dung, N.S. Trong, N.D. Hoa, N.V. Duy, N.V. Hieu, Outstanding gas-sensing performance of graphene/SnO2 nanowire Schottky junctions, Appl. Phys. Lett. 105 (2014) 013107, http://dx.doi.org/10.1063/1.4887486. [25] D.D. Trung, N.V. Toan, P.V. Tong, N.V. Duy, N.D. Hoa, N.V. Hieu, Synthesis of single-crystal SnO2 nanowires for NOx gas sensors application, Ceram. Int. 38 (2012) 6557–6563, http://dx.doi.org/10.1016/j.ceramint.2012.05.039. [26] D.H. Oh, N.D. Hoa, D. Kim, Single-walled carbon nanotube thin film gas sensors controlled by diffusion, J. Nanosci. Nanotechnol. 11 (2011) 1601–1604, http://dx.doi.org/10.1166/jnn.2011.3318. [27] F. Hennrich, R. Krupke, S. Lebedkin, K. Arnold, R. Fischer, D.E. Resasco, M.M. Kappes, Raman spectroscopy of individual single-walled carbon nanotubes from various sources, J. Phys. Chem. B. 109 (2005) 10567–10573, http://dx. doi.org/10.1021/jp0441745. [28] K.D. Schierbaum, U.K. Kirner, J.F. Geiger, W. Göpel, Schottky-barrier and conductivity gas sensors based upon Pd/SnO2 and Pt/TiO2 , Sens. Actuators B 4 (1991) 87–94, http://dx.doi.org/10.1016/0925-4005(91)80181-I. [31] J. Lee, H. Kim, S.-J. Kahng, G. Kim, Y.-W. Son, J. Ihm, H. Kato, Z.W. Wang, T. Okazaki, H. Shinohara, Y. Kuk, Bandgap modulation of carbon nanotubes by encapsulated metallofullerenes, Nature 415 (2002) 1005–1008, http://dx.doi. org/10.1038/4151005a. [32] S.H. Lee, T.W. Uhm, Y.G. You, S.W. Kim, S.H. Jhang, U. Dettlaff-Weglikowska, Y.W. Park, Hall effect in carbon nanotube thin films, Synth. Met. 198 (2014) 84–87, http://dx.doi.org/10.1016/j.synthmet.2014.10.001. [33] B.J. Van Zeghbroeck, Principles of Semiconductor Devices, University of Colorado at Baulder, 1999 http://ece-www.colorado.edu/∼bart/book/.

Biographies Quan Thi Minh Nguyet received Engineer in physics and MSc degree in materials science at Hanoi University of Science and Technology (HUST), in Vietnam, in 2007 and 2010, respectively. She is currently pursuing the PhD degree at International Training Institute for Materials Science ITIMS at HUST, where he is working on gas sensors based on hetero-junction between carbon nanotube and metal oxide nanowires.

1127

Nguyen Van Duy is currently working as a research lecture at International Training Institute for Material Science (ITIMS), Hanoi University of Science and Technology (HUST). He received PhD degree from the Department of Electrical and Electronics Engineering at Sungkyunkwan University, South Korea, in 2011. His current research interests include nanomaterials, nanofabrications, characterizations, and applications to electronic devices, gas sensors, and biosensors. Nguyen Thi Phuong received MSc degree in materials science at the International Training Institute for Materials Science (ITIMS), HUST, in Vietnam, in 2014. Currently, she is a lecture at Ha Nam College of Education, Vietnam. Nguyen Ngoc Trung received MSc and PhD degree in Electrical Engineering at St. Petersburg State Electrotechnical University “LETI”, Russia, in 1992 and 1995, respectively. From 1996 to 1998, he worked as researcher at International Training Institute for Materials. Since 1999, he has joined Department of Optics and Optoelectronics, School of Engineering Physics, Hanoi University of Science and Technology, and became Associate Professor in 2011. His research focuses on process fabrication and characterizations of materials and semiconductor devices and Renewable energy applications. Chu Manh Hung is a research lecturer at International Training Institute for Materials Science (ITIMS), Hanoi University of Science and Technology. He received Master’s degree in electronic materials engineering at Kyungpook National University in South Korea in 2011 and his PhD degree in Materials Science at ESRF-the European synchrotron and University of Grenoble Alpes in France 2014. He worked as a post-doctoral researcher at the Synchrotron SOLEIL in France from 2014–2016. He has joined ITIMS since early 2016. His research focuses on synthesis of nanomaterials for gas sensor applications as well as characterization of nanomaterials using synchrotron source based techniques. Nguyen Duc Hoa obtained his PhD degree in Materials Science and Engineering in 2009 at Chungnam National University in Korea. He awarded JSPS fellowship andconducted the research at National Institute for Materials Science (NIMS, Japan) from 2009 to 2011. His research activity has covered a wide range of nanostructuredmaterials from synthesis, fundamental, and applications. Currently, he is an Associate Professor at Hanoi University of Science and Technology, Vietnam. Nguyen Van Hieu is director and professor at International Training Institute for Materials Science (ITIMS), Hanoi University of Science and Technology. He received MSc. degree in materials science at ITIMS in 1997 and his PhD degree from the Faculty of Electrical Engineering at University of Twente in The Netherlands in 2004. He worked as a post-doctoral fellow at the Korea University from 2006 to 2007. He has joined ITIMS since 2004, and he became associate professor and full professor in 2009 and 2015, respectively. His research focuses on synthesis, characterizations of nanomaterials for gas sensor and biosensor applications. He has published more than 90 ISI papers and his H-index is about 23 from Google scholar.