CONTENTS
12.2 Material Synthesis and Device Fabrication
12.2.1 Material Synthesis and characterization
12.2.4 Surface Modifications of Sensor Samples
12.3.1 Sensitivity and Response Time
12.3.2 Specificity and Stability
12.4 Understanding the Basic Mechanism of the Chemiresistive Response
12.4.1 The Role of Surface Hydroxyl Groups On TiO2-B Nanowires to Response Explosives
12.4.1.1 Varying the Density of Surface Hydroxyl Groups through Plasma Treatment
12.4.2.1 Chemiresistive Response to Different Positional Isomers of Nitro-Compounds
12.4.2.2 Chemiresistive Response to Nitroanilines and Corresponding Nitrotoluenes
12.4.3.1 Sensing Response to Nitrotoluenes of Different Numbers of Nitro Groups
12.4.3.2 Comparison between Aniline and Nitrobenzene
Nearly every terrorist attack involves explosives. To protect society from the increasing terrorist threat, effective techniques to detect high-explosive materials commonly used in terrorist attacks, such as TNT (2,4,6-trinitrotoluene), RDX (cyclotrimethylene trinitramine), PETN (pentaerythritol tetranitrate), and picric acid (2,4,6-trinitrophenol), at trace levels in luggage, vehicles, mail, aircraft, and soils, are very necessary. However, trace explosives are known to be very difficult to detect due to several factors, including the physical state of the sample to be detected (solid, liquid, and gas), very low vapor pressure of explosives [1], complicating or inhibiting the detection due to degradation by-products of explosives, and interferences leading to false signal due to other chemicals in the environment and the lack of selectivity of the detection techniques. In recent years, governments and industries have made a lot of effort in improving existing detecting technologies as well as developing new methods that could allow high sensitivity, selectivity, small size, and low cost.
Techniques to detect concealed explosives can be classified into two categories [2]: bulk detection and trace detection. Bulk detection is typically based on x-ray and gamma ray imaging techniques [3]. Considering extremely low vapor pressure of some explosives, a detection device needs to be very sensitive. There are a wide variety of analytical techniques for trace explosive detection that have been developed in recent years. They include ion mobility spectrometry [4], gas chromatography–mass spectrometry [5], and optical techniques such as fluorescing quenching, Raman, and laser breakdown spectroscopy [6,7]. However, detection systems based on those techniques are usually bulky, expensive, and with high-power consumption. These factors limit their applications in explosives detection [8,9,10,11]. Trained animals like dogs are an extremely sensitive sniffer to detect specific explosives, but these animals only produce qualitative alarms rather than quantitative data, and animals require a lot of effort to maintain. There are fundamental limitations to those techniques in significantly reducing the size, weight, power consumption, and quantification of detecting trace explosives. The development of nanotechnology potentially provides a feasible solution for building substantially smaller, highly sensitive, and selective trace explosive detectors through the use of a variety of nanostructured materials. Nanomaterials also have a high surface-to-volume ratio. This favors the adsorption of gases on the sensor and can increase the sensitivity and response time of the device because the interaction between the analytes and the sensing material is stronger [12,13,14,15,16]. It was found that in a chip-size device with greatly reduced size, weight, and power consumption, both the sensitivity and response speed of the sensor based on the nanostructured materials have surpassed those of current technologies [17,18,19]. Nanostructures of metal-oxide semiconductors such as In2O3 and SnO2 are already widely used as a base material for commercial gas sensors for the detection of toxic (e.g., CO) or dangerous (e.g., CH4) gases, owing to the lower production costs, high sensitivity, and long-term stability [20]. The material characteristics and size effects of metal oxides have been well explored for sensor applications [21].
Compared to other wide-band-gap metal oxides, such as SnO2, Ga2O3, ZnO, and WO2, titania (TiO2) exhibits unique chemical and electrical characteristics, including superior photocatalytic properties and excellent chemical stability [22]. TiO2-B is Trace Explosive Sensor Based on Titanium Oxide-B Nanowires one of the crystal polymorphs of TiO2 [23]. It proves to be the least dense polymorph of TiO2, which has a relatively open structure with significant voids and continuous channels compared to the other titanium dioxide polymorphs [24,25,26]. TiO2-B as a good functional material promising for sensor applications was first suggested in Wang et al.’s paper [27]. Recently, our group reported that the TiO2-B nanowires exhibit sensitive and rapid chemiresistive responses to subtrace concentrations of nitroaromatic and nitroamine explosives (Figure 12.1) [14]. Detection limits below sub-ppb levels and response times below a second were observed. This chapter will present a summary of recent developments of TiO2-B nanowires as explosive sensors. The mechanism of TiO2-B nanowires to effectively detect explosive vapors will be discussed, and in the final section we will point out new potential improvements that can be developed to further increase the sensing properties of TiO2-B nanowires as explosive sensors.
FIGURE 12.1 Schematic of a test sample for the chemiresistive response to nitroaromatic and nitroamino explosives. (Reproduced from D.L. Wang et al., IEEE Sensors Journal, 1352–1358, 2011. Copyright © 2011 IEEE. With permission.)
12.2 MATERIAL SYNTHESIS AND DEVICE FABRICATION
This section describes nanowire synthesis, test sample fabrication, equipment, and procedures to test nanowire samples for chemiresistive responses.
12.2.1 MATERIAL SYNTHESIS AND CHARACTERIZATION
The TiO2-B nanotubes were synthesized with a hydrothermal method [28,29,30,31,32]. Typically, a suspension containing 0.5 g of commercial anatase TiO2 powder (J.T. Baker Chemical Co.) dispersed in 20 ml of 10 M NaOH aqueous solution was prepared as precursor. After vigorous stirring for 5 ~ 10 min, the suspension was transferred to a Teflon vessel and placed in a hermetically sealed autoclave and heated at 180°C for 32 h. The precipitate produced was washed with 0.1 M HCl several times and then with deionized (DI) water until the pH value reached 7. This treatment removes Na+ ions remaining in the titanate nanoproducts and results in the formation of H2TiO3 nanotubes. A following post-treatment at 450°C for 1 h was carried out to promote the phase transformation from H2TiO3 to TiO2-B [28]. The length of the nanowires could be controlled by choosing the ultrasonic treatment time to the suspension solution of TiO2 nanoparticles, as described in literature [30]. The diameter of the nanowires was determined by the temperature during hydrothermal growth.
A Joel JSM-7000F scanning electron microscope (SEM) was used to characterize the morphology of the TiO2(B) nanowires. An x-ray diffraction (XRD) pattern of as-synthesized TiO2 nanowires was obtained with a Bruker F8 Focus Power XRD to confirm the TiO2(B) crystal structure. Fourier transfer infrared (FTIR) spectra were taken using a Bruker Vector 33 spectrophotometer in the 500–4000 cm–1 range with a 0.6 cm–1 resolution, which provided information of chemical bonds on the TiO2 nanowires’ surface. The morphology of the nanowire films was studied with scanning electron microscopy (SEM) in Figure 12.2, and the films have a highly porous structure made of a three-dimensional (3D) mesh of randomly orientated and interconnected nanowires. The length of the nanowires ranges from 1.1 to 2.2 μm, and the diameter of the nanowires is 40–100 nm, adjustable through synthesis conditions. Electric contact pads made of titanium were deposited onto the thin film by sputtering. Titanium forms a good ohmic contact of low junction resistance with TiO2(B). Shown in Figure 12.3 are an energy-dispersive spectrum (EDS) and x-ray diffraction (XRD) of the nanowire film, revealing the composition of titanium oxide and the crystal structure of TiO2-B.
FIGURE 12.2 SEM image of the interconnected 3D mesh structure of a TiO2-B thin film. (Reproduced from D.L. Wang et al., IEEE Sensors Journal, 1352–1358, 2011. Copyright © 2011 IEEE. With permission.)
FIGURE 12.3 (a) EDS spectrum and (b) XRD pattern of synthesized TiO2-B nanowires. θ is the x-ray diffraction angle. The XRD pattern matches the TiO2-B pattern in literature [23, 31]. (Reproduced from D.L. Wang et al., IEEE Sensors Journal, 1352–1358, 2011. Copyright © 2011 IEEE. With permission.)
As-synthesized nanowires were dispersed in ethanol to form a suspension solution. This solution was then drop-casted on glass substrates and heated at 70°C to attain a thin film of nanowires about 10 μm in film thickness. To fabricate a sensor, as shown in Figure 12.1, patterned titanium electrodes were deposited through a shadow mask over the nanowire film by sputtering. The titanium electrodes have a circular shape and are 4 mm in diameter. The spacing between contacts is 1 cm for the convenience of shadow mask patterning and probing during testing. The nominal thickness of the titanium electrodes is 200 nm.
Vapors of equilibrium concentrations at room temperature were generated using glass beads coated with various explosive analytes (Inert Products, LLC) in a vapor generator based on the method described in [33]. The vapor concentrations were confirmed by HP 5797 gas chromatography–mass spectrometry (GC-MS). After changing an analyte, all the tubings in the downstream of the vapor generator were replaced and the vapor generator was operated for several days before the vapor was used for sensor testing. This is to ensure that impurities and water in the generated vapor are purged, and the adsorption and desorption of analyte molecules on the inside walls of the tubing are at equilibrium. As vapor left the generator with the heated analyte, the temperature of the vapor gradually dropped to room temperature and excess explosives condensed on the walls of the tubing. At the exit end of the tubing, air that contained saturated concentration of analyte vapor at room temperature was obtained.
To test the sample for response to explosive vapor, a solenoid valve alternated the gas flow to the sample between air that contained saturated analyte vapor and pure air without the vapor. The cycle time of the valve was a few seconds. The sensitivity and response time of the sensor were determined from the change of the resistance between the electrodes with a Keithley 617 electrometer while the valve that controlled the gas flow to the test sample was cycled. The resistance measurements were made at room temperature and in ambient air.
12.2.4 SURFACE MODIFICATIONS OF SENSOR SAMPLES
Hydrogen and oxygen plasma treatments of nanowires were carried out in a microwave plasma cleaner/etcher (Plasma-preen). The plasma treatments were made at pressure of 1 Torr and radio frequency (RF) power of 300 W for 5 min, when the effect of plasma treatments became saturated.
12.3.1 SENSITIVITY AND RESPONSE TIME
The sensing performance of nanowire film samples was tested in ambient air, except for tests of temperature effects. The response time τ (1/e time constant) is obtained by fitting the resistance change to an exponential function. The chemiresistive response is defined as S = (Rv – R0)/R0, where Rv is the resistance when the sample is exposed to the vapor, and R0 is the resistance of the fresh sample before it is exposed to any explosive vapor [34]. The resistance of the nanowires increases to Rv upon exposure to explosive vapor. After the sample is returned to fresh air, its resistance decreases and reaches a stabilized value Rs, which is usually 0 to 100% higher than R0 and is likely due to incomplete desorption of explosive molecules from the surface of nanowires. Subsequent switching cycles between vapor of the same concentration and pure air make the resistance vary between Rv and Rs, as shown in Figure 12.4. The recovery time of the resistance after the explosive vapor is replaced by pure air is almost the same as the response time to the vapor. The incomplete desorption mentioned above can be eliminated by heating the sample at 80 to 100°C for several minutes, and the resistance of the sample returns to R0.
FIGURE 12.4 Typical resistance change of a TiO2-B nanowire thin film in response to vapors. (a) 1 ppb of 2,4,6-trinitrotoluene (TNT). (b) 5 ppt of 1,3,5-trinitroperhydro-1,3,5-triazine (RDX) at ambient conditions. (Reproduced from D.L. Wang et al., IEEE Sensors Journal, 1352–1358, 2011. Copyright © 2011 IEEE. With permission.)
TABLE 12.1
Response of a TiO2-B Nanowire Thin Film to Explosives
Symbol |
Equilibrium Vapor Concentration |
Percentage Response (100 × (Rv – R0)/R0) |
Response Time (s) |
Molar Mass (g/mol) |
Nitrotoluene (NT) |
130 ppm |
55 |
0.57 |
132 |
2,4-Dinitrotoluene (DNT) |
100 ppb |
58 |
0.64 |
182.13 |
2,4,6-Trinitrotoluene (TNT) |
5 ppb |
57 |
1.67 |
227.13 |
RDX |
5 ppt |
50 |
2.35 |
222.12 |
Note: Response of a TiO2-B nanowire thin-film sample to room temperature equilibrium vapor of common high explosives. ppm = parts per million, ppb = parts per billion; ppt = parts per trillion.
Table 12.1 lists the percentage resistance change and response times of the TiO2-B nanowire thin film for vapor of representative explosive compounds at equilibrium concentrations, where the values of the equilibrium vapor concentration are referenced from literature [36]. Significant and fast changes in resistance have been observed with all major explosives, including cyclotrimethylene trinitramine (RDX), which has extremely low vapor pressure. The concentrations of TNT and RDX vapors were analyzed by HP 5797 GC-MS spectrum and compared with the results in literature [14,35]. Note that molecules of smaller mass produce faster response, because the response time is limited by the diffusion of the vapor molecules through the nanowire thin film, and smaller molecules permeate the 3D matrix of nanowires faster.
12.3.2 SPECIFICITY AND STABILITY
Since the detection of explosives based on TiO2-B nanowires was conducted at room temperature, the interaction between explosive molecules and TiO2-B nanowires should be quite different from previous semiconductor gas sensors, which typically operate at above 400°C. Explosive molecules are unlikely to be capable of producing significant and fast chemical reduction-oxidation reactions at room temperature to produce sensitivity below the ppt level (Figure 12.4) and a fast response on the order of a second. The synthetic TiO2-B nanowires have a relatively higher charge carrier transfer ability than anatase TiO2 [27], less compact structure, and higher level of oxygen vacancies due to Ti4+ ions. The Ti4+ ions can be coordinated by hydroxyl groups and form hydroxyl-terminated surfaces [37,38,39]. The surface hydroxyl groups can trap electrons and facilitate adsorption of explosive compounds via their nitro groups. This indicates that the hydroxyl groups on the TiO2-B nanowires’ surface play an important role in determining the explosive gas-sensing properties at room temperature, and further study is described in the following section. Previous studies have shown that titanium oxide is an n-type semiconductor due to oxygen vacancies [40,41]. It is also known that nitroaromatic compounds and most high explosive compounds are highly electronegative, meaning that they tend to attract electrons from other molecules through charge transfer interactions from nitro groups in explosives to hydroxyl groups on TiO2-B nanowires’ surface. When explosive molecules adsorb on the surface of n-type semiconductor nanowires, the explosive molecules can trap charge carriers via surface hydroxyl groups and create a carrier depletion region near the TiO2-B nanowires’ surface. This could explain the increase of the resistance when the nanowires are exposed to explosive vapors.
In order to determine whether the response is due to charge transport within individual wires or across junctions between connecting wires, test samples of nanowires of the same diameter but different lengths were fabricated and their test results compared. Test results showed that the length of the wires has no significant effect on the chemiresistive response. However, shorter wires respond to the vapor at a slower rate. The slower response can be attributed to the denser packing of shorter nanowires and the consequent slower permeation of the thin film by vapor. Since the test samples have the same separation between the electrodes, the average path length for an electron to travel from one electrode to the other electrode is largely independent of the length of nanowires. However, the average number of junctions between connecting nanowires in the path of the electron is strongly dependent on the length of individual wires. In the film made of shorter wires, electrons need to pass through a greater number of junctions. The fact that the response is not affected by the length of nanowires indicates that the junction between nanowires does not play a significant role in the sensing process. In addition, different metal electrodes, including gold, aluminum copper, and titanium, have been used. Different types of metals did not change sensitivity and response time, indicating that the interface between the metal electrodes and nanowires did not play a significant role in the chemiresistive response. Test samples with titanium electrodes exhibited a more linear current-voltage (I-V) relationship, characteristic of a good ohmic contact between the metal electrode and nanowires. Based on these observations, it can be concluded that the electrical response is dominated by the charge transport within individual nanowires.
The surface depletion being the origin of the chemiresistive response is confirmed by the effect of the wire diameter on the response. Samples made of films with different wire diameters were prepared [42] and tested. The same film thickness and electrode spacing were used for all test samples. Wires of 50 ± 10 nm in diameter produced a response of 30%, higher than the 22% response observed for wires of 100 ± 20 nm in diameter. This observation supports the surface depletion hypothesis because thinner wires are more susceptible to surface depletion [43]. The response time is found to be largely independent of the diameter of nanowires.
Because the band-gap of TiO2 is much greater than the thermal energy, the chemiresistive response of TiO2-B nanowires to explosive trace vapors is found to be reliable and insensitive to temperature. There is only a few percent decrease in (Rv- Rs), and a slightly faster response at 75°C over 25°C. Such stability is the key to reliable sensors for practical applications. Also, as indicated in Figure 12.5, although the resistant baseline has a drift from 0.6 GΩ to 1.5 GΩ, the resistance change between Rv and Rs is found to be highly consistent over a long-term test of 15,000 switching cycles. The material demonstrated the exceptional stability of sensing performance. The sample was at ambient temperature throughout the test.
FIGURE 12.5 Resistance change Rv-Rs over 15,000 continuous test cycles between vapor of 100 ppb of DNT and pure air. Each test cycle consists of 6 s of vapor and 6 s of air. (Reproduced from D.L. Wang et al., IEEE Sensors Journal, 1352–1358, 2011. Copyright © 2011 IEEE. With permission.)
FIGURE 12.6 The sensing response of TiO2-B nanowires to different chemical compounds. (Reproduced from D.L. Wang et al., IEEE Sensors Journal, 1352–1358, 2011. Copyright © 2011 IEEE. With permission.)
On the other hand, the nanowires are found to have good specificity (seen in Figure 12.6) and are not sensitive to chemicals that are unrelated to explosives that cause false positives to other explosive detectors, such as inert chemicals that have high nitrogen content (for example, urea).
12.4 UNDERSTANDING THE BASIC MECHANISM OF THE CHEMIRESISTIVE RESPONSE
According to the above results, the achievement of TiO2-B nanowires in detecting explosives is related to two common characteristics of the nitroaromatic and nitroamine explosive compounds: (1) high electronegativity due to strong oxidizing chemical groups and (2) strong tendency to adsorb on the surfaces of objects exposed to their vapor. Therefore to effectively detect these explosive compounds, it is desirable that the sensing material has a high surface-to-volume ratio as well as a large surface area for the adsorption of explosive molecules, and moreover, it must be able to facilitate a strong surface charge transfer interaction with the molecules. TiO2-B nanowires meet both conditions. Also, the Ti4+ ions in TiO2-B can be coordinated by hydroxyl groups and tend to have hydroxyl-terminated surfaces [38]. The surface hydroxyl groups could effectively interact with nitro groups in explosives at room temperature and facilitate TiO2-B surface physisorption and charge transfer interactions. In order to further study the role of surface hydroxyl groups in the chemiresistive response of explosives, a series of surface treatments to change the density of surface hydroxyl groups on TiO2-B nanowires have been carried out, and the effects of each surface treatment on the chemiresistive sensing response have been characterized. On the other hand, the surface hydroxyl groups mean that TiO2-B nano-wires have a polar surface [38,44,45,46,47]. A stronger dipole-dipole interaction between TiO2-B nanowires and polar analyte molecules can be another way to understand the sensing mechanism of TiO2-B nanowires being the explosives sensor.
12.4.1 THE ROLE OF SURFACE HYDROXYL GROUPS ON TIO2-B NANOWIRES TO RESPONSE EXPLOSIVES
As we mentioned above, the resistant baseline has a drift from 0.6 to 1.5 G in the long-term scan, indicating that TiO2-B nanowires’ chemiresistive responses to explosive vapors are affected by the humidity of the environment to which the nanowires are exposed [14]. This indicates that the hydroxyl groups on the surfaces of nanowires participate in the chemiresistive response [48], as it is known that humidity strongly affects the density of hydroxyl groups on the surface of metal oxides. In this section, a series of surface treatments to change the density of surface hydroxyl groups on TiO2-B nanowires have been carried out in order to show how the density of hydroxyl groups on the TiO2-B nanowires’ surface influences its chemiresistive response.
12.4.1.1 Varying the Density of Surface Hydroxyl Groups through Plasma Treatment
The surface hydroxyl groups on the TiO2-B nanowires and the charge transfer via the surface hydroxyl groups and nitro group in explosives were initially observed via a FTIR spectrum. Figure 12.7 is FTIR spectra of the TiO2 -B nanowires before and after exposure to 2,4-dinitrotulene (DNT) vapor. Three main bands were detected in FTIR, and they were consistent with the bands described by Vargas and Nunez [39]. Compared to the FTIR spectrum of the control sample, the broad band at 3370 cm–1 is due to surface hydroxyl groups (Ti-OH) in the TiO2-B thin film [45,46]. This band is evidence that TiO2-B nanowires’ surface is terminated with hydroxyl groups. The new band at 1625 cm–1 is attributed to the complex between nitro groups of DNT bonded with hydroxyl groups at the TiO2-B surface (-O-Ti-OH-N-O-) [39]. This band did not exist in neat thin films (samples that had not been exposed to DNT). The new band at 1625 cm–1 after adsorption of DNT vapor onto TiO2-B indicates a charge transfer pathway from the TiO2-B surface to nitro groups in explosives via hydroxyl groups. The bands at 2335 and 2364 cm–1 are from carbon dioxide and aromatic groups of DNT. Figure 12.8 shows the sensing responses of nanowires to explosive vapor, 2,4,6-trinitrotoluene (TNT), after oxygen and hydrogen plasma treatments, respectively. A higher response of TiO2-B nanowires to TNT after an oxygen plasma treatment was observed compares to the response of the untreated TiO2-B nanowires and the nanowires after a surface hydrogen plasma treatment. It is noticed that the hydrogen plasma treatment did not change the sensitivity significantly. This is probably due to the competing effects of surface cleaning and reduction of hydroxyl groups.
FIGURE 12.7 FITR spectrum of TiO2-B nanowires after DNT exposure. The background was recorded by as-synthesized TiO2-B nanowires. (Reproduced from D.L. Wang et al., Journal of Materials Chemistry, 7369–7373, 2011. Copyright © 2011 The Royal Society of Chemistry. With permission.)
The water wettability tests also confirmed that plasma treatments change the contact angles of water on the TiO2 surface, and therefore change the density of surface hydroxyl groups. Figure 12.9 shows the contact angles on the TiO2-B nanowires’ surfaces after different surface plasma treatments. It is evident that the contact angle increases when TiO2-B nanowires are treated with hydrogen plasma and decreases after oxygen plasma treatment, indicating a more hydrophilic TiO2-B nanowire surface after an oxygen plasma treatment. These results suggest that TiO2-B nanowires’ surface with an oxygen plasma treatment has a higher density of surface hydroxyl groups. This is because an oxygen-rich nanowire surface created by oxygen plasma could enhance the dissociative adsorption of water molecules in air and increase the density of the hydroxyl groups on the TiO2-B surface [48,49]. The plasma-induced hydroxyl groups (OH–), as additional surface defects, can trap electrons and facilitate the charge transfer between the titania nanowires and explosive compounds further [50].
FIGURE 12.8 Effects of surface plasma treatments on the sensitivities of TiO2-B nanowires to TNT vapor. (Reproduced from D.L. Wang et al., Journal of Materials Chemistry, 7369–7373, 2011. Copyright © 2011 The Royal Society of Chemistry. With permission.)
FIGURE 12.9 Contact angles of TiO2-B nanowire surfaces and water. (a) As-fabricated TiO2-B nanowire surface. (b) The nanowire surface after oxygen plasma treatment. (c) The nanowire surface after hydrogen plasma treatment. (Reproduced from D.L. Wang et al., Journal of Materials Chemistry, 7369–7373, 2011. Copyright © 2011 The Royal Society of Chemistry. With permission.)
12.4.1.2 Varying the Density of Surface Hydroxyl Groups through Self-Assembled Monolayer (SAM) and Water Treatment
The importance of surface hydroxyl groups to the chemiresistive response of TiO2-B nanowire was further confirmed through surface functionalization. Since surface functionalization of TiO2-B nanowires with a self-assembled monolayer of certain hydrophobic acids can reduce the density of surface hydroxyl groups, this treatment should result in lower sensitivity of TiO2-B nanowires to explosives if surface hydroxyl groups contribute to the sensitivity. Surface modification was carried out by immersing the nanowire thin films into stearic and benzoic acids, respectively, and rinsing with de-ionized water before drying the thin films in an oven. As indicated in Figure 12.10(a), as-fabricated TiO2-B nanowires exhibited a sensitivity of 38% to DNT. After surface modification by stearic acid, the sensitivity of the same sample dropped to 10%. A similar decrease in sensitivity was also observed after the nanowire surfaces were modified with a self-assembled hydrophobic monolayer of benzoic acid. To the contrary, sensing responses of a TiO2-B nanowire thin film with water treatment were designed to increase the density of surface hydroxyl groups. Water treatment was achieved by immersing the test sample into de-ionized water and letting it become completely dry in the air. As shown in Figure 12.10(b), the sensitivity of TiO2-B nanowire thin film to DNT vapor after the water treatment was almost two times higher than that before the water treatment. This increase in sensitivity is most likely due to water-treated TiO2-B nanowires having a higher density of surface hydroxyl groups of bonded water molecules [50,51], and hydroxyl groups enhance the sensing properties of the TiO2-B nanowires.
FIGURE 12.10 Sensitivities of TiO2-B nanowires to (a) DNT with surface modification and (b) DNT before and after water treatment of the surfaces. (Reproduced from D.L. Wang et al., Journal of Materials Chemistry, 7369–7373, 2011. Copyright © 2011 The Royal Society of Chemistry. With permission.)
According to the above results, it can be concluded that surface hydroxyl groups on TiO2-B nanowires indeed enhance the charge transfer interactions on the nanowire surfaces and increase the chemiresistive response to nitro-explosive compounds. By functionalizing the surface with a high density of hydroxyl groups strongly bonded to TiO2-B, one can achieve a high chemiresistive response to nitro-explosives and low sensitivity to air humidity.
12.4.2 THE RELATIONSHIP BETWEEN THE DIPOLAR STRENGTH OF THE ANALYTE MOLECULE AND SPEED OF SENSOR RESPONSE
In order to improve the performance of the sensor further, a series of experiments were designed to study how the molecular polarity and electron deficiency influence the charge transfer process between nitro-containing explosives and TiO2-B. Depending on their molecular structures with functional groups, two types of analytes have been chosen. The first is positional isomers of dinitrobenzene (DNB); these positional isomers have varying dipole moments but are similar in electron deficiency. The other series of compounds are nitroanilines (NAs) and nitrotoluenes (NTs), such as 4-nitroaniline and 4-nitrotoluene. Although NAs and NTs have similar molecular structures except for one functional group, their molecular polarities are quite different because the electron-rich amino group (-NH2) in the nitroanilines is a stronger electron donor than the methyl group (-CH3) in nitrotolune.
12.4.2.1 Chemiresistive Response to Different Positional Isomers of Nitro-Compounds
Table 12.2 lists the response time and sensitivity of TiO2-B nanowires to the positional isomers of dinitrobenzene (DNB). The relative position of the two nitro groups of dinitrobenzene determines the polarity of the molecule, which can be estimated theoretically by the density functional theory (DFT) [52]. As is evident in Table 12.2, the response time and the dipole moment of the analyte molecules are strongly correlated. Molecules with higher dipole moments exhibit faster responses. For example, 1,2-dinitrobenzene has the most asymmetric structure in terms of electron-withdrawing group position, and therefore it has the highest dipole moment, and produces the shortest response time of 1.54 s. In contrast, 1,4-dinitrobenzene has a symmetrical molecular structure, and its dipole moment is zero. As a result, its response time is the slowest (6.31 s). Another observation is that the response time is not completely correlated with the calculated LUMO level.
TABLE 12.2
Sensing Response to Positional Isomers of Dinitrobenzene
Note: The dipole moments and LUMO levels were calculated using the DFT method [52].
* P = vapor pressure; Ʈ = response time.
12.4.2.2 Chemiresistive Response to Nitroanilines and Corresponding Nitrotoluenes
Another factor to affect the molecular dipole moment μ is the type of functional groups in analytes. In Table 12.3 a series of nitroanilines and corresponding nitrotoluenes are compared. Nitroanilines have an amino group (-NH2) and a nitro group, while in nitrotoluenes the amino group is replaced by a methyl group (-CH3). Both amino and methyl groups are electron donating; however, the amino group (-NH2) is more electropositive than the methyl group (-CH3). The nitro group is electronegative; therefore the dipole moment of a nitroaniline is greater than the dipole moment of its corresponding nitrotoluene [54,55]. Once again, it is observed that the response times of more polar NAs are faster than their corresponding NTs, consistent with the trend shown in Table 12.2.
The fact that the response time is dominated by the polarity of the sensing molecules supports previous arguments that TiO2-B nanowires have a polar surface. Since a polar surface facilitates a stronger dipole-dipole interaction between surface dipoles and more polar analyte molecules, it can facilitate surface adsorption and result in a faster response.
TABLE 12.3
Comparison of Chemiresistive Responses to Nitroanilines and Corresponding Nitrotoluenes
* P = vapor pressure; Ʈ = response time.
12.4.3 THE RELATIONSHIP BETWEEN ELECTRONEGATIVITY OF THE ANALYTE MOLECULE AND THE LEVEL OF SENSOR RESPONSE
In order to study the effect of electron deficiency of analytes, we have measured and compared chemiresistive responses of the TiO2-B nanowire films with nitrotoluenes with an increasing number of nitro groups, 4-nitrotoluene (NT), 2,4-dinitrotoluene (DNT), and 2,4,6-trinitrotoluene (TNT). By comparing the responses of analytes functionalized with electropositive amino groups and electronegative nitro groups, the surface charge transfer interaction, and the consequent depletion of surface electron charge carrier density, is proposed as the origin of the chemiresistive effect of TiO2-B nanowires. The electron deficiency of nitroaromatic compounds varies with the number of nitro groups in a molecule. For example, in the nitrotoluene family, the electron deficiency increases (indicated by decreases in their LUMO levels) from 4-nitrotoluene (NT) to 2,4-dinitrotoluene (DNT) and 2,4,6-trinitrotoluene (TNT). TNT has the most number of electron-withdrawing nitro groups, and it leads to the lowest LUMO level. TNT is more electron deficient than both NT and DNT and tends to cause a stronger charge transfer interaction with the TiO2-B nanowires.
12.4.3.1 Sensing Response to Nitrotoluenes of Different Numbers of Nitro Groups
The results in Table 12.4 demonstrate this trend between LUMO levels and sensitivity. As shown in Table 12.4, although the vapor pressure of TNT is about 1000 times lower than that of DNT, TNT vapor can still produce the same level of sensitivity as DNT vapor of much higher concentration. This means that TNT can induce stronger charge transfer interactions with the TiO2-B surface than DNT. Amazingly, the sensitivity to TNT is even higher than the sensitivity to NT despite the fact that the vapor pressure of NT is almost 105 higher than TNT. From these results, it is clear that the sensitivity is strongly dependent on the number of nitro groups and thus electron deficiency of analyte.
TABLE 12.4
Comparison of Chemiresistive Responses to Nitroanilines and Corresponding Nitrotoluenes
* P = vapor pressure; Ʈ = response time.
TABLE 12.5
Comparison of Chemiresistive Responses to Nitroanilines and Corresponding Nitrotoluenes
* P = vapor pressure; Ʈ = response time.
12.4.3.2 Comparison between Aniline and Nitrobenzene
A comparison between the effect of aniline, nitrobenzene (NB), and 4-nitroanline (4-NA) on the nanowires in Table 12.5 reveals the mechanistic details of the chemiresistive effect at the interface due to nitro-compounds. As described before, aniline is an electron-rich compound (with a LUMO level of +1.44 eV). Replacing the amino group with an electron-withdrawing nitro group results in electron-deficient nitrobenzene (with a LUMO level of –0.36 eV). The electron-rich aniline causes a decrease of resistance (S = −23%). In contrast, the electron-deficient nitrobenzene causes an increase of resistance (S = +25%). In the case that there is both an amino and a nitro group linked to the aromatic ring, as in 4-nitroaniline, the resistance still increases. Also, the response time (26.7 s) of aniline is much longer than that of nitrobenzene (2.83 s), although they have the similar dipole moment. All those suggest the interaction between the nitro group and TiO2-B prevails over the interaction between the amino group and TiO2-B. The trend can be explained by the orientation of permanent dipole moments of analytes relative to the hydroxylated surface dipoles of TiO2-B nanowires [42]. On the other hand, like most wide-band-gap metal-oxide semiconductors, TiO2-B is an n-type semiconductor with electrons as the majority charge carrier. When electron-deficient nitrobenzene is adsorbed onto TiO2-B nanowires, its electron-withdrawing nitro groups likely create localized surface states that trap TiO2-B to cause surface depletion of electron carriers and result in the decreasing conductivity and faster response. The observations from this study show that the rate of surface adsorption is dominated by the molecular polarity, and the degree of the charge transfer interaction is dominated by the electron deficiency of analytes.
In summary, a large and fast increase of the electrical resistivity of TiO2-B nanowires in response to subtrace vapors of nitroaromatic and nitroamine explosive compounds at room temperature has been observed. Experimental results indicate that the response originates from a depletion of electron carriers by the surface states produced by adsorbed molecules of electronegative explosive compounds via hydroxyl groups on TiO2-B nanowires’ surface. The results of FTIR spectrum, surface plasma treatments, and surface modification have consistently shown that surface hydroxyl groups on TiO2-B nanowires are important functional groups in modulating the sensor properties of TiO2-B nanowires. A stronger response was observed after increasing the density of surface hydroxyl groups. These observations indicate that surface hydroxyl groups provide a major charge transfer pathway between nitro groups in nitroaromatic explosives and TiO2-B nanowires and cause the chemiresistive response. Also, how the molecular polarity and electron deficiency influence the charge transfer process between nitroaromatic explosives and TiO2-B has been studied. The results suggest that the sensing response time is dominated by the dipole moment of the analyte molecules. The sensitivity is dominated by the electron deficiency of analytes. Charge transfer interaction between the analyte and TiO2-B is the major factor of the chemiresistive effect in TiO2-B nanowires, and electron-deficient and electron-rich chemical compounds cause opposite chemiresistive change. In nitro-containing compounds, nitro groups have strong interactions with TiO2-B, and the interaction between nitro groups and TiO2-B prevails over that of amino groups.
TiO2-B nanowires can be chemically synthesized with a low-cost and high-yield hydrothermal method, and explosive sensors based on TiO2-B nanowires are compatible with standard microelectronic manufacturing processes. For these reasons this technology is suitable for the mass production of low-cost microelectronic sensors and has excellent potential for practical application.
This work was supported by Office of Naval Research Grants N00014–05–1-0843 and N00014–09–1-0706, and NSF Center on Materials and Devices for Information Technology Research (CMDITR) Grant DMR-0120967. Danling Wang was supported in part by University of Washington WRF-APL Fellowship.
1. B.C. Dionne, D.P. Roundbehler, E.K. Achter, J.R. Hobbs, and D.H. Fine. Vapor pressure of explosives. Journal of Energetic Materials, 4, 447–472, 1986.
2. A.M. Jimenez and M.J. Navas. Chemiluminescence detection systems for the analysis of explosives. Journal of Hazardous Material, 106, 1–8, 2004.
3. J.E. McFee and A.A. Faust. Nuclear methods for explosive detection. SPIE, Newsroom, 2011. DOI: 10.1117/2.1201108.003761.
4. K.M. Roscioli, E. Davis, W.F. Siems, A. Mariano, W.S. Su, S.K. Guharay, and H.H. Hill. Modular ion mobility spectrometer for explosives detection using corona ionization. Analytical Chemistry, 83, 5965–5971, 2011.
5. R. Batlle, H. Carlsson, P. Tollback, A. Colmsjo, and C. Crescenzi. Enhanced detection of nitroaromatic explosive vapors combining solid-phase extraction-air sampling, supercritical fluid extraction, and large-volume injection-GC. Analytical Chemistry, 75, 3137–3144, 2003.
6. J.S. Yang and T.M. Swager. Porous shape persistent fluorescent polymer films: An approach to TNT sensory materials. Journal of American Chemical Society, 120, 5321–5322, 1998.
7. J.S. Yang and T.M. Swager. Fluorescent porous polymer films as TNT chemosensors: Electronic and structural effects. Journal of American Chemical Society, 120, 11864–11873, 1998.
8. Q.L. Fang, J.L. Geng, B.H. Liu, D.M. Gao, F. Li, Z.Y. Wang, G.J. Guan, and Z.P. Zhang. Inverted opal fluorescent film chemosensor for the detection of explosive nitroaromatic vapors through fluorescence resonance energy transfer. Chemistry—A European Journal, 15, 11507–11514, 2009.
9. F. Rock, N. Barsan, and U. Weimar. Electronic nose: Current status and future trends. Chemical Reviews, 108, 705–725, 2008.
10. R.G. Ewing and C.J. Miller. Detection of volatile vapors emitted from explosives with a handheld ion mobility spectrometer. Field Analytical Chemistry and Technology, 5, 215–221, 2001.
11. H.B. Liu, Y.Q. Chen, G.J. Bastiaans, and X.C. Zhang. Detection and identification of explosive RDX by THz diffuse reflection spectroscopy. Optical Express, 14, 415–423, 2006.
12. A.D. Aguilar, E.S. Forzani, M. Leright, F. Tsow, A. Cagan, R.A. Iglesias, L.A. Nagahara, I. Amlani, R. Tsui, and N.J. Tao. A hybrid nanosensor for TNT vapor detection. Nano Letters, 10, 380–384, 2010.
13. Y.H. Gui, C.S. Xie, J.Q. Xu, and G.Q. Wang. Detection and discrimination of low concentration explosives using MOS nanoparticle sensors. Journal of Hazardous Material, 164, 1030–1035, 2009.
14. D.L. Wang, Q.F. Zhang, G.Z. Cao, and A.T. Chen. Room-temperature chemiresistive effect of TiO2-B nanowires to nitroaromatic and nitroamine explosives. IEEE Sensors Journal, 11, 1352–1358, 2011.
15. O.K. Varghese and C.A. Grimes. Metal oxide nanoarchitectures for environmental sensing. Journal of Nanoscience and Nanotechnology, 3, 277–293, 2003.
16. A. Rothschild and Y. Komem. The effect of grain size on the sensitivity of nanocrystalline metal-oxide gas sensors. Journal of Applied Physics, 95, 6374–6380, 2004.
17. K. Shiraishi, T. Sanji, and M. Tanaka. Trace detection of explosive particulates with a phosphole oxide. ACS Applied Materials and Interfaces, 1, 1379–1382, 2009.
18. S.J. Toal and W.C. Trogler. Polymer sensors for nitroaromatic explosives detection. Journal of Materials Chemistry, 16, 2871–2883, 2006.
19. J.S. Yang and T.M. Swager. Porous shape persistent fluorescent polymer films: An approach to TNT sensory materials. Journal of the American Chemical Society, 120, 5321–5322, 1998.
20. G. Korotcenkov. Metal oxides for solid-state gas sensors: What determines our choice? Materials Science and Engineering B, 139, 1–23, 2007.
21. N. Barsan, D. Koziej, and U. Weimar. Metal oxide-based gas sensor research: How to? Sensors and Actuators B, 121, 18–35, 2007.
22. F. Millot, M.G. Blanchin, R. Tetot, J.F. Marucco, B. Poumellec, C. Picard, and B. Touzelin. High temperature nonstoichiometric rutile TiO2-x. Progress in Solid State Chemistry, 17, 263–293, 1987.
23. R. Marchand, L. Brohan, and M. Tournoux. TiO2(B): A new form of titanium-dioxide and the potassium octatitanate K2Ti8O17. Materials Research Bulletin, 15, 1129–1133, 1980.
24. G. Nuspl, K. Yoshizawa, and T. Yamabe. Lithium intercalation in TiO2 modifications. Journal of Materials Chemistry, 7, 2529–2536, 1997.
25. A.R. Armstrong, G. Armstrong, J. Canales, and P.G. Bruce. TiO2-B nanowires. Angewandte Chemie International Edition, 43, 2286–2288, 2004.
26. T.P. Feist and P.K. Davies. The soft chemical synthesis of TiO2(B) from layered titanates. Journal of Solid State Chemistry, 101, 275–295, 1992.
27. G. Wang, Q. Wang, W. Lu, and J.H. Li. Photoelectrichemical study on charge transfer properties of TiO2-B nanowires with an application as humidity sensors. Journal of Physical Chemistry B, 110, 22029–22034, 2006.
28. J. Jitputti, S. Pavasupree, Y. Suzuki, et al. Synthesis of TiO2 nanotubes and its photo-catalytic activity for H-2 evolution. Japanese Journal of Applied Physics, 47, 751–756, 2008.
29. S. Pavasupree, Y. Suzuki, S. Yoshikawa, and R. Kawahata. Synthesis of titanate, TiO2 (B), and anatase TiO2 nanofibers from natural rutile sand. Journal of Solid State Chemistry, 128, 3110–3116, 2005.
30. N. Viriya-Empikul, N. Sano, T. Charinpanitkul, T. Kikuchi, and W. Tanthapanichakoon. A step towards length control of titanate nanotubes using hydrothermal reaction with sonication pretreatment. Nanotechnology, 19, 035601, 2008.
31. Y.Q. Wang, G.Q. Hu, X.F. Duan, H.L. Sun, and Q.K. Xue. Microstructure and formation mechanism of titanium dioxide nanotubes. Chemical Physics Letters, 365, 427–431, 2002.
32. R. Yoshida, Y. Suzuki, and S. Yoshikawa. Syntheses of TiO2(B) nanowires and TiO2 anatase nanowires by hydrothermal and post-heat treatments. Journal of Solid State Chemistry, 178, 2179–2185, 2005.
33. P.A. Pella. Generator for producing trace vapor concentrations of 2,4,6-trinitrotoluene, 2,4-dinitrotoluene, and ethylene-glycol dinitrate for calibrating explosives vapor detectors. Analytical Chemistry, 48, 1632–1637, 1976.
34. S. Ahlers, G. Muller, and T. Doll. A rate equation approach to the gas sensitivity of thin film metal oxide materials. Sensors and Actuators B—Chemical, 107, 587–599, 2005.
35. M.-W. Ahn, K.-S. Park, J.-H. Heo, J.-G. Park, and D.-W. Kim. Gas sensing properties of defect-controlled ZnO-nanowire gas sensor. Applied Physics Letters, 93, 263103–263106, 2008.
36. NRC, Committee on the Review of Existing and Potential Standoff Explosives Detection Techniques. Existing and Potential Standoff Explosives Detection Techniques. National Academies Press, Washington, DC, 2004.
37. G. Munuera, V. Rivesarnau, and A. Saucedo. Photo-adsorption and photo-desorption of oxygen on highly hydroxylated TiO2 surfaces. 1. Role of hydroxyl-groups in photo-adsorption. Journal of the Chemical Society—Faraday Transactions I, 75, 736–747, 1979.
38. R. Marchand, L. Brohan, and M. Tournoux. TiO2(B): A new form of titanium-dioxide and the potassium octatitanate K2Ti8O17. Materials Research Bulletin, 15, 1129–1133, 1980.
39. R. Vargas and O. Nunez. Hydrogen bond interactions at the TiO2 surface: Their contribution to the pH dependent photo-catalytic degradation of p-nitrophenol. Journal of Molecular Catalysis A—Chemical, 300, 65–71, 2009.
40. M.D. Earle. The electrical conductivity of titanium dioxide. Physical Review, 61, 56–62, 1942.
41. M. Gratzel. Photoelectrochemical cells. Nature, 414, 338–344, 2001.
42. M.H. Seo, M. Yuasa, T. Kida, J.-S. Huh, K. Shimanoe, and N. Yamazoe. Gas sensing characteristics and porosity control of nanostructured films composed of TiO2 nano-tubes. Sensors and Actuators B—Chemical, 137, 513–520, 2009.
43. A. Rothschild and Y. Komem. The effect of grain size on the sensitivity of nanocrystalline metal-oxide gas sensors. Journal of Applied Physics, 95, 6374–6380, 2004.
44. O.A. El Seoud, A.R. Ramadan, B.M. Sato, and P.A.R. Pires. Surface properties of calcinated titanum dioxide probed by solvatochromic indicators: Relevance to catalytic application. Journal of Physical Chemistry C, 114, 10436–10443, 2010.
45. K.S. Finnie, D.J. Cassidy, J.R. Bartlett, and J.L. Woolfrey. IR spectroscopy of surface water and hydroxyl species on nanocrystalline TiO2 films. Langmuir, 17, 816–820, 2001.
46. A.S. Vuk, R. Jese, B. Orel, and G. Drazic. The effect of surface hydroxyl groups on the adsorption properties of nanocrystalline TiO2 films. International Journal of Photoenergy, 7, 163–168, 2005.
47. R. Muller, H.K. Kammler, K. Wegner, and S.E. Prtsinis. OH surface density of SiO2 and TiO2 by thermogravimetric analysis. Langmuir, 19, 160–165, 2003.
48. D.L. Wang, A.T. Chen, S.-H. Jang, H.-L. Yip, and A.K.-Y. Jen. Sensitivity of titania(B) nanowires to nitroaromatic and nitroamino explosives at room temperature via surface hydroxyl groups. Journal of Materials Chemistry, 21, 7369–7373, 2011.
49. B. Meyer, D. Marx, O. Dulub, U. Diebold, M. Kunat, D. Langenberg, and C. Woll. Partial dissociation of water leads to stable superstructures on the surface of zinc oxide. Angewandte Chemie International Edition, 43, 6641–6645, 2004.
50. H.W. Ra, R. Khan, J.T. Kim, B.R. Kang, K.H. Bai, and Y.H. Im. Effects of surface modification of the individual ZnO nanowire with oxygen plasma treatment. Materials Letters, 63, 2516–2519, 2009.
51. J.M. Pan, B.L. Maschhoff, U. Diebold, and T.E. Madey. Interaction of water, oxygen, and hydrogen with TiO2(110) surfaces having different defect densities. Journal of Vacuum Sicence and Technology A, 10, 2470–2476, 1992.
52. R.G. Parr. Density functional theory. Annual Review of Physical Chemistry, 34, 631–656, 1983.
53. D. Ferro, V. Piacente, R. Gigli, and G. D’Ascenzo. Determination of the vapour pressures of o-, m-, and p-dinitrobenzene by the torsion-effusion method. Journal of Chemical Thermodynamics, 8, 1137–1143, 1976.
54. J.E. Abbott, X.Z. Peng, and W. Kong. Symmetry properties of electronically excited states of nitroaromatic compounds. Journal of Chemical Physics, 117, 8670–8675, 2002.
55. T. Tanaka, A. Nakajima, A. Watanabe, T. Ohno, and Y. Ozaki. Surface-enhanced Raman scattering spectroscopy and density functional theory calculation studies on adsorption of o-, m-, and p-nitroaniline on silver and gold colloid. Journal of Molecular Structure, 661–662, 2003.
56. K. Aim. Saturated vapor pressure measurements on isomeric monoitrotoluenes at temperatures between 380 and 460 K. Journal of Chemical and Engineering Data, 39, 591–594, 1994.
57. http://actrav.itcilo.org/actrav-english/telearn/osh/ic/100016.htm.
58. H. Carlsson, G. Robertsson, and A. Colmsjo. Response mechanisms of thermionic detectors with enhanced nitrogen selectivity. Analytical Chemistry, 73, 5698–5703, 2001.
59. A. Kawski, B. Kuklinski, and P. Bojarski. Dipole moment of aniline in the excited S1 state from thermochromic effect on electronic spectra. Chemical Physics Letters, 415, 251–255, 2005.