You are on page 1of 23

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.

3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 1

BlackLining

CHAPTER 30

Gas Sensors Based on One-Dimensional Nanostructures


C. H. Xu,1 S. Q. Shi,1 C. Surya2
1 2

Department of Mechanic Engineering, The Hong Kong Polytechnic University, Hung Hom, Hong Kong Department of Electronic and Information Engineering, The Hong Kong Polytechnic University, Hung Hom, Hong Kong

CONTENTS
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1. Why are 1D Nanostructure Gas Sensors Needed? . . . . . . . . . . . . . . . 1.2. What is a 1D Nanostructure Gas Sensor? . . . . . . . . . . . . . . . . . . . . . . 1.3. Performance Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1D Nanostructure Gas Sensors Based on Electrotransducers . . . . . . . . . . . . 2.1. Field Effect Transistors (FET) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Electro Gas Sensors Manufactured by Microelectromechanical System (MEMS) Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3. Resistive Gas Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1D Nanostructure Sensors Based on Optical Transducers . . . . . . . . . . . . . . 3.1. Chemiluminescence (CL) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2. Photoluminescence (PL) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1D Nanostructure Sensors Based on Quartz Crystal Microbalance Sensors . . Mechanism of 1D Nanostructure Gas Sensors . . . . . . . . . . . . . . . . . . . . . . 5.1. Reaction Between Crystal Defects and Gases . . . . . . . . . . . . . . . . . . . 5.2. Mechanism for 1D Nanostructure Gas Sensors . . . . . . . . . . . . . . . . . 5.3. Electrical Sensitivity of 1D Nanostructure Materials to Manufacture History and Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4. Temperature Dependence of Gas Sensors . . . . . . . . . . . . . . . . . . . . . 5.5. SensitivityConcentration Curves . . . . . . . . . . . . . . . . . . . . . . . . . . . 1D Nanostructure Sensors Based on Composite Sensing Elements . . . . . . . . 6.1. Double Metal Oxide Gas Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2. 1D Nanostructure Composite Metal Oxide Gas Sensors . . . . . . . . . . . Summary and Further Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2 2 3 3 3 6 7 9 9 9 10 12 12 13 15 15 16 16 16 16 20 20

2.

3.

4. 5.

6.

7.

ISBN: 1-58883-114-0 Copyright 2008 by American Scientific Publishers All rights of reproduction in any form reserved.

Handbook of Nanoceramics and Their Based Nanodevices Edited by Tseung-Yuen Tseng and Hari Singh Nalwa Volume XX: Pages (122)

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 2

BlackLining

Gas Sensors Based on One-Dimensional Nanostructures

1. INTRODUCTION
1.1. Why are 1D Nanostructure Gas Sensors Needed?
Modern industrialized society has brought a series of problems to our world. Increasing industrialization makes it necessary to constantly monitor and control pollution in the environment, chemical factories, food processing plants, laboratories, hospitals, and technical installations in general. Nitrogen dioxide (NO2) gas, for example, is one of the most dangerous air pollutants; it contributes to the formation of ozone and acid rain. A safe concentration of NO2 in the air is set as 53 ppb by the U.S. Environmental Protection Agency (EPA). Concentrations of NO2 in the air greater than this limit may cause incidence of acute respiratory illness [1]. Solid-state gas sensors play important roles in monitoring pollution in the areas mentioned previously [2, 3]. Up to now, semiconducting ceramics, especially metal oxide sensors, have been widely deployed because of their small dimensions, low cost, and high compatibility with microelectronic processing. The most important type of gas sensors work based on the change in the electrical conductivity of metal oxides due to the adsorption of gas molecules on the surface of the materials. The semiconductor metal oxides contain lattice defects due to the excess or deficiency of metal or oxygen in the lattice. The association of electrons with these defects following chemisorption allows a certain change of the electrical conductivity of the oxide. For polycrystalline ceramic or thin film devices, only a small fraction of the species adsorbed near the grain boundaries is active in modifying the electrical transport properties. This results in the low sensitivity of the device because of the limited surface-to-volume ratio [4]. Moreover, most thin film devices operate at high temperatures (200600C) to achieve enhanced chemical reactivity between the sensor materials and the surrounding gases [5]. Reported sensing limitations of metal oxide ceramics or thin film sensors are about 1 ppm or even greater [68], which cannot satisfy the requirements of detecting low concentration gas standard, such as 53 ppb for NO2. In recent years, one-dimensional (1D) nanostructure materials such as nanotubes, nanowires, or nanobelts [9] have become an important field of research in both science and technology. To date, 1D nanostructures have been produced using various synthetic techniques such as vaporphase evaporation [10], chemical vapor deposition [11], sol-gel [12], template-based method [13], arc discharge [14], laser ablation [15], solution [16], and thermal oxidation [1719]. One-dimensional nanostructures have attracted much attention because of their great potential in practical applications. In particular, such structures are deemed especially useful in solid-state gas sensors, with great potential for overcoming the limitations encountered by metal oxide ceramics or thin film devices. Studies of the applications of 1D nanostructure materials in gas sensing have recently emerged. Gas sensors with a mesh structure of nanobelts [20], ultraviolet (UV)-activated single nanowire sensors in bare [21] and single nanowire or multinanowires in field effect configuration [22], have been reported. Compared with metal oxide ceramics or film ceramic sensors,

Gas Sensor

Gas signal sensing element transducer signal processor

Figure 1. Schematic diagram of a gas sensor.

1D nanostructure gas sensors offer high sensitivity, low detected limitations, and low operating temperatures. For example, it is reported that 5 ppb NO2 can be detected using In2O3 nanowire gas sensors [23]. Moreover, the electric properties of materials in nanoscale under different gases can be detected, which offers information on sensing mechanisms.

1.2. What is a 1D Nanostructure Gas Sensor?


A gas sensor has two main parts: a gas sensing element and a transducer. When a gas molecule is adsorbed on the surface of the gas sensing element, it produces a response, which is translated into a measurable signal by a transducer. The change in the signal can be detected through a signal processor. Figure 1 shows the schematic diagram of a gas sensor. A 1D nanostructure gas sensor uses 1D nanostructure materials as sensing elements. One type of the 1D structure sensing element is based on ceramics such as zinc oxide (ZnO), tin oxide (SnO2), indium oxide (In2O3), and titanium dioxide, or titania (TiO2). ZnO is an important n-type semiconductor with a wide band gap energy (3.3 eV) and a large exciting binding energy (60 meV) at room temperature [24]. SnO2 is an n-type semiconductor with a roomtemperature band gap of 3.6 eV and high achievable carrier concentration (up to 6 3 1020 cm3) [25]. In2O3 is known to be an n-type semiconductor in its nonstoichiometric form due to oxygen vacancy doping [26]. The gas has to be intimately connected to the sensing elements. Gas adsorption on the surface occurs because the atoms or ions at the surface of the solid cannot fully satisfy their valence or coordination requirements [27]. There are two types of adsorption: (a) physisorption, the weak attraction of gas on a solid due to van der Waals forces, and (b) chemisorption, in which the gas becomes adsorbed on the solid with high surface energy through an exchange of electrons with the surface, forming chemical bonds. The heat of chemisorption is higher that of the physisorption. Generally, the heat of chemisorption is in the range of 15 to 200 kcal/ mol, whereas physisorption is only up to 15 kcal/mol [27]. At present, most 1D nanostructure gas sensors studied are constructed around electrotransducers. Some other 1D nanostructure gas sensors have been reported whose operation is based on changes in the photoluminescence spectroscopy and the mass of the sensing element, which is accurately measured using a quartz crystal microbalance.

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 3

BlackLining

Gas Sensors Based on One-Dimensional Nanostructures

1.3. Performance Factors


The most important figures of merit for gas sensors are sensitivity, response time, minimum detectable gas concentration, repeatability (recovery method, recovery time), and selectivity. Sensitivity: The ratio of measured parameters to sensitivity is set to <1. The measured parameters after (or before) exposure to detected gas is divided by that before (or after) exposure, depending on the relative values of two parameters. Response time: The time duration for the parameter to change by a certain percentage of its original value. The minimum detectable gas concentration: The lowest gas concentration that can be detected. Repeatability: The gas sensors can be reused; this includes the recovery method and time. Recovery method: The method used to allow the sensor to return to the original state after sensing gases. Recovery time: The time duration for the parameter to return to its original value. Selectivity: Whether the sensor can give a significant signal for some specific gases.
Figure 2. (a) E-beam lithography fabricated field-effect transistor using a single ZnO nanoribbon, and (b) the cross-sectional structure and working principle of the FET device. Reprinted with permission from [31], Z. L. Wang, Adv. Mater. 15, 432 (2003). 2003, Wiley-VCH.

2.1.2. Physical Characteristics of FETs

2. 1D NANOSTRUCTURE GAS SENSORS BASED ON ELECTROTRANSDUCERS


The most promising area for nanotechnology is metal oxide 1D nanostructures as gas sensors. Their large surfaceto-volume ratio and their function as quasi-1D conductive elements simultaneously contribute to high sensitivity. The change in the electrical conductivity caused by chemisorption of gas molecules on the surface of 1D nanostructure metal oxides is measured by electrotransducers. The main structures of the electrotransducers are field effect transistors, resistive gas microsensors, and resistive gas sensors.

Typical drain-source current IDS vs drain-source voltage VDS characteristics are shown in Figure 3(a). The transfer characteristics of IDS versus gate voltage VG curves at VDS 2 V are shown in Figure 3(b). The following physical properties of FETs can be obtained from these curves. Conductivity of nanowire: Calculation based on the IDSVDS curve in Figure 3(a) and nanowire size. Leakage current: The current IDS at VG 20 to 10V in Figure 3(b). N-type semiconductor: As VG increased, IDS in Figure 3(a) increased. Threshold voltage Vth: The VG to extrapolate a linear region (dotted line) in Figure 3(b) to VG axis. Current switching ratio: The large and small current on the curve in Figure 3(b). The gate capacitance CG: Estimation from Eq. (1) [34]: CG 2pee0 L ln2h=r 1

2.1. Field Effect Transistors (FET) 2.1.1. Sensor Structure and Fabrication
To fabricate 1D nanostructure FETs, 1D nanostructure materials must be produced by a reliable technique [28]. These 1D nanostructure materials in a solution such as isopropyl alcohol are then sonicated into a suspension. The suspension is deposited onto a degenerately doped silicon wafer covered with 100500 nm SiO2. Photolithography and metallization of the source and drain electrodes are achieved using Ti/Au (50/200 nm) [29, 30] Ti or Au [31, 33]. The back gate consists of a doped Si substrate or a doped Si substrate with a metallic layer [32]. The alternative way of accomplishing electrical contact to the 1D nanostructure materials is achieved by depositing the suspension directly on prefabricated electrodes [33]. The channel length between the two electrodes is 100 nm to 7 lm [33, 30]. Figure 2(a) shows a FET fabricated using a single ZnO nanoribbon. Its cross-sectional structure and working principle are given in Figure 2(b).

where r and L are the nanowire radius and the length of nanowire channel, e0 is permittivity in vacuum, and h and e are the thickness and the average dielectric constant of the device, respectively. Electron density ne per unit length of nanowire: From Eq. (2) [35]: ne CG jVth j eL 2

where e is the electronic charge.

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 4

BlackLining

Gas Sensors Based on One-Dimensional Nanostructures

Figure 3. (a) Typical characteristics of the IDSVDS curves at different gate voltages: from 7 to 3 V in 1 V steps as VDS varies. (b) The transfer characteristics with various VG at VDS 2 V. The transconductance was inferred from the dashed line. Reprinted with permission from [35], Q. H. Li et al., Appl. Phys. Lett. 85, 6389 (2004). 2004, American Institute of Physics. Figure 4. (a) IDSVDS curves measured before and after exposure to 100 ppm NO2 plus Ar. Inset: SEM image of a metal/In2O3-nanowire/ metal device. (b) The IDS VG curves before and after exposure to 100 ppm NO2 plus Ar with VDS 0.3 V. (c) Response time of the In2O3 nanowire FET to various NO2 concentrations. Reprinted with permission from [32], C. Li et al., Appl. Phys. Lett. 82, 1613 (2003). 2003, American Institute of Physics.

Transconductance gm: Calculation from the slope of the dashed line in Figure 3(b), according to Eq. (3) [35]: gm dIDS dVG 3

The mobility l of the transistor: Calculation from parameters in Figure 3(b), using Eq. (4) [34]: dIDS =dVg l CG =L2 VDS 4

2.1.3. Sensor Properties


The FET devices are mounted in a sealed chamber with electrical feedthrough and gas inlet/outlet during a measurement. Sensing experiments are carried out by monitoring the single nanowire conductance under different gas concentrations [36, 37]. As an example, the sensor properties of In2O3 nanowire FET to detect NO2 gas can be drawn from Figure 4 [32]. The system is usually pumped to vacuum first to clean the surface of the nanowires, and then the conductance of the nanowires is monitored while

Subthreshold swing S: [35] assuming completely depleted carriers (VG < Vth) [see Figure 3(b)], s dVG d lgjI DS j 5

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 5

BlackLining

Gas Sensors Based on One-Dimensional Nanostructures


Table 1. Summary of the sensing properties of 1D nanostructure FET sensors. Response time (s) 5 10 600 25 25 25 25 200 Vg/60 s 104 106 mPa UV [86] [41] [35] [33] [31] T (C) 25 25 Recovery method/time UV/30 s UV

Nanowire In2O3 In2O3 multiwires In2O3 ZnO ZnO ZnO SnO2

Gas NO2 NH3 NO2 NH3 NO2 NH3 O2 O2 Ethanol NO2 CO O2 Ethanol

LDC 0.5 ppm 200 ppm 0.02 ppm 0.005 ppm 10000 ppm 1 ppm 5000 ppm 920 Pa 5 mPa 250 ppm 0.5 ppm 5 ppm 5 3 104 torr 4.7V%

Sensitivity 106 105

Ref. [32] [23]

SnO2 SnO2

200 3000 70 250 15

200 >100

200 400

105 Ar

[36] [37]

flowing diluted NO2 (0.5100 ppm) in Ar. Figure 4(a) shows IDSVDS curves recorded before and after exposing the nanowire FET to 100 ppm NO2 plus Ar for 5 minutes at the gate voltage of 0 V. The shapes of IDSVDS curves recorded before and after the exposure are significantly different. The maximum change in IDS of the device occurs at VDS 0.3 V, at which the conductance of the In2O3 nanowire FET reduces by six orders of magnitude after the exposure. If the ratio of currents before and after exposure to NO2 gas is defined as sensitivity, the sensitivity of the FET is about 106 from Figure 4(a). After each exposure, the system was pumped to vacuum, followed by UV illumination to desorb the NO2 molecules. The device was fully recovered to its initial status immediately after the UV was turned on. The sensing properties of devices can also be studied by monitoring the current dependence on the gate bias. Figure 4(b) shows two IDSVG curves recorded before and after exposure to 100 ppm NO2 plus Ar with VDS 0.3 V. The threshold voltage shifted from 48 V before the exposure to 20 V after the exposure. The threshold voltage difference of 68 V before and after exposure can also be regarded as sensitivity to judge the sensor properties. The response to NO2 can be understood by considering the interaction between NO2 and the n-type doped In2O3 nanowires. Adsorption of oxidizing gases (such as NO2) reduces the number of free electrons in the nanowire and thus reduces the conductivity. Besides the sensitivity, another important parameter for gas sensors is the response time for the detected gas. Figure 4(c) shows the response of an In2O3 nanowire FET to 100 ppm NO2 plus Ar with VG 0 V and VDS 0.3 V. If the response time is defined as the time duration for the conductance to change by one order of magnitude, the response time for concentrations of 20, 2, and 0.5 ppm NO2 are 5 s, 5 min, and 1012 min, respectively. The lowest detectable concentration for NO2 is 0.5 ppm, as shown in Figure 4(c). A response time of 5 s for the In2O3 nanowire FET is significantly better than the response time of 50 s for thin-film-based semiconducting oxide sensors operating at elevated temperatures upon exposure to 100 ppm NO2 [38].

Besides the sensitivity, response time, the lowest detectable concentration of the gas, and recovery time discussed previously, the advantages of FET sensors are the changes of electrons in nanowires, based on the change in the physical properties in Section 2.1.2. For example, carrier concentration and mobility can change with various gases and at various gas concentrations. The estimation of the change of carrier contraction (DN) in a FET after the gas exposure can be calculated as in Eq. (6); DN CG jDVth j ep r2 L2 6

where DVth is the shift of the threshold voltage in Figure 4(b), CG is the capacitance of the nanowire with respect to the back gate, e is the electronic charge, and r and L are radius and length of the nanowire channel, respectively. Another advantage of FET type gas sensors is the possibility of the amplification of the signal through the control of the gate voltage as shown in Figure 3(a). Table 1 lists a summary of the sensing properties of 1D structure FET gas sensors. The parameters include the lowest detectable gas concentrations (LDC), sensitivity, and response time, which relate to measured gas concentrations, recovery methods, and measurement temperature (T).

2.1.4. Recovery Methods


Repeatability is another important requirement for a gas sensor. Gas is adsorbed on the surface of the nanowires during sensing. It is important for the gas sensor to return to its original state after sensing. Some methods to refresh the sensors have already been reported. UV illumination results in a jump in the conductance [39] by increasing the net carrier density and decreasing the depletion width. Conductance will decrease when UV illumination is turned off. The reducing rate of conductance, however, is much slower than the increasing rate [33]. Infrared (IR) is also used for recovery. The response of conductance to IR, however, is slower than that to UV [39]. Heating sensors at an elevated temperature in inert gases such as N2 and Ar can be used as a recovery method [40]. Sometimes air can be used as a recovery gas, depending on the sensitivity in the

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 6

BlackLining

6
test [51]. The electric field applied over the back gate electrode of the ZnO nanowire FET significantly affects the sensitivity as it modulates the carrier concentration. A strong negative field is used to refresh the sensors by an electroadsorption mechanism [41]. A high vacuum environment can also be used for deadsorption of gas [35].

Gas Sensors Based on One-Dimensional Nanostructures

2.1.5. Position of 1D Nanostructure Materials on the Electrodes


1D nanostructure FET gas sensors have some advantages, such as good sensitivity, as described in Sections 2.1.2 and 2.1.3. The change in physical properties of the FETs under different gases and concentrations can be used to investigate the sensing mechanism of gas sensors [35]. The fabrication of reliable contacts to the devices is a crucial technical issue in the manufacturing process of the sensors. To increase placement yield, several groups investigated the placement of nanowires on selected positions from a suspended solution [42]. Nanowires can be assembled into parallel arrays with control of the average separation by combining fluidic alignment with surface-patterning techniques [43]. A direct placement method has been reported by applying a DC or AC voltage bias [44, 45]. A solution containing nanowires is dropped on the top of the electrode array. The nanowires are aligned along the direction of electric field. Different species such as a nanowire or contaminant species will respond to different frequencies of AC bias. Therefore, applying an AC bias with a corrective frequency offers the advantage of selectively choosing nanowires from other contaminant species in the solution [46]. Solutions such as vertical field effect transistors [47], field emission growth [48], and self-assembled nanowires [49, 50] have been proposed.

2.2. Electro Gas Sensors Manufactured by Microelectromechanical System (MEMS) Technology


Placing nanowires on electrodes is a time-consuming process in the manufacture of FET sensors. Other types of electrosensors based on MEMS techniques have been reported recently. A voltage is applied on the two electrodes during measurement and the current or resistance is used as the measurement signal.

Figure 5. (a) Nanowire-pattern sensor device completed with circuitry elements; the nanowire pattern is in the middle of the sample (high magnification in right bottom), whereas all around are the visible heater element and parallel electrode contacts for SnO2 nanowires. (b) Functional characterization of 80 nm wire-pattern sensor (solid line) and continuous film sensor (dotted line). Reprinted with permission from [51], P. Candeloro, et al., J. Vac. Sci. Technol. B 23, 2784 (2005). 2005, American Institute of Physics.

2.2.1. E-beam Lithographic Nanowire-Patterns


A SnO2 nanowire-pattern on a SiO2/Si substrate is fabricated by electron beam lithography and lift-off technique [51]. Figure 5(a) shows two Pt electrodes (right side) for measuring the conductance of SnO2 nanowire-pattern and a heater element Pt to operate from room temperatures up to 500C. The measurement is performed by monitoring the changes in the current or the resistance of the device. Figure 5(b) shows the change in current for a thin film sensor, as indicated by dotted lines, and a nanowire-pattern sensor, as indicated by solid lines. The operating temperature was kept at 300C and a relative humidity of 40%. Conductivity of both sensors increases when exposed to reducing gases

such as CO, ethanol, and acetone and decreases when exposed to oxidizing gases such as NO2. The responses of the nanowire-pattern sensor are higher for all tested gases. For example, for acetone, an 80 nm wire-pattern sensor presents a relative response (DG/G0) (where DG is the conductance variation and G0 the original value), increasing from 1.95 to 7.20 when acetone concentration ranges from 10 up to 100 ppm, whereas the continuous film device reports a maximum response of 1.62 [51].

2.2.2. Chemical Oxidation of Metal Ti to Nanostructured Titania


Nanostructured titania (NST) arrays are also used for gas sensors [52]. Sensing devices are fabricated as follows. A 500 nm thick Ti film is first evaporated on a SiO2/Si wafer.

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 7

BlackLining

Gas Sensors Based on One-Dimensional Nanostructures

Figure 7. (a) Top-view SEM image of the fabricated substrate embedded with Pt interdigitating electrodes and Pt heater. (b) Schematic showing the cross-section of the fabricated sensor. Reprinted with permission from [54], Q. Wan et al., Appl. Phys. Lett. 84, 3654 (2004). 2004, American Institute of Physics. Figure 6. (a) Optical micrograph of sensor with nanostructured titania (NST) arrays as sensing elements. (Inset) SEM image of a single metallized nanostructured pad. (b) Cross-sectional TEM of a nanostructured titania pad. Reprinted with permission from [52], A. S. Zuruzi et al., Appl. Phys. Lett. 88, 102904 (2006). 2006, American Institute of Physics.

Another SiO2 layer is then deposited on the Ti film. The SiO2 layer on Ti film is subsequently etched, exposing Ti patterns, which are then oxidized in aqueous 10% hydrogen peroxide at 80C to form NST arrays. After annealing at 300C, Ti/Pt electrodes are evaporated. The top view of the electrodes on the NST arrays is shown in Figure 6(a). The structure of the sensor can be seen clearly on the cross section of the devices, as shown in Figure 6(b). Electrical characterization indicates that contacts are ohmic and nanostructured titania is highly sensitive to O2. Variations of hundreds of oxygen molecules over a pad sensing element are detected at 250C.

deposited using plasma enhanced chemical vapor deposition (PECVD); (3) a Pt interdigitating electrode is prepared on the SiON layer by sputtering; and (4) a window of 1.4 3 1.4 mm2 is formed through the backside etching of silicon with KOH solution. The structure of the sensors is shown in Figure 7(a). The synthesized ZnO nanowires are ultrasonically dispersed in ethanol for 30 min. The resulting ZnO nanowire paste is then deposited onto the silicon-based membrane by spin coating [see Figure 7(b)]. Finally, the sensors are dried at 400C for 1 h. Ethanol sensing characteristics of the sensor have been studied, demonstrating that ZnO nanowire sensors exhibit a high sensitivity to ethanol gas and fast response time at operating temperatures of 300C. The sensing properties of electro gas sensors manufactured by MEMS technology are summarized [5155] in Table 2.

2.3. Resistive Gas Sensors 2.3.1. Structure of Resistive Sensors


The channel length between the two electrodes for resistive sensors is usually in the millimeter scale. The typical structure of a resistive sensor is illustrated in Figure 8(a) [56]. A mixture of 1D nanostructure materials and ethanol is deposited as a thin film by span coating across the two Au electrodes on the outer surface of the alumina tube. After heating at 380C in air for 4 h, a small NiACr alloy coil is

2.2.3. Drop of Nanowire Paste


This type of gas sensor is manufactured by combining MEMS technology with dropping nanowires on the device [53, 54]. For example, a silicon-based membrane is prepared as follows: (1) Pt heater and the temperature sensors are sputtered on Si3N4/SiO2/Si substrates and patterned by liftoff methods; (2) an insulating SiON layer 1 mm thick is

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 8

BlackLining

Gas Sensors Based on One-Dimensional Nanostructures

Table 2. Summary of the sensing properties of 1D nanostructure MEMS electrosensors. Response time seconds T (C) 25 300 Recovery method/time seconds Air

Nanowire V2O5 SnO2

Gas amine CO C2H5OH acetone NO2 NH3 O2 C2H5OH

Concentration 30 ppb 30 ppm 10 ppm 10 ppm 0.4 ppm 100 ppm 0.8mtorr 1200 ppm

Sensitivity 42 1.957.2

Ref. [55] [51]

WO2.72 TiO2 ZnO

260 245

seconds seconds seconds

20250 200600 300

Air/minute N2/seconds Air/seconds

[53] [52] [54]

crossed through the alumina tube as a heater to produce the operating temperature by adjusting the heater voltage Vh. The fabricated sensors are aged at 300C for 96 h. The testing principle is schematically shown in Figure 8(b) [56]. The heating voltage (Vh) is supplied to the coils for heating the sensors. The circuit voltage (Vc) is supplied across the sensors. The load resistor (RL) is connected to the sensor in series. The signal voltage (Vout) across the load is used as the measurement signal, which will change depending on the ambient gas and its concentration. Measurement can also be set as in Figure 8(c). In this case, current (I) is used as the detected signal. A given amount of gas such as SO2 is injected into the chamber by a microinjector. The sensitivity of the sensor can be measured when the detecting gas is mixed with air homogeneously. A similar structure with a plane shape ceramic as substrate has been reported [57, 58]. For the fabrication of sensors, the 1D nanostructure materials are dispersed in terpineol, which is used as a binder to form pastes. Platinum interdigitated electrodes are deposited with shadow masking on an alumina substrate. The alumina plane with platinum electrodes is then dipped into the nanowireterpineol

paste several times to obtain a gas sensing film. Then the device is annealed at 350C for 1 h in an ambient atmosphere to evaporate the terpineol. Finally, the device and an NiCr heater is welded onto a BakeliteTM substrate.

2.3.2. Properties of Resistive Sensors


Vout or Rs can be used as detecting signals. Vout can be measured directly, as shown in Figure 8(b). Sensor resistor (Rs) can obtained from Eq. (7): RS Vc Vout RL Vout 7

Current through the sensor can also be detected directly when the measurement is set up as in Figure 8(c). In this case, current (I) can be used as the detecting signal. The signal can be Vout, Rs, or I in Figure 8(b, c). For example, CO and H2 are measured using a polycrystalline SnO2 nanowire resistive sensor. Air is often used as recovery gas for the resistive sensors. The sensitivity of the sensor S Rair/Rgas, where Rair is the resistance in atmospheric air and Rgas is the resistance of the SnO2 nanowires in a CO (or H2)air mixture. Figure 9 shows the changes in resistance of the sensor at room temperature under the exposure of 20 ppm CO (dotted curve) and 500 ppm H2 (solid curve). Resistive sensors usually show fast recovery, which is nearly equal to the response time. Compared with an FET sensor or a MEMS electrosensor, a resistive sensor contains more nanowires, which results in larger signals and faster recovery. The sensing properties of resistive sensors [5868, 82, 87, 95] are summarized in Table 3.

3. 1D NANOSTRUCTURE SENSORS BASED ON OPTICAL TRANSDUCERS


Optical based sensors have been developed to overcome the difficulty in forming a high quality contact to the nanowire [69]. Optical gas sensors form another important and extensively studied class. Using chemiluminescence (CL) and photoluminescence (PL) for gas detection offers certain advantages: remote sensing, easy reversion, and good selectivity.

Figure 8. Schematic diagrams of (a) the structure of a resistive gas sensor and (b) the working principle, Vh, Vc, Vout, and RL, which represent the heating voltage, circuit voltage, signal voltage, and load resistor, respectively. Reprinted with permission from [56], G. Y. Zhang et al., Sens. Actuators B 114, 402 (2006). 2006, Elsevier. (c) The working principle, A, is a current meter.

3.1. Chemiluminescence (CL)


An experiment setup for CL is illustrated in Figure 10(a), where a-Fe2O3 nanowire materials were put over the

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 9

BlackLining

Gas Sensors Based on One-Dimensional Nanostructures


Table 3. Summary of the sensing properties of 1D nanostructure resistive sensors. Concentration (ppm) 200 200 100 100 1000 0.05 1.7 7 10 10 20 500 60000 20 50 50 50 50 10 500 500 500 100 10 10 10 T (C) 225 225 100 270 20, 350 1.7 20 4.2 1s 2s 10 min minutes 300 150 100 25 Recovery method/time N2

Nanowire ZnO

Gas CO NO2 H2S C2H5OH C2H5OH H2S C2H5OH H2 C2H5OH C2H5OH CO NO2 CO H2 C2H5OH C2H5OH C2H5OH acetone, CS2, NH3, H2S, benzene, ether, CH3CN C2H5OH C2H5OH H2 C2H5OH

Sensitivity 1.7 2 1.4 45 40

Response time

Ref. [58]

ZnO ZnO ZnO a-Fe2O3 SnO2 SnO2 SnO2 polycrystals

Air Air 3 min 320 min Air Air/1 Air/10 min Air/10 min Air

[87] [82] [95] [59] [60] [61] [62] [63] [64] [65]

SnO2 WO3 hollow spheres

In2O3 V2O5 Co3O4

4 3 >10000 22 2 3.5 1.56 1.5 21.8 2.56 2.72 3.18 27 1.9 1.7

50 s 1020 s

300 400

10 s seconds

370 200 25

Air/20 s Air/seconds Air

[66] [67] [68]

bottom of a quartz tube. A mixture of air and H2S was passed through the quartz tube. At a certain temperature, H2S was oxidized at the surface of the a-Fe2O3 sample by the oxygen in the flowing air. The change in CL intensity was measured with a BPCL chemiluminescence analyzer. The gas flow rate and detection wavelength were set at 200 mL/min and 400 nm, respectively. A typical response curve is shown in Figure 10(b). The sensor exhibited a very strong CL intensity of about 2500 absorption units in response to 22 ppm H2S in the tube at 134C [70]. This CL intensity is almost comparable to that of the a-Fe2O3 film sensor in response to 100 ppm H2S at 360C [71]. In addition, the nanowire sensor has a short response time of 15 s and a recovery time of less than 100 s, from Figure 10(b).

3.2. Photoluminescence (PL)


The ZnSe nanowires were passivated in (NH4)2S solution for 60 min. For PL measurement, the nanowire samples were mounted in a small vacuum chamber with ports to allow gases in and out and optical access. The chamber was cyclically filled to 760 torr using gases of H2, N2, Ar and air or was pumped to below 10 m torr. Figure 11(a) shows the PL spectra of the nanowires in vacuum and in gases H2, air, Ar, or N2 at atmospheric pressure [72]. Under normal circumstances, the intensity of the near band edge emissions (NBE) peaks at 2.67 eV. The broad deep level emissions centered at 2.1 eV remain relatively constant. Compared with the spectral

Figure 9. Change in resistance for a resistive sensor based on polycrystalline SnO2 nanowires: 20 ppm CO (dotted curve) and 500 ppm H2 (solid curve) under room temperature. Reprinted with permission from [62], X. H. Jiang et al., J. Mater. Chem. 14, 695 (2004). 2004, The Royal Society of Chemistry.

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 10

BlackLining

10

Gas Sensors Based on One-Dimensional Nanostructures

Figure 10. (a) Schematic diagram of the CL sensor measurement system and (b) the CL response curve. Reprinted with permission from [71], Z. Y. Zhang et al., Sens. Actuators B 102, 155 (2004). 2004, Elsevier.

peaks obtained in vacuum, exposure to H2 increases the NBE intensity, but exposure to N2, O2, and even Ar decreases it. The time dependence of the changes in the intensity of NBE peak in vacuum or various gases is shown in Figure 11(b) [72]. The nanowires indicate fast response and recovery in the pressure of H2. The response to the increase in the pressure of N2 and Ar is as fast as it is to H2, but the response to the decrease it is considerably slower. This suggests a fast response and slower recovery for N2 and Ar. After air is passed through the sample, the intensity of NBE peak cannot return to the value in vacuum when the system was repumped to below 10 mtorr, suggesting the responses to O2 are not reversible. Photodesorption and photoadsorption of gases are found to affect the intensity of only the NBE peak, indicating the participation of free carriers. The sensing properties of 1D nanowire optical sensors are listed in Table 4.

Figure 11. (a) Comparisons of the PL spectra of ZnSe nanowires in vacuum and in different gases at 760 torr and (b) response of the intensity of the NBE emissions to the cyclical changes of pressure between 10 mtorr and 760 torr in different gases. Reprinted with permission from [72], K. M. Ip et al., Nanotechnology 16, 1144 (2005). 2005, IOP Publishing.

4. 1D NANOSTRUCTURE SENSORS BASED ON QUARTZ CRYSTAL MICROBALANCE SENSORS


Quartz crystal microbalance (QCM) is an extremely sensitive measurement device. The principle of the measurement is based on mass change as the gas adsorbs on the surface of the sensing material. It can be measured quantitatively, according Eq. (9): Df af 2 Dm=A 8

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 11

BlackLining

Gas Sensors Based on One-Dimensional Nanostructures


Table 4. Summary of the sensing properties of 1D nanostructure optical sensors. Response time (s) 15 15 5 15 15 15 1 1.8 Recovery methods/time Air/100 s Air/120 s 10 mtorr 5s 3 min 3 min T (C) 134 360 25

11

Nanowire Fe2O3 Fe2O3 WO3 ZnSe

Gas H2S H2S H2 N2 Ar Air NO2

Concentration 22 ppm 3 ppm 760 torr

Sensitivity 2500 unit

PL/CL CL CL PL

Ref. [70] [71] [72]

SnO2

110 ppm

25120

PL

[61]

where a is a constant, is the fundamental frequency of the unloaded piezoelectric crystal, Dm is the mass loading on the surface of the crystal, and A is the surface area of the electrode. Circular-shaped 25 MHz quartz crystals (consisting of two silver electrodes with a diameter of 4.5 mm and a thickness of 0.2 mm) were used to make QCM sensors [73]. The materials for the sensing element were 1D nanostructure ZnO with two different shapes. One was a multileg shape with the average length of 1 lm and a mean diameter of about 300 nm, used as sample I. The other was a nanowire with a mean diameter of 3040 nm and a length of about 25 lm, used as sample II. A 1D nanostructure ZnO in ethanol was deposited on the surface of the QCM electrodes using the spin-coating technique. The samples were dried at temperatures of less than 500C. Humidity sensing experiments were performed in the system shown in Figure 12(a). The various humidity levels were achieved using anhydrous P2O5 and saturated aqueous solutions of CaCl2, LiCl, MgCl2, NaBr, NaCl, KCl, and K2SO4 in a closed glass vessel at room temperature, which yielded approximately 5.0%, 12.0%, 33.2%, 57.6%, 75.8%, 84.3%, and 96.7% relative humidity (RH), respectively. The oscillation generation and frequency test of QCM were carried out using a crystal impedance meter and a digital frequency counter. Figure 12(b) shows the frequency shift (Df) curves of two QCM humidity sensors coated with the different shapes of ZnO films as a function of the RH within the time interval of 180 s. The results showed that the QCM sensor II coated with ZnO nanowires has a larger frequency response than that of the QCM sensor I coated with multileg ZnO. The inset in Figure 12(b) shows curves of the sensitivity (S) vs RH for two QCM sensors with different morphology ZnO nanostructure coatings. Here, the frequency changing rate S is defined as frequency sensitivity: D f Sf 9 Dt where D is the response frequency shift within the time interval of Dt. The sensitivity of sensor II is higher than that of sensor I, concluding that thinner and longer 1D nanostructure ZnO can produce a larger response. The response and recovery times can be found from the frequency shift, about 1.5 min and 2 min, respectively. The sensing properties of 1D nanostructure OCM sensors [73, 74] are listed in Table 5.

Figure 12. (a) Schematic diagram of a humidity testing system and (b) frequency shift curves of two ZnO nanostructure coated QCM humidity sensors as a function of the relative humidity: (I) multileg ZnO; (II) ZnO nanowires. The corresponding insets are curves of the sensitivity vs relative humidity. Reprinted with permission from [73], Y. Zhang et al., Physica B 368, 94 (2005). 2005, Elsevier.

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 12

BlackLining

12
Table 5. The sensing properties of 1D nanostructure OCM sensors.

Gas Sensors Based on One-Dimensional Nanostructures

Nanowire ZnO ZnO

Gas H2O ammonia

Concentration 5%97% RH 401000 (ppm)

Sensitivity 580

Response time 1.5 m 5s

T (C) room room

Recovery time 2 min seconds

Ref. [73] [74]

5. MECHANISM OF 1D NANOSTRUCTURE GAS SENSORS


Most of the 1D nanostructure gas sensors operate on the basis of a change in electrical properties of the active element, which adsorbs gas molecules on the surface of the sensor. The discussions on the mechanism of gas sensors in this section concentrate on electrotransducer sensors.

5.1. Reaction Between Crystal Defects and Gases


An n-type semiconductor can be obtained either through an excess of metal (for example, ZnO) or a deficit of nonmetal (for example, TiO2 and SnO2), as shown in Figure 13. Here

M and O represent metal and oxygen, respectively. To allow extra metal in the oxide, it is necessary to postulate the existence of interstitial cations with an equivalent number of electrons in the conduction band. The structure of ZnO can be represented as that shown in Figure 13(a) [89]. Here, M2 is represented as possible occupiers of interstitial sites. Alternatively, n-type behavior of oxides can be caused by nonmetal deficit. This can be visualized as the discharge and subsequent evaporation of an oxygen ion. The electrons enter the conduction band and a vacancy is created on the anion lattice in Figure 13(b). For n-type oxide single crystals, the intrinsic carrier (or oxygen ion vacancy) concentration is determined primarily by the deviation from stoichiometry in the form of equilibrium interstitial metal ions (or oxygen ion vacancy), which are predominantly atomic defects. The conduction electrons resulting from the point defects play a major role in the gas sensing of the materials. Interstitial metal ion and oxygen vacancies on many oxide surfaces are electrically and chemically active. When interstices and vacancies are created, the electrons left behind are localized in states whose energies lie close to the conduction band and function as n-type donors. As a result, the interstices and vacancy creation increase the conductivity of the oxide. Oxygen adsorbs on the surface of oxides and ionizes to O 2 ads or Oads in the formation of a depletion layer on the oxide surface under an oxygen environment first, according to Eq. (10) [75]: O2 e O 2 ads 10

Oxygen vacancies can be the best positions for oxygen adsorption. As a result, the adsorption of an electron acceptor lowers the conductivity of the n-type oxide. The resistance of a metal oxide increases or decreases depending on the type of gas it is exposed to. If the adsorbed molecules are charged acceptors, they will act as reducing agents (such as CO or H2S). In the absence of an O2 environment, reducing gas molecules such as CO adsorb on the surface of an n-type oxide as COads. Gas molecules interacting with the surface of an oxide tend to bind at the oxygen vacancy sites [76]. Adsorbed COads reacts with the oxide, according to Eq. (11) [77]:
COads , CO ads e

11

In the presence of an oxygen environment, O 2 ads or Oads are the main adsorbed species, and these form a depletion layer on the oxide surface [77]. A reducing gas such as CO reacts with the adsorbed O 2 ads or Oads species as in Eq. (12) to change the thickness of the depletion layer [58]:

Figure 13. The structure of an n-type oxide: (a) interstitial cations and excess electrons and (b) oxygen ion vacancies and excess electrons.

2CO O 2 ads ! 2CO2 e

12

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 13

BlackLining

Gas Sensors Based on One-Dimensional Nanostructures


Table 6. Responses of nanoparticle/nanotube sensors to various gases of 400 ppm at 300C. CuCo2O4 Detected gas CH3COOH SO2 C2H5OH CO Cl2 NO2 Nanoparticle 1.27 1.61 1.38 1.08 1.39 1.07 Nanotube 1.40 41.65 1.11 8.37 2.72 1.37 NiCo2O4 Nanoparticle 1.12 1.12 1.00 1.49 1.01 Nanotube 35.28 3.78 7.63 1.61 1.50 1.51 ZuCo2O4 Nanoparticle 0.944 1.26 1.00 1.35 1.27 1.50 Nanctube 1.455 2.83 6.17 1.69 2.48 1.72

13

Source: Reprinted with permission from [56], G. Y. Zhang et al., Sens. Actuators B 114, 402 (2006). 2006, Elsevier.

In both cases, the reducing action of the CO injects electrons into the conduction band, according to Eqs. (11) and (12), which results in the increase in conductivity. Oxidizing gases such as NO2 react with a negatively charged surface state and are thereby neutralizing. This reduces the depleting region surrounding the surface, resulting in an overall increase in the conductivity of the nanowire [58].

exhibit superior gas sensing capabilities toward H2 and C2H5OH gases than the Co3O4 nanoparticles. The sensitivity of nanowire and nanotube sensors is much higher than that of nanoparticle sensors. This has also been reported in other systems, La0.59Ca0.41CoO3 [81] and a-Fe2O3 [59] for H2 and ethanol.

5.2. Mechanism for 1D Nanostructure Gas Sensors


Various gas sensing mechanisms in semiconductor oxide sensors have been suggested, including the desorption of the oxygen atoms that are adsorbed on the surface and grain boundaries in polyoxides [78]. The exchange of charges between adsorbed gas species and the surface of oxides lead to changes in depletion depth [79] and changes in the surface or grain boundary conduction by gas adsorption/desorption [80]. In all of these cases, the mechanism is surface related. In addition to the previous mechanisms, some mechanisms are used to explain the sensing properties of 1D nanostructure sensing elements.

5.2.1. Nanoparticles and Nanowires


Resistive sensors are fabricated using NiCo2O4, ZnCo2O4, and CuCo2O4 nanoparticles (20 nm) and nanotubes (polycrystals with an outer diameter of approximately 200 nm and a wall thickness of about 20 nm). The sensor response Rair/Rgas, the ratio of the resistance in air and the resistance in detected gas, is measured in different gases at a concentration of 400 ppm and temperature of 300C, as listed in Table 6. Although the surface area of nanoparticles is larger than that of the nanotubes, the response of nanotube sensors is higher than that of nanoparticle sensors. The Co3O4 nanotubes are synthesized via chemical decomposition of Co(NO3)26H2O within the alumina membranes, whereas Co3O4 nanoparticles are prepared by ball-milling the decomposition product of Co(NO3)26H2O in Ar ambient at 500 rpm for 1 h. The average outer diameter of the nanotubes is about 200 nm and the wall thickness is about 2030 nm. The diameter of nanoparticles is about 100 nm. The Co3O4 nanotubes and nanoparticles are then used as sensing elements to make resistive sensors. Figure 14 shows the sensitivities of two types of sensors using nanotubes and nanoparticles as sensing elements to H2 and C2H5OH gases at room temperature. The Co3O4 nanotubes

Figure 14. Sensitivity to (a) H2 and (b) C2H5OH of the sensors made by CO3O4 nanotubes (triangles) and nanoparticles (squares) at room temperature. Reprinted with permission from [68], W. Y. Li et al., Adv. Funct. Mater. 15, 851 (2005). 2005, Wiley-VCH.

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 14

BlackLining

14 5.2.2. Shape of Nanowires


The morphologies of nanowires can also lead to changes in the sensitivity of nanosensors. T-ZnO and M-ZnO nanowires are shown in Figures 15(a) and (b), respectively. These two types of nanowires are used as sensing elements to make resistive sensors. Figure 15(c) shows the typical response curves of the two kinds of ZnO nanowire sensors to ethanol (1000 ppm) in air. For all the ZnO nanowire sensors, the gas sensitivity increases with increasing concentration of ethanol in air [82]. The M-ZnO nanowire sensors have higher ethanol sensitivity than the T-ZnO nanowire devices. The M-ZnO nanowire sensors have a fast response time, approximately 10 s, wherease the response time of the T-ZnO nanowire devices is approximately 16 s. This enhanced gas sensitivity of the M-ZnO nanowire sensors has also been found in their responses to other gases. For the T-ZnO nanowires, the intensity of the UV emission is higher than that of the green emissions, whereas for the M-ZnO nanowires, the intensity of the UV emission is lower than that of the green emissions. The higher green emission for M-ZnO is believed to be due to more oxygen ion vacancies in the nanowires [83].

Gas Sensors Based on One-Dimensional Nanostructures

5.2.3. Mechanism
The total energy of adsorption and reaction of molecular AQ1 oxygen on the ZnO (10 10) surface has been calculated by first principles [84]. The adsorption is fully molecular with a small adsorption energy, and the dissociation is not energetically favored on stoichiometric ZnO (10 10) surfaces. On a partially reduced ZnO (10 10) surface, the dissociative adsorption is energetically preferred. The dissociative adsorption will influence the electronic properties of the ZnO (10 10) surface. The electronic structure of three-dimensional SnO2 nanostructures (aerogels) is studied by soft X-ray absorption near-edge structure (XANES) spectroscopy. Highresolution O K-edge and Sn M3-edge and M4,5-edge XANES spectra of nanocrystalline rutile SnO2 aerogels with surface areas from 320 to 55 (m2/g) are compared with the spectra of bulk rutile SnO2. The result shows that the presence of under-coordinated surface atoms affects the position of Fermi level and the structure of the conduction band by introducing additional Sn-related electronic states close to the conduction band minimum. These additional states arise from oxygen deficiency and are attributed to the surface reconstruction of SnO2 nanoparticles forming the aerogel skeleton [85]. Theoretical analysis suggests that that defects on the surface of an oxide are important for gas sensing. 1D nanostructure metal oxides, nanowires, and nanotubes are able to improve the performance of chemical sensors significantly. The following explains this phenomenon: first, the large surface-to-volume ratio of the 1D nanostructure offers large areas for gas adsorption and reaction. Second, the 1D nanostructure favors the carrier motion in two directions. Third, the hollow structure of the nanotubes is extremely convenient for gases to drift in and out. Finally, the defects in 1D nanostructure facilitate the change of the electrical conductivity of the oxide. For n-type semiconducting metal oxides, the sensing mechanism is usually related to the

Figure 15. (a) SEM image of the T-ZnO nanowires. (b) SEM image of the M-ZnO nanowires. (c) Typical response curves of the two kinds of ZnO nanowire sensors to ethanol (1000 ppm) in air. Reprinted with permission from [82], T. Gao and T. H. Wang, Appl. Phys. A 80, 1451 (2005). 2005, Springer.

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 15

BlackLining

Gas Sensors Based on One-Dimensional Nanostructures

15
in vacuum. The changes in conductivity after annealing are caused by the variations in the number of oxygen species adsorbed on surfaces and by the change in the number of oxygen vacancies on the material bulk due to different oxygen partial pressures in the environment. Annealing in vacuum should decrease the number of adsorbed oxygen species and increase the number bulk oxygen vacancies whereas annealing in oxygen or in air should do the opposite. The equilibrium concentration of intrinsic defects for n-type oxides increases with the oxygen partial pressure [89]. When a 1D nanostructure ZnO FET is annealed in hydrogen gas at 400C, the hydrogen-annealed devices are more conducting because the hydrogen increases the n-type doping in the ZnO [92]. The hydrogen-annealed nanowire sensors were insensitive, however, to the measurement gases, such as N2, O2, H2, N2O, and C2H2. Compared with hydrogen-annealed nanowires, unannealed samples show a strong sensitivity to measurement gases, even hydrogen gas [93]. One might expect the current to be dependent on this ambient gas if the conduction were truly dominated by the surface. Therefore, bulk properties of the nanowire may contribute to conductance change in the sensing of gases [93].

depletion layer (its thickness, Ld) formed on their surface as electrons are trapped by adsorbed gas species such as O2. The electrical conductivity of nanocrystalline oxides depends strongly on the surface states produced by molecular adsorption that result in space charge layer changes and band modulation. In addition to the surface-related mechanisms, bulk properties of the nanowire can contribute to conductance changes in the sensing of gases [93].

5.3. Electrical Sensitivity of 1D Nanostructure Materials to Manufacture History and Environments


Electrical sensitivity of nanowire gas sensors depends strongly on the history of the fabrication of nanowires. For example, In2O3 nanowires with different doping concentrations were investigated by examining the conductance of 1D In2O3 FET gas sensors in NH3 gas environment. The conductance of the FETs changes in an opposite direction when introducing NH3, because of the effect that doping in the nanowire has on the density of oxygen vacancies along the nanowire sensors [86]. ZnO tetrapod samples prepared by evaporating high purity Zn in various ambient gases (humidified Ar, dry Ar, humidified N2, and dry N2) are used as sensing elements to make resistive sensors. The results of experimentation demonstrated that the sensor response was strongly dependent on the ambient of the deposition process. The sensors based on ZnO tetrapods prepared in a humidified Ar atmosphere showed a high response, good selectivity, and short response time to dilute C2H5OH [87]. The history of the fabrication of nanowires can influence the defect type and concentration with a consequent effect on the sensing properties. ZnO with a wurtzite structure is naturally an n-type semiconductor because of the deviation from stoichiometry. The deviation is considered as the pres00 0 or VO ) ence of intrinsic defects such as O vacancies (VO   and Zn interstitials (Zni or Zni ). The equilibrium concentration of intrinsic defects on the surface and bulk oxide will be a function of the environmental oxygen partial pressure and the temperature [88, 89]. Undoped ZnO shows intrinsic n-type conductivity with very high electron densities of about 1021 cm3 [90]. In addition, defect concentration is strongly dependent on the impurity type and concentration. Obtaining n-type doping of ZnO is relatively easy compared to p-type doping. Group III elements Al, Ga, and In as substitutional elements for Zn and group VII elements Cl and I as substitutional elements for O can be used as n-type dopants [91]. P-type doping in ZnO may be possible by substituting either group I elements (Li, Na, and K) for Zn sites or group V elements (N, P, and As) for O sites [24]. The conductivities of nanowire devices are also very sensitive to the annealing process, including temperature, time, and gases. The conductivity of SnO2 nanowire FET increases significantly after the device has undergone annealing in oxygen at 800C for 2 h [28]. When a SnO2 nanowire FET is annealed at 200250C in air or vacuum, generally the conductivity of the device decreases after annealing in air and increases after annealing

5.4. Temperature Dependence of Gas Sensors


The temperature dependence of gas sensors has been studied in detail [58, 66, 87]. The response of the sensor to the presence of the reducing gases depends mainly on two factors. The first is the density of active sites for oxygen and the reducing gases on the surface of the sensor materials. The second is the reactivity of the reducing gases. Operation temperature of a sensor is strongly dependent on the bond energy of the gas. The bond energy of HASH is 381 kJ/mol [94], so it is easy to break the HASH bond at low temperatures. On the other hand, the bond energies of HACH2, HAOC2H5, and HACH in C2H5OH are 473, 436, and 452 kJ/mol [94], respectively, so it is difficult to break the bonds in C2H5OH at low temperatures. Therefore, the ZnO nanowire sensors show high responses to H2S at low temperatures and to C2H5OH at high temperatures [95]. The resistance variation with temperature (T) reflects an activated process with the activation energy. The currentvoltage IDS VDS curves of In2O3 FET sensors are measured in air as a function of temperature at 290, 180, 120, 70, 30, and 10 K with 0 V applied to the gate electrode. The conductance vs 1/T shows a linear relationship, which indicates that the transport through the device was dominated by thermal activation of electrons across the metal-semiconductor. Thermal activation energy Ea was calculated as 6.90 meV, according to the formula   2Ea 13 G exp kT where G is conductance, k is Boltzmanns constant, and T is absolute temperature [30]. The IV characteristics of single ZnO nanowires have also been measured as a function of both temperature and ambient gas. The conductivity of the nanowires can be increased by a postgrowth anneal in hydrogen (400C), and these nanowires show a thermally

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 16

BlackLining

16

Gas Sensors Based on One-Dimensional Nanostructures

Sensitivity (S)

Sensitivity (S)

Concentration (C) (a)

Concentration (C) (b)

Sensitivity (S)

Concentration (C) (c)

Figure 16. Illustration of the relations between gas concentration and responses of a sensor: (a) linear law, (b) power law and (c) saturation.

activated current. The apparent activation energy derived from Eq. (13) is 89 6 0.02 meV. This does not correspond to any of the known donor dopant or native defect ionization energies in ZnO [96].

If the concentration range of the target gas is large, the responseconcentration curve will be linear at low concentrations and will tend to saturate with an increase in the gas concentration, which can be described by the following equation: s a
b 1C

5.5. SensitivityConcentration Curves


The relationship between sensitivity (S) and concentration (C) of detected gases may obey a linear law, a power law, or saturation, as illustrated in Figures 16 (a, b, c), respectively. The relationship between the concentration of target gas and the response of a sensor depends on several factors, such as gas type, the concentration range of target gas, test temperature, and the defect density on n-type semiconductors. If the concentration of active sites on the semiconductor surface is much higher than the adsorbed gas, a linear relation will be expected. If the active sites on the oxide surface are fully covered by the target gas, a saturation situation will occur. Otherwise, the sensitivity vs concentration of gas may show a power law relationship. Some mathematical models have been used to describe these curves. It has been demonstrated that the sensitivity of oxide semiconductors is usually depicted as [54, 97] S aCN 14

15

where a and b are experimental constants and C is the concentration of the gas. The explanation is that the surface coverage of adsorbed molecules follows Langmuir isotherm [100].

6. 1D NANOSTRUCTURE SENSORS BASED ON COMPOSITE SENSING ELEMENTS


1D nanostructures of semiconducting metal oxides such as ZnO, SnO2, TiO2, and WO3 have attracted great attention because of the potential applications in gas sensors, humidity sensors, and nanoelectronic circuits. Nowadays, a great deal of effort is put into the study of double metal oxides such as ZnSnO3, 1D nanostructure composites such as 1D nanostructure ZnOASnO2, and nanoparticle coated nanowires such as Pd particles on SnO2 nanowires because of their novel properties in nanodevices.

where a denotes a constant and C is the concentration of the target gas. N is usually 1, indicating a linear law, or 1/2, indicating a power law, dependent on the charge of the surface species and the stoichiometry of the elementary reactions on the surface [98]. For thin films composed of SnO2 particles with a diameter of 620 nm, N is about 1. N is about 0.5, however, when the diameter of SnO2 increases to longer than 20 nm [99]. For example, ethanol (gas concentration set to less than 300 ppm) is measured at 300C using a SnO2 resistive sensor with nanowires of diameter of 620 nm as sensing elements. The sensitivity of the sensor is directly proportional to the concentration of ethanol gas (linear law) [60]. In this case, N is about 1. In a similar experiment, the In2O3 nanowires had a diameter ranging from 60 to 160 nm. The ethanol (gas concentration range of 1001000 ppm) was measured at 370C. The relationship between the response of the sensor and gas concentration obeyed the power law [66]. In that case, N was about 1/2.

6.1. Double Metal Oxide Gas Sensors


Bulk ZnSnO3 sensing elements have been suggested as having a higher sensitivity to ethanol gas than bulk ZnO or SnO2 sensors [101, 102]. The mixed valences of the cations in these composite oxides may help the reversible adsorption of gases by providing donor-acceptor sites for chemisorption [103]. The properties of double metal oxide 1D nanostructure sensors [56, 81, 104] are summarized in Table 7.

6.2. 1D Nanostructure Composite Metal Oxide Gas Sensors


Metal oxide sensing films have been doped with noble metals (or metal oxides) to increase sensitivity, reduce response

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 17

BlackLining

Gas Sensors Based on One-Dimensional Nanostructures


Table 7. The sensing properties of 1D nanostructure double oxide sensors. Concentration (ppm) 400 Response time (s) T (C) 300 Recovery method/time (s) Air

17

Nanowire CuCO2O4 NiCO2O4 ZnCO2O4 ZnSnO3 La0.59Ca0.41 CoO3

Gas SO2 CH3COOH C2H5OH C2H5OH C2H5OH H2

Sensitivity 41 32 6 2.742 2.7 1.7

Sensor Resistive

Ref. [56]

1500 10 10

300 350

Air/1 Air

Resistive Resistive

[104] [81]

AQ2 time, and selectivity. There are two mechanisms of sensitization by metal or metal oxide additives: chemical and electronic sensitizations. The chemical sensitization is performed by a spillover effect [105], wherease the electronic sensitization is accomplished by the direct exchange of electrons between the semiconductor and the metal additives [106]. Dopants such as Pt or Pd are catalysts that promote chemical reactions between the film and the test gas by reducing the activation energy without being consumed themselves [106]. This allows the reaction to occur at a faster rate, at lower temperature, and at lower gas concentrations. The 1D nanostructure composite materials can be one of the following. The doped metal (or oxide) can be nanoparticles on the surface of oxide nanowires; doped elements can be impurities modifying the 1D nanostructure materials; or two different types of 1D nanostructure materials can be mixed to form a composite. 1D nanostructure composite materials can be used as sensor elements to make sensors. The sensing properties of 1D nanostructure composite material gas sensors [57, 107115] are listed in Table 8.

6.2.1. Chemical Sensitization


The selectivity of gas sensors and catalysts is usually achieved by functionalizing oxides with catalytically active metals. Enhanced gas sensing by SnO2 nanowires functionalized with Pd catalyst particles has been reported [109].
AQ8 Table 8. Sensing properties of 1D nanostructure composite gas sensors. Concentration (ppm) 100 2 2 10 10 5004000 100 103 torr 500 10 500020000 200 3

Either Pd or Au is deposited on a SnO2 nanowire FET, as shown in Figure 17(a). The sensing properties of the FET were measured in situ in at variable temperatures during the deposition. The deposition of Pd or Au on the SnO2 nanowire was monitored during the deposition process by measuring drain-source current, IDS. The change in IDS with increasing Au deposition time is shown in Figure 17(b). In the initial cluster nucleation stage, the drop in conductance implies the formation of depletion regions at the metalsemiconductor interface. Current is small in the stage of cluster growth. The conductance dramatically increases in the stage of cluster percolation, because of shorting out the SnO2 nanowire. The sensing properties of a SnO2 nanowire were measured in oxygen and hydrogen pulses at temperatures of 443 and 473 K before (dashed curves) and after (solid curves) Pd deposition, as shown in Figure 17(c). The SnO2 nanowire surface with Pd shows larger IDS response for both gases than the SnO2 nanowire. The mechanism of the Pt function as a catalyst is illustrated in Figure 17 (d)-(I). In the case of the absence of Pt particles on the SnO2 nanowire, ionosorption of oxygen occurs at defect sites of the nanowire surface, as illustrated in process (1); chemical catalysis of Pt occurs at process (2) and process (3). Oxygen dissociates on Pd nanoparticles followed by spillover onto the oxide surface (spillover effects) in process (2). In process (3), a Pd nanoparticle captures

Materials SnO2 on MWNT SnO2 on MWNT

Gas LPG NO NO2 C2H5OH C2H2 H2 C2H5OH H2 O2 LPG C2H5OH H2 C2H5OH H 2S

Sensitivity 4-7 1.9 1.77 1.83 1.5 90

Response time 1.8 s

T (C) 20, 335 300

Recovery method/time (s) Air/100 Air

Sensor Resistive Resistive

Ref. [57] [107]

SnO2 on NaY zeolite CdS on SnO2 Pd on SO2 PdO on ZnO 3%Sb doped-SnO2 0.9TiO20.1SnO2 Zn2SnO4SnO2 CuOSnO2

seconds 19 25 0.87 0.95 8 5 10 s 5s 2 min seconds seconds

280 400 200 332 300 400 250 25

Air Vacuum Air/30 Air/5 Air/420 Air/s Air/30

Resistive Resistive FET Resistive Resistive Resistive Resistive Resistive

[108] [115] [109] [110] [111] [112] [113] [114]

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 18

BlackLining

18

Gas Sensors Based on One-Dimensional Nanostructures

Figure 17. (a) Schematic view of device used for the current measurements under gas exposure and metal deposition, (b) source-drain current, IDS, through a SnO2 nanowire (VDS 2 V) measured during Au deposition, (c) response of SnO2 nanowire (dashed line) and Pd-functionalized (solid line) nanostructure to sequential O2 and H2 pulses at 473 K (top panel) and 543 K (bottom), and (d) (I) schematic depiction of the three major processes taking place at a SnO2 nanowire/nanobelt surface: (1) ionosorption of oxygen at defect sites of the pristine surface; (2) mechanism of spillover effects; (3) mechanism of back-spillover effects, and (II) band diagram of the pristine SnO2 nanostructure and SnO2 in the vicinity of (and beneath) a Pd nanoparticle. Reprinted with permission from [109], A. Kolmakov et al., Nano Lett. 5, 667 (2005). 2005, American Chemical Society.

weakly adsorbed O2, which has diffused along the SnO2 surface to the Pd nanoparticles vicinity (followed by process [2]). RS is the effective radius of the spillover zone, and RC is the radius of the collection zone. Figure 17(d)-(II) illustrates the band diagram of the pristine SnO2 nanostructure and SnO2 in the vicinity of (and beneath) a Pd nanoparticle.

6.2.2. Electronic Sensitization


The structure of a SnO2 nanobelt is shown in Figure 18(a). The sonochemical synthesis of CdS nanoparticles has been done in a neutral aqueous solution with the SnO2 nanobelts to fabricate a structure of SnO2 nanobelt/CdS

nanoparticles, as shown in Figure 18(b). The SnO2 nanowires and the SnO2ACdS core/shell are used as sensing elements to make resistive sensors. Figure 18(c) shows the response curves of the SnO2 nanobelt sensors and the SnO2ACdS core/shell heterostructured sensors to 100 ppm ethanol vapor in air at an operating temperature of 400C. The ethanol-sensing performance of the SnO2ACdS core/ shell heterostructured sensors shows significant improvement compared to the SnO2 nanobelt sensor. Considering the efficient charge separations in SnO2ACdS core/shell heterostructured sensors, the CdS nanoparticles would serve as additional electron sources by the electronic sensitization, and can greatly improve electron conduction in SnO2 nanobelts [115].

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:47 User ID: bhuvaneswaric Enabled Page Number: 19

BlackLining

Gas Sensors Based on One-Dimensional Nanostructures

19
recovery times to 10 ppm ethanol are approximately only 1 and 5 s, respectively, for the Sb-doped SnO2 nanowire gas sensors [111], whereas the recovery time of pure SnO2 nanowire sensors to ethanol was longer than 10 min. Thus, Sb doping can greatly reduce the recovery time of SnO2 nanostructure gas sensors to ethanol [20]. The possible mechanism is that Sb doping favors or accelerates the adsorption of oxygen molecules and the formation of O2 ions on the surface of SnO2 nanowires, which is of significance in the reduction of the recovery times.

6.2.3. pn Junction Between 1D Nanostructure Materials


The SnO2 (nanowires) and 4mol% CuO (nanoparticles) can be mixed mechanically to perform as sensing materials to make resistive sensors. The sensitivity of the sensor to 3 ppm H2S is as high as 18,000 with a response time of 15 s and recovery time of more than 450 s at room temperature. The mechanism of the reaction of a SnO2ACuO sensor to H2S gas is explained by the formation or distortion of p-n junctions between the SnO2 and CuO interface before (air) and after (H2S) exposure, as shown in Figure 19. CuO and SnO2 are p- and n-type semiconductors, respectively. It is easy to form pn junctions at CuOASnO2 interfaces. Numerous pn junctions at CuOASnO2 interfaces cause high resistance in air, because of the one direction conduction of p-n junctions. On exposure to H2S gas, CuO particles are rapidly converted to CuS by the following chemical reaction: CuO H2 S ! CuS H2 O 16

Figure 18. (a) TEM image of the uncoated SnO2 nanobelts. Inset shows the corresponding SAED patterns. (b) General morphology of the SnO2 nanobelt/CdS nanoparticle core/shell heterostructures. Inset shows the corresponding SAED patterns. (c) Response curves of the SnO2 nanobelt sensors and the SnO2CdS core/shell heterostructured sensors to 100 ppm ethanol vapor in air at a working temperature of 400C. Reprinted with permission from [115], T. Gao and T. H. Wang, Chem. Commun. 22, 2558 (2004). 2004, The Royal Society of Chemistry.

PdO doped ZnO nanowires are used as sensing materials in the manufacture of a single nanowire sensor. The PdO additive to ZnO nanowires acts as a strong acceptor of electrons from the oxide, and induces an enlarged surface space charge layer, which results in depletion of electrons near the interface. When the PdO additive is reduced on contact with the target gas, it relaxes the space charge layer by giving back electrons to the oxide. Such a change in the oxidation state of the additive is responsible for the promotion of the gas response. Such results agree with the electronic sensitization model [110]. The use of SnO2 nanoparticles on carbon nanotubes has been studied as gas sensors. The resistances of this kind of sensor are much lower than those of the SnO2 nanowire sensors. The sensor resistance is dominated by the barriers between the SnO2 grains on the multiwall carbon nanotubes (MWNTs), and the barrier height is controlled by the adsorptive gas molecules, which extract or release electrons to produce the sensing response [107]. Both Sb-doped SnO2 nanowires and SnO2 nanowires are used as sensing elements to fabricate resistive sensors. The resistance change of the Sb-doped SnO2 nanowires is significant when the ambient gas is changed. The response and

CuS is metallic in nature and its formation shorts out the pn junctions existing on the intersurface, causing a large decrease in electrical resistance. The formation of CuS has been confirmed by X-ray photoelectron, X-ray diffraction, and Raman spectroscopic studies [116]. On the other hand, CuS will be oxidized in air and will change back to CuO reversibly through the following reaction: 3 CuS O2 ! CuO SO2 2 17

These two reactive processes dominate the response and recovery of the CuOASnO2 sensors. Reaction rates for Eq. (16) are much higher than those for Eq. (17) at low

Figure 19. Schematic showing the CuOASnO2 nanoparticle/nanoribbon sensor. Reprinted with permission from [114], X. H. Kong et al., Sens. Actuators B 105, 449 (2005). 2005, Elsevier Science SA.

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:48 User ID: bhuvaneswaric Enabled Page Number: 20

BlackLining

20
temperatures, so the CuOASnO2 sensors have a large response at low temperatures. It must be pointed out that the crystal structure of CuS is changeable at 103C and becomes Cu2S, an ionic conductor with higher resistivity, when the temperature is >220C [117]. Finally, 1D nanostructure CuOASnO2 composite gas sensors have good selectivity to H2S gas.

Gas Sensors Based on One-Dimensional Nanostructures

7. SUMMARY AND FURTHER STUDY


Compared with bulk or film ceramic sensors, 1D nanostructure gas sensors offer high sensitivity, low detection limitations, and low operating temperatures. In addition, 1D nanostructure gas sensors also provide a platform to investigate the relationship between electrical transport properties with dimensionality size in various gas environments. At present, 1D nanostructure gas sensors can be constructed around electrotransducers, optical transducers, and quartz crystal microbalance transducers. 1D nanostructure gas sensors using electrotransducers work based on the change in the electrical conductivity of metal oxides due to chemisorption of gas molecules on the surface of the materials. Among these sensors, 1D nanostructure FET gas sensors have advantages such as good sensitivity and the possibility of the amplification of the signal through the control of the gate voltage. The electrical transport properties of 1D nanostructure materials in various gas environments can be identified because of the known size of the nanowires, so 1D nanostructure FET gas sensors often are used to investigate the sensing mechanism. The fabrication of reliable contacts to the devices is a crucial technical issue in the manufacturing process of FET sensors. For resistive sensors, the size of the sensing element is usually in the millimeter scale. Compared with an FET sensor, a resistive sensor contains more nanowires, which results in large signals and fast recovery. Therefore, air is often used as a recovery gas. Moreover, the resistive sensors are easy to fabricate and cheap because of larger detected signals. Therefore, the resistive sensors are suitable for practical applications. The span coating technique is usually used to deposit the mixture of 1D nanostructure materials and ethanol on electrodes to fabricate a resistive sensor. As mentioned, a good connection between electrodes and 1D nanostructure materials is a serious issue for resistive sensors. Some electrosensors based on MEMS techniques, such as e-beam lithographic nanowire-patterns and chemical oxidation of metal Ti to nanostructured titania, have been reported recently; both of these can provide good connections between electrodes and 1D nanostructure sensing materials. Optical based sensors have been developed to overcome the difficulty in forming high quality contacts to the nanowires [118]. Optical gas sensors offer advantages such as remote sensing, easily reversible changes, and high selectivity for some gas species. 1D nanostructure optical sensors are suitable for gas measurement in dangerous environments. The quartz crystal microbalance (QCM) is an extremely sensitive measurement device for mass. The sensing signal is directly proportional to the mass change as the gas adsorbs on the surface of the sensing material. Compared

with the sensors based on the electrotransducers, both chemisorption and physisorption of gas molecules can occur during the measurement process of 1D nanostructure sensing QCM sensors. Therefore, 1D nanostructure sensing QCM sensors are also suitable for detecting the physisorption of gases. Although promising gas sensing results from the performance of 1D nanostructure metal oxides have been reported, studies of the applications of nanowires in gas sensors are still in the preliminary stages. Techniques to maintain good connections between electrodes and 1D nanostructure materials and to obtain stable 1D nanostructure sensing materials need to be improved. The mechanism of improving the gas sensitivity of 1D nanostructures has been unclear up to now. Developing highly selective and controllably sensitized devices remains a future challenge for 1D nanostructure oxide gas sensors.

ACKNOWLEDGMENT
This work was supported by a research grant from the Research Grants Council of Hong Kong (Project no. PolyU 5236/03E).

REFERENCES
1. U.S. Environmental Protection Agency, Air Trends 1995 Summary http://www.epa.gov/oar/aqtrnd95/no2.html. 2. G. Sberveglieri, Sens. Actuators B 23, 103 (1995). 3. N. Yamazoe, Sens. Actuators B 5, 7 (1991). 4. W. Y. Chung, G. Sakai, K. Shimanoe, N. Miura, D. D. Lee, and N. Yamazoe, Sens. Actuators B 46, 139 (1998). 5. J. Shieh, H. M. Feng, M. H. Hon, and H. Y. Juang, Sens. Actuators B 86, 75 (2002). 6. R. Winter, K. Scharnagl, A. Fuchs, T. Doll, and I. Eisele, Sens. Actuators B 66, 85 (2000). 7. H. Steffes, C. Imawan, F. Solzbacher, and E. Obermeier, Sens. Actuators B 78, 106 (2001). 8. T. A. Jones and B. Bott, Sens. Actuators 9, 27 (1986). 9. Z. W. Pan, Z. R. Dai, and Z. L. Wang, Science 291, 1947 (2001). 10. Z. L. Wang, Z. W. Pan, and Z. R. Dai, US Patent No: 0094450 A1 (2002). 11. M. Yazawa, M. Koguchi, A. Muto, M. Ozawa, and K. Hiruma, Appl. Phys. Lett. 61, 2051 (1992). 12. M. Adachi and T. Harada, Langmuir 15, 7097 (1999). 13. E. Braun, Y. Eichen, U. Sivan, and G. Ben-Yoseph, Nature (London) 391, 775 (1998). 14. Y. C. Choi, W. S. Kim, Y. S. Park, S. M. Lee, D. J. Bae, H. Y. Lee, G. S. Park, W. B. Choi, N. S. Lee, and J. M. Kim, Adv. Mater. 12, 746 (2000). 15. A. M. Morales and C. M. Leiber, Science 279, 208 (1998). 16. T. J. Trentler, K. M. Hickman, S. C. Goel, A. M. Viano, P. C. Gibbons, and W. E. Buhro, Science 270, 1791 (1995). 17. X. Jiang, T. Herricks, and Y. Xia, Nano Lett. 2, 1333 (2002). 18. Y. Y. Fu, R. M. Wang, J. Xu, J. Chen, Y. Yan, A. V. Narlikar, and H. Zhang, Chem. Phys. Lett. 379, 373 (2003). 19. G. Gu, B. Zheng, W. Q. Han, S. Roth, and J. Liu, Nano Lett. 2, 849 (2002). 20. E. Comini, G. Faglia, G. Sberveglieri, Z. Pan, and Z. Wang, Appl. Phys. Lett. 81, 1869 (2002). 21. M. Law, H. Kind, B. Messer, F. Kim, and P. Yang, Angew. Chem. Int. Ed. 41, 2405 (2002).

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:48 User ID: bhuvaneswaric Enabled Page Number: 21

BlackLining

Gas Sensors Based on One-Dimensional Nanostructures


22. D. Zhang, C. Li, X. Liu, S. Han, T. Tang, and C. Zhou, in Proceedings of IEEE NANO 2003, San Francisco, (2022 August, 2003). 23. D. H. Zhang, Z. Q. Liu, C. Li, T. Tang, X. L. Liu, S. Han, B. Lei, and C. W. Zhou, Nano Lett. 4, 1919 (2004). r, Y. I. Alivov, C. Liu, A. Teke, M. A. Reshchikov, Ozgu 24. U. S. Do gan, V. Avrutin, S. J. Cho, and H. Morkoc, J. Appl. Phys. 98, 041301 (2005). 25. P. Grosse, F. J. Schmitte, G. Frank, and H. Kostlin, Thin Solid Films 90, 309 (1982). 26. J. R. Bellingham, A. P. Mackenzie, and W. A. Phillips, Appl. Phys. Lett. 58, 2506 (1991). 27. A. Mandelis, Physics, Chemistry and Technology of Solid State Gas Sensor Device (New York: John Wiley & Sons, 1993). 28. Z. L. Wang, Nanowires and Nanobelts Materials, Properties and Device, Vol. 2 (Boston/Dordrecht/London: Kluwer Academic Publishers, 2003). 29. D. Zhang, C. Li, S. Han, X. Liu, T. Tang, W. Jin, and C. Zhou, Appl. Phys. A 77, 163 (2003). 30. D. H. Zhang, C. Li, S. Han, X. L. Liu, T. Tang, W. Jin, and C. W. Zhou, Appl. Phys. Lett. 82, 112 (2003). 31. Z. L. Wang, Adv. Mater. 15, 432 (2003). 32. C. Li, D. H. Zhang, X. L. Liu, S. Han, T. Tang, J. Han, and C. W. Zhou, Appl. Phys. Lett. 82, 1613 (2003). 33. Q. H. Li, Q. Wan, Y. X. Liang, and T. H. Wang, Appl. Phys. Lett. 84, 4556 (2004). 34. R. Martel, T. Schmidt, H. R. Shea, T. Hertel, and P. Avouris, Appl. Phys. Lett. 73, 2447 (1998). 35. Q. H. Li, Y. X. Liang, Q. Wan, and T. H. Wang, Appl. Phys. Lett. 85, 6389 (2004). 36. S. V. Kalinin, J. Shin, S. Jesse, D. Geohegan, A. P. Baddorf, Y. Lilach, M. Moskovits, and A. Kolmakov, J. Appl. Phys. 98, 044503 (2005). 37. Y. Liu and M. Liu, Adv. Funct. Mater. 15, 57 (2005). 38. Y. Shimizu and M. Egashira, MRS Bull. 24, 18 (1999). 39. Q. H. Li, Q. Wan, Y. G. Wang, and T. H. Wang, Appl. Phys. Lett. 86, 263101 (2005). 40. Y. Zhang, A. Kolmakov, S. Chretien, H. Metiu, and M. Moskovits, Nano Lett. 4, 403 (2004). 41. Z. Y. Fan and J. G. Lu, Appl. Phys. Lett. 86, 123510 (2004). 42. D. C. Duffy, J. C. Mcdonsld, J. A. Schueller, and G. M. Whitesides, Anal. Chem. 70, 4974 (1998). 43. Y. Huang, X. Duan, Q. Wei, and C. M. Lieber, Science 291, 630 (2001). 44. A. Bezryadin and C. Dekker, J. Vac. Sci. Technol. B 15, 793 (1997). 45. D. Porath, A. Bezryadin, S. de Vries, and C. Dekker, Nature (London) 403, 635 (2000). 46. L. A. Nagahara, I. Amlani, J. Lewenstein, and R. K. Tsui, Appl. Phys. Lett. 80, 3826 (2002). 47. P. Nguyen, H. T. Ng, T. Yamada, M. Smith, J. Li, J. Han, and M. Meyyappan, Nano Lett. 4, 651 (2004). 48. J. Thong, C. H. Oon, V. Yeadon, and W. D. Zhang, Appl. Phys. Lett. 81, 4823 (2002). 49. T. Kamins, M. Saif Islam, S. Sharma, and R. Stanley Williams, in Proceedings of IEEE NANO 2004, ed. D. Muenchen, (1719 August, 2004). 50. D. S. Hopkins, D. Pekker, P. M. Goldbart, and A. Bezryadin, Science 308, 1762 (2005). 51. P. Candeloro, E. Comini, C. Baratto, G. Faglia, G. Sberveglieri, R. Kumar, A. Carpentiero, and E. Di Fabrizio, J. Vac. Sci. Technol. B 23, 2784 (2005). 52. A. S. Zuruzi, A. Kolmakov, N. C. MacDonald, and M. Moskovits, Appl. Phys. Lett. 88, 102904 (2006). 53. Y. S. Kim, S. C. Ha, K. W. Kim, H. Yang, S. Y. Choi, Y. T. Kim, J. T. Park, C. H. Lee, J. Y. Choi, J. S. Paek, and K. Lee, Appl. Phys. Lett. 86, 213105 (2005).

21
54. Q. Wan, Q. H. Li, Y. J. Chen, T. H. Wang, X. L. He, J. P. Li, and C. L. Lin, Appl. Phys. Lett. 84, 3654 (2004). 55. I. Raiblea, M. Burghardb, U. Schlechtb, A. Yasudaa, and T. Vossmeyer, Sens. Actuators B 106, 730 (2005). 56. G. Y. Zhang, B. Guo, and J. Chen, Sens. Actuators B 114, 402 (2006). 57. Y. L. Liu, H. F. Yang, Y. Yang, Z. M. Liu, G. L. Shen, and R. Q. Yu, Thin Solid Films 497, 355 (2006). 58. S. Kar, B. N. Pal, S. Chaudhuri, and D. Chakravorty, J. Phys. Chem. B 110, 4605 (2006). 59. J. Chen, L. Xu, W. Li and X. Gou, Adv. Mater. 17, 582 (2005). 60. Y. J. Chen, X. Y. Xue, Y. G. Wang, and T. H. Wang, Appl. Phys. Lett. 87, 233503 (2005). 61. C. Baratto, E. Comini, G. Faglia, G. Sberveglieri, M. Zha, and A. Zappettini, Sens. Actuators B 109, 2 (2005). 62. X. H. Jiang, Y. L. Wang, T. Herricks, and Y. N. Xi, J. Mater. Chem. 14, 695 (2004). 63. Y. L. Wang, X. H. Jiang, and Y. N. Xia, J. Am. Chem. Soc. 125, 16176 (2003). 64. Z Ying, Q. Wan, Z T Song, and S. L. Feng, Nanotechnology 15, 1682 (2004). 65. X. L. Li, T. J. Lou, X. M. Sun, and Y. D. Li, Inorg. Chem. 43, 5442 (2004). 66. X. F. Chu, C. H. Wang, D. L. Jiang, and C. M. Zheng, Chem. Phys. Lett. 399, 461 (2004). 67. J. F. Liu, X. Wang, Q. Peng, and Y. D. Li, Adv. Mater. 17, 764 (2005). 68. W. Y. Li, L. N. Xu and J. Chen, Adv. Funct. Mater. 15, 851 (2005). 69. G. Faglia, C. Baratto, G. Sberveglieri, M. Zha, and A. Zappettini, Appl. Phys. Lett. 86, 011923 (2005). 70. Z. Y. Sun, H. Q. Yuan, Z. M. Liu, B. X. Han, and X. R. Zhang, Adv. Mater. 17, 2993 (2005). 71. Z. Y. Zhang, H. J. Jiang, Z. Xing, and X. R. Zhang, Sens. Actuators B 102, 155 (2004). 72. K. M. Ip, Z. Liu, C. M. Ng, and S. K. Hark, Nanotechnology 16, 1144 (2005). 73. Y. Zhang, K. Yu, S. Quang, L. Luo, H. Hu, Q. Zhang, and Z. Zhu, Physica B 368, 94 (2005). 74. X. H. Wang, J. Zhang, and Z. Q. Zhu, Appl. Surface Sci. 252, 2404 (2006). 75. H. Kind, H. Yan, B. Messer, M. Law, and P. Yang, Adv. Mater. 14, 158 (2002). 76. V. E. Henrich and P. A. Cox, Surface Science of Meal Oxides. (Cambridge, U.K.: publisher name, 1996). AQ5 77. S. H. Hahn, N. Barsan, U. Weimar, S. G. Ejakov, J. H. Visser, and R. E. Soltis, Thin Solid Films 436, 17 (2003). 78. K. D. Mitzner, J. Sternhagen, and D. W. Galipeau, Sens. Actuators B 93, 92 (2003). 79. P. Mitra, A. P. Chatterjee, and H. S. Maiti, Mater. Lett. 35, 33 (1998). 80. J. F. Chang, H. H. Kuo, I. C. Leu, and M. H. Hon, Sens. Actuators B 84, 258 (1994). 81. G. Y. Zhang and J. Chen, J. Electrochem. Soc. 152, A2069-A2073 (2005). 82. T. Gao and T. H. Wang, Appl. Phys. A 80, 1451 (2005). 83. B. D. Yao, Y. F. Chan, and N. Wang, Appl Phys Lett. 81, 757 (2002). 84. Y. F. Yan, M. M. Al-Jassim, and S. H. Wei, Phys. Rev. B 72, 161307(R) (2005). 85. S. O. Kucheyev, T. F. Baumann, P. A. Sterne, Y. M. Wang, T. van Buuren, A. V. Hamza, L. J. Terminello, and T. M. Willey, Phys. Rev. B 72, 035404 (2005). 86. D. H. Zhang, C. Li, X. L. Liu, S. Han, T. Tang, and C. W. Zhou, Appl. Phys. Lett. 83, 1845 (2003). 87. X. F. Chu, D. L. Jiang, Aleksandra B. Djuris  ic, and H. L. Yu, Chem. Phys. Lett. 401, 426 (2005). 88. S. Samson and C. G. Fonstaad, J. Appl. Phys. 44, 4618 (1973).

AQ3

AQ4

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:48 User ID: bhuvaneswaric Enabled Page Number: 22

BlackLining

22
89. N. Birks and G. H. Meier, Introduction to high temperature oxidation of Metals. (London: Edward Arnold Ltd., 1983). 90. T. Minami, H. Sato, H. Nanto, and S. Takata, Jpn. J. Appl. Phys. Part 2, 24, L781 (1985). 91. H. Kato, M. Sano, K. Miyamoto, and T. Yao, J. Cryst. Growth 538, 237 (2002). 92. K. Ip, M. E. Overberg, Y. W. Heo, D. P. Norton, S. J. Pearton, C. E. Stutz, B. Luo, F. Ren, D. C. Look, and J. M. Zavada, Appl. Phys. Lett. 82, 385 (2003). 93. Y. W. Heo, L. C. Tien, D. P. Norton, B. S. Kang, F. Ren, B. P. Gila, and S. J. Pearton, Appl. Phys. Lett. 85, 2002 (2004). 94. J. A. Dean, Langes Handbook of Chemistry (Chinese ed.). (LocaAQ6 tion: Science Press, 2003). 95. C. H. Wang, X. F. Chu, and M. M. Wu, Sens. Actuators B 113, 320 (2006). 96. D. C. Look, J. W. Hemsky, and J. R. Sizelove, Phys. Rev. Lett. 82, 2552 (1999). 97. R. W. J. Scott, S. M. Yang, G. Chabanis, N. Coombs, D. E. Williams, and G. A. Ozin, Adv. Mater. 13, 1468 (2001). 98. D. E. Williams, in Solid State Gas Sensors, ed. P. T. Moseley and B. C. Tofield, (Bristol, UK: Adam Hilger, 1987), 71. 99. H. Ogawa, M. Nishikawa, and A. Abe, J. Appl. Phys. 53, 4448 (1982). 100. P. Qi, P, O. Vermesh, M. Grecu, A. Javey, Q. Wang, and H. J. Dai, Nano Lett. 3, 347 (2003). 101. Y. S. Shen and T. S. Zhang, Sens. Actuators B 12, 5 (1993). 102. X. H. Wu, Y. D. Wang, Z. H. Tian, H. L. Liu, Z. L. Zhou, and Y. F. Li, Solid-State Electron. 46, 715 (2002).

Gas Sensors Based on One-Dimensional Nanostructures


103. G. Singh, M. H. Miles, and S. Srinivasan, in Electrocatalysis on Nonmetallic Surfaces, ed. A. D. Franklin. (Washington, DC: U.S. Government Printing Office, 1976) 289. 104. X. Y. Xue, Y. J. Chen, Y. G. Wang, and T. H. Wang, Appl. Phys. Lett. 86. 233101 (2005). 105. S. Khoobiar, J. Phys. Chem. 68, 411 (1964). 106. S. V. Manorama, C. V. G. Reddy, and V. J. Rao, Appl. Surf. Sci. 174, 93 (2001). 107. Y. X. Liang, Y. J. Chen, and T. H. Wang, Appl. Phys. Lett. 85, 5682 (2004). 108. X. W. Xu, J. Wang, and Y. C. Long, Micropor. Mesopor. Mat. 83, 60 (2005). 109. A. Kolmakov, D. O. Klenov, Y. Lilach, S. Stemmer, and M. Moskovits, Nano Lett. 5, 667 (2005). 110. J. Q. Xu, Y. P. Chen, D. Y. Chen, and J. N. Shen, Sens. Actuators B 113, 526 (2006). 111. Q. Wan and T. H. Wang, Chem. Commun. 30, 3841 (2005). 112. C. M. Carney, S. Yoo, and S. A. Akbar, Sens. Actuators B 108, 29 (2005). 113. Z. G. Lu and Y. G. Tang, Mat. Chem. Phys. 92, 5 (2005). 114. X. H. Kong and Y. D. Li, Sens. Actuators B 105, 449 (2005). 115. T. Gao and T. H. Wang, Chem. Commun. 22, 2558 (2004). 116. G. S. Devi, S. Manorama, and V. J. Rao, Sens. Actuators B 28, 31 (1995). 117. X. H. Zhou, Q. X. Cao, H. Huang, P. Yang, and Y. Hu, Mater. Sci. Eng. B 99, 44 (2003). 118. G. Faglia, C. Baratto, G. Sberveglieri, M. Zha, A. Zappettini, Appl. Phys. Lett. 86, 011923 (2005).

Path: K:/ASP-TSENG-07-0601/Application/ASP-TSENG-07-0601-B04.3d Date: 26th March 2008 Time: 11:48 User ID: bhuvaneswaric Enabled Page Number: 23

BlackLining

Gas Sensors Based on One-Dimensional Nanostructures

23
Author Queries

AQ1: Is this correct, or should it be (100)? AQ2: Is the sensitivity changed in my way? e.g., increase sensitivity? AQ3: Please provide publisher and publisher location, and page range. AQ4: Please provide publisher name and location, and page range. AQ5: Who is the publisher? AQ6: What is the location of the publisher? AQ7: Should there be a gas for this LDC in the second column? AQ8: Please provide the missing datas for column head "Response time".

You might also like