On the role of individual metal oxide nanowires in the scaling down of chemical sensors

Abstract

Single-crystalline semiconductor metal oxide nanowires exhibit novel structural and electrical properties attributed to their reduced dimensions, well-defined geometry and the negligible presence of grain boundaries and dislocations in their inside. This favours direct chemical transduction mechanisms at their surfaces upon exposure to gas molecules, making them promising active device elements for a new generation of chemical sensors. Furthermore, metal oxide nanowires can be heated up to the optimal operating temperature for gas sensing applications with extremely low power consumption due to their small mass, giving rise to devices more efficient than their nanoparticle-based counterparts. Here, the current status of development of sensors based on individual metal oxide nanowires is surveyed, and the main technological challenges which act as bottleneck to their potential use in real applications are presented.

---

1. Introduction

Representative metal oxide semiconductors have unique properties perfect for several applications such as sensors, photocatalyzers, and varistors.^1–5 The basic working principle of these materials as gas sensors is based on the chemico-electrical transduction reactions which take place at the interface between molecular adsorbents and the metal oxide’s surfaces.^4–6 However, their final performance is highly determined by manifold experimental parameters such as the concentration of oxygen vacancies, which are originated during the synthesis of metal oxides; and the intrinsic carrier concentration in their inside.^5–8 Commercial sensors are commonly based on thin layers of metal oxide semiconductors deposited onto hotplates which are used to fix the temperature at the optimal values and thus, activate the surface transduction mechanisms necessary to detect the gas species^4,9 (Fig. 1a). Although this design has become a success story from a commercial point of view,^9,10 it presents two major drawbacks: (1) high power consumption which hinders its use in portable and autonomous systems and, (2) poor stability derived from multiple causes such as the grain boundary contribution among nanoparticles inside the sensing layers.¹¹ Consequently, making low-consumption, stable and device-quality metal oxide sensors remains a challenging issue, and puts this technology at a disadvantage to others (i.e.electrochemical cells).⁹

Fig. 1 Schematic diagrams of different types of conductometric gas sensors based on metal oxides. (a) Commercial thin-film sensor formed by a layer of nanoparticles. Here, electrons must go through a network of nanocrystals with different size and shape. From an energy point of view, electrons are to overcome potential barriers [(i) metal–semiconductor barriers (eV_C) and (ii) intergrain boundary barriers (eV_B)]. The overall influence of gas on the height of the barriers determines the final response of the sensor. This is equivalent to a network of resistors [(i) metal–semiconductor contacts (R_C), (ii) grain boundary interfaces (R_B) and (iii) metal-oxide grains (R_G)]. (b) Multi-nanowire sensor. The above mentioned discussion is valid here as well. (c) Single-nanowire sensor. If a nanowire is measured in 4-probe DC configuration, the conductometric response is basically determined by changes of the conduction channel along the nanowire (R_NW). On the contrary, contacts effects are overcome.

In view of the unique properties of metal oxide nanowires, such as well-defined geometry, high surface-to-volume ratio and good crystallinity;^1,5 intensive research efforts have been devoted to develop a new generation of sensors. This situation has given rise to a fast and significant progress in the synthesis and characterization of nanowires; and their final integration in proof-of-concept devices.^12–23 However, the controlled manipulation and characterization of nanowires is not a straightforward process due to the intrinsic problems of working at the nanoscale. For instance, the reduced contact area between metal electrodes and nanowires magnifies the contribution of the contact electrical properties, hiding the phenomena which take place on the surface of nanowires.^22–25

To accomplish electrical measurements free of parasitic effects and develop competitive sensors, different fabrication and characterization strategies have been successfully evaluated up to now, such as following low-current measurement protocols in individual nanowires which allow long-term device operations without degradation of their properties.²⁶ Thus, it can be established that the present state of development of nanowire-based technologies guarantees a complete and well-controlled characterization of proof-of-concept devices, overcoming most of the abovementioned matters which were considered insurmountable obstacles in the past.⁵

One of the main advantages of using individual nanowires as building blocks of sensor prototypes is their potential to provide a deeper comprehension of the fundamental adsorption mechanisms of gas molecules onto metal oxides than the obtained one with thin-film sensors based on layers formed by randomly oriented nanoparticles, since they make up better-defined and easier to model experimental scenarios⁵ (Fig. 2). This gives way to gaining a deeper insight into the chemico-electrical transduction mechanisms which rule the performance of metal-oxide sensors, and may become an extremely helpful tool for improving this technology in the future.²⁷

Fig. 2 (a) Diagram illustrating different types of necks in polycrystalline metal oxide matrices derived from the random nature of the layer. and Thick and thin necks between grains: the thicker the neck, the easier the electron transfer. and Intergrain boundary without material continuity: if the intergrain distances is short enough, electron transfer will take place by tunnel-assisted mechanisms.⁶⁴ Non homogeneous intergrain interface. In this scenario the electric field that causes the band bending near the surface due to the gas interaction (E_bar) and the electrical field externally applied to perform the conductometric measurements (E_bias) are not necessarily orthogonal. (b) Diagram illustrating the gas interaction in individual nanowires: any intergrain necks or boundaries are considered. Moreover, E_bar and E_bias fields are always orthogonal and independent.

On the other hand, the low mass of individual nanowires facilitates heating them up.^28,29 In this work, the design strategies as well as the operating principles of sensors based on individual nanowires with power consumption of only a few tens of microwatts are presented. This means a significant reduction of energy needs compared to thin-film sensors mounted onto micro-hotplates, which usually require milliwatts to be operated in continuous mode.²⁶

However, the present state of technological development only allows the fabrication of single-nanowire prototypes (Fig. 1c), whereas the integration in large-scale and low-cost processes remains an unsolved challenge.

For all these reasons, the main aim of this work is to present the current stage of development of a new line of research in metal oxide sensors, which has been extremely productive in the last years; and to highlight their potential advantages; the use of nanowires as building blocks of innovative but at the same time not fully developed devices may give a chance to find the solution of problems which remain partially unsolved in standard technologies, such as the lack of stability, poor selectivity towards different gas species and power consumption. Here, the recent breakthroughs in nanowire-based sensors as well as the major unsolved technological challenges are clearly identified. Nevertheless, clearing up the issues which hamper the massive use of this technology in commercial devices is a very ambitious goal completely out of the scope of this work, and will require intensive research and development efforts in the future.

2. On the advantages of individual nanowire-based sensors

Fig. 1a shows the typical configuration of commercial metal oxide sensors formed by a film layer of nanoparticles, which are heated in order to activate adsorption/desorption of gas molecules at their surfaces.^4,11 The operating principle of these devices is based on their conductance modulation under exposure to gas. In a simplistic approximation, grain boundaries are described by highly resistive barriers, which represent the main contribution to the overall device resistance (Fig. 1a). Consequently, the modulation of barriers among nanoparticles constitutes the most important gas transduction mechanism in porous-film sensors.^9,11 However, the random aggregation of nanoparticles in their inside as well as the spread in size make difficult the accurate study of the gas transduction phenomena at the nanoscale level. The particular geometry of each neck and boundary, and the unique orientation of the adsorbate-modulated electric field (E_bar) regarding the externally applied bias field (E_bias) present in conductometric measurements (Fig. 2a) make it necessary to describe thin-film sensors with simplified models, and to analyze their responses as the convolution of diverse contributions.⁹ The same conclusions are found if bundles of nanowires are used instead of nanoparticles, since randomly oriented boundary effects are not eliminated either (Fig. 1b). Although this sort of devices has enabled reaching a good comprehension of the fundamental operating principles of metal oxide sensors, their geometry constitutes an impediment to performing fine analysis of the gas–surface interactions, since uncontrolled characteristics of necks and boundaries among nanoparticles has a determining influence in the gas response, as Williams reported elsewhere.^30,31

To contrast, these uncontrolled contributions are circumvented if individual nanowire-based sensors are studied (Fig. 1c).

In this scenario, E_bar is always orthogonal to E_bias (Fig. 2a), and direct responses towards gases are monitored, because gas diffusion among nanograins/nanowires is completely eliminated (see ESI‡). Thus, the sensing principle of individual metal oxide nanowires can be theoretically described by pure surface effects.^5–7 Adsorption of gas molecules at the nanowire modulates the width of a depleted region and the intensity of the associated electrical field E_bar (Fig. 2) close to the external shell. This effect based on the capture and release of electrical charges at the surface modifies the conduction channel through the nanowire and as a consequence the electrical resistance R_NW (Fig. 1c and 3), providing direct and fundamental information of the electrical charge exchange and the role of surface states n_v in the sensing process.^6,7 According to this model, R_NW under exposure to gas is given by:

where ρ is the nanowire resistivity, L the nanowire’s length, r the nanowire’s radius and λ the width of the depletion layer created by adsorbed molecules (Fig. 3). Eqn (1) establishes a direct connection between the nanowire’s radius and the gas response: the thinner the nanowire is, the higher the resistance modulation will be observed (Fig. 2b and 3). Thus, the highest responses are always measured with ultrathin nanowires (radius below 40 nm).^7,32

Fig. 3 Response of SnO₂nanowires to synthetic air/nitrogen pulses as function of their radii measured at T∼ 300 °C. Higher responses are clearly observed with shrinking dimensions in correspondence with eqn (1) (solid line).

It is noteworthy that the reduction of the sensing area of nanowire-based devices does not mean a decrease of their responses towards gases, since the modulation of the electrical resistance R_NW is directly proportional to the number of surface sites n_v per unit area interacting with gas molecules, and regardless of the total effective surface. That is to say, the same relative gas response is monitored with a single nanowire prototype than the one obtained with multiple nanowire-based devices contacted in parallel, provided that the nanowires’ radii are the same in both cases.⁶³

To guarantee the validity of the former assertion, gas concentration in air must be high enough to assure the interaction of gas molecules with most of the surface sites n_v in a time scale shorter than the typical response times towards gas species. With regard to this point, it can be demonstrated that the interval collision between molecules and nanowires (even for the thinnest ones, r = 20 nm) at typical gas concentrations (from 0.1 to 1.000 ppm) is always orders of magnitude shorter than the typical dynamic response of these sensors (see ESI‡).

Finally, it should be highlighted that the absence of nooks and crannies in nanowire-based devices facilitates direct adsorption/desorption of gas molecules, improving the dynamic behaviour of these prototypes in comparison to those observed with porous-film sensors, in which gas diffusion dominates their dynamic response^33–38 (see ESI‡). In this case, interactions between gas molecules and metal oxide nanowires occurs in a simplified scenario, which facilitates an easier comparison between experimental data and the results obtained in simulation works in which almost ideal metal oxide’s surfaces are always considered.^39,40 The good agreement between the experiments and simulated systems found up to now^39,40 paves the way to gaining a deeper comprehension of the fundamental sensing mechanisms ruling the behaviour of metal oxides, and may become an extremely helpful tool to engineer new and better devices in the future.

3. On the issues derived from working with ultrathin individual nanowires

For the reasons described thus far in the former section, ultrathin nanowires (r < 40 nm) are optimal for obtaining the highest gas responses, but how to interface and electrically access these minute materials has become a complex and partially unsolved issue.

On the one hand, the need to obtain good electrical contacts in a controlled and reproducible process has forced to look for innovative nanofabrication approaches. Nowadays, metal stripes with well-defined shapes in the nanometer range and high electrical quality are easily fabricated with different techniques such as focused ion beam (FIB-),^22,26 e-beam-⁴¹ or UV- and shadow-mask-lithography,⁴² enabling the fast engineering of advanced proof-of-concept devices (Fig. 1c). Nevertheless, most of these techniques and particularly the nanofabrication protocols based on them evaluated so far are only suitable for research prototyping, since they are not scalable, and therefore do not fulfil the requirements for taking a leap in industry.

For this reason, new solutions to solve the problem of scalability of nanowire-based devices and to edit the first complex circuit with them are currently under evaluation, such as self-assembly approaches,^43,44 and the use of other techniques like electrospinnig and microcontact/ink-jet printing.^45–49 In the former approach, metal oxide nanowires are directly deposited by a high-voltage-driven injection nozzle onto the electrodes. In the latter one, the metal electrodes are precisely printed on top of the nanowires.

Besides enabling the integration of nanowires in industrial processes, the chosen solution should palliate the poor reproducibility of present proof-of-concept devices (compared to the standards of the microelectronic industry), which is mainly originated by the lack of well-established fabrication and characterization methodologies. However, it should be highlighted that all these fabrication techniques are still in a preliminary stage of development despite some promising results were recently reported.⁵⁰

On the other hand, performing electrical measurements on these nanodevices and modelling the experimental data remains a non-trivial process since the reduced nanocontact area between metal electrodes and the nanowire magnifies the contribution of electrical contacts.^22–25 Rectifying interfaces are usually formed between metal oxide nanowires and typical metal stripes (i.e.Au, Pt, W),^22,24 whereby it becomes essential that their contribution to the overall response is correctly modelled to avoid false interpretations of the electrical parameters of nanowires due to contact resistance underestimations.^22,25 Yet, this limitation can not be considered anymore an insurmountable obstacle in the study of metal oxide nanowires. To date, several models have been reported to clarify the electrical behaviour of nanowires studied in 2-probes configuration, which are commonly described by two back-to-back Schottky diodes in series with the nanowire’s resistance, categorizing the different types of conduction mechanisms through the contacts as well as establishing design strategies capable to minimize the contact contribution to the device’s responses.^22,25 In fact, some fabrication and operation methodologies, such as performing 4-probes measurements or heating the device (Fig. 1c) are strongly recommended to minimize undesired contact contributions.²⁵

The third important issue related to the development of ultrathin nanowire-based sensors is related to the self-heating effects produced during their operation. The high contact resistance at metal–nanowire interfaces and the small geometrical section of nanowires minimize the effective area to dissipate the heat generated by the Joule effect when probing current I flows through them.^25,28,29 This leads to a significant rise of the temperature T, and for this reason it must be taken into account in order to minimize an accelerated degradation of the operational life of prototypes, since dissipated power values in the range of milliwatts are high enough to destroy both the contacts and the nanowire²⁵ (Fig. 4). This issue is usually circumvented if the suitable experimental conditions are selected. That is to say, by applying at any time probing current intensities lower than I = 100 nA the device can operate for weeks without evidence of degradation.^25,26 Although self-heating was considered a major drawback in the past, recent studies demonstrated that it is possible to use it for heating nanowires in a controlled way, and thus to optimize their response towards gases.^28,29 This operation strategy significantly reduces energy consumption in these nanosensors, which are continuously operated with only a few tens of microwatts,²⁸ and envisages new industrial applications for this sort of materials.^51,52 With regard to this subject, it should be noted that intensive efforts to develop room temperature nanowire-based sensors based on technologies different from self-heating are also under evaluation. In particular, light-operated gas sensors are considered the most promising alternative to self-heating, as is demonstrated elsewhere.^53,54

Fig. 4 SEM images of the failure of a single nanowire device due to self-heating effect (r≈ 50 nm). (a) before and (b) after applying an uncontrolled current peak higher than 100 mA.

As far as the selectivity issue is concerned, metal oxide nanowires display the same limitations than conventional metal oxide sensors, which are affected by important crossed-sensitivity effects.^55,56 On the contrary, the high quality of their crystallographic structure as well as the good stability of their surfaces open the door to the design of stable devices with extremely long life spans. Detailed structural analyses of metal oxide nanowires produced with state-of-the-art synthesis routes can be found elsewhere.⁵⁷

In short, nanowire-based sensors face up to some issues to be transferred to industrial production, but they also exhibit some intrinsic advantages compared to their nanoparticle-based counterparts. For instance, their final response level can be accurately selected by choosing the diameter of nanowires (Fig. 3). Unresolved challenges for the future are manifold, such as controlling the doping level in their inside or designing heterostructured geometries. These desirable advances could significantly improve the sensing characteristics of future devices.

4. Proof-of-concept chemical sensors based on individual nanowires

Among the large family of metal oxide nanowires synthesized up to now, SnO₂, In₂O₃ and ZnO nanowires are usually considered the best candidates to demonstrate the feasibility of monitoring their electrical response towards different external stimuli, due to their low cost and well-known properties.^12,15 In particular, SnO₂ has been traditionally used to fabricate metal oxide conductometric gas microsensors.^7,9,11,12

From a technological point of view, it is a major challenge to fabricate a competitive sensor based on individual nanowires. For this reason, the foreseeable advantages related to this potential technology are giving a boost to intensive research in the field.⁵ The first one in comparison to traditional SnO₂ microsensors is the excellent recovery of the initial resistance baseline when the target gas is removed from the chamber, since most of drift causes typical of thin film sensors are eliminated, as was stated in the former section. It should be noted that the recovery time of individual nanowire sensors can be significantly accelerated, illuminating them with UV lamps, since impinging photons clean their surface through a well-controlled photodesorption induced process of adsorbed gas species.^53,58,59

To date, it has been demonstrated that obtaining gas sensors based on metal oxide nanowires to both oxidizing and reducing species is feasible (Fig. 5),^5,11,55 and these first prototypes stand out by their low energy consumption.²⁸ This means a major technological breakthrough in the field of metal oxide sensors, since self-heating effects described in the previous sections can be used to optimise the operating conditions of nanowires and their sensitivity towards gas species as function of the probing current, avoiding the requirements of external heaters and ensuring fast dynamic responses of the sensor system (see ESI‡).

Fig. 5 Sensing response of a metal oxide nanowire-based device. Gas response of a SnO₂nanowire operated at 175 °C towards NO₂ pulses of different concentrations. The operating temperature is modified through self-heated originated by probing current I.

These first self-heated proof-of-concept systems exhibited responses equivalent to those obtained with the conventional approach, but they lower the required power from the milliwatt to the microwatt range.²⁸ This result represents an important advance in power efficiency and miniaturization, making these devices especially attractive to be integrated in a new generation of portable devices.

Finally, it should be noted that the use of metal oxide nanowires is currently not restrained for gas sensing applications. Their integration in optoelectronic devices, such as UV photodetectors, has been successfully evaluated as well.⁶⁰

5. Future challenges

The first generation of portable prototypes based on individual nanowires has already been developed and successfully evaluated.^26,61,62 Nevertheless, these devices are still too complex and expensive to fabricate in scalable processes. Furthermore, they do not fulfil many other industrial requirements to be transferred into the real world, because their state of development is still in its infancy. This assertion is especially significant in the field of gas sensors, since any competitive commercial device has appeared on the market after some years of intensive basic research. In fact, there is an increasing interest in developing low-cost portable devices with integrated nanowires as functional elements to evaluate their market potential. Although different self-assembly and more advanced fabrication techniques have been successfully evaluated, such as dielectrophoresis^43,44 and electrospinning,⁴⁵ the use of nanowires remains restrained to the academic level. For this reason, the integration issue has become one of the main priorities in nanotechnology research. On the other hand, the development of complex electronic architectures such as e-nose devices based on nanowires and their controlled surface functionalization is the second focus of interest, as is shown by recent studies.^5,56,63 These works should help to improve selectivity towards gases of metal oxide chemical sensors in the future.

6. Conclusions

The use of nanowires as chemical sensors has potential advantages compared to traditional thin-film devices due to the intrinsic properties of nanowires. This has brought about intensive research efforts onto this new area to obtain the first proof-of-concept devices. However, the manipulation and characterization of these minute materials is a complex process which requires well-established methodologies not yet fully developed. Nanowire-based chemical sensors may become extremely helpful for gaining a deeper insight into the interaction mechanisms between gas molecules and metal oxide semiconductors; and their integration into portable systems would be feasible due to their low energy consumption, overcoming one of the main limitations of present metal oxide sensors. However, the use of nanowires in real devices is still in a preliminary stage, becoming a major challenge for the future how to integrate them with low-cost industrial processes.

Acknowledgements

This work was partially supported by the Spanish Government [projects N—MOSEN (MAT2007-66741-C02-01), and MAGASENS], the UE [project NAWACS (NAN2006-28568-E), the Human Potential Program, Access to Research Infrastructures]. J.D.P. and R.J.D. are indebted to the MEC for the FPU grant. Thanks are due to the European Aeronautic Defense and Space Company (EADS N.V.) for supplying the suspended micromembranes. S.M. thanks the University of Cologne for financial support.

References

1. M. Law, J. Goldberger and P. Yang, Annu. Rev. Mater. Res., 2004, 34, 83.
2. S. V. N. T. Kuchibhatla, A. S. Karaoti, D. Bera and S. Seal, Prog. Mater. Sci., 2007, 52, 699.
3. G. Eranna, B. C. Joshi, D. P. Runthala and R. P. Gupta, Crit. Rev. Sol. State Mater. Sci., 2004, 29, 188.
4. G. Korotcenkov, Sens. Actuators, B, 2005, 107, 209.
5. E. Comini, C. Baratto, G. Faglia, M. Ferroni, A. Vomiero and G. Sberveglieri, Prog. Mater. Sci., 2009, 54, 1.
6. A. Kolmakov, D. O. Klenov, Y. Lilach, S. Stemmer and M. Moskovits, Nano Lett., 2005, 5, 667.
7. F. Hernandez-Ramirez, J. D. Prades, A. Tarancon, S. Barth, O. Casals, R. Jimenez-Diaz, E. Pellicer, J. Rodriguez, J. R. Morante, M. A. Juli, S. Mathur and A. Romano-Rodriguez, Adv. Funct. Mater., 2008, 18, 2990.
8. Y. Zhang, A. Kolmakov, Y. Lilach and M. Moskovits, J. Phys. Chem. B, 2005, 109, 1923.
9. N. Barsan, D. Koziej and U. Weimar, Sens. Actuators, B, 2007, 121, 18.
10. There are many companies offering this type of sensors, such as Figaro, FIS, MICS, UST, CityTech, Applied-Sensors, NewCosmos, etc. See information provided by the gas sensors manufacturers on their homepages: (i) http://www.figaro.co.jp; (ii) http://www.fisinc.co.jp; (iii) http://www.appliedsensors.com; (iv) http://www.citytech.com; (v) http://www.microchem.com.
11. G. Korotchenkov, Mater. Sci. Eng., R, 2008, 61, 1.
12. S. Mathur, R. Ganesan, I. Grobelsek, H. Shen, T. Ruegamer and S. Barth, Adv. Eng. Mater., 2007, 9, 658.
13. Z. W. Pan, Z. R. Dai and Z. L. Wang, Science, 2001, 291, 1947.
14. Z. L. Wang, J. Phys.: Condens. Matter, 2004, 16, R829.
15. Z. L. Wang, Mater. Today (Oxford, U. K.), 2004 , June 26.
16. X. D. Wang, J. Zhou, J. H. Song, J. Liu, N. S. Xu and Z. L. Wang, Nano Lett., 2006, 6, 2768.
17. J. Zhou, N. S. Xu and Z. L. Wang, Adv. Mater., 2006, 18, 2432.
18. Z. L. Wang and J. H. Song, Science, 2006, 14, 242.
19. J. H. He, C. L. Hsin, J. Liu, L. J. Chen and Z. L. Wang, Adv. Mater., 2007, 19, 781.
20. A. Fort, S. Rocchi, M. B. Serrano-Santos, M. Mugnaini, V. Vignoli, A. Atrei and R. Spinicci, Sens. Actuactors B: Chem., 2006, 116, 43.
21. F. Hernández-Ramírez, A. Tarancón, O. Casals, J. Arbiol, A. Romano-Rodríguez and J. R. Morante, Sens. Actuators, B, 2007, 121, 3.
22. F. Hernandez-Ramirez, A. Tarancon, O. Casals, J. Rodríguez, A. Romano-Rodriguez, J. R. Morante, S. Barth, S. Mathur, T. Y. Choi, D. Poulikakos, V. Callegari and P. M. Nellen, Nanotechnology, 2006, 17, 5577.
23. C. Y. Nam, D. Tham and J. E. Fischer, Nano Lett., 2005, 5, 2029.
24. Y.-F. Lin and W.-B. Jian, Nano Lett., 2008, 8, 3146.
25. F. Hernandez-Ramirez, A. Tarancon, O. Casals, E. Pellicer, J. Rodríguez, A. Romano-Rodriguez, J. R. Morante, S. Barth and S. Mathur, Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 76, 085429.
26. F. Hernandez-Ramirez, J. D. Prades, A. Tarancon, S. Barth, O. Casals, R. Jimenez-Diaz, E. Pellicer, J. Rodríguez, M. A. Juli, A. Romano-Rodriguez, J. R. Morante, S. Mathur, A. Helwig, J. Spannhake and G. Mueller, Nanotechnology, 2007, 18, 495501.
27. Z. L. Wang, Adv. Mater., 2000, 12, 1295.
28. J. D. Prades, R. Jimenez-Diaz, F. Hernandez-Ramirez, S. Barth, A. Cirera, A. Romano-Rodriguez, S. Mathur and J. R. Morante, Appl. Phys. Lett., 2008, 93, 123110.
29. E. Strelcov, S. Dmitriev, B. Button, J. Cothren, V. Sysoev and A. Kolmakov, Nanotechnology, 2008, 19, 355502.
30. R. W. J. Scott, S. M. Yang, G. Chabanis, N. Coombs, D. E. Williams and G. A. Ozin, Adv. Mater., 2001, 13, 1468.
31. R. W. J. Scott, S. M. Yang, N. Coombs, G. A. Ozin and D. E. Williams, Adv. Funct. Mater., 2003, 13, 225.
32. S. Dmitriev, Y. Lilach, B. Button, M. Moskovits and A. Kolmakov, Nanotechnology, 2007, 18, 055707.
33. J. Maier, Festkörper-Fehler und Funktion: Prinzipien der Physikalischen Festkörperchemie, Teubner Studienbücher Chemie, Teubner, Stuttgart, 2000.
34. Y. Min, PhD Thesis, Massachusetts Institute of Technology, 2003.
35. G. Korotcenkov, Sens. Actuators, B, 2007, 121, 664.
36. G. Korotcenkov, V. Brinzari, J. R. Stetter, I. Blinov and V. Blaja, Sens. Actuators, B, 2007, 128, 51.
37. B. Kamp, R. Merkle, R. Lauck and J. Maier, J. Solid State Chem., 2005, 178, 3027.
38. M. A. Ponce, C. M. Aldao and M. S. Castro, J. Eur. Ceram. Soc., 2003, 23, 2105.
39. M. Batzill and U. Diebold, Prog. Surf. Sci., 2005, 79, 47.
40. J. D. Prades, A. Cirera and J. R. Morante, in Applications of DFT Calculations to Chemical Gas Sensors: Designing and Understanding, Quantum Chemical Calculations of Surface and Interfaces of Materials, ed. V. A. Basiuk and P. Ugliengo, American Scientific Publishers, Valencia, CA, 2008, ISBN: 1-58883-138-8.
41. D.-W. Kim, I.-S. Hwang, S. J. Kwon, H.-Y. Kang, K.-S. Park, Y.-J. Choi, K.-J. Choi and J.-G. Park, Nano Lett., 2007, 7, 3041.
42. S. V. Kalinin, J. Shin, S. Jesse, D. Geohegan, P. Baddorf, Y. Lilach, M. Moskovits and A. Kolmakov, J. Appl. Phys., 2005, 98, 044503.
43. S. Kumar, S. Rajaraman, R. A. Gerhardt, Z. L. Wang and P. J. Hesketh, Electrochim. Acta, 2005, 51, 943.
44. D. Wang, R. Zhu, Z. Zhaoying and X. Ye, Appl. Phys. Lett., 2007, 90, 103110.
45. D. Lin, H. Wu, R. Zhang and W. Pan, Nanotechnology, 2007, 18, 465301.
46. K. M. Sawicka, A. K. Prasad and P. I. Gouma, Sensors Lett., 2005, 3, 31.
47. T. H. J. van Osch, J. Perelaer, A. W. M. de Laat and U. S. Schubert, Adv. Mater., 2008, 20, 343.
48. Y.-K. Kim, S. J. Park, J. P. Koo, D.-J. Oh, G. T. Kim, S. Hong and J. S. Ha, Nanotechnology, 2006, 17, 1375.
49. J.-W. Song, J. Kim, Y.-H. Yoon, B.-S. Choi, J.-H. Kim and C.-S. Han, Nanotechnology, 2008, 19, 095702.
50. Z. Fan, J. C. Ho, Z. A. Jacobson, H. Razavi and A. Javey, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 11066.
51. Spanish Patent P200900334.
52. J. D. Prades, R. Jimenez-Diaz, F. Hernandez-Ramirez, A. Cirera, A. Romano-Rodriguez and J. R. Morante, Individual Nanowire Chemical Sensor System Self-powered with Energy Scavenging, 15th International Conference on Solid-State Sensors, Actuators & Microsystems (Transducers’09), 2009.
53. J. D. Prades, R. Jimenez-Diaz, F. Hernandez-Ramirez, S. Barth, J. Pan, A. Cirera, A. Romano-Rodriguez, S. Mathur and J. R. Morante, Sens. Actuators, B DOI:10.1016/j.snb.2009.04.070.
54. H. Kind, H. Yan, B. Messer, M. Law and P. Yang, Adv. Mater., 2002, 14, 158.
55. A. Kolmakov and M. Moskovits, Annu. Rev. Mater. Sci., 2004, 34, 151.
56. V. V. Sysoev, B. K. Button, K. Wepsiec, S. Dimitriev and A. Kolmakov, Nano Lett., 2006, 6, 1584.
57. S. Mathur, S. Barth, H. Shen, J.-C. Pyun and U. Werner, Small, 2005, 1, 713; S. Mathur and S. Barth, Small, 2007, 3, 2070.
58. M. Law, H. Kind, B. Messer, F. Kim and P. Yang, Angew. Chem., Int. Ed., 2002, 41, 2405.
59. J. D. Prades, F. Hernandez-Ramirez, R. Jimenez-Diaz, M. Manzanares, T. Andreu, A. Cirera, A. Romano-Rodriguez and J. R. Morante, Nanotechnology, 2007, 19, 465501.
60. J. D. Prades, R. Jimenez-Diaz, F. Hernandez-Ramirez, L. Fernandez-Romero, T. Andreu, A. Cirera, A. Romano-Rodriguez, A. Cornet, J. R. Morante, S. Barth and S. Mathur, J. Phys. Chem. C, 2008, 112, 14639.
61. D. C. Meier, S. Semancik, B. Button, E. Strelcov and A. Kolmakov, Appl. Phys. Lett., 2007, 91, 063118.
62. K. Ryu, D. Zhang and C. Zhou, Appl. Phys. Lett., 2008, 92, 093111.
63. Y. Zhang, A. Kolmakov, S. Chretien, H. Metieu and M. Moskovits, Nano Lett., 2004, 4, 403.
64. T. Andreu, J. Arbiol, A. Cabot, A. Cirera, J. D. Prades, F. Hernandez-Ramirez, A. Romano-Rodriguez and J. R. Morante, in Nanosensors: Transduction Mechanisms at the Nanoscale Using Metal Oxides and Semiconductors, Sensors Based on Nanostructured Materials, ed. Francisco J. Arregui, Springer Science+Business Media, 2009, ISBN: 978-0-387-77752-8.

---

Footnotes

† This article was submitted as part of a Themed Issue on metal oxide nanostructures. Other papers on this topic can be found in issue 19 of Vol. 11 (2009). This issue can be found from the PCCP hopepage [http://www.rsc.org/PCCP].

‡ Electronic supplementary information (ESI) available: The dynamic response of individual nanowires; interval between gas–surface collisions at the nanowires. See DOI: 10.1039/b905234h

---