Doping of Polymers with ZnO Nanostructures for Optoelectronic and Sensor Applications

2.1. Technological importance of ZnO

ZnO has received much attention in recent years due to its diverse properties. It is a direct wide bandgap semiconductor (E_g = 3.4 eV) with large exciton binding energy (~60 meV), suggesting that it is a promising candidate for stable room temperature luminescent and lasing devices. It exhibits strong ultraviolet (Huang et al., 2001) and visible photoluminescence (Konenkamp et al., 2004). ZnO film can be used as transparent conducting electrode. It is also a piezoelectric material, with potential applications in surface acoustic wave filters (Wang et al., 2008). Ferromagnetism in ZnO has already been reported ( Kim et al., 2009 ) so it is being considered a material for spintronics. With all these important properties, it becomes more attractive because of its abundance in nature. It can be fabricated practically into different forms like thin film and nanowire using a variety of methods such as solution-based approaches, vacuum deposition techniques chemical vapor deposition.

2.2. Fabrication of ZnO nanostructures for dispersion in solution

There are several ways of fabricating ZnO nanostructures on a substrate. Nanowires, nanorods, nanobelts, nanoneedles and nanosprings have been reported by many researchers. For doping polymers, it is advantageous for the nanostructures to be dispersed in a solution since most polymers are prepared via a solution process. ZnO nanostructures grown on a substrate can be scraped and dispersed into a solvent. This approach however does not produce enough amount of nanostructures. Hence, it is not a practical method for obtaining a usable concentration of ZnO in a solvent. Solution-based methods have been widely used to prepare ZnO nanostructures. Nanorods and nanowires with variable-aspect-ratio (length/diameter) have been prepared in alcohol/water solution by reacting a Zn²⁺ precursor with an organic weak base, tetramethylammonium hydroxide (Me₄NOH) at temperature window of 75 – 150 ^oC (Cheng et al., 2006). In another report, a room temperature wet-chemical approach was employed to synthesize highly regulated, monodispersed ZnO nanorods and derived hierarchical nanostructures such as hexagonally branched, reversed umbrella-type and cactus-like (Liu & Zeng, 2004). These hierarchical nanostructures are comprised of individual c-oriented nanorods. In the typical synthesis, the precursor solution of Zn was prepared by dissolving zinc nitrate and sodium hydroxide in deionized water. A sample of pure ethanol was added followed by ethylenediamine. The mixed solution was covered and kept at room temperature under constant stirring for 1-12 days to form the ZnO nanorods. ZnO nanoparticles with size ranging from 2 to 7 nm can also be prepared by addition of lithium hydroxide to an ethanolic zinc acetate solution (Meulenkamp, 1998).

One of the disadvantages of solution-based ZnO fabrication is that it involves successive processes that take a fair amount of time. After the synthesis, the product has to be thoroughly cleaned to remove all traces of chemicals used in the process. In this article, a simple, straightforward method to fabricate ZnO nanostructures with fast turn around is explored. It is based on the oxidation of zinc vapor while a high purity Zn metal is being thermally evaporated. Figure 1 is a diagram that illustrates this method. The Zn source is placed in a crucible, which is then inserted into a vertically-oriented quartz tube. The tube is flushed with argon to remove ambient air and its open end is sealed [Figure 1(a)]. It is inserted into a vertically-oriented furnace set at 600 ^oC to evaporate the Zn. When the source starts to evaporate, air is allowed to diffuse towards the crucible by opening the seal. With the correct diffusion rate of O₂ and evaporation rate of Zn, oxidation can be contained inside the crucible. Thus, the growth of ZnO nanostructures mainly occurs in the crucible. Figure 1(b) shows the actual crucible before and after the growth process. Typically, when the amount of Zn source is 0.03 g, it yields approximately 0.01 g of ZnO, which is the white material that fills the crucible in the picture. The product is like a cotton as shown by the SEM picture in Figure 1(c). It can be easily collected and dispersed in a solution. The entire fabrication process only takes less than 30 minutes. It is highly scalable and can be employed to mass-produce ZnO nanostructures.

The cotton-like ZnO product shown in Figure 1(c) consists of different nanostructures entangled together. They can be separated and size selected if desired. Two solvents were tried to disperse them: water and chlorobenzene. By ultrasonication, they were effectively segregated in both solvents. The solution became whitish after several minutes of agitation. When the procedure was stopped, ZnO structures slowly settled at the bottom and as expected, large entities sank first. This can be used as a simple size-selection method. For example, nanoparticles with lower diameter are the last ones to settle at the bottom. If the solution is allowed to be still for a long time after sonication, then get the top part of it, nanoparticles are obtained. They are shown in Figure 2(a). They are deposited on Si wafer by drop casting. If the large and unwanted entities are already removed, good nanowires can be extracted from the bottom of the solution. These nanowires are shown in Figure 2(b). Their length could exceed 10 m.

3.1. Basic requirements for efficient electroluminescence

Polymeric-based light emitting device is one type of OLED, which stands for organic light emitting diode. Photon is generated in the polymer emissive layer (EL), which should possess excellent photoluminescence (PL) properties. In an efficient device, the EL is sandwiched between an electron transport layer (ETL), which injects electrons into the LUMO (lowest unoccupied molecular orbital) of the EL, and a hole transport layer (HTL), which injects holes into the HOMO (highest occupied molecular orbital) of the EL. The injected electrons and holes undergo radiative recombination due to the applied electric field. The charge transporting layers serve as facilitator for efficient charge transfer from the electrodes to the EL. This is illustrated in Figure 3. When a negative potential is applied to the anode, which is a low work function metal, electrons are injected into the ETL. If the ETL has excellent electron mobility (n-type) and its LUMO align closely with the LUMO of EL and the work function of the anode, efficient electron charge transport towards the EL is achieved. On the other hand, the positive potential applied to the cathode, which is a high work function transparent conductor, results in the injection of holes into the HTL. Efficient hole transport towards the EL is achieved if the HTL has excellent hole mobility (p-type) and its HOMO is well aligned with the HOMO of the EL and the work function of the cathode. It should be realized that the realization of high efficient OLEDs depends not only on the electronic and optical properties of the EL material, but also on the control of charge transport in the device.

3.2. Control of charge transport

One of the benefits of doping polymers in OLED with ZnO nanostructures is the improvement in charge carrier transport. It has been reported that by incorporating ZnO nanorods with the EL material PVK:DCJTB, which represents poly(vinylcarbazole) doped with 4-(dicyanomethylene)-2-t-butyl-6-(1,1,7,7-tetramethyljulolidyl-9-enyl)-4H-pyran at 1%, low threshold voltage and higher electroluminescence efficiency is obtained (Zhang et al., 2006). The authors employed the device structure: ITO/EL/Alq₃/Al, with different concentrations of ZnO (1 wt%, 5 wt%, 10 wt%) in the EL material. They observed that the hole current density increases with increasing concentration of ZnO nanorods. This made the current injection more balanced and consequently enhanced the electroluminescence efficiency. They explained the increase in the hole current in terms of the effect of ZnO on the molecular chain of PVK. The nanorods can make the PVK chain more unwindable and possibly connect aligned molecular chains. Thus, current is not limited by hopping of carriers between chains. In addition, wurtzite ZnO has higher hole mobility (_h) than PVK. For low p-type conductivity at room temperature, it ranges from 5 to 50 cm²/V-s (Norton et al., 2004). For PVK at low electric fields, it is in the order of 10^-7 cm²/V-s (Blom et al., 1997).

A similar observation has been reported on poly[2-methoxy,5-(2-ethylhexyloxy)-1,4-phenylenevinylene], which is commonly called MEH-PPV for short (Xu et al., 2007). By doping the MEH-PPV with ZnO nanocrystals or tetrapods and employing it to build field effect transistor, the authors showed higher p –type mobility in the polymer nanocomposite. Pure MEH-PPV, which only exhibits p– channel behaviour, has _h value in the order of 10^-4 cm²/V-s. This increases with increasing concentration of ZnO nanostructures and saturates at 40 wt%. For nanocrystals and tetrapods, _h saturates at a value of about 0.08 cm²/V-s and 0.15 cm²/V-s respectively. It should be mentioned that doping with ZnO did not change the conductivity type, indicating that charge transport takes place in the polymer. This is not surprising because there is a large energy barrier for holes to be transferred from ZnO to MEH-PPV. The valence band edge for ZnO is 7.6 eV while the HOMO of MEH-PPV is 5.3 eV. When the density of traps was investigated, it was found that it decreases with increasing ZnO concentration. Interestingly, it also saturates at 40 wt%. This is a strong indication that the effect of ZnO is to reduce the density of traps in the polymer and consequently improve its _h. The authors further supported this interpretation by transient current measurement where a constant bias voltage was applied to the gate and the drain. The decrease in channel current with time was more dramatic for the pure polymer than the nanocomposite. This suggests that charge trapping is significantly reduced by ZnO doping. Other authors propose that the enhancement in _h is due to the superposition of several transport mechanisms in the nanocomposite including percolation through polymer-ZnO nanoparticle interface network (Aleshin et al., 2008).

3.3. Control of optical emission

The high surface reactivity of ZnO nanostructures with certain polymers may provide interactions that can modify polymer properties in a positive way. The interaction of ZnO quantum dots (QDs) with poly(vinyl alcohol) (PVA) has already been demonstrated (Sui et al., 2005). The authors prepared PVA/ZnO hybrid nanofibers with ZnO concentration of 10 wt% and employed differential scanning calorimetry (DSC) to study them with respect to pure PVA. It is known that the two exothermic peaks at about 305 ^oC and 500 ^oC are associated with the degradation of side chain (the scission of C-O) and main chain (the scission of C-C) with the delta enthalphy of 3760 and 1188 J g^-1, respectively (Koji et al., 1999). For PVA/ZnO nanofibers, the exothermic peak below 450 ^oC was not observed, while the peak around 500 ^oC was sharp and strong with delta enthalpy of 3305 J g^-1, suggesting that ZnO modifies the property of PVA. Their interaction, which is via the formation of an H bond and an O-Zn-O bond between PVA molecule and ZnO, is believed to be responsible for the novel luminescent properties of PVA/ZnO nanofibers.

Figure 4 shows a typical PL spectrum of ZnO nanowires (NWs) prepared by the physical vapor transport method described in section 2.2. It has two emission bands around 380 and 520 nm, which is consistent with another report using a polymer-assisted route (Li et al., 2003). The UV emission originates primarily from a mixture of free exciton and bound exciton related to impurity or defects (Shan et al., 2005). The visible emission comes from the transition between the electron near the conduction band and the deeply trapped hole at the V₀ ^** center which is an oxygen vacancy containing no electrons (Dijken et al., 2000). It is also attributed to the transition between the electron at [V₀ ^*, electron] or [V₀ ^**, two electrons] and the hole at vacancy associated with the surface defects. For pure PVA, the PL spectrum consists of two emission bands at about 364 and 440 nm, originating from its organic functional groups (Sui et al., 2005). In PVA/ZnO hybrid film, the PL spectrum is a superposition of the PL of each component. The 440 nm emission from PVA overlaps with the 550 nm from ZnO making the visible emission band broader. In hybrid nanofibers however, this visible emission is much intense possibly due to two factors. In the film, ZnO-QDs can easily form aggregates resulting in bigger particle exhibiting bulk behavior. This is in fact one of the issues in polymer nanocomposites and ways to circumvent it are being explored (Sun et al., 2008). In the nanofiber, QDs align along the PVA matrix, resulting in better dispersion. Individual QD also interacts strongly with the PVA molecule. The other factor is the high surface to volume ratio of the nanofiber, which enhances the visible luminescence due to surface defects (Shalish et al., 2004).

The interaction between ZnO and polymer can also provide several new radiative recombination mechanisms that may be utilized for broadband light emission or wavelength tunable emission. Since these mechanisms are influenced by several factors such as the individual properties of the organic and inorganic components as well as their relative concentration, control of emission properties may be done in a number of ways. The existence of new recombination mechanisms is strongly suggested by the generation of new PL emission lines (420, 460 & 480 nm) when MEH-PPV is doped with ZnO nanoparticles (Aleshin et al., 2009). In this work, the authors prepared the polymer nanocomposite film on a Si substrate with a 200 nm-thick SiO₂ layer via drop casting or spin coating. Gold and aluminum electrodes were then deposited on top with separation of 7 – 15 m. With this device structure, they could investigate PL while applying an external electric field. The existence of new emission lines, which could not be generated when ZnO was replaced by Si nanoparticles, was attributed to the formation of interfacial states termed exciplex. These states are formed in a type II heterojunction, such as ZnO/MEH-PPV, where there is an accumulation of electrons and holes on the opposite side. Charge transport is blocked by the barriers that result from the offset of HOMO and conduction band edge of MEH-PPV and ZnO respectively. It was proposed that since the carriers cannot proceed across the heterojunction, they diffuse two-dimensionally along the interface until they encounter an oppositely charged carrier on the opposite side of the interface and these form an exciplex (Morteani et al., 2005). If bulk exciton in the ZnO or MEH-PPV is within the exciton diffusion length, it can excite exciplex emission. It should be mentioned that the influence of ZnO nanoparticles is most dramatic at an optimum concentration of 10 wt%. This could be ascribed to the aggregation of nanoparticles which reduces interaction with polymer.

Another observed phenomena when MEH-PPV is doped with ZnO is the blue shift of its PL emission spectrum accompanied by quenching. At 2 wt% concentration of ZnO nanoparticles, 11 nm shift has been reported ( Ton-That et al., 2008 ). This shift increases with increasing nanoparticle concentration. However, its dependence becomes weaker at concentration greater than 4%. Using Raman spectroscopy, the authors did not observe any changes in the conjugation length, which may have caused the blue shift in PL spectrum. Their results indicate that the chemical structure and the gap energy of MEH-PPV is not affected by the incorporation of ZnO nanoparticles. Thus, they explained their observation in terms of the effect of electric field produced by excess electrons on the nanoparticles surface (Musikhin et al., 2002). The surface of ZnO nanoparticles is a strong perturbation of the lattice where there exists a high concentration of both shallow and deep defect levels ( Ton-That et al., 2008 ). X-ray photoelectron spectroscopy (XPS) analysis of their surface revealed that there is a large number of oxygen vacancies. These vacancies can accept electrons from MEH-PPV, creating an electric field outside the nanoparticles. As a result, the energy of the lowest exciton state is raised resulting in the blue shift of the luminescence.

4.1. Challenges

The highest power conversion efficiency () demonstrated by a polymer-based solar cell is achieved by employing an active layer blend of poly(3-hexylthiophene) and (6,6)-phenyl C61 butyric acid methyl ester (Ma et al., 2005). For convenience, this blend will be referred as P3HT:PCBM. It has a strong optical absorbance in the limited wavelength range of 400 nm to 600 nm mainly due to the P3HT. Because of this narrow band absorption, more than 60% of the photons in the solar spectrum are not harnessed. In fact, even the ones absorbed are not 100% converted to photocurrent because not all electron-hole pairs become free charges. Exciton binding energy of P3HT is so large that it does not lead directly to free charge carriers. However, the high electron affinity of PCBM provides sufficient potential difference that breaks exciton and thus it becomes energetically favourable for the electrons to jump from the LUMO of P3HT to the LUMO of PCBM. This electron donor-acceptor mechanism implies that only excitons near the P3HT/PCBM interface dissociate. The maximum distance for dissociation to occur is determined by the exciton diffusion length (), which is around 10 nm for P3HT. In the blend, phase segregation results in the formation of isolated islands of P3HT and PCBM. The size of P3HT islands could be more than so not all excitons can dissociate. There is another factor in the exciton dissociation process that reduces . Excited electrons in P3HT have excess energy that is dissipated quickly when they transfer to the PCBM. This energy loss, which depends on the energy alignment of the two materials, is reflected in the open circuit voltage (Koster et al., 2006).

In a photovoltaic device, separated holes and electrons need to reach their respective electrodes to have high . This requires the blend to have high values of hole (_h) and electron (_e) mobilities. Good balance of mobilities is obtained using optimum ratio of P3HT and PCBM, which is estimated to be between 1:1 and 1:0.9 (Chirvase et al., 2004). This could yield values of _h ~ 6.5 x 10^-6 cm²/V-s and _e ~ 5.0 x 10^-6 cm²/V-s. However these numbers are still far inferior when compared to their inorganic counterpart, which raises the question of whether doping with inorganic nanostructures such as ZnO would be beneficial. In the blend, PCBM does not provide direct conduction pathways for electrons towards the electrode because of the isolated island formation. Electrons undergo a hopping process between PCBM islands before reaching the electrode. Thus, they become prone to recombination with holes. This limits the thickness of the active layer. High efficient P3HT:PCBM devices are achieved with active layer thickness around 100 nm to 200 nm, which is not the optimum thickness for photon absorption. Increasing the thickness absorbs more photons but it also increases the electron-hole recombination and degrades charge transport. The final stage of charge transport from point of exciton dissociation to the electrode is the crossing of the organic-inorganic interface. Large contact resistance will reduce the fill factor of a device. Potential barriers at the interface originating from impurities or polymer damage during vacuum deposition of the metal electrodes may also limit the open circuit voltage. Controlling the organic-inorganic interfacial properties is an important and urgent challenge in organic electronics. Several ways have already been explored such as insertion of very thin interlayer (Brabec et al., 2002), tuning the work function of the inorganic electrode (Sharma et al., 2009) and surface doping of polymer (Mukherjee et al., 2007).

4.2. ZnO as electron acceptor

PCBM is the most widely used electron acceptor for P3HT in a bulk heterojunction (BHJ) solar cell. It has the right LUMO and HOMO levels for charge separation and exchange. Its electron mobility in the order of 10^-2 cm²/V-s, is sufficient for transporting photogenerated electrons towards the electrodes. However, higher mobilities are still demonstrated by inorganic materials. Wurtzite ZnO for example can have electron mobility of 200 cm²/V-s. Its LUMO and HOMO levels are well positioned to accept electrons from P3HT. Thus, doping of conjugated polymers with ZnO nanostructures can be employed to fabricate BHJ solar cells. For example, P3HT doped with ZnO nanoparticles yielded a photovoltaic device with ~ 0.92% at optimum concentration of 26% by volume (Beek et al., 2006). The values of the open circuit voltage (V_OC), short circuit current density (J_SC) and fill factor (FF) were 0.69 V, 2.19 mA/cm² and 55% respectively. The authors employed the device structure: ITO/PEDOT:PSS/P3HT:ZnO/Al, which is a common structure used by P3HT:PCBM active layer. They identified the major performance limiting factors as the poor contact of P3HT with the ZnO and the coarse morphology. They also observed that thermal annealing of the active layer was very important.

In another work, ZnO nanofibers were grown on the transparent electrode to achieve fixed morphology before the introduction of the P3HT (Olson et al., 2007). By spin-coating the P3HT on top of the ZnO nanostructure followed by annealing, intercalation into the voids between the nanofibers was induced. In this structure, a huge donor-acceptor interface area is formed between the P3HT and ZnO for exciton dissociation. Photogenerated electrons injected into the ZnO are transported directly towards the collecting electrode, providing pathways with higher electron mobility. The growth of ZnO nanofibers on ITO-coated glass was achieved via a low-temperature hydrothermal route from a solution of zinc nitrate precursor. In this device structure in which Ag is used as back electrode instead of Al, electrons are collected by the ITO while holes are collected by the Ag. The best device they obtained had V_OC, J_SC and FFof 0.44 V, 2.17 mA/cm² and 56% respectively corresponding to ~ 0.53%. Although this performance is relatively poorer than P3HT:PCBM devices ( ~ 5%), it is expected to improve when the major mechanisms in the device are better understood. For example, the role of oxygen vacancies in the ZnO, which is the origin of visible emission, is not yet clear. Obtaining a balanced electron and hole mobilities in the active layer may also be a critical issue. Understanding the effect of ZnO on the effective Fermi level of the active layer, which determines the interfacial properties at the metal junction, is important. This may have a direct consequence on the V_OC. Lastly, the mechanism of aging in ambient air may give insights on how to optimize this device. The authors observed that right after fabrication, the device exhibited only little diode behavior. But after storing it in dark ambient air for 24 hours, better device performance was observed. In contrast, the device stored in dark argon environment decreased in performance. After 3 days in argon, the device performance continued to deteriorate whereas the performance of the device stored in air improved substantially. The improvement came from the increase in FFand V_OC. They explained this behavior in terms of the oxygen vacancies. These vacancies, which are intrinsic electron donors in ZnO, may be quenched when exposed to oxygen. As a result, semiconducting properties are improved as both the oxygen vacancy and carrier concentrations are reduced. This aging mechanism makes ZnO a very promising electron acceptor for BHJ solar cells, which require long term operation.

4.3. Effect of ZnO nanowire doping on the Fermi level of P3HT

In polymer-based photovoltaics, it is important to obtain the desired junction characteristic of the polymer-metal interface. Depending on the device structure, it may require an Ohmic or a Schottky junction. The junction characteristic, which has a direct influence on the charge transfer from the active layer to the electrode, is determined by the energy level alignment of the polymer and the metal. Fermi level (E_F) is the most fundamental parameter that is used to understand polymer-metal junction. It determines the work function (energy difference between E_F and vacuum level) of a material. It can be used to predict the built in potential (V_bi) across a junction in thermal equilibrium. In thin film photovoltaic devices, V_bi is an essential parameter. It influences charge dissociation, charge transport and charge collection. Most importantly, it sets the maximum V_oc of a single junction solar cell. Thus, measurement of V_bi is very important practically and it can be a tool to understand how the E_F of a material behaves when it is doped. In most situations, V_bi across a junction is not equal to the value predicted by the difference in the E_F of each component. During the formation of the junction, such as the deposition of the metal on the polymer or the drying of the polymer on a metal surface, chemical reaction may occur. This leads to the existence of new compounds in the middle of the junction, which modifies the expected V_bi. Thus, determination of actual V_bi in real devices is more important practically because it takes into account the effect of processing. In photovoltaics for example, measuring V_bi in P3HT-Al junction formed by thermal evaporation of Al on P3HT film, is more meaningful than using a similar junction formed by spin-coating P3HT on Al film. This is because the first system is the one being used in the development of organic solar cells.

In this manuscript, recent results are presented about the change in the Fermi level of P3HT when doped with ZnO nanowires. This system is already being explored as an active layer together with Al as back electrode (Beek et al., 2006). To measure V_bi, capacitance-voltage (C-V) measurement was employed using an LCR meter. Test diodes were prepared by drop-casting P3HT and doped P3HT (~ 1 m thick) on Cr/Pt film then depositing Al on top via resistive thermal evaporation.

The device structure is shown on Figure 5(a). Both doped and undoped devices exhibited good rectifying behaviour as depicted by their I-Vcurves in Figure 5(b). The rectification is due to the Schottky junction formed by the polymer and Al. The high work function value of Pt (~6.35 eV) results in an Ohmic contact with the polymer. According to inorganic semiconductor theory, the capacitance Cof the depletion layer is related to V_bi by the equation (Sze, 2007):

Where Vis the applied bias voltage, kT/qis the thermal voltage whose value at room temperature is 0.0259 V, _s is the semiconductor permittivity and N_D is the donor impurity density. V_bi is determined by plotting this equation and finding the value of V where 1/C² extrapolates to zero. This is illustrated in Figure 6. The behavior of 1/C² is quite linear for both devices in the bias range of 0.5 V, consistent with equation (1). The solid line is the linear fit for each data set. Its intercept plus the thermal voltage is equal to V_bi. The obtained value for the pure P3HT is 1.243 V. This is fairly close to the expected value based on the E_L offset between P3HT and Al.

The work function of P3HT is around 5.1 to 5.2 eV. For Al, it ranges from 4.0 to 4.3 eV. The ZnO-doped P3HT has V_bi of 1.181V, indicating that E_F of the P3HT moved farther from the valence band edge and closer to the E_F of Al. One possible explanation for this reduction is the increase in the free electron concentration in the polymer nanocomposite provided by the ZnO. In semiconductors, E_F lies closer to the conduction band edge for n-type conductivity and it lies closer to the valence band edge for p – type conductivity. The upward shift in E_F due to the ZnO may be the cause of the drop in V_OC for P3HT:ZnO nanofibers BHJ (Olson et al., 2007). When the authors compared the photovoltaic performance of ZnO bilayer and ZnO nanofiber devices, they observed higher V_OC in the bilayer device (500 mV versus 440 mV). Although the nanofiber device still had better due to its higher J_SC. The results shown in Figure 6 may be helpful in optimizing BHJ solar cell based on P3HT and ZnO nanostructures. High values of J_SC and V_OC may be obtained simultaneously by tailoring the property of the active layer-metal interface or selecting alternative metals or alloys for back electrode.

5.1. Conductivity sensors for chemical detection

Conductivity sensors for chemical detection are very attractive because of their simple operation. The circuit design required to convert their response to electrical signals is relatively straightforward. These sensors are based on the change in the electrical conductivity of the sensing material due to its interaction with the stimuli. Device fabrication is also not difficult. For example, the sensing material may be deposited on interdigitated or two parallel electrodes, which form the connection for measuring the change in electrical resistance. Three common classes of sensing materials for conductivity sensors are conducting polymer composites, intrinsically conducting polymers and metal oxides (Arshak et al., 2004). The mechanism behind the change in conductivity is different for each class.

In conducting polymer composites, conducting particles such as polypyrrole and carbon black (Albert et al., 2000) are embedded in an insulating polymer matrix. The stimuli can cause expansion of the polymer and the polypyrrole particles resulting in the increase of electrical resistance. If the stimuli interact chemically with the polypyrrole, its intrinsic conductivity will also change, making the resistance change in the composite more difficult to interpret. The operation of intrinsically conducting polymers (ICP) such as polypyrrole, polythiophene and polyaniline, is based on the alteration of intrachain conductivity, intermolecular conductivity and ionic conductivity after interacting with the stimuli. Intrachain conductivity is detrmined by the backbone while intermolecular conductivity is due to electron hopping to different chains. Ionic conductivity is affected by proton tunneling induced by hydrogen bond interaction at the backbone and also by ion migration through the polymer. One reported application of ICP employed an array of conducting polymer sensors incorporating 3-Methylthiophene, Aniline and Pyrrole (Guadarrama et al., 2000). The authors evaluated the sensors to different wine varieties and considered the cross sensitivity of the polymeric films to moisture and ethanol. The operation of a metal oxide sensor is based on the dependence of the conductivity to O₂ molecules adsorbed on its surface. For n-type metal oxides such as ZnO or TiO₂, O₂ molecules capture electrons on the surface or at the grain boundaries resulting in the decrease of conductivity. Trapped electrons at the grain boundaries also produce potential barriers between grains that impede current flow. When exposed to reducing gas like H₂, CH₄, CO, C₂H₅ or H₂S, the adsorbed O₂ molecules react with the gas and release the captured electrons. This increases the conductivity of the sensing material. When exposed to oxidizing gas like O₂, NO₂ or Cl₂, the effect is opposite. Conductivity is decreased because the number of adsorbed O₂ molecules that can capture electrons is increased. For p-type metal oxides, their response to oxidizing gas is to increase in conductivity because the capture of electrons produces more holes. In contrast, their conductivity decreases when exposed to reducing gases.

There are still many challenges in conductivity gas sensors. For polymer composites, aging is a concern because this leads to sensor drift. Their applicability is only limited to certain gases because of lack of sensitivity. For ICP, challenges include understanding the mechanism of sensor response, high sensitivity to humidity and drift in conductivity with time. These sensors can also have short lifetimes due to oxidation of the polymer. For metal oxides, one challenge is the low sensitivity at room temperature. They require higher operating temperatures and thus heating element has to be incorporated with the sensor (Roy & Basu, 2002) This increases the power consumption, which is not very appealing for handheld applications. Good selectivity is another challenge for metal oxide conductivity sensors. Their behavior to reducing gases is practically the same. For example, the conductivity of a sensor changed by when exposed to ethanol vapor with concentration C₁. The same response can be realized when it is exposed to methanol with a different concentration C₂. Thus, the identification of a particular reducing vapor, with a background of different others is almost impossible. Due to the mentioned challenges, there is a continuous effort to develop new materials for conductivity sensors. Reliable sensors with appropriate levels of selectivity and sensitivity will always be in demand. In the standpoint of defense and homeland security, new hazardous compounds or biomaterials that need to be detected is expected to arise. New applications such as self-powered sensors will require excellent energy efficient sensing materials without sacrificing sensitivity and selectivity.

5.2. Sensing with P3HT doped with ZnO nanowires

There is a lack of work in the exploration of ZnO doped polymers as conductivity sensors for chemical detection. Interestingly, only P3HT, which is mostly used in photovoltaics, has been reported to have promising sensing behavior when blended with ZnO-NWs (Saxena et al., 2007). As separate materials, P3HT and ZnO-NWs can both be used as a conductivity gas sensor. But due to the difference in their conductivity type, they have contrasting response to a specific stimulus. For example, they both exhibit high sensitivity to NO₂ and H₂S (Saxena et al., 2007). P3HT decreases in resistance upon exposure to NO₂, which is oxidizing gas, while it behaves oppositely for H₂S, a reducing gas. This characteristic is due to its p – type conductivity. The adsorption of an oxidizing compound on P3HT surface captures electrons leading to increased hole carrier concentration. Conversely, the adsorption of a reducing compound leads to the decrease in hole carrier concentration via the release of captured electrons. For n – type ZnO, the effect of oxidizing and reducing compounds is to decrease and increase conductivity respectively.

Doping of P3HT with oxygen defficient ZnO nanowires results in the reduction of P3HT. This is well supported by GI-XRD, XPS and FTIR studies (Saxena et al., 2007). The authors also observed decrease in the P3HT conductivity after ZnO doping suggesting that its hole carrier concentation drops when ZnO donates electrons. This interaction may be used to tailor the response of polymers to different stimuli, opening up the possilibity of tunable polymer-based conductivity sensors. For example, the authors have demonstrated that highly reduced P3HT resulting from ZnO doping, has enhanced sensitivity to NO₂ because this chemical can easily pick up electrons from P3HT to get adsorbed as NO^- ₂ ions. They have obtained room temparature sensors that can detect NO₂ in the 0 – 10 ppm range with very high selectivity. In contrast, the sensitivity of ZnO-doped P3HT to a reducing H₂S is weakened because it is difficult to further reduce the P3HT.