NO2 Gas Sensing Mechanism of ZnO Thin-Film Transducers: Physical

Duca degli Abruzzi, 24, 10129 Turin, Italy. ACS Sens. , 2016, 1 (4), pp 406–412. DOI: 10.1021/acssensors.6b00051. Publication Date (Web): Februa...
0 downloads 0 Views 1021KB Size
Subscriber access provided by NEW YORK UNIV

Article 2

NO Gas Sensing Mechanism of ZnO Thin-Film Transducers: Physical Experiment and Theoretical Correlation Study Athanasios Tamvakos, Kiprono Kiptiemoi Korir, Dimitrios Tamvakos, Davide Calestani, Giancarlo Cicero, and Daniele Pullini ACS Sens., Just Accepted Manuscript • DOI: 10.1021/acssensors.6b00051 • Publication Date (Web): 05 Feb 2016 Downloaded from http://pubs.acs.org on February 6, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Sensors is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

NO2 gas sensing mechanism of ZnO thin-film transducers: physical experiment and theoretical correlation study Athanasios Tamvakos1, Kiprono Korir3*, Dimitrios Tamvakos1, Davide Calestani2, Giancarlo Cicero3, Daniele Pullini1 1

Centro Ricerche Fiat, 50 Strada Torino, 10043 Orbassano (TO), Italy 2

3

IMEM-CNR, Parco Area delle Scienze 37/A, 43100 Parma, Italy

Politecnico di Torino, Corso Duca degli Abruzzi, 24, 10129 Turin, Italy

ABSTRACT.

In this work, ZnO thin films were investigated to sense NO2, a gas exhausted by the most common combustion systems polluting the environment. To this end, ZnO thin films were grown by RF sputtering on properly designed and patterned substrates to allow the measurement of the electrical response of the material when exposed to different concentrations of the gas. XRD was carried out to correlate the material’s electrical response to the morphological and microstructural features of the sensing materials. Electrical conductivity measurements showed that the transducer fabricated in this work exhibits the optimal performance when heated at 200°C, and the detection of 0.1 ppm concentration of NO2 was possible. Ab initio modelling allowed the understanding of the sensing mechanism driven by the competitive adsorption of NO2 and atmospheric oxygen mediated by heat. The combined theoretical and experimental study here reported provides insights into the sensing mechanism which will aid the optimization of ZnO transducer design for the quantitative measurement of NO2 exhausted by combustion systems which will be used, ultimately, for the optimized adjustment of combustion resulting into a reduced pollutants and greenhouse gases emission.

ACS Paragon Plus Environment

ACS Sensors

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

KEYWORDS: GAS SENSING, DFT, ZnO, NO2, Thin-Film, RF-sputtering.

*Corresponding author: K K Korir

Email: [email protected]

Nitrogen oxide gases, commonly referred to collectively as NOX, are common pollutants formed in internal combustion engines and industrial combustion systems by thermal fixation and oxidation of atmospheric nitrogen1. NOX, which includes NO2 and NO gases, have adverse effects on the environment, and are the leading cause of green house effect, acid rain, and photochemical smog. In humans, exposure to more than 3 ppm of NO2 gas for periods longer than 8 hours can cause respiratory and cardiovascular diseases2. Hence, better detection of such gases is of utmost importance, and that calls for better understanding of detection mechanism to facilitate development and optimization of sensing devices. Semiconducting metal oxides based thin films have traditionally been used in detection of NOX gases, with the common measuring arrangement relying on changes in electrical conductivity upon interaction with the gas3-7. In particular, ZnO based gas sensors have been employed

ACS Paragon Plus Environment

Page 2 of 21

Page 3 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

extensively in the detection of both reducing and oxidizing gases due to the superior properties of this material such as enhanced sensitivity, stability, and low cost8. For example, Tomchenko and coworker9 found that a thick ZnO film based sensor is characterized by high sensitivity and selectivity to NO and NO2 gases at 300 °C, and a decrease in conductivity was recorded when the gas was introduced in the chamber containing the analyzed sample. On the theoretical front, density functional theory (DFT) has been employed in a number of studies to address the interaction of NO2 with ZnO surfaces, yet the mechanism leading to a decrease in ZnO conductivity upon NO2 detection has not been proven before. For example, Rodriguez et al.10 showed that adsorption of NO2 on (

)-O terminated surface yields NO3 type species through

binding of NO2 to a surface O atom, while on the (

)-Zn terminated surface, NO2 prefers to

bind to a surface Zn via its O atom. For NO2 adsorption on both O and Zn terminated (

)

surfaces, stability was found to decrease with increase in surface coverage. On the other hand, Spencer et al.11 showed that the stability of NO2 adsorption on non-polar ZnO (

) surface is

improved by the presence of oxygen vacancies, which act as preferred site and promotes dissociative adsorption of NO2. Hence, detection of NOX gases in ZnO nanostructures is governed by surfaces but the processes that drive detection are still not well understood. Sadek et al.12 proposed that NO2 gas reacts directly with ZnO surface rather than with chemisorbed oxygen molecules present on the surface of the sensor, while other studies11 have suggested oxygen vacancies as essential for NO2 detection. In addition, the changes in conductivity of the sensor at operating temperature of 300400 °C has always been associated with dissociation of NO2, but a rigorous understanding on which surface processes explicitly affects ZnO conductivity (i.e. makes detection possible) still

ACS Paragon Plus Environment

ACS Sensors

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

awaits to be given. Particularly previous studies have completely neglected the n-type character of ZnO used as sensing elements. In this work, we combined experimental measurements with DFT calculations to probe the interaction of NO2 gas on ZnO (101) non-polar surface by explicitly considering ZnO n-type character and gain insight in the role played by surfaces during detection. We also studied the effect of temperature via ab initio molecular dynamics (AIMD) calculations for temperature range between 200-400 °C, in which a marked change in conductivity is experimentally observed, while in temperature values of below 200°C the sensor response is lower and much slower.

EXPERIMENTAL SECTION I. Device preparation In order to fabricate the prototype gas sensor, contacts and heating element were patterned onto an alumina substrate: shadow masks of specific shape and dimensions were chosen for patterning of 50 nm Au thin film and 200 nm Pt thin film on 3x3 mm2 alumina: Au is used as contacts while Pt acts as heating element, as shown in Fig. 1.

ACS Paragon Plus Environment

Page 4 of 21

Page 5 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

Fig. 1: Schematics of the ZnO sensor design.

ZnO sensing element was deposited on the alumina patterned substrate (within “deposition zone” in Fig. 1) using RF sputtering technique as described in literature13, 14,. Controlling the distance between the nozzle and substrate, as well as the sputtering power, allowed achieving desired grain size15. In particular, the distance of the substrate from the ZnO target was set at 9 cm and sputtering power of 150W was applied. The deposition rate was set to 6.1 nm/min and a total thickness of 200 nm was obtained. Finally, the sensing element was annealed at 400 °C for 60 minutes.

II. Film characterization The surface morphology of the samples was analyzed in terms of mean grain size and surface roughness by using an Atomic Force Microscope (AFM) dimension 3100. In addition, X-ray diffraction (XRD) patterns were obtained with a powder diffractometer (Bruker AXS D8) equipped with a Cu Kα X-ray tube. III. Test and measurements Gas sensing properties of the fabricated gas-sensing elements were characterized employing a computer-controlled gas-sensing characterization system. The system consisted of two major parts: the gas mixture generation unit and the electrical test unit, as shown in Fig. 2. The gas mixture generation unit mixes air with the testing gas stored in a cylinder. Three gas lines are used to deliver the gases, one containing dry air, a second one with 100 % saturated humid air and a third one with the diluted gas to be tested. By regulating the flows of these three lines via mass controllers located next to the cylinders, desired humidity and testing gas concentration was achieved. Conductance measurements were performed at an overall constant flow rate of 0.5 L/min. The sensing element is housed in a test cell, which consist of a metal box shielding

ACS Paragon Plus Environment

ACS Sensors

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

external electrical noise and keeping the sensor at room temperatures. Other sensors for monitoring humidity, pressure and temperatures are included in the test cell.

Fig. 2: Schematic of the gas sensor set-up which consists of three main components: the gas mixture generation system, the sensor electrical test equipment and the data acquisition system.

COMPUTATIONAL DETAILS All calculations were performed using the Quantum Espresso suite16, which performs first principle simulations within the Density Functional Theory (DFT) in the Kohn-Sham approach17. The generalized gradient approximation to the exchange and correlation functional was employed in the form proposed by Perdew, Burke and Ernzerhof (PBE). The core-electrons were replaced by ultra-soft pseudo potentials following Vanderbilt's formulation18, and the electronic wave functions (charge densities) were expanded in a plane wave basis set with an energy cutoff of 28 Ry (280 Ry). The ZnO surface was modelled in a supercell containing a 12 layers slab characterized by a (3x2) surface periodicity. A thick vacuum layer (~15 Å) was included in the direction perpendicular to the surface to ensure minimum interaction between periodic images.

ACS Paragon Plus Environment

Page 6 of 21

Page 7 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

Integration over Brillouin zone was performed using (4⋅4⋅1) Monhorst and Pack19 grids, and structures were relaxed until forces on all atoms were lower than 0.02 eV/ . Adsorbates were symmetrically adsorbed on both top and bottom surfaces to avoid spurious electrostatic interactions. DFT failure to predict band-gap was addressed by applying Hubbard corrections (DFT+U) to the relaxed structure: U values of 12.0 eV and 6.5 eV were used for Zn 3d orbitals oxygen 2p orbitals respectively, as proposed in some recent work on ZnO20. Finally, spin polarization was included in calculations involving oxygen molecules. AIMD simulations for the most stable structures of NO2 and O2 adsorbed on the ZnO surface (see Fig. 7a-b) were performed at 200°C , 300°C and 400°C , which correspond to temperatures range at which the gas sensor was operated. The systems where initially equilibrated for 2 ps at 27 °C through an NVT molecular dynamics simulations, afterwards the temperature was gradually increased till target temperatures were achieved by use of a Nose-Hoover thermostat21. The overall MD simulation time was 5 ps. A time step of 1 femtosecond was employed for the numerical integration of the equation of motion applying the Verlet velocity algorithm22.

RESULTS AND DISCUSSION The ZnO sample exhibits hexagonal wurtzite structure exposing both polar and non-polar facets. As shown in Fig. 3a, among non-polar surfaces the (101) is the most dominant, thus it is expected to play a major role during exposure of the sensing element to NO2 containing gas. Fig 3b is given for a direct structural comparison23 of the prepared sample with those that have been reported in literature.

ACS Paragon Plus Environment

ACS Sensors

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Fig. 3: Direct comparison of XRD diffractogram of the a) sputtered ZnO film with b) ZnO nanostructures. Reprinted (Adapted or Reprinted in part) with permission from Shi, Feng; Xue, Chengshan. Morphology and growth mechanism of novel zinc oxide nanostructures synthesized by a carbon thermal evaporation process. CrystEngCom. 2012, 14, 5407-5411. Copyright [2012] Royal Society of Chemistry.

Conductometric measurements were performed on the operating sensor using flow-through technique at a constant rate flow of 0.5 L/min for a temperature range of 200-400 °C; the relative humidity during detection was kept at 30 %; this percentage was chosen in order to have a more realistic response from the sensor, without having at the same time an excessive humidity effect. Before the influx of gas into the chamber, the temperature within the sensor and oxygen adsorption on its surface is stabilized. The procedure usually takes 1-3 hours; we note that when stabilization is achieved, the resistance of the sensing element is higher for the lowest temperature of the considered range. The sensor's temperature was previously calibrated by

ACS Paragon Plus Environment

Page 8 of 21

Page 9 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

putting a platinum stripe beside the heater in the same position of the sensing element. The temperature there was measured through the variation of platinum stripe resistance. Resistance versus time plot (see Fig. 4) after the stabilization of the sensor shows that the resistance is directly related to the concentration of NO2 and that complete recovery is attained in air. In addition, highest resistance change is obtained at 200 °C for 0.5 ppm gas concentration while at 400 °C smaller resistance changes are observed for all gas concentration considered.

Fig. 4: Resistance versus NO2 concentration (from 0.2 to 0.5 ppm) and time after sensor’s stabilization.

ACS Paragon Plus Environment

ACS Sensors

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 21

The change in the conductivity observed upon exposure to the gas reveals a typical behavior of the n-type sensors. In the present study, response is defined as follows:

(1) where

and

are the electric resistance in air and NO2, respectively. Ideal gas sensor should

have high response, in this work, response of 1850 % was obtained for 0.5 ppm gas concentration at 200 °C and 100 % for 0.1 ppm, as shown in Fig. 5. Lower response rates has been reported at higher operating temperatures of 300 °C and 400 °C; the maximum response was noticed for 0,5 ppm gas concentration and it was around 130% and 60% respectively.

Fig. 5: Response in respect to NO2 concentration at elevated operating temperature.

To understand why exposure to NO2 molecules leads to the observed increase in resistance of the ZnO sensors, we performed a series of first principle calculations on the interaction NO2 and O2

ACS Paragon Plus Environment

Page 11 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

(the most relevant reactive gas in the atmosphere) with the ZnO (101) non-polar surface. To cope with the fact that ZnO experimentally shows intrinsic n-type behavior, we have intentionally doped our ZnO surface structures with interstitial hydrogen as recently proposed in literature24. The optimized structures for the surface with adsorbates were obtained by placing the gas molecules ~3 Å above the surface and optimizing the systems using the Broyden-FletcherGoldfarb-Shanno (BFGS) algorithm25; the final geometries are shown in Fig. 6. The bonding strength of the gas molecules to the ZnO surface was estimated by calculating the binding energies (BE), defined as follows:

(2) where

is the total energy of relaxed surface with the adsorbate,

free adsorbate, and

is the total energy of

is the total energy of the relaxed surface. The coefficient ½ takes into

account the fact that 2 molecules are adsorbed on the ZnO slab, one for each equivalent surface. Results on the BE for both the H-doped and the undoped ZnO surface interacting with O2 and NO2 are summarized in Table I.

d(Zn-O) (Å)

ZnO ( ZnO: H (

) )

Eads(eV/mol)

O2

NO2

O2

NO2

2.18

2.08

-0.32

-0.57, -0.6120

2.13

2.10

-0.35

-0.59

ACS Paragon Plus Environment

ACS Sensors

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 21

Table I: Adsorption distance (dZn−O) and adsorption energy per molecule (Eads) in eV for the interaction of O2 and NO2 with the undoped and H-Doped ZnO (

) surface.

Fig. 6: Side view of the optimized O2/ZnO ( ) interface (a) and NO2/ZnO ( ) interface (b) represented along the [1-210] (left structures) and [000-1] (right structures) directions. Only two bilayers of the ZnO slab are represented beside the adsorbate molecules.

Fig. 7: Density of State (DOS) for the (a) H-doped ZnO ( ) surface, (b) NO2 molecule and (c) oxygen molecule adsorbed on the H-doped ZnO ( ) surfaces. The Fermi level is indicated by the dotted line for all systems. In panel (c) the spin-up and spin-down DOS are represented by the black and red curves respectively.

ACS Paragon Plus Environment

Page 13 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

The ZnO (

) surface is characterized by the presence of ZnO dimers, which are slightly tilted

(see Fig. 6). Each Zn and O atom at the surface is threefold coordinated, and, as such, presents an unsaturated bond (with respect to bulk coordination) which allows for strong interaction with gas molecules. Indeed, when ZnO is exposed to O2 rich atmosphere, i.e., in ambient, our calculations show that oxygen molecules attach to the ZnO surface by bridging two neighboring Zn surface atoms with a binding energy of -0.35 eV. The equilibrium distance between the oxygen atoms of O2 and the Zn atoms at surface is 2.13 Å, in agreement with previous studies26-28. Similarly to the oxygen molecule, NO2 adsorbs at the ZnO non-polar surface in a bi-dentate configuration: the two oxygen atoms bind to two surface Zn atoms while nitrogen points outwards, as shown in Fig. 6b and in agreement to previous studies11, 29. Upon adsorption, the NO2 molecule was observed to be slightly distorted with respect to the gas phase geometry being the N-O bond elongated by 0.05 Å and the bond angle contracted by 12°. This adsorption configuration corresponds to a binding energy of -0.59 eV. Based on the values of BE and irrespective of the presence of H dopant in ZnO (see Table I), it is predicted that NO2 would bind to the ZnO (101) surface more strongly than O2, thus, as the sensor is exposed to a NO2 containing gas, ZnO would be readily covered with NO2 molecules that would either occupy some free binding site or would displace some of the chemisorbed O2 molecules. To understand how the binding of gas molecules change the electronic properties of the ZnO sensors, we analyzed the density of states (DOS) of the ZnO surface interacting with gas molecules and compared it to that of the uncovered surface. We highlight that explicitly considering n-type conductivity (i.e. unintentional doping) in the simulations, although it has almost no effects on the equilibrium geometry and on the values of BE, it is fundamental to understand NO2 sensing mechanism and the “order of magnitude” change in resistance observed

ACS Paragon Plus Environment

ACS Sensors

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 21

experimentally upon gas detection. In previous theoretical studies regarding NO2 detection, the n-type character of ZnO has never been considered, and sensing was explained only based on the analysis of small amount of charge transfer occurring between gas molecules and surface atoms addressed analyzing either Bader or Mulliken charges9. Indeed, such variations are so local and so small that can hardly justify the measured conductivity changes. Fig. 7, represents the DOS for the H-doped ZnO surface before (panel a) and after adsorption of NO2 (panel b) and O2 (panel c) molecules. It is evident that pristine ZnO surface, when doped with hydrogen, is and n-type conductor since the Fermi level is located in the conduction band (see Fig. 7a). When ZnO is exposed to air containing O2 the resistance of the system increases as demonstrated by the presence in the DOS of molecular states in the ZnO energy gap that traps free electrons from the conduction band and pin the Fermi level within the ZnO gap (see Fig. 7c). This corresponds to the condition in which the sensor active element is found before being exposed to NO2. When NO2 binds to the H-ZnO (

) surface the Fermi level shifts further

down in energy and it is found right at the top of the valence band, as shown in Fig. 7b. This indicates that NO2 generates surface levels deeper than those generated by O2 further reducing free carrier concentration in the conduction band and thus inducing an increase in sensor resistance as found experimentally. The effect of temperature increase during NO2 detection has been analyzed by employing AIMD simulations: the distance of the adsorbed molecules (both O2 and NO2) from the surface binding site as a function of time is reported in Fig. 8 for three annealing temperatures (the zero of time correspond to the molecules adsorbed in the minimum energy configuration). It is evident that, while O2 tents to fly away from the surface already at 200 °C, at this temperature NO2 oscillates above the ZnO surface at a bonding distance of 2.67 Å: the molecule is still attached to the

ACS Paragon Plus Environment

Page 15 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

surface and it traps electrons in deep surface states (in this configuration the DOS is similar to that represented in Fig. 7b). At higher temperatures, NO2 is also found to drift away from the ZnO surface with distances larger than 4 Å after 4 ps. It is worth noticing that in our MD simulations no gas pressure in present, thus systems (molecules adsorbed at the ZnO surface) are heated in vacuum conditions. Yet based on our MD results, it is expected that in the temperature range considered here the highest response of ZnO to NO2 would be at 200 °C.

Fig 8: Evolution of the distance of (a) O2, (b) NO2 molecules from the surface Zn atoms during MD simulations at 200 (black curve), 300 (red curve), and 400 °C (green curve).

ACS Paragon Plus Environment

ACS Sensors

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 21

In conclusion, based on our calculations, NO2 detection mechanism can be summarized as follows: before ZnO in exposed to NO2 the sensor surface is covered with O2 molecules present in the atmosphere in ambient conditions; after exposure to NO2, because of the higher NO2 BE to ZnO compared to O2 (see Table I), ZnO would be partially covered with NO2 and, consequently, its resistance would increase due to trapping of electrons in deep surface states appearing in the gap that shift of the Fermi level close to the top of the valence band. Our MD results also show that the NO2/ZnO bond is stable up to 200 °C, while at higher temperatures NO2 molecules start desorbing from the surface releasing back electrons in the conductions band and, as a consequence, measured conductivity changes would be smaller. This latter finding agrees with the experimental measurements reported in Fig. 9, which reveal that the sensor attains enhanced response (i.e. larger resistance change for a given NO2 concentration) for temperature of about 200 °C . In particular, we found that at 200 °C the sensor response appears to be 10 times higher than the one observed at 300 °C and 15 times higher than that at 400 °C . Furthermore, it is evident that NO2 concentration influences sensor response, in particular, highest response is achieved at 0.5 ppm.

ACS Paragon Plus Environment

Page 17 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

Fig 9: Sensors response at different temperatures (from 200 to 400 °C) and NO2 concentrations (from 0.1 to 0.5 ppm).

CONCLUSION Our results reveal that the sensor had optimum performance at 200 °C, in terms of sensitivity and response, with NO2 concentration as low as 0.1 ppm being detectable. Furthermore, we showed that NO2 detection is driven almost exclusively by competitive adsorption with atmospheric O2. In particular we found that, upon NO2 adsorption at the ZnO surface, deep acceptor levels are generated in the ZnO energy gap which are responsible for free electron trapping and for the resistance increase observed experimentally. Moreover, our combined theoretical and experimental results show that enhanced sensor response is achieved at the temperature for which O2 molecules start flying away from the ZnO surface while NO2 are still strongly bound to it. Our unified study gives insights on detection mechanism, which may provide crucial details essential for improving the performance of ZnO based devices for NO2 detection and other related application.

ACS Paragon Plus Environment

ACS Sensors

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 21

Acknowledgements The authors acknowledge funding from the European Union Seventh Framework Programme under grant agreement ITN-Nanowiring, (Grant N. 265073). We acknowledge CHPC (Capetown) and CINECA (Italy) for high performance computing resources and support.

REFERENCES (1) Scorer, R. Fundamentals of Air Pollution. Quarterly Journal of the Royal Meteorological Society. 1985, 111, 672-672. (2) W. H. O. R. O. for Europe, Air quality guidelines for Europe, 91 (WHO Regional Office Europe, 2000). (3) Comini, E; Faglia, G; Sberveglieri,G. UV light activation of tin oxide thin films for NO2 sensing at low temperatures. Sensors and Actuators B: Chemical. 2001, 78, 73-77. (4) Tanaka, K; Morimoto, S; Sonoda, S; Matsuura, S; Moriya, K; Egashira, M. Combustion monitoring sensor using tin dioxide semiconductor. Sensors and Actuators B: Chemical. 1991, 3, 247-253. (5) White, N; Turner, J. Thick-Film Sensors: Past, Present and Future.. Measurement Science and Technology. 1997, 8, 1 . (6) Sberveglieri, G; Benussi, P; Coccoli, G; Groppelli, S; Nelli, P. Reactively sputtered indium tin oxide polycrystalline thin films as NO and NO2 gas sensors. Thin Solid Films. 1990, 186,349-360. (7) Koshizaki, N; Oyama, T Sensing Characteristics of ZnO-Based NOx Sensor.. Sensors and Actuators B: Chemical. 2000, 66, 119-121. (8) Ozgur, U; Alivov, Y. I; Liu,C; Teke, A; Reshchikov, M; Do_gan, S; Avrutin, V; Cho, S.-J; Morkoc, H. A comprehensive review of ZnO materials and devices. J. Appl. Phys. 2005, 98, 041301. (9) Tomchenko, A. A; Harmer, G. P; Marquis,B. T; Allen, J. W. Semiconducting metal oxide sensor array for the selective detection of combustion gases. Sensors and Actuators B: Chemical. 2003, 93, 126-134.

ACS Paragon Plus Environment

Page 19 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

(10) Rodriguez,J. A; Jirsak, T; Dvorak, J; Sambasivan, S; Fischer, D. Reaction of NO2 with Zn and ZnO: Photoemission, XANES, and Density Functional Studies on the Formation of NO3 J. Phys. Chem. B. 2000, 104, 319-328. (11) Spencer, M. J; Yarovsky, I. J. ZnO Nanostructures for Gas Sensing: Interaction of NO2, NO, O, and N with the ZnO(1010) Surface. J. Phys. Chem. C. 2010, 114, 10881-10893. (12) Sadek, A. Z; Choopun, S; Wlodarski, W; Ippolito, S. J; Kalantar-zadeh, K. Characterization of ZnO nanobelt-based gas sensor for H 2, NO 2, and hydrocarbon sensing. Sensors Journal, IEEE. 2007, 7, 919-924. (13) Carotta, M; Cervi, A; Di Natale, V; Gherardi, S; Giberti, A; Guidi, V; Puzzovio, D; Vendemiati, B; Martinelli, G; Sacerdoti, M. ZnO gas sensors: a comparison between nanoparticles and nanotetrapods-based thick films. Sensors and Actuators B: Chemical. 2009, 137, 164-169. (14) Calestani, D; Mosca, R; Zanichelli, M; Villani, M; Zappettini, A. Aldehyde detection by ZnO tetrapod-based gas sensors. Journal of Materials Chemistry. 2011, 21, 15532-15536. (15)Podobinski, D; Zanin, S; Pruna, A; Pullini, D. Effect of annealing and room temperature sputtering power on optoelectronic properties of pure and Al-doped ZnO thin films. Ceramics International. 2013, 39, 1021-1027. (16) Giannozzi,P; Baroni, S; Bonini, N; Calandra, M; Car, R; Cavazzoni, C; Ceresoli, D; Chiarotti, G; Cococcioni, M; Dabo, I; et al. QUANTUM ESPRESSO: a modular and opensource software project for quantum simulations of materials. J. Phys.: Condens. Matter. 2009, 21, 395502. (17) Kohn, W; Sham, L. J. Self-consistent equations including exchange and correlation effects. Phys. Rev. 1965, 140, A1133. (18) Vanderbilt, D. Soft self-consistent pseudopotentials in a generalized eigenvalue formalism. Phys. Rev. 1990, B 41, 7892. (19) Monkhorst, H. J; Pack, J. D. Special points for Brillouin-zone integrations. Phys. Rev. 1976, B 13, 5188-5192. (20) Calzolari, A; Ruini, A; Catellani, A. Anchor Group versus Conjugation: Toward the GapState Engineering of Functionalized ZnO(101̅0) Surface for Optoelectronic Applications. J. Am. Chem. Soc. 2011, 133, 5893-5899. (21) Hoover, William. G. Canonical dynamics: Equilibrium phase-space distributions. Phys. Rev. 1985, A 31, 1695–1697. (22) Verlet, L. Computer "Experiments" on Classical Fluids. I. Thermodynamical Properties of Lennard-Jones Molecules. Physical Review. 1967, 159, 98–103.

ACS Paragon Plus Environment

ACS Sensors

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 21

(23) Shi, Feng; Xue, Chengshan. Morphology and growth mechanism of novel zinc oxide nanostructures synthesized by a carbon thermal evaporation process. CrystEngCom. 2012, 14, 5407-5411. (24) Korir, K. K; Catellani, A; Cicero, G. Ethanol Gas Sensing Mechanism in ZnO Nanowires: An ab Initio Study. J. Phys. Chem. C. 2014, 118, 24533–24537. (25) Møller, M. F. A Scaled Conjugate Gradient Algorithm For Fast Supervised Learning. Neural networks. 1993, 6, 525-533. (26) Li, Limiao; Wang, Taihong; Yuan, Quanzi; Zhao, Ya-Pu. Ab Initio Study of ZnO-Based Gas-Sensing Mechanisms: Surface Reconstruction and Charge Transfer. J. Phys. Chem. C. 2009, 113, 6107–6113. (27) Yan, Yanfa; Al-Jassim, M. M; Wei, Su-Huai. Oxygen-vacancy mediated adsorption and reactions of molecular oxygen on the ZnO(1010) surface. Phys. Rev. B. 2005, 72, 161307. (28) Xu, H; Zhang, R. Q; Tong, S. Y. Interaction of O2, H2O, N2, and O3 with stoichiometric and reduced ZnO(1010) surface. Phys. Rev. B. 2010, 82, 155326. (29) Prades, J. D; Cirera, A;

Morante, J. R. Ab initio calculations of NO2 and SO2 chemisorption onto non-polar ZnO surfaces. Sensors and Actuators B: Chemical. 2009, 179– 184.

ACS Paragon Plus Environment

Page 21 of 21

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

SYNOPSIS TOC.

Illustration 1: Graphical representation of sensing mechanism of NO2 at 200C

ACS Paragon Plus Environment