Abstract
Rare earth oxycarbonates are potential candidate materials for constructing simple and low-cost chemiresistive sensors for monitoring carbon dioxide (CO2) gas in the living and working environment for personal comfort and health reasons. Also, measurement of CO2 concentrations is needed in many industrial processes. Specifically, sol-gel made nanoparticles of Nd and La oxycarbonates have been studied previously as novel CO2 gas sensor materials. In this paper, pulsed laser deposition of La oxycarbonate (La2O2CO3) thin films was studied and structural properties of obtained thin films were characterized. Also, CO2 gas sensing ability of synthesized films was evaluated. The films deposited under CO2 partial pressure in various conditions were all Raman amorphous. In situ or ex situ annealing procedure at high CO2 partial pressure was needed for obtaining crystalline La2O2CO3 films, whereby hexagonal and monoclinic polymorphs were obtained in ex situ and in situ processes, respectively. Sensor structure, made using in situ process, was sensitive to CO2 gas and showed relatively fast response and recovery characteristics.
1. Introduction
CO2 is one of the harmful pollutant gases in closed rooms and also in the environment. Therefore, there is a demand for effective methods for monitoring CO2 levels. Current industrially available CO2 sensors are relatively complex and expensive, so it is desirable to develop simple and low-cost methods for monitoring CO2 level [1]. La and Nd oxycarbonates have recently been studied as promising candidates for chemiresistive CO2 sensor materials [2–6]. However, in these reports, La and Nd oxycarbonates have been studied in a form of nanorods and there are no investigations of thin film deposition of these promising materials for CO2 gas sensor application. In [2–5], La and Nd oxycarbonate materials were made using respective precursor hydroxides in a form of nanoparticles prepared in a sol-gel process. Conversion of La and Nd hydroxide nanorods to respective oxycarbonates was carried out by heating the materials in the atmosphere containing CO2 gas.
Within the class of oxycarbonate materials, pulsed laser deposition (PLD) has been used for the synthesis of Ba-Ca-Cu, Sr-Cu, and Ba-Sr-Cu oxycarbonate thin films, which have been studied as superconductors [7–10]. In all of these reports [7–10], pure oxide materials were used as pulsed laser deposition targets. Oxycarbonate thin films were obtained by postdeposition furnace annealing [8] or by in situ annealing of deposited films at higher gas pressure [7, 9, 10].
In general, rare earth oxycarbonates are stable compounds in ambient conditions. Their synthesis, thermal stability, and crystalline modifications (metastable polymorphs (I) tetragonal, (Ia) monoclinic, and (II) hexagonal) are described in [11–14].
In this work, lanthanum oxycarbonate thin films were deposited using La2O2CO3 ceramic targets. Two different postgrowth treatments for obtaining crystalline La2O2CO3 films were used. Obtained films were characterized by electron microscopy, X-ray diffraction, and Raman spectroscopy methods. CO2 gas sensing capability of grown La2O2CO3 films was demonstrated.
2. Materials and Methods
Lanthanum oxycarbonate films with ~180 nm thickness were deposited by PLD method on two types of substrates: (a) Si(100)/SiO2 (25 nm thermal oxide) substrates for structural characterization and (b) Windsor Scientific SM160 microelectrode substrates with Pt electrodes (Figure 1). SM160 has seven independently addressable electrode pairs with different electrode separation distances ranging from 4 to 200 μm (a-type areas in Figure 1). The length of six electrode gaps is 0.6 mm. Films were deposited onto electrode substrate using a stainless steel mask that left open only the central part of the substrate. Mask ensured that no film was deposited onto contact pads. Prior to deposition, the substrates were cleaned by processing in an ultrasonic bath in acetone and rinsing in acetone and methanol. La2O2CO3 ceramic targets were made from La2(CO3)3 6H2O (purity 99,9%) powder, which was dried in vacuum oven at 200°C during 24 h and pressed into pellets at 400 MPa. Following thermal treatment of pellets was carried out in two steps: (1) the temperature was raised slowly up to 500°C in air and (2) the pellets were sintered in CO2 and O2 gas mixture (1 : 9) at 800°C for 15 h. A KrF excimer laser (COMPexPro 205, Coherent; wavelength 248 nm, pulse width 25 ns) was used for PLD. During deposition of La2O2CO3 films, the energy density of laser pulse on the target was 1.5 J/cm2, the pulse repetition rate was 10 Hz, and the distance between the substrate and the target was 7.5 cm. Pure CO2 or O2/10% CO2 background gas at 10−2 or 10−1 mbar pressure was used during the deposition process. The substrate temperature was varied between 300 and 700°C. Other details of the deposition process are described elsewhere [15]. Two types of annealing procedures were used for after-growth processing of the films: (a) ex situ annealing or (b) in situ annealing process. Ex situ annealing was carried out in a tube furnace at 550°C for 30 min in flowing humidified O2/10% CO2 gas mixture at normal pressure. Humidification was realized by guiding the gas mixture through the distilled water. In situ annealing was carried out immediately after the PLD process in the deposition chamber under the CO2 gas atmosphere at the pressure of 7 mbar and at the temperature of 660°C. In situ annealing time was 30 min.
The structure of the deposited films was characterized by grazing incidence X-ray diffraction (GIXRD, incidence angle 0.5°), scanning electron microscopy (SEM), Raman spectroscopy, and spectroscopic ellipsometry. The measurements of electrical and gas response characteristics were carried out with a Keithley 2400 SourceMeter, a gas mixing system based on five Brooks mass flow controllers, and a sample chamber with small hotplate heater (see [16]). A mixture of N2 and O2 gases (both 99.999% pure) was used as a carrier gas. The source of CO2 additive was a mixture of 10% CO2 in O2 (99.999% pure). The oxygen concentration (21%) and the flow rate (200 ml/min) of the gas mixture were kept constant during the measurements.
3. Results and Discussion
Figure 2 depicts the Raman spectrum of a PLD target synthesized using pure lanthanum carbonate La2(CO3)3 6H2O. The Raman bands in this spectrum can be attributed either to monoclinic or to hexagonal La2O2CO3, which are commonly observed La oxycarbonate crystal phases and are generally marked as phase (Ia) and phase (II), respectively. Both polymorphs have their distinctive bands in the Raman spectrum. (II) La2O2CO3 is characterized by intense bands at 363 and 393 cm−1, along with a sharp band at 1088 cm−1. (Ia) La2O2CO3 can be easily identified by two bands at 1050 and 1060 cm−1 [17]. All intense peaks in the spectrum of PLD target shown in Figure 2 belong to (II) La2O2CO3 phase. However, weaker bands at 1050 and 1060 cm−1 indicate the presence of Ia type minority polymorph in the target. This is supported also by four weak bands appearing at 223, 297, 442, and 458 cm−1 in the spectrum, which can all be attributed to the monoclinic La2O2CO3 phase. Therefore, according to Raman analysis, our PLD target consisted mainly of hexagonal La2O2CO3 with only small traces of monoclinic phase. No indication of any other phases, for example, La oxide or La hydroxide phases, was found in Raman analysis. It has to be noted that the coexistence of type (Ia) and (II) La2O2CO3 phases is a previously known phenomenon and at equilibrium conditions type Ia transforms to type (II) at temperatures over 400°C [12, 17].
According to the analysis of Raman spectra, thin films grown at various deposition conditions were all amorphous and showed no Raman peaks. Ex situ or in situ annealing at higher CO2 partial pressure was needed for film crystallization. This is illustrated in Figure 3 where (a) is the spectrum of the film grown at 500°C using O2/10% CO2 gas mixture at mbar. Raman bands characteristic to La2O2CO3 are absent in graph (a) of Figure 3. Only Raman peaks originating from Si/SiO2 substrate can be seen. This is easily confirmed by comparing with spectrum taken from a pure substrate, graph (b). Graphs (c) and (d) depict Raman spectra of ex situ and in situ annealed thin films, respectively. Raman spectrum of the film annealed ex situ at 550°C (Figure 3(c)) has a clear band at 1088 cm−1, which is characteristic for hexagonal La2O2CO3 phase. In addition, Raman spectrum in Figure 3(c) also contains two weak Raman bands at 362 and 393 cm−1 which can be assigned to (II) La2O2CO3 phase. The spectrum of the film annealed in situ at 660°C (Figure 3(d)) shows two clear peaks at 1050 and 1060 cm−1, which belong to monoclinic La2O2CO3 phase. Spectrum (d) contains also a weak band at 224 cm−1 assignable to monoclinic phase.
XRD analysis of ex situ and in situ annealed films, in general, confirmed the results of Raman spectroscopy. XRD patterns are depicted in Figure 4. Both diffraction patterns can be assigned to crystalline La2O2CO3, whereas the peaks originating from ex situ and in situ annealed films can be attributed to hexagonal II (ICDD PDF-2 file number 37-0804, space group P63/mmc) or monoclinic (ICDD PDF-2 file number 01-70-5540, space group C12/c1) and tetragonal (ICDD PDF-2 file number 23-0320, space group I4/mmm) or monoclinic (ICDD PDF-2 file number 48-1113, space group unknown) phases, respectively. XRD pattern cannot distinguish between hexagonal and monoclinic phases of ex situ annealed film because both phases show quite similar diffraction patterns (Figure 4, patterns denoted as HEX and MONOC). The space groups P63/mmc and C12/c1 can be distinguished only by very weak extra reflections of monoclinic phase that cannot be detected because of a low average intensity of observed diffraction patterns. However, it is important to note that the monoclinic polymorph is just more refined model structure of type (II) La2O2CO3 [18]. Typically, only average, hexagonal, model of type (II) polymorph is considered since it cannot be distinguished from monoclinic polymorph by applying typical experimental setups for structural analysis. Therefore, it is reasonable to consider that our ex situ annealed films consist of type (II) hexagonal La oxycarbonate crystal phase.
The monoclinic phase is the preferred assignment over tetragonal phase for in situ annealed films because the former gave a better match for all weak reflections in the following ranges of XRD pattern: 25–27°, 30.7–38.7°, and 49–59° (Figure 3). However, it cannot be completely ruled out that some amount of tetragonal phase is present in in situ annealed films. A strong and sharp XRD peak near 56° originates from the substrate and can be assigned to reflection 311 of Si (diffraction angle is 56.12° according to ICDD PDF-2 file number 27-1402).
SEM images of in situ and ex situ annealed La2O2CO3 films are shown in Figure 5. Crystallite grains of film material are clearly noticeable in the SEM images. Also, it can be seen from the SEM images that both films have relatively high concentration of particulates with sizes up to 350 nm in diameter. Note that the average thickness of the films is approximately 180 nm. Micro- and nanoparticle production during PLD process is a usual phenomenon [19, 20]. These particles are formed either from solidified liquid droplets that are expelled from the target during laser-target interaction process or by nucleation and growth of nanoparticles in propagating plasma plume of gas phase species [19]. Droplets originating from the target have typical sizes in the micron or submicron range, while the sizes of the particulates created in PLD plasma are in the nanometer range. The quantity of particulates could possibly be reduced by optimizing process parameters, such as laser fluence and background gas pressure. A noticeable feature is that the particulates in the SEM image of ex situ annealed film (Figure 5(a)) are split into two or more parts, while the particulates in the image of in situ annealed film (Figure 5(b)) are intact. The splitting of particulates is probably related to the hydroxylation and partial carbonation of La2O3, which is probably the phase of as-grown films (see discussion below), under ambient atmosphere. As reports of SEM images of oxycarbonate films are scarce, we cannot compare the appearance of our La2O2CO3 films with previously obtained results.
(a) SEM image of ex situ annealed La2O2CO3 film
(b) SEM image of in situ annealed La2O2CO3 film
Gas sensor measurements were made at 450°C, with CO2 gas concentrations of 750 and 1500 ppm. Relative humidity of the test gas was 21%. The measurements were made using the electrodes with 48 µm gap width and 0.6 mm gap length. The current through the sample was monitored at the applied constant voltage of 3 V. Previously, it is reported that the sensor structures based on La and Nd oxycarbonate nanorods have high electrical resistivity and that restrict their use in commercial sensors [2–4]. In our sensor structures, the resistivity followed the Arrhenius law with an activation energy of 0.56 eV between 250 and 450°C. It shows that the conductivity is determined by deep trap levels, as the bandgap of the material, as concluded from X-ray absorption spectroscopy and X-ray emission spectroscopy studies by Hirsch et al. [6], is approximately 3.7 eV in the case of La2O2CO3 nanoparticles. At 250°C, the resistivity was at the limit of our detection capability as the detected current was slightly below 1 nA. Reasonable signal and fastest response times were obtained at the highest temperature probed, at 450°C, when the resistivity of the material was 340 Ω·m. Time transients of the CO2 gas response of in situ annealed sensor structure are depicted in Figure 6. It can be seen that the relative gas response to 1500 ppm of CO2 is about 2.2; that is, the measured current decreased from to A. The response and recovery times of the sensor signal, determined as the times at which the 90% level of the final signal were achieved, are 165 s and 189 s, respectively. Our thin film sensor structures have considerably faster responses as compared to sensors based on La and Nd oxycarbonate nanorods [2–5]. It is especially notable in the case of recovery times, which have been found to be over 30 min in [2–5].
According to Hirsch et al. [6], CO2 adsorbing onto monoclinic La2O2CO3 interacts directly with La site on the surface and forms surface carbonate at the site of absorption. Adsorbing CO2 acts as an electron acceptor, thus reducing the concentration of free electrons and causing an increase of resistivity of the sensor material. Also, recent studies indicate that surface hydroxyl groups play important role in CO2 adsorption at La2O2CO3, but exact mechanism and the role of OH groups are not known [4, 6].
The difference between the films obtained by in situ and ex situ annealing procedures can be explained as follows. As discussed in the works of Orera et al. and Bakiz et al. [17, 21], the La2O3-H2O-CO2 ternary system is relatively complex. The phases obtained in different experiments and described in the literature vary and depend considerably on the thermal history of the sample and various experimental conditions. However, our results can be clearly interpreted based on the results and discussion of the paper by Orera et al. [17]. According to Raman analysis, our as-grown films do not contain crystalline phases. It can be presumed that the CO2 pressure that was used during the film growth was not sufficient for the formation of a stable oxycarbonate phase. It can be supposed that the as-grown film consists mainly of amorphous lanthanum oxide but it may also be partially carbonated to form a type nonstoichiometric amorphous phase. When deposited film is taken out from the deposition chamber, it gets into the contact with H2O and CO2 in ambient air. When La2O3 is in contact with ambient air, it rapidly forms La(OH)3. According to previous literature reports [17, 22, 23], the aging of La(OH)3 in air at room temperature (RT) results in the formation of partially hydrated carbonates of type. In [17], the aging of La(OH)3 material at RT in ambient air was followed by annealing at 400°C, and, as a result, the sample was crystallized into pure hexagonal La2O2CO3 phase. However, when the high-temperature annealing was performed with La(OH)3, that is, without longer aging of the material at RT, the authors found, as a final product, monoclinic type La2O2CO3. The carbonization of La2O3 under CO2 flow was also studied in [21]. The transformation of La2O3 to monoclinic type La2O2CO3 was observed at the temperatures around 525°C. Our results are in agreement with these previous outcomes as our ex situ annealed films, which got into contact with ambient air, were crystallized into hexagonal phase. In addition, if our laser deposited film was annealed in situ, without exposing it to ambient air, monoclinic La2O2CO3 was obtained.
4. Conclusions
La oxycarbonate thin films were obtained by pulsed laser deposition and postannealing in CO2 containing atmosphere using ceramic La2O2CO3 target. The structure of as-deposited films was amorphous. In order to obtain crystalline oxycarbonate films, in situ or ex situ annealing under high CO2 partial pressure was necessary. The type of crystalline La oxycarbonate phase depended on annealing method as hexagonal, or type (II) phase was obtained by ex situ annealing and monoclinic phase, or type Ia was obtained by in situ annealing procedure. The main difference between two annealing methods was that the films, subjected to ex situ annealing, got into contact with atmospheric water vapor. Two different routes for obtaining La2O2CO3 films in this work can schematically be described as follows: (a) ex situ annealing: La2O3 → La(OH)3 → → hexagonal La2O2CO3; (b) in situ annealing: La2O3 → monoclinic La2O2CO3. The films obtained by in situ annealing showed a relatively fast conductometric response to CO2 gas in the actual concentration range (400–2000 ppm). The conductivity of 180 nm thick thin films in the pure synthetic air was 2.9 mS/m and it decreased over 2 times under the influence of 1500 ppm of CO2. The gas response and recovery times determined at 450°C were within 165–189 s.
Disclosure
The preliminary results have been presented at the E-MRS 2015 Fall Meeting.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
This study was supported by Estonian Research Council (Institutional Projects IUT34-27 and IUT2-24) and by European Regional Development Fund (SA Archimedes Project 3.2.1101.12-0014). The authors gratefully thank Aare Floren for help with technical problems.