Humidity-independent Oxide Semiconductor Chemiresistors Using

Subscriber access provided by - Access paid by the | UCSB Libraries

Functional Inorganic Materials and Devices

Humidity-independent Oxide Semiconductor Chemiresistors Using Terbium-doped SnO2 Yolk-Shell Spheres for Real-time Breath Analysis Chang-Hoon Kwak, Tae-Hyung Kim, Seong-yong Jeong, Ji-Won Yoon, Jun-Sik Kim, and Jong-Heun Lee ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.8b04245 • Publication Date (Web): 16 May 2018 Downloaded from http://pubs.acs.org on May 16, 2018

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 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 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.

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 34 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 Applied Materials & Interfaces

Humidity-independent Oxide Semiconductor Chemiresistors Using Terbium-doped SnO2 YolkShell Spheres for Real-time Breath Analysis Chang-Hoon Kwak,†Tae-Hyung Kim,†Seong-Yong Jeong,†Ji-Won Yoon,†Jun-Sik Kim,† and Jong-Heun Lee*,† †

Department of Materials Science and Engineering, Korea University, Seoul 02841, Republic of

Korea *Author to whom correspondence should be addressed Email: [email protected]; Fax: +82-2-928-3584; Tel: +82-2-3290-3282 Keywords: Tb-doped SnO2, gas sensor, yolk-shell spheres, humidity dependence, acetone, diabetes

ABSTRACT The chemiresistive sensing characteristics of metal oxide gas sensors depend closely on ambient humidity. Herein, we report that gas sensors using Tb-doped SnO2 yolk-shell spheres can be used for reliable acetone detection regardless of the variations in humidity. Pure SnO2 and Tb-doped SnO2 yolk-shell spheres were prepared via ultrasonic spray pyrolysis and their chemiresistive sensing characteristics were studied. The sensor resistance and gas response of the pure SnO2 yolk-shell spheres significantly changed and deteriorated upon exposure to moisture. In stark contrast, the Tb-doped SnO2 yolk-shell spheres exhibited similar gas responses and sensor resistances in both dry and humid (relative humidity (RH) 80%) atmospheres. In addition,

ACS Paragon Plus Environment

1 / 34

ACS Applied Materials & Interfaces 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 2 of 34

the Tb-doped SnO2 yolk-shell sensors showed a high gas response (resistance ratio) of 1.21 to the sub-ppm-levels (50 ppb) of acetone with low responses to the other interference gases. The effects of Tb oxide and the chemical interactions among the Tb oxide, SnO2, and water vapor on this humidity-independent gas-sensing behavior of the Tb-doped SnO2 yolk-shell sensors were investigated. This strategy can provide a new road to achieve highly sensitive, selective, and humidity-independent sensing of acetone, which will facilitate miniaturized and real-time exhaled breath analysis for diagnosing diabetes.

1. INTRODUCTION Recently, the sensing of biomarker gases from human exhaled breath using semiconductortype chemiresistors has attracted great attention because it provides non-invasive, convenient, cost-effective, portable, and rapid disease diagnosis.1-4 In particular, the exhaled acetone is considered as a potential biomarker to diagnose diabetes.1,5 The concentrations of breath acetone in diabetic patients are reported to be higher than 1.8 ppm, whereas those of healthy people are lower than 0.8 ppm.6 Gas chromatography or ion mobility spectrometry have been used to measure the concentrations of breath acetone in a precise manner.7 However, these analytic instruments are not only expensive and bulky but also require sampling and pre-conditioning (e.g. dehumidification) of exhaled breath, which hampers the real-time breath analysis.4 Oxide semiconductor chemiresistors are advantageous for the portable and in situ analysis of exhaled breath because of their high sensitivity, simple structures, facile integration, and low power consumption.8,9 However, for real applications, the following two challenges should be overcome. Firstly, the influence of water vapor on the gas-sensing characteristics of these


2 / 34

Page 3 of 34 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


devices should be decreased or removed because atmospheric water vapor (H2O) forms inactive hydroxyl groups (OH) on the sensing surface by reaction with negatively charged surface oxygen,10,11 which markedly changes the resistance of the sensor and affects its gas response. Secondly, the selectivity toward a specific biomarker gas is essential because the exhaled breath consists of few hundreds of different volatile organic compounds.12 Accordingly, humidityindependent oxide semiconductor gas sensors with high selectivity toward a specific biomarker gas should be designed for reliable exhaled breath analyses. Note that the moisture-induced poisoning of gas-sensing characteristics has been one of the greatest obstacles in designing reliable oxide semiconductor gas sensors since its discovery in the 1960s.13 In addition, for the widespread applications of oxide semiconductor gas sensors, it is imperative to overcome their humidity dependence because the ambient moisture is not only high but can also continuously change depending on the climate, temperature, region, and season. Various oxide chemiresistors such as Pt-SnO2,9 Pt-WO3,14 Si-WO3,15 and Au-In2O316,17 have been investigated in order to fabricate exhaled breath sensors that can be operated under nearly saturated humidity level (e.g., RH > 80%). However, the previous studies in this context did not study the humidity dependence of the chemiresistive sensing behaviors of exhaled breath gas sensors. Various approaches such as controlling the sensor temperature,18 changing the adsorbed oxygen species,19 and loading or doping of Rh,20 NiO,21 and CuO22 additives with high affinity to moisture have been employed to reduce the humidity dependence of the chemiresistive sensing behaviors of oxide semiconductor chemiresistors. It should be noted that the high affinity of additive materials toward moisture may not be sufficient to overcome the humidity dependence of chemiresistive sensing behaviors. This is because the moisture-absorbing capacity of additive materials is limited, whereas moisture is


3 / 34


Page 4 of 34

continuously provided from the ambient atmosphere. To address this issue, in one of our previous studies, we suggested a strategy to remove the water-poisoning effect by refreshing of the In2O3 sensor surface assisted by CeO2 additives with facile redox-reaction due to multivalency of Ce.23 Terbium oxide is also known to be multivalent and thus has been used as a key additive to increase the oxygen storage capability of CeO2-based automotive three-way catalysts by promoting redox reactions.24,25 Moreover, there has been a report on the enhancement of gas response by doping Tb to SnO2.26 In this viewpoint, it is worthwhile investigating the effect of Tb doping on the humidity dependence of the gas-sensing behaviors of oxide chemiresistors. However, the possibility of terbium oxide as an additive to remove the water-poisoning effect in oxide semiconductor chemiresistors have never been investigated. In this study, we developed a novel sensing material, Tb-doped SnO2 yolk-shell spheres, to overcome the water-poisoning effect in oxide semiconductor chemiresistors. To study the effect of Tb doping on the resulting chemiresistor, we synthesized pure and Tb-doped SnO2 yolk-shell spheres by spray pyrolysis and investigated their chemiresistive sensing behaviors in dry and wet atmospheres. The sensors using Tb-doped SnO2 yolk-shell spheres showed similar sensor resistance and gas response both in dry and highly humid (RH 80%) atmospheres. These sensors also showed a high response and good selectivity to sub-ppm acetone. Hence, the sensors using Tb-doped SnO2 yolk-shell spheres can be used for portable and in situ sensing of exhaled breath for the diagnosis of diabetes.


4 / 34

Page 5 of 34 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


2. EXPERIMENTAL Preparation of Tb-doped SnO2 yolk-shell spheres. Pure SnO2 yolk-shell spheres were prepared by ultrasonic spray pyrolysis method and a subsequent heat treatment. An aqueous spray solution was prepared by dissolving of Sn (II) oxalate (6.3285 g, SnC2O4, 98%, SigmaAldrich, USA) and sucrose (51.6030 g, C12H22O11, ≥99.5 %, Sigma-Aldrich, USA) in distilled water (270 mL) was prepared. Hydrogen peroxide solution (30 mL, 30%, Sigma-Aldrich, USA) was added to obtain a clear stock solution.27,28 The SnO2 spheres were prepared by spray pyrolysis of solution (0.5M tin oxalate, 0.1 M sucrose) at 1000 °C and were then annealed at 600 °C for 2 h. Experimental setup is shown in Figure S1 and detailed experimental procedures were shown elsewhere.27 Tb-doped SnO2 yolk-shell spheres were also synthesized by ultrasonic spray pyrolysis. Terbium(III) chloride hexahydrate (0.1120, 0.5601, and 1.6802 g, TbCl3·6H2O, Sigma-Aldrich, USA), Sn(II) oxalate (6.3285 g), and sucrose (51.6030 g) were dissolved in 300 mL of mixed solution (30 mL of 30% hydrogen peroxide solution diluted with 270 mL of distilled water). The [Tb]/[Sn] molar ratios of the three spray solutions were 0.01, 0.05, and 0.15. The Tb-doped SnO2 yolk-shell spheres were prepared by the spray pyrolysis at 1000 °C and the subsequent annealing at 600 °C for 2 h. The specimens obtained after the heat treatment were denoted as 1Tb-SnO2, 5Tb-SnO2, and 15Tb-SnO2. Characterization. The phase and crystal structure of the spheres were investigated by X-ray diffraction (D/MAX-2500 V/PC, Rigaku, Japan; CuKα, λ = 1.5418 Å) The morphology of the spheres was characterized using TEM (Talos F200X, FEI Co., USA) and SEM (S-4800, Hitachi Co. Ltd. Japan). The X-ray photoelectron spectroscopy (Thermo Scientific, MultiLab 2000) was used to analyze the sensing materials. The pore size distribution and surface area were


5 / 34


Page 6 of 34

determined by carrying out the Braunauer–Emmett–Teller (BET) analysis of their nitrogen adsorption measurements (Tristar 3000, Micromeritics, USA). Thermogravimetric (TG) analysis (Perkin-Elmer TGA7, USA) was carried out by heating from 25 to 700 °C at a rate of 5 °C/min under air flow. Gas-sensing characteristics. Pure and Tb-doped SnO2 spheres were dispersed in distilled water and the resulting slurry was coated on alumina substrate with electrodes and heater. The gas concentrations and humidity (dry atmosphere or RH 20-80%) were independently controlled by mixing the gases with dry/humid synthetic air (flow rate: 200 cm3 min-1). Detailed procedure for measuring gas sensing characteristics were shown elsewhere.23 Note that the sensors are locally heated to 350–450 °C using microheater and room-temperature ambient gases with different humidity were provided in order to measure the humidity dependence of the gassensing characteristics. The gas responses (S = Ra/Rg; Ra and Rg: sensor resistances in air and analyte gas) to 20 ppm acetone, ammonia, carbon monoxide, toluene, p-xylene, and benzene were measured at 350–450 °C. Two-probe DC resistance was measured using a computer controlled electrometer.

3. RESULTS AND DISCUSSION The pure SnO2 and Tb-doped SnO2 yolk-shell spheres were prepared by one-pot ultrasonic spray pyrolysis of spray solution at 1000 °C. The development of the yolk-shell morphology in these samples can be explained by the occurrence of the following events: (1) evaporation of solvent; (2) formation of the C-Tb-Sn-precursor (or C-Sn-precursor) sphere by the polymerization/carbonization of sucrose; (3) formation of the Tb-doped SnO2 (or SnO2) shell by the decomposition/oxidation of the outer part of the C-Tb-Sn-precursor (or C-Sn-precursor)


6 / 34

Page 7 of 34 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


sphere; (4) contraction of the inner precursors and sequent converting into the Tb-doped SnO2 (or SnO2) yolk by decomposition/oxidation (Figure S2). Approximately 3% of weight loss was observed below 550°C from the TG analysis of as-prepared powders (Figure S3), which can be attributed to the residual components due to short residence time (~ 30 s) of droplets at high temperature region. Thus, the precursor powders were annealed at 600°C for 2 h. The SEM images of the pure SnO2 spheres showed that they had a yolk-shell morphology with semi-transparent shells. This is consistent with the TEM results (Figure S4). The average diameter of the outer shells was 0.64 ± 0.29 µm and the shell was 96 ± 20 nm thick. Three terbium added SnO2 specimens (1Tb-SnO2, 5Tb-SnO2, and 15Tb-SnO2) also exhibited the yolkshell morphology and had spheres with similar sizes (Figure S5a,b and 1a,b,e,f). The average diameters of the outermost shells for the 1Tb-SnO2, 5Tb-SnO2, and 15Tb-SnO2 spheres were found to be 0.63 ± 0.35, 0.65 ± 0.43, and 0.69 ± 0.34 µm, respectively. The shells of the 1TbSnO2, 5Tb-SnO2, and 15Tb-SnO2 spheres were 87 ± 19, 71 ± 15 and 68 ± 17 nm thick, respectively. The thickness of these shells was substantially lower than that of the pure SnO2 spheres (96 ± 20 nm). The compositional mapping images showed uniform distribution of Sn, Tb, and O over the 1Tb-SnO2 yolk-shell spheres (Figure S5d). The high-angle annular dark-field images (Figures 1d1, h1) of the 5Tb-SnO2 and 15Tb-SnO2 specimens showed ring-like whiter contours in the shells. However, no significant compositional variation was observed in these rings, suggesting that Tb was uniformly distributed over the 5Tb-SnO2 and 15Tb-SnO2 spheres (Figures 1d, h). In the spray pyrolysis reaction, since each droplet becomes a reaction container, the compositional heterogeneity is restricted within a microsphere unless the sphere was disintegrated into smaller nanoparticles. This explains the uniform distribution of Tb over the


7 / 34


Page 8 of 34

entire spheres in the present study, which will facilitate the intimate interaction between moisture and Tb-related component. The XRD pattern of the pure SnO2 yolk-shell spheres showed (110), (101), and (211) peaks, which are characteristic of the rutile phase (JCPDS 41-1445) (Figure 2a). The 1Tb-SnO2 and 5Tb-SnO2 spheres exhibited only the rutile SnO2 phase with no secondary phases (Figures 2b, c). This indicates that Tb ions were successfully incorporated into the SnO2 lattice. This can be attributed to the fact that at the coordination number of 4, the ionic radii of Tb3+ (90 Å) and Sn4+ (83 Å) are quite similar.29 Meanwhile, in the case of the 15Tb-SnO2 yolk-shell spheres (Figure 2d), in addition to the rutile SnO2 phase, the Tb2O3 (JCPDS 86-2478) and Tb2(Sn2O7) (JCPDS 87-1221) phases were also observed. According to the Scherrer equation, the crystallite sizes of pure SnO2, 1Tb-, 5Tb-, and 15Tb-SnO2 spheres were determined to be 22.7, 19.4, 14.5, and 14.6 nm, respectively. The chemical states of Sn and Tb were analyzed by XPS (Figure 3). The C 1s line to 284.6 eV was used as reference to correct binding energy. The Sn 3d3/2 and Sn 3d5/2 peaks of the pure SnO2 spheres (upper curve in Figure 3a) were located at 494.8 and 486.3 eV, respectively.30 The Sn 3d3/2 and Sn 3d5/2 binding energies for 1Tb-SnO2, 5Tb-SnO2, and 15TbSnO2 were 0.2, 0.4, and 0.6 eV, respectively, lower than those for the pure SnO2 yolk-shell spheres. The shifting of the Sn 3d peaks due to Tb addition is attributed to the enhanced electron screening effect because the electronegativity of Tb (χ = 1.1) is lower than Sn (χ = 1.96). This again rationalizes the incorporation of Tb into the SnO2 lattice. The XPS spectra of 1Tb-SnO2, 5Tb-SnO2, and 15Tb-SnO2 showed Tb 4d peaks (Figure 3b). Very few reports are available on the XPS data for Tb (both Tb 3d and Tb 4d). In particular, the analysis of the Tb 3d core level using a conventional spectrometer is difficult because of the insufficient kinetic energies of the photoemitted electrons.31 However, a few studies have reported that in the case of Tb3+, the Tb


8 / 34

Page 9 of 34 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


4d peak is observed at binding energies lower than 150 eV. On the other hand, in the case of Tb4+ this peak is observed at binding energies higher than 150 eV.31,32 Note that the Tb 4d spectra of the 5Tb-SnO2 and 15Tb-SnO2 samples were very broad with substantial shoulders below 150 eV and a tail toward 160 eV (Figure 3b), indicating the co-existence of Tb3+ and Tb4+ in these samples. The sensing transients of the four sensors for 20 ppm acetone at 450 °C in dry and wet (RH 20 − 80%) atmospheres are shown in Figures 4a–d. All the sensors exhibited the typical chemiresistive variation of n-type metal oxide semiconductors. The resistance decreased and recovered when exposed to acetone and air, respectively. The humidity dependence of the chemiresistive sensing behaviors of these sensors will be discussed later. Note that the Ra values in dry air atmosphere increased significantly with Tb doping. The substitution of Tb3+ at the Sn4+ sites in SnO2 can be compensated by the formation of oxygen vacancy or holes. ଶௌ௡ைమ

ᇱ

௑

Tb2O3ሱۛۛሮ 2ܾܶௌ௡ + 3ܱை + Vை•• ଵ

ଶௌ௡ைమ

ᇱ

(1) ௑

Tb2O3 + ܱଶ (݃) ሱۛۛሮ 2ܾܶௌ௡ + 4ܱை + 2ℎ• ଶ

(2)

Because the Ra values increased with Tb doping, the electronic compensation based on reaction (2) was more feasible than the ionic compensation based on reaction (1). And the resistance increase by Tb doping is in line with the results in the literature.26 Figure S6 shows the sensing transients of 5Tb-SnO2 sensor for 20 ppm acetone at 350 – 450 °C in dry and wet (RH 20 − 80%) atmospheres. The sensor resistance and humidity dependence of the chemiresistive gas-sensing behaviors of 5Tb-SnO2 sensor decreased as the operating temperature increased.


9 / 34


Page 10 of 34

The gas responses (S = Ra/Rg) of the sensors to 20 ppm acetone, ammonia, carbon monoxide, toluene, p-xylene, and benzene at 350–450 °C in dry atmosphere were shown in Figures 4e–h. Note that the gas-sensing characteristics below 350 °C were not investigated because pure SnO2, 1Tb- and 5Tb-SnO2 sensors showed significant humidity dependence of the Ra and Ra/Rg values, while the resistance of the 15Tb-SnO2 sensor became too high to measure. In all the sensors, the response to acetone was higher than the response to the other interference gases. The Tb-added sensors exhibited relatively low gas responses compared to those of the pure SnO2 sensors. The pure SnO2 spheres showed the largest specific surface area (27.70 m2/g) followed by 1Tb-SnO2 (24.61 m2/g), 5Tb-SnO2 (22.90 m2/g), and 15Tb-SnO2 (21.84 m2/g) (Figure S7). Note that the modal size of the mesopores gradually decreased from 8.0 to 6.4 nm with the addition of Tb. This is consistent with the change of crystallite size from 22.7 nm to 14.6 nm with increasing the Tb doping concentration, indicating that the coarse particles form the larger mesopores. And the largest specific area of SnO2 (27.70 m2/g) in spite of the largest crystallite size (22.7 nm) can be attributed to the formation of relatively large (8.0 nm) and abundant mesopores. A reduction in the mesopore size with Tb doping can decrease the gas accessibility because the Knudsen diffusion is predominant in the mesopores having a size of 2–50 nm, and the Knudsen diffusion coefficient of a gas is known to be proportional to the radius of pore (r).33 Hence, the lower gas responses of the Tb-SnO2 specimens compared to those of pure SnO2, can be partly attributed to their lower specific surface areas and smaller mesopores. Although the pure SnO2 sensor showed the highest gas responses (Figure 4e), its resistance and gas response were critically dependent on the atmospheric humidity (Figure 4a). This suggests that not only the gas-sensing characteristics of the sensing materials but also the humidity dependence of these characteristics should be taken into account while designing the


10 / 34

Page 11 of 34 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


sensing materials for exhaled breath analysis. To quantify the effect of humidity on the chemiresistive sensing behaviors of the samples, we calculated the ratios of the sensor resistances in air (Ra-wet/Ra-dry) and the gas responses (Swet/Sdry) under wet (R.H. 80%) and dry atmospheres (Figures 5, 6). Sensors with Swet/Sdry = 1 and Ra-wet/Ra-dry = 1 show humidityindependent gas response and sensor resistance in air, respectively. The Swet/Sdry and Ra-wet/Ra-dry values lower than 1 indicate a large decrease in the gas response and sensor resistance by moisture, respectively. The decrease in Ra and S is known to relate with the following waterpoisoning reaction occurring at the sensor surface.34-36 ି (or ଶି ) ‫ܪ‬ଶ ܱ ௚௔௦ + 2ܵ݊௟௔௧ + ܱ௔ௗ ܱ௔ௗ → 2(ܵ݊௟௔௧ − ܱ‫ )ܪ‬+ ݁ ି (‫ ݎ݋‬2݁ ି ) (3)

The moisture-induced electron generation decreases the sensor resistance. On the other hand, a decrease in the number of adsorbed oxygen species (O-ad or O2-ad) deteriorates the gas response. Note that for all the sensors, the Swet/Sdry and Ra-wet/Ra-dry values tended to move closer to unity with increasing the sensor temperature from 350 to 450 °C, indicating that the humidity dependence of the chemiresistive sensing behaviors of the sensors decreased. This is probable because the formation of OH radicals on the sensor surface by the water-poisoning reaction (3) can be suppressed by facilitating the dehydration reaction at elevated temperatures. Nevertheless, the pure SnO2 yolk-shell spheres exhibited relatively low Ra-wet/Ra-dry (0.64) and Swet/Sdry values (0.48) even at 450 °C (Figures 5e, 6e). This can hamper the implementation of these sensors for real-time breath analysis. In the case of the 1Tb-SnO2 sensor, although the Swet/Sdry value at 450 °C was 1.07 (Figure 6f), the Ra-wet/Ra-dry value remained low (0.66) (Figure 5f), thus reducing the overall gas response and acetone selectivity of the sensor (Figure 4f). Hence, the 1Tb-SnO2 sensor also could not be implemented for real-time applications. The 5Tb-SnO2 sensor showed a


11 / 34


Page 12 of 34

relatively high response (15.9–26.0) to 20 ppm acetone with negligible cross-responses (2.5–4.4) to the other gases (Figure 4g). The sensor exhibited the selective detection of acetone over the wide sensing temperature range of 350–450 °C. Moreover, the Ra-wet/Ra-dry and Swet/Sdry values for this sensor at 450 °C were as high as 0.76 (Figure 5g) and 0.80 (Figure 6g), respectively. Hence, the operation of the 5Tb-SnO2 sensor at 450 °C (marked by arrow) can be an optimized condition for the sensitive and selective detection of acetone with negligible water-poisoning effect. The 15Tb-SnO2 sensor showed the lowest water-poisoning effect over the temperature range of 350–450 °C (Figures 5h, 6h). This sensor showed relatively good response and selectivity to acetone at 350 °C (Figure 4h). Thus, the 15Tb-SnO2 sensor showed the potential to detect the acetone exhaled from breath (at 350 °C) under real-time conditions. However, the sensor resistances at 350 °C in air (Figure 5d) was extremely high (7 GΩ). This hampered the measurement of resistance using the conventional electric circuit. Moreover, the Swet/Sdry value (0.62) of the 15Tb-SnO2 sensor at 350 °C (Figure 6h) was substantially lower than that of the 5Tb-SnO2 sensor at 450 °C (0.80) (Figure 6g) although the Ra-wet/Ra-dry value of the 15Tb-SnO2 sensor at 350 °C (0.75) (Figure 5h) was similar to that of the 5Tb-SnO2 sensor at 450 °C (0.76) (Figures 5g). In addition, the responding speed of the sensors should also be also taken into account. Hence, the 90% response time (τres) of the sensors i.e., the time taken to reach 90% variation in the sensor resistance when exposed to the analyte gases, were measured from their sensing transients. The τres value of the 5Tb-SnO2 sensor at 450 °C (τres = 9 s in RH 80%) was significantly lower than that of the 15Tb-SnO2 sensor at 350 °C (τres = 35 s in RH 80%). Thus, taking all the sensing characteristics into consideration, it can be stated that the operation of the


12 / 34

Page 13 of 34 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


5Tb-SnO2 sensor at 450 °C was the optimum condition to achieve highly selective, sensitive, and rapid detection of acetone with negligible humidity dependence of the sensor resistance and gas response. Table 1 summarizes the humidity dependence of the chemiresistive sensing characteristics of the gas sensors in the literature and in this study.21,22,37-39 Although the moisture-induced variation of sensor resistance and gas response in the present 5Tb-SnO2 sensor are similar to those in Ni-SnO237 and CuO-SnO2 sensors22, no acetone selectivity with humidityindependent gas-sensing characteristics was reported. Moreover, the 5Tb-SnO2 sensor showed superior stability against moisture compared to the TiO2-SnO238 and Pd/Sb-SnO2 sensors.39 In our previous work,23 we suggested that the loading of CeO2 nanoclusters can remove the water poisoning effect in In2O3 gas sensors by maintaining the concentration of the adsorbed oxygen species regardless of the dynamically changing humidity. The key idea was to refresh the surface of the sensor by the interaction between the Ce3+/Ce4+ redox pair and the OH groups (or water vapor) on the surface of In2O3. Terbium oxide is another representative multivalent oxide, as confirmed by the XPS results obtained in this study. Note that co-existence of Ce3+ and Ce4+ has been widely employed for oxygen buffering in three-way catalysts40-42 and the addition of multivalent terbium oxide (Tb3+ and Tb4+) to cerium oxide is known to enhance the oxygen buffering capability of these catalysts significantly.24,25 Hence, it can be stated that the regenerative surface-refreshing of the SnO2 sensors by the interaction between the Tb3+/Tb4+ redox pair and the surface OH groups (or water vapor) was also responsible for their humidityindependent gas-sensing characteristics. However, further investigation is necessary to elucidate the underlying sensing mechanism. The gas selectivity of the pure SnO2 and 5Tb-SnO2 sensors under dry and wet (RH 80%) atmospheres were measured at 450 °C (Figure 7). In dry atmosphere, the response of the pure


13 / 34


Page 14 of 34

SnO2 sensor to 20 ppm acetone was as high as 63.8, which is markedly higher than its response to the other interference gases (carbon monoxide, ammonia, toluene, p-xylene, and benzene). However, in humid atmosphere (RH 80%), the gas responses to all the analyte gases deteriorated significantly. This is attributed to the water-poisoning effect (Figure 7a). In contrast, the 5TbSnO2 sensor exhibited similar gas responses and selectivity toward 20 ppm acetone under dry and humid atmospheres (Figure 7b and S8). These results clearly indicate that 5Tb-SnO2 sensor can be implemented for the analysis of exhaled breath even in highly humid environments (RH > 80%). Note that the doping of Tb enhanced neither the acetone response nor acetone selectivity of SnO2 sensor. Thus, the key role of Tb dopants in the present study is not the catalytic activation of acetone sensing but the keeping of the same sensor resistance, gas response, and acetone selectivity regardless of humidity change. Figure 8 shows the gas responses of pure SnO2 and 5Tb-SnO2 to 0.05–5 ppm acetone at 450 °C under both dry and humid conditions (RH 80%). The difference between the gas responses obtained under the dry and humid conditions was smaller in 5Tb-SnO2 sensor than in pure SnO2 sensor. This difference should be low in order to achieve reliable and humidityindependent sensing. The detection limit of the 5Tb-SnO2 sensor in humid atmospheres (RH 80%) was determined to be 50 ppb with Sacetone> 1.2, which was used as the criterion for sensing. Under humid atmosphere, the τres values of the 5Tb-SnO2 sensor upon exposure to 0.05 – 5 ppm ranged from 138 to 15 s, while 90% recovery time upon exposure to air atmosphere ranged from 298 to 526 s (Figure S9). This confirms that the 5Tb-SnO2 sensor could reliably detect the subppm- and ppm-levels of acetone under wide range of ambient humidity with relatively fast responding/recovering speed. We have measured the 30 repetitive sensing transients (Figure 8c)


14 / 34

Page 15 of 34 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


and long-term stability (Figure 8d) of 5Tb-SnO2 sensor for 14 days. This results showed the 5TbSnO2 sensor is very stable and reproducible. Acetone is a volatile and colorless gas which can cause headache, confusion, nausea, and even unconsciousness when inhaled in high concentrations.43 The 5Tb-SnO2 sensor developed in this study could detect even the low concentrations of acetone precisely regardless of the humidity variations. For the diagnosis of diabetes, the acetone concentration (> 1.8 ppm) exhaled by diabetic patients should be discriminated from that (

Humidity-independent Oxide Semiconductor Chemiresistors Using

Recommend Documents