Ethanol-sensing properties of cobalt porphyrin-functionalized titanium dioxide nanoparticles as chemiresistive materials that are integrated into a low power microheater

Gaseous ethanol detection has attracted significant interest owing to its practical applications such as in breath analysis, chemical process monitoring, and safety evaluations of food packaging. In this study, titanium dioxide (TiO2) nanoparticles functionalized with cobalt porphyrin (CoPP) are utilized as resistive ethanol-sensing materials, and are integrated with a suspended micro-heater for low power consumption. The micro-heater with the suspended structure inhibits substrate heat transfer, resulting in power consumption as low as 18 mW when the operating temperature is approximately 300 °C. CoPP functionalization allows an enhanced response (197.8%) to 10 ppm ethanol compared to that of pristine TiO2 nanoparticles. It is confirmed that the sensor response is reliable upon exposure to 10 ppm ethanol for three cycles. In addition, responses of different magnitude are obtained under exposure to ethanol at various concentrations from 9 to 1 ppm, indicating that the resistance change originates from a charge transfer between the sensing materials and target gas. The sensing mechanism of CoPP-functionalized TiO2 in relation to charge transfer is analyzed, and the performance of the proposed sensor with previously reported TiO2-based ethanol sensors is compared. Considering that it is processed by batch fabrication, consumes low power, and offers high sensitivity, the proposed sensor is promising for use as a portable sensor in the distributed monitoring of gaseous ethanol.


Introduction
With the emergence of the Internet of Things and advancements in machine learning, there is significant interest in the collection and processing of physical/ chemical information using various sensors at multiple locations. The sensors must be portable so that vast amounts of data can be obtained from various locations.
In addition, portable sensors must possess several features, including a small size for module integration, fast response/recovery speed to physical/chemical stimuli, productivity and cost-efficiency, low power consumption, and durability for long-term operation.
Among the goals of using distributed sensor networks is monitoring harmful molecules in the gaseous state; to this end, many researchers have focused on gas sensors for decades. Among the many sensors used to detect harmful gas species, those for the detection of ethanol have received significant attention owing to their practical applications, such as in the breath analysis of Open Access *Correspondence: kimjb@yonsei.ac.kr † Kwanhun Kim and Yunsung Kang contributed equally to this work School of Mechanical Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul 03722, Republic of Korea intoxicated drivers [1], chemical process monitoring [2], and safety evaluations of food packaging [3]. Various mechanisms have been proposed to detect gaseous ethanol, such as those that are based on chemiresistive [4], electrochemical [5], and colorimetric sensing [6]. Among them, chemiresistive sensors, which use metal oxide-based materials as ethanol-sensing elements, typically offer advantages such as high sensitivity, costeffectiveness, and a simple electronic setup for measurement [7,8]. In general, metal oxide-based sensors require high-temperature operation above 200 °C to efficiently induce a redox reaction of the target gas on the material's surface. The use of titanium dioxide (TiO 2 ) between various metal oxides has received remarkable attention for gas detection owing to its stability and reliability at high temperatures [9,10]. Several ethanol sensors that utilize the properties of TiO 2 have also been reported [2,[10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25]. However, except for a few studies, it has largely been challenging to achieve the highly sensitive detection (in several ppm) of ethanol [2,[23][24][25]. Moreover, most aforementioned studies merely presented the ethanolsensing characteristics of TiO 2 -based materials without considering their productivity and low-power consumption [26,27], which are required for portable sensors.
In our previous work, we presented a batch-producible benzene, toluene, and xylene sensor using TiO 2 nanoparticles functionalized with cobalt porphyrin (CoPP) as sensing materials [28].
This study presents an ethanol sensor using TiO 2 nanoparticles functionalized with CoPP as sensing materials. We also integrated the sensing materials with a suspended micro-heater, enabling the desired heat generation with low power consumption. CoPP functionalization on TiO 2 nanoparticles allowed an increased sensitivity (197.8%) to 10 ppm ethanol compared to pristine nanoparticles. We also investigated the ethanol-sensing characteristics of TiO 2 nanoparticles functionalized with CoPP based on the operating temperatures. The fabricated sensor showed a reliable response under repeated exposure to 10 ppm ethanol. The sensor responses to ethanol at various concentrations are presented in the paper, along with the suggested sensing mechanisms. Figure 1 depicts the fabrication process of the sensor platform with the suspended micro-heater. First, 500 nm-thick silicon nitride (Si 3 N 4 ) was deposited on a 300 nm-thick silicon dioxide (SiO 2 ) substrate through low-pressure chemical vapor deposition (CVD) (Fig. 1a) to induce low residual stress of a membrane. Subsequently, a 25/250 nm-thick Ti/Pt film was deposited on the Si 3 N 4 /SiO 2 substrate and patterned via photolithography and reactive ion etching (RIE) to obtain the metal heating element and sensor electrodes in the desired shape (Fig. 1b). A 200 nm-thick SiO 2 film was then deposited via plasma-enhanced CVD and patterned by photolithography and RIE, resulting in the formation of a passivation layer for the heating element (Fig. 1c). Photolithography and RIE were performed to etch the SiO 2 / Si 3 N 4 /SiO 2 layers for opening the bare Si layer (Fig. 1d). The opened Si substrate was then wet-etched using tetramethylammonium hydroxide (TMAH) to form the suspended structure (Fig. 1e). We could not observe a fracture of the membrane of the sensor platform when etched in TMAH solution. This would be due to the tri-layered structure (SiO 2 /Si 3 N 4 /SiO 2 ) of membrane, contributing the reduction of residual stress in the membrane with partial compensation of the compressive and tensile stress of the oxide and the nitride, respectively. An optical image of the fabricated sensor platform, including the micro-heater, is presented in Fig. 1 f. The image confirms the formation of the SiO 2 passivation layer except for the part of sensor electrodes. This prevents the undesired effects to the micro-heater arising from the preparation of the sensing materials.

Methods and materials
To prepare the sensing materials, we drop-coated a 3.3-3.7 wt% TiO 2 nanoparticle dispersion (titanium(IV) oxide, Sigma-Aldrich) in DI water using a micropipette. It was uniformly formed at the center circle of the sensor platform. E-beam evaporation of CoPP (5,10,15,20-tetraphenyl-21 H,23 H-porphyrin cobalt(II)) was then performed for functionalizing TiO 2 . Consequently, the originally insulated sensor electrodes were connected with the electrical channel of the CoPP-functionalized TiO 2 nanoparticles. The material characterization of the TiO 2 nanoparticles, CoPP, and CoPP-functionalized TiO 2 is described in detail elsewhere [28].
For evaluating the ethanol detection capability of the sensor, we placed it in a quartz tube at room temperature. Mass flow controllers were connected to one end of the quartz tube to control the flow rates of dry air and 10 ppm ethanol gas. The total flow rate was fixed at 500 sccm, and ethanol gas at various concentrations was generated and injected into the quartz tube by adjusting the flow rate of each gas. The current change in the CoPP-functionalized TiO 2 nanoparticles on exposure to various concentrations of ethanol, with a fixed input voltage of 3.5 V to the sensor electrodes, was recorded with a computer-controlled source meter (2400, Keithley). To generate heat for efficiently inducing the redox reaction between ethanol and sensing materials, a voltage (3 V) to the micro-heater was applied with direct current (DC) power supply (E3647A, Agilent). Figure 2a shows the sensor responses of the pristine TiO 2 nanoparticles and CoPP-functionalized nanoparticles to 10 ppm ethanol. Here, the sensor response is defined as R air /R gas (R gas and R air denote the resistance measured in ethanol and air environment, respectively), because TiO 2 is an n-type semiconducting material, and ethanol is well-known as a reducing gas that provides electrons to sensing materials. Before CoPP functionalization, a sensor using pristine TiO 2 nanoparticles as the sensing material exhibited a response of 6.41 to 10 ppm ethanol. However, the sensor with CoPP functionalization showed a response of 12.68 upon exposure to 10 ppm ethanol, which is approximately twice as high as that of using pristine nanoparticles. These results indicate that CoPP functionalization on TiO 2 nanoparticles is highly advantageous for achieving enhanced sensitivity to ethanol. This result corresponds well with previous studies, where CoPP was suggested to be an effective functionalization material for enhancing the sensitivity to volatile organic compounds, including diverse nanomaterials such as graphene [29], tin dioxide [30], and zinc oxide [31]. Ethanol is also known as a volatile organic compound, and the detailed sensitivity-enhancing mechanism of CoPP-functionalized TiO 2 to ethanol will be discussed later. Figure 2b shows the relationship between the input voltage of the micro-heater and the corresponding temperature. The operating temperature was indirectly estimated using a resistance temperature detector (RTD) with the micro-heater. The RTD was fabricated together with the sensor platform using the same process. A resistor of the RTD was designed and fabricated at the location where the sensor electrodes were originally positioned. We measured the resistance of the resistor in a furnace at various temperatures and compared it for various input voltages of the micro-heater. This made it possible to estimate the relationship between the input voltage of the micro-heater and the corresponding temperature. Based on the temperature estimation, we investigated the response of the sensor to 10 ppm ethanol at various operating temperatures. The responses of the sensor were 3.12, 9.56, 10.85, 12.68, and 9.94 on exposure to 10 ppm ethanol when the operating temperatures of the micro-heater were 172.64, 217.08, 263.32, 308.6, and 341.72 °C, respectively (Fig. 2c). The optimal temperature of the sensor using CoPP-functionalized TiO 2 as the sensing material was achieved with a micro-heater input voltage of 3.0 V. Notably, at 3.0 V, we achieved a power consumption as low as 18 mW. However, at 3.4 V, at which point the operating temperature increased, the sensor response slightly decreased. This is because the rate of desorption of the reactants exceeded that of absorption of the gas at higher temperatures [32]. Therefore, subsequent characterization using the sensor was performed with a heater input voltage of 3.0 V (i.e., at an operating temperature of 308.6 °C). Figure 3a depicts the sensor response under repeated exposure to 10 ppm ethanol; there is no remarkable change in the sensor responses. The sensor response to ethanol at various concentrations from 9 down to 1 ppm is presented in Fig. 3b; the sensor response increases as the ethanol concentration increases. These results indicate that the change in the resistance of the sensing materials originated from charge transfer from ethanol gas. The detailed ethanol-sensing mechanism of the CoPP-functionalized TiO 2 nanoparticles was analyzed and is depicted in Fig. 4. Three mechanisms simultaneously occur in ethanol detection using the CoPPfunctionalized TiO 2 nanoparticles. The first is a direct interaction between ethanol and CoPP. CoPP itself is known as a gas-sensing material, particularly for volatile organic compounds including ethanol; it also acts as a semiconducting material [33], and reacts with ethanol due to adsorption via π-π interactions and hydrogen bonding [34]. This induces electron transfer from ethanol to CoPP. Subsequently, the transferred electrons migrate from CoPP to TiO 2 due to the difference in the Fermi energy between them [28]. The second is the catalytic reaction between ethanol and oxygen ions near the CoPP and TiO 2 nanoparticles. CoPP is known to act as a catalyst for oxidizing volatile compounds [35]. Hence, the CoPP catalyst enables the oxidization of ethanol, which efficiently induces charge transfer from ethanol to TiO 2 . The third mechanism is a redox reaction between ethanol and oxygen ions on TiO 2 nanoparticles without the influence of CoPP, like in other metal oxide-based ethanol sensors. All these mechanisms contribute to a change in the depletion region of the TiO 2 nanoparticles. Thus, an electrical channel composed of TiO 2 nanoparticles is more conductive when exposed to ethanol, enabling the highly sensitive detection of ethanol. Table 1 compares the sensing performances of the TiO 2 -based ethanol sensors. Most previously reported sensors operated at high temperatures with an external Fig. 2 a Sensor response of pristine and CoPP-functionalized TiO 2 nanoparticles to 10 ppm ethanol. b Relationship between the input voltage of the micro-heater and corresponding temperature. c Sensor response to 10 ppm ethanol at various operating temperatures heating system such as a furnace or a separated hotplate. The operation of such sensors involved extremely high power consumption. Meanwhile, some studies reported TiO 2 -based ethanol sensors that could operate at room temperature [2,17,19,21]. However, the TiO 2 -based sensing materials exhibited high resistance (on the order of gigaohms) at room temperature, making it difficult to integrate the sensor with simple and inexpensive circuits. By comprehensively considering the sensitivity, power consumption, and limit of detection, our ethanol sensor offers advantages over previously reported sensors.

Results and discussion
One aspect of the sensor to be considered is long-term stability. We presented the long-term stability of a sensor with the same sensing materials to toluene [28], and the sensitivity reduced to 60% level after about 5 days. Thus, the long-term stability of the presented sensor to ethanol at a similar level to toluene is expected. Meanwhile, anatase is known as the metastable phase and can easily transform to the most stable rutile phase after heating TiO 2 at temperatures of 450-850 ºC [36]. This range is beyond the operating temperatures of our sensor, hence, the decrease in sensitivity after several days might originate from the thermal stability of CoPP. Nevertheless, when considering the sensor can be batch-fabricated at a low cost, the sensors can be easily replaced with another after a certain amount of time. Furthermore, the same sensing materials exhibited a stable long-time operation to toluene for 14 h [28], which means that it can be used with reliability over a certain period of time. Another approach to enhance the thermal stability of CoPP is adding a functional group to increase ionization potential [37]. This can be utilized for enhancing the thermal stability of CoPP.

Conclusions
We demonstrated an ethanol sensor using TiO 2 nanoparticles functionalized with CoPP as sensing materials together with a micro-heater for low power consumption. The sensor platform, including the sensor electrodes and micro-heater on the suspended structure, was batchfabricated by bulk micromachining. The suspended structure of the micro-heater allowed the power consumption of the sensor to be as low as 18 mW, providing the desired temperatures due to the limited heat transfer through the substrate. We confirmed that CoPP functionalization was effective in achieving high sensitivity to 10 ppm ethanol, allowing a response that was approximately twice higher than that of pristine nanoparticles. Various operating temperatures were achieved with different input voltages of the micro-heater, and the sensor response to 10 ppm ethanol was evaluated at different temperatures. The results indicated that the optimal operating temperature for the CoPP-functionalized TiO 2 nanoparticles to detect ethanol was 308.6 °C. The sensor showed a reliable response under repeated exposure  to 10 ppm ethanol. The sensor also exhibited an increase in response as the ethanol concentration increased, indicating that the change in resistance originated from the charge transfer between the sensing material and target gas. The mechanism of charge transfer between ethanol and CoPP-functionalized TiO 2 was analyzed. Considering its high sensitivity, low power consumption, costeffectiveness due to batch fabrication, and small size, our sensor would be promising for portable applications requiring ethanol detection.

Acknowledgements
Not applicable.

Author contributions
KK, YK, KB and JK developed the idea. KK and YK carried out fabrication, measurement, and analysis of the results, and wrote the manuscript. KB supported fabrication process and measurement. JK supervised the research and reviewed the manuscript. All authors read and approved the final manuscript.

Availability of data and materials
All data generated or analyzed during this study are included in this published article.

Competing interests
The authors declare that they have no competing interests.