Solid-State Gas Sensors with Ni-Based Sensing Materials for Highly Selective Detecting NOx

Abstract

Precise monitoring of NO_x concentrations in nitric oxide delivery systems is crucial to ensure the safety and well-being of patients undergoing inhaled nitric oxide (iNO) therapy for pulmonary arterial hypertension. Currently, NO_x sensing in commercialized iNO instruments predominantly relies on chemiluminescence sensors, which not only drives up costs but also limits their portability. Herein, we developed solid-state gas sensors utilizing Ni-based sensing materials for effectively tracking the levels of NO and NO₂ in the NO delivery system. These sensors comprised of NiO-SE or (NiFe₂O₄ + 30 wt.% Fe₂O₃)-SE vs. Mn-based RE demonstrated high selectivity toward 100 ppm NO under the interference of 10 ppm NO₂ or 3 ppm NO₂ under the interference of 100 ppm NO, respectively. Meanwhile, excellent stability, repeatability, and humidity resistance were also verified for the proposed sensors. Sensing mechanisms were thoroughly investigated through assessments of adsorption capabilities and electrochemical reactivity. It turns out that the superior electrochemical reactivity of NiO toward NO, alongside the NO₂ favorable adsorption characteristics of (NiFe₂O₄ + 30 wt.% Fe₂O₃), is the primary reason for the high selectivity to NO_x. These findings indicate a bright future for the application of these NO_x sensors in innovative iNO treatment technologies.

1. Introduction

Pulmonary arterial hypertension (PAH) is a complex and fatal emergency vascular disorder that affects more than 100 million people worldwide, and is mainly found in neonates [1,2]. Compared with oral medicine or injection, inhaled nitric oxide (iNO), a pulmonary vasodilator approved by the Food and Drug Administration (FDA), offers several advantages in the PAH treatment such as noninvasive, targeted, fast onset of action, and less severe systemic adverse effects [3,4,5]. This method, depending on the severity of disease, usually involves the continuous administration of NO gas in the range of 100 ppm by nasal plug or mask [6,7]. However, during treatment, the concentration of NO delivered needs to be tightly controlled in the first place; otherwise, it may cause a rebound in pulmonary arterial pressure and endanger the patient’s life [8]. Additionally, a certain amount of low concentration NO₂ may be generated during the transport of NO, which also should be precisely detected timely because such a toxic gas can directly endanger patients’ health when the concentration is above 3 ppm [9,10].

It is worth noting that commercialized iNO systems generally use chemiluminescence devices to measure NO and NO₂, which limits portability and leads to very high cost [3,11]. In recent years, with the rapid development of gas sensors, NOx gas sensors based on solid-state YSZ (yttria-stabilized zirconia)-based electrolytes have received extensive attention from researchers due to their high selectivity, desirable sensitivity, and satisfactory stability [12,13,14,15,16,17,18]. Moreover, YSZ-based sensors can be fabricated in planar configuration, greatly simplifying the structure of the sensor and facilitating the development of integration for commercially available products [13,19,20]. However, currently reported mixed-potential NOx gas sensors mainly focus on the detection of NO and NO₂ in the middle-to-high concentration range of 100–500 ppm in automotive and industrial contexts [21,22,23,24,25], while the iNO treatment instrument normally requires a relatively low NOx detection limit. Furthermore, it is a great challenge to selectively detect 3 ppm NO₂ with 100 ppm NO interference, tens times the difference in concentration. Herein, we proposed high-performance YSZ-based NOx sensors by screening out optimal sensing materials. The sensing characteristics of the developed sensors were systematically studied, and we compared the sensors’ performance with some typical NOx gas sensors in

Table 1. Our work in contrast with the literature.

Sensor Structure SE |YSZ| RE	Temperature (°C)	Detection Range (ppm)	ΔV/Concentration	Ref.
WO3 |YSZ| Pt	300	30–500	22 mv/30 ppm NO2 (in 5% O2)	[22]
NiFe2O4 |YSZ| Pt	400	100–500	81.3 mv/100 ppm NO2 (in 10% O2)	[23]
CoTiO3 |YSZ| Pt	650	20–200	130 mv/100 ppm NO2 (in 5% O2)	[24]
La0.6Sr0.4CoO3 |YSZ| Pt	400	100–500	125 mv/400 ppm NO2 (in 10% O2)	[25]
ZnO |YSZ| Pt	700	40–400	−12 mv/100 ppm NO (in 21% O2)	[26]
La2CuO4 |YSZ| Pt	450	50–650	20 mv/650 ppm NO (in 3% O2)	[27]
Pt |YSZ pellet| Pt	500	100–800	13 mv/800 ppm NO (in 3% O2)	[28]
LCMF |YSZ| Pt	550	——	−62 mv/400 ppm NO (in 1.5% O2)	[29]
NiO |YSZ| MnO2 NiFe2O4 + 30%Fe2O3 |YSZ| MnO2	310 390	10–100 1.5–10	−22.7 mv/100 ppm NO 13.8 mv/3 ppm NO2 (in 21% O2)	this work

. Moreover, the adsorption capabilities and electrochemical reactivity of these sensors were characteristics to clarify the working principle.

2. Experimental

2.1. Fabrication of Sensors

Figure 1 depicts the schematic components of the planar sensor. Initially, commercial MnO₂ powder (99% purity, Sigma Aldrich Chemie GmbH, Taufkirchen, Germany) was thoroughly mixed with α-terpineol, and the obtained paste was painted on the YSZ plate (length × width × thickness: 2.75 × 0.5 × 0.1 cm³; NIKKATO CORPORATION, Osaka, Japan); after drying at 130 °C for 4 h, the YSZ plate attached with the MnO₂ layer was calcined at 1400 °C for 2 h in a muffle furnace to form the Mn-based reference electrode (RE). Various commercial metal oxides (Sigma Aldrich Chemie GmbH, Taufkirchen, Germany, with a purity of 99%) were fabricated with a similar procedure and both were calcined at high temperature for 2 h to form the sensing electrodes (SE).

2.2. Evaluation of Sensing Characteristics

The fabricated sensor was set in a quartz tube, and alternatively exposed to the base gas (21 vol.% O₂, N₂ balance) or the sample gas containing NO, NO₂ (total flow rate: 100 mL·min⁻¹) for 5 min. As the sensors are expected to be applied in iNO treatment devices in which the process of NO generation and transportation is tightly controlled, as a result, there are few interference gases excepted NO₂ and the concentration of NO₂ is below 10 ppm in reality. Moreover, the NO at specific concentrations will be directly inhaled by the patients. Taking these factors into account, in such a specific application, the NO tested concentration range was 10–100 ppm in 21 vol% O₂, N₂ balance, and the NO₂ tested concentration range was 1.5–10 ppm in 21 vol% O₂, N₂ balance. An adjustable DC power supply (SS-3305D, A-BF, Dongguan, China) controls the Pt heater’s working voltage to provide a high working temperature for the sensors; notably, the temperatures of the experiments mentioned in the following research just represent the temperature of a tiny area (approximately: 0.4 × 0.4 cm²) on the sensor’s surface rather than the whole tests chamber. And there is no significant heat spread, so the temperature of transported gas flow is virtually unaffected. The electrochemical signals were recorded by a Data Acquisition/Switch Unit (34970A, Agilent, Nanjing, China) via Pt wire connected, and the potential difference ΔV (ΔV = V_{Sample gas} − V_{Base gas}) represented the sensor’s response to sample gas. The surface morphology of calcined powders was observed by Scanning Electron Microscope (Gemini 360, Carl Zeiss, Oberkochen, Germany) with an excitation voltage of 15 kV, and the crystalline phase was characterized by X-ray diffraction (D8 DaVinci, Bruker, Berlin, Germany). Finally, microcantilever chips and an integrated Gas Sensing System (LoC-GSS 1000, High-End MEMS Technology, Xiamen, China) were utilized to test the sensing materials’ gas adsorption.

3. Results and Discussion

3.1. Screening of Nitric Oxide Sensing Materials and Developing a High-Performance NO Sensor

The sensing mechanism of mixed-potential type NO_X can be interpreted by mixed potential theory [13], where the electrochemical reactions for NO (NO₂) and O₂ occur simultaneously at the SE/YSZ interface as follows.

For NO:

$$2NO+{2O}^{2-}\to 2N{O}_2+4e^{-} \left(\mathrm{Anodic}\right)$$

(1)

$$O_2+4e^{-} \to 2O^{2-} \left(\mathrm{Cathodic}\right)$$

(2)

For NO₂:

$$2O^{2-} \to O_2+4e^{-} \left(\mathrm{Anodic}\right)$$

(3)

$$2N{O}_2+4e^{-} \to 2NO+2O^{2-} \left(\mathrm{Cathodic}\right)$$

(4)

It can be seen that NO and NO₂ undergo different reactions respectively and the responses to NO and NO₂ for YSZ-based NO_X sensors are opposite to each other. Generally, the response is negative to NO and positive to NO₂.

Several metal oxides were selected as potential candidates for a NO sensing material, and among the examined SEs, In₂O₃−SE not only gave the highest potential value of 100 ppm NO but also a higher potential value of 10 ppm NO₂, resulting in poor NO selectivity (Figure 2). In contrast, NiO-SE showed both desirable NO sensitivity and selectivity in comparison with other examined materials. Consequently, NiO was predicted to be the optimal material for NO detection due to its higher sensitivity and selectivity.

The SEM image in Figure 3 shows that the NiO powder annealed at 1200 °C was plate-like and relatively compact. According to former research showing that plate−like metal oxides normally exhibit better selectivity to NO [30,31], we anticipated desirable selectivity for NiO toward NO when against NO₂.

To confirm the practicability of using NiO tracking a high level of NO with high selectivity against 10 ppm NO₂, sensing characteristics of the YSZ−based gas sensor comprised of NiO-SE vs. Mn−-based RE were studied. Figure 4a depicts the relationship between the sensor’s response and the operating/calcination temperature for the NiO−SE; it can be confirmed that the optimal response signal was observed for the sensor fabricated at 1200 °C. Moreover, it should be noted the operated temperature of YSZ solid electrolyte must be above 300 °C [32]. And it needs to be clarified that when the operating temperature was below 300 °C, the conductivity of YSZ was greatly reduced, which led to poor recovery characteristics and noticeable noise in the sensor’s response curves. Consequently, the optimal operating/calcination temperature was selected at 310 °C and 1200 °C for the following study. As shown in Figure 4b, the response signal of the sensor varied linearly to the logarithm of the NO concentration in the range of 10–100 ppm. According to the repeated response transients of the sensor that were exposed to 100 ppm NO and 10 ppm NO₂ (Figure 4c), it can be concluded that the sensor gave excellent repeatability in the successive runs and high selectivity to 100 ppm NO (24.7 mV) against 10 ppm NO₂ (less than 1 mV). In light of the long-term stability, which is a crucial factor in practical applications. A 35-day test to 100 ppm NO was further measured under 310 °C, as shown in Figure 4d, variation in the response value during the tested period was roughly estimated to be less than 3 mV, indicating satisfactory stability of the sensor using NiO-SE and Mn−based RE.

To further justify the NiO-SE sensor’s ability to differentiate NO in a mixture, Figure 5a compared the responses of the sensor to 100 ppm NO interfered with different concentrations of NO₂. It could be concluded that the response to 100 ppm NO changed little and the NiO-SE sensor showed excellent anti-interference capability. In addition, the sensor may work under relatively high humidity environments in real practical applications, and Figure 5b illustrates the responses and the change in response values of the NiO-SE sensor at 0%, 24%, 67%, and 91% RH separately to 100 ppm NO at 310 °C. The maximum |ΔV| value was 5% under the range of 0–91% RH, proving that the sensor displays good humidity resistance.

3.2. Screening of Nitric Dioxide Sensing Materials and Developing a High-Performance NO₂ Sensor

It is a great challenge to detect low concentration of NO₂ compared to ten times the concentration of NO, and pure metal oxides as sensing materials are hardly able to realize the research target. Recently, spinel−based oxides have proven to be good candidates for NO_x sensing. For instance, NiFe₂O₄ was reported to be a potential NO₂ sensing material but the inadequate detection limit (LOD) down to 5 ppm limits its application in iNO treatment technologies [23,33,34]. In order to improve the sensing behavior of NO₂, it is believed that adding additives or dopants to NiFe₂O₄ would effectively extend the LOD [35,36]. In the meantime, Fe₂O₃ was announced that not only exhibited the same trivalent cation as the NiFe₂O₄’s AB₂O₄ general formula but also demonstrated acceptable capability to detect NO₂ [37]. Thus, we added Fe₂O₃ into NiFe₂O₄ in an attempt to enhance the NO₂ detection capacity.

Figure 6 demonstrated the XRD patterns for NiFe₂O₄, Fe₂O₃, and (NiFe₂O₄ + x wt.% Fe₂O₃) composites after calcined at 1200 °C for 2 h. No impurity peaks were observed, confirming the high purity of the products. However, there were no characteristic diffraction peaks of Fe₂O₃ for the composite materials; the reason for this could be Fe₂O₃’s lower loading content and weak crystallization.

Furthermore, micro-structures of NiFe₂O₄, Fe₂O₃, and their composite (e.g., NiFe₂O₄ + 30 wt.% Fe₂O₃) were characterized via scanning electron microscope (SEM) and presented in Figure 7. In comparison with NiFe₂O₄ (Figure 7a,b) and Fe₂O₃ (Figure 7c,d), a certain amount of grains (with a diameter of around 0.6 μm), which could be assigned to NiFe₂O₄, was observed in the NiFe₂O₄ + 30 wt.% Fe₂O₃ composite. Moreover, it is reasonable to deduce that the formed composite had a more porous and three-dimensional (3D) structure when compared with that of NiFe₂O₄. The porous 3D micro-structure would offer more active reaction sites for NO₂.

Accordingly, sensing behavior to detect 3 ppm NO₂ and 100 ppm NO for the YSZ−based gas sensor utilizing NiFe₂O₄ and its composite-SEs (vs. Mn-based RE) was examined. As shown in Figure 8a, when the amount of Fe₂O₃ reaches 30%, the sensor gave superior selectivity and sensitivity to 3 ppm NO₂ against 100 ppm NO. In other words, the YSZ−based gas sensor using (NiFe₂O₄ + 30 wt.% Fe₂O₃)-SE is a good candidate for tracking the trace amount of NO₂. Figure 8b suggests the optimal operating/calcination temperature was 390 °C/1200 °C.

The dependence of the sensing response on the NO₂ concentration in the range of 1.5–10 ppm was examined and a linear relationship between response value and NO₂ concentration on a logarithmic scale (Figure 9a). Notably, it can be concluded that after aging for several days (2–3 days), the response signal of the sensor to 100 ppm NO significantly decreased, whereas it maintained a high response value of 3 ppm NO₂ (Figure 9b). This phenomenon might be caused by the sensing materials’ gas phase catalytic reaction to NO, whereby a certain amount of NO was converted to NO₂ before arriving at the TPB, and both NO and NO₂ triggered the electrochemical reaction in reality. Based on the aforementioned Reaction (1)–(4), the measured potential signal was the final response after the response of NO offset the response of NO₂. However, in early tests, the mixed materials’ properties were not yet stable, and the conversion efficiency of NO was dynamic. After several days of aging, the NO conversion rate stabilized and the sensor’s response to NO was maintained. Figure 9c,d further indicates the sensor’s excellent repeatability and stability as well as ultra-high selectivity to 3 ppm NO₂.

The selectivity to NO₂ of (NiFe₂O₄ + 30 wt.% Fe₂O₃)-SE could be reinforced in Figure 10a. The responses of the sensor only reduced less than 4 mv even when 3 ppm NO₂ was mixed with 90 ppm NO and it could be concluded that the (NiFe₂O₄ + 30 wt.% Fe₂O₃)-SE sensor showed excellent selectivity to NO₂ at 390 °C. The sensing performance of the (NiFe₂O₄ + 30 wt.% Fe₂O₃)-SE sensor on the variation in the water vapor content (0–91% RH) was also studied, and Figure 10b shows that the sensor’s response values to 3 ppm NO₂ were hardly affected by the change in humidity, which could be ascribed to the desorption of water at such a high operating temperature.

3.3. Study on the Sensing Mechanism

According to the working principle of an electrochemical gas sensor, the target gas initially selectively adsorbs on the surface of the sensing material and then participates in the following electrochemical reaction. In this case, both adsorption capabilities and electrochemical reactivity contribute to the final response selectivity and response value. To understand which step primarily determines the widely concerned selectivity during NO_x sensing, we initially used microgravimetric technology [38,39] to compare the NO_x adsorption capabilities of each sensing material. In this study, changes in the microcantilever’s vibration frequency directly reflected the difference in the adsorption capabilities of NO and NO₂.

Figure 11 shows the frequency shift to 100 ppm NO and 10 ppm NO₂ of NiO, respectively. In repeated tests, the frequency shift to NO was greater than that of NO₂, implying that NiO preferred to adsorb NO molecules when exposed to the NO_x mixture. In a similar manner, the gas adsorption capabilities of NiFe₂O₄(+30 wt.% Fe₂O₃) were examined and are presented in Figure 12. It is reasonable to conclude that (NiFe₂O₄ + 30 wt.% Fe₂O₃) exhibits similar adsorption capability to both 3 ppm NO₂ and 100 ppm NO, since only a minor difference in the frequency shift to NO and NO₂ was observed.

To further understand the mechanism for the sensors’ different selectivity, polarization curves were measured in the sample gas (3 ppm NO₂, 10 ppm NO₂, 10 ppm NO, and 100 ppm NO, in 21% O₂, with N₂ balanced) and air, which could reflect the reaction rate for the electrochemical reaction that occurred in TPB (triple-phase boundary). Figure 13a shows that when the bias voltage was 0 mV, the current value difference between various sample gases at different concentrations and air was small, and it can be inferred that NiO-SE had poor electrochemical catalytic activity to NO and NO₂. On the contrary, an obvious current difference between NO₂ and NO is given in Figure 13b. Furthermore, a higher current value of NO₂ suggests that the (NiFe₂O₄ + 30 wt.% Fe₂O₃)-SE had better electrochemical catalytic activity of NO₂, compared with NO.

Through a comparative analysis of gas adsorption capabilities and electrochemical reactivity, we can conclude that the response selectivity of the sensor using NiO−SE primarily arises from its strong adsorption affinity for NO (confirmed by Figure 11). Although NiO exhibits similar electrochemical catalytic activity toward both NO and NO₂ (Figure 13a), it demonstrates a significant difference in the adsorption of these gases. Conversely, for the sensor utilizing (NiFe₂O₄ + 30 wt.% Fe₂O₃)-SE, the notable selectivity for NO₂ can be attributed to the high electrochemical reactivity of NiFe₂O₄ + 30 wt.% Fe₂O₃ with NO₂, as only a minor difference was observed in its adsorption capabilities (as presented in Figure 12 and Figure 13b).

4. Conclusions

To precisely track the level of NOx during iNO treatment, we developed YSZ-based gas sensors comprised of Ni-based SEs and Mn-based RE. These sensors using NiO or (NiFe₂O₄ + 30 wt.% Fe₂O₃)-SE demonstrated high selectivity, desirable stability, and excellent repeatability to high-level NO or low-level NO₂. The working principle of these sensors can be summarized that NO preferred to be adsorbed on the surface of NiO, while NiFe₂O₄ + 30 wt.% Fe₂O₃ composite demonstrated high electrochemical reactivity to NO₂, resulting in satisfactory sensing characteristics for these sensors in tracking the variation of NO_x. However, it should be noted that due to the high operational temperature of these sensors, it is necessary to overcome the limitations of high power consumption in practical applications. Additionally, the current response times might not be fast enough for real-time monitoring, and some strategies, such as a porous sensing layer to accelerate the diffusion rates of gases and machine learning for sensors’ response characteristics to recognize the differences in short-term transient response processes, will be further investigated to improve sensors’ efficiency. In summary, these findings mark a bright future for the application of robust and high-performance NO_x sensors in innovative iNO treatment technologies.