image
Home Journals Progress in Chemistry
Progress in Chemistry

Abbreviation (ISO4): Prog Chem      Editor in chief: Jincai ZHAO

About  /  Aim & scope  /  Editorial board  /  Indexed  /  Contact  / 
Online first  /  Current issue  /  All issues  /  Top access  /  RSS

Progress in Chemistry >

2026 , Vol. 38 >Issue 3: 561 - 576

DOI: https://doi.org/10.7536/PC20251118

Review

Online Monitoring of Greenhouse Gases for Sustainable Agriculture： The Role and Prospects of Semiconductor Sensing Technology

  • Kunmei Yang ,
  • Bingchen Zhu ,
  • Maojie Xu ,
  • Jia Yan ,
  • Hui Xu ,
  • Zhilong Song , *
Expand
  • Institute of Energy Research （School for Future Technology）， School of Environment and Safety Engineering， School of Agricultural Engineering， Zhenjiang 212013， China
*

† These authors contributed equally to this work

Received date: 2025-11-24

  Revised date: 2026-01-12

  Online published: 2026-03-18

Supported by

Senior Talent Fund of Jiangsu University(23JDG011)

Senior Talent Fund of Jiangsu University(23JDG012)

Fold

Abstract

Agricultural activities constitute a significant source of greenhouse gases including methane （CH4）， nitrous oxide （N2O）， and carbon dioxide （CO2）. Achieving continuous， real-time， and large-scale online monitoring of these gases represents a crucial means of advancing sustainable agriculture and addressing climate change. Although monitoring technologies such as infrared spectroscopy and electrochemical sensing have demonstrated mature performance in terms of accuracy and selectivity， their high cost， energy consumption， and complex deployment methods have limited widespread adoption in agricultural settings. This review highlights that semiconductor gas sensors， with their advantages of low cost， ease of integration， suitability for large-scale deployment， and deep integration with the Internet of things， are emerging as the ideal core technology for constructing future agricultural monitoring networks. The paper systematically reviews recent research advances inenhancing semiconductor sensor sensitivity， selectivity， and stability through strategies including nanomaterial regulation， heterostructure construction， catalytic and surface engineering， and signal processing algorithm integration. It also delves into practical challenges encountered in real agricultural environments—such as environmental interference， humidity effects， cross-sensitivity， and long-term stability—within livestock management and soil monitoring applications. Finally， this paper outlines future development trends for semiconductor gas sensors in agriculture： intelligent design of sensing materials， high integration of sensing nodes with IoT， multi-gas collaborative monitoring， and AI-based gas identification and emission modelling. Collectively， these advancements will drive the formation of future smart agricultural systems integrating precise monitoring， intelligent decision-making， and ecological management.

Contents

1 Introduction

2 Greenhouse gas detection technologies

2.1 Benchmark monitoring technologies

2.2 Semiconductor gas sensors

2.3 Comparison of technical pathways and evolutionary trends

3 Agricultural application scenarios

3.1 Livestock management

3.2 Soil and crop management

3.3 Greenhouse gas monitoring and control

4 Technical challenges and resolution pathways

4.1 Environmental interference

4.2 Long-term stability and power consumption

5 Outlook for sustainable integrated agriculture

5.1 Intelligent sensing and network architecture

5.2 Implementation of management closed-loop systems and comprehensive benefit assessment

Key words： greenhouse gases; sustainable agriculture; continuous online monitoring; semiconductor gas sensors; internet of things; precision agriculture

Cite this article

Kunmei Yang , Bingchen Zhu , Maojie Xu , Jia Yan , Hui Xu , Zhilong Song . Online Monitoring of Greenhouse Gases for Sustainable Agriculture： The Role and Prospects of Semiconductor Sensing Technology[J]. Progress in Chemistry, 2026 , 38(3) : 561 -576 . DOI: 10.7536/PC20251118

1 Introduction

Agricultural production constitutes a significant source of greenhouse gas emissions[1,2]， primarily comprising. Methane originating from ruminant enteric fermentation[3,4] and anaerobic conditions in paddy fields[5]， possessing a global warming potential approximately 28 times that of carbon dioxide over a century timescale[6]. Nitrous oxide from nitrification and denitrification processes associated with nitrogen fertiliser application and manure management[7,8]， whose per-molecule warming effect can reach 298 times that of carbon dioxide[9]. Carbon dioxide from fossil fuel consumption by agricultural machinery[10] and decomposition of soil organic carbon[11]. These emissions not only directly exacerbate climate change but also pose a counterproductive threat to agricultural productivity through their consequences， such as extreme weather events and water resource alterations[12]. Against this backdrop， developing sustainable agriculture centred on emissions reduction and balancing productivity with ecological resilience has become crucial for addressing global climate challenges and safeguarding food security[13-14]. Precise monitoring of agricultural greenhouse gases is an essential prerequisite for achieving this objective.
Given agriculture's significant contribution to greenhouse gas emissions， establishing an effective monitoring system holds dual core significance. Firstly， accurate and continuous emissions data form the foundation for developing scientifically grounded mitigation strategies and advancing precision agricultural practices[15-16]. By identifying key emission pathways， policymakers and producers can optimise production activities such as fertilisation， irrigation， and livestock management， thereby effectively reduce emissions intensity while safeguard productivity[17-18]. Secondly， reliable monitoring data provides essential support for fulfilling international commitments such as the Paris Agreement[19]， whilst offering credible quantitative evidence for market mechanisms like carbon trading. This enables agricultural producers to gain economic returns through emission reduction actions， thereby establishing a long-term incentive mechanism for sustainable development.
Traditional greenhouse gas monitoring primarily relies on periodic sampling and laboratory analysis， presenting limitations in data lag and insufficient completeness. In contrast， continuous online monitoring technology enables precise capture of dynamic emission processes in agricultural environments influenced by production activities and natural conditions through real-time data acquisition[20]. The high temporal resolution data it provides accurately identifies emission peaks and patterns， offering critical evidence for implementing precise emission reduction interventions[17]. Simultaneously， continuous online monitoring effectively circumvents errors arising from intermittent sampling， significantly enhancing the accuracy and reliability of emission inventories[21]. Furthermore， continuous online monitoring technology can be integrated with other smart agricultural technologies， such as precision farming tools and internet of things （IoT） devices. This integration facilitates the development of comprehensive management systems to oversee agricultural practices[22-23]. Against this backdrop， semiconductor gas sensing technology——characterised by low cost， miniaturisation， low power consumption， and ease of integration——is emerging as an ideal solution for achieving large-scale， sophisticated agricultural greenhouse gas monitoring. Despite facing challenges in absolute accuracy and long-term stability， semiconductor gas sensors are regarded as an ideal solution for constructing future universal agricultural sensing networks due to their comprehensive advantages. This review aims to systematically examine the current state of sustainable online continuous greenhouse gas monitoring technologies for agriculture， with semiconductor gas sensing technology serving as the central focus for in-depth discussion. This paper will first compare and analyse the performance limitations of mainstream monitoring technologies， then focus on how semiconductor sensors address the unique challenges of agricultural monitoring through material innovation and structural design. Concurrently， it will explore bottlenecks encountered in practical agricultural applications， including stability， selectivity， and data management. By integrating with the Internet of things and artificial intelligence， it will forecast future trends towards intelligent and green development.

2 Greenhouse gas detection technologies

Over the years， various technologies have been developed and refined to accurately detect and quantify gases such as CH4， N2O， and CO2. This chapter first outlines techniques including infrared spectroscopy and electrochemical sensing， before focusing on the principles， advances， and innovations within semiconductor gas sensing technology.

2.1 Benchmark monitoring technologies

To achieve precise measurement of agricultural greenhouse gases， infrared spectroscopy and electrochemical sensing have emerged as two widely recognised benchmark monitoring methods. However， when applied to complex， expansive and low-cost agricultural environments， their inherent limitations become increasingly apparent.

2.1.1 Infrared detection method

Infrared spectroscopy stands as one of the most mature and reliable greenhouse gas detection technologies. Its principle is based on the absorption of infrared light by gas molecules at specific wavelengths corresponding to their molecular vibrations. Each gas possesses a unique absorption spectrum， enabling precise identification and quantification. Within agricultural greenhouse gas monitoring， this technology has yielded multiple established application protocols. For instance， Bai et al.[24] successfully quantified CH4 and N2O emissions at a cattle farm in southeastern Australia using an FTIR trace gas analyser combined with an inverse dispersion model. However， its open-path measurement mode is susceptible to environmental interference. For example， to address the significant biases in N2O and CO2 measurements caused by diurnal humidity and temperature variations in field settings， researchers have developed chemometric calibration models， the validation setup for which is illustrated in Fig. 1[25]. Nevertheless， the high equipment costs， complex optical systems， and sensitivity to environmental temperature and humidity fluctuations limit its large-scale deployment in the agricultural sector.
图1 用于评估所测定的N2O和CO2浓度准确性的仪器示意图[25]

Fig.1 Schematic of the instrumentation used to assess the accuracy of N2O and CO2 concentration determined[25]. Copyright © Author（s） 2019. This work is distributed under the Creative Commons Attribution 4.0 License

2.1.2 Electrochemical detection technology

Electrochemical sensors measure gas concentrations by detecting electrical signals generated from redox reactions occurring at electrode surfaces. Such sensors typically offer advantages including compact size， low power consumption， and relatively low cost， making them suitable for real-time continuous online monitoring of mixed gas emissions in dynamic environments. For instance， the four-electrode mixed-potential electrochemical sensor developed by Halley et al.[26] achieved a detection limit of 40 ppm for methane in simulated natural gas mixtures， demonstrating potential applications in leak detection. As shown in Fig. 2， Fapyane et al.[27] significantly enhanced the response speed and long-term stability of Clark-type CO2 microsensors by optimising the composition of ionic liquid electrolytes， offering a novel pathway for high-precision micro-area monitoring. Electrochemical sensors offer viable portable solutions for agricultural greenhouse gas monitoring. However， their core disadvantages lie in limited operational lifespans， signal drift due to electrolyte depletion or electrode ageing， and cross-sensitivity to multiple gases. These factors constrain their application in long-term continuous online monitoring scenarios.
图2 采用离子液体-非质子溶剂电解质的二氧化碳快速反应微型传感器示意图[27]

Fig.2 Schematic diagram of a rapid-response carbon dioxide microsensor with ionic liquid-nonprotonic solvent electrolyte[27]. Copyright © 2020 American Chemical Society

2.2 Semiconductor gas sensors

The core principle of semiconductor gas sensors lies in the interaction between gas molecules and the surface of semiconductor materials， typically involving adsorption-reaction-desorption processes. This interaction alters the carrier concentration within the material， leading to significant changes in electrical resistance or conductivity. n-type semiconductors （e.g.， SnO2， ZnO， In2O3） exhibit decreased resistance upon exposure to reducing gases and increased resistance to oxidising gases； p-type semiconductors （e.g.， CuO， NiO， Cr2O3） demonstrate the opposite response behaviour. Traditional MOS sensors typically require high temperatures （300~500 ℃） to ensure sufficient surface reactivity and response speed， which fundamentally conflict with the strict energy budgets of battery-powered， long-term IoT deployments such as LoRaWAN nodes. As a result， conventional high-temperature MOS sensors are currently not directly suitable for continuous operation in energy-constrained IoT systems. Recent research efforts therefore focus on overcoming this limitation through nanoengineering， heterostructure construction， surface modification， and multidimensional sensing strategies， with the aim of reducing operating temperatures， enabling room-temperature-sensitive materials， or developing ultra-low-power microheaters. These approaches represent forward-looking pathways toward， rather than established solutions for， integrating semiconductor sensors into low-power IoT platforms.

2.2.1 CH4 sensing mechanisms

CH4 molecules possess exceptionally high C—H bond energies， and their catalytic oxidation on semiconductor surfaces typically requires substantial thermal activation energy. This necessitates conventional MOS sensors to operate at elevated temperatures with high power consumption， while also being susceptible to interference from complex gas mixtures. Current research frontiers are addressing these challenges through bandgap engineering and microstructural regulation. Loading noble metal nanoparticles （e.g.， Pt， Pd） represents a classical strategy for enhancing CH4 sensitivity and lowering operating temperatures. These particles serve as highly efficient catalytic sites， dissociating and activating CH4 molecules via the ‘spillover effect’ before migrating to the semiconductor surface for reaction. For instance， the palladium-modified indium oxide （Pd-In2O3） nanoflower structure developed by Han et al.[28] exhibits a response value of 1.162 to 1000 ppm（1 ppm=1×10-6） methane at 340 ℃， demonstrating rapid response and excellent repeatability， leveraging its temperature dependent sensing properties. Work by Bunpang et al.[29] demonstrates that metal ion doping can similarly regulate surface chemical states. Their flame-spray synthesised Cr-doped SnO2 nanoparticles exhibited a response of 1268 at 350 ℃ for 1% CH4， alongside exceptional selectivity. Beyond doping strategies， constructing heterojunctions represents another effective approach. The In2O3-CuO nanocomposite developed by Nie et al.[30] exhibits a unique p-type response to CH4 at 350°C while maintaining an n-type response to CO and H2， enabling direct gas species discrimination through this reversal of response types. As shown in Figure 3， More still， the silicon nanowire/TiO2 core-shell structure designed by Liu et al.[31] ingeniously harnesses core-shell interface effects to achieve effective detection of 20~120 ppm CH4 at room temperature， reducing power consumption to the microwatt level. This offers a highly promising solution for long-term， in situ agricultural monitoring. Unless otherwise specified， the detection limit （LOD） mentioned in this document is typically defined as three times the relative standard deviation （RSD） of sensor noise （σ） divided by the linear fit slope： LOD （ppb） = 3×σ/slope[32-33].
图3 （a） p-SiNWs/TiO2和（b） n-SiNWs/TiO2阵列在室温下检测CH4的传感机制[31]

Fig.3 Sensing Mechanism of （a） p-SiNWs/TiO2 and （b） n-SiNWs/TiO2 Array for CH4 Detection at Room Temperature[31]. Copyright © 2017 American Chemical Society

2.2.2 N2O sensing mechanisms

Unlike CH4， which can be detected at ambient temperatures using advanced materials， N2O poses greater challenges due to its high chemical stability.
N2O molecules exhibit high chemical stability， characterised by strong bond energies and reactive inertness， rendering effective adsorption-reaction processes on semiconductor surfaces challenging. Consequently， conventional sensors often require extremely high operating temperatures while suffering from insufficient sensitivity and selectivity. Current research primarily pursues two pathways to activate N2O molecules and enhance sensing performance， surface chemical modification and heterojunction engineering.
Research by Kanazawa et al.[34] revealed that loading 0.5% SrO onto SnO2 enhances its N2O response to approximately three times that of pure SnO2， enabling effective detection of 10~300 ppm N2O concentrations at 500 ℃. In contrast to such high temperature dependent approaches， Turlybekuly's innovative work[35] constructed a CuO/TiO2 p-n heterojunction with nanorods tuned to the double Debye length range. This significantly enhanced interfacial charge separation and transport efficiency， enabling ultra-sensitive detection of N2O at 50 ppb（1 ppb=1×10-9） levels at room temperature with rapid response recovery. It should be noted that the resistance-based sensing mechanism generally relies on the universal surface redox interactions of gas-sensitive materials with oxidizing species rather than on reactions unique to N2O[36-37]. In real agricultural soil environments， background oxygen concentrations can fluctuate at the percentage level due to soil respiration， gas diffusion， and moisture dynamics， which far exceed the typical ppb-level variations of N2O emissions. As a result， oxygen-induced resistance changes often dominate the sensor signal， leading to severe cross-sensitivity， baseline drift， and misinterpretation of N2O responses[38]. Suppressing O2 interference through material design （e.g.， selectively activated or O2-inert active sites） or system-level solutions is therefore a prerequisite for achieving specific N2O detection， yet remains largely unresolved for low-cost semiconductor sensors.
Although research has made progress in enhancing N2O sensing performance under controlled laboratory conditions， developing sensors with both low cost and long-term stability that are truly suitable for agricultural field deployment remains a major bottleneck，， particularly when compared to CH4 sensors[39]. N2O molecules exhibit higher chemical stability， and their activation on common metal oxide surfaces typically requires higher temperatures， resulting in higher power consumption and reduced long-term stability. Furthermore， the coexistence of O2， water vapor， and multiple nitrogen-containing gases in agricultural environments imposes stringent selectivity requirements that most current MOS-based N2O sensors fail to meet[40]. Consequently， the majority of existing studies remain at the laboratory material research stage， lacking field-validated， low-cost N2O sensing solutions capable of reliable long-term operation in complex agricultural settings. This limitation continues to constrain the establishment of comprehensive greenhouse gas monitoring networks in agriculture.

2.2.3 CO2 sensing mechanisms

Accurate detection of CO2 at ambient background concentrations imposes stringent demands on the selectivity of semiconductor sensors， particularly their resistance to water vapour interference. Given CO2’s inherent chemical inertness， sensing mechanisms often rely indirectly on reactions between adsorbed oxygen species and CO2， or on the material surface’s conductive response to carbonate species formation.
The hollow CeO2 nanostructures synthesised by Zito et al.[41] exhibit a sensitivity twice that of solid CeO2 nanoparticles at a low operating temperature of 100 ℃， owing to their enormous specific surface area and efficient gas diffusion pathways. Duan et al.[42] employed a Co-doping strategy to modify LaFeO3， yielding LaCo0.1Fe0.9O3 sensors exhibiting broad-range CO2 responsivity at 220 ℃. Crucially， these sensors demonstrated exceptional tolerance across a wide developed by Li et al.[43] achieves effective CO2 humidity range of 0~86.7% RH. Furthermore， surpassing conventional resistive sensors， the La2O3/GaN heterojunction field-effect transistor sensor detection within the 1~500 ppm range by measuring changes in electrical transport properties induced by the gate-sensitive layer. This demonstrates the novel device’s unique advantages for low-concentration， high-precision detection.

2.2.4 Pathways for optimising semiconductor sensor performance

Breakthroughs in semiconductor gas sensor performance fundamentally stem from synergistic dimensional nanostructuring provides ample design at both material and interfacial scales， with core pathways summarised as follows. Firstly， low adsorption and reaction sites for gas molecules by constructing nanostructures with large specific surface areas， forming the foundation for enhanced sensitivity. Secondly， interface engineering involves creating heterojunctions or applying noble metal modifications to induce band structure modulation and built-in electric fields at interfaces. This enhances signal detection， lowers operating temperatures， and improves selectivity.

2.3 Comparison of technical pathways and evolutionary trends

To sum up， these three core monitoring technologies have different performances and are complementary to each other. As shown in Table 1， infrared spectroscopy offers the highest precision， yet its cost and bulkiness constrain dense deployment.
表1 主要温室气体监测技术对比

Table 1 Comparison of major greenhouse gas monitoring technologies

Technical Indicators Infrared Spectroscopy Electrochemical Sensor Semiconductor Sensor
Cost High Medium Low
Detection Accuracy Extremely High （ppb Level） High （ppm Level） Medium （ppm Level）
Power Consumption High Low Low
Integration Level Low Medium High
Applicable Scenarios Fixed-Point High-Precision Monitoring Portable Real-Time Monitoring Large-Scale Network Deployment
Long-Term Stability High Limited Medium
Electrochemical sensors excel in detecting specific gases， but their lifespan and stability pose bottlenecks for long-term monitoring. In comparison， semiconductor gas sensors， while still requiring improvements in absolute accuracy and long-term stability， demonstrate the most pronounced comprehensive advantages in terms of cost， power consumption， miniaturisation， and integration convenience. This facilitates extensive monitoring network coverage， enabling smallholder farmers' participation in carbon markets and promoting social equity. How do these semiconductors sensors perform in complex real-world environments such as farmland and pastures? What specific challenges arise， and what corresponding solutions exist? These questions form the focus of our subsequent discussion.

3 Agricultural application scenarios

Building upon the technical foundations outlined in Chapter Two， we shall now address the practical application of semiconductor sensors within agricultural environments. Continuous online monitoring technology provides an unprecedented data foundation for understanding and managing greenhouse gas emissions. In this chapter， we shall explore how the cost-effectiveness and miniaturisation advantages of semiconductor sensors can be leveraged for integration into two critical agricultural application scenarios： livestock management and soil-crop systems.
This chapter will focus on two core scenarios， livestock management and soil and crop systems， to conduct an in-depth analysis of their monitoring requirements， exploring the technical suitability and deployment strategies of semiconductor sensors.

3.1 Livestock management

Livestock production constitutes a significant source of CH4 and N2O emissions within agriculture. Continuous online monitoring is essential for precise quantification of emissions and optimisation of management strategies.
In the rumen of ruminant animals such as cattle， deer and camels， feed undergoes microbial fermentation to produce methane[44]. Regarding individual animal emission monitoring， traditional methods such as portable accumulative chambers[45] and respiratory hoods， whilst capable of high-precision measurements， prove challenging for continuous operation and incur high costs. High-end optical equipment based on infrared absorption spectroscopy[46] an open-path respiratory calorimetry systems （e.g.， GreenFeed[47]） are widely employed for research-grade monitoring. However， such systems are prohibitively expensive for widespread adoption. This cost barrier highlights the key advantage of semiconductor gas sensors. Semiconductor gas sensors are inexpensive and enable long-term， continuous online monitoring of emission patterns across entire herds， whereas deploying high-precision systems on such a scale is economically unfeasible. Research by Dida et al.[48] corroborates this direction. A comparison between the low-cost MQ-4 semiconductor sensor and the GreenFeed system revealed a moderate correlation between the two sets of data. This provides empirical support for the low-cost solution， whilst also indicating that the semiconductor sensor requires further refinement in measurement accuracy and stability before it can be deployed in practical applications. As shown in Fig. 4， the storage and treatment processes of manure slurry represent concentrated emission points for CH4 and N2O[49]. Studies such as that by Støckler et al.[50] employ high-precision equipment like cavity ring-down spectroscopy for simultaneous multi-component monitoring， thereby resolving the complex effects of factors such as solid-liquid separation and seasonal temperature variations on CH4 and N2O emissions. However， achieving comprehensive monitoring of dispersed manure storage facilities necessitates deploying sensor networks comprising numerous fixed nodes. This imposes stringent cost constraints on individual sensor nodes， thereby establishing a deployment paradigm where low-cost sensors are not merely an option but an imperative. Semiconductor sensors possess a distinct advantage in addressing this challenge due to their inherent suitability for dense network integration. For instance， semiconductor sensors based on rare earth oxides or perovskites have demonstrated excellent gas sensitivity and superior humidity tolerance at the laboratory level[51,52]， making them highly suitable for the high humidity， high concentration environments typical of manure storage areas.
图4 动物粪便中气体排放和处理流程示意图[49]

Fig.4 Schematic Diagram of Gas Emission and Treatment Process in Animal Manure[49]. Copyright © 2016 American Chemical Society

3.2 Soil and crop management

Soil N2O and CO2 emissions primarily originate from nitrification and denitrification processes stimulated by nitrogen fertilizer application， alongside soil respiration. These emissions are regulated by factors such as soil moisture， temperature， and pH， exhibiting highly spatiotemporal heterogeneity and pulsed emission characteristics[53-54]. Such dynamic variations impose stringent demands on the spatiotemporal resolution and deployment density of monitoring technologies， which current mainstream techniques cannot yet fulfil cost-effectively. Currently， gas analysis systems based on infrared spectroscopy techniques such as non-dispersive infrared （NDIR） remain the benchmark for precision in soil CO2 flux monitoring. For instance， the laser spectroscopy system developed by Stiefvater et al. [55] can capture N2O flux pulses with high accuracy； while the ‘Artemisia’ soil CO2 NDIR probe developed by Anderson et al.[56] represents the evolution of this technology towards in situ， long-term monitoring. Nevertheless， even after optimization， the cost of core infrared components remains a constraint on large-scale， high-density deployment. This constitutes the primary technical bottleneck in realizing ‘carbon-nitrogen flux maps’ for agricultural fields.
This bottleneck represents a critical research direction and application gap for semiconductor gas sensors. Theoretically， semiconductor gas sensors hold potential for low cost， miniaturization， and low power consumption， making them an ideal choice for constructing the high-density sensor networks required to map soil greenhouse gas micro-emission hotspots. Although the large-scale， autonomous deployment of these sensors for greenhouse gas monitoring within soil-crop systems remains an emerging frontier with limited validated data， fundamental material advances discussed in Section 2 provide essential proof-of-concept， such as room-temperature CH4 sensors[31] and moisture-tolerant CO2 sensors[42]. The primary challenge now lies in translating these laboratory-validated technologies into integrated， field-ready systems capable of long-term stable operation within complex soil environments.
Monitoring CH4 in paddy fields presents unique challenges due to the complex wetland environment. Presently， floating chambers coupled with high-precision analysers remain unrivalled in accuracy[57]. Semiconductor CH4 sensors suitable for paddy fields with strong interference resistance represent a key research direction， though their technical maturity lags significantly behind the aforementioned methods. For instance， the In2O3-CuO sensor with p-n response conversion properties[30]， mentioned in Section 2.2， holds potential application value in paddy field monitoring under complex gas backgrounds due to its ability to distinguish CH4 from other gases.
Building upon this foundation， the practical solution for agricultural monitoring lies in the design of integrated systems. The value of portable systems incorporating multiple sensors resides in providing a design paradigm for multi-parameter fusion. Consequently， the ideal future architecture is likely to be a hybrid network combining high-end and low-end components， where high-precision instruments serve for calibration benchmarks and data validation， while low-cost semiconductor sensors enable extensive coverage and high-frequency sampling. Only through such strategic integration can a truly accurate and comprehensive field sensing system be established.

3.3 Greenhouse gas monitoring and control

Greenhouses serve as critical environments for both greenhouse gas emissions and precise regulation. Their enclosed nature offers advantages such as controllable fluctuations in temperature， humidity， and gas concentrations， providing ideal conditions for the initial promotion of semiconductor sensors[58]. Compared to open fields and pastures， greenhouses feature enclosed or semi-enclosed environments where temperature， humidity， and ventilation can be artificially controlled. Gas concentration fluctuations are relatively manageable， and monitoring areas are concentrated， eliminating the need for extensive long-distance deployment. These characteristics make greenhouses an ideal initial application scenario for the early deployment and promotion of semiconductor gas sensors. Compared to the high cost of traditional infrared spectroscopy equipment and the stability limitations of electrochemical sensors， semiconductor sensors——with their low cost and ease of integration——can be rapidly deployed across different greenhouse zones to enable real-time monitoring of multiple gas concentrations[20]. Their compatibility with greenhouse IoT systems provides data support for low-carbon management practices such as precision fertilization and smart ventilation， thereby advancing the development of low-carbon greenhouses.

4 Technical challenges and resolution pathways

Having clarified the technical principles and performance potential of semiconductor gas sensors， we must urgently address a practical question， what formidable challenges will these sensors encounter when transitioning from controlled laboratory conditions to the complex and variable real-world agricultural environment? And how might these challenges be resolved? The open nature of agricultural settings necessitates sensors capable of withstanding extreme temperature fluctuations， high humidity， and cross-sensitivity from multiple gases. These combined factors lead to signal drift， diminished sensitivity， and selectivity overlap. This chapter systematically analyses the core technical bottlenecks confronting semiconductor gas sensors in agricultural applications， focusing on the most promising contemporary solutions and development pathways.

4.1 Environmental interference

The complex agricultural environment poses a major challenge to gas sensors， as fluctuations in temperature， humidity， and gas composition directly affect their performance. For promising semiconductor sensors， overcoming these environmental interferences is a prerequisite for long-term， stable operation.

4.1.1 Temperature fluctuations and their compensation

The electrical characteristics of semiconductor gas sensors are significantly influenced by temperature， posing a fundamental challenge to their application in both conventional and agricultural settings. Whilst strategies such as heterojunction engineering （as discussed in Section 2.2） aim to achieve low-power operation， even at room temperature， thereby inherently mitigating power consumption and thermal drift issues at the source， many high-performance sensing responses still necessitate elevated temperatures. The work of Bunpang et al.[29] vividly illustrates this dependency， identifying 350 ℃ as the optimum operating temperature for a chromium-doped SnO2 CH4 sensor and noting that temperature fluctuations directly compromise response and stability. To address this challenge， researchers have developed compensation strategies combining hardware and software approaches. At the hardware level， Cheng et al.[59] achieved real-time temperature measurement and signal decoupling by integrating miniature thermistors and constructing sensor arrays， enabling direct hardware-level correction. At the software level， Yang et al.[60] employed convolutional neural networks （CNNs） to model response data under temperature variations， successfully reducing the relative error in H2 concentration prediction to 5.7%. These temperature and humidity compensation modules are progressively being integrated into next-generation smart sensor designs， providing more stable performance guarantees for agricultural field monitoring.

4.1.2 Humidity interference and suppression strategies

Humidity ranks among the most significant factors affecting sensor performance， particularly in high-humidity environments such as paddy fields and greenhouses. Water vapour in the air competes with target gas molecules for adsorption on the sensor surface， altering the band structure of semiconductor materials and the chemical properties of their surface-active sites. This leads to drift in the response signal and a reduction in sensitivity[61-62]. To mitigate humidity interference， multiple countermeasures have been developed. Material modification represents a fundamental strategy. For instance， Qu et al.[63] effectively blocked water molecule permeation into the sensing layer by modifying the surface of semiconductor metal oxides （SMOX） with a composite layer of hydrophobic polymers and metal-organic frameworks （MOFs）， enabling the sensor to maintain stable response across a relative humidity range of 0~90%. Active regulation offers another effective approach. Algorithmic compensation represents another effective approach. Mahdavi et al.[64]. employed support vector machines （SVM） and k-NN methods to analyse the response patterns of sensors under temperature-controlled modulation， successfully quantifying and compensating for the interference caused by humidity variations on detection accuracy. Additionally， in environments with saturated humidity， such as rice paddies during early morning hours or enclosed livestock facilities， condensation may form on the sensor housing， protective mesh， or membrane surfaces when the ambient temperature drops below the dew point. Even when hydrophobic polymers or MOF coatings are employed to prevent direct liquid water penetration， condensed water layers on external structures can still reduce effective gas diffusion rates and lead to delayed response or signal attenuation. This condensation-induced mass transfer limitation represents an additional challenge for field deployment that is often overlooked in laboratory studies. Therefore， future sensor packaging and encapsulation strategies must consider not only hydrophobicity at the material level but also structural and thermal designs that mitigate condensation while maintaining sufficient gas permeability.

4.1.3 Enhancing cross-sensitivity and selectivity

Agricultural environments are rich in interfering gases such as ammonia， hydrogen sulphide， and volatile organic compounds[65-66]. The broad-spectrum response characteristics of semiconductor sensors pose significant selectivity challenges[67]. When monitoring target greenhouse gases like methane and nitrous oxide， mitigating interference from other gases becomes critical for practical applications.
To enhance gas recognition selectivity， researchers primarily employ sophisticated material engineering and intelligent data processing to strengthen specific responses to target gases[68]. Material engineering represents the core approach for improving selectivity. As shown in Fig. 5， Wang et al.[69] significantly enhanced the electrostatic interactions and Lewis acidity of the material surface towards CO2 molecules by alkali-metal modification of MOF-74. As shown in Fig. 5c， the optimized adsorption configuration of CO2 on the M-Mg-MOF-74 surface enhances its specific recognition capability. The corresponding Langmuir adsorption isotherm models （Fig. 5d， e） further confirm the improvement in adsorption kinetic parameters， providing a theoretical basis for highly selective sensing This work demonstrates that rationally designing the surface chemistry of sensing materials， rather than relying solely on high-temperature reactions， represents a key direction for developing next-generation， high-precision， low-power greenhouse gas sensors.
图5 （a） Mg-MOF-74、（b） M-Mg-MOF-74（M = Li、Na、K）和 （c） CO2在 M-Mg-MOF-74上的吸附结构图；（d， e）Langmuir吸附等温线模型可确定吸附动力学参数[69]

Fig.5 Structure diagram of （a） Mg-MOF-74， （b） M-Mg-MOF-74 （M = Li， Na， K）， and （c） CO2 adsorption on M-Mg-MOF-74. （d， e） Langmuir adsorption isotherm model determines the adsorption kinetic parameters[69]. Copyright © 2025 American Chemical Society

4.2 Long-term stability and power consumption

The harsh operating conditions encountered in agricultural settings impose heightened demands on sensor long-term stability and maintenance costs. Sensor drift， material ageing， and contaminant accumulation represent key constraints to their large-scale deployment.
Regarding long-term stability， rational material and heterojunction design represent the core approach to enhancing durability[70]. As shown in Fig. 6， Haldar et al.[71] successfully developed a high-performance CO2 sensor operating at room temperature by constructing a CuO/rGO p-p heterojunction. As shown in Fig. 6g， the sensor maintained 98% of its initial response after 30 days of continuous testing at 500 ppm CO2， demonstrating excellent long-term stability. Fig. 6i further indicates its outstanding moisture resistance across a wide humidity range， validating the potential applicability of this heterojunction structure in complex agricultural environments. First-principles studies further elucidated its interaction mechanism with CO2 molecules. This work provides a highly promising design paradigm for developing low-power， long-lifetime greenhouse gas sensors suitable for high-humidity， variable agricultural environments without requiring heating.
图6 （a） 所有样品对400 ppm CO2的温度依赖性响应，（b） CuO/rGO-1、5、10和20的动态CO2传感响应-恢复曲线，（c） CuO/rGO-5在不同CO2含量下的动态电阻变化曲线，（d） CuO/rGO-5在300 ppm CO2含量下的响应-恢复曲线，（e） CuO/rGO-5、10和20异质结构在不同CO2含量下的瞬态响应/恢复曲线，（f） CuO/rGO-5在400和50 ppm CO2下的动态电阻变化曲线，（g） CuO/rGO-5的长期稳定性曲线，（f） CuO/rGO-5在CO2含量为 400 和 50 ppm 时的动态电阻变化曲线，（g） CuO/rGO-5在CO2含量为500 ppm时的长期稳定性曲线，（h） 所有样品对不同分析气体的选择性，（i） 湿度对制备的CuO/rGO-5 CO2传感器的影响[71]

Fig.6 （a） Temperature-dependent CO2 sensing response of all samples to 400 ppm of CO2， （b） dynamic CO2 sensing response-recovery curve of CuO/rGO-1， 5， 10， and 20， （c） dynamic resistance variation curves of the CuO/rGO-5 for different CO2 content， （d） response and recovery curve of CuO/rGO-5 at 300 ppm of CO2 content， （e） transient response/recovery curves of CuO/rGO-5， 10， and 20 heterostructures for different content of CO2， （f） dynamic resistance variation curves of the CuO/rGO-5 at 400 and 50 ppm of CO2， （g） long-term stability curve of the CuO/rGO-5 at 500 ppm of CO2， （h） Selectivity bar diagram of all the samples for different analyte gases， （i） Humidity effect on the prepared CuO/rGO-5 CO2 sensor[71]. Copyright © 2024 The Authors. Published by American Chemical Society.

Reducing the sensor’s own operational power consumption is one of the fundamental methods for minimising its maintenance requirements and enabling long-term deployment. Conventional thermopile sensors typically consume hundreds of milliwatts. Shwetha et al.[72] designed A milliwatt-level suspended SiO2 thin-film microheater has been designed， achieving ultra-low power consumption while ensuring uniform temperature distribution across the heating zone. The integrated sensing material exhibits sensitivity variations of 21% and 70% for 400 ppm and 1000 ppm CO2 respectively， demonstrating the significant value of miniaturisation and low-power design. These advances in low-power hardware， coupled with the rich， multi-variable data generated by sensor arrays （as discussed in Section 4.1.3）， lay crucial foundations for integrating semiconductor sensors into energy-autonomous IoT nodes and data-driven artificial intelligence platforms， which will be explored in the outlook.

5 Outlook for sustainable integrated agriculture

The preceding sections have critically examined the technical principles， performance optimisation pathways， and challenges faced by semiconductor gas sensors as front-end sensing cores within complex agricultural environments. However， to fully unlock the potential of semiconductor sensors and overcome limitations in accuracy and long-term stability， efforts must extend beyond sensor optimisation alone. These sensors must be integrated into a broader systemic architecture. This chapter explores how to construct sustainable agriculture and integrated farm systems based on semiconductor sensor networks. Through intelligent network architecture， multi-source data fusion， and decision algorithms， dispersed sensor signals are transformed into reliable farm-level management insights， ultimately achieving closed-loop management for greenhouse gas monitoring and resource recycling[73-74].
图7 半导体传感器在智慧农业中的应用前景

Fig.7 The application prospects of semiconductor sensors in smart agriculture

5.1 Intelligent sensing and network architecture

LoRaWAN， with its low-power characteristics， emerges as an ideal choice for connecting dispersed sensing nodes. For instance， the AgriLink platform designed by Baraka et al.[75] for Moroccan agriculture successfully utilised LoRaWAN to establish a low-cost， energy-efficient monitoring network， validating the technology's feasibility in field applications. At the data and decision layer， the core challenge lies in transforming raw data into actionable insights. This is typically achieved through edge-cloud collaborative computing and intelligent data fusion. The Farm-Light Seek framework proposed by Jiang et al.[76] exemplifies this approach， processing multi-source data at network edge nodes， such as gas concentrations， imagery， and meteorological information. It performs cross-modal reasoning and disease detection to enable low-latency on-site management decisions. The future of agricultural gas monitoring lies in the deep integration of sensing hardware with intelligent data processing， moving from single-point detection to discriminative intelligence[77]. Concurrently， it collaborates with the cloud for model updates， demonstrating the tight coupling between perception and decision-making. Furthermore， data fusion drives management optimisation by establishing correlation models between environmental parameters and agricultural emission reduction/yield enhancement objectives. Kang et al.[78] employs a recursive segmentation model to deeply integrate IoT environmental data with crop growth states， establishing a closed-loop analytical framework that reveals data's core value in understanding intrinsic patterns within agricultural ecosystems. Concurrently， to address the energy constraints of long-term field monitoring， Lu et al.[79] integrated the Simultaneous Wireless Information and Power Transmission technology into sensor network architectures. Through intelligent energy management， they provided a breakthrough solution for the sustainable operation of permanent field monitoring nodes.

5.2 Implementation of management closed-loop systems and comprehensive benefit assessment

Translating advanced sensing technologies into tangible agricultural production benefits hinges on establishing comprehensive management closed-loop systems and conducting precise evaluations of their integrated efficacy. This section systematically analyses the practical application outcomes and value of greenhouse gas monitoring technologies within agricultural environmental management through representative case studies.
The core of the management loop lies in establishing a complete technical chain encompassing ‘monitoring-analysis-decision-execution’. The multi-sensor IoT system developed by the Balan team[80] stands as a prime example in this field. By integrating atmospheric greenhouse gas monitoring with soil parameter analysis， this system constructs a comprehensive ‘soil-atmosphere’ integrated monitoring solution. More significantly， the system incorporates an intelligent decision support module that translates real-time monitoring data into actionable agronomic recommendations， thereby achieving a complete closed loop from environmental perception to management decision-making. This integrated solution provides scientific rationale for farm managers， markedly enhancing the precision and timeliness of agricultural environmental management.
Regarding economic benefit assessment， Saraswathi et al.[81] offers compelling evidence. The intelligent agricultural system developed by this team not only achieves precise monitoring of environmental parameters， but its embedded AI analysis module can predict crop market prices with 92% accuracy. This innovation effectively combines environmental monitoring data with market intelligence， providing farm managers with decision support that balances environmental and economic benefits. It highlights the immense potential of intelligent monitoring systems in elevating agricultural management standards. At the technical validation and application level， multiple research projects have demonstrated the feasibility of low-cost sensing solutions in real agricultural environments. Suriano et al.[82] successfully achieved continuous online monitoring of gases including NH3， CH4， H2S， and CO2 in poultry manure storage facilities by deploying portable monitoring units based on chemical resistance principles in remote rural areas. This project innovatively employed email protocols to overcome communication challenges posed by unstable wireless signals， demonstrating the practical value of low-cost semiconductor gas sensors in complex agricultural settings. Collectively， these case studies demonstrate that modern agricultural environmental management is undergoing a profound shift from experience-driven to data-driven approaches. By establishing robust management cycles and scientific benefit assessment systems， greenhouse gas monitoring technology will not only make significant contributions to environmental protection but also provide reliable technical support for sustainable agricultural development. This demonstrates broad application prospects.
In summary， semiconductor gas sensors hold significant application potential in agricultural greenhouse gas monitoring due to their low cost， ease of integration， and scalability for large-scale networking. This technological approach aligns closely with China's dual carbon strategy， offering critical support for agricultural emissions monitoring， carbon trading data infrastructure， and low-carbon technology promotion. Looking ahead， through the convergence of materials science， algorithms， and the Internet of things， semiconductor sensing technology is poised to become a key driver in advancing smart agriculture and achieving carbon neutrality goals.

References

Publishing order | Descend order by publishing year | Descend order by cited within
[1]
Chanu P H, Singh A, Bharadwaj S, Pandey S. Kollah B.Climate Smart Agriculture.Eds.: Mohanty S R, Cham: Springer, 2024: 1.

[2]
Liu G, Zhang F, Deng X Z. Commun. Earth Environ., 2023, 4: 161.

[3]
Runkle B R K, Suvočarev K, Reba M L, Reavis C W, Smith S F, Chiu Y L, Fong B. Environ. Sci. Technol., 2019, 53(2): 671.

[4]
Qian H Y, Zhu X C, Huang S, Linquist B, Kuzyakov Y, Wassmann R, Minamikawa K, Martinez-Eixarch M, Yan X Y, Zhou F, Sander B O, Zhang W J, Shang Z Y, Zou J W, Zheng X H, Li G H, Liu Z H, Wang S H, Ding Y F, van Groenigen K J, Jiang Y. Nat. Rev. Earth Environ., 2023, 4(10): 716.

[5]
Jackson R B, Saunois M, Bousquet P, Canadell J G, Poulter B, Stavert A R, Bergamaschi P, Niwa Y, Segers A, Tsuruta A. Environ. Res. Lett., 2020, 15(7): 071002.

[6]
Intergovernmental Panel on Climate C. Climate Change 2014: Mitigation of Climate Change: Working Group III Contribution to the IPCC Fifth Assessment Report. Cambridge: Cambridge University Press, 2015.

[7]
Pelletier A J, Faubert P, Lafond J, Bertrand N, Legault J, Ouimet R, Pelster D E, Pichette A, Ziadi N, Paré M C. Geoderma Reg., 2025, 43: e01024.

[8]
Liao W H, Liu M L, Bian Y, Wang W J, Gao Z L. J. Environ. Manag., 2025, 395: 127786.

[9]
Forster P, Ramaswamy V. Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Eds.: Solomon S, Qin D, Manning M. Cambridge: Cambridge University Press, 2007, 129.

[10]
Zhuang M H, Wang X, Yang Y, Wu Y F, Wang L G, Lu X. Nat. Food, 2025, 6(5): 513.

[11]
Raffa D W, Bogdanski A, Tittonell P. Biomass Bioenergy, 2015, 81: 345.

[12]
Khan A, Ali M, Mandadi A, Anjum A, Balzter H. Remote. Sens., 2025, 17(21): 3615.

[13]
Pisante M, Gogoi N, Farooq M. Sustainable Agriculture and the Environment. Amsterdam: Elsevier, 2023: 3.

[14]
Gulisano G, Strano A, De Luca A I, Falcone G, Iofrida N, Stillitano T. Sustainable Food Systems from Agriculture to Industry. Amsterdam: Elsevier, 2018: 123.

[15]
Delgado J A, Vandenberg B, Kaplan N, Neer D, Wilson G, D’Adamo R, Carter J, O’Gan L, Grow N, Marquez R, Arthur D, Eve M, Del Grosso S J, Johnson J M F, Karlen D L, Durso L, Finley J, Acosta-Martinez V, Knaebel D B, Harmel D, Derner J D. J. Soil Water Conserv., 2018, 73(6): 158A.

[16]
Yan M, Cheng K, Luo T, Yan Y, Pan G X, Clean. Prod., 2015, 104: 130.

[17]
Sha Y Y, Li J Q, Zhang H B, Wei C B. Sci. Rep., 2025, 15: 38421.

[18]
Samal A, Sahu S K, Mishra A, Mangaraj P, Pan S K, Beig G. Aerosol Air Qual. Res., 2024, 24(6): 230204.

[19]
Nyawira S S, Herold M, Mulatu K A, Roman-Cuesta R M, Houghton R A, Grassi G, Pongratz J, Gasser T, Verchot L. Mitig. Adapt. Strateg. Glob. Change, 2024, 29(2): 13.

[20]
Zhang H Z, Ren R, Gao X, Wang H S, Jiang W, Jiang X S, Li Z F, Pan J J, Wang J Y, Wang S H, Ding Y F, Mu Y, Wang X L, Du J Z, Li W T, Xiong Z Q, Zou J W. Water Res., 2025, 268: 122663.

[21]
Fortin M. Mitig. Adapt. Strateg. Glob. Change, 2021, 26(2): 7.

[22]
Song Y C, Bi J P, Wang X. Internet Things, 2024, 25: 101029.

[23]
Barr A G, Landgraf J, Martinez-Velarte M, Vrekoussis M, Sussmann R, Morino I, Strong K, Zhou M Q, Velazco V A, Ohyama H, Warneke T, Hase F, Borsdorff T. Atmos. Meas. Tech., 2025, 18(21): 6093.

[24]
Bai M, Suter H, Lam S K, Davies R, Flesch T K, Chen D L. Agric. For. Meteor., 2018, 258: 50.

[25]
Lin C H, Grant R H, Heber A J, Johnston C T. Atmos. Meas. Tech., 2019, 12(6): 3403.

[26]
Halley S, Ramaiyan K, Garzon F, Tsui L K. Sens. Actuat. B Chem., 2023, 392: 134031.

[27]
Fapyane D, Revsbech N P. ACS Sens., 2020, 5(8): 2604.

[28]
Han L Y, Zhang S S, Zhang B, Zhang B W, Wang Y, Bala H R, Zhang Z Y. J. Materiomics, 2022, 8(3): 545.

[29]
Bunpang K, Wisitsoraat A, Tuantranont A, Singkammo S, Phanichphant S, Liewhiran C. Sens. Actuat. B Chem., 2019, 291: 177.

[30]
Nie S, Li J, He Y X, Yin X T. ACS Sens., 2024, 9(12): 6390.

[31]
Liu D, Lin L M, Chen Q F, Zhou H Z, Wu J M. ACS Sens., 2017, 2(10): 1491.

[32]
Song Z L, Wei Z R, Wang B C, Luo Z, Xu S M, Zhang W K, Yu H X, Li M, Huang Z, Zang J F, Yi F, Liu H. Chem. Mater., 2016, 28(4): 1205.

[33]
Li M, Zhou D X, Zhao J, Zheng Z P, He J G, Hu L, Xia Z, Tang J, Liu H. Sens. Actuat. B Chem., 2015, 217: 198.

[34]
Kanazawa E, Kugishima M, Shimanoe K, Kanmura Y, Teraoka Y, Miura N, Yamazoe N. Sens. Actuat. B Chem., 2001, 75(1/2): 121.

[35]
Turlybekuly A, Sarsembina M, Mentbayeva A, Bakenov Z, Soltabayev B. Sens. Actuat. B Chem., 2023, 397: 134635.

[36]
Chizhov A, Kutukov P, Astafiev A, Rumyantseva M. Sensors, 2023, 23(3): 1055.

[37]
Milam-Guerrero J, Yang B X, To D T, Myung N V. ACS Sens., 2022, 7(12): 3598.

[38]
Yamazoe N, Suematsu K, Shimanoe K. Sens. Actuat. B Chem., 2012, 163(1): 128.

[39]
Rivera-Martinez R, Kumar P, Laurent O, Broquet G, Caldow C, Cropley F, Santaren D, Shah A, Mallet C, Ramonet M, Rivier L, Juery C, Duclaux O, Bouchet C, Allegrini E, Utard H, Ciais P. Atmos. Meas. Tech., 2024, 17(14): 4257.

[40]
Shiraishi N, Lu J, Fauzi F B, Imaizumi R, Tsukahara T, Mogari S, Iida S, Matsukura Y, Teramoto S, Yokoi K, Ichinose I, Kimura M. Biosensors, 2025, 15(1): 55.

[41]
Zito C A, Perfecto T M, Dippel A C, Volanti D P, Koziej D. ACS Appl. Mater. Interfaces, 2020, 12(15): 17745.

[42]
Duan X H, Jiang Y D, Liu B H, Duan Z H, Zhang Y J, Yuan Z, Tai H L. Sens. Actuat. B Chem., 2024, 402: 135136.

[43]
Li W M, Zheng H Z, Sokolovskij R, Wen K Y, Jiang Y, Shi Y Q, Deng C K, Wang Q, Yu H Y. Sens. Actuat. A Phys., 2025, 384: 116286.

[44]
Kumari S, Dahiya R P, Naik S N, Hiloidhari M, Thakur I S, Sharawat I, Kumari N. Environ. Dev., 2016, 20: 31.

[45]
Levrault C M, Difford G F, Steinheim G, Groot Koerkamp P W G, Ogink N W M. Biosyst. Eng., 2023, 236: 201.

[46]
Shen F J, Chen D, Wang G X, Lu J, Hu X Y, Gao X M, Fertein E, Chen W D. J. Quant. Spectrosc. Radiat. Transf., 2024, 327: 109131.

[47]
Tadesse D, Puchala R, Yirga H, Patra A K, Goetsch A L. Smart Agric. Technol., 2025, 11: 100989.

[48]
Dida M F, Shirvan M B, Kawaguchi T, Garcia S C, Gonzalez L A. Smart Agric. Technol., 2025, 10: 100706.

[49]
Hou Y, Velthof G L, Lesschen J P, Staritsky I G, Oenema O. Environ. Sci. Technol., 2017, 51(1): 375.

[50]
Støckler A H, Kamp J N, Feilberg A, Dalby F, Hafner S D, Pedersen J. J. Environ. Manag., 2025, 390: 126333.

[51]
Tri N, Phung Anh N, Ba Long D, Gia Thien Thanh H, Thi Hong Nhung B, Thi Thuy Van N, Anh Ha C, Tien Cuong H, Cam Loc L. Fuel, 2025, 380: 133137.

[52]
Youn J R, Kim M J, Kim K C, Kim M, Jung T, Go K S, Jeon S G, Kim W. Fuel Process. Technol., 2024, 263: 108130.

[53]
Schädel C, Bader M K, Schuur E A G, Biasi C, Bracho R, Čapek P, De Baets S, Diáková K, Ernakovich J, Estop-Aragones C, Graham D E, Hartley I P, Iversen C M, Kane E, Knoblauch C, Lupascu M, Martikainen P J, Natali S M, Norby R J, O’Donnell J A, Chowdhury T R, Šantrůčková H, Shaver G, Sloan V L, Treat C C, Turetsky M R, Waldrop M P, Wickland K P. Nat. Clim. Change, 2016, 6(10): 950.

[54]
Wakhid N, Hirano T, Okimoto Y, Nurzakiah S, Nursyamsi D. Sci. Total Environ., 2017, 581/582: 857.

[55]
Stiefvater G, Hespos Y, Wiedenmann D, Lambrecht A, Brunner R, Wöllenstein J. Sensors, 2023, 23(15): 6686.

[56]
Anderson J F, Huber D P, Walsh O A. Sensors, 2024, 24(18): 6034.

[57]
Ma Y Z, Qiao Y F, Tang Y J, Wu Y, Miao S J. Sci. Total Environ., 2025, 974: 179168.

[58]
Zhang F, Liu F B, Ma X, Guo G Z, Liu B, Cheng T H, Liang T, Tao W L, Chen X P, Clean. Prod., 2021, 317: 128449.

[59]
Cheng Z Z, Zhang Y, Muhammad W, Hao Y H, Tang Y J, Sun L L, Wang D, Zhou J T, Wang R Z, Deng Y G, Li Y J, Geng L, Han W. IEEE Trans. Ind. Electron., 2022, 69(1): 911.

[60]
Yang Y M, Lin S, Hu J H. Sens. Actuat. B Chem., 2023, 385: 133721.

[61]
Ma B, Karimi M S, Mohammed K S, Shahzadi I, Dai J P. J. Clean. Prod., 2024, 450: 141801.

[62]
Liang Z, Hermansen C, Weber P L, Pesch C, Greve M H, de Jonge L W, Mäenpää M, Leifeld J, Elsgaard L. Commun. Earth Environ., 2024, 5: 286.

[63]
Qu F D, Zhang S D, Huang C Z, Guo X Y, Zhu Y, Thomas T, Guo H C, Attfield J P, Yang M H. Angew. Chem. Int. Ed., 2021, 60(12): 6561.

[64]
Mahdavi H, Rahbarpour S, Hosseini-Golgoo S M, Jamaati H. Sens. Actuat. B Chem., 2021, 331: 129091.

[65]
Vergote T L I, Bodé S, De Dobbelaere A E J, Buysse J, Meers E, Volcke E I P. Biosyst. Eng., 2020, 196: 159.

[66]
Tu X S, Wang J, Liu X Y, Liu Y, Zhang Y H, Uwiragiye Y, Elrys A S, Zhang J B, Cai Z C, Cheng Y, Müller C. Environ. Sci. Technol., 2024, 58(14): 6215.

[67]
Peyron M, Bertora C, Pelissetti S, Said-Pullicino D, Celi L, Miniotti E, Romani M, Sacco D. Agric. Ecosyst. Environ., 2016, 232: 17.

[68]
Song Z, Yan J. Chem. Mater., 2023, 35(18): 7750.

[69]
Wang X, Xu X, Zhou T, Zhang T. ACS Sens., 2025, 10(4): 2569.

[70]
Yan J, Kang Y, Fang W H, Zhu B C, Song Z L. ACS Sens., 2025, 10(8): 6084.

[71]
Haldar T, Shiu J W, Yang R X, Wang W Q, Wu H T, Mao H I, Chen C W, Yu C H. ACS Sens., 2024, 9(11): 5856.

[72]
Shwetha H R, Sharath S M, Guruprasad B, Rudraswamy S B. Micro Nano Eng., 2022, 16: 100156.

[73]
Huang X W, Sun W, Li Z H, Shi J Y, Zhang N, Zhang Y, Zhai X D, Hu X T, Zou X B. Food Chem., 2022, 396: 133654.

[74]
Han Z, Ahmad W, Rong Y N, Chen X Y, Zhao S G, Yu J H, Zheng P F, Huang C C, Li H H. Foods, 2024, 13(11): 1721.

[75]
Baraka K, Ezzahar J, Madiafi M, Erraji H, Vivaldi G A, Tallou A. Front. Sustain. Food Syst., 2025, 9: 1542166.

[76]
Jiang D W, Shen Z S, Zheng Q S, Zhang T H, Xiang W, Jin J. IEEE Internet Things Mag., 2025, 8(5): 72.

[77]
Mao H P, Du X X, Yan Y T, Zhang X D, Ma G X, Wang Y F, Liu Y, Wang B, Yang X Y, Shi Q. Int. J. Agric. Biol. Eng., 2022, 15(6): 180.

[78]
Kang C C, Mu X Y, Novaski Seffrin A, Di Gioia F, He L. Comput. Electron. Agric., 2025, 230: 109866.

[79]
Lu W, Xu X, Huang G, Li B, Wu Y, Zhao N, Yu F R. IEEE Trans. Ind. Inform., 2021, 17(6): 4335.

[80]
Balan T, Dumitru C, Dudnik G, Alessi E, Lesecq S, Correvon M, Passaniti F, Licciardello A. Sensors, 2020, 20(15): 4127.

[81]
Saraswathi R V, Rahul L, Lokeswar S, Vathsav Varma R, Ajay U. In 2025 6th International Conference on Mobile Computing and Sustainable Informatics, ICMCSI 2025. Eds.: IEEE. Piscataway: IEEE, 2025, 246-251.

[82]
Suriano D. In The 11th International Electronic Conference on Sensors and Applications, ECSA-11. Eds.: Laheurte J M. Basel: MDPI, 2024: 74.

Options
Outlines

/

〈 〉