Advancements in Chemiresistive and Electrochemical Sensing Materials for Detecting Volatile Organic Compounds in Potato and Tomato Plants

Abstract

Tomatoes (Solanum lycopersicum) and potatoes (Solanum tuberosum) are vital staple crops. They are prone to diseases from pathogens like Ralstonia and Fusarium, which cause significant agricultural losses. Detecting volatile organic compounds (VOCs) emitted by plants under stress offers a promising approach for advanced monitoring of crop health. This study examines sensing materials for wearable plant sensors targeting VOCs as biomarkers under abiotic and biotic stress. Key questions addressed include the specific VOC emission profiles of potato and tomato cultivars, how materials and sensing mechanisms influence sensor performance, and material considerations for agricultural use. The analysis reveals cultivar-specific VOC profiles under stress, challenging the identification of universal biomarkers for specific diseases. Through a literature review, this study reviews VOC responses to fungi, bacteria, and viruses, and compares non-composite and hybrid chemiresistive and electrochemical sensors based on sensitivity, selectivity, detection limits, response time, robustness, cost-effectiveness, and biocompatibility. A superstructure bridging materials science, plant pathology, AI, data science, and manufacturing is proposed, emphasizing three strategies: sensitivity, flexibility, and sustainability. This study identifies recent research trends that involve developing biodegradable wearable sensors for precision agriculture, leveraging flexible biocompatible materials, multi-parameter monitoring, self-healing properties, 3D-printed designs, advanced nanomaterials, and energy-harvesting technologies.

1. Introduction

Climate change has been significantly affecting crop production and yield worldwide, and agricultural systems to sustain food security can mitigate its long-term effects. Crop health monitoring is pivotal to assessing plant physiological and morphological changes. Numerous technologies have been developed for monitoring plants using unmanned aerial vehicles (UAVs), utilizing vegetation indices such as the Normalized Vegetation Index (NDVI) to monitor vegetation health conditions [1], imaging techniques such as hyperspectral [2], RGB, and thermal [3] for crop monitoring and identifying plant stresses, and artificial intelligence (AI) for analyzing agronomic and environmental data acquired from these technologies to distinguish patterns and trace anomalies [3]. While traditional crop monitoring methods rely heavily on visual inspection and field sampling, which is time-consuming, labor-intensive, and subjective, crop conditions are often assessed only when visible symptoms and damages have already surfaced. Hence, the early detection of infection biomarkers such as volatile organic compounds (VOCs) can contribute insights into addressing potential problems in plants.

Plants release a wide array of compositionally diverse volatile organic compounds through the leaf stomata and flowers, such as isoprene, monoterpenes, and their oxygenated derivatives [4,5]. Plant VOCs enable intra- and interspecific plant communications and enable more timely and earlier diagnosis of plant stress and disease, which can promote sustainability in agriculture [6]. VOCs emitted by plants are indicators of plants’ vulnerability, mainly to pathogens and to abiotic stressors. Moreover, aside from early detection of stress and disease, they serve a role in defense signaling and priming when attacked by insects or even pathogens [4,7]. Plants can also detect VOCs in neighboring plants, which can prime and enhance their defense responses [8]. VOC emissions occur naturally, even in the absence of stress. Unstressed plants continually unleash VOCs as an integral part of their physiological functions, some of which include gas exchange processes (net photosynthesis and transpiration) and primary and secondary metabolism, and VOCs can serve as an indicator of plants’ chlorophyll content [9]. In contrast, VOC emissions differ in response to biotic and abiotic stresses, where emission profiles change, signaling immediate intervention [10]. Some of the many VOCs, such as green leaf volatiles (GLVs) and sesquiterpenes, are released by plants due to mechanical damage brought about by wilted leaves due to herbivory and pathogenic attacks such as fungal pathogens [11,12] and in response to abiotic stresses such as drought, cold stress, and temperature variability [11]. Many staple crops, such as rice, corn, and wheat from the Poaceae (grass) family and beans and soybeans from the Fabaceae (legume) family, are important because of their nutritional importance and contribution to food security. These crops are also susceptible to both biotic and abiotic stresses but are proven to be climate-resilient and adaptive, which supports sustainable agriculture [13]. While the Solanaceae family is equally important due to its diverse applications in agriculture, medicine, and nutrition, Solanaceae crops are more vulnerable to climate alternations and fungal and bacterial diseases, implying less resilience compared with the other families [14]. Hence, monitoring and detecting volatile organic compounds can provide crucial information about plant conditions for early intervention and management, especially in staple crops in the Solanaceae family, because of their economic importance and contribution to food security.

Solanaceae, commonly known as the nightshade family, has a worldwide distribution that includes several crops that are considered staple foods. One of these crops is tomato (Solanum lycopersicum L.), which is the second most important crop next to potato (Solanum tuberosum L.) [15]. Potatoes, which grow below ground, are the fourth most important food, following rice, wheat, and corn, and they are vital due to their high yield, distribution, and nutritional value [16]. Potato is also foreseen as one of the solutions to future food crises [16]. In contrast, tomatoes, which grow above ground, account for an annual global production of 160 million tons, making them crucial for agricultural development [17]. However, climate change exacerbates yield losses and the decline of tomatoes and potatoes, escalating fungal vulnerability and increasing pathogenic attacks [18]. These pathogens do not just affect key staple crops in the field early; they also pose a threat during harvest and post-harvest periods, leading to economic losses and agricultural decline [19]. Although tomatoes and potatoes are among the crops with high economic value, there appears to be limited recent VOC profiling for disease detection and monitoring.

Both tomato and potato plants are affected by pathogens such as Phytophthora sp. Potato Virus Y and Fusarium sp. [10,20,21]. Among the most important and quarantine diseases are Ralstonia solanacearum and Fusarium oxysporum inciting vascular wilt on host plants. These pathogens are soil-borne with diverse host ranges causing heavy yield losses of up to 80–90%, making them relatively hard to manage [22,23]. In some cases, tomato fusarium wilt infection is invisible to the naked eye due to latent infection [24]. Also, Ralstonia-infected potato tubers may not be observable to the naked eye but will manifest when used as planting material [25]. Monitoring these crops through non-invasive technologies such as wearable sensors can help mitigate the impending damage caused by aggravating pathogenic attacks.

Precision agriculture (PA) leverages advanced technologies such as sensors, drones, the Internet of Things (IoT), global positioning system (GPS) devices, and geographic information system (GIS) approaches to improve crop yields, agricultural productivity, and sustainability. Among the emerging technologies revolutionizing PA are wearable sensors. Wearable sensor technology has been widely utilized and developed for real-time and non-invasive monitoring of plants’ phytometric activity, growth, and overall health; these sensors can be attached to the leaves or stems of the plant to identify underlying diseases and conditions that may affect crop yield and quality [26,27,28], providing crucial information to combat and prevent the massive effects of biotic and abiotic stressors over time. One of the recent developments in wearable sensors is plant tattoo wearable sensors deployed in plant leaves to monitor leaf relative water content, surface temperature, and bioelectric potential. These parameters are crucial to understanding plant responses under recurring stresses and establishing agricultural management [27]. Another development is the design of a flexible wearable sensor to obtain information about plant growth and leaf transpiration or vapor pressure deficit (VPD), an indicator of air dryness that serves as a driving force of transpiration [29]. Several polymers, such as polyaniline, have also been developed for VOC detection [30]. However, some wearable sensors face several key challenges upon deployment, such as biocompatibility issues that hinder optimal plant growth over time [31], insufficient power to bolster the sensor for extended operations (since traditional power supplies such as batteries have limited lifespan) [32], adhesion problems due to the requirement for additional adhesive to adhere to the plant surface [33], and stability and flexibility issues [34]. Hence, the synergy of composites such as metal oxides [35] and even polymers [36] with base materials has been continuously studied to overcome these limitations. Nevertheless, the integration of wearable sensor technology in agriculture offers a non-invasive method for detecting VOC, reducing the frequent use of gas chromatography–mass spectrometry (GC-MS), which requires complex gas sampling processes, labor-intensive protocols, and expensive instruments that leave it unsuitable for field application [29].

Literature Survey Strategy and Analysis

The literature review was conducted through searching through databases such as Google Scholar and Science Direct with the use of different keywords with “Volatile Organic Compounds”, “Potato”, “Pathogen”, “Environmental factors”, “Disease detection”, “Tomato”, “VOC detection techniques”, and “Gas chromatography-Mass Spectrometry”. Studies that explicitly discussed VOC emission profiles were used and some studies were selected based on their conclusions, which may provide useful insights for future studies on the topic. VOC emission profiles affected by different factors are the focus of Section 2.1, while Section 2.2 and Section 2.3 focus on studies that mention potato and tomato crops. Examples focusing on the VOC emission profiles of other crops were also not included in the study.

The keywords used for Section 3 and Section 4 included “electronic-nose”, “sensing materials”, “VOC sensors”, “agriculture”, “biosensors”, “nanomaterials”, “metal oxide semiconductors”, “conducting polymers”, and “carbon nanomaterials”. The literature discussing invasive sensor designs, quality control, and health monitoring sensors was excluded. Keywords such as “Volatile Organic Compounds”, “Wearable Sensors”, “Artificial Intelligence”, “Internet of Things”, “Challenges”, “biodegradable”, “sustainable”, and “Agriculture” were used for the literature review that contributed to the findings in Section 5. Studies that discussed trends and in plant sensor use for agricultural processes were utilized with a focus on the integration of technology such as artificial intelligence and the Internet of Things. Studies within the last ten years were prioritized as “2015–2025”; however the primary range for studies was 2003–2025.

The bibliometric networks of scientific and technical keywords for sensing materials relating to plant diseases and pathogens from 2003 to 2024 are shown in Figure 1. Four themes were generated based on the map, namely, volatile organic compounds, sensor technology, substrates, and polymers (red); plant leaf health diagnosis, non-destructive wearable sensors and electronic noses (green); gas sensing, detection and nanomaterials (yellow); and plant phenotyping, organic compounds, and environmental monitoring (violet) (Figure 1). “Volatile organic compound” is the main keyword that exhibits thicker links or weights connecting all clusters, which indicates important emerging aspects of VOC measurement in relation to materials science for plant pathogen detection and diagnosis. Sustainability was not identified as one of the keywords for this study, with the different sensor designs not subjected to life cycle assessment (LCA) to identify their overall environmental impact.

The largest nodes corresponded to VOCs and the respective detection technology focusing on sensors and gas chromatography, indicating increasing interest towards the use of precision agriculture technology such as sensors for VOC detection. There is a need to identify clear points of direction due to the variety of specific sensing materials used with the minimal number of VOCs (acetone and benzene). Thus, the goal of this study is to answer the following research questions in relation to the advancements in sensing materials for detecting volatile organic compounds in potato and tomato plants.

• What are the specific VOCs emitted by potato and tomato plants under different stress conditions and how do their emission profiles vary between biotic and abiotic stressors?
• What recent advancements in sensing materials have improved the detection of VOCs in agricultural applications and how do different sensing mechanisms influence the performance of VOC sensors?
• How do different chemiresistive and electrochemical non-composite and hybrid sensing materials compare in terms of sensitivity, selectivity, limit of detection, response time, robustness, cost-effectiveness, biocompatibility, substrate, modifications, sensing mechanism, and detection resolution for VOC detection in agriculture? What criteria make a sensing material ideal for wearable plant sensors, particularly for monitoring VOCs in potato and tomato crops?
• What interdisciplinary opportunities exist between materials science, plant pathology, artificial intelligence, data science, and industrial manufacturing to enhance wearable plant sensor technologies?

This study mainly contributes to the following:

• Elucidation and analysis of existing VOC emission profiles of different potato and tomato cultivars focusing on VOCs emitted due to pathogen-induced stress with corresponding insights for conducting future VOC emission analysis and studies.
• Comparative analysis of sensing materials following various performance metrics and recent advances that could offer valuable insights for selecting suitable materials for future development of sensors for agricultural applications.
• Development of a superstructure pinpointing the roles of various disciplines in advancing wearable plant sensor technologies that could facilitate a strategic approach for agricultural stakeholders. This highlights the need for interdisciplinary collaborations, with emphasis on the integration of artificial intelligence, the Internet of Things, and smart sensing technologies for data-driven agriculture.

This study involves three primary discussion sections, focusing on VOCs in potato and tomato crops (Section 2), advances in sensing materials for VOC detection (Section 3), and comparative analysis of sensing materials for VOC detection (Section 4). Key points of the discussion are outlined in Figure 2. In-depth discussion of different sensor designs is not included within the scope of this review. Section 5 synthesizes the preceding three sections and describes a framework through which advanced technologies and interdisciplinary actions could improve the sensing capabilities of wearable sensors in relation to plant diseases.

2. Volatile Organic Compounds in Potato and Tomato Crops

2.1. Plant Volatile Organic Compounds

Volatile organic compounds (VOCs) are present in both underground and aboveground plant parts, including the tubers and leaves [20,37]. Plant VOCs released from fruits and other parts are responsible for the flavor and aroma of agricultural produce, while serving a role indicating health conditions [4]. While biocontrol agents respond to pathogens through detection of volatile compound emission, plants also activate certain genes upon recognition of infection to initiate defense responses associated with emission of VOCs; for instance, the tomato gene MTS1 is responsible for the synthesis of terpenoids which are associated with its disease resistance [37]. For this reason, profiling of plant VOCs has become a subject of research to aid in the prevention, early detection, monitoring, and management of plant diseases as part of integrated disease management.

Plants emit VOCs that vary in terms of amount and type, depending on biotic, abiotic, and genetic factors, and their stage of development (Figure 3). The plant species itself and differences in composition of even a single gene contribute to variations in VOC emissions. Plants within the same family, such as Solanaceae, and even within the same genus have differences in their VOC emission [38]. In fact, isogenic plant VOCs can change with the insertion of even one gene, and some may be cultivar-specific for a plant, serving to confirm its identity compared with another cultivar [10,37]. Emission profiles also show that VOCs vary across different plant tissues and developmental stages within the same plant [12]. Apart from those plant factors, VOCs are highly influenced by biotic and abiotic factors. Research has demonstrated increase in either the amounts of the same VOC or in the number of VOCs in pathogen-infected plants compared with their uninfected counterparts [12]. This is also true for insect damage and mechanical damage caused by factors other than living pests [33,39,40]. VOCs also show changes as infection progress within time [5,41].

As with pathogens, plants’ responses to insect pest attack induce emissions of green leaf volatiles with multiple functions, that send warnings to neighboring plants when released in the air [8,42]. Among the abiotic factors influencing plant VOCs are temperature, light intensity, pH, fertilization, and external air pollutants. Temperature factors such as, for instance, a rise of up to 30 °C induced increased volatiles in potatoes that further varied in combination with insect feeding [42]. Increased temperature upregulates the expression of certain genes associated with pathways responsible for terpene precursors. Along with this, high temperature is also linked to increased biosynthetic activities, thus affecting VOC diffusion rates. In leaves, temperature affects the stomatal aperture and conductance, thus influencing the rate and quantity of VOC diffusion [42]. Light or darkness and pH have also been demonstrated to influence VOC synthesis. For instance, the synthesis of Z-3-hexanol in maize leaves is optimum at 5–6 pH and decreases with increasing pH. This response to pH is hypothesized to be related to why green leaf volatiles change their emission profiles in the dark, due to simultaneous outflow of H⁺, thus altering the pH within the cells [11]. Soil amendments such as urea also affect VOC profiles, as shown by cucumber fruits. This is hypothesized to be due to the fertilizer’s effect on the plants’ physiological processes [43]. Above ground, atmospheric gasses such as carbon dioxide and tropospheric ozone influence plant VOC profiles to some extent [38,39]. Increased concentrations beyond ambient conditions result in varying amounts of VOC. However, it appears that this has minimal effect on the VOC profile, which rather is highly influenced by other factors. For instance, healthy rice plants exposed to increased carbon dioxide did not show significant difference from those that were exposed to an ambient amount of carbon dioxide but showed variation when exposed to brown leafhopper infestation [39]. Regarding the effects of tropospheric ozone on four Brassica spp., certain VOCs showed variation in their responses to varying levels of ozone; at the same time, each of the species showed varying degrees of sensitivity to the treatments [38]. In conclusion, these factors, genetic and environmental, act singly and interactively to influence the VOC emission profiles of plants.

2.2. Tomato VOCs

Fusarium wilt of tomato, caused by Fusarium spp., is one of the problems in tomato growing areas in the Philippines and in other countries such as Jordan, which is among the world’s top tomato exporters [24]. The pathogen enters the roots, multiplies, and blocks the vascular system, thus causing typical wilting symptoms. At some point, host plants may develop latent infection and will manifest only under favorable conditions [24]. Fusarium spp. naturally emits VOCs even without infection. Several studies have reported hundreds of VOCs emitted from host plants infected with Fusarium spp. [5,42,44,45]. Reported VOCs induced by pathogens can be largely grouped into terpenoids, benzenoids, fatty acid derivatives, and amino acid derivatives, based on their biosynthetic pathways [4,5]. Some of these have been identified as potential markers for detection of infection through differences in the released amount compared to uninfected counterpart or the consistent detection of a unique VOC in the infected host. For instance, methyl propyl sulfide and β-phellandrene are potential biomarkers of Fusarium oxysporum f. sp. ceapeae infection in Allium cepa due to their relatively increased emissions compared with uninfected plants [5].

Potential biomarkers for a certain organism-to-plant interaction may need careful profiling since most VOCs are not specific to a certain organism or infection. In Solanum lycopersicum alone, different VOCs were abundantly emitted in response to different biotic factors even though they were all sourced from the same plant part (

Table 1. Comparison of VOCs from tomato detected through gas chromatography–mass spectrometry.

Tomato Cultivar	Pathogen	VOC	VOC Source	Stressor Type	Reference
Momor, UC82 & UC82 grafted on Manduria Rio Grande	F. oxysporum f. sp. lycopersici, Potato Virus Y, Pseudomonas syringae	Methyl salicylate	Leaves	Fungus, virus, bacterium	[4,21,37]
Momor	F. oxysporum f. sp. lycopersici	Ethyl salicylate, 2-nonenal, 2-hexanone, 2-heptanone, 3-heptanone, 3-buten-2-one, 2-pentanone, butanal, 2-butanone	Leaves	Fungus	[37]
Rio Grande	Pseudomonas syringae	Sesquiterpene (unidentified), Salicyl aldehyde, α-pinene, α-phellandrene, β-phellandrene, limonene, Isoprenoid chloride	Leaves	Bacterium	[4]
UC82 and UC82 grafted on Manduria	Potato Virus Y	2-ethyl-Furan, 3-Pentanol, 1-Penten-3-ol, α-Terpinene, (E)-2-Hexenal, 2-pentyl-Furan, Cymene, 4-Hexen-1-yl acetate, (Z)-2-Penten-1-ol, 6-methyl-5-Hepten-2-one, 1-Hexanol, (E)-3-Hexen-1-ol, (E)-2-Hexen-1-ol, p-Cymenene, 2-ethyl-1-Hexanol, Methyl nonanoate, Linalool, Terpene 4	Leaves	Virus	[21]

). VOCs may also coincide across different categories of plant-to-pathogen interaction. For instance, α-pinene, β-phellandrene, and limonene are relatively high in tomato inoculated with Pseudomonas syringae and have also been detected in potato inoculated with Phytophthora infestans [4,10]. Another VOC, 1-hexanol, was detected from tomato inoculated with Potato Virus Y and P. infestans, in separate studies [10,21]. Interestingly, 1-hexanol was detected from seven strains of the pathogen F. oxysporum [44]. In some cases, VOC may serve as a general indicator of a stressed plant. For instance, methyl salicylate, a VOC reported at high emission levels from tomato infected with Potato Virus Y, F. oxysporum f. sp. Lycopersici, and P. syringae in separate studies, is quite a common response to infection [4,21,37]. On the other hand, some VOCs are emitted only during a certain time of infection. The early stage of susceptible tomato infection with F. oxysporum f. sp. lycopersici was associated with increased emission of methyl salicylate and marked with the exclusive emission of 3-buten-2-one, butanal, and 2-butanone. Ethyl salicylate, on the other hand, was over-emitted during the second week, and the infection ended with exclusive emission of 2-pentanone. Meanwhile, 2-nonenal, 2-hexanone, 2-heptanone, and 3-heptanone were present throughout the infection process [37]. Table 1 lists tomato VOCs with relatively increased emissions induced by infection of representative pathogens from viral, bacterial, and fungal pathogens.

2.3. Potato Volatile Organic Compounds

Some of the common diseases that potato (Solanum tuberosum) is afflicted with include late blight (caused by the fungus Phytophthora infestans), fusarium wilt, dry rot (caused by multiple fungi such as Fusarium ambucaine, F. solani, F. graminearum, F. oxysporum), bacterial wilt (caused by the bacterium Ralstonia solanacearum), and soft rot (caused by the bacteria (Pectobacterium, Pectobacterium brasiliense, Dickeya, R. solanacearum) [46]. In the case of diseases caused by the fungus Fusarium, such as wilt and dry rot, Fusarium generally enters from the root tips and grows in the xylem where it takes up resources such as nutrients and water, eventually leading to wilting [46].

Ralstonia solanacearum, on the other hand, enters through plant wounds and through stomata. It also grows in the plant’s xylem and after it matures, it leaves through the plant roots to transfer to other uninfected plants through water flow. It can affect a wide variety of crops such as potatoes, tomatoes, bananas, tobacco, and more than a hundred other types of plant species [25]. The traditional Ralstonia disease control procedure involves the inspection of crops through visual inspections, testing of possible infected crops, and then quarantining and addressing the infected area [25]. Traditional methods for Ralstonia disease prevention include crop rotation for bacterial control, the use of fertilizers to modify the pH level of the soil, and the use of specific types of crop seeds that are more resistant to disease, with efforts being made to move towards more biologically based control methods [25]. The visible manifestations of wilting caused by Ralstonia solanacearum in potatoes start with the wilting of leaves followed by brown discoloration along the stems eventually leading to bronze discoloration of the leaves, until the plant’s death [25].

The introduction of pathogens such as R. solanacearum and F. sambucinum into potato crops can change the VOC emission profiles of potato plants and is thus a useful tool for the early identification and prevention of crop-related diseases [20,47].

Table 2. Comparison of VOCs from potatoes detected through gas chromatography–mass spectrometry.

Potato Cultivar	Pathogen	VOC	VOC Source	Stressor Type	Reference
Kufry Pukhraj	F. sambucinum	Naphthalene 1-butanol-3-methyl 2-undecanone γ-muurolene	Tubers	Fungus	[20]
Maris Peer, Marfona, Estima, Pentland, Dell, Wilja	R. solanacearum	2-propanone 2-propanol 2-butanol 2-pentanol	Tubers	Bacteria	[47]
1681-11	No pathogen	Δ3-carene α-terpinene	Leaves	None	[10]
Reet	No pathogen	α-pinene
Ando	No pathogen	acetaldehyde
Anti	No pathogen	α-muurolene
Sarme	No pathogen	β-bourbonene
Kuras	No pathogen	β -pinene
Alouette	No pathogen	α-bergamotene
Jogeva Kollane	No pathogen	6-methyl-5-hepten-2-one
Teele	No pathogen	limonene

compares VOCs detected from potatoes through a gas chromatography–mass spectrometry headspace analysis, and their related pathogens. The table does not list the complete VOC emission profile due to the sheer number of detected VOCs. The VOCs were detected around the plant headspace, although the general focus varied, ranging from the plant headspace to more targeted parts of the plant such as the leaves or the potato tuber itself.

Solid-phase microextraction (SPME) combined with gas chromatography and mass spectrometry (GC-MS) was utilized for the early identification of soft rot disease, through identifying VOCs in different sample potato discs extracted from tubers that had been inoculated with five separate strains of bacteria [41]. There were five species of bacteria utilized, the first three of which were Pectobacterium carotovorum subsp. brasilience, P. carotovorum subsp. carotovorum, and P. carotovorum subsp. Parmentieri while the second batch of bacterial strains were Dickeya solani and D. dianthicola; the key VOC biomarkers included dimethyl ether, methyl cyclopentane, (R,R)-2,3 butanediol, 2-pentyl thiophene, (Z)-3-Nonen-1-ol, 3-octanone, 2-pentanone, and 1-hexanol [41]. The biomarkers tied to the Pectobacterium strains were then identified, with the relevant VOCs being 2-methylbutanol, 2-methypropanol, 1-penten-3-ol, (Z)-2-decanal, fenchol, and 2-undecanone [41]. However, the disease detection in this case would happen at the tuber stage; so, it would mainly occur in the harvesting, storage, and transport stages rather than in the growth stage.

The emission blends from leaves of nine uninoculated potato cultivars with varying levels of resistance to Phytophthora infestans (a pathogen linked to late blight disease) included forty-six identified VOCs, of which the majority were sesquiterpenes such as (E)-β caryophyllene, valencene, α-humulene, and δ-cadinene and monoterpenes such as camphene, limonene, and tricyclene [10]. Among the nine types of potato cultivars used, the VOCs present as well as their concentration levels differed greatly, such as in the cases of the 1681-11 cultivar where only 6 of the 23 types of sesquiterpenes were identified, and the Reet cultivar, which was the only cultivar in which the VOC α-terpinene was identified [10]. The sheer variety across the VOC emission profiles of different cultivars within a single plant species emphasizes the need to conduct an actual VOC detection experiment on a specific potato cultivar to create an accurate VOC emission profile.

Another important point to note in VOC detection is that some VOCs appear only at certain points after inoculation, such as in the case of 1-butanol, 3-methyl which was only detected on the day of inoculation with F. sambucinum, whereas VOCs such as γ-muurolene and α-isomethyl ionone were detected both 2 and 6 days after inoculation [20].

2.4. VOC Detection Techniques

Traditional VOC detection techniques involve the use of gas chromatography and mass spectrometry (GC-MS), proton transfer reaction mass spectrometry (PTR-MS), gravimetric methods based on mass sensors, electrical methods that consider different electrical reactions between the sensor design, material configuration, and VOC vapor, and optical methods that rely on the analysis of the intensity of radiation [48,49].

Table 1 and Table 2 both utilize GC-MS as it is the traditional and most reliable method for the identification of volatile materials and is used for purposes such as research and development and quality control across a wide array of fields, such as in medicine, fuel development, biotechnology, and environmental analysis [50].

GC-MS is highly accurate in VOC detection due to its high sensitivity and reproducibility, quick analysis times, and ability to determine the composition of unknown organic compounds [50]. There is also less concern about extraneous factors such as humidity and temperature affecting the analysis of VOCs due to the nature of the headspace analysis process wherein the sample is placed in an enclosed space to prevent external influences. This is reflected in the studies by Ray et al. [20] and Stinson et al. [47] where the potato samples to be analyzed were sealed in glass containers and maintained under the desired conditions.

While it does have many advantages, GC-MS does come with its own set of drawbacks due to its high investment costs, the need for trained personnel, lack of real-time analysis capabilities, equipment and power consumption, and limitations in analyzing the vapor pressure of compounds below 10⁻¹⁰ torr [41,48,49,50]. There are two primary extraction methods for GC-MS, namely, solid phase-microextraction (SPME) and supercritical fluid extraction (SFE) [48]. SPME utilizes the concept of adsorption of VOCs through the use of fibers, whereas the SFE technique is more capable of detecting specific VOCs at the expense of being less efficient in detecting large quantities of VOCs [48].

The PTR-MS method offers several advantages for real-time monitoring of volatile organic compounds (VOCs) such as rapid sampling, elimination of the need for sample preparation, and high sensitivity [48]. Despite this, the PTR-MS method struggles to distinguish between isomeric compounds and it cannot identify compounds with a proton affinity higher than that of water [51].

Another technique for VOC detection is the field asymmetric ion mobility spectrometry (FAIMS) method, which can be used to identify the presence of volatile organic compounds, focusing more on the mobility of ions rather than chemical composition [52]. This method of VOC analysis, while effective in the early detection of soft rot in potato tubers, is only capable of detecting the presence of VOCs and is unable to identify specific VOCs due to its focus on ion mobility [52]. Gas chromatography–ion mobility spectrometry (GC-IMS) is another method for VOC detection, which combines the two methods and leads to some benefits such as the shortening of analysis time and improvement in detection performance [52,53]. There have been efforts made to implement GC-IMS for agricultural purposes, ranging from the early identification of diseases such as soft rot to the monitoring and management of agricultural food products [53].

Photoionization detectors (PIDs) are also capable of detecting VOCs at different concentrations, through the use of photon energies from UV lamps to ionize VOCs; different types of VOCs provide different signal responses for detection [54,55,56]. PIDs can be utilized alongside gas chromatography techniques (GC-PID) to aid in the identification of VOCs, with low-cost PID sensors becoming more commercially available and representing a viable possibility as real-time, portable VOC detectors for purposes such as chemical hazard detection and compound identification in agricultural products [56].

It should be noted that these detection methods are not mutually exclusive, with different studies exploring the use of methods such as PTR-MS and FAIMS as complementary detection methods to GC-MS that may help make up for some of the limitations of GC-MS, such as its focus on qualitative analysis [51,57].

Table 3. Advantages and challenges in different VOC detection techniques.

VOC Detection Technique	Advantages	Limitations	Reference
GC-MS	High accuracy; high sensitivity; high reproducibility; able to determine composition of unknown organic compounds;	High investment costs; need for specialized personnel; lack of real-time analysis capabilities; requires a large quantity of VOC samples; equipment and power consumption; vapor pressure; less efficient in quantitative analysis	[41,48,49,50,51]
PTR-MS	Quick sampling time; high sensitivity; low detection limit; field portable; capable of real-time analysis;	Cannot identify certain compounds (isomeric, etc.); limited to compounds with a certain level of proton affinity	[48,51]
FAIMS	field portable; customizable; lower costs; suitable for agricultural products; capable of real-time analysis	Cannot identify specific compounds; relatively new technology with limited sampling techniques;	[52,57]
GC-IMS	High sensitivity; relatively high accuracy; cost-efficient; automatable	poor resolution; limited dynamic range; limited ability to identify unknown compounds; limited compound database;	[53,54]
GC-PID	Portable; capable of real-time analysis	low selectivity; ionization energy may not match that of target VOC;	[55,56]

describes the notable advantages and limitations of the different techniques previously discussed. The gravimetric, electrical, and optical methods previously discussed cover a wide range of sensors with their own individual advantages and limitations [49]. The materials used for the development of these sensors also serve as a key factor in their performance, making it important to consider advances in research as the use of more effective materials can serve to improve the overall efficiency of the sensors.

3. Advances in Sensing Materials for VOC Detection

Many sensors continue to be useful in the field of agriculture, such as smart soil sensors, which are just one of the many sensors that continue to revolutionize agriculture by providing real-time monitoring of crucial soil parameters such as temperature, pH levels, moisture, salinity, NPK (nitrogen–phosphorus–potassium) levels, and electrical conductivity (EC). These sensors have been utilized to optimize rice management and determine how the rice crop metabolism was influenced [58]. Low-cost plant-based sensors have been used for optical sensing to monitor physiological responses of plants synchronized in real time, based on seasonal changes in leaf color [59]. Also, environmental sensors such as temperature sensors, relative humidity sensors, and CO₂ sensors have been used to assess water stress in crops and identify prevailing diseases and plant anomalies [60,61]. However, principal concerns that need to be addressed in relation to sensor technology include sensitivity and stability, with some sensors suffering from low sensitivity and signal stability, which affects their accuracy and reliability [34]. Environmental durability is also an issue, with some sensors struggling to withstand harsh environmental conditions, such as extreme temperatures and humidity [34].

Electric nose (e-nose) devices have garnered popularity as a sensing technique offering non-invasive and inexpensive methods for detecting VOCs. Consisting of different gas sensors with varying sensitivities, utilizing adsorption or binding to molecules with diverse functional groups, e-noses collect broad responses, otherwise known as fingerprint data, which can be computed and processed with techniques such as artificial intelligence (AI) [62]. VOC sensing materials for e-noses primarily use chemiresistive, potentiometric, and amperometric sensors to measure responses. Despite advances in chemiresistive VOC detection techniques, there is a notable gap in research specifically focusing on electrochemical gas sensors for agricultural applications. This could be due to the complexity of electrochemical systems or their specificity towards agricultural VOCs. However, electrochemical sensors offer advantages including high sensitivity and selectivity, which are crucial for early detection of disease. The sensing materials for VOC detection can be classified into three groups based on their material composition, namely, polymeric, carbon nanomaterials, and metal oxide semiconductors [63]. Electrochemical sensors as described in various publications and used in the field utilize hybrid composites of these materials to target specific analytes or compounds.

Figure 4 and

Table 4. Comparison of chemiresistive and electrochemical VOC-sensing material types.

Material Type	Sensitivity	Selectivity	LOD	Response Time	Material Cost	Biocompatibility	Reference
Conducting polymer	High	High (tunable by functionalization)	ppb to ppm	seconds	Low	Variable	[64,65]
Carbon nanotubes	Very High	High (dependent on functionalization)	sub-ppb	seconds	High	Variable	[64,65]
Graphene derivatives	High	High (dependent on functionalization)	sub-ppb	seconds	Moderate	Variable	[64,65]
Carbon dots	High	High	sub-ppb to ppm	seconds	Low	High	[64,65]
Metal oxide semiconductors	Very High	Variable	low ppm	seconds	Low	Low	[64,65]

summarize the comparison of the performance metrics of each material for VOC detection. The performance metrics are based on sensitivity, selectivity, limit of detection (LOD), response time, robustness, cost-effectiveness, and biocompatibility. Sensitivity refers to the ability of the sensing material to detect low concentrations of the target VOC; selectivity is the material’s capability to distinguish between VOCs; LOD is the lowest concentration of a VOC that the sensing material can detect; response time is how quickly the sensor can detect changes in VOC concentration; robustness is the ability of the material to maintain a consistent performance over various conditions and extended periods; and biocompatibility is the compatibility of the materials with the organism.

3.1. Polymeric Materials

Conducting polymers are a group of organic polymers that exhibit electrical conductivity due to their conjugated nature, allowing the delocalization of π-electrons which enable the movement of charge [66]. Materials such as polyaniline (PANI), polypyrrole (PPy), and polythiophene (PT), among many others, are good examples of conducting polymers utilized in biosensors, chemical sensors, and gas/VOC sensors. Moreover, chemical modifications such as doping and functionalization further increase conductivity and specificity towards specific target compounds. Doping is the process by which the conducting polymer is oxidized or reduced to add or remove charge carriers, modifying its electrical conductivity [67]; it can be classified as p-type or n-type doping. P-type doping involves oxidation by protonation, adding charge carriers to the conjugated π-electron system of the polymer, enhancing conductivity, whereas N-type doping involves reduction, introducing electrons to the polymer and decreasing conductivity. Further modifications include functionalization, wherein functional groups or nanomaterials are added to the polymer backbone to improve interaction between the polymer matrix and the target analyte. This process enhances binding, sensitivity, and selectivity towards the target analyte [68].

3.2. Carbon-Based Nanomaterials

Carbon, with four valence electrons, possess three different hybridization states, namely, sp, sp2, and sp3, bonding covalently with other carbon atoms and other non-metallic elements [69]. This allows matrices to assume various nanoallotropes based on the predominant type of covalent bonds linking C- atoms or their morphology. Primarily consisting of sp2 carbons arranged in a hexagonal network, some of these nanoallotropes exhibit excellent electrical conductivity, high surface-area-to-volume ratio, mechanical strength, chemical stability, and good optical properties [69,70]. Due to these properties, carbon-based nanomaterials have been used as an ideal material for sensing applications, utilized across a wide range of applications such as biomedicine, energy storage, electronics, composite materials, and agriculture.

Carbon nanotubes (CNTs) were among the first allotropes to be discovered and characterized, consisting of a hexagonal framework of sp2 hybridized carbon atoms. Described as a graphene sheet rolled up into a tube, CNTs possess excellent thermal, electric, elastic, and mechanical properties [71]. CNTs can be classified into two types, single-walled carbon nanotubes (SWCNTs) and multi-walled carbon nanotubes (MWCNTs) with diameters ranging from 0.4 to 10 nm and 4 to 100 nm, respectively. SWCNTs consist of a single layer of graphene, making them more pliable, while MWCNTs, on the other hand, may consist of multiple layers of graphene in a more complex and rigid structure [72]. Graphene is a single-atom-thick planar film of hexagonal carbon atoms, exhibiting high electrical conductivity, a large surface area, and mechanical flexibility [73]. These factors make graphene well suited for sensing applications. The development of functionalized graphene-based materials such as graphene oxides (GOs) and reduced graphene oxides (rGOs) further enhance this sensing potential, enhancing reactivity, solubility, and conductivity. Carbon dots (CDs) are spherical carbon nanoparticles primarily composed of sp3 hybridized carbons, with diameters ranging from 2 to 10 nm, exhibiting biocompatibility and relatively strong photoluminescence, making them suitable for sensing and imaging techniques [70,74]. Graphene quantum dots (GQDs) are nanoscale fragments of a graphene monolayer, composed mainly of sp2 hybridized carbons exhibiting crystalline morphology as well as oxygen-containing functional groups. They exhibit characteristics of both graphene and quantum dots, having good conductivity, photoluminescence, and fluorescence [70]. While most of these carbon-based nanomaterials are chemically inert, being difficult to disperse or process, modifications such as doping and functionalization increase matrix/solvent interactions, enhancing reactivity towards a specific target or analyte.

3.3. Metal Oxide Semiconductors

Metal oxide semiconductors (MOSs) are a group of chemiresistive gas sensors that function depending on the electrical conductivity of polycrystalline oxide semiconductors within the surrounding atmosphere [75]. These are valence compounds that encompass a high level of ionic bonding with negative oxygen ions; their band edges are composed mainly of metal ns orbital and oxygen 2p orbital. MOSs exhibit various morphologies that significantly influence their gas sensing performance, such as quantum dots (QDs) which are zero-dimensional semiconductor nanoparticles where the material properties are highly dependent on the particle size, allowing bandgap tuning [76], and 1D structures such as nanowires (NWs) with high surface-area-to-volume ratios and high crystallinity which make them highly effective for gas detection [77]. Nanofibers (NFs) also possess a high surface area that produces numerous adsorption sites [78]. Hierarchical nanostructures include nanorods, nanotubes, or nanosheets which yield more adsorption sites for gases [79,80].

Metal oxide semiconductors have unique optical properties, excellent chemical and thermal stability, tunable energy band gaps, and a very high dielectric constant [81,82], making them effective for gas sensing. Furthermore, MOSs can also be n-type doped, where electrons are the major current carriers leading to a decrease in conductivity upon interactions of semiconductor metal oxides and ionic oxygen species [83], or p-type doped, where holes are the major current carriers and increased conductivity is obtained [83]. Some metal oxide semiconductors, such as indium gallium zinc oxide (IGZO), zinc oxide (ZnO), and tin dioxide (SnO₂), are of n-type conductivity. In contrast, tin monoxide (SnO) and CuxO are of p-type conductivity. Through redox reactions between the sensing material and gas molecules, metal oxide semiconductors can be applied to detect target gases [83]. MOS was utilized to detect trace levels of volatile organic compounds in temperature-cycled operation, detecting formaldehyde, benzene, and naphthalene in ppb and sub-ppb concentrations [84]. In another study, a gas sensor detection system was developed based on 13 metal oxide semiconductors sensitive to aromas during Oolong tea production, detecting aldehydes, alcohols, and olefins during oxidation [85]. MOSs were also utilized to inspect VOCs emitted at room temperature for linalool detection in aging rice [86]. Meanwhile, a combination scheme of p-type and n-type MOSs was employed for the detection of benzene, toluene, ethylbenzene, xylene, and styrene (BTEXS), which are harmful gases [87]. However, despite the good performance metrics of MOS sensors, their low biocompatibility may hinder plant health over time. A possible approach to address this issue involves coating MOS sensors with biocompatible polymers or their integration with carbon-based composites to enhance biocompatibility. Nevertheless, metal oxide semiconductors for plant-wearable sensor technology can be explored further, and their sensitivity can be improved through adjustments to microstructure, catalysts, and heterojunctions.

4. Comparative Analysis of Sensing Materials for VOC Detection

To enhance the selectivity and sensitivity of VOC sensors towards a target analyte, various sensing materials and elements can be combined or used in tandem as a hybrid or composite material. The combination of sensing materials can compensate for the shortcomings of a single material, helping to overcome the limitations of single-material sensors by improving conductivity, increasing binding affinity for target analytes, and enhancing stability.

Table 5. Comparison of hybrid/composite chemiresistive and electrochemical VOC sensing materials.

VOC/Target Compound	Nanomaterial	Material Type	Modification	Sensing Mechanism	LOD	Environment	Stability	Pathogen	Application	Reference
Methanol	poly (ATD) (2-amino-1,3,4-thiadiazole) & PtNPs	Polymer and metal oxide-based	none	Current	<1 ppm	Field-tested	60 °C working temperature, 100% relative humidity	none	Agricultural	[88]
Ammonia	PANI and SnO2	Polymer and metal oxide-based	none	Resistance	≥1.8 ppm	Controlled	room temperature	none	Environmental	[89]
Ammonia	PANI and MWCNT	Polymer and carbon nanomaterial-based	none	Resistance	2 ppm	Controlled	room temperature	none	Environmental	[90]
SDE1 Biomarker	SWCNT and SiO2	Carbon and metal oxide-based	Functionalization with (3-aminopropyl) triethoxysilane &PBASE, Blocking with BSA & Tween-20	Resistance	<1 nm	Controlled	none	Citrus greening (huanglongbing)	Agricultural	[91]
VOCs from Aspergillus and Rhizopus	MWCNT	Carbon-based	Boron & Nitrogen Doping	Conductance	none	Controlled	28 °C working temp., 70% relative humidity	Aspergillus and Rhizopu	Agricultural	[92]
Acetone, Hexane, Methanol, Benzene, Diisopropanolamine	SWCNT and Metalloporphyrin	Carbon nanomaterial-based	Metalloporphyrin	Current	varying	None	none	none	Environmental	[93]
Various VOCs	rGO and AuNPs	Carbon nanomaterial and metal oxide-based	Functionalization with AuNps & thiourea	Resistance	varying	Field-tested	25 °C working temp., 50% relative humidity	Phytophora infestans	Agricultural	[94]
Ethanol	rGO and SnO2	Carbon nanomaterial and metal oxide- based	none	Resistance	5–500 ppm	Controlled	300 °C working temp., 98% relative humidity	none	Environmental	[95]
p-ethylguaiacol	SnO2 and TiO2 NPs	Metal oxide-based	none	Current	82 nM	Controlled	none	Phytophora cactorum	Agricultural	[96]
Ethanol and acetone	Mn-doped ZnO	Metal oxide-based	CdO modification	Resistance	5 ppm	Controlled	240 °C working temp.	none	Environmental	[97]
Plant virus (ssDNA)	AuNPs	Metal oxide-based	Thiolated ssDNA probes	Resistance	100 nM	Ambient conditions	none	Citrus tristeza	Agricultural	[98]

provides a detailed comparison of various studies that employ hybrid/composite materials for VOC detection, along with the specific composite, functionalization for VOC targeting, and the working environment of the sensor.

4.1. Conducting Polymer-Based Composites

Conducting polymers, which are often combined with metal oxides or carbon-based nanomaterials, have been widely studied for their role as sensing materials. A plant-wearable sensor was fabricated using a composite of the conducting polymer poly(2-amino-1,3,4-thiadiazole), (poly(ATD)), and platinum nanoparticles (PtNPs), producing a highly sensitive, robust, and portable methanol sensor for maize plants, with minimal effects from temperature and humidity [88]. A hybrid structure of PANI SnO₂ composite was synthesized to detect ammonia, which performed with high sensitivity and selectivity towards the target analyte [89]. Moreover, the material proved to be flexible and stable as a wearable sensor. A composite material of PANI and MWCNT was successful in detecting trace levels of ammonia and showed excellent response and recovery characteristics as a room temperature sensor [90].

4.2. Carbon Nanomaterial-Based Composites

Carbon-based nanomaterials such as SWCNTs, MWCNTs, graphene, and rGO have been integrated with various other materials and functional groups to enhance VOC sensing. SWCNTs were paired with SnO₂ along with functionalization and blocking mechanisms to specifically target the protein biomarker SDE1, which proved successful in the detection of the target analyte [91]; possible optimizations and improvements were suggested to enable the sensor to be used in field applications. Boron and nitrogen-doped MWCNTs were utilized to detect various VOCs produced by Aspergillus and Rhizopus from strawberries, utilizing tristimulus analysis to distinguish plants infected by either fungus [92]. Five groups of VOCs namely, alkanes, ketones, alcohols, and aromatic hydrocarbons, were distinguished through intermolecular forces between the porphyrin complexes and the analytes when utilizing MWCNTs functionalized with porphyrin group complexes [93]. Reduced graphene oxides (rGOs) were utilized alongside gold nanoparticles functionalized with various ligands to target thirteen individual VOCs emitted by plants in response to stress [94]. Furthermore, that study implemented a precalibrated function for temperature and humidity to minimize interference; the plant patch sensor was cost-effective and used a kirigami-based structure in the mechanistic design of the plant patch as a long-term health monitoring tool. A similar approach was deployed by fabricating a nanocomposite of rGOs with SnO2 and was used to detect ethanol [95]. The material exhibited a significant improvement in sensitivity, response, recovery time, and detection limits in comparison to pure SnO₂ sensors.

4.3. Metal Oxide-Based Sensors

Metal oxides (MO) are a fundamental component in hybrid sensors, often combined with polymers or carbon nanomaterials for improved performance. SnO₂ and TiO₂ nanoparticles were cast on screen-printed electrodes for the detection of p-ethylguaiacol, as a key marker of the fungal pathogen Phytophthora cactorum [96]. The sensor proved its sensitivity and selectivity towards the target VOC, which promised real-time and field-based applications. Mn-doped ZnO nanomaterials was synthesized for the detection of ethanol and acetone gases in comparison with undoped ZnO sensors [97]. The material was further modified by adding CdO, which further enhanced the response to ethanol. An electrochemical sensor was fabricated, capable of detecting the citrus tristeza virus using gold nanoparticles (AuNP) along with thiolated ssDNA probes [98]. The sensor, however, had a prolonged response time of 1–2 h due to hybridization of DNA.

5. Future Directions and Research Trends

The general trend of plant disease identification is targeted towards the detection of diseases through either hyperspectral imaging with probes, satellites, or drones or the identification of volatile organic compounds as biomarkers of disease [2]. One such method of plant disease identification is the implementation of sensors in agricultural land for a variety of reasons, including the observation of soil properties such as moisture content and composition, bacterial presence, and detection of conditions under which plant diseases occur, for prevention and early identification [2]. The detection of these phenotypic traits is also beneficial in other fields such as improving crop productivity through selective crop breeding [99].

One trend in the development of sensors for plant disease identification is the use of wearable sensors, which have been applied in phenotyping and in the monitoring of plant properties and growth as well as various responses generated in response to certain stressors, including the detection of VOCs and monitoring of the microenvironment to identify the conditions under which these stressors appear [26]. The general research trends aiming towards sustainability have led to the design of self-powered wearable sensors with greater self-reliance for wireless connectivity, and degradable sensors to minimize waste [26].

The implementation of wearable sensors in precision agriculture comes with its own set of challenges, although the specific limitations are affected by the sensor design and materials used. Wearable sensors, while considered a non-destructive method of plant monitoring, do influence plant growth when attached to plants due to their weight and the way they are attached [100]. The wearable sensor needs to be designed in such a way as to allow the sensor to remain attached while affected by environmental conditions such as weather and to ensure the sensing mechanisms are close enough to detect the relevant VOCs while not restricting the plant’s growth [100].

Recent studies have reported advancements in novel designs for flexible wearable sensors in hydrogel-based [101,102] and bio-inspired sensor systems. Bio-inspired sensor systems, as the name suggests, are inspired by the mechanisms of biological organisms including animals such as birds and arthropods, and plants [103]. The article by Cui et al. [103] describes different instances where bio-inspired energy harvesters have been implemented in sensor systems to advance self-powered agricultural monitoring systems. While there have been multiple studies on bio-inspired sensors in the field of human health, there are limited studies on bio-inspired sensors rather than systems. Searches using the keywords “Bio-inspired” AND “Agriculture” on both ScienceDirect and Scopus resulted in 27 and 3 results, respectively, among which none discussed actual bio-inspired sensors for agricultural purposes.

Hydrogels are three-dimensional materials that exhibit hydrophilic properties with great flexibility, elasticity, mechanical strength, and biocompatibility, making them suitable for a variety of purposes such as in triboelectric energy harvesters and electrochemical biosensors [102]. Hydrogels face challenges such as difficulties in the identification of suitable polymer compositions, challenging preparation process, and poor conductivity [102]. An example of the implementation of hydrogels in plant-wearable sensors was described [101] where researchers utilized a polyacrylic acid hydrogel with reduced graphene oxide and polyaniline in tandem with a triboelectric energy harvester, suitable for purposes such as plant growth monitoring and ammonia detection in smart farming systems, with the sensor demonstrating great mechanical strength, customization options due to its fabrication process, and high sensitivity.

Alongside the growing use of sensors, concerns have arisen regarding the increase in electronic waste (e-waste) pollution, with global e-waste production predicted to reach 74.7 metric tons by 2030 [104]. E-waste is generally comprised of diverse types of materials ranging from metals such as aluminum and iron to metal oxides, ceramics, and plastics, depending on the type of electronic item, thus contributing to environmental impacts of varying degrees [104]. E-waste also negatively influences plant life; the accumulation of heavy metals in the soil not only changes the soil composition but also poses the risk of being absorbed by plants and hindering their growth [104].

Efforts to lower the accumulation of e-waste have been made with the development of biodegradable sensors. Biodegradable sensors can be created based on a variety of materials, but the two primary categories are paper-based sensors and bioplastic-based sensors [105]. Wearable sensors have also utilized biomaterials such as pollen, cellulose, and silk as well as degradable synthetic polymers in their design [26]. There have been multiple studies conducted on the implementation of biodegradable sensors across a wide variety of fields, with some notable uses seen in fields such as medical bioanalysis, including the use of colorimetric devices, graphene- and polyethylene terephthalate (PET)-based electrodes, and wearable devices; forensic analysis of different compounds such as alcohol content in breath analyzers; and environmental analysis for the detection of different organic compounds, gases, bacteria, and heavy metals [105].

The implementation of biodegradable sensors in agriculture o has seen more focus in terms of research into more sustainable forms of technology. Biodegradable sensors have been applied in fields such as soil moisture monitoring [106,107], pathogen detection [107], and plant-wearable sensors [108]. It has been observed that these approaches have generated some success with the use of cellulose acetate, a plant-based material with advantages such as affordability and availability, as a substrate for a wearable sensor for pesticide detection in agricultural products [108].

While the products released by biodegradable sensors reduce the overall negative impact on the environment during their degradation [107], there are some key challenges in implementing biodegradable sensors in agricultural systems, such as their overall costs, shortened operational life due to their degradability, limitations on the overall sustainability of sensors due to limited availability of materials for components such as conductive inks, and difficulties in integration with remote monitoring and self-powering systems due to the necessary electrical components [106,108].

Trends are also leaning towards the use of artificial intelligence (AI), which enables more efficient use of the different types of data that these sensors can gather. Then, depending on the type of AI algorithm implemented, certain beneficial actions can be performed. There has been a noticeable increase in the number of AI-related studies in recent years, from 6 studies in 2015 that used AI in the field of sensor imaging for the investigation of plant stress response to 65 studies on the topic in 2020, representing a noticeable increase in the number of studies implementing AI [3]. There have, in fact, been a wide variety of studies focused on the implementation of AI in terms of plant imaging, with about 145 studies utilizing deep learning, 84 utilizing support vector machines, and 52 implementing artificial neural networks, with other types of AI algorithms being used at a lower rate [3]. Despite the popularity of deep learning algorithms in plant stress imaging, this method does have its limitations with the need for a large amount of data and the cost of acquiring relevant images as well as rigidness when encountering situations that the network has not been trained to address [3]. Sensors can be used for the detection of relevant phenotypic data for the training of the relevant AI algorithm in place of UAVs and satellite imaging, by attaching them to vehicles such as carts or tractors [99].

The concept of multimodal wearable sensors is a possibility for the automation of a comprehensive crop monitoring system capable of monitoring all relevant plant information. However, depending on the scale of the wearable sensors, there may be concerns about the sensor components affecting each other depending on material sensitivity [26,109]. The implementation of multimodal sensors and AI algorithms also helps in neutralizing the limitations of individual sensors that are only able to operate at a smaller scale and focus on individual types of data, leading to an incomplete analysis of the plant’s overall condition [100], and these algorithms can be useful for handling larger datasets that the sensor may gather throughout its operational life.

Sensor devices can also be connected to devices for everyday use, such as phones, through systems such as cloud connectivity, although this may also be limited by the data bandwidth as well as issues with signal connectivity in the area of use, with plant-wearable sensors being especially sensitive to this due to design limitations for minimizing plant interference [100]. The introduction of the IoT into the detection process enables the more efficient distribution of information, such as in the case of soil moisture sensors where the real-time monitoring of soil moisture content can help farmers in creating more informed decisions regarding, for example, the distribution of water resources, or in the case of soil pollutant sensors for addressing possible issues in the soil that may negatively impact agricultural yield [110].

According to the results reported in Section 2, there are a vast array of VOCs that are produced by plants in general in response to different stress conditions as well as natural factors caused by location, life cycle, and soil composition [26]. In the identification of VOCs in tomatoes caused by a single pathogen, hundreds of VOCs may be detected at differing concentrations. Thus, it is necessary for sensor designers to cooperate with experts in plant pathology to successfully identify which VOCs the sensor needs to detect. Depending on the location where the sensor is designed to be placed, it is also important to understand the way the plant component behaves; for example, more flexible sensors are necessary in the stem, which expands and contracts by varying amounts [26].

Cooperation with a materials science expert is then necessary to identify the most suitable sensing materials to be used in the design of the relevant sensors, and it is also necessary to consider the aspect of sensor degradability to minimize environmental impact. Other opportunities for collaboration may involve experts in artificial intelligence and machine learning, such as data scientists and engineers, for the optimization of design and detection processes, as well as experts in the IoT to ensure connectivity and ease of monitoring for farmers or other relevant staff [109]. The superstructure framework in Figure 5 focuses on the application of plant-wearable sensors in plant disease detection through VOC detection and the roles of experts from different disciplines, namely, materials scientists (yellow), plant pathologists (pink), data engineers (peach), and manufacturing engineers (light blue). Figure 5 includes three primary divisions which represent the key priorities (gray boxes), the research directions (green circles), and the strategic advantages (blue circles) [2,26,64,65,67,68,69,83,85,89,91,97,99,100,105,109,110,111]. The three key priorities for plant-wearable sensors are flexibility, sensitivity, and sustainability. Sensitivity involves the ability of the plant-wearable sensor to efficiently identify VOCs. Thus, the cooperation of plant pathologists is necessary for the selection of biomarkers to be identified, that of materials scientists is required for the selection and treatment of the material to be used, and that of manufacturing engineers is needed for the design and manufacturing of the sensor. On the other hand, flexibility is focused more on the sensor’s durability, universality, and adaptability, making it a more comprehensive opportunity for collaboration due to the need for expertise on aspects such as plant morphology and integration of data storage and collection. Lastly, sustainability is focused on ensuring minimal negative impact on the environment through different improvements throughout the sensor’s life cycle, from design, operation, and monitoring, all the way through to its eventual disposal, making this aspect the focus of manufacturing engineers and data engineers.

The different strategic advantages in the integration of key priorities have also been emphasized, with the agriculture sector being the primary beneficiary for advances in plant-wearable sensor research. The strategic advantages between sensitivity and sustainability involve focusing on the advantages of technological integration, allowing autonomous monitoring and operation of the sensor system, which enables agricultural stakeholders to more efficiently manage agricultural resources as well as further research, leading to improved efficiency in VOC detection while minimizing negative environmental impact. The integration of sensitivity and flexibility allows the use of plant-wearable sensors on different types of plants while maintaining data accuracy and minimizing any impact on the plants’ growth. Lastly, the combination of sustainability and flexibility allows overall sensor performance to be improved through technological integration while also considering any environmental ramifications. It should be noted that while strategic advantages are connected to two of the key priorities, they can also be connected to the third priority, such as in the case of non-intrusive monitoring, where sustainability may also apply. As such, it is integral to achieve a balance across all three key priorities to maximize the benefits from the key advantages. The potential research directions in Figure 5 detail the general trends of plant-wearable sensor research, ranging from materials research (biocompatible and flexible materials, self-healing materials, and advanced nanomaterials) to sensor design innovations (3D-printed and customizable patch design and energy harvesting sensors) and technological integration (multiple parameter monitoring, data-driven agriculture, smart sensing, and IoT integration).

6. Conclusions and Recommendations

This study examined the distinct VOC emission profiles of potato and tomato cultivars under various stressors, highlighting the challenges in identifying universal biomarkers for stress differentiation. It emphasizes advancements in sensing materials, including conductive polymers, carbon nanomaterials, and metal oxide semiconductors, showcasing their potential in VOC detection through hybrid/composite designs. Key trends include the adoption of biodegradable materials, AI, and IoT for sustainable, data-driven agriculture. Through bridging materials science, agriculture, AI, and data science, and the manufacturing industry, this research offers insights into developing wearable plant sensors for early stress detection and enhanced crop management. This study identified several critical aspects regarding advancements in sensing materials for agriculture and plant pathology, as follows:

• The VOC emission profiles of different potato and tomato cultivars were analyzed with the identification of biomarkers; however, no universal VOC profile was identified, making it difficult to identify clear differences between VOCs from each type of stressor.
• Novel doping and functionalization techniques have enhanced the sensitivity and selectivity of various sensing materials, such as conducting polymers, carbon nanomaterials, and metal oxides, which can be combined into hybrid or composite materials that could further positively influence sensing performance.
• The most effective sensing materials are composite materials that can be put through processes such as doping or functionalization for detecting the target VOCs, with future research directions focusing on the identification of more efficient composite materials and treatment methods for detection of specific VOCs.
• The different interdisciplinary opportunities between materials science and agriculture can be summarized in three key priorities, which are anchored on sensitivity and flexibility, focusing more on the sensor design and manufacturing process innovations, and sustainability, focusing on minimizing negative environmental impact.

Future studies should analyze VOC emission profiles under key stressors, considering temporal dynamics to minimize risks. Interdisciplinary collaborations are recommended to optimize sensor designs by integrating advanced, sustainable materials that enhance selectivity, sensitivity, stability, and scalability. This approach will maximize agricultural benefits while minimizing environmental impacts.