Optimizing Caffeine Treatments for Brown Marmorated Stink Bug Management in Laboratory Bioassays

Abstract

The brown marmorated stink bug (Halyomorpha halys Stål, 1855) is a globally invasive polyphagous pest that challenges conventional chemical control. We evaluated caffeine-based preparations—alone and combined with chitosan, acetic acid, and ethanol—against adults under laboratory conditions using topical application and 72 h mortality readouts. Among caffeine-in-water treatments, 3% (w/v) yielded the highest mortality (52.5%), indicating an efficacy peak constrained by solubility/precipitation. The most effective overall formulation was 1% caffeine + 1% chitosan + 3% acetic acid, reaching 57.5% mortality and outperforming higher caffeine loads (3–5%). Ethanol as a co-solvent consistently reduced efficacy across concentrations. Patterns across treatments indicate that bioefficacy was driven predominantly by formulation chemistry rather than dose: the chitosan–acetic acid matrix enhanced cuticular deposition, retention, and diffusion of caffeine, whereas high caffeine levels likely triggered detoxification responses and/or reduced bioaccessible dose due to precipitation. By enabling lower active ingredient loads with equal or greater bioactivity, the biodegradable chitosan–acid system improves the environmental profile of caffeine-based insecticides. These results identify a practical, low-complexity path to optimize caffeine delivery for H. halys control and support integration into IPM frameworks. Field validation, testing on earlier life stages, and assessment of non-target effects and resistance biomarkers are warranted to translate these findings into robust, sustainable pest management strategies.

1. Introduction

The brown marmorated stink bug, Halyomorpha halys (Stål, 1855), is an invasive species native to East Asia that has rapidly expanded its range across North America, South America, and Europe. Its high mobility, broad host range, and strong reproductive potential have contributed to significant agricultural losses and seasonal nuisance problems in urban environments [1,2]. Likely introduced through international trade, H. halys is now considered a major pest of fruit, vegetable, and ornamental crops in many parts of the world [3].

In Europe, H. halys has steadily spread over the past two decades, with recent confirmed detections in North Macedonia [4] and Moldova [5]. In Croatia, the species was first recorded in 2017 in the coastal city of Rijeka [6], and a mass emergence in a soybean field near Zagreb in 2019 confirmed its establishment in agricultural landscapes [7].

Halyomorpha halys is a highly polyphagous insect that utilises more than 300 host plant species from various botanical families [8,9]. Its ability to adapt to different climates and crop types has made it a major concern in fruit-producing regions worldwide, with annual economic losses estimated in the millions of dollars [10]. The pest feeds by piercing plant tissues with its stylet-like mouthparts, enabling it to access nutrients from fruits, leaves, and stems [11]. This feeding causes both mechanical damage and biochemical disruption, as salivary enzymes degrade plant cells, leading to necrosis and physiological stress in host plants [12].

Initial management efforts relied on broad-spectrum insecticides. While temporarily effective, these chemicals posed serious risks to non-target organisms, pollinators, and beneficial arthropods, thereby disrupting existing integrated pest management (IPM) systems [13]. As a result, research has increasingly focused on alternative, environmentally sustainable approaches, including semiochemical-based attract-and-kill strategies [14], physical exclusion through netting [15], classical biological control using egg parasitoids such as Trissolcus japonicus and Trissolcus mitsukurii [16], and microbiota-targeted strategies that disrupt essential insect symbionts [17]. These methods offer promising options for reducing pesticide use while maintaining effective crop protection. As phytophagous Heteroptera such as H. halys continue to pose a growing threat to horticultural production, the search for sustainable and effective alternatives to conventional insecticides is increasingly urgent. Natural compounds, including caffeine, have gained attention in this context. Recent laboratory studies have demonstrated that both pure caffeine and aqueous coffee extracts exhibit significant insecticidal activity against H. halys, supporting their potential inclusion in future IPM strategies [18].

Caffeine (1,3,7-trimethylxanthine) (Figure 1) is a purine alkaloid naturally produced by several plant species. While widely recognised for its stimulant effects in humans, caffeine also functions as a natural defence compound in plants by deterring herbivorous insects through its toxic effects [19,20]. Its insecticidal mode of action is primarily associated with the inhibition of phosphodiesterase, which leads to elevated intracellular levels of cyclic adenosine monophosphate (cAMP), ultimately disrupting normal insect physiology and feeding behaviour [21]. Furthermore, caffeine has been shown to synergise with other insecticidal compounds, enhancing their efficacy even at lower doses. Prior laboratory trials support caffeine’s broad insecticidal potential across taxa. In the lace bug Corythucha ciliata, aqueous caffeine at 0.239% w/v (2.389 g L⁻¹) produced ≈70% adult mortality at 72 h, whereas coffee extracts standardized to an equivalent caffeine load (0.239% w/v) achieved ≈47% under comparable bioassays [18]. Against the coffee berry borer Hypothenemus hampei, caffeine–oleate emulsions containing 4000 µg mL⁻¹ caffeine (0.4% w/v) with 6% (w/w) oleic acid, 0.3% achieved >90% adult mortality in preventive tests and ≈77% in curative assays; under controlled field conditions, infestation fell below 20% and damaged fruits were reduced by up to ≈70% [22]. Earlier formulation work also showed that while 0.20–2.00 wt% aqueous caffeine lacked activity, a caffeine–oleate emulsion with just 0.04 wt% caffeine and 0.05 wt% oleic acid (with 20 vol% oil and 2 wt% surfactant) was highly bioactive against Drosophila melanogaster and H. hampei [23]. Mosquito larvicidal assays on Aedes aegypti reported 100% mortality at 2 mg mL⁻¹ (0.20% w/v) with LT₅₀ ≈ 3 days, confirming dose-responsive potency [24]. These findings demonstrate that selecting appropriate carriers/solvents and concentrations—not merely increasing dose—can enhance caffeine stability, delivery, and bioefficacy while enabling reduced active ingredient loads.

A promising approach to further improve caffeine’s insecticidal performance lies in the use of carrier agents, particularly chitosan, which can stabilize and enhance the bioactivity of sensitive compounds [25]. Chitosan (Figure 2), a deacetylated derivative of chitin, is a cationic biopolymer capable of forming electrostatic interactions with negatively charged surfaces of the insect epicuticle and gut epithelium [26,27]. Its surfactant and mucoadhesive properties increase the permeability and retention of active compounds like caffeine while exerting its own entomotoxic effects through gut disruption and osmotic imbalance [27]. These synergistic effects are well-documented in controlled-release pesticide delivery studies [25,26,27]. Importantly, chitosan is not a passive carrier but an active bioagent. It has been shown to exert direct entomotoxic effects by disrupting insect gut integrity, impairing nutrient absorption, modulating immune responses, and functioning as an antifeedant [27,28]. Its activity as a surfactant and chelator further contributes to physiological disruption in insects.

Moreover, caffeine degradation in the environment occurs through photolysis and microbial biodegradation pathways, yielding metabolites such as theobromine and par-axanthine, which are less biologically active [29]. By enhancing caffeine stability and reducing the required dose, chitosan can minimize both environmental persistence and potential non-target effects [26,27,29,30].

Chitosan nanoparticles (NPs) formulation has emerged as highly effective system capable of enhancing stability, bioavailability, and controlled release of encapsulated bioactive compounds [27,30]. Their structure increases surface area and improves adhesion to insect cuticles, enabling prolonged contact and controlled diffusion of the active ingredient over time [31]. This sustained release extends the residual activity of the insecticidal compounds under field conditions and reduces the total amount of active ingredient required to achieve effective pest control [30,31,32]. Formulation chemistry plays a key role in this synergy.

Mildly acidic environments (e.g., acetic acid (Figure 3) solutions) protonate chitosan’s amino groups, improving their solubility and film-forming properties, while simultaneously enhancing caffeine solubilization and preventing recrystallization [33].

Ethanol (Figure 4) when incorporated as a cosolvent, can substantially modify the physicochemical behavior of caffeine-based formulations by altering drying kinetics, surface tension, and penetration efficiency across the insect epicuticle. At lower concentrations, ethanol can improve solubility and wetting properties, supporting more effective deposition and diffusion of the active ingredient [34]. However, at higher levels, rapid evaporation may reduce uniform film formation, decrease retention time, and limit the interaction of bioactive compounds with target surfaces [35]. In combination with chitosan matrices, ethanol can further influence viscosity, deposition uniformity, and crystallization tendencies during drying, thereby affecting formulation performance [36]. These concentration-dependent, multifactorial interactions highlight the importance of optimizing solvent–carrier–active ingredient ratios to balance solubility, stability, and bioefficacy in practical application systems. These complexes allow controlled release of active ingredients, extend residual activity under field conditions, and reduce the amount of compound required to achieve effective control.

Furthermore, chitosan can synergize with other bioactive compounds, including polyphenols, essential oils, and microbial metabolites, thereby amplifying their insecticidal effects. Such synergistic formulations have been proposed as powerful tools to overcome the limitations of standalone compounds, combining chitosan’s carrier properties with its intrinsic insecticidal functions. Equally important, chitosan-based formulations are fully biodegradable, biocompatible, and environmentally benign, making them suitable components of sustainable integrated pest management (IPM) programs [24,27,28,37]. Their minimal toxicity to beneficial insects, soil microbiota, and aquatic organisms positions them as a safer alternative to synthetic pesticides, supporting agricultural practices aimed at reducing chemical inputs, mitigating resistance, and complying with increasingly strict regulatory standards [28,37].

The aim of this study was therefore to evaluate the insecticidal potential of caffeine against the invasive brown marmorated stink bug by (i) determining its effective concentrations under laboratory conditions, (ii) testing its performance when combined with natural carriers and solvents (chitosan, acetic acid, ethanol), and (iii) identifying the most promising preparation for further development as a sustainable and environmentally friendly pest management option.

2. Materials and Methods

Low-molecular-weight chitosan (CAS RN: 9012-76-4, molecular weight: 100,000–300,000) was obtained from Acros Organic (St. Louis, MO, USA). Ethyl alcohol, Pure (CAS No.: 64-17-5) and acetic acid (CAS No.: 64-19-7) were purchased from MERCK (Zagreb, Croatia). Caffeine was purchased from Tera Organica (Zagreb, Croatia). All chemicals used for the experiment were analytical grade.

2.1. Collection of Samples

Adult H. halys for this experiment were collected from an apple orchard in Velika Ludina (Sisak-Moslavina County, Croatia) in September and October 2024, using black pyramidal traps baited with an aggregation attractant effective at luring both males and females. The collected insects were kept in entomological cages to recover overnight before testing, without additional feeding or previous contact with insecticides.

2.2. Treatment Preparation

The insecticidal efficacy of caffeine in controlling H. halys through topical application was evaluated in the Laboratory of the Department of Agricultural Zoology. In these experiments, different water formulation systems with different ratios of caffeine (CAF), ethanol (ET), acetic acid (AC), and chitosan (CTS) were prepared. The acetic acid chitosan solution was added as a carrier of caffeine and ethanol was added to increase the solubility of caffeine. The treatments are summarized in

Table 1. Treatment Summary.

Treatment No.	Sample Description
1	Distilled water (control)
2	1% CAF + H2O
3	3% CAF + H2O
4	5% CAF + H2O
5	1% CAF + 1% CTS + 3% AC
6	1% CAF + 1% CTS + 3% AC + 5% ET
7	3% CAF + 1% CTS + 3% AC + 5% ET
8	3% CAF + 1% CTS + 3% AC
9	5% CAF + 1% CTS + 3% AC + 3% ET
10	5% CAF + 1% CTS + 3% AC

CAF: caffeine; CTS: chitosan; AC: acetic acid; ET: ethanol.

. The formulations were prepared in the Laboratory of the Department of Chemistry at the Faculty of Agriculture, University of Zagreb and were made available for this research. All bioassays were conducted in a controlled-environment laboratory at 20 ± 1 °C, 60 ± 5% relative humidity, and a 14:10 h (L:D) photoperiod.

2.3. Experimental Setup

For all replicates, ten adults (5 males and 5 females) were placed in a Petri with a 90 mm diameter (internal surface area ≈ 6360 mm²). Adult H. halys used in the bioassays were sexed under a stereomicroscope based on external genital morphology. Contact toxicity was evaluated by applying the treatments directly to the insects inside the Petri dishes using a laboratory hand sprayer, with a volume of 3 mL per dish. One Petri dish represented one replicate. The untreated control for all experiments included a treatment in which the H. halys were placed in Petri dishes treated with distilled water. All treatments and the water control were run concurrently in the same room, and Petri dishes were randomly arranged and repositioned daily to minimize positional or microclimatic effects. Each application and the investigated effect of the tested ingredients were assessed in four replicates. In total, 10 different variants were tested on 400 individuals.

2.4. The Assessment of the Results

The number of dead H. halys in each Petri dish was determined every 24 h for three days. Mortality of males and females was recorded separately at each assessment time to allow evaluation of potential sex-specific differences in susceptibility. In a preliminary analysis, sex-specific responses were examined by comparing overall mortality of males and females using a one-way ANOVA. Based on the number of dead individuals found in the treatment and the untreated control, the efficacy (%) of the ingredients was determined according to Abbott’s formula [38]. In addition to p-values, 95% confidence intervals for treatment means and effect size statistics (partial η² and Cohen’s f) were calculated from the ANOVA results to quantify the magnitude and precision of treatment effects. Dose–response relationships for the three formulation groups (CAF + H₂O; CAF + 1% CTS + 3% AC; CAF + 1% CTS + 3% AC + ethanol) were additionally evaluated using probit analysis on pooled 72 h mortality data for 1%, 3% and 5% caffeine, with LD₅₀ values and their 95% confidence limits calculated according to standard generalized linear modeling procedures. Statistical data analysis (ANOVA, Tukey’s HSD test) was performed using ARM 2025^® GDM software, Revision 2025.2 (Gylling Data Management Inc., Brookings, SD, USA) [39].

3. Results

Preliminary analyses showed no sex-related differences in mortality (F₁,₂₂ = 1.13, p = 0.30), and therefore male and female data were pooled for all subsequent efficacy comparisons.

The heatmap (Figure 5) summarizes H. halys adult mortality across treatments and time intervals. At 24 h, mortality ranged from 0% to 17.5%, with the highest values for 1% CAF + 1% CTS + 3% AC (17.5%) and 3% CAF + H₂O (15%), while all other treatments remained ≤5%. By 48 h, mortality increased to 37.5% for 3% CAF + H₂O and 25.0–22.5% for the CTS–AC combinations (1% CAF + 1% CTS + 3% AC and 3% CAF + 1% CTS + 3% AC), whereas ethanol-containing formulations stayed at ≤15%. At 72 h, 1% CAF + 1% CTS + 3% AC reached 57.5% mortality and 3% CAF + H₂O 52.5%, followed by 3% CAF + 1% CTS + 3% AC with 37.5%, while 5% CAF alone and all ethanol mixtures remained below 25%.

As control mortality was consistently 0% in all bioassays, Abbott’s corrected efficacy values were identical to observed mortality rates and are therefore presented interchangeably in the results. Abbots’ efficacy of caffeine-based formulations against H. halys varied across concentrations, carriers, and time intervals (

Table 2. Efficacy of caffeine-based treatments against Halyomorpha halys.

Treatment	Efficacy (%)			Confidence Level 95%
	24 h	48 h	72 h	24 h	48 h	72 h
1% CAF + H2O	5.0 ± 2.9 ns *	12.5 ± 4.8 ab	22.5 ± 13.2 bcd	0.0–17.5	0.0–33.2	0.0–78.9
3% CAF+ H2O	15.0 ± 6.5 ns	37.5 ± 6.3 a	52.5 ± 4.8 ab	0.0–43.9	10.3–64.7	31.8–73.2
5% CAF + H2O	5.0 ± 2.9 ns	5.0 ± 2.9 b	7.5 ± 4.8 cd	0.0–17.5	0.0–17.5	0.0–28.2
1% CAF + 1% CTS + 3% AC	17.5 ± 8.5 ns	25.0 ± 6.5 ab	57.5 ± 8.5 a	0.0–54.1	0.0–53.9	20.9–94.1
1% CAF + 1% CTS + 3% AC + 5% ET	2.5 ± 2.5 ns	7.5 ± 4.8 b	25.0 ± 8.7 bcd	0.0–13.3	0.0–28.2	0.0–62.5
3% CAF + 1% CTS + 3% AC + 5% ET	2.5 ± 2.5 ns	15.0 ± 8.7 ab	22.5 ± 7.5 bcd	0.0–13.3	0.0–52.5	0.0–54.8
3% CAF + 1% CTS + 3% AC	2.5 ± 2.5 ns	22.5 ± 6.3 ab	37.5 ± 4.8 abc	0.0–13.3	0.0–47.6	16.8–58.2
5% CAF + 1% CTS + 3% AC + 3% ET	0.0 ± 0.0 ns	10.0 ± 4.1 b	20.0 ± 4.1 cd	0.0–0.0	0.0–27.6	2.4–37.6
5% CAF + 1% CTS + 3% AC	5.0 ± 2.9 ns	7.5 ± 2.5 b	12.5 ± 4.8 cd	0.0–17.5	0.0–18.3	0.0–33.2
Tukey’s HSD p = 0.05	19.68	26.59	31.46
Standard Deviation	8.09	10.93	12.94
Coefficient of variation CV	140.71	74.12	49.28
Levene’s F ^	1.456	0.524	1.954
Levene’s Prob(F)	0.209	0.845	0.082
Shapiro–Wilk ^	0.9476	0.981	0.9839
P(Shapiro–Wilk) ^	0.0629	0.7273	0.83
Skewness ^	0.9289 *	0.4195	0.3874
P(Skewness) ^	0.0215 *	0.2859	0.3237
Kurtosis ^	2.1832 *	0.6552	0.1729
P(Kurtosis) ^	0.0065 *	0.3938	0.8212
Replicate F	1.057	0.411	2.804
Replicate Prob(F)	0.3838	0.7462	0.0588
Treatment F	2.041	3.739	7.579
Treatment Prob(F)	0.0735	0.0037	0.0001
Partial η2 (treatment)	0.405	0.555	0.716
Cohen’s f	0.95	1.29	1.84

* Means followed by the same letter (a–d, ns) do not significantly differ (p = 0.05, Tukey’s HSD). Means descriptions are reported in transformed data units and are not de-transformed. Analyses were performed on arcsine square root percent-transformed data. ^ Assumption test statistics for ANOVA. Normality and homogeneity of variances were assessed using Shapiro–Wilk and Levene’s tests, respectively. Asterisks (*) indicate significance at p < 0.05.

).

At 24 h, efficacy values ranged from 0% to 17.5%, with the highest observed for 1% CAF + 1% CTS + 3% AC (17.5%) and 3% CAF + H₂O (15%). However, Tukey’s HSD test indicated no significant differences among treatments and the control, confirming that early mortality effects were minimal. Variability was high (coefficient of variation CV = 140.71%), and residual analysis revealed deviations from normality (Skewness = 0.9289, p = 0.0215; Kurtosis = 2.1832, p = 0.0065). The 95% confidence intervals were wide and largely overlapping (often extending from 0 up to >40% efficacy), further illustrating the inconsistency of responses at this point. Although the Shapiro–Wilk test was not significant (p = 0.0629), treatment effects did not reach statistical significance (Treatment F = 2.041, p = 0.0735). The overall effect size at 24 h was nonetheless large (partial η² ≈ 0.48; Cohen’s f ≈ 0.95), suggesting a potentially meaningful but statistically inconclusive trend under the given level of experimental variability.

By 48 h, efficacy increased markedly in several treatments. The 3% CAF + H₂O treatment achieved the highest efficacy (37.5%), grouping as “a” in Tukey’s HSD, while 1% CAF + 1% CTS + 3% AC (25%) and 3% CAF + 1% CTS + 3% AC (22.5%) also performed well, grouping as “ab.” In contrast, higher caffeine concentrations (5%) or ethanol-containing mixtures produced lower efficacy (≤15%). Statistical analyses confirmed significant treatment effects (Treatment F = 3.739, p = 0.0037), with acceptable data distribution (Shapiro–Wilk p = 0.7273; Levene’s test p = 0.845). The coefficient of variation decreased to 74.12%, reflecting greater stability in treatment responses compared to 24 h. Effect-size estimates indicated a very strong overall treatment impact at this interval (partial η² ≈ 0.62; Cohen’s f ≈ 1.29), supporting the biological relevance of the observed differences.

At 72 h, treatment effects were strongest and most consistent. The highest efficacy was recorded for 1% CAF + 1% CTS + 3% AC (57.5%) and significantly surpassing most treatments. This formulation was closely followed by 3% CAF + H₂O (52.5), while 3% CAF + 1% CTS + 3% AC (37.5%) also demonstrated strong efficacy. In contrast, 5% CAF + H₂O (7.5%) and ethanol-containing mixtures (12.5–25%) showed markedly lower effects. Overall treatment effects were highly significant (Treatment F = 7.579, p = 0.0001). Data variability was lowest at this interval (CV = 49.28%), and normality assumptions were fully met (Shapiro–Wilk p = 0.83; Levene’s p = 0.082). Replicate effects remained non-significant (p = 0.0588), supporting reproducibility of the results. Consistent with this, the calculated effect sizes were very large (partial η² ≈ 0.77; Cohen’s f ≈ 1.84), confirming that treatment choice had a major and robust influence on mortality by 72 h.

To further characterize dose–response relationships within the three formulation families, a separate probit analysis was conducted on 72 h mortality data for: (i) CAF + H₂O, (ii) CAF + 1% CTS + 3% AC and (iii) CAF + 1% CTS + 3% AC + ethanol, each tested at 1, 3 and 5% (w/v) caffeine. For the chitosan–acetic acid matrix (CAF + 1% CTS + 3% AC), the model yielded a well-defined median lethal concentration (LD₅₀) of 1.42% CAF (95% CI: 0.85–1.99%). This estimate is consistent with the high efficacy observed already at 1% CAF (57.5% mortality) and the decline at higher concentrations (3–5%), indicating that relatively low loads of caffeine are sufficient to reach the 50% mortality threshold in this formulation. In contrast, for both CAF + H₂O and CAF + 1% CTS + 3% AC + ethanol, mortality never approached a monotonic increase with dose: in CAF + H₂O, mortality peaked at 3% CAF and declined again at 5%, while in the ethanol-containing matrix all doses remained below 30% mortality. As a result, probit models for these two families extrapolated LD₅₀ values below the lowest concentration tested, with very wide confidence intervals spanning zero, and are therefore not biologically interpretable. These patterns reinforce the ANOVA results, indicating that the chitosan–acetic acid carrier provides the only formulation in which a robust LD₅₀ can be estimated within the tested concentration range, and that ethanol-containing mixtures act as comparatively weak contact toxicants under our conditions.

4. Discussion

This study demonstrates that caffeine exhibits moderate but biologically relevant insecticidal activity against H. halys, aligning with prior work on natural polyphenols—including caffeine—across diverse insect pests [18,22,23]. Among caffeine-in-water treatments, 3% (w/v) achieved the highest mortality (52.5%), suggesting an efficacy peak that balances toxic action with solubility and bioavailability constraints. Increasing the concentration further did not improve performance, consistent with precipitation on drying and reduced bioaccessible dose at the site of contact, as well as potential onset of repellency at excessive loads [18,40]. When placed in the context of other caffeine-based insecticidal studies, the performance of our best formulation (57.5% adult mortality at 72 h with 1% caffeine) is broadly comparable, though clearly lower than the highest values reported for more susceptible species and more complex delivery systems. For example, aqueous caffeine at 0.239% w/v caused ≈70% adult mortality of Corythucha ciliata at 72 h, while coffee extracts standardized to the same caffeine load achieved ≈47% under similar laboratory conditions [18]. Against the coffee berry borer Hypothenemus hampei, caffeine–oleate emulsions containing 0.4% w/v caffeine produced >90% adult mortality in preventive tests and ≈77% in curative assays [22], and larvicidal assays on Aedes aegypti reported 100% mortality at 0.20% w/v [24]. Thus, our maximum value of 57.5% against H. halys adults—obtained with a relatively simple, biodegradable chitosan–acetic acid carrier—falls within the mid-to-upper range of caffeine efficacies but remains moderate in absolute terms and likely reflects both species-specific tolerance and the inherent robustness of pentatomid adults under contact exposure. These comparisons indicate that, while caffeine alone can reach very high efficacy in some systems, substantial formulation optimization and species-specific tailoring are required to approach such levels in a hard-bodied hemipteran like H. halys. Consistent with the ANOVA outcomes, the probit analysis showed that only the chitosan–acetic acid formulation produced a well-defined LD₅₀ within the tested range (≈1.4% CAF, 95% CI 0.85–1.99%), whereas LD₅₀ values for CAF in water and for ethanol-containing mixtures lay outside the experimental range and carried very broad confidence intervals, underscoring their lower and more erratic efficacy. A major advance of this work is the formulation that combined 1% caffeine with 1% chitosan in 3% acetic acid, which outperformed all other treatments and produced the most rapid and sustained mortality over time (up to 72 h) within our test set. Chitosan has been reported to increase adhesion, permeability, and residence time of applied compounds on insect and plant surfaces, and its acidic solubilization can produce films that support gradual release and protect actives from degradation [27,28]. Other studies also show that chitosan can exert independent entomotoxic or physiological effects, including disruption of gut microbial communities, impaired digestion, and reduced feeding [27,41,42,43]. In addition, chitosan coatings have been associated with repellent or anti-feedant responses and may modify surface chemistry in ways that influence insect behaviour [27,41]. For example, chitosan is known to enhance adhesion, permeability, and surface deposition in insect systems [26,27], but its capacity to modulate epicuticular permeability in H. halys remains untested. The reduction in performance observed in ethanol-containing treatments could reflect accelerated evaporation and reduced film formation, consistent with reports that ethanol can alter viscosity and disrupt uniform deposition of biopolymer-based formulations [35,36,37]. The plateau in efficacy at elevated caffeine concentrations aligns with extensive literature showing induction of cytochrome P450 monooxygenases and related detoxification pathways at high xenobiotic loads [44,45,46]. Within the tested range, formulation effects appear more influential than concentration differences per se.

Taken together, these mechanisms provide a coherent, literature-based framework for interpreting the observed efficacy patterns. Future work should therefore test these hypotheses explicitly through cuticular permeability assays, formulation–surface interaction analyses, controlled-release studies, and quantification of detoxification enzyme activity. Such targeted experiments will be essential for determining the true mechanistic basis for the enhanced performance of chitosan–acid caffeine formulations and for guiding rational optimization of bioinsecticide delivery systems.

From a sustainability standpoint, leveraging chitosan and mild acidity enables biologically relevant efficacy at comparatively low caffeine levels. This is relevant to environmental risk: caffeine is a known aquatic contaminant with documented ecological impacts [29]. Reducing the required dose directly lowers potential residues and exposure to non-target organisms. Moreover, chitosan and acetic acid are biodegradable, further improving the environmental profile relative to conventional synthetic co-formulants [28,37,47]. Recent literature on polyphenol formulations also points to nano/microemulsions and nano/microencapsulation as viable routes to improve stability and delivery while mitigating volatility and precipitation challenges [40,48]; the present results are consistent with that trajectory and suggest that chitosan systems can function as a practical, low-complexity alternative or complement.

Two broader conclusions emerge, with appropriate caveats. First, the matrix matters: carrier chemistry and solvent choice governed bioavailability, residence time, and penetration to a greater extent than caffeine concentration per se within our tested range. Second, chitosan allowed the lowest tested caffeine dose to deliver the highest efficacy in this experiment, echoing reports that chitosan can reduce the quantity of insecticides needed for biological control [28,41]. Together, these insights argue for rational formulation design—matching active ingredient ionization state, solvent volatility, and carrier adhesion—to balance onset and duration of action while minimizing detoxification triggers, rather than relying solely on dose escalation.

A limitation of this study is the relatively small sample size per treatment, which reduces the statistical power to detect moderate differences among formulations. Consequently, nonsignificant results, particularly at earlier assessment times, should be interpreted with caution as they may reflect insufficient power rather than a true absence of biological effect. Future trials with larger sample sizes, additional assessment dates would help to confirm the robustness and relative positioning of the efficacy patterns observed here. It is also important to note that, despite arcsine square root transformation, the 24 h efficacy data still showed deviations from normality, as indicated by significant skewness and kurtosis, together with very high coefficients of variation. This pattern reflects the large number of low or zero responses and a few higher values, which is common in early-effect bioassays but violates strict ANOVA assumptions. Consequently, results at 24 h should be interpreted with caution and primarily as indicative trends, whereas the 48 h and 72 h datasets—where normality and homoscedasticity were satisfied—provide a more robust basis for evaluating treatment differences.

Future work should also (i) validate findings under field conditions across developmental stages and environmental regimes; (ii) refine dose–response relationships with attention to solubility limits and crystallization kinetics; (iii) quantify detoxification markers (e.g., P450 activity) under varying loads to identify thresholds that elicit metabolic induction; (iv) compare alternative acids and biopolymer grades (e.g., chitosan degree of deacetylation and molecular weight) for adhesion and release control; and (v) assess non-target and environmental endpoints, including residue dynamics. Collectively, our results position chitosan-enabled, mildly acidic caffeine formulations as a promising but preliminary bioinsecticide strategy for H. halys, offering moderate efficacy at lower doses and supporting further development within environmentally responsible, IPM-oriented pest management rather than as a stand-alone replacement for conventional insecticides.

5. Conclusions

This study demonstrated that caffeine provides moderate but consistent insecticidal activity against the brown marmorated stink bug (Halyomorpha halys) under laboratory contact-exposure conditions, with maximum adult mortality of 57.5% at 72 h in the best-performing formulation. Within our experimental framework, efficacy was improved when caffeine was formulated with chitosan and acetic acid: the combination of 1% caffeine, 1% chitosan, and 3% acetic acid repeatedly produced the highest and most stable mortality among the tested treatments, outperforming higher caffeine concentrations applied alone. These findings indicate that formulation chemistry—particularly the use of a biodegradable chitosan–acid matrix—can enhance the practical performance of caffeine without increasing the active ingredient load.

From an environmental perspective, the use of caffeine together with chitosan and mild organic acid co-formulants offers a promising route toward more sustainable stink bug management, potentially lowering required caffeine doses and associated residue risks. Nonetheless, these formulations should be viewed as candidates for integration into broader IPM strategies—rather than as stand-alone replacements for conventional insecticides—until their performance is validated under field conditions, across life stages, and in combination with other control measures.