Fabrication of gelatin microspheres containing ammonium hydrogen carbonate for the tunable release of herbicide

The major challenge in utilizing pesticides lies in identifying the precise application that would improve the efficiency of these pesticides and decline their environmental and health hazards at the same time. Such application requires the development of specific formulations that enable controlled, stimuli-responsive release of the pesticides. Gelatin is a relatively cheap material characterized by temperature-sensitivity and abundant amino acid groups, which makes it suitable for the storage and controlled release of pesticides. In this study, gelatin microspheres were prepared by emulsion and cross-linking, then they were loaded with 2,4-dichlorophenoxyacetic acid sodium (2,4-D Na) as a model herbicide. To achieve temperature-tunable release of 2,4-D Na from the microspheres, NH4HCO3 was added to the formulations at different concentrations. The prepared formulations were characterized by SEM, FTIR, and size distribution analyzes, and their drug loading capacities were determined. Based on bioassay experiments, the 2,4-D Na-NH4HCO3-loaded gelatin microspheres can effectively control the spread of dicotyledonous weeds. Therefore, the strategy proposed herein can be used to develop novel, effective herbicide formulations.


Introduction
Agricultural chemicals (ACs) are extensively used in modern crop production to control the spread of weeds, pests, and plant diseases. However, AC residues that remain in the soil or on the crops after application have an adverse impact on the environment (Yolanda et al. 2014;Andreia et al. 2020). Moreover, the stability and efficacy of ACs depend strongly on meteorological factors. For example, herbicides are more effectively absorbed at higher temperatures, leading to improved weeding effect (Park et al. 2017) but quick evaporation, which dramatically decline their efficiency (Brown et al. 2000). Hence, the major challenge in AC utilization lies in identifying the precise application that would simultaneously enhance the efficacy of the pesticides and decrease the associated environmental and health hazards. Such application requires the use of controlled and stimuli-responsive release formulation of ACs.
Strategies are developed to protect ACs and improve their efficiency by sustained and controlled delivery systems (Amrita et al. 2020). In addition to enhancing the environmental stability and response ability of ACs, these systems decrease environmental toxicity and pollution by reducing the amount of ACs discharged into the environment (Zhao et al. 2018(Zhao et al. , 2019Liu et al. 2019;Xu et al. 2017;Gao et al. 2019). Sheng et al. proposed a thermo controlled pesticide release system composed of a thin film of poly(2-(dimethylamino)ethyl methacrylate) (PDMAEMA) grafted with polydopamine (PDA) microcapsules for the loading of avermectin (Sheng et al. 2015). Zhang et al. also reported an AC delivery system. Their system consists of temperature-responsive mixed micelles (MMs-Pys-7) that can be used to control the release of pyrethrins (Zhang et al. 2018). Despite their effectiveness, the proposed systems composed of synthetic composite materials cannot be readily prepared.
Gelatin is a cheap, temperature-sensitive, and biocompatible protein which is rich in amino acid groups (glycine, alanine, and proline), consequently used as a sustained-release carrier of ACs (Rampurna and Gullapalli 2010; Kim et al. 2018;Su and Wang 2015). In this study, we analyze the efficiency of gelatin microspheres loaded with Ammonium hydrogen carbonate (NH 4 HCO 3 ) in delivering 2,4dichlorobenzene oxygen ethanoic acid sodium (2,4-D Na), the first herbicide developed in the world (Molnar et al. 2015). NH 4 HCO 3 is relatively unstable and readily decomposes into ammonia and carbon dioxide at high temperatures (Chi et al. 2017). Also, ammonia and carbon can promote the growth of plants (Drake 1997). 2,4-D Na is usually used to inhibit the growth of barley, wheat, rice, corn, and other gramineous crops in urban pastures (Islam et al. 2017). However, the application of this herbicide is limited by its low adsorption coefficient and high solubility in water (Lerro et al. 2017), leading to serious lost in the field by rain, which damages the adjacent dicotyledonous crops and trees (Mountassif et al. 2010;Brito et al. 2015). At high temperatures, 2,4-D Na volatilizes quickly and diffuses, resulting in weakend effectiveness.
Herein, 2,4-D Na was encapsulated in gelatin microspheres by emulsifying and cross-linking. To control the release of 2,4-D Na, the content of NH 4 HCO 3 in the temperature-responsive microspheres was adjusted. The results obtained in this study provide an experimental basis for the development of new temperature-responsive formulations that can applied in the sustained release of herbicides, which decreases the negative impact of these chemicals on the environment.

Materials and methods
Preparation of gelatin microspheres, 2,4-D Naloaded gelatin microspheres and 2,4-D Na-NH 4 HCO 3 -loaded gelatin microspheres According to the method of emulsification and crosslinking (Chen et al. 2017;Akin and Hasirci 2010), gelatin solution was heated at 50°C and stirred for 1 h. Then liquid paraffin (V/V water = 3/10) and 1.5% span 80 were added respectively, strring for 30 min at 400 rpm to form water-in-oil droplets, the microspheres were fixed with 3% formaldehyde. Finally, the gelatin microspheres were dried in a lyophilizer.

Characterization
Size Distribution, FTIR and SEM were used to characterize the gelatin microspheres (see Supplemental Methods).
Drug loading capacity of 2,4-D Na-loaded gelatin microspheres 0.01 g gelatin microspheres were dispersed in 5 mL of deionized water and sonicated for 30 min in order to fully break the microspheres, then the absorbance of the supernatant was measured at 282 nm. The experiment was repeated three times to obtain the mean value and standard deviation. The drug loading capacity (Q, mg/g)) and the embedding ratio (R) were subsequently calculated using Eqs. (1) and (2), respectively (Maryam et al. 2017).
where C is the concentration of 2,4-D Na in the supernatant (mg/mL), V is the volume of the supernatant solution (mL), M is the mass of the drug-loaded microspheres (mg), and M 0 is the initial mass of input 2,4-D Na (mg).
Sustained release of 2,4-D Na in water at different temperatures 0.2 g of each microsphere sample (2,4-D Na-and 2,4-D Na-NH 4 HCO 3 -loaded gelatin microspheres) was transferred to dialysis bag containing 8 mL of deionized water, subsequently placed in a water bath (200 mL) at 25, 35, and 45°C. After agitation for 12 h, 1 mL of releasing solution was taken out for measurement then replace with an equivalent amount of distilled water. The process was repeated every 2 h first and every 12 h beyond 24 h. The experiment was repeated three times to obtain the mean value and standard deviation. The cumulative release rate of 2,4-D Na (%) was calculated using Eq. (3).
where Cq is the concentration of 2,4-D Na in the release medium at time n (mg/mL), Vq is the volume of the release solution (mL), Q is the drug loading capacity of carriers with respect to 2,4-D Na (mg/g), and M is the mass of the drug-loaded microspheres.

Bioassay experiment
Cucumber seeds (Tangshan Fengyou Agricultural Technology Co., LTD.) were soaked at 25°C for 4 h. The sterilized filter paper was placed in each petri dish, then adding 8 mL of sterilized water and 15 seeds, keeping at 25°C with 14 h light (13,200 lx)/10 h dark cycles. After 3 days, the germinated seeds with the similar growth rate were selected. Subsequently, 0.3 g of 2,4-D Na-or 2,4-D Na-NH 4 HCO 3 -loaded gelatin microspheres were added and incubated at 30°C. The root length and fresh weight were measured at the 5th day. The experiment was repeated three times to obtain the mean value and standard deviation.

Results and discussion
Characterization of the 2,4-D Na-loaded gelatin microspheres

Particle size
The particle size of 2,4-D Na-loaded gelatin microspheres becomes larger after loading drugs (Supplementary Table 1). The addition of 2,4-D Na during emulsification increases the diameters of the water/oil droplets, which weakens the volume reduction effect of the freeze-drying process.

FTIR analysis of 2,4-D Na-loaded gelatin microspheres
The surfaces of the analyzed materials comprise various groups (Fig. 1) via FTIR spectra. The spectrum of gelatin microspheres does not have a peak corresponding to benzene. Comparatively, 2,4-D Naloaded gelatin microspheres present a characteristic peak of benzene rings in the range of 750-800 cm -1 , which is similar to the peak observed in the 2,4-D Na Fig. 1 Infrared spectra of 2,4-D Na-loaded gelatin microspheres and unloaded gelatin microspheres in the range of 700-950 cm -1 spectrum, indicating that 2,4-D Na had been successfully loaded in the gelatin microspheres.

SEM analysis
The morphology of unloaded gelatin microspheres remains intact (Fig. 2a). However, the 2,4-D Naloaded gelatin microspheres exhibit a rough and porous morphology (Fig. 2b). Such variation might be attributed to the sublimation of surface and internal moisture in the drug granules coated with water/oil emulsiondroplets, during the process of freeze-drying. Moreover, the variation confirms the successful loading of the herbicide.
Drug loading capacity of 2,4-D Na-loaded gelatin microspheres The drug loading capacity and embedding ratio of 2,4-D Na-NH 4 HCO 3 -loaded gelatin microspheres were calculated at different concentrations of NH 4 HCO 3 (0, 3, and 6%). As shown in Table 1, the drug loading capacity of the gelatin microspheres loaded with 10 mg/mL 2,4-D Na only (no NH 4 HCO 3 ) is 77.95 mg/ g. This capacity decreases when both, NH 4 HCO 3 and 2,4-D Na are loaded into the microspheres, and it decreases even further when the concentration of NH 4 HCO 3 is increased, probably due to the limited availability of hollow space in the microspheres. The lowest drug loading capacity is 39.1 mg/g, which is higher than the common clinoptilolite and montmorillonite nanoparticles (Tomar et al. 2015) due to the cavity inside the gelatin microspheres improving encapsulation of the 2,4-D Na.
Similarly, the embedding ratio of gelatin microspheres co-loaded with NH 4 HCO 3 and 2,4-D Na is less than that of the microspheres loaded with 10 mg/mL 2,4-D Na only (61.76%). At higher concentrations of NH 4 HCO 3 , the amount of microsphere-embedded 2,4-D Na is further decreased.
Sustained release of 2,4-D Na in water at different temperature The results (Fig. 3) of sustained release behaviour of 2,4-D Na-and 2,4-D Na-NH 4 HCO 3 -loaded gelatin microspheres demonstrate that both temperature and NH 4 HCO 3 content have a significant effect on the sustained release of 2,4-D Na from gelatin microspheres. Obviously, the rate of release increases at higher temperatures and higher concentrations of NH 4 HCO 3 .
The release rate profiles of 2,4-D Na-loaded gelatin microspheres may be divided into three stages (Fig. 3a). First, the drug release process is relatively fast, with 5.70% 2,4-D Na released within 1 h of dialysis at 35°C. In the second stage, the release rate slows down significantly, and it reaches 12.86% by the 12th h. The decreasing rate may be attributed to the absorption of water by the microspheres. Finally, the amount of the released herbicide peaks at around the 9th day. This indicates that the stability of the microspheres can be maintained during 10 days, despite a low melting point of gelatin. Owing to cross-linking between gelatin and formaldehyde, 2,4-D Na-loaded gelatin microspheres's release period can be up to 10 days, and it's longer than some of the 2,4-D Na-loaded zeolite and bentonite minerals (Shirvani After 1 day of dialysis, the release rate of the herbicide from 2,4-D Na-3% NH 4 HCO 3 -loaded gelatin microspheres reaches 22.56% at 45°C, 12.35% at 35°C, and 11.17% at 25°C. Considering that the melting temperature (Tm) of gelatin is normally below 35°C, the increased drug release at 35°C and 45°C compared to 25°C may be attributed to the melting of the gelatin, as well as to the decomposition of NH 4 HCO 3 . However, due to cross-linking between  gelatin and formaldehyde, the degradation of gelatin microspheres is slow. Therefore, the faster release at higher temperatures is mainly a consequence of greater NH 4 HCO 3 decomposition. Similar to 2,4-D Na-3% NH 4 HCO 3 -loaded gelatin microspheres, the release rate of microspheres loaded with 6% NH 4 HCO 3 increases significantly with increasing temperature during the first day of dialysis (Fig. 3c). At 35°C and 45°C, the release rate is 56-57% after 6 and 8 h of dialysis, respectively. However, at 25°C, 53.52% release rate is reached after much longer time (4 days). This confirms that the prepared gelatin microspheres exhibit controlled release performance. Notably, when 6% NH 4 HCO 3 is added to the microspheres, the drug is completely released within 12 h.
By comparing the release rate profiles of different materials at 35°C (Fig. 3d), it is evident that the addition of NH 4 HCO 3 achieves adjustable release of 2,4-D Na from the loaded microspheres. 2,4-D Na NH 4 HCO 3 -loaded gelatin microspheres show the adjustable temperature-response release, which stands out from some other traditional carriers such as activated carbons (Yang et al. 2019). NH 4 HCO 3 concentration varied formulations may provide various selections according to different application scenarios. As a foaming agent, NH 4 HCO 3 can produce CO 2 and NH 3 , both of which promote plant growth and induce the deconstruction of microspheres, resulting in drug release.
Only 20-60% release rate could be achieved using any of the prepared materials after 10 days. This is probably due to the incomplete dissolution of microspheres at the end point. In practical application, 2,4-D Na residues inside the microspheres may damage the crops after reaping or cause non-point pollution. Since the microspheres co-loaded with NH 4 HCO 3 and 2,4-D Na exhibit more efficient drug release behavior than those loaded with 2,4-D Na only, it is suggested that the addition of NH 4 HCO 3 is possible to achieve a more precisely controlled pesticide release during growing season.

Bioassay experiment
Knowing that 2,4-D Na can kill dicotyledonous weeds, cucumber was selected as a model dicotyledonous weed to test the weeding ability of 2,4-D Na-and Na-NH 4 HCO 3 -loaded gelatin microspheres. The obtained results indicate that both materials inhibit the growth of the cucumber seedlings and decreasing root length (Fig. 4) (Supplementary Table 2). At higher concentrations of NH 4 HCO 3 the inhibition effect is observed at earlier times (Fig. 4b). As shown in Fig. 4d, the seedlings in Groups 3 and 4 are almost dead, which proves that the tested microspheres can sustainably release the herbicide and exhibit potent weeding capacity. Compared to the control group (Group 1), the fresh weight and main root length of the cucumber seedlings in the groups treated with 2,4-D Na-, 2,4-D Na-3% NH 4 HCO 3 , and 2,4-D Na-6% NH 4 HCO 3loaded gelatin microspheres (Groups 2, 3, and 4, respectively) are significantly decreasing on the 5th day. The greatest decrease in weight and root length is observed for the cucumber seedlings in Group 4, probably due to the effect of quick NH 4 HCO 3 decomposition in promoting faster herbicide release.

Conclusion
In this study, 2,4-D Na-and 2,4-D Na-NH 4 HCO 3loaded gelatin microspheres were prepared by means of emulsification and cross-linking. Based on experimental analysis and calculations, the drug loading capacity and embedding ratio of the prepared materials are up to 77.95% and 61.7%, respectively. Moreover, both formulations show sustained and temperature-responsive release behaviour. However, the gelatin microspheres containing NH 4 HCO 3 exhibit higher release rates. The bioassay results prove that the release of 2,4-D Na herbicide from the microspheres can be tuned by adjusting the content of NH 4 HCO 3 . Overall, the data reported in this study provide a good basis for the development of pesticide formulations that are suited to specific weed-control needs. It provides experimental basis for the research and development of new temperature-responsive sustained-release herbicide formulations and decrease the negative effects of herbicides application in the future. Further studies are needed to assess the effect of the prepared formulations on herbicide residues in the soils and to evaluate the permissible limit of these formulations. Fig. 4 Inhibitory effects of the tested formulations on cucumber seedlings at 33°C after a 1 b 2 c 3 d 5 days of treatment. The numbers À-ˆrefer to different groups. Group 1 is blank control; Groups 2, 3 and 4 are treated with 2,4-D Na-, 2,4-D Na-3% NH 4 HCO 3 and 2,4-D Na-6% NH 4 HCO 3 -loaded gelatin microspheres, respectively