Optimization of the Nanoparticle Yield and Physicochemical Properties by a Structured Study of Nanoprecipitation Process Parameters

Abstract

Nanoprecipitation is a widely used and highly effective method for synthesizing polymer nanoparticles, especially in pharmaceutical and biomedical applications, due to its reproducibility and ease of use. However, controlling the nanoparticle yield, size, size distribution, and surface charge remains a significant challenge. To overcome these challenges and optimize the nanoprecipitation method, it is critical to understand the effects of process parameters on the yield and physicochemical properties of nanoparticles. To address these challenges, we systematically investigated five critical process parameters: surfactant (Pluronic F-127) concentration, stirring speed, solvent evaporation duration, organic-phase flow rate, and ultrasonication time. These parameters were optimized to enhance the yield of poly­(d,l-lactic-co-glycolic acid) (PLGA) nanoparticles while maintaining controlled physicochemical properties. Twenty different process conditions were evaluated for the yield, hydrodynamic size, polydispersity index (PDI), and ζ-potentials. Further, empirical modeling using a nonlinear polynomial fit was performed to fit the reaction yield and physicochemical properties. Empirical second- and third-order polynomial models described (R ² > 0.90 for most parameters) the process parameter response relationships with the reaction yield, size, PDI, and ζ-potential sufficiently well. Finally, we identified that process condition PC10 (7.5 h solvent evaporation, 1500 rpm stirring speed, 0.01 g/mL Pluronic F-127 (PF-127), 0.3 mL/min organic-phase flow rate, no ultrasonication) achieved a yield of 90.35%, a 9-fold improvement over the conventional method (10.06 ± 0.67%). This work provides a quantitative framework for rational PLGA nanoparticle production, enabling high-yield, monodisperse, and stable formulations, and supporting future translational pipelines.

Introduction

Polymer nanoparticles have been employed in various biomedical applications, including drug delivery, , in vivo bioimaging, diagnostics, gene therapy, tissue engineering, vaccine delivery, and cancer therapy. Drug-loaded nanoformulations address the limitations of traditional medicine by enabling controlled and sustained drug release, targeted delivery, improved bioavailability, enhanced drug stability, better penetration across biological barriers, and improved therapeutic outcomes. − In particular, polylactic acid (PLA) has served as the backbone for polymeric drug carriers. It contains an asymmetric α-carbon, commonly designated as the d or l form according to stereochemical nomenclature, or alternatively as the R and S forms, respectively. The polymer PLA exists in two enantiomeric forms: poly-d-lactic acid (PDLA) and poly-l-lactic acid (PLLA). ,

Poly-d,l-lactic-co-glycolic acid (PLGA) is a copolymer in which the d and l-lactic acid units are present in different ratios, including 50:50 and 85:15. This versatile material can be molded into virtually any shape or size, and it has the ability to encapsulate molecules of varying sizes. Additionally, PLGA exhibits solubility in a wide range of solvents, including chlorinated solvents, tetrahydrofuran, acetone, and ethyl acetate. It degrades in water via the hydrolysis of its ester bonds. The presence of methyl side groups in PLA makes it more hydrophobic than PGA; hence, lactide-rich PLGA copolymers are less hydrophilic, absorb less water, and degrade more slowly than glycolide-rich copolymers. , Hydrolysis of PLGA leads to changes in properties that are typically considered stable in solid formulations, including glass transition temperature (Tg), moisture content, and molecular weight. , The impact of these polymer characteristics on the rate of drug release from biodegradable polymeric matrices has been extensively investigated. The change in the PLGA characteristics during polymer biodegradation affects the release rates of the encapsulated drug molecules. The mechanical strength of PLGA is affected by physical properties, including the molecular weight and the L to G ratio. These properties also affect the ability to formulate them as drug delivery devices and may control the device’s degradation rate and hydrolysis. Recent studies have found, however, that the type of drug also plays a role in determining the release rate. Crystalline PGA, when copolymerized with PLA, reduces the degree of crystallinity of PLGA and, as a result, increases the rate of hydration and hydrolysis. As the PGA content increases, the degradation rate of PLGA increases, with a 50:50 PLA/PGA ratio demonstrating the most rapid breakdown. However, when PGA exceeds 50%, the degradation process decelerates due to enhanced polymer matrix stability, prolonging the degradation period.

The cornea is the major refractive element of the eye, and its transparency is critical for perfect vision. The posterior monolayer of the corneal endothelium controls corneal transparency via a pump-leak mechanism owing to the presence of tight junctions between cells. Hypothermic storage of the donor corneal endothelium before transplantation disrupts the tight junctional barrier integrity and results in the loss of transparency. Pharmacological pretreatment of the donor corneal endothelium with microtubule stabilizers is a clear choice against potential adverse effects induced by hypothermia. To achieve this, a sustained intracellular release of microtubule stabilizers from PLGA nanoparticles has been demonstrated to protect the tight junctional integrity of the donor corneal endothelium under prolonged cold storage by our group. Moreover, a sustained intracellular delivery will also resist microtubule disassembly that might be induced by cytokines in the event of allograft rejection after transplantation of the cornea. Sustained intracellular delivery of microtubule stabilizers was achieved using PLGA nanoparticles, which are employed in several FDA-approved formulations because of their biocompatibility and biodegradability. , The nanoparticles were prepared by using PLGA via a nanoprecipitation method. However, the conventional nanoprecipitation method employed by us resulted in a very low % yield of nanoparticles, which has also been reported in the literature (15% w/w).

To address this issue, the parameters associated with the nanoprecipitation process, viz., solvent evaporation duration, stirring speed, Pluronic F-127 (PF-127) concentration, ultrasonication of the emulsion, and flow rate of the organic phase, were optimized. The focus of this study was to understand the effect of each of these parameters on the nanoparticle yield, to achieve maximum yield through optimization of parameters, and to develop an empirical formula to describe the reaction yield of PLGA nanoparticles. This proposed empirical relationship will serve to precisely portray the corresponding yield on the basis of the adjustments made to these parameters. The intention was to rely on this empirical formula as a valuable guide in the pursuit of achieving the highest possible yield of PLGA nanoparticles, thereby reducing the overall loss.

Experimental Section

Materials

Poly­(lactic-co-glycolic) acid (PLGA; lactide and glycolide ratio: 50:50; MW 07–17 kDa; CAS No: 26780-50-7; inherent viscosity: 0.16–0.24 dL/g; end group: acid; nature: hygroscopic) was purchased from Nomisma Healthcare, India. Pluronic F-127 (PF-127) was procured from Sigma-Aldrich, India. Acetone (AR) was purchased from Molychem, India. Beaker (volume 50 mL and 3 cm diameter), polytetrafluoroethylene-coated magnetic stir bar (20 × 5 mm), and ultrasonicator (LAB MAN scientific instruments; Serial No: L6807; Model No: LMUG-4; capacity: 4 L; frequency: 40 kHz; ultrasonic power: 100 W; tank volume: 300 × 155 × 100 mm³). Ultrapure, deionized water was obtained from the Center of Applied Research and Nanotechnology, Siddaganga Institute of Technology, Tumkur.

Synthesis of PLGA Nanoparticles Using the Nanoprecipitation Method

A foundational study by Fessi et al. (1989) reported a nanoprecipitation method for the production of polymeric nanoparticles. Based on this, PLGA nanoparticles were prepared by the conventional nanoprecipitation method (Table ) described by Thanuja et al.(2021). Briefly, an organic solution of 50 mg of PLGA (50:50) dissolved in 2 mL of acetone was added dropwise (at different flow rates, Table ) to 10 mL of an aqueous phase containing PF-127 at different concentrations (Table ), while stirring at different speeds (Table ). The resulting nanoemulsion was continuously stirred for different time periods (solvent evaporation duration, Table ) at room temperature until the complete removal of the organic solvent. This was followed by centrifugation at 15,000 rpm for 30 min at 4 °C. The obtained nanoparticle pellet was suspended in ultrapure water and washed three times with deionized water to remove any residual surfactant, which was confirmed using phosphomolybdic acid–poly­(vinyl alcohol)–glycine (PAG) colorimetric assay (Supporting Information) using a PF-127 standard curve. Finally, the PLGA nanoparticles were lyophilized for 48 h and stored at 4 °C under anhydrous conditions. The emulsion was ultrasonicated for different durations (the last four schemes are shown in Table ).

1. Process Parameters for the Preparation and Optimization of PLGA Nanoparticles Using the Nanoprecipitation Method .

	optimized process conditions
process condition names	PF-127 Concentration (g/mL)	stirring speed (rpm)	solvent evaporation duration (h)	organic-phase flow rate (mL/min)	ultrasonication time (min)
CM	0.01	1500	5	0.3
PC1	0.005	1500	5	0.3
PC2	0.02	1500	5	0.3
PC3	0.03	1500	5	0.3
PC4	0.05	1500	5	0.3
PC5	0.01	1000	5	0.3
PC6	0.01	1750	5	0.3
PC7	0.01	2000	5	0.3
PC8	0.01	2300	5	0.3
PC9	0.01	1500	2.5	0.3
PC10	0.01	1500	7.5	0.3
PC11	0.01	1500	10	0.3
PC12	0.01	1500	15	0.3
PC13	0.01	1500	5	0.16
PC14	0.01	1500	5	0.45
PC15	0.01	1500	5	0.72
PC16	0.01	1500	5	0.3	15
PC17	0.01	1500	5	0.3	30
PC18	0.01	1500	5	0.3	40
PC19	0.01	1500	5	0.3	60

^aCM, conventional method; PC, process condition.

The present study involved a series of experiments exploring various process parameters, including the concentration of PF-127, stirring speed, solvent evaporation duration, organic-phase flow rate, and ultrasonication time. The impact of these factors on the reaction yield, mean Z-average particle size, polydispersity index (PDI), and ζ-potential was analyzed.

Reaction Yield of PLGA Nanoparticles

Nanoparticle preparation methods that achieve high reaction yields can minimize chemical waste and lower production costs. The reaction yield of the PLGA nanoparticles was calculated gravimetrically using formula .

$$\mathrm{r}\mathrm{eaction}\mathrm{yield}(\%)=\frac{\mathrm{actual}\mathrm{mass}\mathrm{of}\mathrm{nanoparticles}\mathrm{obtained}\mathrm{in}\mathrm{mg}}{\mathrm{theoretical}\mathrm{mass}\mathrm{of}\mathrm{nanoparticles}\mathrm{in}\mathrm{mg}}\times 100$$ 1

The actual mass of the nanoparticles obtained did not include the residual surfactant.

Characterization of Size, PDI, and ζ-Potential of the Nanoparticles

The Z-average size, PDI, and ζ-potential of PLGA nanoparticles were analyzed using Dynamic Light Scattering analysis (ZSNano90, Malvern, United Kingdom; 90° scattering angle; 0.8872 cP viscosity and 1.330 refractive index at 25 °C). Briefly, the nanoparticles were dispersed to a final concentration of 1 mg/mL in deionized water (pH 7.0) prior to measurements. The particle size and PDI were measured using polystyrene macrocuvettes (5100022) at 25 ± 1 °C, while ζ-potential measurements were performed using disposable folded capillary cells (Malvern, DTS1070). All measurements were conducted at 25 ± 1 °C and performed in triplicate. The nanoparticles obtained under PC10 were redispersed in deionized water, and a drop of the suspension was placed onto a clean aluminum stub and allowed to dry at room temperature (25 °C). The dried samples were mounted using conductive carbon tape to ensure adequate electrical conductivity. The surface morphology of the nanoparticles was examined using field emission scanning electron microscopy (FE-SEM; SUPRA 55, Carl Zeiss, Germany) operated at an accelerating voltage of 5–10 kV under high vacuum conditions.

Data Analysis

The Graph Pad Prism 10.2.3 software was used for statistical analysis and plotting graphs. An ordinary one-way ANOVA (Tukey’s post hoc test) analysis was conducted on the mean ± SE (n = 3) for comparison with the conventional method. *p < 0.05, **p < 0.01, and ***p < 0.001 indicate significance. The second- and third-order polynomial equations were employed using OriginPro 8.5 to determine the effect of independent variables on dependent variables. A one-factor-at-a-time (OFAT) approach was employed to evaluate the effect of each process parameter independently.

Results and Discussion

Improving the yield and quality of nanoparticles requires a mechanistic understanding of process parameter behavior, as well as the extent of their variability and impact on the nanoprecipitation process. Most of the experimental designs focus on altering one variable at a time while keeping all other variables constant. This type of experimentation has several limitations: it is often unreliable, time-consuming, demands a large number of runs and significant resources, and may also give false data because it ignores the interaction effects between the investigated process parameters. Empirical models relate the inputs and outputs by empirical correlations that fit reasonably well with the experimental results and are reliable only within the region of input–output of these experimental results. Because of the complexity of the physical and chemical phenomena involved, each nanoparticle system should first be characterized precisely before carrying out the above-mentioned empirical modeling. Therefore, the desired physicochemical characteristics and yield of the nanoparticles can be obtained by optimizing process parameters. Hence, the present study focused on optimizing the process parameters of the nanoprecipitation method involved in the preparation of PLGA nanoparticles.

A total of 20 different process conditions were used to optimize PLGA nanoparticles, involving five independent variables (concentration of PF-127, stirring speed, solvent evaporation duration, organic-phase flow rate, and ultrasonication time) and one dependent variable (% yield, size, PDI, or ζ-potential). An empirical model depicting the effect of each independent variable on the dependent variable was generated by fitting the experimental data of % yield, size, PDI, and ζ-potential to a polynomial equation of the order 2 or 3. These equations represent a univariate empirical polynomial fit describing the relationship between individual process parameters and response variables. These fits are descriptive within the studied experimental range and are not intended for multivariate prediction.

Effects of Process Parameters on the Nanoparticle Yield

The effects of various process parameters, such as PF-127 concentration, stirring speed, solvent evaporation duration, organic-phase flow rate, and ultrasonication time on the % yield of PLGA nanoparticles are illustrated in Figure . The conventional method (CM) resulted in a reaction yield of 10.06 ± 0.67% under the process parameters described in Table .

1.

Effect of process parameters on the reaction yield (%) of PLGA nanoparticles. (a) PF-127 concentration, (b) stirring speed, (c) solvent evaporation duration, (d) organic-phase flow rate, and (e) ultrasonication time. Data are expressed as mean ± standard deviation (n = 3) with significant differences compared to the conventional method (CM) indicated by *p < 0.05 and ***p < 0.001 via ordinary one-way ANOVA (Tukey’s post hoc test).

With an increase in the PF-127 concentration, the yield increased, reaching a maximum of 69.86 ± 0.065% at 0.03 g/mL, followed by a subsequent decrease. The use of 0.005 g/mL PF-127 did not significantly affect the reaction yield compared to the conventional method (0.01 g/mL; p = 0.93; Figure a). In contrast, higher concentrations of PF-127 (0.02, 0.03, and 0.05 g/mL) led to a significant increase in % yield (6.40-, 6.94-, and 2.91-fold, respectively; p < 0.001), as shown in Figure a.

PF-127, a nonionic surfactant and stabilizer, plays a crucial role in nanoparticle formation because of its amphiphilic nature. At optimal concentrations, PF-127 prevents nanoparticle aggregation and enhances colloidal stability, thereby promoting uniform particle formation and improving reaction yield. Additionally, PF-127 forms micelles in aqueous media, and higher concentrations can lead to increased micelle formation, which may further improve the nanoparticle formation and yield. In corroboration with previous studies, the present study observed that increasing the PF-127 concentration initially led to an improved reaction yield. However, beyond a certain threshold, a further increase in PF-127 concentration resulted in a decline in yield. This reduction may be attributable to excessive PF-127 adsorption on the nanoparticle surface, which promotes flocculation, aggregation, and loss of particles during purification or washing steps.

A stirring speed of 1750 rpm resulted in the highest reaction yield of 85.26 ± 1.63%; however, a subsequent increase in the stirring speed reduced the reaction yield. The stirring speed of 1000 rpm did not produce a yield whose difference from that of the conventional method (1500 rpm) was statistically significant (p > 0.99; Figure b). In contrast, increasing the stirring speed to 1750, 2000, and 2300 rpm significantly enhanced the reaction yields by 8.47-, 7.79-, and 4.27-fold, respectively (p < 0.001; Figure b).

Increasing the stirring speed may enhance the reaction yield by improving the emulsion quality, mass transfer, and particle stability. Faster stirring may create fine emulsions with a larger surface area, facilitating more efficient solvent evaporation, polymerization, and also rapid reactant diffusion, which improves the reaction kinetics and yield. Additionally, vigorous stirring reduces nanoparticle aggregation and promotes a uniform particle size, minimizing losses during purification and enhancing the overall yield. Our study found that increasing the stirring speed initially led to an improvement in the reaction yield. However, at higher speeds of 2000 and 2300 rpm, a significant decrease in the yield was observed. This may be attributable to high stirring speed disrupting localized concentration gradients, poor mixing and fragmentation, or incomplete polymerization.

The duration of solvent evaporation had a similar effect; the reaction yield reached a maximum of 90.35 ± 0.89% at 7.5 h. Compared to the conventional method (5 h), the evaporation durations significantly enhanced the reaction yield by 2.77-, 8.98-, 7.75-, and 3.56-fold, respectively (p < 0.001; Figure c). Solvent evaporation duration plays a crucial role in nanoparticle formation and directly impacts the reaction yield, aligning with our findings that increased stirring speed also enhances the reaction yield.

Efficient solvent removal promotes uniform solidification and well-defined polymer nanoparticle morphology. Moderately slow evaporation allows sufficient time for polymer interactions, thereby improving the reaction yields. An optimized evaporation duration helps to prevent premature aggregation or phase separation. These factors helped to increase the reaction yield. However, an extended evaporation time, such as 15 h in our study, was associated with reduced yield. This may be due to excessive polymer hardening, accelerated phase separation, and potential degradation of the polymer or surfactant, leading to particle aggregation and a decreased yield.

The organic-phase flow rate of 0.45 mL/min produced a maximum yield of 63.56 ± 5.13; however, further increases in the flow rate led to a noticeable decline in the yield. Compared to the conventional method (0.3 mL/min), 0.16 mL/min did not significantly affect the reaction yield (p = 0.82; Figure d). However, increasing the flow rates to 0.45 and 0.72 mL/min resulted in substantial enhancement in the reaction yield by 6.31- and 3.90-fold, respectively (p < 0.001; Figure d).

An optimized organic-phase flow rate promotes efficient emulsification and uniform particle size distribution, enhances reproducibility, and improves the yield. When the polymer-containing solvent is introduced into aqueous media, rapid mixing occurs due to interfacial turbulence, resulting in uniform emulsification. A well-balanced ratio of organic phase and aqueous phase obtained by controlling the flow rate of the organic phase ensured sufficient interaction time between the polymer chains and the nonsolvent, facilitating controlled nucleation and particle growth, which ultimately contributed to a higher reaction yield. The current study observed that an organic-phase flow rate of 0.45 mL/min resulted in the maximum reaction yield. Further increases in the flow rate led to a decline in the yield. This may be due to poor mixing and unstable nucleation processes.

Compared to the conventional method (no ultrasonication), an increase in the ultrasonication time (30, 40, and 60 min) significantly increased the reaction yield by 2.76- (p < 0.001), 3.02- (p < 0.001), and 1.66-fold (p < 0.05), respectively (Figure e), while reaching a maximum of 30.44 ± 0.81% at 40 min. Ultrasonication for 15 min did not show a statistically significant difference compared to the conventional method (p = 0.26; Figure e). Ultrasonication enhances mixing and mass transfer by inducing the formation and collapse of vapor bubbles within the liquid medium. This process facilitates efficient emulsification followed by uniform dispersion of nanoparticles within the polymer matrix and minimization of agglomeration, thereby contributing to increased reaction yield. Our study demonstrated that increasing the ultrasonication time initially improved the yield; however, prolonged sonication for 60 min resulted in a decline in the yield. Excessive sonication time may lead to breakage of polymer chains, reduction in molecular weight, and polymer degradation.

Collectively, negligible residual solvent in the nanoparticles (assumed) after solvent evaporation, negligible mass of PLGA nanoparticles (estimated), and residual surfactant in the supernatant (estimated) after washing are assumed to not contribute to the total mass of PLGA nanoparticles obtained that was used to calculate the yield. Hence, a comprehensive mass balance accounting analysis was not performed in the present study. This may be considered a limitation and could be addressed in future studies focusing on process-scale validation and recovery efficiency.

Effects of Process Parameters on the Nanoparticle Size

Figure illustrates the impact of various process parameters (PF-127 concentration, stirring speed, solvent evaporation duration, organic-phase flow rate, and ultrasonication time) on the size of the PLGA nanoparticles. Each of these factors significantly influences the particle size, underscoring their importance in optimizing nanoparticle formation. Using the conventional method (CM), a particle size of 77.81 ± 0.35 nm was obtained under the process conditions outlined in Table . Interestingly, a very low concentration of PF-127 (0.005 g/mL) produced relatively large nanoparticles (258.49 ± 2.95 nm), which were larger than those obtained using the conventional method.

2.

Effect of process parameters on the size of PLGA nanoparticles. (a) PF-127 concentration, (b) stirring speed, (c) solvent evaporation duration, (d) organic-phase flow rate, and (e) ultrasonication time. Data are expressed as mean ± standard deviation (n = 3) with significant differences compared to the conventional method (CM) indicated by ***p < 0.001 via ordinary one-way ANOVA (Tukey’s post hoc test).

Further increasing the PF-127 concentration to 0.2 g/mL resulted in an increase in the particle size to 140.9 ± 4.43 nm (Figure a). However, a subsequent increase in PF-127 concentration led to a notable reduction in the particle size. However, the nanoparticle size ranged from 100 to 200 nm under all process conditions. Compared to the conventional method (0.1 g/mL), PF-127 at concentrations of 0.005, 0.02, 0.03, and 0.05 g/mL significantly improved the particle size by 3.32-, 1.18-, 1.57-, and 1.30-fold, respectively (p < 0.001; Figure a).

PF-127 has a relatively high hydrophilic–lipophilic balance (HLB), which leads to the formation of large micelle cores during nanoparticle synthesis. These micelles act as templates and influence the formation of larger particles. Furthermore, although PF-127 provides steric stabilization, at higher concentrations it may promote aggregation, which can result in the formation of larger particles. On the contrary, at very low concentrations, PF-127 is unable to form micelles effectively. This results in insufficient steric stabilization and incomplete coating of nanoparticle surfaces, which in turn promotes particles to aggregate and grow, ultimately producing larger particle sizes. , In the current study, we observed that very low concentrations of PF-127 increased the particle size. 0.02 g/mL PF-127 was found to be the optimum concentration since the nanoparticle size obtained ranged between 100 and 200 nm, indicating that this concentration was ideal for our desired application. Further reduction in the size at higher PF-127 concentrations was also observed, which may be attributed to enhanced micelle formation, increased viscosity, and improved steric stabilization.

The stirring speed had a significant influence on the particle size. From the conventional method (1500 rpm), a particle size of 77.81 ± 0.35 nm was obtained under the process conditions outlined in Table , but a much larger average particle size of 204.41 ± 7.70 nm was obtained at a lower stirring speed of 1000 rpm. Doubling the stirring speed to 2000 rpm resulted in an increase of the particle size to 133.35 ± 3.65 nm (Figure b). A subsequent increase in the stirring speed beyond this point led to a gradual reduction in the particle size. Compared to the conventional method (1500 rpm), stirring at 1750 and 2000 rpm resulted in a significant increase in the particle size by 1.42- and 1.71-fold, respectively (p < 0.001; Figure b). In contrast, compared to the conventional method, a further increase in the stirring speed to 2300 rpm did not show any significant changes in the nanoparticle size (p = 0.211; Figure b).

A higher stirring speed increases the collision frequency between the reactant molecules, generating more nucleation sites and promoting the formation of a larger number of nuclei. As a consequence, each nucleus has less material available for growth, leading to smaller particle sizes. Additionally, higher stirring speeds introduce stronger shear forces, which can inhibit the formation of nanoparticles and prevent agglomeration. Finally, this helps in the formation of smaller and more uniformly dispersed nanoparticles. ,, In our study, we observed that low stirring speeds also resulted in larger particle sizes; this may be due to reduced collision frequency between reactant molecules and insufficient shear forces, which is in corroboration with previous studies.

The solvent evaporation duration had a direct impact on controlling the particle size. A particle size of 77.81 ± 0.35 nm was obtained using the conventional method (5 h), as detailed in the process parameters listed in Table . Interestingly, a shorter solvent evaporation duration (2.5 h) produced a larger particle size, measuring 247.83 ± 5.89 nm (Figure c). Increasing the solvent evaporation duration to 10 h produced a relatively large particle size of 152.94 ± 4.98 nm as compared to the conventional method. However, further increase in the solvent evaporation duration led to a significant decrease in the particle size (99.76 ± 1.18 nm). Compared to the conventional method (5 h), solvent evaporation for 7.5 and 10 h significantly increased the particle size by 1.78- and 1.96-fold, respectively (p < 0.001; Figure c). In contrast, a solvent evaporation duration of 15 h did not significantly affect the particle size compared to the conventional method (5 h; p = 0.201; Figure c).

The optimal solvent evaporation duration depends on factors such as the type of polymer, solvent volatility, and stirring conditions. Prolonged evaporation times may allow for controlled solidification, improved surfactant action, and enhanced polymer chain entanglement, which can result in smaller and less uniform particle sizes. Similarly, a short solvent evaporation duration leads to increased particle size due to incomplete solidification, limited surfactant action, residual solvent effects, and thermal and mechanical stresses. The present study highlighted all of these phenomena; a short solvent evaporation time of 2.5 h resulted in larger particle sizes, while an extended evaporation duration of 15 h produced smaller particles compared to the conventional method.

The organic-phase flow rate of 0.45 mL/min produced a relatively large particle size, with an average size of 132.54 ± 2.94 nm. As the organic-phase flow rate increased, the particle size progressively decreased. Compared to the conventional method (0.3 mL/min), the organic-phase flow rates of 0.16, 0.45, and 0.72 mL/min resulted in significantly increased particle size by 1.60-, 1.70-, and 1.46-fold, respectively (p < 0.001; Figure d). However, the particle sizes obtained at four different flow rates were well within the desired size range of 100 to 200 nm.

A higher organic-phase flow rate leads to slower and less efficient mixing, and hence, it allows polymer chains more time to aggregate before solidifying, resulting in large particle sizes. Additionally, the rapid injection of the organic phase quickly dilutes in the aqueous phase, reducing the supersaturation levels and promoting excessive nucleation, leading to the formation of larger particles. The present study demonstrated that an increase in the organic-phase flow rate produces larger particle sizes. However, at an optimal flow rate (0.45 mL/min), the desired particle size was obtained. This may be attributed to fast laminar mixing, which arrests particle growth early and minimizes aggregation.

Compared to the conventional method (no ultrasonication; 77.81 ± 0.35 nm), ultrasonication of the emulsion for 15, 30, 40, and 60 min produced the desired particle size between 100 and 200 nm with an increase in the particle size by 1.35-, 1.59-, 1.71-, and 1.38-fold, respectively (p < 0.001; Figure e). However, the particle size decreased beyond the optimal ultrasonication time (40 min).

Recent studies have highlighted that shorter ultrasonication times result in larger particle formation, but prolonged ultrasonication can cause degradation or agglomeration of the particles. In this study, we found that longer ultrasonication contributed to larger particle sizes, which may be attributable to polymer chain degradation, thermal effects, and particle agglomeration. This study also highlighted that extended ultrasonication resulted in a smaller particle size. Prolonged sonication may generate microscopic bubbles that collapse, leading to the generation of shockwaves, producing strong mechanical forces and enhancing dispersion, which directly contribute to the formation of smaller particles.

Effects of Process Parameters on the Nanoparticle PDI

Figure illustrates the impact of process parameters, such as PF-127 concentration, stirring speed, solvent evaporation duration, organic-phase flow rate, and ultrasonication time, on the polydispersity index (PDI) of the PLGA nanoparticles. Under the process parameters specified in Table , the PLGA nanoparticles prepared using the conventional method exhibited a PDI of 0.354 ± 0.80, indicating their polydispersity.

3.

Effect of process parameters on the PDI of PLGA nanoparticles. (a) PF-127 concentration, (b) stirring speed, (c) solvent evaporation duration, (d) organic-phase flow rate, and (e) ultrasonication time. Data are expressed as mean ± standard deviation (n = 3) with significant differences compared to the conventional method (CM) indicated by *p < 0.05, **p < 0.01, ***p < 0.001 via ordinary one-way ANOVA (Tukey’s post hoc test).

With an increase in PF-127 concentration, a PDI of 0.147 ± 0.01 at 0.03 g/mL, representing a high degree of monodispersity, was initially observed, followed by a subsequent transition to polydispersity. Compared to the conventional method (0.01 g/mL), PF-127 concentrations of 0.005 (p = 0.79) and 0.05 g/mL (p = 0.11) did not significantly affect the PDI (Figure a) and remained polydisperse. In contrast, concentrations of 0.02 and 0.03 g/mL led to a significant decrease in PDI, by 1.54= and 2.40-fold, respectively, compared to the conventional method (p < 0.001; Figure a), resulting in monodispersity.

The PF-127 triblock copolymer consists of hydrophilic PEO and hydrophobic PPO blocks arranged in a PEO–PPO–PEO manner. At the optimal concentration, PF-127 forms stable micelles with a hydrophobic PPO core and a hydrophilic PEO shell. These stable micelles prevent aggregation and promote uniform nucleation and growth of nanoparticles, resulting in a narrow size distribution reflected by a low PDI. In the current study, the optimal PF-127 concentration was found to be 0.03 g/mL, with the lowest PDI of 0.14 ± 0.01. Similarly, low concentrations of PF-127 promote an increase in polydispersity due to poor surface coverage, aggregation, nonuniform nucleation, and variable particle growth. The current study observed both increasing and decreasing trends in PDI. Interestingly, excessive concentrations of PF-127 may cause a transition from monodispersity to polydispersity, which may be attributed to particle overcoating by PF-127, thermoreversible gelation, increased viscosity, and micelle overcrowding. ,

A stirring speed of 1750 rpm produced high monodispersity, as indicated by a PDI of 0.09 ± 0.04. However, a further increase in the stirring speed led to a transition to polydispersity. Low and high stirring speeds of 1000 (p > 0.26) and 2300 rpm (p > 0.99), respectively, did not significantly alter the PDI compared to the conventional method (1500 rpm; Figure b) and remained polydispersed. The moderate stirring speeds of 1750 (p < 0.001) and 2000 rpm (p < 0.05) significantly decreased PDI compared to the conventional method (1500 rpm; Figure b) and were found to have a very high degree of monodispersity (PDI < 0.2).

Stirring speed is a very important process parameter to achieve monodisperse PLGA nanoparticles, especially in the nanoprecipitation and emulsion-based methods. Moderate stirring speed maintains a homogeneous distribution of reactants, leading to simultaneous nucleation events, resulting in a uniform particle size. In our study, we observed polydispersity at a low stirring speed of 1000 rpm. Lower stirring may cause delayed nucleation and poor mixing of the reactants, which leads to increased nanoparticle aggregation. This may have contributed to variable particle growth rates and consequently to the polydispersity of the nanoparticles. At a high stirring speed of 2300 rpm, we also observed a polydispersity index of PLGA nanoparticles. An extremely high stirring speed may influence turbulent mixing and destabilization in the emulsions, leading to the fragmentation of droplets and inconsistent nanoparticle formation. Additionally, a higher stirring speed may not homogenize the organic and aqueous phases effectively, leading to heterogeneous nucleation and the generation of heat, which alters the solvent evaporation and polymer precipitation dynamics. These factors may significantly affect the polydispersity.

A solvent evaporation duration of 10 h resulted in a PDI of 0.103 ± 0.0045, indicating high monodispersity. However, a further increase in the evaporation time led to a transition toward polydispersity. Compared to the conventional (5 h) method, a shorter evaporation duration of 2.5 h did not show a statistically significant difference in PDI (p > 0.99; Figure c) and remained polydisperse. In contrast, evaporation durations of 7.5, 10, and 15 h resulted in significant reductions in PDI by 1.95- (p < 0.001), 3.43- (p < 0.001), and 1.41-fold (p < 0.05), respectively, compared to the conventional method. Furthermore, they all resulted in a high degree of monodispersity with PDI < 0.25 (Figure c).

Shorter solvent evaporation durations cause incomplete solvent removal, insufficient stabilization time, and trigger premature nucleation. These factors contribute to higher PDI values. In our study, we also observed that prolonged solvent evaporation led to the polydispersity of the nanoparticles. Extended solvent exposure to the organic phase may enhance particle collisions and fusion, particularly with stabilizers. This prolonged duration may influence polymer chain entanglement and diffusion, thereby altering particle morphology and size uniformity, ultimately impacting the PDI. Our findings suggest that a solvent evaporation time between 7.5 and 10 h is optimal.

The organic-phase flow rate significantly influences the PDI, particularly in microfluidic and emulsification-based synthesis methods. An organic-phase flow rate of 0.72 mL/min resulted in high monodispersity, as indicated by a PDI of 0.105 ± 0.002. Compared to the conventional method (0.3 mL/min), organic-phase flow rates of 0.45 and 0.72 mL/min resulted in a significant reduction in PDI by 1.71- and 3.37-fold, respectively (p < 0.001; Figure d) and a high degree of monodispersity (PDI < 0.20).

High flow rates promote rapid and uniform mixing, controlled nucleation, and reduced aggregation by shortening the residence time and minimizing the diffusion gradients. They also support a stable dripping regime, resulting in uniform droplet formation, which ultimately contributes to the high monodispersity of the polymer nanoparticles. This finding is in corroboration with our study. , We also observed that a slow organic-phase flow rate leads to polydispersity, likely due to inefficient mixing dynamics, unstable laminar flow, and increased particle aggregation, as reported by other studies. ,

Compared to the conventional method (PDI of 0.354 ± 0.02), ultrasonication of the emulsion for 15 and 60 min resulted in increased PDI values of 0.55 ± 0.007 (p < 0.001; Figure e) and 0.40 ± 0.01 (p < 0.05; Figure e), respectively, indicating a very high polydispersity. In contrast, ultrasonication of the emulsion for 30 and 40 min significantly decreased the PDI to 0.22 ± 0.003 and 0.21 ± 0.002, respectively (p < 0.001; Figure e) compared to the conventional method.

A short sonication time causes insufficient energy input, incomplete dispersion, and nonuniform cavitation. Together, these factors lead to particle agglomeration, resulting in a broad size distribution and, consequently, higher polydispersity. , Extended ultrasonication time can lead to uneven particle breakdown and excessive heat generation, which may destabilize emulsions and promote aggregation. Continuous cavitation can also cause structural damage to the particles, resulting in particle size variability and an increased PDI. In our study, we observed that both shorter and extended ultrasonication times produced high PDI. The current study suggested that an ultrasonication time of 30 to 40 min is optimal.

Effect of Process Parameters on the ζ-Potential of the Nanoparticles

ζ-potential is a key physicochemical parameter that influences nanoparticle stability and interactions with biological systems. , The effects of various process parameters, such as PF-127 concentration, stirring speed, solvent evaporation duration, organic-phase flow rate, and ultrasonication time on the ζ-potential of the PLGA nanoparticles are illustrated in Figure . Based on the process parameters detailed in Table , the ζ-potential of the PLGA nanoparticles synthesized using the conventional method was found to be −10.72 ± 0.8 mV.

4.

Effect of process parameters on the ζ-potential of the PLGA nanoparticles. (a) PF-127 concentration, (b) stirring speed, (c) solvent evaporation duration, (d) organic-phase flow rate, and (e) ultrasonication time. Data are expressed as mean ± standard deviation (n = 3) with significant differences compared to the conventional method (CM) indicated by *p < 0.05, **p < 0.01, ***p < 0.001 via ordinary one-way ANOVA (Tukey’s post hoc test).

As the concentration of PF-127 increased, a corresponding increase in the stability of the nanoparticles was observed with a maximum ζ-potential of −17.11 ± 0.13 mV g/mL. The reduced PF-127 concentration (0.005 g/mL) did not significantly affect the ζ-potential compared to the conventional method (0.01 g/mL; p > 0.97; Figure a). In contrast, higher PF-127 concentrations of 0.02, 0.03, and 0.05 g/mL resulted in significant increases in the ζ-potential by 1.18- (p < 0.05), 1.36- (p < 0.001), and 1.59-fold (p < 0.001), respectively, compared to the conventional method, as shown in Figure a.

PF-127 is a widely used nonionic surfactant that plays a crucial role in stabilizing polymer nanoparticles by controlling their surface properties, preventing aggregation, and influencing their size and dispersity. At high PF-127 concentrations, more polymer chains of the surfactant adsorb onto the surface of the nanoparticles, altering their surface charge environment. , The hydrophilic poly­(ethylene oxide) segments of PF-127 extend into the aqueous phase, forming a stabilizing layer that affects the electrokinetic potential. As the concentration increases, PF-127 tends to form micelles, which can interact with or encapsulate nanoparticles, thereby modifying the surface charge distribution. Additionally, PF-127 has the ability to neutralize surface charges, potentially shifting the ζ-potential toward more negative or even positive values, depending on the system. These factors collectively contribute to changes in ζ-potential with increasing PF-127 concentration. It is evident from the current study that increasing the concentration of PF-127 led to a corresponding increase in ζ-potential, which is in correlation with the findings reported in previous studies.

From the conventional method (stirring speed of 1500 rpm), the ζ-potential obtained was −10.72 ± 0.8 mV under the conditions outlined in Table . An increase in the stirring speed led to a corresponding increase in the stability of the nanoparticles, with a speed of 2300 rpm resulting in the maximum ζ-potential of −18.95 ± 0.39 mV. Compared to the conventional method, a lower stirring speed of 1000 rpm noticeably reduced nanoparticle stability (−7.37 ± 0.26 mV; p < 0.001; Figure b). In contrast, higher stirring speeds of 1750, 2000, and 2300 rpm significantly improved nanoparticle stability, with an increase by 1.65- (p < 0.001), 1.19- (p < 0.001), and 1.76-fold (p < 0.001), respectively, compared to the conventional method (Figure b).

High stirring speeds produced more stable nanoparticles by promoting the uniform dispersion of surfactants and polymers, ensuring consistent surface coverage and improved charge stabilization. Rapid stirring reduces particle–particle interactions, helping maintain distinct surface charges and enhancing electrostatic repulsion. Additionally, a higher stirring speed improves the interactions between the charged components, leading to the development of a stronger surface charge. Furthermore, higher stirring speeds tend to produce smaller nanoparticles with a larger surface area to volume ratio, which amplifies the impact of the surface charge. In the current study, we observed that a higher stirring speed increased the stability of the nanoparticles.

Next, a solvent evaporation duration of 5 h (conventional method) resulted in a ζ-potential of −10.72 ± 0.8 mV. Prolonged solvent evaporation for 15 h resulted in an increase in the ζ-potential to −16.06 ± 0.07 mV. In contrast, a shorter evaporation time of 2.5 h led to reduced nanoparticle stability, with a ζ-potential of −6.44 ± 0.23 mV (p < 0.001; Figure c). On the other hand, extended evaporation durations of 7.5, 10, and 15 h significantly enhanced the nanoparticle stability, with corresponding increases in the ζ-potential by 1.43- (p < 0.001), 1.21- (p < 0.001), and 1.49-fold (p < 0.001), respectively, relative to the conventional method (Figure c).

The time allowed for solvent evaporation during the preparation of PLGA nanoparticles using the nanoprecipitation method significantly influences the ζ-potential of the nanoparticles. Increased solvent removal leads to tighter polymer packing, which reduces the particle size, increases the surface area, and alters the charge distribution. This may enhance particle stability by making the surface charge more strongly positive or negative. Extended solvent evaporation may lead to the removal of residual solvent that may mask the surface charge, revealing the true ζ-potential. In our study, we observed that extending the solvent evaporation time to 15 h resulted in a highly stable nanoparticle suspension. Conversely, we also observed that shorter evaporation times led to lower nanoparticle stability due to incomplete solvent removal, poor polymer packing, and poor colloidal stability. ,

The organic flow rate of 0.3 mL/min used in the conventional method yielded a ζ-potential of −10.72 ± 0.8 mV. A stable nanoparticle dispersion with a ζ-potential of −14.33 ± 0.33 mV was achieved at an organic-phase flow rate of 0.45 mL/min. However, increasing the flow rate to 0.72 mL/min led to reduced stability, as indicated by a lower ζ-potential of −9.74 ± 0.10 mV. Compared to the conventional method (0.3 mL/min), an organic-phase flow rate of 0.16 mL/min significantly decreased the nanoparticle stability (−6.95 ± 0.28; p < 0.01; Figure d). In contrast, a flow rate of 0.45 mL/min resulted in a 1.33-fold increase in stability relative to that of the conventional method, which was statistically significant (p < 0.05; Figure d). Meanwhile, a flow rate of 0.72 mL/min did not produce a statistically significant difference in stability compared to that of the conventional method (p > 0.52; Figure d).

A high organic flow rate promotes rapid mixing, which facilitates the formation of smaller particles and a more uniform distribution. This may enhance the surface charge exposure and potentially increase the ζ-potential, either positively or negatively. Additionally, increased stirring speed improves electrostatic repulsion, helping to reduce aggregation and enhance colloidal stability. In our study, we found that an extended and slower organic-phase flow rate reduced the stability of nanoparticles due to the formation of larger or uneven particles, surface charge masking, and increased aggregation. ,, The current study suggests that an organic-phase flow rate of 0. 45 mL/min is optimal.

Compared to the conventional method (−10.72 ± 0.80 mV), ultrasonication of the emulsion for 60 min led to enhanced nanoparticle stability, with a ζ-potential of −14.98 ± 0.49 mV (p > 0.001; Figure e). In contrast, ultrasonication of the emulsion for 15 min significantly decreased the nanoparticle stability, with a ζ-potential of −8.76 ± 0.17 mV (p < 0.05). Additionally, ultrasonication of 30 and 40 min showed no significant change in the ζ-potential (−11.18 ± 0.36 and −12.33 ± 0.51 mV, respectively) compared to the conventional method (p > 0.92 and p > 0.102, respectively), as indicated in Figure e. In this study, it was found that 60 min of ultrasonication was optimal for obtaining the highest ζ-potential.

Ultrasonication significantly improves the dispersion and colloidal stability of nanoparticles by breaking particle agglomerates and decreasing particle size. This process may lead to the formation of smaller and more uniform particles that display enhanced colloidal properties. The reduction in size and increase in the surface area expose more charged functional groups on the nanoparticle surfaces, thereby leading to an increase in ζ-potential. The mechanical energy generated during sonication may promote the exposure of functional groups on the surfaces of nanoparticles. This increased surface charge may enhance the electrostatic repulsion among the particles, thereby minimizing their tendency to aggregate. Consequently, the nanoparticles demonstrate higher ζ-potential values. The combination of ultrasonication with surfactants further enhances the nanoparticle stability. Surfactants aid in forming a uniform coating around nanoparticles, preventing agglomeration through steric and electrostatic stabilization mechanisms. The combination of ultrasonication and surfactants notably enhances the ζ-potential and promotes prolonged stability. Our study suggests that an ultrasonication time of around 60 min is optimal for achieving a better ζ-potential.

Development of Empirical Models for the Reaction Yield (%), Nanoparticle Size, PDI, and ζ-Potential

Polymer nanoparticles exhibit physicochemical properties, including size, size distribution, and ζ-potential, which play a critical role in their interaction with biological systems. In the context of polymer-based nanoparticles, empirical modeling approaches have been extensively used to optimize synthesis conditions, allowing for systematic control of crucial attributes. Empirical and statistical modeling methods play a crucial role in enhancing the encapsulation efficiency (EE) and yield of block copolymer nanoparticles produced through nanoprecipitation. They enable researchers to efficiently determine optimal formulation parameters, minimize the need for extensive experimentation, and improve consistency in results. Previous studies have discussed how empirical and statistical models guide the design of nanoparticles for targeted therapy, imaging, and reaction yield. −

In the present study, key process parameters, such as PF-127 concentration, stirring speed, solvent evaporation duration, organic-phase flow rate, and ultrasonication time, were systematically optimized to evaluate their impact on the reaction yield, particle size, PDI, and ζ-potential of the PLGA nanoparticles. These variables were quantitatively modeled using a nonlinear polynomial fit that demonstrates empirical modeling approaches to describe nanoparticle characteristics by correlating process conditions with physicochemical outcomes. ,,

Figure illustrates the nonlinear curve fitting between the experimental data of the reaction yield versus the different process parameters and the empirical polynomial fit model. From curve fitting, we obtained the relation between PF-127 and reaction yield as

$$y=-58.33+8162.91x_1-127932.89x_1^2$$ 2

where x ₁ is PF-127 concentration (g/mL), and y is the reaction yield (%).

5.

Curve fitting between the experimental reaction yield data and nonlinear polynomial models. (a) Pluronic F-127 concentration and (b) stirring speed fitted with a second-order polynomial model, (c) solvent evaporation duration, and (d) ultrasonication time fitted with a third-order polynomial model. Data are expressed as mean ± standard deviation (n = 3).

From curve fitting, we obtained the relation between the stirring speed and reaction yield as

$$y=-1338.03+1.46x_2-0.000375x_2^2$$ 3

where x ₂ is the stirring speed (rpm), and y is the reaction yield (%).

From curve fitting, we obtained the relation between the solvent evaporation duration and reaction yield as

$$y=-318.91+92.02x_3-5.24x_3^2$$ 4

where x ₃ is the solvent evaporation duration (h), and y is the reaction yield (%).

From curve fitting, we obtained the relation between the organic-phase flow rate and reaction yield as

$$y=35.48-187.488x_4+276.049x_4^2$$ 5

where x ₄ is the organic-phase flow rate (mL/min), and y is the reaction yield (%).

From curve fitting, we obtained the relation between the ultrasonication time and reaction yield as

$$y=9.36-1.60x_5+0.118x_5^2-0.0015x_5^3$$ 6

where x ₅ is the ultrasonication time (min), and y is the reaction yield (%).

Since each parameter was varied independently while keeping the others constant, these models do not capture interaction effects and should be interpreted as single-factor empirical relationships rather than multivariable predictive models.

As shown in Figure , the PF-127 concentration, stirring speed, and solvent evaporation time exhibited the best fit with a second-order polynomial model, achieving R ² values of 0.965, 0.945, and 0.984, respectively. Meanwhile, the ultrasonication time was more accurately represented by a third-order polynomial model, yielding an R ² value of 0.964. However, the organic-phase flow rate did not fit the polynomial model of any order (Figure S1).

The second-order fit for the PF-127 concentration supports earlier findings that its impact on micelle formation and encapsulation efficiency is nonlinear, driven by the balance between micelle saturation and viscosity. Similarly, our results show for the first time that the reaction yield is nonlinearly dependent on the PF-127 concentration. The second-order fit for stirring speed aligns with previous evidence, indicating an optimal range, beyond which excessive agitation may reduce the yield due to shear effects. The second-order relationship observed for the solvent evaporation duration underscores the importance of optimizing the evaporation time to maintain structural integrity and avoid defects, which may be attributable to rapid or extended solvent loss. The third-order fit for the ultrasonication time captures its dual role, initially improving dispersion, but potentially decreasing the yield with excessive sonication. Lastly, the absence of a polynomial fit for the flow rate suggests a more complex, interaction-driven influence on the yield, likely involving threshold behaviors not adequately modeled by standard polynomial equations.

The nonlinear curve fitting between the experimental data of the PLGA nanoparticle size vs different process parameters and the empirical fit model is shown in Figure .

6.

Curve fitting between the experimental nanoparticle size data and nonlinear polynomial models. (a) Pluronic F-127 concentration fitted with a third-order polynomial model, (b) stirring speed, (c) solvent evaporation duration fitted with a second-order polynomial model, (d) organic-phase flow rate fitted with a third-order polynomial model, and (e) ultrasonication time fitted with a second-order polynomial model. Data are expressed as mean ± standard deviation (n = 3).

As shown in the figure, the stirring speed, solvent evaporation duration, and organic-phase flow rate fit best to a second-order polynomial model, achieving R ² values of 0.9104, 0.9889, and 0.9203, respectively. In contrast, the PF-127 concentration and ultrasonication time showed a perfect fit with a third-order polynomial model, each achieving an R ² value of 1.

The third-order fit for the PF-127 concentration relationship showed an optimum size at a PF concentration of 0.02 g/mL. The PF-127 concentration-dependent increase in the Z-average size suggests micelle formation or aggregation. The second-order fit for stirring speed revealed a parabolic relationship with the nanoparticle size, where an optimal stirring speed resulted in a favorable particle size, likely due to improved mixing without causing shear-induced aggregation. The second-order fit for solvent evaporation duration affects the particle size due to the gradual solvent loss, influencing the nucleation rates and particle growth kinetics. Meanwhile, the third-degree polynomial fit for the organic-phase flow rate revealed a nonlinear relationship, highlighting its crucial role in controlling the droplet formation and diffusion. Both slower and faster flow rates may result in inefficient particle formation or cause destabilization. The second-order fit for the ultrasonication time highlighted that the particle size consistently decreased with increasing sonication time up to a certain point, likely as a result of acoustic cavitation breaking up the particle clusters; beyond this point, prolonged sonication could lead to system destabilization.

The relationship between the size distribution of the PLGA nanoparticles (PDI) and the different process parameters was modeled by using nonlinear polynomial fitting, as shown in Figure . As illustrated in the figure, the stirring speed, solvent evaporation duration, and organic-phase flow rate were best fitted with a second-order polynomial model, achieving R ² values of 0.992, 0.997, and 0.963, respectively. Polynomial modeling failed to adequately describe the relationship between PF-127 concentration (Figure S2A), ultrasonication time (Figure S2B), and PDI.

7.

Curve fitting between experimental PDI data and nonlinear polynomial models. (a) Stirring speed, (b) solvent evaporation duration, and (c) solvent flow rate fitted with a second-order polynomial model. Data are expressed as mean ± standard deviation (n = 3).

The strong second-order fit for the stirring speed indicates that increasing the stirring speed initially reduces the PDI, likely enhancing the uniformity. However, at higher speeds, the PDI levels slightly increased again because of turbulence, causing droplet breakup. An R ² value of 0.997 for the solvent evaporation duration aligns with earlier studies, highlighting that factors such as droplet stabilization, solvent diffusion, and polymer precipitation have a direct influence on the nanoparticle PDI. If evaporation is too long or too short, droplets may deform, causing an increase in the PDI. The R ² value suggests that the organic-phase flow rate has a predictable and significant impact on the PDI, and the second-order polynomial fit suggests an optimal flow rate at which the nanoparticle distribution is monodisperse. Deviating from the optimal flow rate resulted in inefficient mass transfer and a broader distribution of solute particles, which contribute to an increased PDI. The relationship between PDI and both PF-127 concentration and ultrasonication time shows a generally weak correlation, as reflected by the low adjusted R ² values of the polynomial model. However, the PDI decreases significantly as the PF-127 concentration increases to about 0.02 g/mL, suggesting an enhanced nanoparticle uniformity within this range. Meanwhile, the trend indicates an optimal ultrasonication period of 40 min, during which a slight reduction in the PDI is observed.

The relationship between ζ-potential and various process parameters was modeled using nonlinear polynomial regression, as illustrated in Figure . Among the parameters, the PF-127 concentration and organic-phase flow rate showed the best fit with a second-order polynomial model, achieving high R ² values of 0.997 and 0.980, respectively. In contrast, the solvent evaporation duration and ultrasonication time were better described by a third-order polynomial model, with corresponding R ² values of 0.960 and 0.952, respectively. However, the stirring speed did not fit the polynomial model of any order (Figure S3).

8.

Curve fitting between the experimental ζ-potential data and nonlinear polynomial models. (a) Pluronic F-127 concentration fitted with a second-order polynomial model, (b) solvent evaporation duration fitted with a third-order polynomial model, (c) organic-phase flow rate fitted with a second-order polynomial model, and (d) ultrasonication time fitted with a third-order polynomial model. Data are expressed as mean ± standard deviation (n = 3).

The strong second-order polynomial fit for the PF-127 concentration suggests a predictable influence of PF-127 on the PLGA nanoparticle surface charge. As the concentration of PF-127 increases, the ζ-potential becomes more negative, which is closely linked to enhanced nanoparticle stability due to increased electrostatic repulsion. PF-127, a nonionic surfactant, stabilizes the nanoparticles primarily through steric hindrance, while its concentration also influences the electrostatic environment surrounding the particles. The third-order polynomial fit for the solvent evaporation duration demonstrates a strong correlation and good empirical fit regarding its effect on the ζ-potential. As the evaporation time increases, the ζ-potential becomes more negative, reaching its lowest value at 7.5 h. This trend suggests improved nanoparticle stability, likely due to more efficient removal of the residual solvent and enhanced adsorption of the surfactant onto the nanoparticle surface, leading to stronger electrostatic repulsion.

The relationship between the organic-phase flow rate and surface charge was effectively captured by a second-order polynomial model with an R ² value of 0.980, indicating a strong correlation. At a flow rate of 0.45 mL/min, the ζ-potential reached its most negative value, suggesting enhanced nanoparticle stability due to greater electrostatic repulsion. The high R ² value indicates that the ultrasonication time has a significant and predictable effect on the ζ-potential. The third-order polynomial model shows that the ζ-potential gradually becomes more negative with increasing ultrasonication time, reaching around −15 mV at 60 min. This pattern suggests enhanced nanoparticle stability, likely due to improved dispersion and more consistent surface modification as sonication continued.

Surface Morphology of the PLGA Nanoparticles

Compared to the other 19 process conditions and the conventional method, process condition 10 (PC10) demonstrated a significantly improved reaction yield, optimal particle size, and narrow polydispersity index (PDI). In this context, the shape, surface morphology, and particle size distribution of the PC10 PLGA nanoparticles were analyzed using field emission scanning electron microscopy (FE-SEM), as shown in Figure S4. The FE-SEM micrograph reveals that the PC10 PLGA nanoparticles exhibit a predominantly spherical morphology with a relatively uniform distribution. The particle sizes appear to be uniformly dispersed. The observed rough surface morphology may be attributed to rapid solvent evaporation and polymer precipitation.

In this study, we employed biomedical-grade PLGA 50:50 (lactide:glycolide) with an acidic end group (7–17 kDa). This specific composition was chosen because PLGA 50:50 is widely recognized for its rapid hydration and degradation kinetics, which promote efficient nanoprecipitation and consistently yield nanoparticles with small diameters, narrow polydispersity indices (PDI), and moderately negative ζ-potentials. These physicochemical attributes arise from the balanced ratio of lactide to glycolide, which optimizes chain mobility and water uptake during particle formation.

The influence of PLGA composition on nanoparticle properties is well established, with systematic variations in the lactide and glycolide ratios directly impacting hydrophobicity, precipitation dynamics, and degradation behavior. Relative to PLGA 50:50, polymers with higher lactide contents, such as PLGA 55:45 and 65:35, exhibit increased hydrophobicity, resulting in slower solvent–nonsolvent exchange, larger particle sizes, and slightly reduced yields. Moreover, the enhanced exposure of terminal carboxyl groups at the particle surface contributes to more negative ζ-potentials. Further increasing the lactide fraction, as in PLGA 75:25 and 85:15, accentuates hydrophobicity and chain rigidity, typically producing larger nanoparticles with broader PDI distributions and diminished yields, unless compensated by elevated surfactant concentrations or higher energy input during nanoprecipitation. − Together, these observations underscore a consistent trend: increasing lactide content systematically shifts nanoparticle characteristics toward larger size, lower yield, and more negative ζ-potential. While the qualitative relationships remain conserved across PLGA grades, the absolute values of size, PDI, and surface charge are predictably modulated by the lactide and glycolide ratio, reflecting the fundamental interplay between polymer hydrophobicity, chain dynamics, and degradation kinetics.

Conclusion

In the current study, the process parameters controlling the nanoprecipitation method for producing PLGA nanoparticles were optimized using experimental and empirical approaches. To identify the optimal nanoprecipitation conditions for achieving a high yield, desirable particle size, low PDI, and stability (ζ-potential), the effects of surfactant (PF-127) concentration, stirring speed, solvent evaporation duration, organic-phase flow rate, and ultrasonication time were evaluated. The conventional method resulted in a yield of 10.06 ± 0.67%. However, by adjusting the process parameters to a PF-127 concentration of 0.03 g/mL, stirring speed of 1750 rpm, solvent evaporation duration of 7.5 h, organic-phase flow rate of 0.45 mL/min, and ultrasonication time of 40 min, maximum yields of 69.86, 85.26, 90.35, 63.56, and 30.44% were obtained, respectively, which were significantly higher than those obtained with the conventional method. Under all five process conditions, the size of the produced nanoparticles was in the desirable size range of 100 to 150 nm. Higher values of the process parameters resulted in a PDI below 0.3, which indicated higher monodispersity of the nanoparticles compared to that obtained by the conventional method. Similarly, at higher process parameter values, the nanoparticles possessed a larger ζ-potential (>−10 mV), suggesting the stability of the suspension.

Nonlinear polynomial fitting was used to obtain empirical descriptions of individual parameter–response relationships for the reaction yield and physicochemical characteristics of the PLGA nanoparticles. Second- and third-order polynomial fits exhibited strong predictive power (R ² > 0.9) for variables, such as the reaction yield, particle size, PDI, and ζ-potential, highlighting the PF-127 concentration, stirring speed, solvent evaporation duration, organic-phase flow rate, and ultrasonication time as critical factors. However, certain parameters, specifically the organic-phase flow rate (for yield), PF-127 concentration (for PDI), ultrasonication time (for PDI), and stirring speed (for ζ-potential), did not conform to any polynomial model, indicating potentially more complex nonlinear interactions. The present study does not evaluate the interaction or additive effects between process parameters, as the experiments were conducted using a one-factor-at-a-time approach. Design of experiments (DoE) or Response Surface Methodology (RSM) will be required to capture interaction effects and develop robust multivariate predictive models.

Finally, process condition PC10 (Table : 7.5 h of solvent evaporation, a stirring speed of 1500 rpm, 0.01 g/mL PF-127 concentration, organic-phase flow rate of 0.3 mL/min, and no ultrasonication) was identified as the best condition to produce the highest nanoparticle yield of 90.35%, which was significantly higher than that obtained using the conventional method (10.063 ± 0.67%). These findings provide a robust framework for the rational design and optimization of PLGA nanoparticle production processes.

Glossary

Abbreviations

PLGA

poly-d,l-lactic-co-glycolic acid

PDI

polydispersity index

PLA

polylactic acid

PDLA

poly-d-lactic acid

PLLA

poly-l-lactic acid

PF-127

Pluronic F-127

CM

conventional method

PC

process condition

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.6c00706.

• Curve fitting between the experimental reaction yield, PDI, and ζ-potential with nonlinear polynomial models; organic solvent flow rate fit with a second-order polynomial model (Figure S1); PF-127 concentration and ultrasonication time fit with a third-order polynomial model (Figure S2); PF-127 stirring speed fit with a third-order polynomial model (Figure S3); representative FE-SEM micrograph of the PLGA nanoparticles prepared using optimized nanoprecipitation process condition 10 (PC 10) (Figure S4) (PDF)

†.

S.K.S., N.R.G., and R.H. contributed equally to this work. The manuscript was written through the contributions of all authors. All authors have given approval to the final version of the manuscript.

This review was supported by the Department of Biotechnology (DBT), Government of India, under the Cancer Disease Biology Research Grant (BT/PR40470/MED/30/2295/2020), awarded to S.H.R.

The authors declare no competing financial interest.