Polymeric nanoparticles: A study on the preparation variables and characterization methods

1. Introduction

Polymeric nanoparticles (PNPs) have attracted considerable interest over the last few years due to their unique properties and behaviors resulting from their small size [1]. As asserted by different authors, these nanoparticulate materials show potential for a wide range of applications such as diagnostics and drug delivery [2], [3], [4]. Advantages of PNPs as carriers include controlled release, the ability to combine both, therapy and imaging (theranostics), protection of drug molecules and its specific targeting, facilitating improvements in the therapeutic index [5], [6], [7].

The uptake of nanoparticles into cells usually involves endocytotic processes, which depend primarily on their size and surface characteristics [8], [9]. These properties can be tuned by the nanoparticle preparation method [10], [11]. Depending upon the preparation method and composition of the organic phase, nanocapsules or nanospheres can be obtained [3]. A nanocapsule particle has core-shell morphology with an aqueous or oily cavity in which the active compounds are confined and surrounded by a polymer shell. Nanospheres have a matrix-like structure in which the active compounds and the polymer are uniformly dispersed.

The classic mechanism for controlling drug release from PNPs is achieved by regulation of the rates of polymer biodegradation and drug diffusion out of the polymer matrix [12]. More recently, exogenous and endogenous stimuli triggered drug release is of particular interest, as it is selective to the microenvironment of specific diseases [3], [12].

The technology of polymeric drug delivery relies heavily in the biodegradability and biocompatibility of the polymers. Biodegradable polymers have advantages since they are completely eliminated from the body by natural metabolic pathways [13]. Natural polymers are usually biocompatible and biodegradable; however their use has been limited due to batch-to-batch variations in properties and could be mildly immunogenic. On the other hand, synthetic polymers are well-known for their controlled chemical composition. Several synthetic or natural polymers have been used for the preparation of PNPs, such as proteins, sugars or other natural macromolecules, biodegradable polymers and non-biodegradable, but pharmaceutically acceptable polymers [14], [15], [16]. To self-assemble these materials into PNPs, several preparation techniques has been successfully developed [17]. The choice of a specific method is usually determined by the type of polymer, the drug's physicochemical properties and the final desired characteristics of the PNPs. Nevertheless, all these methods share a common step which is polymer precipitation. This can occur either through addition of a non-solvent or after a decrease of polymer solubility. Besides, in most of these methods PNPs are produced in an aqueous suspension form, in which their chemical and physical stability is poor. Freeze-drying has been the most commonly drying method used to overcome the instability of nanoparticles suspension, improving their long-term stability and facilitating handling and storage [18]. As this process is highly stressful for nanoparticles, cryoprotectants during freezing and lyoprotectants during drying may be used to minimize the damages [18].

Effective post-synthesis purification of nanoparticles is also an important step for controlling their quality and characteristics and therefore their suitability for a biomedical application. Depending on the preparation method, several impurities will be present in the nanoparticle suspension and adsorbed to the nanoparticles. These potentially toxic impurities include organic solvents, salts, particle aggregates, and reagent residues. Filtration, centrifugation and dialysis techniques are commonly used purification methods [19].

Herein the preparation techniques and characterization of PNPs are reviewed. The aim is to collect and compile the information, and also to update and highlight recent developments. We hope the information covered in this review will stimulate the development of novel nanomedicines that are easier to handle, compatible with physiological media and suitable for further clinical developments.

2. Preparation of polymeric nanoparticles

Polymeric nanoparticles have been synthesized by several methods depending on the requirements of their application and the physicochemical characteristics of the drug [20]. The choice of the most suitable method plays a vital role in order to obtain PNPs with the desired properties for a particular application.

Several preparation methods have been developed and these can be divided into two groups, namely, those based on the polymerization of monomers and those taking advantage of preformed polymers (Fig. 1) [17]. These methods can be further classified into two categories: two-step procedures involving the preparation of an emulsification system followed by formation of nanoparticles in the second step of the process and one-step procedures where emulsification is not required for the formation of nanoparticles.

For the polymerization methods, the monomers are polymerized to form the encapsulating polymer. This process can be carried out in two ways, either using emulsion polymerization techniques or interfacial polymerization [10], [11]. Some drawbacks have been reported which have limited the use of polymerization methods for the synthesis of PNPs [10], [11]. Not only are most PNPs formed from slowly biodegradable or nonbiodegradable monomers, but also non-biocompatible byproducts may be generated with these methods. Toxic residues such as monomers and initiators may persist which require extensive purification work to result in a pharmaceutically acceptable product. Another challenge is the requirement for free-radical polymerization or UV light to trigger polymerization, which prevents the addition of proteins or peptides during polymerization [21]. Considering the limitations of polymerization techniques, attention is focused on describing the methods involving preformed polymers, as many of the problems involved in the former method can be avoided.

2.1. Two-step procedures based on emulsification

Emulsion based colloidal delivery systems are widely used in the food and pharmaceutical industries to encapsulate, protect, and deliver bioactive components.

The term emulsion is defined basically as a mixing of one liquid phase into another totally or partially immiscible, through the use of amphiphilic surface-active molecules (surfactants) that reduce the interfacial tension between the two liquids in order to achieve stability.

Emulsions can also be classified based on their composition (oil, water, surfactants) or morphology [22]. Generally, emulsions may be of the oil-in-water (o/w) or water-in-oil (w/o) types depending on whether the oil is dispersed as droplets in water, or vice versa. In addition, more complex systems such as (water-in-oil)-in-water (w/o/w) can also be obtained. Depending on the droplet size, the emulsion formed can be classified into three main categories: nanoemulsions, miniemulsions and macroemulsions (Fig. 2) [22]. In contrast to the more common microscale emulsions, nanoemulsions have particular interesting physical properties. For instance, nanoemulsions appear visibly different and exhibit optical transparency since the droplets are much smaller than optical wavelengths. On the other hand, microscale emulsions typically exhibit multiple scattering of visible light, and, as a result, they appear white [22].

A common mistake reported in the scientific literature is to describe nanoemulsions as “microemulsions”, but the latter is a thermodynamic phase and not an emulsion system [22]. Microemulsions are not emulsions in the classical sense, but rather nanoscale self-assembled equilibrium phases (lyotropic phases). In contrast, nanoemulsions do not form spontaneously, as an external shear must be applied so the droplets will break up into smaller ones. Although these two kinds of colloidal dispersion can be comprised of three simple components (oil, water and surfactants) and have many structural similarities, the distinguishing difference is not one of composition, but rather one of thermodynamics: microemulsions are thermodynamically stable, whereas nanoemulsions are not [23].

In two-step emulsification/solvent removal methods the polymer organicsolution is emulsified in an aqueous phase. Low- and high-energy emulsification techniques can be used to produce nanodroplets and consequently nanoparticles [17]. In emulsion method the droplet formation step is fundamental because it determines the size and size distribution of the resulting PNPs. Polymer precipitation on preformed nanodroplets is achieved by removing the organic solvent by different methods such as solvent evaporation, fast diffusion after dilution or salting out. A similarity between these techniques is the drug encapsulation process in which the drug is generally added in the polymer solution.

2.1.1. Emulsification-solvent evaporation

Solvent evaporation was the first method developed to prepare PNPs from a preformed polymer [24]. In a pioneer study, Gurny et al. applied this technique successfully in the preparation of drug carriers from biocompatible polymers [25]. For this method, the polymer is first dissolved in a volatile solvent (Fig. 3). Dichloromethane and chloroform have been widely used in the past. However, due to their toxicity they have been replaced by ethyl acetate which displays a better toxicological profile and therefore more suitable for biomedical applications [26]. The resulting organic solution is emulsified in the aqueous phase and the mixture is typically processed using a surfactant and high-speed homogenization or ultrasonication, yielding a dispersion of nanodroplets. A suspension of nanoparticles is formed by evaporation of the polymer solvent, which is allowed to diffuse through the continuous phase of the emulsion [26]. The solvent is evaporated either by continuous magnetic stirring at room temperature or under reduced pressure, which is a slow process. After the solvent has evaporated, the solidified nanoparticles can be washed and collected by centrifuging, followed by freeze-drying for long term storage [18].

The emulsification-solvent evaporation method has been widely applied to prepare PNPs with the desired characteristics by adjusting different experimental parameters. For example, poly(lactic acid) (PLA) PNPs with an average size of 200 nm and a low polydispersity index (PDI) (< 0.1), were prepared by Zambaux et al. using dichloromethane and polyvinyl alcohol (PVA) as the solvent and stabilizing agent, respectively [27]. The authors adopted the double-emulsion method (w/o/w) to prepare the PNPs and studied the influence of experimental parameters such as the preparation temperature, solvent evaporation method, internal aqueous phase volume, surfactant concentration and polymer molecular weight on the physicochemical properties of the obtained PNPs. They observed that an increase of the surfactant concentration resulted in a marked decrease of the nanoparticle size and of the PDI due to a good emulsification process. Likewise, the particle size decreased when the polymer concentration decreases. In another report, Bilati et al. studied the effect of the sonication process on the characteristics of poly(D,L-lactic-co-glycolic acid) (PLGA) nanocapsules prepared by the w/o/w solvent evaporation method [28]. They concluded from their study that the duration of the second mixing step (leading to the w/o/w emulsion), had a greater influence on the final mean particle size than the first step (w/o emulsion). In order to study the effect of the polymer solvent on the PNP properties, Mainardes et al. prepared PLGA nanoparticles applying either of two organic solvents (dichloromethane and ethyl acetate) as the dispersed phase [29]. They demonstrated that the size of the PNPs prepared with dichloromethane was larger than those prepared with ethyl acetate. Thus, the best encapsulation efficiency was obtained for dichloromethane. This behavior for smaller particles could be explained by the larger surface area of the droplets during the emulsification step, resulting in a higher drug loss by diffusion towards the suspension medium. Manchanda and coworkers formulated PLGA nanoparticles using methanol and dichloromethane (1:2 v/v) as the solvent system and PVA as the stabilizing agent [30]. It was evidenced that an increase in polymer amount leads to larger nanoparticles. This was attributed to an increase in the viscous resistance of the emulsion mixture, thereby absorbing the agitation energy which in turn leads to reduction in shear stress resulting in droplets with larger size. When the concentration of PVA was taken into account, the particle size decreased from 159 nm (1% PVA) to 113 nm (5% PVA) due to an improvement in the emulsification process. PLGA and PLA nanoparticles were prepared by Budhian et al. by employing dichloromethane as the solvent and PVA as the stabilizing agent [31]. It was shown that decreasing the organic solvent volume resulted generally in a decrease in particle size. They adopted both homogenization and sonication to formulate the emulsion and observed that the particle size distribution produced by sonication was narrower that the particles obtained by homogenization. Ashjari et al. studied the influence of the method used to remove the organic solvent on the final properties of the PNPs [32]. They prepared magnetic/cisplatin loaded PLGA nanocapsules through a w/o/w double emulsion-solvent evaporation technique and observed a change in the morphology and a decrease in the particle size of nanocapsules when slow evaporation at room temperature was used in comparison with evaporation at reduced pressure (Fig. 4). An extensive study was reported by Khoee et al. on the physicochemical properties of cisplatin loaded polybutyladipate (PBA) nanoparticles prepared from w/o/w emulsion [33]. The obtained PNPs showed a size dependence on polymer concentration, decreasing when the polymer concentration decreases. The size of PNPs was also influenced by other process parameters such as volume of oil phase, power of sonication and drug concentration in the internal water phase. Ethyl cellulose (EC) nanospheres were prepared by Wachsmann et al. to study the influence of surface properties on the accumulation selectivity of nanoparticles in murine experimental colitis [34]. For this purpose, PNPs of similar sizes (212–258 nm) were prepared using different surfactants(polysorbate 20 (P20), SDS, sodium cholate (SC), cetyltrimethylammonium bromide (CTAB) and PVA). It was shown that the accumulation of PNPs in the inflamed areas as well as in the healthy tissue was dependent on surfactant type. The targeting pattern for P20 and CTAB particles showed a distinctly increased accumulation in the inflamed tissue compared to SDS particles with slightly higher values for P20. However, as CTAB particles also exhibited a significantly higher accumulation in healthy tissue compared to the other two preparations, the highest selectivity was obtained with P20 particles. Barba and coworkers developed a preparation technique based on multiple emulsion system to produce polymeric nano- and microparticles [35]. Different polymers, such as polyesters (PCL, PLA and poly(D,L-lactide-co-caprolactone) 70:30 (PLC)) and poly(methylmethacrylate-acrylic acid) (Poly (MMA-AA)) 73/27, loaded with different model molecules, were explored. Depending on type of polymer and, consequently, on solvent used for solubilization (dichloromethane), polyester-based nanoparticles with round shape and smooth surface were obtained (Fig. 5). On the other hand, polyacrylates yielded microparticles with high porosity and lower yield of encapsulation, due to the presence of hydrophilic co-solvents (ethanol and isopropyl alcohol) that caused an easy coalescence between the oil and the water phases (Fig. 5). More recently, PCL nanoparticles were produced by Iqbal et al. by applying dichloromethane as the solvent and PVA as the stabilizing agent [36]. They adopted the double- emulsion solvent evaporation-like process using power ultrasound. The mean particle size has been found to be related to sonication amplitude in the second step of emulsification. Additionally, particle size was also related to outer aqueous phase volume, and to PVA and PCL concentration. In another report, PLGA microparticles were prepared by microfluidic techniques which allowed the control of size and size distribution of the droplets formed in the emulsification step [37]. This resulted in a narrower particle size distribution when compared to conventional methods.

Over the past decade, methods for emulsion preparation with nanoscale droplets have been considerably developed due to the technological improvement of emulsification devices, which has prompted the development of the solvent evaporation technique. Though this method is simple and versatile, it can only be applied mainly to liposoluble drugs, it is time consuming and there is also the possibility of nanoparticle coalescence during evaporation. In addition, for scale-up production, alternative methods using low-energy requirements in homogenization, such as the phase inversion composition (PIC) method, are preferred [38].

2.1.2. Emulsification-solvent diffusion

The emulsification-solvent diffusion method was first introduced by Leroux et al. [39]. It consists in the formation of a conventional o/w emulsion between a partially water-miscible solvent containing the polymer and the drug, and an aqueous solution, containing a surfactant (Fig. 6). For the success of this method, the polymer solvent and water are mutually saturated at room temperature to ensure the initial thermodynamic equilibrium of both liquids. The subsequent dilution with an extensive amount of water induces solvent diffusion from the dispersed droplets into the external phase, resulting in the formation of colloidal particles. Such diffusion process is milder than the direct evaporation of the organic solvent from the nanodroplets. In contrast with methods based on solvent evaporation, in this technique the droplet size decreases suddenly in a millisecond time scale during solvent diffusion [40]. Generally, nanospheres are produced by this method but nanocapsules can be obtained just by adding a small amount of oil, for example miglyol, in the organic phase. Finally, depending on its boiling point, the solvent can be eliminated by evaporation or filtration.

Several formulation parameters can affect the size of the obtained PNPs by solvent diffusion. For example, it has been shown that most properties of nanocapsules are determined at the emulsification step [40]. Guinebretière et al. prepared PCL nanocapsules using ethyl acetate and PVA as the solvent and stabilizing agent, respectively [41]. They observed that the final size of the nanocapsules was influenced by the concentration of oil in the organic phase, volume of the solvent in the emulsion and nature and concentration of the surfactant. Also, the thickness of the nanocapsule was linked to the polymer concentration in the organic phase. In another study, PLA nanospheres were prepared by Quintanar-Guerrero et al. by employing propylene carbonate as the solvent and PVA or Pluronic F68 as surfactants [42]. It was evidenced that high concentrations of polymer leads to larger particle sizes with an increase in the polydispersity index. On the other hand, an increase in stirring rate and in the surfactant concentration were found to reduce moderately the size of the PNPs. Similar results were obtained by Trimaille et al. who prepared PLA nanoparticles using ethyl acetate and Pluronic F68 as the solvent and stabilizing agent, respectively [43]. It was shown that increasing the PLA concentration resulted in an increase of the mean particle size from approximately 260 to 530 nm. The same experimental parameters were applied by Surassmo et al. for the preparation PCL nanocapsules [44]. The authors observed that an increase of the surfactant amount resulted in a decrease of the mean particle size. Although, it seems that above some level further significant size reduction is no longer possible as the excess surfactant remains in the continuous phase, and does not play any significantly role in the emulsification. In a different approach, Colombo and coworkers designed with chemical engineering equipment, a pilot plant to study the process of emulsification-diffusion [45]. It was demonstrated that the agitation time, stirrer type and rotational speed were the most important parameters in the emulsification step. In contrast, during the dilution step, the agitation has no influence on the final size distribution. Only sufficient mixing is needed in order to homogenize the mixture. In another report, Sahana and coworkers used a modified emulsification- diffusion method by using a non-saturated organic solvent and water [46]. In their work they studied the influence of the type of solvent and surfactant on particle size distribution and entrapment efficiency of PLGA nanoparticles. When didodecyldimethylammonium bromide (DMAB) and PVA were compared, the former gave the smallest particles but with the lowest encapsulation efficiency, regardless of the type of organic solvent. For both stabilizers, dichloromethane in combination with ethyl acetate yielded the highest entrapment efficiency. Similar results were obtained by Jain et al. who compared the influence of different stabilizers (Pluronic F68, DMAB and PVA) in the preparation of PLGA PNPs from ethyl acetate organic solutions [47]. DMAB, when used as surfactant, led to smaller particle sizes as compared to PVA, but on the other hand, PVA produced particles with higher entrapment efficiency (Table 1). The authors also studied the effect of droplet size reduction by using both homogenization and sonication. It was demonstrated that sonication resulted in smaller particles (165 nm) as compared to homogenization (ca.225 nm). Hallouard et al. studied the oil nature on the physicochemical characteristics of nanocapsules [48]. For PCL-mPEG diblock copolymer, it was shown that the nature of the oil had no influence on the encapsulation rate. Though, it influenced the particle size and polydispersity, with macroglycerides, appearing to be the lipid structure best suited to obtain the smallest monodisperse nanocapsules. Musazzi and coworkers prepared PNPs by a modified solvent-diffusion method without surfactant from a combination of PLGA and PCL-PEG copolymer [49]. The latter was used due to its amphiphilic nature, which stabilizes the nanodroplets surface in the emulsification step. More recently, Chen et al. prepared curcumin-loaded PLGA nanoparticles with an average size of 190 nm using a modified spontaneous emulsification solvent diffusion method (SESD) [50]. The influence of main preparation parameters such as the volume ratio of binary organic solvents (acetone and ethanol) and the concentration of surfactant were studied.

Table 1. Effect of stabilizer type on particle size, PDI and entrapment efficiency, for PLGA nanoparticles prepared by Jain et al. [47].

Surfactants	Size (nm)	Polydispersity index (PDI)	Entrapment efficiency (%)
2% PVA	166	0.085	86.20
1% DMAB	120	0.015	15.30
Pluronic F68	130	0.136	38.56

The experimental parameters related to the solvent diffusion step do not seem to affect particle size [51]. Free solvent diffusion is guaranteed as long as the organic solvent solubility condition is satisfied. This could explain some contradictory results reported by Song et al. in which the highest particle size was obtained at the lowest volume of water for dilution [52]. In their study, the lowest volumes of water used did not lead to complete solubility of the organic solvent. In addition, difficult solvent diffusion can be expected due to the barrier effect of the stabilizing agent on the emulsion droplet. This could also explain the results reported by Kwon et al. where the size of submicron particles prepared using PVA as a stabilizing agent is influenced by the temperature of the dilution water [53]. It was shown that the particle size decreased as the temperature of dilution water increases. In this case, reducing the viscosity of the external phase promoted solvent diffusion, and consequently a decrease in particle size.

The emulsification-diffusion technique presents clear advantages such as high yields, easy in scaling-up, no need for high-pressure homogenizers or ultrasonication, batch-to-batch reproducibility and generally good encapsulation efficiencies [54]. However, there are also disadvantages. For instance, high volumes of water to be eliminated from the suspension and possible leakage of water soluble drugs into the external phase throughout emulsification step.

2.1.3. Emulsification–reverse salting-out

The emulsification solvent diffusion procedure previously described can be considered as a modification of the emulsification-reverse salting-out method [55]. The main difference comes from the composition of the emulsion which is formulated from a water-miscible polymer solvent like acetone and an aqueous gel containing the salting-out agent and a colloidal stabilizer (Fig. 7). Examples of suitable salting-out agents include electrolytes such as magnesium chloride, calcium chloride or magnesium acetate, and non-electrolytes such as sucrose. The emulsification is achieved due to an Ouzo-effect, without employing any high-shear forces [56]. The miscibility of acetone and water is reduced by saturating the aqueous phase, which allows the formation of an o/w emulsion from the otherwise miscible phases. A reverse salting-out effect is obtained by dilution of the formed o/w emulsion with an excess of water to promote the diffusion of acetone into the aqueous phase, which leads to the precipitation of the polymer dissolved in the emulsified nanodroplets. The remaining polymer solvent and salting-out agent are eliminated by cross-flow filtration [57]. The condition of complete miscibility between the organic solvent and water is not essential, but simplifies the execution process [55]. If it is not the case, there is a need for a greater water/solvent volume ratio during the formation of the nanoparticles.

PNPs have been prepared successfully by employing the emulsification-reverse salting-out method with several polymers, solvents and salting-out agents. The only condition which should be met is the need for a two-phase system in the presence of the salting-out agent. In a typical process carried out by Allémann et al. the influence of several process parameters on particle size was studied [58]. Eudragit® S PNPs were first prepared using acetone and magnesium chloride as the solvent and salting-out agent, respectively. In addition, PVA was also added to the aqueous phase as a viscosity-increasing agent and emulsion stabilizer. It was shown that by changing the molecular weight and concentration of PVA in the external phase, the size of the PNPs could be controlled within a wide range (186–1130 nm). Additionally, an increase in stirring rate also allowed a slight decrease in particle size. PNPs with a particle size of 200 to 500 nm were obtained by varying the polymer concentration or the internal/external phase ratio. When Eudragit® E was studied, magnesium acetate was selected as salting-out agent due to the solubility of the polymer in acidic medium. For PLA PNPs when an acidic aqueous phase (magnesium chloride) was changed to a basic one (magnesium acetate), a slight increase on the particle size was observed (228 to 247 nm). Zweers and coworkers also concluded in their studies that the particle size can be best controlled by adjusting the polymer concentration in the external phase [59]. In another report, Song et al. prepared PLGA nanoparticles by employing sodium chloride as the salting-out agent instead of magnesium chloride or magnesium acetate [60]. Although acetone is the most commonly used organic solvent, other solvents have been reported. For example, Konan et al. prepared PLGA and PLA nanoparticles with a mean particle size below 200 nm using tetrahydrofuran (THF) as the polymer solvent [57].

The main advantages of the salting-out method are the avoidance of chlorinated solvents, which are hazardous to the environment as well as to the physiological systems. The greatest disadvantages are the exclusive application in encapsulating lipophilic drugs and the need for intensive purification steps due to the use of salts. The latter can be responsible for the few reports that have been published in the recent years quoting the salting-out method.

2.2. One-step procedures

2.2.1. Nanoprecipitation method

The nanoprecipitation method, also called solvent displacement, was firstly developed by Fessi et al. [61]. The basic principle of this technique is based on the interfacial deposition of a polymer after displacement of the organic solvent from a lipophilic solution to the aqueous phase (Fig. 8). The polymer is dissolved in a water-miscible solvent of intermediate polarity and this solution is added into a stirred aqueous solution in one shot, stepwise, dropwise or by controlled addition rate [51]. Due to the fast spontaneous diffusion of the polymer solution into the aqueous phase, the nanoparticles form instantaneously in an attempt to avoid the water molecules. This process appears to be governed by the Marangoni effect, wherein a decrease in the interfacial tension between the two phases, increases the surface area due to the rapid diffusion and leads to formation of small droplets of organic solvent [62]. As the solvent diffuses out from the nanodroplets, the polymer precipitates in the form of nanocapsules or nanospheres. In general, the organic phase is added to the aqueous phase but the protocol could also be reversed without compromising the nanoparticle formation. The most common used organic solvent is acetone, because it is miscible with water and easy to remove by evaporation. Though, ethanol and binary solvent blends, such as acetone with a small amount of water, ethanol or methanol can also be used [63], [64]. It is also possible to use either two organic phases or two aqueous phases as long as solubility, insolubility and miscibility conditions are satisfied [51]. Usually, surfactants could be included in the process to guarantee the stability of the colloidal suspension, but their presence is not required to ensure formation of nanoparticles. The obtained nanoparticles are typically characterized by a well-defined size and a narrow size distribution, which is better than those produced by the emulsification solvent evaporation procedure. The key variables that are conditioning the final nanoparticle properties are those related with the experimental design. By carefully adjusting the nature and concentration of the components, organic phase/aqueous phase ratio, organic phase injection rate, fluid dynamics and mixing speed, it is possible to control the PNP physicochemical properties. For example, increasing the polymer concentration or the polymer molecular weight generally results in an increase on particle size. These findings are explained by a higher organic phase viscosity, which hinders solvent diffusion and results in larger nanodroplets [65]. In a typical process carried out by Chancón et al. the polymer concentration, organic phase injection rate and needle gauge, were identified as the principal size determinants on the preparation of PLGA nanoparticles [66]. The smallest particles (46 nm) were obtained by using the lowest polymer concentration, the highest injection rate and the lowest needle gauge. Similar results were obtained by Simsek and coworkers in the preparation of PLGA-b-PEG nanoparticles [67]. The average hydrodynamic diameter of these particles could be controlled between 30 and 172 nm by the choice of polymer concentration and PEG content. Chorny et al. prepared nanospheres of PLA by employing acetone and dichloromethane (39:1 v/v) as the solvent system and Pluronic F-68 as the stabilizing agent [65]. It was demonstrated that replacing acetone in the organic phase by equal volumes of ethanol resulted in particle size reduction from 115 to 70 nm. The authors also observed that the organic solvent evaporation rate and an increase of aqueous phase volume had no influence on the nanosphere size. In another report, Dong et al. prepared PEG-PLA nanoparticles by using acetonitrile as the solvent and Pluronic F-68 as the surfactant [68]. They found that the surfactant concentration slightly influenced the nanoparticle size. Likewise, a slight smaller size was obtained by increasing the organic phase volume. PNPs were prepared by Özcan and coworkers by a modified nanoprecipitation method using several poly(γ-benzyl-L-glutamate) (PBLG) derivatives [69]. Briefly, the polymers were dissolved in tetrahydrofuran at 30 °C and this solution was added to an aqueous phase by dripping without the presence of any surfactant. The polymer concentration in the organic phase strongly influenced the mean diameter of the nonpegylated nanoparticles. In contrast, very small PNPs were obtained for the PBLG-PEG copolymer and almost no influence of the concentration was observed due to the amphiphilic nature of this copolymer. Asadi et al. studied the influence of the mixing rate in the preparation of PNPs by nanoprecipitation [70]. They prepared nanoparticles from PLA-PEG-PLA tri-block copolymer. It was shown that increasing the mixing rate led to a decrease in particle size. Furthermore, smaller particles were also obtained by decreasing organic/aqueous phase ratio due to a better dispersion of the solvent and faster diffusion rate. More recently, Bukhari et al. studied the effect of solvent/non-solvent dispersion medium on the preparation of polystyrene (PS) nanoparticles [71]. Chloroform and tetrahydrofuran were explored as solvents for polystyrene and several dispersion phases (methanol, chloroform, acetone and water) were investigated. The results revealed that the combination of tetrahydrofuran with acetone and water, as well as chloroform with methanol and acetone leads to the formation of nanoparticles. They concluded from their study that the dielectric constant difference plus the affinity of the solvent for the non-solvent were responsible for nanoprecipitation. In addition, the morphology was remarkably dependent on the nature of the dispersion phase. In our research group, we prepared surfactant-free PNPs based on amphiphilic polymeric conjugates composed of cholic acid, a sucrose moiety and PEG [72]. Owing to their amphiphilic characteristics in aqueous solution, PNPs were prepared by the nanoprecipitation method with a mean particle size of ca. 90–120 nm. All the obtained PNPs showed negative surface charge and no size dependence on the polymer concentration forming stable nanoparticle suspensions. More recently, we synthesized sucrose and cholic acid functionalized PLGA PNPs and focused our study on the formulation strategies to control the physicochemical characteristics of the PNPs during both the synthesis and post-synthesis processing [73]. To evaluate the relationship between PNPs size and solvent miscibility with water, three organic solvents (acetone, acetonitrile, tetrahydrofuran) were used for the preparation of PNPs. Smaller particles were obtained using the most water miscible solvent (acetone), which could be related to a more efficient solvent diffusion and faster polymer dispersion into water. The influence of the evaporation rate on the mean particle size of PNPs was also explored. We concluded that evaporation under reduced-pressure resulted in the smallest particle size, suggesting that increasing the evaporation rate reduces the likelihood of PNPs coalescence. Regarding the effect of initial polymer concentration on the PNP size distribution, it was evidenced that the mean particle size increased with increasing polymer concentration, which could be attributed to the increase in organic phase viscosity. We also demonstrated that the centrifugation step was responsible for the aggregation behavior of the PNPs during post-synthesis treatment. Conflicting conclusions have been obtained throughout the literature, where the aggregation behavior is mainly attributed to lyophilization. In fact, centrifugation can cause caking, coalescence and difficulties in redispersing nanoparticles [19].

Overall, the challenge in nanoprecipitation is to find a suitable drug/polymer/solvent/non-solvent system, which allows successful nanoparticle production and drug encapsulation. Though, this method is widely used due to its simplicity, quickness and reproducibility. One of the difficulties is the mixing process during nanoprecipitation. A microfluidic platform could be a promising tool for the controlled synthesis of PNPs, where the hydrodynamic flow ensures a fast and tunable mixing of solvent/non-solvent in the microfluidic channels [74], [75]. Other recent development is the advent of automation to nanoprecipitation with high throughput experimentation (pipetting robot, inkjet printing) [76]. Another drawback is the poor encapsulation efficacy of hydrophilic drugs, because the drug can diffuse to the aqueous phase during polymer precipitation. By modifying the solubility of the drug through changes in the pH or varying the solvent composition are among other means to improve encapsulation efficiency [77], [78].

2.2.2. Dialysis

The dialysis method has been applied successfully in the preparation of small PNPs with narrow size distribution [79]. It is governed by a mechanism resembling that previously described for the nanoprecipitation technique, but with a slightly different experimental setup. In this method, dialysis tubes or semipermeable membranes with a suitable molecular weight cut-off (MWCO) are used as a physical barrier for the polymer [10]. Generally, the polymer is dissolved in an organic solvent, placed inside the dialysis membrane and dialyzed against a non-solvent (Fig. 9). Basic prerequisites are the miscibility of the solvents and the existence of dilute polymer solutions. The displacement of the solvent inside the membrane causes the mixture to be progressively less able to dissolve the polymer. In addition, an increase in interfacial tension results in polymer aggregation and leads to the formation of a colloidal suspension of nanoparticles. Although dialysis is a simple and common method, the large volume of counter dialyzing medium could arouse a premature release of the nanoparticle payload due to the long duration of the process.

The morphology and particle size distribution of the obtained PNPs can be modulated by several experimental parameters, such as solvent/non-solvent pair, dialysis MWCO, the temperature at which the procedure is carried out, polymer concentration and speed of solvent mixing [80]. The influence of the solvent was examined in a work of Akagi et al. in which either of four organic solvents (dimethylsulfoxide (DMSO), dimethylformamide (DMF), dimethylacetamide (DMAc), N-methyl-2-pyrrolidinone (NMPy)) were applied as the polymer solvent to prepare PNPs based on poly(γ-glutamic acid) (PGGA) [81]. They concluded that the particles prepared with DMSO were smaller, with a narrower size distribution than those prepared with NMPy. A similar approach was used by Jeong et al. to prepare PLGA nanoparticles from DMAc, DMF, DMSO and acetone as the polymer solvent [82]. The size of the PNPs prepared from DMAc, DMF and DMSO were in the range of 200–300 nm and not significantly different. On the other hand, acetone yielded larger particles with a mean size of 642 nm. This change in particle size could be explained by the difference in solvent viscosity, miscibility with water and in the solubility behavior of the polymer. In another report, Chronopoulou et al. studied the influence of several experimental parameters on the size and morphology of nanoparticles prepared from natural and synthetic polymers [80]. For poly(methyl methacrylate) (PMMA) nanoparticles, a linear correlation between polymer concentration and the size of the nanospheres was determined. The same behavior was observed for poly(phenyl acetylene) (PPA) nanospheres. In addition, at low concentrations, PPA nanoparticles with smaller diameters were obtained at low temperature than those obtained at room temperature. The influence of the MWCO was also studied for this synthetic polymer. It was shown, that reducing the membrane MWCO induces a decrease of the mean particle size. The biopolymers under study also followed the same trend. This is due to a decrease in the mixing rate of the solvents, thus favoring thermodynamic factors over kinetic ones. Different morphologies were observed for hyaluronic acid based nanostructures by changing the chemical properties of the solvent/nonsolvent pair (Fig. 10).