Investigating the Particle to Fibre Transition Threshold during Electrohydrodynamic Atomization of a Polymer Solution

Electrohydrodynamic Atomization (EHDA) is a key research area for producing micro and nano-sized structures. This process can be categorized into two main operating regimes: electrospraying for particle generation and electrospinning for fibre production. Producing particles/fibres of the desired size or morphology depends on two main factors; properties of the polymeric solution used and the processing conditions including flow rate, applied voltage and collection distance. In this work the particle-fibre transition region was analyzed by changing the polymer concentration of PLGA poly (lactic-co-glycolic acid) in acetone between 2-25wt%. Subsequently the processing conditions were adjusted to study the optimum transition parameters. Additionally the EHDA configuration was also modified by adding a metallic plate to observe the deposition area. The diameter and the distance of the plate from the capillary tip were adjusted to investigate variations in particle and fibre morphologies as well. It was found that complete transition from particles to fibres occurs at 20wt% indicating concentration to be the dominant criterion. Low flow rates yielded fibres without beads. However the applied voltage and distance between the tip of the nozzle jetting the polymer solution and collector (working distance) did not yield definitive results. Reducing the collector distance and increasing applied voltages produces smooth as well as beaded fibres. Addition of a metal plate reduces particle size by ~ 1μm; the fibre size increases especially with increasing plate diameter while bead density and size reduces when the disc is fixed closer to the capillary tip. Additionally, the deposition area is reduced by 70 % and 57% with the addition of metal plates of 30mm and 60mm, respectively. The results indicate that a metal plate can be utilized further to tune the particle/fibre size and morphology and this also significantly increases the yield of EHDA process which is currently a limitation in adopting it as a mass production technique.


Introduction
Polymeric particles and fibres are finding uses in a range of industries from drug delivery and tissue engineering to food additives and even textiles [1]. A range of techniques exist for the production of these structures including spray drying and emulsion evaporation for particles [2] and thermal induced phase evaporation and self-assembly for fibres [3,4]. Included along with these processes is also the electrohydrodynamic atomization (EHDA) technique which as the name suggests is principally governed by electro-mechanical and hydrodynamic forces.
EHDA has been the subject of wide research for more than a century with several of its benefits being realized. It is well accepted that this process is suitable in terms of providing a well-controlled size and morphology [5,6] which is increasingly becoming a requirement for better efficiency. For example in the case of tissue engineering, electrospun collagen scaffolds possesses more homogeneous pore structures as opposed to freeze dried scaffolds [7]. Similarly electrospraying offers several advantages such as higher drug loading efficiency and narrower particle size distribution over other particle fabrication processes [8].
Electrospraying and electrospinning are two main electrohydrodynamic atomization techniques. Fibres via electrospinning and particles via electrospraying can be produced through proper selection of the processing parameters and solution properties. The basic difference between these two processes lies mainly in the concentration of the solutions [9].
In electrospraying, low viscosity solutions are used, as the concentration is sufficiently low to destabilize the electrified jet that breaks down into small droplets that further solidify through rapid evaporation process and form particles. In electrospinning, more viscous solutions with higher polymer concentration are used to form the charged jet, which after an initial straight path, undergoes "bending instability" and characteristic "whipping motion", and as a result fibres are formed [10].
Additionally, EHDA has proven to be a cost effective technology [11,12] and one of the main advantages of this process is the wide range of materials that can be processed from solutions and also melts, emulsions etc. [4,13]. There has been a significant amount of interest in recent years in developing new printing methods for patterning materials in nanoscale [14,15]. EHD printing has many unique features such as formation of 3D products with high resolutions and the ability of using coarser nozzles in order to avoid clogging [16].
This study derives its need from the current limitation EHDA faces in terms of reproducibility and accuracy [17]. The overall process is dependent on multiple aspects, all of which have an interdependent relationship further complicating the dynamics of EHDA. As a result this proves to be a major hurdle in upscaling the process to an industrial level where consistency would be necessary for mass production. As a result, EHDA has become a flourishing research area for micro/nano level particle and fibre production [18].
The main aim of this work was to define the effects of the parameters on which the particle and fibre formation process is based in electrohydrodynamic processing. Primarily this included the effect of solution parameters in terms of concentration and the process parameters, specifically flow rate, applied voltage and nozzle tip-collector distance.
Additionally, another area investigated within the scope of this work was the effect of an auxiliary metal plate. The main aim was to comprehend whether this EHDA set-up modification induced any changes in relation to the produced particle/ fibre morphology.
Transition from particle to fibre formation has a high correlation with respect to the polymer concentration in the solution. Quantitatively this transition process for different polymeric solutions has been characterized using the Berry number [19] which is a product of the intrinsic viscosity and the solution concentration as described in equation 1. (1) Where is the Berry number, ( ) is the intrinsic viscosity, which is calculated by determining the reduced viscosity ( where is the specific viscosity) and extrapolating to infinite dilution, and is the polymer concentration.
The Berry number has a quantitative significance as it maps out the transition regime between electrospraying and electrospinning. As a result different regions have been mapped out as summarized in equation 2. (2) Where is the polymer concentration, is the critical overlap concentration and is the chain entanglement concentration. The first region is defined as the dilute region up to the point where , and the working polymer concentration is < as defined in equation 3.
The intermolecular entanglement is significantly weak, limiting viscous effects, resulting in electrospraying. This region can be further subdivided into different regimes, the first being the semi dilute un-entangled regime where increasing concentration leads to significant chain overlap. The values in this region correspond to . However the chain overlapping/polymer-polymer interaction does not prove to be sufficient enough to inhibit the molecular motion due to which the occurrence of particles, and as a result beaded and non-continuous fibres are common.
Beyond values of , researchers [19][20][21] have generally observed the formation of pure fibres completing the transition to what is termed as the semi-dilute entangled regime.
Specifically, in terms of concentration, this transition is signified by . Hong et al. [21] observed a steep rise in viscosity at resulting in a strong chain overlap. Because of this the chain entanglement proves to be competent enough to hinder the molecular motion and ultimately the visco-elastic stresses are strong enough to prevent fibre fracture.

Materials and Solution characterization
All the experiments were carried out with PLGA (co-polymer 50:50, Resomer RG503H, Hamburg, Germany) was used to measure the surface tension using the Wilhelmy's plate method. Electrical conductivity of each solution prepared was estimated using a conductivity probe (Jenway 3540pH/conductivity meter). All the measurements, presented in Table 1, were conducted at the ambient temperature (21°C) and relative humidity of 40-50% after calibrating the equipment with distilled water.

Particle/fibre characterization
All the samples were collected on microscopic glass slides. These were analyzed mainly under an optical microscope (Nikon Eclipse ME 600) fitted with a camera (Micropublisher supplied the required electric field to the solution. In order to investigate the effect of the addition of the auxiliary plate on the particle/fibre formation and size, the EHDA setup was modified by inserting metal plates of various size (30mm and 60mm in diameter). The plates were attached at two fixed distances of 5mm and 25mm from the capillary tip as shown in figure 1. All the experiments were conducted at the ambient temperature of 21°C and relative humidity was controlled between 40-50%.

Transition regime
Changing the concentration proved to be a clear differentiator between particle and fibre formation. Effects of concentration have been widely tested and characterized [22,23]  While 10wt% (figure 2c) appears to be a significant increase in concentration compared with 4wt%, electrospraying was still the dominant phenomenon. This is expected since the measured viscosity shows a rise of approximately 1mPa s which is still relatively small.
Although the visco-elastic stresses are not strong enough, the rise in viscous effects does have an effect on the morphological variation, mainly manifested by the formation of tailed particle structures. As the viscous effects become more significant at 3mPa s at 15wt%, the formation of particles and significant beads on string structures can also be seen (figure 2d).
Simultaneously the observed increase in surface tension at higher concentrations shows that it is still strong enough to compete with the low visco-elastic stresses in generating particles and beads.
Beyond 15wt% the viscosity rises steeply from 3 to 7mPa s explaining the formation of only fibres at 20wt% (figure 2e). The significant amount of chain entanglement and overlap at this concentration prevents jet fracture caused by the electrostatic field. Resistance offered by the visco-elastic stresses to the surface tension plays a key role in reducing the bead density.
Qualitatively this is supported by the shape of beads which are more spindle-like, generally observed at higher viscosities [24].
As the viscosity further increases to 15mPa s at 25wt% (figure 2f), polymer chain interaction is quite high for the specific processing condition, leading to the formation of mostly smooth fibres. While a few beads are still observed, the beads are highly elongated relative to those seen at 20wt%.

Quantification of the transition process
Prior to estimating the , the values for the specific viscosity were extrapolated to estimate the intrinsic viscosity as shown in figure 3a. The R 2 value of the extrapolated trend line to zero concentration gives a value of 0.82 which appears to show a good fit with the actual data. As indicated in figure 3a, the intrinsic viscosity estimated for PLGA in acetone is approximately 0.3. This is in good agreement with the values that have been reported in literature [25]. The intrinsic viscosity is deemed to be an important variable as it could allow for further correlation with the degree to which a polymer forms inter-or intra-molecular chain entanglements at different concentrations [25].
Quantitative mapping of the transition mechanism in terms of the shows that >6 is required as indicated in figure 3b  It would be reasonable to expect variations in the reported values as the polymer concentration is just one factor which affects the particle/fibre formation process. For example, changing the molecular weight of the polymer can either increase or decrease the threshold for fibre formation by affecting the polymer-solvent interaction as observed by Koski et al. [28]. Similarly the type and solubility of the polymer in the solvent used would be another governing factor affecting the intrinsic viscosity.
One aspect where the estimated value for the PLGA-acetone system is in agreement with values of different solutions is the minimum threshold required for fibre formation.
Obtaining complete fibres or even a fibre dominant morphology is highly unlikely at and particles and beads on strings would be more prevalent.

Size distribution of particles and fibres
In the electrospraying regime (2-15wt%) the particle size appears to be steadily increasing with concentration ( figure 4a). The increase in concentration has raised the average particle size from 4.36±0.9μm at 2wt% to 8.99 ±2.4μm at 15wt%. Concentration also shows an effect on the particle size distribution. 2wt% appears to have a narrower size distribution ( figure   4b). The variation in particle size becomes more pronounced with increasing the concentration as can be observed from the size distribution graphs at 4wt%, 10wt% and 15wt% (figure 4c, 4d and 4e, respectively). In electrospraying viscous effects are less insignificant, leaving the surface tension and electrical conductivity as the main competing forces. Generally the surface tension would act to reduce the surface area of the jet promoting larger particle size [30]. The electric field on the other hand would induce greater charge repulsion increasing the surface area [31]. This charge repulsion is the main reason for the particle breakdown as it flows down on to the collector.
In electrospinning, visco-elastic stress resists any further change in fibre morphology. Apart from bead formation, surface tension has minor effects on the fibre size [32] but the electric field competes with the visco-elastic stress to promote tension within the fibres leading to

A C C E P T E D M A N U S C R I P T
ACCEPTED MANUSCRIPT 9 fibre thinning. At 25wt% the visco-elastic stress is relatively more dominant than at 20wt% resulting in thicker fibres.

Electrospraying
Optical microscopic images of particles electrosprayed with and without an auxiliary metal plate are shown in figure 7a with their corresponding size distribution. Electrospraying in the presence of a metal plate clearly shows a reduction in particle size as can be seen from figure   7b and the size distribution is narrower when making use of a metal plate at the different applied voltages tested. Additionally, it indicates that the frequency of obtaining particles 5μm is much higher when electrospraying under the plate configuration.

Electrospinning
Changing the system configuration showed significant variations in size and morphology of the fibres as well as the shape of the jet. Fibres electrospun without an auxiliary metal plate were highly beaded in nature (both for 20 and 25wt% concentrations as figures 8a(i, ii), 9a depict). For the 20wt% concentration, the addition of both 30mm and 60mm diameter auxiliary plates placed at a distance of 5mm above the needle tip resulted in smother fibres.
On the other hand, increasing the plate-tip distance to 25mm led to the generation of beaded fibres. Figure 8b, demonstrates the variation of average fibre size with applied voltage at different plate-tip distances and different auxiliary plate diameter.
In the case of the average fibre size, for the 25wt% concentration ( figure 9a,b), configurations with a metal plate seem to have an effect on the fibre diameter which is generally greater for the 60mm metal plate compared to the fibres electrospun without an auxiliary metal plate.
Further, from figure 9b it can be inferred that the 60mm metal plate yields larger diameter fibres in comparison with the 30mm metal plate, irrespective of the plate-tip distance.
The increase in the average field strength in the presence of a plate as reported by Yang et al. [33] leads to greater bending speeds in the whipping instability region sufficiently stretching the fibres. As a result when both 30 and 60mm plates were fixed at 5mm from the tip of the nozzle, fibres were much smoother with reduced bead sizes.

A C C E P T E D M A N U S C R I P T
ACCEPTED MANUSCRIPT 10 At 25mm however, the greater distance between the plate and the tip reduces the effect of the plate on the electric field and thus the uniformity, resulting in larger bead sizes. At both working distances the average bead size is greater for the 60mm plate even if the bead density is reduced. This would be explainable on the basis of the larger cone jet, which is formed under the 60mm plate.
The results also show that the effect of the plate does not appear to be limited along the working distance direction. In the absence of an auxiliary plate the electric field is highly focused in the vertical direction as opposed to the presence of the plate where the electric field is distributed in the horizontal plane as well, generating a uniform electric field gradient [34]. Following on from this argument, at a plate-tip distance of 25mm using a smaller plate diameter of 30mm is possibly less effective in distributing the electric field along the horizontal direction in comparison to the 60mm plate. As a result the combined effect of the reduced plate diameter when using the 30mm plate and the increased plate-tip distance possibly does not increase the overall average electric field strength to a level necessary to reduce bead formation.

Collection area
As part of the experiments performed with the auxiliary metal plate, the collection area was also examined. Samples were collected at working distance of 150mm. While examining the collection area under both configurations, the fibres were electrospun for a time period of 120s and collected on an aluminum foil. Figure 10 shows the collection of fibres under normal conditions as well as electrospinning performed in the presence of an auxiliary metal plate. The collection area of the fibres was further examined by using ImageJ. The deposition area is reduced by 70% using the 60mm metal plate while by incorporating the 30mm metal plate, the reduction is approximately 57% in the presence of the metal plate. A second advantage that a uniform electric field offers can be inferred from the reduced fibre deposition area. Under the plate the electric field is much more concentrated to a specific area and therefore allowing the fibres to be collected more efficiently. Potentially this is one manner through which yield can be increased which is currently a limitation in existing EHDA setups [35].

Conclusions
In this study the transition mechanism between formation of particles and fibres by EHDA was investigated by varying the solution properties and processing conditions. It was seen that transitions showed a primary dependency on the polymer concentration. Low polymer concentrations between 2wt%-4wt% resulted in electrospraying, while a mixture of particles and beads-on strings was dominant at intermediate concentrations of 10wt% and 15wt%. at higher concentrations of 20wt% and 25wt%, only fibres were obtained. At higher concentrations the bead density reduced and the beads appeared more spindle shaped.
Smoother transition to uniform non-beaded fibres was facilitated by lower flow rates (e. g. higher flow rates produced beads with the shape gradually changing from spindle-like to spherical as the flow rate was increased).
Results on changing the working distance and applied voltage at different polymer concentrations were not in mutual agreement. Lower working distances and higher applied voltage produced smooth as well as beaded fibres. Furthermore even the average fibre size did not show any specific trend on changing these variables. More precisely, particle diameter was found to be significantly reduced and a narrower size distribution was more prevalent. In electrospinning, the plate dimension and the distance of the plate from the capillary tip were observed to affect the morphological formation of fibres and their sizes. Beads in between fibres were reduced when the auxiliary plate was attached closer to the capillary tip. However, at large distances from the capillary tip, a minimum plate dimension was necessary to influence the formation of fibres without beads. Fibre diameter was found to be dependent primarily on the plate dimension with a larger plate diameter generating larger sizes. The collection area was also observed to be concentrated in the presence of a plate, thus significantly improving the product yield. Electrospinning stability and conditions were found to be improved in the presence of a metal plate as well as on further increasing the plate diameter.
Using a plate to concentrate the electric field could be one manner to address the issue of productivity and yield apart from other more complicated techniques used such as multi-needle electrospinning. However, further research would still be required on its applicability and feasibility for an industrial level production capacity. On a mid-market level, EHDA has the potential to be combined with a desktop 3D printer to cater for public consumers rather than industrial customers. The desktop 3D printing market is a huge market in itself and additionally the production level both in terms of time and cost needed for this is less demanding. This could possibly allow for better integration of EHDA systems within 3D printing through the mentioned techniques of multi-needle electrospinning or the presence of a uniform electrical field as shown in this study.       : Optical micrographs of EHDA products obtained at applied voltage of 15kV a(i) without metal plate and a(ii) with 30mm diameter metal plate and b) particle size variation of EHDA products generated without metal plate and with 30mm diameter metal plate (the flow rate and working distance used were 50μL/min and 200mm, respectively).   Wai (Keith) Lau obtained a degree in Biomedical Science from King's College London and a Masters in Biomaterials and Tissue Engineering at University College London for which he obtained a distinction and was mentioned on the Dean's list of commendations. He was awarded funding by the EPSRC and the IBME at UCL to undertake his PhD in Prof. Edirisinghe's laboratory. His current research topic investigates novel methods of utilising electrohydrodynamic processing for biomedical applications. He is currently under the employment of AtoCap Ltd. to spearhead the development of co-axial electrospraying technology for industry.
Professor Mohan Edirisinghe FREng has pioneered both fundamental and user-inspired engineering research, leading to remarkable progress and international excellence in the novel forming of advanced materials. He has invented many processes and devices for the preparation of bubbles, vesicles, particles, capsules and fibres and has been awarded numerous prizes, including the Royal Society Brian Mercer Feasibility Award for an unprecedented three times. His seminal work has resulted in significantly advancing areas crucial to the betterment of life, such as optimal drug delivery, smart orthopaedic coatings and the translation of novel medical technologies from the laboratory to clinical practice.
Dr Maryam Parhizkar studied Mechanical Engineering at Brunel University and achieved a MEng degree in 2007. She joined Jacobs Engineering in 2008 as a junior consultant, working on a number of engineering and design projects. In 2010, she started her PhD degree in Prof. Edirisinghe's lab at University College London. Her research focus was production of microbubbles/scaffold using microfluidics and electrohydrodynamic techniques. Since 2014, she has been working as a postdoctoral research associate on a EPSRC funded project investigating new methods of delivery of Cisplatin utilising particles and microbubbles.

Graphical Abstract
A C C E P T E D M A N U S C R I P T