Available online at www.sciencedirect.com Colloids and Surfaces A: Physicochem. Eng. Aspects 311 (2007) 67–76 Controlled size polymer particle production via electrohydrodynamic atomization Christopher J. Hogan Jr. b , Ki Myoung Yun a , Da-Ren Chen b , I. Wuled Lenggoro c , Pratim Biswas b , Kikuo Okuyama a,∗ a Department of Chemical Engineering, Graduate School of Engineering, Hiroshima University, 1-4-1 Kagamiyama, Higashi Hiroshima 739-8527, Japan b Department of Energy, Environmental, & Chemical Engineering, One Brookings Drive, Box 1180, Washington University in St. Louis, St. Louis, MO 63130, USA c Institute of Symbiotic Science & Technology, Tokyo University of Agriculture & Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8588, Japan Received 22 March 2007; received in revised form 18 May 2007; accepted 31 May 2007 Available online 7 June 2007 Abstract Electrohydrodynamic atomization was used for the production of controlled size and controlled physical property polymer particles. Nearmonodisperse submicrometer and supermicrometer polymer particles were produced by the electrohydrodynamic atomization of both water soluble and water insoluble polymer particles. A suitable scaling law, which allows for prediction of the size of the produced polymer particles based on the polymer volume fraction and electrohydrodynamic atomization process parameters, was tested and verified. The scaling law is valid for polymer solutions for which electrohydrodynamic atomization produces pure droplets (no fibers) that do not undergo multiple explosions or give rise to coalescing particles upon collection. Electrohydrodynamic atomization can be used to produce not only controlled size polymer particles but also particles with controlled physical properties. Photoluminescent Cerium doped Y3 Al5 O12 encapsulated poly(methyl methacrylate) composite particles were produced by electrohydrodynamic atomization, demonstrating that this simple technique can be used for advanced materials synthesis. © 2007 Elsevier B.V. All rights reserved. Keywords: Electrospray; Electrospinning; Polymer particles; Yttrium aluminum garnet (YAG); Aerosol synthesis 1. Introduction Polymer particles are widely used as carrier particles in drug delivery systems [1,2]. Recently, the production of novel polymer particles, i.e. polymer–nanoparticle composite particles [3,4] with a core–shell morphology, has expanded interest in polymer particles to many fields of physics, chemistry, and biology, as such particles can be made with unique mechanical [5], magnetic [6], and optical properties [7]. During particle production, it is essential to control the size of the particles produced. The mobility of particles in both gas and liquid systems is size dependent; thus, particles must be of a known, controlled size to effectively control their motion and translocation (e.g. in ∗ Corresponding author. Tel.: +81 82 424 7716; fax: +81 82 424 7850. E-mail address: okuyama@hiroshima-u.ac.jp (K. Okuyama). 0927-7757/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfa.2007.05.072 drug delivery systems). Likewise, in applications where polymer particles with unique properties are desired, the intensity of the desired property (e.g. photoluminescence) must be controlled. It is therefore necessary to develop a general method to produce polymer particles of controlled size as well as controlled properties for wide variety of polymers. Electrohydrodynamic atomization is frequently used in analytical chemistry for the production of highly charged, gas phase ions [8]. Because size controllable near-monodisperse droplets [9,10] are produced in electrohydrodynamic atomization, it has the potential for use in the production of controlled size polymer particles. While several studies [1,11–14] have used electrohydrodynamic atomization and related technologies [15] for the production of polymer particles, their production by electrohydrodynamic atomization across a wide range of electrohydrodynamic atomization conditions has not been examined previously. Furthermore, the ability to predict and control 68 C.J. Hogan Jr. et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 311 (2007) 67–76 the size of polymer particles produced has not been demonstrated. The purpose of this study is to use electrohydrodynamic atomization to produce near-monodisperse, submicrometer and supermicrometer sized water soluble and water insoluble polymers and determine the effects of process parameters and polymer properties on the size distribution of the polymer particles. Electrohydrodynamic atomization of polymer solutions for polymer particle production differs from most cases where electrohydrodynamic atomization is used for particle production [16]. The properties of the polymer itself will greatly influence the electrohydrodynamic atomization process such that previously derived scaling laws [9,10] for particles produced by electrohydrodynamic atomization may not be valid. Furthermore, for certain polymer solutions, Taylor cone-jets [17] may not break up into droplets, i.e. electrohydrodynamic atomization may not be possible. In this study, a suitable polymer particle scaling law [9,18] is tested for a variety of polymer solution properties and electrohydrodynamic atomization process parameters (flowrate and electrical conductivity). The effects of molecular weight and volume fraction of the polymers used on the use of scaling laws for predicting the size of the polymer particles are discussed. Electrohydrodynamic atomization is a robust method for particle production in that it not only allows for the production of controlled size particles, but also particles with controlled properties. This is shown here through the preparation of photoluminescent composites of optically translucent poly(methyl methacrylate) (PMMA) and yellow phosphor Cerium doped Y3 Al5 O12 (YAG:Ce) particles with a core–shell morphology. Overall, this study serves as a guide to production of controlled size polymer particles using electrohydrodynamic atomization. 2. Experimental 2.1. Polymer solutions Polymer particles were prepared from solutions of water soluble polyethylene glycol (PEG, Pure Waco Chemicals Ltd., Japan) and polyvinylpyrrolidone (PVP, Sigma–Aldrich, Saint Louis, MO, USA) and water insoluble poly(methyl methacrylate) (Sigma–Aldrich). PEG and PVP were dissolved in a 1:1 mixture of water and ethanol (by volume), and PMMA was dissolved in N,N-dimethylformamide (DMF). The size and charge on droplets produced by electrohydrodynamic atomization are functions of the density, surface tension, viscosity, electrical conductivity, and dielectric constant of the atomization solution [10,19,20]. Surface tension measurements were made using an automatic surface tensiometer (CBVP-2, Kyowa Interface Science Co., Ltd.). Dynamic viscosity measurements were made using a Brookfield DV-III rheometer (Brookfield, Middleboro, MA, USA) and solution electrical conductivities were measured using a TCX-90i conductivity meter (Toko Chemical Laboratories Co., Ltd., Japan). While the dielectric constant of liquids was not measured, the dielectric constants of polar liquids suitable for electrohydrodynamic atomization are between approximately 20 and 100, in which the dielectric constant has little effect on the predicted droplet diameter [18]. Prior to electrohydrody- Fig. 1. Schematic of the system used for electrohydrodynamic atomization. namic atomization, all polymer solutions were mixed using a magnetic stirrer for a period of at least 24 h. 2.2. Electrohydrodynamic atomization and particle size measurement A schematic diagram of the system used for electrohydrodynamic atomization is shown in Fig. 1. Taylor cones were generated using a stainless-steel capillary needle with an inner diameter of 200 ␮m and an outer diameter of 400 ␮m. A high voltage power supply was connected directly the capillary needle and a positive voltage of 4.5–6.0 kV was applied for operation. A CCD camera was connected to a monitor and was used to visually observe the Taylor cone at the capillary outlet. The liquid flowrate was controlled using a syringe-pump (PHD 2000, Harvard Apparatus Inc., Holliston, MA, USA). Dry N2 continuously flowed through the system to control the relative humidity and allow for rapid evaporation of electrohydrodynamically atomized droplets. The relative humidity in the system was below 30% for all experiments, which allowed for complete evaporation of solvent prior to particle deposition on the ground plate. The highly charged polymer particles produced deposited on the ground plate, which was 6 cm away from the tip of the capillary for all experiments. Collected particles were viewed using a Field Emission Scanning Electron Microscope (FE-SEM, S5000, Hitachi Ltd., Japan). Polymer particle size distribution functions were determined by measuring over 100 particles for each experiment and then binning the particle sizes. 2.3. Cerium doped Y3 Al5 O12 –PMMA composites To produce controlled size, controlled photoluminescence polymer particles, submicrometer YAG:Ce particles were added to a solution of PMMA in DMF with a solution volume fraction of 0.1691 PMMA. PMMA was used for composite production because of its low absorbance of visible light; thus, the photoluminescent properties of YAG:Ce would be retained within a composite of PMMA and YAG:Ce. The YAG:Ce particles were synthesized as described previously [21] and had a primary C.J. Hogan Jr. et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 311 (2007) 67–76 particle diameter of approximately 200 nm. YAG:Ce particles were added to the polymer solution such that there was a mass ratio of 1:10 YAG:PMMA in the suspension. The YAG:PMMA suspensions were kept stable by continuous mixing with a magnetic stir bar, as YAG:Ce particles tended to agglomerate and settle out of the DMF solution without mixing. The photoluminescence of YAG:PMMA composite particles was determined using a Spectrofluorophotometer (RF-5300PC, Shimadzu Corp., Japan). Photoluminescence spectra were measured using an excitation wavelength of 450 nm. 3. Theory and calculations For controlled size particle production by electrohydrodynamic atomization, stable system operation is critical. Fig. 2 shows digital images of the electrohydrodynamic atomization source operated in pulsating, silver-bullet, and cone-jet mode [9,17,22]. Images of these modes have been shown in previous studies [9] but are shown here to emphasize the fact that stable electrohydrodynamic atomization is essential to be able to predict the size of particles produced. The silver-bullet and cone-jet modes are stable, i.e. droplets produced in these modes have near-monodisperse diameters. The silver-bullet mode was used in all experiments in this study, as the solutions used all had relatively high surface tensions [9]. Although it is possible that high frequency pulsations of the silver-bullet mode were present, it appeared, from observations with the CCD camera, that system operation was stable with no pulsations. Several scaling laws, which relate process parameters and solution properties to the size of the droplets produced, have been developed for droplets produced by electrohydrodynamic atomization in the silver-bullet and cone-jet mode [18,23,24]. Rosell-Llompart and Fernandez de la Mora [10] as well as Chen and Pui [18] developed a semi-empirical model to predict the initial droplet size for electrohydrodynamic atomization in the cone-jet mode:   Q 1/3 Ddrop = G(κ) κε0 (1) K Ddrop is the predicted diameter of the droplet, κ the dielectric constant of the solution, ε0 the permittivity of a vacuum, Q the solution flowrate, K the solution electrical, and G(κ) is given by[18]: G(κ) = −10.9κ−6/5 + 4.08κ−1/3 (2) 69 If the diameter of the droplets produced can be predicted by Eq. (1) and droplet evaporation occurs without substantial mass loss by other mechanisms, then the diameter of a polymer particle produced by electrohydrodynamic atomization can be predicted from the equation:   Q 1/3 (3) Dpoly = G(κ) φκε0 K Dpoly is the diameter of the polymer particle and φ is the volume fraction of polymer in the solution. However, the jet breakup process in electrohydrodynamic atomization to form droplets can also be influenced by the viscosity of the solution [10,16] and Eqs. (1) and (2) are typically only valid for solutions with viscosity numbers, Π µ , much greater than unity, where Π µ is Πµ =  σ 2 ρκε0 Kµ3 1/3 (4) ρ, σ, and µ are the density, surface tension, and dynamic viscosity of the liquid, respectively. While Eqs. (1) and (2) have been tested for relatively low viscosity solutions[9,16], few studies have examined the droplet formation process for solutions with Π µ less than unity [10,25]; thus, electrohydrodynamic atomization of polymer solutions may form droplets with sizes not predictable by Eqs. (1) and (2). The evaporating, highly charged droplets produced by electrohydrodynamic atomization will reduce in size to a point at which the repulsion between charges on the surface of the droplet is greater than the force of surface tension holding the droplet together. The point at which the attractive surface tension forces balance with the repulsive coloumbic forces on a droplet surface is called the Rayleigh limit, and is mathematically expressed as [26]:  1/3 q2 Dcrit = (5) 8π2 ε0 σ Dcrit is the diameter at which the Rayleigh limit is reached and q is the net charge on the droplet. Once a droplet evaporates to Dcrit , it will undergo a coloumbic explosion [26,27], which reduces the charge and causes the droplet to lose mass. In terms of controlled size polymer particle production, coloumbic explosions will increase the polydispersity of the polymer particle size distribution function and therefore must be prevented. Fig. 2. Characteristic spray-modes found during electrohydrodynamic atomization of polymer solutions. 70 C.J. Hogan Jr. et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 311 (2007) 67–76 In addition to viscous effects and droplet explosions reducing the degree of predictability and control on polymer particle production, some polymer solutions will not produce droplets by electrohydrodynamic atomization. Cone-jets of polymer solutions with sufficient resistance to shear stress will not breakup to form droplets, rather, a polymer fiber forms from them after solvent evaporation by a process called electrospinning [28]. Resistance to jet breakup in cone-jets of many polymer solutions has been correlated with the number of polymer entanglements (overlaps between polymer chains) in the solution. Shenoy et al. [29,30] have developed a semi-empirical theory where the critical parameter determining the “electrospinnability” of a solution is the entanglement number, (ne )soln , defined as (ne )soln = φMw Me (6) Mw is the molecular weight of the polymer in the solution and Me is the polymer entanglement molecular weight, which is a function of polymer chain topology and morphology [29]. An entanglement number of 2 for a polymer solution corresponds to approximately 1 entanglement per chain and for pure fiber formation by electrospinning, it is typically necessary to use a polymer solution with an entanglement number greater than 2. At an entanglement number less than 2, there will be a transition from pure electrohydrodynamic atomization to an intermediate process between electrohydrodynamic atomization and electrospinning in which neither pure fibers nor pure particles, but particles and beaded fibers are produced [31,32]. While it is necessary to maximize the entanglement number for electrospinning to ensure fiber formation, the entanglement number must be kept to a minimum for controlled size particle production. Based on scaling laws, the size of polymer particles produced by electrohydrodynamic atomization will be a function of the solution flowrate, electrical conductivity, dielectric constant, polymer volume fraction, and possibly the viscosity, solution density, surface tension, and molecular weight. The measured and calculated properties of all polymer solutions and conditions examined for particle production in this study are shown in Table 1, as well as the viscosity number, entanglement number, predicted droplet size (Eq. (1)), and predicted polymer particle size (Eq. (3)). A range is given for the dielectric constant and subsequently the viscosity number, droplet size, and predicted polymer particle size. Results of the experiments performed here are presented for each polymer type examined, highlighting the key observed phenomena. The scaling law given for particles produced via electrohydrodynamic atomization is then evaluated. Finally, the production of polymer–nanoparticle composites with controlled size and pho- Table 1 Properties of polymer solutions, and parameters for electrohydrodynamic scaling laws for all experiments K (mS/cm) Q (␮L/min) σ (N/m) µ (Pa s) ρ (kg/m3 ) ne Πµ Droplet size (␮m) Polymer particle size (␮m) PEG, MW = 20 kDa 0.0375 21–78 0.0375 21–78 0.0375 21–78 0.0759 21–78 0.0759 21–78 0.0759 21–78 0.1154 21–78 0.1154 21–78 0.1154 21–78 0.0024 0.0024 0.0024 0.0037 0.0037 0.0037 0.0048 0.0048 0.0048 4 6.67 10 3 5 8 2 5 8 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.030 0.0092 0.0092 0.0092 0.0209 0.0209 0.0209 0.0421 0.0421 0.0421 906.3 906.3 906.3 918.4 918.4 918.4 930.9 930.9 930.9 0.0426 0.0426 0.0426 0.0863 0.0863 0.0863 0.1311 0.1311 0.1311 0.93–1.44 0.93–1.44 0.93–1.44 0.35–0.55 0.35–0.55 0.35–0.55 0.16–0.25 0.16–0.25 0.16–0.25 4.44–5.16 5.27–6.12 6.04–7.00 3.51–4.08 4.17–4.84 4.88–5.66 2.81–3.27 3.82–4.43 4.47–5.18 1.48–1.73 1.77–2.05 2.02–2.34 1.49–1.73 1.76–2.05 2.07–2.40 1.37–1.59 1.86–2.16 2.18–2.52 PVP, MW = 29 kDa 0.0565 24–78 0.0565 24–78 0.0565 24–78 0.1187 24–78 0.1187 24–78 0.1187 24–78 0.076 0.076 0.076 0.105 0.105 0.105 1 2 4 2 4 7 0.030 0.030 0.030 0.030 0.030 0.030 0.0098 0.0098 0.0098 0.0267 0.0267 0.0267 914.6 914.6 914.6 936.7 936.7 936.7 0.0673 0.0673 0.0673 0.1413 0.1413 0.1413 0.29–0.43 0.29–0.43 0.29–0.43 0.10–0.14 0.10–0.14 0.10–0.14 0.91–1.02 1.15–1.30 1.45–1.63 1.03–1.16 1.30–1.47 1.57–1.77 0.35–0.40 0.44–0.50 0.55–0.63 0.51–0.57 0.64–0.72 0.77–0.87 PVP, MW = 56 kDa 0.0565 24–78 0.0565 24–78 0.0565 24–78 0.058 0.058 0.058 1 2 7 0.030 0.030 0.030 0.0142 0.0142 0.0142 914.6 914.6 914.6 0.1300 0.1300 0.1300 0.22–0.32 0.22–0.32 0.22–0.32 1.00–1.12 1.26–1.42 1.91–2.15 0.38–0.43 0.48–0.54 0.73–0.83 PMMA, MW = 31 kDa 0.0833 37–84 0.0833 37–84 0.0833 37–84 0.1691 37–84 0.1691 37–84 0.1691 37–84 0.008 0.008 0.008 0.008 0.008 0.008 2 4 8 2 4 7 0.037 0.037 0.037 0.037 0.037 0.037 0.0026 0.0026 0.0026 0.0069 0.0069 0.0069 966.6 966.6 966.6 984.7 984.7 984.7 0.1666 0.1666 0.1666 0.3382 0.3382 0.3382 3.10–4.08 3.10–4.08 3.10–4.08 1.19–1.56 1.19–1.56 1.19–1.56 2.55–2.72 3.22–3.43 4.05–4.32 2.58–2.76 3.26–3.48 3.92–4.19 1.12–1.19 1.40–1.50 1.77–1.89 1.43–1.53 1.80–1.92 2.17–2.31 φ k φ = polymer volume fraction, k = dielectric constant, Q = solution flowrate, σ = surface tension, µ = dynamic viscosity, ρ = solution density, ne = entanglement number, Π µ = viscosity number. Droplet size and polymer particle size are predicted sizes based on Eqs. (1) and (3). C.J. Hogan Jr. et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 311 (2007) 67–76 toluminescence via electrohydrodynamic atomization is shown and discussed, demonstrating that with the proper control of polymer solution and process parameters, polymers particles with controlled size and controlled properties can be produced by electrohydrodynamic atomization. 4. Results and discussion 4.1. Polyethylene glycol Scanning electron micrographs and measured size distribution functions of PEG particles are shown in Fig. 3a and b, respectively. Also shown in Fig. 3b are solid vertical lines showing the predicted particle sizes based on Eq. (3). Solutions with three different polyethylene glycol volume fractions were used for particle production. For all tested volume fractions, polyethylene glycol particles were near-spherical and their size distribution functions were near-monodisperse. Particles produced from solutions with volume fractions of 0.0759 and 0.1154 had diameters which agreed well with the predicted diameter from Eq. (3). Polymer particles produced from solutions with volume fractions of 0.0375, however, were smaller than predicted. The deviation from theory for low volume fraction polymer solutions was presumably due to multiple coloumbic explosions [26,27] occurring prior to complete evaporation of the solvent. To prevent coloumbic explosions from occurring, Dpoly from Eq. (3) must be greater than Dcrit from Eq. (5). Previous measurements of droplet charge have shown that initial electrohydrodynamic atomization droplets have approximately 70% of the maximum possible charge as defined by the Rayleigh limit [33]. Assuming that no charge reduction by ion evaporation [34] occurs from supermicrometer droplets, the necessary volume fraction of polymer in solution for Dpoly to be greater than Dcrit is approximately 0.49. Therefore, prior to complete evaporation, all droplets from all polymer solutions 71 used would have undergone at least one explosion. For droplets with dielectric constants in the range used here, explosions create a bidisperse distribution of droplets [26,27], with most of the mass remaining in a single large, mother droplet. Mass loss by a single explosion event is typically not large enough to affect the diameter of the resulting polymer particle produced [27]. However, after a droplet has undergone several explosions, the resulting polymer particle size would presumably deviate from the diameter predicted by Eq. (3). This was likely the case for PEG particles produced from solutions with a polymer volume fraction of 0.0375. The smaller daughter droplets resulting from explosions would be prone to subsequent explosions and would give rise to smaller polymer particles. Such small particles migrate to the outer regions of the spray in EHDA [35], and would not be collected in the same region of the ground plate as larger polymer particles. Small polymer particles from daughter droplets were not visible in FE-SEM images, implying that either the polymer particles produced were too small to resolve in images, or that daughter droplets deposited on the outer regions of the ground plate. 4.2. Polyvinylpyrrolidone Scanning electron micrographs and size distribution functions of PVP particles are shown in Fig. 4a and b, respectively. Unlike PEG and PMMA, PVP was found to increase the electrical conductivity of the solution, and, as a consequence, PVP particles were expected to be submicrometer in size. Two different molecular weights of PVP, 29 and 56 kDa, were used. Particles produced with the 29 kDa PVP coalesced upon deposition onto the ground plate, which resulted in polydisperse, bimodal or trimodal size distribution functions with some visible submicrometer primary particles, and supermicrometer, aspherical polymer sheets [36]. A submicrometer peak in the size distributions, presumably corresponding to PVP particles Fig. 3. (a) FE-SEM images of PEG particles. Scale bars are 10 ␮m. (b) Size distribution functions of PEG particles. 72 C.J. Hogan Jr. et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 311 (2007) 67–76 Fig. 4. (a) FE-SEM images of PVP particles. Scale bars are 10 ␮m. (b) Size distribution functions of PVP particles. which did not coalesce with other PVP particles upon deposition, agreed well with the predicted PVP particle size. Although approximate calculations of the time necessary for droplets to reach the ground plate [37] showed that solvent evaporation should have completed prior to particle deposition on the ground plate, some solvent may still have remained on some droplets, promoting coalesence upon collection. Experimental evidence suggests that in addition to the possible presence of solvent, polymer properties play a role in coalescence during collection. Coalescence of polymer particles was reduced for polymer particles made from higher molecular weight polymers. Although the impact velocity of polymer particles was on the order of 10 m/s and considerably lower than velocities necessary for inertial particle collection [38], impact stresses were still sufficient to give rise to substantial compressive failure in depositing particles, causing particle–particle coalescence. Presumably, polymer particles made from higher molecular weight polymers have higher breaking stresses and would be less likely to coalesce. Less polydisperse size distributions with decreased particle coalescence were found when 56 kDa PVP was used. The higher molecular weight PVP, however, had an increased viscosity and the produced particles were slightly larger in diameter than was predicted, which agrees with the results of Ku and Kim [25] for high viscosity solutions. The major caveat of using higher molecular weight polymers is that solutions of high molecular weight polymers have high (ne )soln values and can give rise to incomplete jet breakup and fiber formation, i.e. increasing the entanglement number of a polymer solution will change the process from electrohydrodynamic atomization to electrospinning. Fig. 5 shows a scanning electron micrograph of the deposit from a solution with 56 kDa PVP with a volume fraction of 0.2374 ((ne )soln = 0.791). Although particle coalescence to form polymer sheets was not apparent, produced particles were connected to each other by polymer fibers as the result of incomplete droplet breakup [31,32]. As stated previously, recent work [29,30] has shown that the entanglement number must be greater than 2 for non-beaded fibers (no particles) to form by the electrospinning process. However, the entanglement number is a measure primarily of the elasticity of a polymer solution. While the elasticity may be the primary parameter governing the formation of pure fibers versus beaded fibers, it may not be the only parameter affecting the formation of pure particles versus beaded fibers. Solution viscosity, evaporation rate, and needle to ground plate distance may also affect particle formation, and further study will be needed to fully understand pure particle versus beaded fiber formation. Fig. 5. FE-SEM image of PVP mixed particles and fibers produced from a solution with (ne )soln = 0.791. C.J. Hogan Jr. et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 311 (2007) 67–76 There is a tradeoff in using high and low molecular weight polymers for polymer particle production by electrohydrodynamic atomization. Sufficiently high volume fractions of low molecular weight polymers can be used in solution to prevent coloumbic explosions from occurring while having a low entanglement number. However, particles produced from low molecular weight polymers more often coalesce during deposition with particles already deposited on the ground plate, creating polymer sheets. Solutions with high molecular weight 73 polymers at high volume fraction have an increased number of molecular entanglements and formation of a Taylor cone from such solutions may not result in particle production but rather mixed fibers and particles. Rather than increase the polymer molecular weight and promote fiber formation, alternative collection methods may also reduce particle coalescence, such as liquid impingement [39], or charge reduction using corona charger [40] or a bipolar charger [41] prior to collection on the ground plate. Fig. 6. (a) FE-SEM images of PMMA particles. Scale bars are 10 ␮m. (b) Size distribution functions of PMMA particles. 74 C.J. Hogan Jr. et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 311 (2007) 67–76 4.3. Poly(methyl methacrylate) With proper control of the polymer volume fraction in solution, molecular weight of polymer used, solution flowrate, and electrical conductivity, near-monodisperse particles of a controllable size can be produced by electrohydrodynamic atomization. Fig. 6a shows scanning electron micrographs of PMMA particles and Fig. 6b shows the PMMA particle size distribution functions. Particles were produced from solutions with two different PMMA volume fractions. Good agreement was observed between the predicted polymer particle diameter (Eq. (3)) and measured polymer particle diameters for particles produced from a solution with a PMMA volume fraction of 0.0833. Particles produced from this solution were slightly aspherical, containing nanostructures similar to those found on the surfaces of particles and fibers measured previously [1,42], which form when the solvent evaporation rate is high. Solutions with a PMMA volume fraction of 0.1691 had viscosity numbers close to 1 and, like higher molecular weight PVP, had slightly larger diameters than predicted. Overall, the size of PMMA particles produced was predictable and controllable. Size predictability and control is not unique to PMMA particles but is in fact possible for all polymer particles. Table 2 shows the predicted particle diameter, geometric mean particle diameter, and geometric standard deviation of all polymer particles produced in this study. With the exception of particles produced from droplets undergoing multiple explosions, low molecular weight polymers (PVP), Table 2 Predicted particle diameters, and measured geometric mean diameters and geometric standard deviations for all experiments Predicted particle dia. (␮m) Geo. mean dia. (␮m) σg PEG 1.48–1.73 1.77–2.05 2.02–2.34 1.49–1.73 1.76–2.05 2.07–2.40 1.37–1.59 1.86–2.16 2.18–2.52 1.10 1.65 1.52 1.72 1.66 2.04 1.33 2.23 2.62 1.15 1.12 1.33 1.12 1.12 1.19 1.13 1.13 1.14 PVP 0.35–0.40 0.44–0.50 0.55–0.63 0.38–0.43 0.48–0.54 0.73–0.83 0.51–0.57 0.64–0.72 0.77–0.87 0.35 0.52 0.59 0.56 0.59 0.79 0.51 0.66 0.91 1.32 1.45 1.42 1.13 1.29 1.23 1.20 1.33 1.31 PMMA 1.12–1.19 1.40–1.50 1.77–1.89 1.43–1.53 1.80–1.92 2.17–2.31 1.18 1.47 1.64 1.74 2.19 2.71 1.08 1.08 1.32 1.17 1.10 1.12 Fig. 7. Experimental determined and theoretically predicted initial droplet sizes during electrohydrodynamic atomization. and highly viscous solutions, particle size distribution functions had geometric standard deviations in the 1.1–1.2 range and geometric mean diameters were in excellent agreement with the predicted diameters. Further verification of the scaling law used to control the size of produced polymer particles can shown by calculating the size of the droplets [43] produced during each experiment and comparing it to the theoretically predicted droplet sizes. For droplet size calculation, it was assumed that produced particles had zero porosity, which is a reasonable assumption based on transmission electron microscopy images of polymer particles produced in a similar manner to particles in this study [44]. Fig. 7 shows the calculated droplet diameter from all experiments as a function of the liquid flowrate to electrical conductivity ratio (Q/K). Also shown are the theoretical lines for the droplet size based on Eq. (1). Both κ = 20 and 90 were used, as 20 < κ < 90 for all solutions. Over three orders of magnitude in Q/K, the calculated initial droplet diameters for the electrohydrodynamically atomized polymer solutions were in excellent agreement with the predictions from Eq. (1); thus, the assumption the particles were non-porous was reasonable, and the scaling laws given here can be used to predict and control the produced polymer particle size over a wide range of electrohydrodynamic atomization conditions. The electrohydrodynamic atomization method is robust in that it allows for the production of not only controlled size polymer particles, but also controlled-property polymer particles by the addition of other materials (e.g. nanoparticles, pharmaceuticals) to the electrohydrodynamic atomization solution. Fig. 8 shows scanning electron micrographs of pure PMMA particles, pure YAG:Ce particles, as well as PMMA particles with submicrometer YAG:Ce particles enclosed within the PMMA. YAG:Ce particles were nonspherical and partially agglomerated but with agglomerate sizes in the submicrometer range and primary particles diameters ranging from 100 to 300 nm. PMMA particles prepared without YAG:Ce were spherical and no change in particle surface morphology was observed when YAG:Ce was added to the precursor solution, indicating that C.J. Hogan Jr. et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 311 (2007) 67–76 75 Fig. 8. FE-SEM images of PMMA particles with and without submicrometer YAG:Ce particles enclosed within the PMMA. The scale bars are 20 ␮m. YAG:Ce particles alone are shown in the upper left, with a 1 ␮m scale bar. Photoluminescence spectra of YAG–PMMA composites under 450 nm light are shown in the upper right. A reference photoluminescence spectra of YAG:Ce is also shown. the YAG:Ce particles were fully enclosed within the PMMA polymer. This agreed well with transmission electron microscope images of PMMA particles with dye nanoparticles within their interior produced by electrohydrodynamic atomization, described elsewhere [44]. Also shown in Fig. 8 is the photoluminescence intensity spectrum of the YAG–PMMA core–shell particles with excitation by 450 nm light. The characteristic broad emission spectrum in the 500–650 nm range for YAG:Ce phosphors [21] is clearly seen visible. The number of YAG:Ce per droplet follow a Poisson distribution [43]; thus, the number of YAG:Ce particles per composite particle will not be the same for all composite particles. Although precise control over the properties of individual particles is not possible, by adjusting the amount of YAG:Ce in the polymer solution, the intensity of the photoluminescence of a collection particles can be controlled, and by adjusting electrohydrodynamic atomization process parameters (flowrate, electrical conductivity, polymer volume fraction) the size of the composite particles can be controlled. 5. Conclusions It was demonstrated that water soluble and water insoluble, low dispersity polymer particles can be readily prepared by electrohydrodynamic atomization with geometric mean diameters in the 0.35–2.71 ␮m size range. Polymer particle production by electrohydrodynamic atomization requires generation of controlled size monodisperse droplets (using a suitable flowrate and conductivity), a sufficient polymer volume fraction in the solution to prevent multiple droplet explosions, but a low number of polymer entanglements to avoid beaded fiber formation. Electrohydrodynamic atomization can not only be used for the production of polymer particles with controlled size, but also polymer particles with controlled properties by adding nanoparticles to polymer solutions. Coupled with recent advances in controlling the dispersion of nanoparticles in solution [45] as well as multiplexing electrohydrodynamic atomization systems for increased production rates [46], electrohydrodynamic atomization can be used as a simple technique for production of advanced polymer materials, as such particles can be produced from essentially an infinite combination of polymers, copolymers, and nanoparticles. Acknowledgements This work was supported by a Grant-in-Aid for Scientific Research (A) (No. 18206079) from the Japan Society for the Promotion of Science (JSPS) and the Ministry of Education, Culture, Sports, Science and Technology. CJH acknowledges support from a National Science Foundation (NSF) graduate research fellowship. The authors thank Akihiro Kinoshita for assistance in preparation of Cerium doped Y3 Al5 O12 particles. References [1] J. Zie, L.K. Lim, Y. Phua, J. Hua, C.-H. Wang, Electrohydrodynamic atomization for biogredable polymeric particle production, J. Colloid Interf. Sci. 302 (2006) 103–112. [2] H. Kawaguchi, Functional polymer microspheres, Prog. Polym. Sci. 25 (2000) 1171–1210. [3] G. Schmidt, M.M. Malwitz, Properties of polymer–nanoparticle composites, Curr. Opin. Colloid Interf. Sci. 8 (2003) 103–108. [4] D.-G. Yu, J.H. An, J.Y. Bae, S. Kim, Y.E. Lee, S.D. Ahn, S.-Y. Kang, K.S. Suh, Carboxylic acid functional group containing inorganic core/polymer shell hybrid composite particles prepared by two-step dispersion polymerization, Colloids Surf. A 245 (2004) 29–34. [5] Z.H. Guo, T. Pereira, O. Choi, Y. Wang, H.T. Hahn., Surface functionalized alumina nanoparticle filled polymeric nanocomposites with enhanced mechanical properties, J. Mater. Chem. 16 (2006) 2800–2808. [6] Z.H. Guo, L.L. Henry, V. Palshin, E.J. Podlaha, Synthesis of poly(methyl methacrylate) stabilized colloidal zero-valence metallic nanoparticles, J. Mater. Chem. 16 (2006) 1772–1777. [7] M. Benaissa, K.E. Gonsalves, S.P. Rangarajan, AlGaN nanoparticle/polymer composite: synthesis, optical, and structural characterization, Appl. Phys. Lett. 71 (1997) 3685–3687. [8] J.B. Fenn, M. Mann, C.K. Meng, S.F. Wong, C.M. Whitehouse, Electrospray ionisation—principles and practice, Mass Spectrom. Rev. 9 (1990) 37–70. [9] D.R. Chen, D.Y.H. Pui, S.L. Kaufman, Electrospraying of conducting liquids for monodisperse aerosol generation in the 4 nm to 18 ␮m diameter range, J. Aerosol Sci. 26 (1995) 963–977. [10] J. Rosell-Llompart, J. Fernandez de la Mora, Generation of monodisperse droplet 0.3 to 4 ␮m in diameter from electrified cone-jets of highly conducting and viscous liquids, J. Aerosol Sci. 25 (1994) 1093–1119. [11] A. Gomez, D. Bingham, L. de Juan, K. Tang, Production of protein nanoparticles by electrospray drying, J. Aerosol Sci. 29 (1998) 561–574. 76 C.J. Hogan Jr. et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 311 (2007) 67–76 [12] L.Y. Yeo, Z. Gagnon, H.C. Chang, AC electrospray biomaterials synthesis, Biomaterials 26 (2005) 6122–6128. [13] J. Xie, J.C.M. Marijnissen, C.-H. Wang, Microparticles developed by electrohydrodynamic atomization for the local delivery of anticancer drug to treat C6 glioma in vitro, Biomaterials 27 (2006) 3321–3332. [14] K.-H. Roh, D.C. Martin, J. Lahann, Biphasic Janus particles with nanoscale anisotropy, Nat. Mater. 4 (2005) 759–763. [15] M.R. Bohmer, R. Schroeders, J.A.M. Steenbakkers, S.H.P.M. de Winter, P.A. Duineveld, J. Lub, W.P.M. Nijssen, J.A. Pikkemaat, H.R. Stapert, Preparation of monodisperse polymer particles and capsules by ink-jet printing, Colloids Surf. A 289 (2006) 96–104. [16] I.W. Lenggoro, K. Okuyama, J. Fernandez de la Mora, N. Tohge, Preparation of ZnS nanoparticles by electrospray pyrolysis, J. Aerosol Sci. 31 (2000) 121–136. [17] M. Cloupeau, B. Prunetfoch, Electrostatic spraying of liquids in cone-jet mode, J. Electrostat. 22 (1989) 135–159. [18] D.R. Chen, D.Y.H. Pui, Experimental investigation of scaling laws for electrospraying: dielectric constant effect, Aerosol Sci. Technol. 27 (1997) 367–380. [19] S.N. Jayasinghe, M.J. Edirisinghe, Effect of viscosity on the size of relics produced by electrostatic atomization, J. Aerosol Sci. 33 (2002) 1379–1388. [20] J. Fernandez de la Mora, I.G. Loscertales, The current emitted by highly conducting Taylor Cones, J. Fluid Mech. 260 (1994) 155–184. [21] M. Abdullah, K. Okuyama, I.W. Lenggoro, S. Taya, A polymer solution process for synthesis of (Y,Gd)3 Al5 O12 :Ce phosphor particles, J. NonCryst. Solids 351 (2005) 697–704. [22] M. Cloupeau, B. Prunetfoch, Electrostatic spraying of liquids—main functioning modes, J. Electrostat. 25 (1990) 165–184. [23] A.M. Ganan-Calvo, The surface charge in electrospraying: its nature and its universal scaling laws, J. Aerosol Sci. 30 (1999) 863–872. [24] R.P.A. Hartman, D.J. Brunner, K.B. Geerse, J.C.M. Marijnissen, B. Scarlett, Scaling laws for droplet size and current produced in the cone-jet mode, J. Electrostat. 163 (1999) 115–118. [25] B.K. Ku, S.S. Kim, Electrospray characteristics of highly viscous liquids, J. Aerosol Sci. 33 (2002) 1361–1378. [26] J. Fernandez de la Mora, On the outcome of the coloumbic fission of a charged isolated drop, J. Colloid Interf. Sci. 178 (1996) 209– 218. [27] K.Y. Li, H.H. Tu, A.K. Ray, Charge limits on droplets during evaporation, Langmuir 21 (2005) 3786–3794. [28] D. Li, Y.N. Xia, Electrospinning of nanofibers: reinventing the wheel? Adv. Mater. 16 (2004) 1151–1170. [29] S.L. Shenoy, W.D. Bates, H.L. Frisch, G.E. Wnek, Role of chain entanglements on fiber formation during electrospinning of polymer solutions: good solvent, non-specific polymer–polymer interaction limit, Polymer 46 (2005) 3372–3384. [30] S.L. Shenoy, W.D. Bates, G. Wnek, Correlations between electrospinnability and physical gelation, Polymer 46 (2005) 8990–9004. [31] H. Fong, I. Chun, D.H. Reneker, Beaded nanofibers formed during electrospinning, Polymer 40 (1999) 4585–4592. [32] K.H. Lee, H.Y. Kim, H.J. Bang, Y.H. Jung, S.G. Lee, The change of bead morphology formed on electrospun polystyrene fibers, Polymer 44 (2003) 4029–4034. [33] L. de Juan, J. Fernandez de la Mora, Charge and size distributions of electrospray drops, J. Colloid Interf. Sci. 186 (1997) 280–293. [34] J.V. Iribarne, B.A. Thomson, On the evaporation of small ions from charged droplets, J. Chem. Phys. 64 (1976) 2287–2294. [35] A.M. Ganan-Calvo, J.C. Lasheras, J. Davila, A. Barrero, The electrostatic spray emitted from an electrified conical meniscus, J. Aerosol Sci. 25 (1994) 1121–1142. [36] I.B. Rietveld, K. Kobayashi, H. Yamada, K. Matsushige, Morphology control of poly(vinylidene fluoride) thin film made with electrospray, J. Colloid Interf. Sci. 298 (2006) 639–651. [37] I.B. Rietveld, K. Kobayashi, H. Yamada, K. Matsushige, Electrospray deposition, model, and experiment: toward general control of film morphology, J. Phys. Chem. B 110 (2006) 23351–23364. [38] P. Biswas, R.C. Flagan, High-velocity inertial impactors, Environ. Sci. Technol. 18 (1984) 611–616. [39] K. Willeke, X.J. Lin, S.A. Grinshpun, Improved aerosol collection by combined impaction and centrifugal motion, Aerosol Sci. Technol. 28 (1998) 439–456. [40] D.D. Ebeling, M.S. Westphall, M. Scalf, L.M. Smith, Corona discharge in charge reduction electrospray mass spectrometry, Anal. Chem. 72 (2000) 5158–5161. [41] M. Scalf, M.S. Westphall, L.M. Smith, Charge reduction electrospray mass spectrometry, Anal. Chem. 72 (2000) 52–60. [42] S. Megelski, J.S. Stephens, D.B. Chase, J.F. Rabolt, Micro- and nanostructured surface morphology on electrospun polymer fibers, Macromolecules 35 (2002) 8456–8466. [43] C.J. Hogan, E.M. Kettleson, B. Ramaswami, D.-R. Chen, P. Biswas, Charge reduced electrospray size spectrometry of mega- and gigadalton complexes: whole viruses and virus fragments, Anal. Chem. 78 (2006) 844–852. [44] H. Widiyandari, C.J. Hogan, K-.M. Yun, F. Iskandar, P. Biswas, K. Okuyama, Production of narrow size distribution polymer–pigment nanoparticle composites via electrohydrodynamic atomization, Macromol. Mater. Eng. 292 (2007) 495–502. [45] M. Inkyo, T. Tahara, T. Iwaki, F. Iskandar, C.J. Hogan, K. Okuyama, Experimental investigation of nanoparticle dispersion by beads milling with centrifugal bead separation, J. Colloid Interf. Sci. 304 (2006) 535–540. [46] W.W. Deng, J.F. Klemic, X.H. Li, M.A. Reed, A. Gomez, Increase of electrospray throughput using multiplexed microfabricated sources for the scalable generation of monodisperse droplets, J. Aerosol Sci. 37 (2006) 696–714.