Particulate matter is ubiquitous in everyday life, most noticeably in manufactured products that support our lives in the form of foodstuffs, pharmaceutical formulations, fine chemicals and fuels, personal and household products. Particulate systems constitute a wide range of materials such as polymer composites, ceramics and minerals. Such systems may be solid particulates suspended in gases or liquids, or liquid particles (droplets) in other liquids, gases or even solids. To take a common example, the manufacture of scratch-and-sniff capsules (Figure 1) that pervade the weekend newspapers and magazines requires some relatively sophisticated design to:
create a carrier particle (normally a polymer gel) that contains a fragrance;
formulate a method of making a dispersion of these capsules;
manufacture these materials in large quantities;
enable the materials to be coated at high shear rates on to paper and dried;
release an adequate amount of fragrance on demand – when the enquirer raises a paper flap enclosing the capsules, thus inducing rupture (under tension), or when the capsules are pressed or scratched (in compression);
avoid any inadvertent damage to the other newsprint.
Other more sophisticated or ‘smart’ particle designs may require the particles to respond to external stimuli such as chemicals, environment, temperature, light or to have different behaviour according to their age. Chemical triggers are common (Figure 2); these have significant applications in environmental remediation and in delivering functionality in the performance of pharmaceutical, personal and household products.
The imperfect particle
Problems such as those described above are increasingly referred to as ‘product formulation engineering’ since there is a need to apply fundamental understanding to allow particles to be synthesised and then manufactured. This has to be done in response to user demand and often in as short a time as possible in order to gain business advantage. Such issues are well known to the pharmaceutical industry and in the development of household and personal products. However, a cursory glance at classical research monographs – and even some modern student texts – advocates development routes based on statistical experiments to achieve optimisation rather than logical deduction from first principles coupled with engineering acumen. If we are to achieve real progress we have to do better! Prior to the mid 1980s the subject of particle science and technology was often regarded as wholly empirical. It was less favoured for fundamental study than the modelling and simulation of molecular liquids (for which thermodynamic and kinetic properties were generally available or could be readily measured). In fact the complexities of particulate systems defied simulation since the defining processes occurred across a range of length scales and for different time scales including molecular, microscopic, mesoscopic and macroscopic phenomena. Not surprisingly the real challenge of predicting the behaviour of such systems is only now being addressed with any seriousness; this has been made possible by technological developments in instrumentation for characterising these systems and radical improvements in hybrid modelling methods that can be applied to them. The quest is on for the best means of making the perfect particle.
The perfect particle
Our understanding of how to create new functional particulate materials is setting the rate of progress of sunrise industries in modern electronics, communications, biomaterials, drug delivery, gene therapy, functional foodstuffs and across a range of other sectors. This has been recognised – only quite recently – by significant moves within the UK to fund multidisciplinary research in molecular engineering and nanotechnology. The USA and Japan already have notable research programmes in these fields. The modern applications place far greater demands on our ability to make clever particulates than ever before – in short, the challenges are substantial and at a very different level to those that could even be envisaged two decades ago. In essence the task is how to ‘engineer’ very sophisticated particulate products so that they can be manufactured reliably and in large quantities. Demanding tolerances may be required on several properties (size, shape, mechanical properties, chemistry, etc.) simultaneously.
This process consists of at least three dependent stages (Figure 3). At each stage three core disciplines are likely to be involved, spanning biological and molecular chemistry, interfacial science and engineering, and process technology (the interaction of particles and multiphase fluids in equipment). These disciplines combine, bounded by business pressures, to develop specific products to meet known and expected markets as quickly as possible. Knowledge of molecular, thermodynamic and interaction properties and forces is vital at each stage.
Creation is the design of new particulate materials (measured in terms of functionality). It may occur through one of two routes. The first is direct construction via molecular assembly, in which atoms and molecules are crafted together to construct the required entity (this may be in a liquid state or gas phase synthesis). Direct assembly may not be possible for some inorganic materials. The second route is the more ‘traditional’ method of the modification of existing materials, for example, through physical processing (controlled breakage) or chemical and thermal reaction (laser processing, controlled surface chemical reactions, etc.). The use of natural phenomena of formation based on biomimetics is gaining momentum because the observation of, for instance, natural laws of atom incorporation into a crystalline lattice can yield self-ordered systems that possess useful properties, such as for catalysis (Figure 4).
Manufacture is the means of producing the particulate materials in sufficient quantities to meet market demand, both in terms of specification and market size. This can be extremely problematic since ‘designer particles’ are not readily made in large quantities and this severely hinders the widespread adoption of smart particles. The need to understand molecular phenomena in a multiphase and complex flow environment is intensely demanding. Some progress has been made with the molecular manufacturing of polymers and drugs using supercritical fluids, but mostly for very low tonnages and at high cost. Some significant advances are being made in the use of aerosol processes for the manufacture of optical fibres and related materials; this will have longterm implications for new materials. Other routes for molecular manufacturing make use of miniaturised reactors (see later) and microfactories (lab-on-a-chip) for at-site analysis, separation and manufacture.
Application is the deployment of the particulate materials or device; this also presents challenges. The product functionality for consumer products in bulk use can be sophisticated (for example, the release of combinations of different active compounds contained in different sized particles and triggered by different external stimuli). Functionality is very specific to each implementation. Some examples of emerging and future application areas are summarised in the table below.
Selected examples of emerging applications for smart particles
Magnetic carrier particles for targeted in vivo drug delivery
In vitro cellular separations and gene therapies using antibody-coated particles
Magnetic particles for immobilised enzyme reactors
Triggered controlled release of chemicals from within particulates
Environmental clean-up based on carrier immobilisation
Intelligent tracer particles (to track location, feel forces and measure chemical environments)
Intelligent micro-sensor panels for localised measurements in
Optical fibres and materials
Designer nanostructures ceramics and composites
Wholly reconfigurable materials, suited for recycling and total reuse
Supercritical fluid processing
Emulsions – can perfection be achieved from chaos?
Recent developments in the precision manufacturing of liquid droplets will now be considered. Here the goal is often (simply!) to produce particles that have identical sizes or a controlled distribution of sizes.
Emulsions are formed by the creation of droplets of one liquid (such as an oil) dispersed in another liquid phase (such as water), together with surfactant molecules that may be adsorbed at the droplet oil/water interface. Traditional methods of emulsification – those relying on shear that is induced by stirring and shaking – typically result in an emulsion with a poorly controlled droplet size distribution and significant polydispersity. These systems employ a high shear and use turbulent fluid mixing regimes in which the two immiscible liquids are contained within a single vessel. Because of turbulent eddies produced by mixing, the discontinuous phase is broken up into droplets and becomes suspended in the continuous phase.
Emulsification is often performed in rotor–stator systems or colloid mills under strongly shearing flows in a narrow gap between a high-speed rotor and a smooth or roughened stator surface. The geometry of the rotor and gap determines the degree of shear and elongation flow and hence the nature of the rupture process(es). The rotor is dynamically balanced and can rotate at speeds of 1000 to 20,000 rpm and the gap between the rotor and stator surfaces may be as small as 25 µm. Strong shear forces are set up and the liquid surfaces are torn violently apart. Particle diameters of the order of 2 µm are easily obtainable with colloid mills.
High-pressure homogeniser systems can also be deployed, in which dispersion of the liquids is achieved by forcing the mixture through a small orifice under very high pressures. This results in very fine particle sizes of 1 µm or smaller but requires very high pressures.
The principle of using methods that may be regarded as essentially ‘chaotic’ (characterised by the random appearance of turbulent eddies) to generate uniform (size) properties is not an ideal means of ensuring product quality, even after judicious selection of process equipment and operation. Scale-up (or -down) of the manufacturing process is difficult since the very processes that cause the droplets to be formed do not scale up or down as the equipment size and geometry changes! Hence process design is notoriously difficult. However, several new methods for the production of monodisperse emulsions are now emerging. The question of producing highly monodispersed droplets has implications for a wide range of industrial applications. For example, in the food industry emulsion technology may be applied to the production of oil/water emulsions such as salad dressings and UHT products, whilst water/oil emulsions are vital to the formation of margarine.
The desire to improve manufacturing methods has been driven by several factors: the desire for new formulation properties; the need to improve product consistency and quality; the need for greater versatility in manufactured batch size and ability to change products; the lower costs attainable by better plant utilisation; and the corporate desire to roll out technologies into multiple sites within globally extended enterprises.
Droplets for perfect emulsions
To enter the world of making perfect particles, it is necessary to return to the fundamental science of the droplet formation process. It soon becomes evident that the development of precision emulsification methods must be rooted in one of two possible manufacturing approaches:
The reduction of the process length scales of the turbulent perturbations, and their uniformity in the shear and elongation mixing processes that rupture the liquids.
The manufacture of droplets individually (drop-by-drop) using microfabricated systems.
Constraining the length scales of the shear fields rupturing the droplets
This can sometimes be achieved by designing specific shear and elongational flows in a homogenisation chamber. More recently microreactors, in which micromixing takes place in thousands of interdigitised chambers, have been used to produce emulsions and to perform reactions (Figure 5). Such devices can utilise diffusion mixing between thin fluid layers which are usually generated by division of the principal stream into many smaller subflows, by reduction of the channel width across the flow axis or by hydrodynamic focusing. Mixing times in the range of milliseconds to microseconds are achievable, corresponding to fluid layers with a thickness of a few tens of nanometres to micrometres.
The production of specific droplet size and size distributions depends on the geometric parameters and operating conditions of the micromixer (Figure 6). For example, an investigation into the formation of a hydrophilic cosmetic cream using a micromixer reported mean droplet sizes of 0.8–2.5 µm, with standard deviations ranging from 0.4–1.3 µm. This is a fascinating area of technology but as yet the narrowness of the droplets is not significantly better than with bespoke homogenisers or alternative technologies. For some configurations thousands of such units can be assembled into arrays (Figure 5(c)). Such devices are now sold commercially and are beginning to find application.
Nanolitre droplet injection using piezoelectric devices
Ideally, if the exact volume of a substance could be discharged through a pore by rapid squeezing (dispensing) of a tube (for example, by a piezoelectric device) then rapid production of droplets akin to inkjet printing could be achieved. Unfortunately, experiments have shown that it is very difficult to produce such a system without also giving rise to satellite droplets that skew the resultant product size distribution to an unacceptable level.
Cross membrane emulsification
This represents a compromise: droplets are extruded from the outlet of many individual pores. The droplets become detached by a subtle combination of transmembrane pressure (giving liquid flux through the pores) and shearing across the pore surface (by movement of the continuous phase that scours the membrane surface). Increasing the cross-flow velocity reduces the droplet size and can be used in combination with an increase in the transmembrane pressure to enhance productivity. The ability of the stabilised surfactant to diffuse to the newly formed interface sometimes represents a limitation on very high rates of production, unless high surfactant levels are tolerated in the continuous phase. Various membrane materials can be used (ceramic, glass, Teflon, etc.). A crucial aspect of this process is very close control of the pore size and spacing to avoid coalescence and other unwanted effects. Apparatus for pilot-scale manufacture using ceramic membranes is shown in Figure 7. Key advantages of this method are:
Droplet size and distribution are often fully controllable.
Products can be made reproducibly.
The process can be operated as a batch or continuous process.
Scale-up can be achieved with known confidence.
The process requires lower energy and uses it more efficiently.
In work at the University of Leeds we used tubular ceramic membranes with uniform pore sizes varying from 10 µm to 0.2 µm. Figure 7(b) shows experimental results for monodispersed droplet formations from 0.2 and 0.5 µm membrane tubes. Work in Japan continues on the Shirasu Porous Glass (SPG) technique, in which the glass is prepared by the phase separation from the primary glass of CaO–B2O3–Al2O3–SiO2, resulting in a membrane with an extremely narrow distribution of pore size. Since this pioneering water/oil emulsion work, SPG has been used successfully to produce a variety of monodispersed microspheres and, most recently, targeted drug products for in vivo treatment of cancer.
The ‘ultimate’ in precision emulsification deploys the advantages of a cross-flow operation with a clever array of pores based on microfabricated devices. The method is subtly different from others since the production of droplets via channel and terrace regions (Figure 8) is spontaneous. As the droplet collects on the terrace region the asymmetry is believed to result in an imbalance of the Laplace pressure across the droplet, hence any small perturbations in the cross-flow channel result in automatic detachment of droplets. Such systems have been fabricated with hundreds of channels and used to make food emulsions and biological emulsions. The lesson learnt is that consideration of the fundamental phenomena can result in radical improvements in performance. Extending the observations gained here suggests that use of non-asymmetric pores should result in much greater productivity – and this indeed has been shown to be the case. This method produces the most uniform particles yet observed and seems likely to gain a place as one of the first microfabricated technologies to be used in the industrial manufacture of emulsions.
The engineering of precision particle assembly methods presents challenges of science, mechanical fabrication and materials selection but offers great potential, since the prize is the ability to make new products. The route to success lies in gaining and applying knowledge of the fundamental principles of the process. Then the consequences for the design of equipment must be dealt with in order to control the underlying phenomena and to ensure practical operability. The ability of engineers to succeed lies in the formation of multiskilled teams who can address the various aspects of molecular assembly, modelling and process design. The result will be a range of new functional particulate products based on molecular manufacturing methods that are scaleable and robust.