Medical Device Link.

MANUFACTURING

Micromoulding defined

Microinjection moulding (micromoulding) is not directly comparable with conventional injection moulding. The product dimensions can be extremely small (weights of approximately 1 milligram are typical), and the thermal and flow conditions within a micromoulding process are significantly different from the conventional injection moulding process, from which the technology was derived. Product surface areas are large compared with product volumes. This results in rapid heat loss from the molten material as the cavity fills and requires high injection speeds to avoid premature solidification and incomplete products. This high-speed injection can also cause stresses and strain rates in the material that are orders of magnitude higher than those found in conventional injection moulding and result in unusual material integrity and morphology. These factors mean that the operating window is small and process variations, temperature fluctuations, material inconsistency and ingress of impurities can result in substandard products that will not function in the manner for which they were designed. These factors pose interesting challenges for process measurement and control, which are required to create a successful manufacturing system (Figure 1).

Figure 1: Manufacturing process monitoring.

The first large-scale production of micromouldings began in the late 1980s. A number of subsequent enabling technologies provided the processing and measurement facilities that are essential for this specialised area of production. The processing requirements have led to the development of a number of dedicated moulding machines. Some are described below and have adopted different approaches:

• A servo electric micromoulding machine from Battenfeld (www.battenfeld.co.uk) has a three-stage injection system, integrated part handling and inspection and the ability to undertake postprocesses such as packing or assembly. Typical product weights range from 0.1 gm down to submilligram components.
• The existing servo electric machines from Fanuc (www.fanuc.co.jp) have been enhanced by adding an ultra-fast injection unit (up to 800 mm/s) that allows production of micro and nano features on larger components.
• The servo hydraulic approach has been followed by Juken (www.juken.com) with small, dedicated moulding machines. A spur gear weighing

1 millionth of a gram has been produced for exhibition purposes and although this is not a functional product it serves to illustrate the processing extremes that are now possible.
Some essential enabling technologies for assisting fundamental research into micromoulding process behaviour are

• simulation software specifically dedicated to micromoulding
• flow visualisation in real time with high-speed video camera
• ultrasound flow and cooling measurement
• inline three-dimensional (3D) video camera inspection within the cycle time.

Figure 2: Sapphire window mould with high-speed video camera.

Simulation software

Flow simulation software programs, which have proved so successful with conventional mouldings, cannot be applied to micromoulding. Conventional computational fluid dynamics (CFD) solvers, based on Euler and Navier-Stokes models lose accuracy when considering microscale flows. The continuum assumptions on which they are based become less valid and molecular effects and interactions become important. Molecular dynamics (MD) and variants have emerged as techniques offering greater accuracy in microscale regions, but are highly computationally expensive and can only easily be adopted for simple geometries. A compromise exists whereby a viable approximation of an MD solution can be achieved through modification of continuum-based CFD solvers. This will provide improved accuracy in the near-wall regions where phenomena such as wall slip and temperature discontinuities are known to exist under certain conditions. Work is ongoing in this area to produce commercial software that offers a powerful tool for investigating the feasibility of a micromoulding process for component manufacture without a costly research and development moulding trial.

Figure 3: Flow front recorded through sapphire window.

Flow visualisation

Further theoretical work is required to improve current understanding of microscale flow phenomena, which needs to be validated against high quality experimental data. Flow visualisation techniques offer accurate, full field evaluation of a range of flow phenomena and are regularly employed for characterisation of macroscale flows. The high magnification required for imaging of micromoulding flows, combined with the short filling times (typically submillisecond) requires the use of microscope objectives and high speed complementary metal-oxide semiconductor cameras that are able to capture images in excess of 10000 frames/s.

A system of this type has been created to study polymer flows in a range of micromoulding geometries. A transparent sapphire window forms the fixed surface of the mould cavity and a 45° first surface mirror is mounted behind to enable operators to view the cavity directly. A Mikrotron (www.mikrotron.de) 1310 full speed CameraLink camera is mated with a beam splitter and light source to allow coaxial illumination of the subject and data are collected and archived using a high specification PC (Figure 2). The speed of acquisition is limited by the data transfer between camera and computer. For example, full 1.3 megapixel images can be captured at 500 fps, but much higher speeds can be achieved by capturing data over a smaller pixel area (A frame 2403160 pixels allows a capture speed of over 17000 fps). The system currently offers a number of different process measurements, including:

• Flow front shape/velocity. The flow front geometry and location can be tracked from frame to frame and from this the flow front velocity can be calculated. Unexpected filling patterns and flow behaviours can also be evaluated (Figure 3).
• Shrinkage measurement. As the material cools in the cavity after filling has taken place, it shrinks away from the cavity wall. The sapphire surface is optically flat and the small air gap that is formed provides a striking Newton’s Rings pattern when using a monochromatic light source. These fringes can be reconstructed to provide a time-resolved contour that directly quantifies shrinkage behaviour.
• Birefringence. Incorporation of polarising elements in the optical train allows viewing of birefringence fringes caused by internal stresses in polymer materials, which are typically highly birefringent. This technique enables the measurement of stress-induced birefringence during the filling and cooling phases of the moulding cycle (Figure 4).

Figure 4: Cascade of sequential images showing stress birefringence forming during cooling (top). Fringes appearing on the left of each image are Newton’s rings caused by an air gap forming between the sapphire and the polymer as a result of material shrinkage.

Ultrasound flow and cooling measurement

Measurement of process conditions using a range of sensor technologies can help to optimise machine parameters to reduce cycle-to-cycle variability and highlight process failure. However, one of the most sensitive measurements, that is, melt pressure in the mould cavity using pressure sensors can be difficult to employ because of the physical size of the cavity and/or the complexity of the mould tooling. Research activity is in progress to develop novel sensor techniques that are optimised for the study of the micromoulding process and relevant materials.

Figure 5: Ultrasound sensor sprayed on rear face of mould insert.

Examples include piezoelectric film-based sensors that can be deposited directly onto the back of a mould cavity insert in a sol-gel solution (Figure 5). When hit with a square wave signal, these sensors resonate at frequencies of approximately 15 MHz. This creates an ultrasonic wave that propagates through the tool steel and is reflected by the cavity surface because of the high attenuation of high frequency ultrasound in air. The reflected sound wave is absorbed by the piezoelectric sensor to provide an electrical signal for measurement. When polymer flows into the cavity, a proportion of the ultrasonic signal is transmitted through the polymer and the amplitude of the first reflection drops and provides an indicator of the time of arrival of polymer at the sensor location. The signal transmitted though the polymer can also be detected and provides useful information on the temperature, pressure and morphology of the material. When the polymer melt cools and the material shrinks away from the mould surface, the amplitude of the ultrasound signal at the cavity surface rises back to its initial state and provides a guide for optimising cycle times to increase productivity. These sensors can be as small as 1 mm in diameter, with thicknesses of the order of tens of microns. Their compactness combined with the range of measurements that are achievable make them ideally suited for micromoulding applications.

Inline 3D product inspection

Optical product assessment (machine vision) systems can be employed to provide 100% inspection of mouldings and allow precise quality control. Recent developments in high-speed camera technologies allow 3D product measurement within the cycle time of the process. By taking a series of images at different heights on the product, a 3D image can be reconstructed to represent the actual form and dimensions of the product (Figure 6). These data can then be used for automated measurement of critical dimensions. which are referenced with the desired specifications. Products failing to meet the criteria can be ejected at this stage of the process to avoid any successive assembly and packaging operations and associated extra cost to the manufacturer.

Figure 6: Inline 3D product inspection system and reconstructed product data (inset).

Future developments

Research work performed thus far has highlighted the need for a greater understanding of the impact of the extreme processing environment on the flow behaviour and final properties of micromoulded components. Work is now ongoing to further develop measurement technologies and tooling that will provide information on the fundamental physical behaviour of a range of materials and processing parameters. This knowledge will be applied to advance the technology and satisfy the future demands of industry and result in even smaller devices and better control of surface features for medical applications.

Dr Ben Whiteside* is Technical Manager of the Centre for Micro- and Nanomoulding in the School of Engineering, Design and Technology., at Bradford University, Bradford BD7 1DP, UK,
tel. +44 1274 236 266 e-mail: b.r.whiteside@bradford.ac.uk, www.microandnanomoulding.co.uk

Peter Manser is Industry Associate at Bradford University, Accrutek, Leeds, UK, tel. +44 7702 785 001, e-mail: info@accrutek.co.uk.

The University of Bradford hosts the Centre for Micro- and Nanomoulding. The Centre has been expanded through open-access funding from Yorkshire Forward and the UK’s Department of Trade and Industry so that commercial development projects can be managed effectively alongside pure science based research and development. The processing laboratory is being extended to include an isolated area for medical clean-room activities. A dedicated characterisation laboratory allows product evaluation using a variety of advanced techniques such as scanning electron microscopy, atomic force microscopy and white light interferometry.

* To whom all correspondence should be addressed