Interview with Melinda Hale: Low-cost manufacturing of microfluidic devices

October 4, 2010

 Melinda Hale

 

Image credit: Melinda Hale

 

Hot embossing system. Image credit: Melinda Hale

 

Recently I was lucky to meet Melinda Hale, a graduate student at MIT in David Hardt’s group.  Melinda’s research focuses on developing low-cost systems for manufacturing microfluidic devices.  Here Melinda generously answers a few questions on the hot embossing system she developed.

 

What is hot embossing?  Why is hot embossing important for microfluidics?

Hot embossing is a simple method of manufacturing thermoplastic parts, consisting of three steps. First, a sheet of polymer material is heated (to make it soft), then a tool with the desired pattern is pressed into the substrate to form the part features. Finally, the part is cooled and removed from the tool. Hot embossing is one of a family of replication techniques that can be used to produce micron-scale features in polymers, and thus are well suited for manufacturing microfluidic devices. Other techniques include soft lithography (or casting), micro-injection molding, and ultraviolet embossing.

 

Hot embossing is important for microfluidics because it provides a low-cost way to both prototype and manufacture microfluidic devices. If you only need to make a few devices, most researchers will choose to make them using PDMS casting. If you are an established company needing to make ten thousand devices, often injection molding is used. But there are many applications that require hundreds to thousands of devices. In that range, hot embossing may be the best choice because you can make very good quality parts with equipment that has low capital cost, a high rate of production, and a flexible process to allow for quick and simple tooling changes.

 

What makes your hot embossing system lower cost?

A careful, minimal design using simple, readily-available components wherever possible. The equipment was designed to make one part (up to 1″ by 3″ size) at a time, as opposed to an entire wafer which would need to be diced apart later. This allows the equipment to have a smaller overall footprint, and requires less force so that the components can be smaller and cheaper. The finished equipment costs $10k, fits easily on a desktop, and produces one embossed part every two minutes with submicron variation.

 

What have been some of the greatest challenges in developing your system?  How have you addressed these challenges?

The first challenge was to reduce the cycle time of the hot embossing process, so it is more competitive with injection molding. The limiting factor in how fast parts can be produced is generally the speed of the heating and cooling cycle. To minimize the cycle time, two ideas were used. First, the cooling system was designed to be dynamic – the cooling block is actuated so that it does not touch the heated components during the heating cycle, then comes into contact only during the cooling step. Second, the number and size of machine components that have to be heated is minimized. The less heat energy that is put into the system, the faster it can be cooled.

 

The other challenge was the alignment of the tool to the substrate. Errors in either lateral and angular alignment can lead to unacceptable parts. This is actually helped by the “one part at a time” strategy. It is easier to align a 1″ by 1″ tool than across an entire 6″ wafer. In addition, the equipment incorporates kinematic material handling, to make sure the substrate is always aligned the same way every time a new part is placed into the machine.

 

In general, what do you see as the main challenges in manufacturing of microfluidic devices?

There have been many good perspectives on this topic – especially from Holger Becker in the “Focus” series in Lab on a Chip. Also, from Nathan Blow in “Microfluidics: the great divide.” I can only add that from my own experience and discussions with  microfluidic technology startups, one difficult issue seems to be transitioning from prototype manufacturing to full-scale production. Often the basic microfluidic technology has been proven in PDMS chip prototypes, but the commercial device needs to be in a thermoplastic material. Many manufacturing issues seem to arise when trying to switch manufacturing processes between prototype and finished device. My hope is that with a low cost, rapid-cycle hot embossing system available to researchers and startups, the same manufacturing process can be used both to prototype and to produce products at high-volume, and eliminate some of the challenge in manufacturing microfluidic devices.

 

If a researcher or startup is interested in implementing your system, what would the next steps be?

I would be happy to point them to my thesis, which includes the design and instructions for fabrication of the equipment. Or they are welcome to contact me, and I am happy to discuss.

 

 

Please reload