3248
A N A LY T I C A L C H E M I S T R Y / M A Y 1 , 2 0 0 7
When PDMS isn’t the best What are its weaknesses, and which other polymers can researchers add to their toolboxes?
RAJENDRANI MUKHOPADHYAY
I
f a popularity contest for materials were held in the microfluidics community, the silicone PDMS would win hands down. It’s easy to pattern by soft lithography, optically transparent, flexible, gas-permeable, and cheap enough for students to use in copious quantities without denting lab budgets. These qualities make PDMS an excellent material for the rapid prototyping of microfluidic devices. In that regard, it’s practically irreplaceable. But experts say that although the material is attractive for quickly testing the fluidics of new device designs and for cell-based studies, it has problems. “There’s no perfect material. There’s no perfect instrument. There’s no perfect technique. PDMS is not an exception,” explains Daniel Chiu at the University of Washington. “It’s a great material, but you need to know what you’re using it for and know its properties.” Issues with PDMS include absorption of organic solvents and small molecules, its innate hydrophobicity, and evaporation of water. Some experts say that PDMS’s shortcomings become obvious as you spend time working with it. You soon learn to design around the limitations and use the silicone to its best advantage. Other experts argue that not everyone is keenly aware of PDMS’s shortcomings and that many use it simply because it’s convenient, not because it is the wisest choice for the job at hand. Andrew Kamholz of Edge Embossing LLC says, “There are so many papers out there where PDMS is being used. You can get some good results with
© 2007 AMERICAN CHEMICAL SOCIETY
it, so people are quick to say, ‘Look, everyone else is doing it, and I’m going to do it too.’ But I’m not sure that everyone goes through the process of thinking about whether PDMS is going to be an issue for them.”
Quick ’n’ easy The microfluidics community has embraced PDMS for many reasons. It allows simple, planar systems to be replicated and produced easily. Complex 3D structures and microchannel networks can be fabricated quickly in PDMS by multilayer prototyping approaches. The material is transparent from 240 to 1100 nm, so various optical detection schemes can be used; even optical elements can be created out of it. Because of PDMS’s elastic properties, micromechanical valves, developed by groups headed by Stephen Quake at Stanford University and Richard Mathies at the University of California Berkeley, are best made with it. For cell-based applications, the silicone is attractive because it’s nontoxic and gaspermeable. “One reason why people like PDMS is that it doesn’t break,” points out George Whitesides of Harvard University. “When you’re trying to do things with glass, you’re always worried about sharps and disposal problems. None of this is an issue with PDMS because it’s soft.” Christopher Culbertson of Kansas State University adds that using PDMS “requires substantially less skill than making glass chips, is cheaper, and if you drop the chip, it’s going to bounce, not break, on you.” M A Y 1 , 2 0 0 7 / A N A LY T I C A L C H E M I S T R Y
3249
PDMS is unique in that, when freshly plasma-oxidized, it can be sealed to itself and other materials without an adhesive. “As far as I know, every other polymer requires an adhesive,” says Whitesides. “Using an adhesive in a system with 50-µm channels on its surface can be very challenging and may not lend itself well to manufacturing.” For all these reasons, researchers like to use PDMS even when they know it’s not the best material. “For example, for electrophoretic separations, it’s not the greatest surface, but it enables you to quickly turn over designs and do rapid tests,” says Nancy Allbritton of the University of California Irvine. “I use PDMS as an education tool for students,” says Andreas Manz at the Institute of Analytical Sciences (Germany). “Every single Ph.D. student will have to undergo that education process and get a few chips out. That’s why the publication will feature PDMS—not because I think it’s relevant but because I think it’s easy-access, quick-and-dirty. The students can make all the mistakes at low cost and gain some experience.”
The absorption problem
PMMA
But some worry that the very fact that PDMS is easy to work with leads researchers to overlook its drawbacks. Take, for example, the shortcoming that David Beebe and Michael Toepke at the University of Wisconsin Madison focused on in a Lab on a Chip paper: PDMS’s penchant for sucking up small, hydrophobic molecules from solution (1). The fact that PDMS absorbs molecules isn’t a new discovery—the material has been previously used, for example, as an extraction matrix to remove trace organic compounds from solutions. Beebe says researchers in his lab and other groups in the microfluidics community were aware of PDMS’s absorption properties, but “we, like everyone else, just ignored it because it was inconvenient.” A few years ago, Beebe was granted a National Institutes of Health K25 award that allowed him to add to his electrical engineering skills by training in cancer biology (2). The trip back to school made Beebe aware that the microfluidics community couldn’t afford to continue ignoring issues like absorption. Many microfluidics efforts focus on cell biology, and Beebe realized that biological molecules and drugs were just the types of molecules that would fall victim to the silicone’s tendency to soak things up. “If you’re going to look at cell signaling, it’s just a big black box,” says Beebe. If PDMS is the device material for a cell-signaling assay, “you’re compounding the black box because there’s a whole segment of molecules you might be interested in looking at that are just disappearing on you.” To get an idea how much of an effect absorption had on assays, Beebe and Toepke quantified the absorption of small, hydrophobic molecules by PDMS microchannels. They tested Nile red, a low-molecular-weight, hydrophobic fluorophore, and quinine, a small hydrophobic drug. They saw that Nile red rapidly partitioned into the PDMS around the inlet port and didn’t fill the channel. Rinses with water and detergent couldn’t make the Nile red budge from the PDMS. With quinine, the investigators 3250
A N A LY T I C A L C H E M I S T R Y / M A Y 1 , 2 0 0 7
showed that its absorption depended on pH. Beebe and Toepke concluded that the absorbance of small molecules into PDMS could have a profound effect on the outcome of drug screening studies with microfluidic devices, particularly when fixed volumes of drug were used to determine dose–response. PDMS also absorbs hydrocarbon solvents—and swells up like a sponge when it does (3). “It is, after all, a ‘solid solution’, if you want to think about it that way,” says Whitesides. “It’s just a hydrocarbon. Any molecule that you would expect to dissolve in a hydrocarbon will dissolve well in PDMS.” Joseph DeSimone at the University of North Carolina Chapel Hill says PDMS is not a very solvent-resistant material. “So, when you think about PDMS and microfluidics, it’s relegated to water-based chemistry, which cuts out a lot of applications. Then, even in the water-based chemistry, there are a number of solutes with a slight lipidic or hydrophobic-like character that will partition into the silicone.”
Other problems
PDMS’s hydrophobicity also has been a noted stumbling block. Although the surface can be made hydrophilic, for instance through a plasma treatment, this is not a natural act for the polymer, and it will revert back to its hydrophobic self in air. The relapse can be a major problem because unwanted phenomena, such as adsorption of proteins to the surface, start to happen. But in some cases the relapse proves to be useful. Chiu says, “Sometimes, when we do work with droplets, we take advantage of the fact that PDMS can revert back from hydrophilic, sustaining EOF [electroosmotic flow], to hydrophobic because droplets can form in a hydrophobic channel.” Another issue is the evaporation of water. “Frequent users conceptually appreciated, but not as much quantitatively analyzed and utilized, the fact that water evaporates through PDMS,” says Shuichi Takayama at the University of Michigan. The silicone is thought to be a perfect material for cell-based studies because it allows oxygen and carbon dioxide to freely diffuse through it. But evaporation through PDMS becomes critical because several cell types are sensitive to changes of osmolality in the cell culture media. Takayama says, “To be able to use PDMS in cellular studies, you have to understand this.” To prove the point, he and colleagues quantified the effect. They studied the rate of osmolality shifts with PDMS membranes of various thicknesses inside a humidified incubator over a 4-day period. In spite of the humidified environment, significant evaporation occurred from the devices that contained a thinner PDMS membrane; in these cases, the shift in osmolality was more dramatic. The researchers observed a correspondingly worse development rate for mouse embryos housed in the chips—most of them stopped developing at the two-cell stage, or in ~24 h (4). Stephen Jacobson’s group at Indiana University ran into the problem of getting exact replicas of features in PDMS that were
