Micromilling: a method for ultra-rapid prototyping of plastic microfluidic devices

Abstract

This tutorial review offers protocols, tips, insight, and considerations for practitioners interested in using micromilling to create microfluidic devices. The objective is to provide a potential user with information to guide them on whether micromilling would fill a specific need within their overall fabrication strategy. Comparisons are made between micromilling and other common fabrication methods for plastics in terms of technical capabilities and cost. The main discussion focuses on "how-to" aspects of micromilling, to enable a user to select proper equipment and tools, and obtain usable microfluidic parts with minimal start-up time and effort. The supplementary information provides more extensive discussion on CNC mill setup, alignment, and programming. We aim to reach an audience with minimal prior experience in milling, but with strong interests in fabrication of microfluidic devices.

Figure 1

A schematic showing the basic components of a CNC mill, which can use computer-aided design (CAD) models to produce finished devices. The mill consists of a worktable (to provide motion in the XY-plane), a cutting tool (to remove material from the workpiece), and a spindle (to hold the cutting tool, spin the cutting tool, and provide motion along the Z-axis).

Figure 2

A comparison between milling and other microfabrication methods for plastics, in terms of: (A) material compatibility, feature capability, and quality; and (B) cost. In (A), three filled circles = “excellent”, three open circles = “impractical” or “inadequate”; see legend (bottom left of A). In (B), for process times, “Time” represents the time of fabrication for one device for both on-site and outsourced devices. “Cost” (in USD) is an estimate, where onsite fabrication is calculated based on cost of goods (not including labor) used (estimated from the labs of the authors), and outsourced fabrication is based on the lowest quoted price we obtained for the different quantities. N/A = not applicable.

Figure 3

(A) CNC mills from several manufacturers are compared and categorized into price ranges. Costs were assessed based on quotes of the lowest level mill from each manufacturer, except for the Tormach mill, which was quoted to be comparable in terms of capabilities to the other CNCs. Unlisted specifications were not given by the manufacturer. (B) Endmills – the most common cutting tool for milling – are available in many profiles, in a variety of materials, and with a variety of coatings. Mills are also compatible with a variety of other cutting tools, some of which are shown. (C) Endmills are defined by several characteristics, each of which contributes to the endmill capabilities and feature quality.

Figure 4

Surface roughness and resolution using an entry-level CNC mill. (A) Colored contour plots showing surface roughness as a function of spindle speed (y-axis) and feed rate (x-axis). Surface roughness was measured by interferometry (see SI), and ranged from 0.420 to 1.52 µm (root-mean-squared averages, color legend, right). Black dots are speed and feed conditions tested (n = 3 samples per dot), while colored contours are interpolated data. Speed and feed conditions that resulted in broken endmills are marked with a red “X”. Graphs are arranged in a 3×3 matrix representing data for three different plastics (PS, PMMA, COC), each tested with three different endmill sizes (127, 254, 508-µm diameters). Resolution in the (B) XY-plane and (C) the vertical z-axis were assessed by comparing the actual size of a fabricated feature to its target “nominal” size (i.e., tolerance or accuracy). (n = 3 samples for all conditions; error bars = standard deviation, represents precision; p = 0.79 via Bartlett test for (B)). (D) SEM micrographs of the features used to characterize the resolution. Red arrows point out the ability to make sharp internal corners via embossing, while rounded fillets form for a pocket made via micromilling.

Figure 5

Cell culture and image analysis in milled microchannels. (A) Channels are assembled in three configurations: (1) A milled channel with ports is bonded to a cover layer, (2) a milled port layer is bonded to a milled channel, and (3) an embossed channel with ports is bonded to a cover layer. A microtiter plate is used as a control. Mammalian cell lines are cultured for 48 hours in each configuration, then assayed for cell viability (error bars represent one standard deviation, N=3). No statistically significant difference was observed between culture methods (p > 0.14 in all cases, Students T-test). (B) Phase contrast and fluorescent images were taken of HS-5 stromal cells in each channel configuration using 4, 10, and 20× magnifications. (cell culture described in SI)