3D printing of microfluidic devices: An experimental evaluation of selected printers
29 January 2019
Figure 3: Microchips incorporating classic microfluidic features, designed using Fusion 360 software: (a) full-block enclosed channels with sharp bends, (b) full-block, enclosed, rectangular channels which change in dimension (from 1 mm high x 1 mm wide to 750 μm x 750 μm) as fluid moves through them, and Y-channels (inlet: 1 mm high x 500 μm wide; outlet: 1 mm high x 1 mm wide); (c) half-block straight rectangular channels and Y-channels, as for (b).
Katie Kilcoyne summarises the experimental evaluation of three, commercially available, 3D printers, each employing a different mode of printing, with reference to printing of simple open, simple enclosed and complex enclosed microfluidic channels, for microfluidic (MF) devices for potential lab-on-a-chip (LOC) and/or continuous pharma manufacturing applications.
Pharmaceuticals are typically produced via large-scale, equipment-intensive, multi-stage, batch processes. Continuous processing, in which raw materials are continuously added to a production system and products are continuously removed from it, potentially represents a major paradigm shift for pharma manufacturing (Lee et al., 2015), offering more flexible processing schedules, significant reductions in equipment size and the ability to more quickly respond to changes in product demand.
Central to continuous processing is continuous flow technology. Micro-fluidics encompasses both the study of the flow of small volumes (typically, nL-mL) of fluid through small (typically, 101nm-103 μm) channels, and the manufacture of the associated channelled devices, (flow cells).
This technology has been exploited for lab-on-a-chip (LOC) applications (Waheed et al. 2016), involving the miniaturisation of chemical and biological analytical steps, implemented in ‘chips’ ranging in size from mm2-cm2.
For chemical reactions, flow cells could overcome common chemical engineering problems associated with scale-up: to increase the size of a process train, more chips would simply be added, scaling processes out, rather than up, while offering greater scope for automation and in-line monitoring of process variables.
To date, the potential of LOC technology and mass manufacturing of drugs has not been realised, largely due to limitations MF device manufacturing. Existing methods are, in general, slow, expensive and not amenable to mass production.
Recent developments in 3D printing offer a new path to quick, cheap production of a complete microfluidic device in a single step from a computer model. The ability to create fully realised 3D structures facilitates new microfluidic capabilities by exploiting the z-axis dimension in a manner not previously possible.
However, not all 3D printers are amenable to such applications. The printing technology selected must produce high-resolution, highly detailed structures, including 3D internal microfluidic features. Extraneous support material should be easy to remove, surface-roughness should be sufficiently low to preclude irregular channel dimensions and the prints should be consistent and replicable.
Three, commercially available 3D printers, each using a different printing technique, were investigated, in order to evaluate their potential for manufacturing MF devices: the Ultimaker 2, FFF (fused filament fabrication); the Objet Eden 260 V (polyjet printing) and the Form 2, SLA (stereolithography).
The devices were compared on the basis of performance in printing (A) open, (B) enclosed and (C) intricate-feature microfluidic channels; full factorial design was employed, where appropriate. Channel designs were developed using Fusion 360 SolidsWorks software and exported in STL format, compatible with printer software.
Printed chips were compared, visually and via SEM images, on the basis of channel integrity, print accuracy, consistency and surface roughness.
A./ Open-channel printing eliminates the use and removal of support material and facilitates comparison of printing resolution and accuracy across printing techniques.
On the basis of a full factorial design, six samples (chips, 33 x 25 x 2 mm (x, y, z)), each incorporating 10 rectangular microchannels, were printed in each printer. On a single chip, channel width was fixed (in the range 100 μm -1 mm) and channel height varied (in the same range).
Figure 1 shows SEM images of printed microchannels (nominally, 500 μm wide, 500 μm high) from each of the printers. It is immediately apparent that, although intended to have identical dimensions, channel features varied significantly with printing technique.
Figure 1: SEM images of simple open channels (model dimensions: width 500 μm, height 500 μm) printed using the (a) Form 2 (b) Objet Eden 260 V and (c) Ultimaker 2.
Of the three printers, the Form 2 (Figure 1(a)) best preserved the rectangular structure; prints were characterised by high accuracy, low surface-roughness and least channel-width variation.
The Objet Eden260V (Figure 1(b)) yielded the lowest surface-roughness; however, because of the shrinkage of the Veroclear™ material and forces of attraction at the top of the microchannels, there was significant channel sloping, with loss of the desired rectangular structure.
The Ultimaker 2 (Figure 1(c)) produced channels 150−200 μm smaller in width than specified, due to thermoplastic ‘spilling’ of the thermoplastic over the print lines.
Additionally, the prints were characterised by high surface-roughness and channel width variation associated with the visible print lines.
Based on analysis of the accumulated open channel data (results not shown), the Form 2 performed best, reliably printing open channels of width 500 μm and height 100 μm, with average errors in both dimensions of less of 25 per cent. To reduce these errors, channel dimensions must be increased.
B./ Closed channel printing ability was tested using the Form 2 (which requires no internal support material) and the Objet Eden 206 V (which uses a water-soluble support mixture of propylene, polyethylene, acrylic monomer and glycerine; soaking in a two per cent NaOH solution can improve removal (Stratasys, 2013)).
The same factorial design and same chip dimensions as for open channels (A) were employed. The Ultimaker 2 was excluded, as it prints a solid support material (ABS Yellow), impossible to remove from enclosed microchannels; additionally, enclosed channels cannot be readily visualised in the non-transparent ABS.
From the sample chips shown in Figure 2, it can be seen that problems were encountered with both devices. With the Form 2 (Figure 2(a)), although there is no support material, resin entrained within the channels during printing and during post-print curing seals the channel.
The smallest micro-channel successfully printed and (manually) cleared of resin was 1 mm high and 500 μm wide. With the Objet Eden260V (Figure 2(b)), obstacles were associated with removal of support material.
After 72 hours of sonicated immersion in a two per cent NaOH solution, followed by removal of dislodged material with compressed air, the smallest channel successfully cleared was 1 mm high and 400 μm wide.
C./ Complex Closed Printing, including printing of features that could truly exploit fluid properties and their interactions on the microscale, must be possible if 3D printing is to surpass current methods for MF cell manufacture.
Figure 3 (main image) shows samples of the enclosed, complex channel features with which both the Form 2 and the Objet Eden 206V were challenged. In preliminary trials with the Form 2 (results not shown) channels did not form, most likely due to retention of entrained resin during curing.
For this reason, half-block designs (Figure 3 (c)) were developed; the blocks were assembled using o-rings with nut-and-bolt clamping.
Sample completed chips are shown in Figure 4. Although leaking occurred with several Form 2 half-block prints (Figure 4(a)), the results showed potential for further investigation.
With the Objet Eden260V prints, partially cleared channels (Figure 4 (b)) reflect difficult in removing support material.
Conclusions and recommendations
On the basis of this work, manufacturing of microfluidic devices using additive manufacturing methods is feasible. Channels in the micron range were successfully printed using each of the 3D printers compared in Table 1.
At $2,500, the Ultimaker 2 (FFF) is a very affordable option; however, high surface-roughness and ‘spilling’ effects limits its potential for MF devices.
The Objet Eden 260 V ($20,000) showed most promise for closed microfluidic structures (> 100 μm), although diminished by challenges in clearing support material from smaller channels: the smallest straight microchannel successfully cleared was 1 mm high x 400 μm wide. The Form 2 (SLA) (about $4,000) is a low-cost desktop solution.
With high-resolution and no requirements for support material, it is an attractive prospect for MF device printing; its smallest, successfully cleared channel was 1 mm high x 500 μm wide.
For all devices, further research is required: for the Ultimaker 2, investigation of dual-head printing for easily removable support material and smaller nozzle sizes to reduce surface roughness and thermoplastic ‘spilling’; determining a more effective way to remove support material from Objet Eden260V prints; exploring the Form 2 failure to produce smaller channels under the conditions investigated and addressing pre-curing cleaning issues.
3D printing of MF devices is promising, but more work is essential before it can be applied to mass-production of devices for LOC and pharma manufacturing applications.
But, given the potential for reduced space-time demands, reduced costs and greater flexibility relative to current alternatives, it is most certainly worth the investment.
1.) Casquillas, G. & Houssin, T. (2015). Introduction to lab-on-a-chip 2015: review, history and future [online]. Elveflow. Available at: www.elveflow.com/microfluidic-tutorials/microfluidic-reviews-and-tutorials/introduction-to-lab-on-a-chip-2015-review-history-and-future/ [Accessed 19 Apr. 2018].
2.) Kotz, F., Arnold, K., Bauer, W., et al. (2017). Three-dimensional printing of transparent fused silica glass, Nature, 544: 337-337.
3.) Lee, S.L., O’Connor, T.F., Yang, X. et al. (2015). Modernising Pharmaceutical Manufacturing: from Batch to Continuous Production. J Pharm Innov, 10: 191-199. DOI 10.1007/s12247-015-9215-8.
4.) Stratasys. (2013). Eden260V 3D Printer System, User Guide. DOC-32011 Rev. A. Available at: www.aetlabs.com/wp-content/uploads/2015/12/DOC-32011_A_Eden260V-UserGuide-.pdf [Accessed 19 Apr. 2018].
5.) Waheed, S., Cabot, J. M., MacDonald, N. P., et al. (2016). 3D printed microfluidic devices: enablers and barriers, Lab Chip, 16(11): 1993-2013. DOI 10.1039/c6lc00284f.
6.) Yazdi, A.A., Popma, A., Wong, W. et al. (2016) 3D printing: an emerging tool for novel microfluidics and lab‐on‐a‐chip applications. Microfluid Nanofluid, 20: 50. DOI 10.1007/s10404-016-1715-4
Author: Katie Kilcoyne is a 2018 BE graduate in chemical and bioprocess engineering from University College Dublin, currently working with Jacobs, Dublin. On the basis of the work described in this paper, she was one of two 2018 BE chemical and bioprocess engineering graduates jointly awarded the Carthy Graduate Research Project Award 2018, kindly endowed by Mark Carthy, a 1982 UCD chemical engineering graduate, currently managing partner, Orion Healthcare Equity Partners. Kilcoyne is the winner of the 2018 Talking Engineers competition, organised by Engineers Ireland Young Engineers.http://www.engineersjournal.ie/2019/01/29/3d-printing-of-microfluidic-devices-an-experimental-evaluation-of-selected-printers/http://www.engineersjournal.ie/wp-content/uploads/2019/01/a-aaaaaabc4.jpghttp://www.engineersjournal.ie/wp-content/uploads/2019/01/a-aaaaaabc4-300x300.jpgChem3D Printing,chemical and process,UCD