OpenMFDA Foundry

Hardware and process

What our chips are printed on, and what with. Every number on this page says where it came from, so you can tell what we measured ourselves, what we took from someone else's published work, and what we simply do not know.

How to read this page

Published
A number from a published scientific paper, or from the specification a printer's maker publishes. We did not measure it ourselves. It describes someone else's printer, resin or process. We give it because it is the best description available, not because we have repeated it here.
Our records
A number written down in one of the lab's own project files: a design file, a configuration, a script or a report. We name the file and give the number here: most of these files are in the lab's working copy of the design software, not on this site, so you may not be able to open them.
Our account
Something we know, but have not written down in any project file. It lives in people's memory and in conversations, so there is no record to point you at and no way for you to check it. These are the facts most likely to be lost. One of them survives only as long as the person who remembers it.

Nothing published on this site has been proven correct on any printer. This page exists so that the figures which are in circulation can be read with their provenance attached. Our records items name a file: one on this site is linked, and any other is given relative to the lab's working copy of the design software, openmfda_flow. Our account items have no project file to name — that absence is exactly what the label is telling you, and it is the whole reason the label exists. Where a figure records something that physically happened, the path it cites is docs/process-record/, which is this project's empirical record.

1. Printers

Two printers appear on this page. They work at very different scales, so a number that applies to one tells you nothing about the other. Ten dots of light measure about a fifth of a millimetre on the Phrozen Sonic Mighty 12K, and less than a tenth of a millimetre on the BYU printer.

Technical details

Ten pixels is 0.19 mm across and 0.24 mm down on the Phrozen Sonic Mighty 12K, and 0.076 mm either way on the BYU printer.

Each printer below is described in the same three parts, in the same order: What it is, Pixel size and What is not known. The BYU printer's section then has a fourth part, the OS1: a printer design the Nordin group has released as open hardware. Its figures are BYU's published figures for the OS1.

Phrozen Sonic Mighty 12K

Where it is: in the Gale Lab. What it is used for: it is the printer this project's chips are meant for. Twenty of the library's 21 reference chips name it as their printer (the twenty-first, the BYU pump-mixer, is laid out for the BYU printer), and the project's current printer profile is written for it. Today the lab uses it for its early test prints of pumps. Our account

What it is

propertyvaluetiersource
Model Phrozen Sonic Mighty 12K Our account The badge on the machine, read and confirmed by the lab. No tracked file names it; see the history below.
XY pixel pitch 19 µm (X) × 24 µm (Y) — non-square Our account Derived from the model: 11520 × 5120 over a 218 × 123 mm build area. Labelled "Our account" rather than "Published" because it follows from knowing the model, and the model comes from reading the badge on the machine rather than from a file.
Native resolution 11520 × 5120 Published Vendor and multiple retailer specifications for the Phrozen Sonic Mighty 12K agree.
Build area 218 × 123 × 235 mm Published Vendor and multiple retailer specifications agree.
Layer height in use 10 µm (0.01 mm) Our records openmfda_flow/flow/platforms/h.r.3.3/config.mk — LAYER_VAL = 0.01
Light source 405 nm Our account Stated by the lab. physical_interface.json carries the pixel pitch but not the wavelength, and no other tracked file records it.
Slicer profile in use a Phrozen Sonic Mighty 8K profile, in the one slicer file on record Our records see the slicer profile below

Pixel size

Three pitches, and which one is real. The machine was identified from the badge on it. Until that happened, three different pixel pitches were in circulation in this project, attached to three different machine names. All three are recorded here, because the history explains where the figures in the older files came from.

namepitchwhere it came fromstatus
Phrozen Sonic Mini 8K 22 µm A 7.1″ machine with a 165 × 72 mm build area. Its specification is where the project's 22 µm figure came from — the lab's printer was misidentified as this model. The name still appears in May 2026 files. not the machine
Phrozen Sonic Mighty 8K 28 µm The profile the one slicer file on record was produced through. A third machine again. not the machine
Phrozen Sonic Mighty 12K 19 × 24 µm The badge on the machine in the lab. Our account

This is a documented fact about the project, not a confession. A pixel pitch that travels between files without a machine attached to it is exactly the failure this page was built to make visible, and it is the reason every figure here carries where it came from.

What the corrected pitch does and does not affect

The obvious reading of a wrong pixel pitch is that everything printed came out the wrong size. That is not what happened, and it is worth being precise about.

Physical dimensions are not affected. The SCAD and the STL are in millimetres. The printer rasterises millimetres using its own pitch, whatever that pitch is. A chamber the tool reports as 8.8 mm prints at 8.8 mm. Volumes are correct. Nothing printed was the wrong size.

Non-square pixels do not distort parts. A feature designed 880 × 880 µm prints 880 × 880 µm. It simply occupies a different pixel count on each axis — about 46 across and about 37 down, rather than 40 and 40.

What is affected is every figure expressed in pixels, because the pixel is the wrong size. A count, a clearance or a minimum feature quoted in px means something different at 19 or 24 µm than it did at 22.

The files give lengths in millimetres, and the printer rounds them to its own pixels. The lab has accepted rounding to the nearest printable position. Worst-case quantisation error is half a pixel: ±9.5 µm in X and ±12 µm in Y.

Layer-height-based dimensions are unaffected entirely. Membrane thickness and chamber depth are counted in layers, and the layer height has not changed.

The slicer profile

The one slicer file on record used a Phrozen Sonic Mighty 8K profile

The one slicer file on record was produced through a Phrozen Sonic Mighty 8K profile — a 28 µm machine, not the 19 × 24 µm machine in the lab.

If the slicer is still set that way, files are being rasterised for a different machine than the one printing them. Parts still come out the right physical size — millimetres are millimetres — but the pixel grid the geometry is quantised onto, and the exposure schedule applied to it, belong to another printer.

Custom DLP-SLA printer, built by the Nordin group at Brigham Young University

Where it is: at Brigham Young University (BYU), where the Nordin group built it. It is a custom research instrument, not a commercial product; on this page it is called the BYU printer for short. What it is used for: this project keeps it for its finest, high-accuracy prints. In June 2026 a collaborator at BYU printed five pumps on it, and the lab tested them: some worked and some did not (see §8). The lab calls this printer the H.R.3.3, the printer the README of the lab’s process design kit names (utah-MFDA/h.r.3.3_pdk, read 2026-09-30). Our account This project's h.r.3.3 platform also keeps a profile for it. The OS1 (below) is the printer design the Nordin group has released as open hardware; what BYU publishes about it is given there as BYU's. The Nordin group prints microfluidic chips with very small channels on it: their papers report flow channels as small as 18 µm × 20 µm, and those published figures are quoted elsewhere on this page.

The rows marked Published come from the papers describing that instrument, and none of them has been reproduced here.

What it is

propertyvaluetiersource
Type Custom DLP-SLA printer Published Gong, Bickham, Woolley, Nordin, “Custom 3D printer and resin for 18 µm × 20 µm microfluidic flow channels”, Lab on a Chip, 2017 [1]
Light source 385 nm LED Published Gong 2017 [1], read 2026-09-30; the 2021 paper [2] gives 385 nm as well. Chosen over 405 nm because 385 nm greatly widens the range of usable UV absorbers. Goenner et al. 2025 [9] give 365 nm for the printer used in that paper, as the OS1's Specs page [7] does for the OS1. For both figures, see the OS1's wavelength.
XY pixel pitch 7.6 µm in the projected image plane Published Gong 2017 [1]. See the pitch note below.
XY pixel pitch as this project stores it 7.6 µm (0.0076 mm) Our records openmfda_flow/flow/platforms/h.r.3.3/config.mk — PX_VAL = 0.0076; physical_interface.json profile byu_hr33_7p6um
Later designation “High Resolution 2” (Generation 2) printer Published Nature Communications 12, 2021, s41467-021-25788-w [2]
Light engine Visitech (Lier, Norway) Published Gong 2017 [1]; ASME, “Regime Change in 3D Printed Microfluidics”, 2017 [3]
Optical path 45° turning mirror, three axes of adjustment Published Gong 2017 [1]; ASME 2017 [3]
Mechanism Heavily modified Solus DLP-SLA (Junction 3D, Santa Clarita, CA), custom mounts Published Gong 2017 [1]; ASME 2017 [3]
Smallest demonstrated flow-channel cross section 18 µm × 20 µm Published Gong 2017 [1]
Substrate 25 mm square silanized glass slides Published Nat Commun 2021 [2] — rinsed with acetone and IPA, then immersed 2 h in toluene with 10% 3-(trimethoxysilyl)propyl methacrylate

Pixel size

Pitch note. The pixel pitch is given as 7.6 µm in the papers and 7.56 µm in Gong's dissertation (BYU ScholarsArchive etd/7690 [4]). This project's files use 7.6 µm. The discrepancy is recorded here.

The OS1: a printer design BYU has released as open hardware

BYU's description

Everything in this part is BYU's description of the OS1, in BYU's figures. The BYU printer this project designs for, the H.R.3.3, is described above; the five pumps in §8 were printed on it.

What it is. The OS1 (Open Source 1) is a DLP printer for microfluidic devices, designed by the Nordin group at BYU and published as open hardware. Like both printers above, it is a vat photopolymerization printer: an image projected into a vat of liquid resin cures the part one layer at a time. BYU presents it as the result of more than ten years of the group's work on its own printers. The features its repository lists are automatic levelling of the build surface, automatic focus calibration, focus kept across the build area by a confocal distance sensor, the light engine's output corrected with a fibre-optic photodiode, and greyscale correction of the projected image for a more even exposure [8].

Where to get it. The design files are in the Nordin group's repository on GitHub, the OS1 repository of 3D-Printing-for-Microfluidics [8]. BYU's website for the printer is os1.byu.edu [7]. This site copies nothing from either: the files, the text and the pictures are BYU's, and the links lead to them.

What the repository holds, as read on 2026-09-30, at commit 4d22c93 [8]:

  • CAD models of the whole printer and of each part made for it, as STEP files; a 3MF file for each part to be printed on a filament (FDM) printer; and DXF cutting layouts for the aluminium and polycarbonate parts, which are cut on a waterjet;
  • the design and manufacturing files for two circuit boards made for it: a load-cell board, drawn in Eagle, and a motor board for levelling the build surface, drawn in KiCad;
  • the bill of materials, as an Excel workbook, grouped to follow the build manual;
  • the build manual, and a guide to setting up the Raspberry Pi that runs the printer, both as PDF;
  • a Supplemental folder with a STEP model and a second build manual for the printer fitted with a newer light engine: see the caveat below.

The printer's control software is not in it. The Raspberry Pi guide installs it from a separate BYU repository, 3D_printer_control.

Licence. The repository is licensed under the CERN Open Hardware Licence version 2, strongly reciprocal, CERN-OHL-S-2.0, as read on 2026-09-30 [8]. In brief: anyone may study, change, make and share the design, and whoever shares it, or a product made from it, must share the complete design files under the same licence. That summary is ours, and the licence text is what governs. It is not this site's MIT licence, which does not extend to BYU's files.

BYU's caveat: the light engine is discontinued

The OS1 is built around a Visitech LRS WQ light engine, and the repository's README says Visitech has discontinued it. Its replacement, the LRS WQ Plus, talks to the computer in a new way: it needs a new driver, one that does not send the image from the Raspberry Pi over HDMI, and the current OS1 software cannot drive it. The Supplemental folder has a STEP model of the printer with the WQ Plus and a second build manual for it [8].

BYU's figures. Each figure below is BYU's published claim about the OS1, with the date it was read. BYU's Specs page gives all its values as approximate.

propertyvaluetiersource
Pixel size 7.6 µm Published BYU, OS1 Specs page [7], read 2026-09-30. BYU labels this figure “Validated”.
Projected image 2560 × 1600 pixels Published BYU, OS1 Specs page [7], read 2026-09-30. On the BYU grid the design tools compare a block with this, in pixels, either way round, and flag one bigger both ways without refusing the files.
Layer height 10 µm Published BYU, OS1 Specs page [7], read 2026-09-30
Build area 19.5 × 12.2 mm Published BYU, OS1 Specs page [7], read 2026-09-30. 2560 × 1600 pixels of 7.6 µm are 19.456 × 12.16 mm, which this figure rounds; the design tools show it beside the pixel count and compare the pixels.
Wavelength 365 nm Published BYU, OS1 Specs page [7], read 2026-09-30. Goenner et al. 2025 [9], read 2026-09-30, give 365 nm for the printer used in that paper.
Wavelength, in the group's earlier papers 385 nm Published Gong et al. 2017 [1], read 2026-09-30, and the 2021 paper [2]. Both describe earlier printers of the group, not the OS1.
UV absorbers in the resin NPS and avobenzone Published BYU, OS1 Specs page [7], read 2026-09-30. The same two absorbers are in the resin of the group's 2026 multiresolution paper [10], read 2026-09-30, which BYU's website links. That paper describes a different printer, with two light engines, and does not mention the OS1.
Printing accuracy ±2 µm Published BYU, OS1 Specs page [7], read 2026-09-30. BYU labels this figure “Validated”.
Minimum channel width about 18 µm Published BYU, OS1 Specs page [7], read 2026-09-30. BYU labels this figure “Validated”.

Which wavelength. Two are published for the Nordin group's printers: 365 nm, on the OS1's Specs page [7] and in Goenner et al. 2025 [9], and 385 nm, in Gong et al. 2017 [1].

BYU's own prints and papers. BYU's OS1 page titled “Validation” lists papers by the Nordin group from 2015 to 2026, and its Gallery shows the printer, renderings of it, and photographs and micrographs of chips the group has printed, with no captions saying which printer made each [7]. Both are the Nordin group's own work, published by BYU. None of it is a print made for this project: this project's prints are in §8, and BYU's are not counted there.

2. Resin

There is no proprietary formulation here. Both recipes below are public.

formulationcompositiontiersource
Gale Lab resin Phrozen Sonic Mighty 12K 98% PEGDA + 1% NPS + 1% Irgacure 819 Our account Stated by the lab; the formulation appears in no tracked project file. NPS = 2-nitrophenyl phenyl sulfide (UV absorber); Irgacure 819 = phenylbis(2,4,6-trimethylbenzoyl)phosphine oxide (photoinitiator).
BYU published resin BYU printer PEGDA MW258 + 1% (w/w) Irgacure 819 + 2% (w/w) NPS Published Gong 2017 [1]; Nat Commun 2021 [2]
Optical penetration depth Dp BYU published resin ≈ 11.2 µm at 2% NPS Published Gong 2017 [1], section 3.2: the paper’s ha, 11.2 µm for its resin with 2% NPS, the BYU published resin above. Not the Gale Lab resin, which has 1% NPS. Read 2026-10-01.
Optical penetration depth Dp Gong 2017, 3% NPS ≈ 8 µm at 3% NPS Published Gong 2017 [1], section 3.3.1: ha = 8 µm for its resin with 3% NPS. Not the Gale Lab resin, which has 1% NPS. Read 2026-10-01.

The two recipes are not the same system

The Gale Lab resin runs NPS at 1% where the published formulation uses 2%, and runs on a 405 nm engine where the published work uses 385 nm. NPS absorbs weakly at 405 nm, while Irgacure 819 remains sensitive to roughly 440 nm. Optical penetration into already-printed layers is therefore larger than in the published system.

This is a documented property of the combination, not a fault. It means the published BYU figures — the 18 µm × 20 µm channel, the penetration depths above — describe a more strongly absorbing resin under a shorter wavelength than the Phrozen Sonic Mighty 12K runs, and should not be read across to it.

Safety, protective equipment and disposal

Handle the resin as its supplier's safety data sheet directs.

Work from the suppliers' safety data sheets for the specific PEGDA, NPS and Irgacure 819 products you hold.

3. Supporting equipment

What is recorded is the chip-side interface geometry.

itemwhat is recordedtiersource
Fluid interface Top +Z fluid cups, 2.5 mm diameter × 3.0 mm deep, six-cup grid Our records openmfda_flow/flow/platforms/h.r.3.3/physical_interface.json — fluid_interface
Fluid loading Syringe-loaded, not manifold-mated Our records physical_interface.json — substrate_cup_mode: minimal_active_cups
Pneumatic face +Y face, 32 annular gasket seats — OD 2.00 / ID 1.70 / depth 0.30 mm Our records openmfda_flow/flow/platforms/h.r.3.3/pdk/Components/pumps/block_3pumps_91px_22um_0/ — topology record
Of those 32 seats 18 carry a 0.50 × 0.50 mm bore; 14 are alignment-only Our records same topology record

Chip-side ports are simplified straight bores. A chip from this library does not arrive with a standard-compliant port.

4. Post-processing

stepwhat is recordedtiersource
Flush IPA, syringe port-to-port until effluent runs clear Our records A design report in the lab’s working copy, for reference chip chamberchip_f584, a CAD model: §7, its advice for printing. openmfda_flow/flow/reference_chips/chamberchip_f584/chamberchip_f584_report.md
Ordering — hard rule Flush the cavity clear BEFORE any UV post-cure Our records The same design report, §7: resin left inside cures permanently and cannot be removed
Alternative clearing method IPA ×3 under vacuum; IPA rinse Published Beauchamp 2018 [5], section 2.5 (three IPA flushes under vacuum); Gong 2017 [1], section 2.4 (an IPA rinse). Read 2026-10-01. The BYU printer's published process.

The ordering rule is the one post-processing step that has a consequence attached.

6. Known process behaviour

This is the most useful part of the page: it records what has actually happened, rather than what we expect to happen. Each item says how it turned out, including the ones that went wrong.

1. Resin clearing

observed failure Our account

Prints that had previously cleared began trapping resin unclearably in fluid channels and chambers. Pneumatic channels in the same parts remained clearable.

The diagnosis the lab gives attributes this to optical penetration at 405 nm with 1% NPS: light continues to reach resin in already-printed layers and gels it in place.

A geometric dependence was observed alongside it. Wide chambers gel before narrow channels, because every ceiling layer illuminates the whole chamber footprint and deposits dose into the liquid below it. The larger the roofed area, the more dose the trapped liquid receives.

Stated by the lab. This entire record — the failure, the diagnosis and the geometric dependence — appears in no tracked project file. It is the longest "Our account" item on this page, and the one whose loss would cost the most.

2. Membrane thickness

printed failure Our records

A CAD-intact 70 µm (7-layer) membrane leaked into the pneumatic chamber.

Attributed to cure-through and sag over a 440 µm void — not a CAD defect. The drawn wall was intact; the printed wall was not.

Downstream consequence: the generator now enforces a 100 µm floor on membrane thickness.

The leak: written down on 2026-09-11 in docs/process-record/record-01-membrane-leak.md, on this site. The enforced floor is in the lab’s own generator script, openmfda_flow/flow/platforms/h.r.3.3/pdk/py_scripts/gen_centerfed_block.py.

3. Interface mismatch

printed failure Our records

The first 3-pump centerfed block print did not mate with the reference manifold. The +Y shared fluid manifolds did not match the manifold's port layout.

Fix: the shared +Y manifolds were replaced with six independent +Z fluid cups, 2.5 mm diameter × 3.0 mm deep — one inlet and one outlet per pump, so each pump is fluidically independent.

The mismatch: written down on 2026-09-11 in docs/process-record/record-02-interface-mismatch.md, on this site. The cups as drawn, in the lab’s own interface file, openmfda_flow/flow/platforms/h.r.3.3/physical_interface.json — fluid_interface: six, 2.5 mm across and 3.0 mm deep.

4. Features reported as printing successfully

reported success Our account

0.44 × 0.44 mm pneumatic Y-channels on the +Y face of pump cells.

This is the one geometry the lab reports as having printed the way it was drawn. It is a short, wide, well-drained feature, and it is the only empirical anchor on this page for what the process can do rather than what it cannot.

The lab’s account.

5. Minimum-feature figures

Our account

Every minimum-feature figure in circulation on this project rests on literature or on a non-print experiment.

Where the minimum-feature figures rest: the lab’s account, for the Phrozen and its resin and for the BYU printer.

6. A printed single-pump cell, and its recipe

printed Our records

A single-pump cell, one_pump_v1_sw, has been printed. It is not the only print: five pumps were printed on the BYU printer in June 2026 (§8).

A slicer file for it exists — one_pump_v1_sw_mem120_fl150, whose name encodes a 120 µm membrane and a 150 µm fluid chamber. That 120 µm is the thickness recommended after the leak in record 1, which makes this the first sign that the recommendation was not only written down but built.

The file is encrypted CTB. Its settings block is under AES. What is plaintext is the machine name: “Phrozen Sonic Mighty 8K” — a 28 µm machine, neither the 22 µm the project assumed nor the 19 × 24 µm machine in the lab.

The slicer file, held by the lab.

7. Process limits

The limits below are the ones on record for these printers: how small a feature can be, how thin a wall can be, how wide a roof can span. They are the numbers people quote, and each one says what it actually rests on. If a design tool sent you here from its not checked list, this is all there is.

limitfigurewhat it rests onsource
Channel height must be less than its width h < w Our records — a fluid-model constraint, not a printability rule. See below. openmfda_flow/flow/platforms/h.r.3.3/pdk/docs/Rectangular Channel.docx — “Component Geometric parameters”
Lateral minimum feature, Phrozen Sonic Mighty 12K 10 px = 0.19 mm (X) / 0.24 mm (Y) Our records — has a written derivation, but from a routing-clearance experiment, not a print test openmfda_flow/flow/platforms/h.r.3.3/physical_interface.json — clearance._derivation
Membrane floor 100 µm Our records — enforced in the generator. openmfda_flow/flow/platforms/h.r.3.3/pdk/py_scripts/gen_centerfed_block.py (100 µm); gen_pump.py and gen_mixer.py (20 µm)
Edge clearance — hard floor 0.50 mm Our records — enforced in code; the build aborts below it openmfda_flow/flow/platforms/h.r.3.3/pdk/py_scripts/gen_centerfed_block.py — --edge-clearance-min
Edge clearance — preferred 0.75 mm Our records — enforced in code; warns below it gen_centerfed_block.py — --edge-clearance-pref

What h < w actually is

h < w is the one documented rule the design tools enforce, and it is the most misread number in the project. It comes from the h.r.3.3 PDK's rectangular-channel component document, where it sits under “Component Geometric parameters” alongside the hydraulic-resistance model.

It is the validity condition of the Bruus Hagen–Poiseuille series — the 0.63 · h/w term — and therefore a fluid-model constraint, not a printability rule. A channel taller than it is wide does not fail to print because of this rule. It falls outside the range over which the resistance approximation is written, so the number the calculator returns for it is computed on the same duct with the two cross-section dimensions exchanged.

8. What has been printed

Nothing generated by /design/channel/, /design/chamber/ or /design/droplet/ has ever been printed. Those tools produce geometry and arithmetic; they say so on their own pages.

All 21 reference chips in the library carry status not-tested: designs checked in CAD, 20 for the Phrozen Sonic Mighty 12K and the BYU pump+mixer for the H.R.3.3. Every chip and component in the library is designed for a specific printer: BYU's H.R.3.3 or the Phrozen Sonic Mighty 12K.

One single-pump cell has been printed. A slicer file for it exists, and its machine profile is readable: see §6.6 above.

Five pumps have been printed on the BYU printer. Our account In June 2026 a collaborator at BYU printed five pumps on the BYU printer, and the lab tested them. Some worked and some did not. By the lab's account that printer is the H.R.3.3 (§1). The prints BYU shows on its own OS1 pages are the Nordin group's, not this project's, and are not counted here.

Sources: each design tool's own “what is checked, and what is not” section; data/library.json status_enum and the 21 reference-chip entries. The five pumps printed at BYU: the lab’s account, given on 2026-09-29, and no file. That the printer was the H.R.3.3: the lab’s account, given on 2026-09-30, and no file.

References

Works cited by the Published rows above. They describe the Nordin group's printers and resins, the OS1 among them. None of the figures taken from them has been reproduced in this lab.

  1. Gong, Bickham, Woolley, Nordin. “Custom 3D printer and resin for 18 µm × 20 µm microfluidic flow channels.” Lab on a Chip 17(17), 2899–2909, 2017. doi:10.1039/c7lc00644f. Its penetration depths (sections 3.2 and 3.3.1) and IPA rinse (section 2.4) read 2026-10-01.
  2. Sanchez Noriega, Chartrand, Valdoz, Cribbs, Jacobs, Poulson, Viglione, Woolley, Van Ry, Christensen, Nordin. “Spatially and optically tailored 3D printing for highly miniaturized and integrated microfluidics.” Nature Communications 12, 5509, 2021. doi:10.1038/s41467-021-25788-w. Identified from its article number through Crossref, 2026-10-01.
  3. MacRae. “Regime Change in 3D Printed Microfluidics.” ASME, 8 November 2017. asme.org. A magazine article about Gong 2017 [1], not a paper. Read 2026-10-01.
  4. Gong. “3D Printing for Microfluidics.” PhD dissertation, Brigham Young University, 2018. BYU ScholarsArchive etd/7690, read 2026-10-01. (Gives the pixel pitch as 7.56 µm; see the pitch note in §1.)
  5. Beauchamp, Gong, Woolley, Nordin. “3D printed microfluidic features using dose control in X, Y, and Z dimensions.” Micromachines 9(7), 326, 2018. doi:10.3390/mi9070326. Section 2.5: the liquid resin left in each print was cleared with three IPA flushes under vacuum. Read 2026-10-01.
  6. Bickham, Pang, George, Topham, Nielsen, Nordin, Woolley. “3D printed microfluidic devices for solid-phase extraction and on-chip fluorescent labeling of preterm birth risk biomarkers.” Analytical Chemistry 92(18), 12322–12329, 2020. doi:10.1021/acs.analchem.0c01970. Its methods (“3D Printing”) give the 600 ms, 200 ms and 21 mW/cm² figures quoted in §5 as context only, for the group’s custom printer with a 385 nm light source and a resin of 2% NPS and 1% Irgacure 819 in PEGDA. Read 2026-10-01.
  7. BYU. The OS1 website, os1.byu.edu: its Specs, “Validation” and Gallery pages, read 2026-09-30. © BYU: nothing on this page is copied from it.
  8. Nordin group, BYU. The OS1 repository on GitHub, 3D-Printing-for-Microfluidics: its README, LICENSE (CERN-OHL-S-2.0) and file list at commit 4d22c93, read 2026-09-30.
  9. Goenner, Temple, Zapata, Wakeham, Snelgrove, Gaillardon, Nordin, Gale. “An open source platform to automate the design, verification, and manufacture of 3D printed microfluidic devices.” Scientific Reports 15, 33077, 2025. doi:10.1038/s41598-025-15976-9. Read 2026-09-30.
  10. Miner, Viglione, Hooper, Woolley, Nordin. “Fast multi-resolution 3D printing of microfluidics: enabling 2 µm channels and ultra-compact mixers.” Microsystems & Nanoengineering 12, 66, 2026. doi:10.1038/s41378-026-01194-4. Read 2026-09-30. The paper BYU's OS1 website links as its multiresolution paper; it does not mention the OS1.

← Back to the design tools · Home