Direct 3D printing of sterile parts




While most laboratories in the developed world are well-stocked with insturments and basic equipment, this is not a global universal. There are many settings in which laboratory supplies and equippment are needed, but either unavailable, unaffordable or both.


  • Test for growth on richer media than LB :

    • Blood, serum, chocolate agar

    • SOB

    • AYE (ACES Yeast Extract broth)

    • grow with CO2, in “candle can” + anaerobically

    • stimulate w/ known germinants to induce spore germination

    • let sit in broth on bench for a few weeks, then plate for CFU

  • Plate onto

    • LB

    • blood agar

    • fungal media (Sabouraud?)

    • chocolate, maconkey, TSA, etc agar (See above)

    • test if leeching from or PLA itself affects tissue cell culture growth, both cell lines and bone marrow derived macrophages

    • see if potential microbes can be resusitated in amoebic or other hosts

  • See if this technique can possibly be compliant with CLIA (Emily’s suggestion)



Experiment Started Tweeted Material Part Media Time Temperature Gas Fab. facility Cult. facility Replicates Result
Preliminary 1/23/2014 (Neches 2014, Neches 2014a, Neches 2014b) Orange PLA blob LB 96 37 aerobic UC Davis UC Davis 1 -
First trial 1/25/2014 Orange PLA tube LB 96 37 aerobic UC Davis UC Davis 6 -

Preliminary experiment


A sterile glass beaker contianing roughly 20ml of LB media was placed under the nozzle of a fused deposition modeling (FDM) 3D printer. The nozzle was heated to \(220\,^{\circ}\mathrm{C}\), and the extruder drive motor was driven forward until into the nozzle until about 20mm of polylactide (PLA) filament had been fused and expelled through the nozzle and into the beaker. After 20mm, a tangle of molten and cooled PLA detached from the nozzle and fell into the beaker. The mouth of the beaker was then covered with sterile aluminum foil. An unopened sterile beaker of LB was prepared as the negative control. A positive control was prepared by removing a length of several centimeters of un-melted PLA filament from the spool and placing into a beaker of sterile LB media. The three beakers were placed into a shaking incubator at \(37\,^{\circ}\mathrm{C}\) for 96 hours. No growth was observed in the negative control, robust growth was observed in the positive control, and the beaker containing extruded PLA displayed no evidence of growth.

Growth after 96 hours at \(37\,^{\circ}\mathrm{C}\) in a shaking incubator. The leftmost beaker contains LB media inoculated with PLA plastic extruded from the printer nozzle at \(220\,^{\circ}\mathrm{C}\). The center beaker contains LB media inoculated with a segment of unextruded PLA plastic filament from the same spool. The leftmost beaker contains uninoculated LB.

Growth experiment on printed component


The preliminary experiment seemed to indicate a potentially useful killing effect from the nozzle’s heat and pressure, and so a slightmy more realistic assay was conducted. A simple model was created using the OpenSCAD modeling language consisting of a cylinder of radius 4mm and height 10mm.

cylinder( r=4, h=10 );

The model was exported in Standard Tessellation Language (STL) format. (Burns 1993)

A very simple model part was created in OpenSCAD and exported in STL format.

The manifold was then converted into G-code commands (Kramer 2000) using Cura (version 13.12-test on Linux), using a wall width of 0.4mm (equal to the nozzle diameter), cooling fans inactive, no infill, a top and bottom layer height of zero, and a spiralized outer wall (“Joris mode,” after Joris van Tubergen) to produce a small, open tube.

The G-code toolpath visualization of test part in Cura. The slicing engine was set to a 0.4mm wall width (equal to the diameter of the nozzle), cooling fans inactive, no infill a top and bottom layer height of zero, and a spiralized “Joris Mode” outer wall.

The G-code was stored on a SD card and printed on an Ultimaker kit-based FDM 3D printer. A small patch of aluminum foil was lightly abraded with fine-grit sandpaper to improve surface ashesion properties, and flamed over a Bunsen burner until signs of melting appeard. The foil patch was then affixed to the build platform, so that the build area indicated in the G-code toolpath would be entirely within the untouched center of the patch. The G-code toolpath was also examined to insure that the nozzle would contact no surface except the build are on the foil. Printing was then initiated with a feed rate of 50 mm/sec at \(220\,^{\circ}\mathrm{C}\).

3D printing a test object onto sterilized aluminum foil.

Once printing was complete, finished parts were immediately removed from the build area using flamed forceps and transfered to culture tubes prepared with LB media. The tubes were then incubated for 96 hours in a shaking incubator at \(37\,^{\circ}\mathrm{C}\).

After 96 hours incubating in a shaking incubator at \(37\,^{\circ}\mathrm{C}\), none of the six replicates showed evidence of growth.

Independent reproduction of growth experiment on printed component

\label{sec:indpendent_rep} The above experiment using printed components was replicated at Michigan State University. We used a similar kit-built Ultimaker 3D printer modified with an E3D all metal print head with a 0.4mm nozzle (citation not found: A cylindar was designed using OpenSCAD with a radius of 4mm and a height of 12mm. The model was exported to an STL object and sliced with Cura SteamEngine 13.12. The cylinder was printed with a wall thickness of 0.4mm, a feed-rate of 10mm/second (the effective speed with the minumum layer cooling time set to 5 seconds), and a nozzle temperature of \(225^{\circ}\mathrm{C}\). The printbed was prepared with 3M Scotch Blue painters tape and was lightly wiped with ethanol before printing began.

Two printed cylinders were transfered to sterile glass tubes filled with 4mL of Luria-Burtani (LB) broth with flamed tweezers. A fragment of unused fillament was used as a positive control, and an uninoculated tube was used as a negative control. Tubes were transfered to a shaking incubator set at \(30^{\circ}\mathrm{C}\). No growth was observed after 24 hours in any of the tubes with printed parts, while the unused fillament contaminated the media. After two days, another cylinder was printed and incubated in LB broth. Again, after 24 hours no growth was observed. None of the tubes with printed parts showed signs of growth after 96 hours.

After 48 hours, only the positive control (left) was contaminated. Printed cylindars in LB did not appear to contaimate media.

Continued Growth Experiments at UMich

AYE broth (ACES Yeast Extract)

Quick pilot experiment with Luis’ positive control plastic (tube 1) and “Foil + EtOH + fire #2” plastic sample (tube 3) added to 5mL ACES-buffered yeast extract broth supplemented with FeNO3, L-cys and thymidine (standard culture medium for Legionella pneumophila studies. ph = 6.9 (blank tube 2)

After 24h on roller at 37C, no growth was observed. Examination under light microscope reveals only bacteria in positive control. Returned to wheel for the weekend.

No growth was observed in any tube besides positive control at 48 and 72h post innoculation. At 96h post innoculation, bacterial growth was seen in the “sterile” PLA cylinder tube. Under the light microscope, cells appeared to differ between the positive control and the PLA tube. 10ul was plated onto solid AYE media (CYE- contains charcoal) for each positive, PLA and negative control.

10ul of each AYE tube (positive control, PLA plastic and negative control) was struck out on Charcoal Yeast Extract solid media and put at 37C to grow for 24h. Growth revealed that the PLA tube (top, white) appeared to contain a different bacterial species than the positive control tube (bottom, yellow). The negative control was plated on the right (no growth). Under light microscope, both bacterial growths appear coccoid, with the yellow colonies forming clumps more often. The tube experiment was repeated and set up including parts from Luis and Russell, plus controls. So far, no growth observed for any 3D parts at 24h.

Terrific Broth Experiments

3D printed parts from Russell were suspended in sterile Terrific Broth supplemented with potassium salts (methods). After 24h at 37C, no growth was observed for either +/- UV parts. On the right, tubes from the above experiment at 48h.

As of 2/4/14 (96h post innoculation), no biotic growth has been observed outside of the positive control in Terrific Broth for Russell’s parts. There does appear to be a bit of filament or abiotic material floating around the broth.

After 2 weeks in anaerobic chamber at 37C in “meat broth”, only Russell’s “no UV” part had growth on it. All others were clean (besides positive control). Plated on BHI + blood overnight and will analyze via 16S rRNA sequencing (Nicholas Pudlo).


Bacterial strains, culture conditions and reagents

  • For AYE growth experiments, 3D parts were cultured on a rolling spinner at 37°C in N- (2-acetamido)-2- aminoethanesulfonic acid (ACES; Sigma)-buffered yeast extract (AYE) broth supplemented with 100μg/ml thymidine (Sigma). To quantify colony forming units (CFU), aliquots were plated on ACES-buffered charcoal-yeast extract agar (CYE) supplemented with 100μg/ml thymidine (T) and incubated at 37°C for 4 days.

  • Terrific Broth (TB) experiments were conducted on a rolling spinner at 37°C in media containing yeast extract, tryptone and glycerol supplemented with 0.17 M KH_2PO_4, 0.72 M K_2HPO_4.

  • Meat broth experiments were performed in media containing yeast extract, beef extract, glucose, maltose, fructose, xylose in an anaerobic chamber at 37°C.

Printed parts from UC Davis

The following materials were prepared in the Eisen Lab at UC Davis and shipped to the Swanson Lab at the University of Michigan. All printed parts were printed using Printbl Orange 3mm PLA filament at \(220\,^{\circ}\mathrm{C}\) with a feed rate of 50 mm/sec, using the same G-code files described in section \ref{sec:experiment2}.

  • Test objects, printed under biosafety hood (10x)

  • Test objects, printed under biosafety hood with UV (10x)

  • Test objects, printed under biosafety hood with UV, then dropped onto no-sterile surface during handling (2x)

  • Test object, printed under biosafety hood with UV (1x)

  • Test object, printed under biosafety hood without UV, dropped during handling (1x)

  • Empty, unopened conical tube

  • Test object, printed under biosafety hood without UV (1x)

  • Test object, printed under biosafety hood without UV (1x)

  • Test object, printed under biosafety hood with UV (1x)

  • Test object, printed under biosafety hood with UV (1x)

  • Test object, printed under biosafety hood with UV, handled with ungloved hands (1x)

  • Test objects, printed on open bench and left on lab bench overnight (2x)

  • Unused Printbl Orange 3mm PLA filament (3x)

  • Unused Laywoo-D3 cherrywood 3mm printable wood filament (3x)

  • Unused Protoparadigm White 3mm PLA filament (3x)

  • Unused Printbl Crystal Blue 3mm PLA filament (3x)

In addition, several printed parts were prepared at Michigan State University and sent to the Swenson lab at the University of Michigan. Cylindars were printed using the same G-code and parameters described in \ref{sec:indpendent_rep}. All printed parts from Michigan State University were printed using Ultimaker translucent blue PLA (citation not found:

Three cylindars were printed on blue painters tape that was wiped down with ethanol, and three cylindars were printed on abraded foil that was wiped with ethanol and flamed. Each part was removed from the printbed using flamed tweezers and transfered to a sterile 15mL plastic tube. In addition to the printed parts, three pieces of unused fillament were packaged independently in sterile tubes.


  1. Russell Neches. Conjecture : 3D prints are sterile. Let’s find out. (2014). Link

  2. Russell Neches. No growth in culture inoculated with 3D printer extruded plastic, pos. & neg. controls worked. So far, so good! (2014). Link

  3. Russell Neches. Still no growth in media inoculated with plastic extruded from the 3D printer. w00t! (2014). Link

  4. Marshall Burns. Automated fabrication : improving productivity in manufacturing. PTR Prentice Hall, 1993.

  5. Thomas R. Kramer, Frederick M. Proctor, Elena R. Messina. The NIST RS274NGC Interpreter - Version 3. (2000).

[Someone else is editing this]

You are editing this file