Researchers develop revolutionary 3D printing technology

March 17, 2015

A 3D printing technology developed by Silicon Valley startup, Carbon3D Inc., enables objects to rise from a liquid media continuously rather than being built layer by layer as they have been for the past 25 years, representing a fundamentally new approach to 3D printing. The technology, to appear as the cover article in the March 20 print issue of Science, allows ready-to-use products to be made 25 to 100 times faster than other methods and creates previously unachievable geometries that open opportunities for innovation not only in health care and medicine, but also in other major industries such as automotive and aviation.

Joseph M. DeSimone, professor of chemistry at UNC-Chapel Hill and of chemical engineering at N.C. State, is currently CEO of Carbon3D where he co-invented the method with colleagues Alex Ermoshkin, at Carbon 3D and Edward T. Samulski, also professor of chemistry at UNC. Currently on sabbatical from the University, DeSimone has focused on bringing the technology to market, while also creating new opportunities for graduate students to use the technique for research in science and at UNC and NCSU.

The technology, called CLIP - for Continuous Liquid Interface Production - manipulates light and oxygen to fuse objects in liquid media, creating the first 3D printing process that uses tunable photochemistry instead of the layer-by-layer approach that has defined the technology for decades. It works by projecting beams of light through an oxygen-permeable window into a liquid resin. Working in tandem, light and oxygen control the solidification of the resin, creating commercially viable objects that can have feature sizes below 20 microns, or less than one-quarter of the width of a piece of paper.

The video will load shortly

"By rethinking the whole approach to 3D printing, and the chemistry and physics behind the process, we have developed a new technology that can create parts radically faster than traditional technologies by essentially 'growing' them in a pool of liquid," said DeSimone, who revealed the technology at a TED talk on March 16 in the opening session of the conference in Vancouver, British Columbia.

Through a sponsored research agreement between UNC-Chapel Hill and Carbon 3D, the team is currently pursuing advances to the , as well as new materials that are compatible with it. CLIP enables a very wide range of material to be used to make 3D parts with novel properties, including elastomers, silicones, nylon-like materials, ceramics and biodegradable materials. The technique itself provides a blueprint for synthesizing novel materials that can further research in .

Rima Janusziewicz and Ashley R. Johnson, graduate students in DeSimone's academic lab, are co-authors on the paper and are working on novel applications in drug delivery and other areas.

"In addition to using new materials, CLIP can allow us to make stronger objects with unique geometries that other techniques cannot achieve, such as cardiac stents personally tailored to meet the needs of a specific patient," said DeSimone. "Since CLIP facilitates 3D polymeric object fabrication in a matter of minutes instead of hours or days, it would not be impossible within coming years to enable personalized coronary stents, dental implants or prosthetics to be 3D printed on-demand in a medical setting."

CLIP's debut coincides with the United Nation designating 2015 as the International Year of Light and Light-Based Technologies, which recognizes important anniversaries of scientific advances enabled with light.

Explore further: Novel 3-D printing process enables metal additive manufacturing for consumer market

More information: Continuous liquid interface production of 3D objects, Published Online March 16 2015. Science DOI: 10.1126/science.aaa2397

Related Stories

Lower-cost metal 3-D printing solution available

February 10, 2015

3D printing of plastic parts to prototype or manufacture goods is becoming commonplace in industry, but there is an urgent need for lower-cost 3D printing technology to produce metal parts. New substrate release solutions ...

Princeton team explores 3D-printed quantum dot LEDs

November 6, 2014

Some of the most important developments marking advances in the 3D printing industry are in the realm of material science, notes 3d Printing Industry. "The more materials we can print, the more useful the technology becomes." ...

Recommended for you

Samsung to disable Note 7 phones in recall effort

December 9, 2016

Samsung announced Friday it would disable its Galaxy Note 7 smartphones in the US market to force remaining owners to stop using the devices, which were recalled for safety reasons.

Swiss unveil stratospheric solar plane

December 7, 2016

Just months after two Swiss pilots completed a historic round-the-world trip in a Sun-powered plane, another Swiss adventurer on Wednesday unveiled a solar plane aimed at reaching the stratosphere.

Solar panels repay their energy 'debt': study

December 6, 2016

The climate-friendly electricity generated by solar panels in the past 40 years has all but cancelled out the polluting energy used to produce them, a study said Tuesday.

36 comments

Adjust slider to filter visible comments by rank

Display comments: newest first

jalmy
5 / 5 (7) Mar 17, 2015
That is one of the coolest things I have ever seen.
betterexists
4 / 5 (1) Mar 17, 2015
They should make it so big that a house is constructed /formed right away just like in that Video.
People can use them on vacations or even for permanent residence. They have to scale it up like Dinosaurs were scaled up from the tiny lizards!
I wish they make mammals that big so that we can have awful lot of milk to drink or for any other similar purpose! I am sure Latest Genetics can be used. Since Artificial Eggs rather than Wombs are more easily available, we should be able to scale up the chicken ASAP!
LariAnn
5 / 5 (3) Mar 17, 2015
To me, this is the true future of manufacturing. If stock in this company is available, it seems a sure bet. Next, think about when they will be able to do this using liquid metal (steel or alloys) and parts would rise up out of the liquid. It will happen, and perhaps not too long from now.
hillmeister
not rated yet Mar 17, 2015
Now we know how Superman's fortress of solitude was created!

The other video at their website certainly looks like it, heh. :p

http://carbon3d.com/
shavera
5 / 5 (1) Mar 17, 2015
@LariAnn: "Next, think about when they will be able to do this using liquid metal (steel or alloys)"

The problem is that this specifically uses the UV light to cause chemical reactions in the plastic that make it set into a solid.

But perhaps some other system could be devised with a vat of molten metal and some cooling... jet. Not sure it would be quite so precise... but it could be interesting to look into it.
antialias_physorg
not rated yet Mar 17, 2015
Is there audio with the movie? I don't quite get how this works.
Is there a UV screen below that basically projects slice-images of the object to be printed through this "oxygen window" (not sure what that is, exactly)?

But I agree: this is seriously cool (though probably doesn't lend itself to multi-material printing like the traditional 'top-down' approach).
Rustybolts
not rated yet Mar 17, 2015
Now that is awesome!
JRi
not rated yet Mar 17, 2015
Material properties are probably quite limited compared to conventional methods, but they do say they can print rubber, too. It's certainly good to see something new.
Dethe
1 / 5 (5) Mar 17, 2015
For laymen everything is awesome, but I'm not quite sure, how/why the speed of CLIP should be significantly higher than another continuous photo-polymerization methods. It can be only faster than discontinuous methods.
Tektrix
5 / 5 (3) Mar 17, 2015
"For laymen everything is awesome"

For sanctimonious prigs, every forum is an opportunity to be a jerk.
Dethe
1 / 5 (2) Mar 17, 2015
So don't be a jerk.
lantzjason
not rated yet Mar 17, 2015
oldworldlabs
DonGateley
5 / 5 (1) Mar 17, 2015
@antialias_physorg: 'Is there a UV screen below that basically projects slice-images of the object to be printed through this "oxygen window"'

That was my understanding except that the probable frame rate of the projection divided by the feed rate of the puller would give frames/mm that would be continuous for all practical purposes rather than sliced and stepped as we are used to seeing.
Eikka
5 / 5 (1) Mar 18, 2015
The article is misleading on all counts.

Continuous liquid interface production is achieved with an oxygen-permeable window below the ultraviolet image projection plane, which creates a "dead zone" (persistent liquid interface) where photopolymerization is inhibited between the window and the polymerizing part.


The object is still being formed by slicing a 3D model into layers and then projecting the layers into a tank of UV-hardening resin. This technology has existed for years.

The difference is that in previous DLP 3D printers the resin cures right down to the bottom of the tank and the object sticks to the glass. The glass bed is then tilted slightly to peel it off and allow the object to become unstuck with new resin underneat the previous surface. This harms accuracy and speed.

The new invention employs a chemical passivation of the resin at the surface of the glass window to allow the object to be continously lifted without sticking to the glass.
Eikka
not rated yet Mar 18, 2015
Another similiar means of "continous" 3D printing is turning the system around and filling a tank with the resin liquid, projecting the image on the surface of the rising liquid.

probable frame rate of the projection divided by the feed rate of the puller would give frames/mm that would be continuous for all practical purposes rather than sliced and stepped as we are used to seeing.


The resolution of the projection optics still makes the object pixelated in the X-Y dimension, so you won't actually get rid of the typical stepped banding effect you see on these printers. It's always been possible to use arbitrary layer thickness limited only by how much you need to lift the object at minimum to get new resin to flow underneath it between the glass.

In effect, the surface tension and viscosity of the resin determines your minimum layer thickness, unless you want to start lifting and lowering the object down in a discontinuous manner like before.
antialias_physorg
5 / 5 (1) Mar 18, 2015
but I'm not quite sure, how/why the speed of CLIP should be significantly higher

Because you can do the entire layer in one go instead of point by point, (and don't have the time delay of covering the top again with material like in stereo-litography)

The resolution of the projection optics still makes the object pixelated in the X-Y dimension

At 20 microns that isn't noticeable. The spatial resolution at the fingertip is roughly 0.9mm (blind people roughly 0.8mm). Vision-wise it is very close to the resolution limit which is somewhere areound 10-12 micrometers. So I'd think you could get away without any kind of post processing. 20 microns is certainly better than the roughness of many conventionally manufactured goods.
Eikka
5 / 5 (1) Mar 18, 2015
At 20 microns that isn't noticeable.


20 microns is actually quite bad for a surface finish. It's worse than what can be attained by sandcasting parts.

http://en.wikiped...ring.png

The spatial resolution at the fingertip is roughly 0.9mm (blind people roughly 0.8mm).


Actually, we feel features down to the nanometers due to the friction and vibration they cause. We just can't distinguish them individually.

Vision-wise it is very close to the resolution limit which is somewhere areound 10-12 micrometers.


It's very visible due to reflections and scattering from the surface. It's nowhere close to a smooth finish.
Eikka
5 / 5 (1) Mar 18, 2015
Here you can see an FDM print with 20 µm layer height:

https://www.flick...0925213/

It's possible to go down to sub-micron resolutions with DLP printers, but the working volume of the printer becomes tiny and you can't print objects much larger than a coin:

https://www.youtu...1K8RNhIk
Eikka
not rated yet Mar 18, 2015
The physical limitation is that DLP projectors have arrays of movable mirrors that form the picture. The mirrors are 10-20 µm in size, and the physical size of the array is about 3x2 cm for a typical projector chip.

Since you're projecting the image of this mosaic array onto the resin, to get better resolution you must use reducing optics, but then the object you're printing becomes smaller than the array of mirrors. Similiarily, to make objects larger than about 1" square you sacrifice resolution by making the image larger. That's why many print at 50-100 µm or more.

Producing larger arrays to maintain resolution is exponentially expensive because there's fewer chips you can fit per standard 300 mm silicon disc, and the yield goes down because larger chips catch more defects, resulting in more duds. Producing smaller mirrors has a limit at about 1 µm because of the wavelenghts of light they're trying to project. At smaller and smaller sizes they're no longer mirrors.

antialias_physorg
not rated yet Mar 18, 2015
A finished product will have a coat of paint or even some inactivation layer added. Even if the 20 microns were visible or you could sense them any kind of finishing would easily smooth that out.

I do agree with LariAnn: This is the future of manufacturing (and maybe even the end of the manufacturing _industry_ as such)
SamB
5 / 5 (2) Mar 18, 2015
For a few runs this may do well but for true large scale production you just can't beat injection molding for efficiency. They can pop out thousands of copies in an hour.
Eikka
not rated yet Mar 18, 2015
A finished product will have a coat of paint or even some inactivation layer added.


Most plastic products are made by pressure molding and they come out readily smooth and shiny without extra processing. Coating processes such as electrostatic plating to make a mirror or metallic look, or simply painting the surface also depend on the surface being essentially flat or they get a mottled appearance without gloss.

There's an emerging process of spraying fine solvent mist or gaseous solvents onto 3D printed parts to smooth out the surfaces, but they also smooth out over any fine details and provide for less control over the dimensions, and may weaken or cause parts to sag.

Not all products of course need to be smooth and shiny, but the particular problem with 3D printing is the clearly stratified banding effect. It wouldn't matter as much if the pattern was random, but adding dithering at the edges of the layers would make the surface quality even worse.

jalmy
not rated yet Mar 18, 2015
A finished product will have a coat of paint or even some inactivation layer added. Even if the 20 microns were visible or you could sense them any kind of finishing would easily smooth that out.

I do agree with LariAnn: This is the future of manufacturing (and maybe even the end of the manufacturing _industry_ as such)


Agree, although very complicated or large parts may require the precision/size of large machines that cost too much for the average person today. But the current molding industry would adapt to using these newer machines to produce quality automotive etc. parts in bulk for a potentially large savings in energy and polymer. Which gets passed down to consumers.
DonGateley
not rated yet Mar 18, 2015
Since the observed patterning with prior printing methods is predominantly orthogonal to the direction of deposition I thought it was primarily pass registration error due to the mechanics of the printer (hysteresis and such.) Is that not so? If it is, this method eliminates such registration error.
Eikka
not rated yet Mar 18, 2015
Since the observed patterning with prior printing methods is predominantly orthogonal to the direction of deposition I thought it was primarily pass registration error due to the mechanics of the printer (hysteresis and such.) Is that not so? If it is, this method eliminates such registration error.


The patterning is because you can't print a slope with arbitrary resolution in the X-Y direction. You're limited to the pixel size and magnification of your DLP projecting optics, which means that while your Z resolution may be 20 microns or however finely you lift the object, your X-Y resolution will be fixed. This may be 50 - 100 microns for any practically sized DLP printer.

If the object consists of nothing but orthogonal vertical and horizontal faces aligned to the major axis, then of course you have no banding. As soon as you have a diagonal line anywhere, you get jaggies.
SteveGinGTO
not rated yet Mar 20, 2015
WOW, now THIS is freaking GREAT!

This is almost an analog of laser printing vs dot-matrix printing, especially in terms of speed. But wait until they learn how to speed it up.

Star Trek, we are getting there, one step at a time.
gkam
1 / 5 (2) Mar 20, 2015
"For a few runs this may do well but for true large scale production you just can't beat injection molding for efficiency. They can pop out thousands of copies in an hour."
---------------------------------------

Oh, yeah. Got that 350-ton molding press available? Know what it takes to run it? Cooling? Power? Feedstock and product handling?

BTW, the complex geometry of the sample would make it very difficult to injection mold, since it would need several core-pulls and retractable elements.
Eikka
not rated yet Mar 21, 2015
Oh, yeah. Got that 350-ton molding press available? Know what it takes to run it? Cooling? Power? Feedstock and product handling?


Not all injection molding requires great pressure or complex molds made out of exotic materials. You can even make a disposable injection mold out of plastic and it will last you just long enough to do a prototype run.

it depends entirely on the product. In general, injection molding becomes economical at a couple hundred copies.

3D printing meanwhile becomes uneconomical beyond a few dozen copies because the time per copy scales linearily even when printing multiple objects on the same printer, because the laser/extruder has to scan more surface. DLP printers don't have this effect, but their print volumes are limited because of the resolution problem: you have to choose between printing one object per machine at fine resolution, or multiple at poor resolution.

gkam
1 / 5 (2) Mar 21, 2015
No, Eikka, you are not going to mold that with a plastic mold. Production machines usually operate at about 2,000 psi injection pressure, or did so when I ran two 500-ton Cincinatti's.

Yes, Eikka, I was VP and plant manager of Plasticon, from 1974 to 1975. Don't tell otto.
Eikka
5 / 5 (2) Mar 21, 2015
No, Eikka, you are not going to mold that with a plastic mold.


Yes. Yes I am.

The point is using high-temperature high density thermoplastic or epoxy resin to create disposable moulds for low temperature low density thermoplastics as a cheaper alternative to machining prototypes out of steel. They last between 10-50 uses before the material creeps out of tolerances.

That's good enough for the prototype or fitting run of a new product without the expense of machining a set of very expensive dies out of metal.

Plastics have advanced since 1975. Even amateurs can do it today:

http://www.instru...ALLSTEPS

Yes, Eikka, I was VP and plant manager of Plasticon, from 1974 to 1975.


And what's that supposed to prove again?
gkam
1 / 5 (2) Mar 21, 2015
You are talking about engineering, not production.

That other comment was for otto.

And other wiki-warriors.
TheGhostofOtto1923
1 / 5 (2) Mar 22, 2015
Robotically-controlled lasers can be used to refine the finish on a printed object.

"To improve the overall surface quality of Selective Laser Sintering parts, a robotic finishing system has been developed as a part of an ongoing research project. A finishing tool is held by a robot and moved according to programmed paths generated from the original CAD model data. This paper describes the experimental system in detail and shows that the surface roughness, dimensional accuracy, and geometrical accuracy can be improved."

-They can be used to either smooth or roughen, or produce custom patterns at most any scale.

-Actually that article is for mechanical finishing of 3D parts. Here are laser finishing examples:
https://scholar.g...kQgQMwAA
TheGhostofOtto1923
1 / 5 (2) Mar 22, 2015
Yes, Eikka, I was VP and plant manager of Plasticon, from 1974 to 1975. Don't tell otto
Yes because I would have to conclude that that was one more failed job that you bullshitted yourself into, and out of. Because it's obvious you know little about the field...
plastics have advanced since 1975
-See? My god you're full of shit. It took your employers there less than a year to figure that out didnt it?
katesisco
not rated yet Mar 22, 2015
Dental implants are only and ever titanium.
Is this a refinement on injection molding?
gkam
1 / 5 (2) Mar 22, 2015
Kate, it is a different process, hardening liquids to make shapes.
Dug
5 / 5 (1) Mar 23, 2015
Sounds like fun even with limited applications and as long as energy, petroluem and petrochemical industry economics are intact and favorable.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.