74 Science & technology The Economist March 19th 2022
3DprintingA Gutenberg moment
E
arlyformsofadditivemanufacturing,
or 3d printing as it is popularly called,
began to emerge in the 1980s. But it took
more than a decade for the technology to
start taking off. Initially, it was used to
make prototypes. Now, intricate compo
nents are routinely 3dprinted in plastic
and metal, for use in products ranging
from jet engines and robots to cars.
Sales of 3dprinting services and ma
chines grew by more than 17% in 2021, to
reach around $15bn, according to prelimi
nary estimates for a report by Wohlers As
sociates, a firm that tracks the industry.
However, as useful as additive manufac
turing has become, it struggles to compete
on cost and speed with more established
ways of making things, such as injecting
molten plastic into moulds or stamping
out metal parts with a giant press.
As a result, most manufacturers use 3d
printers to produce lowvolume, highval
ue parts. The extra time and expense this
takes can be worth it for certain items.
Making things additively produces objects
layer by layer, so tricky internal structures
can be incorporated more easily into a de
sign. Shapes can also be optimised for
strength and lightness, saving materials.
But what if these advantages could be had
at the speed and cost of conventional fac
tory processes? A new form of additive
manufacturing aims to do just that.
The origin of this process, trademarked
“Area Printing”, goes back to 2009. That was
when James DeMuth, having finished his
master’s degree in mechanical engineering
at Stanford University, started work at the
National Ignition Facility, part of the
American Department of Energy’s Law
rence Livermore National Laboratory
(llnl). This uses some of the world’s most
powerful lasers to study nuclear fusion.
One of the challenges Mr DeMuth was
given was to find a way to use a highly spe
cialised type of steel to manufacture a 12
metre wide fusion chamber containing
many complex features. He considered a
form of 3dprinting, called Laser Powder
Bed Fusion (l-pbf), for the job. This em
ploys a laser beam to weld together parti
cles on a thin bed of powdered metal, to
form the required shape of the object’s first
layer. Then more powder is added and a
second layer is welded on top of the first.
And so on, until the item is complete.
The problem is that, as with most other
forms of 3dprinting, there is an inverse relationshipbetweenresolution, which gov
erns the level of detail that can be printed,
and the speed of the process. Hence, some
large components with fine details can
take days, if not months, to print. Produc
ing the chamber looked as if it might take
decades. l-pbfwas clearly unfeasible for
such an application.
This got Mr DeMuth and a group of col
leagues thinking about how to speed
things up without compromising quality.
After some work, they started using a de
vice called an optically addressed light
valve, which had been developed at llnl.
This permits a pulsed infrared laser, with
its beam shaped to have a square crosssec
tion, to be patterned with a highresolu
tion image. Working a bit like a photo
graphic negative, the image can block or
pass light, creating millions of tiny laser
spots, much like the pixels that make up a
digital image.
When projected onto a bed of powder,
this patterned laser light can weld a com
plete area in one go. Mr DeMuth likens the
process to producing documents with a
printing press instead of writing them out
individually with a pen. Not such a dotty idea
In 2015 Mr DeMuth cofounded Seurat
Technologies, to commercialise the tech
nology. This Massachusettsbased firm is
named after Georges Seurat, a postim
pressionist French artist who pioneered a
painting style called pointillism that
builds pictures up from dots. Several com
panies, including gmand Volkswagen, apair of carmakers, Siemens Energy, a divi
sion of a large German group, and Denso, a
big Japanese components firm, have part
nered with Seurat to explore the use of its
first prototype areaprinting machine.
This prototype produces a series of
small, patternable squares on the powder
bed. Their size depends on the material.
Aluminium requires 15mm squares. Tita
nium requires 13mm. Steel requires 10mm.
Individually, these squares might seem
small. But 40 of them can be printed adja
cent to each other every second, so a large
area can be covered quickly. The prototype
was designed to work at this scale to keep
the size of the laser and the amount of en
ergy it consumes to a practical level.
With the equivalent of 2.4m pixels pro
jected in each square, the machine can
print parts with layers just 25 microns
(millionths of a metre) thick at a rate of 3kg
an hour. This is ten times faster than a typ
ical l-pbfmachine at such a fine resolu
tion, says Mr DeMuth. Production versions
of the area printer are now being built, and
future generations of the machine should
end up being 100 times faster.
All that, says Mr DeMuth, means area
printing will be competitive with mass
production factory processes, such as ma
chining, stamping and casting. As an ex
ample, he believes that by 2030 it will be
possible to produce silverware (utensils
that nowadays are made from stainless
steel) for $25 a kilo. “That means we could
actually print silverware cheaper than you
could stamp them out,” he adds.
Other laserbased 3dprinters are get
ting faster, too. l-pbfmachines, for exam
ple, may be fitted with several beams—
though the complexity involved could lim
it their number. And many nonlaser ways
to print things are improving as well, using
all manner of materials to make items
ranging from buildings to bridges to bis
cuits. One way or another, then, 3dprint
ing seems at last to be ready to givetradi
tional factories a run for their money.nA new approach to 3dprinting may bring it into the mainstream