Does Material Choice Drive Sustainability of 3D Printing?

Abstract

Environmental impacts of six 3D printers using various materials were compared to determine if material choice drove sustainability, or if other factors such as machine type, machine size, or machine utilization dominate. Cradle-to-grave life-cycle assessments were performed, comparing a commercial-scale FDM machine printing in ABS plastic, a desktop FDM machine printing in ABS, a desktop FDM machine printing in PET and PLA plastics, a polyjet machine printing in its proprietary polymer, an SLA machine printing in its polymer, and an inkjet machine hacked to print in salt and dextrose. All scenarios were scored using ReCiPe Endpoint H methodology to combine multiple impact categories, comparing environmental impacts per part made for several scenarios per machine. Results showed that most printers’ ecological impacts were dominated by electricity use, not materials, and the changes in electricity use due to different plastics was not significant compared to variation from one machine to another. Variation in machine idle time determined impacts per part most strongly. However, material impacts were quite important for the inkjet printer hacked to print in salt: In its optimal scenario, it had up to 1/38th the impacts coreper part as the worst-performing machine in the same scenario. If salt parts were infused with epoxy to make them more physically robust, then much of this advantage disappeared, and material impacts actually dominated or equaled electricity use. Future studies should also measure DMLS and SLS processes / materials.

Full text

 Abstract—Environmental impacts of six 3D printers using
various materials were compared to determine if material choice
drove sustainability, or if other factors such as machine type, machine
size, or machine utilization dominate. Cradle-to-grave life-cycle
assessments were performed, comparing a commercial-scale FDM
machine printing in ABS plastic, a desktop FDM machine printing in
ABS, a desktop FDM machine printing in PET and PLA plastics, a
polyjet machine printing in its proprietary polymer, an SLA machine
printing in its polymer, and an inkjet machine hacked to print in salt
and dextrose. All scenarios were scored using ReCiPe Endpoint H
methodology to combine multiple impact categories, comparing
environmental impacts per part made for several scenarios per
machine. Results showed that most printers’ ecological impacts were
dominated by electricity use, not materials, and the changes in
electricity use due to different plastics was not significant compared
to variation from one machine to another. Variation in machine idle
time determined impacts per part most strongly. However, material
impacts were quite important for the inkjet printer hacked to print in
salt: In its optimal scenario, it had up to 1/38th the impacts coreper
part as the worst-performing machine in the same scenario. If salt
parts were infused with epoxy to make them more physically robust,
then much of this advantage disappeared, and material impacts
actually dominated or equaled electricity use. Future studies should
also measure DMLS and SLS processes / materials.
Keywords—3D printing, Additive Manufacturing, Sustainability,
Life-cycle assessment, Design for Environment.
I. INTRODUCTION
D printing is revolutionizing some fields of manufacturing,
especially prototyping [1]. It is sometimes assumed to be a
more sustainable way to manufacture, but such blanket
statements are unrealistic for any manufacturing technology,
as production methods for different kinds of finished products
vary so widely. For some kinds of products it can be a great
improvement, and indeed it enables the production of some
products that could not be economically produced any other
way. GE is printing jet engine nozzles predicted to save
millions of gallons of fuel per year due to geometries enabled
by 3D printing, which were not economically viable through
previous manufacturing methods [2]. Many people assume 3D
printing virtually eliminates waste, but this is only true for
some circumstances, such as FDM machines not using support
material; other 3D printers can produce as much as 43%
material waste, even before support material is counted (see
Results section). Many people also assume that 3D printing is
Jeremy Faludi, Zhongyin Hu, Shahd Alrashed, Christopher Braunholz,
Suneesh Kaul, and Leulekal Kassaye are with the Department of Mechanical
Engineering at University of California Berkeley, Berkeley USA 94720
(phone: +1 206 659 9537; e-mail: faludi@berkeley.edu).
more sustainable because it can eliminate transportation of
consumer goods [3]. Unfortunately, this is misguided because
transportation only represents a small fraction of lifetime
ecological impacts for most products [4], even ignoring the
fact that 3D printers still require raw materials to be
transported. On the other hand, Markus Kayser's "solar sinter"
demonstrated 3D printing of glass from desert sand, an
abundant, non-toxic, local material fused together directly by
sunlight in a printer run entirely from solar power [5]. One
could hardly ask for a more sustainable manufacturing method
(assuming the resulting printed objects are robust). As a result
of all these issues, there is not one simple answer. Recent
studies [6], [7] have shown that even for the relatively limited
scope of prototyping plastic parts, 3D printing can be either
better or worse than status-quo methods such as machining,
depending on multiple factors.
To drive the 3D printing industry toward a future where it
does become an inherently more sustainable manufacturing
method than other options, we should study where the biggest
impacts of 3D printing lie and how to minimize them.
Moreover, we should communicate these results in a way that
is easy for industry to understand and make decisions based on
it. This study examined whether material choice was the most
important factor determining the sustainability of 3D printing,
or if other factors such as machine size or utilization
frequency were dominant. Some types of 3D printing allow
for very “green” material choices—ones which are renewable
or abundant, non-toxic, recyclable or compostable, and which
have little embodied energy or resources. A modest example is
PLA bioplastic (an improvement compared to ABS); more
daring examples include salt, sugar [8], starch [9], or sawdust
[10]. Some of these materials also enable low-energy printing
processes, because they rely on chemical adhesion as opposed
to melting plastic or curing photopolymers with UV light. This
study also measured such factors, as they are usually
inextricable from material choice. An SLA machine can only
print in photopolymers, an inkjet machine cannot melt
plastics, and so on. So for a complete picture, whole-system
printer performance must be considered, as well as the
different materials.
II. BACKGROUND
Some specific environmental impacts of 3D printing have
been studied in depth—usually energy use [11], [12], [13], but
occasionally also toxicity [14]. Even when researchers do
specifically study health impacts from 3D printing, such as
evaporated plastic particles in the air [15], they rarely compare
these to energy use or other impacts to find top priorities for
Jeremy Faludi, Zhongyin Hu, Shahd Alrashed, Christopher Braunholz, Suneesh Kaul, Leulekal Kassaye
Does Material Choice Drive Sustainability of
3D Printing?
3
World Academy of Science, Engineering and Technology
International Journal of Mechanical, Aerospace, Industrial and Mechatronics Engineering Vol:9 No:2, 2015
144International Scholarly and Scientific Research & Innovation 9(2) 2015
International Science Index Vol:9, No:2, 2015 waset.org/Publication/10000327

sustainability. Only one study was found to have measured
multiple kinds of ecological impacts together to balance the
effects of material use, waste, toxins, and other factors against
energy use in a life-cycle assessment (“LCA”) with combined
single-score measurements, comparing several 3D printer
types [16]. That study was from 1999, so even without the
current project's new focus on materials, the older study
should be updated for changes in 3D printer technology,
available 3D printing materials, and LCA tools. Several of the
machine types and materials measured here were not in use
then.
III. METHODS
A. LCA Scope and Functional Units
This project extends the work of recent studies [6], [7] by
measuring more machines and testing variations in material
choice. For this study, the printers measured were a large
commercial-scale Dimension 1200BST fused-deposition
modeling (“FDM”) machine, a small desktop-scale Afinia
H480 FDM machine, a small desktop Type A Machines Series
1 FDM machine, an Objet Connex 350 polyjet machine, a 3D
Systems Projet 6000 stereolithography (“SLA”) machine, and
a Zcorp 310 inkjet machine.
LCAs were conducted in SimaPro software, with data
primarily from the EcoInvent database, but some data coming
from US Franklin LCI and other standard databases. ReCiPe
Endpoint H methodology [17] was used to combine 17
different categories of ecological impact (including climate
change, toxicity, resource depletion, and other factors) into
unified single scores. LCA scope was cradle-to-grave,
including electricity used to print parts, material comprising
the parts printed, and waste material generated during printing,
as well as electricity use while machines idle or start up,
embodied impacts of raw materials and manufacture of the
machines themselves, transportation of the machines to and
from UC Berkeley, and disposal of the machines at their end
of life, conservatively assumed to be five years, since no 3D
printer manufacturer was willing to provide lifetime estimates,
and estimates from an informal survey of prototypers
produced few and highly varying answers.
Masses and manufacturing processes of printer components
were not provided by the manufacturers, so they had to be
estimated by measuring the dimensions of every one of the
dozens of components that could be accessed, and calculating
their masses by standard densities of steel, aluminum, glass,
polyurethane, ABS, copper wire and motor windings, etc.
Electronics were estimated by area of circuit board, length of
cable, or by approximate equivalence to existing items in the
databases (for example, 1 desktop computer for the SLA
machine’s control and interface electronics, since the actual
electronics were inaccessible).These component estimates are
uncertain, but environmental impacts of the entire machines’
materials and manufacturing was usually less than 10% of
lifetime impacts, so further precision was not deemed
necessary. Electricity use was measured with a WattsUp Pro
ES power datalogger, except where raw data was already
available from previous studies. Ecological impacts from
electricity were modeled as average US electricity grid mix.
Disposal was modeled with a standard combination of landfill
and recycling, the EcoInvent process “Durable goods waste
scenario/US S.”
These different printers work in very different ways, with
different kinds of environmental impacts, so to create a fair
“apples-to-apples” comparison, ecological impacts of different
materials and printers were compared per object printed. The
functional unit was the printing of a single thin-walled part,
designed to be representative of a typical prototyping job—see
Fig. 1. Industry representatives told us that roughly “90%” of
their customers’ prototyping jobs were thin-walled plastic
enclosures for consumer products.
Fig. 1 Two units of the printed part, showing inside and outside
B. Materials
The Dimension (large FDM) and Afinia (one of the desktop
FDMs) printed in ABS plastic. The Type A (the other desktop
FDM) printed in PET plastic and PLA bioplastic. These are all
fairly standard plastics today. LCAs and toxicological studies
alike have found that PLA has the lowest health and
environmental impacts of the three, followed by PET and then
ABS [18], [19]. PLA is notable because it is a bioplastic, made
from agricultural sources such as corn rather than fossil fuels,
and it has a significantly lower melting point, allowing
printers to extrude it with less energy use. In addition, neither
PLA nor PET requires a 3D printer to have a heated bed to
avoid curling as ABS does [20], which should save significant
energy.
The Zcorpprinter generally uses a proprietary plaster
powder bonded with proprietary inkjet ink. However,
measurements here were performed with a Zcorp printer
hacked to print in many alternative materials, including salt,
sawdust, and concrete. Such hacking is done by a small but
growing community of people pursuing both eco-friendly
materials and cheaper materials than the proprietary ones sold
by printer manufacturers. The Zcorp machine measured was
hacked by UC Berkeley architecture professor Ron Rael and
his students, working with their own proprietary formulations,
so a public-domain recipe was taken from an internet forum
where people trade recipes for do-it-yourself 3D printer
materials [21], and Raelstatedit was similar enough for
accurate modeling. This “salt” printing recipe was a powder
mixture of 88% fine-ground salt and 12% maltodextrin,
bonded with a liquid mixture of 280 mL isopropyl alcohol,
920 mL distilled water, and 45 mL food coloring per inkjet
bottle. (One bottle lasts for many print jobs, so the actual
amount of liquid per print is a fraction of this.)Since this
World Academy of Science, Engineering and Technology
International Journal of Mechanical, Aerospace, Industrial and Mechatronics Engineering Vol:9 No:2, 2015
145International Scholarly and Scientific Research & Innovation 9(2) 2015
International Science Index Vol:9, No:2, 2015 waset.org/Publication/10000327

material by itself is fragile, parts are very often strengthened
after printing by soaking epoxy, cyanoacrylate, or other
bonding agents into the salt printout. Since the ecological
impacts of epoxy are roughly 47 times larger than the salt /
dextrose / isopropyl material (as measured in ReCiPe
Endpoint H points per unit mass), LCA scores with and
without epoxy were both calculated for each scenario of the
inkjet. This range of scores with and without epoxy should
cover the whole range of materials the inkjet printer is likely
to use, from proprietary plaster formulations to hacker
formulations of sawdust or concrete or other materials.
The Projet used a proprietary SLA resin called Accura ABS
White SL7810, a polymer that hardens with exposure to UV
light. Despite its name, it was not ABS. Its somewhat vague
Materials Safety Data Sheet (MSDS) said it was composed of
hydrogenated bisphenolA epoxy resin, 3-ethyloxetane-3-
methanol, propylene carbonate, “sulfonium salt mixture,” and
bisphenol A epoxy resin. While epoxy resin was in the
EcoInvent database, the other chemicals largely did not match
chemicals in the EcoInvent or other LCA databases available
to this team, so a sensitivity analysis compared 15 different
chemicals considered most likely to match these ingredients’
environmental impacts. Extreme high and low ReCiPe point
values were eliminated, and final LCAs included two
scenarios each—a high estimate assuming the material was
entirely epoxy resin, and a low estimate using “acrylic acid, at
plant”. Resulting differences in total ReCiPe Endpoint H
points per part printed in the four different SLA machine
utilization scenarios ranged from 16% (running 24 hrs/day, 7
days/wk, printing 4 parts at a time) to a 0.2% difference
(printing 1 part/wk but left idling when not in use). Final
results shown later in the Results section use the high
estimates, as the MSDS did explicitly list epoxy resin as
comprising 30–60% of the material.
The Objet used a proprietary “polyjet”UV-curing polymer
called Fullcure 720, whose MSDS listed the ingredients exo-
1,7,7-trimethylbicyclo[2.2.1]-hept-2-yl acrylate, acrylic
monomer, urethane acrylate oligomer, acrylate oligomer, and
epoxy acrylate. Again, exact matches for all these chemicals
were not available in the databases, but sensitivity analysis
was performed, so each scenario had a high estimate (epoxy
resin again) and low estimate (“acrylic acid, at plant” again)
for material impacts. Resulting differences in total ReCiPe
Endpoint H points per part printed ranged from 9% (running
24 hrs/day, 7 days/wk, printing 4 parts at a time) to a 0.2%
difference (printing 1 part/wk but left idling when not in use).
Final results shown here use the high estimates, for
consistency with the SLA machine. High estimates were also
chosen because the purpose of this study was to see how large
variations due to material choice could be, and even the lower-
impact scenarios for these materials these materials were at the
higher-impact end compared to salt and dextrose.
While all of these machine types (FDM, polyjet, inkjet, and
SLA) can print in different materials, the materials listed
above were the only materials made available to us by the
machine operators. Only the Type A machine was measured
using two different materials; for all other machines, the type
of machine was tied to one type of material, and any variation
was from theoretical calculations of sensitivity analysis. While
this is certainly a limitation of the study, we believe the results
show that this does not affect the validity of the conclusions
(see Results section).
C. Machine Utilization
3D printer utilization varies widely in industry—some
machines run nearly 24 hours/day, 7 days/week, especially
those used for manufacturing finished parts (as opposed to
prototypes), or those run by contractors who print for hire
(“job shops”). Other machines may go for days or weeks (even
months) between print jobs, especially small inexpensive
desktop units used by design firms for occasional prototypes,
or used by home hobbyists. An informal utilization survey
sent to nearly a thousand product design practitioners provided
little insight, with few responses and a wide range of answers,
so no defensible “average” utilization could be determined.
Therefore, a range of scenarios was calculated. Maximum
utilization was defined as printing parts 24 hrs/day, 7 days/wk,
for a machine’s entire life (which is not actually possible, but
represents the asymptotic “best case” scenario).
Some printers can only print one part at a time (large and
small FDM machines), but some printers can print several
parts in almost exactly the same time it takes to print a single
part, without using noticeably more energy (polyjet, inkjet,
and SLA machines). Therefore, maximum utilization for
polyjet, inkjet, and SLA machines is not only printing parts 24
hrs/day, 7 days/week, but also printing multiple parts at once.
The number of parts that can be printed at once without adding
more print time (thus adding energy use and higher
environmental impacts) was not clearly defined for any of the
machines, and surely varies from machine to machine, since
the SLA machine can print parts throughout its entire print bed
without adding much more time, while the inkjet and polyjet
machines can only print parts within the width of their moving
print heads without adding more time. Budget and time
constraints did not allow the printing of large numbers of parts
to test the limits of these improved efficiencies, nor did any of
the company representatives provide hard data on the number
of additional parts printed before print times increased, but
informal discussions with machine operators indicated that for
the scale of parts being used as the functional unit here, at
least four parts could be printed in almost the exact same time,
with almost the exact same energy use, as one part. Perhaps
even more parts may be printed simultaneously before the
additional time and energy use would become appreciable
(one company representative suggested ten parts or more at a
time), but such changes would create such extreme
improvements to the ecological impact scores that they should
be backed by real empirical data, not mere estimations.
Here, “minimum” utilization was defined as printing one
part per week, since results of the utilization survey indicated
that was a common (if not necessarily typical or dominant)
rate around the low end of professional use. However, this
minimum utilization was split into two separate scenarios,
because the amount of electricity used by idle machines left on
World Academy of Science, Engineering and Technology
International Journal of Mechanical, Aerospace, Industrial and Mechatronics Engineering Vol:9 No:2, 2015
146International Scholarly and Scientific Research & Innovation 9(2) 2015
International Science Index Vol:9, No:2, 2015 waset.org/Publication/10000327

h
a
t
h
a
r
w
(t
h
g
r
w
w
p
r
h
o
fr
o
d
e
n
o
e
q
fr
o
fr
o
f
a
F
i
A
i
m
p
r
u
t
h
u
c
a
c
h
l
a
c
a
a
r
a
s an enormo
u
h
e machines
m
r
e sitting idle
a
w
atts idle vs.
2
h
e majority),
r
eatly multip
l
w
as for each p
r
w
hen not in u
s
r
int one part
p
o
urs. Other u
t
o
m the data
e
scribed there
A. Ecologica
As mention
e
o
t merely on
e
q
uivalent emi
s
o
m energy
u
o
m resource
a
ctors.
i
g. 2 Ecologica
l
A
BS, operatin
g
are normali
z
Fig. 2 show
s
m
pacts domi
n
r
inting in
A
t
ilization. Fo
s
u
man health,
a
using both
h
h
ange damag
e
a
rgest types
a
tegories (su
c
r
e so small
a
u
s effect on
e
m
easured use
a
s they do w
h
2
80 watts ru
n
the sheer
n
l
y electricity
r
inter to print
s
e, the other
p
er week an
d
t
ilization rate
s
shown in th
e
.
IV.
l Impacts per
e
d earlier, one
e
kind of envi
r
s
sions), but t
o
u
se can be m
e
use and w
a
l
impacts per j
o
g
at maximum
u
z
ed and weight
e
s
that just thr
e
n
ate for the
l
A
BS, operati
n
s
sil fuel depl
e
particulate
m
h
uman and e
c
e
to ecosyst
e
of impacts,
c
h as ionizin
g
a
s to be invi
s
e
cological im
p
nearly as mu
h
en they are p
r
n
ning). Even
f
n
umber of h
o
impacts. Th
e
one part per
scenario wa
s
d
be left on f
o
s
could be cal
c
e
Results sect
R
ESULTS
Part
of this study’
r
onmental im
p
o
measure se
v
eaningfully
c
a
ste, material
o
b for the large
u
tilization. Diff
e
e
d into ReCiPe
e
e to five cat
e
l
arge comme
r
n
g at its t
h
e
tion, climat
e
m
atter format
i
c
osystem hea
l
e
ms, and hu
m
in order.
M
g
radiation, o
z
s
ible. Climat
e
p
act scores.
S
ch power wh
r
inting (at wo
r
f
or those that
o
urs spent i
d
e
refore, one
s
week but be
s
s
for each pr
i
o
r the remain
i
c
ulated
b
y th
e
ion, using a
s goals is to
m
p
act (e.g. kg
o
v
eral, so that
i
c
ompared to
i
toxicity, an
d
FDM machine
e
rent types of i
m
Endpoint H po
e
gories of ec
o
r
cial FDM
m
h
eoretical m
a
e
change da
m
i
on (a kind
o
l
th damage),
m
an toxicity
M
ost other
z
one depleti
o
e
change, fos
S
ome of
en they
r
st, 260
do not
d
le will
s
cenario
s
hut off
i
nter to
i
ng idle
e
reader
method
m
easure
o
f CO2
i
mpacts
i
mpacts
d
other
printing
m
pacts
ints
o
logical
m
achine
a
ximum
m
age to
o
f smog
climate
are the
impact
o
n, etc.)
sil fuel
de
p
d
o
Fi
g
pa
b
e
e
n
us
e
[1
5
bu
m
u
ca
t
in
th
e
co
bu
R
e
n
o
sc
o
b
e
m
f
re
s
m
a
pr
o
m
o
m
a
th
e
m
a
ea
c
“a
l
di
s
ap
p
th
e
as
s
lo
n
es
t
i
m
te
n
st
u
in
ex
(c
o
in
c
pr
i
di
s
p
o
pa
de
p
pr
i
pr
i
a
n
w
e
ap
p
p
o
pr
i
is
pletion, and
o
minant impa
c
g
. 2 also sh
o
a
rts is the do
m
expected, si
n
n
ergy but by c
o
e
rs. The ultr
a
5
] may not b
e
u
t even if h
u
u
ltiplied tenf
o
t
egories. Eve
n
the final pro
e
machines,
uld be argue
d
u
t a previous
e
CiPe Endp
o
o
rmalization a
n
Combining a
l
o
res, we get
F
comes one s
f
g” includes
s
ource use, t
o
a
nufacturing
o
cessing into
o
lded plasti
c
a
chines). The
e
sense that
a
chine, so onl
c
h 3D print
e
l
located dis
p
s
posal of the
m
propriate fra
c
e
Methods
s
s
umed to be
f
n
ger than th
i
t
imates of pri
n
m
pacts accord
i
n
years woul
d
u
dy are divid
e
the printed
p
traction and
o
nservatively
c
ludes both s
u
i
nting that d
o
s
posed of in
s
o
lyjet printe
r
.)
a
rts, as well
pending on t
h
Fig. 3 com
p
i
nting with
d
i
nting one pa
r
n
d idling for
t
e
ek but powe
propriate sta
r
o
lyjet printer,
i
nting one pa
r
because ope
r
smog largel
y
c
t categories
o
ws that elec
t
m
inant cause
o
n
ce 3D print
i
o
mparison ca
u
a
fine plastic
e
adequately
c
u
man toxicit
y
o
ld, they wo
u
n
for machine
ducts printed
these impact
d
this is due
t
study by s
o
o
int H res
u
n
d weighting
l
l the differen
t
F
igs. 3 and
4
egment in t
h
all the clim
a
o
xicity, and
(both raw
the bent st
e
c
parts, and
se manufact
u
they are a
m
y the correct
f
ed
part. Lik
e
p
osal” includ
e
m
achines (no
c
tion allocate
d
s
ection, prin
t
f
ive years; si
n
i
s, readers a
r
n
ter lifetime
a
i
ngly. (For in
s
d
mean all the
e
d by two.) “
M
p
arts themse
l
disposal a
t
assumed t
o
u
pport materi
a
o
es not end
s
tea
d
. (This
“Electricity
u
as to sit idl
e
h
e scenario.
p
ares ecologi
c
d
ifferent mate
r
t per week b
u
t
he rest of th
e
ring complet
e
r
tup and shut
d
which is sh
o
r
t per week b
u
r
ators inform
e
y
dominate b
e
of status-qu
o
t
ricity use du
o
f ecological
i
i
ng uses a si
g
u
ses little dir
e
particles exp
c
aptured by t
h
y
and partic
u
u
ld still not
d
manufacturi
n
and material
categories r
to bias in th
e
o
me of the
a
u
lts against
[22]; they we
t
categories o
4
. Each stack
e
h
e new stack
e
a
te change,
p
other ecolog
i
material e
x
e
el struts, gla
electronics
u
ring impacts
m
ortized acr
o
f
raction of th
e
e
wise, “alloc
a
e
transportat
i
t the printed
p
d
to each part
p
t
er lifetime
n
ce many use
r
r
e welcome
t
a
nd reduce th
e
s
tance, assu
m
“allocated” i
m
M
aterial use”
i
l
ves, as well
t
the end
o
o
always be
a
l and any m
o
up in the f
i
was primari
l
u
se” includes
e
, power up
,
c
al impacts
o
rials, in two
u
t with the pri
n
e
time, and p
e
ly off betwe
d
own times).
T
o
wn in both
u
t left idling,
e
d us that t
h
e
cause they
a
o
US electrici
t
ring the prin
t
i
mpacts, whi
c
g
nificant am
o
e
ct toxin exp
o
osure menti
o
h
e LCA mod
e
u
late impact
s
d
omina
t
e the
n
g, the materi
a
waste produ
emain domi
n
e
weighting
m
a
uthors [6] c
h
IMPACT
re nearly ide
n
f impacts int
o
e
d column in
e
d bars: “Al
l
p
ollution em
i
i
cal impacts
x
traction an
d
ss plates, inj
that compri
s
were "alloca
t
o
ss the life
e
m were allo
c
a
ted transpo
r
i
on and end
-
p
arts), with o
n
p
rinte
d
. As s
t
was conser
v
r
s keep their
p
t
o make the
i
e
allocation
o
m
ing a printer
m
pacts show
n
i
ncludes the
m
as its raw
m
o
f the parts
’
landfill). “
W
o
del material
u
i
nished part,
l
y an issue
f
energy use
d
t
,
and power
o
f different
p
different sce
n
ter left pow
e
rinting one p
en prints (in
c
T
he exceptio
n
areas of the
not turned o
f
h
e machine i
s
a
re the
t
y use.
t
ing of
c
h may
o
unt of
o
sure to
o
ned in
e
l here,
s
were
impact
a
l used
ced by
n
ant. It
m
etho
d
,
h
ecked
2002+
n
tical.
o
single
Fig. 2
l
ocated
i
ssions,
due to
d
their
ection-
s
e the
t
ed" in
of the
c
ated to
r
t” and
-
of-life
n
ly the
t
ated in
v
atively
p
rinters
i
r own
o
f these
life of
n
in this
m
aterial
m
aterial
’
lives
W
aste”
u
sed in
but is
f
o
r
the
t
o print
down,
p
rinters
narios:
e
red on
art per
c
luding
n
is the
graph
f
f. This
s
never
World Academy of Science, Engineering and Technology
International Journal of Mechanical, Aerospace, Industrial and Mechatronics Engineering Vol:9 No:2, 2015
147International Scholarly and Scientific Research & Innovation 9(2) 2015
International Science Index Vol:9, No:2, 2015 waset.org/Publication/10000327

s
h
p
r
a
v
(
W
t
h
m
p
r
n
o
e
l
p
r
w
m
t
h
a
p
P
E
c
o
c
o
a
s
c
A
p
o
c
o
a
r
T
h
e
a
i
d
be
p
a
T
h
t
h
H
F
i
br
p
t
h
ut down bet
w
r
int jobs, due
v
oid clogs f
r
W
hen the m
a
h
rough the li
n
m
achine is sh
o
r
inting in PE
T
Fig. 3 Ecolog
i
machines le
f
Fig. 3 show
s
o
t dominant
l
ectricity use
r
inters thems
e
w
aste impacts
a
m
ate
r
ials can
a
h
e bulk of the
p
pears to be a
E
T in the sa
m
o
mpared to t
h
o
mmercial F
D
small deskt
o
c
enarios is
m
A
BS. The diff
e
o
lyjet, SLA,
o
o
nfounding f
a
r
e larger com
m
h
e most obvi
o
a
ch of these p
d
ling than wh
e
Changes in
e
e
come visibl
e
a
rts are bein
g
h
is is shown
i
h
e difference
i
H
points per jo
b
i
g. 4. The s
m
r
eakout box
f
t
s/job is visib
l
Fig. 4’s gra
p
w
een prints,
to the hassle
r
om resin p
o
a
chine sits id
l
n
es to avoid
c
o
wn twice in
T
, once for pri
n
i
cal impacts pe
r
f
t idling or tur
n
s
that the eco
l
for any of
and allocat
e
e
lves are so
a
re hardly ev
e
a
lso change t
h
machine req
u
minor effect.
m
e desktop
F
h
e difference
b
D
M machine.
o
p FDM of
A
m
iniscule com
p
e
rence betwee
n
o
r inkjet is u
n
a
cto
r
—the p
o
m
ercial-scale
u
o
usly domina
n
rinters has fa
r
e
n turned off
b
e
cological i
m
e
in the maxi
m
g
printed 24
h
i
n Fig. 4. It
w
i
n scale is so
b
go from a
m
m
allest four
b
f
or readabilit
y
l
e.
p
h of impact
s
even if two
w
involved in
p
o
tentially har
d
l
e powered
o
c
logs.) Also
n
each area of
n
ting in PLA.
r
job for low ut
n
ing machines
o
l
ogical impac
t
the scenario
s
e
d impacts o
f
dominant th
e
n visible on
h
e amount o
f
u
ired to print t
h
.
The differen
F
DM machin
e
b
etween that
Likewise, th
e
A
BS and the
p
ared to the
n
printing by
n
clear, becau
s
o
lyjet, SLA,
a
u
nits like the
n
t factor in Fi
g
r
larger impa
c
b
etween print
s
m
pacts due to
m
um utilizat
i
h
ours per da
y
w
as not inclu
d
large—note
t
m
aximum of 5
b
ars in Fig.
y
, so the mi
n
s
at maximu
m
w
eeks pass
b
p
urging fluid
d
ening in th
e
o
n, pumps r
u
n
ote that the
T
the graph—
o
ilization, eithe
r
o
ff when not in
t
s of material
s
picture
d
—
i
f
manufactur
i
at material
u
the graph. C
h
f
electricity u
s
h
e materials;
ce between P
L
e
is almost i
n
machine and
e
difference
b
other deskto
p
large FDM
p
FDM vs. pri
n
s
e machine s
i
a
nd inkjet m
large FDM
m
g
. 3, howeve
r
c
ts per part w
h
s
.
material choi
c
i
on scenarios
y
, 7 days per
d
ed in Fig. 3
b
t
he ReCiPe E
n
.5 in Fig. 3 to
4 are repeat
e
n
imum score
m
utilization
p
b
etween
lines to
e
lines.
u
n fluid
T
ype A
o
nce for
r
with
use
use are
i
n fact,
i
ng the
u
se and
h
oice of
s
ed and
but this
L
A and
n
visible
a large
b
etween
p
FDM
p
rinting
n
ting by
i
ze is a
achines
m
achine.
r
, is that
h
en left
c
e only
(where
wee
k
).
b
ecause
n
dpoint
0.25 in
e
d in a
of .002
p
arallels
th
e
ut
i
se
n
o
n
p
o
ti
m
fo
u
T
h
th
e
si
g
ne
m
e
sc
o
es
t
pr
i
m
a
us
a
ot
h
w
a
F
i
S
c
T
h
de
m
a
te
c
d
w
a
n
a
n
al
o
co
to
i
m
pr
o
e
material an
d
i
lization, bu
t
n
sitivity anal
y
n
ly one part
w
o
lyjet machin
e
m
e as one pa
r
u
r parts are p
r
h
ese scenario
s
e
oretical res
u
g
nificantly m
o
arly the sam
e
e
asuring suc
h
o
pe of this
t
imates of i
m
i
nted simulta
n
a
ximum utili
z
a
ge by the c
h
h
er impact s
o
a
ste, and disp
o
ig. 4 Ecologic
a
c
enarios denot
e
Fig. 4 does s
h
h
e fairest co
m
sktop FDM
m
a
chines wer
e
c
hnology. H
o
w
arfed by the
nd
the large
c
n
d polyjet or
o
ne, because
mmercial uni
t
the desktop
F
m
pacts per pa
r
o
duced a lar
g
d
machine sc
e
t
also adds
y
sis. As me
n
w
as printed p
e
e
s can print
s
r
t. Thus, Fig.
r
inte
d
with t
h
s
are denoted
u
lts. These
p
o
re parts (per
h
e
time with
n
h
variations i
n
study. If re
a
m
proved eco-
e
n
eously, the
y
z
ation graph
,
h
osen factor
o
urces (manu
f
o
sal) constant
a
l impacts per j
o
e
d by (*) are fo
u
h
ow variatio
n
m
parison of
d
m
achines prin
t
e
mos
t
simi
l
o
wever,
t
hei
r
difference b
c
ommercial
F
SLA cannot
the polyjet
t
s like the lar
g
F
DMs. While
t
r
t (it not onl
y
g
e amount of
e
narios in Fig
four additi
o
n
tioned in t
h
e
r machine,
b
s
everal parts
4 also inclu
d
h
e same ener
g
by (*) to in
d
p
rinters ma
y
h
aps ten or
m
n
early the sa
m
n
mass-printi
a
ders wish
t
e
fficiency fro
m
y
can do so
e
,
dividing i
m
of improve
m
f
acturing, tra
n
.
o
b at maximu
m
u
r parts being
p
n
in impacts f
r
d
ifferent mate
r
t
ing PLA, PE
T
l
ar to each
r
difference
etween the s
m
F
DM. Differe
n
be ascribed
and SLA
p
g
e FDM, not
t
he polyjet pr
i
y
used the m
o
waste—roug
h
. 3 running a
t
o
nal variatio
n
h
e Methods
s
b
ut SLA, inkj
e
in nearly th
e
d
es scenarios
g
y usage as o
n
d
icate they ar
e
y
be able t
o
ore parts at o
n
m
e energy u
s
ng was bey
o
t
o make the
i
m
more part
s
e
asily by us
i
m
pacts from
m
ent, and lea
v
n
sport, mater
i
m
utilizatio
n
sce
n
p
rinted simulta
n
r
om material
c
r
ials is in th
e
T
, and ABS,
a
other in si
z
in impacts
i
m
all desktop
n
ces betwee
n
to material
p
rinters wer
e
directly com
p
i
nter had the
h
o
st energy, b
u
h
ly 43% by
m
t
lower
n
s for
s
ection,
e
t, and
e
same
where
n
e part.
e
more
o
print
n
ce) in
s
e, but
o
nd the
i
r own
s
being
i
ng the
energy
v
ing all
i
al use,
n
arios.
n
eously
c
hoice.
e
small
a
s these
z
e and
i
s still
FDMs
n
FDM
choice
e
large
p
arable
h
ighest
u
t also
m
ass of
World Academy of Science, Engineering and Technology
International Journal of Mechanical, Aerospace, Industrial and Mechatronics Engineering Vol:9 No:2, 2015
148International Scholarly and Scientific Research & Innovation 9(2) 2015
International Science Index Vol:9, No:2, 2015 waset.org/Publication/10000327

a
l
br
v
a
p
r
F
D
o
f
m
p
r
“
g
i
m
p
r
f
o
n
e
T
h
j
o
p
a
s
o
p
r
i
m
e
p
p
r
p
r
m
t
o
o
f
e
x
m
j
o
p
e
r
e
tr
a
F
i
m
v
a
u
t
s
e
o
t
o
f
c
o
fr
o
l
l liquid resin
)
r
ings its imp
a
a
lues for co
m
r
inter had hi
D
M, dependi
n
f
four parts.
m
aterial choic
e
Most notabl
y
r
inting techn
o
g
reen” materi
a
m
pact scores
r
inters at ma
x
o
ur parts toge
t
e
xt-
b
est tech
n
h
e inkjet has
o
b as the poly
j
a
rt at a time
o
o
aked into s
a
r
edictably sk
y
m
pac
t
scores
r
p
oxy, impact
r
inting salt w
i
r
inting in P
L
m
aterials in th
e
o
cause ecolo
g
f
salt and the
u
As mention
e
x
trapolate to
e
m
aximal utiliz
a
o
b/week, idlin
e
r week, then
e
scaling the a
m
a
nsport, and
e
i
g. 5 Range of
v
scenarios
B.
R
anges o
f
Ecological
i
m
aterials on
a
riation shoul
t
ilization. Ev
e
e
veral parts
a
t
her variables
f
3D printin
g
o
mpare the r
a
o
m material
c
)
, the scenari
o
a
cts to withi
n
m
mercial F
D
gher or low
e
n
g on wheth
e
Here agai
n
e
or machine t
y
y
, Fig. 4 als
o
o
logy can do
m
a
l of salt doe
s
than all oth
e
x
imum utiliz
a
t
he
r
, it has 1/
5
n
ology, PLA
roughly 1/3
8
j
et, regardles
s
o
r four parts
a
a
lt parts to
h
y
rocket. Print
i
r
oughly doub
l
scores rou
g
i
th epoxy sc
o
L
A. As ment
i
e
inkjet (such
a
g
ical impacts
u
pper bound
o
e
d in the
M
e
ven lower u
t
a
tion energy i
m
g” scores in
F
multiplying
m
ortized imp
a
e
nd of life) ac
c
v
ariation betw
e
of different ut
i
f
Variation
i
mpac
t
score
s
different m
a
d be compar
e
e
n without h
e
a
t once, mac
h
in having th
e
g
. Fig. 5 use
s
a
nge of vari
a
c
hoice agains
t
o
where it pri
n
n
the extreme
D
M impacts.
e
r impacts t
h
e
r it printed s
i
n
machine
u
y
pe.
o
shows wh
e
m
inate: The i
s
in fact have
e
r materials
p
a
tion. When
t
5
th the impac
t
printed by
s
8
th to 1/40th
t
s
of whether
b
a
t a time. Ho
w
h
arden them,
i
ng one part
a
l
e; printing fo
g
hly quintup
l
o
res better th
a
i
oned in Me
t
a
s sawdust, p
l
varying betw
o
f salt with ep
M
ethods secti
o
t
ilization sce
n
m
pact scores
F
ig. 3 to find
that by the d
e
a
cts of the pr
i
c
ordingly.
e
en scenarios o
f
i
lization for di
ff
s
vary greatl
y
a
chines,
b
ut
e
d to the vari
a
e
roic improv
e
h
ine utilizatio
n
e
most influe
n
s
the data fr
o
a
tion in ecol
o
t
the range o
f
n
ts four parts
ends of unc
e
Likewise, t
h
h
an the co
m
i
ngle parts or
u
tilization do
m
e
re material
c
nkjet printin
g
far lower ec
o
p
rinted
b
y a
l
t
he inkjet is
p
t
score per jo
b
s
mall deskto
p
t
he impact s
c
b
oth are print
i
w
ever, when
e
ecological
i
a
t a time with
ur parts at o
n
l
e. Neither
s
a
n the deskto
p
t
hods, printin
g
l
aster, etc.) a
r
een the lowe
r
oxy.
o
n, the read
e
n
arios by sub
t
in Fig. 4 fro
m
idling energy
e
sired idle ti
m
i
nter (manufa
c
f
different mat
e
ff
erent machine
s
y
b
etween
d
as mention
e
a
tion due to
m
e
ments from
p
n
already do
m
n
ce on sustai
n
o
m Figs. 3 a
n
o
gical impac
t
f
variation in
i
at once
e
rtainty
h
e SLA
m
mercial
groups
m
inates
c
hoice /
g
in the
o
logical
l
l other
p
rinting
b
as the
p
FDM.
c
ore per
i
ng one
e
poxy is
i
mpacts
epoxy,
n
ce with
s
cenario
p
FDM
g
other
r
e likely
r
bound
e
r may
t
racting
m
the “1
impact
m
e, and
c
turing,
e
rials vs.
s
d
ifferent
e
d, this
m
achine
p
rinting
m
inates
n
ability
n
d 4 to
t
scores
i
mpacts
fr
o
(P
L
F
D
i
m
ty
p
i
m
(T
si
m
m
a
th
a
ut
i
sa
m
ut
i
m
i
ev
ea
s
U
n
q
u
b
e
T
h
Fi
g
pa
pr
i
te
n
F
D
F
D
ha
w
i
al
s
c
u
de
o
p
w
h
pa
in
to
Fi
g
q
u
ca
u
in
f
o
m machine
u
As Fig. 5 sh
o
L
A, PET, A
B
D
M) gives a
h
m
pact score.
V
p
e
b
ut opera
t
m
pact score i
s
his does n
m
ultaneously,
a
terials.) For
e
a
t machine
i
lization are r
o
m
e machine
i
lization. So
i
nimized by
c
en more cruc
i
C. Print Qua
l
Choosing w
h
s
y if enviro
n
n
fortunately i
t
u
ality, and the
tween print
q
h
is can be se
e
g
. 6’s deskto
p
a
rt. Polyjet an
i
nt quality (h
i
n
d to have h
i
D
M had mid
r
D
Ms and the
i
a
d the lowest
q
i
th less smoo
t
s
o had errors.
u
rved surfac
e
tachment fro
m
p
erator said t
h
h
ich can re
q
a
rameters in o
r
salt had min
o
its lower res
o
g
. 6 Quality an
o
In addition t
o
u
ality. Parts p
u
sing two s
m
f
used with
e
u
tilization.
o
ws, varying
B
S) within th
e
h
ighest impa
c
V
arying the 3
D
t
ing only at
m
s
roughly 35
ot include
as that is
e
ach individ
u
and that
m
o
ughly 45 to
and same
m
although
e
c
hoosing goo
d
i
al first step.
l
ity
h
ich 3D print
e
n
mental imp
a
t
is not. A ve
r
re appears to
q
uality and
e
e
n by compar
i
p
FDM-printe
d
d SLA prints
i
ghest resolu
t
i
gher impact
s
r
ange quality
i
nkjet printin
g
q
uality. Thei
r
t
hness in the
The Afinia
F
e
was slig
h
m
support m
a
h
is is not ver
y
q
uire multip
l
r
der to avoid
o
r s
u
rface irr
e
o
lution.
o
malies in a de
pri
n
o
surface finis
h
rinted in salt
m
all pieces of
t
e
poxy (not s
material amo
e
same type
o
c
t score mere
l
D
printing
m
m
aximum uti
l
times the l
o
machines p
a change
u
al machine, t
h
m
aterial oper
a
95 times the
i
m
aterial ope
r
e
nvironmenta
l
d
materials, g
o
e
r and mater
i
a
ct were the
r
y important
c
be a roughly
e
cological im
p
i
ng Fig. 1’s
S
d
PET part a
n
unquestiona
b
t
ion and smo
o
s
per part. T
h
and midran
g
g
in salt had l
r
prints were
a
curved surfa
c
F
DM’s parts
h
h
tly mangle
a
terial (see F
i
y
common, b
u
l
e test print
s
such marring
e
gularities (se
sktop FDM pri
n
n
t (right)
h
quality, the
r
alone on th
e
t
he part to br
e
hown in Fi
g
ng different
p
o
f machine (
d
l
y double the
m
aterial and
m
l
ization, the
h
o
west impact
rinting four
in utilizatio
n
h
e impact sc
o
a
ting at mi
n
i
mpact score
s
r
ating at ma
x
l
impacts c
o
od utilizatio
n
i
al to use w
o
only consid
e
c
onsideration
i
inverse relat
i
p
act score p
e
S
LA-printed
p
n
d inkjet-prin
t
b
ly have the
h
o
thest surfac
e
h
e large com
m
g
e impacts.
D
ow impacts
b
a
ll lower res
o
c
es, and som
e
h
ad places w
h
d from i
m
i
g. 6). The
m
u
t is a know
n
s
tuning th
e
. The inkjet
p
e Fig. 6) in a
d
n
t (left) and in
k
r
e is also mec
h
e
inkjet were
e
ak off befor
e
g
. 6). While
p
lastics
d
esktop
lowest
m
achine
h
ighest
score.
parts
n
, not
o
res for
n
imum
s
of the
x
imum
an be
n
is an
o
uld be
e
ration.
i
s print
i
onship
e
r part.
p
arts to
t
ed salt
h
ighest
e
s), but
m
ercial
D
esktop
b
ut also
o
lution,
e
prints
h
ere the
m
proper
m
achine
n
issue
e
print
p
rinting
d
dition
k
jet salt
h
anical
brittle,
e
being
many
World Academy of Science, Engineering and Technology
International Journal of Mechanical, Aerospace, Industrial and Mechatronics Engineering Vol:9 No:2, 2015
149International Scholarly and Scientific Research & Innovation 9(2) 2015
International Science Index Vol:9, No:2, 2015 waset.org/Publication/10000327

prototypes do not need physical strength or durability, it can
be a requirement for functional prototypes, so this could be a
significant decision point for some users.
V. LIMITATIONS
For this study, access was available to a limited number of
materials and machines compared to the vast variety that
exists in the market today. We believe it does not harm the
validity of conclusions here, but more data would improve
confidence. The lack of a direct metal laser sinterer (“DMLS”
printer) is significant, as DMLS uses significantly more
energy to print parts in metal than the printers here use to print
parts in plastic. This would increase the variation in
environmental impacts due to material choice. Access to such
machines was unavailable, but readers trying to minimize their
environmental impacts per part made will be content with the
data here, as DMLS will only have higher impacts compared
to printing in plastic or salt. Selective laser sintering (“SLS”)
of plastics would also be useful to measure. For the sake of
completeness, future studies should measure more machine
types and machine sizes.
Machine access was also limited in the number of parts that
could be printed, not allowing finer-grained study of
maximum utilization in machines that could print multiple
parts at once. However, as mentioned in Methods and Results,
reduced eco-impacts from increases in utilization can be easily
estimated by the reader.
VI. CONCLUSIONS
As 3D printing rapidly becomes a large industry, the
industry’s sustainability rapidly becomes important. Part of
this is determining what role material choices play in the
sustainability of 3D printing—whether they dominate impacts,
are insignificant, or somewhere in between. Today, 3D
printing does not commonly use “green” materials which
cause few ecological impacts in their extraction or production.
The possible exception is PLA bioplastic, which is commonly
used, and which this study shows to lower printer energy use
as well as having lower embodied impacts than ABS plastic.
Innovative approaches, such as printing salt with an inkjet 3D
printer, can lower ecological impacts per part even further.
Printing this material on this machine reduced the ReCiPe
Endpoint H impact score per part to as much as 1/35th the
score of the highest-impact printer and material at maximum
utilization (printing parts 24 hrs/day, 7 days/week). Other low-
impact materials could include sawdust, plaster, or other
relatively inert substances that can be bonded with low-
toxicity adhesives. When higher-toxicity adhesives such as the
epoxy studied here are required to give such materials
adequate physical strength, they can eliminate the advantages
of the “greener” material. Here, an inkjet printing salt parts
later infused with epoxy scored worse than a desktop FDM
printing PLA, and similar to a desktop FDM printing PET.
As much of a difference as “green” materials and printers
can make, these advantages can only be realized if machine
utilization is also optimized, to avoid wasting electricity
through powered-up idling between prints, or inefficient print
setups. Idling is particularly important. A printer running at
low utilization (printing one part per week but sitting
powered-on for all its idle time) can have up to roughly 95
times the ecological impact score as the same printer running
at maximum utilization (printing 24 hrs/day, 7 days/wk, 4
parts/print).
With such huge gains possible, 3D printing can be a highly
sustainable manufacturing method if printer manufacturers,
operators, and researchers focus their efforts. Future work
should experiment with and measure the impacts of 3D
printing with more alternative materials that both have low
environmental impacts themselves and also enable low-energy
printing processes. Industry should design printer interfaces
that help maximize printer utilization to avoid idle time and
amortize impacts of machines. For example, interfaces to
encourage sharing printers among multiple users, interfaces to
minimize material use (and thus also print time) in FDM
machines, or interfaces to maximize the number of parts
printed together for SLA, polyjet, and inkjet machines.
Printers should also allow automatic power-saving standby
modes to avoid the impacts of idle power consumption.
Ideally, industry should also steer away from business models
where proprietary materials are the primary profit source, with
printers merely a vehicle for material demand, so that more
material experimentation is enabled. 3D printing can already
be a more sustainable manufacturing method for some
products; with efforts such as these, it might become a greener
way to make most products.
ACKNOWLEDGMENTS
Thanks to EspenSivertsen and Miloh Alexander of Type A
Machines, Patrick Dunne and Marco Teran of 3D Systems,
Ron Rael and Kent Wilson of UC Berkeley architecture dept.,
and Chris Myers of UC Berkeley Invention lab, for access to
machines and helpful information.
REFERENCES
[1] 3D Hubs. “Trend Report June,” Accessed 13 Jun 2014 from
http://www.3dhub s.com/trends/2014-june.
[2] D. Freedman, "Layer by layer," Technology Review 115.1, pp. 50-53,
2012.
[3] C. Reynders, “3D printers create a blueprint for future of sustainable
design and production,” The Guardian, Friday 21 March 2014. Accessed
Sep 15 2014 from http://www.theguardian.com/sustainable-business/3d-
printing-blueprint-future-sustainable-design-production .
[4] M. Huijbregts et al., “Ecological footprint accounting in the life cycle
assessment of products,” Ecological Economics 64.4, pp. 798-807, 2008.
[5] R. Armstrong, “Is There Something Beyond ‘Outside of the Box’?”
Architectural Design 81.6, pp. 130-133, 2011.
[6] J. Faludi, C. Bayley, M. Iribane, S. Bhogal, “Comparing Environmental
Impacts of Additive Manufacturing vs. Traditional Machining via Life-
Cycle Assessment,” Journal of Rapid Prototyping.to be published 2015.
[7] J. Faludi, R. Ganeriwala, B. Kelly, T. Rygg, T. Yang, “Sustainability of
3D Printing vs. Machining: Do Machine Type & Size Matter?”
Accepted for publication in Proceedings of EcoBalance Conference,
Japan 2014.
[8] D. Southerland, P. Walters, and D. Huson, “Edible 3D printing,” NIP &
Digital Fabrication Conference, Vol. 2011 No. 2, Society for Imaging
Science and Technology, 2011.
[9] T. Anderson and J. Bredt, “Method of three dimensional printing,” U.S.
Patent No. 5,902,441, 11 May 1999.
World Academy of Science, Engineering and Technology
International Journal of Mechanical, Aerospace, Industrial and Mechatronics Engineering Vol:9 No:2, 2015
150International Scholarly and Scientific Research & Innovation 9(2) 2015
International Science Index Vol:9, No:2, 2015 waset.org/Publication/10000327

[10] H. Lipson and M. Kurman, Fabricated: The new world of 3D printing,
John Wiley & Sons, 2013.
[11] P. Mognol et al., “Rapid prototyping: energy and environment in the
spotlight,” Rapid Prototyping Journal 12.1, pp. 26-34, 2006.
[12] M. Baumers et al. “Sustainability of additive manufacturing: measuring
the energy consumption of the laser sintering process,” Proceedings of
the Institution of Mechanical Engineers, Part B: Journal of Engineering
Manufacture 225.12, pp. 2228-2239, 2011.
[13] C. Telenko and C. Seepersad, “A comparison of the energy efficiency of
selective laser sintering and injection molding of nylon parts,” Rapid
Prototyping Journal 18.6, pp. 472-481, 2012.
[14] A. Drizo, and J. Pegna, “Environmental impacts of rapid prototyping: an
overview of research to date,” Rapid Prototyping Journal 12.2, pp. 64-
71, 2006.
[15] B. Stephens et al., “Ultrafine particle emissions from desktop 3D
printers,” Atmospheric Environment 79, pp. 334-339, 2013.
[16] Y. Luo et al. “Environmental performance analysis of solid freedom
fabrication processes,” Proceedings of the 1999 IEEE International
Symposium on Electronics and the Environment, pp. 1-6, 1999.
[17] M. Goedkoop et al. ReCiPe 2008: A life cycle impact assessment method
which comprises harmonised category indicators at the midpoint and the
endpoint level, Pré Consultants, 2009.
[18] M. Tabone et al., “Sustainability metrics: life cycle assessment and
green design in polymers,” Environmental Science & Technology 44.21,
pp. 8264-8269, 2010.
[19] M. Rossi et al., “Design for the Next Generation: Incorporating Cradle-
to-Cradle Design into Herman Miller Products,” Journal of Industrial
Ecology 10.4, pp. 193-210, 2006.
[20] B. Evans, Practical 3D Printers, Apress, 2012.
[21] RepRap community, “Powder Printer Recipes,” RepRap Wiki. Accessed
Aug 24 2014 from http://reprap.org/wiki/Powder_Printer Recipes.
[22] O. Jolliet et al., “IMPACT 2002+: a new life cycle impact assessment
methodology,” International Journal of Life Cycle Assessment 8.6, pp.
324-330, 2003.
[1] 3D Hubs. “Trend Report June,” Accessed 13 Jun 2014 from
http://www.3dhub s.com/trends/2014-june.
[2] D. Freedman, "Layer by layer," Technology Review 115.1, pp. 50-53,
2012.
[3] C. Reynders, “3D printers create a blueprint for future of sustainable
design and production,” The Guardian, Friday 21 March 2014. Accessed
Sep 15 2014 from http://www.theguardian.com/sustainable-business/3d-
printing-blueprint-future-sustainable-design-production .
[4] M. Huijbregts et al., “Ecological footprint accounting in the life cycle
assessment of products,” Ecological Economics 64.4, pp. 798-807, 2008.
[5] R. Armstrong, “Is There Something Beyond ‘Outside of the Box’?”
Architectural Design 81.6, pp. 130-133, 2011.
[6] J. Faludi, C. Bayley, M. Iribane, S. Bhogal, “Comparing Environmental
Impacts of Additive Manufacturing vs. Traditional Machining via Life-
Cycle Assessment,” Journal of Rapid Prototyping. to be published 2015.
[7] J. Faludi, R. Ganeriwala, B. Kelly, T. Rygg, T. Yang, “Sustainability of
3D Printing vs. Machining: Do Machine Type & Size Matter?”
Accepted for publication in Proceedings of EcoBalance Conference,
Japan 2014.
[8] D. Southerland, P. Walters, and D. Huson, “Edible 3D printing,” NIP &
Digital Fabrication Conference, Vol. 2011 No. 2, Society for Imaging
Science and Technology, 2011.
[9] T. Anderson and J. Bredt, “Method of three dimensional printing,” U.S.
Patent No. 5,902,441, 11 May 1999.
[10] H. Lipson and M. Kurman, Fabricated: The new world of 3D printing,
John Wiley & Sons, 2013.
[11] P. Mognol et al., “Rapid prototyping: energy and environment in the
spotlight,” Rapid Prototyping Journal 12.1, pp. 26-34, 2006.
[12] M. Baumers et al. “Sustainability of additive manufacturing: measuring
the energy consumption of the laser sintering process,” Proceedings of
the Institution of Mechanical Engineers, Part B: Journal of Engineering
Manufacture 225.12, pp. 2228-2239, 2011.
[13] C. Telenko and C. Seepersad, “A comparison of the energy efficiency of
selective laser sintering and injection molding of nylon parts,” Rapid
Prototyping Journal 18.6, pp. 472-481, 2012.
[14] A. Drizo, and J. Pegna, “Environmental impacts of rapid prototyping: an
overview of research to date,” Rapid Prototyping Journal 12.2, pp. 64-
71, 2006.
[15] B. Stephens et al., “Ultrafine particle emissions from desktop 3D
printers,” Atmospheric Environment 79, pp. 334-339, 2013.
[16] Y. Luo et al. “Environmental performance analysis of solid freedom
fabrication processes,” Proceedings of the 1999 IEEE International
Symposium on Electronics and the Environment, pp. 1-6, 1999.
[17] M. Goedkoop et al. ReCiPe 2008: A life cycle impact assessment method
which comprises harmonised category indicators at the midpoint and the
endpoint level, Pré Consultants, 2009.
[18] M. Tabone et al., “Sustainability metrics: life cycle assessment and
green design in polymers,” Environmental Science & Technology 44.21,
pp. 8264-8269, 2010.
[19] M. Rossi et al., “Design for the Next Generation: Incorporating
Cradle‐to‐Cradle Design into Herman Miller Products,” Journal of
Industrial Ecology 10.4, pp. 193-210, 2006.
[20] B. Evans, Practical 3D Printers, Apress, 2012.
[21] RepRap community, “Powder Printer Recipes,” RepRap Wiki. Accessed
Aug 24 2014 from http://reprap.org/wiki/Powder_Printer Recipes.
[22] O. Jolliet et al., “IMPACT 2002+: a new life cycle impact assessment
methodology,” International Journal of Life Cycle Assessment 8.6, pp.
324-330, 2003.
World Academy of Science, Engineering and Technology
International Journal of Mechanical, Aerospace, Industrial and Mechatronics Engineering Vol:9 No:2, 2015
151International Scholarly and Scientific Research & Innovation 9(2) 2015
International Science Index Vol:9, No:2, 2015 waset.org/Publication/10000327