Ultrafine particle emissions from desktop 3D printers

Abstract

The development of low-cost desktop versions of three-dimensional (3D)
printers has made these devices widely accessible for rapid prototyping
and small-scale manufacturing in home and office settings. Many desktop
3D printers rely on heated thermoplastic extrusion and deposition, which
is a process that has been shown to have significant aerosol emissions
in industrial environments. However, we are not aware of any data on
particle emissions from commercially available desktop 3D printers.
Therefore, we report on measurements of size-resolved and total
ultrafine particle (UFP) concentrations resulting from the operation of
two types of commercially available desktop 3D printers inside a
commercial office space. We also estimate size-resolved
(11.5 nm-116 nm) and total UFP (<100 nm)
emission rates and compare them to emission rates from other desktop
devices and indoor activities known to emit fine and ultrafine
particles. Estimates of emission rates of total UFPs were large, ranging
from
˜2.0 × 1010 # min-1
for a 3D printer utilizing a polylactic acid (PLA) feedstock to
˜1.9 × 1011 # min-1
for the same type of 3D printer utilizing a higher temperature
acrylonitrile butadiene styrene (ABS) thermoplastic feedstock. Because
most of these devices are currently sold as standalone devices without
any exhaust ventilation or filtration accessories, results herein
suggest caution should be used when operating in inadequately ventilated
or unfiltered indoor environments. Additionally, these results suggest
that more controlled experiments should be conducted to more
fundamentally evaluate particle emissions from a wider arrange of
desktop 3D printers.

Full text

Technical note
Ultrafine particle emissions from desktop 3D printers
q
Brent Stephens
a
,
*
, Parham Azimi
a
, Zeineb El Orch
a
,
b
, Tiffanie Ramos
a
a
Department of Civil, Architectural and Environmental Engineering, Illinois Institute of Technology, Chicago, IL, USA
b
National Institute of Applied Sciences (INSA de Lyon), Lyon, France
article info
Article history:
Received 22 April 2013
Received in revised form
24 June 2013
Accepted 26 June 2013
Keywords:
Indoor aerosols
Three-dimensional printers
Thermoplastic emission
Molten extrusion deposition
abstract
The development of low-cost desktop versions of three-dimensional (3D) printers has made these de-
vices widely accessible for rapid prototyping and small-scale manufacturing in home and office settings.
Many desktop 3D printers rely on heated thermoplastic extrusion and deposition, which is a process that
has been shown to have significant aerosol emissions in industrial environments. However, we are not
aware of any data on particle emissions from commercially available desktop 3D printers. Therefore, we
report on measurements of size-resolved and total ultrafine particle (UFP) concentrations resulting from
the operation of two types of commercially available desktop 3D printers inside a commercial office
space. We also estimate size-resolved (11.5 nme116 nm) and total UFP (<100 nm) emission rates and
compare them to emission rates from other desktop devices and indoor activities known to emit fine and
ultrafine particles. Estimates of emission rates of total UFPs were large, ranging from w2.0 10
10
#min
1
for a 3D printer utilizing a polylactic acid (PLA) feedstock to w1.9 10
11
# min
1
for the same type of
3D printer utilizing a higher temperature acrylonitrile butadiene styrene (ABS) thermoplastic feedstock.
Because most of these devices are currently sold as standalone devices without any exhaust ventilation
or filtration accessories, results herein suggest caution should be used when operating in inadequately
ventilated or unfiltered indoor environments. Additionally, these results suggest that more controlled
experiments should be conducted to more fundamentally evaluate particle emissions from a wider
arrange of desktop 3D printers.
Ó2013 The Authors. Published by Elsevier Ltd. All rights reserved.
1. Introduction
Three-dimensional (3D) printers are gaining popularity as
rapid prototyping and small scale manufacturing devices. The
development of low-cost desktop versions has made this tech-
nology widely accessible for use in home and office settings. The
majority of commercially available 3D printers utilize an additive
manufacturing technique known as molten polymer deposition
(MPD), whereby a solid thermoplastic filament is forced through a
computer-driven extrusion nozzle (Bumgarner, 2013). The heated
nozzle melts the thermoplastic feedstock and deposits streams of
extruded plastic in thin layers across a moving baseplate. As the
material hardens and the baseplate moves to the next layer, a
three-dimensional solid shape is rapidly formed.
Several types of thermoplastics are commonly used in a variety
of these commercially available desktop MPD devices. Most
desktop 3D printers currently utilize either acrylonitrile butadiene
styrene (ABS) or polylactic acid (PLA) as a thermoplastic feedstock
(Ragan, 2013). Primary differences between ABS and PLA based
printers are feedstock origin and nozzle and baseplate tempera-
tures during operation. PLA is a biodegradable, corn-based plastic
that prints at nozzle temperatures of w180

C and baseplate
temperatures near room temperature. ABS is a stronger thermo-
plastic that typically prints at w220

C nozzle temperatures and
w80

C baseplate temperatures in most commercially available
devices (Weinhoffer, 2012). Other thermoplastic feedstock sources
include polyvinyl alcohol (PVA), polycarbonate (PC), and high-
density polyethylene (HDPE), although they are not widely used
in commercially available devices (Ragan, 2013).
Previous studies of moderately high temperature (e.g., 170e
240

C nozzle temperatures) thermal processing of thermoplastics
in large scale industrial extrusion equipment have shown that both
gases and particles are emitted during operation (Contos et al.,
1995; Unwin et al., 2012). Primary gas-phase products of ABS
thermal decomposition at very high temperatures have been
shown to include carbon monoxide and hydrogen cyanide, as well
q
This is an open-access article distributed under the terms of the Creative
Commons Attribution-NonCommercial-No Derivative Works License, which per-
mits non-commercial use, distribution, and reproduction in any medium, provided
the original author and source are credited.
*Corresponding author.
E-mail address: brent@iit.edu (B. Stephens).
Contents lists available at SciVerse ScienceDirect
Atmospheric Environment
journal homepage: www.elsevier.com/locate/atmosenv
1352-2310/$ esee front matter Ó2013 The Authors. Published by Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.atmosenv.2013.06.050
Atmospheric Environment 79 (2013) 334e339

as a variety of volatile organics (Rutkowski and Levin, 1986).
Exposure to thermal decomposition products from ABS has also
been shown to have toxic effects in both rats (Zitting and
Savolainen, 1980) and mice (Schaper et al., 1994).
Other studies have shown that exposure to fumes from thermal
decomposition of other thermoplastics, such as polytetrafluoro-
ethylene (PTFE), is also acutely toxic to mammals, including
humans (Oberdörster et al., 2005 and references therein). More-
over, ultrafine particles (UFPs: particles less than 100 nm) may be of
particular importance for toxicity of fumes emitted from melting of
some thermoplastics. For example, in a previous study of high
temperature melting of PTFE at w480

C, UFPs with a count median
diameter of 18 nm were produced, which were also shown to be
highly toxic to rats (Oberdörster et al., 1995). Additional studies
revealed that the gas phase of PTFE fumes alone was not acutely
toxic (Johnston et al., 2000), which suggests that ultrafine aerosols
emitted from thermal degradation of thermoplastic materials may
be of concern among emission products generated in these high
temperature applications.
Despite the rapid commercial uptake of desktop 3D printers that
rely on similar moderate or high temperature thermoplastic melting
and extrusion technology, we are not aware of any data on particle
emissions from these commercially available devices. Indoor
emissions from these devices may be particularly important
because they are most often sold as standalone devices without
mechanical ventilation or filtration accessories. Therefore, we
report on the first measurements of which we are aware of size-
resolved and total UFP concentrations resulting from the opera-
tion of several models of a single type of commercially available
desktop 3D printer utilizing two types of thermoplastic feedstocks
and operating inside a small office space. We also estimate indi-
vidual size-resolved and total UFP emission rates and compare them
to other desktop devices and indoor activities known to emit UFPs.
2. Methods
Measurements were conducted in a 45 m
3
furnished and
conditioned office space belonging to a company who specializes in
3D printer education, training, and sales for the general public. Nine
3D printers were installed on tables in this particular space; only
five adjacent printers were used in this study. Particle number
concentrations were measured in the closed room using a TSI
NanoScan SMPS Model 3910 logging at 1-min intervals. The SMPS
utilizes an isopropanol-based CPC and a radial DMA for size reso-
lution across 13 bins from 10 to 420 nm. It was placed on a table
inside the closed room approximately 2 m away from the nearest
printer. The door remained closed during the testing procedure
except during periods when printers were reset.
We were granted access to the office space only for a limited
time and thus relied on an ad hoc experimental design consisting of
four distinct operational periods over a period of approximately
2.5 h: (1) background measurements without printers operating for
approximately 25 min; (2) two identical 3D printers using PLA
thermoplastic feedstocks operating for approximately 20 min to
print small plastic figures (followed by a short decay period while
the next printing period was setup); (3) the same two PLA-based
printers operating simultaneously with three of the same make
and model printers operating with higher temperature ABS feed-
stocks for approximately 20 min to print another set of small plastic
figures; and (4) a concentration decay period lasting approximately
40 min. The measured concentration data were used to solve for
size-resolved and total UFP emission rates using a combination of
methods from the various monitoring periods, as described below.
Increases in particle concentrations during either printer oper-
ation period were observed only for particles smaller than the
150 nm bin; therefore we only use data from the first nine particle
size bins (11.5 nme116 nm) in this work. In addition to size-
resolved measurements, we also describe total UFPs as the sum
of the first eight particle size bins smaller than 100 nm, per existing
nomenclature in the field (Oberdörster et al., 2005). The room was
assumed to be well-mixed for all periods, which should be
reasonable for a small room with several high temperature devices
operating over a 2.5 h time period (Baughman et al., 1994; Klepeis,
1999), although we cannot confirm this assumption. However,
similar approaches for estimating emission rates from even larger
spaces have been used successfully in previous studies (Wallace
et al., 2004; Buonanno et al., 2009).
2.1. Period 1: background measurements
Upon arrival to the room the printers had not been operating
since the previous day. Particle concentrations were first measured
inside the room during this time for a period of w25 min. Because
the resulting concentrations were stable, data from this period was
used as the representative steady-state background concentration
for each particle size, including total UFPs (i.e., the sum of all par-
ticle concentrations less than 100 nm). The background concen-
tration is a function of the fundamental parameters described in
Equation (1), although only concentrations were directly measured.
C
i;in;ss;bg
¼P
i
l
C
i;out
L
i
(1)
where C
i,in,ss,bg
is the mean baseline size-resolved (or total UFP)
indoor particle concentration (# cm
3
); C
i,out
is the size-resolved
outdoor particle concentration (# cm
3
, not measured in this
study); L
i
is the size-resolved total loss rate in the room due to the
combined effects of air exchange with outside the room, deposi-
tion to surfaces, and removal by any operating HVAC system and
filter (min
1
); P
i
is the room penetration factor (dimensionless, not
measured); and
l
is the air exchange rate in the room (# min
1
,
not measured). Because of the short access window for the field
measurements and limited instrument availability, we were not
able to measure the air exchange rate in the office space. However,
the total loss rate L
i
was estimated from the decay period in this
study according to a procedure outlined in the description of
Period 4. Total loss rates are then used for estimating emission
rates in both Periods 2 and 3. This same procedure has been used
successfully in other previous studies of indoor UFP emissions
from a variety of devices and appliances (e.g., Wallace et al., 2004;
Buonanno et al., 2009).
2.2. Period 2: two PLA printers operating
During Period 2, two 3D printers were operated simultaneously
to print sample plastic products. Printer operation continued until
the objects were fully printed, which took approximately 20 min.
These identical make and model printers utilized a polylactic acid
(PLA) feedstock and operated at an extruder temperature of 200

C
and baseplate temperature of 18

C. One printer assembled a small
plastic frog (shown in Fig. 1) and the other assembled a plastic
chain link.
UFP concentrations measured during Period 2 were shown to
reach approximately steady state before the print jobs were com-
plete (see Fig. 2). Therefore, individual emission rates were esti-
mated using steady-state indoor concentrations (C
i,in,ss,2PLA
)as
shown in Equation (2). Emission rates were assumed to be the same
for each PLA printer because they were the same make and model,
although we were not able to test the validity of this assumption.
B. Stephens et al. / Atmospheric Environment 79 (2013) 334e339 335

C
i;in;ss;2PLA
¼C
i;in;ss;bg
þ2E
i;PLA
=V
L
i
(2)
where E
i,PLA
is the individual size-resolved (or total) UFP emission
rate from each of the two PLA printers (# min
1
) and Vis the vol-
ume of the room (cm
3
). The mean and standard deviation of the
background concentrations from Period 1 were used in conjunction
with steady-state data from Period 2 and estimates of L
i
from Period
4 to solve for E
i,PLA
. The procedure was performed using data from
each of the nine particle size bins (11.5e116 nm), as well as for the
total UFP number concentrations (the sum of all eight UFP bins).
Uncertainty in E
i,PLA
was estimated as the relative standard de-
viations from the means for both C
i,in,ss,bg
and C
i,in,ss,2PLA
added in
quadrature with the relative uncertainty in L
i
. Period 2 ended when
the 3D printers finished printing the models; subsequently there
was a short decay period while we prepared additional printers for
Period 3 tasks.
2.3. Period 3: two PLA printers and three ABS printers operating
During Period 3, the same two printers from Period 2 utilizing
PLA feedstocks were operated again along with three additional
printers of the same make and model. However, the three addi-
tional printers utilized an ABS feedstock and operated at higher
extruder and baseplate temperatures of 220

C and 118

C,
respectively. Each printer assembled another small plastic frog,
with the exception of one of the PLA printers that again assembled a
small chain link. Because of the time-varying nature of the data
measured during this period, emission rates were estimated using
the analytical solution to a dynamic mass balance on the space, as
shown in Equation (3).
C
i;in
ðtÞ¼C
i;in;t¼0
e
L
i
t
þC
i;in;ss;bg
þ2E
i;PLA
=Vþ3E
i;ABS
=V
L
i

1e
L
i
t
(3)
A two-parameter nonlinear least squares regression (Stata
Version 11) was performed to estimate the size-resolved (and total)
UFP emission rates from the combination of all five printers during
this period, where E
i,total
¼2E
i,PLA
/Vþ3E
i,ABS
/V. Both the initial
concentration (C
i,in,t¼0
) and the total emission rates (E
i,total
)were
treated as unknowns for each particle size (and total UFPs) in the
regression analysis. The same values of L
i
from Period 4 decay data
(which were also used in Period 2) were used in conjunction with
mean values of C
i,in,ss,bg
from Period 1. Uncertainty in emission rates
was estimated as the relative standard error of the regression co-
efficients for both E
i,total
and L
i
added in quadrature with the rela-
tive standard deviation in C
i,in,ss,bg
from Period 1. Size-resolved UFP
emission rates for the two PLA printers (E
i,PLA
) were assumed to be
the same as in Period 2 and size-resolved UFP emission rates for
each of the three ABS printers (E
i,ABS
) were also assumed to be
equal. As a check on our emission rate estimates, we also estimated
size-resolved and total UFP emission rates assuming there were no
losses during this brief period of high emissions (where
E
i,total
¼
D
C
i
V/
D
t). This served to provide a reasonable lower bound
of emission rates for the printers. This procedure for solving for
emission rates neglects any particle coagulation or growth by
condensation, which may introduce some additional uncertainty in
our estimates of size-resolved emission rates. However, emission
rates of total UFPs based on this approach are not affected by
coagulation or condensation growth, as these mechanisms only
impact individual bins within the total UFP range. Other recent
studies have also used a similar methodology neglecting coagula-
tion or condensation, as source strengths are often large enough to
overwhelm other impacts within a short period of testing (e.g.,
Wallace et al., 2004; Afshari et al., 2005; He et al., 2007).
0
50000
100000
150000
UFP(#cm
-3
)
0
5000
10000
15000
20000
25000
30000
Number concentration (# cm
-3
)
0 30 60 90 120 150
Time
(
minutes
)
11. 5 n m
15.4 nm
20.5 nm
27.4 nm
36.5 nm
48.7 nm
64.9 nm
86.6 nm
116 n m
2PLA+
3ABS
Printers off
Total UFPs
11. 5 n m
15.4 nm
20.5 nm
27.4 nm
36.5 nm
48.7 nm
64.9 nm
86.6 nm
116 n m
Background 2PLA
printers
Printers
off
Fig. 2. Size-resolved and total (<100 nm) ultrafine particle (UFP) concentrations measured in the office space during the sampling campaign.
Fig. 1. Example of a three-dimensionally printed frog in this study.
B. Stephens et al. / Atmospheric Environment 79 (2013) 334e339336

2.4. Period 4: printers off (decay)
Finally, during Period 4, all printers were turned off and indoor
concentrations decayed back toward normal background levels.
Total loss rates (L
i
) were estimated for each particle size and for
total UFPs using the time-varying data that fit a straight line on a
log-linear plot, as shown in Equation (4).
ln C
i;in
ðtÞC
i;in;ss bg
C
i;in
ðt¼0ÞC
i;in ss bg
!L
i
t(4)
Estimating loss rates in this manner provides lumped loss rates
that account for the combined effects of air exchange, deposition to
indoor surfaces, and removal by any HVAC filtration. Unfortunately
due to both equipment and access limitations, relative contribu-
tions of each of these were not directly measured. However, pre-
vious studies have shown this lumped approach to be valid for
estimating emission rates (Wallace et al., 2004; Buonanno et al.,
2009). Total particle loss rates were assumed to be constant dur-
ing the relatively short field-testing period of 2.5 h for use in solving
for emission rates from the previous periods.
3. Results
Fig. 2 shows resulting time-resolved UFP concentrations
measured in the office space throughout the sampling campaign.
The bottom portion shows size-resolved number concentrations
for the 11.5 nme116 nm particle size bins (as previously
mentioned, no elevations in particle concentrations were observed
in particle bins larger than 116 nm and thus are not shown for
graphical clarity). The top portion of Fig. 2 also shows total UFP
concentrations summed across the first 8 particle size bins smaller
than 100 nm.
The operation of the two printers utilizing PLA as a feedstock
increased concentrations primarily for particles larger than 20 nm.
Indoor concentrations during PLA printer operation also reached
approximately steady-state conditions for a period of w15 min.
Subsequently, the operation of the two PLA printers in conjunction
with three of the same make and model printers (albeit utilizing a
higher temperature ABS feedstock) resulted in substantial increases
in all UFP sizes. Table 1 summarizes mean (s.d.) size-resolved and
total UFP concentrations measured during both background (Period
1) and steady-state operation of two PLA-based printers (Period 2),
along with peak concentrations from the operation of all five 3D
printers (Period 3). Only peak concentrations are shown for each
particle size for Period 3 measurements because steady-state
conditions were not achieved before the printers finished their
print jobs. Table 1 also summarizes loss rates and associated un-
certainty estimated from Period 4 data. Regression coefficients
showed good correlation between measured and modeled con-
centrations during the decay period for most particle sizes (except
for the smallest and largest size bins).
Mean size-resolved particle concentrations during the opera-
tion of the two printers utilizing PLA feedstock were a factor of
w1e4 times higher than during background periods, depending on
particle size. Total UFP concentrations were almost three times
higher (w27,800 cm
3
vs. w9700 cm
3
). The largest increases were
observed in the 36e86 nm size ranges. During operation of the
same two PLA printers combined with three additional printers
utilizing ABS feedstocks, size-resolved particle concentrations
rapidly elevated to as high as 9e56 times background and 3.6e31
times that with only two PLA-based printers operating, depending
on particle size. Peak total UFP concentrations with all five printers
operating (w142,200 cm
3
) were five times higher than with only
two PLA-based printers operating and nearly 15 times higher than
background conditions.
Lumped loss rates estimated from all of the Period 4 decay data
ranged from w2.5 h
1
to w5.6 h
1
. Total UFP loss rates were
approximately 3 h
1
. The largest uncertainty in loss rates was
associated with the largest and smallest particle size bins, likely
due to relatively low peak concentrations from which decay
occurred. Although the relative contributions of air exchange,
deposition to surfaces, and control by any HVAC filtration are not
known, they are not necessary to solve for emission rates herein
(Wallace et al., 2004; Buonanno et al., 2009).
Fig. 3 shows size-resolved and total UFP emission rates and
associated uncertainty for individual 3D printers estimated from the
measured concentration data following the methodology described
in Section 2. The higher temperature ABS-based printers had total
UFP emission rates nearly an order of magnitude higher than the
lower temperature PLA-based printers (1.8e2.0 10
11
# min
1
compared to 1.9e2.0 10
10
# min
1
). Peak emission rates from the
PLA-based printers occurred in the 48e65 nm size range while peak
emission rates from the higher temperature ABS-based printers
occurred in a smaller size range (w15e49 nm). Table 2 describes the
same central estimates of UFP emission rates from Fig. 3 along with
ranges of uncertainty estimated for each size bin. Additionally,
minimum estimates of emission rates made by ignoring particle
losses were 30e51% lower than our central estimates; for example,
the minimum estimate for the total UFP emission rate from a single
ABS printer was 9.7 10
10
# min
1
compared to 1.9 10
11
# min
1
for our best estimate. Therefore, even if there is additional uncer-
tainty in our estimates of emission rates, the ABS printers still have a
total UFP emission rate on the order of 10
11
# min
1
.
Several recent studies have also reported size-resolved and/or
total UFP emission rates from a variety of other consumer devices,
appliances, and activities such as laser printers, candles, cigarettes,
irons, radiators, and cooking on gas and electric stoves (e.g.,
Dennekamp, 2001; Wallace et al., 2004, 2008; Afshari et al., 2005;
Buonanno et al., 2009; He et al., 2010). Unfortunately, it is not
straightforward to compare our results directly to results from
many of these studies because they have varied in both their
minimum and maximum measured particle sizes, as well as in their
definition of UFPs. However, Buonanno et al. (2009) reported total
UFP emission rates over the same size range as ours measured
during various cooking activities. For comparison, our estimate of
the total UFP emission rate for a single PLA-based 3D printer (1.9e
2.0 10
10
# min
1
) was similar to that reported during cooking
with an electric frying pan (1.1e2.7 10
10
# min
1
). The same 3D
printer utilizing a higher temperature ABS feedstock had an emis-
sion rate estimate (1.8e2.0 10
11
# min
1
) similar to that reported
during grilling food on gas or electric stoves at low power (1.2e
Table 1
Summary statistics for each test period.
Particle
size
Period 1 Period 2 Period 3 Period 4
C
i,ss,in,bg
(# cm
3
)
C
i,ss,in,2PLA
(# cm
3
)
C
i,in,2PLA þ3ABS
(# cm
3
)
L
i
(h
1
)R
2
Mean (s.d.) Mean (s.d.) Peak
11.5 nm 158 (25) 495 (33) 8977 4.01 0.20 0.91
15.4 nm 544 (27) 557 (33) 17,438 5.65 0.08 0.99
20.5 nm 975 (25) 1085 (63) 14,356 4.19 0.06 0.99
27.4 nm 1779 (64) 3712 (99) 22,475 2.89 0.05 0.99
36.5 nm 2025 (76) 5976 (164) 27,185 2.58 0.05 0.98
48.7 nm 1793 (76) 6671 (229) 25,473 2.54 0.06 0.98
64.9 nm 1372 (66) 5649 (218) 20,265 2.54 0.06 0.98
86.6 nm 1038 (45) 3695 (161) 13,479 3.09 0.09 0.97
116 nm 695 (58) 1575 (137) 142,211 4.42 0.35 0.80
UFP 9684 (248) 27,838 (658) 6345 2.97 0.04 0.99
B. Stephens et al. / Atmospheric Environment 79 (2013) 334e339 337

2.9 10
11
# min
1
), but approximately an order of magnitude
lower than gas or electric stoves operating at high power (1.2e
3.4 10
12
# min
1
). Regardless, the desktop 3D printers measured
herein can all be classified as “high emitters”with UFP emission
rates greater than 10
10
particles per min, according to criteria set
forth in He et al. (2007).
4. Discussion
UFPs are particularly relevant from a health perspective because
they deposit efficiently in both the pulmonary and alveolar regions
of the lung (Hinds, 1999; Chalupa et al., 2004), as well as in head
airways. Deposition in head airways can also lead to translocation
to the brain via the olfactory nerve (Oberdörster et al., 2004). The
high surface areas associated with UFPs also lead to high concen-
trations of other adsorbed or condensed compounds (Delfino et al.,
2005; Sioutas et al., 2005). Several recent epidemiological studies
have shown that elevated UFP number concentrations are associ-
ated with adverse health effects, including total and cardio-
respiratory mortality (Stölzel et al., 2007), hospital admissions for
stroke (Andersen et al., 2010), and asthma symptoms (Peters et al.,
1997; Penttinen et al., 2001; Von Klot et al., 2002). Therefore, re-
sults herein suggest that caution should be used when operating
these 3D printing instruments inside unvented or unfiltered indoor
environments due to their large emissions of UFPs.
One important limitation to this study is that we have no in-
formation about the chemical constituents of the UFPs emitted
from either type of 3D printer, although condensation of synthetic
organic vapors from the thermoplastic feedstocks are likely a large
contributor (Morawska et al., 2009). In addition to large differences
in emission rates observed between PLA- and ABS-based 3D
printers, there may also be differences in toxicity because of dif-
ferences in chemical composition. As mentioned, thermal decom-
position products from ABS have been shown to have toxic effects
(Zitting and Savolainen, 1980; Schaper et al., 1994); however, PLA is
known for its biocompatibility and PLA nanoparticles are widely
used in drug delivery (Anderson and Shive, 1997; Hans and
Lowman, 2002).
Another important limitation to this study is that we did not
explicitly account for particle coagulation or growth by condensa-
tion in our methodology for estimating emission rates. Coagulation
has been shown to be a factor in indoor environments primarily
during the first few minutes of high concentration periods (e.g.,
>20,000 cm
3
) for particles smaller than 20 nm, and particularly
for those smaller than 10 nm (Wallace et al., 2008; Rim et al., 2012).
Fig. 2 shows that there may have been some additional losses from
the smallest size bin due to coagulation and/or condensation
growth, as peak concentrations lagged other sizes by 3e5 min.
However, it is not clear whether this is due to particle growth or
mixing issues or, alternatively, that the difference is meaningful.
Because we could not conduct controlled experiments in the test
space with limited access, this potential evolution is difficult to
evaluate precisely and is not explicitly accounted for herein, which
may introduce some additional uncertainty in our estimates of size-
resolved emission rates. However, emission rates of total UFPs
based on the lumped loss rate approach are not affected by coag-
ulation or condensation growth, as these mechanisms only impact
individual bins within the total UFP range and no emissions were
observed for particles greater than the 116 nm size bin. Regardless,
particle growth and/or coagulation should be explored in more
controlled environments in future studies.
5. Conclusions
In this work, we present some of the first known measurements
of which we are aware of UFP emissions from commercially avail-
able desktop 3D printers. Emission rates of total UFPs were
approximately an order of magnitude higher for 3D printers uti-
lizing an ABS thermoplastic feedstock relative to a PLA feedstock:
w1.9 10
11
# min
1
compared to w2.0 10
10
# min
1
. However,
Table 2
Estimates of emission rates from individual 3D printers utilizing different thermo-
plastic feedstocks.
Particle
size bin
UFP emission rates per printer (# min
1
)
PLA (Period 2) ABS (Period 3)
Central
estimate
Range Central
estimate
Range
11.5 nm 5.8 10
8
[4.7e6.8] 10
8
1.4 10
10
[1.1e1.7] 10
10
15.4 nm 2.8 10
7
[2.6e3.0] 10
7
3.2 10
10
[2.8e3.6] 10
10
20.5 nm 1.7 10
8
[1.6e1.8] 10
8
2.5 10
10
[2.2e2.7] 10
10
27.4 nm 2.4 10
9
[2.2e2.5] 10
9
3.0 10
10
[2.8e3.3] 10
10
36.5 nm 3.6 10
9
[3.4e3.7] 10
9
3.4 10
10
[3.2e3.7] 10
10
48.7 nm 4.5 10
9
[4.2e4.7] 10
9
3.1 10
10
[2.8e3.3] 10
10
64.9 nm 4.0 10
9
[3.8e4.3] 10
9
2.2 10
10
[2.0e2.4] 10
10
86.6 nm 3.0 10
9
[2.8e3.2] 10
9
1.5 10
10
[1.3e1.6] 10
10
116 nm 1.5 10
9
[1.2e1.7] 10
9
6.9 10
9
[5.4e8.5] 10
9
Total UFPs
(<100 nm)
2.0 10
10
[1.9e2.0] 10
10
1.9 10
11
[1.8e2.0] 10
11
Fig. 3. Individual UFP emission rates from 3D printers utilizing two types of thermoplastic feedstocks in this study: (a) size-resolved emission rates (11.5e116 nm) and (b) total UFP
(<100 nm) emission rates.
B. Stephens et al. / Atmospheric Environment 79 (2013) 334e339338

both can be characterized as “high emitters”of UFPs. These results
suggests caution should be used when operating some commer-
cially available 3D printers in unvented or inadequately filtered
indoor environments. Additionally, more controlled experiments
should be conducted to more fundamentally evaluate aerosol
emissions from a wider range of desktop 3D printers and
feedstocks.
Acknowledgments
The authors thank Robert Zylstra and Julie Friedman Steele from
The Metaspace and The 3D Printer Experience, respectively, for
providing access to the test space.
References
Afshari, A., Matson, U., Ekberg, L.E., 2005. Characterization of indoor sources of fine
and ultrafine particles: a study conducted in a full-scale chamber. Indoor Air 15,
141e150.
Andersen, Z.J., Olsen, T.S., Andersen, K.K., Loft, S., Ketzel, M., Raaschou-Nielsen, O.,
2010. Association between short-term exposure to ultrafine particles and
hospital admissions for stroke in Copenhagen, Denmark. European Heart
Journal 31, 2034e2040.
Anderson, J.M., Shive, M.S., 1997. Biodegradation and biocompatibility of PLA and
PLGA microspheres. Advanced Drug Delivery Reviews 28, 5e24.
Baughman, A.V., Gadgil, A.J., Nazaroff, W.W., 1994. Mixing of a point source
pollutant by natural convection flow within a room. Indoor Air 4, 114e122.
Bumgarner, B., 2013. Getting started with a 3D printer. Make, 12e16.
Buonanno, G., Morawska, L., Stabile, L., 2009. Particle emission factors during
cooking activities. Atmospheric Environment 43, 3235e3242.
Chalupa, D.C., Morrow, P.E., Oberdörster, G., Utell, M.J., Frampton, M.W., 2004. Ul-
trafine particle deposition in subjects with asthma. Environmental Health
Perspectives 112, 879e882.
Contos, D.A., Holdren, M.W., Smith, D.L., Brooke, R.C., Rhodes, V.L., Rainey, M.L.,
1995. Sampling and analysis of volatile organic compounds evolved during
thermal processing of acrylonitrile butadiene styrene composite resins. Journal
of the Air & Waste Management Association 45, 686e694.
Delfino, R.J., Sioutas, C., Malik, S., 2005. Potential role of ultrafine particles in as-
sociations between airborne particle mass and cardiovascular health. Envi-
ronmental Health Perspectives 113, 934e946.
Dennekamp, M., 2001. Ultrafine particles and nitrogen oxides generated by gas and
electric cooking. Occupational and Environmental Medicine 58, 511e516.
Hans, M.., Lowman, A., 2002. Biodegradable nanoparticles for drug delivery and
targeting. Current Opinion in Solid State and Materials Science 6, 319e327.
He, C., Morawska, L., Taplin, L., 2007. Particle emission characteristics of office
printers. Environmental Science & Technology 41, 6039e6045.
He, C., Morawska, L., Wang, H., Jayaratne, R., McGarry, P., Richard Johnson, G.,
Bostrom, T., Gonthier, J., Authemayou, S., Ayoko, G., 2010. Quantification of the
relationship between fuser roller temperature and laser printer emissions.
Journal of Aerosol Science 41, 523e530.
Hinds, W.C., 1999. Aerosol Technology: Properties, Behavior, and Measurement of
Airborne Particles. Wiley-Interscience.
Johnston, C., Finkelstein, J.N., Mercer, P., Corson, N., Gelein, R., Oberdörster, G., 2000.
Pulmonary effects induced by ultrafine PTFE particles. Toxicology and Applied
Pharmacology 168, 208e215.
Klepeis, N.E., 1999. Validity of the uniform mixing assumption: determining human
exposure to environmental tobacco smoke. Environmental Health Perspectives
107 (Suppl. 2), 357e363.
Von Klot, S., Wölke, G., Tuch, T., Heinrich, J., Dockery, D.W., Schwartz, J.,
Kreyling, W.G., Wichmann, H.E., Peters, A., 2002. Increased asthma medication
use in association with ambient fine and ultrafine particles. European Respi-
ratory Journal 20, 691e702.
Morawska, L., He, C., Johnson, G., Jayaratne, R., Salthammer, T., Wang, H., Uhde, E.,
Bostrom, T., Modini, R., Ayoko, G., McGarry, P., Wensing, M., 2009. An investi-
gation into the characteristics and formation mechanisms of particles origi-
nating from the operation of laser printers. Environmental Science &
Technology 43, 1015e1022.
Oberdörster, G., Celein, R.M., Ferin, J., Weiss, B., 1995. Association of particulate air
pollution and acute mortality: involvement of ultrafine particles? Inhalation
Toxicology 7, 111e124.
Oberdörster, G., Oberdörster, E., Oberdörster, J., 2005. Nanotoxicology: an emerging
discipline evolving from studies of ultrafine particles. Environmental Health
Perspectives 113, 823e839.
Oberdörster, G., Sharp, Z., Atudorei, V., Elder, A., Gelein, R., Kreyling, W., Cox, C.,
2004. Translocation of inhaled ultrafine particles to the brain. Inhalation Toxi-
cology 16, 437e445.
Penttinen, P., Timonen, K.L., Tiittanen, P., Mirme, A., Ruuskanen, J., Pekkanen, J.,
2001. Ultrafine particles in urban air and respiratory health among adult
asthmatics. European Respiratory Journal 17, 428e435.
Peters, A., Wichmann, H.E., Tuch, T., Heinrich, J., Heyder, J., 1997. Respiratory effects
are associated with the number of ultrafine particles. American Journal of
Respiratory and Critical Care Medicine 155, 1376e1383.
Ragan, S., 2013. Plastics for 3D printing. Make, 22.
Rim, D., Green, M., Wallace, L., Persily, A., Choi, J., 2012. Evolution of ultrafine par-
ticle size distributions following indoor episodic releases: relative importance
of coagulation, deposition and ventilation. Aerosol Science and Technology 46,
494e503.
Rutkowski, J.V., Levin, B.C., 1986. Acrylonitrile-butadiene-styrene copolymers (ABS):
pyrolysis and combustion products and their toxicityea review of the literature.
Fire and Materials 10, 93e105.
Schaper, M.M., Thompson, R.D., Detwiler-Okabayashi, K.A., 1994. Respiratory re-
sponses of mice exposed to thermal decomposition products from polymers
heated at and above workplace processing temperatures. American Industrial
Hygiene Association Journal 55, 924e934.
Sioutas, C., Delfino, R.J., Singh, M., 2005. Exposure assessment for atmospheric ul-
trafine particles (ufps) and implications in epidemiologic research. Environ-
mental Health Perspectives 113, 947e955.
Stölzel, M., Breitner, S., Cyrys, J., Pitz, M., Wölke, G., Kreyling, W., Heinrich, J.,
Wichmann, H.-E., Peters, A., 2007. Daily mortality and particulate matter in
different size classes in Erfurt, Germany. Journal of Exposure Science & Envi-
ronmental Epidemiology 17, 458e467.
Unwin, J., Coldwell, M.R., Keen, C., McAlinden, J.J., 2012. Airborne emissions of
carcinogens and respiratory sensitizers during thermal processing of plastics.
Annals of Occupational Hygiene 57, 399e406.
Wallace, L., Wang, F., Howard-Reed, C., Persily, A., 2008. Contribution of gas and
electric stoves to residential ultrafine particle concentrations between 2 and
64 nm: size distributions and emission and coagulation rates. Environmental
Science & Technology 42, 8641e8647.
Wallace, L.A., Emmerich, S.J., Howard-Reed, C., 2004. Source strengths of ultrafine
and fine particles due to cooking with a gas stove. Environmental Science &
Technology 38, 2304e2311.
Weinhoffer, E., 2012. 3D printing FAQ. Make. http://makezine.com/magazine/make-
ultimate-guide-to-3d-printing/3d-printing-faqs/, (accessed 12.07.13.).
Zitting, A., Savolainen, H., 1980. Effects of single and repeated exposures to thermo-
oxidative degradation products of poly(acrylonitrile-butadiene-styrene) (ABS)
on rat lung, liver, kidney, and brain. Archives of Toxicology 46, 295e304.
B. Stephens et al. / Atmospheric Environment 79 (2013) 334e339 339