3D-Printing inside the Glovebox: A Versatile Tool for Inert-Gas Chemistry Combined with Spectroscopy

Article abstract of DOI:10.1002/hlca.201500502

3D-Printing with the well-established 'Fused Deposition Modeling' technology was used to print totally gas-tight reaction vessels, combined with printed cuvettes, inside the inert-gas atmosphere of a glovebox. During pauses of the print, the reaction flasks out of acrylonitrile butadiene styrene were filled with various reactants. After the basic test reactions to proof the oxygen tightness and investigations of the influence of printing within an inert-gas atmosphere, scope and limitations of the method are presented by syntheses of new compounds with highly reactive reagents, such as trimethylaluminium, and reaction monitoring via UV/VIS, IR, and NMR spectroscopy. The applicable temperature range, the choice of solvents, the reaction times, and the analytical methods have been investigated in detail. A set of reaction flasks is presented, which allow routine inert-gas syntheses and combined spectroscopy without modifications of the glovebox, the 3D-printer, or the spectrometers. Overall, this demonstrates the potential of 3D-printed reaction cuvettes to become a complementary standard method in inert-gas chemistry.

Full text of DOI:10.1002/hlca.201500502

Helv. Chim. Acta 2016, 99, 255 – 266
255
FULL PAPER
3D-Printing inside the Glovebox: A Versatile Tool for Inert-Gas Chemistry Combined
with Spectroscopy
€
by Felix Lederle, Christian Kaldun, Jan C. Namyslo, and Eike G. Hubner*
Institute of Organic Chemistry, Clausthal University of Technology, Leibnizstr. 6, DE-38678 Clausthal-Zellerfeld
(phone: +49-5323-723834; e-mail: eike.huebner@tu-clausthal.de)
€
Dedicated to Prof. Dr. rer. nat. Dr. h.c. mult. Hansjorg Sinn on the occasion of his 86th birthday
3D-Printing with the well-established ‘Fused Deposition Modeling’ technology was used to print totally gas-tight reaction
vessels, combined with printed cuvettes, inside the inert-gas atmosphere of a glovebox. During pauses of the print, the
reaction flasks out of acrylonitrile butadiene styrene were filled with various reactants. After the basic test reactions to proof
the oxygen tightness and investigations of the influence of printing within an inert-gas atmosphere, scope and limitations of
the method are presented by syntheses of new compounds with highly reactive reagents, such as trimethylaluminium, and
reaction monitoring via UV/VIS, IR, and NMR spectroscopy. The applicable temperature range, the choice of solvents, the
reaction times, and the analytical methods have been investigated in detail. A set of reaction flasks is presented, which allow
routine inert-gas syntheses and combined spectroscopy without modifications of the glovebox, the 3D-printer, or the
spectrometers. Overall, this demonstrates the potential of 3D-printed reaction cuvettes to become a complementary standard
method in inert-gas chemistry.
Keywords: 3D-Printing, Online analytics, Inert-gas synthesis, Trimethylaluminium, NMR Spectroscopy.
allowed to distinguish between two different products
yielded from the same starting materials by the geometry
Introduction
3D-Printing has found its way into daily life and evolved
from more experimental and specialized setups towards
the mass market and is used for efficient production set-
ups. Besides spectacular demonstrations of 3D-printing
(ranging from food to concrete), the main impact of
of the printed reactor [3][4]. Closely packed tubes printed
out of PP were used for the hydrothermal syntheses of
metal–organic frameworks [5][6]. The thin channels in a
matrix of PP or PET were used as flow reactors for the
synthesis and further analytical applications of different
3D-printers on our daily life probably is their rapidly nanoparticles [7 – 9]. A combination of PP and silicones
increasing dissemination (beginning with the availability
in market stores up to the use in small-scale productions)
nowadays [1][2]. Most notably, printers based on ‘Fused
Deposition Modeling’ (FDM), which is realized by build-
was used to print ‘reaction cubes’ in which multistep
reactions were realized by overturning the appropriate
sides of the cube to mix the desired starting materials
step-by-step, while catalysts were embedded into the
ing up the desired objects layer by layer of molten poly- silicon matrix [10].
mer, have arisen as most cost-efficient and fast
The concept for inert-gas chemistry within 3D-printed
technology. They are available in the price range around reaction vessels and cuvettes presented here is based on a
$1000. Primarily, acrylonitrile butadiene styrene (ABS)
and polylactide (PLA) are used as printing material, of
which ABS shows excellent mechanical properties. The
commercially available 3D-printer without any modifica-
tions. The 3D-printer is ready for operation just after
insertion into a glovebox, which is not modified either.
use of various other polymers, such as polyethylene The set of reaction flasks presented here (Fig. 1) allows
terephthalate (PET), polyamides, and the chemically inert for inert-gas chemistry at elevated temperatures by main-
polypropylene (PP) as well as silicones is increasing,
although more complex printing setups may be necessary, glovebox. Accompanying measurements of UV/VIS, IR,
as well as reduced mechanical stability may be
and NMR spectra are possible without opening the reac-
achieved. Consequently, 3D-printing has successfully been tion flasks. Additionally, after completion of the reaction,
taining the nitrogen atmosphere after removal from the
a
applied for the construction of chemical reactions vessels.
For example, microreactors printed out of silicone
pure products can easily be isolated without contamina-
tion from the ABS matrix material.
DOI: 10.1002/hlca.201500502 © 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.
This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium,
provided the original work is properly cited.

256
Helv. Chim. Acta 2016, 99, 255 – 266
polymers after printing was found to be around 200 ppm
(0.028 ꢁ 0.004% for ABS-K and 0.024 ꢁ 0.005% for
ABS-N) and therefore the amount of water released from
the inner surface of printed reaction vessels can be con-
sidered to be extremely low and does not interfere with
the use of highly reactive reagents.
Tensile tests have been performed to show the influ-
ence of printing within an inert-gas atmosphere. Printed
plates made of ABS-N showed a significant increase in
the elongation at break from eB(air) = 4.4 ꢁ 1.5% to
eB(N2) = 10.7 ꢁ 2.6% which is an indication for a signifi-
cantly improved layer adhesion if printed within an
inert-gas atmosphere. This is an interesting result for 3D-
printing in general, and furthermore, an important aspect
for the impermeability of printed reactions flasks (Fig. 2).
To get an idea of the impermeability of the printed
flasks against oxygen and H2O, the intrusion was roughly
estimated by the surface area (200 cm2), the wall thick-
ness (3 mm), and the permeability of ABS against oxygen
Fig. 1. 3D-Printed reaction flasks, cuvettes, and objects. From left:
round-bottom flask made of ABS-N (F1), conical flask made of ABS-N
(F2), UV cuvette made of ABS-K (F3), inert-gas flask with UV cuvette
made of ABS-K (F4), IR cuvette made of ABS-N (F5), inert-gas flask
with IR cuvette made of ABS-N (F6), NMR tube made of ABS-N
(F7), and inert-gas flask/spinner with NMR tube made of ABS-N (F8).
Results and Discussion
For all prints, an appropriate sized FDM-3D-printer was
used, which was inserted into the glovebox via the vac- (approx. 80 ml mm/m2/day/atm O2) and water (approx.
uum chamber over night. To communicate with the prin- 8 g mm/m2/day H2O) to 5 lmol O2/day and 3 lmol H2O/
ter inside the glovebox, the USB signal was transmitted
day [11]. Although imperfect printed walls and imperfect
via ethernet and the ethernet signal was passed into the layer adhesion are neglected in this estimation, it is a
glovebox via PowerLAN using the existing power supply promising result for gas-tight reaction flasks. For first tests
of the glovebox, so no additional wiring was necessary. of the leak tightness, the flasks F1 and F2 (Fig. 1) were
After insertion into the inert-gas atmosphere, the general filled with water and isooctane (containing rhodamine
aspects of printing inside an oxygen-free atmosphere were
investigated. The polymer used for all prints was ABS
due to its extraordinary mechanical stability. In general,
ABS is a graft copolymer of styrene and acrylonitrile
(SAN) on polybutadiene. As a result of phase separation
and rubrene, respectively, as dyes) during short pauses of
the print. The flasks were cooled down to ꢀ40 °C and
heated up to 95 °C for 24 h subsequently, and no weight
loss (indicating a leakage) was noticeable. Since the print-
ing of the round upper shell of the flask F1 was found to
(phases of different refractive indices) it is usually turbid be challenging, all following experiments have been per-
[11]. With the help of ‘contrast matching’ via the addition
of acrylates, this effect may be avoided and a clear poly-
mer is obtained. Consequently, such a polymer (ABS-K)
was used to print the UV/VIS cuvettes. All other flasks
were printed out of an unmodified ABS (ABS-N) to keep
formed in flask F2, whose conical closing can be printed
precisely and reproducibly. The dissolution properties of
ABS allow for the usage of strong polar solvents, such as
water and lower alcohols, as well as strong nonpolar sol-
vents, such as n-alkanes. To check the possibility of mea-
the number of reactive residues within the matrix polymer suring accurate spectra directly in printed reaction flasks,
as low as possible. Additionally, the C=O band of the UV/VIS, IR, and NMR spectra of pure substances have
acrylates would block a significant region of the IR spec-
trum and ABS-N was the choice for IR cuvettes, too. The
been measured before setting up reactions inside the
glovebox. All cuvettes have been dimensioned according
glass transition temperature (Tg) of printed stripes of both to their glass or metal counterparts and directly fit into
polymers was determined via differential scanning standard spectrometers (Fig. 1).
calorimetry to 98 °C (ABS-K) and 111 °C (ABS-N) and
therefore the temperature range for heating up the
The UV/VIS spectrum of the empty cuvette F3 made
of ABS-K (vs. air) shows a strongly increasing absorption
printed reaction vessels is limited to roughly 100 °C. Sam- above 350 nm and consequently UV/VIS spectra can be
ples of both polymers were dissolved in acetone, the sol-
uble part isolated and the molecular weight of the matrix
material determined to Mn = 55,000 g/mol (ABS-K) with
measured in the range from 900 – 350 nm. The measure-
ment of a rhodamine B solution in the printed cuvette F3
almost perfectly matches the spectrum measured in the
quartz glass cuvette (Fig. 3a) and the extinction coeffi-
cient was determined to e = 1.26 9 105 l/mol/cm at
kmax = 545 nm, which is close to the specifications of the
a
molecular weight distribution of 2.08 and
Mn = 67,000 g/mol (ABS-N) with a molecular weight dis-
tribution of 2.25. The presence of C=O groups was
checked via ATR-IR and they were found at 1730 cmꢀ1 manufacturer. The excitation–emission matrix of a plate
in case of ABS-K, while no absorption in this region was
noticeable for ABS-N. Finally, a set of printed stripes was
of ABS-K shows a weak fluorescence beginning at an
excitation wavelength of kexc = 410 nm with a maximum
analyzed via Karl–Fischer titration with regard to the at kexc = 380 nm. Consequently, the emission maximum
total residual water content. The water content of both
of a 0.001 mM solution of rhodamine B in water was cor-
www.helv.wiley.com
© 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.

Helv. Chim. Acta 2016, 99, 255 – 266
257
Fig. 2. Load–strain curves of plates made of ABS-N printed within a N2 atmosphere (solid line) and at air (dashed line). Elastic modulus:
E(air) = 0.87 ꢁ 0.03 kN/mm2, E(N2) = 0.85 ꢁ 0.04 kN/mm2; tensile strength: rM(air) = 19.4 ꢁ 0.8 N/mm2, rM(N2) = 21.6 ꢁ 0.2 N/mm2.
rectly determined to kem = 579 nm (kexc = 545 nm) in tion for oxygen. The reaction flask F4 with an attached
absolute accordance with the measurement in a quartz
glass cuvette. The detected fluorescence intensity was low-
ered by a factor of 5 compared to the glass cuvette. The
ATR-IR spectra could be recorded in the range from
and printed UV cuvette was printed within the glovebox
and filled with an approx. 0.15M solution of pyrogallol (1)
in 0.65M aqueous NaOH during a short pause. After fin-
ishing the print, flask F4 was discharged out of the glove-
1500 – 2800 cmꢀ1 and > 3100 cmꢀ1 using cuvette F5 box and UV/VIS spectra were recorded on a daily basis.
printed out of ABS-N, which is in accordance with the In a basic solution, pyrogallol decomposes extremely fast
absorption spectrum of the pure polymer (Fig. 3b). The into intensively colored reaction products in the presence
C=O band of acetoneꢀ1in nonpolar pentane was deter-
of oxygen. Therefore, it is well-known and used as a
quantitative and qualitative test reaction for oxygen
~
mined to v = 1735 cm
which is a hypsochromic shift
compared to the measurement of the pure substance (Fig. 4a) [15][16].
(v = 1712 cmꢀ1) and is in good accordance with results
reported in the literature for diluted solutions in nonpolar
solvents [12]. The OH-stretching vibratiꢀo1n of MeOH in
The absorption at 420 nm remains unchanged during
6 days (Fig. 4b), and consequently, flask F4 shows an
impressive impermeability against oxygen. After punctur-
~
~
pentane was determined to v = 3390 cm in accordance ing the flask to allow air to insert via a small hole, an
with literature results again, and is dominated by H-
bridges formed at the given concentration [13]. Finally,
intense brown color of the solution was observed within
few minutes which was reflected in the fast increase in
the 13C-NMR spectrum of pure ABS-N dissolved in the absorption at 420 nm.
CDCl3 shows only comparably weak and narrow signals
close to 130 ppm (primarily aromatic carbons) and at
25 – 45 ppm (aliphatic carbons). Consequently, it was
After the reaction flask has proven to be sufficiently
oxygen tight, it was used to monitor the equilibrium
between copper(I) and copper(II). This allows to check
possible to record a very neat spectrum of CD3OD in the the usability for inorganic aqueous reactions and is an
1
printed NMR tube F7 and the JC,D coupling constant additional proof for the impermeability. In the presence
was correctly determined to 20 Hz (Fig. 3c) [14]. of non-chelating ligands, the equilibrium between copper
After the general measurements of spectra have (I) and copper(II) is located strongly on the side of copper
proven to be possible, the actual impermeability against (I) (~3 mol-% Cu(II) in the presence of 1,5-diaminopentane)
oxygen was checked via an UV spectrometric test reac-
(Fig. 4a) [17][18]. Reaction flask F4 was filled with CuCl
© 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.
www.helv.wiley.com

258
Helv. Chim. Acta 2016, 99, 255 – 266
Fig. 3. Spectroscopy in 3D-printes cuvettes. a) UV/VIS Spectra of rhodamine B (0.01 mM) in MeOH (vs. MeOH background) in glass cuvettes
(dashed line) and cuvettes F3 (solid line). b) IR Spectra of pentane (bottom), 5 wt% acetone in pentane (middle), and 5 wt% EtOH in pentane (top)
(vs. pentane background in all cases) in cuvette F5. c) 13C-NMR Spectrum of ABS-N in CDCl3 (bottom) and CD3OD in NMR tube F7.
and a solution of 1,5-diaminopentane in H2O during two strated the UV/VIS spectroscopic opportunities and inert-
short pauses of the print and UV/VIS spectra were ness of the flasks against various reagents, the synthetic
recorded on a daily basis (Fig. 4b). A slight indication for potential of printed reaction vessels in the presence of
a disproportionation at the first day was noticeable by a
slight increase in the characteristic absorption at 600 nm
and by the visible formation of tiny amounts of elemental
copper. During the following days, equilibration (proba-
bly by reaction of the copper(II) with dispersed elemental
copper) took place and the absorption at 600 nm
remained unchanged. After puncturing the flask, an
immediate occurrence of a deep blue color indicated the
irreversible formation of copper(II) within a few minutes
(Fig. 4b). Based on these experiments, which demon-
highly reactive starting materials was evaluated. Conse-
quently, the reductive methylation of alcohols and
carboxylic acids by trimethylaluminum (AlMe3) was
investigated (Scheme 1) [19][20].
In contrast to literature, octane was chosen as solvent
and therefore 4-fluorobenzoic acid instead of benzoic acid
was used as catalyst, as it shows a good solubility in
octane, the F–aryl bond tolerates AlMe3 and only compa-
rably low-boiling products are formed out of the catalyst
[21]. The known reduction of triphenylmethanol (2a) was
www.helv.wiley.com
© 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.

Helv. Chim. Acta 2016, 99, 255 – 266
259
Fig. 4. Inert-gas chemistry combined with UV/VIS spectroscopy performed in reaction cuvette F4. a) Top: Test Reaction of pyrogallol with oxy-
gen under basic conditions. Bottom: Disproportionation of copper(I) in aqueous solution in the presence of the non-chelating ligand 1,5-diamino-
pentane and oxidation of copper(I) to copper(II) in the presence of oxygen. b) Plot of the absorption at 420 nm vs. initial absorption at 420 nm of
the pyrogallol solution (black) and plot of the absorption at 600 nm vs. initial absorption at 600 nm of the CuCl reaction mixture (gray). Solid
lines are a guide for the eye.
performed in the reaction cuvette F6, which was filled
with the starting material, the catalyst, and a solution of
AlMe3 in octane during two short pauses. After discharg-
showed a significant shift of the C–O C-atom from
d = 75.2 – 81.8 ppm in the 13C-NMR spectrum compared
to the starting material. While it was not possible to
ing from the glovebox, the flask was heated to 85 °C for detect a signal for aluminum via 27Al-NMR after acid
6 days, and after opening the reaction vessel (and aque-
ous workup), pure triphenylethane (3a) was received in be determined via ICP-OES to 10.19 wt%. This is in
69% yield. Based on this promising result, the reduction accordance with the dimeric dimethylaluminum com-
digestion with H2SO4/H2O2, the aluminum content could
of the aromatic alcohol 2b under analogous conditions in pound 3d. The characteristic strong Al–C vibrational band
the reaction flask F2 was tested and yielded the neat
reduction product 3b as new and fully characterized com-
pound. Owed to the limited reaction temperature of
about 100 °C, a reduction of the tertiary aliphatic alcohols
2c and 2d could not be observed. Instead, the reaction of
(+)-cedrol (2d) with AlMe3 using the flask F2 and aque-
ous workup afforded an air-stable compound, which
was detected at 690 cmꢀ1 in the ATR-IR spectrum [22]
and the Al–CH3 Me protons were detected at ꢀ2.9 ppm
1
((D8)toluene, 70 °C) in the H-NMR spectra as a unique
singlet. High-resolution mass spectra doubtlessly con-
firmed the structure of 3d by the mass peak of
[M – Me]+, which is typical for compounds of the type
[Me2Al-(l-OR)]2
[23].
Furthermore,
the
known
© 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.
www.helv.wiley.com

260
Helv. Chim. Acta 2016, 99, 255 – 266
Scheme 1. Reaction products obtained by reduction of alcohols and carboxylic acids with trimethylaluminium in flasks F2 and F6.
dimethylaluminum compound 3c was prepared for com-
parison and has been characterized as the assumed dimer
via mass spectrometry [20]. The extraordinary stability of
serves as a convenient experimental system to be exam-
ined via the C=O vibrational band. The reduction of the
carboxylic acid to the corresponding tert-butyl group
the new enantiopure dimethylaluminum compound 3d is (dtBu = 1.41 ppm), expected according to literature, could
worth to be pointed out, since the Al–CH3 Me groups not not be observed at the comparably low reaction tempera-
only withstand the aqueous workup and subsequent atmo-
spheric water and oxygen for a longer period of time, but
also treatment with semiconcentrated HCl. The dimethy-
ture of 85 °C. Furthermore, the formation of the ace-
tophenone derivative (dMe = 2.49 ppm) could be excluded
according to the H-NMR spectra of the isolated reaction
1
laluminum species 3d showed to be significantly more products after opening the flask (Scheme 1) [19][24 – 27].
stable than the comparable compound 3c. Consequently,
it is currently under investigation as Me transfer agent
toward transition metals.
Instead, the tertiary alcohol 3e was isolated as main pro-
duct; consequently lowering the process temperature to
85 °C can be considered as a convenient method to
reduce carboxylic acids to the corresponding tertiary
The printed reaction cuvette F6 offers the opportunity
to monitor the reduction process with highly reactive alcohol. Due to the long reaction time of 6 days (which
AlMe3 via IR spectroscopy without the need to open the has not been optimized to increase the yield of alcohol), a
flask. Obviously, the reduction of the carboxylic acid 2e slow elimination to yield the alkene 4e was noticeable. The
www.helv.wiley.com
© 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.

Helv. Chim. Acta 2016, 99, 255 – 266
261
main interesting aspect of the reduction process was the
fast disappearance of the C=O vibrational band in the IR
spectra, which was nearly undetectable directly after dis-
solvent and subsequently with triethylamine during two
short pauses of the print. After discharging flask F8
from the glovebox and the measurement of initial NMR
charging flask F6 from the glovebox and before heating up spectra, the flask was heated in an aluminum bead bath
the reaction mixture. All together, this leads to the rather (to prevent contamination of the flask with silicon oil)
detailed reaction sequence presented in Scheme 2: After to 85 °C (internal temperature 70 °C) and NMR spectra
the instantaneous deprotonation of the carboxylic acid
(which is backed by the absence of OH vibrational bands in
the IR spectrum measured directly after discharging from
the glovebox), a fast transfer of the first Me group can be
assumed, which is indicated by the immediate loss of the
were recorded on a regular basis. With TMSCl, no indi-
cation of enolisation was noticeable after several days of
reaction time, which is in accordance with the compara-
bly low reactivity [31]. In contrast, by treatment of 5
with TMSBr, which is more reactive by a factor of ~105,
C=O vibrational band. Since the acetophenone derivative after a few hours at 70 °C, the characteristic olefinic sig-
could not be isolated, the second Me group has to be trans- nals of 6 at d = 150.9 (C(1)) and 103.4 ppm (C(2)) as
ferred comparably fast, too. This allows the alcohol 3e to well as the set of signals of the aliphatic C-atoms at
be isolated after aqueous workup. Slowly, the elimination
of the aluminum alcoholate to the isolated alkene 4e takes clearly be detected in the 13C-NMR spectra. The con-
place, while the low temperature prevents formation of the sumption of the starting material 5 was monitored via
d = 24.4 (C(3)), 23.8 (C(4)) and 23.1 ppm (C(5)) could
alkane [28][29].
the decrease in the signal of the C-atoms at d = 42.1
(C(2)/C(20)) and 26.0 ppm (C(4)). Since the silyl ether
formation proceeds via a rate-determining pre-equilibrium
of TMSBr, NEt3, and the ketone to an intermediate,
After having proven the practicability of printed flasks
for monitoring inert-gas reactions with UV/VIS and IR
spectroscopy, NMR spectroscopy was investigated as
more challenging analytical technology. To maintain the which decomposes to the enolised product following a
possibility of larger scaled reactions inside the printed
reactions vessels as synthetical approach, the typical spin-
ner of a NMR setup was designed as hollow reaction
first-order rate law, it was possible to determine the
reaction rate constant to k = 9.4 ꢁ 0.2 9 10ꢀ5 sꢀ1
(70 °C, octane) (Fig. 5a). Compared to known results in
chamber. The directly attached and printed NMR tube literature, this nicely matches the expected range [31].
was used for the measurement, while the whole spinner/
Additionally, 29Si-NMR spectra were recorded in the sil-
tube combination fits into an unmodified NMR spectrom- icon-free NMR tube/spinner combination. With the help
eter. The reaction flask F8 was used to follow the enolisa- of comparative NMR measurements of the treatment of
tion of cyclohexanone (5) toward the silyl ether 6 (Fig. 5) TMSCl with EtOH and H2O in (D12)cyclohexane, the
by 13C- and 29Si-NMR spectroscopy [30][31].
F8 was filled with a solution of approx. 0.1 g of the
29Si-NMR signals were doubtlessly assigned (dSi = 24.3
(Me3SiBr), 14.0 (6) and 6.8 ppm Me3Si–O–SiMe3). The
starting material 5 and trimethylsilyl chloride (TMSCl) NMR spectra of the reaction mixture in flask F8
or bromide (TMSBr), respectively, in (D18)octane as a
revealed a minor amount of hydrolysis (~15 mol % of
the initial TMSBr) after 300 min parallel to the product
formation. After total consumption of the starting mate-
rial 5 was achieved, flask F8 was opened and a control
measurement recorded in a conventional NMR tube
(Fig. 5b).
Scheme 2. Stepwise reduction of the carboxylic acid 2e with AlMe3
and reaction mechanism discussed on the basis of isolated products
and IR spectroscopic measurements in reaction flask F6.
After examining the different analytical methods in
3D-printed reaction cuvettes and their stability against
highly reactive reagents, the choice of suitable solvents
was finally investigated by the controlled atom transfer
radical polymerization (ATRP) of styrene in the reaction
flask F2 [32][33]. Besides the necessary absence of oxy-
gen, the rather similar solubility of the reaction product,
polystyrene (PS), and the reaction flask printed out of
ABS is the challenging aspect in this case. (R)-Limonene
is known to dissolve PS, but not ABS [34]. Nevertheless,
it cannot be used as reaction medium in this case, since it
is known to polymerize by a radical mechanism itself [35].
Fortunately, the structurally related methylcyclohexane
and cyclohexane turned out to be suited as solvents as
well according to solvation tests of small stripes of ABS-N
at 80 °C. Both solvents are just at the edge to a non-sol-
vent for polystyrene with an upper critical solution tem-
perature (UCST) of approx. 72 °C (methylcyclohexane)
© 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.
www.helv.wiley.com

262
Helv. Chim. Acta 2016, 99, 255 – 266
Fig. 5. Reaction of cyclohexanone (5) with TMSBr. a) Determination of the reaction rate constant k. Error bars are calculated from uncertain-
ties of the integral values in the NMR spectra. b) 13C-NMR Spectra measured in flask F8 directly after printing and discharging the flask from the
glovebox (bottom), and after 120, 210, and 300 min as well as control measurement of the isolated reaction solution in a glass NMR tube (top).
and 35 °C (cyclohexane), respectively [36]. To minimize to 70 °C (internal temperature) for 21 and 56 h, respec-
the risk of falling below the theta temperature from the
start of the polymerization until precipitation of the poly-
mer, cyclohexane was chosen as solvent and the catalytic
system adapted to these highly nonpolar reaction condi-
tions (Fig. 6).
tively, after discharging from the glovebox. Subsequently,
the reaction vessel was opened and without allowing it
to cool down, the dark red solution was poured into
MeOH and the polymer precipitated. The ATRP
proceeded well under the nonpolar reaction conditions
Flask F2 was filled with CuBr and the ligand 4,40- and an excellently narrow molecular weight distribution
dinonyl-2,20-bipyridine during a first and a solution of
was achieved (Fig. 6b), while with increasing reaction
styrene and the initiator dodecyl 2-bromisobutyrate in time the molar mass of the polystyrene increased as well.
cyclohexane during a second pause. Flask F2 was heated The size-exclusion chromatography traces did not show
www.helv.wiley.com
© 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.

Helv. Chim. Acta 2016, 99, 255 – 266
263
Fig. 6. ATRP of styrene in cyclohexane in flask F2. a) Reaction conditions adapted to the nonpolar medium. b) Molecular weight distribution
achieved. Ratio [cyclohexane]:[styrene]:[ligand]:[Cu(I)]:[initiator]: 1000:100:2:1:0.2. Polymerization time: 21 h (P1, solid line) and 56 h (P2, dashed line).
the slightest indication of impurities dissolved from the cuvettes were designed, which allow to perform chemical
reaction flasks. The molar masses achieved are rather reactions under inert-gas atmosphere with highly reactive
low, as expected, although the reaction times chosen
were not too short. This correlates with a slow polymer-
ization process and can be explained by the fact of a
reagents in a strongly polar or nonpolar reaction medium
with accompanying analytical measurements via UV/VIS,
IR, and NMR spectroscopy. For example, the reduction
polymerization in a solution containing only approx. with AlMe3 performed in printed reaction cuvettes, which
10 mol-% of monomer.
led to new compounds and revealed details about the
reaction mechanism, demonstrates the usefulness of the
method. The concept presented here can easily be
extended to further analytical methods and does not
require any modifications of the 3D-printer, the glovebox,
or the analytical instruments. The transfer of all necessary
equipment inside the glovebox is realized within one
night, and reaction flasks and cuvettes are printed within
These experiments illustrate the limit of suitable sol-
vents for flasks and cuvettes printed out of the ABS
matrix. Additionally, they demonstrate the impermeability
of the reactions flasks at higher temperatures, hence the
diffusion of just ~0.1 mmol O2 into the flask would have
been sufficient to oxidize the total amount of Cu(I) to
Cu(II), and terminate the polymerization.
a
few hours, allowing the technique to become a
complementary routine method for inert-gas chemistry.
Currently, the extension towards further polymers available
for 3D-printing is investigated, to enlarge the spectrum of
suitable solvents as well as the use of overpressure.
Conclusions
In conclusion, 3D-printing within an inert-gas atmosphere
was investigated and a set of gas-tight reaction flasks and
© 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.
www.helv.wiley.com

264
Helv. Chim. Acta 2016, 99, 255 – 266
We thank G. P. Brunotte and Dr. L. Steuernagel of the and only the middle fraction (approx. 80%) was used. Sil-
Institute of Polymer Materials and Plastics Engineering ica gel 60 was used for column chromatography (CC).
€
for the tensile tests, M. Heinz and U. Kocher of the Insti- UV/VIS Spectra: Jasco V 640 spectrophotometer (Jasco
tute of Technical Chemistry for SEC and DSC measure-
ments, PD Dr. J. Adams of the Institute of Physical
Chemistry for fluorescence measurements and P. Sommer
of the Institute of Mineral and Waste Processing, Waste
Germany GmbH, Gross-Umstadt, Germany). Fluores-
cence spectra: Jasco FP-8500 fluorescence spectrometer
(Jasco Germany GmbH). If not noted differently, spectra
were recorded with a speed of 100 nm/min. IR Spectra:
Disposal and Geomechanics for the ICP-OES analyses. Bruker Alpha-T FT-IR spectrometer (Bruker Coporation,
€
We thank Dr. G. Drager of the Institute of Organic
Chemistry of the Leibniz University Hannover and
Billerica, Massachusetts, USA). If not noted differently,
16 spectra were accumulated per measurement. For ATR
measurements, a Platinum diamond-ATR unit was used
Dr. H. Frauendorf of the Institute of Organic and Biomole-
€
for the measurement of the high-resolution mass spectra.
cular Chemistry of the Georg-August University Gottingen (Bruker Corporation). NMR Spectra: Bruker Avance 400
(Bruker Corporation) (400 MHz (1H), 100 MHz (13C)
and Avance III 600 (600 MHz (1H), 150 MHz (13C) FT-
NMR spectrometer. Chemical shifts are given in ppm rel-
ative to tetramethylsilane (d = 0.0) or the residual solvent
signal of the deuterated solvent. Mass spectra: Varian 320
MS TQ mass spectrometer at 20 eV and 70 eV for EI
mass spectra (Agilent Technologies, Inc., Santa Clara, CA,
USA). Additionally, CI ionization with CH4 was applied.
High-resolution mass spectra were measured externally
Supporting Information
Additional Supporting Information may be found in the
online version of this article:
Appendix S1. 3D-Printing inside the glovebox: A versa-
tile tool for inert gas chemistry combined with spectro-
scopy.
€
by Dr. Gerald Drager at the Institute of Organic
Chemistry of the Leibniz University Hannover and Dr.
Holm Frauendorf at the Institute of Organic and
Biomolecular Chemistry of the Georg August University
Experimental Part
Detailed information of the 3D-printing setup and the
3D-printing procedure, as well as additional spectroscopic
measurements, further analytics and detailed procedures
for experimental setups monitored by UV/VIS, NMR,
and IR spectroscopy are given in the Appendix S1. STL
files of all presented reaction flasks and cuvettes are
gladly provided on request.
€
Gottingen. SEC measurements were performed with a
setup equipped with a Waters 515 HPLC pump (Waters
GmbH, Eschborn, Germany), Knauer Smartline RI detec-
¨
tor 2300 (Knauer Wissenschaftliche Gerate GmbH, Berlin,
ꢀ
Germany) and 2 9 5 lm mixed-C and 1 PLgel 1000 A
column from Polymer Laboratories (Agilent Technologies,
Inc.). THF with a flow rate of 1 ml/min at 25 °C was used
as eluent. Molecular weights are given relative to
polystyrene calibration. Glass transition temperatures
have been measured with a Mettler-Toledo DSC-1 appara-
tus under N2 with a heating speed of 10 K/min (Mettler-
Toledo GmbH, Gießen, Germany). ICP-OES analyses
have been performed on a Varian Vista MPX instrument
in 0.5 vol % HNO3 (Agilent Technologies, Inc.). Samples
have been digested as follows. The pure substance (ap-
prox. 5 mg) has been weighed in exactly, dissolved in
conc. H2SO4, and heated 2 h at 140 °C. The solution
turned dark brown. After addition of 1.5 ml H2O2 (30%),
General
All operations with air sensitive compounds were carried
out at a high-vacuum line (< 10ꢀ6 mbar) using Schlenk
techniques and in a glovebox (M. Braun, Labmaster 130,
M. Braun Inertgas-Systeme GmbH, Garching, Germany)
under N2. If not noted differently, all chemicals
were bought from Sigma–Aldrich (St. Louis, Missouri,
USA) and used as received. Cyclohexane and octane
(Acros Organics, Thermo-Fisher Scientific, Geel, Belgium)
were purchased dry (< 0.002% H2O) and used as
received. H2O was degassed three times before use. All
solids used for air-sensitive operations were degassed for
one night at high in vacuo and transferred via flasks
equipped with Teflon stopcocks into the glovebox. AlMe3
(Witco GmbH, Bergkamen, Germany) was used as
received. 1,5-Diaminopentane was degassed three times
before use. Styrene was degassed and dried two times
over CaH2 and distilled before use. (D18)octane (Deutero
GmbH, Kastellaun, Germany) was degassed and dried
with BuLi and distilled before use. Triethylamine was
dried two times over CaH2 and distilled before use.
Cyclohexanone was dried over CaH2 and distilled before
use. Bromotrimethylsilane was distilled at the vacuum line
heating was continued for 12 h.
A
quantity of
0.5 ml H2O2 (30%) was added to the nearly clear solu-
tion, and heating was continued for 2 h. The totally color-
less solution was cooled down and diluted with 0.5 vol %
HNO3 to exactly 25.0 ml. Tensile tests were performed
with a Zwick/Roell BZ1-MM14450.ZW05 (Zwick GmbH
& Co. KG, Ulm, Germany) universal testing machine
equipped with a 10 kN load cell and with a speed of
1 mm/min at 23 °C/50% rel. humidity. A Heraeus Labo-
fuge 400R at 20 °C was used for centrifugation (Thermo
Fisher Scientific Messtechnik GmbH, Oberhausen, Ger-
many). The residual H2O content of the polymer samples
was measured externally via Karl–Fischer titration of the
condensed residue emitted from molten polymer samples
www.helv.wiley.com
© 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.

Helv. Chim. Acta 2016, 99, 255 – 266
265
at 150 °C under N2 flow by the ‘Mikroanalytisches Labor TMA solution (5.743 g, 2.1 mmol). The crude product
Pascher’ in Remagen, Germany.
was washed with acetonitrile and after removal of the sol-
vent [Me2Al(l-O-cedrane)]2 (3d) was obtained. Yield:
61 mg (28%). White solid. M.p. 152 – 154°. IR (ATR):
3008w, 2947w, 2931w, 2912w, 2883w, 1456w (br.), 1381w,
1198m, 1112w, 1044w, 922m, 899s, 856w, 796w, 753m,
690s (Al–C), 671m, 643m, 566w. 1H-NMR (600 MHz,
CDCl3): ꢀ0.63 (br. s, 4 Me(13)); 0.83 (d, J = 7.1, 2 Me
General Procedure for Reductions with
Trimethylaluminium
The reaction vessel combined with an IR cuvette (flask
F6) or the reaction flask F2 was printed with natural
ABS-N. The solid starting materials (5 mol-% p-fluoro- (10)); 0.98 (s, 2 Me(12b)); 1.26 (dddd, J = 11.8, J = 8.6,
benzoic acid as catalyst and approx. 50 – 100 mg of the 7.7, 6.0, CH2(1); 1.31 (s, 2 Me(12a)); 1.33 – 1.42 (m,
corresponding alcohol or carboxylic acid) were inserted CH2(2,4,5)); 1.40 (s, 2 Me(11)); 1.45 (dddd, J = 13.1, 6.4,
during a first pause of the print. A soln. of AlMe3 2.3, 1.9, CH2(50)); 1.49 – 1.55 (m, CH2(40)); 1.60 (ddd,
(1.000 g, 13.9 mmol) in dry octane (37.870 g) was pre-
pared in a separate flask and used for all reductions. The
appropriate amount (5 equiv. AlMe3 in case of alcohols, 8
equiv. in case of carboxylic acids) of the TMA stock soln.
J = 12.3, 5.5, 2.9, CH2(20)); 1.64 (q, J = 7.0, 2 H, H–C(3));
1.73 (dd, J = 8.1, 2 H, H–C(8a)); 1.74 (dd, J = 8.1, 2 H,
H–C(7)); 1.88 (dd, J = 11.8, 5.5, CH2(10)); 1.88 (dd,
J = 11.8, 8.1, CH2(9)); 2.12 (ddd, J = 11.8, 8.1, 8.0, 2
was added via a syringe during the second pause of the CH2(90)). 13C-NMR (150 MHz, CDCl3): ꢀ3.6 (C(13));
3D-print, and after cooling down flask F2 or F6 was dis-
charged out of the glovebox. Immediately after discharg-
ing, an initial IR spectrum was recorded after turning the
flask upside down when flask F6 was used. The reaction
15.6 (C(10)); 25.4 (C(5)); 26.9 (C(12b)); 28.5 (C(12a));
30.7 (C(11)); 31.5 (C(4)); 34.2 (C(9)); 37.1 (C(1)); 41.2
(C(3)); 41.4 (C(2)); 43.1 (C(8)); 52.9 (C(3a)); 57.9 (C(8a));
1
60.7 (C(7)); 81.8 (C(6)). H-NMR (600 MHz, (D8)toluene,
flask was heated to 85 °C using an aluminum bead heat- 70°): ꢀ0.32 (s, 4 Me(13)); 13C-NMR (150 MHz, (D8)-
ing bath and IR spectra were recorded every 2 days.
After 6 days, the reaction flask was opened with a saw
and emptied, the flask washed three times with heptane.
H2O (50 ml) was added to the combined org. solns. and
stirred for 5 min. After phase separation, the aq. phase
was extracted three times with heptane (3 9 20 ml), the
toluene, 70°): ꢀ2.9 (C(13)). CI-MS: 441 ([M – Me]+, 90),
337 (M – Me – cedrane]+, 15), 205 (cedrane+, 100). HR-
EI-MS: 541.4147 (C33H59Al2O+2, [M – Me]+; calc.
541.4151). ICP-OES: Al 10.19% (C34H62Al2O2; calc.
9.69%).
Details for the synthesis of 1,1,1-triphenylethane (3b),
combined org. phases were dried with Na2SO4, and the [Me2Al(l-O-adamantyl)]2 (3d) and the reduction of
solvent was removed under reduced pressure. For all IR
measurements, printed cuvette F5 (ABS-N) filled with
heptane was used as reference cell. Spectra were recorded
with a resolution of 2 cmꢀ1. Thirty-two spectra have been
accumulated per measurement and the final spectrum has
been averaged over 35 data points.
4-butylbenzoic acid are given in the Appendix S1.
Reaction of Cyclohexanone with Trimethylsilyl
Bromide
The reaction flask F8 combined with a printed NMR tube
1,4-Bis(1,1-diphenylethyl)benzene (3b). Following the was printed with natural ABS-N. A stock solution of dry
general procedure (flask F2) benzene-1,4-diylbis(diphenyl- cyclohexanone (122 mg, 1.2 mmol) in dry (D18)octane
methanol) (62 mg, 0.14 mmol, 0.28 mmol OH) and p- (1.602 g) was prepared separately. A solution of freshly
fluorobenzoic acid (2 mg, 5 mol-%) were treated with
TMA solution (6.508 g, 2.3 mmol). The crude product
distilled trimethylsilyl bromide (TMSBr) (258 mg,
1.7 mmol, 3.2 eq.) in (D18)octane (738 mg) was mixed
was purified by column chromatography (Rf (PE/EE 2:1) with the cyclohexanone stock solution (733 mg,
0.77) and after removal of the solvent 1,4-bis(1,1-diphenyl- 0.5 mmol). The combined clear solution was inserted to
ethyl)benzene (3b) was obtained. Yield: 27 mg (44%). flask F8 via a syringe during a first pause of the print.
White solid. M.p. 222 – 224°. IR (ATR): 3083w, 3057w, Dry triethylamine (184 mg, 1.8 mmol, 3.4 eq.) was added
3018w, 2975w, 2935w, 2876w, 1596w, 1489w, 1441w, via a microsyringe during a second pause of the print.
1023w, 1014w, 860w, 811w, 756m, 697s, 671w, 628m, After finishing the print, flask F8 was discharged from the
1
583m, 534w, 522w. H-NMR (600 MHz, CD2Cl2): 2.16 (s, glovebox and carefully checked for irregularities. Initial
2 Me(1)); 6.99 (s, 4 H, H–C(8)); 7.09 – 7.16 (m, 8 H, H– NMR spectra (13C, 29Si) were recorded directly after dis-
C(4)); 7.17 – 7.21 (m, 4 H, H–C(6)); 7.22 – 7.31 (m, 8 H, charging and flask F8 was heated upside down at 85 °C in
H–C(5)). 13C-NMR (150 MHz, CD2Cl2): 30.8 (C(1)); 52.8 an aluminum bead bath. The flask was cooled down for
(C(2)); 126.5 (C(6)); 128.4 (C(5)); 128.7 (C(8)); 129.2
the repeated measurement of NMR spectra and heating
(C(4)); 147.2 (C(7)); 149.8 (C(3)). EI-MS (70 eV): 438 continued after each measurement. After completion of
(M+, 51), 423 ([M – Me]+, 100), 408 ([M – 2 Me]+, 10). the reaction, flask F8 was opened and the reaction solu-
HR-EI-MS: 438.2350 (C34H+30; calc. 438.2348).
tion transferred into a NMR tube for a final measure-
[Me2Al(l-O-cedrane)]2 (3d). Following the general ment. All NMR spectra have been recorded on the
procedure (flask F2) (+)-cedrol (89 mg, 0.4 mmol) and p- Bruker Avance III 600 spectrometer without spinning the
fluorobenzoic acid (3 mg, 5 mol-%) were treated with
spinner/tube combination (flask F8). For all acquisitions a
© 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.
www.helv.wiley.com

266
Helv. Chim. Acta 2016, 99, 255 – 266
[6] P. J. Kitson, R. J. Marshall, D. Long, R. S. Forgan, L. Cronin,
Angew. Chem., Int. Ed. 2014, 53, 12723.
deuterium lock on the (D18)octane signal has been
applied. Before each measurement, the automatic shim-
ming routine was performed. A flip angle of 30° (13C:
zgpg30, 29Si: zgig30) and a relaxation delay of D1 = 2 s
(13C), D1 = 0.2 s (29Si) has been used. A total of 512
measurements have been accumulated for each FID.
Exponential multiplication with a line broadening (lb) of
factor 2 has been applied. Spectra are referenced to the
C(2)-carbon of (D18)octane at 22.0 ppm (13C) or TMSBr
at 24.3 ppm (29Si) which was measured in the presence of
tetramethylsilane in (D18)octane separately.
[7] V. Dragone, V. Sans, M. H. Rosnes, P. J. Kitson, L. Cronin,
Beilstein J. Org. Chem. 2013, 9, 951.
[8] P. J. Kitson, R. H. Rosnes, V. Sans, V. Dragone, L. Cronin,
Lab Chip 2012, 12, 3267.
[9] G. W. Bishop, J. E. Satterwhite, S. Bhakta, K. Kadimisetty, K.
M. Gillete, E. Chen, J. F. Rusling, Anal. Chem. 2015, 87, 5437.
[10] P. J. Kitson, M. D. Symes, V. Dragone, L. Cronin, Chem. Sci.
2013, 4, 3099.
[11] L. W. McKeen, ‘Permeability Properties of Plastics and Elas-
tomers’, Elsevier, Amsterdam, 2012.
[12] J.-J. Max, C. J. Chapados, J. Chem. Phys. 2007, 126, 154511.
[13] Y. Yokozeki, D. J. Kasprzak, M. B. Shiflett, Appl. Energy 2007,
84, 863.
General Procedure for Polymerization
[14] H. O. Kalinowski, S. Berger, S. Braun, ‘13C-NMR-Spektrosko-
pie’, Georg Thieme, Stuttgart, 1984.
The reaction flask F2 was printed with natural ABS-N. The
solid educts (CuBr (approx. 0.175 mmol), 4,40-dinonyl-2,20-
bibyridine (dnnbipy) (approx. 0.350 mmol) were inserted
during a first pause of the print. A solution of styrene
(5.368 g, 51.5 mmol) in dry cyclohexane (42.436 g,
0.504 mol) was prepared in a separated flask and used for
all polymerizations. The appropriate amount (approx.
16.2 g stock solution, 17.5 mmol styrene, 0.17 mol cyclo-
hexane) of the stock solution was transferred to a mixing
flask; dodecyl 2-bromoisobutyrate (dbib) (approx.
0.035 mmol) was added via a microsyringe and the solution
shaken intensively. The mixture of initiator and styrene
was added to flask F2 via a syringe during the second pause
of the 3D-print and after finishing the print flask F2 was dis-
charged out of the glovebox. The reaction flask was heated
to 85 °C using an aluminum bead heating bath. After heat-
ing up, the flask was held well above the upper critical solu-
tion temperature of polystyrene in cyclohexane (35 °C)
until precipitation of the polymer. After the desired reac-
tion time, the warm reaction flask was opened with a saw,
emptied, the flask washed two times with warm cyclohex-
ane and the combined solutions precipitated in 160 ml
MeOH. After centrifugation for 15 min (2800g), the super-
nate solution was decanted, the residue dissolved in 5 ml
(R)-limonene via sonication, filtered through a 0.45 lm
syringe filter and precipitated in 160 ml MeOH. After
centrifugation, the white residue was dried in vacuo.
[15] F. Schoedler, ‘Das Buch der Natur’, Friedrich Vieweg und
Sohn, Braunschweig, 1860.
[16] T. Ramasarma, A. V. S. Rao, M. M. Devi, R. V. Omkumar, K.
S. Bhagyashree, S. V. Bhat, Mol. Cell. Biochem. 2015, 400, 277.
[17] F. A. Cotton, G. Wilkinson, ‘Advanced Inorganic Chemistry’,
John Wiley & Sons, New York, 1972.
[18] Y. Inaki, S. Nakagawa, K. Kimura, K. Takemoto, Angew.
Makromol. Chem. 1975, 48, 29.
[19] A. Meisters, T. Mole, J. Chem. Soc., Chem. Commun. 1972, 595.
[20] D. W. Harney, A. Meisters, T. Mole, Aust. J. Chem. 1974, 27, 1639.
[21] W. Gu, M. R. Haneline, C. Douvris, O. V. Ozerov, J. Am.
Chem. Soc. 2009, 131, 11203.
[22] S. Kvisle, E. Rytter, Spectrochim Acta, Part A 1984, 40, 939.
[23] J. H. Rogers, A. W. Apblett, W. M. Cleaver, A. N. Tyler, A. R.
Barron, J. Chem. Soc., Dalton Trans. 1992, 3179.
[24] W. K. Chow, C. M. So, C. P. Lau, F. Y. Kwong, J. Org. Chem.
2010, 75, 5109.
~
[25] M. Pena-Lopez, M. Ayan-Varela, L. A. Sarandeses, J. P. Ses-
ꢁ
ꢁ
telo, Chem. Eur. J. 2010, 16, 9905.
[26] B. Wang, H.-X. Sun, Z.-H. Sun, G.-Q. Lin, Adv. Synth. Catal.
2009, 351, 415.
[27] J.-P. Pitteloud, Y. Liang, S. F. Wnuk, Chem. Lett. 2011, 40, 967.
[28] A. Meisters, T. Mole, Aust. J. Chem. 1974, 27, 1665.
[29] E. C. Ashby, J. Laemmle, H. M. Neumann, J. Am. Chem. Soc.
1968, 90, 5179.
[30] J.-M. Lin, B.-S. Liu, Synth. Commun. 1997, 27, 739.
[31] H. H. Hergott, G. Simchen, Liebigs Ann. Chem. 1980, 1718.
[32] K. Matyjaszewski, W. A. Braunecker, ‘Radical Polymerization’
in ‘Macromolecular Engineering: Precise Synthesis, Materials
Properties, Applications’, Eds. K. Matyjaszewski, Y. Gnanou
and L. Leibler, Wiley-VCH, Weinheim, 2007, pp. 161 – 215.
[33] J. Pyun, S. Jia, T. Kowalewski, G. D. Patterson, K. Maty-
jaszewski, Macromolecules 2003, 36, 5094.
REFERENCES
[34] R. D. Pascoe, Waste Management 2006, 26, 1126.
[35] A. Singh, M. Kamal, J. Appl. Polym. Sci. 2012, 125, 1456.
[36] A. Imre, W. A. van Hook, J. Phys. Chem. Ref. Data 1996, 25,
637.
[1] P. Marks, New Sci. 2011, 221, 17.
[2] P. Marks, New Sci. 2011, 221, 18.
[3] R. D. Johnson, Nat. Chem. 2012, 4, 338.
[4] M. D. Symes, P. J. Kitson, J. Yan, C. J. Rickmond, G. J. T.
Cooper, R. W. Bowman, T. Vilbrandt, L. Cronin, Nat. Chem.
2012, 4, 349.
Received November 17, 2015
Accepted December 21, 2015
[5] I. D. Williams, Nat. Chem. 2014, 6, 953.
www.helv.wiley.com
© 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.