A comparative solid state 13C NMR and thermal study of CO 2 capture by amidines PMDBD and DBN

Article abstract of DOI:10.1039/c1gc15457e

The present work shows study of the CO₂ capture by amidines DBN and PMDBD using ¹³C solid-state NMR and thermal techniques. The solid state ¹³C NMR analyses demonstrate the formation of a single PMDBD-CO₂ product which was assigned to stable bicarbonate. In the case of DBN, it is shown that two DBN-CO₂ products are formed, which are suggested to be stable bicarbonate and unstable carbamate. The role of water in the DBN-CO₂ capture as well as the stability of the products to environmental moisture was also investigated. The results suggest that the carbamate formation is favored in dry DBN, but in the presence of water it decompose to form bicarbonate. Thermal analysis shows a good gravimetric CO ₂ absorption of DBN. Release of CO₂ was found to be almost quantitative from the PMDBDH⁺ bicarbonate about 110 °C.

Full text of DOI:10.1039/c1gc15457e

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Green Chemistry
Cite this: Green Chem., 2011, 13, 2146
www.rsc.org/greenchem
PAPER
A comparative solid state 13C NMR and thermal study of CO2 capture by
amidines PMDBD and DBN
Fernanda Stuani Pereira,a,b Deuber Lincon da Silva Agostini,a,b Rafael Dias do Esp´ırito Santo,a,b
Eduardo Ribeiro deAzevedo,c Tito Jose´ Bonagamba,c Aldo Eloizo Joba and Eduardo Rene´ Pe´rez Gonza´lez*a
Received 25th April 2011, Accepted 12th May 2011
DOI: 10.1039/c1gc15457e
The present work shows study of the CO2 capture by amidines DBN and PMDBD using 13
C
solid-state NMR and thermal techniques. The solid state 13C NMR analyses demonstrate the
formation of a single PMDBD-CO2 product which was assigned to stable bicarbonate. In the case
of DBN, it is shown that two DBN-CO2 products are formed, which are suggested to be stable
bicarbonate and unstable carbamate. The role of water in the DBN-CO2 capture as well as the
stability of the products to environmental moisture was also investigated. The results suggest that
the carbamate formation is favored in dry DBN, but in the presence of water it decompose to form
bicarbonate. Thermal analysis shows a good gravimetric CO2 absorption of DBN. Release of CO2
was found to be almost quantitative from the PMDBDH+ bicarbonate about 110 ◦C.
can be recycled from several mixtures. On the other hand, the
1. Introduction
formation of carbamates showing a covalent bond between a
nucleophilic nitrogen center and electrophilic carbon of CO2
molecule is especially interesting for biological processes in very
low water concentration.
While CO2 presents an environmental issue as a greenhouse gas,
it is also environmentally benign in comparison with many other
chemical substrates. The development of suitable methods for
the preparation of interesting CO2-containing compounds, like
organic carbonates and urethanes could be an alternative to
recycling CO2 and using it as a substitute for the highly toxic
phosgene and its derivatives.1–5
On the other hand, the recycling of CO2 requires a previous
knowledge of the possible mechanisms for CO2 capture and
release in several reaction media. Therefore, methodologies have
been developed for investigation and characterization of the CO2
chemical capture. The selectivity and nature of the capture prod-
ucts are two of the most important aspects of the CO2 fixation.6–10
In most of the amine compounds considered for CO2 trap, two
types of reaction seem to take place. One is a thermodynamic
basic attack towards a proton-donor (water, other acidic solvent)
and the other is a nucleophilic attack towards a carbon atom
which is a moderated electrophilic center.
In this article we describe the results of a comparative thermal
and spectroscopic study of two amidines for CO2/H2O trapping.
Solid state 13C NMR measurements allowed proposal of the
formation of two products resulting from CO2 capture by DBN
◦
amidine (water traces at low temperature £-10 C), a DBNH+
bicarbonate and a DBN-CO2 carbamic adduct.
2. Experimental
2.1. Materials
Amidines used for the present study were 3,3,6,9,9-
pentametil-2,10-diazabiciclo[4.4.O]dec-1-en (PMDBD) from
Sigma-Aldrich and 1,5-diazabiciclo-[3.4.0]non-5-en (DBN)
from Fluka (≥98% by GC; <1% H2O) without any previous
purification.
Capture of CO2 in aqueous media seems to occur with
simultaneous capture of water molecules to yield bicarbonates.
This is an interesting and promising process since CO2 and water
2.2. Typical procedure for CO2 capture with amidines PMDBD
10 mmol (2.08 g) of PMDBD were placed into a 50 mL two-
neck round bottom in a minimal fresh acetonitrile volume.
CO2 was introduced at normal pressure and constant flow
under magnetic stirring (eventually, acetonitrile solution can be
previously saturated with CO2). Reaction was kept for two hours
at 0–5 ◦C.6
aUniversidade Estadual Paulista, Campus de Presidente Prudente,
Departamento de F´ısica, Qu´ımica e Biologia, Laborato´rio de Qu´ımica
Orgaˆnica Fina – C.P. 467, Presidente Prudente, 19060-900, SP, Brazil.
E-mail: eperez@fct.unesp.br
bPrograma de Po´s-Graduac¸a˜o em Cieˆncia e Tecnologia de Materiais
(POSMAT), Universidade Estadual Paulista, Sa˜o Paulo, Brazil
cInstituto de F´ısica de Sa˜o Carlos, Universidade de Sa˜o Paulo, C.P. 369,
Sa˜o Carlos, 13560-970, SP, Brazil
Reaction of PMDBD with CO2 in the presence of adventitious
water yielded the corresponding bicarbonate salt. However, if
2146 | Green Chem., 2011, 13, 2146–2153
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
acetonitrile or other solvents are previously dried, no solid
product is formed.
CPMAS and DPMAS pulse sequences. Both techniques are used
for observing 13C solid-state NMR spectra, however, due to the
different excitation schemes used to produce the NMR signal,
in the case of DPMAS the signals arise from both mobile (liquid
phase) and rigid (solid phase) molecules and in the CPMAS the
signals from the solid-phases are privileged.
2.3. Procedure for CO2 capture with DBN
10 mmol (1.24 g) of fresh DBN were placed in a 50 mL round-
bottomed two-necked flask under a constant flow (25 mL min-1)
of CO2 for 5, 10 and 30 min with magnetic stirring. Reaction was
externally cooled using ice-salt-liquid N2 open bath (temperature
was about -15 to -10 ◦C). Apparently all DBN was converted to
a white solid which was stored overnight at -8 ◦C for solid-state
13C NMR analysis.
Fig. 1a shows the CPMAS (top) and DPMAS (bottom)
spectra of the PMDBD-CO2 sample. The CPMAS and DPMAS
spectra are very similar; showing that in the product there is only
the formation of a solid-phase product. As shown in Fig. 1, it is
possible to distinguish at least six lines in the 20 to 40 ppm region,
with chemical shifts of 23.5, 29.7, 30.2, 31.0, 31.8 and 33.7 ppm.
These chemical shift values are quite similar to those obtained in
the solution spectrum (not shown), being the differences mostly
due to solvent or solid-state effects (changes in the chemical
shifts due to relative conformations). Three lines with chemical
shifts of 53.5, 54.0 and 54.4 ppm are observed in the 50–55 ppm
region. Based on the chemical structure shown in Fig. 1a, this
spectral region is attributed to carbons C4 and C8, in which
the signal split is associated to the different conformations of
2.4. 13C solid-state NMR
13C solid-state NMR experiments were performed using a VAR-
1
IAN INOVA spectrometer at 13C and H frequencies of 100.5
and 400.0 MHz, respectively. A VARIAN 7-mm MAS double-
resonance probe head with variable temperature (VT) was used.
The excitation used to acquire each spectrum is mentioned in
the test. The pulse sequence used to acquire the 13C NMR solid-
state spectra were the well-known cross-polarization and direct-
polarization (or single-pulse) techniques, both performed under
magic-angle spinning (MAS) and proton dipolar decoupling,
identified by the acronyms CPMAS and DPMAS, respectively.
The experiments were performed at the MAS speed of 5 kHz
and -30 ◦C. Chemical shifts were referenced to TMS using the
carbonyl line of a Glycine sample as external reference (176.2
ppm). The precision in the Chemical shift determination is about
0.1 ppm. Typical p/2 pulse lengths of 3.5 and 4.5 ms were applied
for 13C and 1H, respectively. Time proportional phase modulated
(TPPM) proton dipolar decoupling with field strength of 70 kHz,
cross-polarization time of 1 ms, acquisition time of 30 ms and
recycle delays of 15 s were used, so that all the CPMAS spectra
presented are fully relaxed. The DPMAS spectra were acquired
with 60 s recycle delays.
2.5. Thermal analysis
The study was carried out using a coupled TG/FT-IR system,
in which all gas evaporated in the TG chamber is channeled
to the FTIR spectrometer. The equipment used was Netzsch
(model 209). The samples (5.0 mg) were heated in an alumina
crucible using pure N2 and CO2 as gas carriers (15 mL min-1)
◦
and a heating rate of 10 C/min. FT-IR spectra was recorded
in the wavenumber range 400–4000 cm-1 with 4 cm-1 of spectral
resolution, 32 scans, and a DTGS detector.
3. Results and discussion
3.1. 13C solid-state NMR studies of amidine-CO2 products
After the CO2 fixation, the solid phase of both PMDBD-
CO2 and DNB-CO2 were obtained by freeze drying. The solid
products were then used for performing the Solid-State NMR
experiments. Since a portion of liquid phase may remain in
the sample, it is necessary to use an excitation scheme that
allows selectivity observing only the solid portion. Despite the
PMDBD-CO2 and DBN-CO2 has a solid aspect, the presence of
a minor liquid-like component cannot be ruled out. This can be
clarified by acquiring 13C solid-state NMR spectra using both
Fig. 1 13C solid-state (a) CPMAS (top) and DPMAS (bottom) 13C
solid-state NMR spectrum of the PMDBD-CO2 product. (b) Zoom of
the 20–60 ppm region of the DPMAS spectrum.
This journal is
The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 2146–2153 | 2147
©

View Article Online
Table 1 13C spin–lattice relaxation times measured for each line in the
the attached CH3. The spectral region corresponding to amidic,
carbonyl and carboxyl carbons shows two lines at 162.2 and
167.0 ppm. The lines observed at 162.2 and 167.0 ppm can be
attributed to a carboxyl group of a bicarbonate structure and
to the amidic carbon (C3), respectively.11 Thus, both the 13C
solid-state and the solution NMR spectra are consistent with
the formation of a stable bicarbonate product with the chemical
structure depicted in Fig. 1a.
spectral range of 150 –160 ppm
Line position (ppm)
T1 (s)
165.8
163.4
162.0
149.0
120 10
35
40
5
5
115
5
Fig. 2a shows the CPMAS spectra corresponding to three
samples of products from reactions of DBN with CO2 by 5
(sample C), 10 (sample B) and 30 min (sample A). As in the
case of PMDBD, most of the peaks are concentrated in the
spectral regions from 145 to 170 ppm and 20 to 60 ppm, but
there is a higher number of peaks, suggesting the formation of
more than one chemical product. This is confirmed by noting
that the intensities of the lines assigned as product B in Fig. 2
increase all together as the reaction time increases, suggesting
that the formation of product B is favored when the reaction
time is increased. Further evidence showing that two products
are formed is provided by the 13C spin lattice relaxation times
(T1) measured by a Torchia sequence (CPT1),12 as shown in Table
1. As can be observed, while the two lines attributed to product
A have T1 values of 35 and 45 s, the lines assigned to product B
have T1 values of 115 and 120 s. This difference in the T1 values
evidences that the local relaxation environment of the two groups
of carbons is different, consistent with the formation of two
distinct products in the reaction. The occurrence of four lines in
the 145 to 170 ppm region suggests the presence of at least two
amidic and two carbonyl carbons. Interestingly, the spectrum
also presents two lines at 162.0 and 163.4 ppm, suggesting
that product A is a bicarbonate species similar to the case of
PMDBD-CO2. Moreover, the line at 149.0 ppm is consistent
with the presence of a carbonyl carbon chemically bonded to
nitrogen. In addition, the line at 165.8 ppm due to product B
shows that the amidinic carbon of this component is less shielded
than that of product A, which would also correspond to the
presence of nitrogen bonded to a carbon rather to hydrogen as
in the bicarbonate. Recently, the formation of carbamates in
TBD-CO2 products was reported.8 In this article, the authors
assigned the signal at 154.0 ppm to a carbonyl carbon attached
to a nitrogen atom in a solution 13C NMR spectrum. Thus, the
presence of the line at 149.0 ppm might be attributed to the
carbonyl carbon of a carbamate structure, such as shown in
the inset of Fig. 2a. The longer T1 relaxation times observed
for product B is also consistent with the assignment of these
peaks as being from a carbamate, since it is expected that the
absence of a close by 1H lead to a less efficient dipolar relaxation
mechanism.
The 13C DPMAS spectrum of sample C is shown in Fig. 2b.
Unlike the CPMAS spectrum observed for PMDBD-CO2, the
DPMAS spectrum of sample A shows the presence of a liquid-
like phase in the product. In fact, as shown in Fig. 2b, the
DPMAS spectrum can be fully assigned to a liquid DBN sample,
showing that, at 30 min of reaction time, there was a substantial
amount of liquid DBN in the sample.
Despite the results shown in Fig. 2, which are quite consistent
with the presence of both bicarbonate and carbamate com-
pounds in the reaction product, there is still the possibility that
product B is not a carbamate but a carbonate, which would also
present 13C lines in the 160–170 ppm region. DBN is highly basic
and likely to react with the bicarbonate to form a carbonate.
To check this possibility, three additional experiments were
performed. First, we mixed DBN in aqueous pH 14 solution
and then added CO2. The solid precipitated product was then
analyzed by 13C CPMAS NMR. The resulting spectrum is shown
in Fig. 3b. In this case, the 20 to 60 ppm region shows only 5 lines,
which is an indication of the formation of a single major product.
Interestingly, the chemical shifts of these lines are very close to
those attributed to the bicarbonate product. Moreover, from
the 160 to 170 ppm region there are two evident lines (at 160.9
and 163.3 ppm) and tiny lines at 166.7 and 171.0 ppm. In the
strongly basic conditions in which the reaction is performed, the
Fig. 2 (a) 13C solid-state CPMAS NMR spectra of the DBN-CO2 prod-
ucts. The proposed bicarbonate (product A) and carbamate (product B)
structures consistent with the observed spectrum is shown as insets. (b)
DPMAS spectrum of sample A.
2148 | Green Chem., 2011, 13, 2146–2153
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
ular sieves (sample D). The water amount decreased from 0.12%
in freshly DBN to 0.04% in dried DBN, as determined by Karl
Fischer coulometric titration (Karl Fischer C30X apparatus).
CO2 was added to the DBN during 30 min and the solid product
was readily analyzed by solid-state NMR. Fig. 4a shows the
resulting 13C CPMAS spectra observed for the two new samples.
Fig. 3 13C solid-state CPMAS NMR spectra of the products resulting
from the reaction of DBN and CO2 under different pH conditions: (a)
DBN + CO2 with water traces. (b) DBN diluted in pH 14 water + CO2.
(c) Precipitated from DBN + CO2 diluted in pH 14 water. (d) DBN dilute
in pH 7 water + CO2. The chemical shifts of the lines assigned with an
asterisk in Fig. 3d match perfectly with those of the peaks assigned to
bicarbonate in Fig. 2a and 3a.
formation of carbonate ions should predominate. The chemical
shift of the carbonate and bicarbonate species should be very
similar except for the carbonyl peak, which should be more
shielded due to the presence of the extra negative charge. This is
exactly what is observed by comparing Fig. 3a and 3b, where the
most prominent shift difference is observed for the carbonyl line.
The tiny lines observed might be due to some impurity in the
sample or the formation of some intermediate in the reaction.
More striking is the complete absence of a peak at 149 ppm
in the spectrum of Fig. 2b. This clearly shows that product B
can hardly be assigned to a carbonate ion, which strengthens the
hypothesis of the formation of a carbamate product. In a second
experiment, the DBN-CO2 product (the same used for obtaining
the spectrum of Fig. 3a) was dissolved in pH 14 water. Thus, if
product B was a carbonate instead of a carbamate, we should
observe a 13C CPMAS spectrum that resembles that assigned to
product B in Fig. 2a. The resulting spectrum is shown in Fig.
3c. As can be observed, the resulting spectrum is completely
different, again ruling out the attribution of product B to a
carbonate ion. As a last experiment, DBN was dissolved in pH
7 water and then reacted with CO2. Again, the solid product
was analyzed by 13C CPMAS NMR and the resulting spectrum
is shown in Fig. 3d. In this case, under weaker basic conditions,
the formation of bicarbonate ions should occur. The spectrum
shown in Fig. 3d cannot be attributed only to bicarbonate ions
and the full assignment of this product is beyond the scope of
this article. However, the chemical shift of the lines assigned
with an asterisk in Fig. 3d matches perfectly with those of the
peaks assigned to bicarbonate in Fig. 2a and 3a. Since under this
reaction condition bicarbonate ions would certainly be present,
this is further evidence pointing to the attribution of product A
as bicarbonate.
Fig. 4 (a) 13C CPMAS spectra of the products resulting from the
reaction DBN with CO2. Top: sample D, dried DBN; Bottom: sample A¢,
fresh DBN. The spectra were acquired right after the sample preparation
(t = 0 h). (b) DPMAS spectra of the products resulting from the reaction
of DBN with CO2. Top: sample D, dried DBN; bottom: Sample A¢, fresh
DBN. The spectra were acquired right after the sample preparation (t =
0 h). (c) Spectra of sample D 6 h after preparation. Top: CPMAS spectra;
Bottom: DPMAS spectra. (d) Time evolution of the lines intensities
of the signals at 149 and 162 ppm (CPMAS spectra) and 158 ppm
(DPMAS) as representatives of product B, product A, and liquid DBN,
respectively. The dashed lines are just guides for the eyes.
Another interesting point for discussion is the role of the water
in the reaction. To provide some experimental data on that, two
additional samples were prepared, the first using fresh DBN
(sample A¢) and the second using a DBN dried using A4 molec-
This journal is
The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 2146–2153 | 2149
©

View Article Online
It is clearly observed that the amount of the product A (assigned
as bicarbonate) is reduced in sample D as compared with sample
A¢. This indicates that water plays a crucial role, favoring the
formation the product A (bicarbonate). The DPMAS spectra
of the two samples are shown in Fig. 4b. As it can be seen,
the CPMAS and the DPMAS spectra of sample D are mostly
identical, presenting the signals of products A and B. In contrast
for sample A¢, besides the characteristic signals of products A
and B, the liquid DBN signals, as shown in Fig. 2b, are also
present. This indicates that the presence of water in the system
is somehow involved in the observation of the liquid DBN. This
result seems to be contradictory with the DPMAS spectrum of
sample A, observed in Fig. 2b, which depicted only the signals
from the liquid DBN. However, it should be pointed out that,
for achieving a good signal-to-noise ratio, the total acquisition
time of the DPMAS spectrum shown in Fig. 2b was about 4
h and this spectrum was obtained after the acquisition of the
corresponding CPMAS spectrum, which took about 2 h. In this
sense, it was not possible to assure that the liquid phase observed
in the DPMAS spectrum of sample A¢, was due to the initial
excess of water present in the sample or to the transformation
of the sample due to the exposure to the environment. To clarify
that, the dried sample, sample D, was measured as a function of
time using an NMR rotor with a holed cap to insure the presence
of moisture in the system. CPMAS and DPMAS spectra were
alternately acquired every 7 min. The CPMAS and DPMAS
spectra obtained after 6 h are shown in Fig. 4c. As it can be
observed, in the CPMAS spectrum, only the signal from product
A is observed, while in the DPMAS spectrum only the signals of
liquid DBN are present. Thus, it became clear that the reason for
the DPMAS spectrum showed in Fig. 2b contain only signals of
liquid DBN is the long period between the sample preparation
and the measurement. In order to show in more details the
sample transformation, Fig. 4d shows a plot of the line intensities
corresponding to the signals at 162 ppm (product A) and 149
ppm (product B) in the CPMAS spectrum as well as the signal
at 159 ppm in the DPMAS spectrum (liquid DBN). It is clearly
seen that the decrease in the amount of product B is correlated
with the appearance of liquid DBN in the system, suggesting
that the product B is unstable in the presence of moisture,
probably reacting with water, releasing CO2 and producing
liquid DBN. However, it can also be noticed that the amount
of product A initially increase, passing through a maximum
and then stabilizing. The interpretation for this behavior is
that initially the CO2 and liquid DBN added in the system by
the decomposition of product B are consumed in the reaction
(under the presence of water) to form the bicarbonate (product
A). When the decomposition of product B ends, consequently
no more CO2 is released and formation of bicarbonate stops.
Therefore the amount of bicarbonate in the system becomes
constant. In the absence of moisture the reaction is much slower,
as we noticed by obtaining the same NMR spectra for a sample
that stayed ~5 h in a sealed rotor. All this behavior would be in
agreement with product B as being a carbamate.
water is still a controversial issue and, despite our experiments
strongly suggests that, it is not yet a definitive answer. Further
experiments to clarify this issue as well as to follow more closely
the kinetics of the product degradation are being designed, and
we intend to present the results in a future publication.
3.2. GC-MS study of DBN/H2O +CO2 system
Finally, a GC-MS study of the DBN hydrolysis showed al-
most quantitative conversion of DBN to the corresponding 3-
(aminopropyl)-2-pyrrolidone (APP) product by using 2 to 25
equivalents of water at room temperature over 12 h. Then,
reaction with CO2 (at 5–10 ◦C) for 2 h allowed formation
of APP-CO2 carbamate trapped as APP-C(O)OEt carbamate
(~13% GC) by the addition of ethyl iodide and stirring for 20 h
at room temperature (Scheme 1).
Scheme 1 DBN hydrolysis and APP carbamate formation by CO2
capture.
By-products were mono and di-alkylated APP. The carbamate
was identified by GC-MS analysis by observation of the total
fragment ions generated by EI ionization and particularly the
peak observed at m/z 214 which corresponds to molecular ion
of product APP-C(O)OEt (Fig. 5).
Fig. 5 GC-MS total fragment ions of APP-C(O)OEt carbamate. EI
ionization mode using 70 eV.
3.3. CO2 capture and release by PMDBD and DBN. Thermal
study
The results of the thermal studies concerning the capture of
CO2 using the amidine bases PMDBD and DBN are shown
sequentially in Fig. 6 through 13 below.
Fig. 6 shows the thermogravimetry analysis of the PMDBD
samples submitted to three isothermal treatments at 25, 35 and
50 ◦C, and using CO2 atmosphere for 180 min. For the sample
◦
submitted to the treatment at 25 C, a mass increase of 1.63%
can be observed at 35 min. Samples studied at 35 and 50 ◦C,
show a decrease in the mass of -0.87 and -2.41% respectively.
Therefore, it can be clearly noticed that the CO2 capture by
PMDBD is influenced by temperature.
The CO2 capture process by PMDBD was found to be
selective in the presence of N2 stream as shown in Fig. 7.
The amidines PMDBD were kept at 25 ◦C for 120 min,
under CO2 atmosphere. During the exposition time a 0.48%
of mass increase was observed confirming the CO2 capture.
Subsequent dynamics treatment under N2 gas shows an abrupt
mass decrease, characteristic of the degradation of pure material
In summary, the 13C solid-state NMR results point to
the presence of two compounds in the composition of the
DBN-CO2 product, the first being a bicarbonate salt and
the second possibly corresponding to a carbamate DBN-CO2
adduct. However, the formation of carbamate with traces of
2150 | Green Chem., 2011, 13, 2146–2153
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
PMDBDH+ hydroxide formed by water re-absorption. For
sample under CO2 atmosphere the endothermic peak with
maximum at 122.8 ◦C confirms that the sample really captures
CO2 molecules, in accordance with the TG measures where a
loss of mass of 21% was attributed to CO2 release.
◦
Presumably, thermal transitions at 219 and 224 C could be
associated to vaporization of PMDBD liberated by thermal
degradation of bicarbonate represented by structure I, in
according to Scheme 2.
Fig. 6 Isothermal CO2 absorption by PMDBD at three temperatures.
Scheme 2 CO2 capture and release by PMDBD.
In the presence of CO2, the corresponding fixation product is
formed and releases the CO2 molecule when CO2 atmosphere is
replaced by N2 (Fig. 9).
Fig. 7 Selective CO2 absorption by PMDBD amidines.
like PMDBD. This process occurred in one step, as confirmed
by DTG curve, finishing in about 180 min at 250 ◦C with a loss
of mass of 98.83%.
As observed in Fig. 8, in a DSC experiment of PMDBD under
CO2 atmosphere, a thermal transition occurred at 122.8 ◦C
associated with the melting and degradation of PMDBD-CO2
product, which confirmed the CO2 capture. Results clearly
show that the sample under N2 presents an endothermic peak
about 75 ◦C characteristic of water molecules weakly absorbed
on the surface of the sample material. Another well accented
endothermic peak appeared at 219 ◦C which could be attributed
to the melting and degradation of PMDBD and a possible
Fig. 9 Reversibility of CO2 fixation with the amidine PMDBD.
Thermogravimetric analysis for a sample of PMDBD-CO2
product (obtained under undried conditions) is illustrated in Fig.
10 and shows loss of solvent molecules (acetonitrile), water and
CO2 from the bicarbonate moiety starting around 100 ◦C with a
maximum at 110.3 ◦C, this loss occurring without melting of the
Fig. 8 Comparative DSC of PMDBD molecule under both, CO2 (curve
4) and N2 (curve 5) flows.
Fig. 10 TG and DTG of PMDBD-CO2.
This journal is
The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 2146–2153 | 2151
©

View Article Online
solid product (presumably by sublimation) and represents ~23%
of the initial mass. At 122.9 ◦C a loss of mass corresponding
to ~21% of the initial mass is observed and could be associated
to a melting process with thermal degradation and subsequent
release of CO2 seems to have happened to the PMDBD-CO2
product in concordance with DSC analysis of PMDBD under
CO2 environment (Fig. 8). Clearly, the quantitative release of
CO2 can be observed in the temperature range of ~100 to
123 ◦C.
with bicarbonate formation. The bands at 2974, 2931, 2857,
2824, 2785, 1450, 1376 and 890 cm-1 are attributed to PMDBD
molecule in the vapor phase.
Thermal study was preliminarily performed for DBN-CO2
system. The initials results were found to be promising (Fig. 12
and 13).
From Fig. 11 it can be observed that at 110 ◦C the CO2 release
is almost selective, however, at 153 ◦C only bands attributed to
the PMDBD molecule have appeared, that is, a complete CO2
release from PMDBD-CO2 product has occurred.
Fig. 13 DSC study of DBN-CO2 solid product (obtained after 3 h of
reaction time).
Fig. 12 shows a high efficiency of CO2 gravimetric fixation
(40% of mass increase relative to initial mass of DBN) and
CO2 release occurring at moderate temperatures (starting at
~40 ◦C) with a peak at about 80–90 ◦C. Increase of about
18% of mass occurred rapidly at ~10 min in accordance with
solid state 13C NMR experiments. It is possible to assume the
formation of DBNH+ bicarbonate by the splitting of water
contained in starting DBN and subsequent reaction with CO2
to form the corresponding bicarbonate salt. After around 10 to
15 min free DBN reacts with CO2, yielding around 22% of DBN-
CO2 carbamate adduct, completing this way 40% of gravimetric
absorption of CO2 or of DBN conversion to DBN-CO2 products
under conditions of the experiment described in Fig. 12(A). On
the other hand, pointed thermal degradation until 80–90 ◦C
(Figure1 12B) is associated to a loss of mass about 25% which
could be related to CO2 release from DBN-CO2 carbamate.
DSC curve for DBN-CO2 solid displayed in Fig. 13 shows a
low glass transition temperature at -57.8 ◦C typical for some
urethane systems. On the other hand, the solid seems to melt
with decomposition (CO2 release) at a well defined temperature
of 73.4 ◦C (sharp transition curve). A second thermal transition
Fig. 11 FT-IR spectra from TG/FT-IR analysis of vapor phase at
temperatures between 110 and 153 ◦C.
The absorption bands of FT-IR spectrum recorded at 110 ◦C
proved the thermal degradation of PMDBD-CO2 product show-
ing two intense superposed bands at 2360–2340 cm-1 attributed
to free CO2. The spectra show bands in the region of 3960–3490
cm-1 associated to the presence of steam, which is in accordance
Fig. 12 (A) Isothermal CO2 absorption (25 ◦C) by DBN. A mass increase of 40% can be observed. (B) DBN-CO2 thermal degradation with CO2
release.
2152 | Green Chem., 2011, 13, 2146–2153
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
◦
can be observed at 94.8 C. Since the boiling point of DBN is
Acknowledgements
95–98 ◦C at 7.5 mmHg, this transition could be better associated
to thermal degradation of another compound present in DBN-
CO2 product mixture with possible liberation of water, CO2 and
liquid DBN in agreement to NMR results and CO2 absorption–
release study shown in Fig. 12.
The behavior of the amidine DBN towards CO2 molecule
at water concentration less than 1%, low temperature and in
solventless conditions could be associated with both the high
nucleophilicity of DBN toward Csp2 center and the higher Lewis
basicity of this amidine among other bases. The nucleophilicity
and Lewis basicity of DBU and DBN toward Csp2 centers have
been measured: nucleophilicities increase in the series DMAP<
DBU< DBN< DABCO while Lewis basicities are DABCO<
DMAP< DBU< DBN.13
The authors are thankful to FAPESP for a research funding
(contract 2006/51987-6), CNPq and POSMAT post-graduation
program. The authors are also very grateful to Dr David
J. Heldebrant of Pacific Northwest National Laboratory for
his very important suggestions and discussions of the topics
described in the present manuscript. Authors are indebted to
Polymer Laboratory at IFSC-USP for lab facilities. We would
like to acknowledge the Chemistry Bacharel Aline Cordeiro Re-
cacho from Mettler Toledo at Sa˜o Paulo, Brazil, for performing
Karl Fischer coulometric titrations.
References
1 E. R. Pe´rez, M. O. da Silva, V. C. Costa, U. P. Rodrigues-Filho and
D. W. Franco, Tetrahedron Lett., 2002, 43, 4091–4093.
2 P. Alessio, D. M. Ferreira, A. E. Job, R. F. Aroca, A. Riul, C. J.
L. Constantino and E. R. P. Gonzalez, Langmuir, 2008, 24, 4729–
4737.
4. Conclusions
PMDBD amidine capture CO2 forming thermodynamically
stable products (bicarbonates) which can release CO2 selectively
at moderated temperatures. This amidine base has shown a
selective CO2 capture capacity in the presence of N2 (none
PMDBD-N2 products were formed) which induces its use as
a model compound for future study of selective CO2 capture
from several gas mixtures. The preference for a proton transfer
reaction also converts PMDBD (better as thermodynamic base)
as a model for simulation of enzyme activity in catalysis of CO2
– proton transfer reactions.
Independent of being focused on the study of CO2 capture, the
PMDBD molecule can fix at the same time water molecules as
stable bicarbonates, which can maintain its properties for long
periods without special storage conditions.
The DBN molecule shows the highest capacity of CO2 gravi-
metric capture among several amidine bases (DBU, PMDBD)
that can be justified for DBN capacity of forming a DBN-CO2
carbamate adduct. This high reactivity toward CO2 molecule
and the fastest response to CO2 presence in gas phase and
solution give reason for more investigation of CO2 – DBN
complexes using thermal and spectroscopic techniques.
3 C. R. Gomes, D. M. Ferreira, C. J. L. Constantino and E. R. P.
Gonzalez, Tetrahedron Lett., 2008, 49, 6879–6881.
4 J. M. Hooker, A. T. Reibel, S. M. Hill, M. J. Schueller and J. S. Fowler,
Angew. Chem., Int. Ed., 2009, 48, 3482–3485.
5 D. J. Heldebrant, P. K. Koech, C. Trisha, M. Ang, C. Liang, J. E.
Rainbolt, C. R. Yonker and P. G. Jessop, Green Chem., 2010, 12,
713–721.
6 E. R. Perez, R. H. A. Santos, M. T. P. Gambardella, L. G. M. de
Macedo, U. P. Rodrigues-Filho, J. C. Launay and D. W. Franco, J.
Org. Chem., 2004, 69, 8005–8011.
7 B. Ochiai, K. Yokota, A. Fujii, D. Nagai and T. Endo, Macro-
molecules, 2008, 41, 1229–1236.
8 C. Villiers, J.-P. Dognon, R. Pollet, P. Thuery and M. Ephritikhine,
Angew. Chem., Int. Ed., 2010, 49, 3465–3468.
9 F. S. Pereira, E. R. DeAzevedo, E. F. da Silva, T. J. Bonagamba, D.
Agost´ıni, A. Magalha˜es, A. Job and E. R. Pe´rez, Tetrahedron, 2008,
64, 10097–10106.
10 D. J. Heldebrant, C. R. Yonker, P. G. Jessop and L. Phan, Energy
Environ. Sci., 2008, 1, 487.
11 (a) J.-Y. Park, S. J. Yoon and H. Lee, Environ. Sci. Technol., 2003, 37,
1670–1675; (b) P. J. Hajduk, R. P. Meadows and S. W. Fesik, Q. Rev.
Biophys., 1999, 32, 211–240.
12 L. E. Kay, L. K. Nicholson, F. Delagio, A. Bax and D. Torchia, J.
Magn. Reson., 1992, 97, 359–375.
13 M. Baidya and H. Mayr, Chem. Commun., 2008, 1792–
1794.
This journal is
The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 2146–2153 | 2153
©