International Journal of Greenhouse Gas Control 16 (2013) 217–223

Contents lists available at SciVerse ScienceDirect

International Journal of Greenhouse Gas Control

journal homepage: www.elsevier.com/locate/ijggc

Efficient CO2 absorption and low temperature desorption with

non-aqueous solvents based on 2-amino-2-methyl-1-propanol (AMP)
Francesco Barzagli a , Fabrizio Mani a,b,∗ , Maurizio Peruzzini a
a
ICCOM CNR, via Madonna del Piano 10, 50019 Sesto Fiorentino, Firenze, Italy
b
University of Florence, Department of Chemistry, via della Lastruccia 3, 50019 Sesto Fiorentino, Firenze, Italy

a r t i c l e i n f o a b s t r a c t

Article history: In the present study, an efficient process for CO2 capture using some AMP–alkanolamine blends, namely
Received 7 December 2012 DEA, MDEA, MMEA, DIPA, in non-aqueous solvents has been described. The solvents used are mixtures of
Received in revised form 22 March 2013 ethylene glycol or 1,2-propandiol with either methanol or ethanol. Additionally, either pure 1-propanol
Accepted 23 March 2013
or diethylene glycol monomethyl ether have been also employed. Throughout the whole set of exper-
iments, the amine solutions were continuously circulated in a closed cycle between the absorber (set
Keywords:
at 20 ◦ C) and the desorber (set at 65, 70, 75, 80 ◦ C). The CO2 absorption efficiency at equilibrium is in
CO2 capture
13 the range 73.1–95.9%. The carbon containing species in the amine–CO2 –alcohol equilibria were ana-
C NMR speciation
2-Amino-2-methyl-1-propanol
lyzed by 13 C NMR spectroscopy showing that they originate from alcohol and/or amine carboxylation.
Alkyl carbonates The AMP–DEA–1-propanol–CO2 system gives a solid compound that has been identified as the carba-
Amine carbamates mate derivative of AMP, [(AMPH)(AMPCO2 )], from its 13 C solid state NMR spectrum and confirmed by
single crystal X-ray structure determination. [(AMPH)(AMPCO2 )] instantaneously dissolves in methanol,
ethanol and ethylene glycol affording the respective monoalkyl carbonate. The lower stripping tempera-
ture of the studied absorbents offers several potential advantages with respect to conventional aqueous
solutions of the same alkanolamines, yet preserving comparable efficiency.
© 2013 Elsevier Ltd. All rights reserved.

1. Introduction 2005; Jamal et al., 2006; Oyenekan and Rochelle, 2006; Idem
et al., 2006; Choi et al., 2007; Böttinger et al., 2008). However,
Chemical capture of CO2 by aqueous alkanolamines has drawbacks such as the energy cost for the solvent regenera-
been investigated for a long time as a possible technique for tion, the solvent loss due to evaporation, oxidative and thermal
the separation and removal of CO2 from flue gas (Bottoms, degradation in the absorption–desorption cycles and corrosion
1931; Teller and Ford, 1958; Caplow, 1968; Astarita et al., problems, still represent serious problems in view of a widespread
1984). For instance, monoethanolamine (MEA, 2-aminoethanol), use of this technique for CCS applications (Goff and Rochelle,
diethanolamine (DEA, 2,2 -iminodiethanol), methylethanolamine 2006; Rao and Rubin, 2006; Jassim and Rochelle, 2006; Davison,
(MMEA, N-methyl-2-aminoethanol), methyldiethanolamine 2007; Lepaumier et al., 2009; Reynolds et al., 2012). Therefore,
(MDEA, N-methyl-2,2 -iminodiethanol), aminomethylpropanol it is mandatory to formulate new absorbent systems capable of
(AMP, 2-amino-2-methyl-1-propanol), diethylene glycolamine maximizing the net balance CO2(captured) − CO2(emitted) . CO2(emitted)
(DGA, 2-(2-aminoethoxy)ethanol), diisopropanolamine (DIPA, represents the overall amount of CO2 released in the atmosphere
bis(2-hydroxypropyl)amine) and their blends, have been thor- by burning fossil fuels to produce all forms of energy (electrical,
ougly studied in lab-scale experiments. Aqueous MEA and DEA, thermal and mechanical) necessary to sustain the entire CO2 -
in particular, have had industrial scale application to CO2 cap- removal cycle, from the production of the reagents to the final
ture from natural gas extraction, gas refinery and exhaust gases transport of CO2 and its disposal. This objective has to be taken
produced by fossil fuel fired power generation by virtue of their into account for a reliable cost-to-benefit assessment, other-
high CO2 absorption efficiency and loading capacity (Yeh et al., wise any CCS technology is highly questionable for large-scale
2001; Bonenfant et al., 2003; Mandal et al., 2004; Sun et al., applications.
In aqueous alkanolamine systems, CO2 is captured in solution as
HCO3 (only to a much lesser extent as CO3 2− ) and, for primary and
secondary unhindered amines, also as thermally stable carbamates.
∗ Corresponding author at: University of Florence, Department of Chemistry, via 13 C NMR analysis has allowed to establish that the relative amount
della Lastruccia 3, 50019 Sesto Fiorentino, Firenze, Italy. Tel.: +39 055 4573279;
fax: +39 055 4573385.
of the two species depends on both the amine concentration and
E-mail address: fabrizio.mani@unifi.it (F. Mani). CO2 /amine molar ratio (Hook, 1997; Yoon and Lee, 2003; Park et al.,

1750-5836/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijggc.2013.03.026
218 F. Barzagli et al. / International Journal of Greenhouse Gas Control 16 (2013) 217–223

2003; Barzagli et al., 2009, 2010, 2011; Ballard et al., 2011). The 2. Materials and methods
release of CO2 from stable amine carbamates and from HCO3 −
requires high stripping temperatures, typically about 110–140 ◦ C 2.1. General information
at pressure over 1 bar and a huge amount of energy is consumed
in the amine regeneration step. Replacing water with organic sol- All used chemicals were of analytical grade. The alka-
vents may provide significant advantages in regard to the reduced nolamines (>98%) and the alcohols (>99%) were purchased from
solvent decomposition and to the energy saving in the regeneration Sigma–Aldrich and were used as received without further purifica-
step due to the lower heat capacity (about half), the lower heat of tion. Pure CO2 (99%) and air (Rivoira Spa) were used to simulate flue
vaporization of organic solvents and the higher boiling temperature gas. Flow rates of air and CO2 were measured with gas mass flow
compared to water (Oexmann and Kather, 2010). meters (Aalborg) equipped with gas controllers (Cole Parmer). The
Recently, specific organic liquids have been designed and tested inlet and outlet CO2 concentrations in the flue gas mixture were
for CO2 capture including CO2 binding organic liquids, CO2 BOLs measured with a Varian CP-4900 gas chromatograph calibrated
(Heldebrant et al., 2008; Liu et al., 2006) and room temperature with a 10% (v/v) CO2 /N2 reference gas (Rivoira Spa).
ionic liquids, RTILs (Bates et al., 2002; Bara et al., 2009, 2010; Hart The bench-scale experiments of CO2 absorption were carried
et al., 2010). These absorbents have several advantages over the out in a continuous cycle, as previously described for aqueous
aqueous alkanolamines being high-boiling, thermally stable liquids alkanolamines (Barzagli et al., 2010). In each experiment both
with lower heat capacity than water. However, the expensive syn- the CO2 -loaded (carbonated) and the regenerated amine solu-
thesis of the starting compounds (for example, amidine, guanidine, tions were continuously circulated between the absorber and the
amine functionalized imidazole) with respect to alkanolamines desorber in a closed cycle, set at the appropriate temperatures.
and the high viscosity of the carbonated derivatives, prevent their The closed cycle absorption–desorption apparatus consists of a
application to industrial processes. Equimolar mixtures of alka- properly designed absorber (Mani et al., 2008) connected to a
nolamines and RTILs have been tested as effective solvents for desorber unit and equipped with a double head peristaltic pump
reversible CO2 capture (Camper et al., 2008). Upon reaction with (Masterflex) allowing the solutions to circulate continuously in a
CO2 solid products are formed, which, represent a neat drawback closed loop between the absorber and the desorber at the flow
for any industrial application irrespectively of the improved CO2 rate of 0.600 dm3 h−1 . The absorber (set at 20 ◦ C) and the desor-
capture efficieny. ber (set at 65, 70, 75 and 80 ◦ C) are maintained at the desired
We have devised a different protocol based on the replacement temperature by two thermostatted baths (Julabo model F33-MC
of water with alcohols either alone or in mixtures. We selected refrigerated bath). The carbonated amine solution exiting from
AMP based solvents as suitable candidates to test the efficiency the absorber is preheated by a cross heat exchanger with hot
of this technique, because of their relatively low regeneration regenerated amine solution exiting from the desorber which is,
energy and high thermal stability (Barzagli et al., 2010; Freeman in turn, cooled to room temperature before being recycled to
et al., 2011). AMP indeed cannot form carbamate in water solu- the absorber. The starting solutions of blended amines were at
tion due to steric hindrance at the amine functionality (Sartori overall concentrations 2.00 mol dm−3 . AMP–amine molar ratios
and Savage, 1983). Finally, we preferred testing AMP blended with were 1:1 but AMP–DEA 2:1 in DEGMME. Mixed solvents were 1:1
different amines as the performances of the amine blends are gen- on volume scale, except the 1:2 volume ratio of EG:1-propanol.
erally superior over their single components, due to the faster CO2 Both the absorber and the desorber were charged with the same
absorption of the ancillary amines (Mandal et al., 2003; Sun et al., volume (0.300 dm3 ) of the appropriate blend solution that was
2005). previously 50% saturated with CO2 . To prepare these solutions,
The study here reported was aimed at investigating the CO2 0.300 dm3 of the appropriate amine solution were pre-saturated
capture properties of systems formed by AMP blended with DEA, with pure CO2 and then mixed with the required volume of the
MDEA, MMEA, DIPA, dissolved in different alcohol mixtures (eth- same free amine solution to obtain an overall 0.600 dm3 solu-
ylene glycol, EG, and either methanol or ethanol; 1,2-propandiol tion.
and ethanol, in 1:1 volume ratio) or in single 1-propanol and To mimic the flue gas, a gas mixture containing 12% (v/v) CO2
diethylene glycol monomethyl ether (DEGMME). The absorption in air was continuously fed into the absorber through a sintered
experiments were carried out in a continuous cycle with CO2 cap- glass diffuser (16–40 ␮m pores) at the bottom of the absorbent
ture and simultaneous amine regeneration, as previously described solution with a flow rate of 14.0 dm3 h−1 (0.0689 mol h−1 of CO2 at
for aqueous alkanolamines (Barzagli et al., 2010). The carbon con- 24 ◦ C). Three polyethylene disks threaded on a 2-mm glass rod were
taining species participating in the amine–CO2 –alcohol equilibria placed within the liquid to increase turbulence, therefore provid-
were analyzed by our 13 C NMR spectroscopy procedure (Barzagli ing the reaction mixture with a sufficient residence time. The vent
et al., 2009, 2011), and were found to originate from carbonatation gas exited from the top of the absorber. The inlet gas mixture was
of both alcohol and amine. Systematic experimental investigations humidified by bubbling through water before absorption. The out-
of amine–CO2 reactions in organic solvents are still scarce, unlike let gas was dried by in turn flowing through a condenser cooled at
in aqueous solutions, and have been limited to kinetic and solubil- −5 ◦ C, a concentrated H2 SO4 solution and a gas purification tower
ity studies of MEA, DEA, MDEA in a variety of organic solvents (Xu filled with P2 O5 , before being GC analyzed. The stripped CO2 was
et al., 2002; Park et al., 2005). not recovered. A complete cyclic experiment lasted 24–36 h and,
The CO2 capture by AMP–DEA solution in 1-propanol formed a for each desorption temperature, it was stopped when the reac-
solid compound identified as the carbamate derivative of AMP. The tions of CO2 capture and amine regeneration reach a steady state.
peculiar reactivity of this compound toward some alcohols was also The uncertainty of the apparatus is given by the error in the solu-
investigated. tion preparation: it is estimated less than 1%. All of the experiments
The potential advantages of this technology, in comparison to were repeated in duplicate or triplicate with good reproducibility
conventional aqueous solutions of identical alkanolamines, would of the results: the deviation from the average values was always in
include lower amine decomposition and evaporation as well as the range 1.5–3%.
reduced equipment corrosion due to the reduced stripping tem- A summary of the operating conditions is reported in Table 1.
perature. Additionally, the lower heat capacity and evaporation Absolute viscosity values (at 27 ◦ C) of the starting solutions have
enthalpy of the organic solvents compared to water should con- been reported in Table 2. Reference values (mPa s) are: aqueous
tribute to decrease the regeneration heat penalty. AMP/DEA, 1.78; water 0.92; EG 16.2.
 F. Barzagli et al. / International Journal of Greenhouse Gas Control 16 (2013) 217–223 219

Table 1 was used as internal reference (CH3 , ı = 1.47). To provide enough

Operating conditions in the absorption–desorption experiments.
signal for deuterium lock, D2 O (Aldrich) contained in a sealed
Absorption solution 0.300 dm3 50% carbonated solution glass capillary was introduced into the NMR tube containing the
Desorption solution 0.300 dm3 50% carbonated solution amine solution. To obtain quantitative 13 C NMR spectra, the pulse
Absorption temperature 20 ◦ C
sequence with proton decoupling and NOE suppression was used
Desorption temperature/pressure 65–70–75–80 ◦ C/1 bar
Amine concentration 2.00 mol dm–3 ; 16.8–22.6 wt%
to acquire the 13 C{1 H} with the following acquisition parameters:
Liquid flow rate 0.600 dm3 h−1 pulse angle, 90.0◦ ; acquisition time, 1.36 s; delay time, 2–30 s; data
Gas flow rate 14 dm3 h−1 points, 64 K; number of scans, 500–2000. The data were processed
by using Bruker-Biospin Topspin software. Increasing the acqui-
sition time and/or the relaxation delay time (up to 60 s) does not
The maximum loading capacity of each amine solution was mea-
produce substantial changes in the relative peak areas of the CH2
sured by fluxing continuously pure CO2 until no more gas was
carbon atoms with the same number of protons (Breitmaier and
absorbed. The CO2 uptake was obtained from the weight differ-
Voelter, 1990). The speciation of the carbon-containing species in
ence between the unloaded and CO2 saturated solution to constant
the CO2 saturated solutions, in the starting solutions (50% CO2
weight. The saturation of each amine solution with pure CO2 is
saturation), in the carbonated (absorber) and regenerated (desor-
a very exothermic reaction and was carried out at thermostatted
ber) solutions at increasing temperatures (65–80 ◦ C) was analyzed
20 ◦ C. To prevent loss of absorbent; CO2 was passed through the
by running 13 C NMR spectra of the solutions at successive steps
appropriate alcohol (or alcohol mixture) before being admitted into
of each experiment. The relative amounts of carbamate and the
the amine solution. The absorption was stopped when no more CO2
rapidly equilibrating (free amine)/(protonated amine) have been
was absorbed.
estimated by peak integration for each CH2 resonance, as already
The saturation with 12% CO2 of 0.300 dm3 of 2.0 mol dm−3
reported (Barzagli et al., 2009, 2010, 2011). In order to quantify
AMP–DEA in 1-propanol for 40 min causes the slow formation of
the relative amounts of carbamate and alkyl carbonate, we care-
a voluminous white precipitate which was separated by filtration,
fully integrated the carbon resonances of carbamate R-CO2 − and
washed in turn with the same solvent, a mixture of the solvent
carbonate R -OCO2 − falling in the range 164–159 ppm. The quater-
with diethyl ether (1:1, v/v) and diethyl ether alone, before being
nary 13 C atoms of R-OCO2 − and of R -CO2 − functionalities exhibit
dried at room temperature with a dry N2 stream. The mass of the
much longer relaxation times than CH2 groups, thus resulting in
solid was 64% with respect to the mass of the starting AMP. In
lower intensity resonances. Although some uncertainty in compar-
an analogous CO2 saturation experiment, the precipitate was not
ing the different integrals is unavoidable (estimated error < 10%),
separated from the solution and the slurry was decomposed at
nonetheless the method is certainly reliable and provides a first
increasing temperatures between 50 and 65 ◦ C, at room pressure,
estimation of the concentration changes of the species originating
until the release of CO2 ended and the solid phase disappeared.
from the adsorbed CO2 in solution and of their dependence from the
The volume of the overall desorbed CO2 was measured at room
desorption temperature. In the CO2 absorption experiments yield-
temperature and pressure using the previously described gas-tight
ing solid precipitates, 13 C NMR spectra of the solutions were carried
apparatus (Barzagli et al., 2009).
out either before or after precipitation of the solid compounds.
13 C Cross polarization magic angle spinning (CP-MAS) 13 C NMR
2.2. NMR spectroscopy
spectra were recorded at room temperature on a Bruker Avance
DRX-400 spectrometer equipped with a 4 mm BB CP-MAS probe at a
The 13 C NMR spectra of the solutions were obtained at room
working frequency of 100.62 MHz. The spectra were recorded using
temperature with a Bruker AvanceIII 400 spectrometer operating
a cross polarization pulse sequence under magic angle spinning at
at 100.61271 MHz. Chemical shifts are to high frequency rela-
a spinning rate of 10.0 kHz.
tive to tetramethylsilane as external standard at 0.00 ppm. CH3 CN

Table 2
CO2 loading capacity at 20 ◦ C and absorption efficiency by the different amines at increasing desorption temperatures. The overall amine concentration is 2.0 mol dm−3 .

Entry Amine Solventa Amine Viscosity Loading Loading Average absorpt. efficiencyd and
conc. (wt%) (mPa s, capacityb capacityc (wt%) desorpt. temp. (◦ C)
27 ◦ C) (molar ratio)

65 70 75 80
e
1 AMP/DEA 1,2-PD /ethanol 20.6 8.65 0.65 29.4 75.7 84.7 91.8 93.8
2 AMP/DEA EGf /methanol 19.3 4.28 0.77 34.9 73.5 84.9 92.4
3 AMP/DEA 1-Propanolg 22.1 3.40 0.65 29.4 81.7 87.9 92.9
4 AMP/DEAh DEGMMEi 18.3 5.76 0.70 31.7 73.1 77.4 89.5 91.6
5 AMP/MDEA EG/methanol 20.7 3.33 0.70 28.7 81.3 88.5 93.5
6 AMP/MMEA EG/methanol 16.9 3.73 0.81 43.4 76.7 79.0 91.3 95.9
7 AMP/MMEA EG/ethanol 16.8 5.51 0.76 40.7 62.5 86.4 92.6
8 AMP/DIPA EG/ethanol 22.6 8.05 0.69 27.3 73.8 87.2 93.1
a
Solvent mixtures are 1:1 in volume, unless otherwise stated.
b
Molar ratio between absorbed CO2 and starting amine; absorption carried out with pure CO2 .
c
Percent ratio between the masses of absorbed CO2 and starting amine.
d
Percent ratio between absorbed and flowed CO2 ; absorption carried out with 12% (v/v) CO2 in air; the deviation from the average values was always in the range 1.5–3%.
e
1,2-Propandiol.
f
Ethylene glycol.
g
A small amount of solid formed during absorption.
h
AMP/DEA = 2:1 on molar basis.
i
Diethylene glycol monomethyl ether.
220 F. Barzagli et al. / International Journal of Greenhouse Gas Control 16 (2013) 217–223

3. Results and discussion

3.1. Continuous cycles of absorption–desorption

The choice of the alcohols to be mixed with amines was dic-

tated by four main considerations: (i) low cost, (ii) high boiling
temperature, (iii) solubility of the compounds which form upon
reaction between amine and CO2 and (iiii) avoidance of foaming
problems and immiscible liquid phases. Concerning point (iii), eth-
ylene glycol (hereafter indicated as EG) was found necessary to
avoid the formation of precipitates during CO2 absorption. Unfor-
tunately, pure EG is unsuitable for CO2 capture in spite of its high
boiling temperature and low cost, due to its high absolute viscosity
( = 19.9 mPa s, at 20 ◦ C) that reduces the mass transfer, therefore
lowering the absorption efficiency. Consequently, in most experi-
ments we employed 1:1 (v/v) mixtures of EG with either methanol
or ethanol ( = 0.54–1.07 mPa s, at 20 ◦ C) to reduce the viscosity
of the solvent yet maintaining a boiling temperature up to 150 ◦ C.
For comparison purposes, we also used amine solutions in pure
1-propanol and diethylene glycol monomethyl ether (DEGMME).
In order to verify how the absorption efficiency depends Scheme 1.

on the desorber temperature and to find the best compromise

between regeneration temperature and CO2 absorption efficiency, et al., 2010). However, the fast CO2 alkylation in very mild condi-
the experiments were carried out at increasing desorber temper- tion (20 ◦ C and P(CO2 ) = 0.12 bar) promoted by a common amine is
ature (65, 70, 75 and 80 ◦ C). Each experiment was stopped when noteworthy. The stoichiometric loading capacity is 1, according to
the reactions of CO2 capture and amine regeneration reached the reaction (3), and 0.5 according to reactions (1) and (2). In our exper-
steady state and the absorption efficiency did not change with time. iments, the measured loading capacity was much greater than 0.5,
The results of some representative experiments are reported in indicating clearly that reaction (3) was competing with reaction (2)
Table 2. As expected, increasing the desorber temperature increases in the absorption step.
the amine regeneration and, consequently, the CO2 absorption effi- To collect further confirmatory evidence for this hypothesis, we
ciency. Taking into account both the loading capacity, 43.4 wt% analyzed the CO2 loaded solutions by 13 C NMR spectroscopy, as
(the maximum loading capacity, CO2 /amine molar ratio, is 0.81) already reported for aqueous solutions (Barzagli et al., 2009, 2010,
and the lowest desorption temperature required to keep the 2011). The peak assignment was done according to literature data
absorption efficiency greater than 90%, the blended AMP–MMEA (Stueber et al., 2001; Park et al., 2003; Ma’mun et al., 2006). Besides
solution in EG–methanol mixture (Table 2, entry 6) showed the best the most intense signals due to the carbon atoms of the alcohols
performance. A comparison of the absorption efficiency (greater and of both protonated and free amines that are fast exchanging
than 90%) of the AMP–DEA and AMP–MDEA in organic solvents on the NMR time scale, a couple of resonances at about 67.0 and
with aqueous DEA (69.1%), MDEA (69.0%), AMP–DEA (77.1%) and 60.9 ppm can be assigned to the carbon atoms of the monoalkyl
AMP–MDEA (78.8%) (Barzagli et al., 2010) under similar conditions carbonate derivative of ethylene glycol (HO CH2 CH2 OCO2 − ).
(amine concentration, absorption temperature), clearly shows that Furthermore, a resonance at 159.2 ppm identifies the alkyl car-
amine regeneration is easier in non-aqueous solvents, despite the bonate quaternary carbon (R-OCO2 − ). Additionally, 13 C resonances
higher desorption temperature of the aqueous amine solutions ascribable to the monoalkyl carbonates derivatives of methanol
(90 ◦ C) with respect to organic solvents (80 ◦ C). (160.1, 52.6 ppm), ethanol (159.5, 61.0, 14.7 ppm) or 1-propanol
In the absence of water, CO2 reacts with both primary and (159.4, 67.1, 22.9, 10.4 ppm), respectively, were observed. Finally,
secondary amines forming unstable carbamic acids [Eq. (1); AmH the resonances of the carbamate derivative of DEA (164.24, 61.75,
denotes either primary or secondary amine]: 51.06 ppm), MMEA (164.02, 60.53, 51.07, 35.15 ppm) and DIPA
(164.88, 67.90, 54.70, 20.74 ppm) were found in the spectra. On the
AmH + CO2  AmCO2 H (1)
contrary, the AMP carbamate could not be detected. The 13 C res-
Once formed, the carbamic acid may react with an excess of onance assignments for the AMP–DEA–EG–methanol are reported
amine [Eq. (2)]: in Table 3.
A possible mechanism accounting for the formation of
AmCO2 H + AmH  AmCO2 − + AmH2 + (2)
monoalkyl carbonate [reaction (3)] is shown in Scheme 1A and B.
Although tertiary amines such MDEA cannot react with CO2 in The reaction was initiated by the nucleophilic attack of the
anhydrous conditions, they can still enter in the CO2 capture pro- AMP amine lone pair to the CO2 carbon followed by amine depro-
cess when mixed with primary or secondary amines when they tonation, resulting in the formation of both AMP carbamate and
behave as bases for the carbamic acid deprotonation [Eq. (2)]. Addi- protonated AMP (Scheme 1A). Subsequently, the unstable AMP
tionally, the alcohol itself could take part in the CO2 uptake in the carbamate intermediate may undergo the nucleophilic attack at
presence of a base (i.e. the amine) forming an alkyl carbonate [Eq. the carbonyl atom by the oxygen of alcohol, which gives back the
(3); R denotes an alkyl group]: amine while forming the monoalkyl carbonate (Scheme 1B). An
indirect support for the above mechanism is given by the complete
AmH + CO2 + ROH  AmH2 + + ROCO2 − (3)
conversion of the solid AMP carbamate to alkyl carbonate upon dis-
Reaction (3) has been reported to occur with saturated alcohols solution in alcohols (see later). The above mechanism suggests that
in the presence of alkylbromide and Cs2 CO3 (Kim et al., 1999), the direct nucleophilic attack of the ROH oxygen atom to the CO2
dibutyltin(IV)oxide (Jimil et al., 2009) and of both amidine and carbon is disfavored by a greater activation energy.
guanidine (Heldebrant et al., 2008). It also occurs with unsatu- The maximum loading capacity of alcoholic AMP–amines in the
rated alcohols in the presence of tert-butyl ipoiodide (Minakata range 0.65–0.81 (Table 2) follows from the CO2 capture as alkyl
 F. Barzagli et al. / International Journal of Greenhouse Gas Control 16 (2013) 217–223 221

Table 3
13
C NMR resonances (ppm, carbon atoms in italics) of AMP–DEA–EG–methanol solution in the CO2 saturated solution (entry 1), in the starting solution (50% CO2 saturated)
for cyclic experiments (entry 2), in both the absorber (entry 3) and desorber solutions (entry 4) at the end of each experiment (desorption at 75 ◦ C). Percentages of DEA
carbamate with respect to the total carbonated species are reported in parentheses.

Entry Carbon atoms AMP/AMPH+ DEA/DEAH+ DEACO2 − EG EG–CO3 − MeOH MeCO3 −

R-CO2 − 164.30
R-CO3 − 159.45a 160.10a
CH2 OH 67.31 57.54 61.73 (31) 63.32 67.10
1 CH2 O 61.03
C NH2 54.99
C NH 49.76 51.03
CH3 22.12 49.16 52.65

R-CO2 − 164.40
R-CO3 − 159.31 159.48a 160.13a
CH2 OH 68.67 59.61 61.88 (76) 63.34 67.12
2
CH2 O 61.08
C NH2 53.65
C NH 50.68 51.15
CH3 23.28 49.17 52.65

R-CO2 − 164.43
R-CO3 – 159.42a 160.06a
CH2 OH 68.59 59.54 61.79 (80) 63.28 67.08
3
CH2 O 61.01
C NH2 53.63
C NH 50.60 51.07
CH3 23.25 49.15 52.64

R-CO2 – 164.45
R-CO3 – 159.43a 160.07a
CH2 OH 69.28 59.93 61.82 (93) 63.31 67.09
4
CH2 O 61.03
CNH 50.79 51.11
C NH2 52.95
CH3 23.82 17.87 14.81
a
The summed percentage of both carbonates is the difference with respect to DEA carbamate percentage.

carbonates that may overcome the formation of amine carbamate MDEA, unable to produce a carbamate derivative. A possible path-
(see, for example Fig. 1 and Table 3, entry 1: DEA carbamate 31% and way for the alkyl carbonate transformation to amine carbamate is
summed alkyl carbonates 69%). One should notice that the amine depicted in the Scheme 1C. The conversion of alkyl carbonate to
carbamate cannot react with CO2 in the absence of water and, con- carbamate caused by an excess of free amine is just the reverse of
sequently, cannot increase its 0.5 loading capacity, according to the conversion of carbamate to alkyl carbonate in the CO2 saturated
the stoichiometry of reactions (1) and (2). Interestingly, the addi- solutions (Scheme 1B) where most of the free amine has been con-
tion of free amine to each solution just containing the prevalence of sumed. The relative amount of the secondary amine (DEA, MMEA,
monoalkyl carbonates, results in the formation of greater amounts DIPA) carbamate, in turn, increases with the desorption temper-
of the amine carbamates (amine = DEA, MMEA, DIPA; see Table 3, ature and, on going from absorption to desorption, at any step of
entry 2; see also Fig. 1) and causes a decrease of the amount of desorption temperature, contrary to alkyl carbonates (Fig. 1). These
R-OCO2 − [R = CH3 , C2 H5 , C3 H7 , (CH2 )2 OH], with the exception of features indicate that the thermal decomposition of alkyl carbo-
nates [inverse of reaction (3)] occurs at a greater extent compared to
that of amine carbamate [reaction (4); AmH stands for DEA, MMEA,
DIPA)]:

AmCO2 − + AmH2 +  2AmH + CO2 (4)

Finally, we can notice that CO2 is captured by

AMP–MDEA–EG–methanol solution as alkyl carbonates, being
both AMP and MDEA unable to produce amine carbamate, at least
in solution (see Section 3.2).

3.2. Formation and reactivity of solid AMP carbamate

The solid compound which separated out from AMP–DEA–1-

propanol–CO2 system (Table 2, entry 3) was identified as the AMP
carbamate, [(AMPH)(AMPCO2 )], from its 13 C solid state NMR spec-
Fig. 1. CO2 absorption in the AMP–MMEA–EG–CH3 OH system: change of the rel- trum (Fig. 2). The nature of the compound was then confirmed by
ative amounts (% on molar scale) of MMEA carbamate and of R-OCO2 − [R = CH3 , single crystal X-ray structure determination. However, we do not
(CH2 )2 OH] as a function of the desorption temperature and CO2 /amine ratio. LOAD
further discuss here any crystallographic details because, after we
stands for the CO2 fully saturated solution and MIX for the CO2 50% saturated start-
ing solution. In the inset, ABS and DES stand for absorption and desorption solutions, had established the solid-state structure of AMP carbamate, the
respectively. crystal structure of this compound was independently published
222 F. Barzagli et al. / International Journal of Greenhouse Gas Control 16 (2013) 217–223

DEA) which give stable carbamates, 13 C NMR spectroscopy indi-

cates that AMP–CO2 − is completely converted to free AMP, whereas
AMPH+ is neutralized by the excess of the added amine [reaction
(6); AmH = MEA, DEA, MMEA]:

AMPH+ + AMPCO2 − + 2AmH → AmH2 + + AmCO2 − + 2AMP (6)

4. Conclusions

In the present study we have described a technique of CO2

separation based on chemisorption by non-aqueous solvents con-
taining AMP blends with DEA, MDEA, MMEA, DIPA and pure
alcohols (1-propanol or diethylene glycol monomethyl ether) or
mixtures of ethylene glycol with methanol or ethanol. 13 C NMR
analysis indicates that CO2 is reversibly captured in solution as
monoalkyl carbonate, R-OCO2 − , (R = CH2 CH2 OH, CH3 , C2 H5 , n-
C3 H7 ) and amine (DEA, MMEA, DIPA) carbamate. Due to the lower
stability of monoalkyl carbonates with respect to HCO3 − and to
the carbamates which are formed in the aqueous solutions, strip-
ping temperatures of 75–80 ◦ C at room pressure are sufficient to
attain absorption efficiencies greater than 90%. The relatively low
amine concentration (17–22 wt%) and the lower stripping tem-
peratures, as compared with the aqueous alkanolamines, should
also have beneficial effects by reducing the amine loss by evapo-
ration, their thermal and oxidative degradation and, additionally,
the equipment corrosion rate (Oexmann and Kather, 2010). More-
Fig. 2. Solid state 13 C CP-MAS NMR spectrum of [(AMPH)(AMPCO2 )] (100.62 MHz). over, a disadvantage of the technology based on aqueous media,
The inset expands the signal due to the CH3 groups (f). The inequivalence of the is due to the high energy demand of sorbent regeneration. It is
two methyl groups in each AMP residue results in four chemical shifts. generally accepted that about half of the total energy of the CCS
process is due to the sorbent regeneration (Rao and Rubin, 2006;
(Jo et al., 2010). Surprisingly, the CO2 -saturated solutions, either Raynal et al., 2011). The lower heat capacity of alcohols and the
during the precipitation and after the separation of the precipitate reduced evaporation of the absorbents combined with the lower
by filtration, do not contain detectable amounts of the AMP car- stripping temperature, all together should co-operate to reduce the
bamate, as inferred from inspection of the 13 C NMR spectra that energy demand of sorbent regeneration. Additionally, the higher
display only resonances ascribable to the DEA carbamate and 1- CO2 loading in the organic sorbents results in a lower circula-
propanol carbonate, besides those of the amines and the unreacted tion rate and lower sensible heat loss. In this respect, it is worth
alcohol. The proposed mechanism for the carbamate formation is stressing that any industrial scale application of this technology,
reported in Scheme 1A. A perusal of these results indicate that: (i) based on alcohol-amine absorbents, should require an accurate
the AMP carbamate is less soluble than either DEA carbamate and preliminary study of the operating and investment costs and ben-
1-propanol carbonate; (ii) the lattice energy due to the ionic inter- efits of the process, including the CO2 compression which asks
actions stabilizes [(AMPH)(AMPCO2 )] in the solid state in spite of for a substantial amount of energy (about 10% of the total energy
its unstability in solution (Sartori and Savage, 1983). CO2 desorp- cost).
tion from such slurry takes place in a separate step at increasing In the course of some absorption experiments, we have obtained
temperatures between 50 and 65 ◦ C and at room pressure until no the carbamate of AMP in the solid state whose lattice energy due
more CO2 was released and the solid phase disappeared. The aver- to the ionic bonding stabilizes this otherwise quite unstable com-
age efficiency of CO2 desorption, measured as the percent ratio pound. As a matter of fact, this salt is completely and quickly
between the desorbed CO2 and that absorbed in the saturation converted either into the monoalkyl carbonate or the bicarbonate
step, is 94% and corresponds to the regeneration efficiency. The derivatives of the protonated amine, once dissolved in alcohol or,
overall performance of this slurry appears superior to any of the respectively, in water.
solutions without precipitates, reported in Table 2, as a high amine We could admittedly conclude that a major drawback of the
regeneration efficiency is obtained at lower desorption tempera- amine-alcohol absorbent technique here described could be the
ture. However, a more detailed comparison between the two types cost of alcohols. Presumably, the greater capital cost due to alco-
of experiments cannot be carried out due to the different proce- hols, that are completely recycled, could be compensated by the
dures employed. The solid state stabilization of AMP carbamate lower amine concentration (about 20 wt%) and by the reduced
due to the lattice energy, is further supported by its instability in evaporation/thermal degradation, due to the lower operating tem-
solvents like methanol, ethanol and ethylene glycol where it instan- peratures.
taneously converts to the monoalkyl carbonate [reaction (5)], while
HCO3 − is produced in water solution:
Acknowledgements
AMPCO2 − + CH3 CH2 OH  AMP + CH3 CH2 OCO2 − (5)
The proposed conversion mechanism of AMPCO2 − to alkyl car- Financial support from DSCTM CNR (Rome) is gratefully
bonate is described by Scheme 1B. Reaction (5) can be substantially acknowledged. Thanks are also expressed to Florence Hydrolab
shifted forward or backward by adding an excess of alcohol or of (ECRF) and ENOTRIA CNR projects for supporting this research
free amine, respectively. Additionally, when solid AMP carbamate activity. Finally, thanks are expressed to Mr. Maurizio Passaponti
is dissolved in either primary or secondary amines (MEA, MMEA, for technical assistance.
 F. Barzagli et al. / International Journal of Greenhouse Gas Control 16 (2013) 217–223 223

References Jimil, G., Yogesh, P., Muthukumaru, P.S., Pradip, M., 2009. Methanol assisted selec-
tive formation of 1,2-glycerol carbonate from glycerol and carbon dioxide using
Astarita, G., Savage, D.W., Bisio, A., 1984. Gas Treating with Chemical Solvents. John Bu2 SnO as a catalyst. Journal of Molecular Catalysis A 304, 1–7.
Wiley & Sons, New York. Jo, E., Jhon, Y.H., Choi, S.B., Shim, J.-H., Lee, J.H., Jang, K.-R., Kim, J., 2010. Crys-
Ballard, M., Bown, M., James, S., Yang, Q., 2011. NMR studies of mixed amines. Energy tal structure and electronic properties of 2-amino-2-methyl-1-propanol (AMP)
Procedia 4, 291–298. carbamate. Chemical Communications 46, 9158–9160.
Bara, J.E., Carlisle, K.T., Gabriel, C.J., Christopher, J., Camper, D., Finotello, A., Gin, Kim, S.I., Chu, F., Dueno, E.E., Jung, K.W., 1999. Alkyl carbonates: efficient three
L.D., Noble, R.D., 2009. Guide to CO2 separations in imidazolium-based room- component coupling of aliphatic alcohols, CO2 , and alkyl halides in the presence
temperature ionic liquids. Industrial and Engineering Chemistry Research 48, of Cs2 CO3 . Journal of Organic Chemistry 64, 4578–4579.
2739–2751. Lepaumier, H., Picq, D., Carrette, P.L., 2009. New amines for CO2 capture, I. Mecha-
Bara, J.E., Camper, D.E., Gin, D.L., Noble, R.D., 2010. Room-temperature ionic liquids nisms of amine degradation in the presence of CO2 . Industrial and Engineering
and composite materials: platform technologies for CO2 capture. Accounts of Chemistry Research 48, 9061–9067.
Chemical Research 43, 152–159. Liu, Y.X., Jessop, P.G., Cunningham, M., Eckert, C.A., Liotta, C.L., 2006. Switchable
Barzagli, F., Mani, F., Peruzzini, M., 2009. A 13 C NMR study of the carbon diox- surfactants. Science 313, 958–960.
ide absorption and desorption equilibria by aqueous 2-aminoethanol and Ma’mun, S., Jakobsen, J.P., Svendsen, F., Juliussen, O., 2006. Experimental and
N-methyl-substituted 2-aminoethanol. Energy & Environmental Science 2, modeling study of the solubility of carbon dioxide in aqueous 30% mass 2-
322–330. ((2-aminoethyl)amino)ethanol solution. Industrial and Engineering Chemistry
Barzagli, F., Mani, F., Peruzzini, M., 2010. Continuous cycles of CO2 absorption Research 45, 2505–2512.
and amine regeneration with aqueous alkanolamines: a comparison of the effi- Mandal, B.P., Biswas, A.K., Bandyopadhyay, S.S., 2003. Absorption of carbon diox-
ciency between pure and blended DEA, MDEA and AMP solutions by 13 C NMR ide into aqueous blends of 2-amino-2-methyl-1-propanol and diethanolamine.
spectroscopy. Energy & Environmental Science 3, 772–779. Chemical Engineering Science 58, 4137–4144.
Barzagli, F., Mani, F., Peruzzini, M., 2011. A 13 C NMR investigation of CO2 Mandal, B.P., Biswas, A.K., Bandyopadhyay, S.S., 2004. Selective absorption of
absorption and desorption in aqueous 2,2 -iminodiethanol and N-methyl-2,2 - H2 S from gas streams containing H2 S and CO2 in aqueous solutions of
iminodiethanol. International Journal of Greenhouse Gas Control 5, 448–456. N-methyldiethanolamine and 2-amino-2-methyl-1-propanol. Separation and
Bates, E.D., Mayton, R.D., Ntai, L., Davis, J.H., 2002. CO2 capture by a task specific Purification Technology 35, 191–202.
ionic liquid. Journal of the American Chemical Society 124, 926–927. Mani, F., Peruzzini, M., Stoppioni, P., 2008. Combined process of CO2 capture by
Bonenfant, D., Mimeault, M., Hausler, R., 2003. Determination of the structural potassium carbonate and production of basic zinc(II) carbonates: CO2 release
features of distinct amines important for the absorption of CO2 and regener- from bicarbonate solutions at room temperature and pressure. Energy Fuels.
ation in aqueous solution. Industrial and Engineering Chemistry Research 42, 21, 1714–1719.
3179–3184. Minakata, S., Sasaki, I., Ide, T., 2010. Atmospheric CO2 fixation by unsaturated alco-
Böttinger, W., Maiwald, M., Hasse, H., 2008. Online NMR spectroscopic study of hols using t-BuOI under neutral conditions. Angewandte Chemie International
species distribution in MDEA–H2 O–CO2 and MDEA–PIP–H2 O–CO2 . Industrial Edition 49, 1309–1311.
and Engineering Chemistry Research 47, 7917–7926. Oexmann, J., Kather, A., 2010. Minimising the regeneration heat duty of post-
Bottoms, R.R., 1931. Organic bases for gas purification. Industrial and Engineering combustion CO2 capture by wet chemical absorption: the misguided focus on
Chemistry 23, 501–504. low heat of absorption solvents. International Journal of Greenhouse Gas Control
Breitmaier, E., Voelter, W., 1990. Carbon-13 NMR Spectroscopy, 3rd ed. VCH- 4, 36–43.
Weinheim, Germany. Oyenekan, B.A., Rochelle, G.T., 2006. Energy performance of stripper configurations
Camper, D., Bara, J.E., Gin, D.L., Noble, R.D., 2008. Room-temperature ionic liquid- for CO2 capture by aqueous amines. Industrial & Engineering Chemistry Research
amine solutions: tunable solvents for efficient and reversible capture of CO2 . 45, 2457–2464.
Industrial and Engineering Chemistry Research 47, 8496–8498. Park, S.W., Lee, J.W., Choi, B.S., Lee, J.W., 2005. Kinetics of absorption of carbon diox-
Caplow, M., 1968. Kinetics of carbamate formation and breakdown. Journal of the ide in monoethanolamine solution of polar organic solvents. Journal of Industrial
American Chemical Society 90, 6795–6803. and Engineering Chemistry 11, 202–209.
Choi, W.-J., Cho, K.-C., Lee, S.-S., Shim, J.-G., Hwang, H.-R., Park, S.-W., Oh, K.-J., Park, J.Y., Yoon, S.J., Lee, H., 2003. Effect of steric hindrance on carbon dioxide absorp-
2007. Removal of carbon dioxide by absorption into blended amines: kinet- tion into new amine solutions: thermodynamic and spectroscopic verification
ics of absorption into aqueous AMP/HMDA, AMP/MDEA, and AMP/piperazine through solubility and NMR analysis. Environmental Science and Technology
solutions. Green Chemistry 9, 594–598. 37, 1670–1675.
Davison, J., 2007. Performance and costs of power plants with capture and storage Rao, A.B., Rubin, E.S., 2006. Identifying cost-effective CO2 control levels from amine-
of CO2 . Energy 32, 1163–1176. based CO2 capture systems. Industrial and Engineering Chemistry Research 45,
Freeman, S.A., Davis, J., Rochelle, G.T., 2011. Degradation of aqueous piperazine 2421–2429.
in carbon dioxide capture. International Journal of Greenhouse Gas Control 4, Raynal, L., Bouillon, P.-A., Gomez, A., Broutin, P., 2011. From MEA to demixing sol-
756–761. vents and future steps, a road map for lowering the cost of post-combustion
Goff, G.S., Rochelle, G.T., 2006. Oxidation inhibitors for copper and iron catalyzed carbon capture. Chemical Engineering Journal 171, 742–752.
degradation of monoethanolamine in CO2 capture processes. Industrial and Reynolds, A.J., Verheyen, T.V., Adeloju, S.B., Meuleman, E., Feron, P., 2012. Towards
Engineering Chemistry Research 45, 2513–2521. commercial scale postcombustion capture of CO2 with monoethanolamine sol-
Hart, R., Pollet, P., Hahne, D.J., John, E., Llopis-Mestre, V., Blasucci, V., Huttenhower, vent: key considerations for solvent management and environmental impacts.
H., Leitner, W., Eckert, C.A., Liotta, C.L., 2010. Benign coupling of reactions and Environmental Science and Technology 46, 3643–3654.
separations with reversible ionic liquids. Tetrahedron 66, 1082–1090. Sartori, G., Savage, D.W., 1983. Sterically hindered amines for CO2 removal from
Heldebrant, D.J., Yonker, C.R., Jessop, P.G., Phan, L., 2008. Organic liquid CO2 capture gases. Industrial and Engineering Chemistry Fundamentals 22, 239–249.
agents with high gravimetric CO2 capacity. Energy & Environmental Science 1, Stueber, D., Patterson, D., Mayne, C.L., Orendt, A.M., Grant, D.M., Parry, R.W., 2001.
487–493. Carbonates, thiocarbonates and the corresponding monoalkyl derivatives, 1.
Hook, R.J., 1997. An investigation of some sterically hindered amines as poten- Their preparation and isotropic 13 C NMR chemical shifts. Inorganic Chemistry
tial carbon dioxide scrubbing compounds. Industrial and Engineering Chemistry 40, 1902–1911.
Research 36, 1779–1790. Sun, W.-C., Yong, C.-B., Li, M.-H., 2005. Kinetics of the absorption of carbon dioxide
Idem, R., Wilson, M., Tontiwachwuthikul, P., Chakma, A., Veawab, A., Aroonwilas, A., into mixed aqueous solutions of 2-amino-2-methyl-l-propanol and piperazine.
Gelowitz, D., 2006. Pilot plant studies of the CO2 capture performance of aque- Chemical Engineering Science 60, 503–516.
ous MEA and mixed MEA/MDEA solvents at the University of Regina CO2 Capture Teller, A.J., Ford, H.E., 1958. Packed column performance of carbon dioxide-
Technology Development Plant and the Boundary Dam CO2 Capture Demonstra- monoethanolamine system. Industrial and Engineering Chemistry 50,
tion Plant. Industrial and Engineering Chemistry Research 45, 2414–2420. 1201–1206.
Jamal, A., Meisen, A., Lim, C.J., 2006. Kinetics of carbon dioxide absorption and Xu, H.J., Zhang, C.F., Zheng, Z.S., 2002. Solubility of hydrogen sulfide and carbon diox-
desorption in aqueous alkanolamine solutions using a novel hemispherical con- ide in a solution of methyldiethanolamine mixed with ethylene glycol. Industrial
tactor. II. Experimental results and parameter estimation. Chemical Engineering and Engineering Chemistry Research 41, 6175–6180.
Science 61, 6590–6603. Yeh, J.T., Pennline, H.W., Resnik, K.P., 2001. Study of CO2 absorption and desorption
Jassim, M.S., Rochelle, G.T., 2006. Innovative absorber/stripper configurations for in a packed column. Energy & Fuels 15, 274–278.
CO2 capture by aqueous monoethanolamine. Industrial and Engineering Chem- Yoon, S.J., Lee, H., 2003. Substituent effect in amine–CO2 interaction investigated
istry Research 45, 2465–2472. by NMR and IR spectroscopies. Chemistry Letters 32, 344–345.