Environ. Sci. Technol.
2009, 43, 2575–2580
H2CO3(s): A New Candidate for CO2
Capture and Sequestration
J. A. TOSSELL*
Department of Chemistry and Biochemistry, University of
Maryland College Park, Maryland 20742, and The Renewable
Energy Institute, George Washington University,
Washington, D.C. 20052
Received August 26, 2008. Revised manuscript received
February 4, 2009. Accepted February 6, 2009.
To reduce the magnitude of anthropogenic global warming it
is necessary to remove CO2(g) from the effluent streams of coal-
fired power plants and to sequester the CO2 either as a liquid
or by reaction with other compounds. A major difficulty in achieving
this goal arises from the very weak acidity of CO2(g), causing
it to react only incompletely with weak bases, although this weak
interaction does provide a means for “stripping” the CO2 from
the acid-base complex at high temperatures. Reaction with strong
bases like NaOH yields more stable complexes, but massive
amounts of chemical reactants would need to be purchased and
chemical products like NaHCO3 then stored. However, when gas-
phase CO2 reacts with the weak base water (or when
bicarbonate reacts with strong acid) the unstable product
monomeric “H2CO3” can be formed. The free energy required is
about 16 kcal/mol in the gas phase and about 10 kcal/mol in
aqueous solution. This energy can be supplied by particle or
photon excitation and is only a small fraction of the energy released
when a mole of CH4 is converted to a mole of CO2. Although
this monomeric compound is highly unstable, its oligomers are
considerably more stable, due to internal H-bonding, with
free energies for the larger oligomers in the gas phase which
are about 4 kcal/(mol of H2CO3) lower, only about 6 kcal/mol H2CO3
higher than the gas-phase combination of CO2 and H2O at
room temperature. Also, at lower temperature the entropic penalty
for the oligomer is less and oligomeric H2CO3 becomes stable
around the sublimation temperature of dry ice. This indicates that
it may be possible to capture gas-phase CO2 directly, using
only cheap and abundant H2O as a reactant, and to store the
resulting (H2CO3)n as a oligomeric solid at only moderately cold
temperatures. These conclusions are based on quantum
computations that accurately reproduce the structures, spectra,
and stabilities of H2CO3 oligomers. Methods for producing and
characterizing the H2CO3 oligomers are discussed. However, some
aspects of the proposed scheme are quite speculative and
will require additional investigation. Several important questions
need to be answered before the feasibility of this procedure
on a planetary scale can be assessed, particularly those involving
the vapor pressure curve, heat of sublimation, density, and
compressibility of (H2CO3)n.
Introduction
The existence of H2CO3 as a discrete molecular species has
long been controversial. It has been tentatively identified in
* Author phone: 301 405 1868; fax: 301 314 9121; e-mail: Tossell@
umd.edu.
10.1021/es802393s CCC: $40.75  2009 American Chemical Society
Published on Web 03/03/2009
the gas phase, but only by a peak in the mass spectrum (1)
of the vapor existing over heated solid NH4HCO3. There is,
however, indirect evidence that solid H2CO3 can be sublimed
or otherwise decomposed and then reconstituted from the
gas phase (2). A feature in the Raman spectrum of acidified
bicarbonate solutions has been assigned (3) to H2CO3, but
this feature is almost coincident with one of the frequencies
of HCO3-, a somewhat unexpected result. A recent study of
the Raman spectrum of carbonate systems species in solution
(4) found no peak that could be attributed to H2CO3(aq). An
early 13C NMR study (5) of solutions of CO2 in water did not
identify any peaks other than those attributable to CO2,
HCO3-, and CO3-2, although the NMR shift expected for
H2CO3 could only be estimated at that time. The 13C NMR
of carbonates in superacid solutions has been determined
and some of the observed features have been attributed to
H2CO3 and H3CO3+, but the calculated shieldings for the
assigned species did not match the experimental trends (6).
Recently, the 13C NMR of monomethycarbonic acid (but not
carbonic acid itself) has been determined (7).
It is commonly accepted within the aqueous geochemistry
community that “carbonic acid” solution, produced by
dissolving CO2 in water, contains only a fraction of a percent
of actual H2CO3 molecules (8). An acidity value has been
measured (9) for a transient species produced by the
acidification of sodium carbonate solutions, which may be
monomeric H2CO3. This species is considerably more acidic
(with a pKa around 3) than is carbonic acid itself (with a pKa
around 6). This smaller pKa of the transient is consistent
with its instability in solution and with our recent calculations
(10).
Previous Experiments on Solid H2CO3. In the early 1990s,
proton irradiation of icy mixtures of CO2 and H2O produced
(11) a solid whose IR features seemed qualitatively consistent
with those expected for H2CO3, particularly in the CdO
stretching and O-H stretching regions. Subsequent studies
showed that a number of different synthetic approaches
(12-16), some based on irradiation of cryogenic ices and
some based on acidification of aqueous bicarbonates under
cryogenic conditions, could produce a solid with the same
spectral characteristics, although small quantitative differ-
ences suggested that there were actually two different
polymorphs of the solid material, which have been labeled
R- and β-H2CO3(s). Recently H2CO3-like species have been
identified on the surface of acid-treated wet calcite, primarily
on the basis of their vibrational spectra (17, 18). More recently,
UV irradiation has been used to produce solid H2CO3 from
mixed CO2:H2O ices using 10 eV photons (19). Electron
irradiation of CO2:H2O ices has also been shown (20) to
produce a material with the characteristic IR spectrum of
solid H2CO3. There is also strong evidence (21) that the two
different crystal forms of solid H2CO3, R and β, can be formed
from two different amorphous forms of the material, sug-
gesting little structural change between the amorphous and
crystalline materials.
Previous Calculations on H2CO3. Several quantum theo-
retical studies (22-26) have shown that gas-phase monomeric
H2CO3 has a less negative free energy than the reactants CO2
+ H2O, but quantum studies have also indicated a very
substantial activation energy barrier (more than 40 kcal/
mol) for the decomposition of H2CO3 in the absence of water
(27), which was greatly reduced (to around 20 kcal/mol) when
several explicit water molecules were added (27-30). Thus,
the H2CO3 monomer, while not thermodynamically stable,
may well be kinetically stable at reasonably low temperatures
and/or in the absence of water. In fact, in ref 27, entitled “On
VOL. 43, NO. 7, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 2575
TABLE 1. Calculated Relative Energies at T = 298 K per H2CO3 Unit (in kcal/mol) for H2CO3 Monomer, Oligomers, and 1D Crystal
in Gas Phase and Aqueous Solution Compared to the s-a Isomer of Monomeric H2CO3
∆Eg CBSB7 B3LYP ∆GVRT ∆Eg CBSB7 B3LYP + ∆GVRT
H2CO3 s-a isomer 0 0
(H2CO3)2 -11.9 +6.6
(H2CO3)3 -14.6 +8.3
(H2CO3)4 -15.9 +9.3
(H2CO3)∞ 1D solid -19.1 - -
a
na ) not attempted due to computational limitations.
the surprising kinetic stability of carbonic acid (H2CO3)”, a
series of pictures is shown of the solid compound held on
a fingertip for a period of about 2 min.
One study calculated the (harmonic) vibrational spectrum
(26) of H2CO3 at a high quantum mechanical level (6-311+G**
MP2) and uncovered systematic quantitative disparities with
the experimental data for solid H2CO3, which the authors
attributed to unspecified medium effects. Note that the only
data available to characterize the solid H2CO3 species were
these IR spectra. The stability of some oligomers of H2CO3
has also recently been studied (30, 31) using both correlated
post-Hartree-Fock and DFT methods. The small H-bonded
oligomers studied were found to have energies per H2CO3
unit that were generally lower than that for monomeric
H2CO3. The DFT study (31) also showed that the calculated
CdO and O-H stretching frequencies were lower in the
oligomers than in the monomer, but the accuracy of the
results was not considered sufficient to definitively reassign
the experimental spectra.
We carried out a computational study (32) of the acidity
of a number of aqueous oxyacids, such as B(OH)3, H2CO3,
and HNO3. We found that the complex CO2 · · · H2O was the
predominant species in “carbonic acid”, roughly a thousand
times more abundant than H2CO3, in agreement with
experiment. Of the three conformers of monomeric H2CO3,
s-s, s-a, and a-a, we found the a-a to be most stable in the
gas-phase, but in aqueous solution the a-a and s-a conformers
were within a couple of kilocalories/mole of each other, due
to a more negative hydration energy for the less symmetric
s-a isomer. The stability of the monomer was increased in
aqueous solution, compared to hydrated CO2 and H2O,
because of strong H-bonding.
Our recent studies (10) of H2CO3 oligomers using high-
level methods have established that the oligomers are indeed
of lower free energy (per H2CO3 unit) than the corresponding
monomers in the gas phase. We presented energetic results
for the monomer, dimer, trimer, and infinite periodic 1D
crystal. These results will be updated and expanded below.
For the monomer and dimer, we could employ the CBS-QB3
method to determine that the free energy per H2CO3 unit
was lowered by about 4.5 kcal by dimerization in the gas
phase. However, in aqueous solution, simulated by the PCM,
this additional stability is lost, because the hydration free
energies (per H2CO3 unit) of the oligomers are of smaller
magnitude than that of the monomer. This is a consequence
of the change from intermolecular H2CO3-water to in-
tramolecular H bonding. The O-H groups that form in-
tramolecular H bonds to hold the oligomers together are less
available for H-bonding to the solvent. Although we were
restricted to less accurate methods for the trimer and 1-D
polymer, the calculated internal energies indicate additional
stablilization compared to the dimer. This suggests that the
polymeric material may be stable enough to capture ad-
ditional CO2 and provide a stable storage site for it.
Our calculations of the IR and Raman spectra, employing
both high level harmonic methods and anharmonic cor-
rections, established that the spectra of oligomers like the
dimer or trimer give a much better match to the spectrum
2576 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 7, 2009
∆GCBS-QB3 ∆GCBS-Lq ∆GCPCM ∆Gaq ) ∆GCBS-Lq + ∆GCPCM
0 0 0 -14.5 -14.5
-5.3 -4.5 -3.1 -6.7 -9.8
-6.3 naa -3.6 -6.9 -10.5
-6.6 na -3.9 -3.4 -7.3
- - - -
of H2CO3(s) than does that of the monomer. In particular, to
match the CdO stretching and O-H stretching frequencies
observed it is necessary to employ an oligomeric model and
also to utilize high-level methods for determining the
harmonic frequency and to incorporate anharmonic effects.
Since the accuracy of the methods applied to the trimer could
not be as high as that applied to the dimer, we could not
determine if the fit of calculated and experimental vibrational
spectra actually improved for the higher oligomers compared
to the dimer. However, we did establish definitively that
H2CO3(s) observed experimentally have oligomers as their
basic local structural unit.
Computational Methods
For a detailed description of the computational methods,
see the Supporting Information.
Results and Discussion
New Energetic Results for Oligomers. In the present study
we extend our calculations on the carbonic acid system to
higher oligomers of H2CO3, including (H2CO3)3 and two
different isomers of (H2CO3)4. Pictures of several oligomers
are shown in Figure 1 of the Supporting Information. Most
of our energetic results are presented in Table 1. Since some
of the molecules are quite large, we had to use the less
computationally demanding but also less accurate version
of the CBS method, CBS-Lq, to obtain energetics across the
series (H2CO3)n, n ) 1-4.. Using this method, we find that
there is still a substantial lowering of free energy per H2CO3
unit in the first oligomerization step (although slightly smaller
than for CBS-QB3) and slightly larger additional lowerings
(per H2CO3 unit) for the later steps. Thus, the overall energy
lowering in the oligomer compared to a collection of
monomers increases slightly faster than linear with n. The
electronic energy (per H2CO3 unit) also becomes more
negative as n increases, with a lowering of 3.2 kcal/(mol of
H2CO3) between (H2CO3)4 and the 1D infinite polymer of
H2CO3. Perhaps more relevant is the sum of the electronic
energy Eg and ∆GVRT, the VRT contribution to the free energy,
which incorporates both zero point effects and the entropic
penalty arising from the oligomerization. This becomes
continuously more negative as we go to the higher oligomers.
We have also calculated a number of the components of the
free energy, to demonstrate the competition between the
negative enthalpy changes (∆HCBSLq) and the positive VRT
contributions to the free energy (∆G VRT,CBSLq), arising from
the reduced entropy of the oligomers compared to a collection
of monomers. The change in gas-phase ∆GCBSLq with size of
the oligomer is thus a result of a compensation between
enthalpic and entropic effects, as for a normal addition
polymerization. But in this case, the lowering of enthalpy is
driven by H-bonding rather than covalent bond formation.
The CPCM solution corrections to the free energy also become
less negative as the oligomer increases in size as a larger
fraction of the H-bonds become intermolecular, as do the
aqueous free energies of reaction, ∆GCBSLq + ∆GCPCM,CBSLq,
which combine the gas phase and hydration contributions.

TABLE 2. Calculated Variation with Temperature of Gas-Phase
Vibrational, Translation and Rotational Contributions to Free
Energy, ∆GVRT,CBSLq, for Oligomers vs That for the s-a
Monomera
n 0K 100 K 200 K 273 K 298.15 K 373 K
2 1.00 2.87 4.94 6.44 6.95 8.48
3 1.19 3.60 6.25 8.16 8.81 10.76
4 1.27 3.97 7.08 9.12 9.86 12.03
a
All energies in kcal/(mol of H2CO3) for (H2CO3)n, n)2, 3,
4.
All the oligomers are less stable than the monomer in aqueous
solution because of the loss of hydration free energy. For
example, the total calculated CPCM hydration energy of
H2CO3 is -14.5 kcal/mol, while the total calculated value for
(H2CO3)4 is only -13.6 kcal/mol (or -3.4 kcal/(mol of H2CO3)).
The equilibrium between the monomer and the oligomers
is also calculated to be strongly T-dependent, since the
entropic penalty for the formation of the oligomers increases
rapidly with T. This is shown in Table 2, where we evaluate
the vibrational, rotational, and translational contribution to
the gas-phase free energy within the CBS-Lq method,
∆GVRT,CBSLq (per H2CO3 unit), minus that for the s-a monomer
as a function of T for the n ) 2, 3, and 4 oligomers. For
example, we calculate ∆GVRT,CBSLq of 9.60 and 33.10 kcal/mol
for the monomer and the dimer at T ) 298.15, giving a
difference in the normalized value of this free energy
contribution of 6.95 kcal/(mol of H2CO3), as seen in Table
2. It is apparent that this quantity increases by about 6-9
kcal between 0 K and room temperature. Thus, the oligomers
are strongly destabilized compared to the monomer at high
T. This of course explains why solid oligomeric H2CO3 is stable
at very low temperatures and dissociates to the gas-phase
monomer at temperatures above 200-240 K.
For the gas-phase reaction
1 (50). The resultant OH radicals can react with CO2 to form
(H CO ) chain isomer ⇒ H2CO3 s-a isomer
4 2 34
we obtain total free energy changes of +17.1, +12.4, +6.6,
and +0.6 kcal as we increase T along the series 0, 100, 200,
and 298.15 K. Our calculated numbers suggest that the
oligomer is stable to about 300 K. Experimentally, solid H2CO3
in a cryoscopic environment on a metal surface is observed
to be stable up to at least 200-240 K, implying that the free
energy for a reaction like that above changes sign from
positive to negative around this value. Thus, the energetics
we calculate for the above reaction probably differ from the
experimental observation by on the order of 5 kcal/(mol of
H2CO3). This implies that either (1) CBS-Lq is simply not
sufficiently accurate for describing the energetics of oligo-
merization, (2) the real H2CO3(s) is a more stable compound
than the simple tetramer (H2CO3)4, or (3) the (H2CO3)4 is
stabilized within the cryoscopic environment, either by
intermolecular H-bonding or interaction with the substrate.
In general, then, stabilization of carbonic acid oligomers
could be accomplished at low T in the same way as for solid
CO2 (dry ice, which sublimes at about 200 K at P ) 1 atm)
and similar energetic and containment considerations would
apply.
Implications of H2CO3 Oligomer Stability. We think that
the stability of H2CO3 may be of more than merely academic
interest, since it represents a thermodynamically, or at least
kinetically stable, reaction product of CO2 and H2O. Methods
have been used in industry for some time to remove CO2
dissolved in aqueous systems (45), using amines like mo-
noethanolamine (MEA). However, theoretical understanding
of such processes is poor; the reaction products and
mechanisms have yet to be definitively determined (46). The
process of breakdown for this product to produce pure CO2
requires elevated temperatures, such that the complex
becomes unfavored entropically at the high temperatures
employed. The CO2(g) extracted must then be converted to
a condensed phase at high P and low T before it can be
transported and/or stored. The generation of both high
temperature and high pressure requires energy. The increase
in cost of electrical power generation associated with this
type of capture and sequestration is often estimated to be
in the range of at least 30-40%, but since this process has
never been applied on a massive scale, even this estimate
remains uncertain (47). Whiteside and Crabtree (48) have
suggested that the magnitude of environmental problems
arising from increased CO2 might be reduced by a more
complete understanding of the chemical transformations of
CO2. Indeed, a number of researchers are systematically
exploring the chemical reactivity of CO2 (e.g., ref 49), so as
to expand the range of useful products obtainable from it,
increasing its economic value and making its removal more
favorable economically.
It is now worthwhile to explore the energy requirements
for the formation of H2CO3(s) starting with gas-phase CO2 +
H2O or with CO2:H2O ices. This was of little consequence in
laboratory experiments focused upon producing small
amounts of a unique material, but it becomes critically
important if such a laboratory process is to be scaled up to
industrial (or planetary) size. Since the H2CO3 monomer has
a less negative free energy than CO2 + H2O in the gas phase,
energy must be supplied in order to produce it. Although the
oligomers and H2CO3(s) are more stable than the H2CO3
monomer, they are still unstable with respect to CO2:H2O
ices, since these ices do not spontaneously react to form
H2CO3(s). The energy may come from irradiation by protons
or other particles or by absorption of photons. Experimental
UV absorption studies have used 10 eV photons. These have
about twice the energy of 5.10 eV needed to homolytically
dissociate water to H + OH in either the gas or liquid phase
HCO3. Our calculations indicate this process has a free energy
change close to zero. The H radicals formed can then reduce
HCO3, giving H2CO3. It may well be that the limiting energy
needed for UV irradiation to produce solid H2CO3 from CO2:
H2O ices may be that needed for the homolysis of water, but
the dependence of yield on the light frequency employed
has not yet been measured.
Alternatively, one can start with HCO3- in aqueous
solution obtained from the reaction of CO2 and an alkali
metal carbonate. This could then be protonated with strong
acid and irradiated. However, the energy and materials costs
of this process may well be very large for a planetary scale
process, due to the need to purchase the alkali carbonate
reactant and the HCl.
However, all these energies should be weighed against
that obtained from the oxidation of 1 mol of CH4 (with
atmospheric O2) to produce 1 mol of CO2. The calculated ∆G
for this reaction is given in Table 3 and has a value of -190.8
kcal/mol at the CBS-QB3 level. We can think of this as our
original energy output and analyze what reductions in
magnitude from this value are necessary for the creation of
(H2CO3)n and its storage. If this is only a small percentage of
energy obtained in forming the CO2, then the overall process
of formation, capture, and sequestration will be economically
feasible. Qualitatively, to reach hydrated H2CO3 from gas-
phase H2O and CO2 (by some more efficient catalysis method)
requires about 10 kcal/mol, while going over the activation
energy barrier in aqueous solution would require about 20
kcal/mol and going over the free-molecule activation energy
barrier would require almost 50 kcal/mol. For the aqueous
VOL. 43, NO. 7, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 2577
TABLE 3. Calculated Free Energies (in kcal/mol at 298K) for Indicated Reactions in Aqueous Solutiona
reaction
CH4 + 2 O2 w CO2 + 2 H2O
CO2 + H2 w HCOOH
2HCOOH w (HCOOH)2
CO2 + H2O w H2CO3
CO2 + NH3 + H2O w NH4+ + HCO3-
CO2 + C2H4OHNH2 w C2H4OHNHCO2H
a
Experimental free energy changes from refs 54-56 and NIST database are in parentheses.
solution case, we are thus losing about 10% of the energy
generated in forming CO2 by oxidation of CH4.
The energies of other reactions, obtained at the CBS-QB3
+ CPCM level with CPCM corrections for hydration, are also
given in Table 3. The reaction of CO2 with H2 to produce
HCOOH and the dimerization of HCOOH are used primarily
as a test of our methodology. For the HCOOH formation
reaction agreement of calculation with tabulated experi-
mental thermodynamic data for both the gas phase and
aqueous solution is reasonable. For the dimerization of
HCOOH, we calculate ∆H ) -14.9 kcal, while the range of
experimental values is quoted as -13.0 to -15.3 for two
slightly different methods (see Table 3 for references). Of
course, the gas-phase ∆G is much more positive than ∆H.
There is also strong differential stabilization of the monomer
compared to the dimer in the CPCM calculations, so that the
dimerization reaction in aqueous solution has a positive free
energy change. Thus, HCOOH is a dimer in the gas phase
and a monomer in aqueous solution. As described earlier,
the reaction of CO2 + H2O to produce H2CO3 has a
significantly positive free energy in the gas-phase, but this
is reduced somewhat in aqueous solution, since the H2CO3
is strongly hydrated.
Synthesis of H2CO3 Oligomers. The energy requirements
for the formation of H2CO3 oligomers from monomeric H2CO3
are much more favorable. In particular, for each step the free
energy per H2CO3 unit in the gas phase (at the CBS-Lq level)
goes down (see Table 1). Once an oligomeric H2CO3 template
has been established, addition of further monomers becomes
increasingly favorable energetically, at least for n ) 2-4.
How might one create such a template? A speculative
possibility is to create a self-assembled monolayer (SAM) of
alkanethiol groups, connected to a metal surface on one end
by a M-S bond and with a carboxylic acid group of the other
end. H2CO3 H-bonds more strongly to carboxylic acid groups
than to itself. For example, we calculate the gas-phase free
energy for the reaction of HCOOH and H2CO3 (a-a isomer)
to create a H-bonded complex as -15.0 kcal. Thus, a SAM
terminated by COOH groups might be a useful template for
the oligomer, in the laboratory or perhaps in industrial
production.
Although we have calculated properties for only a fragment
of the crystal structure of H2CO3, we can crudely estimate
the density of the material. We can use the VOLUME keyword
in Gaussian to obtain a molar volume for (H2CO3)2 and
combine that with the molecular mass to obtain a density
of about 1.7 g cm-3. This is somewhat larger than the density
of 1.18 g cm-3 tabulated for liquid CO2. Thus, the volume
required to store the solid H2CO3 oligomer would be on the
order of magnitude of that required for liquid CO2 (admittedly
a quite enormous figure for all coal-fired power plants).
However, this is a very crude estimate and is one of the most
important parameters to be determined experimentally.
A chemical process that reacted CO2 to produce a more
stable product like (H2CO3)n would incur only the costs of
purchasing the necessary reactant(s) and storing the prod-
uct(s) produced. If the other reactant was very cheap, like
H2O, and the product of reasonably high density, like
2578 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 7, 2009
∆G CBS-QB3 (gas-phase) ∆GCPCM ∆Gaq
-190.8 -
+11.8 (+10.3) -10.3 +1.5 (-0.9)
-3.8 (-3.3) +13.3 +9.5
+16.5 -6.4 +10.1
+148.5 -143.8 +4.7
+5.9 -6.2 -0.3
TABLE 4. Calculated Values of 13C NMR Shieldings for
Carbonate C Atoms in Carbonic Acid Monomers and
Oligomers, (H2CO3)n, Obtained Using the GIAO HF Method with
a 6-311+G(2d,p) Basis Set
n σC (ppm)
1, s-a 37.5
2 25.3
3 inner 21.2, outer 25.6
4, chain inner 21.4, outer 26.2
4, dimer of dimers 24.4, 24.4, 25.6, 25.6
Si(CH3)4, reference zero for C shieldings 192.3
H2CO3(s), than onsite storage, rather than transportation to
a remote storage site, might be possible. This would reduce
transportation costs (although cooling and containment costs
would still be incurred) and would not require the use of
storage sites in the Earth’s ocean or crust. Since storage in
abandoned wells, aquifers, etc., while probably of low risk,
might invoke a strong local political response, onsite storage
could vastly reduce societal and political opposition to the
overall CO2 sequestration process.
Further Characterization of Oligomers. In order to
facilitate such a process, it would be valuable to more
completely specify the midrange order of the various forms
of H2CO3(s). This could help to clarify the possible ranges of
density of the solid material. Although vibrational spectros-
copy can be very valuable in identifying structural features
in such compounds, the information given is quite local in
nature. Determining the degree of oligomerization from
vibrational spectra is difficult, as discussed in our previous
study (10) on H2CO3 oligomers. Changes in vibrational spectra
with degree of oligomerization are subtle, and calculating
spectra of comparable accuracy for oligomers of different
size can be difficult. In constrast, NMR spectroscopy has
often demonstrated its potential to give information on the
midrange structure of materials, such as second nearest
neighbor effects, modifications due to ring or cage size, etc.
(51). Our calculations on the H2CO3 oligomers indicate that
the 13C shift of the C atom is somewhat diagnostic of the size
of the oligomer, as shown in Table 4. It at least allows us to
distinguish between the dimer and the higher oligomers,
and the intensity ratios of the inner and outer C peaks in the
higher oligomers could give further information on oligomer
size. Thus, if we generate a range of different size oligomers,
we could see a range of shieldings in the carbonate region
of the 13C NMR spectra. This would allow us to test our
calculations of the relative energetics of the oligomers and
to refine our basic structural concept for (H2CO3)n(s).
Stability and Density of Oligomers. Of course, one general
difficulty with the creation of carbonic acid oligomers is that
any oligomeric H2CO3(s) would likely still remain slightly
higher in free energy energy than gas-phase CO2 + H2O, even
if it were kinetically stable. Therefore, decomposition of this
species to CO2 + H2O would release energy.
Our calculations of the energies for the interactions of
dimers indicate unfavorable but rather weak interactions,
such that the dimers could be stored rather tightly. For
example a “dimer of dimers”, as shown in Figure 1 of the
Supporting Information, is higher in free energy than two
isolated dimers by only about 2.1 kcal/(mol of (H2CO3)2) at
the CBS-Lq level. However, this is a point that needs to be
further explored for other geometries and at higher quantum
mechanical levelssother approach geometries than the one
considered may be more favorable. A critical consideration
would be whether linear chains of H2CO3 could be packed
more closely (i.e., compressed) without incurring a large
energy penalty. Minimizing the storage space required for
liquid or supercritical CO2 is a similar problem, but of less
importance if large geological features are to be used for
storage, rather than on-site storage. Of course, a fallback
position would be to use geologic storage to hold the (H2CO3)n.
Effect of Release of H2CO3 Oligomers to the Atmosphere.
A very different approach would be to create solid H2CO3
using whichever is the most energy efficient of the low-
temperature synthesis techniques and then to simply allow
it to warm and volatilize. As we have noted, some experi-
mental evidence (ref 2) indicates that the gas phase species
produced can be recondensed to the solid, indicating that
it remains as H2CO3 and does not quickly decompose to CO2
+ H2O in the gas phase. Theoretical studies indicate that
H2CO3 has a high activation energy barrier to decomposition
in the absence of water. Thus, H2CO3 and its oligomers may
have a long lifetime in some regions of the Earth’s atmo-
sphere. Our calculations (10) indicate that H2CO3 and
(H2CO3)2 would both have pKa’s around 3, rather than the
values near 6-7 characteristic of CO2(aq), the dominant acidic
species in normal rainwater. Thus, H2CO3 and its oligomers
would be moderately strong weak acids, comparable to HF.
At a pH of 5-6 in unpolluted rainwater the dominant species
arising from H2CO3 would be HCO3-.
However, inspection of the calculated IR spectrum of
H2CO3 or its dimer indicates that they have substantial
absorption in the region around 500-600 cm-1, as is shown
in Tables 1 and 3 of ref 10. CO2 of course shows a O-C-O
bending mode absorption around 660 cm-1, which is
responsible for most of its effect on the Earth’s radiation
balance (52). In both H2CO3 and (H2CO3)2 such frequencies
are also associated with O-C-O bending vibrations. The
calculated IR spectrum of HCO3- also shows a C-O-C bend
around 660 cm-1, although it is of low intensity. There is also
an intense absorption in HCO3- at 933 cm-1 (within the
900-1000 cm-1 atmospheric window (53), corresponding to
C-O stretching combined with C-O-C bending. If HCO3-
interacts with cations such as NH4+, it could conceivably
precipitate as a solid.
The great advantage of this approach is that the H2CO3
oligomer species would not actually have to be sequestered
but would be allowed to escape into the atmosphere, so no
expenses or problems arising from transportation and storage
in the geosphere or hydrosphere would arise. However, our
calculations indicate that the H2CO3 monomer and dimer
and HCO3- all absorb at several frequencies within the
blackbody radiation spectrum from the Earth’s surface
(between about 200 and 1500 cm-1) and specifically within
the window (between about 830 and 1430 cm-1) where CO2
and H2O do not absorb. Thus, these species might contribute
as much or more to radiative forcing as would CO2 itself.
We have calculated the free energy for the transition state
in the decomposition of (H2CO3)2 to 2H2CO3. This is a
symmetric structure with partial H-bonds (see Figure 1 in
Supporting Information) that is only 4.8 kcal above the
equilibrium structure of (H2CO3)2 and only 0.3 kcal above
the free energy of two H2CO3 s-a. Thus, the activation energy
in the absence of water would be 4.8 kcal/mol for dissociation
to the monomer. This is only a small fraction of the gas-
phase activation energy for dissociation of H2CO3 to CO2 +
H2O.
Thus, our results indicate that (H2CO3)2 or a larger oligomer
is the material previously identified as solid H2CO3. This
material is stable at temperatures up to around 200-240 K,
at least in the absence of water. It can be formed directly
from CO2(g) and H2O by freezing a CO2:H2O mixture,
irradiating it with electrons, ions, or photons, and then slightly
rewarming it to allow reaction to H2CO3(s). It is only a few
kilocalories/(mole of H2CO3) less stable (in terms of free
energy) than CO2(g) plus H2O(g). Although the energy barrier
to decomposition of the dimer to monomers is fairly small,
the free energy of activation for the decomposition of H2CO3
to CO2 + H2O is quite large in the absence of water. Thus,
this material may be kinetically stable for long periods. Our
estimated density for the H2CO3 oligomer is large enough
that it could occupy a space compatible with on-site storage
of the CO2 from coal-fired power plants.
Acknowledgments
This work was supported by DOE grant DE-FG02-94ER14467
and NSF grant EAR0539109.
Supporting Information Available
Computational methods and Figure 1. This information is
available free of charge via the Internet at http://pubs.acs.org.
Literature Cited
(1) Terlouw, J. K.; Lebrilla, C. B.; Schwarz, H. Thermolysis of
NH4HCO3: A simple route to the formation of free carbonic-
acid (H2CO3) in the gas-phase. Angew. Chem., Int. Ed. Engl.
1987, 26, 354–355.
(2) Hage, W.; Liedl, K. R.; Hallbrucker, A.; Mayer, E. Carbonic acid
in the gas phase and its astrophysical relevance. Science 1998,
279, 1332–1335.
(3) Falcke, H.; Eberle, S. H. Raman-spectroscopic identification of
carbonic-acid. Wat. Res. 1990, 24, 685–688.
(4) Rudolph, W. W.; Fischer, D.; Irmer, G. Vibrational spectroscopic
studies and density functional theory calculations of speciation
in the CO2-water system. Appl. Spectrosc. 2006, 60, 130–134.
(5) Abbott, T. M.; Buchanan, G. W.; Kruus, P.; Lee, K. C. C-13 nuclear
magnetic resonance and Raman investigations of aqueous
carbon-dioxide systems. Can. J. Chem. 1982, 60, 1000–1006.
(6) Rasul, G.; Reddy, V. P.; Zdunek, L. Z.; Surya Prakash, G. K.; Olah,
G. A. Chemistry in superacids. 12. Carbonic acid and its
monoprotonation and diprotonationsNMR, ab initio and IGLO
investigation. J. Am. Chem. Soc. 1993, 115, 2236–2238.
(7) Dibenedetto, A.; Aresta, M.; Giannoccaro, P.; Pastore, C.; Papai,
I.; Schubert, G. On the existence of the elusive monomethyl
ester of carbonic acid [CH3OC(O)OH] in 300K: 1H- and 13C
NMR measurements and DFT calculations. Eur. J. Inorg. Chem.
2006, 908–913.
(8) Stumm, W.; Morgan, J. J. Aquatic Chemistry: Chemical Equilibria
and Rates in Natural Waters; Wiley: New York, 1995.
(9) Soli, A. L.; Byrne, R. H. CO2 system hydration and dehydration
kinetics and the equilibrium CO2/H2CO3 ratio in aqueous NaCl
solution. Marine Chem. 2002, 78, 65–71.
(10) Tossell, J. A. H2CO3 and its oligomers: Structures, stabilities,
vibrational and NMR spectra, and acidities. Inorg. Chem. 2006,
45, 5961–5970.
(11) Moore, M. H.; Khanna, R. K. Infrared and mass-spectral studies
of proton irradiated H2O + CO2 icesEvidence for carbonic-
acid. Spectrochim. Acta 1991, 47A, 255–262.
(12) Hage, W.; Hallbrucker, A.; Mayer, E. Carbonic acid: Synthesis
by protonation of bicarbonate and FTIR spectroscopic char-
acterization via a new cryogenic technique. J. Am. Chem. Soc.
1993, 115, 8427–8431.
(13) Hage, W.; Hallbrucker, A.; Mayer, E. A polymorph of carbonic
acid and its possible astrophysical relevance. J. Chem. Soc.,
Faraday Trans. 1995, 91, 2823–2826.
(14) Hage, W.; Hallbrucker, A.; Mayer, E. J. Metastable intermediates
from glassy solutions. 3. FTIR spectra of alpha-carbonic acid
and its H-2 and C-13 isotopic forms, isolated from methanolic
solution. J. Chem. Soc. Faraday Trans. 1996, 92, 3183–3195.
VOL. 43, NO. 7, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 2579
(15) Hage, W.; Hallbrucker, A.; Mayer, E. Metastable intermediates
from glassy solutions. 4. FTIR spectra of beta-carbonic acid and
its H-2 and C-13 isotopic forms, isolated from aqueous solution.
J. Chem. Soc. Faraday Trans. 1996, 92, 3197–3209.
(16) Brucato, J. R.; Palumbo, M. E.; Strazzulla., G. Carbonic acid by
ion implantation in water/carbon dioxide ice mixtures. Icarus
1997, 125, 135–144.
(17) Al-Hosney, H. A.; Grassian, V. H. Carbonic acid: An important
intermediate in the surface chemistry of calcium carbonate.
J. Am. Chem. Soc. 2004, 126, 8068–8069.
(18) Al-Hosney, H. A.; Grassian, V. H. Water, sulfur dioxide and nitric
acid adsorption on calcium carbonate. A transmission and ATR-
FTIR study. Phys. Chem. Chem. Phys. 2005, 7, 1266–1276.
(19) Gerakines, P. A.; Moore, M. H.; Hudson, R. L. Carbonic acid
production in H2O:CO2 ices: UV photolysis vs. proton bom-
bardment. Astron. Astrophys. 2000, 357, 793–800.
(20) Zheng, W.; Kaiser, R. I. On the formation of carbonic acid (H2CO3)
in solar system ices. Chem. Phys. Lett. 2007, 450, 55–60.
(21) Winkel, K.; Hage, W.; Loerting, T.; Price, S. L.; Mayer, E. Carbonic
acid: From polyamorphism to polymorphism. J. Am. Chem. Soc.
2006, 129, 13863–13871.
(22) Nguyen, M. T.; Ha, T. K. A theoretical study of the formation
of arbonic-acid from the hydration of carbon-dioxidesA case
of active solvent catalysis. J. Am. Chem. Soc. 1984, 106, 599–602.
(23) Nguyen, M. T.; Hegarty, A. F.; Na, T.-K. Ab-initio study of the
hydration of carbon-dioxide - additional comments based on
refined calculations. J. Molec. Struct. THEOCHEM 1987, 150,
319–325.
(24) Jonsson, B.; Karlstrom, G.; Wennerstrom, H.; Roos, B. Ab initio
molecular-orbital calculations on water-carbon dioxide sys-
tems
carbonic-acid. Chem. Phys. Lett. 1976, 41, 317–320.
(25) Khanna, R. K.; Tossell, J. A.; Fox, K. On the stability of carbonic
acid under conditions in the atmosphere of Venus. Icarus 1994,
112, 541–544.
(26) Wight, C. A.; Boldyrev, A. I. Potential-energy surface and
vibrational frequencies of carbonic-acid. J. Phys. Chem. A 1995,
99, 12125–12130.
(27) Loerting, T.; Tautermann, C.; Kroemer, R. T.; Kohl, I.; Hallbrucker,
A.; Mayer, E.; Liedl, K. R. On the surprising kinetic stability of
carbonic acid (H2CO3). Angew. Chem., Int. Ed. Engl. 2000, 39,
892–894.
(28) Merz, K. M., Jr. Gas-phase and solution-phase potential-energy
surfaces for CO2 + nH2O (n ) 1, 2). J. Am. Chem. Soc. 1990, 112,
7973–7980.
(29) Nguyen, M. T.; Raspoet, G.; VanQuickenborne, L. G.; van
Duijnen, P. T. How many water molecules are actively involved
in the neutral hydration of carbon dioxide. J. Phys. Chem. A
1997, 101, 7379–7388.
(30) Liedl, K. R.; Sekusak, S.; Mayer, E. Has the dimer of carbonic
acid a lower energy than its constituents water and carbon
dioxide. J. Am. Chem. Soc. 1997, 119, 3782–3784.
(31) Ballone, P.; Montanari, B.; Jones, R. O. Density functional theory
of carbonic acid clusters. J. Chem. Phys. 2000, 112, 6571–6575.
(32) Tossell, J. A. Boric acid, “carbonic” acid, and N-containing
oxyacids in aqueous solution: Ab initiio studies of structure,
pKa, NMR shifts and isotopic fractionations. Geochim. Cos-
mochim. Acta 2005, 69, 5647–5658.
(33) Montgomery, J. A.; Frisch, M. J.; Ochterski, J. W.; Petersson,
G. A. A complete basis set model chemistry. VI. Use of density
functional geometries and frequencies. J. Chem. Phys. 1999,
110, 2822–2827.
(34) Scuseria, G. E.; Schaefer, H. F., III. An efficient reformulation
of the closed-shell coupled cluster single and double excitation
(CCSD) equations. J. Chem. Phys. 1988, 89, 7382–7387.
(35) Petersson, G. A.; Tensfeldt, T. G.; Montgomery, J. A. A complete
basis set model chemistry. 3. The complete basis set-quadratic
configuration-interaction family of methods. J. Chem. Phys.
1991, 94, 6091–6101.
2580 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 7, 2009
(36) Clabo, D. A., Jr.; Allen, W. D.; Remington, R. B.; Yamaguchi, Y.;
Schaefer, H. F. III A systematic study of molecular vibrational
anharmonicity and vibration-rotation interaction by self-
consistent-field higher-derivative methodssAsymmetric-top
molecules. Chem. Phys. 1988, 123, 187–239.
(37) Carbonniere, P.; Lucca, T.; Pouchan, C.; Regia, N.; Barone, V.
Vibrational calculations beyond the harmonic approximation:
Performance of the B3LYP density functional for semirigid
molecules. J. Comput. Chem. 2005, 26, 384–388.
(38) Barone, V. Anharmonic vibrational properties by a fully
automated second-order perturbative approach. J. Chem. Phys.
2005, 122, 014108.
(39) Truong, T. N.; Stefanovitch, E. V. A new method for incorporating
solvent effect into the classical ab-initio molecular- orbital and
density-functional theory frameworks for arbitrary shape cavity.
Chem. Phys. Lett. 1995, 240, 253–260.
(40) Tomasi, J.; Cammi, R.; Mennucci, B.; Cappelli, C.; Corni, S.
Molecular properties in solution described with a continuum
solvation model. Phys. Chem. Chem. Phys. 2002, 4, 5697–5712.
(41) Wolinski, K.; Hinton, J. F.; Pulay, P. Efficient implementation
of the gauge-independent atomic orbital method for NMR
chemical shift calculations. J. Am. Chem. Soc. 1990, 112, 8251–
8260.
(42) Bauernschmitt, R.; Ahlrichs, R. Treatment of electronic excita-
tions within the adiabatic approximation of time dependent
density functional theory. Chem. Phys. Lett. 1996, 256, 454–464.
(43) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R. ; Montgomery, J. A., Jr.; Vreven, T.;
Kudin, K. N.; Burant, J. C.; et al. Gaussian 03, Rev. A.1; Gaussian,
Inc.: Pittsburgh, PA, 2003.
(44) Frisch, A. E.; Dennington, R. D.; Keith, T. A.; Nielsen, A. B.;
Holder, A. J. GaussView, Rev. 3.0; Gaussian, Inc.: Pittsburgh, PA,
2003.
(45) Intergovernmental Panel on Climate Change, Special Report on
Carbon Dioxide Capture and Storage; Cambridge University
Press: Cambridge UK, 2005.
(46) Arstad, B.; Blom, R.; Swang, O. CO2 absorption in aqueous
solutions of alkanolamines: Mechanistic insight from quantum
chemical calculations. J. Phys. Chem. A 2007, 111, 1222–1228.
(47) Kintisch, E. Power generationsMaking dirty coal plants cleaner.
Science 2007, 317, 184–186.
(48) Whitesides, G. M.; Crabtree, G. W. Don’t forget long-term
fundamental research in energy. Science 2007, 315, 796–801.
(49) Aresta, M.; Dibenedetto, A. Utilisation of CO2 as a chemical
feedstock: Opportunities and challenges. Dalton Trans. 2007,
2975–2992.
(50) Ruscic, B.; Feller, D.; Dixon, D. A.; Peterson, K. A.; Harding,
L. B.; Asher, R. L.; Wagner, A. F. Evidence for a lower enthalpy
of formation of hydroxyl radical and a lower gas-phase bond
dissociation energy of water. J. Phys. Chem. A 2001, 105, 1–4.
(51) Tossell, J. A. Calculation of B and O NMR parameters in molecular
models for B2O3 and alkali borate glasses. J. Non-Cryst. Solids
1997, 215, 236–243.
(52) see e.g. Fig. 3.14 from Girard, J. E. Principles of Environmental
Chemistry; Jones and Bartlett Publishers: Sudbury, MA, 2005.
(53) Marley, N. A.; Gaffney, J. S.; Cunningham, M. M. Aqueous
greenhouse species in clouds, fogs and aerosols. Environ. Sci.
Technol. 1993, 27, 2864–2869.
(54) Wagman, D. D.; Evans, D. H.; Parker, V. B.; Schumm, R. H.;
Halow, I.; Bailey, S M.; Churnay, K. L.; Nuttall, R. L. The NBS
Tables of Chemical Thermodynamic Properties: Selected Values
for Inorganic and C1 and C2 Organic Substances in SI Units;
ACS and AIP: Washington, DC, 1982.
(55) Auwera, J. V.; Didriche, K.; Perrin, A.; Keller, F. Absolute line
intensities for formic acid and dissociation constant of the dimer.
J. Chem. Phys. 2007, 126, 124311.
(56) Clague, A. D. H.; Bernstein, H. J. Heat of dimerization of some
carboxylic acids in vapour phase determined by a spectroscopic
method. Spectrochim. Acta 1969, 25A, 593–596.
ES802393S