10 1021@acs Jpca 9b11744

Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE
Feature Article
How Collective Phenomena Impact CO
2
Reactivity and Speciation in Different Media

Daniela Polino, Emanuele Grifoni, Roger Rousseau, Michele Parrinello, and Vassiliki-Alexandra Glezakou
J. Phys. Chem. A, Just Accepted Manuscript • DOI: 10.1021/acs.jpca.9b11744 • Publication Date (Web): 17 Apr 2020
Downloaded from pubs.acs.org on April 23, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,

Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.
Page 1 of 41 The Journal of Physical Chemistry
1
2
3
4
5
6
7
8
How Collective Phenomena Impact CO2
9
10
11
12
Reactivity and Speciation in Different Media
13
14
15
16 Daniela Polinoa, b, Emanuele Grifonia, b, Roger Rousseauc, Michele Parrinelloa, b, d and Vassiliki-
17
18
19 Alexandra Glezakou*, c
20
21
(a)Department of Chemistry and Applied Biosciences, ETH Zurich, c/o USI Campus
22
23
24 Via Giuseppe Buffi 13, CH-6900 Lugano Switzerland.
25
26 (b)Facoltà di Informatica, Istituto di Scienze Computazionali, Università della Svizzera
27
28 Italiana,Via Giuseppe Buffi 13, CH-6900 Lugano Switzerland.
29
30 (c) Pacific Northwest National Laboratory, 902 Battelle Blvd, PO Box 999, MSIN K1-83,
31
32 Richland WA 99352.
33
34 (d) Istituto Italiano di Tecnologia, Via Morego 30, 16163 Genova, Italy.
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 ACS Paragon Plus Environment
1
The Journal of Physical Chemistry Page 2 of 41
1
2
3 ABSTRACT: CO2 has attracted considerable attention in the recent years due to its role in the
4
5
6 greenhouse effect and environmental management. While its reaction with water has been
7
8 studied extensively, the same cannot be said for reactivity in supercritical CO2 phase, where the
9
10 conjugate acid/base equilibria proceed through different mechanisms and activation barriers. In
11
12
13
spite of the apparent simplicity of the CO2 + H2O reaction, the collective effect of different
14
15 environments has drastic influence on the free energy profile. Enhanced sampling techniques and
16
17 well-tailored collective variables provide a detailed picture of the enthalpic and entropic drivers
18
19
underscoring the differences in the formation mechanism of carbonic acid in the gas, aqueous
20
21
22 and supercritical CO2 phases.
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
2
1
2
3
4
5
1. INTRODUCTION
6
7
8
In recent years, a number of CO2 management technologies has been developed to mitigate the
9
10 adverse effects of anthropogenic emissions. Promising solutions involve carbon capture and
11
12 sequestration (CCS) practices,1 where geological formations are exposed to captured CO2 that is
13
14
chemically and permanently fixated. In the past decade, a number of solvent systems has been
15
16
17 developed for selective removal of CO22 from point sources in the form of supercritical CO2
18
19 (scCO2). Supercritical fluids possess intriguing properties due to their short- and long-range
20
21 inhomogeneities that researchers believe are at the center of their unusual reactivity.3 In contrast
22
23
24 to water mediated reactions,4 reactivity in scCO2 has not received much attention. The well-
25
26 defined concepts of solubility or activity in water do not necessarily have corresponding
27
28 thermodynamic meaning, and the widely accepted reaction between H2O and CO2 in aqueous
29
30
31
environments, eq. (1), may not be proceeding the same way in scCO2, eq. (2), due to the local
32
33 inhomogeneities of this condensed phase:
34
35
36 CO2(aq) + H2O(aq) H2CO3(aq) HCO3-(aq) CO32- (1)
37
38
39
40
41 scCO2 + H2O(sc) H2CO3(sc) (2)
42
43
44
45 Theory and simulation are uniquely equipped to provide molecular level understanding of
46
47
48
the structural subtleties due to both local and long range fluctuations. The microsolvation of CO2
49
50 with H2O has been studied quite extensively by means of quantum mechanical methods such as
51
52 density functional theory (DFT), Hartree-Fock (HF) or post-Hartree-Fock methods (e.g.,
53
54 perturbation or coupled cluster theories). Jena and Mishra5 studied the CO2-H2O formation and
55
56
57
58
59
3
1
2
3 rotational barriers in water clusters, concluding that CO2 solvation is likely to include more than
4
5
6 8 H2O molecules, and that use of higher correlation methods and basis sets alone does not
7
8 improve the agreement with experiment. Nguyen et al.6 also studied the microsolvation of CO2
9
10 by nH2O (n=1-4) with MP2 and CCSD(T) methods, and found that the barrier to H2CO3
11
12
13
formation with 1 H2O is as high as 50 kcal/mol. Addition of 4 H2O molecules successively
14
15 lowers the energy barrier to ~20–27 kcal/mol and the free energy barrier to ~19–23 kcal/mol
16
17 depending on the local geometries of the waters, where three of them directly participate in the
18
19
transition state. Loerting et al.7,8 reported similar values, 52.6 kcal/mol and 43.55 kcal/mol for
20
21
22 H2CO3 dissociation, while adding one H2O reduced the barriers to 34.7 and 27.13 kcal/mol, and
23
24 adding two H2O to 31.6 and 24.01 kcal/mol, respectively. Tautermann et al.9 studied the stability
25
26 and decomposition rate of H2CO3 and determined barriers of 44.8 kcal/mol and 52.8 kcal/mol for
27
28
29 the H2CO3 formation. As other studies have concluded, addition of more waters helps reduce
30
31 both the forward and reverse barriers, where the additional waters assist in a Grotthus-like proton
32
33 transport.
34
35
36
37
In all studies, it is obvious that the dynamic role of actively partaking and spectator water
38
39 molecules greatly influences both the forward (H2CO3 formation) and reverse (H2CO3
40
41 decomposition) reactions. Kumar et al.10 used metadynamics to evaluate the energetics,
42
43
conformational changes and gas phase dissociation of carbonic acid yielding a free energy
44
45
46 barrier of 37.1 kcal/mol. Finally, Gallet et al.11 used metadynamics to study the interactions of
47
48 CO2 with water molecules and the dynamic role of participating waters in the
49
50 formation/decomposition of carbonic acid. These studies, albeit first of their kind, provided only
51
52
53 a qualitative sampling of the free energy surface due to the relatively low number of reactive
54
55 events. In fact, Kumar et al. observed only one reactive event per simulation, while Gallet et al.
56
57
58
59
4
1
2
3 performed a longer simulation that allowed them to observe about 5 re-crossing events
4
5
6 (10 reactive events) in 200 ps of simulation time.
7
8
9 The present work was motivated by recent studies of water-bearing supercritical CO2. In
10
11 2007, Saharay and Balasubramanian first used ab initio molecular dynamics to study the
12
13
14
structure and electrostatics of a solitary water molecule in supercritical CO2 at three different
15
16 densities.12 Their study shows that the instantaneous CO2-H2O interactions give rise to the
17
18 formation of H-bonds, an increase in the water dipole moment, and larger CO2 deviations from
19
20
linearity. ScCO2 has a characteristic structure where each CO2 is surrounded by ~6 other CO2
21
22
23 molecules around the carbon in a distorted T-shaped configuration and 3+3 CO2 molecules
24
25 coordinating the oxygen poles. In 2009, Saharay and Balasubramanian performed a study of one
26
27 H2O in scCO2 and reported on the structure, electronic and dynamic properties of H2O in this
28
29
30 condensed phase, but did not examine any reactivity.13 In 2010 Glezakou et al. examined the
31
32 structure, dynamics, and vibrational spectra of scCO2/(H2O)n, where n=0-4 at the carbon
33
34 sequestration relevant conditions (density of 0.74 g/cm3 and T=318.15 K) using ab initio
35
36
37
molecular dynamics.14 The analysis showed that the waters do not disrupt the scCO2 structure
38
39 and that the strongest interactions between CO2/H2O occur in the case of monomeric water,
40
41 where the water is mostly found in between the equatorial and polar coordinating CO2
42
43
molecules, forming dynamic hydrogen bonds with the oxygens. These findings led us to believe
44
45
46 that the interactions and chemistry of CO2/H2O in scCO2 could be quite distinct and different
47
48 than in aqueous environments.
49
50
51 This work examines and compares the CO2/H2O reactivity in gas, scCO2 and aqueous
52
53
54 phases by means of metadynamics to sample the collective fluctuations that lead to a reaction
55
56 between CO2 and H2O and form carbonic acid and its derivative ions. As we will show, the
57
58
59
5
1
2
3 choice of the appropriate collective variables that can describe both the formation and
4
5
6 decomposition of carbonic acid and re-crossing events, is a non-trivial process. In the case of
7
8 scCO2, it has to properly describe the oxygen scrambling15 and in the case of water, both the
9
10 oxygen scrambling and proton hopping.16 At the same time, this method allows us to compute
11
12
13
the free energy landscape for this reaction, and assess the role and source of entropy for the same
14
15 reaction in the different environments. To our knowledge, this is the first comprehensive study of
16
17 this fundamental reaction that compares the reactivity in these distinct environments relevant to
18
19
common environmental processes, such as CO2 sequestration, CO2 conversion, and corrosion in
20
21
22 scCO2 environments.
23
24
25 2. COMPUTATIONAL MODELS AND METHODS
26
27
28 2.1. Molecular dynamics. Ab initio molecular dynamics simulations were carried out
29
30
31
using the CP2K package17 driven by the PLUMED2 code.18,19 The Born-Oppenheimer forces
32
33 were used to propagate the dynamics of nuclei, with a convergence criterion of 5  10-6 a.u. for
34
35 the optimization of the wavefunction. In this scheme, only valence electrons were explicitly
36
37
38 considered. More specifically, the Khon-Sham orbitals were represented by a Gaussian basis set,
39
40 while a supplementary plane-wave basis was used to represent the electron density. We opted for
41
42 the MOLOPT-DZVP (2s2p1d/2s1p Gaussian basis set)20 and a plane wave cutoff of 400 Ry.
43
44
45
Core electrons were treated using the Goedecker-Teter-Hutter (GTH) pseudopotentials.21 The
46
47 revPBE exchange-correlation density functional22 augmented by Grimme’s third generation
48
49 dispersion corrections23 was used to account for non-covalent effects. The time-step for the
50
51 integration of the equations of motion was set to 1 fs. Simulations were carried out within the
52
53
54 NVT ensemble at 323 K using the stochastic velocity-rescaling of Bussi et al.24 For the three
55
56 environments investigated (gas, scCO2, and water phase), we used a periodic cubic box of 12 Å
57
58
59
6
1
2
3 with 1 CO2 and 1 H2O, a 20.5 Å with 1 H2O and 63 CO2 molecules to simulate super-critical
4
5
6 CO2 (density 0.74 g/cm3 and T=318.15 K) and a box of 12.57 Å with 1 CO2 and 63 H2O to
7
8 reproduce the water environment (density 0.998 g/cm3). Two additional simulations were also
9
10 performed starting with the formed product (H2CO3) in both scCO2 and water boxes to explore
11
12
13
the phase space of the product. All systems were initially equilibrated for ~20 ps (gas phase
14
15 system) and ~100 ps for the condensed systems and analyses on the unbiased systems after
16
17 discarding the first 20–30 ps. We used the equilibrated systems as starting points for the
18
19
metadynamics simulations. Analysis was performed on trajectories of 1.5 ns for the gas phase
20
21
22 system, ~600 ps for the scCO2, and ~2 ns for the water system, after we observed 8–10 crossings
23
24 per system, and after discarding the initial steps before the first crossing.
25
26
27 2.2 Construction of collective variables. Since the formation of carbonic acid is an
28
29
30 activated process, we accelerated the dynamics of this event by means of metadynamics.25-28
31
32 Metadynamics enhances the sampling of rare events by introducing a history dependent bias
33
34 potential function of a set of collective variables (CVs). This repulsive potential can be written as
35
36
37
a sum of Gaussians deposited along the system trajectory in the CV space, thus discouraging the
38
39 system from revisiting already sampled configurational space.
40
41
42 Scheme 1 depicts the gas phase reaction between CO2 and H2O, along with the transition
43
44
state and products to illustrate the basic features of the reactants, products and definition of the
45
46
47 collective variables. The product, H2CO3 can exist in different conformers that are identifiable as
48
49 stationary structures,7,10,29,30 and by spectroscopic methods.31,32
50
51
52
53
54
55
56
57
58
59
7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29 Scheme 1. (a) Reaction of carbonic acid formation CO2+H2O  H2CO3 and (b) carbonic acid
30
31
conformers, where Ci defines the coordination number for the water O and H with the C and O of
32
33
34 CO2 respectively.
35
36
37 The definition of a set of collective variables that is able to describe the process in both
38
39 directions is non-trivial. The coordination number-based path collective variables introduced by
40
41 Pietrucci and Saitta33,34 are considered prime candidates for this task. Recently, in fact, they have
42
43
44 been successfully applied both to formamide decomposition33 and urea decomposition in water.35
45
46 Briefly, we defined state A and B as the reactants (CO2 + H2O) and the product (H2CO3) basins,
47
48 respectively, and formulated s and z to be s=1 in state A and s=2 in state B, as follows:
49
50
51
52
53 (3)
54
55
56
57
58
59
8
1
2
3
4
5 (4)
6
7
8
9
10
The key ingredient of the path CVs is the definition of distance between the atomic
11
12 configuration at time t, , and the reference structure, or . Considering the different
13
14
15 characteristics of the three media, we had to design specific distances in the three cases, which
16
17
18 resulted in including different coordination numbers to distinguish state A and B.
19
20
21 In the case of the gas-phase and sc-CO2 environments, we defined:
22
23
24
25 (5)
26
27
28
29
30 (6)
31
32
33
34 Here, the reference values for is 0 in state A and 1 in state B (see Scheme 1 (a)). The
35
36
37
38 coordination number, , was calculated as follow:
39
40
41
42
43
44
45 (7)
46
47
48
49
50 where rOw,i is the distance between the water-oxygen and the ith carbon, while r0=2.0 Å, m=6, and
51
52
53
n=12. In both the gas-phase and sc-CO2 environments, a problem of indistinguishability of
54
55 oxygens arises when H2CO3 decomposes back to CO2 and H2O. For this reason, we had to define
56
57
58
59
9
1
2
3
the coordination number based on the a priori identification of the reacting water oxygen.
4
5
6
7 More specifically, the position of the water oxygen was identified as a weighted average position
8
9 based on the value of the coordination number distributions, :
10
11
12
13
14
15 (8)
16
17
18
19
20 This calculation is activated by the keyword CENTER_OF_MULTICOLVAR
21
22 implemented in PLUMED 2 from version 2.3 on. Here, the values correspond to the
23
24
25
26 position of the ith oxygen atom at time t, while is defined as the number of hydrogen atoms
27
28
29 around the ith oxygen:
30
31
32
33
34
35 (9)
36
37
38
39
40
where ri,j is the distance between atom i and atom j, while r0, m, and n are the parameters of the
41
42
43 switching function that were chosen equal to 1.5 Å, 6, and 12, respectively. Knowing the
44
45 distribution of , we can then assign the weights wi. , thus, takes into account the
46
47
48
49 coordinates of the oxygen atom together with those of closely coordinating hydrogens. Once the
50
51 water-oxygen position is identified, we can calculate .
52
53
54
55
56
57
58
59
10
1
2
3 To follow the same reaction in a water environment, the distances used to build the path
4
5
6 CVs are:
7
8
9
10 (10)
11
12
13
14 (11)
15
16
17
18 where consists of the (rescaled) number of oxygens around the carbon atom of the only CO2
19
20
21
22 present in the box:
23
24
25
26
27 (12)
28
29
30
31
32
33 here rC,i is the distance between the carbon atom and the ith oxygen, while r0=2.0 Å, m=12 and
34
35 n=24. In this fashion, the reference value for is 0 in state A and 1 in state B (see
36
37
38
39 Scheme 1 (a)).
40
41
42 A more natural and transparent representation of the chemistry of the reaction was then
43
44 obtained by projecting the FES onto two variables CV1 and CV2. These are meant in general
45
46
47
terms to represent the number of C-O and O-H bonds involved in the reaction, respectively.
48
49
50 For the gas-phase and sc-CO2 (where we have only one H2O), CV1 is = , and gives
51
52
53
the number of carbons around the water-oxygen. To determine CV2, we calculated again the
54
55
56 distribution of the coordination numbers of all the oxygens with all the hydrogens in the system,
57
58
59
11
1
2
3 , and then we counted the number of oxygens with more than 1.5 hydrogens bonded, as it
4
5
6
7 follows:
8
9
10
11
12
13 (13)
14
15
16
17
18
19 where N=6 and M=12 and =1.5. With this CV we could discriminate between state A ,
20
21
22
23
characterized by CV2=1, and state B, for which CV2=0.
24
25
26 In the case of water, we used CV1= , which is the normalized number of oxygens
27
28
29 around the carbon atom of CO2. Whereas, to describe correctly the number of hydrogens bound
30
31
32 to any oxygen needed in the computation of CV2, we used a hydrogen-oxygen coordination
33
34 distribution calculated by Voronoi tessellation as implemented by Grifoni et al.36 Briefly, for
35
36 each i-th oxygen, the hydrogen coordination number can be calculated as:
37
38
39
40
41 (14)
42
43
44
45
46 where =4. Similarly to the previous case, we can apply a switching function to this distribution
47
48
49
50 ( ) to set the OH coordination ( ) equal to 1 in state A and 0 in state B counting the
51
52
53 number of oxygens with more than 1.5 hydrogens bonded, as follows:
54
55
56
57
58
59
12
1
2
3
4
5
6 (15)
7
8
9
10
11
12
13 where N=6 and M=12 and =1.5.
14
15
16
17 2.3. Collective Variables for the reaction H2CO3  HCO3- and conformers. In
18
19
20
acid-base reactions, such as the deprotonation of carbonic acid, one forms either a hydronium
21
22 (H3O+) or a hydroxyl (OH-) ion.37-39 The structure of the solvated ions is rather elusive, as they
23
24 can rapidly diffuse in the medium via a Grotthuss mechanism.16 They are highly fluxional and
25
26
the identity of the atoms involved in their structure changes continuously. As a result, it is very
27
28
29 difficult to capture the nature of these species with an explicit analytic function of the atomic
30
31 coordinates. Recently, Grifoni et al.36 developed CVs that can follow the protonation state of the
32
33 system (sp) and the distance between the two conjugate acid-base sites (sd). This CVs enabled us
34
35
36 to correctly assign the reactive centers during the carbonic acid deprotonation reaction and
37
38 accurately compute the corresponding reaction free energy.
39
40
41 Carbonic acid has three possible conformers, depending on the relative orientation of the
42
43
44
hydroxyl groups, commonly called cis-trans (CT), trans-trans (TT), and cis-cis (CC), illustrated
45
46 in Scheme 1 (b). To monitor the presence of the different conformers of H2CO3 in either the sc-
47
48 CO2 or water environments a more complex variable was needed than just the dihedral angles of
49
50
the molecule, because of the oxygen exchange (scrambling) between H2CO3 and CO2 (scCO2) or
51
52
53 H2O (aqueous). To follow this exchange, yet another collective variable, dC-HC, had to be defined,
54
55 as the distance between the carbon atom of the H2CO3 and Hc, the mid-distance of the H atoms
56
57
58
59
13
1
2
3 in H2CO3, see Scheme 1 (b). Still, the identification of C and Hc is not straightforward in the sc-
4
5
6 CO2 and water environments, but rather exhibits a very dynamic behavior.
7
8
9 In sc-CO2, the reacting CO2, and hence the C center, cannot be identified a priori.
10
11 Following the same approach as in the case of water, we defined the C center as the weighted
12
13
14 average position based on the coordination number distribution :
15
16
17
18
19
(16)
20
21
22
23
24 and
25
26
27
28
29
30 (17)
31
32
33
34
35 where represents the number of hydrogen atoms around the ith carbon, ri,j is the distance
36
37
38
39 between atom i and atom j, and r0, m, and n are the parameters of the switching function with
40
41 values set to 1.7 Å, 24 and 48, respectively, such that the two basins are well separated. As a
42
43
result, , represents the C center with the larger number of nearby hydrogens. Once the C
44
45
46
47 center of the carbonic acid is identified, we can define as collective variable the distance between
48
49 this atom and the center of the hydroxyl-hydrogens, dC-Hc, as illustrated in Scheme 1 (b).
50
51
52
53 In water, we have an opposite scenario: whereas in this case of CO2 the carbon atom is
54
55 well identified since it is the only carbon present in the box, the Hc cannot be defined a priori,
56
57
58
59
14
1
2
3 due to the of the rapid proton shuttling of the H2CO3 hydroxyl hydrogens. Hence, we define Hc
4
5
6 as:
7
8
9
10 (18)
11
12
13
14
15 and
16
17
18
19
20
(19)
21
22
23
24
25 where gi represents the number of carbon atoms around the ith hydrogen and ri,j is the distance
26
27
between atom i and atom j, while r0, m, and n are the parameters of the switching function
28
29
30 chosen to be 2.0 Å, 16 and 32, respectively. represents the weighted average position of the
31
32
33 hydrogens closer to the carbon atom, which is exactly the mid-distance between the two
34
35
36
hydroxyl hydrogens (Hc). Again, once the mid-distance (HC) is identified, we can define as
37
38 collective variable the distance between the carbon atom and HC, dC-Hc.
39
40
41 Given the number of processes investigated, in Table 1 we summarize all the reported the
42
43
CVs used for all the different reactions studied in this work.
44
45
46
47 Table 1. Summary of CV used to perform each metadynamic run and to calculate the
48
49 corresponding FES.
50
51
52 Metadynamics
53 Reaction: Environment performed on: FES reweighted on:
54 path variables CV1=COwC
55 CO2 + H2O → H2CO3 gas phase
s(t) and z(t) CV2=COH
56
57
58
59
15
1
2
3 path variables CV1=COw-C
4 scCO2
5
s(t) and z(t) CV2= COH
6 CV1=CCO
path variables
7 water CV2=
8 s(t) and z(t)
9
10
H2CO3 → HCO3- water sP and sD sP and sD
11  and   and 
12
gas phase
(see Fig. 5) (see Fig. 5)
13
CV1=COwC
14 H2CO3: conformational analysis scCO2
15 dC-Hc
16 CV1=CCO
17 water
dC-Hc
18
19
20
21 2.4. Free Energy, Internal Energy and Entropy Maps. The internal energy and
22
23 entropy maps were built following the procedure described in a recent publication by Salvalaglio
24
25
26 and coworkers.40 The ensemble average of the internal energy is mapped on the CV space
27
28
29 :
30
31
32
33
34
35 (20)
36
37
38
39
40 At convergence, the ensemble average of the free energy can be written as:
41
42
43
44
45
(21)
46
47
48
49
50 with,
51
52
53
54
55
56
57
58
59
16
1
2
3
4 (22)
5
6
7
8 Here , is the bias deposited along the metadynamics run, and is the
9
10
11
reweighting factor that can be computed as proposed in the paper by Tiwary and Parrinello:41
12
13
14
15
16 (23)
17
18
19
20
21 Knowing and it is possible to compute the ensemble average of entropy
22
23
24 as:
25
26
27
28
29
(22)
30
31
32
33 Quasi harmonic calculations were also carried out to compare the entropic contribution to
34
35
36 the reaction free energy in the different environments, with the one obtained by metadynamics
37
38 calculation.
39
40
41
42
3. RESULTS AND DISCUSSION
43
44
Figure 1 summarizes the free-energy landscapes for the formation of carbonic acid from its
45
46
47 constituent molecules CO2 and H2O, in the three different environments, (I) gas phase,
48
49 (II) scCO2, and (III) water. These were based on the metadynamics simulations with the
50
51 collective variables described in the previous section. From our unbiased ab initio molecular
52
53
54 dynamics simulations performed for both A (reactants, CO2+H2O) and B (products, H2CO3)
55
56 basins, we computed the radial distribution functions to assess the local structure in scCO2 and
57
58
59
17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44 water, see sections S1 and S2 in SI. In Fig. 2 we plot the minimum free energy path that clearly
45
46
47 shows the barriers for the forward and reverse reactions. The gas phase reaction proceeds
48
49 through a narrow, steep channel with a free-energy barrier of 51.4 kcal/mol. The gas phase
50
51
52
53
54
55
56
57
58
59
18
1
2
3 Figure 1. Free energy surface (FES) for the reaction CO + H2O H2CO3 in (I) gas-phase, (II)
4
5
6
7 sc-CO2 and (III) in water. The white dashed lines depict the minimum energy path (MEP)
8
9 along each FES. Units are in kcal/mol.
10
11
12
13 energetics of the CO2/H2O association has been studied extensively. The geometries and relative
14
15
16 energetics between the CO2…H2O pre-complex, transition state and product computed here are in
17
18 excellent agreement with experimental measurements42 or theoretical calculations, see for
19
20 example Wight and Boldyrev,30 Loerting et al.,8 Kumar et al.,10 Nguyen et al,6 and Gallet et al.11
21
22
23 and references therein. In the gas phase, the pre-complex is a loosely bound van der Waals
24
25 complex (A2) with a binding energy of ~2 kcal/mol where CO2 and H2O are roughly on the same
26
27 plane and the C and Ow centers are aligned. On the free energy surface, the pre-complex appears
28
29
to be iso-energetic with the dissociated CO2/H2O state. The transition state (TS) occurs after
30
31
32 significant charge transfer from the Ow to C resulting in a bent CO243 and strong H-bonding
33
34 between one of the H2O hydrogens and one of the CO2 oxygens. The reverse barrier in the gas-
35
36 phase of 42.3 kcal/mol is also in very good agreement with other published studies. Next, we
37
38
39 compare how the FES in scCO2 and water reactions to the gas-phase picture.
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
19
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22 Figure 2. Minimum energy pathway (MEP) from the computed FES for the reaction CO +
23
24
25 H2O H2CO3 in (I) gas-phase, (II) sc-CO2 and (III) in water. Units are in kcal/mol.
26
27
28
29
30
31 In scCO2, each CO2 is coordinated by several others along the equatorial plane and the
32
33
34
oxygen poles and their geometry shows instantaneous, yet significant deviations from linearity,
35
36 indicative of induced dipoles attributed to the inhomogeneous near-neighbor environment.12 The
37
38 structure and dynamics of monomeric H2O,13,14 dispersed and small water clusters14 in scCO2
39
40
have been examined before by means of AIMD. All of the studies show that the distorted
41
42
43 T-shaped configuration of scCO2 is not perturbed by the presence of neither dispersed
44
45 monomeric H2O nor by small H2O clusters. The H2O/CO2 interactions are fairly weak and the
46
47 strongest interactions occur with the monomeric H2O that forms H-bonds with the oxygens of
48
49
50 CO2. Given the presence of the equatorial coordinating CO2s at ~3.4 Å, we hypothesized that the
51
52 reaction between CO2 and H2O to form H2CO3 would likely have a high barrier, since H2O
53
54 would have to displace the surrounding CO2’s out of the equatorial positions and align with the
55
56
57
58
59
20
1
2
3 reacting C to approach at distances < 3.0 Å. Our hypothesis was confirmed by the result of the
4
5
6 metadynamics AIMD simulations: the computed forward free energy barrier is 50 kcal/mol, only
7
8 1.4 kcal/mol lower than the gas phase, Fig. 2. The reverse barrier for the scCO2-mediated H2CO3
9
10 decomposition is 30.8 kcal/mol, about 10 kcal/mol lower than the gas phase. Finally, to confirm
11
12
13
the assignment of the “transition state” geometry, we performed committor analysis for the
14
15 reaction in scCO2, since this is the least studied system, while both gas and aqueous phase
16
17 reactions have been studied in much detail. We selected 32 different geometries in the proximity
18
19
of the “transition state”, and for each of them , we ran 5 different MD simulations initialized with
20
21
22 random velocities. Of these, 86 simulations fell in reactant basin (54%) and 74 simulations fell in
23
24 product basin (46%). This analysis confirms our conclusion regarding the “transition state”
25
26 species on the free energy surface.
27
28
29
30 We will now discuss the carbonic acid formation and decomposition in water. Table 2
31
32 summarizes the forward/reverse barriers for this reaction from this work, as well as a
33
34 representative subset of the literature on the subject. Several studies have shown that addition of
35
36
37
even 1 additional H2O in the gas-phase helps considerably reduce the barrier, by 18–20 kcal/mol,
38
39 see Table 2. Addition of more H2O molecules (3 or 4) that micro-solvate the gas-phase system
40
41 further reduce the barrier almost by half. Nonetheless, the reported results show that adding 2 or
42
43
more extra H2O molecules to the systems is not enough to obtain agreement with experiments.
44
45
46 Reasonable agreement, though, was found by Stirling and Papai44 adopting enhanced ab initio
47
48 molecular dynamics simulations on extended models. This emphasizes the importance of
49
50 including the proper collective motions of the environment to correctly describe the properties of
51
52
53 condensed phase reactions.
54
55
56
57
58
59
21
1
2
3
4 Table 2. Forward/Reverse (first/second row) energy barriers of carbonic acid formation in
5
6 different media and reaction free energies, activation barriers and pKA for carbonic acid
7
8 dissociation in water.
9
10
CO2 + H2O → H2CO3
11
microsolvation
12 gas water sc-CO2
13 (n = 1,2,3 assisting H2O molecules)
phase
14 1 H2O 2 H2O 3 H2O n>3 H2O
15 Theoretical
16 This work 51.4 17.4 50.0
17 42.3 14.9 30.8
18 49.2 15.5
Nguyen et al.45
19 54.0 5.0
20 52.6 34.7 31.6
Loerting et al.46
21 43.6 27.13 24.0
22 52.8 34.2 29.1
Tautermann et al.9
23 44.8 27.7 20.9
24 50.6 31.8
Jena and Mishra5
25 -- --
26 --
27 Kumar et al.10
37.1
28 33.0
29 51.6 31.2a 25.6 22.3a 19.9 19.0a
30 Nguyen et al.6
41.0 24.4 20.2 17.4a --
31 23.3a
32 54 31 24
33 Gallet et al.11
38 19 18
34 18.8
35 Stirling and Papai44
15.5
36 48.9 33 21.6 20.3 19
37 Wang and Cao47
44.4 24.9 17.0 10.9 15
38
Experimental
39
Meier and --
40
41 Schwarzenbach48 16.1
42 17.7
Magid and Turbeck49
43 14.6
44 19.5
Pocker and Bjorkquist37
45 --
46 19.3
Wang et al.38
47 17.1
48 H2CO3 → HCO3-
49 ΔGR pKA EA
50 Theoretical
51 This work 4.0 2.7 (323 K)
52 Stirling and Papai44 5.1 3.7 (350 K)
53 Galib and Hanna39,50 5.5 9.3
54 Wang and Cao47 8.9–7.5 8.6–11.1
55 Experimental
56
57
58
59
22
1
2
3 Adamczyk et al.51 4.7 3.45 (298 K)
4 Pines et al.52 4.8 3.49 ( 298 K)
5 a Solvation correction calculated with COSMO-PCM.
6
7
8
9 The reported free energy barrier of 18.8 kcal/mol calculated at 350 K is consistent with
10
11 the experimental value of 21.8 kcal/mol measured at the same temperature. The minimum energy
12
13 path determined on this surface at 323 K proceeds with a barrier of 17.4 kcal/mol with the
14
15
16 concerted addition of H2O to CO2 and immediate formation of an ion pair HCO3-/H3O+. The
17
18 difference between this work and that of Stirling and Pápai is likely due to our choice of CVs,
19
20 which allows for a better sampling of reactive events (several crossings/re-crossing across the
21
22
23
transition state) and consequently a more accurate free energy surface. The same ion pair
24
25 formation mechanism has been discussed before by Stirling and Pápai44 and Adamczyk et al.51 as
26
27 the potential culprit for the unusual kinetic stability of H2CO3 and lower than expected pKa
28
29
values. The free energy barrier estimated for the formation of the carbonic acid is in good
30
31
32 agreement with values measured at the same temperature, 17.4 kcal/mol vs 17.749 and
33
34 19.3 kcal/mol.38
35
36
37 It is worth mentioning that during the simulation we detected also the formation of the
38
39
40 carbonate ion (CO32-), whose geometry is depicted in Fig. 3. This structure corresponds to the
41
42 relative minimum observed in the (1, 1) region of the CV space on the FES reported in Fig. 1.III.
43
44 The CVs adopted in this metadynamics simulation were not designed to follow the equilibrium
45
46
47
reaction between H2CO3, HCO3- and CO32-. Still, we obtained useful qualitative information on
48
49 the relative stability of HCO3-, H2CO3, and CO32- with the latter being the least stable. A more
50
51 detailed analysis of the first deprotonation reaction of H2CO3 is presented below.
52
53
54
55
56
57
58
59
23
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18 Figure 3. Geometry of carbonate CO32- ion in water.
19
20
21
22
23 To quantify the H2CO3 HCO3- equilibrium reaction, we used the CVs proposed by
24
25
26
27
Grifoni et al.36, sp and sd, which have been designed specifically to follow acid-base equilibria
28
29 reactions. The FES obtained is illustrated in Fig. 4 (a) together with the corresponding free
30
31 energy profile along the minimum energy path, in Fig. 4 (b). During the deprotonation reaction,
32
33
the system passes through an intermediate complex (B*) formed by the ionic pair HCO3-/H3O+
34
35
36 which is only 2.6 kcal/mol higher in energy than H2CO3. Only then is the pair completely
37
38 separated (sd > 6.0 Å) and the relative stability between carbonic acid and bicarbonate can be
39
40 measured. The reaction free energy determined is equal to ~4.0 kcal/mol with a corresponding
41
42
43 pKA of 2.7. Both values are in reasonable agreement with experiments51,52 and previous
44
45 theoretical calculations44 (see Table 2).
46
47
48
49
50
51
52
53
54
55
56
57
58
59
24
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37 Figure 4. Free energy surface in water a) and MEP b) for the HCO3- ↔ H2CO3 reaction.
38
39 B* corresponds to a metastable complex formed by the ionic couple [HCO3]-  H3O+.
40
41
42
43
44
45
Finally, we also investigated the conformational equilibrium between the H2CO3
46
47 conformers, CC, TT, and CT (see Scheme 1 (b)) in the different environments. In gas phase, the
48
49 different conformers cannot be accessed because the barriers dividing them are much larger than
50
51 1 kBT (0.642 kcal/mol at 323 K). Spontaneous conformational isomerization is thus unlikely, and
52
53
54 no conformational changes were detected during the metadynamics simulation of carbonic acid
55
56 formation in the gas phase. To further interrogate this conformational transformation, we carried
57
58
59
25
1
2
3
out a second metadynamics run, placing a bias along the dihedral angles  and . With this
4
5
6 approach, we recovered the free energy surface that connects the three states CC, CT, and TT
7
8 (see Fig. 5). CC is the global minimum, with CT being ~1.5 kcal/mol higher, whereas the TT
9
10
metastable state is found at ~8.5 kcal/mol in our FES. The TS dividing CC and CT is
11
12
13 ~9 kcal/mol. Loerting and Bernard report similar values with CT 1–2 kcal/mol and TT is
14
15 ~10 kcal/mol higher than CC, while the TS can be found at ~12 kcal/mol. More recently, also
16
17 Schwedtfeger and Mazziotti29 reported that CT is 1.2 kcal/mol higher in energy than CC and that
18
19
20 the barrier dividing CC and CT is of 9.5 kcal/mol. Bernard et al.31 carried out spectroscopy
21
22 measurements that confirmed this theoretical results. In particular, they have been able to detect
23
24 both conformers (CC and CT) in the gas phase at 210 K with a 1:10 ratio between CC and CT
25
26
27
which corresponds to a difference in Gibbs energies of 1 kcal/mol.
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
Figure 5. Free energy surface of the CC ↔ CT ↔ TT isomerism of H2CO3 in gas phase.
54
55
56
57
58
59
26
1
2
3 In Fig. 6 we report the results of the conformational isomerism analysis in scCO2 and in
4
5
6 H2O. In both environments, we found that the barrier dividing conformers CC and CT is much
7
8 lower than the barrier computed in gas phase and comparable to 1 kBT. In scCO2, we found CC
9
10 and CT to be almost equi-energetic, while in H2O, the CC structure is ~4.0 kcal/mol higher. This
11
12
13
is in contrast with gas phase results predicting the CC isomer to be the global minimum. On the
14
15 contrary, we found the CT conformer to be more stable than CC by ~2.4 kcal/mol, and this
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52 Figure 6. Free energy surface of the CT ↔ TT isomerism of H2CO3 in (a) sc-CO2 and
53
54 (b) water.
55
56
57
58
59
27
1
2
3 agrees well with the experimental observations of Loerting and Bernard7 where only the CT
4
5
6 conformer has been detected in water vapor environment.
7
8
9 Finally, the entropy contributions (<TΔS>) and internal energy (<ΔU>) to the free energy
10
11 (<ΔF>) maps are reported in Fig. 7 for the three different media. The reaction free energy,
12
13
14
energy and entropy calculated using eqs. (20)–(24) are reported in Table 3. Results from quasi-
15
16 harmonic approximation analysis14,53,54 were carried out for comparison and both methods
17
18 provide very similar estimates with errors < 20%. In gas and sc-CO2 phase, <ΔU> increases as
19
20
the systems depart from the minima. Contrarily, in water, this behavior is true only for H2CO3
21
22
23 but not for CO2 + H2O, where the internal energy and entropy terms cancel out. Further analysis
24
25 of the entropic contributions reveals a noticeable difference in the entropy between the reactants
26
27 and the products, because the formation of H2CO3 in all media reduces the degrees of freedom
28
29
30 and hence the entropy term. This difference is significantly larger in the condensed phases
31
32 compared to the gas-phase. However, while in water this effect is balanced out by the enthalpic
33
34 stabilization of H2CO3 through the network of water hydrogen bonds, in sc-CO2 there is no
35
36
37
stabilization of the product and hence H2CO3 formation is less probable.
38
39
40 Table 3. Free energy , internal energy and entropy between state A and
41
42
43
44 state B (B.1 in water) in kcal/mol adopting quasi-harmonic approximation and
45
46 metadynamics.
47
48
gas-phase sc-CO2 water
49
50 Quasi-harmonic approximation
51
52
53 11.7 7.0 -4.7 19.0 1.3 -17.7 0.2 -13.6 -13.8
54
55
Metadynamics
56
57
58
59
28
1
2
3
4
5
6 9.8 8.9 -1.7 23.2 1.6 -22.0 0.2 -13.5 -13.6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39 Figure 7. Free energy , internal energy and entropy maps as function of the
40
41
42 collective variables CV1 and CV2 for the reaction CO2 + H2O  H2CO3 in: (I) gas-phase, (II)
43
44
45
sc-CO2 and (III) water (III).
46
47
48
49
50 Following the free energy decomposition maps in Fig. 7, it becomes obvious that the role
51
52 of entropy in the transition state in the three different environments is very distinct. In the gas
53
54
55
phase, the entropy term is essentially constant along the reaction path up to the transition point
56
57
58
59
29
1
2
3 and it only involves the relative rotation of the reactants about the C-Ow distance. As a result, the
4
5
6 activation energy U† is very similar to the activation free energy F† (see Fig. 8). In contrast, in
7
8 scCO2, while the activation free energy is similar to the gas phase, the activation energy is
9
10
appreciably larger ~ 150 kcal/mol, and compensated by an increased activation entropy, owing to
11
12
13 the disruption of local CO2 structure by H2O at the TS. In water, the overall picture is completely
14
15 different: the microsolvation helps lowering the overall activation free energy but also stabilize
16
17 the TS as the charge separation ensues, by providing multiple channels for proton shuttling. The
18
19
20 same trend holds true for the product basin, where the proton shuttling and formation of
21
22 hydronium species (H3O+) helps stabilize the formation of the conjugate bases (HCO3- and
23
24 CO32-), but also strongly contributing to a large entropy term at both the TS and the products.
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
30
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48 Figure 8. Free energy , internal energy and entropy profiles projected on the
49
50
51
52 MEP for the reaction CO2 + H2O  H2CO3 in: (I) gas-phase, (II) sc-CO2 and (III) water (III).
53
54
55
56
57
58
59
31
1
2
3
4
4. OUTLOOK
5
6
7
In this work, we probed the collective solvent motions that impact the reaction profile of a
8
9 prototypical and environmentally relevant system, H2O + CO2. CO2 chemistry and reactivity
10
11 constitutes an active research direction in many fields, such as separations, catalysis and carbon
12
13
remediation. The CO2/H2O interaction is invariably and readily assumed to lead to carbonic acid
14
15
16 formation and subsequent decomposition to its carbonate anions regardless of the nature of the
17
18 solvent, i.e., water or scCO2. The low solubility of water in scCO2 and solvent microstructure
19
20 create a still unexplored reactive environment that is very different from the aqueous phase.4 Our
21
22
23 simulations present evidence that the prototypical reaction in scCO2 has an activation barrier as
24
25 high as that in gas phase, a notion that was never before entertained in the literature. However,
26
27 decomposition of the free energy landscape shows that while the enthalpy and entropy terms are
28
29
30
comparable in scCO2, they are markedly different than in gas phase. Formation of H2CO3 in
31
32 scCO2 is thermodynamically unfavorable, but once formed, below water saturation, it will most
33
34 likely be preserved in its undissociated form. In water, the activation barrier is considerably
35
36
lower, and entropy stabilizes both the transition state and the products by providing multiple
37
38
39 channels for H+ shuttling and exchange. Based on these results, we posit that the product of H2O
40
41 and CO2 and speciation in scCO2 (bulk solvent) is likely to be very different than the same
42
43 reaction in water. However, at a solid/liquid interface, CO2 activation can also happen without a
44
45
46 direct interaction with H2O and still result in formation of carbonate species.55
47
48
49 In closing, this work highlights the importance of sampling methods56 which capture the
50
51 conformational and configurational complexity of the solvent media as well as the potential side
52
53
54
reactions related to reactants and products. In this case, well-tailored collective variables are able
55
56
57
58
59
32
1
2
3 to represent the overall reactive landscape and reveal novel insights into solvent effects and
4
5
6 unanticipated reactivity.
7
8
9 ASSOCIATED CONTENT
10
11
12 Supporting Information. Supporting Information is available free of charge. The file contains
13
14 the following information:
15
16 1. Radial Distribution Functions: 1.1 Supercritical CO2 (scCO2); 1.2 Water
17
18 2. Metadynamics Convergence.
19
20
21 3. Error Analysis in the Estimation of Free Energy Barriers
22
23
24
25
26 AUTHOR INFORMATION
27
28
29 Corresponding Author
30
31
32 Vassiliki-Alexandra Glezakou, Vanda.Glezakou@pnnl.gov
33
34
35 Author Contributions
36
37
38 D.P. designed the collective variables and performed the majority of metadynamics simulations
39
40 and analyses and E.G. provided technical assistance with the Grifoni variables. V.-A.G. scoped
41
42 the research and together with R.R. performed some of the simulations. All authors contributed
43
44
45 to the writing of the manuscript and have given approval to the final version of the manuscript.
46
47
48 ACKNOWLEDGMENTS
49
50
51 V.-A.G. and R.R. were supported by the U.S. Department of Energy (DOE), Office of Science,
52
53
Office of Basic Energy Sciences, Division of Chemistry, Geochemistry and Biological Sciences,
54
55
56 and located at Pacific Northwest National Laboratory (PNNL). Computational resources were
57
58
59
33
1
2
3 provided by National Energy Research Scientific Computing Center (NERSC), a DOE Office of
4
5
6 Science User Facility located at Lawrence Berkeley National Laboratory (LBNL). PNNL is
7
8 operated by Battelle for the US Department of Energy under Contract DE-AC05-76RL01830.
9
10 D.P., E.G. and M.P. thankfully acknowledge the financial support provided by the NCCR
11
12
13
MARVEL, funded by the Swiss National Science Foundation, and the European Union Grant
14
15 No. ERC-2014-AdG-670227/VARMET. Computational resources were provided by the Swiss
16
17 National Supercomputing Centre (CSCS) under project IDs p503, s768 and s910.
18
19
20
21
22
23
24 REFERENCES
25
26
27 1. Tan, Y.; Nookuea, W.; Li, H.; Thorin, E.; Yan, J., Property Impacts on Carbon Capture and
28 Storage (CCS) Processes: A Review. Energy Convers. Manage. 2016, 118, 204-222. DOI:
29 http://dx.doi.org/10.1016/j.enconman.2016.03.0797.
30 2. Brennecke, J. F.; Chateauneuf, J. E., Homogeneous Organic Reactions as Mechanistic
31
Probes in Supercritical Fluids. Chemical Reviews 1999, 99, 433-452. DOI:
32
33
http://dx.doi.org/10.1021/cr970035q.
34 3. Levelt Sengers, J. M. H., Thermodynamics of Solutions Near the Solvent’s Critical Point.
35 In Supercritical Fluid Technology: Reviews in Modern Theory and Applications; Bruno, T.
36 J., Ely, J. F., Eds.; CRC Press: Boca Raton, 1991; pp. 1-56.
37 4. McGrail, B. P.; Schaef, H. T.; Glezakou, V. A.; Dang, L. X.; Owen, A. T., Water Reactivity
38 in the Liquid and Supercritical CO2 Phase: Has Half the Story Been Neglected? Energy
39
Procedia 2009, 1, 3415-3418. DOI: http://dx.doi.org/10.1016/j.egypro.2009.02.131.
40
41 5. Jena, N. R.; Mishra, P. C., An ab initio and Density Functional Study of Microsolvation of
42 Carbon Dioxide in Water Clusters and Formation of Carbonic Acid. Theor. Chem. Acc.
43 2005, 114, 189-199. DOI: http://dx.doi.org/10.1007/s00214-005-0660-1.
44 6. Nguyen, M. T.; Matus, M. H.; Jackson, V. E.; Ngan, V. T.; Rustad, J. R.; Dixon, D. A.,
45 Mechanism of the Hydration of Carbon Dioxide: Direct Participation of H2O versus
46 Microsolvation. J. Phys. Chem. A 2008, 112, 10386-10398. DOI:
47
48
http://dx.doi.org/10.1021/jp804715j.
49 7. Loerting, T.; Bernard, J., Aqueous Carbonic Acid (H2CO3). ChemPhysChem 2010, 11,
50 2305-2309. DOI: http://dx.doi.org/10.1002/cphc.201000220.
51 8. Loerting, T.; Tautermann, C.; Kroemer, R. T.; Kohl, I.; Hallbrucker, A.; Mayer, E.; Liedl,
52 K. R., On the Surprising Kinetic Stability of Carbonic Acid (H2CO3). Angew. Chem. Int.
53 Ed. 2000, 39, 891-894. DOI: http://dx.doi.org/10.1002/(sici)1521-3773(20000303)39:5
54
9. Tautermann, C. S.; Voegele, A. F.; Loerting, T.; Kohl, I.; Hallbrucker, A.; Mayer, E.; Liedl,
55
56
K. R., Towards the Experimental Decomposition Rate of Carbonic Acid (H2CO3) in
57
58
59
34
1
2
3 Aqueous Solution. Chem. Eur. J. 2002, 8, 66-73. DOI: http://dx.doi.org/10.1002/1521-
4
5
3765(20020104)8:1.
6 10. Kumar, P. P.; Kalinichev, A. G.; Kirkpatrick, R. J., Dissociation of Carbonic Acid: Gas
7 Phase Energetics and Mechanism from ab initio Metadynamics Simulations. J. Chem. Phys.
8 2007, 126, 204315-204315. DOI: http://dx.doi.org/10.1063/1.2741552.
9 11. Gallet, G. A.; Pietrucci, F.; Andreoni, W., Bridging Static and Dynamical Descriptions of
10 Chemical Reactions: An ab Initio Study of CO2 Interacting with Water Molecules. J. Chem.
11
Theory Comput. 2012, 8, 4029-4039. DOI: http://dx.doi.org/10.1021/ct300581n.
12
13
12. Saharay, M.; Balasubramanian, S., Evolution of Intermolecular Structure and Dynamics in
14 Supercritical Carbon Dioxide with Pressure: An ab Initio Molecular Dynamics Study. J.
15 Phys. Chem. B 2007, 111, 387-392. DOI: http://dx.doi.org/10.1021/jp065679t.
16 13. Saharay, M.; Balasubramanian, S., Ab Initio Molecular Dynamics Investigations of
17 Structural, Electronic and Dynamical Properties of Water in Supercritical Carbon Dioxide.
18 Indian J. Phys. 2009, 83, 13-29. DOI: http://dx.doi.org/10.1007/s12648-009-0001-7.
19
14. Glezakou, V.-A.; Rousseau, R.; Dang, L. X.; McGrail, B. P., Structure, Dynamics and
20
21 Vibrational Spectrum of Supercritical CO2/H2O Mixtures from ab initio Molecular
22 Dynamics as a Function of Water Cluster Formation. Phys. Chem. Chem. Phys. 2010, 12,
23 8759-8759. DOI: http://dx.doi.org/10.1039/b923306g.
24 15. Windisch, C. F.; Glezakou, V.-A.; Martin, P. F.; McGrail, B. P.; Schaef, H. T., Raman
25 Spectrum of Supercritical C18O2 and Re-evaluation of the Fermi Resonance. Phys. Chem.
26 Chem. Phys. 2012, 14, 2560-2566. DOI: http://dx.doi.org/10.1039/C1CP22349F.
27
28
16. Cukierman, S., Et tu, Grotthuss! and Other Unfinished Stories. Biochim. Biophys. Acta,
29 Bioenerg. 2006, 1757, 876-885. DOI: http://dx.doi.org/10.1016/j.bbabio.2005.12.001.
30 17. CP2K Developers Group. CP2K, version 6.1; http://www.cp2k.org/.
31 18. Bonomi, M.; Branduardi, D.; Bussi, G.; Camilloni, C.; Provasi, D.; Raiteri, P.; Donadio, D.;
32 Marinelli, F.; Pietrucci, F.; Broglia, R. A.; Parrinello, M., PLUMED: A Portable Plugin for
33 Free-Energy Calculations with Molecular Dynamics. Comput. Phys.Commun. 2009, 180,
34
1961-1972. DOI: http://dx.doi.org/10.1016/j.cpc.2009.05.011.
35
36
19. Tribello, G. A.; Bonomi, M.; Branduardi, D.; Camilloni, C.; Bussi, G., PLUMED 2: New
37 Feathers for an Old Bird. Comput. Phys.Commun. 2014, 185, 604-613. DOI:
38 http://dx.doi.org/10.1016/j.cpc.2013.09.018.
39 20. VandeVondele, J.; Hutter, J., Gaussian Basis Sets for Accurate Calculations on Molecular
40 Systems in Gas and Condensed Phases. J. Chem. Phys. 2007, 127, 114105. DOI:
41 http://dx.doi.org/10.1063/1.2770708.
42
21. Goedecker, S.; Teter, M.; Hutter, J., Separable Dual-Space Gaussian Pseudopotentials.
43
44 Phys. Rev. B 1996, 54, 1703-1710. DOI: http://dx.doi.org/10.1103/PhysRevB.54.1703.
45 22. Perdew, J. P.; Burke, K.; Ernzerhof, M., Generalized Gradient Approximation Made
46 Simple. Phys. Rev. Lett. 1996, 77, 3865-3868. DOI:
47 http://dx.doi.org/10.1103/PhysRevLett.77.3865.
48 23. Grimme, S.; Bannwarth, C.; Shushkov, P., A Robust and Accurate Tight-Binding Quantum
49 Chemical Method for Structures, Vibrational Frequencies, and Noncovalent Interactions of
50
51
Large Molecular Systems Parametrized for All spd-Block Elements (Z = 1–86). J. Chem.
52 Theory Comput. 2017, 13, 1989-2009. DOI: http://dx.doi.org/10.1021/acs.jctc.7b00118.
53 24. Bussi, G.; Donadio, D.; Parrinello, M., Canonical Sampling Through Velocity Rescaling. J.
54 Chem. Phys. 2007, 126, 014101. DOI: http://dx.doi.org/10.1063/1.2408420.
55
56
57
58
59
35
1
2
3 25. Barducci, A.; Bonomi, M.; Parrinello, M., Metadynamics. WIREs Comput. Mol. Sci. 2011,
4
5
1, 826-843. DOI: http://dx.doi.org/10.1002/wcms.31.
6 26. Barducci, A.; Bussi, G.; Parrinello, M., Well-Tempered Metadynamics: A Smoothly
7 Converging and Tunable Free-Energy Method. Phys. Rev. Lett. 2008, 100, 020603. DOI:
8 http://dx.doi.org/10.1103/PhysRevLett.100.020603.
9 27. Bussi, G.; Laio, A.; Parrinello, M., Equilibrium Free Energies from Nonequilibrium
10 Metadynamics. Phys. Rev. Lett. 2006, 96, 090601. DOI:
11
http://dx.doi.org/10.1103/PhysRevLett.96.090601.
12
13
28. Valsson, O.; Tiwary, P.; Parrinello, M., Enhancing Important Fluctuations: Rare Events and
14 Metadynamics from a Conceptual Viewpoint. Annu. Rev. Phys. Chem. 2016, 67, 159-184.
15 DOI: http://dx.doi.org/10.1146/annurev-physchem-040215-112229.
16 29. Schwerdtfeger, C. A.; Mazziotti, D. A., Populations of Carbonic Acid Isomers at 210 K
17 from a Fast Two-Electron Reduced-Density Matrix Theory. J. Phys. Chem. A 2011, 115,
18 12011-12016. DOI: http://dx.doi.org/10.1021/jp2057805.
19
30. Wight, C. A.; Boldyrev, A. I., Potential Energy Surface and Vibrational Frequencies of
20
21 Carbonic Acid. J. Phys. Chem. 1995, 99, 12125-12130. DOI:
22 http://dx.doi.org/10.1021/j100032a012.
23 31. Bernard, J.; Seidl, M.; Kohl, I.; Liedl, K. R.; Mayer, E.; Gálvez, Ó.; Grothe, H.; Loerting,
24 T., Back Cover: Spectroscopic Observation of Matrix-Isolated Carbonic Acid Trapped from
25 the Gas Phase (Angew. Chem. Int. Ed. 8/2011). Angew. Chem. Int. Ed. 2011, 50, 1946-
26 1946. DOI: http://dx.doi.org/10.1002/anie.201008066.
27
28
32. Mori, T.; Suma, K.; Sumiyoshi, Y.; Endo, Y., Spectroscopic Detection of Isolated Carbonic
29 Acid. J. Chem. Phys. 2009, 130, 204308. DOI: http://dx.doi.org/10.1063/1.3141405.
30 33. Pietrucci, F.; Saitta, A. M., Formamide Reaction Network in Gas Phase and Solution Via a
31 Unified Theoretical Approach: Toward a Reconciliation of Different Prebiotic Scenarios.
32 Proc. Nat. Acad. Sci. U.S.A. 2015, 112, 15030-15035. DOI:
33 http://dx.doi.org/10.1073/pnas.1512486112.
34
34. Pietrucci, F., Strategies for the Exploration of Free Energy Landscapes: Unity in Diversity
35
36
and Challenges Ahead. Rev. Phys. 2017, 2, 32-45. DOI:
37 http://dx.doi.org/10.1016/j.revip.2017.05.001.
38 35. Polino, D.; Parrinello, M., Kinetics of Aqueous Media Reactions via Ab Initio Enhanced
39 Molecular Dynamics: The Case of Urea Decomposition. J. Phys. Chem. B 2019, 123, 6851-
40 6856. DOI: http://dx.doi.org/10.1021/acs.jpcb.9b05271.
41 36. Grifoni, E.; Piccini, G.; Parrinello, M., Microscopic Description of Acid–Base Equilibrium.
42
Proc. Nat. Acad. Sci. 2019, 116, 4054-4057. DOI:
43
44 http://dx.doi.org/10.1073/pnas.1819771116.
45 37. Pocker, Y.; Bjorkquist, D. W., Stopped-Flow Studies of Carbon Dioxide Hydration and
46 Bicarbonate Dehydration in Water and Water-d2. Acid-Base and Metal Ion Catalysis. J.
47 Am. Chem. Soc. 1977, 99, 6537-6543. DOI: http://dx.doi.org/10.1021/ja00462a012.
48 38. Wang, X.; Conway, W.; Burns, R.; McCann, N.; Maeder, M., Comprehensive Study of the
49 Hydration and Dehydration Reactions of Carbon Dioxide in Aqueous Solution. J. Phys.
50
51
Chem. A 2010, 114, 1734-1740. DOI: http://dx.doi.org/10.1021/jp909019u.
52 39. Galib, M.; Hanna, G., Mechanistic Insights into the Dissociation and Decomposition of
53 Carbonic Acid in Water via the Hydroxide Route: An Ab Initio Metadynamics Study. J.
54 Phys. Chem. B 2011, 115, 15024-15035. DOI: http://dx.doi.org/10.1021/jp207752m.
55
56
57
58
59
36
1
2
3 40. Gimondi, I.; Tribello, G. A.; Salvalaglio, M., Building Maps in Collective Variable Space.
4
5
J. Chem. Phys. 2018, 149, 104104. DOI: http://dx.doi.org/10.1063/1.5027528.
6 41. Tiwary, P.; Parrinello, M., A Time-Independent Free Energy Estimator for Metadynamics.
7 J. Phys. Chem. B 2015, 119, 736-742. DOI: http://dx.doi.org/10.1021/jp504920s.
8 42. Peterson, K. I.; Klemperer, W., Structure and Internal Rotation of H2O–CO2, HDO–CO2,
9 and D2O–CO2 van der Waals Complexes. J. Chem. Phys. 1984, 80, 2439-2445. DOI:
10 http://dx.doi.org/10.1063/1.446993.
11
43. Walsh, A. D., 467. The Electronic Orbitals, Shapes, and Spectra of Polyatomic Molecules.
12
13
Part II. Non-hydride AB2 and BAC Molecules. J. Chem. Soc. 1953, 2266-2288. DOI:
14 http://dx.doi.org/10.1039/JR9530002266.
15 44. Stirling, A. s.; Pápai, I., H2CO3 Forms via HCO3− in Water. J. Phys. Chem. B 2010, 114,
16 16854-16859. DOI: http://dx.doi.org/10.1021/jp1099909.
17 45. Tho, N. M.; Ha, T. K., A Theoretical Study of the Formation of Carbonic Acid from the
18 Hydration of Carbon Dioxide: A Case of Active Solvent Catalysis. J. Am. Chem. Soc. 1984,
19
106, 599-602. DOI: http://dx.doi.org/10.1021/ja00315a023.
20
21 46. Ludwig, R.; Kornath, A., In Spite of the Chemist's Belief: Carbonic Acid is Surprisingly
22 Stable. Angew. Chem. Int. Ed. 2000, 39, 1421-1423. DOI:
23 http://dx.doi.org/10.1002/(SICI)1521-3773(20000417)39:8.
24 47. Wang, B.; Cao, Z., How Water Molecules Modulate the Hydration of CO2 in Water
25 Solution: Insight from the Cluster-Continuum Model Calculations. J. Comput. Chem. 2013,
26 34, 372-378. DOI: http://dx.doi.org/10.1002/jcc.23144.
27
28
48. Meier, J.; Schwarzenbach, G., Eine mit Glaselektroden ausgerüstete Strömungsapparatur
29 (Über die Aciditätskonstante der wahren Kohlensäure und deren
30 Dehydratationsgeschwindigkeit). Helv. Chim. Acta 1957, 40, 907-917. DOI:
31 http://dx.doi.org/10.1002/hlca.19570400405.
32 49. Magid, E.; Turbeck, B. O., The Rates of the Spontaneous Hydration of CO2 and the
33 Reciprocal Reaction in Neutral Aqueous Solutions between 0° and 38°. Biochim. Biophys.
34
Acta, Gen. Subj. 1968, 165, 515-524. DOI: http://dx.doi.org/10.1016/0304-4165(68)90232-
35
36
8.
37 50. Galib, M.; Hanna, G., The Role of Hydrogen Bonding in the Decomposition of H2CO3 in
38 Water: Mechanistic Insights from Ab Initio Metadynamics Studies of Aqueous Clusters. J.
39 Phys. Chem. B 2014, 118, 5983-5993. DOI: http://dx.doi.org/10.1021/jp5029195.
40 51. Adamczyk, K.; Prémont-Schwarz, M.; Pines, D.; Pines, E.; Nibbering, E. T. J., Real-Time
41 Observation of Carbonic Acid Formation in Aqueous Solution. Science 2009, 326, 1690-
42
1694. DOI: http://dx.doi.org/10.1126/science.1180060.
43
44 52. Pines, D.; Ditkovich, J.; Mukra, T.; Miller, Y.; Kiefer, P. M.; Daschakraborty, S.; Hynes, J.
45 T.; Pines, E., How Acidic Is Carbonic Acid? J. Phys. Chem. B 2016, 120, 2440-2451. DOI:
46 http://dx.doi.org/10.1021/acs.jpcb.5b12428.
47 53. De Sousa, R.; Alves, H., Ab initio Calculation of the Dynamical Properties of PPP and
48 PPV. Braz. J. Phys. 2006, 36, 501-504. DOI: http://dx.doi.org/10.1590/S0103-
49 97332006000300072.
50
51
54. Ndongmouo, U. F. T.; Lee, M. S.; Rousseau, R.; Baletto, F.; Scandolo, S., Finite-
52 Temperature Effects on the Stability and Infrared Spectra of HCl(H2O)6 Clusters. J. Phys.
53 Chem. A 2007, 111, 12810-12815. DOI: http://dx.doi.org/10.1021/jp0765603.
54
55
56
57
58
59
37
1
2
3 55. Lee, M.-S.; Peter McGrail, B.; Rousseau, R.; Glezakou, V.-A., Structure, Dynamics and
4
5
Stability of Water/scCO2/Mineral Interfaces from ab initio Molecular Dynamics
6 Simulations. Sci. Rep. 2015, 5, 14857-14857. DOI: http://dx.doi.org/10.1038/srep14857.
7 56. Invernizzi, M.; Parrinello, M., Rethinking Metadynamics. arXiv:1909.07250 2019.
8 https://arxiv.org/abs/1909.07250.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
38
1
2
3 TOC
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23 Free energy barriers for the reaction CO2 + H2O in three different environments.
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
39
1
2
3 Daniela Polino
4
5
6 Daniela Polino received her MS in 2009 and her PhD in 2013 from the
7 Politecnico di Milano under the supervision of Professor C. Cavallotti. In
8 2012, she was a visiting scholar at Argonne National Laboratory working
9 with Dr. S. J. Klippenstein. She is currently a postdoctoral fellow in
10 Professor M. Parrinello’s group at the Università della Svizzera Italiana
11
12
USI, Faculty of Informatics. Her research interests are focused on the
13 molecular modeling of complex chemical systems (in gas or condensed
14 phase) that are relevant to the chemical industry.
15
16
17
18 Emanuele Grifoni
19
20 Emanuele Grifoni received his MS in Chemistry from the University of
21
22
Florence in 2017 under the supervision of Professors R. Chelli and G.
23 Cardini. After graduating, he joined the group of Professor M. Parrinello at
24 ETH Zurich as a PhD student. In 2019 he spent six months as an AFS at the
25 Pacific Northwest National Laboratory working with Dr. V.-A. Glezakou and
26 Dr R. Rousseau. His current research interests are focused on the
27 development of new approaches to study rare events via molecular dynamics
28
simulations.
29
30
31
32
33 Roger Rousseau
34
35 Dr. Roger Rousseau received his Ph.D. from the University of Michigan in
36 1995 (Chemistry) and followed up with a postdoctoral research
37 appointment at MPI-Stuttgart (1996–1998). Prior to joining PNNL, he was
38
39
an associate professor of theoretical condensed matter physics at the
40 International School for Advanced Studies in Trieste, Italy (2003–2007).
41 He currently studies theory and simulation of complex reactive systems for
42 application in the development of energy efficient technologies.
43
44
45
46 Michele Parrinello
47
48
49
Michele Parrinello is currently Professor at ETH Zurich, and the Università
50 della Svizzera italiana Lugano, Switzerland. He is known for his many
51 technical innovations in the field of atomistic simulations and for a wealth of
52 interdisciplinary applications ranging from materials science to chemistry and
53 biology. For his work he has been awarded the 2011 Prix Benoist, the 2017
54 Dreyfus Prize, the European Chemistry Gold Metal and Benjamin Franklin
55
Medal in 2020 along with many other prizes and honorary degrees. He is a
56
57
58
59
40
1
2
3 member of numerous academies and learned societies, including the National Academy of
4
5
Science, the British Royal Society and the Italian Accademia Nazionale dei Lincei. He is the
6 author of more than 600 papers and his work is highly cited.
7
8
9
10 Vassiliki-Alexandra Glezakou
11
12 Dr. Vassiliki-Alexandra Glezakou obtained her BS in Chemistry from the
13
National Kapodistrian University of Athens, Greece and her Ph.D. in theoretical
14
15
physical chemistry from Iowa State University in 2000. After a postdoctoral
16 appointment at University of California San Diego (2000–2002) and a staff
17 research position at Caltech (2003), she joined PNNL in 2004, where she is a
18 Chief Scientist. Her current research activities focus on modeling structure,
19 spectroscopy, and reactivity of complex interfaces with emphasis on
20 separations, clean energy processes and catalysis.
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
41

10 1021@acs Jpca 9b11744

Uploaded by

Copyright:

Available Formats

10 1021@acs Jpca 9b11744

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

10 1021@acs Jpca 9b11744

Uploaded by

Copyright:

Available Formats

Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE

Reactivity and Speciation in Different Media

is published by the American Chemical Society. 1155 Sixteenth Street N.W.,

You might also like