Rigorous pKa Estimation of Amine Species Using ... - ACS Publications

Article Cite This: J. Chem. Theory Comput. 2018, 14, 351−356

pubs.acs.org/JCTC

Rigorous pKa Estimation of Amine Species Using Density-Functional Tight-Binding-Based Metadynamics Simulations Aditya Wibawa Sakti,† Yoshifumi Nishimura,‡ and Hiromi Nakai*,†,‡,§,∥

Downloaded via UNIV OF THE SUNSHINE COAST on June 25, 2018 at 17:40:58 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

†

Department of Chemistry and Biochemistry, School of Advanced Science and Engineering, Waseda University, Tokyo 169-8555, Japan ‡ Research Institute for Science and Engineering, Waseda University, Tokyo 169-8555, Japan § Core Research for Evolutional Science and Technology (CREST), Japan Science and Technology Agency (JST), Saitama 332-0012, Japan ∥ Elements Strategy Initiative for Catalysts and Batteries (ESICB), Kyoto University, Kyoto 615-8520, Japan S Supporting Information *

ABSTRACT: Predicting pKa values for different types of amine species with high accuracy and efficiency is of critical importance for the design of high performance and economical solvents in carbon capture and storage with aqueous amine solutions. In this study, we demonstrate that densityfunctional tight-binding-based metadynamics simulations are a promising approach to calculate the free energy difference between the protonated and neutral states of amines in aqueous solution with inexpensive computational cost. The calculated pKa values were in satisfactory agreement with the experimental values, the mean absolute deviation being only 0.09 pKa units for 34 amines commonly used in CO2 scrubbing. Such superior reproducibility and correlation compared to estimations by static quantum mechanical calculations highlight the significant effect of dynamical proton transfer processes in explicit solvent molecules for the improvement of the estimation accuracy.

1. INTRODUCTION In the present era, the increasing abundance of CO2 in the atmosphere is arguably leading to global warming.1 To reduce the CO2 levels emitted by fossil fuel combustion sources, consideration efforts have been devoted to the development of carbon capture and storage (CCS) systems.2−4 Aqueous amine solutions play a central role as advanced materials in CCS by chemical absorption.5,6 The pKa values of the conjugate acids of amines are considered to be an indicator of absorption efficiency.7,8 Experimental measurements have thus been performed for a wide range of amine species.8−11 On the theoretical side, quantitative evaluation of pKa values is still challenging. The routinely adopted procedure is a combination of thermodynamic cycles and free energy calculations of associated molecules in individual steps.12,13 Since the pioneering study by da Silva and Svendsen,14 a great deal of attention has been paid to the pKa calculation of amines, in which solvent effects are included implicitly and often explicitly.15−21 In these early theoretical studies, a mean absolute deviation (MAD) of 0.2−0.4 pKa units relative to the selected experimental values was considered an elaborated achievement. Unfortunately, static theoretical calculations are unable to process the various possible solvation patterns in deprotonation reactions. The delicate structural changes and proton transfer © 2017 American Chemical Society

processes involved in acid dissociation reactions can be favorably described by full quantum mechanical molecular dynamics (MD) simulations. Density functional theory (DFT)based MD simulations have provided not only mechanistic insight22−24 but also the ability to predict pKa values25−33 for typical acid−base reactions. Most of those investigations have adopted the metadynamics sampling technique34−36 to study such infrequent deprotonation events within accessible simulation times. The free energy difference between the protonated and deprotonated species calculated from reconstructed free energy profiles in metadynamics has shown an error of approximately 0.2 pK a units relative to the experimentally determined values.26−29 The density-functional tight-binding (DFTB)37−39 method is a highly efficient approximation to the Kohn−Sham DFT, which uses atomic and diatomic parameters obtained from DFT calculations. Computational speeds 2−3 orders magnitude faster than that of conventional DFT implementation have been achieved.37,39 Furthermore, the DFTB method retains the robustness of the chemical reaction dynamics by virtue of the self-consistent determination of charge distributions.40 The authors’ group has implemented an efficient program of the Received: August 11, 2017 Published: December 5, 2017 351

DOI: 10.1021/acs.jctc.7b00855 J. Chem. Theory Comput. 2018, 14, 351−356

Article

Journal of Chemical Theory and Computation DFTB-MD simulation,41 especially, by combining with the linear scaling divide-and-conquer technique.42−45 The successes of describing several chemical phenomena in aqueous solutions46−48 have stimulated us to investigate whether the DFTB-based simulations can become an alternative for predicting pKa values of amines that enables fast screening for new compounds. In the present study, we have estimated the pKa values of amine species in solution using metadynamics simulations at the DFTB level. Although the combination of the DFTB method and the metadynamics technique has attracted recent interest for application in complex environments and yielded encouraging results,49−51 evaluation of DFTB accuracy for various types of amine species needs to be addressed for prediction purpose. To this end, the calculated pKa values are compared with available experimental data for 34 amine species. We also examine the effect of system size to the quality of theoretical predictions, which is hard to obtain from computationally more demanding DFT-based MD simulations. The paper is organized as follows. The computational details of the DFTB metadynamics simulations are provided in section 2. Section 3 presents the accuracy and efficiency of the investigated approach. Finally, concluding remarks are described in section 4.

where rNHi is the distance between the nitrogen and hydrogen atoms, r0 is the cutoff distance set to 1.6 Å, and the summation runs over the NH hydrogen atoms initially coordinated to the nitrogen atom. In the DFTB metadynamics simulations, the height and width of the Gaussian function to describe the bias potential were set to 1.31 kJ/mol and 0.1, respectively. Note that the CV is dimensionless. The Gaussian bias potential was deposited at time intervals of 0.04 ps. After reaching the convergence of free energy profile, the pKa value at temperature T was calculated using the relationship

pK a =

ΔG 2.303RT

(3)

where R is the universal gas constant and ΔG is the free energy difference between the two states in eq 1, directly calculated from the sampled free energy profile.

3. RESULTS AND DISCUSSION 3.1. pKa Estimation of Aqueous MEA Solution. First, we discuss the accuracy and efficiency of DFTB metadynamics simulations taking MEA, which is one of the most commonly used amines in CO2 capture process,5 as representative example. The time-course change of the CV during the

2. COMPUTATIONAL DETAILS Canonical ensemble simulations were performed for a cubic cell with three-dimensional periodic boundary conditions unless noted otherwise. In the initial condition, the system consisted of one protonated amine, one hydroxide ion, and 50 water molecules. For monoethanolamine (MEA), three larger simulation boxes containing 100, 200, and 500 water molecules were additionally prepared to investigate the system size dependence on the calculated pKa values. The density of each system was determined from classical MD simulations using the general CHARMM force field52 and NAMD package.53 This pre-equilibration procedure was performed for 3 ns with a time step of 1.0 fs under NVE, NVT, and NPT ensembles. The classically equilibrated structure was further equilibrated at the third-order DFTB (DFTB3)54 level with 3ob parameter set55 for 50 ps with a time step of 1.0 fs. The choice of DFTB model and parameter was based on the best estimation of pKa values for MEA at two different temperatures among several DFTB variants considered in preliminary simulations. During the DFTB equilibration, the temperature was controlled using an Andersen thermostat at 298.15 or 313.15 K. The structure and velocity of last equilibration step were adopted as initial conditions of the DFTB metadynamics simulations. In the production run, the time step was 0.5 fs. All DFTB calculations were performed using the DC-DFTB-K program.41 To observe the deprotonation reaction given by [amine(H)]+ + H 2O ⇌ amine + H3O+

(1)

the number of hydrogen atoms coordinated to the nitrogen atom of the amine species (nNH) was chosen as the collective variable (CV) for the DFTB metadynamics simulations. In accordance with previous successful free energy samplings,26−29,56 the rational form of nNH was adopted NH

nNH =

∑ i=1

1 − (rNHi /r0)6 1 − (rNHi /r0)12

Figure 1. Time-course changes of the (a) CV, nNH, and (b) free energy difference, and (c) the free energy profile as a function of the CV for the deprotonation of MEA(P) at 313.15 K.

(2) 352

DOI: 10.1021/acs.jctc.7b00855 J. Chem. Theory Comput. 2018, 14, 351−356

Article

Journal of Chemical Theory and Computation

Table 1. Comparison of the pKa Values Calculated from DFTB Metadynamics Simulations with the Experimental Results for 34 Amine Species in Aqueous Solution entry

amine

abbreviation

typea

T (K)

pKa (calcd)

pKa (exptl.)

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 MAD

monoethanolamine 2-amino-2-methyl-1-propanol 2-(2-aminoethoxy)ethanol 2-amino-2-ethyl-1,3-propanediol 3-amino-1-propanol serinol(2-aminopropane-1,3-diol) tris(hydroxymethyl)aminomethane methylmonoethanolamine 2-(ethylamino)ethanol 2-(tert-butylamino)ethanol diisopropanolamine n-cyclohexylethanolamine diethanolamine n-cyclopentylethanolamine tricine n-cyclopropylethanolamine methyldiethanolamine 2-(diisopropylamino)ethanol triethanolamine tris[2-(2-methoxyethoxy)ethyl]amine triethylamine 3-dimethylamino-1-propanol ethyldiethanolamine N,N-dimethylisopropanolamine tert-butyldiethanolamine piperazine 2-piperidinemethanol 2-piperidineethanol 3-piperidinemethanol 1-piperidinepropanolamine 3-piperidino-1,2-propanediol 3-quinoclidinol 4,2-hydroxyethylmorpholine 3-morpholinopropylamine

MEA AMP 2-AEE AEPD MPA SAPD THMAM MMEA EAE TBAE DIPA n-CHEA DEA n-CPEA TRC n-CPREA MDEA 2-DIPA TEA TMEEA TREA 3-DMAP EDEA DMIPA t-BDEA PZ 2-PPM 2-PPE 3-PPM 1-PPP 3-PPPD 3-QCD 4,2-HEMO 3-MOPA

1° 1° 1° 1° 1° 1° 1° 2° 2° 2° 2° 2° 2° 2° 2° 2° 3° 3° 3° 3° 3° 3° 3° 3° 3° c-2° c-2° c-2° c-2° c-3° c-3° c-3° c-3° c-3°

313.15 313.15 298.15 298.15 298.15 298.15 298.15 298.15 298.15 298.15 298.15 298.15 298.15 298.15 298.15 298.15 313.15 313.15 313.15 313.15 313.15 298.15 298.15 298.15 298.15 313.15 298.15 298.15 298.15 298.15 298.15 298.15 298.15 298.15

9.14 9.16 9.49 8.78 9.86 8.58 8.16 9.94 10.00 9.76 8.80 9.89 8.80 10.25 8.00 8.50 8.43 9.27 7.41 6.67 10.30 9.30 8.66 9.45 8.92 9.30 10.47 11.08 10.98 9.32 8.79 10.22 6.82 9.89 0.09

9.09b 9.23c 9.42c 8.82c 9.96c 8.55d 8.08d 9.85c 10.00d 9.70d 8.88d 10.10d 8.88d 10.10d 8.10d 8.40d 8.31b 9.11b 7.45c 6.69b 10.32c 9.27d 8.80e 9.47e 9.03e 9.38b 10.60d 10.90d 10.80d 9.43d 8.91d 10.10d 6.90e 9.95e

a Primary, secondary, tertiary, cyclic secondary, and cyclic tertiary amines are denoted as 1°, 2°, 3°, c-2°, and c-3°, respectively. bRef 10. cRef 11. dRef 8. eRef 9.

at 3.5−4.8 ps. The fluctuation at 5.2−5.9 ps is a manifestation of the instantaneous formation of a MEA(P)−water pair. The CV starts to be diffusive from ∼6.5 ps, at which the metadynamics sampling was stopped. In the diffusive region, the proton dissociated from MEA(P) migrates through the hydrogen bond network of water molecules, the so-called Grotthuss shuttling.57,58 The relatively short simulation time to complete the reaction in eq 1 is a consequence of the sufficient equilibration achieved before the production run.26−29 Figure 1b shows the time-course change of the free energy difference between the protonated and neutral states of MEA at 313.15 K. Sampling of the diffusive region hardly changes the free energy difference, demonstrating that the reconstructed free energy profile is well converged. Figure 1c shows the free energy profile for the deprotonation of MEA(P) at 313.15 K. Two minima are found at nNH = ∼1.9 and ∼2.8, representing the MEA and MEA(P) states, respectively. The calculated free energy difference is 54.19 kJ/ mol, equivalent to a pKa value of 9.15. This estimation is in good agreement with the experimental value of 9.09.10

Figure 2. Correlation between the calculated pKa values from DFTB metadynamics simulations and the experimental results8−11 for 34 amine species in aqueous solution.

deprotonation process of protonated MEA (MEA(P)) at 313.15 K is shown in Figure 1a. The nNH values of ∼1.9 and ∼2.8 correspond to MEA and MEA(P), respectively. A proton is shuttling between MEA(P) and an adjacent water molecule 353

DOI: 10.1021/acs.jctc.7b00855 J. Chem. Theory Comput. 2018, 14, 351−356

Article

Journal of Chemical Theory and Computation

4. CONCLUSION In conclusion, DFTB-based metadynamics simulations are able to estimate pKa values of amine species in aqueous solution with remarkable accuracy. The calculated MAD for 34 amines commonly used in CO2 capture was only 0.09 pKa units with respect to the experimental results, superior by a factor of 2−3 to common static theoretical calculations. The observed accuracy may be attributed to the inclusion of Grotthuss shuttling effects and the sampling of solvation shells around the reaction centers during the deprotonation process. In contrast, the single-point calculations at the DFTB3/3OB level underestimate the gas-phase acidities and proton affinities of amines, as shown in Table S1. Incorporation of polarizable continuum model cannot improve the pKa values for single-point calculations, as shown in Table S2. We also found that the pKa values obtained with DFTB metadynamics simulations are less sensitive with respect to the system size while temperature effect can be incorporated adequately. Although possible error sources exist, such as metadynamics sampling errors, DFTB model/parameter dependence, and the neglect quantum effects, those negative factors seem to be canceled out in the resultant free energy difference results.26−29,59 The required computational demands are small with no deterioration of the prediction accuracy; hence, the present approach presents a great advantage over DFT- and/or wave function-based protocols in that prospective candidates for efficient chemical absorption of CO2 can be explored in a feasible manner. The accurate determination of acid dissociation constants for different groups of compounds may be feasible after careful inspection of the DFTB performance.

We also performed the DFTB metadynamics simulations of MEA at different conditions. One is to treat different unit cell sizes containing a larger number of water molecules. The timecourse changes of CV and calculated free energy profiles for 100, 200, and 500 water molecules are presented in Figures S1−S3. The estimated pKa values of 100, 200, and 500 water molecules system are 9.19, 9.03, and 9.07, respectively. Considering the insensitivity of calculated pKa value with respect to the system size, 50 water molecules to solvate amine species seems to be sufficient for prediction purpose and was used throughout this study. The other is to calculate the pKa value at 298.15 K (for corresponding simulation results, see Figure S4). The agreement of estimation (9.40) with experimental data (9.47)10 is again excellent, which indicates that the metadynamics sampling is able to capture the effect of slight change in temperature reasonably. The efficiency of the DFTB metadynamics simulation is remarkable. Calculation of a 60 ps long trajectory took less than an hour using an 8-core Intel Xeon E5−2637 v3 (3.50 GHz) workstation. The results demonstrate that the region shown in Figure 1 can be generated within minutes. In contrast, previous DFT-based metadynamics simulations of systems containing a similar number of solvent water molecules have been reported to require more expensive computational resources (i.e., an IBM Blue Gene cluster with 256 processors).26,29 3.2. pKa Estimation of Various Types of Amines. The obtained results in section 3.1 give us confidence to evaluate the performance of DFTB metadynamics simulations for amines being in common use in the CCS technology. Specifically, the number of target amines was extended to 34 in total, including seven primary (1°), nine secondary (2°), nine tertiary (3°), four cyclic secondary (c-2°), and five cyclic tertiary (c-3°) species. A comparison of the calculated pKa values with the experimental results8−11 is shown in Table 1. The time-course changes of the CV and the simulated free energy profiles are given Figures S5−S37, except for MEA at 313.15 K, already shown in Figure 1. In all cases, sampling of reactive region to fulfill the free energy profile with bias potential was continued no longer than 12 ps and protonated and neutral states were clearly distinguished from each other in the free energy profile. Overall consistency was observed between the estimated and experimental pKa values. Moreover, the obtained MAD was only 0.09 pKa units. The largest error was 0.21 pKa units (1.20 kJ/mol) for n-cyclohexylethanolamine. It should be noted that this specific group of amines does not present an obvious trend toward a certain range of pKa values. The averaged statistical error bar based on five MD trajectories for each amine is 0.26 pKa units, with the maximum deviation of 0.53 pKa units for 3-piperidino-1,2-propanediol. More details of the estimated pKa values and the simulated free energy profiles for individual metadynamics runs are given in Table S3 and Figures S38−S71. The correlation of pKa values in Table 1 is graphically displayed in Figure 2. An excellent correlation coefficient of R2 = 0.990 was obtained for the 34 amines, with pKa values ranging from 6.6 to 11.1. For comparison with previous studies using static theoretical calculations, see Figures S72−S74. The outperformance of the present approach strongly supports the need to account for the multiple arrangements of the solvent water molecules and the forward/backward proton shuttling processes to accurately determine pKa values.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jctc.7b00855. Single-point DFTB calculation results for gas-phase acidities, proton affinities, and pKa values with polarizable continuum model of amines, time-course changes of collective variable and free energy profile for deprotonation reaction of 34 amine species obtained from production run of DFTB metadynamics simulations, and correlations between estimated pKa values obtained from previous static theoretical calculations and the experimental results (ZIP)

■

AUTHOR INFORMATION

Corresponding Author

*Phone: +81-3-5286-3452. E-mail: [email protected]. ORCID

Hiromi Nakai: 0000-0001-5646-2931 Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This study was supported in part by a Grant-in-Aid for Scientific Research (A) “KAKENHI Grant Number JP26248009” from the Japan Society for the Promotion of 354

DOI: 10.1021/acs.jctc.7b00855 J. Chem. Theory Comput. 2018, 14, 351−356

Article

Journal of Chemical Theory and Computation

(18) Gupta, M.; da Silva, E. F.; Svendsen, H. F. Modeling Temperature Dependency of Amine Basicity Using PCM and SM8T Implicit Solvation Models. J. Phys. Chem. B 2012, 116, 1865−1875. (19) Zhang, S. A Reliable and Efficient First Principles-Based Method for Predicting pKa Values. 4. Organic Bases. J. Comput. Chem. 2012, 33, 2469−2482. (20) Gangarapu, S.; Marcelis, A. T. M.; Zuilhof, H. Accurate pKa Calculation of the Conjugate Acids of Alkanolamines, Alkaloids and Nucleotide Bases by Quantum Chemical Methods. ChemPhysChem 2013, 14, 990−995. (21) Gupta, M.; da Silva, E. F.; Svendsen, H. F. Postcombustion CO2 Capture Solvent Characterization Employing the Explicit Solvation Shell Model and Continuum Solvation Models. J. Phys. Chem. B 2016, 120, 9034−9050. (22) Park, J. M.; Laio, A.; Iannuzzi, M.; Parrinello, M. Dissociation Mechanism of Acetic Acid in Water. J. Am. Chem. Soc. 2006, 128, 11318−11319. (23) Lee, J.-G.; Asciutto, E.; Babin, V.; Sagui, C.; Darden, T.; Roland, C. Deprotonation of Solvated Formic Acid: Car-Parrinello and Metadynamics Simulations. J. Phys. Chem. B 2006, 110, 2325−2331. (24) Galib, M.; Hanna, G. Mechanistic Insights into the Dissociation and Decomposition of Carbonic Acid in Water via the Hydroxide Route: An Ab Initio Metadynamics Study. J. Phys. Chem. B 2011, 115, 15024−15035. (25) Ivanov, I.; Chen, B.; Raugei, S.; Klein, M. L. Relative pKa Values from First-Principles Molecular Dynamics: The Case of Histidine Deprotonation. J. Phys. Chem. B 2006, 110, 6365−6371. (26) Tummanapelli, A. K.; Vasudevan, S. Dissociation Constants of Weak Acids from Ab Initio Molecular Dynamics Using Metadynamics: Influence of the Inductive Effect and Hydrogen Bonding on pKa Values. J. Phys. Chem. B 2014, 118, 13651−13657. (27) Tummanapelli, A. K.; Vasudevan, S. Ab Initio Molecular Dynamics Simulations of Amino Acids in Aqueous Solutions: Estimating pKa Values from Metadynamics Sampling. J. Phys. Chem. B 2015, 119, 12249−12255. (28) Tummanapelli, A. K.; Vasudevan, S. Ab Initio MD Simulations of the Brønsted Acidity of Glutathione in Aqueous Solutions: Predicting pKa Shifts of the Cysteine Residue. J. Phys. Chem. B 2015, 119, 15353−15358. (29) Tummanapelli, A. K.; Vasudevan, S. Estimating Successive pKa Values of Polyprotic Acids from Ab Initio Molecular Dynamics Using Metadynamics: The Dissociation of Phthalic Acid and Its Isomers. Phys. Chem. Chem. Phys. 2015, 17, 6383−6388. (30) Leung, K.; Nielsen, I. M. B.; Criscenti, L. J. Elucidating the Bimodal Acid-Base Behavior of the Water-Silica Interface from First Principles. J. Am. Chem. Soc. 2009, 131, 18358−18365. (31) Sulpizi, M.; Sprik, M. Acidity Constants from DFT-Based Molecular Dynamics Simulations. J. Phys.: Condens. Matter 2010, 22, 284116. (32) Mangold, M.; Rolland, L.; Costanzo, F.; Sprik, M.; Sulpizi, M.; Blumberger, J. Absolute pKa Values and Solvation Structure of Amino Acids from Density Functional Based Molecular Dynamics Simulation. J. Chem. Theory Comput. 2011, 7, 1951−1961. (33) Gaigeot, M.-P.; Sprik, M.; Sulpizi, M. Oxide/Water Interfaces: How the Surface Chemistry Modifies Interfacial Water Properties. J. Phys.: Condens. Matter 2012, 24, 124106. (34) Laio, A.; Parrinello, M. Escaping Free-Energy Minima. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 12562−12566. (35) Ensing, B.; De Vivo, M.; Liu, Z.; Moore, P.; Klein, M. L. Metadynamics as a Tool for Exploring Free Energy Landscapes of Chemical Reactions. Acc. Chem. Res. 2006, 39, 73−81. (36) Barducci, A.; Bonomi, M.; Parrinello, M. Metadynamics. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2011, 1, 826−843. (37) Seifert, G.; Joswig, J.-O. Density-Functional Tight Binding−An Approximate Density-Functional Theory Method. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2012, 2, 456−465. (38) Elstner, M.; Seifert, G. Density Functional Tight Binding. Philos. Trans. R. Soc., A 2014, 372, 20120483.

Science (JSPS), a Grand-in-Aid for Challenging and Exploratory Research “KAKENHI 15K13629” from the Ministry of Education, Culture, Sports, Science, and Technology (MEXT), Japan, and by MEXT as “Priority Issue on Post-K computer” (Development of new fundamental technologies for highefficiency energy creation, conversion/storage, and use). One of the authors (A.W.S.) acknowledges financial support from the Yoshida Scholarship Foundation (YSF). Some of the calculations were performed at the Research Center for Computational Science (RCCS), Okazaki, Japan.

■

REFERENCES

(1) Stern, N. Stern Review on the Economics of Climate Change; Cambridge University Press: Cambridge, 2006. (2) D’Alessandro, D. M.; Smit, B.; Long, J. R. Carbon Dioxide Capture: Prospects for New Materials. Angew. Chem., Int. Ed. 2010, 49, 6058−6082. (3) MacDowell, N.; Florin, N.; Buchard, A.; Hallett, J.; Galindo, A.; Jackson, G.; Adjiman, C. S.; Williams, C. K.; Shah, N.; Fennell, P. An Overview of CO2 Capture Technologies. Energy Environ. Sci. 2010, 3, 1645−1669. (4) Kenarsari, S. D.; Yang, D.; Jiang, G.; Zhang, S.; Wang, J.; Russell, A. G.; Wei, Q.; Fan, M. Review of Recent Advances in Carbon Dioxide Separation and Capture. RSC Adv. 2013, 3, 22739−22773. (5) Rochelle, G. T. Amine Scrubbing for CO2 Capture. Science 2009, 325, 1652−1654. (6) Mumford, K. A.; Wu, Y.; Smith, K. H.; Stevens, G. W. Review of Solvent Based Carbon-Dioxide Capture Technologies. Front. Chem. Sci. Eng. 2015, 9, 125−141. (7) Versteeg, G. F.; van Dijck, L. A. J.; van Swaaij, W. P. M. On the Kinetics between CO2 and Alkanolamines both in Aqueous and NonAqueous Solutions. An Overview. Chem. Eng. Commun. 1996, 144, 113−158. (8) Puxty, G.; Rowland, R.; Allport, A.; Yang, Q.; Bown, M.; Burns, R.; Maeder, M.; Attalla, M. Carbon Dioxide Postcombustion Capture: A Novel Screening Study of the Carbon Dioxide Absorption Performance of 76 Amines. Environ. Sci. Technol. 2009, 43, 6427− 6433. (9) Rayer, A. V.; Sumon, K. Z.; Jaffari, L.; Henni, A. Dissociation Constants (pKa) of Tertiary and Cyclic Amines: Structural and Temperature Dependences. J. Chem. Eng. Data 2014, 59, 3805−3813. (10) Tagiuri, A.; Mohamedali, M.; Henni, A. Dissociation Constant (pKa) and Thermodynamic Properties of Some Tertiary and Cyclic Amines from (298 to 333) K. J. Chem. Eng. Data 2016, 61, 247−254. (11) Hamborg, E. S.; Versteeg, G. F. Dissociation Constants and Thermodynamic Properties of Amines and Alkanolamines from (293 to 353) K. J. Chem. Eng. Data 2009, 54, 1318−1328. (12) Alongi, K. S.; Shields, G. C. Theoretical Calculations of Acid Dissociation Constants: A Review Article. Annu. Rep. Comput. Chem. 2010, 6, 113−138. (13) Ho, J.; Coote, M. L. A Universal Approach for Continuum Solvent pKa Calculations: Are We There Yet? Theor. Chem. Acc. 2010, 125, 3−21. (14) Da Silva, E. F.; Svendsen, H. F. Prediction of the pKa Values of Amines Using Ab Initio Methods and Free-Energy Perturbations. Ind. Eng. Chem. Res. 2003, 42, 4414−4421. (15) Bryantsev, V. S.; Diallo, M. S.; Goddard, W. A., III pKa Calculations of Aliphatic Amines, Diamines, and Aminoamides via Density Functional Theory with a Poisson-Boltzmann Continuum Solvent Model. J. Phys. Chem. A 2007, 111, 4422−4430. (16) Khalili, F.; Henni, A.; East, A. L. L. Entropy Contributions in pKa Computation: Application to Alkanolamines and Piperazines. J. Mol. Struct.: THEOCHEM 2009, 916, 1−9. (17) Yamada, H.; Shimizu, S.; Okabe, H.; Matsuzaki, Y.; Chowdhury, F. A.; Fujioka, Y. Prediction of the Basicity of Aqueous Amine Solutions and the Species Distribution in the Amine-H2O-CO2 System Using the COSMO-RS Method. Ind. Eng. Chem. Res. 2010, 49, 2449− 2455. 355

DOI: 10.1021/acs.jctc.7b00855 J. Chem. Theory Comput. 2018, 14, 351−356

Article

Journal of Chemical Theory and Computation

(57) Agmon, N. The Grotthuss Mechanism. Chem. Phys. Lett. 1995, 244, 456−462. (58) Marx, D. Proton Transfer 200 Years after von Grotthuss: Insights from Ab Initio Simulations. ChemPhysChem 2006, 7, 1848− 1870. (59) Stirling, A.; Pápai, I. H2CO3 Forms via HCO3− in Water. J. Phys. Chem. B 2010, 114, 16854−16859.

(39) Gaus, M.; Cui, Q.; Elstner, M. Density Functional Tight Binding: Application to Organic and Biological Molecules. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2014, 4, 49−61. (40) Elstner, M.; Porezag, D.; Jungnickel, G.; Elsner, J.; Haugk, M.; Frauenheim, T.; Suhai, S.; Seifert, G. Self-Consistent-Charge DensityFunctional Tight-Binding Method for Simulations of Complex Materials Properties. Phys. Rev. B: Condens. Matter Mater. Phys. 1998, 58, 7260−7268. (41) Nishizawa, H.; Nishimura, Y.; Kobayashi, M.; Irle, S.; Nakai, H. Three Pillars for Achieving Quantum Mechanical Molecular Dynamics Simulations of Huge Systems: Divide-and-Conquer, Density-Functional Tight-Binding, and Massively Parallel Computation. J. Comput. Chem. 2016, 37, 1983−1992. (42) Yang, W.; Lee, T.-S. A Density-Matrix Divide-and-Conquer Approach for Electronic Structure Calculations of Large Molecules. J. Chem. Phys. 1995, 103, 5674−5678. (43) Akama, T.; Kobayashi, M.; Nakai, H. Implementation of Divideand-Conquer Method Including Hartree-Fock Exchange Interaction. J. Comput. Chem. 2007, 28, 2003−2012. (44) Kobayashi, M.; Nakai, H. In Linear-Scaling Techniques in Computational Chemistry and Physics; Springer: Dordrecht, the Netherlands, 2011; pp 97−127. (45) Kobayashi, M.; Nakai, H. How Does It Become Possible to Treat Delocalized and/or Open-Shell Systems in Fragmentation-Based Linear-Scaling Electronic Structure Calculations? The Case of the Divide-and-Conquer Method. Phys. Chem. Chem. Phys. 2012, 14, 7629−7639. (46) Nakai, H.; Nishimura, Y.; Kaiho, T.; Kubota, T.; Sato, H. Contrasting Mechanisms for CO2 Absorption and Regeneration Processes in Aqueous Amine Solutions: Insights from DensityFunctional Tight-Binding Molecular Dynamics Simulations. Chem. Phys. Lett. 2016, 647, 127−131. (47) Nakai, H.; Sakti, A. W.; Nishimura, Y. Divide-and-ConquerType Density-Functional Tight-Binding Molecular Dynamics Simulations of Proton Diffusion in a Bulk Water System. J. Phys. Chem. B 2016, 120, 217−221. (48) Sakti, A. W.; Nishimura, Y.; Nakai, H. Divide-and-ConquerType Density-Functional Tight-Binding Simulations of Hydroxide Ion Diffusion in Bulk Water. J. Phys. Chem. B 2017, 121, 1362−1371. (49) Selli, D.; Baburin, I. A.; Martoňaḱ , R.; Leoni, S. Novel Metastable Metallic and Semiconducting Germaniums. Sci. Rep. 2013, 3, 1466. (50) Knaup, J. M.; Wehlau, M.; Frauenheim, T. PermutationInvariant Collective Variable to Track and Drive Vacancy Dynamics in Simulations of Solids. Phys. Rev. B: Condens. Matter Mater. Phys. 2013, 88, 220101. (51) Cuny, J.; Korchagina, K.; Menakbi, C.; Mineva, T. Metadynamics Combined with Auxiliary Density Functional and Density Functional Tight-Binding Methods: Alanine Dipeptide as a Case Study. J. Mol. Model. 2017, 23, 72. (52) Vanommeslaeghe, K.; Hatcher, E.; Acharya, C.; Kundu, S.; Zhong, S.; Shim, J.; Darian, E.; Guvench, O.; Lopes, P.; Vorobyov, I.; Mackerell, A. D., Jr. CHARMM General Force Field: A Force Field for Drug-Like Molecules Compatible with the CHARMM All-Atom Additive Biological Force Fields. J. Comput. Chem. 2009, 31, 671−690. (53) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.; Villa, E.; Chipot, C.; Skeel, R. D.; Kalé, L.; Schulten, K. Scalable Molecular Dynamics with NAMD. J. Comput. Chem. 2005, 26, 1781− 1802. (54) Gaus, M.; Cui, Q.; Elstner, M. DFTB3: Extension of the SelfConsistent-Charge Density-Functional Tight-Binding Method (SCCDFTB). J. Chem. Theory Comput. 2011, 7, 931−948. (55) Gaus, M.; Goez, A.; Elstner, M. Parametrization and Benchmark of DFTB3 for Organic Molecules. J. Chem. Theory Comput. 2013, 9, 338−354. (56) Iannuzzi, M.; Laio, A.; Parrinello, M. Efficient Exploration of Reactive Potential Energy Surfaces Using Car-Parrinello Molecular Dynamics. Phys. Rev. Lett. 2003, 90, 238302. 356

DOI: 10.1021/acs.jctc.7b00855 J. Chem. Theory Comput. 2018, 14, 351−356