Guest-induced symmetry lowering of an ionic clathrate material for carbon capture

Abstract

We report a new lattice structure of the ionic clathrate hydrate of tetra-n-butylammonium bromide induced by guest CO₂ molecules, which is found to provide high CO₂ storage capacity. The structure was characterized by a set of methods, including single crystal X-ray diffraction, NMR, and MD simulations.

---

Gas capture by guest–host compounds is a technology which has a possibility to mitigate a specific gas from a gas mixture with low energy consumption and not accompanied by an irreversible chemical reaction.¹ Ionic clathrate hydrates are a class of guest–host materials in which large ionic guest species, primarily substituted ammonium or phosphonium salts, are incorporated into a solid ice-like hydrogen-bonded water framework.² They are generally formed from aqueous solutions, and have higher melting temperatures than ice, which makes these materials suitable for applications such as cool-energy storage and gas separation/storage.^3–6 The interest in ionic clathrate hydrates lies in the fact that the small dodecahedral (D) water cages formed in the lattice are available for occupancy by small gas-phase guest species, e.g., CO₂, CH₄ and H₂, generally under milder conditions as compared to the pure small-guest canonical clathrate hydrates. The tetra-n-butylammonium bromide (TBAB) + CO₂ hydrate has been a very popular system for investigation and there have been a number of reports on the modeling of the TBAB + CO₂ hydrate.⁷ However, the work reported here reinforces the need to have good structural information as the currently available models for the ionic hydrates depend on a number of unfounded assumptions. For example, at a CO₂ gas pressure of 4 MPa, the TBAB ionic hydrate, encapsulating CO₂ guests, is formed at 291 K which is ∼10 K higher than the temperature needed to synthesize the pure CO₂ clathrate hydrate at the same pressure.⁴ In the canonical clathrate hydrates, several guest-induced modifications of the host water framework have been reported.⁸ Such host-framework rearrangements in the presence of a gaseous guest allow the most efficient incorporation of the guest in the resulting structure. The occupancy of CO₂ as a guest substance in hydrate cages is affected by guest–host and/or guest–guest interactions,^9,10 however, a lattice structural transition induced by CO₂ has not been reported.

Herein, we report a new CO₂-driven ionic clathrate-hydrate structure transformation which leads to efficient CO₂ encapsulation. To see how the guest gas molecules stabilize/destabilize the ionic clathrate hydrate, we performed single crystal X-ray diffraction (SCXRD) measurements at 100 K on the TBAB + CO₂ ionic clathrate hydrate. The crystals were synthesized from an aqueous TBAB solution with mole fraction x_TBAB = 0.0064 under CO₂ pressure P = 1.08 MPa at temperature T = 282.65 K. The experimental details and a figure of the crystal structure are given in the ESI.† The crystallographic information is summarized in the footnote.‡ The TBA cation is encapsulated in super cages derived from four of the large canonical clathrate cages, and Br⁻ replaces water molecules in the lattice.² Unexpectedly, in the presence of CO₂, only the orthorhombic Imma crystal structure was found, whereas the TBAB hydrate formed without a guest gas has the tetragonal (TBAB·26H₂O) and orthorhombic (TBAB·38H₂O) phases as the most stable phases, with yet other phases also observed. The determined space group Imma has the same point group but a different lattice from Pmma observed for TBAB·38H₂O.¹¹ The symmetry of the Imma structure is lower because of the doubling of the unit cell size in the direction of the b-axis compared to the pure Pmma hydrate. This lattice change evidently arises from the asymmetrical distributions of the incorporated CO₂ molecules in the dodecahedral (D) cages of the Imma hydrate and is not related to structural changes in the TBAB and water framework upon CO₂ encapsulation. The CO₂ molecules occupy the two symmetry-distinct dodecahedral-cages as shown in Fig. 1 (and Fig. S1 of the ESI†), and include the highly distorted cages D_A and more regular dodecahedral cages D_B. The chemical formula of the empty host framework can be written as TBAB·38H₂O·D_A·2D_B.

Fig. 1 The highly distorted D_A and regular shaped D_B dodecahedral cages in the TBAB + CO₂ structure. The arrow denotes the direction of the unit cell axes.

In the crystal structure, the fraction of D_A cages occupied by CO₂ is almost twice that of the D_B cages, i.e., 0.867 and 0.490, respectively. As shown in Fig. 1, the water molecules forming the D_A cage are displaced by the n-butyl chains of the [TBA]⁺ ion and pushed inward by the Br⁻ ions. As a result, the lattice is highly distorted along a direction perpendicular to the plane of the confined CO₂ molecules. This clearly differs from the structure of D cages holding CO₂ in canonical clathrate hydrate phases. The D_A and D_B cages have similar volumes, i.e., 153 and 160 ų, respectively. As reported in Table S2 of the ESI,† three crystal samples of the TBAB + CO₂ hydrate formed under the different P-T-x conditions had the same orthorhombic Imma structure. The preference of an orthorhombic hydrate structure over the tetragonal structure in the presence of CO₂ (or small-guest) can be rationalized by considering the unit cell volume per 5¹² cages, i.e., 76H₂O/6 D cages for the orthorhombic phase and 164H₂O/10 D cages for the tetragonal TBAB phase.^2,6 Thus application of Le Chatelier's principle suggests that the orthorhombic structure gives the largest pressure drop and the greatest stabilization effect as a result of CO₂ incorporation into the D cages.

To study the CO₂ guest environments and occupancies in the D_A and D_B cages, CP-MAS ¹³C NMR spectra of the TBAB + CO₂ solid phase were obtained, as shown in Fig. 2. This sample was made by exposing the solid TBAB hydrate to CO₂ gas under pressure; see the ESI† for details of the sample preparation. The spectra demonstrate a complex overlap of the isotropic signals and spinning sidebands near 125 ppm of CO₂ guests in three distinct sites and TBAB in the lattice of the ionic clathrate. The ratio of the integrated intensities for the hydrate cage signals with isotropic chemical shifts at 124.87 and 125.13 ppm is less than unity, i.e., 0.54 (30/55), which is consistent with the relationship between the total number of CO₂ guests in the D_A and D_B cages determined by the SCXRD, N(D_A)/N(D_B) = 0.867/(2 × 0.490) = 0.88. Both the signal intensities in the NMR spectrum and the CO₂ occupancies in the SCXRD structure indicate a smaller CO₂ population in the more abundant D_B cages.

Fig. 2 (A) Experimental ¹³C CP MAS spectrum of the ¹³CO₂·TBAB·38H₂O hydrate obtained under a field of 9.4 T (Larmor frequency of 100.67 MHz) at 173 K and a spinning speed of 2975 Hz. (B) Simulation of the experimental spectrum with a model accounting for the spectral intensity spread in the spinning sidebands.

The isotropic chemical shifts for CO₂ residing in D cages in this work are close to the ∼128 ppm chemical shifts of CO₂ guests previously reported in other clathrate hydrate phases.⁷ The difference in chemical shifts from 128 ppm in D cages reported for other hydrates to ∼125 ppm in the presently found D_A and D_B cages can be related to the smaller sizes of the latter cages, or to the presence of the charged Br⁻ and [TBA]⁺ groups in close proximity, and also to the fact that previous chemical shift data were reported for stationary samples thus offering lower accuracy because of the broad powder patterns. The assignment of the low intensity signal at 125.40 ppm is less certain, and could arise from a different ionic clathrate phase formed under the non-equilibrium sample preparation conditions of the powder NMR sample. This is verified by powder X-ray diffraction (PXRD) measurements of the TBAB + CO₂ phases formed under conditions similar to the preparation of the NMR samples. The PXRD patterns given in the ESI† show that the sample contained both orthorhombic and tetragonal phases, and thus the two main signals for CO₂ (natural abundance) appearing at 124.85 and 125.06 ppm of the ¹³C NMR spectrum shown in Fig. S5 (ESI†) should originate from the orthorhombic phase and the additional small peak at 125.30 ppm from a small fraction of the tetragonal phase, respectively.

Guest dynamics were also studied using the MD simulation methodology with details outlined in the ESI.† The CO₂ guests were found to have restricted, but different spatial distributions/motions in D_A and D_B cages. The sampled orientations of the CO₂ guest molecules in the D_A and D_B cages with respect to the lattice b-direction (corresponding to the flattened direction of the D_A cage) are reported in the ESI.†

The dynamics of the CO₂ guest motion were also studied using the orientation autocorrelation function (OACF) around the molecular axis. If ϕ(t) is the rotation angle of the unit vector along the direction of the CO₂ molecular axis at time t, μ(t), with respect to the vector orientation at time 0, μ(0), the OACF defined as the ensemble average of the first moment of the Legendre polynomial for the rotation angle, 〈P₁(t)〉, is presented as,

〈P₁(t)〉 = 〈μ(t)·μ(0)〉 = 〈cos θ(t)〉 (1)

where the brackets 〈〉 indicate and ensemble the average over all CO₂ molecules in the separate D_A or D_B cages. The behaviors of 〈P₁(t)〉 for CO₂ guests in the two cages are given in Fig. 3. The D_A cages are smaller and the orientation correlations for CO₂ guests in these cages decay more slowly at all temperatures.

Fig. 3 The OACF for CO₂ molecules in (a) the D_A cages, (b) the D_B cages. The size of the D_A cages is smaller and the orientation correlations decay more slowly for molecules in these cages.

The simulations show that the lattice vibrations of tetra-n-butylammonium and bromide groups about their equilibrium positions are small and they are kept in place by electrostatic and van der Waals interactions with the neighboring water molecules. This is particularly interesting with regard to the Br⁻ lattice vibrations as this ion may seem to be relatively free to move in the ionic hydrate lattice. The lattice site of injection of the Br⁻ anion can be the site of Bjerrum L-defects¹² in the neighboring water molecule lattice. In the simulation, water molecules show relatively slow rotational dynamics in the lattice. This is consistent with experiments since the ¹³C NMR powder patterns are broad and look distributed at around 200 K, much the same as for the large cage in the sI hydrate at this temperature. The water dynamics in the double hydrate of tetrahydrofuran + CO₂ are much faster at temperatures around 200 K.¹⁰

The presence of CO₂ guests lowered the symmetry of the original TBAB hydrate structure formed without small guest molecules. The CO₂ gas storage capacities in the two types of D cages in the structure differ significantly because of the size and shape differences of the cages. This demonstrates that the shape of the guest molecules is a secondary factor for determining the gas storage capacity of the ionic clathrates in addition to the primary factor of molecular size. From a practical point of view, our experiments show that upon exposure to relatively mild CO₂ pressures and controlled temperature/composition conditions, it is possible to selectively drive the formation of the ionic clathrate hydrate towards the orthorhombic TBAB·38H₂O·nCO₂ phase which is the most effective one for carbon capture. We note that the large difference in gas-holding capacity of the D_A and D_B small cage types in the hydrate is of key importance in developing satisfactory models that will account for a variety.

This study was subsidized in part by the Keio University Global Center of Excellence Program “Center for Education and Research of Symbiotic, Safe and Secure System Design”. SM acknowledges the financial support provided by the Japan Society for the Promotion of Science (JSPS) through the program “Grant-in-Aid for JSPS Fellows of the Ministry of Education, Culture, Sports, Science, and Technology (MEXT)” (Grant No. 23-56572).

Notes and references

1. Comprehensive Supramolecular Chemistry, Supramolecular Technology, ed. D. N. Reinhoudt, Pergamon, Oxford, 1996, vol. 10; D. M. D'Alessandro, B. Smit and J. R. Long, Angew. Chem., Int. Ed., 2010, 49, 6038; S. D. Kenarsari, D. Yang, G. Jiang, S. Zhang, J. Wang, A. G. Russell, Q. Wei and M. Fan, RSC Adv., 2013, 3, 22739.
2. M. Bonamico, G. A. Jeffrey and R. K. McMullan, J. Chem. Phys., 1962, 37, 2219; R. K. McMullan, G. A. Jeffrey and M. Bonamico, J. Chem. Phys., 1963, 39, 3295; D. W. Davidson, in Water. A Comprehensive Treatise, ed. F. Franks, Plenum Press, New York, NY, 1973; G. A. Jeffrey, in Inclusion Compd., ed. J. L. Atwood, J. E. D. Davies and D. D. MacNicol, Academic Press, London, 1984, ch. 5, vol. 1.
3. A. Chapoy, R. Anderson and B. Tohidi, J. Am. Chem. Soc., 2007, 129, 746; N. H. Duc, F. Chauvy and J.-M. Herri, Energy Convers. Manage., 2007, 48, 1313; D. Zhong and P. Englezos, Energy Fuels, 2012, 26, 2098.
4. N. Ye and P. Zhang, J. Chem. Eng. Data, 2012, 57, 1557.
5. H. Oyama, W. Shimada, T. Ebinuma, Y. Kamata, S. Takeya, T. Uchida, J. Nagao and H. Narita, Fluid Phase Equilib., 2005, 234, 131; K. Sato, H. Tokutomi and R. Ohmura, Fluid Phase Equilib., 2013, 337, 115.
6. S. Muromachi, S. Takeya, Y. Yamamoto and R. Ohmura, CrystEngComm, 2014, 16, 2056.
7. X. Liao, X. Guo, Y. Zhao, Y. Wang, Q. Sun, A. Liu, C. Sun and G. Chen, Ind. Eng. Chem. Res., 2013, 52, 18440; W. Lin, D. Dalmazzone, W. Fürst, A. Delahaye, L. Fournaison and P. Clain, J. Chem. Eng. Data, 2013, 58, 2233; A. Joshi, P. Mekala and J. S. Sangwai, J. Nat. Gas Chem., 2012, 21, 459; P. Paricaud, J. Phys. Chem. B, 2011, 115, 288; M. Kwaterski and J.-M. Herri, Fluid Phase Equilib., 2014, 371, 22; P. Babu, W. I. Chin, R. Kumar and P. Linga, Ind. Eng. Chem. Res., 2014, 53, 4878.
8. K. Shin, S. Choi, J.-H. Cha and H. Lee, J. Am. Chem. Soc., 2008, 130, 7180; K. A. Udachin, C. I. Ratcliffe, G. D. Enright and J. A. Ripmeester, Angew. Chem., Int. Ed., 2008, 47, 9704; K. Shin, Y. Kim, T. A. Strobel, P. S. R. Prasad, T. Sugahara, H. Lee, E. D. Sloan, A. K. Sum and C. A. Koh, J. Phys. Chem. A, 2009, 113, 6415; K. Shin, K. A. Udachin, I. L. Moudrakovski, D. M. Leek, S. Alavi, C. I. Ratcliffe and J. A. Ripmeester, Proc. Natl. Acad. Sci. U. S. A., 2013, 110, 8441; K. Shin, R. Kumar, K. A. Udachin, S. Alavi and J. A. Ripmeester, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 14785; K. Shin, I. L. Moudrakovski, M. D. Davari, S. Alavi, C. I. Ratcliffe and J. A. Ripmeester, CrystEngComm, 2014 10.1039/c3ce41661e.
9. J. A. Ripmeester and C. I. Ratcliffe, Energy Fuels, 1998, 12, 197; K. A. Udachin, C. I. Ratcliffe and J. A. Ripmeester, J. Phys. Chem. B, 2001, 105, 4200; S. Alavi, P. Dornan and T. K. Woo, ChemPhysChem, 2008, 9, 911.
10. S. Alavi and J. A. Ripmeester, J. Chem. Phys., 2012, 137, 054712; K. Tezuka, R. Shen, T. Watanabe, S. Takeya, S. Alavi, J. A. Ripmeester and R. Ohmura, Chem. Commun., 2013, 49, 505.
11. L. A. Gaponenko, S. F. Solodovnikov, Y. A. Dyadin, L. S. Aladko and T. M. Polyanskaya, Zh. Strukt. Khim., 1984, 25, 175; W. Shimada, M. Shiro, H. Kondo, S. Takeya, H. Oyama, T. Ebinuma and H. Narita, Acta Crystallogr., Sect. C, 2005, 61, o65.
12. N. Bjerrum, Science, 1952, 115, 385.
13. G. M. Sheldrick, Acta Crystallogr., Sect. A, 1990, 46, 467; G. M. Sheldrick, Acta Crystallogr., Sect. A, 1993, 49(suppl), C53.
14. K. Momma and F. Izumi, J. Appl. Crystallogr., 2011, 44, 1272.

---

Footnotes

† Electronic supplementary information (ESI) available: Experimental and computational methods and supplementary results of single crystal and powder XRD, NMR and MD simulations. CCDC 963037. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c4cc02111h

‡ C_17.85H₁₁₁BrNO_41.69, M_r = 1087.28, crystal dimensions: 0.6 × 0.4 × 0.2, orthorhombic Imma, a = 21.0197(7), b = 25.2728(8), c = 12.0096(4) Å, V = 6379.8(4) ų, Z = 4, T = 100.0(1) K, ρ_calcd = 1.132 g cm⁻³, μ = 0.730, radiation and wavelength: Mo Kα and 0.71070 Å, 2θ_max = 72.76, no. of measured and independent reflections: 87 521 and 8002, R_int = 0.0363, final R(I > 2σ(I)) indices: R₁ = 0.0266, wR₂ = 0.0589. The structure was solved by direct methods using the SHELXTL suite of programs¹³ and visualized by VESTA.¹⁴ CCDC 963037.

---