Transient Phase of Ice Observed by Sum Frequency Generation at the

Feb 2, 2017 - water/mineral interfaces during freezing, which enhanced the intensity of the ... stable form of ice I under atmospheric conditions is h...
0 downloads 0 Views 1MB Size
Download PDF
Letter pubs.acs.org/JPCL

Transient Phase of Ice Observed by Sum Frequency Generation at the Water/Mineral Interface During Freezing Kaitlin A. Lovering, Allan K. Bertram,* and Keng C. Chou* Department of Chemistry, University of British Columbia, Vancouver, British Columbia V6T 1Z1, Canada S Supporting Information *

ABSTRACT: We observed a transient noncentrosymmetric phase of ice at water/mineral interfaces during freezing, which enhanced the intensity of the IR-visible sum frequency generation intensity by up to 20-fold. The lifetime of the transient phase was several minutes. Since the most stable form of ice, hexagonal and cubic ice, are centrosymmetric, our study suggests the transient existence of stacking-disordered ice during the freezing process at water/ mineral interfaces. Stacking-disordered ice, which has only been observed in bulk ice at temperatures lower than −20 °C, is a random mixture of layers of hexagonal ice and cubic ice. However, the transient phase at the ice/mineral interface was observed at temperatures as high as −1 °C. It suggests that the mineral surface may play a role in promoting and stabilizing the formation of stacking-disordered ice at the interface.

T

the water/solid interface. Here we specifically investigate whether Isd ice forms when ice nucleates at the water/solid interface. Measurements at a buried surface are difficult because the surface is the minority component of most systems. Traditional methods used to identify Isd, such as diffraction or calorimetry fail, for these buried interfaces. Raman spectroscopy has also been used recently to study Isd in the bulk,11 but this technique does not have the sensitivity to monitoring surface structures. One of the most obvious differences between Ih, Ic, and Isd ice phases is their space groups. Both Ih and Ic ice phases belong to centrosymmetric space groups, P63/mmc and Fd3̅m, respectively, while Isd ice is noncentrosymmetric.5 This symmetry difference can be harnessed by nonlinear optical techniques that probe the second-order susceptibility. Under the electric-dipole approximation, this tensor is nonzero only when inversion symmetry is broken.15 For some materials this requirement means only a few surface layers, where the symmetry is necessarily broken, are visible. Other materials, however, inherently lack centrosymmetry; these materials have larger second-order responses from the bulk.16 In this study, IR-visible sum frequency generation (SFG) vibrational spectroscopy is used to monitor the phase transition of water from liquid to solid on silica surfaces. We found that when water underwent this phase change, there was a 5 to 20fold increase in the SFG intensity that could best be explained by the transient formation of Isd ice. Cooling Cell. In order to ensure freezing occurs from the surface of the silica, a copper block that fits over the prism was

he liquid-to-solid phase transition of water is a quotidian phenomenon that impacts atmospheric, geological, and biological processes.1−3 Ice has a complex phase diagram,4 and under standard atmospheric conditions ice I forms. The most stable form of ice I under atmospheric conditions is hexagonal ice, ice Ih, which has hydrogen-bonded hexagonal ring structures in an ABAB stacking arrangement. A closely related but less stable phase of ice I is called cubic ice, ice Ic. Ice Ic also has six-membered hydrogen-bonded rings, but the rings are stacked in an ABCABC arrangement.5 Although ice Ic is metastable with respect to ice Ih, X-ray diffraction,1,5−7 electron microscopy,8 and neutron diffraction5,8−10 measurements have shown that these two phases can coexist under certain conditions, depending on the freezing pathway as well as the ambient temperature and pressure. Under no conditions, however, has Ic ice been found to exist as a pure phases; Ih ice is always mixed in with the Ic ice.11 Recently, these mixtures of Ic and Ih ice have become known as stacking-disordered ice, Isd ice.5 It was suggested that the so-called cubic ice actually refers to an ensemble of stacking-disordered ice.5,12 Isd ice has been seen in both theoretical and experimental studies of bulk phase transitions.7,9,12−14 Experimental studies have also shown that colder temperatures and faster cooling rates lead to more stacking faults and lock in the faults as Isd ice.6,13 Experimentally, bulk Isd ice has only been observed at temperatures lower than approximately −20 °C since annealing to Ih ice takes place at warmer temperatures.8 A recent study using large-scale molecular dynamics simulations indicates that Isd ice is formed even at warmer temperatures, but the relatively high temperature allows annealing before the presence of Isd ice can be observed.12 Although the formation of Isd ice has been investigated in the bulk, the formation of this phase has not been investigated at © 2017 American Chemical Society

Received: December 13, 2016 Accepted: February 2, 2017 Published: February 2, 2017 871

DOI: 10.1021/acs.jpclett.6b02920 J. Phys. Chem. Lett. 2017, 8, 871−875

Letter

The Journal of Physical Chemistry Letters constructed. The block was cooled by a circulation chiller that circulates cold octamethyltrisiloxane through the block. Holes were drilled into the block to allow the ingress and egress of laser beams (Figure 1). Dried N2 gas was purged through the

Figure 2. Spectrum of liquid water (red) and hexagonal ice (blue) taken at the interface with silica. The water spectrum was taken at 22 °C, and the ice spectrum was taken at −6 °C roughly 30 min after freezing. The polarizations of the beams were s-, s-, and p-polarized for SFG, visible, and IR, respectively. The intensity of the spectra has been normalized by the Fresnel factor.

Figure 1. Experimental apparatus. A copper cooling block was used to cool the silica prism. The prism is attached to a Teflon flow cell, and water is introduced to the flow cell by opening the inlet valve. The laser beams are aligned to overlap at the base of the prism. The temperature of the prism is recorded by a thermocouple inserted into a hole near the base of the prism.

higher intensity of ice may be due to dipole−dipole coupling,22 charge transfer to the H-bond partner,23 or coordinated motion.24 Figure 3 shows the SFG spectra of water/silica interface recorded during a slow cooling process (∼0.001 °C/s). The

holes to prevent water condensation on the prism surface. The temperature was measured by a thermocouple (Omega, RTD PT100) placed within a hole drilled one millimeter from the surface of the prism. In order to assess the accuracy of the temperature reading, a second thermocouple was placed inside the cell, at the base of the prism. At freezing, the difference in temperature recorded at each thermocouple was around −2 °C, with the thermocouple inside the prism recording the colder temperature. The temperatures reported in this work are from the thermocouple inside the prism because the presence of a thermocouple inside the cell during freezing could affect the water structure during the freezing process by providing an additional site for nucleation. The refractive index difference between water and ice is negligible.17 Freezing Experiments. Two different types of freezing experiments were carried out. Each freezing experiment began at room temperature, around 22 °C. In the first type, the IR beam was scanned from 2800 to 3800 cm−1 during the freezing process with a step size of 20 cm−1 and each step was averaged for 1 s (10 laser shots), and consequently a complete spectrum was collected in 50 s. The cooling rate of this experiment was ∼0.001 °C/s. In the second type, rather than scanning the IR frequency across the range of hydroxyl stretching modes, the IR laser was parked at a fixed wavelength, and the SFG signal was collected as the temperature changed. Each data point is an average over 5 s (50 laser shots). The cooling rate typically started at ∼0.015 °C/s and decreased to ∼0.008 °C/s after freezing. Shown in Figure 2 are SFG spectra of liquid water and ice at the silica surface. The ice spectrum, which was recorded after ∼30 min when the temperature had stabilized, was similar to those at the surface of hexagonal ice (0001).18 This agreement, and the fact that Ih ice is the stable form of ice, leads to the conclusion that hexagonal ice is observed as the stable phase at the silica surface. The liquid water phase has two main peaks, one near 3200 cm−1 and one near 3400 cm−1.19,20 A recent study by Myalitsin et al. showed that water molecules at the water/silica interface experience a broad continuum of stretching modes due to vibrational coupling.21 The SFG spectrum of hexagonal ice at silica has a higher intensity than the liquid spectrum (Figure 2). It has been proposed that the

Figure 3. SFG spectra taken at the water/silica interface during the freezing process. The average temperatures are listed in the legend. The temperature at the beginning and end of each spectral collection are (a) 16.4 to 13.7, (b) 5.2 to 3.7, (c) −0.8 to −1.2, (d) 0.3 to −1.2, (e) −2.0 to −5.2, (f) −5.0 to −5.5, (g) −5.7 to −8.4 °C.

water structure at the silica surface is stable until the freezing process begins at ∼ −4 °C. Figure 3e,f were taken during a transient transition. Therefore, they do not reflect the actual spectral profile. However, it is clear that water undergoes a transient state that has a higher SFG intensity than those of both liquid water and hexagonal ice. This transient state diminishes in intensity at −6 °C (Figure 3f) until the stable hexagonal ice spectrum is observed at −7 °C (Figure 3g). To better capture the dynamics of the transient state, the SFG intensity was recorded during freezing at a fixed wavenumber. The onset of freezing is apparent by a sudden increase in SFG intensity, as shown in Figure 4. The SFG intensity has a more than 10-fold transient increase before stabilization at a lower temperature. Since the SFG intensity is proportional to the square of the second-order susceptibility, this increase in the SFG intensity corresponds to a 3−4 times increase in the second-order susceptibility. The transient state 872

DOI: 10.1021/acs.jpclett.6b02920 J. Phys. Chem. Lett. 2017, 8, 871−875

Letter

The Journal of Physical Chemistry Letters

understanding the origin of the signal difficult.28 A transient increase in SHG signal was not observed after freezing,29 and, given that SHG is less sensitive to the ordering of interfacial water, this is not surprising. Moreover, the authors note that nucleation began at the joint of their cooling cell and not in the area probed by SHG. In the current study, the temperature at which the stackingdisordered ice was observed is significantly higher than previously reported temperatures observed in bulk water. Bulk Isd ice has only been observed experimentally at temperatures lower than −20 °C since annealing to Ih ice takes place at warmer temperatures.8 Our results suggest that the mineral surface may play a role in stabilizing the formation of the stacking-disordered ice. Because ice nucleation is a stochastic process, even two identical systems may freeze at different times and temperatures. Figure 5 shows 35 freezing processes with the freezing

Figure 4. Typical data sets collected during cooling experiments. Results are shown for silica taken with the IR laser parked at 3100 cm−1.

exists for a period of ∼3 min when the temperature of the system (blue curve in Figure 4) is relatively constant between 1000 and 1200 s, which is an indication of a phase transition. Similar SFG increases were observed when the IR laser was parked at 3200, 3100, 3000, and 2900 cm−1 (see the Supporting Information), reinforcing the presence of a broad and intense resonance during the transition. This transient SFG enhancement was also observed in the anisotropic tensor elements of the second-order susceptibility monitored in the psp and spp polarization combinations (data shown in the Supporting Information), indicating that the growth of the transient phase is anisotropic. Data present in Figure 4 suggests that a noncentrosymmetric structure of water molecules is formed in the ice-growth process at the water/mineral interfaces, since both liquid water and Ih ice are centrosymmetric, in which the SFG from the bulk media is forbidden under the electric dipole approximation, and the SFG intensity is dominated by the SFG generated at the water (or Ih ice) interfaces.18 The current study supports the transient formation of stacking-disordered ice at the mineral surface in the freezing process. Since both Ih ice and Ic ice are centrosymmetric, the large increase in SFG intensity can be explained by the transition from seeing only the interface of the water/silica to seeing the bulk Isd ice and then to seeing only the interface of Ih ice/silica. There are a few other studies that monitored ice formation at mineral surfaces using nonlinear spectroscopy. Amin-Danso et al. studied the surface charge effect on the structure of interfacial ice using SFG.25 They found that at a negatively charged surface, the intensity of the SFG signal decreased with freezing, while at a positively charged surface the signal increased.25 Zhang et al. performed SFG experiments at the sapphire and mica surface.26 They found that salt hydrates have differing propensities to segregate at the surface, depending on the identity of the cation.26 A transient, although shorter lived, increase in signal was also observed by Amin-Danso et al. at the sapphire surface at pH 9.8, but in their study the transient change in SFG signal was explained by the presence of Na+ ions, which affect the bilayer stitching of the ice.27 Contrary to our work, their experimental setup did not initiate freezing at the mineral surface, but rather from the bulk water, and their experiments included ions, making interpretation of the transient change more difficult. Abdelmonem et al. performed a second harmonic generation (SHG) study monitoring freezing at silica and mica. The SHG is a nonresonant process, and the inclusion of nonresonant contributions makes

Figure 5. Onset freezing temperatures of the water are plotted in bins of one degree. The median freezing temperature is −3.3 °C, while the average is −4.5 °C.

point in the range of −12 and −1.2 °C; the majority froze between −3 and −1 °C. All 35 freezing process show the existence of stacking-disordered ice at the water/mineral interface is ∼10−20 °C higher than those in the bulk media. In addition to the positive evidence presented supporting the formation of Isd ice, other possible explanations for the large SFG intensity increase are unsatisfactory. For example, although there is evidence of a liquid−liquid phase transition prior to freezing, and that the prefreeze liquid has a higher order parameter, the liquid−liquid transition occurs at super cooled conditions that are not reached in this study.30 Is it possible that another phase of ice, not Isd ice, is forming before Ih ice? Russo et al. defined a new phase of ice called Ice 0, which also lacks inversion symmetry.31 This phase, however, exists on the femtosecond time scale and would be too shortlived to account for the relatively long lifetime of the transient phase seen in this study.31 Moreover, their simulations were carried out for homogeneous nucleation when no other structure-inducing substrate was present. There are also phases of ice in which the protons are locked in place and therefore are more ordered;4 some of these phases also lack inversion symmetry. However, these proton-ordered phases occur typically at high pressure and are not expected to form under the present experimental conditions.4 Ferroelectric ice is a proton-ordered phase of ice that also lacks inversion symmetry. The formation of ferroelectric ice has recently been observed on Pt(111) surfaces at temperatures less than −98°C 873

DOI: 10.1021/acs.jpclett.6b02920 J. Phys. Chem. Lett. 2017, 8, 871−875

The Journal of Physical Chemistry Letters

■

and when formed by vapor deposition.32 Ferroelectric ice has also been made in one dimension via growth within nanochannels.33,34 Although water has a large molecular electric dipole moment, formation of hydrogen bonds suppresses the ferroelectric ordering of ice. Therefore, ferroelectric ice has been observed only when water is confined, and the temperature is significantly below 0 °C. It has been suggested that a hydroxylated silica surface could provide a template for growth of ferroelectric ice, similarly to the Pt(111) surface.35 However, it requires depositional growth at much lower temperatures.32,36 Proton-ordered ice, including ferroelectric ice, is not expected to form at the warm temperatures and high dimensional system of the current study. Of the other studies where the freezing process of water is monitored by SFG spectroscopy, which is done on a sapphire surface, one also sees a transient increase in SFG signal after freezing but only at pH 9.8. The transient increase is attributed to the presence of Na+ ions in the solution that disrupts the charge transfer and the stitching bilayer.27 This explanation is insufficient for our study because no ions are present in the current system. In summary, we have observed the formation and destruction of a transient noncentrosymmetric phase of ice at water/ mineral interfaces during freezing. While the observed data can be explained by the formation of stacking-disordered ice, the temperature observed at the interface is ∼20 °C higher than those observed in bulk ice. Our results suggest that the mineral surface may play a role in promoting and stabilizing the formation of the stacking-disordered ice at the interface.

Letter

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpclett.6b02920. Additional spectra of the freezing process at different wave numbers, and additional cooling curves at PSP and SPP polarization configurations (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*Allan K. Bertram: [email protected]. *Keng C. Chou: [email protected]. ORCID

Keng C. Chou: 0000-0002-8782-5253 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was financially supported by the Natural Sciences and Engineering Research Council of Canada.

■

REFERENCES

(1) Murray, B. J.; Knopf, D. A.; Bertram, A. K. The formation of cubic ice under conditions relevant to Earth’s atmosphere. Nature 2005, 434 (7030), 202−205. (2) Zeng, Q.; Li, K. F.; Fen-Chong, T. Heterogeneous nucleation of ice from supercooled NaCl solution confined in porous cement paste. J. Cryst. Growth 2015, 409, 1−9. (3) Maki, L. R.; Galyan, E. L.; Chang-Chien, M.-M.; Caldwell, D. R. Ice Nucleation Induced by Pseudomonas-Syringae. Appl. Microbiol. 1974, 28 (3), 456−459. (4) Salzmann, C. G.; Radaelli, P. G.; Slater, B.; Finney, J. L. The polymorphism of ice: five unresolved questions. Phys. Chem. Chem. Phys. 2011, 13 (41), 18468−18480. (5) Malkin, T. L.; Murray, B. J.; Salzmann, C. G.; Molinero, V.; Pickering, S. J.; Whale, T. F. Stacking disorder in ice I. Phys. Chem. Chem. Phys. 2015, 17 (1), 60−76. (6) Murray, B. J.; Bertram, A. K. Formation and stability of cubic ice in water droplets. Phys. Chem. Chem. Phys. 2006, 8 (1), 186−192. (7) Mayer, E.; Hallbrucker, A. Cubic Ice From Liquid Water. Nature 1987, 325 (6105), 601−602. (8) Kuhs, W. F.; Sippel, C.; Falenty, A.; Hansen, T. C. Extent and relevance of stacking disorder in ″ice I-c″. Proc. Natl. Acad. Sci. U. S. A. 2012, 109 (52), 21259−21264. (9) Hansen, T. C.; Koza, M. M.; Lindner, P.; Kuhs, W. F. Formation and annealing of cubic ice: II. Kinetic study. J. Phys.: Condens. Matter 2008, 20 (28), 285105. (10) Hansen, T. C.; Koza, M. M.; Kuhs, W. F. Formation and annealing of cubic ice: I. Modelling of stacking faults. J. Phys.: Condens. Matter 2008, 20 (28), 285104. (11) Carr, T. H. G.; Shephard, J. J.; Salzmann, C. G. Spectroscopic Signature of Stacking Disorder in Ice I. J. Phys. Chem. Lett. 2014, 5 (14), 2469−2473. (12) Hudait, A.; Qiu, S.; Lupi, L.; Molinero, V. Free energy contributions and structural characterization of stacking disordered ices. Phys. Chem. Chem. Phys. 2016, 18 (14), 9544−9553. (13) Reinhardt, A.; Doye, J. P. K. Free energy landscapes for homogeneous nucleation of ice for a monatomic water model. J. Chem. Phys. 2012, 136 (5), 054501. (14) Moore, E. B.; Molinero, V. Is it cubic? Ice crystallization from deeply supercooled water. Phys. Chem. Chem. Phys. 2011, 13 (44), 20008−20016. (15) Shen, Y. R. Surface-Properties Probed by 2nd-Harmonic and Sum-Frequency Generation. Nature 1989, 337 (6207), 519−525.

■

EXPERIMENTAL METHODS The surface of a fused silica equilateral IR-grade fused silica prism was washed with Extran AP12 cleaning agents, rinsed with Millipore water (resistivity >18.2 MΩ·cm), and left to soak in concentrated sulfuric acid containing NOCHROMIX (Godax Laboratories, Inc. USA) for several hours. The prism was subsequently rinsed with Millipore water several times, soaking in the Millipore water for at least 30 min between each rinse. To prevent exposure to air during the course of the experiment, the prism was attached to a flow cell that allowed the addition of liquid without removing the prism. The visible and IR laser beams for SFG vibrational spectroscopy were obtained from a Nd:YAG (yttrium aluminum garnet) laser with output wavelength of 1064 nm (30 ps, 40 mJ/pulse, and 10 Hz). The laser was used to generate a second harmonic beam at 532 nm in a KTiOPO4 (KTP) crystal. The tunable IR beam was produced by difference frequency mixing of the 1064 nm beam with the output of a KTP optical parametric generator/amplifier pumped by the 532 nm beam. The 532 nm and IR beams were overlapped, both spatially and temporally, on the sample. The energy of the laser beams were both ∼200 μJ/pulse for the visible and IR beams. The polarizations of the beams were s-, s-, and p-polarized for SFG, visible, and IR, respectively. The polarization combinations psp and spp were also used to monitor the anisotropic tensor element of the second-order susceptibility. (Data shown in the Supporting Information.)37 The incident angles of the IR and the visible beam are 50 and 55 degrees, respectively. The SFG intensity was detected by a photomultiplier tube after spatial filtering by an aperture and spectral filtering by a bandpass filter. In typical SFG spectra of water, the OH vibrational modes are probed by scanning between 2800 and 3800 cm−1. 874

DOI: 10.1021/acs.jpclett.6b02920 J. Phys. Chem. Lett. 2017, 8, 871−875

Letter

The Journal of Physical Chemistry Letters (16) Shoji, I.; Kondo, T.; Kitamoto, A.; Shirane, M.; Ito, R. Absolute scale of second-order nonlinear-optical coefficients. J. Opt. Soc. Am. B 1997, 14 (9), 2268−2294. (17) Schiebener, P.; Straub, J.; Sengers, J.; Gallagher, J. S. RefractiveIndex of Water and Steam as Function of Wavelength, Temperature and Density. J. Phys. Chem. Ref. Data 1990, 19 (3), 677−717. (18) Wei, X.; Miranda, P. B.; Zhang, C.; Shen, Y. R. Sum-frequency spectroscopic studies of ice interfaces. Phys. Rev. B: Condens. Matter Mater. Phys. 2002, 66 (8), 085401. (19) Ostroverkhov, V.; Waychunas, G. A.; Shen, Y. R. New information on water interfacial structure revealed by phase-sensitive surface spectroscopy. Phys. Rev. Lett. 2005, 94 (4), 046102. (20) Du, Q.; Freysz, E.; Shen, Y. R. Vibrational-Spectra of WaterMolecules at Quartz Water Interfaces. Phys. Rev. Lett. 1994, 72 (2), 238−241. (21) Myalitsin, A.; Urashirma, S. H.; Nihonyanagi, S.; Yamaguchi, S.; Tahara, T. Water Structure at the Buried Silica/Aqueous Interface Studied by Heterodyne-Detected Vibrational Sum-Frequency Generation. J. Phys. Chem. C 2016, 120 (17), 9357−9363. (22) Shultz, M. J.; Bisson, P.; Vu, T. H. Insights into hydrogen bonding via ice interfaces and isolated water. J. Chem. Phys. 2014, 141 (18), 18C521. (23) Ishiyama, T.; Takahashi, H.; Morita, A. Origin of Vibrational Spectroscopic Response at Ice Surface. J. Phys. Chem. Lett. 2012, 3 (20), 3001−3006. (24) Groenzin, H.; Li, I.; Buch, V.; Shultz, M. J. The single-crystal, basal face of ice I(h) investigated with sum frequency generation. J. Chem. Phys. 2007, 127 (21), 214502. (25) Anim-Danso, E.; Zhang, Y.; Dhinojwala, A. Surface Charge Affects the Structure of Interfacial Ice. J. Phys. Chem. C 2016, 120 (7), 3741−3748. (26) Zhang, Y.; Anim-Danso, E.; Dhinojwala, A. The Effect of a Solid Surface on the Segregation and Melting of Salt Hydrates. J. Am. Chem. Soc. 2014, 136 (42), 14811−14820. (27) Anim-Danso, E.; Zhang, Y.; Alizadeh, A.; Dhinojwala, A. Freezing of Water Next to Solid Surfaces Probed by Infrared-Visible Sum Frequency Generation Spectroscopy. J. Am. Chem. Soc. 2013, 135 (7), 2734−2740. (28) Han, Y.; Raghunathan, V.; Feng, R. R.; Maekawa, H.; Chung, C. Y.; Feng, Y.; Potma, E. O.; Ge, N. H. Mapping Molecular Orientation with Phase Sensitive Vibrationally Resonant Sum-Frequency Generation Microscopy. J. Phys. Chem. B 2013, 117 (20), 6149−6156. (29) Abdelmonem, A.; Luetzenkirchen, J.; Leisner, T. Probing icenucleation processes on the molecular level using second harmonic generation spectroscopy. Atmos. Meas. Tech. 2015, 8 (8), 3519−3526. (30) Palmer, J. C.; Martelli, F.; Liu, Y.; Car, R.; Panagiotopoulos, A. Z.; Debenedetti, P. G. Metastable liquid-liquid transition in a molecular model of water. Nature 2014, 510 (7505), 385−388. (31) Russo, J.; Romano, F.; Tanaka, H. New metastable form of ice and its role in the homogeneous crystallization of water. Nat. Mater. 2014, 13 (7), 733−739. (32) Sugimoto, T.; Aiga, N.; Otsuki, Y.; Watanabe, K.; Matsumoto, Y. Emergent high-T-c ferroelectric ordering of strongly correlated and frustrated protons in a heteroepitaxial ice film. Nat. Phys. 2016, 12 (11), 1063−1069. (33) Zhao, H. X.; Kong, X. J.; Li, H.; Jin, Y. C.; Long, L. S.; Zeng, X. C.; Huang, R. B.; Zheng, L. S. Transition from one-dimensional water to ferroelectric ice within a supramolecular architecture. Proc. Natl. Acad. Sci. U. S. A. 2011, 108 (9), 3481−3486. (34) Gorshunov, P.; Torgashev, V. I.; Zhukova, E. S.; Thomas, V. G.; Belyanchikov, M. A.; Kadlec, C.; Savinov, M.; Ostapchuk, T.; Petzelt, J.; Prokleska, J.; Tomas, P. V.; Pestrjakov, E. V.; Fursenko, D. A.; Shakurov, G. S.; Prokhorov, A. S.; Gorelik, V. S.; Kadyrov, L. S.; Uskov, V. V.; Kremer, R. K.; Dressel, M. Incipient ferroelectricity of water molecules confined to nano-channels of beryl. Nat. Commun. 2016, 7, 12842. (35) Parkkinen, P.; Riikonen, S.; Halonen, L. Ice XI: Not That Ferroelectric. J. Phys. Chem. C 2014, 118 (45), 26264−26275.

(36) Iedema, M. J.; Dresser, M. J.; Doering, D. L.; Rowland, J. B.; Hess, W. P.; Tsekouras, A. A.; Cowin, J. P. Ferroelectricity in water ice. J. Phys. Chem. B 1998, 102 (46), 9203−9214. (37) Guyotsionnest, P.; Chen, W.; Shen, Y. R. General Considerations on Optical 2ND-Harmonic Generation from Surfaces and Interfaces. Phys. Rev. B: Condens. Matter Mater. Phys. 1986, 33 (12), 8254−8263.

875

DOI: 10.1021/acs.jpclett.6b02920 J. Phys. Chem. Lett. 2017, 8, 871−875