Controlling Nonclassical Content of Clathrate Hydrates Through the

Dec 20, 2010 - Low-temperature, low-pressure studies of clathrate hydrates (CHs) have revealed that small ether and other proton-acceptor guests great...
1 downloads 14 Views 4MB Size
ARTICLE pubs.acs.org/JPCA

Controlling Nonclassical Content of Clathrate Hydrates Through the Choice of Molecular Guests and Temperature I. Abrrey Monreal and J. Paul Devlin* Department of Chemistry, Oklahoma State University, Stillwater, Oklahoma 74078, United States

Zafer Mas-lakcı, M. Bora C-ic- ek, and Nevin Uras-Aytemiz Department of Chemistry, Suleyman Demirel University, 32260 Isparta, Turkey

bS Supporting Information ABSTRACT: Low-temperature, low-pressure studies of clathrate hydrates (CHs) have revealed that small ether and other protonacceptor guests greatly enhance rates of clathrate hydrate nucleation and growth; rapid formation and transformations are enabled at temperatures as low as 110 K, and cool moist vapors containing small ether molecules convert to mixed-gas CHs on a subsecond time scale. More recently, FTIR spectroscopic studies of the tetrahydrofuran (THF)-HCN double clathrate hydrate revealed a sizable frequency shift accompanied by a four-fold intensification of the C-N stretch-mode absorption of the small cage HCN, behavior that is enhanced by cooling and which correlates precisely with similar significant changes of the ether C-O/C-C stretch modes. These temperature-dependent correlated changes in the infrared spectra have been attributed to equilibrated extensive hydrogen bonding of neighboring large- and small-cage guest molecules with water molecules of the intervening wall. An ether guest functions as a proton acceptor, particularly so when complemented by the action of a proton-donor (HCN)/electron-acceptor (SO2) small-cage guest. Because guest molecules of the classic clathrate hydrates do not participate in hydrogen bonds with the host water, this H-bonding of guests has been labeled “nonclassical”. The present study has been enriched by comparing observed FTIR spectra with high-level molecular orbital computational results for guests and hydrogen-bonded guest-water dimers. Vibrational frequency shifts, from heterodimerization of ethers and water, correlate well with the corresponding observed classical to nonclassical shifts. The new spectroscopic data reveal that the nonclassical structures can contribute at observable levels to CH infrared spectra for a remarkable range of temperatures and choice of guest molecules. By the choice of guest molecules, it is now possible to select the abundance levels of nonclassical configurations, ranging from ∼0 to 100%, for a given temperature. This ability is expected to hasten understanding of the role of guest-induced nonclassical structures in the acceleration or inhibition of the rates of CH formation and transformation.

’ INTRODUCTION Clathrate hydrates (CHs) exist as “crystalline” nonstoichiometric solids including two common “classic” structural types, identified here as structures I and II (s-I and s-II), with each type containing both small cages (sc) and large cages (lc) that form the cubic host lattice structure. The 46-water-molecule unit cell of the s-I host lattice is composed of 8 cages, of which 2 are small and 6 large. The s-II structure contains more water molecules per unit cell (136) that make up 16 sc and 8 lc. In both cases, the crystal H2O host lattice is stabilized by appropriately sized guest molecules within the cages. For a display of CH structures, including a hexagonal one, with summaries of their properties, see http:// www.lsbu.ac.uk/water/clathrate.html, as well as refs 1-4. It has long been presumed that certain properties of ether CHs reflect the formation of transient hydrogen bonds of the protonacceptor ether molecules with the cage walls. Such bonding was r 2010 American Chemical Society

first invoked to explain the orders of magnitude greater dielectric relaxation rates of the ether CHs compared to the observed values for ice and CHs having nonpolar or weakly polar guests.1,4,5 Similarly, the unusual ability of ether CHs to form rapidly in a variety of manners at temperatures in the 100-140 K range was interpreted as facilitated by molecular mobilities triggered by transient H-bonds to the host walls.6,7 Such a H-bond redirects an O-H group from the wall structure, thereby generating an orientational (Bjerrum) L defect. Such point defects are generally regarded as the source of molecular mobilities in ice-like Special Issue: Victoria Buch Memorial Received: October 7, 2010 Revised: December 3, 2010 Published: December 20, 2010 5822

dx.doi.org/10.1021/jp109620b | J. Phys. Chem. A 2011, 115, 5822–5832

The Journal of Physical Chemistry A structures,1,8 particularly at low temperatures, mobilities required for dielectric relaxation5 or structural rearrangements6 associated with crystal growth. Most recently, high populations of such defect structures have been inferred.9-12 Because guest molecules of classical clathrate hydrate structures have been viewed as lacking specific bonding to host molecules, such as a hydrogen bond, local configurations with guests H-bonded to a cage wall have been identified as “nonclassical”.11 As an extreme example, the dimethylether (DME)-HCN double CH s-II structure is composed dominantly of nonclassical components over a wide range of temperatures (60-200 K).11 The current study of transformations between classical and nonclassical structures was directed to the s-II CHs with only limited results for type s-I. Analysis of the FTIR data for new systems can be based firmly on the reported results for the s-II ether-HCN double CHs for which the ether and HCN occupy the lc and sc, respectively. This is because, regardless of the CH systems considered, the spectroscopic data reflect transitions from classical cage positions of the guest molecule to new configurations with guests H-bonded to cage wall sites. Consequently, the impact on a vibrational mode of a particular guest molecule is similar for any classical to nonclassical transformation (and may also be largely independent of the shift direction of the temperature effect). For example, as has been noted,11 the 19 cm-1 shift of the asymmetric C-O stretch mode of THF is the same whether the sc guest is a proton donor (HCN) or the electron acceptor SO2. Similarly, the downshift and intensification of the nonclassical CN stretch mode of sc HCN is similar for the different ether guests, as well as acetone. Further, as discussed below, observed shifts are in the same direction and of the same magnitude as computed shifts for both sc-guest and lc-guest dimerizations with water. These pervasive characteristics of the infrared band shifts are a valuable tool that is used extensively in analysis of the new results for the varied systems considered here.

’ EXPERIMENTAL METHODS Clathrate hydrate nanocrystal arrays have been prepared from pulse/pump cycles with vapor mixtures using a small doublewalled Ewing-type13 cold condensation cell supported on the stem of an ARS closed cycle helium refrigerator. The cell was held at 75-80 K during deposition of mixed guest and ice particle arrays and later at 124-136 K for complete conversion of the ice to CH nanocrystals.14 Ether and second-guest nanoparticles were deposited on the inner-cell windows from a ∼1% mixture in N2(g) by pulse loading aliquots of the guest mixture stored at 200 Torr in a 1 L bulb. After loading and evacuation of a 90 mL aliquot of the guest mixture, a 1% D2O or H2O in N2(g) mixture, containing a trace of HDO and maintained at 300 Torr in a 1 L bulb, was similarly cycled through the cell. This four-stage cycle (pulse/pump/pulse/pump) was repeated eight times to produce an intimately mixed array of the ice and guest particles deposited on the ZnS cell windows. During the subsequent warmup, the guest particles vaporized with the guest molecules absorbing on ice particle surfaces so that CH formation begins near 110 K. Sample FTIR spectra were monitored at 2 cm-1 resolution during particle deposition, CH formation, and, finally, to observe the temperature effects on the relative classicalnonclassical content of a CH in the 60-140 K range. In some instances, the CH composition was adjusted after formation to examine dilution and guest exchange effects on the nonclassical content.14 The lc guests studied included DME, ethylene oxide (EO),

ARTICLE

Figure 1. Cartoon representation of one sc and one lc of structure-II CH with HCN in a sc and DME in a lc, both of which are interacting with one water molecule of the intervening wall.

Figure 2. Heterodimers of water with DME, EO, TMO, THF, HCN, and C2H2 obtained at the MP2/aug-cc-pvdz level of theory. See Table 1 and the Supporting Information for structural and vibrational mode parameters.

trimethylene oxide (TMO), THF, and acetone along with sc guests HCN, H2S, acetylene (C2H2), and CO2. Ice and CH average particle sizes ranged from ∼12 to 40 nm as estimated from the infrared band intensity of modes of the dangling-H surface sites relative to that of the interior O-H/O-D vibrational modes.15 Deuterated water was used throughout much of the present study to allow a clear view of critical guest bands in the 900 cm-1 region.

’ COMPUTATIONAL METHODS The full geometry optimizations (with tight criteria) have been conducted for monomers of DME, EO, TMO, THF, HCN, and C2H2, heterodimers of these molecules with water and 5823

dx.doi.org/10.1021/jp109620b |J. Phys. Chem. A 2011, 115, 5822–5832

The Journal of Physical Chemistry A

ARTICLE

Table 1. Cluster Values from Optimized B3LYP and MP2 Levels with the aug-cc-pvDZ Basis Seta system H2O

method B3LYP

bond

frequencies (cm-1) and

frequency shift ratios

intensity change

lengths (Å)

(intensities) (km/mol)

comp./obs.

ratios

0.964 O-H

3904 (60) (O-H) asym 3794 (4) (O-H) sym

MP2

0.965 O-H

3937 (67) (O-H) asym 3803 (4) (O-H) sym

DME

B3LYP

MP2

1.41 C1-O8

3017 (130) (C-H) asym

2.35 C1-C5

2976 (66) (C-H) sym

1.09 C1-H2

1184 (102) (C-O) asym 934 (35) (C-O) sym

1.423 C1-O8

3084 (108) (C-H) asym

2.341 C1-C5

3020 (58) (C-H) sym

1.097 C1-H2

1188 (96) (C-O) asym 939 (34) (C-O) sym

EO

B3LYP

1.438 C1-O3

3196 (40) (C-H) asym

1.470 C1-C2

3093 (35) (C-H) sym

1.092 C1-H5

1296 (15) (C-C) 880 (71) (C-O) sym

1.453 C1-O3

3266 (27) (C-H) asym

1.475 C1-C2

3148 (25) (C-H) sym

831 (10) (C-O) asym MP2

1.093 C1-H5

1295 (10) (C-C) 879 (68) (C-O) sym 809 (7) (C-O) asym

TMO

B3LYP

1.492 C2-O10 1.476 C1-C2 1.070 C2-H6

3070 (68) (C-H) asym 3023 (159) (C-H) sym 1041 (5) (C-C) 1020 (95) (C-O) asym 918 (32) (C-O) sym

MP2

1.453 C2-O10

3145 (32) (C-H) asym

1.544 C1-C2

3064 (105) (C-H) sym

1.100 C2-H6

1049 (3) (C-C) 1004 (72) (C-O) asym 915 (26) (C-O) sym

THF

B3LYP

1.428 C1-O2

3123 (68) (C-H) asym

1.542 C1-C3

2980 (103) (C-H) sym

1.105 C1-H6

1092 (79) (C-O) asym 933 (28) (C-O) sym

MP2

HCN

B3LYP MP2

C2H2 (acetylene)

B3LYP

MP2

1.438 C1-O2

3173 (54) (C-H) asym

1.550 C1-C3

3034 (66) (C-H) sym

1.106 C1-H6

1094 (71) (C-O) asym 932 (22) (C-O) sym

1.156 C1-N3

2186 (1.5) (C-N)

1.074 C1-H2

3450 (67) (C-H)

1.182 C1-N3

1990 (0.5) (C-N)

1.077 C1-H2

3456 (74) (C-H)

1.209 C1-C3

2062 (0) (C-C) sym

1.070 C1-H2

3417 (95) (C-H) asym

1.231 C1-C3

3521 (0) (C-H) sym 1942 (0) (C-C) sym

1.075 C1-H2

3431 (93) (C-H) asym 3518 (0) (C-H) sym

DME 3 3 3 H2O

B3LYP

1.422 C1-O8

3871 (87) (O-H) free

2.368 C1-C5

3624 (479) (O-H) bonded

1.095 C1-H2

3047 (96) (C-H) asym 5824

(-) 225/240

dx.doi.org/10.1021/jp109620b |J. Phys. Chem. A 2011, 115, 5822–5832

The Journal of Physical Chemistry A

ARTICLE

Table 1. Continued system

method

bond

frequencies (cm-1) and

frequency shift ratios

lengths (Å)

(intensities) (km/mol)

comp./obs.

1.892 O8 3 3 3 H11 0.975 O10-H11 MP2

EO 3 3 3 H2O

B3LYP

1.432 C1-O8

ratios

2996 (64) (C-H) sym

(þ) 20/18

0.96/1

1175 (83) (C-O) asym 923 (49) (C-O) sym

(-) 9/5 (-) 11/10

1.4/1

3895 (104) (O-H) free

2.354 C1-C5

3626 (395) (O-H) bonded

1.097 C1-H2

3113 (74) (C-H) asym

1.872 O8 3 3 3 H11 0.976 O10-H11

3037 (47) (C-H) sym

(þ) 17/18

1175 (74) (C-O) asym

(-) 13/5

923 (39) (C-O) sym

(-) 16/10

1.446 C1-O3 1.469 C1-C2

intensity change

(-) 244/240 0.81/1 1.15/1

3876 (88) (O-H) free 3626 (383) (O-H) bonded

1.091 C1-H5

3217 (19) (C-H) asym

1.905 O3 3 3 3 H9 0.960 O8-H9

3108 (22) (C-H) sym 1295 (16) (C-C)

(-) 1/1

874 (83) (C-O) sym 814

(-) 6/8

1.10/1

814 (11) (C-O) asym MP2

1.463 C1-O3

3898 (102) (O-H) free

1.474 C1-C2 1.092 C1-H5

3631 (325) (O-H) bonded 3281 (11) (C-H) asym

1.888 O3 3 3 3 H9 0.976 O8-H9

3158 (15) (C-H) sym 1294 (13) (C-C)

(-) 1/1

871 (76) (C-O) sym

(-) 8/8

1.14/1

793 (8) (C-O) asym TMO 3 3 3 H2O

B3LYP

1.462 C2-O10

3871 (78) (O-H) free

1.542 C2-C3

3572 (548) (O-H) bonded

1.098 C2-H6 1.848 O10 3 3 3 H11

3097 (36) (C-H) asym 3089 (26) (C-H) sym

0.973 O12-H11

1036 (7) (C-C)

(-) 5/3

1002 (87) (C-O) asym

(-) 18/14

0.95/1

913 (45) (C-O) sym MP2

1.475 C2-O10

3893 (93) (O-H) free

1.542 C2-C3

3572 (470) (O-H) bonded

1.099 C2-H6

3199 (24) (C-H) asym

1.828 O10 3 3 3 H11 0.979 O12-H11

3123 (21) (C-H) sym 1044 (3) (C-C)

(-) 5/3

990 (48) (C-O) asym

(-) 14/14

0.66/1

908 (31) (C-O) sym THF 3 3 3 H2O

B3LYP

MP2

1.437 C1-O2

3871 (90) (O-H) free

1.539 C1-C3

3611 (566) (O-H)bonded

1.103 C1-H6

3129 (57) (C-H) asym

1.889 O2 3 3 3 H14 0.975 O15-H14

3008 (58) (C-H) sym

(þ) 28/25

0.56/1

1077 (76) (C-O) asym 921 (47) (C-O) sym

(-) 15/19

0.96/1

1.460 C1-O2

3890 (91) (O-H) free

1.539 C1-C5

3572 (446) (O-H) bonded

1.101 C1-H6

3170 (34) (C-H) asym

1.830 O2 3 3 3 H14 0.978 O15-H14

3052 (63) (C-H) sym

(þ) 18/25

0.95/1

1069 (85) (C-O) asym

(-) 25/19

1.20/1

(-) 14/8

26/4

892 (23) (C-O) sym HCN 3 3 3 H2O (HCN is proton donor)

B3LYP

1.157 C1-N3 1.083 C1-H2

3904 (85) (O-H) asym 3797 (14) (O-H) sym

0.964 O4-H5

2172 (39) (C-N)

2.035 O4 3 3 3 H2

3322 (376) (C-H) 5825

dx.doi.org/10.1021/jp109620b |J. Phys. Chem. A 2011, 115, 5822–5832

The Journal of Physical Chemistry A

ARTICLE

Table 1. Continued system

method MP2

C2H2 3 3 3 H2O (C2H2 is the proton donor)

B3LYP

MP2

a

bond

frequencies (cm-1) and

frequency shift ratios

intensity change

lengths (Å)

(intensities) (km/mol)

comp./obs.

ratios

1.183 C1-N3

3933 (91) (O-H) asym

1.085 C1-H2 0.965 O4-H5

3802 (13) (O-H) sym 1989 (9) (C-N)

2.049 O4 3 3 3 H2 1.070 C1-H2

3352 (362) (C-H)

1.210 C1-C3

3798 (9) (O-H) sym

1.076 C3-H4

2051 (11) (C-C)

(-) 1/8

18/4

(-) 46/27

2.9/2

(-) 41/27

2.8/2

3906 (73) (O-H) asym

0.964 O5-H6

3371 (273) (C-H) asym

2.191 O5 3 3 3 H4 1.074 C1-H2 1.232 C1-C3

3498 (1.4) (C-H) sym 3935 (80) (O-H) asym 3802 (9) (O-H) sym

1.066 C3-H4

1938 (8) (C-C)

0.965 O5-H6

3390 (262) (C-H) asym

2.173 O5 3 3 3 H4

3497 (1.0) (C-H) sym

The labeling of atoms follows Figure 2. The two right columns compare computed band frequency shift values and band intensity ratios for monomers versus heterodimers to the corresponding experimental quantities for classical versus nonclassical bands. The ratios given are computed over observed values; (-) and (þ) indicate decreasing or increasing frequencies. Experimental band intensity ratios between 0.5 and 1.5, which are not significant because of measurement errors, are represented by unity in the table.

ternary complexes (i.e., DME 3 3 3 water 3 3 3 HCN/H2C2). Density functional theory (DFT) with Becke's three-parameter exchange potential16 and the Lee-Yang-Parr correlation functional B3LYP17 and the second-order Møller-Plesset (MP2) methods with the augmented correlation-consistent polarized valence sets of double-ζ quality (aug-cc-pvDZ) basis sets18 were used. This basis set has already been applied and discussed for some systems that are a subject of this article, for example, THF 3 3 3 H2O,19 H2C2 3 3 3 H2O,20 and HCN 3 3 3 H2O.21 Furthermore, it was shown that this basis set produces quite well the geometries, frequencies, and electric properties of hydrogenbonded clusters.22,23 The electronic structure calculations have been performed using the GAUSSIAN 03 program.24

’ COMPUTATIONAL RESULTS It is concluded from the observed spectra that extensive hydrogen bonding of neighboring lc and sc guest molecules with water molecules of the intervening wall can occur in nonclassical configurations of ether CHs11 (see, for example, Figure 1 and the Experimental Methods section). As discussed further in the next section, in the context of the experimental results, this transition to nonclassical configurations is characterized by significant change in the spectra of the guest molecules. In order to obtain simple direct insight into the meaning of the observed frequency shifts and band intensity changes, we performed first-principle electronic structure calculations on the respective dimers and trimers. The aim was to compare the extent of frequency and intensity shifts of the guest molecules interacting with water in the heterodimer form, as in Figure 2a-f (or the heterotrimer form as in Figure S1a-b, Supporting Information), with the experimental CH values. The ternary complex computational results are complicated by a tendency to form a cyclic heterotrimer not representative of the encaged molecules of Figure 1. For that reason, the monomer and heterodimer results (Table 1) will be used as the primary source of comparative values, while the computational results for the ternary complexes are included in the Supporting Information in Table S1.

’ EXPERIMENTAL RESULTS AND DISCUSSION Throughout this section, the values in the two right-hand columns of Table 1 will be used to relate observed frequency and intensity changes to the corresponding computed values. It can be noted that the experimental ether molecule frequency shifts break cleanly into ones that red shift upon the transition to nonclassical configurations, the CO and CC modes in the 800-1200 cm-1 range, and those that blue shift, the CH stretch modes between 2800 and 3100 cm-1. The one exception is the acetylene C-H stretch mode that red shifts by 27 cm-1. In all cases, this precise pattern is also present in the corresponding computed frequency shifts (Table 1). The experimental band intensity changes are approximate because of overlapping bands; therefore, comparison with computed values is only significant where H-bonding upon dimerization leads to a large intensity enhancement factor. More complete experimental mode frequency data are included in the Supporting Information in Table S2. THF-HCN and THF-SO2 Double CH Nonclassical Behavior. The frequency of the THF C-O asymmetric stretch for the simple s-II CH of THF is ∼1073 cm-1.25,26 For this and other classical THF CHs, the infrared band of this mode shows only minor position or intensity dependence on the identity of sc guest molecules or the sample temperature, sharpening only slightly when cooled from 130 to 70 K, as in Figure 3. Further, this temperature response does not change significantly when sc guests CO2 or H2S are present in a double CH with THF. However, a remarkable ∼19 cm-1 shift in the position of the THF C-O band, identified with a switch to a nonclassical structure, occurs for both the THF-HCN and the THF-SO2 s-II double CHs.11 For the s-II THF-HCN CH, the THF C-O stretch shifts in a single step from 1073 to 1054 cm-1 (Figure 4) in a change correlated with a shift of the CN stretch mode from ∼2093 to 2085 cm-1. The CN mode shift is also accompanied by an increase in integrated band intensity by a factor of ∼4, a strong enhancement that can also be noted in the heterodimer computational results and which is often associated with H-bond formation (see Figure 9 for an example of this shift and 5826

dx.doi.org/10.1021/jp109620b |J. Phys. Chem. A 2011, 115, 5822–5832

The Journal of Physical Chemistry A

Figure 3. Spectra showing the insensitivity to temperature change, from 70 to 130 K, of the C-O asymmetric stretch mode of THF as a guest in the s-II simple THF CH. The weak broad band near 1055 cm-1 is from THF adsorbed on the surface of the CH nanocrystals.

ARTICLE

Figure 5. Comparison of the peak intensity change in absorbance units (au) as a function of temperature at ∼1073 and ∼1054 cm-1 for the THF-HCN and THF-SO2 double s-II CHs. Data are from samples with D2O host molecules.

Figure 4. Infrared spectra showing the band intensity shift with decreasing temperature, from the classical frequency of 1073 to the nonclassical value of 1054 cm-1, for the asymmetric C-O stretch mode of guest THF in the double THF-HCN CH. Here and in subsequent figures, au signifies absorbance units.

Figure 6. Variation of the cooling-induced (peak) intensity transfer of the THF classical 1073 cm-1 to nonclassical ∼1054 cm-1 band from dilution of the sc HCN with CO2 at levels ranging from 50 to 85%. Note, by comparison with Figure 5, that 50% decrease in the HCN sc content has only a trivial effect.

intensification of the CN stretch mode). The 19 cm-1 THF shift also closely matches the monomer-dimer computed shift of average value 20 cm-1. The peak intensity of the THF 1054 cm-1 band relative to that at 1073 cm-1 increases steadily as the temperature decreases from 130 to 80 K (Figure 4), with a maximum rate from 110 to 90 K as shown in Figure 5. A similar transfer of band intensity, from 1073 to 1054 cm-1, is observed when SO2 rather than HCN is present in the sc. The THF-SO2 double CH was formed at 136 K from the simple THF s-II CH by exposure to the saturation vapor of SO2. While the resulting CH gives a similar band frequency shift, the net intensity transfer with cooling is only ∼50% of that observed with HCN as the sc guest (Figure 5). This difference is centered in the 90-60 K range as intensity transfer stops for the THF-SO2 CH below ∼95 K, suggesting onset of a kinetic factor resulting from lower SO2 mobility within the CH cage. SO2 is a potent Lewis acid, perhaps capable of significant association with oxygens of the cage wall, as proposed for Cl2.27 Binding to sc wall oxygen molecules, other than those exposed to an available ether molecule, could restrict SO2 involvement in the nonclassical configurations below 90 K. The more limited nature of this transition to a nonclassical structure induced by sc SO2 has been mimicked in the THF-

HCN CH through dilution of the sc population of HCN with CO2 (Figure 6), which, as noted above, does not act with THF to induce an observable nonclassical component. With the sc population evenly divided between CO2 and HCN, a comparison of Figures 6 and 5 shows that the CO2 causes no major change in the THF-HCN classical to nonclassical effect. This is expected because, for the s-II CH with a 2:1 sc to lc ratio, only 50% of the original sc HCN molecules are required for a full nonclassical transition, that is, a lc THF molecule can only interact indirectly with one sc HCN molecule in the manner depicted for DME in Figure 1, while there are approximately twice as many HCN as ether molecules. In fact, as Figure 6 shows, the sc population of CO2 approaches 85% before the form of the band intensity shift, from 1073 to 1054 cm-1, resembles that of the s-II THF-SO2 CH. This dilution effect, which can be reversed by replacing CO2 with HCN, is consistent with the SO2 effect also being limited by a population of ineffectual SO2 guest molecules, that is, ineffectual SO2 molecules bound to the cage walls at 90% nonclassical below 100 K, but unlike that for the “pure” DME-HCN double CH, a significant classical component appears at higher temperatures. Clearly, the CH structural distortion of the nonclassical DME enables neighboring acetone to further engage in nonclassical H-bonds, while acetone suppresses somewhat the nonclassical tendencies of DME; that is, the energetics of the classical-nonclassical transition of one lc guest is sensitive to the identity of a neighboring lc guest. Defining the Unique Classical to Nonclassical Transitions. The infrared spectra help identify three additional distinct variations on the nature of classical to nonclassical structural transitions for ether CHs. Because these are now included with the original (type 1) classical to nonclassical transitions described previously,11 there is a need to define and label the four possibilities. Type 1 is defined as a transition of CHs with the combination of lc proton-acceptor guests with sc proton-donor or electron-acceptor molecules. Such transitions display increasing populations of nonclassical configurations with decreasing sample temperature. The shifted nonclassical bands for type 1 transitions are generally characterized by bandwidths greater than but similar to bands of the classical configurations (i.e., < 10 cm-1;

ARTICLE

Figure 13. Infrared spectrum of CN stretch of HCN in the sc of a s-I double CH with EO showing the shift of intensity from the classical to nonclassical band from warming above 100 K. The 2085 cm-1 band is distorted on the low-frequency side at 110 K by adsorbed HCN that desorbs at higher temperatures.

see Figures 4, 7-10, and 12). Since ether guest mode shift values resemble computed shifts for the H-bonding of ethers in dimers with water (Table 1) and with an ice or CH particle surface,32,33 the type 1 transition is attributed to an increasing population of guest-host H-bonds, as also computed for the DME-HCN double CH.11 A type 2 transition, like type 1, depends on a guest sc molecule to help stabilize nonclassical structures at low temperatures, but in addition to displaying type 1 behavior, it also transitions to nonclassical rather than mostly classical configurations at elevated temperatures. This double action is displayed by TMO-HCN in Figure 7 and DME-C2H2 in Figure 10a, that is, a transition to a type 1 nonclassical form at low T with a type 2 transition at higher T. Rather than forming a sharp intense 990 cm-1 classical band, as in the case of the THF 1073 cm-1 band of Figure 6, warming of the TMO-HCN as well as the DME-C2H2 hydrates promotes the evolution of a broad band that extends from the classical frequency to beyond the type 1 nonclassical position. It is not that the nonclassical 974 cm-1 TMO band of Figure 7 is strongly intensified but rather that the “990” band peak intensity is weak at 144 K because of its unusual breadth; for example, the absorbance at 968 cm-1 is actually greater at 144 K than that at 65 K. The TMO band at 144 K resembles that noted for DME in Figure 11 at 130 K, with a bandwidth of ∼20 cm-1 and with much of the broad adsorption near the position of the sharper low-temperature nonclassical band. At this time, only EO with sc HCN has been confirmed to display the type 3 transition. Like type 1 transitions, it leads to a nonclassical CN band of increased breadth shifted ∼9 cm-1 from the classical value (Figure 13). However, the 2085 cm-1 band only emerges and grows with increased temperatures. Despite the presence of HCN small guests, this s-I double CH does not show any type 1 transition during cooling. Several of the EO bands broaden above 110 K but experience only minor shifts (