69451 Weinheim, Germany - Wiley-VCH · 3 was 20 s and the detector distance was 6 cm. All CF 3 groups were disordered a nd refined using bond distance constraints. S1.2 Reaction of

Supporting Information © Wiley-VCH 2006

69451 Weinheim, Germany

1

Cyclic Dimethylsiloxanes as Pseudo Crown Ethers; Syntheses and Characterization of Li(Me2SiO)5[Al(OC(CF3)3)4], Li(Me2SiO)6[Al(OC(CF3)3)4] and Li(Me2SiO)6[Al(OC(CF3)2Ph)4]** Andreas Decken, Jack Passmore,* Xinping Wang S1 Experimental section S1.1 General experimental technique S1.2 Reaction of Li[AlPhF] ([AlPhF] = [Al(OC(CF3)2Ph)4]) with D5 in CH2Cl2 S1.3 Reaction of Se4(AsF6)2, 2Se2Ph2, with 2 Li[AlF] ([AlF] = Al(OC(CF3)3)4) leading to crystals LiD6[AlF] S2 FT-IR, FT-Raman spectra and Multiple Nuclear NMR chemical shifts S3 Crystal structures S4 Calculations

2

S1 Experimental section S1.1 General experimental technique

All manipulations were performed by using standard Schlenk techniques under nitrogen

atmosphere, grease free metal apparatus, and dry box techniques under nitrogen atmosphere. The

compounds Li[AlF] ([AlF] = Al(OC(CF3)3)4) and Li[AlPhF] ([AlPhF] = Al(OC(CF3)2Ph)4) were

prepared by literature1 methods and their purity were checked by H1, 19F, 7Li, 27Al and 13C NMR,

and FT-Raman spectroscopy. D5 (Aldrich, 97%) and D6 (Gelest, >95%) (D = Me2SiO) were applied

once arrived. Sulfur dioxide (Matheson, anhydrous, 99.85%) was vacuum-distilled and stored over

CaH2 before use. CH2Cl2 and n-hexane were dried over CaH2 and degassed. NMR spectra were

recorded on a Varian 400 NMR spectrometer. 1H, 13C and 29Si[1H] chemical shifts were reported in δ

units downfield from Me4Si in SO2 as the reference signal. CFCl3 (in D2O), AlCl3 (in D2O), and

LiAsF6 (in D2O) were used as references for measuring 19F, 27Al and 7Li, respectively. NMR samples

were prepared in 10 mm thick walled NMR tubes fitted with J. Young valves using SO2 as solvent.

FT-IR spectra were recorded using a Thermo Nicolet spectrometer (Nexus 470 FT-IR). FT-Raman

spectra were obtained from neat samples, sealed under a nitrogen atmosphere in glass capillaries,

using an FT-IR spectrometer (Bruker IFS66) equipped with an FT-Raman accessory (Bruker FRA

106) incorporating a Nd-YAG laser. Chemical analyses were performed by Galbraith Laboratories,

Inc. (U. S.). Mass spectra were recorded using a KRATOS ms 50 TC mass spectrometer equipped

with an EI source, 70 eV or 30 eV, from samples sealed in dried glass melting point tubes by the

direct inlet method. Melting points were measured in sealed capillary tubes under nitrogen.

Single crystals were coated with Paratone-N oil, mounted using a glass fibre and frozen in the

cold nitrogen stream of the goniometer. For the crystal LiD6[AlPhF], a hemisphere of data was

collected on a Bruker AXS P4/SMART 1000 diffractometer using ω and θ scans with a scan width

of 0.3 ° and 10 s exposure times. The detector distance was 5 cm. The data were reduced (SAINT)2

and corrected for absorption (SADABS).3 The structure was solved by direct methods and refined

by full-matrix least squares on F2(SHELXTL). 4 All non-hydrogen atoms were refined

anisotropically. Hydrogen atoms were placed in calculated positions and refined using a riding

model. For the crystal LiD6[AlF], the operation was similar except that the crystal was a multiple

twin and the orientation matrix for the major component was determined (RLATT).5 Exposure time

3

was 20 s and the detector distance was 6 cm. All CF3 groups were disordered and refined using bond

distance constraints.

S1.2 Reaction of Li[AlPhF] with D5 in CH2Cl2

CH2Cl2 (10 ml) was transferred onto (Me2SiO)5 (0.39 ml, 1.01 mmol) over solid Li[AlPhF](

0.925g, 0.92 mmol) in a 100-ml Schlenk flask. N-hexane (25 ml) was added into the yellowish clear

solution after stirred overnight at room temperature. A small amount of crystals afforded from the

solution at –20 oC after 1 day and removed by filtration. The volatile in the filtrate was removed and

the solid product was rinsed by CH2Cl2 ten times followed by n-hexane three times. Together with

crystals, the total amount for the product is 0.905 g. FT-IR (KBr, neat, cm-1, RT): 1497w, 1449m,

1324w, 1302m, 1280m, 1255s, 1227s, 1200s, 1171s, 1140s, 1107m, 1080m, 1035m, 1002w, 969s,

948s, 932s, 920m, 860m, 829m, 793m, 760m, 734m, 715s, 710s, 692m, 667m, 657m, 571m, 554m,

534m, 507w, 493w, 473w, 456w, 411w, 390w, 370w. Single crystal X-ray diffraction: Li[AlPhF],

triclinic, P-1, a = 9.3358(13), b = 12.7227(17), c = 17.741(2) Å; a = 80.465(2), ß = 76.681(2), ? =

71.473(2)o; Z = 2. Previously reported: Li[AlPhF], monoclinic, C2/c, a = 42.297(6), b = 10.641(1), c

= 19.132(2) Å; a = 90, ß = 114.808(9), ? = 90o; Z = 8.1

Conclusion: FT-IR (Figure S2.3) and single crystal X-ray diffraction (Figure S3.1) showed that

the product was the starting material Li[AlPhF]. The structure of Li[AlPhF] is similar to the reported

one except for unit cell parameters. Thus the reaction of Li[AlPhF] with D5 in CH2Cl2 did not afford

the expected product LiD5[AlPhF] in CH2Cl2.

S1.3 Reaction of Se4(AsF6)2, 2Se2Ph2, with 2 Li[AlF] leading to crystals LiD6[AlF]

Liquid SO2 (9.65 g) was transferred onto Se4(AsF6)2 (0.36 g, 0.52 mmol), Se2Ph2 (0.16 g, 0.51

mmol) and Li[AlF] (1.02 g, 1.05 mmol) in a bulb of a two bulb, two valve vessel fitted with a

stirring bar magnet and a medium frit. The resulted bright red solution over a small amount of

colorless product was filtered after a day and the filtrate was condensed to 2 ml. No crystals were

found after several days. SO2 was removed and CH2Cl2 (3 ml) was added. The resulted bright red

solution was transfer into a schlenk vessel and was set at a cold room (-30 oC) for several days.

Some colorless crystals were found and single crystal X-ray diffraction showed they are LiD6[AlF].

4

S2 FT-IR, FT-Raman spectra and Multiple Nuclear NMR chemical shifts

1000 1500 2000 2500 3000 Wavenumbers (cm-1)

**

*

* *

D5

Li[AlF]

LiD5[AlF]

LiD5[AlF],raman

Figure S2.1 FT-IR of D5, Li[AlF], and LiD5[AlF]; FT-Raman of LiD5[AlF] (Peaks marked by * are due to nujol mull)

5

500 1000 1500 2000 2500 3000 Wavenumbers (cm-1)

*

*

* *

*

D6

Li[AlF]

LiD6[AlF]

LiD6[AlF], Raman

Figure S2.2 FT-IR of D6, Li[AlF], and LiD6[AlF]; FT-Raman of LiD6[AlF]

(Peaks marked by * are due to nujol mull)

6

Figure S2.3 FT-IR spectra of D5, Li[AlPhF] and the solid product

from the reaction of Li[AlPhF] with D5 in CH2Cl2

500 1000 1500 2000 2500 3000

Wavenumbers (cm-1)

solid product of the reaction of Li[AlPhF] with D5 in CH2Cl2

D5

Li[AlPhF]

7

500 1000 1500 2000 2500 3000

Wavenumbers (cm-1)

D6

Li[AlPhF]

LiD6[AlPhF]

LiD6[AlPhF],Raman

Figure S2.4 FT-IR of D6, Li[AlPhF], and LiD6[AlPhF]; FT-Raman of LiD6[AlPhF]

8

Table S2.1 Chemical Shifts of LiD6[AlF], LiD6[AlPhF], D6, Li[AlF]a

and Li[AlPhF]a in SO2 solution at room temperature

a.Decken, A.; Jenkins, H D B.; Nikiforov, G B.; Passmore, J. Dalton Trans. 2004, 2496; b. relative to ligands and anions.

δ D6 Li[AlF] Li[AlPhF] LiD6[AlF] LiD6[AlPhF] ? δb LiD6[AlF] LiD6[AlPhF] 19F -74.0 -73.8 -75.2 -74.0 -1.2 0 27Al 36.8 41.9 35.0 29.6 -1.8 -12.3 7Li -0.36 -0.12 0.19 0.18 0.55 0.30 1H -0.054 7.0-7.8(Ph) 0.296 0.29(36H) 0.35 0.344 7.2-7.9(Ph, 20H) 29Si -22.69 -9.22 -10.14 13.47 12.55

9

S3 Crystal structures

OAl

Li

Figure S3.1 Structure of Li[AlPhF] obtained from

the reaction of Li[AlPhF] with D5 in CH2Cl2

10

(a) (b)

( c ) (d)

11

(e) (f)

Figure S3. 2 Structures of LiD6+, LiD6[AlF] and the related cyclophosphozene metal complex. (a)

Structure of LiD6[AlphF]; (b) a side-view of LiD6+; (c) Bond distances in LiD6

+ ; (d) Bond valences

in LiD6+ ; (e) Bond angles in LiD6

+; (f) Structure of [N6P6(NMe2)12CuCl]+ (Cl atom and Me groups

were omitted for clarity).6

12

(a) (b)

(c)

Figure S3. 3 Structures of LiD6+ and LiD6[AlPhF]: (a) Structure of LiD6[AlF]; (b) Bond distances in

LiD6+; (c) Bond angles in LiD6

+; (d) Bond valences in LiD6+.

13

(a)

Si O2 O1

Si Si

O3

Si

O5

O4

Si

(b) (c) (from CCDC, 2005)* Figure S3.4 Structure of D5 and preliminary structures of LiD5+ in LiD5[AlF] (a) Structure of LiD5[AlF]; (b) Structure of LiD5

+; (c) Structure of D5 (*S. Parson, private communication, CCDC, 2004).

14

Table S3.1 Structural parameters of [AlPhF]- in LiD6[AlphF] and Li[AlPhF]

LiD6[AlphF] Li[AlPhF]

Al-O (Å) 1.711(1)-1.734(1) 1.687(3)-1.773(2)

C-O-Al (o) 145.5(1)-154.8(1) 151.5(2)-168.1(2)

O-Al-O (o) 106.3(1)-111.8(1) 91.8(1)-116.7(1)

15

S4 Calculations Computational details:

All calculations have been carried out with the Gaussian 03 program.[7] Geometry optimizations

were performed at HF/3-21G, HF/6-31G*, and B3LYP/6-31G* levels. All structures optimized at

HF level were identified as true local minima on the energy potential surface unless specified. For

the sake of minimizing computational cost, only two structures, LiD5+(Cs) and D6 (C2), were selected

and characterized as local minima by harmonic frequency calculations at B3LYP/6-31G* level.

Binding energies were evaluated relative to free metal cations and free D5 (Cs) or D6 (C2) and were

counterpoise corrected except Lanl2dz basis set. Enthalpy corrections were determined using HF/3-

21G harmonic vibrational frequencies. The orbital and bond analyses were performed with the

natural bond orbital (NBO) methods at B3LYP/6-31G*.

16

(1) D5 and its alkali metal complexes

Table S4.1 Total Energies, Binding Energies, and Binding Enthalpies of free and Complexed D5

a. Total energy in au; b. Binding energy and binding enthalpy in kJ.mol-1.

Sym method energya ? Eb ? H298b D5 Cs HF/3-21G -2204.09160 Cs HF/6-31G* -2215.73152 Cs B3LYP/6-31G* -2223.36372 C5v HF/3-21G -2204.09113 C5v HF/6-31G* -2215.72957 C5v B3LYP/6-31G* -2223.36055 LiD5

+ C2v HF/3-21G -2211.42197 -376 C2v HF/6-31G* -2223.08758 -301 C2v B3LYP/6-31G* -2230.78463 -343 Cs HF/6-31G* -2223.08784 -303 -294 Cs B3LYP/6-31G* -2230.78469 -344 -335 D5h HF/3-21G -2211.43506 -375 D5h HF/6-31G* -2223.08704 -301 D5h B3LYP/6-31G* -2230.78322 -340 NaD5

+ C5v HF/3-21G -2364.87309 -231 -224 C5v HF/6-31G* -2377.47358 -208 -200 C5v B3LYP/6-31G* -2385.54214 -244 -237 D5h HF/3-21G -2364.86690 -198 D5h HF/6-31G* -2377.45416 -152 D5h B3LYP/6-31G* -2385.52710 -201 KD5

+ C5v HF/3-21G -2800.15842 -122 -115 C5v HF/3-21G/Lanl2dz -2231.85146 -141 -134 C5v HF/6-31G* -2814.75457 -128 -120 C5v B3LYP/6-31G* -2823.15168 -156 -149 D5h HF/3-21G -2800.04548 +202 D5h HF/6-31G* -2814.62695 +214 D5h B3LYP/6-31G* -2823.05062 +117

17

Table S4.2 Summary of geometries parameters for D5 and its complexes with the alkali metals (bond distance Å; bond angle o)

a. σ referring to the plane defined by all oxygen atoms and dM-σ referring to the distance from metal atom (M) to the σ plane; b. H. Oberhammer, W. Zeil, G. Fogarasi, Journal of molecular structure. 1973, 18, 309; c. S. Parson, et al.. private communication, CCDC, 2004; d. Lanl2dz for K, 3-21G for Si, C, O and H; *noncoordinating O.

Sym method M-O Si-O Si-O* Si-C Si-O*-Si Si-O-Si dM-σa

D5 Cs HF/3-21G 1.647 1.893 174.2 Cs HF/6-31G* 1.632 1.874 159.0 Cs B3LYP/6-31G* 1.654 1.876 152.7 C5v HF/3-21G 1.646 1.893 177.9 C5v HF/6-31G* 1.626 1.875 178.1

C5v B3LYP/6-31G* 1.644 1.877 178.0 D5d B3LYP/6-31G* 1.646 1.877 167.2 D5d ? Gas electr.b 1.620 1.845 146.5 C1 X-rayc 1.629 1.845 148.0 LiD5

+ C2v HF/3-21G 2.038 1.678 1.654 1.882 170.5 165.8 0 C2v HF/6-31G* 2.024 1.654 1.638 1.861 168.4 166.5 0 C2v B3LYP/6-31G* 2.019 1.683 1.655 1.863 165.5 164.6 0 Cs HF/6-31G* 2.034 1.664 1.638 1.862 159.3 159.0 0.332 Cs B3LYP/6-31G* 2.023 1.684 1.656 1.863 157.5 159.2 0.271

D5h HF/3-21G 2.141 1.670 1.882 169.8 0 D5h HF/6-31G* 2.133 1.654 1.863 170.6 0

D5h B3LYP/6-31G* 2.150 1.672 1.864 170.2 0 NaD5

+ C5v HF/3-21G 2.353 1.674 1.882 161.2 0.867 C5v HF/6-31G* 2.434 1.657 1.864 155.1 1.076 C5v B3LYP/6-31G* 2.420 1.677 1.866 153.5 0.990 D5h HF/3-21G 2.196 1.687 1.881 172.3 0

D5h HF/6-31G* 2.210 1.667 1.862 174.4 0 D5h B3LYP/6-31G* 2.233 1.684 1.864 174.3 0 KD5

+ C5v HF/3-21G 2.864 1.667 1.884 159.3 1.796 C5v HF/3-21G/Lanl2dzd 2.863 1.667 1.884 158.3 1.796 C5v HF/6-31G* 2.900 1.652 1.866 153.1 1.875 C5v B3LYP/6-31G* 2.855 1.673 1.868 150.4 1.774 D5h HF/3-21G 2.335 1.710 1.878 178.7 0

D5h HF/6-31G* 2.365 1.701 1.861 178.4 0 D5h B3LYP/6-31G* 2.388 1.716 1.863 178.2 0

18

Geometries Uncomplexed D5 ring

Ab initio calculations on several conformation of polysiloxane, (H2SiO)n (n =3, 4, 5) have been

reported. But for dimethylcyclicsiloxanes, only D3 (D = Me2SiO) has been calculated by ab initio

method. Calculation on larger rings by semi-empirical methods such as AM1 and PM3 were

performed but resulted in large discrepancy with experimental results (Table S4.3).

Table S4. 3 Comparison of Geometry parameters of the D5 ring PM3a AM1a HF/3-21Gb HF/6-31G*b B3LYP/6-31G*b Exp. (gas)c Exp. (X-ray)d

Si-O 1.67 1.71 1.647 1.632 1.654 1.620(2) 1.629(2) Si-C 1.91 1.80 1.893 1.874 1.876 1.845(4) 1.845(4) Si-O-Si 134.06 168.72 174.2 159.0 152.7 146.5(1) 148.0(1) a. Field, R. J.; Olson, E. W. journal of non-crystalline solids, 2001, 285, 194 b. this work c. Oberhammer, H; Zeil, W.; Fogarasi, G. Journal of molecular structure, 1973, 18, 309 d. Simon Parson, private communication, CCDC, 2004

Two conformations of D5 were calculated at HF/3-21G, HF/6-31G* and B3LYP/6-31G* levels:

Cs and C5v forms. The Cs geometry has the lowest energy and the (SiO)5 framework is puckered with

some methyl groups pointing inward (Figure S4.1a and b). The C5v geometry has the highest

symmetry with all oxygen and silicon atoms in a plane, forming a cavity for a guest (Figure S4. 1c

and d). The C5v geometry is a saddle point as could be expected. The Si-O bond length of the Cs

form calculated at HF/6-31G* level is very close to experimental structures (gas phase electron

diffraction and X-ray single crystal diffraction). The SiOSi angle at B3LYP/6-31G* level is in good

agreement with experimental result while the Si-O bond length is longer (Table S4.2 and 3).

Geometry parameters seem to be overestimated at HF/3-21G level (Table S4.2 and 3). Cs geometry

is 1-8 kJ.mol-1 more stable than the C5v form.

19

1.89

2

C. 1.651

A. 1.647

C. 1.654

C. 1.652A. 167.4

B. 163.7

A. 1.647

C. 152.8

A. 1.647

C. 153.8

B. 1.630

B. 1.631

B. 167.4

B. 165.7A. 165.7

C. 160.4B. 1.630

B. 1.628

C. 1.650

A. 1.647

A. 167.6

B. 1.630C. 1.655

A. 1.647

a. D5 (Cs, top-view, b. D5 (Cs, side-view) A. HF/3-21G; B. HF/6-31G*; C. B3LYP/6-31G*)

A. 1 kJ.mol-1

B. 5 kJ.mol-1

C. 8 kJ.mol-1

B. 1.626

C. 178.0

A. 1.646

A. 177.9B. 178.1

C. 1.644

c. D5 (C5v, top-view, d. D5 (C5v, side-view) A. HF/3-21G; B. HF/6-31G*; C. B3LYP/6-31G*)

20

1.631(3)1.637(2)

138.7

1.624

(3)

1.61

9(2)

1.62

7(2)

156.8

1.631(3)

1.630(3)

1.628(3)

1.621(3)155.4

1.628(3)

144.2

144.9

e. X-ray (C1, top-view) f. X-ray (C1, side-view) Figure S4.1 D5 geometry

21

LiD5+

A C2v structure was optimized at HF/3-21G level. The Li+ cation was located in the (SiO)5 ring

plane and coordinated with four oxygen atoms (Figure S4. 2e and f). This C2v geometry is saddle

points at HF/6-31G* and B3LYP/6-31G* levels (Figure S4. 2e and f). A HF/6-31G* and B3LYP/6-

31G* minima were achieved only with a lower symmetry (Cs) in which all five oxygen atoms (but

not with Si atoms) are in a plane (σ) and the Li atom was out of (SiO)5 framwork (dLiσ = 0.332 Å

HF/6-31G* level, 0.271 Å, B3LYP/6-31G*) (Figure S4. 2g and h). This Cs structure at HF/6-31G*

and B3LYP/6-31G* levels confirmed X-ray single crystal structure of LiD5+ (with [AlF]- anion) that

has not been well solved (Figure S4.2a and b). A small energy barrier (-2 kJ.mol-1, HF/6-31G*; -1

kJ.mol-1, B3LYP/6-31G*) for Li atom to move from the C2v structure to Cs structure through the

cavity was evaluated as a difference between total energies. We also optimized a D5h geometry at

three levels in which the Li atom is located at the centre of the (Si5O5) plane and equally coordinates

to five oxygen atoms (Figure S4. 2c and d). The Li-O distance (2.133-2.150 Å) appears a little longer

than that of the Cs (2.023 – 2.034 Å) and C2v (2.019 – 2.038 Å) geometries (see Table S4. 2). This

geometry was characterized as a saddle point and is 0.4-3 kJ.mol-1 less stable than C2v form.

Li

a. LiD5

+ (C1, X-ray, top-view) b. LiD5+ (C1, X-ray, side-view)

22

Li

C. 170.2

A. 169.8

A. 1.670

B. 2.805

B. 1.654

C. 2.834

B. 170.6

C. 1.672

A. 2.830

c. LiD5

+ (D5h, top-view, d. LiD5+ (D5h, side-view)

A. HF/3-21G; B. HF/6-31G*; C. B3LYP/6-31G*)

A. -1 kJ mol-1

B. -0.4 kJ mol-1

C. -3 kJ mol-1

A. 1.654

A. 2.569

A. 170.5

C. 1.938

A. 1.689

A. 1.672

C. 176.6

B. 2.092

C. 1.655

A. 1.686A. 1.958

A. 156.9

A. 2.118

C. 2.100

C. 164.6

A. 1.667

C. 154.3

A. 174.6

C. 1.689

C. 1.675

C. 1.671

C. 1.697

B. 1.638

B. 175.9

B. 1.956

B. 1.653

B. 157.1

B. 2.609

B. 1.638

B. 168.4

B. 1.657

B. 1.668

C. 2.718

e. LiD5

+ (C2v, top-view, f. LiD5+ (C2v, side-view)

A. HF/3-21G; B. HF/6-31G*; C. B3LYP/6-31G*)

23

A. 1.654

A. 2.569

A. 170.5

C. 1.938

A. 1.689

A. 1.672

C. 176.6

B. 2.092

C. 1.655

A. 1.686A. 1.958

A. 156.9

A. 2.118

C. 2.100

C. 164.6

A. 1.667

C. 154.3

A. 174.6

C. 1.689

C. 1.675

C. 1.671

C. 1.697

B. 1.638

B. 175.9

B. 1.956

B. 1.653

B. 157.1

B. 2.609

B. 1.638

B. 168.4

B. 1.657

B. 1.668

C. 2.718

e. LiD5

+ (C2v, top-view, f. LiD5+ (C2v, side-view)

A. HF/3-21G; B. HF/6-31G*; C. B3LYP/6-31G*)

B. - 2 kJ.mol-1

C. - 1 kJ.mol-1

1.086

1.08

6

1.086

C. 1.672

C. 1.689

C. 166.0

1.863

C. 2.766

B. 1.669

B. 1.654

B. 164.3

B. 1.658

Si1

B. 2.107

B. 1.638

B. 1.677

B. 1.960

B. 153.7

B. 159.3

B. 2.710

O1

Li

C. 1.677

C. 1.940

C. 152.5

C. 1.698

C. 1.656

C. 2.108

C. 157.5

Li

g. LiD5

+ (Cs, top-view , h. LiD5+ (Cs, side-view

B. HF/6-31G*; dLi-σ: 0.332 HF/6-31G*; C. B3LYP/6-31G*) 0.271 B3LYP/6-31G*) Figure S4. 2 LiD5

+ geometry

24

NaD5+ and KD5

+

Two structures were calculated for each of the NaD5+ and KD5

+ complexes. One of these

structures corresponds to the D5h conformation with the metal cations residing at the centre of the

cavity (Figure S4.3a and b). Since Na+ and K+ cations are rather big, it seems unlikely that these

structures are stable corresponding to minima on the potential energy surface. This was confirmed by

normal mode analysis which revealed negative frequencies. We therefore calculated a structure of a

lower symmetry. As a result, a stable C5v geometry for NaD5+ and KD5

+ was obtained with the cations

out of the (SiO)5 framework (Figure S4. 3c and d). The K-O distance (2.855-2.900 Å) in KD5+ is

comparable to that in KD7+ (2.93 (4) Å) in the solid state.8 The K atom in C5v geometry was farther

away from the σ plane defined by five oxygen atoms than Na atom. For NaD5+, the C5v structure is 33

- 56 kJ.mol-1 more stable than the D5h geometry; for KD5+, the difference is even bigger (273 - 342

kJ.mol-1). Actually, binding of D5 with K+ in a D5h form is impossible as the estimated binding energy

is positive (+117 - +214 kJ.mol-1) (Table S4.1). To access our ability to accurately compute the KD5+

geometry and binding energy, we carried out one calculation with an ECP basis set (Lanl2dz for K)

which includs some relativistic effects. As can be seen in Table S4.1 and Table S4. 2, the 3-21G

binding energy is 19 kJ.mol-1 less than Lanl2dz basis set but the structural parameters including

lengths and angles predicted by lanl2dz (for K) basis set did not show any significant difference from

3-21G basis set. The 19 kJ.mol-1 difference does not affect our comparison in binging energies of D5

with alkali metal cations.

25

1.681

ababcc

a. NaD5

+ (D5h, top-view, b. NaD5+ (D5h, side-view)

A. HF/3-21G; B. HF/6-31G*; C. B3LYP/6-31G*)

A. -33 kJ.mol-1

B. -56 kJ.mol-1

C. -43 kJ.mol-1

C. 1.677

B. 2.434

C. 153.5

C. 2.420

A. 1.674

A. 161.2

B. 1.657

B. 155.1

A. 2.353

c. NaD5

+ (C5v, A. HF/3-21G; d. NaD5+ (C5v, dNa-σ: 0.867 HF/3-21G;

B. HF/6-31G*; 1.076 HF/6-31G*; C. B3LYP/6-31G*, top-view) 0.990 B3LYP/6-31G*; side-view)

Figure S4. 3 NaD5+ geometries

26

1.681

ababcc

a. KD5

+ (C2v, top-view, b. KD5+ (C2v, side-view)

A. HF/3-21G; B. HF/6-31G*; C. B3LYP/6-31G*)

A. -324 kJ.mol-1

B. -342 kJ.mol-1

C. -273 kJ.mol-1

C. 1.673B. 1.652

C. 2.855

A. 159.3

C. 150.4

A. 2.864B. 2.900

D. 1.667

A. 1.667

D. 158.3

B. 153.1

D. 2.863

K

c. KD5

+ (C5v, A. HF/3-21G; d. KD5

+ (C5v, dK-σ: 1.796 HF/3-21G;

B. HF/6-31G*; 1.875 HF/6-31G*; C. B3LYP/6-31G* ; 1.774 B3LYP/6-31G*; D. HF/3-21G/lanl2dz, top-view) 1.796 HF/3-21G/lanl2dz side-view) Figure S4. 4 KD5

+ geometries

27

NBO charge analysis

NBO and Natural energy decomposition analysis (NEDA) of M+/18-crown-6 complexes

indicated that the M+/18-crown-6 interaction is dominated by electrostatic interaction (ES) and

polarization (POL). Charge transfer (CT) contribution is much less important.9 In the present M+/D5

complexes, NBO analysis also revealed strong polarization of the D5 ring by the cation (Table S4.4).

The oxygen and silyl charges in MD5+ complexes showed strong polarization of the electron density

from the silyls (SiMe2) towards the oxygen centres compared to that in neutral D5 ring, e.g. the

charge on the silyl groups in the D5 structures is +1.275, significantly less than +1.329 in the LiD5+

complex. A transfer of 0.054 from each silyl group to the near oxygen atom has been clearly

reflected. Polarization is the strongest in LiD5+ complex and became weaker as going down Group

1. A weak charge transfer from the metal to D5 ring is reflected in the atomic charge (e.g. 0.920 at

Li), 0.08e less than +1 charge of the Li+ cation. The charge transfer also decreases as the metal

cation becomes bigger.

Table S4.4 Atomic and fragment charges for D5, D6 and their alkali metal complexes(B3LYP/6-31G*)

D5 LiD5+ NaD5

+ KD5+

Sym Cs Cs C5v C5v q(M+) 0.920 0.962 0.981 q(O) -1.275 -1.312 -1.305 -1.300 q(SiMe2) 1.275 1.329 1.313 1.303 ?q(SiMe2) 0 0.054 0.038 0.028

28

Frequencies, binding energies and gas phase selectivity of alkli metal

with D5 ring

Normal mode frequencies for D5 and LiD5

+ calculated at HF/3-21G level (scale factor 0.9085)

were in good agreement with that by FT-IR and FT-Raman spectroscopies (Table S4. 5). Figure S4.5

shows the counter poise (CP) corrected binding energies for MD5+ at the three levels (also see Table

S4.1). All levels of theory predicted that the binding affinity of D5 decreased with increasing cation

size. This suggests that binding energies in gas phase of D5 with alkali metals correspond to cation-

ring cavity size matching principle. Thus D5 has the highest selectivity for Li+ as the Li+ cation

matches the cavity of D5 better than other alkali metals. The gas phase selectivity of alkali metals of

D5 is very similar to that of 12-crown-4 based on gas phase binding energy calculation.10

Table S4.5 Calculated and experimental vibrational bands of D5 and LiD5+

D5 D5 (HF/3-21G) LiD5+ in LiD5[AlF] LiD5

+ (HF/3-21G) Assign.

IR Raman IR Raman IR Raman IR Raman

2963s 2956w 2957m 2972m 2975s 2960w 2960s υa(CH) 2899s 2891w 2894s 2907w 2914vs 2894w 2897s υs(CH) 1487w 1483w 1494w 1479w 1482w δa(CH3) 1408m 1466w 1470w δa(CH3) 1349m 1356w 1363m δa(CH3) 1258s 1265m 1275s δs(CH3) 1082s,br 1172s,br 1054s,br 1133s υas(SiOSi) 1061s υas(SiOSi) 859s 875w 858m 855s 876s ρ (CH3) 835w 825s ρ (CH3) ρ (CH3) 799s 795w 808s 797w υa(SiC2) 752m 745w υa(SiC2) 714s 711w 726s 720w 723w υs(SiC2) 696s 686m 703w 697w υs(SiC2) 666s 671m 688w 666w υs(SiOSi) 611m υs(SiOSi)

29

-400

-350

-300

-250

-200

-150

-100∆E

( k

Jmol

-1 )

Li+ Na+ K+

HF/3-21G HF/6-31G*

B3LYP/6-31G*

Figure S4.5 Counter poise corrected binding

energies of D5 with alkali metalcations

30

Reaction energy , reaction enthalpy, and proposed reaction pathway

For calculation of reaction enthalpy, a Born-Haber cycle was used (Scheme S4.1). The

calculated Lithium affinity of D5 was applied. Vaperation energy of D5 was estimated using ? Hvap =

22.8 + 5.6 x n (kJ.mol-1) ( n = number of D unit in Dn, D = Me2SiO; for D5, ? Hvap = 51 kJ.mol-1).11

Lattice energy values for Li[AlF] and LiD5[AlF] were estimated using Jenkins and Passmore’s

volume-based relationship.12 The binding procees of Li+ to D5 could be: in order to combine Li+,

puckered D5 (Cs, Figure S4. 1a and b) probably first adjusted to be planar (C5v, Figure S4. 1c and d)

so all methyl groups pointed outward to form a nucleophilic cavity for Li+ cation. Such geometry

change could not be too hard as the energy gap is rather small (5 - 8 kJ.mol-1). The cavity of D5 (C5v,

Figure S4. 1c and d) appears too big for Li+, thus the resulted planar complex cation LiD5+ (D5h,

Figure S4.2c and d) which has five-coordination mode is unstable thus acting as a transition state

on the way via another transition state (C2v, Figure S4.2e and f) in which the Li+ cation coordinates

to four oxygen atoms, and finally stabilizes as a minimum (Cs, Figure S 4.2g and h) in which the Li+

cation is slightly out of the (SiO)5 framework (0.332 Å at HF/6-31G* and 0.271 Å at B3LYP/6-

31G*). The energy of the reaction of Li[AlF] (s) with D5 (l) was estimated as - 207 (HF/6-31G*) or

-248 (B3LYP/6-31G*) kJ.mol-1 (? H298 = -198 and -239 kJ.mol-1, respectively) and thus the reaction

is exothermic.

31

B. HF/6-31G*; C. B3LYP/6-31G*

Li[AlF] (s) + D5 (l) LiD5[AlF] (s)

Li+ (g) + [AlF]- (g) + D5 (g) LiD5+ (g) + [Al]- (g)

(Cs, puckered)

Li+ (g) + D5 (g) (C5v, planar)

LiD5+ (g)

(Cs, Li+ out of plane)

(C2v, Li+ in the plane)

∆ ΕL ∆ ΕL

∆ Ε

+50+368 -322

B. 5C. 8 B. -2

C. -1

B. -306C. -351

B. -303C. -344

B. -207C. -247

Li D5+ (g)

(D5h, planar)

B. -305C. -348

B. -0.4C. -3

Scheme S4.1 Born-Haber cycle of the reaction of Li[AlF] with D5

32

(2) D6 and its alkali metal complexes

Table S4. 6 Total Energies, Binding Energies, and Binding Enthalpies of free and Complexed D6

a. Total energy in au; b. Energy and enthalpy in kJ.mol-1.

Sym method energya ?Eb ? H298b D6 C2 HF/3-21G -2644.91097

C2 HF/6-31G* -2658.87802 C2 B3LYP/6-31G* -2668.03743

D6h HF/3-21G -2644.90235 D6h HF/6-31G* -2658.87215 D6h B3LYP/6-31G* -2668.02929

LiD6+

C2 HF/3-21G -2652.44599 -389 -377 C2 HF/6-31G* -2666.24441 -344 -332 C2 B3LYP/6-31G* -2675.46740 -382 -370

D6h HF/3-21G -2652.22232 -357 D6h HF/6-31G* -2666.22718 -299 D6h B3LYP/6-31G* -2675.44668 -327

NaD6+

D6h HF/3-21G -2805.69135 -280 -271 D6h HF/6-31G* -2820.63391 -254 -245

D6h B3LYP/6-31G* -2830.22607 -284 -275 KD6

+ C6v HF/3-21G -3240.96631 -149 -140 C6v HF/6-31G* -3257.89351 -160 -151

C6v B3LYP/6-31G* -3267.82098 -186 -177 D6h HF/3-21G -3240.96631 -129

D6h HF/6-31G* -3257.89352 -114 D6h B3LYP/6-31G* -3267.82098 -154

RbD6+

C6v HF/3-21G -5569.53327 -122 -113 D6h HF/3-21G -5569.49622 -25 -16

33

Table S4.7 Summary of geometries parameters for D6 and its complexes with the alkali metals

a. σ referring to the plane defined by all oxygen atoms, dM-σ referring to the distance from metal atom (M) to the σ plane; b. H. Oberhammer, W. Zeil, G. Fogarasi, Journal of molecular structure, 1973, 18, 309; *noncoordinating O.

Sym method M-O Si-O Si-O* Si-C Si-O*-Si Si-O-Si dM-sa

D6 C2 HF/3-21G 1.647 1.893 176.4 C2 HF/6-31G* 1.632 1.874 158.2 C2 B3LYP/6-31G* 1.656 1.876 153.7

D6h HF/3-21G 1.645 1.895 171.5 D6h HF/6-31G* 1.625 1.876 169.4 D6h B3LYP/6-31G* 1.643 1.878 169.0 D3d ?b gas electron 1.622(1) 1.846(1) 149.6(1)

LiD6+

C2 HF/3-21G 2.002 1.688 1.654 1.884 149.5 147.2 C2 HF/6-31G* 2.042 1.671 1.631 1.865 151.5 143.5

C2 B3LYP/6-31G* 2.023 1.692 1.651 1.867 149.3 142.9 D6h HF/3-21G 2.555 1.659 1.886 160.7 0

D6h HF/6-31G* 2.510 1.645 1.866 159.4 0 D6h B3LYP/6-31G* 2.523 1.663 1.867 158.1 0 C1 X-ray ([AlF]-) 2.058 1.658 1.617 1.837 151.1 141.1 0.423 C1 X-ray ([AlphF]-) 2.078 1.655 1.620 1.836 146.5 141.7 0.131

NaD6+

D6h HF/3-21G 2.580 1.663 1.886 161.7 0 D6h HF/6-31G* 2.539 1.647 1.866 160.8 0

D6h B3LYP/6-31G* 2.555 1.664 1.868 160.3 0 KD6

+ C6v HF/3-21G 2.848 1.664 1.886 160.1 1.101 C6v HF/6-31G* 2.880 1.649 1.867 155.2 1.267

C6v B3LYP/6-31G* 2.874 1.669 1.868 153.2 1.218 D6h HF/3-21G 2.638 1.671 1.884 164.2 0

D6h HF/6-31G* 2.621 1.658 1.866 164.4 0 D6h B3LYP/6-31G* 2.638 1.675 1.867 163.9 0

RbD6+

C6v HF/3-21G 3.065 1.663 1.887 159.5 1.553 D6h HF/3-21G 2.683 1.678 1.884 166.1 0

34

Uncomplexed D6 ring Two conformations of D6 were calculated at HF/3-21G, HF/6-31G* and B3LYP/6-31G* levels.

The C2 geometry has the lowest energy and the (SiO)6 framework is puckered with some methyl

groups pointing inward (Figure S4a and b). The D6h geometry has the highest symmetry with all

oxygen and silicon atoms in a plane, forming a cavity for a guest (Figure S4c and d). The D6h

geometry is a saddle point as could be expected. The Si-O bond length of C2 form calculated at

HF/6-31G* level is very close to experimental structures (gas phase electron diffraction). The SiOSi

angle at B3LYP/6-31G* level is in good agreement with experimental result while the Si-O bond

length is longer (Table S4.7). C2 geometry is 15 - 22 kJ.mol-1 more stable than the D6h form.

35

B. 156.7C. 156.4

A. 176.9

B. 160.9A. 1.647 A. 176.8

C. 1.652B. 1.633

A. 1.647

C. 156.6

B. 1.631C. 1.652

a. D6 (C2, top-view) A. HF/3-21G b. D6 (C2, side-view) B. HF/6-31G* C. B3LYP/6-31G*

A. 22 kJ.mol-1

B. 15 kJ.mol-1

C. 21 kJ.mol-1

C. 1.643

A. 171.5

C. 169.0B. 169.4

B. 1.625A. 1.645

c. D6 (D6h, top-view) d. D6 (D6h, side-view) A. HF/3-21G B. HF/6-31G* C. B3LYP/6-31G* Figure S4.6 D6 geometries

36

LiD6+ and NaD6

+

A C2 structure was optimized at three levels (HF/3-21G, HF/6-31G*, B3LYp/6-31G*) (Figure

S4.7a). The geometry at B3LYP/6-31G* (Figure S4. 7b) was close to single crystal structure of

LiD6+ (Figure S4. 7e and f) in LiD6[AlF] in which the Li atom was coordinated to four oxygen

atoms and basically is coplanar with the (SiO)6 framework. While the (SiO)6 framework in the C2

geometry at HF level was twisted and deviated from a plane. The deviation degree became smaller

as basis set became bigger. However the geometry parameters at HF/6-31G* level are in good

agreement with that in solid state. We also optimized a D6h geometry at three levels in which Li

atom is at the centre of (SiO)6 framework and coordinated to six oxygen atoms (Figure S4. 7c and d).

The Li-O distance is much longer than that of C2 geometry. The D6h form is not a minimum as

confirmed by harmonic frequency calcualation which has one imaginary frequency at HF/3-21G

level, and is 32 - 55 kJ.mol-1 less stable than C2 geometry. Evidently the cavity formed in D6h

geometry is too big for Li+ cation. Therefore when the Li+ cation was replaced by Na+, calculation

gave a perfect D6h geometry which was characterized as a local minima corresponding to the

potential energy surface where the Na+ cation was located at the centre of (SiO)6 plane and equally

coordinated to six oxygen atoms (Figure S4. 7i and j).

To check how the CF3-H and CF3-Li contacts in LiD6[AlPhF] affected the geometry of LiD6+

cation, we ran the geometry optimization calculation for a hypothesis complex LiD6F (Figure S4.

7g). The initial geometry of LiD6F contains the planar LiD6+ cation and the F anion which connects

the Li atom. The resulted bent geometry confirmed the crystal structures of LiD6[AlPhF] (Figure S4.

7h), that is, the geometry of LiD6+ cation in LiD6[AlPhF] has been remarkably changed by strong F-

H and F-Li contacts.

37

B. 1.681C. 1.651

C. 1.683

A. 1.694

B. 151.8

C. 142.9

C. 1.683

A. 148.0

A. 148.0

C. 2.023

B. 106.6C. 106.7

B. 2.042

A. 1.654

C. 149.3

C. 1.651A. 2.002

Li

A. 1.683

C. 1.700

A. 2.002

A. 149.5

C. 1.700

C. 2.024

A. 110.5

B. 2.038

A. 1.682

B. 1.662

C. 142.9

B. 143.6

B. 1.661

B. 143.3

B. 1.631

A. 1.654

A. 1.694B. 1.680

B. 1.631

Li

a. LiD6+, C2, top-view b. LiD6

+, C2, side-view (B3LYP/6-31G*) A. HF/3-21G B. HF/6-31G* C. B3LYP/6-31G*

A. 32 kJ.mol-1

B. 45 kJ.mol-1

C. 55 kJ.mol-1

A. 1.659

A. 2.555

A. 1.659

A. 2.555

A. 160.7

B. 1.645

C. 158.7

C. 2.523

B. 159.4

B. 2.510

C. 1.663

c. LiD6

+, D6h, top-view d. LiD6+, D6h, side-view

A. HF/3-21G B. HF/6-31G* C. B3LYP/6-31G*

38

H3A

H5C

Li233

Li

e. LiD6

+ (with [AlF]-), exp. f. LiD6+ (with [AlF]-), exp.

2.279

2.274

1.667

Li

2.331

2.177

F

Si3

C3

O3Si4

O2Si2

F

g. LiD6F, HF/3-21G h. LiD6

+ (with [AlphF]-), exp.

39

A. 1.663

A. 2.580

A. 1.663

A. 2.580

Na

A. 161.7

B. 1.647

C. 160.3

C. 2.555

B. 160.8

B. 2.539

C. 1.664

Na

i. NaD6

+, D6h, top-view j. NaD6+, D6h, side-view

A. HF/3-21G B. HF/6-31G* C. B3LYP/6-31G*

Figure S4.7 LiD6+ and NaD6

+ geometries

40

KD6+ and RbD6

+ Two structures were calculated for each of the KD6

+ and RbD6+ complexes. One of these

structures corresponds to the D6h conformation with the metal cations residing at the centre of the

cavity (Figure S4. 8c, d; Figure S4. 9c, d). Since K+ and Rb+ cations are rather big, it seems unlikely

that these structures are stable. We therefore calculated for each complex a structure of a lower

symmetry. As a result, a stable C6v geometry for each of KD6+ and RbD6

+ was obtained with the

cations out of the (SiO)5 framework (Figure S4. 8a, b; Figure S4. 9a, b). The Rb atom was farther

away from the σ plane defined by six oxygen atoms than K atom. For KD6+, the C5v structure is 20 -

46 kJ.mol-1 more stable than the D6h geometry; for RbD6+, much more ( 97 kJ.mol-1).

41

B. 155.2C. 153.2

K

C. 1.669

A. 1.664

A. 160.1

B. 2.880C. 2.874

B. 1.649

A. 2.848

K

a. KD6

+, C6v, top-view b. KD6+, C6v, side-view

A. HF/3-21G; dNa-σ: 1.101 HF/3-21G; B. HF/6-31G*; 1.267 HF/6-31G*; C. B3LYP/6-31G* 1.218 B3LYP/6-31G*

A. 20 kJ.mol-1

B. 46 kJ.mol-1

C. 32 kJ.mol-1

K

B. 2.621

A. 1.671

C. 2.638

B. 1.658

A. 164.2

C. 163.9A. 2.638

C. 1.675

B. 164.4

c. KD6

+, D6h, top-view d. KD6+, D6h, side-view

A. HF/3-21G B. HF/6-31G* C. B3LYP/6-31G* Figure S4.8 KD6

+ geometries

42

159.5

1.663

Rb

3.065

Rb

a. RbD6

+, C6v, top-view b. RbD6+, C6v, side-view

HF/3-21G dNa-σ: 1.553 HF/3-21G;

97 kJ.mol-1

1.678

166.1

2.683

c. RbD6

+, D6h, top-view d. RbD6+, D6h, side-view

HF/3-21G Figure S4. 9 RbD6

+ Geometries

43

NBO charge analysis Similar to MD5

+, NBO analysis revealed polarization of the D6 ring by the cation (Table S4.8) in

the MD6+ complexes. Compared to neutral D6 ring, the oxygen and silyl charges in MD6

+ complexes

showed strong polarization of the electron density from the silyls toward the oxygen centers which

are involved in coordination, e.g. the charge on the silyl groups in the D6 structure is -1.275, similar

to -1.271 on uncoordinated O atoms but significantly less than -1.316 on coordinated oxygen atoms

in the LiD6+ complex. A transfer of 0.038 from each silyl group to the near coordinated oxygen atom

has been clearly reflected. Polarization is the strongest in LiD6+ complex and became weaker as

going down Group 1. A weak charge transfer from the metal to D6 ring is reflected in the atomic

charge (e.g. 0.932 at Li), 0.068e less than +1 charge of the Li+ cation. The charge transfer also

decreased as the metal cation became bigger.

Table S4.8 Atomic and fragment charges for D6 and

their alkali metal complexes (B3LYP/6-31G*)

D6 LiD6+ NaD6

+ KD6+

Sym C2 C2 D3h C6v q(M+) 0.932 0.962 0.978 q(O) -1.275 -1.271 q(O*) -1.316 -1.303 -1.297 q(SiMe2) 1.275 1.313 1.309 1.301 ? q(SiMe2) 0 0.038 0.034 0.026

44

Frequencies, binding energies and gas phase selectivity of alkli metal

with D6 ring Normal mode frequencies for D5 and LiD6

+ at calculated HF/3-21G level (scale factor 0.9085)

were in good agreement with that by FT-IR and FT-Raman spectroscopies (Table S4.9). Figure

S4.10 shows the counter poise (CP) corrected binding energies for MD6+ at the three levels (also see

Table S4.6). It has been well investigated that in polar solvents the selectivity of 18-crown-6 with

alkali metal cations are corresponding to cation-ring cavity size match principle, that is, the better

the match, the higher the selectivity. Thus one may expect that the selectivity sequence for D6 could

be: Na+ > Li+ as Na+ cation matches the cavity of D6 better than the Li+ cation (Figure S4.7a, i, and

j). However Figure S4.10 shows the gas-phase selectivity sequence for D6 is: Li+ > Na+ > K+ > Rb+,

distinct from above expectation. This unusual behavior of D6 with alkali metals in gas phase has

also been found during calculation of gas phase selectivity of alkli metal cations with 18-crown-6.

Table S4.9 Calculated and experimental vibrational bands of D6 and LiD6+

D6 D6 (HF/3-21G) LiD6+ in LiD6[AlF] LiD6+ in LiD6[AlPhF] LiD6

+ (HF/3-21G) Assign.

IR Raman IR Raman IR Raman IR Raman IR Raman

2972w 2975m 2954w 2967m 2965m 2974s 2963w 2967s 2953w 2962s υa(CH) 2907w 2914s 2893w 2897s 2916w 2915vs 2899w 2909vs 2894w 2897vs υs(CH) 1442m 1489w 1482w 1494w 1446m 1485w 1474w δa(CH3) 1412s 1408m 1412w 1466w 1466w δa(CH3) 1356s 1354m 1377w 1358m δa(CH3) 1275s 1261w 1276s 1266s δs(CH3) 1090s.br 1168s 1087s 1078s 1121s υas(SiOSi) 925w 914w 1001s 1005m 1032vs υas(SiOSi) 855m 878m 853s 855s 849s ρ (CH3) 825s 831w 822s 821s ρ (CH3) 808s ρ (CH3) 797w 794s 797w 795s υa(SiC2) 752m 745w 752m 745w υa(SiC2) 726s 724s 717w 724w υs(SiC2) 709m 703w 688m 706w υs(SiC2) 653m 669w 665w 658m 660w υs(SiOSi) 614w 619m 619w 600w υs(SiOSi)

45

-500

-400

-300

-200

-100

∆E (

kJm

ol-1 )

Li+ Na+ K+ Rb+

HF/3-21G HF/6-31G*

B3LYP/6-31G*

HF/3-21G (18-crown-6)

Figure S4. 10 Counter poise corrected binding energies of D6 and 18-crown-6 with alkali metals

46

Reaction energy, reaction enthalpy and proposed reaction pathway A Born-Haber cycle was used (Scheme S4.2). The calculated lithium affinity of D6 was

applied. Vaperation energy of D6 was estimated using ? Hvap = 22.8 + 5.6 x n (kJ.mol-1) (n = number

of D unit in Dn, D = Me2SiO; for D6, ? Hvap = 57kJ.mol-1). Lattice energy values for Li[AlF][ and

LiD6[AlF] were estimated using the method of Jenkins and Passmore’s volume-based relationship.

The binding procees of Li+ to D6 could be: electrostatically induced by Li+, puckered D6 (C2, Figure

S4.6a and b) probably first adjusted to be planar D6h, Figure S4.6c and d] so all methyl groups

pointed outward to form a nucleophilic cavity for Li+ cation. This geometry transfer (15 - 22 kJ.mol-

1) may be more difficult than that for D5 (5 – 8 kJ.mol-1). The resulted planar complex cation LiD6+

(D6h, Figure S4.7c and d) acts as an intermediate in which the Li+ cation is coplanar with (SiO)6

framework and equally coordinated to six oxygen atoms, then is stabilized to a minima (C2, Figure

S4.7a and b). The energy of the reaction of Li[AlF] (s) with D6 (l) was estimated as -237 – -282

kJ.mol-1 (? H298 = -225 - -267 kJ.mol-1, respectively) and thus the reaction is exothermic. The

enthalpy of this reaction is about 30 kJ.mol-1 more than that of the reaction of Li[AlF] (s) with D5 (l)

at HF/6-31G* and B3LYP/6-31G* levels, which may explain the reason why LiD5[AlF] dissociated

in liquid SO2 but LiD6[AlF] not.

47

Li[AlF] (s) + D6 (l) LiD6[AlF] (s)

Li+ (g) + [AlF]- (g) + D6 (g) LiD6+ (g) + [Al]- (g)

(C2, puckered)

Li+ (g) + D6 (g) (D6h, planar)

LiD6+ (g)

( D6h, planar )

∆ ΕL ∆ ΕL

∆ Ε

+57+368 -318

A. 22B. 15C. 21

A. -32B. -45C. -55

A. -379B. -314C. -349

A. -389B. -344C. -383

A. -282B. -237C. -276

( C2, planar )

A. HF/3-21G; B. HF/6-31G*; C. B3LYP/6-31G*

Scheme S4.2 Born-Haber cycle of the reaction of Li[AlF] with D6

48

References: 1. a) T. J. Barbarich; S. T. Handy; S. M. Miller; O. P. Anderson; P. A. Grieco; S. H. Strauss. Organometallics, 1996, 15, 3776; b) T. J. Barbarich; S. M. Miller; O. P. Anderson; ; S. H. Strauss. J. Mol. Catal. 1998, 128, 289; c) N. Kuhn; H. Lanfermann; P. Schmitz; Liebigs. Ann. Chem. 1987, 727. 2. SAINT 6.02, 1997-1999, Bruker AXS, Inc., Madison, Wisconsin, USA. 3. SADABS George Sheldrick, 1999, Bruker AXS, Inc., Madison,Wisconsin,USA. 4. SHELXTL 6.14, 2000-2003, Bruker AXS, Inc., Madison,Wisconsin,USA. 5. RLATT 2.72, 1999, Bruker AXS, Inc., Madison, Wisconsin, USA. 6. W. C. Marsh, N. L. Paddock, C. J. Stewart, J. Trotter, J. Chem. Soc. Chem. Commun. 1970, 1190. 7. Gaussian 03, Revision C.02, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomery, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C.Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick,

A. D. Rabuck, K.Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez, and J. A. Pople, Gaussian, Inc., Wallingford CT, 2004.

8 a) M. R. Churchill, C. H. Lake, S. H. Chao, O. T. J. Beachley, J. Chem. Soc. Chem. Commun. 1993, 1577; b) C. Eaborn, P. B. Hitchcock, K. Izod, J. D. Smith, Angew. Chem. Int. Ed. Engl. 1995, 34, 2679; Angew. Chemie. 1995, 23, 2936. 9 Glending, E. D.; Feller, D.; Thompson, M. A. J. Am. Chem. Soc. 1994, 116, 10657. 10 Hill, S. E.; Feller, D.; Glending, E. D. J. Phys. Chem. A.1998, 102, 3813. 11. D. F. Wilcock, J. Am. Chem. Soc. 1946, 68, 691. 12. H. D. B. Jenkins, H. K. Roobottom, J. Passmore, L. Glasser, Inorg. Chem. 1999, 38, 3609.