Halogen Bonding-Based âCatch and Releaseâ: Reversible Solid-State

Article pubs.acs.org/crystal

Halogen Bonding-Based “Catch and Release”: Reversible Solid-State Entrapment of Elemental Iodine with Monoalkylated DABCO Salts Anssi Peuronen,† Arto Valkonen,‡ Minna Kortelainen,‡ Kari Rissanen,‡ and Manu Lahtinen*,† †

Department of Chemistry, and ‡Nanoscience Center, University of Jyväskylä, P.O. Box 35, Jyväskylä, FI-40014 JY, Finland S Supporting Information *

ABSTRACT: The halogen bonding (XB) between elemental iodine (I 2 ) and neutral 1,4-diazabicyclo[2.2.2]octane (DABCO) and its monoalkylated PF6− salts was studied by X-ray crystallographic, thermoanalytical, and computational methods. DABCO was found to form both 1:1 and 1:2 complexes with I2 showing an exceptionally strong halogen bond (ΔEcp = −73.0 kJ/mol) with extremely short N···I distance (2.37 Å) in the 1:1 complex (1a). In the more favored 1:2 complex (1b), the XB interaction was found to be slightly weaker [ΔEcp = −64.4 kJ/mol and d(N···I) = 2.42 Å] as compared to 1a. The monoalkylated DABCO salts (2PF6−7PF6) form corresponding 1:1 XB complexes with I2 {[2···I2]PF6−([7···I2]PF6} similarly to the parent free base DABCO, but both X-ray diffraction and calculated (M05-2X/def2-TZVPP) geometrical parameters indicate that the XB interactions are somewhat weaker than with DABCO itself but can nonetheless be considered as moderately strong halogen bonds. The solid -state packing of the monoalkyl DABCO complexes is greatly affected by the length of the lipophilic hydrocarbon chain as the long-tail cations show increasing amphiphilic character. However, partly as a consequence of the amphiphilic nature of parent monoalkyl DABCO PF6− salts, their I2 complexes exhibit a reversible binding of I2 into their originally nonporous crystal lattices. This was verified by thermal analysis and X-ray powder diffraction studies of 2PF6−7PF6 and their corresponding I2 complexes. By varying the length of the alkyl chain, the release temperature of I2 can be tuned from 75 °C ([4···I2]PF6) to 100 °C ([7···I2]PF6). Furthermore, these highly stable (preservable for months in normal laboratory conditions) I2 complexes can be prepared with three different routes: by mixing in solution, by mechanochemical grinding of the components, and via gas-to-solid reaction (i.e., I2 vapor to solid PF6− salts).

■

INTRODUCTION Halogen bonding (XB) as a concept describes a noncovalent interaction between an electron-deficient halogen atom X (Lewis acid, halogen-bond donor) and an electron density-rich species D (Lewis base, halogen-bond acceptor) in a YX···D scheme where Y can be a halogen, carbon, nitrogen, etc. Hence, in great respect, halogen bonding resembles hydrogen bonding (HB) as both of these interactions involve donation of electrondeficient species to an electron-rich acceptor.1 Furthermore, like HB interactions, XB interactions are highly directional, and therefore their expected geometries in the solid state are very predictable, making halogen bonding an excellent tool in supramolecular chemistry.2 The origin of the directionality of the XB interaction arises from the electropositive area at the halogen atom also known as the “σ-hole”.3 The σ-hole is formed when a half-filled p-orbital of a halogen atom participates in covalent bond, thus making the electron density around the halogen atom anisotropic and the orbital lobe opposite to this bond electron deficient. The electron deficiency of the halogen atom and the donor capabilities of the electron-rich species dictate the strength of the XB interaction (depending on the exact definition, can exist roughly in the range of 1−180 kJ/mol).2 The strongest XB interactions are typically found between classical Lewis bases © 2012 American Chemical Society

and dihalogens, in addition to different polyiodides and -bromides. These are in fact the earliest examples of halogenbonded systems long before the establishment of the term “halogen bonding” to describe the weak interaction between electron-deficient halogen and electron donor.4 Later, it was discovered that also halogen species other than dihalogens, such as halocarbons, can function as acceptors of electronic charge providing that the halogen atom is situated near an electron-withdrawing group.5 In recent years, a range of new motifs, containing electropositive halogen atom(s) that can function as halogenbond donors, have been reported. These include, for example, iodoperfluoroalkanes, fluorinated iodobenzenes, iodoalkynes, N-halosuccimides, and N-halosaccharides.6 Whereas such XB donors are considered as rather exotic molecules, because of the challenges regarding their synthesis (halogen atoms must reside near electronegative or polarizable atoms or groups, as in CF3I or in I2, to make them capable in XB interactions), halogenbond acceptors are more accessible as they commonly consist of tertiary or aromatic nitrogen compounds, chalcogens, and Received: May 16, 2012 Revised: July 10, 2012 Published: July 11, 2012 4157

dx.doi.org/10.1021/cg300669t | Cryst. Growth Des. 2012, 12, 4157−4169

Crystal Growth & Design

Article

contributor to the bonding properties and behavior of these monocationic complexes.

particularly halides. Alternatively, almost any electronegative molecule can act as XB acceptor, a point proven by Arduengo et al. in one of the early papers describing complexes of Nheterocyclic carbenes.6b Therefore, because of the capability of halogen atoms acting as both Lewis base and Lewis acid due to the anisotropic charge distribution around the halogen, the concept of “halogen bonding” is convenient as it unambiguously explains the role of the halogen atom as XB donor (Lewis acid) and thus prevents confusion. Altogether, the recognition of XB interactions has already led to some important discoveries regarding its use in supramolecular chemistry and molecular recognition. For example, Metrangolo et al. have shown the capability of alkyldiammonium iodides to resolve mixtures of diiodoperfluoroalkanes.6a Furthermore, using the same quaternary ammonium iodides with a specific alkyl chain length, they were able to isolate the relatively rare I42− dianion.7 In these two studies, iodide anions, situated near the positive region of the quaternary ammonium centers, act as the XB acceptors. There are, however, only a few studies illustrating a cationic moiety that participates in the halogen-bonding interactions by acting as a XB acceptor.8 One of our research interests lies in the study of structural and thermal properties of quaternary ammonium salts, more specifically salts based on caged amines such as 1,4diazabicyclo[2.2.2]octane (DABCO). DABCO is an important diamine used, for example, as a base catalyst in organic synthesis9 and a ligand in coordination chemistry.10 Furthermore, its mono- and diquaternized salts are known to exhibit liquid crystalline phases and/or show exceptionally low melting points and are therefore categorized as room temperature ionic liquids (salts with melting points below 25 °C).11 Being a strong base, DABCO is an excellent electron-donating species for halogen bonding as is shown in reports describing its XB complexes with halocarbons.6c,12 However, to our knowledge, no investigations concerning halogen bonding with DABCO salts, which would include N···X interactions, have been carried out. Moreover, no crystal structures of I2 complexes of either neutral DABCO or its monoalkylated salts have been reported (certain spectroscopic studies of DABCO···X2 complexes, where X = Br/I, have been performed).13 Therefore, because a broad range of workable functionalities can be incorporated to the DABCO by monoquaternization, we were eager to investigate whether monoalkylated DABCO is a strong enough electron donor to engage in halogen bonding via its tertiary amine. The few reports describing solid-state structures of monocationic DABCO as a ligand in transition metal complexes support this assumption.14 However, no straightforward conclusion can be drawn from these data because the interaction strength between a metal cation and an amine can be expected to exceed that of a halogen bond. Nonetheless, we assumed that monocationic DABCO has reasonably high Lewis basicity and can thus function as halogen-bond acceptor particularly if a strong enough halogen-bond donor is employed. Furthermore, because the selective monoquaternization of the DABCO core is fairly straightforward, DABCO salts could be exploited in creating new types of supramolecular XB synthons. The XB interactions between iodine and XB acceptors are typically very strong, and therefore I2 seemed as a suitable candidate for this investigation. In this Article, we present the structures, thermal behavior, and theoretical bonding analysis of the first 1-alkyl-4-aza-1-azoniabicyclo[2.2.2]octane complexes with iodine. The second aspect of this study includes the investigation of amphiphilicity as a

■

EXPERIMENTAL SECTION

General. All reagents and solvents were purchased from SigmaAldrich and were used as received. The purity of products was confirmed by 1H NMR (see Supporting Information Figure S6) and elemental analysis (Vario EL III CHN). The monocationic 1-alkyl-4aza-1-azonia-bicyclo[2.2.2]octane bromides (2−7) were synthesized as described by Chiappe et al.11c The bromide salts were converted to hexafluorophosphates (2PF6−7PF6) by metathesis reaction in water (50 mL) using 10% molar excess of KPF6 salt. The resulting slurries were stirred 18 h at room temperature and were filtered and washed with water until a negative AgNO3 test was achieved after which they were washed once with diethyl ether and dried in vacuo. Single crystals of acceptable quality of compounds 2PF6−4PF6 and 7PF6 were obtained from water/acetone solutions. Synthesis. Bulk Samples. I2 complexes of the salts 2−7PF6 {i.e., [2···I2]PF6−[7···I2]PF6} were prepared by adding 2 mL of 0.05 M iodine solution (in chloroform) to 0.1 mmol of 2−7PF6 in approximately 2 mL of dichloromethane under constant mixing (Scheme 1). A precipitate formed immediately, which was separated

Scheme 1. Synthesis of Monocationic 1-Alkyl-4-aza-1-azoniabicyclo[2.2.2]octane Hexafluorophosphate Halogen-Bonded Complexes with I2 (n = 6,8,...,16)

by filtration, washed with a small amount of dichloromethane, and dried in air. The same method was used for 1b but with 0.2 mmol amounts of both DABCO and iodine. The crystallographic similarity of bulk materials as compared to the single crystals was confirmed by PXRD, whereas chemical composition was verified by elemental analysis (Table 1). Another batch of the same complexes was prepared by hand-grinding equimolar amounts of I2 and corresponding PF6− salt in a mortar for 5 min. These materials were also analyzed by PXRD (Supporting Information Figure S3). Single Crystals. 1a: Iodine (120 mg, 0.5 mmol) was dissolved in CHCl3 (15 mL) by gently heating in large test tube. Into this red solution was added 0.5 mmol of DABCO, and the solution turned cloudy. After a few minutes, a fine red solid deposited on the bottom

Table 1. Yields and Elemental Analysis of Synthesized I2 Complexes CHN (observed) 1b [2···I2] PF6 [3···I2] PF6 [4···I2] PF6 [5···I2] PF6 [6···I2] PF6 [7···I2] PF6 4158

CHN (calculated)

yield

N

C

H

N

C

H

mg

%

4.45 4.50

11.80 24.39

2.30 4.31

4.52 4.70

11.63 24.18

1.95 4.23

50 50

40 84

4.19

27.60

5.07

4.49

26.94

4.68

47

75

4.12

30.11

5.12

4.30

29.46

5.10

39

60

3.91

32.62

5.55

4.12

31.78

5.48

42

62

3.75

34.46

5.86

3.95

33.91

5.83

54

76

3.53

36.41

6.31

3.80

35.88

6.16

62

84



Article

Table 2. Crystallographic Data for Single-Crystal X-ray Measurements identification code empirical formula formula weight/g mol−1 temperature/K crystal size/mm3

2PF6 C12H25F6N2P 342.31 293 0.28 × 0.12 × 0.12 orthorhombic Pbca 15.163(3) 11.876(2) 18.546(4) 90 90 90 3339.7(12) 8 0.218 1440 3.87−25.00 15 927 2812 2812/35/227 1.016 0.0769 0.0640, 0.1535

3PF6 C14H29F6N2P 370.36 293 0.35 × 0.30 × 0.22 monoclinic P21/n 8.6260(5) 14.4220(7) 14.6980(7) 90 90.573(2) 90 1828.40(16) 4 0.205 784 7.34−21.74 4000 2068 2068/0/228 1.113 0.0341 0.0657, 0.1400

4PF6 C16H33F6N2P 398.41 293 0.35 × 0.30 × 0.30 orthorhombic Pnma 7.9020(3) 6.9120(4) 37.3530(19) 90 90 90 2040.17(18) 4 0.189 848 1.09−24.99 8028 1889 1889/0/139 1.071 0.042 0.0730, 0.1982

7PF6 C22H45F6N2P 482.57 293 0.30 × 0.20 × 0.20 orthorhombic Pnma 6.9392(3) 7.9029(4) 47.448(2) 90 90 90 2602.0(2) 4 0.16 1040 0.86−24.99 12 987 2408 2408/0/176 1.117 0.0398 0.0652, 0.1737

1a C6H12I2N2 365.98 123 0.17 × 0.10 × 0.08 monoclinic P21/c 9.4933(3) 8.1792(2) 13.4645(4) 90 109.156(2) 90 987.60(5) 4 6.309 672 2.96−25.25 5717 1785 1785/0/91 1.079 0.0338 0.0237, 0.0460

1b C6H12I4N2 619.78 123 0.31 × 0.26 × 0.03 monoclinic Cm 13.4915(5) 39.8655(17) 6.4622(2) 90 107.925(2) 90 3307.0(2) 10 9.385 2740 1.67−25.00 8617 5655 5655/73/168 1.044 0.0317 0.0424, 0.0912

crystal system space group a/Å b/Å c/Å α/deg β/deg γ/deg V/Å3 Z μ/mm−1 F(000) θ/deg reflns collected independent reflns data/restraints/parameters GOF R(int) final R indices [I > 2σ(I)], R1/wR2 R indices (all data), R1/wR2 0.1427, 0.1961 0.1026, 0.1660 0.1146, 0.2375 0.0976, 0.2014 0.0291, 0.0474 0.0627, 0.1004 (largest diff. peak/hole)/e 0.157/−0.148 0.236/−0.188 0.329/−0.353 0.377/−0.372 0.489/−0.538 1.146/−1.291 Å−3 identification code [3···I2]PF6 [3···I2]PF6-RT [4···I2]PF6 [5···I2]PF6 [6···I2]PF6 empirical formula C14H29F6I2N2P C14H29F6I2N2P C16H33F6I2N2P C18H37F6I2N2P C20H41F6I2N2P formula weight/g mol−1 624.16 624.16 652.21 680.27 708.32 temperature/K 123 293 123 123 123 crystal size/mm3 0.20 × 0.12 × 0.06 0.20 × 0.15 × 0.15 0.30 × 0.28 × 0.12 0.20 × 0.20 × 0.08 0.12 × 0.08 × 0.08 crystal system monoclinic monoclinic triclinic triclinic triclinic space group P21/c P21/m P1̅ P1̅ P1̅ a/Å 9.9339(6) 9.996(5) 7.2800(4) 7.2650(2) 7.3000(2) b/Å 8.5824(6) 8.762(5) 7.7750(5) 7.7670(3) 7.7590(3) c/Å 26.3690(15) 13.644(5) 20.8670(11) 22.4580(6) 23.9360(9) α/deg 90 90 99.297(3) 80.302(2) 97.593(2) β/deg 104.980(4) 108.112(5) 95.362(4) 89.118(2) 93.512(2) γ/deg 90 90 91.801(4) 88.207(2) 91.793(2) V/Å3 2171.7(2) 1135.8(10) 1159.20(12) 1248.46(7) 1340.28(8) Z 4 2 2 2 2 μ/mm−1 3.02 2.887 2.833 2.635 2.458 F(000) 1208 604 636 668 700 θ/deg 2.30−25.00 2.81−25.00 0.99−25.00 1.84−25.00 1.72−25.00 reflns collected 18 440 15 714 11 500 14 382 16 006 independent reflns 3818 2127 4070 4405 4730 data/restraints/parameters 3818/0/218 2127/85/163 4070/0/245 4405/0/263 4730/0/281 GOF 1.047 1.023 1.156 1.186 1.075 R(int) 0.1399 0.058 0.0564 0.0405 0.0501 final R indices [I > 2σ(I)], R1/ 0.0679, 0.0988 0.0488, 0.1123 0.0489, 0.1321 0.0322, 0.0642 0.0387, 0.0753 wR2 R indices (all data), R1/wR2 0.1632, 0.1259 0.0916, 0.1334 0.0607, 0.1485 0.0422, 0.0778 0.0548, 0.0917 (largest diff. peak/hole)/e Å−3 0.670/−0.884 0.741/−0.734 1.306/−1.546 0.522/−0.589 0.596/−0.699

[2···I2]PF6 C12H25F6I2N2P 596.11 123 0.20 × 0.14 × 0.08 orthorhombic Pnma 17.5640(7) 8.4479(4) 13.0119(5) 90 90 90 1930.69(14) 4 3.392 1144 3.10−25.00 16 934 1801 1801/0/136 1.179 0.0749 0.0488, 0.1027 0.0543, 0.1048 0.968/−1.092 [7···I2]PF6 C22H45F6I2N2P 736.37 123 0.40 × 0.30 × 0.08 triclinic P1̅ 7.2822(7) 7.7561(7) 25.708(3) 95.617(3) 97.403(3) 91.763(4) 1431.6(2) Å3 2 2.305 732 2.41−25.00 12 013 4810 4810/0/299 1.1 0.0541 0.0527, 0.1428 0.0572, 0.1469 2.310/−1.374

1.4 mL of CH2Cl2 while avoiding the mixing of two layers as much as possible. Iodine solution (0.5 mL) in CH3CN (43 mg in 3 mL) was also carefully placed uppermost. After a couple of days, the red leaf-like crystals were obtained and used in single-crystal analysis. [2···I2]PF6−[7···I2]PF6: Single-crystal samples were prepared according to the same procedure as described for the bulk samples

of the tube. This solid, mostly containing (I2)···DABCO···(I2), was removed by filtration, and the remaining solution was left to stand in a sealed tube. After a week, red block-like crystals were obtained and used in single-crystal analysis. 1b: DABCO solution (0.5 mL) in CHCl3 (20 mg in 3 mL) was added into a test tube. On top of that solution was carefully inserted 4159



Article

Figure 1. Asymmetric units of crystal structures of (a) 1a and (b) 1b. Thermal ellipsoids are at the 50% probability level. but using dilute CH2Cl2/CHCl3 (1:4) solutions. Crystals emerged from these solutions after few days by slow evaporation of the solvent. X-ray Analysis. Single-crystal X-ray diffraction measurements were carried out with a Bruker-Nonius KappaCCD diffractometer with Apex II detector using Mo Kα (λ = 0.71069 Å) radiation. Nylon loops were used in low temperature measurements together with perfluoropolyether oil, whereas crystals measured at room temperature were either glued on top of glass capillary or mounted on a MiTeGen MicroMount. Data collection was done with COLLECT15 software, and processing was carried out using DENZO-SMN.16 An absorption correction was applied on all data using SADABS.17 Structures were solved using direct methods (either SIR2002,18 SIR2004,19 or SHELXS-9720) and refined on F2 by full-matrix least-squares techniques (SHELXL-9720) with WinGX software.21 Crystallographic data are presented in Table 2. The X-ray powder diffraction measurements were made with a PANalytical X́ Pert PRO diffractometer using a Johansson monochromator to produce pure Cu Kα1 radiation (1.5406 Å; 45 kV, 35 mA). The data were collected by an X́ Celerator detector with a 2θ range of 2−70° using a step size of 0.017°, counting times of 10−200 s per step (sample dependent), and rotating sample holder. Lightly hand-ground powder samples were prepared on a silicon-made zerobackground holder using petroleum jelly as an adhesive in case of ambient conditions. For variable-temperature measurements, an Anton Paar TTK450 temperature-controlled nonambient-temperature chamber with automated sample-stage height-controller, vacuum pump system, and liquid N2 cooling unit was used. A sample was prepared in a sample cavity, placed into the chamber, and kept under vacuum (0.1 mbar) during the measurements. Diffraction patterns were recorded at desired temperatures of 25, 60, 90, 120, and 140 °C and finally again at 25 °C. The program X́ pert HighScore Plus v. 2.2d was used for data analysis, and the simulated PXRD patterns were generated by the program Mercury.22 For [3···I2]PF6, the comparison was done against simulated pattern obtained from the room-temperature single-crystal

X-ray data collection. Other patterns were simulated from the lowtemperature (123 K) measurements. Thermal Analysis. TG measurements were carried out using Perkin-Elmer STA 6000. Samples (4−8 mg) were placed in an open platinum pan and were heated under air atmosphere using a constant heating rate of 5 °C min−1 in the temperature range 25−650 °C. DSC measurements for salts 2−7PF6 were carried out using Perking Elmer Pyris 1 DSC. Samples (3−6 mg) were measured using a closed pan with small capillary holes in the cover by heating (10 °C min−1) under nitrogen atmosphere starting at 25 °C until 200 °C and were then cooled (5 °C min−1) to −40 °C. Heating−cooling cycles were repeated for each sample, and second heating scans were found virtually identical to the first scan. The reported temperatures of both DSC and TG measurements are onset temperatures of observed events calculated using standard procedures. For TG- and DSC-curves, see Supporting Information Figures S4 and S5, respectively. Numerical data derived from DSC measurements are available in Supporting Information Table S2. Computational Details. All theoretical calculations were carried out using Gaussian 0923 and Firefly QC package,24 which is partially based on the GAMESS (US)25 source code. Because of the huge size of monocationic salts studied here, theoretical calculations involving their I2 complexes could not be done using time-consuming MP2 methods. Therefore, we chose to carry out the analyses only to the salts with less than 10 carbon atoms in their alkyl chains {[2···I2]PF6− [4···I2]PF6} using density functional theory. M05-2X functional with def2-TZVPP basis set (we also experimented with basis sets including added diffuse functions, such as def2-TZVPD, but witnessed only minor improvements in either geometries or interaction energies) seemed to perform adequately well with both neutral and monocationic DABCO···I2 complexes and was thus used in all subsequent calculations. Hence, geometries of 1a, 1b, and [2···I2]PF6− [4···I2]PF6 were optimized at the M05-2X/def2-TZVPP level and confirmed as minima by vibrational analysis. These geometries (Supporting Information Table S3) were used in natural bonding 4160



Article

Table 3. Experimental (Single-Crystal X-ray) and Theoretical N···I and I−I Distances (Å)a calculatedc

experimental 1a 1b [2···I2]PF6 [3···I2]PF6 [4···I2]PF6 [5···I2]PF6 [6···I2]PF6 [7···I2]PF6

N···I

I−I

N···I

I−I

2.366(3) 2.4213b 2.4790(78) 2.5185(90) 2.5321(59) 2.5429(38) 2.5449(50) 2.5425(49)

2.8536(4) 2.8305b 2.8009(10) 2.7791(11) 2.7682(6) 2.7688(4) 2.7680(6) 2.7670(5)

2.5001 (2.424) 2.5562 (2.469) 2.6687 2.6699 2.6682

2.7506 (2.782) 2.7269 (2.759) 2.7010 2.7010 2.7012

a Standard deviations are presented for experimental data only. bN···I and I−I distances are mean values calculated from the three complexes in the asymmetric unit (all distinct distances are reported in Supporting Information Table S1). cGeometry optimizations are carried out for 1a, 1b, and [2···I2]PF6−[4···I2]PF6 with M05-2X/def2-TZVPP method (MP2/aug-cc-pVTZ values in parentheses).

Figure 2. I2···I2 interactions in crystal structures of (a) 1a and (b) 1b. For 1b, the I2···I2 distances are reported as mean values deduced from the three distinct complexes found in the asymmetric unit of the crystal structure. orbital (NBO)26 and natural energy decomposition (NEDA)27 analyses (carried out with Firefly using PBE0 density functional with def2-TZVPP basis sets) and in calculations of interaction energies between the fragments. In addition to M05-2X, counterpoise corrected interaction energies (Boys−Bernardi formulation28) were also calculated at the MP2/aug-cc-pVTZ (aug-cc-pVTZ-PP for iodine) level as single point calculations with Firefly MP2 code using M05-2X/ def-TZVPP geometries. All calculations were carried out using effective core potential for iodine atoms. Basis sets were obtained from EMSL basis set library.29

crystallographic data, DABCO binds to iodine very tightly; 1a shows a remarkably short N···I distance of 2.37 Å, making it only ∼0.29 Å longer that the N−I covalent radius. This translates to 33% shortening of the sum of van der Waals radii of respective atoms (henceforth, normalized vdW contact; Nc = 0.67, vdW radii of 1.55 and 1.98 have been used for N and I, respectively). Indeed, to our knowledge, this is one of the shortest N···I distances among similar types of N···I 2 complexes.31,32 Introducing a second iodine molecule to DABCO (1b) increases the N−I distance by 0.06 Å (N···I distance of 1a as compared to the averaged N···I distance of 1b), which indicates that the electron donor capabilities of the nitrogen in the 1:1 complex are affected but not significantly diminished by the addition of a second I2 to the DABCO diamine. In addition to N···I distance, the strength of the halogen bond can be evaluated from the lengthening of the Y− X bond, which occurs when electron acceptor (XB donor) interacts with a strong electron donor (XB acceptor). This phenomenon is caused by the partial occupation of the σ* antibonding orbital of Y−X, which weakens the Y−X bond.32 In both of these complexes, the I−I bond is clearly lengthened as I−I distances in 1a and 1b are 2.85 and 2.83 Å (av), respectively (cf., the solid-state structure of iodine where d(I−I) = 2.715 Å33). These bonding parameters found in crystal structures of 1a and 1b further suggest that the N···I2 XB interaction strengths in both 1:1 and 1:2 complexes of DABCO and I2 are fairly similar. In addition to halogen-bonding interactions between iodine and nitrogen, both 1a and 1b exhibit I2···I2 contacts with average distances of 3.82 Å (Nc =

■

RESULTS AND DISCUSSION Single-Crystal Studies. To evaluate the bonding capabilities of monoquaternized DABCO, investigation of halogenbonded I2 complexes of its neutral form was first carried out. Hence, we start by describing the results of our complexation attempts with neutral DABCO and I2. These attempts afforded crystals from 1:1 and 1:2 complexes of DABCO and I2, 1a, and 1b, respectively. The crystals were subjected to single-crystal Xray diffraction analyses, and their crystal structures were successfully determined. Whereas 1a crystallizes with one DABCO···I2 complex in the asymmetric unit, the asymmetric unit of 1b contains three distinct 1:2 (I2)···DABCO···(I2) units (Figure 1). These pairs are structurally very similar but differ slightly by their bonding parameters as compared to one another (Table 3). The DABCO core has a very shallow twisting potential,30 and therefore its backbone is easily distorted depending on the group interacting with its tertiary amine groups. In both 1a and 1b, the DABCO cage is C3symmetric, and the corresponding averaged N−C−C−N torsion angles are 3.5° for 1a and 1.5° for 1b. According to 4161



Article

Figure 3. Monocationic complexes (a) [2···I2]PF6 and (b) [3···I2]PF6. Disordered CH2- and CH3-groups are omitted from the figure.

ambient lab conditions for months without any detectable loss of iodine. This is not necessarily the case for many known I2 complexes with neutral XB acceptors.34 Complex [3···I2]PF6 was obtained as a small batch of very thin yellow plates, which were sometimes accompanied by an unknown dark purple/brown microcrystalline powder in our crystallization experiments. In many ways, the crystal structure of [3···I2]PF6 resembles that of [2···I2]PF6 (Figure 3b). The anions lie in the vicinity of the quaternary ammonium center, and, in addition to N···I2 XB bonding, no other interactions, in which I2 would participate, are observed. Furthermore, the alkyl tail is disordered over the crystallographic mirror plane. From the N···I and I−I distances, 2.52 Å (Nc = 0.71) and 2.78 Å, respectively, can be deduced that the halogen bond strength is perhaps slightly weakened as compared to [2···I2]PF6 as the N···I distance is increased by only 0.04 Å and the I−I bond is consequently shortened by 0.02 Å. The deviations are nonetheless rather small and can be due to packing effects. During PXRD studies (see below), it was discovered that [3···I2]PF6 undergoes a solid−solid phase transition upon cooling. Therefore, its structure was determined also at room temperature. This yielded structure belonging to monoclinic space group P21/m instead of P21/c with one-half the cell volume as compared to the data at 123 K (see Supporting Information Figure S2 for details). In contrast to the PF6− salts of 2 and 3, salts 4PF6−7PF6 have considerable amphiphilic character due to their long alkyl chains. This has a significant effect on their structural and thermal properties. For instance, amphiphilic compounds often self-assemble to form layers by separating the hydrophilic and hydrophobic parts of the molecule. Upon heating, the high structural order is gradually lost, leading to complex thermal behavior. In this group of PF6− salts, the effects of the increasing amphiphilicity are clearly seen when the n-alkyl chain length is increased from hexyl and octyl to decyl, at which point the packing of the cations changes dramatically. Contrary to 2PF6 and 3PF6, where the solid-state packing seems to be

0.96) and 3.68 Å (Nc = 0.93) for 1a and 1b, respectively (Figure 2). Encouraged by the observations made from the two solidstate structures obtained with neutral DABCO and I2, crystallization experiments were made using monoalkylated DABCOs (2−7). These salts were first converted from bromides to hexafluorophosphates, PF6− (Scheme 1), to prevent formation of (poly)halides such as [BrI2]− or I3− that would hamper the inspection of N···I2 interactions. Even so, when using strongly polar and/or protic solvents, we occasionally obtained iodide and triiodide species possibly due to reaction between I2 and PF6−. Crystallization attempts from mixtures of chloroform and dichloromethane solutions afforded crystals of 1-alkyl-4-aza-1-azonia-bicyclo[2.2.2]octane (alkyl = CnH2n+1, where n = 6, 8, 10, 12, and 16) hexafluorophosphates with I2 (henceforth, [2···I2]PF 6− [7···I2]PF6). Their structures were successfully solved with single-crystal X-ray diffraction, and their refined bonding parameters are shown in Table 3. Complex [2···I2]PF6 crystallized as large yellow plates. The crystal structure exhibits halogen-bonding interaction between the tertiary amine of the monocationic DABCO and I2 (Figure 3a). The alkyl tail of the cation is disordered over the crystallographic mirror plane on which both the ion pair and the I2 reside. Crystal packing is dominated by anion−cation interactions, and consequently no interactions other than four H···I vdW contacts [d(H−I) = 3.11 and 3.08 Å to calculated symmetry related hydrogens] are observed between the I2 and cations or anions. Thus, the bonding is dominated by N···I interactions, and the N···I bonding scheme is in this respect comparable to 1a and 1b. The N···I and I−I distances in [2···I2]PF6 are 2.48 Å (Nc = 0.70) and 2.80 Å, respectively, which correspond to 0.11 Å lengthening and 0.05 Å shortening of the respective distances as compared to 1a. Hence, it can be deduced that the halogen-bonding interactions are slightly weakened when DABCO is monoalkylated. Nevertheless, crystals of [2···I2]PF6 can be stored at room temperature in 4162



Article

Figure 4. Solid-state packing of (a) 2PF6, (b) 3PF6, (c) 4PF6, and (d) 7PF6. Disordered atoms and all hydrogens are omitted from the pictures. Color codes (henceforth): cation head groups (red), anions (blue), and alkyl tails (gray).

alkyl chains interdigitate, causing them to separate from the cationic head groups that reside on their own layer where the halogen bonding also occurs. As a result, the iodine molecules lie as layers between the head groups and form sheets of only one I2 molecule in thickness. This iodine layer resembles that found in crystalline iodine, but because the packing is dominated by N···I XB interactions, iodine molecules orientate parallel to one another, and no I2···I2 vdW contacts can be observed. The halogen-bonding parameters closely comparable those observed for [3···I2]PF6, although a minor lengthening of the N···I distances and subsequent contraction of I−I distances is observed (Table 3). Most likely, this behavior in [4···I2]PF6− [7···I2]PF6 originates from the interdigitated packing that is caused by steric effects, which in turn force the rest of the cation to adapt its geometry and interactions accordingly. The lengthening of the n-alkyl chain itself does not have any direct effect on the strength of N···I2 interactions (see theoretical analysis below for a more detailed discussion). On the other hand, vdW contacts {d(F···I) ≈ 3.41−3.44 Å in [4···I2]PF6− [7···I2]PF6} between I2 and PF6− may have a very minor contribution to the slightly longer N···I distances (Nc = 0.72) in [4···I2]PF6−[7···I2]PF6. The overall bonding scheme in [4···I2]PF6−[7···I2]PF6 including the PF6−···I2 contacts is shown in Figure 7. Characterization of Bulk Materials. Many amphiphilic salts exhibit interesting and diverse thermal behavior. Furthermore, as iodine is easily sublimated already at low temperatures, thermal degradation paths of these complexes may reveal information regarding the halogen-bonding behavior, which is not necessarily deducible from the crystallographic data. Therefore, bulk samples of salts 2PF6− 7PF6 and subsequent I2 complexes were examined by means of powder X-ray diffraction (PXRD), thermogravimetry (TG) and differential scanning calorimetry (DSC). Despite our efforts, we were unable to obtain sufficient quantities of neat 1a because equilibrium between the 1:1 and 1:2 complexes of DABCO and I2 was always observed regardless of attempts to use different molar ratios or solvents. However, synthesis of 1b and [2···I2]PF6−[7···I2]PF6 carried out in CH2Cl2/CHCl3 solutions afforded bulk materials crystallographically equivalent to corresponding single-crystal structures confirmed by PXRD

controlled by the most favorable packing of the ion pair, in 4− 7PF6, the self-assembly of the hydrocarbon chains dictates the solid-state ordering (Figure 4). More specifically, the ionic ammonium head groups and the interdigitated hydrophobic hydrocarbon chains separate into two distinct layers, whereas PF6− anions reside near the ammonium centers. Interestingly, all four amphiphilic salts 4PF6−7PF6 exhibit identical packing and are crystallographically different predominantly by the length of the unit cell c-axis corresponding to the lengthening of the alkyl chain (Figures 4 and 5). In other words, further lengthening of the alkyl chain beyond n-decyl does not seem to affect the packing scheme of the salts. On the basis of the crystallographic data of 4PF6−7PF6, it is not surprising that also their I2 complexes, [4···I2]PF6− [7···I2]PF6, have distinct solid-state ordering different from their nonamphiphilic counterparts, [2···I2]PF6 and [3···I2]PF6 (Figure 6). As in 4PF6−7PF6, in the ensuing I2 complexes the

Figure 5. Powder diffraction patterns for salts 2PF6−7PF6. Patterns for 4PF6−7PF6 show distinct similarities due to analogous layered solidstate packing induced by the hydrophobic alkyl chains. 4163



Article

Figure 6. (a) Packing of single iodine sheet and (b) sheets of iodine ordered between the amphiphilic layers in [4···I2]PF6 (all [4···I2]PF6− [7···I2]PF6 have almost identical solid-state packing). Parts (c) and (d) show the ordering of I2 into the crystal lattices of [2···I2] and [3···I2], respectively.

Figure 7. [5···I2]PF6 as an example in representing the interactions of I2 with cation and anion in [4···I2]PF6−[7···I2]PF6.

measurement. The thermal decomposition and transition temperatures of 1b and [2···I2]PF6−[7···I2]PF6 are presented in Table 4. TG curve of 1b shows the thermal degradation of the sample starting at 139 °C and occurring in three successive steps (Figure 8). The first major step includes removal of 1 mol of I2. The beginning of the next two steps coexists with the former and stems from the concurrent degradation of DABCO and the Table 4. Thermoanalytical Results for [2···I2]PF6−[7···I2]PF6 Salts T(I2)/°C 1b [2···I2]PF6 [2···I2]PF6c [3···I2]PF6 [4···I2]PF6 [5···I2]PF6 [6···I2]PF6 [7···I2]PF6

TDSC/°C 121.7

100.0 100.2 74.7 94.1 91.6 96.7

98.4 69.9 86.4 93.9 104.1

Tdec/°C

Δwt %

Δwt % (I2 calc)

139.5a 109.9a 311.5 318.2 319.1 317.1 317.2 269.1

48.6 39.0 42.4 38.4 39.5 35.9 32.8 33.3

40.9b 42.6 42.6 40.7 38.9 37.3 35.8 34.4

Figure 8. TG curves of (a) 1b, (b) [2···I2]PF6, and (c) [5···I2]PF6 illustrating the difference in thermal behavior of neutral, nonamphiphilic, and amphiphilic I2 complexes, respectively, under dynamic heating conditions. In (d), a 60 min isothermal step at 110 °C is incorporated into the measurement of [2···I2]PF6, leading to full recoverability of the neat salt.

a

I2 released along with sample decomposition. bCorresponds to 1 mol of I2; the second mole releases concurrently with decomposition. c Isothermal measurement. Temperature T(I2) corresponds to the loss of I2, whereas at Tdec the thermal degradation of the sample initiates. TDSC is the temperature where the parent salts (2PF6−7PF6) undergo a phase transition from solid to a soft-solid phase (see further details in Supporting Information Table S2).

elimination of remaining iodine (the endo- and exothermic transitions on DTA signal suggest the existence of short-lived intermediate products while decomposing). The nonamphiphilic salt [2···I2]PF6 acts in a somewhat similar fashion under 4164



Article

components. In the case of 3PF6−7PF6, the PF6− salt undergoes a transition to a soft-solid phase (visual inspection) at which point iodine is vaporized leaving the parent PF6− salt to its original composition. However, if the TG measurement of [2···I2]PF6 is made with a 60 min isothermal step set at the temperature corresponding to the first solid-phase transition observed for 2PF6 salt, a quantitative loss of I2 is observed, and sequel heating of the sample reveals thermal behavior analogous to that of neat 2PF6. This feature is most likely caused by a combined effect of the phase transition process of the cationic part (reorientation of alkyl chains) and ionic interactions prevailing in the crystal lattice. Thereby showing a similar chain of events reported recently for series of bis(trimethylammonium) alkane diiodide salts but due to the phase transition in the amphiphilic complexes [4···I2]PF6− [7···I2]PF6, release of iodine occurs more abruptly in contrast to the perfluorinated diiodoalkanes.6a To investigate the events occurring in the heating process more closely, and to ensure that the salts stay intact after the loss of I2, a variable-temperature PXRD measurement of [5···I2]PF6 was carried out and is presented here as an example. In this measurement, the diffraction pattern of the bulk sample was first measured at room temperature. Next, the temperature was raised, and four consecutive patterns were recorded at desired temperatures between 60 and 140 °C, at which point the yellow sample turned colorless. The last diffraction pattern was collected from the sample cooled back to room temperature. As anticipated in the basis of TG measurements, the last recorded PXRD pattern is a perfect match with bulk salt, 5PF6 (Figure 10), thus demonstrating that the crystalline PF6− salt is indeed recovered in the heating−cooling process. An interesting feature was stumbled upon while conducting solid-state reactions between the PF6− salts and I2, further enriching the halogen-bonding abilities of these amphiphilic systems. It is well-known that both nonporous neutral amines and ammonium salts can capture I2 directly from the gaseous phase or via solid-state reaction.6a,34 Therefore, we attempted to obtain complexes of [2PF6···I2]−[7PF6···I2] also via a mechanochemical route. Light hand-grinding of equimolar amounts of 3PF6−7PF6 and I2 in a mortar affords a yellow powder crystallographically identical (confirmed by PXRD) to single-crystal structures of corresponding complexes (Figure 10, Supporting Information Figure S3). Interestingly, the described

dynamic heating conditions as degradation starts at 110 °C and occurs gradually in a fairly broad temperature range (Figure 8). In contrast to these two salts, complexes [3···I2]PF6− [7···I2]PF6 act very differently while heated. According to calculated versus observed weight losses, it can be deduced that iodine is released readily from the crystal lattice of [3···I2]PF6− [7···I2]PF6 at temperatures between 75 and 100 °C, and the resulting neat salts remain stable until decomposing at significantly higher temperatures. Consequently, in [3···I2]PF6−[7···I2]PF6, iodine is released from the structure in a way that the parent salt remains intact, whereas in [2···I2]PF6 the loss of iodine initiates the decomposition of the parental 2PF6 salt concurrently. The main reason for this behavior lies in the thermal behavior of the hydrophobic part of the salt. DSC measurements of samples 2PF6−7PF6 reveal sharp endothermic peaks at temperatures equal to the temperatures wherein corresponding I2 complexes start to relinquish their iodine molecules (Figure 9). In 2PF6, this

Figure 9. The first phase transition temperatures (DSC) of 2PF6− 7PF6 upon heating as compared to the I2 vaporization temperatures (TG) of [2···I2]PF6−[7···I2]PF6.

transition is followed by a series of phase transitions and almost instantly melting to the liquid state. In a dynamic TG measurement, this behavior hinders the vaporization of I2 possibly due to solvation of I2 into molten 2PF6 and furthermore may induce a chemical reaction between the two

Figure 10. Left: PXRD measurements carried out for [5···I2]PF6 at several different temperatures as compared to the PXRD pattern of neat 5PF6. Right: (a) The simulated diffraction pattern (single-crystal X-ray) of [5···I2]PF6 as compared to PXRD measurements carried out in bulk obtained by (b) crystallization from CHCl3 solution, (c) grinding, and (d) gas-to-solid reaction. 4165



Article

respective I−I bond, evident in the crystallographic data (Figure 11), is reproduced nicely by DFT (experimental data

reaction is near-quantitative for [4···I2]PF6 and [5···I2]PF6, whereas ground samples of [6···I 2 ]PF 6 and especially [7···I2]PF6 show a fair amount of unreacted PF6− salt still remaining (higher yields may be obtained by prolonged grinding). A similar procedure using 2PF6 yields a dark brown powder crystallographically different from the respective I2 complex measured by single-crystal X-ray. It thus seems that iodine is difficult both to introduce to the crystal matrix of the nonamphiphilic salt 2PF6 and to extract from the subsequent complex. We also made an experiment where a sample of 5PF6 was kept in a closed vial together, but not in contact, with iodine crystals for 3 days at room temperature. This gas-to-solid reaction also yields [5···I2]PF6 complex with only trace amounts of 5PF6 left. Hence, according to the above findings, solvent-free routes to prepare I2 complexes of monocationic DABCO salts can be used, but the length of the alkyl chain plays a major role in the complex formation. Computational Analysis. Although the structural ordering caused by the interdigitation of alkyl chains seems to be the decisive factor in explaining the differences of I2 uptake and release capabilities between the different monocationic salts, we also wanted to consider other possibilities. Therefore, we also conducted theoretical analysis of the nature and strength of halogen-bonding interactions of 1a and 1b and compare them to [2···I2]PF6−[7···I2]PF6. Furthermore, possible differences in the bonding interactions between [2···I2]PF6, [3···I2]PF6, and [4···I2]PF6−[7···I2]PF6 may give an insight into the differences in their thermal behavior. Because of the high computational costs of the long-chain complexes and the fact that complexes [4···I2]PF6−[7···I2]PF6 have almost identical experimental bonding parameters, we decided to limit the calculations to concern only salts [2···I2]PF6−[4···I2]PF6 in addition to 1a and 1b. The systems 1a and 1b were small enough to be optimized with the high level method (MP2/aug-cc-pVTZ, aug-cc-pVTZPP for iodine), and these calculations were used as “upperlimit” references in search of suitable DFT method, which would be used in calculations concerning salts [2···I2]PF6− [4···I2]PF6. After experimenting with different density functionals (see Supporting Information Table S4), we decided to use M05-2X with def2-TZVPP basis sets, a combination that was found to perform adequately well (XB bonding parameters were found to deviate less than 0.1 Å from the MP2 results, and the twisting of the DABCO, apparent in the crystal structures, was successfully reproduced) with reasonably small computational expenses (see Computational Details for more information). Bonding parameters of the optimized geometries are presented in Table 3. The optimization of geometries with the selected DFT method results in overestimation of the N···I distances and consequently underestimation of the I−I bond distances by roughly 6% and 4%, respectively, as compared to the experimental values extracted from the crystallographic data. The differences between calculated and experimental values can however be easily explained by the shape of the potential energy surface (disruption of the N···I2 bond from its equilibrium geometry by ±0.1 Å results only in less than 2 kJ/mol change in energy) along with the fact that calculations carried out to single molecular units in the gas phase do not take account of any packing effects present in the solid state (see Supporting Information Figure S7 for a potential energy scan along the N···I2 bond). Although the calculated and experimental bonding parameters differ slightly, the shortening of the N···I2 bond and the subsequent lengthening of the

Figure 11. Correlation (R2 = 0.996) between I−I bond lengths and N···I distances in complexes 1a, 1b, and [2···I2]PF6−[7···I2]PF6.

show 2−5% increase in the I−I distances as compared to 2−4% given by DFT calculations in which the bond distance is compared to an optimized value of 2.649 Å). While DFT describes the differences in XB bonding between 1a and 1b nicely (N···I bond is roughly 0.06 Å longer for 1b both theoretically and experimentally), the DFT optimized geometries for the ionic complexes [2···I2]PF6−[4···I2]PF6 are almost identical to each other in respect to N···I and I−I distances (Table 3), which is in contradiction with the solidstate observations. Therefore, it is likely that the different experimental N···I distances of [2···I2]PF6−[4···I2]PF6 are due to packing effects, as discussed above, or other weak interactions, such as vdW interactions between I2 and PF6−. Hence, the alkyl chain length would only have an indirect effect on the N···I 2 interaction strength. To confirm these assumptions, counterpoise corrected single point calculations were performed to the M05-2X optimized geometries of all studied complexes at M05-2X/def2-TZVPP and MP2/aug-ccpVTZ levels to yield interaction energies of the complex monomers in 1a, 1b, and [2···I2]PF6−[4···I2]PF6. The calculated interaction energies (ΔEcp) (Table 5) confirm the presumptions made from the optimized I−I and N···I distances. The interaction between DABCO and I2 in the 1:1 complex (1a) is the strongest followed by the 1:2 complex (1b) and then the salts [2···I2]PF6−[4···I2]PF6, which have practically equal interaction energies as compared to one another. The N···I2 interaction in 1a and 1b falls into the category of fairly strong halogen-bonding interactions. In fact, ΔEcp(M05-2X) for 1a is roughly 3 kJ/mol more negative (corresponding to stronger interaction) than for the trimethylamine···I231a complex and 22 kJ/mol more negative than what is observed for the 4-picoline···I2 complex.31b These both exhibit exceptionally short N···I distances in the solid state: 2.27 Å (Nc = 0.64) and 2.31 Å (Nc = 0.65), respectively.31 To exceed this value with amine N···I2 interactions, one might have to use a superbase such as 1,5-diazabicyclo[4.3.0]non-5-ene (DBN) whose estimated interaction energy with I2 is in turn ca. 8 kJ/mol smaller than in 1a (M05-2X/def2-TZVPP level).35 As mentioned above, the monocationic salts exhibit notably weaker interactions with I2 than neutral DABCO. Both MP2 and M05-2X derived interaction energies predict that the N···I2 interaction in [2···I2]PF6−[4···I2]PF6 is weakened by roughly 30 kJ/mol as compared to 1a and consequently some 20 kJ/ mol as compared to 1b. Clearly, both donation of electron 4166



Article

Table 5. Counterpoise Corrected MP2/aug-cc-pVTZ and M05-2X/def2-TZVPP Interaction Energies (ΔEcp) and Results from the Natural Bonding Orbital (NBO) and Natural Energy Decomposition Analyses (NEDA) ΔEcp (MP2)a 1a 1b [2···I2]PF6 [3···I2]PF6 [4···I2]PF6

−73.0 −64.4 −45.2 −45.2 −45.3

(11.8) (11.1) (10.5) (10.5) (10.5)

ΔEcp (M05-2X)a −66.7 −57.2 −36.8 −36.8 −36.9

(2.8) (2.7) (2.5) (2.4) (2.5)

ΔEint

CT

EL

POL

qCTb

−59.7 −51.2 −32.6 −32.6 −32.6

−233.5 −191.9 −135.1 −134.6 −135.3

−166.7 −139.9 −96.8 −96.5 −96.9

−153.7 −131.8 −107.0 −106.8 −107.2

−0.175 −0.144 −0.107 −0.107 −0.108

a Basis set superposition error (BSSE) energies in parentheses. bqCT refers to the charge transferred from XB acceptor to I2. For complete NEDA results, see Supporting Information Table S5. All energy values are in kJ/mol.

to the interaction energy reveals that, according to NEDA, both electrostatic and charge transfer interactions play a significant role in the attractive interactions. These findings agree well with the results of recent computational investigations.38 In the NEDA scheme, the electrostatic and charge transfer interactions are balanced by the CORE term, which arises from the Pauli repulsion between the filled orbitals of the complex monomers and is therefore always positive. The cancellations are fairly equal, but result in a slight favoring of electrostatic interactions with longer N···I distances (value of |EL|/CORE increases when comparing 1a to ionic complexes) and charge transfer interactions when N···I distance is short (value |CT|/ CORE is greater for neutral complexes). CT interactions are also frequently quantified by analyzing the amount of electronic charge transferred between the complex monomers. Table 5 shows the I2 monomer charges (qCT) derived from NBO calculations, which supports the trends apparent in NEDA results.

density from the lone-pair of tertiary amine of DABCO to a hydrogen bond (see Supporting Information Figure S1 for a chloroform solvate of 1a) or halogen bond (1b) and its covalent sharing {[2···I2]PF6−[4···I2]PF6} weaken the basicity of the second amine and diminish its capability to act as a halogen-bond acceptor. To extend the analysis, XB interactions were studied using the energy decomposition analysis (NEDA), which allows the partition of interaction energy into charge transfer (CT), electrical (EL), and repulsive (CORE) components (i.e., ΔE = CT + EL + CORE) based on the natural bonding orbital (NBO)36 method. This provides a convenient way to quantify the different contributions affecting the XB interactions.37 Table 5 summarizes the NEDA results. First, regarding the interaction energy, NEDA reproduces the trends observed in counterpoise corrected MP2 and M05-2X interaction energies. Second, the separate energy components (CT, EL, and CORE) computed for molecular complexes (1a and 1b) correlate well with the corresponding values obtained for their ionic counterparts {[2···I2]PF6−[4···I2]PF6}; that is, the CT and EL increase while the CORE term decreases in proportion to the corresponding interaction energies (Figure 12). These findings suggest that, although the XB interaction

■

CONCLUSION In this Article, we have described the structures, thermal behavior, and bonding analysis of complexes of elemental iodine with neutral and monocationic DABCO salts. The long tail amphiphilic monocationic I2 complexes exhibit interesting solid-state structures due to the interdigitation of n-alkyl chains leading to layered packing. It was found that the alkyl chain length has a clear impact also on the thermal behavior of these complexes as the evaporization/release of I2 from the crystal lattices of [2···I2]PF6−[7···I2]PF6 is dependent on the solid phase transitions occurring for the corresponding monocationic DABCOs. This in itself is an interesting feature as it enables the control over the reversible binding of I2. Furthermore, because the binding of I2 occurs quantitatively in most cases and the resulting I2 complexes are very stable at room temperature, salts of this type could be used as recyclable storage material, for example, for radioactive iodine. In many similar systems, where the iodine is trapped into the crystal lattice in the form of I2 or as I42−, they either lose the iodine slowly already in mild conditions or require high temperatures to release all of the bound iodine. The amphiphilic monoalkyl DABCO salts also have the ability to capture gaseous iodine to crystalline bulk, although they do not have any apparently accessible voids in their crystal lattices. Thus, they can be described as dynamically porous materials. The theoretical bonding analyses confirm that the monocationic DABCO core is a moderately strong halogenbond acceptor but clearly weaker as compared to the parent neutral DABCO. The bonding interactions, however, are very similar. The suitable bonding strength is most likely also a significant factor why reversible binding of iodine is observed for the monocationic DABCO salts but not for the neutral DABCO. The role of the anion in the structure and behavior of

Figure 12. Graph showing the decomposition of interaction energies (ΔEint in kJ mol−1) of studied XB interactions in molecular (1a and 1b) and ionic {[2···I2]PF6−[4···I2]PF6} complexes into repulsive (CORE, upper column), charge transfer (CT, middle column), and electrical (EL, bottom column) components as given by natural energy decomposition analysis (NEDA). For more numerical data, see Table 5.

strength is somewhat diminished in the ionic complexes, the nature of the interaction remains more or less the same despite the obvious differences, for example, charge distribution, between neutral and monocationic DABCOs. The diminished interaction strength of the ionic complexes may be a part of the explanation why the monocationic salts are capable of binding and releasing iodine, and do it in a relatively low temperature, whereas the neutral species of the same diamine are not capable for a similar process. The closer inspection of the contributions 4167



Article

Melai, B; Sanzone, A.; Valentini, G. Pure Appl. Chem. 2009, 81, 2035. (d) Ohta, K.; Sugiyama, T.; Nogami, T. J. Mater. Chem. 2000, 10, 613. (12) (a) Cinčić, D.; Frišcǐ ć, T.; Jones, W. Chem. Mater. 2008, 20, 6623. (b) Raatikainen, K.; Rissanen, K. CrystEngComm 2009, 11, 750. (c) Cinčić, D.; Frišcǐ ć, T.; Jones, W. Chem.-Eur. J. 2008, 14, 747. (13) (a) Santos, P. S.; Mello, M. T. S. J. Mol. Struct. 1988, 178, 121. (b) Halpern, A. M.; Weiss, K. J. Am. Chem. Soc. 1968, 90, 6297. (c) Nagakura, S. J. Am. Chem. Soc. 1958, 80, 520. (14) (a) Rozell, W. J.; Wood, J. S. Inorg. Chem. 1977, 16, 1827. (b) Konarev, D. V.; Khasanov, S. S.; Saito, G.; Lyubovskaya, R. N. Cryst. Growth Des. 2009, 9, 1170. (15) COLLECT; Bruker AXS, Inc.: Madison, WI, 2008. (16) Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307. (17) Sheldrick, G. M. SADABS; University of Göttingen: Göttingen, Germany, 2008. (18) Burla, M. C.; Carrozzini, B.; Cascarano, G. L.; Giacovazzo, C.; Polidori, G. Z. Kristallogr. 2002, 217, 629. (19) Burla, M. C.; Caliandro, R.; Camalli, M.; Carrozzini, B.; Cascarano, G. L.; De Caro, L.; Giacovazzo, C.; Polidori, G.; Spagna, R. J. Appl. Crystallogr. 2005, 38, 381. (20) Sheldrick, G. Acta Crystallogr., Sect. A 2008, 64, 112. (21) Farrugia, L. J. J. Appl. Crystallogr. 1999, 32, 837. (22) Macrae, C. F.; Bruno, I. J.; Chisholm, J. A.; Edgington, P. R.; McCabe, P.; Pidcock, E.; Rodriguez-Monge, L.; Taylor, R.; van de Streek, J.; Wood, P. A. J. Appl. Crystallogr. 2008, 41, 466. (23) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, Ö .; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, revision C.1; Gaussian, Inc.: Wallingford, CT, 2009. (24) Granovsky, A. A. Firefly version 7.1.G, www http://classic.chem. msu.su/gran/firefly/index.html. (25) Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.; Gordon, M. S.; Jensen, J. H.; Koseki, S.; Matsunaga, N.; Nguyen, K. A.; Su, S.; Windus, T. L.; Dupuis, M.; Montgomery, J. A. J. Comput. Chem. 1993, 14, 1347. (26) Glendening, E. D.; Badenhoop, J. K.; Reed, A. E.; Carpenter, J. E.; Bohmann, J. A.; Morales, C. M.; Weinhold, F. NBO 5.0; Theoretical Chemistry Institute, University of Wisconsin: Madison, WI, 2001. (27) (a) Glendening, E. D.; Streitwieser, A. J. Chem. Phys. 1994, 100, 2900. (b) Schenter, G. K.; Glendening, E. D. J. Phys. Chem. 1996, 100, 17152. (28) Boys, S. F.; Bernardi, F. Mol. Phys. 1970, 19, 553. (29) (a) Feller, D. J. Comput. Chem. 1996, 17, 1571. (b) Schuchardt, K. .L.; Didier, B. T.; Elsethagen, T.; Sun, L.; Gurumoorthi, V.; Chase, J.; Li, J.; Windus, T. L. J. Chem. Inf. Model. 2007, 47, 1045. (30) C3- and D3h-symmetric geometries are divided by only 0.4 kJ/ mol in energy at the M05-2X/TZVP level. However, according to vibrational analysis, C3-symmetric geometry represents a minimum in the potential energy hypersurface, whereas D3h-symmetric geometry has one imaginary frequency corresponding to twisting of the DABCO core towards C3-symmetry. (31) (a) Stromme, K. O. Acta Chem. Scand. 1959, 13, 268. (b) Hassel, O.; Romming, C.; Tufte, T. Acta Chem. Scand. 1961, 15, 967.

these salts cannot be overemphasized. Anions that are bulky (such as tetraphenylborate) or exhibit more localized negative charge (e.g., carboxylates) would most likely alter the packing and/or have an influence on the XB interactions, respectively. We are currently investigating several new types of XB acceptors based on the monoquaternized DABCO and the effects of different anions on the XB interactions in such systems.

■

ASSOCIATED CONTENT

* Supporting Information S

Additional tables, structure depictions, XRD patterns, DSC and TG curves, and xyz-coordinates. This material is available free of charge via the Internet at http://pubs.acs.org. CCDC reference numbers: 879929−879942.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: manu.k.lahtinen@jyu.fi. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We would like to thank Ms. Elina Hautakangas for the elemental analyses. We gratefully acknowledge the financial support of the University of Jyväskylä, K.R. for the funding from the Academy of Finland (grant nos. 130629, 122350, and 140718), and A.P. for the funding from the Inorganic Materials Chemistry Graduate Program.

■

REFERENCES

(1) (a) Metrangolo, P.; Neukirch, H.; Pilati, T.; Resnati, G. Acc. Chem. Res. 2005, 38, 386. (b) Rissanen, K. CrystEngComm 2008, 10, 1107. (2) Metrangolo, P.; Meyer, F.; Pilati, T.; Resnati, G.; Terraneo, G. Angew. Chem., Int. Ed. 2008, 47, 6114. (3) Clark, T.; Hennemann, M.; Murray, J. S.; Politzer, P. J. Mol. Model. 2007, 13, 291. (4) (a) Guthrie, F. J. Chem. Soc. 1863, 16, 239. (b) Remsen, I.; Norris, J. F. Am. Chem. J. 1896, 18, 90. (5) (a) Bent, H. A. Chem. Rev. 1968, 68, 587. (b) Hassel, O. Science 1970, 170, 497. (6) (a) Metrangolo, P.; Carcenac, Y.; Lahtinen, M.; Pilati, T.; Rissanen, K.; Vij, A.; Resnati, G. Science 2009, 323, 1461. (b) Arduengo, A. J., III; Kline, M.; Calabrese, J. C.; Davidson, F. J. Am. Chem. Soc. 1991, 113, 9704. (c) Raatikainen, K.; Rissanen, K. CrystEngComm 2011, 13, 6972. (d) Raatikainen, K.; Rissanen, K. Chem. Sci. 2012, 2, 1235. (e) Dolenc, D.; Modec, B. New J. Chem. 2009, 33, 2344. (f) Perkins, C.; Libri, S.; Adams, H.; Brammer, L. CrystEngComm 2012, 14, 3033. (7) (a) Abate, A.; Brischetto, M.; Cavallo, G.; Lahtinen, M.; Metrangolo, P.; Pilati, T.; Radice, S.; Resnati, G.; Rissanen, K.; Terraneo, G. Chem. Commun. 2010, 46, 2724. (b) Müller, M.; Albrecht, M.; Gossen, V.; Peters, T.; Hoffmann, A.; Raabe, G.; Valkonen, A.; Rissanen, K. Chem.-Eur. J. 2010, 16, 12446. (8) (a) Aakeröy, C. B.; Chopade, P. D.; Ganser, C.; Desper, J. Chem. Commun. 2011, 47, 4688. (b) Aakeröy, C. B.; Schultheiss, N. C.; Rajbanshi, A.; Desper, J.; Moore, C. Cryst. Growth Des. 2009, 9, 432. (9) Baghernejad, B. Eur. J. Chem. 2010, 1, 54. (10) (a) Oliveri, C. G.; Gianneschi, N. C.; Nguyen, S. T.; Mirkin, C. A.; Stern, C. L.; Wawrzak, Z.; Pink, M. J. Am. Chem. Soc. 2006, 128, 16286. (b) Gu, J.-M.; Kim, W.-S.; Huh, S. Dalton Trans. 2011, 40, 10826. (11) (a) Wykes, A.; MacNeil, S. L. Synlett 2007, 107. (b) Xu, D.-Z.; Liu, Y.; Shi, S.; Wang, Y. Green Chem. 2010, 12, 514. (c) Chiappe, C.; 4168



Article

(32) Metrangolo, P., Resnati, G., Eds. Halogen Bonding: Fundamentals and Applications. Structure and Bonding; Springer: Berlin, Germany, 2008; Vol. 126, p 65. (33) van Bolhuis, F.; Koster, P. B.; Migchelsen, T. Acta Crystallogr. 1967, 23, 90. (34) Bailey, R. D.; Drake, G. W.; Grabarczyk, M.; Hanks, T. W.; Hook, L. L.; Pennington, W. T. J. Chem. Soc., Perkin Trans. 2 1997, 2773. (35) Counterpoise corrected interaction energies were obtained at the M05-2X/def2-TZVPP level using optimized geometries with the corresponding method. The optimized geometries of the I2 complexes of trimethylamine and 4-picoline yielded longer N···I distances than what was obtained for 1a. The interaction energies, calculated at the M05-2X level, only give fair estimates on the strengths of the XB interactions between the chosen monomers and should by no means be taken as highly accurate values. (36) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. Rev. 1988, 88, 899. (37) For a recent study of trends in halogen bonding, studied using natural energy decomposition analysis, see: Huber, S. M.; JimenezIzalb, E.; Ugaldeb, J. M.; Infante, I. Chem. Commun. 2012, 48, 7708. (38) (a) Zou, J.-W.; Jiang, Y.-J.; Guo, M.; Hu, G.-X.; Zhang, B.; Liu, H.-C.; Yu, Q.-S. Chem.-Eur. J. 2005, 11, 740. (b) Palusiak, M. Comput. Theor. Chem. 2010, 945, 89.

4169


Halogen Bonding-Based âCatch and Releaseâ: Reversible Solid-State

Recommend Documents

Halogen Bonding-Based âCatch and Releaseâ: Reversible Solid-State