Nature of Supramolecular Complexes Controlled by the Structure of

Jun 4, 2009 - Department of Chemistry, University of Miami, Coral Gables, Florida ... of complexes has been noted: the one which we call cavitandplex ...
0 downloads 0 Views 1MB Size
Download PDF
pubs.acs.org/Langmuir © 2009 American Chemical Society

Nature of Supramolecular Complexes Controlled by the Structure of the Guest Molecules: Formation of Octa Acid Based Capsuleplex and Cavitandplex Nithyanandhan Jayaraj,† Yaopeng Zhao,†,‡ Anand Parthasarathy,† Mintu Porel,† Robert S. H. Liu,‡ and V. Ramamurthy*,† †

Department of Chemistry, University of Miami, Coral Gables, Florida 33124, and ‡Department of Chemistry, University of Hawaii, Honolulu, Hawaii 96822 Received April 17, 2009. Revised Manuscript Received May 10, 2009

Factors that govern inclusion of organic molecules within octa acid (OA), a synthetic deep cavity cavitand, have been delineated by examining the complexation behavior of a number of organic molecules with varying dimensions and functionalities with OA. The formation of two types of complexes has been noted: the one which we call cavitandplex is a partially open complex in which a part of the guest molecule remains exposed to water, and the other termed capsuleplex is formed through assembly of two OA molecules. In capsuleplex, the guest is protected from water. Generally, guest molecules that possess ionic head groups form cavitandplex, and all others form capsuleplex. Capsuleplex may contain one or two guest molecules within the capsule. Small organic molecules (12 A˚) preferentially form 2:1 capsuleplex. Extensive 1H NMR experiments have been carried out to characterize host-guest complexes. In the absence of the guest, OA tends to aggregate in water. The extent of aggregation depends on the concentration of OA and the presence of salts in solution. We expect the information obtained from this study to be of great value in predicting the nature of complexes with a given guest and facilitating appropriate guest chosen by researchers.

Introduction Following the discovery of crown ethers, molecules with a “hole” or a “cavity” that have fascinated chemists have facilitated exploration of weak intermolecular interactions, mechanism of substrate-enzyme binding, and manipulation of molecules in confined spaces.1-14 To the extensively investigated list of molecules such as cyclodextrins, calixarenes, and calixarene-based *To whom correspondence should be addressed. E-mail: murthy1@ miami.edu. (1) D’Souza, V. T.; Bender, M., L. Acc. Chem. Res. 1987, 20, 146–152. (2) Breslow, R. Chem. Br. 1983, 19, 126; 128; 130-131. (3) Breslow, R. Pure Appl. Chem. 1994, 66, 1573–1582. (4) Breslow, R.; Dong, S. D. Chem. Rev. 1998, 98, 1997–2011. (5) Paulini, R.; Muller, K.; Diederich, F. Angew. Chem., Int. Ed. 2005, 44, 1788–1805. (6) Meyer, E. A.; Castellano, R. K.; Diederich, F. Angew. Chem., Int. Ed. 2003, 42, 1210–1250. (7) Weiss, R. G. Tetrahedron 1988, 44, 3413–3475. (8) Whitten, D. G. Acc. Chem. Res. 1993, 26, 502–509. (9) Whitten, D. G.; Russell, J. C.; Schmehl, R. H. Tetrahedron 1982, 38, 2455–2487. (10) Ramamurthy, V. Tetrahedron 1986, 42, 5753–5839. (11) Tung, C.-H.; Wu, L.-Z.; Zhang, L.-P.; Chen, B. Acc. Chem. Res. 2003, 36, 39–47. (12) Turro, N. J. Acc. Chem. Res. 2000, 33, 637–646. (13) Turro, N. J.; Graetzel, M.; Braun, A. M. Angew. Chem., Int. Ed. Engl. 1980, 19, 675–696. (14) Scaiano, J. C.; Garcia, H. Acc. Chem. Res. 1999, 32, 783–793. (15) Gutsche, C. D. Calixarenes Revisited; Royal Society of Chemistry: Oxford, 1998. (16) Mandolini, L.; Ungaro, R. Calixarenes in Action; Imperial College Press: London, 2000. (17) Rudkevich, D. M. Bull. Chem. Soc. Jpn. 2002, 75, 393–413. (18) Biros, S. M.; Rebek, J. Chem. Soc. Rev. 2007, 36, 93–104. (19) Middel, O.; Verboom, W.; Reinhoudt, D. N. Eur. J. Org. Chem. 2002, 2587–2597. (20) Dodziuk, H. Cyclodextrins and Their Complexes; Wiley-VCH: Weinheim, 2006. (21) Atwood, J. L.; Davies, J. E. D.; MacNicol, D. D.; Vogtle, F. In Comprehensive Supramolecular Chemistry; Szejtli, J., Osa, T., Eds.; Pergamon: 1996; Vol. 3, pp 154-188.

Langmuir 2009, 25(18), 10575–10586

cavitands of the past 5 decades,15-21 cucurbiturils and synthetically designed inorganic hosts have been added during the past decade.22-26 Cyclodextrins and carboxylic or sulfonic acid group functionalized calixarenes and cavitands are water-soluble. This paper deals with one such water-soluble synthetic cavitand based host, octa acid (OA), a molecule as the name implies with eight carboxylic acid groups, in the form of a traditional “lampshade” arranged with four acid groups each at the top and the bottom.27 The features of the cavity of this molecule are similar to cyclodextrins (CD), calixarenes (CA), and cucurbiturils (CB) (Scheme 1). Our interest in exploring the photochemical and photophysical properties of small organic molecules confined in small spaces drew our attention to OA.28-30 Intelligent exploitation of the cavity of any host requires knowledge of (a) the cavity dimensions, (b) the nature of the cavity’s interior (polar or nonpolar), (c) the stoichiometry of the host-guest complex, (d) the types of interactions holding a guest within a host, (e) the strength of the host-guest complexes (binding constants), (f) the dynamics of the host-guest complex, and so on. In this paper, we explore the (22) Lee, J. W.; Samal, S.; Selvapalam, N.; Kim, H.-J.; Kim, K. Acc. Chem. Res. 2003, 36, 621–630. (23) Lagona, J.; Mukhopadhyay, P.; Chakrabarti, S.; Isaacs, L. Angew. Chem., Int. Ed. 2005, 44, 4844–4870. (24) Pluth, M. D.; Raymond, K. N. Chem. Soc. Rev. 2007, 36, 161–171. (25) Fiedler, D.; Leung, D. H.; Bergman, R. G.; Raymond, K. N. Acc. Chem. Res. 2005, 38, 351–360. (26) Fujita, M.; Tominaga, M.; Hori, A.; Therrien, B. Acc. Chem. Res. 2005, 38, 371–380. (27) Gibb, C. L. D.; Gibb, B. C. J. Am. Chem. Soc. 2004, 126, 11408–11409. (28) Sivaguru, J.; Natarajan, A.; Kaanumalle, L. S.; Shailaja, J.; Uppili, S.; Joy, A.; Ramamurthy, V. Acc. Chem. Res. 2003, 36, 509–521. (29) Ramamurthy, V.; Weiss, R. G.; Hammond, G. S. In Advances in Photochemistry; Volman, D., Hammond, G. S., Neckers, D. C., Eds.; John Wiley & Sons: 1993; Vol. 18, pp 67-236. (30) Weiss, R. G.; Ramamurthy, V.; Hammond, G. S. Acc. Chem. Res. 1993, 26, 530–536.

Published on Web 06/04/2009

DOI: 10.1021/la901367k

10575

Article

Jayaraj et al. Scheme 1. Structures and Dimensions of Various Molecular Containersa

a

Dimensions noted below each structure refer to atom-to-atom distance and do not include van der Walls radii.

Scheme 2. Structures of Guests Used in This Investigationa

a Dimensions noted below each structure refer to atom-to-atom distance and do not include van der Walls radii. Distances estimated using either Spartan or Chemdraw programs.

correlation between the structure of a guest and the stoichiometry of OA-guest complexes employing a variety of guest molecules of varying dimensions and functional groups (Scheme 2). Generally, host molecules with functionalities at the periphery that serve dual roles of hydrogen bond donor-acceptor tend to form a capsular assembly by association of two host molecules.31-35 These host molecules could accommodate one or two guest molecules, depending on their cavity size. Thus 1:1, 1:2, 2:1, and 2:2 (in this paper, we shall always refer to host/guest ratio) (31) Conn, M. M.; Rebek, J. Chem. Rev. 1997, 97, 1647–1668. (32) Higler, I.; Timmerman, P.; Verboom, W.; Reinhoudt, D. N. Eur. J. Org. Chem. 1998, 2689–2702. (33) Corbellini, F.; Knegtel, R. M. A.; Grootenhuis, P. D. J.; Crego-Calama, M.; Reinhoudt, D. N. Chem.;Eur. J. 2005, 11, 298–307. (34) Lutzen, A. Angew. Chem., Int. Ed. 2005, 44, 1000–1002. (35) Schmuck, C. Angew. Chem., Int. Ed. 2007, 46, 5830–5833.

10576 DOI: 10.1021/la901367k

complexes are known with cyclodextrins and calixarenes. However, cucurbiturils with only the hydrogen bond acceptor (CdO) at the periphery form 1:1 or 1:2 complexes. Despite the drawback of the not very optimal placement of the eight carboxylic acid groups of OA to bring two molecules of OA together through hydrogen bonds, it tends to form a capsular assembly with a large variety of guest molecules. Thus, the complexation properties of OA are not a simple extension of behavior of the analogous hosts, CD, CB, and CA.36-38 We have recently established OA to be far superior to the latter hosts in bringing to light latent photochemical and photophysical properties of organic guest molecules. (36) Gibb, C. L. D.; Gibb, B. C. J. Am. Chem. Soc. 2006, 128, 16498–16499. (37) Gibb, C. L. D.; Gibb, B. C. Chem. Commun. 2007, 1635–1637. (38) Sun, H.; Gibb, C. L. D.; Gibb, B. C. Supramol. Chem. 2008, 20, 141–147.

Langmuir 2009, 25(18), 10575–10586

Jayaraj et al.

Article

These preliminary results, we believe, also pointed out that the full potential of this host is yet to be realized. The results of a systematic investigation prompted by the imperative understanding of the complexation properties of OA with a variety of organic molecules necessary to realize the full potential of OA are presented below.

Results and Discussion Choice of Guests. All guests we investigated during the last 5 years in the context of excited state behavior formed either a 2:1 or a 2:2 capsular complex (both types of complexes are termed capsuleplex).39-42 No 1:1 complex (cavitandplex) was detected.43 We were therefore curious to understand why only capsuleplexes were formed with the investigated guests and also wished to identify conditions that would promote the formation of cavitandplex. At this stage when our interest is focused on manipulating reactions using a confined space, an understanding of factors controlling the formation of 1:1, 2:1, and 2:2 complexes seemed prudent. The molecules we have employed as guests and their dimensions are listed in Scheme 2. The dimensions of the OA cavity are provided in Scheme 1. The depth that could be occupied by a guest is about 7 A˚ in a cavitand, and in a capsule it would be 14 A˚. Several of the guests used in this study are much longer than the capsule length and therefore would be forced to coil or fold within the capsule. Adamantane and its derivatives (1-4) were the first group of molecules chosen for the investigation. A bulky small molecule, adamantane, known to fit snugly within a CD cavity would be expected to do so within the OA cavity as well to form a 1:1 cavitandplex.44 Believing the presence of an ionic head group to help prevent capsular assembly and favor the formation of 1:1 cavitandplex, adamantane substituted with negatively charged carboxylate anion (2) and positively charged ammonium ion (3) were chosen as guests. Adamantane with charged groups formed 1:1 cavitandplexes, and neutral adamantane preferred 2:1 and 2:2 capsuleplexes. The different behavior of the two adamantane systems 1-3 (neutral and charged) that prompted us to extend the study to butyl adamantane carboxylate 4, a longer molecule with a polar head group, unlike adamantane with an ionic head group, resulted in a 2:1 capsuleplex. To understand the compexation process further, we investigated a series of symmetrical 1,4-dialkyl benzenes (5-10). Depending upon the alkyl chain, the extended lengths of these molecules vary between 9 and 22 A˚. Although dialkyl benzenes are not bulky (like adamantane), they could fill the cavity by coiling/folding the alkyl chains. In principle, shorter dialkylbenzenes 5 and 6 could form 1:1 cavitandplex. As will be discussed later, neither of these molecules formed 1:1 cavitandplex; 5 similar to adamantane formed both 2:1 and 2:2 capsuleplexes, while all others (6-10) formed only 2:1 capsuleplexes. We also believed that long alkyl chain would allow us to explore the extent of involvement of CH3-π interaction in the complexation process and examine whether those in especially 8-10 would coil/fold (39) Kaanumalle, L. S.; Gibb, C. L. D.; Gibb, B. C.; Ramamurthy, V. J. Am. Chem. Soc. 2004, 126, 14366–14367. (40) Natarajan, A.; Kaanumalle, L. S.; Jockusch, S.; Gibb, C. L. D.; Gibb, B. C.; Turro, N. J.; Ramamurthy, V. J. Am. Chem. Soc. 2007, 129, 4132–4133. (41) Kaanumalle, L. S.; Gibb, C. L. D.; Gibb, B. C.; Ramamurthy, V. Org. Biomol. Chem. 2007, 5, 236–238. (42) Sundaresan, A. K.; Ramamurthy, V. Photochem. Photobiol. Sci. 2008, 7, 1555–1564. (43) Porel, M.; Jayaraj, N.; Kaanumalle, L. S.; Maddipatla, M. V. S. N.; Parthasarathy, A.; Ramamurthy, V. Langmuir 2009, 25, 3473–3481. (44) Breslow, R.; Czarniecki, M. F.; Emert, J.; Hamaguchi, H. J. Am. Chem. Soc. 1980, 102, 762–770.

Langmuir 2009, 25(18), 10575–10586

Figure 1. 1H NMR titration spectra of adamantane derivatives with OA, (i) OA (1 mM, 10 mM buffer), (ii) OA:(1-adamantyl) ethyltrimethylammonium iodide (1:1), (iii) OA:1-adamantaneacetic acid (1:1), and (iv) OA:butyl-1-adamantanecarboxylate (2:1). All spectra recorded in borate buffer in D2O. See Scheme 2 for the assignment of adamantyl hydrogens, R, β, γ, and so on, and Scheme 1 for assignment of OA hydrogens.

within the capsule. Finally, adamantane and 1,4-diethylbenzene that formed both 2:1 and 2:2 capsuleplexes suggested that anthracene and naphthalene (molecules similar or smaller sized than 1,4-diethylbenzene) might form 2:1 capsuleplex. We were interested in examining the possibility of forming 2:1 capsuleplexes with anthracene and naphthalene with OA, molecules we had earlier reported to form 2:2 capsuleplexes.45 Results of these studies are presented in this report. Techniques Used. The current study consists of 1H NMR titration (to identify host-guest stoichiometry), 2D-COSY (to assign the host and guest signals), 2D-NOESY (to probe the relative orientation of the guest within the host), 2D-DOSY (to determine the diffusion constants of the complexes), and variable temperature 1H NMR (to examine possible equilibrium between 2:1 and 2:2 complexes) experiments. All 2D-COSY spectra are provided as Supporting Information and are not discussed in the text. These COSY spectra helped us assign 1H NMR signals of the host and guest in the NMR spectra. In the case of anthracene and naphthalene, fluorescence emissions were recorded at various host-guest ratios to complement the conclusions drawn from NMR experiments. It is important to note that, for all host-guest systems investigated, with increasing addition of the guest, 1H NMR signals due to free guest were seen only when no free host remained. During the titration experiments, the absence of observable peak shifts indicating the chemical shifts of the complexed guest peaks were independent of the concentrations of host and guest in solution precluded us from determining the binding constants of host-guest complexes investigated. (45) Kaanumalle, L. S.; Gibb, C. L. D.; Gibb, B. C.; Ramamurthy, V. J. Am. Chem. Soc. 2005, 127, 3674–3675.

DOI: 10.1021/la901367k

10577

Article

Jayaraj et al.

Figure 2. 2D-NOESY correlations of adamantane derivatives with OA (5 mM in 50 mM borate buffer); (i) OA:1-adamantaneacetic acid (1:1), (ii) OA:1-(1-adamantyl)ethyltrimethylammonium iodide (1:1), and (iii) OA:butyl-1-adamantanecarboxylate (2:1). See Scheme 2 for the assignment of adamantyl hydrogens, R, β, γ, and so on, and Scheme 1 for assignment of OA hydrogens.

Based on 1H NMR data, we conclude that, in the NMR time scale, the complexes were stable. Complexation of Adamantane Derivatives. The first two molecules we discuss are the cationic and anionic adamantane derivatives 2 and 3. Several lines of evidence suggest that these molecules form 1:1 stoichiometric complexes with OA in water at pH 8.5 (borate buffer). Inclusion of these guests within the OA cavity was evident from the upfield shift of the signals due to the adamantyl hydrogens (Figure 1). In Figure 1, the signals due to host and guest hydrogens are marked. First, the disturbed host signals are indicative of interaction between the host and the guest molecules. Second, upfield shifted guest signals due to the diamagnetic shielding by the aromatic units lining the interior of the OA cavity confirm the guest’s location within the cavity. Importantly, the lack of splitting of the signals due to OA confirms that 2 and 3 did not form 2:1 capsuleplex where the top and bottom halves of the capsule would be different and hence expected to show different signals (assuming that there is no tumbling within the capsule) for equivalent hydrogens on the two halves. The absence of change in the 1H NMR titration spectra (Figures S1 and S2 in the Supporting Information) beyond addition of one equivalent of guest to OA in water with 2 and 3 suggesting a 1:1 stoichiometry for the complex could mean the formation of either 1:1 cavitandplex or 2:2 capsuleplex. In the case of cationic derivative 3, although the signals due to adamantyl hydrogens are shifted upfield, the signals due to methyl groups (trimethyl ammonium and R-methyl) are unaffected by complexation, suggesting that this part of the molecule is exposed to 10578 DOI: 10.1021/la901367k

water and is not shielded by the host cavity. This is possible only for a 1:1 cavitandplex with the adamantyl part buried within the cavity and the rest facing water. A similar observation was made with anionic derivative 2. Confirmatory evidence in favor of 1:1 cavitandplex came from measurement of diffusion constants through 2D-DOSY experiments. The diffusion constants in borate buffer for free OA at 25 °C (1.88  10-6 cm2/s) upon complexation of the host with 2 and 3 were reduced to 1.73  10-6 and 1.80  10-6 cm2/s, respectively. The slight decrease in the diffusion constant of OA suggests minimal change in the hydrodynamic radius of OA upon binding with either guests. These diffusion constants of OA (free or as complex with guests) are consistent with 1:1 cavitandplex and not with 2:2 capsuleplex (2:1 or 2:2) where it would be expected to be in the range of 1.2  10-6 to 1.5  10-6 cm2/s. The diffusion constants of free 2 (5.14  10-6 cm2/s) and 3 (4.47  10-6 cm2/s) in borate buffer were reduced upon complexation with OA to 2.2  10-6 and 2.09  10-6 cm2/s, respectively. A significant decrease in diffusion constants of guests upon complexation suggests that hydrodynamic radii of the guests have increased and also confirms complex formation with OA. The above conclusion that the adamantanes substituted with an ionic head group form a 1:1 cavitandplex is consistent with the 2DNOESY correlation spectra provided in Figure 2 where the correlations seen for guest and host signals are marked. While both 2 and 3 show NOE correlations between the guest and host signals, they are limited to the adamantyl hydrogens in the case of 2 while more extensive in 3. In the latter case, correlations between Langmuir 2009, 25(18), 10575–10586

Jayaraj et al.

Figure 3. 1H NMR titration spectra of adamantane with OA, (i) OA (1 mM, 10 mM buffer), (ii) H:G (10:0.05), (iii) H:G (10:1.5), (iv) H:G (10:4), and (v) H:G (10:10). Inset: the bound guest proton resonance region is expanded.

the four methyl signals (trimethyl ammonium and R0 -methyl) and the hydrogens at the broader end of OA (Ha, Hc, Hd, and He) are clearly noticeable. Based on extensive NOESY correlations seen for 3, we believe that this molecule, compared to 2, is buried deeper within the OA cavity probably due to the attractive interaction between its cationic head group and the host’s four carboxylate anions at its broader end. Adamantane (1) is not soluble in water. However, upon addition of adamantane to OA in water, signals due to adamantane could be seen in 1H NMR spectra (Figure 3). Titration spectra for OA-adamantane are provided in Figure 3, and a more extensive one in the Supporting Information (Figure S3). As would be expected, hydrogen signals of adamantane included within OA were shifted upfield. More importantly, the chemical shifts of the adamantane signals change with the relative amounts of OA to adamantane (host to guest ratio) in solution. A set of signals present at high OA-adamantane ratio (low adamantane concentration) is replaced by another when their concentration ratio reaches 1:1. We believe these two sets of signals represent two different complexes; those due to the preferred 2:1 capsuleplex at low concentrations of adamantane are replaced by the ones due to 2:2 capsuleplex at high adamantane concentrations. The change in the nature of the complex is also reflected in the signals due to the host OA. In Figure 3, it must be noted that signals due to hydrogens marked (a) and (c) (Scheme 1) for OA are not identical at these two concentrations. The diffusion constants obtained for OA and the guest adamantane at 2:1 and 2:2 host to guest ratios were 1.34  10-6 and 1.33  10-6 cm2/ sec, respectively, significantly lower than that for free OA and consistent with formation of capsuleplex but not cavitandplex. Thus, unlike the ionic adamantane derivatives 2 and 3, adamantane does not form cavitandplex; it only forms capsuleplex with Langmuir 2009, 25(18), 10575–10586

Article

the number of adamantane molecules present in the capsule depending on the relative concentrations of the host and guest. Having concluded that adamantane forms both 2:1 and 2:2 capusleplexes, we were interested to examine the behavior of a neutral adamantane derivative slightly longer than the parent molecule and containing a polar head group (COOR). In this context, we examined the behavior of butyl-1-adamantanecarboxylate 4. With 4, 1H NMR titration experiments suggested the formation of 2:1 capsuleplex (Figure S4 in the Supporting Information). Expectedly, the guest proton resonances were upfield shifted upon binding to OA (Figure 1). The aromatic resonances of the host provided clear evidence in favor of 2:1 capsuleplex. The unsymmetrical nature of the guest rendering the two halves of the capsule different resulted in the equivalent hydrogens of the two capsular halves to have different chemical shifts. Consistent with the capsuleplex formation the diffusion constant of 4@(OA)2 was 1.20  10-6 cm2/s. Thus, 4 longer than adamantane did not form 2:2 capsuleplex but preferred 2:1 capsuleplex. From comparison of the behavior of 2-4, it is clear that 1:1 cavitandplex is preferred only when the ionic head group is present and a neutral polar group as in 4 is insufficient to favor 1:1 cavitandplex. Complexation of 1,4-Dialkyl Benzenes. In this section, we discuss inclusion of molecules longer and less bulkier than adamantyl derivatives, namely, the symmetrical 1,4-dialkyl benzenes 5-10. The results on 6-10 whose length in extended conformation varies between ∼10-22 A˚ are presented first followed by molecule 5 showing a different behavior. We believed complexation behavior of guests 6-10 would provide information on the limit of the length of guest molecules that would fit within the capsular assembly of OA.37 The 1H NMR titration spectra of 6-10 are provided in the Supporting Information (Figure S5-S9). These molecules prefer to form 2:1 capsuleplexes. The diffusion rates for 6-10 complexed to OA measured through 2D-DOSY spectra are close to 1.26  10-6 cm2/s. The measured diffusion rates, although not differentiating between 2:1 and 2:2, are consistent with formation of a capsuleplex. The 1H NMR spectra of 2:1 capsuleplex of 6-10 with OA are provided in Figure 4. While guests 6-9 could be included easily within OA by simple combined stirring of the host and guest in water (borate buffer) for less than 15 min, the longest molecule 10 does not easily complex to OA. In the case of 10, 1H NMR signals due to free guest could be seen in the spectrum even after sonication of the guest and host at 55 °C for more than 2 h (Figure 4, trace VII). The observed negative chemical shifts of the alkyl chain in all cases (6-10) suggest the occupancy of the phenyl group and the alkyl chains in the broader middle region and at the narrower top and bottom parts of the capsule, respectively, with the latter being magnetically shielded by the aromatic groups of the cavity’s walls. The single signal for each set of hydrogen on the two alkyl chains (attached at the 1 and 4 positions of the phenyl group) suggest identical average conformation of the alkyl chains at both the top and bottom halves of the capsule. The most important result relates to the methyl groups’ chemical shift that depends on the alkyl chain length (Figure 4). It is interesting to note that in a plot of the chemical shifts of the CH3- and the adjacent two CH2- groups (CH3-CH2-CH2-) against the extended length of the molecule (Figure 5) the chemical shifts increase from 5 to 8 (∼10 to 17 A˚) but decrease in guests with extended length >19 A˚ (9 and 10). All three sets of hydrogens (CH3-CH2-CH2-) show the same trend. First, let us analyze the trend observed for 6-8. The CH3- and the two adjacent CH2- groups of guests whose linear dimensions are smaller than the capsule as in 6 would be occupying the middle DOI: 10.1021/la901367k

10579

Article

Jayaraj et al.

Figure 4. 1H NMR spectra of the 2:1 complexes of OA and linear guests; the peaks marked a-h represent the proton resonances of free OA; the peaks of a0 -g0 represent the proton resonances of OA in 2:1 capsular complexes. (I) OA (1 mM, 10 mM buffer), (II) OA:5 = 2:1; (III) OA:6 = 2:1; (IV) OA:7 = 2:1; (V) OA:8 = 2:1; (VI) OA:9 = 2:1; (VII) OA:10 = 2:1, and the free guest and host.

Figure 5. Chemical shift of methyl and methylene groups for 5-10 @OA2 versus the molecular length in their extended conformation.

region of the capsule, leaving considerable free space above and below the alkyl chain. In this case, the guest would most likely be in “free tumbling motion” moving up and down the capsule. 10580 DOI: 10.1021/la901367k

This would result in average shielding (of the alkyl hydrogens) that would be lesser than if it had been stationary and occupied the narrow end of the capsule. As the chain gets longer as in 7 and 8, these (CH3 and the two CH2) groups would occupy narrower ends of the capsule as is evident from their further upfield shifts (Figure 5). Interestingly, when the length of the molecule is longer than 19 A˚ as in 9 and 10, these groups experience less shielding than 8, suggesting that they no longer occupy the narrower end. On the basis of the similar effects of all three groups of molecules, we believe the terminal alkyl chains (CH3-CH2-CH2-) are held in a coiled/folded fashion when they reach lengths that would not fit linearly within the capsule. Such conformation would place the CH3 and the two CH2 groups well above the narrower ends of the capsule that provide maximum magnetic shielding. Similar coiling/folding of long alkanes within water insoluble synthetic receptor capsules have been reported in the literature.46,47 Results reported here complement the recent observations of the inclusion of alkanes within OA.37 (46) Scarso, A.; Trembleau, L.; Rebek, J. Angew. Chem., Int. Ed. 2003, 42, 5499– 5502. (47) Scarso, A.; Trembleau, L.; Rebek, J. J. Am. Chem. Soc. 2004, 126, 13512– 13518.

Langmuir 2009, 25(18), 10575–10586

Jayaraj et al.

Article

Figure 6. 2D-NOESY NMR spectrum of (a) 2:1 (OA:9) and (b) 2:1 (OA:7) capsuleplexes. [OA] = 5.0 mM, [9] = 2.5 mM, and [7] = 2.5 mM.

The above model is supported by correlations seen in 2DNOESY spectra provided in Figure 6 (see Figure S10 in the Supporting Information for 2D-NOESY spectra of 8). From Figure 6, the strong correlations of the terminal methyls of 7 with Hg and Hf of OA are clear (see Figure 1 for the notations of OA hydrogens). In contrast, the stronger correlation with Hg and Hf of OA and 9 is not with the terminal methyl groups but with adjacent methylenes (CH3-CH2-CH2-). These different correlations indicate the differences in the relative locations of the methyls and methylenes in the guests 7 and 9 within OA. Thus, 2D-NOESY data are consistent with the model that in 7 the methyl groups occupy the narrower end of the capsule and it is clearly not in the same position in the case of 9. The smallest guest molecule 1,4-diethylbenzene (5) behaved quite differently from the rest of the members in the series 6-10. 1 H NMR titration spectra of 5 with OA are shown in Figure 7. Depending on the ratio of OA to 5 in solution, at least two types of complexes were formed. When the ratio of OA to 5 is more than 2:1 (i.e., excess host), the signals due to methyl and methylene hydrogens are sharp and upfield shifted, suggesting inclusion of 5 within OA. Integration of signals due to methyl hydrogens of 5 and Hf of OA is consistent with a host to guest ratio of 2:1. The 2D-NOESY spectrum provided in Figure 8a indicating stronger correlation of the CH3 group with He and Hd is different from 7 with strong correlation with Hg suggestive of them being located at different depths within the capsule. A plot of the chemical shift of the CH3 group of the guest included in OA in water shown in Figure 5 includes the data for guest 5 as well. The linear increase in the extent of chemical shift of CH3 from 5 to 8 is consistent with the model that the average depth of penetration of the CH3 group into the capsule is linearly related to the length of the alkyl chain. Based on the above data, we conclude that when guest 5 occupies the capsule (2:1 complex), the methyl group is far from the widest part of the capsule. When the relative concentration of the guest was increased, two additional broad peaks appeared in the 1H NMR titration spectra (Figure 7) and the chemical shift due to methyl was more upfield shifted than in the case of 2:1 capsuleplex. When excess guest (host to guest, 2:3) was used, initial sharp signals were completely replaced by two broad signals, one due to methyl and the other due to methylene hydrogens. The broad nature of the signals precluded determination of the host to guest ratio by integration Langmuir 2009, 25(18), 10575–10586

Figure 7. 1H NMR titration spectra of 5 with OA. (I) OA:5 = 2.0:1.1; (II) OA:5 = 2.0:1.8; and (III) OA:5 = 2.0:3.0. [OA] = 1.0 mM.

of host and guest NMR signals. However, we believe the spectrum represents a 2:2 capsuleplex. The strong correlation between methyl hydrogens of 5 and Hg of OA as seen in the 2D-NOESY correlation spectrum provided in Figure 8b is quite different from that seen with 2:1 capsuleplex (Figure 8a). We postulate that in 2:2 capsuleplex two molecules of 5 are aligned one on top of another as in the case of anthracene within OA (like H-type aggregates). Most likely, the two molecules slide on top of each other in the NMR time scale, leading to a spread in the chemical shift and broadening of the guest signals. If this were true, the slowing of the sliding process would result in nonidentical methyl groups on either end of 5, one below and the other away from the phenyl unit, wherein they would be expected to have different chemical shifts. The slowing of the sliding could be achieved by lowering the temperature. Indeed, when the temperature is lowered to 5 °C, the methyl signals are split (Figure 9b), supporting our model. Figure 9a showing 1H NMR spectra recorded at a H:G ratio DOI: 10.1021/la901367k

10581

Article

Jayaraj et al.

Figure 8. 2D-NOESY NMR spectrum of (a) OA:5 = 2:1 ([5] = 2.6 mM) and (b) OA:5 = 2:2 ([5] = 5.2 mM) capsular complex. [OA] = 5.0 mM.

Figure 9. Partial 1H NMR spectra of (a) 2:1 and (b) 2:2 capsuleplex of 5 within OA at different temperatures (OA, 1.0 mM; 5, 0.6 mM; buffer,

10.0 mM). (I) 5 °C; (II) 15 °C; (III) 25 °C; (IV) 35 °C; and (V) 41 °C.

< 2:1 at 5 °C supports the expectation of no effect of temperature in 2:1 capsuleplex having a single olefin within the capsule. The dual behavior of 1,4-diethylbenzene is somewhat surprising and also interesting. The behavior of these molecules is similar to that of octane that is also reported to form 2:1 and 2:2 complexes.37 In conclusion, examination of linear 1,4-dialkyl benzenes 5-10 has brought to light interesting features of OA-guest complexation: (a) None of the 1,4-dialkyl benzenes that lack ionic head groups formed cavitandplex; all molecules formed only capsuleplex. (b) Molecules that are longer than 12 A˚ preferred a 2:1 rather than a 2:2 capsuleplex. (c) 5, the smallest member of the series, formed both 2:1 and 2:2 capsuleplexes depending on the relative 10582 DOI: 10.1021/la901367k

concentrations of host and guest in solution. This observation is very useful when planning reactions of small molecules within OA. (d) When the guest molecule is too long to fit in a linear fashion, they adopt a coiled or folded conformation. Anthracene and Naphthalene. We showed in our earlier studies in collaboration with Gibb’s group on the inclusion of aromatic molecules within OA that anthracene (11) and naphthalene (12) formed 2:2 capsuleplex while pyrene and tetracene formed 2:1 capsuleplex.45 Given the similarity in size of anthracene and naphthalene to 1,4-diethylbenezene, we re-examined their complexation with OA, particularly at relatively lower concentrations of the aromatic molecules. Results with anthracene Langmuir 2009, 25(18), 10575–10586

Jayaraj et al.

are presented first. The 1H NMR titration spectra of anthracene with OA are presented in Figure 10 (for a more extensive titration curves, see Figure S11 in the Supporting Information). In Figure 10, the five signals due to anthracene noncovalent dimer are marked with a triangle (trace v). However, at very low concentrations of anthracene, none of these signals are seen (Figure 10, trace ii); instead, new 1H signals due to OA (in addition to free OA) are seen, suggesting a complex formation with anthracene but not a 2:2 capsuleplex. Careful examination of the spectra reveals three signals due to anthracene (marked with filled circles) that we believe correspond to a 2:1 OA-anthracene capsuleplex. Upon further addition of anthracene, these signals and the expected five signals due to 2:2 capsuleplex are seen in the 1 H NMR spectrum (Figure 11, trace iii); likewise, 1H OA signals due to both 2:2 and 2:1 capsuleplexes are visible. A comparison of traces iii and iv in Figure 11 representing anthracene H10 and anthracene D10, respectively, brings out the existence of the two sets of OA signals. Note the absence of anthracene 1H signals in trace iv. The 2D-COSY correlation spectrum shown in Figure 11 illustrates that, at an OA to antharcene concentration ratio of 1.7 to 1, the two types of capsuleplexes (2:1 and 2:2) coexist. From the signals marked with filled circles (2:1 complex) and triangles (2:2 complex), it is clear that the capsuleplexes correlate well among themselves but not with each other. Emission spectra of anthracene recorded at different OA-anthracene concentrations provided the final confirmation for the existence of 2:1 capsuleplex (Figure 12). The emission observed mainly from excimer with OA to 11 at a concentration ratio of 1:1 that gave way to monomer emission with further addition of OA changed to the mostly monomer emission at 5:1 ratio. This change suggests that, at high and low concentrations of anthracene relative to OA, 2:2 and 2:1 capsuleplexes are preferred, respectively. We believe that the NMR and emission data presented above suggest that anthracene such as adamantane and 1,4-diethylbenezene forms two types of capsuleplexs (2:1 and 2:2) depending on the hostguest concentrations. Thus, we have established that the three guest molecules, all smaller/closer to the depth of the OA cavity, namely, adamantane (5 A˚), 1,4-diethylbenezene (9.8 A˚), and anthracene (9.2 A˚), form both 2:1 and 2:2 capsuleplexes. Based on these results, contrary to our expectation that naphthalene (6.8 A˚), smaller than 1,4-diethylbenezene and anthracene, would form both 2:1 and 2:2 capsuleplexes, only the latter type was formed (according to NMR spectra) at all host-guest concentrations (Figure S12 in the Supporting Information) with a mixture of monomer and excimer emissions. These observations suggested that the molecular dimension alone is insufficient to predict whether a given molecule would form 2:1 or 2:2 capsuleplex. We believe weak interactions between host-guest, guest-guest, and guest-medium have a likely role in determining the nature of the complex formed. A full understanding of this phenomenon requires correlation of the structure of guests with binding constants between OA and guest molecules. The stable complexes with all guests at least in the NMR time scale that we have thus far investigated do not allow these measurements. We hope our future investigations will help better understand factors governing host-guest binding for full exploitation of OA as a reaction cavity. We have examined a number of 2-alkoxynaphthalnes with varying chain length and have shown that 2-methoxynaphthalne forms a 2:2 complex while 2-hexyloxynaphthalene forms a 2:1 complex.43 As discussed with 1,4-dialkylbenzenes, the length of the molecule determines whether a given guest forms a 2:1 or 2:2 complex. What is not obvious is whether all molecules would form both types of complexes, especially smaller ones. Why naphthalene does not Langmuir 2009, 25(18), 10575–10586

Article

Figure 10. 1H NMR spectra of anthracene in octa acid. (i) OA (1 mM in 10 mM buffered D2O), (ii) OA:anthracene 10:2, (iii) OA: anthracene 10:6, (iv) OA:anthracene-d10 10:6, (v) OA:anthracene 10:10, and (vi) OA:anthracene-d10 10:10. The aromatic host resonances of free OA, 2:1 (H:G), and 2:2 capsular assemblies are denoted by a-h, (b) represents the guest resonances of 2:1 (H:G) assembly, and (2) represents the guest resonances of 2:2 capsular assembly.

form a 2:1 complex is not obvious. We speculate that its size being too small does not provide enough van der Waals energy to stabilize the 2:1 complex. On the other hand, anthracene being slightly longer is able to offer enough stabilization energy to form both 2:1 and 2:2 complexes. We recognize that further experimentation alone can provide support to this speculation. Aggregation of Octa Acid in Water. Aggregation of hydrophobic organic molecules in water is a known phenomenon.48-54 However, structures of these aggregates have not been well characterized. On the other hand, amphiphilic organic molecules such as cholic acid and sodium dodecyl sulfate (and related detergent molecules) assemble in water to yield structurally well-defined water-soluble aggregates.55,56 Such aggregates (also known as micelles) serve as hosts to solubilize organic guests in water. OA with eight COOH groups that could be considered a large amphiphilic organic molecule might show a tendency to aggregate. During our studies with OA, we noticed that, in (48) Breslow, R. Acc. Chem. Res. 1991, 24, 159–164. (49) Syamala, M. S.; Ramamurthy, V. J. Org. Chem. 1986, 51, 3712–3715. (50) Ito, Y.; Kajita, T.; Kunimoto, K.; Matsuura, T. J. Org. Chem. 1989, 54, 587–591. (51) Catalan, J.; Zimanyi, L.; Saltiel, J. J. Am. Chem. Soc. 2000, 122, 2377–2378. (52) Bikadi, Z.; Kurdi, R.; Balogh, S.; Szeman, J.; Hazai, E. Chem. Biodiversity 2006, 3, 1266–1278. (53) Ebbing, M. H. K.; Villa, M.-J.; Valpuesta, J.-M.; Prados, P.; de Mendoza, J. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 4962–4966. (54) Li, G.; McGown, L. B. J. Phys. Chem. 1994, 98, 13711–13719. (55) Bohne, C. Langmuir 2006, 22, 9100–9111. (56) Fendler, J. H.; Fendler, E. J. Catalysis in Micellar and Macromolecular Systems; Academic Press, Inc.: New York, 1975.

DOI: 10.1021/la901367k

10583

Article

Jayaraj et al.

Figure 11. 2D-COSY NMR spectrum of a solution containing 2:1 (H:G) and 2:2 capsular assemblies of OA:anthracene. [OA] = 5 mM and [anthracene] = 3 mM.

Figure 12. Emission spectra of anthracene in octa acid. (i) OA: anthracene 1:1, (ii) OA:anthracene 2:1, (iii) OA:anthracene 3:1, (iii) OA:anthracene 4:1, and (iv) OA:anthracene 5:1. [Anthracene] = 10-5 M. λex = 350 nm.

the absence of guests, it showed a tendency to aggregate in water. Recognizing the importance of self-aggregation in exploiting OA as a reaction vessel, we probed it through 1H NMR experiments. In Figure 13vi providing the 1H NMR spectrum of 0.5 mM OA in 50 mM Na2B4O7 buffer, the aromatic and aliphatic proton resonances of OA are reasonably sharp. However, when the concentration of OA was increased to 5.0 mM, although the solution remained transparent suggesting OA is still soluble in water, as shown in Figure 13v to i, the 1H NMR spectra became broader, suggesting aggregation of OA molecules in water. Consistent with the above inference, the diffusion constant of OA decreased from 1.77  10-6 cm2/s at 0.5 mM (50 mM buffer) to 1.16  10-6 cm2/s at 5 mM (50 mM buffer). From 1H NMR spectra taken at various concentrations, we conclude that OA aggregation occurs beyond 1 mM in water. The exact structure of the aggregate and the cause for aggregation are yet to be understood.57 (57) Podkoscienlny, D.; Philip, I.; Gibb, C. L. D.; Gibb, B. C.; Kaifer, A. E. Chem.;Eur. J. 2008, 14, 4704–4710.

10584 DOI: 10.1021/la901367k

Figure 13. 1H NMR spectra of octa acid (OA) at different concentrations in 50 mM borate buffer (Na2B4O7, D2O, 500 MHz, 27 °C, Normalized intensity), diffusion constants are given in brackets; (i) 5 mM (1.1610-6 cm2/s), (ii) 4 mM (1.2510-6 cm2/s), (iii) 3 mM (1.4610-6 cm2/s), (iv) 2 mM (1.6010-6 cm2/s), (v) 1 mM (1.8010-6 cm2/s) and (vi) 0.5 mM (1.7710-6 cm2/s). Proton resonances of the host are denoted (a-j). Langmuir 2009, 25(18), 10575–10586

Jayaraj et al.

Article Scheme 3. Schematic Representations of Guest Inclusion within a Cavitanda

a Guests with ionic substituents form 1:1 cavitandplex, long guests form 2:1 capsuleplex, and small guests form 2:1 and 2:2 capsuleplexes.

Figure 14. 1H NMR spectra of 4,40 -dimethoxybenzophenone (DMBP)@(OA)2 at different concentrations in 50 mM borate buffer (Na2B4O7, D2O, 500 MHz, 27 °C, normalized intensity); diffusion constants are given in brackets. (i) 5 mM (1.38  10-6 cm2/s), (ii) 4 mM (1.37  10-6 cm2/s), (iii) 3 mM (1.39  10-6 cm2/ s), (iv) 2 mM (1.38  10-6 cm2/s), and (v) 1 mM (1.41  10-6 cm2/s). Proton resonances of the host (a-j) and guest (/) are denoted.

Figure 15. 1H NMR spectra of octa acid (1 mM) in the presence of electrolytes (i) 10 mM Na2B4O7/D2O, (ii) 50 mM Na2B4O7/D2O, and (iii) 10 mM Na2B4O7/D2O plus 80 mM NaCl. Note the broadening of signals with increased concentration of the electrolytes sodium tetraborate and sodium chloride. Langmuir 2009, 25(18), 10575–10586

Although OA alone (in the absence of a guest) tended to aggregate, surprisingly, when it was complexed to a guest, it did not aggregate even at concentrations where OA in the absence of a guest aggregated (5 mM). We illustrate this by using 4, 40 -dimethoxybenzophenone (DMBP) as a guest. DMBP formed a 2:1 capsule with OA. The 1H NMR spectrum shown in Figure 14 is suggestive of capsule formation. Although host protons were slightly shifted, none of them were split, suggesting that the two halves of the capsule, in the presence of DMBP, are equivalent. The upfield shift (δ -0.8) of the -OCH3 group is a clear indication of the inclusion of DMBP within the OA capsule. An important observation to note is that the 1H NMR spectra of the complex at 0.5 mM of OA and 5 mM of OA remained identical (Figure 14i to v). No broadening was observed, suggesting that the guest-host complex does not aggregate in the concentration range investigated. Consistent with this, the diffusion constant for DMBP@(OA)2 was measured by PGSE technique to be in the range of 1.37  10-6-1.41  10-6 cm2/s. From these two experiments, it is clear that while free OA remain as aggregates at 5 mM concentration in water, when it holds a guest within it remains as an individual complex. We find this phenomenon to be general, and we have found that all neutral molecules investigated deaggregated the OA clusters. It is also important to note that no aggregation of guest@OA occurred at least up to 3 mM host-guest complexes. We plan to further investigate this process at higher concentrations of guest@OA complexes. Aggregates of free OA formed at 5 mM could be dispersed when a neutral guest molecule is added to the aqueous solution. For example, with gradual addition of DMBP to aggregated OA (at high 5 mM in 50 mM sodium tetraborate buffer), the 1H NMR spectrum became sharper, suggesting progressive deaggregation of the host from encapsulation of the guest (see the Supporting Information, Figure S13). This observation suggests that the existence of aggregates of OA is not of concern, as they could easily be disrupted (by the use of a neutral guest) to form individual guest-host complexes. We conclude this section by pointing out that aggregation of OA even at low concentrations could be induced by addition of inorganic salts to the aqueous solution. The sharp 1H NMR spectrum of OA at 1 mM in the presence of sodium tetraborate (10 mM) was replaced by a broad spectrum at 50 mM sodium tetraborate (Figure 15) and also occurred when 80 mM sodium chloride was added to the OA solution containing 8 mM buffer, suggesting that excess inorganic salts in solution favor DOI: 10.1021/la901367k

10585

Article

Jayaraj et al.

aggregation of free OA. Consistent with this conclusion, the diffusion constant of the OA (1 mM, 10 mM buffer) decreased from 1.80  10-6 to 1.15  10-6 cm2/s (for further details, see Figures S16 and S17 in the Supporting Information) upon addition of sodium chloride (80 mM). It should be of interest to note that DMBP@OA2 ([OA] = 1 mM in 10 mM borate buffer) could not be induced to aggregate by such high concentrations of sodium chloride or sodium tetraborate (Figure S18 in the Supporting Information); no change was noticeable either in the 1H NMR spectrum or in the diffusion constant (1.48  10-6 cm2/s). The results presented in this section clearly suggest that free OA has a tendency to aggregate while the OA-guest complex does not. Although the exact mechanism of aggregation or structure of OA aggregates is yet to be deciphered, the current studies provide the useful caution on the possibility of formation of aggregates.

Summary We have demonstrated that organic molecules with ionic as well as neutral head groups form complexes with OA by examining the complexation behavior of several guest molecules with OA. Two types of complexes are formed: One of these we term as cavitandplex, an open complex in which a part of the guest is exposed to water; in the second one called capsuleplex, the host completely surrounds the guest and protects it from the aqueous exterior. Two types of capusuleplexes, depending on the number of guest molecules (two or one) within a capsule, have been observed. As illustrated in Scheme 3, molecules with ionic head groups tend to form cavitandplex and all other molecules form a

10586 DOI: 10.1021/la901367k

capsuleplex. Molecules longer than 11 A˚ form 2:1 capsuleplex, and those shorter than 10 A˚ form both 2:1 and 2: 2 complexes. The extent of these complexes depends on the concentration of the guest and host in solution. Naphthalene is an exception in this context. Smaller molecules having considerable free space within the capsule experience substantial freedom of motion while the larger ones with very little freedom are often forced to adopt folded/coiled conformations. Free OA (in the absence of guest molecules) has a tendency to aggregate, whose extent is dependent on its concentration and presence of electrolytes in solution. We expect the information obtained from this study to be of great value in predicting the nature of complexes with a given guest and facilitating appropriate guest chosen by researchers. Having demonstrated the nonpolar nature of the interior of the guest occupied OA capsule in an earlier study, the results of our studies on the dynamics of guests within the capsule and the guest-host complexes will be presented in a forthcoming publication. Acknowledgment. V.R. and R.S.H.L. thank the National Science Foundation, USA for financial support (CHE-0848017 and CHE-0600795, respectively). Supporting Information Available: Experimental details, NOESY correlation spectra for 8@OA, 1H NMR titration spectra of host/guest complexes, and 1H NMR spectra of OA under various conditions demonstrating the aggregation phenomenon. This material is available free of charge via the Internet at http://pubs.acs.org.

Langmuir 2009, 25(18), 10575–10586