Investigation of the Interactions Involved in the Formation of

Article pubs.acs.org/Langmuir

Investigation of the Interactions Involved in the Formation of Nanotubes from Organogelators Ahmad Nawaz Khan,† Thi-Thanh-Tam Nguyen,‡ Larisa Dobircau, Marc Schmutz, Philippe J. Mesini,* and Jean-Michel Guenet* Institut Charles Sadron, CNRS-Université de Strasbourg 23 rue du Loess BP84047, F-67034 Strasbourg Cedex 02, France S Supporting Information *

ABSTRACT: Investigations into the formation of nanosized structures, particularly nanotubes, by a diamide ester compound are reported. Two aspects are concurrently examined: the role of the solvent and the role of the alkyl chain. The former is addressed by using a benzene derivative (o-xylene) and a totally saturated double ring (trans-decahydronaphthalene) whereas the latter is achieved by replacing the hydrogenous alkyl chain with its fluorinated counterpart while keeping the overall architecture the same. The thermodynamic behavior by differential scanning calorimetry, the morphology by transmission electron microscopy, and the structure by X-ray scattering and small-angle neutron scattering are studied. Despite the identical architecture, the fluorinated molecule does not produce any nanotubes, unlike its totally hydrogenous counterpart. Also, o-xylene prevents the hydrogenous molecule from forming nanotubes, while nanotapes are produced instead. Conversely, the fluorinated molecule produces regularly twisted protostructures in either solvent. Neutron scattering experiments show that the fluorinated alky chain is located within the core of this structure. This suggests that the prerequisite for forming nanotubes relies on the necessity of the alkyl group to point outward.

■

INTRODUCTION Controlling the self-assembly of small molecules in generating nanosized architectures, using the supramolecular chemistry approach, is still a topical and challenging research field.1−9 The effect of noncovalent interactions such as hydrogen bonding, π stacking, and dipolar, van der Waals, and electrostatic interactions are the driving forces to play with in controlling molecular self-assembly, providing a variety of structures with reversible properties. Therefore, certain organic molecules tend to form organogels in appropriate solvents, thus forming nearly 1D structures such as fibrils, ribbons, nanotubes, and the like. The growing interest in these systems lies in their potentiality for elaborating functional hybrid materials. Among the library of a wide variety of molecules, diamide compounds self-assemble in organic solvents through the formation of hydrogen-bonding interactions.9,10 Mésini et al.11,12 have recently synthesized a diamide ester containing a bulky aromatic decyl ester (Scheme 1), which has been found to form nanotubes with extremely well-defined inner and outer diameters in the range of 25−30 nm depending on the chemical structure. The self-assembly of these diamide esters is known to be controlled by the intermolecular and intramolecular hydrogen bonds among the amide groups and the π stacking between the aromatic parts. The alkyl chains might as well play a decisive role, which is something that remains unclear. The solvent type is also largely involved in the genesis of nanotubes with even conformer effects as reported by Dasgupta et al.13 between trans-decahydronaphthalene wherein © 2013 American Chemical Society

Scheme 1. Chemical Structures of BHPB-10 (Top) and BHPBF (Bottom)

nanotubes are produced and cis-decahydronaphthalene wherein lamellar structure are obtained. Some authors have already reported the effect of the solvent type on the morphology of self-assembled systems.14−17 The reason that nanotubes are formed is still not understood. Investigations by Mésini and coworkers have shown that they are generated by the warping and the edge-to-edge fusion of ribbons,11,12 yet the molecular structure responsible for this effect is not elucidated. Received: October 16, 2013 Revised: December 5, 2013 Published: December 9, 2013 16127

dx.doi.org/10.1021/la404002k | Langmuir 2013, 29, 16127−16134

Langmuir

Article

SAXS and WAXS. Investigations were performed using a Nanostar diffractometer (Bruker-Anton Paar) that operates with a pinhole collimator and a wire proportional gas 2D detector. A monochromatic (λ = 1.54 Å, Cu Kα1) and almost parallel beam (divergence = 0.03°) was obtained through a confocal mirror with an advanced W/Si multilayer coating (XENOCS, SA). The size of the beam incident on the sample was close to 300 μm. The sample−detector distance was set at 22 cm, giving scattering vectors ranging from q = 0.5 to 8 nm−1, with q = 4π/λ sin(θ/2), where λ and θ are the wavelength and the scattering angle, respectively. The XRD data as a function of temperature was obtained by placing the samples into the oven set in the Nanostar diffractometer. A radioactive iron source was used for the correction of detector efficiency and solid angles whereas a silver behenate sample allowed one to calibrate the whole q range. Solvent scattering was subtracted from the samples in order to extract the signal scattered by the molecular assemblies. Neutron Scattering. Small-angle neutron scattering experiments (SANS) were performed on D11, a camera located at Institute LaueLangevin (Grenoble, France). A wavelength of λm = 0.6 nm was used with Δλ/λm = 9%. A built-in 2D sensitive detector was used (further details available on ILL Web site http://www.ill.fr). By varying the sample−detector distance (34, 10, and 4 m), we accessed the following q range: 0.03 < q (nm−1) < 2.5, where q = (4π/λ) sin(θ/2), with θ being the scattering angle. The cell efficiency correction was achieved by dividing the scattered intensities by means of the light water spectrum, which also allows for absolute calibration through a knowledge of the light water cross section dΣ/dΩ determined experimentally for D11 at the used wavelength (dΣ/dΩ = 0.985 cm−1 for λ = 0.6 nm). Correction for transmission and subtraction from the intensity scattered by the empty cell were systematically performed prior to solvent scattering and incoherent background subtraction. The incoherent scattering arising from the hydrogen atoms of the hydrogenous species was calculated by means of an experimentally derived relation13

This article aims at gaining further insight into the understanding of the nanotube formation by playing with the solvent type as well as the chemical composition of the diamide compounds while keeping the molecular architecture the same. This has been done by replacing the alkyl group with a fluorinated counterpart. In doing so, part of the diamide compound is expected to experience differing interaction with the solvent molecules. In particular, one of the pending questions in these systems is whether the alkyl part lies inside or outside the nanotube. We further intend to determine whether fluorination of part of the diamide molecule permits the formation of nanotubes. To shed some light on these issues, we have studied the thermodynamic behavior (T−C phase diagram), the morphology (transmission electron microscopy), and the structure (neutron and X-ray scattering) of the two diamide compounds portrayed in Scheme 1.

■

EXPERIMENTAL SECTION

Materials. The molecules used for this study were synthesized in our group. One is 3,5-bis(5-hexylcarbamoylpentyloxy)-benzoic acid decyl ester, abbreviated as BHPB-10 and shown in Scheme 1 (top). The second is a fluorinated molecule obtained by replacing the hydrogenous ester moiety by a fluorinated analogue. This will be designated as BHPBF in what follows (Scheme 1 bottom). The synthesis of BHPB-10 is described in detail elsewhere11,12 whereas fluorinated molecule BHPBF was prepared according to the method developed for other analogues.11,12 The solvents used were o-xylene and trans-decahydronaphthalene. All solvents of high-purity grades were purchased from Aldrich and used without further purification. Deuterated counterparts were purchased from Aldrich (D10-o-xylene) and from Cambridge Isotopes (D18 trans-decahydronaphthalene), respectively. Typically, the systems were prepared by adding a weighed amount of BHPB-10 or BHPBF to the solvent and heating to 100 °C for a few minutes to form homogeneous solutions while gently stirring the solution. They were then cooled to an appropriate temperature so as to form fibrillar structures and, in most cases, an organogel gel. For each set of experiments, the exact preparation conditions will be detailed accordingly. Techniques. Differential Scanning Calorimetry. Gel melting and gel formation were investigated using Diamond DSC from PerkinElmer. Heating and cooling rates ranging from 2.5 to 20 °C/ min were used. Approximately 30 mg of gels was put into stainless steel sample pans that were hermetically sealed by means of an O-ring so as to prevent solvent evaporation. For the lowest concentration (0.001 g/cm3), a μDSC by Setaram was used. About 600 mg of sample was scanned at a heating and cooling rate of 0.1 °C/min. In the T−C phase diagram, insets display the variation in the enthalpy as a function of concentration (Tamman’s diagram). As is customary, enthalpies are then always given per gram of gel or solution, never per gram of solute. This is what is actually measured. Transmission Electron Microscopy. A freeze−fracture replication technique was used to perform TEM investigations. Pictures were taken on a Tecnai G2 (FEI) microscope operating at 200 kV using an Eagle 2k (FEI) ssCCD camera. For high gel concentrations (∼3 wt %), cryo fracture was performed. The gels were inserted between two copper cupules, frozen in liquid nitrogen, and then cleaved under ultrahigh vacuum by means of a homemade cryo-fracturing apparatus. We used liquid nitrogen as a cryogenic fluid because the standard one (liquid ethane) is miscible with the solvents used to form the gels. The cooling rate is around 10 000 K/s, allowing vitrification of the solvent and preventing structural alteration. The inner surface of the gel is shadowed with a thin (2 nm) layer of platinum at 45° and finally a thick (20 nm) carbon layer was deposited at 90°. These samples were rinsed with chloroform to remove the original molecules, and the replicas were placed onto 400 mesh copper grids.18

Iinc = φH × 8.65

NH VH

(1)

in which φH is the hydrogenous species volume fraction, NH is the number of protons per hydrogenous species, and VH is their molar volume. Absolute calibration was achieved by calculating the contrast factors for the molecules under study. These contrast factors are more complex for the fluorinated molecules because the two types of moieties display differing scattering amplitudes. Conversely, for hydrogenous BHPB-10 the contrast factor is expressed through

K=

(ai − yas)2 NA mi 2

(2)

where a represents the scattering amplitudes of the molecules and the solvents, y is the ratio of the molar volumes of the same, mi is the molecular weight of the molecules, and NA is Avogadro’s number. The absolute intensity was simply obtained by dividing the normalized experimental intensity by K.

■

RESULTS AND DISCUSSION Results obtained for the thermodynamic properties of these systems are presented first, particularly the phase diagrams in the low-concentration range, which is the domain of interest for forming fibrillar organogels. The choice of solvents relies on their propensity to interact with the molecules: o-xylene is prone to interact with the benzyl-type ring moiety whereas trans-decahydronaphthalene would rather interact with the aliphatic arms moiety. It has been reported in a previous paper that BHPB-10 forms hollow nanotubules in the latter solvent.13 Thermodynamic Properties: Temperature−Concentration Phase Diagrams. Systems in ortho-Xylene. For BHPB-10/o-xylene systems (Figure 1c), the melting endo16128


Langmuir

Article

Figure 1. DSC thermograms recorded on cooling and on heating for (a) BHPBF/o-xylene, (b) BHPBF/trans-decahydronaphthalene, and (c) BHPB-10/o-xylene.

therms and the formation exotherms display a single peak unlike BHPBF/o-xylene systems for which two peaks are observed (Figure 1a). It is worth emphasizing that the surface ratio between the two endothermic peaks in BHPBF/o-xylene systems is independent of the heating rate, which leads us to disregard possible annealing effects such as a melting− recrystallization process19 and rather points to the existence of either two types of structures or a phase transition. We shall discuss this further. Notably, melting and the formation of BHPBF/o-xylene systems are shifted to higher temperatures than for the BHPB-10/o-xylene systems. This suggests that oxylene is a poorer solvent for BHPBF than for BHPB-10. The corresponding temperature−concentration (T−C) phase diagrams mapped out in the concentration range of 0.001 to 0.05 (w/w) are shown in Figure 2a,b The insets display the corresponding enthalpies related to the melting endotherms and the formation exotherms. As expected, endothermic and exothermic enthalpies are virtually identical for each system yet slightly smaller for BHPBF than for BHPB10, implying that the associated melting entropies are significantly smaller for BHPBF than for BHPB-10. Systems in trans-Decahydronaphthalene. In the BHPBF/ trans-decahydronaphthalene binary system, two endothermic peaks appear above C = 0.5 wt % whereas a single endothermic peak occurs below C = 0.5 wt % concentration (Figure 1b). Similarly, two exothermic peaks occur above C = 0.75 wt % whereas a single exothermic peak is seen below. Results previously published have shown that only one endotherm/ exotherm is seen for BHPB-10/trans-decahydronaphthalene systems. The T−C phase diagram of BHPBF/trans-decahydronaphthalene systems for up to 5 wt % is shown in Figure 3, together with the melting temperatures reported elsewhere for BHPB-10/trans-decahydronaphthalene systems.13 The melting and formation temperatures of BHPBF/trans-decahydronaphthalene are very high as compared to those of other binary systems, suggesting that trans-decahydronaphthalene is the poorest solvent for BHPBF among those studied here. In this solvent, the enthalpies and entropies are smaller for BHPBF than for BHPB-10. In the case of BHPBF/trans-decahydronaphthalene, it is worth noting that the melting/formation temperature are virtually nonvariant with concentration. Such a situation is reminiscent of the occurrence of a phase transition. In this range of rather low concentrations, the most relevant type of phase separation worth contemplating is a liquid−liquid phase separation.20,21 The first peak occurring on heating would be then related to a monotectic transition due to the presence of a miscibility gap, followed by a second peak that represents the

Figure 2. Temperature−concentration phase diagrams: (A) BHPB10/o-xylene and (B) BHPBF/o-xylene. Insets show the corresponding formation and melting enthalpies. Open symbols correspond to the cooling process, and full symbols correspond to the heating process.

Figure 3. Temperature−concentration phase diagram BHPBF/transdecahydronaphthalene. For the sake of comparison, the diamond symbols stand for the melting temperatures of BHPB-10/transdecahydronaphthalene systems from previously published results.13 The inset shows the corresponding formation and melting enthalpies. Open symbols correspond to the cooling process, and full symbols correspond to the heating process.

final melting of all of the structures.20 Below a given concentration, the system will not cross the miscibility gap, which would account for the observation of a single peak. 16129


Langmuir

Article

That the BHPBF molecules seem to undergo a liquid−liquid phase transition in both solvents is most probably because of the poor interaction with the fluorinated moiety. We find that this statement receives further support from neutron scattering experiments. Morphologies by TEM on Freeze-Fractured Samples. The observation by TEM of replicas prepared after freeze− fracture will give an image of the structures found in the gels in their solvated state. This technique, which is mainly used in biology or soft matter in water, has been successfully developed in organic solvent in our group.22 Systems in o-Xylene. In the case of BHPB-10/o-xylene systems’ TEM images (Figure 4a), a different organization of individual fibrils is observed. The arrow points to a single fiber (with a diameter of 7 ± 2 nm) that can form two different larger organized structures. One type is a high-aspect-ratio tapelike object (shown by arrowheads in Figure 4a), and the other type consists of entanglements of single fibers, thus producing twisted bundles (double arrow in Figure 4a). No formation of nanotubes is detectable, unlike what was reported in other solvents for this molecule. In the case of BHPBF/o-xylene systems, the morphology turns out to be significantly different (Figure 4b). Basically, large tapes are assemblies of regularly twisted fibers with a chevron aspect (arrowhead). These twisted fibers display a cross-sectional diameter of about 7 nm, whereas the pitch associated with the regular twist is of about 6 nm. Unlike BHPB-10, BHPBF therefore self-assembles in a conspicuous helical superstructure, which again is related to the presence of a fluorinated moiety. Systems in trans-Decahydronaphthalene. Recently published AFM investigations have revealed hollow nanotube structures formed by BHPB-10 in trans-decahydronaphthalene.13 Their outer (25.4 nm) and inner (18 nm) diameters are very well defined as further ascertained by small-angle scattering experiments with very little dispersion around these values. AFM and TEM evidence indicates that these nanotubes most probably arise from the warping of ribbons, leading to ringlet-type features. The BHPBF/trans-decahydronaphthalene systems investigated herein do not exhibit the same structure. Rather, morphologies similar to those seen in o-xylene are observed (Figure 4c): regularly twisted protofibrils again of a chevron aspect with a pitch of about 8 nm and a rather well-defined cross-sectional diameter of about 8 nm (arrowheads) and ropelike structures formed through the supertwisting of the protofibrils (double arrow). The open star shows solvent areas. Protofibrils seen in BHPBF trans-decahydronaphthalene therefore resemble to a large extent those in BHPBF/o-xylene. WAXS. The thermodynamic properties and morphology are significantly altered when the alkyl group is replaced with a fluorinated group. X-ray diffraction experiments drawn in Figure 5 for BHPBF and BHPB-10 in o-xylene systems reveal conspicuously different patterns. Note that these patterns are independent of concentration within the range investigated, namely, 0.01to 0.05 w/w. For BHPB-10/o-xylene gel system, three narrow peaks all related to intramolecular order, appearimg at q = 2.56, 3.66, and 5.1 nm−1, yield d spacings of 2.45, 1.72, and 1.23 nm, respectively. The peak at q = 5.1 nm−1 may well be a secondorder reflection of the first peak, which may suggest the existence of long-range order of the molecular organization in BHPB-10 systems.

Figure 4. TEM micrographs from freeze−fracture samples: (a) BHPB10/o-xylene systems, (b) BHPBF/o-xylene systems, and (c) BHPBF/ trans-decahydronaphthalene. Concentration C = 0.04 w/w. For details, see the text. 16130


Langmuir

Article

Figure 5. X-ray diffraction data plotted as q2I(q) vs q. ○ = BHPBF/oxylene (C = 3% w/w); ● = BHPB-10/o-xylene (C = 0.02 w/w). (Inset) Theoretical intensity calculated by means of eq A4 with P = 6 nm and rH = 3 nm.

solvents have been used. Indeed, fluorine possesses a neutron scattering length close to that of deuterium (bF = 0.574 × 10−12 cm, bD = 0.667 × 10−12 cm) and also very little incoherent scattering.27 Clearly, in either type of labeled solvent these molecules will display differing scattering power, thus providing one with more information on the molecular structure than can be obtained from SAXS. For BHPBF/o-xylene systems, the results do not depend on concentration in the range of 0.01 to 0.05 w/w. However, they depend strongly on the nature of the solvent labeling as a result of a contrast effect that is quite expected. The intensity plotted by means of a Kratky plot (q2I(q) vs q) shows a strong upturn that arises from the difference in contrast between the hydrogenous and the fluorinated moiety (Figure 6). This result is typical of what is often designated as a

For BHPBF/o-xylene systems, broader maxima are seen at q = 1.42, 2.1, 2.84, and 4.2 nm−1. This may imply the occurrence of a lower order in these systems. However, these results must be examined in light of the morphologies observed, particularly the twisted structures with a very large pitch revealed in these systems. The intensity scattered by continuous helices has been derived by Pringle and Schmitt.23 Their approach pertains to the present case and is detailed in the Supporting Information. Diffraction by helices gives rise to oscillations due to the Bessel functions in the theoretical expression of the intensity. In the case of helices with a small pitch, only the first term in the Pringle and Schmitt expression is relevant in the q range explored, namely, at low resolution, and corresponds to the scattering function of cylindrical structures.25,26 In the case of a large pitch, a second term should be considered, which is expressed as (Supporting Information)

Figure 6. Neutron scattering curve plotted by means of a Kratky plot (q2I(q) vs q). BHPBF/o-xylene-C8H10 (C = 0.02 w/w). (Inset) Same data plotted by means of a Porod plot, log qI(q) vs q2. The solid line shows the linear part of this plot, allowing one to determine the crosssectional radius.

g2(qrH , P) ∝

rH ⎤ 2 ⎡ ⎢2 − 2Jo (rH B ) − P B J1(rH B )⎥⎦ 2⎣ BrH q2P 2 − 16π 2

negative apparent square cross-section radius.19 If one considers a model of two concentric solid cylinders, the intensity is in the range of qrout < 1 (with rout being the external radius of the outer cylinder)20,28

157.9 (3) P P2 Because this term also contains Bessel functions, oscillations related to the pitch and the cross-sectional radius are again expected as seen in the inset of Figure 5. Although a much better fit would require us to take into account the contrast factors of each moieties and thus rather consider discontinuous helices, this simple approach allows one to suggest that only the peak at q = 2.84 nm−1, namely, d = 2.21 nm, may stand for the molecular arrangement of the BHPBF molecules whereas the other maxima arise from the oscillations of the Bessel functions. This is roughly the same distance as that determined for BHPB10/o-xylene systems. The BHPBF/trans-decahydronaphthalene system displays exactly the same diffraction patterns (data not shown), which suggests that the molecular arrangement is not significantly influenced by the solvent type but is essentially linked to the fluorinated moiety. X-ray patterns do not exhibit any significant alteration with increasing temperature (data not shown). This implies that the occurrence of two peaks in DSC experiments is not linked to any solid−solid phase transition but rather to a monotectic transition as already discussed above. SANS. Small-angle neutron scattering experiments have been carried out for samples in the concentration range of 0.01 to 0.05 w/w. All experiments have been performed at T = 23 ± 1 °C. In the case of BHPBF, hydrogenous and deuterated with B =

2

= q2 −

⎛ q2rout 2 ⎡ A in γ 4 + Aout (1 − γ 4) ⎤⎞ ⎥⎟⎟ ⎢ q2I(q) = πqCμL ⎜⎜1 − 4 ⎣ A in γ 2 + Aout (1 − γ 2) ⎦⎠ ⎝ (4)

where Ain and Aout are the scattering amplitudes of the inner and outer cylinders with respect to the solvent and γ = rin/rout where rin is the inner radius, C is the concentration, and μL is the mass per unit length. Under certain conditions that are fulfilled here, the contrast term in relation 4 can be negative so that departure from linearity in the Kratky representation occurs upwards instead of downwards. Indeed, if one assumes that the inner cylinder consists of the fluorinated moiety, then one can estimate Ain = 11.7 × 10−12 cm and Aout = −0.39 × 10−12 cm.29 This assumption relies on the fact that the fluorinated moiety interacts poorly with the solvents and so prefers to be hidden in the core of the composite fibrils. By using a Porod plot, log qI(q) vs q2, the intensity is expressed through log(qI(q)) ∝ log(πμL C) −

q2rout 2 ⎡ A in γ 4 + Aout (1 − γ 4) ⎤ ⎥ ⎢ 4 ⎣ A in γ 2 + Aout (1 − γ 2) ⎦ (5)

This representation yields a straight line in the low-q range (inset of Figure 6) whose slope together with the calculated 16131


Langmuir

Article

Figure 7. Sketch of the possible way in which the fluorine moieties interact (piling up onto one another or head-to-head), together with the associated lengths.

considering a mixture of concentric cylinders20 and very large ribbons (Supporting Information). When the difference in scattering amplitudes of each moiety with respect to the solvent is taken into account, the intensity is written as a Kratky representation:29

scattering amplitudes Ain and Aout allows one to derive rout provided that γ is available. With this aim, two possible arrangements, where fluorinated moieties are located in the core of the structure, are considered that allows one to estimate the values of γ and rout (Figure 7). If the fluorinated moieties pile up on top of one another, then the following parameters are obtained: γ ≈ 0.21, rout ≈ 2.9 nm, and rin ≈ 0.61 nm. By inserting the value of γ in relation 4, one then derives rout≈ 3.3 nm. If the fluorinated chains interact head-to-head, this leads to γ ≈ 0.33, rout ≈ 3.7 nm, and rin ≈ 1.2 nm, and then one derives from relation 4 rout ≈ 11 nm. The case where fluorinated chains pile up on top of one another provides one with an outer radius in far better agreement with the estimated value and with the value determined by TEM. It is worth emphasizing that no “negative apparent square cross-section radius” would be observed if the inner “cylinder” consisted of the hydrogenous moiety of the molecule. For BHPBF/o-xylene-d10, the intensity plotted by means of a Kratky plot (q2I(q) vs q) reveals conspicuous features (Figure 8). Fitting this curve is not straightforward. After numerous attempts with several models, the best fit has been obtained by

q2I(q) =

XπqCμL qrout 2

[2.13J1(qrout) − 1.97γJ1(qγrout)]2

+ (1 − X )2πA r 2 CμS

4 sin 2(qδr) q2δr 2

(6)

where X and (1 − X) are the weight fractions of the respective structures, δr is the thickness of the ribbon, and μS is the mass per unit area of a ribbon with respect to its length and width (μS = ρδr, where ρ is the density expressed in g/nm3 mol). Ar is the relative average scattering amplitude of the ribbons with respect to the solvent. As can be seen in Figure 8, the fit reproduces the features of the scattering curve in the range of q = 0.2 to 2.2 nm−1 relatively well but fails at very small q. Actually, the expression for a thick ribbon used here is no longer valid in the small-q range, but also long-range interactions are not included in this approach. However, the outcomes are in agreement with the TEM observations. The values derived from relation 5 are rout = 2.7 nm and δr = 29 nm with X = 0.18. The value of rout is also in agreement with that derived from the scattering curve obtained in hydrogenous o-xylene. For BHPB-10/o-xylene-d10 systems, the results depend on concentration but are in line with the DSC findings. For C = 0.01 w/w, part of the scattering curve is best fitted by considering rod-like structures with a very small crosssection (Figure 9). The following equation for the absolute intensity Iabs(q) has been used:20

Figure 8. Neutron scattering curve plotted by means of a Kratky plot (q2I(q) vs q). BHPBF/o-xylene-C8D10 (C = 0.02 w/w). The solid line represents a fit by means of eq 7.

q2Iabs(q) = πCμL × 16132

J12 (qrc) ⎡ 2 ⎤ π q − ⎥ ⎢ L ⎦ q2rc 2 ⎣

(7)


Langmuir

Article

Figure 9. Neutron scattering curve plotted by means of a Kratky plot (q2I(q) vs q). ● = BHPB-10/o-xylene for C = 0.01 w/w; the dotted line stands for a fit by means of eq 7 with rc = 0.6 nm; ○ = BHPB-10/ o-xylene for C = 0.05 w/w. Figure 10. Neutron scattering curve plotted by means of a Kratky plot (q2I(q) vs q). BHPB-10-trans-decahydronaphthalene-C10D18. Fit obtained by considering a hollow cylinder with rout = 12.7 nm and rin= 9 nm (eq 8). (Inset) Scattering curve for BHPBF/o-xylene-C8D10 (C = 2% w/w).

where rc is the rod cross-section radius, ⟨L⟩ is their average length, μL is its mass per unit length, C is the concentration, and J1 is the Bessel function of the first kind and first order. Values derived from the scattering curve yield μL ≈ 344 g· nm−1·mol−1, rc ≈ 0.6 nm, and ⟨L⟩ ≈ 25 nm. Neutron results therefore suggest that tiny, rodlike aggregates still exist although at T = 23 °C the T−C phase diagram indicates that the system is liquid for this concentration. The value of rc corresponds approximately to the BHPB-10 molecule cross-section, and the mass per unit length would be compatible with an ensemble of BHPB-10 molecules interacting in a head-to-tail fashion. However, the reason that it should be so remains obscure. For C = 0.05 w/w, the scattering curve is totally different and clearly shows that large aggregates are now present, an outcome that agrees both with the TEM findings and the T−C phase diagram (Figure 9). Fitting this curve is not straightforward because it contains no special features, unlike those that have been presented above. For BHPB-10/trans-decahydronahpthelene-d18, the scattering curve is again consistent with previous findings and therefore shows the occurrence of hollow nanotubes as ascertained by the fit performed with (Figure 10)13

atoms replace hydrogen atoms, nanotubes are no longer formed. It would then be of interest to perform experiments with BHPBF in fluorinated solvents because the fluorinated moiety of the molecule might then point outward and to find out whether nanotubes are formed.

■

ASSOCIATED CONTENT

S Supporting Information *

Equations for helices and ribbons. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Fax: +33 388414099. Present Addresses

⎤2 ⎡ 2 q2I(q) = 2πqCμL ⎢ { J ( qr ) J ( q r )} × − γ γ out outext ⎥ 1 1 ⎣ (1 − γ 2)rout ⎦

†

School of Chemical and Materials Engineering, National University of Sciences and Technology, H-12 Islamabad, Pakistan. ‡ LFCS, Faculté de Pharmacie, Université de Strasbourg, 67400 Illkirch, France.

(8)

For the sake of comparison, the inset in Figure 10 shows the scattering curve of BHPB-10/o-xylene.

■

Notes

The authors declare no competing financial interest.

CONCLUDING REMARKS Whether nanotubes are formed depends on the nature of the solvent, as already reported by Nguyen et al.,30 and correspondingly on the interactions with a given moiety of the organogelators used here. Aromatic solvents clearly prevent BHPB-10 from forming nanotubes, most probably arising from strong π−π interactions with the aromatic part of the molecule. Seemingly, aliphatic solvents are the best fit for producing nanotubes possibly because they favor π−π interactions and hydrogen bond formation between BHPB molecules. Also, keeping the molecular architecture unchanged while replacing hydrogen atoms on the alkyl chain by fluorine atoms totally alters the resulting molecular structure and particularly the propensity for forming nanotubes. Results herein again suggest that the alkyl chain moiety of BHPB must be allowed to point outward to generate the nanotube structure, something already suspected.11,12 When located inside, as is the case once fluorine

■

ACKNOWLEDGMENTS The technical assistance of C. Saettel (DSC), G. Fleith (SAXS and WAXS), C. Blanck (TEM), and R. Schweins (Neutrons on D11, ILL, Grenoble) was highly appreciated. The grant under the joint venture of the French government and the Higher Education Commission (HEC) of Pakistan (2-2(7) PDFPFrance/HEC/2012/01) for the postdoctoral fellowship of A.N.K. is gratefully acknowledged. A.N.K. is also indebted to CNRS for additional funding (370478). This work has also been performed as part of ANR funding through grant ANR11-BS08-001 (MATISSE project).

■

REFERENCES

(1) Lehn, J.-M. Supramolecular Chemistry: Concepts and Perspectives; VCH: Weinheim, Germany, 1995. 16133


Langmuir

Article

(2) Schneider, H.-J. ; Yatsimirsky, A. Principles and Methods in Supramolecular Chemistry; Wiley: Chichester, U.K., 2000. (3) Terech, P.; Weiss, R. G. Low Molecular Mass Gelators of Organic Liquids and the Properties of Their Gels. Chem. Rev. 1997, 97, 3133. (4) Molecular Gels: Materials with Self-Assembled Fibrillar Networks; Weiss, R. G., Terech, P., Eds.; Springer-Verlag: Dordrecht, The Netherlands, 2006. (5) Hirschberg, J. H. K. K.; Brunsveld, L.; Ramzi, A.; Vekemans, J. A. J. M.; Sijbesma, R. P.; Meijer, E. W. Helical Self-Assembled Polymers from Cooperative Stacking of Hydrogen-Bonded Pairs. Nature 2000, 407, 167. (6) Organic Light Emitting Devices; Müllen, K., Scherf, U., Eds.; VCH: Weinheim, Germany, 2006. (7) Ajayaghosh, A.; Praveen, V. K. π-Organogels of Self-Assembled pPhenylenevinylenes: Soft Materials with Distinct Size, Shape, and Functions. Acc. Chem. Res. 2007, 40, 644. (8) Ajayaghosh, A.; Praveen, V. K.; Vijayakumar, C. Organogels as Scaffolds for Excitation Energy Transfer and Light Harvesting. Chem. Soc. Rev. 2008, 37, 109. (9) Schneider, J.; Messerschmidt, C.; Schulz, A.; Gnade, M.; Schade, B.; Luger, P.; Bombicz, P.; Hubert, V.; Fuhrhop, J. H. Odd−Even Effects in Supramolecular Assemblies of Diamide Bolaamphiphiles. Langmuir 2000, 16, 8575. (10) Tomioka, K.; Sumiyoshi, T.; Narui, S.; Nagaoka, Y.; Iida, A.; Miwa, Y.; Taga, T.; Nakano, M.; Handa, T. Molecular Assembly and Gelating Behavior of Didodecanoylamides of α,ω-Alkylidenediamines. J. Am. Chem. Soc. 2001, 123, 11817. (11) Diaz, N.; Simon, F. X.; Schmutz, M.; Rawiso, M.; Decher, G.; Jestin, J.; Mesini, P. J. Self-Assembled Diamide Nanotubes in Organic Solvents. Angew. Chem., Int. Ed. 2005, 44, 3260. (12) Nguyen, T.-T.-T.; Simon, F.-X.; Khelfallah, N. S.; Schmutz, M.; Mésini, P. J. Mesoporous Polymeric Catalysts Synthesized from SelfAssembled Organic Nanotubes As Templates. J. Mater. Chem. 2010, 20, 3831. (13) Dasgupta, D.; Kamar, Z.; Rochas, C.; Dahmani, M.; Mesini, P. J.; Guenet, J. M. Design of Hybrid Networks by Sheathing Polymer Fibrils with Self-Assembled Nanotubules. Soft Matter 2010, 6, 3573. (14) Hoeben, F. J. M.; Jonkheijm, P.; Meijer, E. W.; Schenning, A. P. H. J. About Supramolecular Assemblies of π-Conjugated Systems. Chem. Rev. 2005, 105, 1491. (15) Cantekin, S.; Nakano, Y.; Everts, J. C.; Schoot, P.; Meijer, E. W.; Palmans, A. A Stereoselectively Deuterated Supramolecular Motif to Probe the Role of Solvent during Self-Assembly Processes. Chem. Commun. 2012, 48, 3803. (16) Yang, X.; Xu, X.; Ji, H.-F. Solvent Effect on the Self-Assembled Structure of an Amphiphilic Perylene Diimide Derivative. J. Phys. Chem. B 2008, 112, 7196−7202. (17) Dasgupta, D.; Thierry, A.; Rochas, C.; Ajayaghosh, A.; Guenet, J. M. Key Role of Solvent in Organogelation. Soft Matter 2012, 8, 8714. (18) Schmutz, M.; Mésini, P. Cryo-Fracture of Self-Assembled Systems in Organic Solvent. In Handbook of Cryo-Preparation Methods for Electron Microscopy; Cavalier, A., Spehner, D., Humbel, B. M., Eds.; CRC Press/Taylor and Francis: New York, 2008; pp 411−430. (19) Blundell, D. J. On the Interpretation of Multiple Melting Peaks in Poly(Ether-Ether Ketone). Polymer 1987, 28, 2248. (20) Guenet, J. M. Polymer-Solvent Molecular Compounds; Elsevier: London, 2008. (21) Reisman, A. Phase Equilibria; Academic Press: New York, 1970. (22) Schmidt, R.; Schmutz, M.; Michel, M.; Decher, G.; Mésini, P. J. Organogelation Properties of a Series of Oligoamides. Langmuir 2002, 18, 5668−5672. (23) Pringle, O. A.; Schmidt, P. W. Scattering from Helical Filaments. J. Appl. Crystallogr. 1971, 4, 290. (24) Fournet, G. Fonctions de Diffusion pour des Formes Géométriques. Bull. Soc. Fr. Mineral. Cristallogr. 1951, 74, 39. (25) Oster, G.; Riley, D. P. Scattering from Cylindrically Symmetric Systems. Acta Crystallogr. 1952, 5, 272. (26) Mittelbach, P.; Porod, G. Zur Röntgenkleinwinkelstreuung Verdünnter Kolloiden Systeme. Acta Phys. Austriaca 1961, 14, 185.

(27) Bacon, G. E. Coherent Neutron-Scattering Amplitudes. Acta Crystallogr., Sect. A 1972, A28, 357. (28) Higgins, J. S., Benoit, H. C. Polymer and Neutron Scattering; Clarendon Press: Oxford, U.K., 1994. (29) The scattering amplitudes are calculated by considering a fluorinated moiety of chemical formula C8F17 with an estimated molar volume of vF ≈ 254 cm3/mol whereas the hydrogenous moiety is C31H50N2O6 with a molar volume of vH ≈ 558 cm3/mol. (30) Simon, F.-X.; Nguyen, T. T. T.; Díaz, N.; Schmutz, M.; Demé, B.; Jestin, J.; Combet, J.; Mésini, P. J. Self-Assembling Properties of a Series of Homologous Ester-Diamides−From Ribbons to Nanotubes. Soft Matter 2013, 9, 8483−8493.

16134


Investigation of the Interactions Involved in the Formation of

Recommend Documents