Opening a Can of Worm(‐like Micelle)s: The Effect of Temperature of Solutions of Functionalized Dipeptides

Abstract A simple heat/cool cycle can be used to significantly affect the properties of a solution of a low‐molecular‐weight gelator at high pH. The viscosity and extensional viscosity are increased markedly, leading to materials with very different properties than when the native solution is used.

The process by which agel is formed using aself-assembled low-molecular-weight gelator (LMWG) is critical in determining the final properties of the gel. [1] There are many reports attempting to link the molecular structure of the gelator to the gel properties;h owever, it is possible to produce gels with very different properties from as ingle gelator by varying the process. [2] Since predicting new gelators from first principles is still difficult, [3] it is extremely useful to know that many properties,s uch as the stiffness,r ecoverability,a nd opacity can all be varied significantly using asingle,robust gelator.
Gel formation requires assembly into one dimensional structures,followed by the entanglement and/or cross-linking of these structures. [4] Hence,i ti sc ommon for this process to be kinetically driven and the method of gelation has asignificant effect on the outcome. [1a] Afurther complication is that the gelator must first be added to the solvent preassembly.F or organic solvents,i ti sc ommon to suspend the insoluble gelator in the solvent, and heat to dissolve the gelator;cooling then leads to assembly.Inwater,this process can also sometimes be achieved, but it is more usual to utilize at rigger,s uch as ac hange in pH, to go from a" dissolved" gelator to the gel. [5] In some cases,the "dissolved" gelator is in fact dispersed as asurfactant-like aggregate. [6] Thet ransition from this aggregate to gel fibers is largely process-controlled and therefore very unlikely to reach at hermodynamic minimum.
Functionalized dipeptides are highly effective hydrogelators. [7] Dipeptides functionalized at the N-terminus with an aphthalene are robust and effective gelators. [8] Thef ree C-terminus allows dispersion in water, and at rigger for gelation. At high pH, these can be dispersed by adding base to deprotonate the free carboxylic acid. Ty pically,s pherical or worm-like micelles are formed, [6b] although this is concentration dependent. [9] We have described this scenario in detail and have shown that solutions of the worm-like micelles can be gelled at high pH by adding adivalent cation. [6] Herein, we show that the properties of the solutions and resulting gels are strongly affected by asimple pre-heating step.T his effect has important implications for these gelators.
Solutions of 2NapFF (2NapFF = 2-(naphthalen-2-yl)acetamido)-3-phenylpropanamido)-3-phenylpropanoic acid; Figure 1a)w ere prepared at ac oncentration of 10 mg mL À1 by adding sodium hydroxide to the dipeptide in water. After stirring overnight, as lightly viscous solution was formed at ap Ho fa pproximately 11. We have previously shown that these solutions contain worm-like micelles. [6b, 9] These solutions are susceptible to shear forces,e xhibiting enhanced extensional viscosity after exposure to shear rates of between 10 3 and 10 6 s À1 . [10] This effect was weak and difficult to reproduce exactly.
Herein, we report that heating to elevated temperatures does not change the viscosity of the solutions (Figure 1b). On recooling,t he samples became significantly more viscous (Figure 1b), and formed materials that were self-supporting for at least 8h( Figure 1c). Overnight, the samples fell to the bottom of the tube.V ery similar materials were formed if the heating was between 40 and 70 8 8C, and samples that could be inverted were also formed after cooling from ashort period of heating with ah ot air gun. Whilst the samples become selfsupporting,itisclear from the rheological data that these are not true gels.T he storage modulus (G')d oes not dominate over the loss modulus (G'')byone order of magnitude;rather the values are very close,i ndicating am ore elastic type of material (Figure 1d), and there is some frequency dependence.T he G'' for the two solutions also highlights the significant difference in the viscosity of the two solutions near the rest state.T his zero-shear viscosity can be determined by the slope of G'' versus w in the limit of vanishing frequency or, as this terminal regime is not approached for the heat/cool cycle sample,roughly estimated with G''/w at w = 1 rad s À1 . [11] Before heating, G00 w % 0.5 Pa s, whereas after the heat/ cool cycle G00 w % 50 Pa s; that is,t wo orders of magnitude higher. All of this is clearly important as it is common for tube inversion to be shown as the only test of gelation. [12] Thei ncrease in viscosity was observed over ar ange of concentrations (Supporting Information, Figure S1), with materials that could be inverted forming at concentrations as low as 5mgmL À1 (solutions at 3and 4mgmL À1 were visibly more viscous after ah eat/cool cycle,b ut did not support inversion). Thec oncentration range coincides with our previous work on the phase diagram for 2NapFF,w here we showed that worm-like micelles are formed above around 1mgmL À1 ,w ith as ignificant increase in the viscosity above 5mgmL À1 . [9] Hence,this heat-induced behavior requires that specific aggregates exist at room temperature,rather than this being aheat-induced structural transition.
Thep Hi sacritical parameter;a t1 0mgmL À1 ,s amples that could be inverted were prepared after aheat/cool cycle as long as the pH was above 9.48 (Supporting Information, Figure S2). Ty pically,f or this kind of molecule,t he apparent pK a of the carboxylic acid is around 6, [13] and can be concentration dependent. At 10 mg mL À1 ,t itration of the 2NapFF shows an apparent pK a at around 8.5 (Supporting Information, Figure S3). After ah eat/cool cycle,t he pK a slightly increased to around 9.0, but it should be stressed that the high viscosity of this solution made this measurement difficult. We note that elsewhere heat/cool cycles have been used during ap Ht itration to establish the pK a and it is possible that this leads to errors on the basis of the current work. Nonetheless,f or the viscosity increase to be observed, the pH needs to be above the apparent pK a of the 2NapFF, implying that significant charge is required on the worm-like micelles.
Thes amples exhibited significant extensional viscosity after heating and cooling,becoming very "stringy" (Supporting Information, Figure S4). We therefore undertook experiments to quantify the resistance of the solutions to extensional/elongational deformations using ac apillary break-up extensional rheometer ("CaBER"). [14] In this device,aliquid bridge (2 mm in length) is formed between two circular discs 4mmi nd iameter,w hich are then rapidly pulled apart (ca. 50 ms). Ther esulting unstable fluid filament consequently thins down under the action of surface tension until finally breaking. Thed iameter of the filament (D) is observed as afunction of time (t)using the equipmentslaser micrometer (resolution ca. 10 microns). Although the filament diameter data can be post-processed into an (apparent) extensional viscosity,t he standard method to quantify extensional effects [15] is by an exponential fit to the filament diameter as afunction of time in the elasto-capillary regime to determine acharacteristic relaxation time (l;more correctly,acharacteristic time for extensional stress growth). Representative plots are shown in Figure 2a where the effect of the heat/cool cycle can be seen to increase this characteristic time l by almost two orders of magnitude,from 25 ms before heating to arelaxation time of 1.35 sa fter aheat/cool cycle.
To explain this behavior, we used small-angle X-ray scattering (SAXS). Before heating, the scattering data fitted  . Key:afreshly prepared solution (10 mg mL À1 ;p H11; (*)), asample after heating and cooling (* *), exponential fits (c c). b) SAXS data (intensity vs. scattering vector (Q)) for asolution of 2NapFF (10 mg mL À1 ;pH11) before heating ((*); fit (c c)) and after heating and cooling ((* *); fit (c c)). c,d) Cryo-TEM of asolution of 2NapFF before heating (c) and after aheat/coolc ycle (d);s cale bar = 200 nm.  Figure S5). After aheat/cool cycle,t he scattering data still fitted best to the same model, but with ar adius of 34.0 AE 0.1 ,aKuhn length of 339.7 AE 3.4 and alength of 4976.4 AE 232.8 (Figure 2b;Supporting Information, Figure S5). In both cases,t he absolute length is beyond the resolution of the fit, but it is clear that the length increases after the heat/cool cycle,a sdoes the Kuhn length, showing that the flexibility decreases.T he radius decreases after the heat/cool cycle.S mall-angle neutron scattering (SANS) data was also collected (Supporting Information, Figure S6 and Table S2). In agreement with our previous work, [9] thedata before heating is best fit to ahollow cylinder with aradius of 36.9 AE 0.1 and acore radius of 19.1 AE 0.2 . After ah eat/cool cycle,t he radius decreases to 31.2 AE 0.1 (in line with the SAXS data) and the core radius also decreases to 9.2 AE 0.3 .These data lead us to suggest that the 2NapFF assembles into acoiled structure with acentral core (Figure 3a); the heat/cool cycle leads to partial dehydration of the core and an extension in the length. Thec ryotransmission electron microscopy (cryo-TEM) images of the solution before and after heating (Figure 2c and d, respectively;Supporting Information, Figures S7 and S8), show that both solutions contain long anisotropic structures with similar diameters (Supporting Information, Figure S9).
Further insight into the heat/cool cycle comes from fluorescence,i nfrared (IR), and nuclear magnetic resonance (NMR) spectroscopy.I mmediately after heating, fluorescence is significantly quenched as compared to the initial solution (see fluorescence spectra in the Supporting Information, Figure S10). There is no red-shift in the peak, which implies that there is no significant change in the local packing. On cooling, the intensity increases again, but does not reach the original value.Normalization of the data shows that there is aslight increase in the tail of the fluorescence,although the position of the main peak has not changed significantly.These data therefore correlate with as imilar molecular packing of the naphthalene rings.T he IR data (Supporting Information, Figure S11) shows relatively little change before and after heating, with the main difference being the shift of ap eak from 1647 cm À1 to 1654 cm À1 on heating;this peak is then lost on cooling. This data implies as ubtle change in the packing after aheat/cool cycle.Inthe NMR spectra, peaks are initially broad and integrate at al ower intensity than might be expected (Supporting Information, Figure S12);this is due to the presence of the worm-like micelles,a sd iscussed previously.Onheating, the peaks become better resolved and the integral of the aforementioned peaks increases with respect to the internal standard. This implies that the 2NapFF is more soluble at these higher temperatures.F or as ample that has been through ah eat/cool cycle,t he peak integral returns to that found before heating, implying av ery similar solubility and exchange rate between the worm-like micelles and free molecule.I nc ombination, these data correlate with our model above;t he heat/cool cycle leads to subtle changes in packing as opposed to the formation of anew structure.
Theh eat/cool cycle results in interesting effects.F or example,s olutions of 2NapFF were loaded into as yringe either without or with ah eat/cool cycle.A fter 18 h, the solutions behave very differently on being pushed out of the syringe (Figure 3b;see video in the Supporting Information). Thes olution without ah eat/cool cycle flows as expected for aslightly viscous solution. Thesolution that has been exposed to ah eat/cool cycle can be pushed out as as ingle strand essentially.S olutions of 2NapFF at high pH can be gelled by adding ad ivalent cation. [6a, 9] Ah omogeneous gel is formed after standing overnight (initially,local gelation occurs where the salt solution contacts the 2NapFF solution). As described previously, [2a] addition of asolution of CaCl 2 to an as-prepared solution of the gelator leads to formation of at urbid gel. A significantly more transparent gel is formed if the solution of CaCl 2 is added to as olution removed immediately after the heating step,ortoapre-heated solution that has been allowed to cool and rest at room temperature (Figure 4a).
Thec losest analogy we can find for the behavior that we observe here is reported by Stupp et al. with respect to peptide amphiphiles, [16] where heating and cooling led to ac hange in the properties of the solution and greater alignment of structures.T he heating and cooling was found to be necessary to allow "noodling", for example.T heir data suggested that the local packing was not changed by heating and cooling,b ut that the aggregates were dehydrated by heating, leading to filaments with significantly greater diam- right, after ca. 0.7 mL has been pushed out. In both cases, the solution on the left has been exposed to ah eat/cool cycle and the solution on the right has not. Scale bar = 2cm. . a) Gels formed by adding asolution of CaCl 2 to asolution of 2NapFF (10 mg mL À1 ;p H11). From left to right:g els formed from asolution as-prepared, from aheated solution with the CaCl 2 added just after heating, and aheated solution with the CaCl 2 added after cooling. b) Strain sweeps for gels formed by adding CaCl 2 to an asprepared solution of 2NapFF (10 mg mL À1 ;pH11, (**)), from ah eated solution with the CaCl 2 added just after heating (* ** *), and eters.T his increase in diameter was backed up by changes in the SAXS data. [16] Hence,o ur systems are behaving differently.
Gels formed from the pre-heated solutions behaved as described previously,w ith a G' of 18.9 AE 3.4 kPa, and at an d (G''/G')o f0 .17 AE 0.02 (Figure 4b). Interestingly,t he gels formed when the solution of the CaCl 2 was added, either immediately after heating or after ac ooling period, were significantly stiffer and very similar to one other. G' was 119.4 AE 1.4 kPa with atan d of 0.20 AE 0.01 for the gels formed by adding CaCl 2 immediately,and G' was 122.7 AE 4.1 kPa with at an d of 0.22 AE 0.01 for the gels formed by adding CaCl 2 after cooling (Figure 4b). These are some of the highest reported values for such ag el. Cryo-TEM of the gels shows that lateral association of the fibers occurs on addition of the calcium salt (Supporting Information, Figures S13-15).
This behavior is not restricted to 2NapFF.Anumber of other LMWGs show similar results (Supporting Information, Figures S16-S18). Ap rerequisite seems to be that the solutions form viscous solutions at high pH, implying that worm-like micelles formed from this class of molecule behave in this manner.Aswehave described previously, [6b] these are the more hydrophobic examples.L ess hydrophobic LMWGs form solutions that are not viscous and show no tendency to either become self-supporting or to become more viscous (Supporting Information, Figures S19-S21).
In summary,wehave shown that asimple heat/cool cycle results in as ignificant change in the physical properties of as olution of aL MWG.T he heat/cool cycle results in an increase in the length of the worm-like micelles and allows interesting new materials to be prepared. Ther heological properties of the gels formed from solutions before heating or after ah eat-cool cycle are very different. Thei ncrease in extensional viscosity,f or example,p otentially allows electrospinning to be carried out. [17] As well as the opportunities that this offers,w eh ighlight that this may have important implications for experiments carried out in labs with different operating temperatures.