Controlling the formation and alignment of low molecular weight gel ‘noodles’

We show how to control the formation and alignment of gel ‘noodles’. Nanostructure alignment can be achieved reproducibly by extensional deformation as the filaments form. Using a spinning technique, very long and highly aligned filaments can be made. The Young's moduli of the gel noodles are similar to that of a bulk gel. By using two syringe pumps in a concentric flow setup, we show that a filament-in-filament morphology can be created.


Materials
The synthesis and characterisation of 1ThNapFF has been reported previously. 1 CaCl 2 was obtained from Alfa Aesar and used with no further purification. Nile blue A was obtained from Sigma Aldrich and a solution in deionised water at a concentration of 8000 ppm was prepared from this.

Solution preparation
Aqueous solutions of 1ThNapFF were prepared as follows. 1ThNapFF was weighed into a 40 mL plastic Falcon tube followed by the addition of deionised water and then 2 equivalents of 2M NaOH in deionised water. The amounts used varied depending on the volume and concentration required.
Solutions were typically prepared on a 10 mL scale. The solutions were stirred overnight and then checked for undissolved solids the following morning and stirred further if required. Once dissolved, the solution pH was checked. The solution pH was adjusted to pH 11.3 ± 0.1 as required with 2M HCl and/or 2M NaOH. Generally, the initial pH >12 and was then lowered to pH 11.3. Care was taken to ensure homogeneity of the sample when pH adjusting. The viscous nature of 1ThNapFF solutions meant that insufficient mixing could result in localised pH differences. Once pH adjusted the solution was ready to use.
To stain the 1ThNapFF solutions blue for enhanced filament visibility when required, a calculated volume of 8000 ppm Nile blue A stock solution was added to achieve a final concentration of 80 ppm Nile blue A in the gelator solution.
1.2.1 Heat/cool cycle Some 1ThNapFF solutions were subjected to a heat/cool cycle. 10 mL 1ThNapFF solutions were pH adjusted to the required value, then transferred to 14 mL glass vials. The glass vial was placed in a preheated and temperature equilibrated oven at the desired temperature (either 60 or 80±2°C). The solutions were kept in the oven for 1 hour, removed and allowed to cool at room temperature for 3 hours.
For forming gel filaments with heat/cooled solutions, the hot gelator solutions were removed from the oven and immediately loaded into a plastic syringe to be used on the syringe pump. The syringe tip was sealed with Parafilm and the solution filled syringe allowed to cool for 3 hours.

Bulk gel preparation
Bulk gels were prepared for nanoindentation from 10 mg/mL 1ThNapFF solutions at pH 11.3. 2 mL of pre-gel solution was pipetted into an upside down plastic syringe with the top cut off, in a fashion as S3 had previously reported. 2 This method is favourable because it means the gels can be easily removed once formed and taken for study using nanoindentation. Two equivalents of Ca 2+ ions (in the form of an 8.9 μL drop of 200 mg/mL CaCl 2 in deionised water) was added near the centre of the pre-gel solution. The open top of the syringe was covered with Parafilm. Localised gelation where the CaCl 2 droplet was placed occurred instantly. For gelation to occur throughout the sample, the Ca 2+ diffuses through the solution. For this particular system the diffusion was an extremely slow process and even when left for many months a completely homogeneous gel is not formed. For the nanoindentation experiments two bulk gels were studied. An inhomogeneous bulk gel, for which diffusion had been left for 7 days, and a homogeneous gel that was left for 6 months. Although it is noted that the homogeneous gel was still not entirely homogeneous by eye. For the 6-month diffusing gel, this was instead prepared on a 2 mL scale but in a 7 mL Sterilin vial to allow for an effective seal and prevent drying of the gel.
The inverted syringe technique cannot be sealed effectively, especially for many months. Instead, the gels dry out.

Viscosity measurements
Viscosity measurements were made using an Anton Paar 301 rotational rheometer with a 50 mm cone and plate geometry. All measurements were performed at 25°C and a shear rate range of 1-1000 s -1 . To minimise the shear that the 1ThNapFF solutions are subjected to prior to the experiment, they were loaded onto the rheometer by pouring instead of using a pipette. Using a pipette can shear-align the sample and change the measured viscosity.

Pipette filament formation
All gel filaments were analysed on the day they were made unless stated otherwise.   3 but it was seemingly not performed with the intention of inducing alignment in that work. Here, a faster dragging than shown in that video was performed.

Syringe pump filament formation
All gel filaments were analysed on the day that they were made unless stated otherwise.

Syringe pump
An Alaris Carefusion syringe pump was used to control the flow rate of the 1ThNapFF pre-gel solutions.
A 10 mL syringe was attached via a Luer lock fittings to 20 cm rubber tubing with a flat-headed needle (413 µm inner diameter) at the end. To load the syringe pump, the syringe was filled directly with gelator solution. The tubing was attached to the syringe and the pre-gel solution pumped manually until the tubing was full with pre-gel solution. The needle was then attached to the end of the tubing and the syringe loaded into the syringe pump. The rubber tubing and needle enabled movement of the needle to position it into the trigger medium. Next, the syringe pump was started until the pre-gel was coming out. During this process, care was taken to avoid the formation of bubbles within the tubing.

Static filament formation
To make gel filaments, a cylindrical Pyrex dish with a microscope slide at the bottom was filled with 100 mL 50 mM CaCl 2 trigger solution. A flow rate was selected and the infusion started. The first drops of pre-gel solution was dabbed onto tissue paper and then the needle moved and positioned vertically downwards into the trigger medium ~5 mm below the surface. The infusion was carried out for 15 s.
To finish the needle was slowly moved to the edge of the dish and then dragged up the side to break the connection to the rest of the structure. As the filaments were generally one continuous structure, simply pulling the needle out may pull the whole structure and alter/damage it. After removing the needle, the syringe pump was stopped. Once formed the filament gel structures became more opaque over the course of a minute. The excess trigger solution was carefully removed manually from the dish with a 50 mL syringe. When the level of trigger solution was reduced to around 20 mL the floating gel structures were carefully positioned above the microscope slide. The remaining trigger solution was then slowly removed so that the gel structure "landed" on the microscope slide. The microscope slide was then taken for microscope imaging. The filaments were never left to dry out and a few drops of trigger medium added atop the microscope slide to keep them hydrated.
The 1ThNapFF solutions, in ambient room lighting, were colourless and the filaments were difficult to see for the first few seconds but became slightly opaque as gelation occurred. A flashlight positioned at S6 the side shining into the dish allowed for viewing of the filaments as they form. The syringe pump was always started before placing the needle into trigger solution. If the infusion was not started when the needle was placed in trigger solution the trigger could travel up the needle, forming a gel and blockage inside the needle. As the infusion was started before immersion in the trigger, there was often a droplet of gelator solution on the needle when it was placed into the trigger. When immersed into the trigger this droplet gelled resulting in a ball shaped gel structure on the tip. A small and fast shake of the needle tip in solution dislodged this allowing for the fluid to elute into solution and form filament structures.
This shake did not affect the formation of the subsequent filament structures.  Figure S3). Care was taken to assure the two flows were concentrically aligned. This was done S7 with the front camera of a mobile phone and a clip-on mcarolense to help focus on the small object.
Pictures were taken to show the good alignment of the two nozzles ( Figure S4c and S5c). The exit nozzles of the concentric flows were submerged in a bath of trigger medium by raising the 50 mM CaCl 2 bath up to the static nozzles using a mechanical jack ( Figure S4a and S5a). As before, the syringe pump flow was started and run until drops were coming out prior to submersion in the trigger medium to prevent blockages. Again, droplet of solution that may be present upon submersion and consequently gelled was removed by a spatula. After this the filament formation would start given sufficient flow rates. 1.6.6 Concentric flow setup: Filament within a filament Using a similar setup but having both the sheath and inner flows as 1ThNapFF solutions, a "filament-infilament" could be formed. Here, a 1 mL pipette tip was used for the sheath flow ( Figure S5c) as opposed to a Luer-lock connector as above. The pipette tip was cut approximately 1 cm from the exit of the tip. This shortened tip was fitted into the rubber tubing so the exit nozzle from the tubing was the plastic pipette tip. Then the inner flow needle was pierced through the tubing and aligned within the pipette tip from which the sheath flow was dispensed. For this setup an inner flow of 5 mg/mL 1ThNapFF solution and a sheath flow of 10 mg/mL 1ThNapFF solution were selected. The 5 mg/mL 1ThNapFF solution was stained blue with 80 ppm of Nile blue A to distinguish the two layers. Once formation of the filament occurred, the inner flow could be adjusted to vary the thickness of the blue, 5 mg/mL, inner filament within the 10 mg/mL sheath filament.

Optical microscopy
A Nikon Eclipse LV100 optical microscope with an Infinity 2 camera connected to a computer was

S11
Once the gel structures had been formed and collected on a microscope slide they were imaged. The underside of the slide was dried thoroughly with paper towel and placed on the microscope stage.
Droplets of trigger medium was placed on the filament structures to form a pool to keep them hydrated.
The gels shrink and lose alignment if not re-hydrated over a 40 minute period ( Figure S6). Additionally, this removed any curved water/air boundaries from near the filament structures which resulted in poor images due to the refraction of light (see Figure S6c for this effect visualised).

ImageJ analysis
ImageJ (version 1.50e) 4 analysis was used to measure the structure diameters in all of the microscope images. This was done using a 1 mm scale bar supplied with the microscope which is imaged at the same magnification, with which the global scale (pixels/mm) can be set in ImageJ. Out of focus structures were not measured. In each microscope image the diameter of every in focus filament structure was measured at a point approximately in the middle of the structure in view. This allowed for a histogram of filament diameter to be plotted for each sample.

Nanoindentation
Nanoindentation was used to measure the Young's modulus of the low molecular weight gel filaments and bulk gels. The Young's modulus can vary depending on the experimental setup used. 5 During a nanoindentation experiment, the tip performs a set distance indentation into the sample and the resistance to the deformation is measured. This allows for a force-indentation curve to be created, which is then fitted using a Hertzian model to obtain a Young's modulus. The technique is localised, therefore S12 the tip is scanned across the surface of the sample and many individual measurements collected. These are then processed and presented all together for each sample. A Chiaro nanoindenter (from Optics11Life) was used for nanoindentation under the indentation control operation mode. Tip radii of 8.5-10 μm, probe stiffness of 0.46-0.51 N/m and a Geo factor range of 1.21-1.25 were used. A maximum displacement of 10,000 nm was used, which includes the whole range including the out of contact region (flat baseline prior to reaching the surface). The indentation was performed at a constant speed of 5 μm/s. The forward segment of the force-distance curves were analysed and filtered for high frequency noise and evaluated with a custom-build software 6 developed using python and available online under open source license. 7 The fitting procedure was applied to the first 2000 nm after the contact point, keeping within 10% of the radius of the spherical indenter tip and 10% the thickness of the sample. A contact point method in the analysis software called 'Goodness of fit' was used. 8 The measurements were performed in 50 mM CaCl 2 solutions.
The thin (<0.2 mm), spun gel filaments could not be measured using the specific setup available. were subtracted. The 20 frames were averaged and, for samples that scattered isotropically, a full azimuthal integration performed. Where necessary, flare from capillary mis-alignment was masked out.

Small angle X-ray scattering
The azimuthal integration yielded the data as I vs q plots. These data were then fitted to structural models in the SasView software (version 5.0.2). 11 Many of the samples exhibited anisotropic scattering, which is consistent with aligned structures within the capillaries. 12 This is attributed to shear-aligning when loading the sample into the capillaries. The anisotropic data cannot be azimuthally integrated as is done for a normal isotropic data set. Instead, bow-tie integrations were performed over the regions of high scattering intensity and low scattering intensity. This results in two separate I vs q plots. The I vs q plots for the high scattering intensity regions were fitted to structural models in SasView. When performing a bow-tie integration over the high intensity (vertical) regions, there are kinks in the data S13 due to boundaries between the detector modules. This results in anomalous data points which have been manually removed and gaps in the curve. There are some limitations in this approach because the form factor fitting assumes that the self-assembled structures are non-interacting and randomly oriented. For aligned samples the structures are not randomly oriented. While this approach is not ideal and may lead to some small errors in the fitting, it is not significant enough to not fit the data. Indeed, by virtue of the fact that the particles have aligned, this reinforces the application of a structural model consistent with 1D structures. For the fitting X-ray scattering length densities (SLDs) were calculated using the NIST neutron activation and scattering calculator. 13 An SLD of 14.025x10 -6 Å 2 was used for 1ThNapFF and a solvent SLD of 9.469x10 -6 Å 2 was used. Figure S7. Viscosity of 1ThNapFF solutions at different concentrations at pH 11.3 across a shear rate of 1 to 1000 s -1 . Error bars are the standard deviation of triplicate measurements.

Rheology
The influence of solution pH on the shear viscosity was investigated and showed that across a relatively broad pH range (9.2 -12.0) the viscosity is broadly unchanged (Figure S8). At pH 8.6 there is a S14 significant drop in viscosity which is likely attributed to a change in the self-assembled structures present, or a change in the interactions between them. Figure S8. Influence of solution pH of 10 mg/mL 1ThNapFF on viscosity.
The shear viscosity of 10 mg/mL 1ThNapFF solutions at pH 11.3 after heat/cool cycles was investigated and showed that after 60°C the shear viscosity was slightly reduced but after 80°C the viscosity was increased. S15 Figure S9. Influence of solution a heat/cool cycle at different temperatures on 1ThNapFF on viscosity.

Small angle X-ray scattering
All samples at a concentration of >1.5 mg/mL scattered strongly but many show strong anisotropic scattering. The 2D scattering patterns of all data collected are shown in Figure S10 and S11. 10 mg/mL at pH 9.2 and (c) 10 mg/mL at pH 12.0.

Fitting data
Data that exhibited isotropic scattering were azimuthally integrated and the resultant plots fitted structural models in SasView as has commonly done before for these type of systems. 14,15 The data for the 1.0 mg/mL and 1.5 mg/mL solutions exhibited low scattering intensity and were not fitted to any structural models. Many of the data show anisotropic scattering patterns in the 2D data. This is attributed to shear-alignment of 1D structures in the 1ThNapFF solution as it is injected into the capillary. As discussed in the experimental, bow-tie integrations were performed. An example of the two I vs q plots obtained from this process is shown in Figure S12. For many of the data a large upturn in intensity at the lowest Q was seen. This has previously been attributed to scattering from the network for selfassembled functionalised dipeptides. 16 For each fit, this low Q region was initially excluded from the fitting and the rest of the data fitted to different cylinder models until a good fit was achieved. For these data typically a cylinder or flexible cylinder model was required to achieve a good fit. This is consistent with the presence of self-assembled worm-like micelles. After this, the values obtained with this flexible cylinder fit were input into a combined flexible cylinder + power law model and the whole Q range fitted. The power law model generally captured this upturn in intensity at low Q. S17 Figure S12. 10 mg/mL pH 11.3 solution bow-tie integration data. (i) the low intensity scattering regions obtained from a horizontal bow-tie integration and (ii) the high intensity scattering regions obtain from a vertical bow-tie integration.
The fitting data for a range of samples in the concentration range of 1 -15 mg/mL are shown in Figure   S13. Importantly, some of these data sets were from isotropic data and others from anisotropic data.
Details are given in Table S1. S18 Figure S13. SAXS data (black hollow triangles) and fits (red lines) from the vertical bow-tie integrations of the 1ThNapFF concentration series at pH 11.3. Concentrations of (a) 1.0 mg/mL, (b) 1.5 mg/mL, (c) 2.0 mg/mL, (d) 2.5 mg/mL, (e) 3 mg/mL, (f) 4 mg/mL, (g) 5 mg/mL, (h) 7.5 mg/mL, (i) 10 mg/mL, (j) 15 mg/mL. S19 Table S1. Processing details and fitting parameters for the concentration series of 1ThNapFF.     A with the fit shown in Figure S16 a. Table S4. Flexible cylinder combined with a power law Fit parameters obtained for 2.0 mg/mL 1ThNapFF with 80 ppm Nile blue A with the fit shown in Figure S17 b. Scale bars represent 0.5 mm.

S27
Filaments were formed whilst dragging the pipette through trigger medium. This technique was challenging to control and resulted in a large variation in filament diameters however the thin, stretched regions typically showed birefringence. Due to the irreproducibility of these filaments, a mean filament diameter was not measured.

Syringe pump filament formation
A photograph of some gel filaments formed using a syringe pump and needle are shown below ( Figure   S23). S28 Figure S23. Photograph of gel filaments formed in bulk trigger medium using different flow rates. The sample labelled "300 mL/hr" can be seen to have formed very thick and inhomogeneous gel filaments, highlighting the need for an optimised flow rate. Figure S24. Photograph of (a) an approximately 5 cm 10 mg/mL 1ThNapFF gel filament and (b) a 5 mg/mL 1ThNapFF (Nile blue A stained) gel filament. Both have been lifted out of the trigger medium S29 on a spatula. They are the result of lifting a filament at the middle to balance around the spatula, such the images show two filaments hanging next to each other.

Effect of flow rate under static formation
With the syringe pump, the filament diameter could be controlled by changing the flow rate. This allows for fine tuning of the structures formed.  Figure S26. A selection of representative optical microscope images (both normal and cross polarised light) to compliment the data in Figure S25. The gel structures are formed from 10 mg/mL 1ThNapFF at pH 11.3 using the static procedure investigating the effect of flow rate. (a) 10 mL/hr; (b) 25 mL/hr; (c) 50 mL/hr and (d) 100 mL/hr. Scale bars represent 0.6 mm.

Effect of 1ThNapFF concentration under static formation
The effect of 1ThNapFF concentration was investigated. The solution concentration influenced the filament diameter at 10 mL/hr, with the lower concentrations forming thinner filaments. These differences are attributed to the difference shear viscosities of the solutions at difference concentrations.
The SAXS data shows similar structures are present, but it is hypothesised that these differences are due to other factors (such as the concentration of self-assembled structures present) not determined from the SAXS analysis. At 100 mL/hr the noodles formed are the same diameter.

Effect of solution pH
The effect of 1ThNapFF solution pH was investigated. The pH will influence the surface charge on the worm-like micelles and consequently could affect the interactions between them.   Figure S31). A 10 mg/mL 1ThNapFF pre-gel solution at pH 11.3 was used. (a -c) three repeats at 10 mL/hr; (d -f) three repeats at 25 mL/hr; (g -i) three repeats at 50 mL/hr and (j -l) three repeats at 100 mL/hr. Scale bars represent 0.6 mm.

Effect of a heat/cool cycle
For 1ThNapFF, a heat/cool was not required to form aligned gel filaments. The influence of a heat/cool cycle was investigated with temperatures of both 60°C and 80°C. A 60°C heat/cool had little impact on the shear rheology of the 1ThNapFF solution ( Figure S9) as well as the gel filaments that are formed ( Figure S33b, S34b, S35b). The 80°C heat/cool resulted in solutions with an increased viscosity ( Figure   S9) but resulted in gel filaments with no alignment seen using CPOM ( Figure S33c, S34c, S35c).

Nanoindentation
Nanoindentation was used to measure the mechanical properties of the gel filaments formed and bulk 1ThNapFF gels formed with CaCl 2 . As a highly localised technique, a wide range of different values are obtained for each sample due to local differences in the mechanical strength. To best represent these data, violin plots were created. In the violin plots the white dots represent the median Young's modulus, the thick dark blue bars represent the 1st and 3rd quartiles, the thin dark blue lines represent all the data excluding outliers and the light blue coloured violin-shaped regions represent the whole distribution of the Young's moduli values for each data set. The thicker the light blue region, the more data is present around that Young's modulus value. The effect of aging on the gel filaments was investigated ( Figure S38). The aging was only studied over a 24 hours period but showed that the Young's modulus fell over 24 hours from around 5 kPa to 2 kPa. S39 Figure S38. Violin plots for a gel filament formed with a syringe pump at a flow rate of 10 mL/hr on the day it was made (day 0) and 24 hours later (day 1).
A cross-section of a filament-in-filament was cut using a scalpel blade and the mechanical properties of the two regions probed using nanoindentation ( Figure S39). For this sample, the Young's moduli of the two regions are similar and most of the data overlap. Figure S39. Violin plots of data collected from the cross-section of a filament-in-filament, allowing both the 10 mg/mL 1ThNapFF and 5 mg/mL 1ThNapFF (Nile blue A stained) regions to be measured S40

Aligned filament formation
It was hypothesised that a sufficiently fast sheath flow of 50 mM CaCl 2 would stretch and align the gel filaments formed from a 10 mg/mL 1ThNapFF solution inner flow. With the flow rates accessible with the syringe pumps, a fast enough 50 mM CaCl 2 sheath flow could not be achieved to significantly stretch the gel filaments as they formed from the inner flow. Instead, the inner flow was started with no sheath flow and a length of filament formed ( Figure S40a). Then the sheath flow syringe was rapidly driven by hand, resulting in a significant thinning and enhanced birefringence of the filaments being formed ( Figure S40b). All of the filament regions formed were collected in the glass dish containing 50 mM CaCl 2 and the filaments taken for imaging. These data reinforce the conclusion that a stretching process results in enhanced alignment within the gel structures. This could, in theory, be used to form very long aligned filaments, with the limiting factors being the volume of the syringes used.

Filament-in-filament morphology
Using the concentric flow setup a filament-in-filament could be formed. The below flow rates were used to create two filament-in-filaments with differing widths of the inner Nile blue A stained gels.
Corresponding photographs of these are shown below.