Lars Jørgensena,
Klaus Qvortrupb and
Ioannis S. Chronakis*a
aTechnical University of Denmark, DTU-Food, Nano-BioScience Research Group, B227, DK-2800 Kgs, Lyngby, Denmark. E-mail: ioach@food.dtu.dk
bUniversity of Copenhagen, Faculty of Health and Medical Sciences, Department of Biomedical Sciences, Panum Institute, DK-2200, Copenhagen N, Denmark
First published on 12th June 2015
Asolectin phospholipid nano-microfibers were prepared using electrospinning processing. The asolectin fibers were studied by scanning electron microscopy, and the fiber morphology was found to be strongly dependent on the phospholipid concentration and the solvents used. The solvents studied were chloroform:
dimethylformamide (CHCl3
:
DMF, 3
:
2 v/v), isooctane, cyclohexane and limonene, producing phospholipid fibers with average diameters in the range of 2.57 ± 0.59 μm, ∼3–8 μm, ∼4–5 μm and 14.3 ± 2.7 μm, respectively. The diameter of asolectin electrospun fibers does not follow the theoretically predicted value of ∼0.35 μm because of the intermolecular aggregation between the reverse micelles formed in the highly concentrated asolectin solutions. However, when co-axial solvent electrospinning was applied, where the outer needle contains a pure solvent and the inner needle contains the asolectin solution in CHCl3:DMF, a substantial reduction in the average fiber diameter was observed. In particular, the average diameter of the fibers when DMF (a solvent with a high dielectric constant) was used as a sheath solvent was reduced by a factor of about 7 and was at the nano-size range, as theoretically predicted. The dielectric constant of the solvents had a strong influence on the jet split properties and affected the morphology of the electrospun asolectin fibers.
The electrospinning technique is a straightforward and versatile top–down manufacturing process for preparing functional nano-microfibers that uses an electrostatically driven jet of polymer solution.4,5 The fibers obtained have diameters usually ranging from several nanometers to a few micrometers. McKee et al.,6 showed that a phospholipid mixture of asolectin solution above 35% w/w inCHCl3
:
DMF [3
:
2 v/v] is capable of forming continuous fibers with an electrospinning processing in normal atmospheric conditions with an average phospholipid fiber diameter of ∼3.3 μm for 45% w/w. It was proposed that, with the increase of the phospholipid concentration, a transition from single molecules to micelles formation and then to the formation of rod-like structures and a creation of elongated cylindrical aggregates occurs, that can be electrospun. These authors also found that the diameter size (D) did not follow the empirical equation (D [μm] = 0.18(C/Ce)2.7) that applies for other polymers based on concentration entanglement (Ce) and solution concentration (C). This is because the micelles aggregated, making the solution highly dependent on viscosity.7 This was confirmed by studying another self-associating polymer modified with quadruple hydrogen bonding capabilities; a linear dependence between diameter and the viscosity parameter ɲ was verified in a log–log plot.
Moreover, Hunley et al.,8 showed the ability to electrospin pure 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine in the melted state at a temperature around 200 °C, due to a rod-forming structure at this temperature. The average fiber size was reported to be 6.5 ± 2.0 μm and fibers were straight without interconnected points between them.
Further, Yu and co-workers,9 electrospun a fibrous network using a mixture of polyvinylpyrrolidone (PVP) with soybean lecithin. Upon immersion into water, this fibrous network spontaneously created liposomes with a very narrow distribution and vesicles between 120 and 370 nm depending on the parameters applied for electrospinning.
Recently, mixtures of water with other organic solvents have also been reported for the electrospinning of synthesized phosphonium gemini surfactants by Hemp et al.10 In particular, the authors tested various spacer link distances and found that spacers longer than three methylenes did not form fibers, while spacer links of two to three methylenes did. Using CHCl3 as solvent, the average diameter of the fibers was about 1.1 μm whereas in H2O:
MeOH (1
:
1 v/v) the diameter was increased to 4–5 μm owing to the increased evaporation point of the solvents. Other approaches use modified lipids such as cholesterol derivatives and cross-linking of the head group to create highly stable lipid-mimetic nanostructures. Zhang et al.,11 demonstrated the ability to create a highly stable lipid-mimetic nanostructure by attaching cholesterol as the head group to a succinyl silane, following a cross-linking of the head group to the polymer structure.
However, no studies have been carried out to investigate how to modulate the morphology and the average diameter of phospholipid electrospun systems. This work reports the utilization of three organogel solvents namely, cyclohexane, isooctane and limonene (citrus flavor), capable of creating electrospun phospholipid fibers of asolectin, and compares their morphology with the electrospun asolectin fibers using CHCl3:
DMF [3
:
2 v/v]. Both cyclohexane and isooctane are well characterized solvents for lecithin solutions and are known promoters of elongated cylindrical reverse micelles.12 Furthermore, to tune the fiber diameter and the morphology of the electrospun phospholipid fibers, we also investigated a co-axial electrospinning processing, using a pure solvent as the outer phase and the phospholipid solution as the inner phase.
Solvent | Conductivity [S m−1] | Dielectric constant | Surface tension [mN m−1] | Evaporation point [°C] |
---|---|---|---|---|
DMF | 2.50 × 10−4 | 36.7 | 35 | 152 |
CHCl3 | 2.00 × 10−6 | 4.8 | 27.16 | 61 |
Isooctane | 1.00 × 10−12 | 1.92 | 18.33 | 99 |
Cyclohexane | 7.00 × 10−13 | 2.01 | 24.98 | 81 |
Limonene | 2.30 | 176 |
Sample name | Concentration [% w/w] | Inner needle solvent | Feed rate [ml min−1] | Distance [cm] | Voltage [kV] | Outer needle solvent | Average diameter [μm] | SD [μm] |
---|---|---|---|---|---|---|---|---|
C-A | 50 | Cyclohexane | 0.01 | 10 | 40 | 5.41 | 2.16 | |
C-B | 50 | Cyclohexane | 0.02 | 10 | 40 | 5.28 | 1.86 | |
C-C | 50 | Cyclohexane | 0.05 | 10 | 40 | 4.01 | 1.71 | |
C-D | 50 | Cyclohexane | 0.1 | 10 | 40 | 4.66 | 2.95 | |
I-A | 60 | Isooctane | 0.02 | 10 | 40 | 3.53 | 1.51 | |
I-B | 60 | Isooctane | 0.05 | 10 | 40 | 8.49 | 2.79 | |
I-C | 60 | Isooctane | 0.1 | 10 | 40 | 6.40 | 2.11 | |
I-D | 60 | Isooctane | 0.2 | 10 | 40 | 6.55 | 3.37 | |
L-A | 60 | Limonene | 0.01 | 5 | 15 | 14.27 | 2.25 | |
CD | 45 | CHCl3![]() ![]() ![]() ![]() |
0.01–0.1 | 10–20 | 20–40 | 2.57 | 0.59 | |
Co-A | 45 | CHCl3![]() ![]() ![]() ![]() |
0.01 | 8 | 15 | DMF | 0.38 | 0.14 |
Co-B | 45 | CHCl3![]() ![]() ![]() ![]() |
0.01 | 8 | 15 | CHCl3 | 0.66 | 0.45 |
Co-C | 45 | CHCl3![]() ![]() ![]() ![]() |
0.02 | 6 | 20 | Isooctane | 1.01 | 0.37 |
Co-D | 45 | CHCl3![]() ![]() ![]() ![]() |
0.01 | 12 | 20 | Cyclohexane | 1.54 | 0.48 |
Concentration [% w/w] | CHCl3![]() ![]() ![]() ![]() |
Cyclohexane | Limonene | Isooctane |
---|---|---|---|---|
35 | Beads | — | — | — |
40 | Beaded-short fibers | Beads | Beads | — |
45 | Fibers | Beaded-short fibers | Beads | Beads |
50 | Fibers | Fiber | Beads | Beads |
55 | — | Gelation | Beads | Beaded-short fibers |
60 | — | — | Fiber | Fiber |
65 | — | — | Gelation | Gelation |
Electrospun asolectin fibers were formed by CHCl3:
DMF [3
:
2 v/v], cyclohexane, isooctane and limonene at concentrations of 45, 50, 60 and 60% w/w, respectively (Fig. 1). All solvents yield uniform fibers without beads given that optimal phospholipid concentrations and conditions for electrospinning processing were used. The average diameters of the samples are shown in Table 2, and a histogram overview of the distribution of their diameter is shown in Fig. 1.
The morphology and the average diameter of the fibers produced using CHCl3:DMF are in accordance with what was previous reported,6 by McKee et al. The average fiber diameter was about 2.57 ± 0.59 μm when electrospinning a 45% w/w asolectin solution. No major change in the average fiber diameter was observed by changing the feed rate, the distance and the electrical field (voltage), although, the concentration appeared to have a major impact, in accordance with previous studies.6 As shown in Fig. 2A, a formation of beaded short-length fibers was observed by electrospinning 40% w/w asolectin dissolved in CHCl3:DMF. For the 45% w/w asolectin solution in cyclohexane, beaded, short-length and thick fibers were formed with an average diameter of about 16–20 μm (Fig. 2B). When the concentration was increased to 50% w/w, uniform electrospun phospholipid fibers without beads were produced that had an average diameter between 4 and 6 μm, depending on the processing settings. The use of isooctane as a solvent for the phospholipid solution promoted the formation of even and elongated asolectin fibers at a concentration of 60% w/w, with a very uniform Gaussian distribution and an average diameter between 3 and 9 μm depending on the electrospinning settings. At a lower concentration of 55% w/w, the asolectin solution in isooctane also formed uneven short-length fibers and beaded structures (Fig. 2C). Moreover, a limonene solution of asolectin of 60% w/w was successful in being electrospun and yielded uniform fibers with an average diameter of 14.27 ± 2.25 μm (Table 2). Recently, Greenfeld, and Zussman experimentally and theoretically studied the disentanglement mechanism that leads to such a jet fragmentation.14 They suggested that the jet fragmentation is caused by disentanglement of chains, to the extent that the network loses its continuity and breaks up at intervals. Short nanofibers formed when the jet breaks into fragments before complete solvent evaporation and fiber solidification.
Hence, it is reasonable to suggest that the phospholipid concentration and the solvent type dominate the formation of the self-assembled nanostructures and promote the electrospinning of phospholipid fibers with different morphologies and average diameters. This agrees with previous studies for purified lecithin supporting that various solvents promote different changes in the surface curvature of the phospholipids,15–17 and that they can thus alter the branching and the aggregation of nanostructures and hence the morphology of the electrospun fibers shown here.
Furthermore, it is well known that both the dielectric constant (εsolvent) and the evaporation point (Bpsolvent) of the solvent impact the final diameter of the electrospun polymers.18–22 The higher the solvent evaporation point, the slower the evaporation will occur, allowing more time for the electrostatic forces to thin the jet, while solvents with a high dielectric constant result in lower diameter fibers. We have observed that the average diameters of the phospholipid fibers are correlated linearly with the evaporation point of the solvent (Fig. 3A). However, the slope of this linearity is opposite to what has been found for various polymers such as PS, PCL, PMPS, PEO, etc.18–22 In addition, as shown at Fig. 3B, the average asolectin fiber diameter (D) can be correlated with both the boiling point and the dielectric constant of the solvent with an allometric function .
![]() | ||
Fig. 3 Average diameter of electrospun phospholipid fibers as a function of: (A) solvent evaporation rate and (B) boiling point and the dielectric constant of the solvent. |
The coefficient n is found to be 3.05 ± 0.24 (R2-value of 0.99). We have assumed that the solvent properties of the CHCl3:
DMF mixture are mainly due to the presence of DMF, as DMF will be retained longer during evaporation.23
Thus, the final diameters of the asolectin electrospun fibers are also closely related to the dielectric constant and the evaporation properties of the solvent utilised. Moreover, at a certain evaporation point, the viscoelastic properties of the solution will become stronger, so that the electrostatic forces can no longer thin the jet, and this will also affect the final diameter of the electrospun fibers. It should be noted, that the asolectin solutions in cyclohexane, isooctane and limonene all have an onset of gelation close to their concentration required for fiber production, with limonene being the closest with a gelation point at just 62% w/w (Table 3). On the other hand, the concentration of 45% v/v asolectin in CHCl3:
DMF that was used, was much lower than its gelation concentration (no gelation was found even up to 65% w/w).
It should be noted that the electrospun phospholipid fibers were stable for one year at −20 °C, without any changes in the average diameter or morphology. Moreover, Quanta 3D dual beam SEM was used to investigate the lower layers of 50–2000 μm thick webs that had been stored for one year at −20 °C. It is confirmed that the structures were intact, even at a lower fiber depth, without collapsing together, even after prolonged storage (Fig. 4).
![]() | ||
Fig. 4 FIB SEM of electrospun fibers from a solution of 45% w/w asolectin in CHCl3![]() ![]() ![]() ![]() |
All the solvents used with the co-axial set-up resulted in a clear reduction of the average fiber diameter compared to the single-needle diameter using the same solution. The lowest average fiber diameter of 0.38 ± 0.14 μm was achieved, using DMF as the outer solvent (Fig. 5). Thus, a reduction factor of ∼7 of the fiber diameter was observed (2.57 μm to 0.38 μm). To the best of our knowledge, no average diameter on the nanoscale for electrospun phospholipid fibers has been reported in the literature. This average diameter is close to the theoretical prediction of the empirical equation (D [μm] = 0.18(C/Ce)2.7), which gives a predicted value for the electrospun phospholipid fibers of ∼0.35 μm in the absence of an intermolecular association between the micelles.
When CHCl3 was used as a solvent in the outer needle (same parameter settings), an average diameter size of 0.66 ± 0.45 μm was observed (which is ∼4 times less than the single-needle setup). However, beads were also present and the fibers show a large SD of the average diameter as well as diameters above 2 μm (Fig. 5B). In the case when isooctane or cyclohexane was used as a solvent in the outer needle (with an inner solution of 45% asolectin in CHCl3:
DMF [3
:
2 v/v]), average diameters of 1.01 ± 0.37 μm and 1.54 ± 0.48 μm were observed, respectively (Fig. 5C and D). Thus, a reduction of the average fiber diameter was achieved by factors of ∼2.5 and ∼1.67, respectively. Consequently, only DMF showed the ability to process phospholipid fibers at the average fiber diameter predicted by the empirical equation.
As discussed above, McKee et al.,6 pointed out the intermolecular association between the micelles to be the main reason behind the increased diameter. Due to this extensive aggregation, it is most probable that an incomplete split of the jet flow occurs at the single needle set-up. Moreover, it is well known that the dielectric constant of the solvent has a significant impact on the bending instability and the structure of the resulting electrospun fibers.18–22 In particular, the accumulation of electric charges passes non-axial instability to the jet flow in the electric field that breaks the jet flow split into thinner ones. Thus, electrospinning of a solution with higher dielectric constant solvents, gives a stronger capability for a jet split in the high electric field. For our systems, DMF has a much higher dielectric constant than that of chloroform, isooctane, and cyclohexane (Table 1). This might play a pivotal rule in the reduction of the average fiber diameter with the use of the co-axial set-up, due to their significant influence on the jet breakdown properties. Therefore, DMF solvent at the out needle could induce extensive bending instability and enhance Coulombic stretching of the jet, which resulted in extremely finer nanostructures and improved uniformity of the electrospun phospholipids.
It is then reasonable to support that the other solvents (chloroform, isooctane, cyclohexane), although they slightly reduce the average diameter of the fibers using the co-axial process, have dielectric constants that are too low to hold the electric charges and to assist asolectin's jet in splitting into nanofibers. This is in agreement with the relatively high average fiber diameters, and the wider fiber diameter distribution originating from an incomplete split of the jet flow produced using the above solvents applying the single needle set-up.
In addition to the dielectric constant, other parameters such as the current intensity, conductivity and the surface tension of the solutions are also crucial for the final diameter of the electrospun fibers.24–26 The following approximate scaling law, described with eqn (1), has been suggested for the cone-jet mode:
![]() | (1) |
![]() | ||
Fig. 6 Average diameter of electrospun phospholipid fibers as a function of current intensity (I) using the co-axial set-up (dotted lines are to guide the eye). |
Yu and co-workers have also used a co-axial solvent set-up for the electrospinning of various concentrated polymer solutions.27–31 For example, they have studied a high concentrated PVP solution (35% w/v) using ethanol as inner solvent and water, chloroform, methanol, formic acid and DMAc as outer needle solvents. By using a single needle approach of the same PVP solution, it will typically result in needle clogging or development of fibers with large diameters. They found that the evaporation point of the sheath (outer) solvent controlled the average diameter of the electrospun PVP fibers. In particularly, an inverse relationship between evaporation point of the solvent and average fiber diameter was observed, except in the case of formic acid; (this was explained due to its poor evaporation and subsequently beads on string morphology formation by solvent–solvent interactions at the stage of the bending instability). Furthermore, it was suggested that the conductivity and the surface tension of the solvents could also play a role, as in the case of methanol and chloroform as the outer solvents for PVP electrospinning, their fiber diameter differences could not be purely correlated with their evaporation points. Moreover, for the electrospinning of asolectin solutions, we have observed that the evaporation point of the sheath solvent is a significant factor and correlated with an inverse relationship of the fibers diameter, (predominantly for the low conductive and dielectric constants solvents such as cyclohexane and isooctane). However, we found that the fiber diameter could be closely correlated using the parameters of eqn (1), rather than the evaporation point of the sheath solvents.
Overall, the sheath solvent approach appears to be capable of reducing the diameter and controlling the morphology of electrospun fibers. It could be an alternative way to produce electrospun fibers from biomaterials which usually are not spinnable without the utilization of additives. For instance, the electrospinnability of biomaterials such as cyclodextrins,32–34 proteins,35–38 polypeptides,39 supramolecular molecules,10,40,41 among other, could be improved.
This journal is © The Royal Society of Chemistry 2015 |