- Email: [email protected]
gelatine nanofibres
European Polymer Journal 49 (2013) 2052–2061
Contents lists available at SciVerse ScienceDirect
European Polymer Journal journal homepage: www.elsevier.com/locate/europolj
Macromolecular Nanotechnology
Structure and morphology of electrospun polycaprolactone/ gelatine nanofibres D. Kołbuk a,b, P. Sajkiewicz a, K. Maniura-Weber c, G. Fortunato d,⇑ a
Institute of Fundamental Technological Research, Polish Academy of Sciences, Pawinskiego 5B, 02-106 Warsaw, Poland Advanced Fibers, Empa, Swiss Federal Laboratories for Materials Testing and Research, Lerchenfeldstrasse 5, CH-9014 St. Gallen, Switzerland c Materials-Biology Interactions, Empa, Swiss Federal Laboratories for Materials Testing and Research, Lerchenfeldstrasse 5, CH-9014 St. Gallen, Switzerland d Protection and Physiology, Empa, Swiss Federal Laboratories for Materials Testing and Research, Lerchenfeldstrasse 5, CH-9014 St. Gallen, Switzerland
MACROMOLECULAR NANOTECHNOLOGY
b
a r t i c l e
i n f o
Article history: Received 9 January 2013 Received in revised form 11 April 2013 Accepted 29 April 2013 Available online 22 May 2013 Keywords: Electrospinning Nanofibres Blend Gelatine polycaprolactone Molecular structure
a b s t r a c t Blends of polycaprolactone (PCL) and gelatine (Ge), being effective materials for tissue engineering strategies, were electrospun at various conditions and polymer weight ratios. The morphology, the supermolecular structure as well as the mechanical properties of resulting submicron sized fibres have been analyzed in relation to electrospinning conditions and PCL/Ge weight ratio. Compared to pure PCL, Ge addition leads to large reduction of fibre diameter and finally to changes of fibre morphology. For parallelised fibres collected on a rotating drum, preferred molecular orientation of PCL crystals is found. With increasing Ge content a general reduction of molecular orientation is observed. In addition, there is peculiar dependence of polycaprolactone crystallinity on the content of Ge, showing maximum at low Ge concentration (20%) as determined by differential scanning calorimetry (DSC) and wide angle X-ray scattering (WAXS). Such a trend can be explained by hydrophobic interactions in the system containing PCL, gelatine and water, being additional driving forces for crystallization of nonpolar PCL molecules. The presence of water within investigated blend systems has been evidenced experimentally using thermal gravimetric analysis (TGA). Young’s modulus of nonwovens, as determined by uniaxial tensile testing, indicates the effect of additivity of the stiffness of both polymers as well as the influence of preferred molecular orientation. Additional experiments were performed using collagen (Col) as a biopolymeric alternative to Ge. WAXS results show evidently amorphous structure of Col within the blended fibres, indicating strong tendency for denaturation of collagen into gelatine under the influence of hexafluoroisopropanol as a solvent. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Electrospinning is one of the most common methods to produce fibrous substrates for tissue engineering applications by use of natural and synthetic polymers. Ideally, a substrate should mimic both the form and functionality of the native extracellular matrix (ECM) [1]. It is known ⇑ Corresponding author. Address: Empa, St. Gallen, Abteilung 271, F1.43, Lerchenfeldstrasse 5, 9014 St. Gallen, Switzerland. Tel.: +41 (0)58 765 76 77; fax: +41 (0)58 765 78 62. E-mail address: [email protected] (G. Fortunato). 0014-3057/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.eurpolymj.2013.04.036
that morphology, architecture and surface properties have strong influence on cells growth, spreading, activity and functions [2]. Thereby, scaffolds formed by electrospinning are gaining increasing interest due to their tunable properties to provide ideal microenvironments for various kinds of cell types. Naturally occurring polymers such as collagen (Col) and gelatine (Ge) do not cause foreign body response [3,4]. PCL does elicit such a response, though minimal. As major drawbacks the poor mechanical properties and their variable physical properties have been identified [5]. In contrast, synthetic biopolymers have predictable and
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
2. Materials and methodology 2.1. Materials PCL having a molecular weight of Mw = 80,000 g/mol (Sigma–Aldrich, Switzerland) and gelatine (Ge) from por-
cine skin Type A, having average molecular weight of 50,000–100,000 g/mol (Sigma–Aldrich, Switzerland), were used. Collagen (Col) from tail tendons of young rats, characterised in [22–25] was used. The polymers were dissolved in 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) (Fluka, Switzerland). 2.2. Electrospinning process The electrospinning equipment was operated in horizontal mode and consisted of a medical infusion pump, a static plate or a rotating drum collector and two high voltage generators. The positive voltage was connected to stainless steel needle (0.6 mm inner diameter 26 mm length) and the negative one to the collector. A voltage of ±10 kV and flow rate of 7 ll/min was applied. Solutions were prepared at a total polymer concentration of 5 wt.% with following weight ratios for PCL/Ge: 80:20 (PG4:1), 50:50 (PG1:1), 20:80 (PG1:4) and PCL/Col 80:20 (PC4:1). Fibres were collected on static plate and on a rotating drum (approximately 1000 min1). The temperature and relative humidity were recorded during electrospinning process, being in the range of 22– 25 °C, and 35–45%, respectively. 2.3. Characterisation A rheometer (Anton Paar Physica MCR 300, Austria) equipped with a plate – cone system was applied in controlled shear rate mode to assess the shear viscosities in function of the shear rate. Flow curves with shear rates varying from 0.01–500 1/s were recorded at 20 °C. Electrical conductivity of solutions was measured using a Metrohm 660, (Switzerland), conductometer. The calibration of conductivity measurement was performed using 0.001, 0.01 and 0.1 mol/l potassium chloride solutions (KCl, Sigma–Aldrich, puriss). Before use KCL was dried for 24 h at 100 °C. Morphology of fibres was determined by scanning electron microscopy (SEM) (Hitachi S-4800, Japan) using 2 kV accelerating voltage at 10 mA flow current. Small pieces of fibre patches were coated with gold for 2 min before SEM imaging. Mean fibre diameter was calculated from SEM micrographs by averaging 30 measurements. Supermolecular structure in terms of crystallinity and molecular orientation was analysed by wide angle X-ray scattering (WAXS) and differential scanning calorimetry (DSC) supported by thermal gravimetric analysis (TGA). In the case of DSC, a Mettler T 28E (Switzerland) apparatus was used. Samples (ca. 5 mg) were heated from 10 °C to 190 °C at a rate of 10 K/min. Crystallinity of PCL was evaluated while specific heat of fusion of 100% crystalline PCL was taken as 142,9 J/g [26]. Temperature and heat calibration was performed using indium as reference material. Melting temperature and crystallinity were taken usually as an average value from three measurements. Results are presented as mean ± standard deviation. The thermograms were fitted by use of Pearson functions to separate the signals for PCL and Ge. WAXS (Oxford Diffraction, United Kingdom) was performed using X-ray diffractometry using Mo Ka1 radiation
MACROMOLECULAR NANOTECHNOLOGY
reproducible mechanical and physical properties (e.g. tensile strength, elastic moduli and degradation rate) and they can be synthesised with high reproducibility. Most synthetic polymers incorporate strong hydrophobic surface properties which present poor cell adhesion sites and thus promote a low level of cell attachment. Thereby, different approaches have been realised to adjust surface wettability properties e.g. by chemical and coating routes. A timely fashioned way is the use of hydrophilic and water soluble native extracellular matrix materials such as Col or Ge to form a hybrid of natural and synthetic polymers. In that way, scaffolds are generated which combine the advantages of both types of material and mainly alter materials surface properties. Up to now, various research works describe influencing parameters such as morphology and architecture of electrospun patches with respect to cell growth and proliferation (e.g. [5–13]). Focusing on blend nanofibers, Col/elastin [14,15], polyurethane/polycaprolactone (PCL) [16], and PCL/Ge [17], PCL/Col [18,19] have been described. However most of them concentrate on morphological topics, chemical structure and biological investigations (cell morphology, cytotoxicity and proliferation). For instance, Chong et al. showed reasonable growth of human dermal fibroblast on polyurethane – PCL/Ge sandwiched structures being promising candidates for wound healing applications [19]. Zhang et al. prepared PCL/Ge 1:1 electrospun nanofibers and found improved bone marrow stromal cell attachment compared to pure synthetic PCL nanofiber scaffolds. Cellular infiltration was demonstrated on the blend fibres of up to 114 lm in-depth [20]. PCL/Col electrospun fibres were investigated by Lee et al. [18]. Compared to pure PCL, the Col-added nanofibers showed better cell responses (adhesion and growth) in vitro and penetration in vivo. However, there are only few papers describing molecular structure of the two polymers for electrospun blend fibres, which influences surface properties and mechanical properties of the materials. It is known that crystallinity and molecular orientation are important parameters on physical properties and cell-scaffold adhesion and interaction [21]. The aim of this work is to investigate mutual effect of gelatine introduction into PCL electrospun fibres on nonwoven architecture, fibre molecular structure and mechanical properties. Analytical tools such as scanning electron microscopy (SEM), wide angle X-ray scattering (WAXS) and differential scanning calorimetry (DSC) were used to investigate fibre morphology, molecular structure and surface properties. Young’s modulus was evaluated from stress–strain measurements. In comparison, results from Col addition on molecular structure of the blend fibres are presented.
2053
2054
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
gle using Polanyi equation. We used the strongest (1 1 0) reflection, being in the radial angular range of amorphous halo. Assuming that crystal phase is usually oriented much higher than amorphous part, our procedure of orientation calculation without subtraction of amorphous part, can be treated as an estimation of crystal orientation. Thermogravimetric analysis was performed using Mettler Toledo TC50, in the range of 40–600 °C at a heating rate of 10 °C/min, under 30 ml/min nitrogen atmosphere. Mechanical properties have been measured by uniaxial extension (Instron 4205, United States) with a 100 N load cell under a cross-head speed of 10 mm/min. The samples were first cut into 10 40 mm samples. At least three samples were tested for each type of electrospun patch. The thickness was averaged from three measurements performed by cross-sectional SEM method. Porosity was not taken into consideration.
(k = 0.70926 Å). Analysis in transmission geometry was performed for the determination of crystallinity and molecular orientation. WAXS radial profiles registered up to 2h = 33°, were fitted using Pearson VII function both for crystal peaks and amorphous halo, after previous subtraction of the background. Crystallinity of particular component was determined by division of the area of crystal peaks by the total area (crystal and amorphous) of this particular component. In the case of azimuthal angle dependence, indicating preferred molecular orientation, crystallinity was determined from radial profile after azimuthal averaging of intensity. Molecular orientation was estimated for nanofibres collected on a drum using bundles of fibres being well parallelised. Crystal molecular orientation was estimated from azimuthal scans of the strongest reflection (1 1 0). Orientation factor, fc, of c axis was determined from:
fc ¼
1 ð3hcos2 Uc i 1Þ 2
ð1Þ
3. Results and discussion
MACROMOLECULAR NANOTECHNOLOGY
2
where hcos Uci is an average square of the cosine of orientation angle Uc of c axis (angle between c axis and fibre axis). In the case of PCL crystallizing in orthorhombic system, orientation factor fc, can be determined from:
hcos2 Uc i ¼ 1 2hcos2 Uhk0 i
3.1. Solution properties All of the pure polymers as well as the blends gave clear and transparent solutions. Fig. 1 shows the plots of viscosity vs. shear rate for the pure and blend solutions indicating Non-Newtonian behaviour for all of the investigated solutions. In the case of PCL there is an evident decrease of viscosity starting from 20 s1 being most probably related to molecular orientation. It was investigated, according to Pakravan et al. [27], that shear rate inside our needle is about 5.4 s1. For pure Ge as well as blends of PCL/Ge there is a shear thickening behaviour at very low shear rates with a strong local maximum of viscosity; this maximum can be attributed to temporary gelation occurring in gelatine in the presence of water [28]. Neglecting this local maximum of viscosity at very low shear rates, it is evident from Fig. 1 that the viscosity generally decreases with increasing content of Ge, approaching lowest value for the blend PG1:4, being practically the same as for pure Ge (Table 1). The analysis of electrical conductivity (Fig. 3) indicates an increase of conductivity with increasing content of Ge, approaching highest value for pure Ge. Conductivity for pure PCL solution (1.82 lS/cm for solutions with 5% concentration) is even lower than the value for the pure solvent which is equal to 2.73lS/cm, what is connected with the non-polar character of PCL. Electrical conductivity for blends increases with Ge addition to achieve a maximum of 384 lS/cm for pure Ge solution. Polymer solution viscosity and conductivity are the most critical factors regarding morphology, particularly the diameter of electrospun fibres. For fibre formation, polymer viscosity should be in a particular range, depend-
ð2Þ
2
where hcos Uhk0i is an average square of the cosine of orientation angle Uhk0 of (h k 0) plate (angle between normal to (h k 0) plate and fibre axis). The value of hcos2 Uhk0i is determined from equation:
hcos2 Uhk0 i ¼
R 2p 0
Ihk0 ðUÞ cos2 Uhk0 sin Uhk0 dU R 2p Ihk0 ðUÞ sin Uhk0 dU 0
ð3Þ
where Ihk0(U) is the dependence of intensity of h k 0 reflection on orientation angle recalculated from azimuthal an-
Fig. 1. Viscosity vs. shear rate for 5 wt.% solutions with PCL, Ge and their blends.
Table 1 Viscosity at shear rate 5.4 s1 and electrical conductivity of 5% solutions with PCL, Ge and their blends.
Viscosity (Pa s) at shear rate 5.4 s1 Conductivity (lS/cm)
PCL
PG4:1
PG1:1
PG1:4
Ge
5.7 1.82
1.9 95
1.8 182
0.5 330
0.5 384
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
ent in low concentration, beads or beaded fibres are usually obtained [30]. Regarding electrical conductivity, a
MACROMOLECULAR NANOTECHNOLOGY
ing on a type of polymer and solvent used [29]. If the viscosity is too low and thus polymer entanglements are pres-
2055
Fig. 2. Morphology of nanofibers with various compositions (a) PCL, (b) PG4:1, (c) PG1:1, (d) PG1:4, (e) PC4:1 and (f) Ge spun on a plate (left column) and on a drum (right column, arrows indicate direction of the drum rotation).
MACROMOLECULAR NANOTECHNOLOGY
2056
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
Fig. 2 (continued)
more conductive polymer solution carries more electric charge during the electrospinning process, generating a stronger repulsive force, which provides higher drawing ratios and thus facilitates the formation of thinner fibres. 3.2. Morphology Fig. 2 illustrates morphology of fibres: PCL, PG4:1, PC4:1, PG1:1, PG1:4, and Ge electrospun on a static plate and on a rotating drum. In general, the applied conditions allow for spinning of fibres with smooth surface, free of beads or other imperfections. First, it is evident from Fig. 2 that as expected, spinning on a drum introduces parallelization of nanofibres compared to random architecture obtained using plate collector. Comparison of mean fibre diameter spun from solutions with total concentration of 5% indicates that addition of Ge results in a drastic reduction of mean diameter in comparison to pure PCL nanofibres, approaching 2– 3 times reduction (Fig. 3). This reduction of diameter is related to an increase of solution electrical conductivity as well as to reduction of the solution viscosity in the presence of Ge. It is seen from Fig. 3 that the lowest diameter was obtained for fibres PG4:1, PC4:1, and PG1:1. The fibre cross sections of PG1:4 and pure Ge fibres are flat and ribbon like, whereas for all other samples circular fibre cross sections were obtained (Fig. 2). The larger diameters as measured for PG1:4 and pure Ge are due to the ribbon like fibre shape (the width of these fibres was measured for those sample) (Fig. 3).
Fig. 3. Mean diameter of various nanofibres collected on plate and on drum (for PG1:4 and Ge the width of the ribbons were used as the ‘‘diameter’’).
Several fine fibres observed in Fig. 2d are most probably related to the effect of splitting of main jet; this phenomena is known in literature [31–33]. Additionally, the average diameter for fibres electrospun on drum decreases in comparison to fibres collected on a static plate electrode. This fact is related to a cold drawing process of the as-spun fibres during collection on rotating drum [34]. 3.3. Supermolecular structure Supermolecular structure was analysed using WAXS and DSC. Figs. 4 and 5 illustrate typical WAXS patterns reg-
2057
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
Table 2 WAXS crystal molecular orientation of c axis, fc, of PCL crystals for various nanofibers spun on a rotating drum. PCL
PG4:1
PG1:1
PG1:4 PC4:1
fc 0.424 ± 0.021 0.307 ± 0.015 0.051 ± 0.003 0
0.397 ± 0.02
Fig. 5. WAXS patterns for PC4:1 fibres collected on drum and plate (arrow indicates the direction of a fibre bundle for oriented fibres).
istered for various nanofibres, collected both on a plate and on a rotating drum. The analysis of azimuthal WAXS profiles (Fig. 6, Table 2) of registered reflections clearly indicate that drum collected fibres have a preferred molecular orientation of PCL crystal phase. The lack of dependence of intensity on azimuthal angle for PCL reflections for plate collected fibres cannot be an evidence of the lack of molecular orientation because of random orientation of fibres. However, our previous results obtained by interference-polarizing microscopy on single nanofibres [35] indicate that the global molecular orientation on plate
Fig. 6. An example of WAXS azimuthal profile of (1 1 0) reflection for PCL and PG4:1 collected on a drum (local decrease of intensity at 45° is due to beam stop).
Fig. 7. Typical WAXS radial profiles for investigated fibres collected on a drum.
fibres collected is indeed very weak. The Herman’s orientation factor, fc, for c axis of PCL crystals in nanofibres collected on a drum is 0.424. Addition of Ge or Col leads to a reduction of orientation of PCL crystals, approaching very weak orientation for a blend of PG1:1 or no orientation for PG1:4 (Table 2). Figs. 7 and 8 show corresponding radial profiles (azimuthally averaged in the case of fibres with preferred molecular orientation). The radial profile of pure PCL fibres shows several peaks attributed to orthorhombic crystal structure: (1 1 0) at 2h = 9.65°; (2 0 0) at 2h = 10.75°; (2 1 0) at 2h = 13.4°; (1 2 0) at 2h = 17.28°; (3 1 0) at 18.40°; (2 2 0) at 19.50°; (2 0 7) at 20.0° and amorphous halo with a maximum at 2h = 9.0°. Compared to PCL, WAXS radial profile of pure Ge exhibits only broad halo, being an evidence of the amorphous character of the structure (Fig. 8b). It is important to
MACROMOLECULAR NANOTECHNOLOGY
Fig. 4. WAXS patterns for fibres collected on drum and plate (arrow indicates the direction of a fibre bundle for oriented fibres).
MACROMOLECULAR NANOTECHNOLOGY
2058
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
Fig. 8. Examples of fittings of WAXS radial profiles (a) PCL, (b) Ge, (c) PG4:1, and (d) PC4:1.
note, that the WAXS radial profile of Col found by us in PCL/Col blends, is practically the same as registered for PCL/Ge, indicating an amorphous structure of collagen. The amorphous structure of collagen strongly suggests that Col transforms through denaturation process to the structure of Ge under the influence of the solvent HFIP. The WAXS profiles of blended fibres contain both reflections from PCL and Ge or Col, respectively. They were deconvoluted using the same method, taking into account both the crystal reflections of PCL together with amorphous halo from PCL and Ge or Col (Fig. 8). Examples of the fittings of WAXS radial profiles of investigated nanofibres using Pearson VII function is shown in Fig. 8, while Fig. 9 shows crystallinity of PCL determined from WAXS experiments.
Fig. 9. Crystallinity of PCL for pure PCL and blended fibres as measured by WAXS.
It is seen from Fig. 9 that relatively small addition of Ge or Col leads to an increase of PCL crystallinity (PG4:1 or PC4:1), while higher content of biopolymers results in a reduced crystallinity value. This trend is observed both for fibres collected on the rotating drum and on the plate collector. Moreover, it is evident from Fig. 9 that for pure PCL fibres, crystallinity is as expected higher for drum collected fibres than for plate collected fibres as a result of a colddrawing process caused by the high rotation speed of the drum. For some reason, unknown at the moment, the situation is opposite for blended fibres (Fig. 9). The presence of water in the investigated system is expected on the basis of the hygroscopic nature of Ge, being capable to incorporate water molecules. The experimental evidence of water in nanofibres containing Ge can be drawn from TGA experiments showing the loss of sample mass in the temperature range characteristic for water desorption, 40–100 °C (Fig. 10). Normalization of blend fibres to the Ge mass in the sample allows estimating water content related to Ge as 4% by weight. Our attempt of explanation of the peculiar trend for PCL crystallinity as a function of Ge content is based on the mechanism of hydrophobic interactions. The hydrophobic effect is in general known, being repulsion between water and nonpolar substances, just like PCL molecules, resulting in a tendency of nonpolar substances to aggregate in order to minimise their interphase with water, and exclude water molecules [36]. Such hydrophobic interactions can act as additional driving force for agglomeration (crystallization) of PCL molecules, which is seen at least for fibres with relatively small amount of biopolymers. Hydrophobic
2059
Fig. 10. Water content from TGA experiments for PCL, Ge and blended nanofibers collected on plate.
Fig. 12. Crystallinity of PCL for pure PCL and blended fibres as determined by DSC.
interaction can also explain higher crystallinity of PCL seen in randomly collected blended fibres than in drum collected fibres (Fig. 9). In the latter case, molecular orientation can disturb PCL molecular aggregation caused by hydrophobic interactions. DSC results support in general the data obtained by WAXS. Fig. 11 illustrates typical DSC scans. For pure PCL the endothermic peak related to the melting of the crystalline phase is observed in a temperature range between 30–70 °C, with a maximum at 59.5 °C. One broad endothermic peak is observed for pure Ge and Col fibres in a range 30–130 °C and 30–170 °C, respectively, with corresponding maximum at 75 °C, and 85 °C. Considering that our WAXS results showed no evidence of crystalline structure in both Ge and Col nanofibres, we suggest that this broad endothermic peak is related to water evaporation together with helix-to-coil transition. This conclusion is done on the basis of analysis of the heat, being much higher than evaporation heat of water being present in Ge in an amount as deduced from mass loss by TGA (Fig. 10). Both transitions are interrelated since the water molecules stabilize the helix structure of Ge and the loss of water is a natural driving force towards helix–coil transition. This kind of explanation of the large endotherm for the used biopolymers coincides with literature data [37].
In the case of blended fibres, thermograms are superpositions of the transitions from both components. This is evident for compositions with relatively high content of Ge or Col (P50) (Fig. 11). For smaller amounts of Ge or Col, the endothermic effect due to water evaporation and helix-to-coil transition is visible only as a shoulder. Fig. 12 shows DSC crystallinity of PCL determined from thermograms registered for various compositions. There is very clear tendency of an increase of PCL crystallinity at addition of 20% of Ge or Col over the value for pure PCL and further decrease of PCL crystallinity at higher content of Ge. This is in full accordance with WAXS results. Our previous results, determined by optical birefringence [37], indicate additionally, that PCL crystallinity within nanofibres is lower than in casted films or raw pellets. This is consistent with literature data for various polymers spun either from solution or melt, indicating that crystallization during electrospinning is limited by fast rate of solvent evaporation, leading to formation of small and/or defective crystals [38,39]. DSC results in Fig. 12 do not indicate the same tendency between crystallinity for plate and drum collected fibres as it was for WAXS data (see Fig. 9). This difference between WAXS and DSC results can be attributed to the fact that preferred molecular orientation for drum collected fibres may change as a result of heating during DSC measurements, changing the initial structure and thus heat of melting. 3.4. Mechanical properties
Fig. 11. DSC scans for various nanofibres collected on a plate.
The analysis of the mechanical properties was focused to the Young’s modulus (Fig. 13) determined in the linear range of the stress–strain curves, typically between 2% and 10% of strain. It is seen in Fig. 13 that the general trend is an increase of mechanical stiffness with Ge content. This trend can be explained as a simple effect of additivity of the stiffness from both components, low for PCL and high for Ge. Similar trends of changes are described by Zhang et al. [39] who analysed blends of PCL/Ge 70:30, 50:50 by stress–strain measurements. They also observed an increase of Young’s modulus with increasing gelatine content. Relatively high mechanical stiffness for PG4:1, being almost the same as for PG1:1, can be explained by high crystallinity observed
MACROMOLECULAR NANOTECHNOLOGY
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
2060
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
stances to aggregate in order to minimise their interphase with water, and thus exclude water molecules [41,42]. The presence of water within our investigated blend systems is expected on the basis of the hygroscopic nature of gelatine, being capable of binding water molecules, and evidenced experimentally using TGA. WAXS results show evidently the amorphous structure of both gelatine and collagen within the blended and pure biopolymer fibres. These results indicate strong tendency for denaturation of collagen into gelatine under the influence of the solvent hexafluoroisopropanol. This is consistent with previous results obtained using other methods [43–45]. Fig. 13. Young’s modulus for various nanofibres.
MACROMOLECULAR NANOTECHNOLOGY
Acknowledgements at this gelatine content (see Fig. 9). Increase of stiffness for on drum collected fibres over the values for plate collected material was observed for PCL, Ge and for blended fibres. This increase is most probably attributed to preferred molecular orientation of PCL for drum collected fibres. Samples with Col as an additive show a little bit lower modulus than those with Ge in the same ratio to PCL. This can be related to lower stiffness of pure Col than Ge [40]. Higher modulus of PCL/Col fibres collected on a drum than those collected on a plate is due to higher molecular orientation of PCL like in the case of fibres with Ge.
4. Conclusions Addition of biopolymers like gelatine and collagen to PCL introduces not only specific sequences of amino acids but affects strongly the solution properties and thus electrospinning process, leading to changes in molecular and supermolecular structure of the fibres. This in turn is expected to affect bioactivity of blended nanofibres. First, it is evident that addition of gelatine leads to large reduction of fibre diameter and finally, at high gelatine content to change of transverse shape of fibres. It is observed for rotating drum and thus parallized collected fibres, which indicate preferred molecular orientation of PCL crystals, general tendency for molecular orientation reduction at the presence of biopolymers. This tendency could be explained by the existence of a biopolymer gel structure, being a molecular network, which hinders process of orientation during spinning. In addition, there is particular dependence of PCL crystallinity on the content of gelatine. It is anticipated that increasing amounts of biopolymer should result in reduction of PCL crystallinity due to increasing spatial scattering of PCL molecules. Such a trend is observed only at relatively high content of gelatine (50 to 80%). At low gelatine concentration (20%), an unexpected increase of PCL crystallinity, determined both from WAXS and DSC measurements, was found. In our opinion, it can be explained by hydrophobic interactions in the system containing PCL, gelatine (collagen) and water, being an additional driving force for agglomeration (crystallization) of PCL. The hydrophobic effect is in general known, being repulsion between water and nonpolar substances, just like PCL molecules, resulting in tendency of nonpolar sub-
This research was supported by Program 10.065-SciexN3. The authors thank Prof. Alina Sionkowska (for supplying collagen) and Felix Reifler (for the help in collecting of WAXS data).
References [1] Sell SA, Wolfe PS, Garg K, McCool JM, Rodriguez IA, Bowlin GL. The use of natural polymers in tissue engineering: a focus on electrospun extracellular matrix analogues. Polymer 2010;2:522–53. [2] Cort JR, Andersen NH. Formation of a molten-globule-like state of myoglobin in aqueous hexafluoroisopropanol. Biochem Biophys Res Co 1997;233(3):687–91. [3] Yaylaog˘lu MB, Korkusuz P, Ors U, Korkusuz F, Hasirci V. Development of a calcium phosphate–gelatin composite as a bone substitute and its use in drug release. Biomaterials 1999;20(9):711–9. [4] Callegari A, Bollini S, Iop L, Chiavegato A, Torregrossa G, Pozzobon M, et al. Neovascularization induced by porous collagen scaffold implanted on intact and cryoinjured rat hearts. Biomaterials 2007;28(36):5449–61. [5] Meng ZX, Wang YS, Ma C, Zheng W, Li L, Zheng YF. Electrospinning of PLGA/gelatin randomly-oriented and aligned nanofibers as potential scaffold in tissue engineering. Mater Sci Eng C 2010;30:1204–10. [6] Chew SY, Mi R, Hoke A, Kam A, Leong W. The effect of the alignment of electrospun fibrous scaffolds on Schwann cell maturation. Biomaterials 2008;29(6):653–61. [7] Chen B, Fowler A, Bhowmick S. Forced and natural convective drying of trehalose/water thin ilms: implication in the desiccation preservation of mammalian cells. ASME J Biomechl Eng 2006;128(3):335–46. [8] Chen M, Patra P, Warner S, Bhowmick S. Optimization of electrospinning process parameters for tissue engineering scaffolds. Biophys Rev Lett 2006;1(2):189–214. [9] Lee J, Tae G, Kim YH, Park IS, Kim SH, Kim SH. The effect of gelatin incorporation into electrospun poly(l-lactide-co-3-caprolactone) fibres on mechanical properties and cytocompatibility. Biomaterials 2008;29(12):1872–9. [10] Yang X, Xu Q, Sui G, Ca Q, Deng X. Structure and wettability relationship of coelectrospun poly (l-lactic acid)/gelatin. Polym Adv Technol 2011;22(12):2222–30. [11] Lee JH, Jung HW, Kang IK, Lee HB. Cell behaviour on polymer surfaces with different functional groups. Biomaterials 1994;15(9):705–71. [12] Bhardwaj T, Pilliar RM, Grynpas MD, Kandel RA. Effect of material geometry on cartilagenous tissue formation in-vitro. J Biomed Mater Res 2001;57(2):190–9. [13] Zhang S. Fabrication of novel biomaterials through molecular selfassembly. Nat Biotechnol 2003;21(10):1171–8. [14] Boland ED, Matthews JA, Pawlowski KJ, Simpson DG, Wnek GE, Bowlin GL. Electrospinning collagen and elastin: preliminary vascular tissue engineering. Front Biosci 2004;9:1422–32. [15] Buttafoco L, Kolkman NG, Engbers-Buijtenhuijs P, Poot AA, Dijkstra PJ, Vermes I, et al. Electrospinning of collagen and elastin for tissue engineering applications. Biomaterials 2006;27(5):724–34.
[16] Williamson MR, Black R, Kielty C. PCL-PU composite vascular scaffold production for vascular tissue engineering: attachment, proliferation and bioactivity of human vascular endothelial cells. Biomaterials 2006;27(19):3608–16. [17] Chong EJ, Phan TT, Lim IJ, Zhang YZ, Bay BH, Ramakrishna S, et al. Evaluation of electrospun PCL/gelatin nanofibrous scaffold for wound healing and layered dermal reconstitution. Acta Biomater 2007;3(3):321–30. [18] Lee JJ, Yu HS, Hong SJ, Jeong I, Jang JH, Kim HW. Nanofibrous membrane of collagen–polycaprolactone for cell growth and tissue regeneration. J Mater Sci Mater Med 2009;20(9):1927–35. [19] Jose MV, Thomas V, Dean DR, Nyairo E. Fabrication and characterization of aligned nanofibrous PLGA/Collagen blends as bone tissue scaffolds. Polymer 2009;50(15):3778–85. [20] Zhang Y, Ouyang H, Lim CT, Ramakrishna S, Huang ZM. Electrospinning of gelatin fibers and gelatin/PCL composite fibrous scaffolds. J Biomed Mater Res B Appl Biomater 2005;72B(1):156–65. [21] Hartman O, Zhang Ch, Adams E, Farach-Carson MC, Petrelli NJ, Chasea BD, et al. Biofunctionalization of electrospun PCL-based scaffolds with perlecan domain IV peptide to create a 3-D pharmacokinetic cancer model. Biomaterials 2010;31(21):5700–18. [22] Sionkowska A, Kozłowska J. Characterization of collagen/ hydroxyapatite composite sponges as a potential bone substitute. Int J Biol Macromol 2010;47(4):483–7. [23] Sionkowska A, Skopinska-Wisniewska J, Wisniewski M. Collagen– synthetic polymer interactions in solution and in thin films. J Mol Liq 2009;145(3):135–8. [24] Sionkowska A, Skopinska J, Wisniewski M. Photochemical stability of collagen/poly (vinyl alcohol) blends. Polym Degrad Stab 2004;83(1):117–25. [25] Wisniewski M, Sionkowska A, Kaczmarek H, Lazare S, Tokarev V, Belin C. Spectroscopic study of a KRF excimer laser treated surface of the thin collagen films. J Photochem Photobiol A Chem 2007;188(2– 3):192–9. [26] Kunugi T, Watanabe H, Hashimoto M. Dynamic mechanical properties of poly(p-phenyleneterephthalamide) fibre. J Appl Polym Sci 1979;24(4):1039–51. [27] Pakravan M, Heuzey MC, Ajji A. Coreshell structured PEO-chitosan nanofibers by coaxial electrospinning. Biomacromolecules 2012;13(2):412–42. [28] Schrieber R, Gareis H. Gelatine handbook. Weinheim: Wiley-Vch; 2007. p. 45–118. [29] Li WJ, Shanti RM, Tuan RS. Tissue, cell and organ engineering. In: Kumar ChSSR, editor. Weinheim: Wiley-VCH; 2007. p. 135–87. [30] Tao J, Shivkumar S. Molecular weight dependent structural regimes during the electrospinning of PVA. Mater Lett 2007;61(11– 12):2325–8.
2061
[31] Sanders EH, Kloefkorn R, Bowlin GL, Simpson DG, Wnek GE. Twophase electrospinning from a single electrified jet: microencapsulation of aqueous reservoirs in poly(ethylene-covinyl acetate) fibers. Biomacromolecules 2003;36:3803–5. [32] Garg K, Bowlin LG. Electrospinning jets and nanofibrous structures. Biomicrofluidics 2011;5(1):0134031-19. [33] Yarin LA. Branching in electrospinning of nanofibres. J App Phys 2005;98:064501–64512. [34] Guex AG, Kocher FM, Fortunato G, Körner E, Hegemann D, Carrel TP, et al. Fine-tuning of substrate architecture and surface chemistry promotes muscle tissue development. Acta Biomater 2012;8(4):1481–9. [35] Kolbuk D, Sajkiewicz P, Kowalewski TA. Optical birefringence and molecular orientation of electrospun polycaprolactone fibers by polarizing-interference microscopy. Eur Polym J 2012;2(48):275–83. [36] Panzavolta S, Gioffrè M, Letizia M, Gualandi C, Foroni L, Bigi A. Electrospun gelatin nanofibers: optimization of genipin crosslinking to preserve fiber morphology after exposure to water. Acta Biomater 2011;7(4):1702–9. [37] Zhang Y, Ouyang H, Lim CT, Ramakrishna S, Huang ZM. Electrospinning of gelatin fibers and gelatin/PCL composite fibrous scaffolds. J Biomed Mater Res B: Appl Biomater 2005;72B:156–65. [38] Arinstein A, Burman M, Gendelman O, Zussman E. Effect of supramolecular structure on polymer nanofibre elasticity. Nature Nanotechnol 2007;2:59–62. [39] Hohmann MM, Shin M, Rutledge G, Brenner MP. Electrospinning and electrically forced jets I stability theory. Phys Fluid 2001;13(8):2201–20. [40] Li M, Mondrinos MJ, Gandhi MR, Ko FK, Weiss AS, Lelkes PI. Electrospun protein fibers as matrices for tissue engineering. Biomaterials 2005;26(30):5999–6008. [41] Schmidt P, Dybal J, Trchova M. Investigations of the hydrophobic and hydrophilic interactions in polymer–water systems by ATR FTIR and Raman spectroscopy. Vib Spectro 2006;42(2):278–83. [42] Chou DH, Morr ChV. Protein-water interactions and functional properties. J Am Oil Chem Soc 1979;56:53A–63A. [43] Hong DP, Hoshino M, Kuboi R, Goto Y. Clustering of fluorinesubstituted alcohols as a factor responsible for their marked effects on proteins and peptides. J Am Chem Soc 1999;12:8427–33. [44] Cort JR, Andersen NH. Formation of a molten-globule-like state of myoglobin in aqueous hexafluoroisopropanol. Biochem Biophys Res Commun 1997;233(3):687–91. [45] Zeugolis DI, Khwe ST, Yew ESY, Ekaputra AK, Tong YW, Yung LYL, et al. Electro-spinning of pure collagen nano-fibres – just an expensive way to make gelatin? Biomaterials 2008;29(15):2293–305.
MACROMOLECULAR NANOTECHNOLOGY
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
Contents lists available at SciVerse ScienceDirect
European Polymer Journal journal homepage: www.elsevier.com/locate/europolj
Macromolecular Nanotechnology
Structure and morphology of electrospun polycaprolactone/ gelatine nanofibres D. Kołbuk a,b, P. Sajkiewicz a, K. Maniura-Weber c, G. Fortunato d,⇑ a
Institute of Fundamental Technological Research, Polish Academy of Sciences, Pawinskiego 5B, 02-106 Warsaw, Poland Advanced Fibers, Empa, Swiss Federal Laboratories for Materials Testing and Research, Lerchenfeldstrasse 5, CH-9014 St. Gallen, Switzerland c Materials-Biology Interactions, Empa, Swiss Federal Laboratories for Materials Testing and Research, Lerchenfeldstrasse 5, CH-9014 St. Gallen, Switzerland d Protection and Physiology, Empa, Swiss Federal Laboratories for Materials Testing and Research, Lerchenfeldstrasse 5, CH-9014 St. Gallen, Switzerland
MACROMOLECULAR NANOTECHNOLOGY
b
a r t i c l e
i n f o
Article history: Received 9 January 2013 Received in revised form 11 April 2013 Accepted 29 April 2013 Available online 22 May 2013 Keywords: Electrospinning Nanofibres Blend Gelatine polycaprolactone Molecular structure
a b s t r a c t Blends of polycaprolactone (PCL) and gelatine (Ge), being effective materials for tissue engineering strategies, were electrospun at various conditions and polymer weight ratios. The morphology, the supermolecular structure as well as the mechanical properties of resulting submicron sized fibres have been analyzed in relation to electrospinning conditions and PCL/Ge weight ratio. Compared to pure PCL, Ge addition leads to large reduction of fibre diameter and finally to changes of fibre morphology. For parallelised fibres collected on a rotating drum, preferred molecular orientation of PCL crystals is found. With increasing Ge content a general reduction of molecular orientation is observed. In addition, there is peculiar dependence of polycaprolactone crystallinity on the content of Ge, showing maximum at low Ge concentration (20%) as determined by differential scanning calorimetry (DSC) and wide angle X-ray scattering (WAXS). Such a trend can be explained by hydrophobic interactions in the system containing PCL, gelatine and water, being additional driving forces for crystallization of nonpolar PCL molecules. The presence of water within investigated blend systems has been evidenced experimentally using thermal gravimetric analysis (TGA). Young’s modulus of nonwovens, as determined by uniaxial tensile testing, indicates the effect of additivity of the stiffness of both polymers as well as the influence of preferred molecular orientation. Additional experiments were performed using collagen (Col) as a biopolymeric alternative to Ge. WAXS results show evidently amorphous structure of Col within the blended fibres, indicating strong tendency for denaturation of collagen into gelatine under the influence of hexafluoroisopropanol as a solvent. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Electrospinning is one of the most common methods to produce fibrous substrates for tissue engineering applications by use of natural and synthetic polymers. Ideally, a substrate should mimic both the form and functionality of the native extracellular matrix (ECM) [1]. It is known ⇑ Corresponding author. Address: Empa, St. Gallen, Abteilung 271, F1.43, Lerchenfeldstrasse 5, 9014 St. Gallen, Switzerland. Tel.: +41 (0)58 765 76 77; fax: +41 (0)58 765 78 62. E-mail address: [email protected] (G. Fortunato). 0014-3057/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.eurpolymj.2013.04.036
that morphology, architecture and surface properties have strong influence on cells growth, spreading, activity and functions [2]. Thereby, scaffolds formed by electrospinning are gaining increasing interest due to their tunable properties to provide ideal microenvironments for various kinds of cell types. Naturally occurring polymers such as collagen (Col) and gelatine (Ge) do not cause foreign body response [3,4]. PCL does elicit such a response, though minimal. As major drawbacks the poor mechanical properties and their variable physical properties have been identified [5]. In contrast, synthetic biopolymers have predictable and
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
2. Materials and methodology 2.1. Materials PCL having a molecular weight of Mw = 80,000 g/mol (Sigma–Aldrich, Switzerland) and gelatine (Ge) from por-
cine skin Type A, having average molecular weight of 50,000–100,000 g/mol (Sigma–Aldrich, Switzerland), were used. Collagen (Col) from tail tendons of young rats, characterised in [22–25] was used. The polymers were dissolved in 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) (Fluka, Switzerland). 2.2. Electrospinning process The electrospinning equipment was operated in horizontal mode and consisted of a medical infusion pump, a static plate or a rotating drum collector and two high voltage generators. The positive voltage was connected to stainless steel needle (0.6 mm inner diameter 26 mm length) and the negative one to the collector. A voltage of ±10 kV and flow rate of 7 ll/min was applied. Solutions were prepared at a total polymer concentration of 5 wt.% with following weight ratios for PCL/Ge: 80:20 (PG4:1), 50:50 (PG1:1), 20:80 (PG1:4) and PCL/Col 80:20 (PC4:1). Fibres were collected on static plate and on a rotating drum (approximately 1000 min1). The temperature and relative humidity were recorded during electrospinning process, being in the range of 22– 25 °C, and 35–45%, respectively. 2.3. Characterisation A rheometer (Anton Paar Physica MCR 300, Austria) equipped with a plate – cone system was applied in controlled shear rate mode to assess the shear viscosities in function of the shear rate. Flow curves with shear rates varying from 0.01–500 1/s were recorded at 20 °C. Electrical conductivity of solutions was measured using a Metrohm 660, (Switzerland), conductometer. The calibration of conductivity measurement was performed using 0.001, 0.01 and 0.1 mol/l potassium chloride solutions (KCl, Sigma–Aldrich, puriss). Before use KCL was dried for 24 h at 100 °C. Morphology of fibres was determined by scanning electron microscopy (SEM) (Hitachi S-4800, Japan) using 2 kV accelerating voltage at 10 mA flow current. Small pieces of fibre patches were coated with gold for 2 min before SEM imaging. Mean fibre diameter was calculated from SEM micrographs by averaging 30 measurements. Supermolecular structure in terms of crystallinity and molecular orientation was analysed by wide angle X-ray scattering (WAXS) and differential scanning calorimetry (DSC) supported by thermal gravimetric analysis (TGA). In the case of DSC, a Mettler T 28E (Switzerland) apparatus was used. Samples (ca. 5 mg) were heated from 10 °C to 190 °C at a rate of 10 K/min. Crystallinity of PCL was evaluated while specific heat of fusion of 100% crystalline PCL was taken as 142,9 J/g [26]. Temperature and heat calibration was performed using indium as reference material. Melting temperature and crystallinity were taken usually as an average value from three measurements. Results are presented as mean ± standard deviation. The thermograms were fitted by use of Pearson functions to separate the signals for PCL and Ge. WAXS (Oxford Diffraction, United Kingdom) was performed using X-ray diffractometry using Mo Ka1 radiation
MACROMOLECULAR NANOTECHNOLOGY
reproducible mechanical and physical properties (e.g. tensile strength, elastic moduli and degradation rate) and they can be synthesised with high reproducibility. Most synthetic polymers incorporate strong hydrophobic surface properties which present poor cell adhesion sites and thus promote a low level of cell attachment. Thereby, different approaches have been realised to adjust surface wettability properties e.g. by chemical and coating routes. A timely fashioned way is the use of hydrophilic and water soluble native extracellular matrix materials such as Col or Ge to form a hybrid of natural and synthetic polymers. In that way, scaffolds are generated which combine the advantages of both types of material and mainly alter materials surface properties. Up to now, various research works describe influencing parameters such as morphology and architecture of electrospun patches with respect to cell growth and proliferation (e.g. [5–13]). Focusing on blend nanofibers, Col/elastin [14,15], polyurethane/polycaprolactone (PCL) [16], and PCL/Ge [17], PCL/Col [18,19] have been described. However most of them concentrate on morphological topics, chemical structure and biological investigations (cell morphology, cytotoxicity and proliferation). For instance, Chong et al. showed reasonable growth of human dermal fibroblast on polyurethane – PCL/Ge sandwiched structures being promising candidates for wound healing applications [19]. Zhang et al. prepared PCL/Ge 1:1 electrospun nanofibers and found improved bone marrow stromal cell attachment compared to pure synthetic PCL nanofiber scaffolds. Cellular infiltration was demonstrated on the blend fibres of up to 114 lm in-depth [20]. PCL/Col electrospun fibres were investigated by Lee et al. [18]. Compared to pure PCL, the Col-added nanofibers showed better cell responses (adhesion and growth) in vitro and penetration in vivo. However, there are only few papers describing molecular structure of the two polymers for electrospun blend fibres, which influences surface properties and mechanical properties of the materials. It is known that crystallinity and molecular orientation are important parameters on physical properties and cell-scaffold adhesion and interaction [21]. The aim of this work is to investigate mutual effect of gelatine introduction into PCL electrospun fibres on nonwoven architecture, fibre molecular structure and mechanical properties. Analytical tools such as scanning electron microscopy (SEM), wide angle X-ray scattering (WAXS) and differential scanning calorimetry (DSC) were used to investigate fibre morphology, molecular structure and surface properties. Young’s modulus was evaluated from stress–strain measurements. In comparison, results from Col addition on molecular structure of the blend fibres are presented.
2053
2054
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
gle using Polanyi equation. We used the strongest (1 1 0) reflection, being in the radial angular range of amorphous halo. Assuming that crystal phase is usually oriented much higher than amorphous part, our procedure of orientation calculation without subtraction of amorphous part, can be treated as an estimation of crystal orientation. Thermogravimetric analysis was performed using Mettler Toledo TC50, in the range of 40–600 °C at a heating rate of 10 °C/min, under 30 ml/min nitrogen atmosphere. Mechanical properties have been measured by uniaxial extension (Instron 4205, United States) with a 100 N load cell under a cross-head speed of 10 mm/min. The samples were first cut into 10 40 mm samples. At least three samples were tested for each type of electrospun patch. The thickness was averaged from three measurements performed by cross-sectional SEM method. Porosity was not taken into consideration.
(k = 0.70926 Å). Analysis in transmission geometry was performed for the determination of crystallinity and molecular orientation. WAXS radial profiles registered up to 2h = 33°, were fitted using Pearson VII function both for crystal peaks and amorphous halo, after previous subtraction of the background. Crystallinity of particular component was determined by division of the area of crystal peaks by the total area (crystal and amorphous) of this particular component. In the case of azimuthal angle dependence, indicating preferred molecular orientation, crystallinity was determined from radial profile after azimuthal averaging of intensity. Molecular orientation was estimated for nanofibres collected on a drum using bundles of fibres being well parallelised. Crystal molecular orientation was estimated from azimuthal scans of the strongest reflection (1 1 0). Orientation factor, fc, of c axis was determined from:
fc ¼
1 ð3hcos2 Uc i 1Þ 2
ð1Þ
3. Results and discussion
MACROMOLECULAR NANOTECHNOLOGY
2
where hcos Uci is an average square of the cosine of orientation angle Uc of c axis (angle between c axis and fibre axis). In the case of PCL crystallizing in orthorhombic system, orientation factor fc, can be determined from:
hcos2 Uc i ¼ 1 2hcos2 Uhk0 i
3.1. Solution properties All of the pure polymers as well as the blends gave clear and transparent solutions. Fig. 1 shows the plots of viscosity vs. shear rate for the pure and blend solutions indicating Non-Newtonian behaviour for all of the investigated solutions. In the case of PCL there is an evident decrease of viscosity starting from 20 s1 being most probably related to molecular orientation. It was investigated, according to Pakravan et al. [27], that shear rate inside our needle is about 5.4 s1. For pure Ge as well as blends of PCL/Ge there is a shear thickening behaviour at very low shear rates with a strong local maximum of viscosity; this maximum can be attributed to temporary gelation occurring in gelatine in the presence of water [28]. Neglecting this local maximum of viscosity at very low shear rates, it is evident from Fig. 1 that the viscosity generally decreases with increasing content of Ge, approaching lowest value for the blend PG1:4, being practically the same as for pure Ge (Table 1). The analysis of electrical conductivity (Fig. 3) indicates an increase of conductivity with increasing content of Ge, approaching highest value for pure Ge. Conductivity for pure PCL solution (1.82 lS/cm for solutions with 5% concentration) is even lower than the value for the pure solvent which is equal to 2.73lS/cm, what is connected with the non-polar character of PCL. Electrical conductivity for blends increases with Ge addition to achieve a maximum of 384 lS/cm for pure Ge solution. Polymer solution viscosity and conductivity are the most critical factors regarding morphology, particularly the diameter of electrospun fibres. For fibre formation, polymer viscosity should be in a particular range, depend-
ð2Þ
2
where hcos Uhk0i is an average square of the cosine of orientation angle Uhk0 of (h k 0) plate (angle between normal to (h k 0) plate and fibre axis). The value of hcos2 Uhk0i is determined from equation:
hcos2 Uhk0 i ¼
R 2p 0
Ihk0 ðUÞ cos2 Uhk0 sin Uhk0 dU R 2p Ihk0 ðUÞ sin Uhk0 dU 0
ð3Þ
where Ihk0(U) is the dependence of intensity of h k 0 reflection on orientation angle recalculated from azimuthal an-
Fig. 1. Viscosity vs. shear rate for 5 wt.% solutions with PCL, Ge and their blends.
Table 1 Viscosity at shear rate 5.4 s1 and electrical conductivity of 5% solutions with PCL, Ge and their blends.
Viscosity (Pa s) at shear rate 5.4 s1 Conductivity (lS/cm)
PCL
PG4:1
PG1:1
PG1:4
Ge
5.7 1.82
1.9 95
1.8 182
0.5 330
0.5 384
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
ent in low concentration, beads or beaded fibres are usually obtained [30]. Regarding electrical conductivity, a
MACROMOLECULAR NANOTECHNOLOGY
ing on a type of polymer and solvent used [29]. If the viscosity is too low and thus polymer entanglements are pres-
2055
Fig. 2. Morphology of nanofibers with various compositions (a) PCL, (b) PG4:1, (c) PG1:1, (d) PG1:4, (e) PC4:1 and (f) Ge spun on a plate (left column) and on a drum (right column, arrows indicate direction of the drum rotation).
MACROMOLECULAR NANOTECHNOLOGY
2056
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
Fig. 2 (continued)
more conductive polymer solution carries more electric charge during the electrospinning process, generating a stronger repulsive force, which provides higher drawing ratios and thus facilitates the formation of thinner fibres. 3.2. Morphology Fig. 2 illustrates morphology of fibres: PCL, PG4:1, PC4:1, PG1:1, PG1:4, and Ge electrospun on a static plate and on a rotating drum. In general, the applied conditions allow for spinning of fibres with smooth surface, free of beads or other imperfections. First, it is evident from Fig. 2 that as expected, spinning on a drum introduces parallelization of nanofibres compared to random architecture obtained using plate collector. Comparison of mean fibre diameter spun from solutions with total concentration of 5% indicates that addition of Ge results in a drastic reduction of mean diameter in comparison to pure PCL nanofibres, approaching 2– 3 times reduction (Fig. 3). This reduction of diameter is related to an increase of solution electrical conductivity as well as to reduction of the solution viscosity in the presence of Ge. It is seen from Fig. 3 that the lowest diameter was obtained for fibres PG4:1, PC4:1, and PG1:1. The fibre cross sections of PG1:4 and pure Ge fibres are flat and ribbon like, whereas for all other samples circular fibre cross sections were obtained (Fig. 2). The larger diameters as measured for PG1:4 and pure Ge are due to the ribbon like fibre shape (the width of these fibres was measured for those sample) (Fig. 3).
Fig. 3. Mean diameter of various nanofibres collected on plate and on drum (for PG1:4 and Ge the width of the ribbons were used as the ‘‘diameter’’).
Several fine fibres observed in Fig. 2d are most probably related to the effect of splitting of main jet; this phenomena is known in literature [31–33]. Additionally, the average diameter for fibres electrospun on drum decreases in comparison to fibres collected on a static plate electrode. This fact is related to a cold drawing process of the as-spun fibres during collection on rotating drum [34]. 3.3. Supermolecular structure Supermolecular structure was analysed using WAXS and DSC. Figs. 4 and 5 illustrate typical WAXS patterns reg-
2057
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
Table 2 WAXS crystal molecular orientation of c axis, fc, of PCL crystals for various nanofibers spun on a rotating drum. PCL
PG4:1
PG1:1
PG1:4 PC4:1
fc 0.424 ± 0.021 0.307 ± 0.015 0.051 ± 0.003 0
0.397 ± 0.02
Fig. 5. WAXS patterns for PC4:1 fibres collected on drum and plate (arrow indicates the direction of a fibre bundle for oriented fibres).
istered for various nanofibres, collected both on a plate and on a rotating drum. The analysis of azimuthal WAXS profiles (Fig. 6, Table 2) of registered reflections clearly indicate that drum collected fibres have a preferred molecular orientation of PCL crystal phase. The lack of dependence of intensity on azimuthal angle for PCL reflections for plate collected fibres cannot be an evidence of the lack of molecular orientation because of random orientation of fibres. However, our previous results obtained by interference-polarizing microscopy on single nanofibres [35] indicate that the global molecular orientation on plate
Fig. 6. An example of WAXS azimuthal profile of (1 1 0) reflection for PCL and PG4:1 collected on a drum (local decrease of intensity at 45° is due to beam stop).
Fig. 7. Typical WAXS radial profiles for investigated fibres collected on a drum.
fibres collected is indeed very weak. The Herman’s orientation factor, fc, for c axis of PCL crystals in nanofibres collected on a drum is 0.424. Addition of Ge or Col leads to a reduction of orientation of PCL crystals, approaching very weak orientation for a blend of PG1:1 or no orientation for PG1:4 (Table 2). Figs. 7 and 8 show corresponding radial profiles (azimuthally averaged in the case of fibres with preferred molecular orientation). The radial profile of pure PCL fibres shows several peaks attributed to orthorhombic crystal structure: (1 1 0) at 2h = 9.65°; (2 0 0) at 2h = 10.75°; (2 1 0) at 2h = 13.4°; (1 2 0) at 2h = 17.28°; (3 1 0) at 18.40°; (2 2 0) at 19.50°; (2 0 7) at 20.0° and amorphous halo with a maximum at 2h = 9.0°. Compared to PCL, WAXS radial profile of pure Ge exhibits only broad halo, being an evidence of the amorphous character of the structure (Fig. 8b). It is important to
MACROMOLECULAR NANOTECHNOLOGY
Fig. 4. WAXS patterns for fibres collected on drum and plate (arrow indicates the direction of a fibre bundle for oriented fibres).
MACROMOLECULAR NANOTECHNOLOGY
2058
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
Fig. 8. Examples of fittings of WAXS radial profiles (a) PCL, (b) Ge, (c) PG4:1, and (d) PC4:1.
note, that the WAXS radial profile of Col found by us in PCL/Col blends, is practically the same as registered for PCL/Ge, indicating an amorphous structure of collagen. The amorphous structure of collagen strongly suggests that Col transforms through denaturation process to the structure of Ge under the influence of the solvent HFIP. The WAXS profiles of blended fibres contain both reflections from PCL and Ge or Col, respectively. They were deconvoluted using the same method, taking into account both the crystal reflections of PCL together with amorphous halo from PCL and Ge or Col (Fig. 8). Examples of the fittings of WAXS radial profiles of investigated nanofibres using Pearson VII function is shown in Fig. 8, while Fig. 9 shows crystallinity of PCL determined from WAXS experiments.
Fig. 9. Crystallinity of PCL for pure PCL and blended fibres as measured by WAXS.
It is seen from Fig. 9 that relatively small addition of Ge or Col leads to an increase of PCL crystallinity (PG4:1 or PC4:1), while higher content of biopolymers results in a reduced crystallinity value. This trend is observed both for fibres collected on the rotating drum and on the plate collector. Moreover, it is evident from Fig. 9 that for pure PCL fibres, crystallinity is as expected higher for drum collected fibres than for plate collected fibres as a result of a colddrawing process caused by the high rotation speed of the drum. For some reason, unknown at the moment, the situation is opposite for blended fibres (Fig. 9). The presence of water in the investigated system is expected on the basis of the hygroscopic nature of Ge, being capable to incorporate water molecules. The experimental evidence of water in nanofibres containing Ge can be drawn from TGA experiments showing the loss of sample mass in the temperature range characteristic for water desorption, 40–100 °C (Fig. 10). Normalization of blend fibres to the Ge mass in the sample allows estimating water content related to Ge as 4% by weight. Our attempt of explanation of the peculiar trend for PCL crystallinity as a function of Ge content is based on the mechanism of hydrophobic interactions. The hydrophobic effect is in general known, being repulsion between water and nonpolar substances, just like PCL molecules, resulting in a tendency of nonpolar substances to aggregate in order to minimise their interphase with water, and exclude water molecules [36]. Such hydrophobic interactions can act as additional driving force for agglomeration (crystallization) of PCL molecules, which is seen at least for fibres with relatively small amount of biopolymers. Hydrophobic
2059
Fig. 10. Water content from TGA experiments for PCL, Ge and blended nanofibers collected on plate.
Fig. 12. Crystallinity of PCL for pure PCL and blended fibres as determined by DSC.
interaction can also explain higher crystallinity of PCL seen in randomly collected blended fibres than in drum collected fibres (Fig. 9). In the latter case, molecular orientation can disturb PCL molecular aggregation caused by hydrophobic interactions. DSC results support in general the data obtained by WAXS. Fig. 11 illustrates typical DSC scans. For pure PCL the endothermic peak related to the melting of the crystalline phase is observed in a temperature range between 30–70 °C, with a maximum at 59.5 °C. One broad endothermic peak is observed for pure Ge and Col fibres in a range 30–130 °C and 30–170 °C, respectively, with corresponding maximum at 75 °C, and 85 °C. Considering that our WAXS results showed no evidence of crystalline structure in both Ge and Col nanofibres, we suggest that this broad endothermic peak is related to water evaporation together with helix-to-coil transition. This conclusion is done on the basis of analysis of the heat, being much higher than evaporation heat of water being present in Ge in an amount as deduced from mass loss by TGA (Fig. 10). Both transitions are interrelated since the water molecules stabilize the helix structure of Ge and the loss of water is a natural driving force towards helix–coil transition. This kind of explanation of the large endotherm for the used biopolymers coincides with literature data [37].
In the case of blended fibres, thermograms are superpositions of the transitions from both components. This is evident for compositions with relatively high content of Ge or Col (P50) (Fig. 11). For smaller amounts of Ge or Col, the endothermic effect due to water evaporation and helix-to-coil transition is visible only as a shoulder. Fig. 12 shows DSC crystallinity of PCL determined from thermograms registered for various compositions. There is very clear tendency of an increase of PCL crystallinity at addition of 20% of Ge or Col over the value for pure PCL and further decrease of PCL crystallinity at higher content of Ge. This is in full accordance with WAXS results. Our previous results, determined by optical birefringence [37], indicate additionally, that PCL crystallinity within nanofibres is lower than in casted films or raw pellets. This is consistent with literature data for various polymers spun either from solution or melt, indicating that crystallization during electrospinning is limited by fast rate of solvent evaporation, leading to formation of small and/or defective crystals [38,39]. DSC results in Fig. 12 do not indicate the same tendency between crystallinity for plate and drum collected fibres as it was for WAXS data (see Fig. 9). This difference between WAXS and DSC results can be attributed to the fact that preferred molecular orientation for drum collected fibres may change as a result of heating during DSC measurements, changing the initial structure and thus heat of melting. 3.4. Mechanical properties
Fig. 11. DSC scans for various nanofibres collected on a plate.
The analysis of the mechanical properties was focused to the Young’s modulus (Fig. 13) determined in the linear range of the stress–strain curves, typically between 2% and 10% of strain. It is seen in Fig. 13 that the general trend is an increase of mechanical stiffness with Ge content. This trend can be explained as a simple effect of additivity of the stiffness from both components, low for PCL and high for Ge. Similar trends of changes are described by Zhang et al. [39] who analysed blends of PCL/Ge 70:30, 50:50 by stress–strain measurements. They also observed an increase of Young’s modulus with increasing gelatine content. Relatively high mechanical stiffness for PG4:1, being almost the same as for PG1:1, can be explained by high crystallinity observed
MACROMOLECULAR NANOTECHNOLOGY
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
2060
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061
stances to aggregate in order to minimise their interphase with water, and thus exclude water molecules [41,42]. The presence of water within our investigated blend systems is expected on the basis of the hygroscopic nature of gelatine, being capable of binding water molecules, and evidenced experimentally using TGA. WAXS results show evidently the amorphous structure of both gelatine and collagen within the blended and pure biopolymer fibres. These results indicate strong tendency for denaturation of collagen into gelatine under the influence of the solvent hexafluoroisopropanol. This is consistent with previous results obtained using other methods [43–45]. Fig. 13. Young’s modulus for various nanofibres.
MACROMOLECULAR NANOTECHNOLOGY
Acknowledgements at this gelatine content (see Fig. 9). Increase of stiffness for on drum collected fibres over the values for plate collected material was observed for PCL, Ge and for blended fibres. This increase is most probably attributed to preferred molecular orientation of PCL for drum collected fibres. Samples with Col as an additive show a little bit lower modulus than those with Ge in the same ratio to PCL. This can be related to lower stiffness of pure Col than Ge [40]. Higher modulus of PCL/Col fibres collected on a drum than those collected on a plate is due to higher molecular orientation of PCL like in the case of fibres with Ge.
4. Conclusions Addition of biopolymers like gelatine and collagen to PCL introduces not only specific sequences of amino acids but affects strongly the solution properties and thus electrospinning process, leading to changes in molecular and supermolecular structure of the fibres. This in turn is expected to affect bioactivity of blended nanofibres. First, it is evident that addition of gelatine leads to large reduction of fibre diameter and finally, at high gelatine content to change of transverse shape of fibres. It is observed for rotating drum and thus parallized collected fibres, which indicate preferred molecular orientation of PCL crystals, general tendency for molecular orientation reduction at the presence of biopolymers. This tendency could be explained by the existence of a biopolymer gel structure, being a molecular network, which hinders process of orientation during spinning. In addition, there is particular dependence of PCL crystallinity on the content of gelatine. It is anticipated that increasing amounts of biopolymer should result in reduction of PCL crystallinity due to increasing spatial scattering of PCL molecules. Such a trend is observed only at relatively high content of gelatine (50 to 80%). At low gelatine concentration (20%), an unexpected increase of PCL crystallinity, determined both from WAXS and DSC measurements, was found. In our opinion, it can be explained by hydrophobic interactions in the system containing PCL, gelatine (collagen) and water, being an additional driving force for agglomeration (crystallization) of PCL. The hydrophobic effect is in general known, being repulsion between water and nonpolar substances, just like PCL molecules, resulting in tendency of nonpolar sub-
This research was supported by Program 10.065-SciexN3. The authors thank Prof. Alina Sionkowska (for supplying collagen) and Felix Reifler (for the help in collecting of WAXS data).
References [1] Sell SA, Wolfe PS, Garg K, McCool JM, Rodriguez IA, Bowlin GL. The use of natural polymers in tissue engineering: a focus on electrospun extracellular matrix analogues. Polymer 2010;2:522–53. [2] Cort JR, Andersen NH. Formation of a molten-globule-like state of myoglobin in aqueous hexafluoroisopropanol. Biochem Biophys Res Co 1997;233(3):687–91. [3] Yaylaog˘lu MB, Korkusuz P, Ors U, Korkusuz F, Hasirci V. Development of a calcium phosphate–gelatin composite as a bone substitute and its use in drug release. Biomaterials 1999;20(9):711–9. [4] Callegari A, Bollini S, Iop L, Chiavegato A, Torregrossa G, Pozzobon M, et al. Neovascularization induced by porous collagen scaffold implanted on intact and cryoinjured rat hearts. Biomaterials 2007;28(36):5449–61. [5] Meng ZX, Wang YS, Ma C, Zheng W, Li L, Zheng YF. Electrospinning of PLGA/gelatin randomly-oriented and aligned nanofibers as potential scaffold in tissue engineering. Mater Sci Eng C 2010;30:1204–10. [6] Chew SY, Mi R, Hoke A, Kam A, Leong W. The effect of the alignment of electrospun fibrous scaffolds on Schwann cell maturation. Biomaterials 2008;29(6):653–61. [7] Chen B, Fowler A, Bhowmick S. Forced and natural convective drying of trehalose/water thin ilms: implication in the desiccation preservation of mammalian cells. ASME J Biomechl Eng 2006;128(3):335–46. [8] Chen M, Patra P, Warner S, Bhowmick S. Optimization of electrospinning process parameters for tissue engineering scaffolds. Biophys Rev Lett 2006;1(2):189–214. [9] Lee J, Tae G, Kim YH, Park IS, Kim SH, Kim SH. The effect of gelatin incorporation into electrospun poly(l-lactide-co-3-caprolactone) fibres on mechanical properties and cytocompatibility. Biomaterials 2008;29(12):1872–9. [10] Yang X, Xu Q, Sui G, Ca Q, Deng X. Structure and wettability relationship of coelectrospun poly (l-lactic acid)/gelatin. Polym Adv Technol 2011;22(12):2222–30. [11] Lee JH, Jung HW, Kang IK, Lee HB. Cell behaviour on polymer surfaces with different functional groups. Biomaterials 1994;15(9):705–71. [12] Bhardwaj T, Pilliar RM, Grynpas MD, Kandel RA. Effect of material geometry on cartilagenous tissue formation in-vitro. J Biomed Mater Res 2001;57(2):190–9. [13] Zhang S. Fabrication of novel biomaterials through molecular selfassembly. Nat Biotechnol 2003;21(10):1171–8. [14] Boland ED, Matthews JA, Pawlowski KJ, Simpson DG, Wnek GE, Bowlin GL. Electrospinning collagen and elastin: preliminary vascular tissue engineering. Front Biosci 2004;9:1422–32. [15] Buttafoco L, Kolkman NG, Engbers-Buijtenhuijs P, Poot AA, Dijkstra PJ, Vermes I, et al. Electrospinning of collagen and elastin for tissue engineering applications. Biomaterials 2006;27(5):724–34.
[16] Williamson MR, Black R, Kielty C. PCL-PU composite vascular scaffold production for vascular tissue engineering: attachment, proliferation and bioactivity of human vascular endothelial cells. Biomaterials 2006;27(19):3608–16. [17] Chong EJ, Phan TT, Lim IJ, Zhang YZ, Bay BH, Ramakrishna S, et al. Evaluation of electrospun PCL/gelatin nanofibrous scaffold for wound healing and layered dermal reconstitution. Acta Biomater 2007;3(3):321–30. [18] Lee JJ, Yu HS, Hong SJ, Jeong I, Jang JH, Kim HW. Nanofibrous membrane of collagen–polycaprolactone for cell growth and tissue regeneration. J Mater Sci Mater Med 2009;20(9):1927–35. [19] Jose MV, Thomas V, Dean DR, Nyairo E. Fabrication and characterization of aligned nanofibrous PLGA/Collagen blends as bone tissue scaffolds. Polymer 2009;50(15):3778–85. [20] Zhang Y, Ouyang H, Lim CT, Ramakrishna S, Huang ZM. Electrospinning of gelatin fibers and gelatin/PCL composite fibrous scaffolds. J Biomed Mater Res B Appl Biomater 2005;72B(1):156–65. [21] Hartman O, Zhang Ch, Adams E, Farach-Carson MC, Petrelli NJ, Chasea BD, et al. Biofunctionalization of electrospun PCL-based scaffolds with perlecan domain IV peptide to create a 3-D pharmacokinetic cancer model. Biomaterials 2010;31(21):5700–18. [22] Sionkowska A, Kozłowska J. Characterization of collagen/ hydroxyapatite composite sponges as a potential bone substitute. Int J Biol Macromol 2010;47(4):483–7. [23] Sionkowska A, Skopinska-Wisniewska J, Wisniewski M. Collagen– synthetic polymer interactions in solution and in thin films. J Mol Liq 2009;145(3):135–8. [24] Sionkowska A, Skopinska J, Wisniewski M. Photochemical stability of collagen/poly (vinyl alcohol) blends. Polym Degrad Stab 2004;83(1):117–25. [25] Wisniewski M, Sionkowska A, Kaczmarek H, Lazare S, Tokarev V, Belin C. Spectroscopic study of a KRF excimer laser treated surface of the thin collagen films. J Photochem Photobiol A Chem 2007;188(2– 3):192–9. [26] Kunugi T, Watanabe H, Hashimoto M. Dynamic mechanical properties of poly(p-phenyleneterephthalamide) fibre. J Appl Polym Sci 1979;24(4):1039–51. [27] Pakravan M, Heuzey MC, Ajji A. Coreshell structured PEO-chitosan nanofibers by coaxial electrospinning. Biomacromolecules 2012;13(2):412–42. [28] Schrieber R, Gareis H. Gelatine handbook. Weinheim: Wiley-Vch; 2007. p. 45–118. [29] Li WJ, Shanti RM, Tuan RS. Tissue, cell and organ engineering. In: Kumar ChSSR, editor. Weinheim: Wiley-VCH; 2007. p. 135–87. [30] Tao J, Shivkumar S. Molecular weight dependent structural regimes during the electrospinning of PVA. Mater Lett 2007;61(11– 12):2325–8.
2061
[31] Sanders EH, Kloefkorn R, Bowlin GL, Simpson DG, Wnek GE. Twophase electrospinning from a single electrified jet: microencapsulation of aqueous reservoirs in poly(ethylene-covinyl acetate) fibers. Biomacromolecules 2003;36:3803–5. [32] Garg K, Bowlin LG. Electrospinning jets and nanofibrous structures. Biomicrofluidics 2011;5(1):0134031-19. [33] Yarin LA. Branching in electrospinning of nanofibres. J App Phys 2005;98:064501–64512. [34] Guex AG, Kocher FM, Fortunato G, Körner E, Hegemann D, Carrel TP, et al. Fine-tuning of substrate architecture and surface chemistry promotes muscle tissue development. Acta Biomater 2012;8(4):1481–9. [35] Kolbuk D, Sajkiewicz P, Kowalewski TA. Optical birefringence and molecular orientation of electrospun polycaprolactone fibers by polarizing-interference microscopy. Eur Polym J 2012;2(48):275–83. [36] Panzavolta S, Gioffrè M, Letizia M, Gualandi C, Foroni L, Bigi A. Electrospun gelatin nanofibers: optimization of genipin crosslinking to preserve fiber morphology after exposure to water. Acta Biomater 2011;7(4):1702–9. [37] Zhang Y, Ouyang H, Lim CT, Ramakrishna S, Huang ZM. Electrospinning of gelatin fibers and gelatin/PCL composite fibrous scaffolds. J Biomed Mater Res B: Appl Biomater 2005;72B:156–65. [38] Arinstein A, Burman M, Gendelman O, Zussman E. Effect of supramolecular structure on polymer nanofibre elasticity. Nature Nanotechnol 2007;2:59–62. [39] Hohmann MM, Shin M, Rutledge G, Brenner MP. Electrospinning and electrically forced jets I stability theory. Phys Fluid 2001;13(8):2201–20. [40] Li M, Mondrinos MJ, Gandhi MR, Ko FK, Weiss AS, Lelkes PI. Electrospun protein fibers as matrices for tissue engineering. Biomaterials 2005;26(30):5999–6008. [41] Schmidt P, Dybal J, Trchova M. Investigations of the hydrophobic and hydrophilic interactions in polymer–water systems by ATR FTIR and Raman spectroscopy. Vib Spectro 2006;42(2):278–83. [42] Chou DH, Morr ChV. Protein-water interactions and functional properties. J Am Oil Chem Soc 1979;56:53A–63A. [43] Hong DP, Hoshino M, Kuboi R, Goto Y. Clustering of fluorinesubstituted alcohols as a factor responsible for their marked effects on proteins and peptides. J Am Chem Soc 1999;12:8427–33. [44] Cort JR, Andersen NH. Formation of a molten-globule-like state of myoglobin in aqueous hexafluoroisopropanol. Biochem Biophys Res Commun 1997;233(3):687–91. [45] Zeugolis DI, Khwe ST, Yew ESY, Ekaputra AK, Tong YW, Yung LYL, et al. Electro-spinning of pure collagen nano-fibres – just an expensive way to make gelatin? Biomaterials 2008;29(15):2293–305.
MACROMOLECULAR NANOTECHNOLOGY
D. Kołbuk et al. / European Polymer Journal 49 (2013) 2052–2061