Introduction
Synthetic textiles dominate global textile production due to their low production cost. However, the unsustainable production processes for synthetic textiles, their lack of biodegradability and the resulting pollution by microplastic fibers are of global concern (Stone et al.
2020; Salvador et al.
2017; Royer et al.
2021). To promote the circular economy and sustainable textile production, synthetic textile should be gradually replaced by eco-friendly textiles. However, this implies that biodegradable textile fibers should have some properties that can currently only be achieved using synthetic fibers.
Man-made cellulose fibers produced from wood pulp derived from sustainable forestry or other lignocellulosic raw materials are more sustainable than synthetic fibers and cotton (Shen et al.
2010; Moriam et al.
2021a). With the development of lyocell fiber technology, a much more sustainable technology is now available to produce high-quality textiles compared to the viscose process (Sayyed et al.
2019b,
2019a; Jiang et al.
2020). Additionally, lyocell fiber technology also has the potential to upcycle cellulose-based waste textiles (Haslinger et al.
2019a,
b; Wedin et al.
2019; Ma et al.
2018a,
b), especially those made of natural cellulose fibers or viscose, into higher added value textile fibers.
The high toughness of cellulose-based textile fibers enables the development of sustainable textiles and upcycling of cellulosic waste textiles into durable textile fibers. However, commercially available high tenacity cellulose fibers compromise the good elongation property and vice versa (Sixta et al.
2015). Hence, a simultaneous increase of elongation and tenacity properties is required to enhance the toughness (the integral over the tenacity-elongation curve).
Ioncell® technology is a recently developed lyocell process to produce man-made cellulosic fibers from ionic liquid solutions by dry-jet wet spinning (Sixta et al.
2015; Asaadi et al.
2016; Michud et al.
2015). Ioncell is an environmentally friendly process and considered as an alternative to the viscose and
N-methylmorpholine
N-oxide (NMMO)-based Lyocell processes (Michud et al.
2015; Elsayed et al.
2020). The ionic liquid ‘1,5-diazabicyclo [4.3.0] non-5-ene acetate [DBNH][OAc]’ is used as an excellent cellulose solvent allowing for a rapid dissolution at moderate temperatures and subsequent shaping into continuous filaments (Hummel et al.
2016). Highly orientated cellulose fibers with high tenacity are obtained by coagulation in a cold-water bath (Hauru et al.
2014).
The process parameters, including extrusion velocity, draw ratio, airgap, spinneret geometry, and coagulation bath temperature, considerably influence the mechanical properties of man-made cellulose fiber produced via lyocell process (Hauru et al.
2014; Michud et al.
2016). Michud et al. previously reported the effect of spinneret diameter (150 vs. 100 μm), draw ratio, and coagulation bath temperature on the structural and mechanical properties of the cellulose fiber (Michud et al.
2016). Guizani et al. presented the effect of air gap conditioning on the mechanical properties of the fibers using a monofilament spinning plant (Guizani et al.
2020b). Previously, two studies were performed to observe the effect of spinneret aspect ratio (L/D) on the Ioncell fiber properties. Hauru et al. reported that a L/D of 2 improved the Ioncell fiber orientation compared to a shorter L/D of 0.2 (Hauru et al.
2014). Later, Guizani et al. further justified that the increase in L/D leads to better mechanical properties of the spun fibers (Guizani et al.
2020a). However, both studies were done with a monofilament spinning system, which was only conditionally representative of real spinning conditions. Moreover, there is a lack of systematic knowledge about the geometry of the spinneret capillaries, such as the length-to-diameter ratio of the capillary, the entrance cone geometry, and the use of multi-hole spinnerets.
Spinneret geometry is one of the important spinning parameters that affects the spinnability, the structural properties, and the fiber mechanical properties (Michud et al.
2016; Wang et al.
2018). The influence of the spinneret diameter on the mechanical properties of the NMMO- Lyocell fibers was investigated. Mortimer et al. reported that the narrower the spinneret, the shorter the draw length. In addition, the die swell effect is smaller for the spinneret with holes of a lower diameter due to the efficient cooling (Mortimer and Péguy
1995).
The effect of the spinneret geometry in the solution flow during spinning was investigated rigorously in a series of publications by Xia et al. for the dry-jet wet spinning of the cellulose/1-butyl-3-methylimidazolium [BMIM]CL solution (Xia et al.
2014,
2015,
2016). Xia et al. studied the effect of the entrance angle and the influence of the capillary length in the flow behavior of the spinning solution using numerical simulation (Xia et al.
2014). According to their study, longer spinneret aspect ratio helped to stabilize the viscosity of the cellulose/IL solution promoting the smooth solution flow. Furthermore, the study reported that the die swell effect was reduced by a shorter spinneret entrance angle and an increased length of the exit channel. A shorter entrance angle reduced the velocity gradient along the flow direction and a longer capillary length promoted the relaxation of the polymer chain. In addition, the longer capillary promoted a stable viscosity of the cellulose solution compared to the shorter capillary. The apparent viscosity remained stable due to the molecular orientation in the outlet channel.
Kirichenko et al. suggested the use of rectangular spinnerets instead of the circular spinnerets to reduce the irregularities of the diameter of the spun Capron yarn (Kirichenko et al.
1977). According to their study, the fluctuation of the airflow and the difference in heat transfer from the filaments to the air depend on the arrangement of the spinneret hole. The cooling conditions of the filaments were more uniform when the spinneret holes were arranged in straight lines compared to the concentric circles due to the transverse unilateral air flow. Due to the uniform cooling in the airgap, the rectangular arrangement of spinneret hole produced yarn that had more regularity in the yarn diameter.
We have previously shown that a longer capillary length of the spinneret for a given diameter significantly increased the fiber toughness under otherwise identical spinning conditions (Moriam et al.
2021b).According to that study, Ioncell fiber achieved a modulus of toughness of 83.3 MPa, which presented the potential of Ioncell fibers to have higher strength and toughness than any other existing cellulose fibers. The long-term goal of the study was to reach the toughness of polyester fibers (128 MPa), a common synthetic textile fiber (Moriam et al.
2021b). As a continuation of the previous study, we have used different spinneret geometries with varying spinneret shape (circular to rectangular hole arrangement), entrance angle, spinneret diameter, hole number, and capillary length to observe their effects in mechanical properties in the final fiber. In addition, the influence of spinneret geometries on the solution flow was investigated using numerical simulation. Furthermore, we investigated the combined effect of the optimized spinneret geometry and different pulps (standard grade and high purity pulp) on the mechanical properties of the spun fibers. Finally, the structural properties of the high toughness fibers and their morphological properties were reported.
The effect of spinneret geometry in the spun fibers’ mechanical properties has not been published before systematically for dry-jet wet spinning. This work will present new insight of improving toughness of man-made cellulose fibers toughness by tuning the spinneret geometries.
Experimental part
Raw materials
This study used high-purity pine pre-hydrolysis Kraft pulp additionally purified with cold caustic extraction produced by Georgia Pacific denoted as high purity grade pulp or HPG-pulp (intrinsic viscosity, CUEN, ISO 5351/1, of 594 ml/g; M
n =96.5 kg/mol and M
w = 209 kg/mol, calculated by GPC-MALLS, hemicellulose 0.7%) and birch pre-hydrolysis Kraft pulp denoted as STG-pulp (Standard grade Pulp) (intrinsic viscosity 482 ml/g; M
n =49.3 kg/mol and M
w = 161.7 kg/mol; hemicellulose 6.8%) (Moriam et al.
2021b). Equimolar amount of 1,5-diazabicyclo [4.3.0] non-5-ene (99% Fluorochem, UK) and acetic acid (100% Merck, Germany) were used to synthesize the Ionic liquid 1,5-diazabicyclo [4.3.0] non-5-ene acetate [DBNH][OAc] for the dissolution of cellulose. Elmer GmbH, Austria provided 10 different spinneret geometries having rectangular hole distributions (Fig. S1) and, as a reference, the spinneret with circular hole distribution (Fig. S1) obtained from Enka Tecnica, Germany (Table
1).
Table 1
The list of spinnerets
504 × 100 × 0.2 | 0.2 | 100 | 504 | 221 | 8° | Rectangular (Fig. S1) |
504 × 100 × 1.0 | 1.0 | 100 | 504 |
216 × 100 × 1.5 | 1.5 | 100 | 216 | 95 |
216 × 100 × 2.0 | 2.0 | 100 |
216 × 100 × 4.0 | 4.0 | 100 |
216 × 100 × 5.0 | 5.0 | 100 |
216 × 80 × 0.2 | 0.2 | 80 |
216 × 80 × 1.0 | 1.0 | 80 |
216 × 50 × 0.2 | 0.2 | 50 |
216 × 50 × 1.0 | 1.0 | 50 |
400 × 100 × 0.2 | 0.2 | 100 | 400 | 58 | 60° | Circular (Fig. S1) |
Dope preparation and rheology
The grinded pulps were dissolved in melted [DBNH][OAc] for 90 min at 80 °C using a vertical kneader system. The cellulose concentration of the dope was 13–15 wt%. After dissolution, the dopes were filtered using a press filtration unit (metal filter fleece, absolute fineness of 5–6 µm). The rheological properties of the dopes were measured using the Anton Paar Physica MCR 302 rheometer containing parallel plate geometry (plate diameter 25 mm and gap size 1 mm). A dynamic sweep test (100–0.1 s
−1) was performed to record the complex viscosity and the dynamic moduli (G' and G''). Zero shear viscosity was calculated using a cross viscosity model (Michud et al.
2016) assuming the validity of the Cox-Merz rule.
Dry-jet wet spinning
Dry-jet wet spinning of the cellulose solutions (dopes) were performed using a customized laboratory piston spinning unit (Fourné Polymertechnik, Germany). The molten dope was extruded through spinnerets. The filaments were coagulated in a water bath (8–10 ºC) keeping an airgap of 1 cm and further guided by Teflon rollers to the godet (velocity varied from 5 to 95 m min−1 depending on the obtained draw ratios). The fibers were washed with hot water (80 ºC) for 2 h under continuous magnetic stirring (the water was changed every 30 min). Fibers were air-dried after washing.
Numerical simulation
The velocity and shear rate distributions affecting the molecular alignment of the cellulose strands in the cellulose solutions were evaluated by simulating the flow of the cellulose solution using the COMSOL Multiphysics software using the Computational Fluid Dynamics module. Prior to this, the viscosity cross-model parameters were estimated from shear rate viscosity data of the dope by fitting the model to the experimental data using Wolfram Mathematica (Moriam et al.
2021b). Accordingly, the simulation of velocity and shear rate distributions in a spinneret channel was done in the Computational Fluid Dynamics module of COMSOL, assuming laminar flow and employing the fitted cross-model to describe the shear rate dependence of the viscosity. By virtue of the axisymmetric spinneret channels, it suffices to create a corresponding 2D geometry in which to carry out the simulation.
In the simulation, the integration of the scalar product of the shear stress and shear rate over the channel geometry gives a quantity that summarizes the work per unit time accomplished by the shear forces on the viscous fluid in the channel- the shear power.
Considering an infinitesimal portion of the flow, we can conclude the shear power absorbed to be
$$dP = \sigma \cdot \dot{\gamma }dV.$$
(1)
Here
dP is the (infinitesimal) power absorbed by the volume element
dV whereas
σ and
\(\dot{\gamma }\) are the shear stress and shear rate, respectively. Expressing the shear stress as the viscosity
\(\mu \left( {\dot{\gamma }} \right)\) times the shear rate and integrating over the volume of the channel renders
$$P = \smallint \mu \left( {\dot{\gamma }} \right)\dot{\gamma }^{2} dV$$
(2)
Equipped with the previous equation, we can obtain estimates for the power absorbed in the channels of the different spinneret types by COMSOL simulation.
Tensile properties and birefringence
The tensile properties (tenacity (cN/tex); elongation at break (%) and linear density (dtex)) of the spun fiber samples were measured both in conditioned and wet state using a Favigraph device (Textechno, Germany). Prior to measurement, the fibers were conditioned in 20 ± 2 °C and relative humidity of 65 ± 2%. For each measurement, a total of 20 individual fibers were measured using 20 cN load cell and 20 mm gauge length. The modulus of toughness was calculated using the initial slope of the individual stress strain curves.
The total orientation of the fibers was measured using a polarized light microscope (Zeiss Axio Scope) connected with a 5λ. Berek compensator. The birefringence Δn was calculated from the retardation of the polarized light divided by the thickness of the fiber assuming cellulose density of 1.5 g/cm3 (Männer et al.
2009). The total orientation was calculated from birefringence Δn divided by the maximum value (0.062) of the cellulose’s birefringence (Adusumalli et al.
2009; Lenz et al.
1994).
Wide angle X-ray Scattering (WAXS)
X-ray diffraction data of the precursor fibers were collected in the transmission setting of a SmartLab instrument (RIGAKU) operated at 45 kV and 200 mA. Cellulose fibers were grinded using a Wiley mill with 60 µm mesh size. The samples were then pressed into disks using a pellet press instrument with constant force for 30 s. Powder diffraction data were collected from 5° to 60° 2θ by θ/2θ setting. Scattering profile were corrected for smoothing, subtracting air scattering, and subtracting inelastic contribution. Guizani et al. presented the detailed smoothing procedure which was used to subtract the amorphous cellulose contribution from remained elastic scattering profile (Guizani et al.
2020a). Thus, estimated amorphous contribution (
\({I}_{bkg}\left(2\theta \right)\)) is used to estimate the crystallinity index (CI) using a range from 9° to 50° 2θ:
$$CI = 100*\frac{{\smallint I\left( {2\theta } \right)d2\theta - \smallint I_{bkg} \left( {2\theta } \right)d2\theta }}{{\smallint I\left( {2\theta } \right)d2\theta }}$$
(3)
The background corrected profiles were fitted with four pseudo-Voigt functions for (
\(1\overline{1}0\)), (110), (020), and (002) peaks for cellulose II (Langan et al.
2001) using a LMFIT software (Newville et al.
2016). The (002) peak was added because there was a visible shoulder for the peak in this θ/2θ and transmission geometry. The Scherrer equation was used to estimate the crystal widths (CW
hkl) as follows:
$$CW_{hkl} = \frac{K \lambda }{{\beta_{hkl} \cos \theta }}$$
(4)
where K = 0.90 is the shape factor, λ is the X-ray wavelength, β
hkl is the full width at half maximum (FWHM) of the diffraction peak in radians, and θ is the diffraction angle of the peak. Due to the significant overlap of (110) and (020) peaks, the crystal widths were reported as average values.
The orientation distribution between fiber axis and the crystallographic (020) lattice plane
\(\left( {\cos^{2} \varphi_{020} } \right)\) was estimated using azimuthal scans obtained from the (020) lattice plane of cellulose II allomorph at 21.9° 2θ.
$$\cos^{2} \varphi_{020} = \frac{{\mathop \smallint \nolimits_{0}^{\pi /2} I\left( {\varphi_{020} } \right)\sin \varphi_{020} \cos^{2} \varphi_{020} d\varphi }}{{\mathop \smallint \nolimits_{0}^{\pi /2} I\left( {\varphi_{020} } \right)\sin \varphi_{020} d\varphi }}$$
(5)
The orientation distribution from equatorial diffraction further converted to the orientation distribution between fiber and crystallographic c-axis
\(\cos^{2} \varphi_{c}\) assuming cylindrical symmetry:
$$\cos^{2} \varphi_{c} = 1 - 2 \cos^{2} \varphi_{020}$$
(6)
Hermans orientation parameter
\(\left( {f_{WAXD} } \right)\) was estimated by:
$${\varvec{f}}_{{{\varvec{WAXD}}}} = \frac{{3\cos^{2} \user2{\varphi }_{{\varvec{c}}} - 1}}{2}.$$
(7)
Small angle X-ray Scattering (SAXS)
SAXS experiment was performed in a transmission mode of Xeuss 3.0 (Xenocs) CuKα X-ray instrument operated at 50 kV and 0.6 mA. Fiber samples were placed vertically on the sample holder and X-ray was radiated on the longitudinal side of the fiber. The sample chamber was kept under vacuum condition (= 0.16 mbar) during the measurement. The scattering intensity was recorded on a 2D detector Eiger2 R 1 M (Detris) and later corrected for the cosmic background.
The azimuthal profile was obtained in a scattering vector Q-range of 0.016–0.002 Å
−1 with intensity corrected for the detector orientation, acquisition time, and transmitted flux. The data were not corrected for the sample thickness. The obtained azimuthal profiles were fitted with a pseudo-Voigt function with a constant background. The fitted azimuthal profile was used to estimate the orientation distribution of the equatorial streak
\(\cos^{2} \varphi_{streak} :\)$$\cos^{2} \varphi_{streak} = \frac{{\mathop \smallint \nolimits_{0}^{\pi /2} I\left( {\varphi_{streak} } \right)\sin \varphi_{streak} \cos^{2} \varphi_{streak} d\varphi }}{{\mathop \smallint \nolimits_{0}^{\pi /2} I\left( {\varphi_{streak} } \right)\sin \varphi_{streak} d\varphi }}$$
(8)
Then it was converted into the orientation parameter (
fSAXS) in the same fashion as the Hermans orientation parameter is estimated from the equatorial diffraction assuming cylindrical symmetry:
$$\cos^{2} \varphi_{c} = 1 - 2{ }\cos^{2} \varphi_{streak}$$
(9)
$$f_{SAXS} = \frac{{3\cos^{2} \varphi_{c} - 1}}{2}$$
(10)
Scanning electron microscopy (SEM) imaging
SEM imaging was performed by using a Zeiss Sigma VP (manufactured country Germany). For cross-section imaging, the fiber was prepared by cryofracture in liquid nitrogen. A bundle of fiber was first dipped into liquid nitrogen and snapped. The fractured fiber bundle was then glued onto the conductive support. The samples were sputter-coated with gold to ensure electric conductivity (20 mA current with 90 s coating time). The secondary electron images were taken at a 3 kV operating voltage at ~ 5 k to 10 k magnification.
Conclusion
The development of a circular economy requires strategies to produce cellulosic fibers with high toughness, as fiber toughness is the core for improving the longevity of cellulose-based textiles. Biodegradable and cellulose-based fibers from closed loop operations are the sustainable solution to meet challenges in the textile industry such as microplastic pollution and non-recyclable textiles. Ioncell technology has already shown tremendous potential for textile recycling and the production of cellulose-based textile fibers with superior strength properties compared to any other cellulose-based fibers on the market. In this study, certain process settings were investigated to enhance fiber toughness, giving priority to the spinneret geometry. The spinneret geometry exhibited a significant effect on the mechanical properties of the fiber spun from the ionic liquid [DBNH][OAc]. Numerical simulations showed that the flow properties of the cellulose solution changed with different spinneret geometries. The spinneret with a capillary aspect ratio L/D 1–2 provided optimized shear power of the cellulose solution, which is required for a better mechanical property of the spun fiber. Spinneret L/D 1–2 promoted the highest tenacity and toughness of the fibers. Both D100 and D80 showed comparable fiber properties. The combination of the use of high-purity pulp (13 wt%) and the capillary aspect ratio L/D 1 with a capillary entry cone 8° of the D100 spinneret resulted in fibers with the highest toughness (93 MPa) and a tenacity of 60 cN/tex. With the new strategies presented in this study, Ioncell fibers were able to achieve approximately 12% higher toughness as compared to the previous study (from 83.3 to 93 MPa) (Moriam et al.
2021b). Both total and the crystalline orientation showed no significant changes for the high-toughness fibers spun from different geometries or raw materials. The microvoid orientations were comparatively lower for the HPG-based fibers than for the STG-based fibers, while the fiber morphologies for the spun fibers were very similar at a standard titer, but still showed qualitatively larger pore structures for the HPG fibers.
Finally, this research created a new perspective for improving the toughness of cellulose-based textile fiber which will contribute to the circular economy and sustainability of textile production. Further studies will include the effect of different spinneret geometries at even larger variations on the mechanical properties of Ioncell fibers made from other ionic liquids.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.