Production of porous cellulose aerogel fibers by an extrusion process

G Model

ARTICLE IN PRESS

SUPFLU-3359; No. of Pages 10

J. of Supercritical Fluids xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

The Journal of Supercritical Fluids journal homepage: www.elsevier.com/locate/supflu

Production of porous cellulose aerogel fibers by an extrusion process Ilknur Karadagli a,∗ , Björn Schulz b , Maria Schestakow a , Barbara Milow a , Thomas Gries b , Lorenz Ratke a a b

Institute of Materials Research, German Aerospace Center (DLR), 51170 Cologne, Germany Lehrstuhl für Textilmaschinenbau und Institut für Textiltechnik (ITA) of RWTH Aachen University, Otto-Blumenthal-Str. 1, 52074 Aachen, Germany

a r t i c l e

i n f o

Article history: Received 11 February 2015 Received in revised form 10 June 2015 Accepted 10 June 2015 Available online xxx Keywords: Twin screw extrusion Salt hydrate melt Cellulose aerogel Aerogel fiber Supercritical drying

a b s t r a c t The preparation, production and properties of light weight, porous cellulose aerogels in the form of thin extruded fibers is compared to monolithic pieces. The cellulose aerogels were synthesized from microcrystalline cellulose in a hydrated calciumthiocyanate salt melt, which upon cooling forms a gel at around 80 ◦ C. Twin screw extrusion experiments were performed systematically yielding thin and wet cellulose filaments. Washing and coagulation of the wet gels in ethanol was followed by supercritical drying with CO2 yielding cellulose aerogel filaments. These were characterized with regard to envelope density, nitrogen adsorption-desorption (BET) analysis, thermal conductivity measurements, tensile and compression tests and scanning electron microscopy (SEM). The microstructure can be described as an open porous network of nano-fibrils with pore sizes ranging from 10 to 100 nm and fibril diameters of around 10 to 25 nm. The densities of supercritically dried (SCD) cellulose aerogels were in the range of 0.009–0.137 g/cm3 and the BET specific surface areas (SSABET ) were between 120 and 230 m2 /g. The cellulose aerogel possessed thermal conductivities from 0.04 to 0.075 W/m.K and compressive moduli up to 16.2 MPa. The tensile strength of aerogel filaments increases with the increasing cellulose amount in the spin dope. Extruded cellulose aerogel filaments show a dependency of their specific surface area on the extrusion temperature: the higher the spinning temperature the higher the surface area. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Cellulose aerogels are multifunctional solid materials with low density, low thermal conductivity, large internal surface area and high porosity. Their structure can be described as a random threedimensional network of cellulose fibrils of a few ten nanometers in diameter. They are promising materials for various applications such as nanostructured, bio-based materials for thermal insulators, insulation materials for the aerospace and aviation industry, storing media for gases in fuel cells, filter materials for extremely fine particles, drug delivery systems, drug coating purposes, hygiene products, highly functional fibers in textile and filtering applications were recently reviewed in [1,2]. Several methods are described in the literature to synthesize cellulose aerogels. These aerogels can be synthesized by dissolving cellulose in suitable media, such as sodium hydroxide [3–5], an aqueous alkali hydroxide/urea solution [6–8], salt hydrate melts

∗ Corresponding author at: Institute of Materials Research, German Aerospace Center (DLR) Linder Hoehe, 51170 Cologne, Germany. E-mail addresses: [email protected] (I. Karadagli), [email protected] (B. Milow).

[9–12], NMMO (N-Methylmorpholin-N-oxide) [3,13–15] or other ionic liquids [16–18]. Synthesis followed by washing, coagulation and finally drying by using special techniques, like supercritical or freeze drying in order to preserve their solid nano-porous network structure. The first cellulose aerogels were prepared by Tan et al. [19]. They used cellulose acetate as a starting material and de-esterified it. The cellulose ester was cross-linked in an acetone solution with toluene-2,4-di-isocyanate. Jin and co-workers [11] developed another technique to produce high-quality cellulose aerogels. They used a so-called salt-hydrate melt as a dissolving agent, being a mixture of water and Ca(SCN)2 at a composition close to the coordination number of the salt cation (Ca(SCN)2 ·4H2 O). Their technique avoids the utilization of toxic isocyanates and allows in contrast to the method of Tan et al. [19] to use lower amounts of cellulose. The best structure preservation results have been obtained by supercritical CO2 drying (SCD). Supercritically dried cellulose aerogels can present porosities as high as 99 % [12]. It is also possible to produce open porous and nano-structured cellulose aerogel fibers [20]. This paper mainly focuses on production and characterization of cellulose aerogel filaments or fibers, a term we synonymously use, based on the work of Jin et al. [11] and our previous works

http://dx.doi.org/10.1016/j.supflu.2015.06.011 0896-8446/© 2015 Elsevier B.V. All rights reserved.

Please cite this article in press as: I. Karadagli, et al., Production of porous cellulose aerogel fibers by an extrusion process, J. Supercrit. Fluids (2015), http://dx.doi.org/10.1016/j.supflu.2015.06.011

G Model SUPFLU-3359; No. of Pages 10

ARTICLE IN PRESS I. Karadagli et al. / J. of Supercritical Fluids xxx (2015) xxx–xxx

2

Table 1 Extrusion parameters. Sample

Cellulose content (wt.%)

Extrusion conditions (temperature, twin screw rotation speed, nozzle hole diameter)

1 2 3

2 3 3 3 3 3 3 4 4.5 5 6

95 ◦ C, 200 rpm, Ø 0.5 mm 93 ◦ C, 200 rpm, Ø 0.5 mm 95 ◦ C, 200 rpm, Ø 0.5 mm 100 ◦ C, 200 rpm, Ø 0.5 mm 105 ◦ C, 100 rpm, Ø 1 mm 105 ◦ C, 200 rpm, Ø 0.5 mm 115 ◦ C, 250 rpm, Ø 0.5 mm 95 ◦ C, 200 rpm, Ø 0.5 mm 95 ◦ C, 200 rpm, Ø 0.5 mm 95 ◦ C, 200 rpm, Ø 0.5 mm 110 ◦ C, 50 rpm, Ø 1 mm

4 5 6 7 8 9 10

of the samples, they were covered with ethanol. During cooling the viscosity rises and below 80 ◦ C a stable gel is formed. 2.3. Production of cellulose aerogel fibers Fig. 1. Schematic diagram illustrating the preparation of cellulose aerogels.

[2,12,20] using supercritical CO2 drying. The results are compared with those of cellulose aerogel monoliths. 2. Materials and methods 2.1. Materials The chemicals used were calciumthiocyanate tetrahydrate (Ca(SCN)2 ·4H2 O) with a purity of 95% and cellulose fibers (medium) powder both from Sigma–Aldrich, product numbers 20,144 and C6288. The degree of polymerisation of cellulose used was determined by us as DPV = 211 having a molecular weight of 32,998. The Cuoxam-method was used based on the change in viscosity of cellulose in Cuoxam solution [21]. The cellulose fiber powder was dried at 105 ◦ C before use to get rid of adsorbed humidity. Ethanol (euro denatured 96%, TechniSolv.) from VWR Chemicals was used as received for washing and coagulation steps. Later on before the supercritical drying step, samples were washed with absolute ethanol (≥99,8 %, with ca. 1% MEK, Carl Roth). Carbondioxide (CO2 ) with a purity of 99.9 % was used as received. 2.2. Preparation of cellulose wet gels The preparation procedure of cellulose aerogels is summarized in Fig. 1. A certain amount (25 g) of calciumthiocyanate tetrahydrate was filled into a beaker, cellulose (0.14–1.74 g) was added in different amounts (0.5–6 wt.%) and two additional moles of deionized water (3,94 g) added finally. The amount of deionized water was calculated due to the coordination number of the cation, calcium(II), so that 6 H2 O molecules deliver the best composition of the dope. The turbid suspension is heated up to 110 ◦ C while stirring. Dissolution of cellulose takes 10–30 min depending on the cellulose concentration and the final temperature. The larger the amount of cellulose the longer it takes to dissolve the cellulose in the salt hydrate melt. When the solution became homogeneous and transparent the cellulose was assumed to be completely dissolved. This was also supported by DP measurements. The DP remains constant up to 60 min and then decreases considerably. The homogeneous viscous solution was transferred to appropriate polypropylene molds for the production of monoliths, or to the twin-screw extrusion apparatus for fiber production. To avoid crystallization and the building of a solid cellulose skin at the surface

Cellulose aerogel fiber production was conducted with a twin screw extrusion facility. After dissolving cellulose in the salt melt hydrate, the homogeneous viscous solution was transferred to a micro-extruder Xplore 15 (DSM, Geleen, Netherlands) for fiber production (Fig. 2). This instrument is utilized for formulation development and screening of polymer materials and polymer blends and polymer composites, and it is possible to produce fibers with it. The micro-compounder can process batch volumes up to 15 ml. The compounder gives the opportunity to select batch volumes of 3, 7 or 15 ml, via multiple recirculation channels. It is formed by a divisible, fluid tight mixing compartment containing two detachable, conical mixing screws to ensure a maximal homogenization (Fig. 2a and b). The main drive is continuously digitally variable. It allows for vertical force and rheological data measurement and constantly controls the pressure for film and fiber productions, by throughput control. The processing temperature can be controlled in separate barrel heating zones, which also enables to process with a temperature gradient over the barrel, or directly via an additional melt thermocouple. Residence time of the spin dope in the extruder can be varied via recirculation of the melt. No screw optimization is needed. Mixing and dispersion are superb, preventing agglomeration. Shear can be set by adjusting the temperature, rotation speed of the screws and gap between screw flank and barrel [22]. 15 ml of viscous cellulose-salt melt hydrate solution were filled into the extruder by loading over a heated hopper in each experiment (Fig. 2c). The complete extruder apparatus was heated up to 250 ◦ C. After reaching 250 ◦ C and waiting for warming the hopper, the extruder was cooled down to reach the extrusion temperature for the experiments. Various extrusion parameters such as temperatures in the range of 95–115 ◦ C, twin screw rotation speeds (50, 100 and 200 rpm) and diameters of the spinnerets (0.5 and 1 mm) were used (Table 1). The extruded wet cellulose gel fibers were directly spun into an ethanol bath at room temperature. The gels were washed and coagulated within an ethanol bath to get rid of the calciumthiocyanate salts from the wet gel body. Washing was performed by using a Soxhlet apparatus for three days. It could be replaced with ionexchange process for large scale implementation which we are also using currently to coagulate such gels within this lab scale process. Residual salt traces in the gel body were determined with conductometry until the conductivity of the washing solution is negligible low (<1 ␮S) and a spot test with 1% iron(III) nitrate solution is negative. The dry gels were also analyzed by EDX in a scanning electron microscope to check the salt melt hydrate content and to prove

Please cite this article in press as: I. Karadagli, et al., Production of porous cellulose aerogel fibers by an extrusion process, J. Supercrit. Fluids (2015), http://dx.doi.org/10.1016/j.supflu.2015.06.011

G Model SUPFLU-3359; No. of Pages 10

ARTICLE IN PRESS I. Karadagli et al. / J. of Supercritical Fluids xxx (2015) xxx–xxx

3

Fig. 2. Scheme of the extrusion facility: (a, b) two conical, fully intermeshing co-rotating twin screws, (c) Filling of a viscous cellulose-salt melt hydrate solution into an extruder.

that there are no traces of calcium thiocyanate in the materials (see below). 2.4. Supercritical drying (SCD) with CO2 All gels washed with ethanol were dried using supercritical CO2 in a house built 1.6 L capacity batch-wise tubular autoclave developed at DLR. The alcogel was placed on a stainless steel filter plate inside the tubular autoclave with a known excess of ethanol. The autoclave has a water heating and cooling system controlled by a Julabo F12 controller and temperature of cooling was first fixed at 290 K. The apparatus was closed and the system pressurized with CO2 up to 45 bar until the samples were completely immersed in the liquid. Then, an outlet flow of CO2 through the column was set at 0.5 bar/min manually. After 30 min of equilibration, the outlet CO2 stream was vented to ambient where the ethanol-rich liquid stream was collected in plastic vials. With several washing steps, extraction of ethanol run for around a week with a renewal of CO2 at almost every 2 h to replaced continuously the ethanol in the gel pores by liquid CO2 . Before going to the supercritical conditions it was ensured that the ethanol was fully exchanged with liquid CO2 . After extracting all excess amounts of ethanol, tubular autoclave was step wise heated by increasing the temperature of the heating system (298, 303, 308, 313 and 318 K) over the critical point of CO2 stopping at a pressure of around 90 bar. After staying at supercritical conditions for 2 h of equilibration, the fluid is then slowly vented at constant temperature, which results in a drop in pressure. Depressurizing was done slowly manually by a venting CO2 to reach atmospheric pressure. When ambient pressure is reached, the vessel is cooled to room temperature and then opened. Prior and after SCD CO2 drying, mass and volume of the cellulose wet gels and aerogels, were recorded to calculate the density of the materials and the extent of shrinking upon the drying step. 2.5. Characterization The cellulose aerogels were characterized with various methods. The envelope density of monolith shape cellulose aerogels was determined using the GeoPyc 1360 envelope density analyzer (Micromeritics). The envelope density of the fiber forms of cellulose aerogels was estimated by dividing their weight by their volume as measured by digital caliper for the diameter at several points of a fiber and the fiber length and diameter and height for monoliths. The helium pycnometry (AccuPyc II 1340 from Micromeritics) was used to determine the skeletal density of cellulose aerogels and the value was found to be 1.501 g/cm3 . This value well agrees with some other given in the literature [23,24]. Specific surface area (SSABET ) of aerogels was determined by N2 adsorption at 77 K (using a Tristar II 3020, from Micromeritics, Norcross, USA) by using

the BET (Brunauer-Emmett-Teller) method. Prior to the adsorption measurements, the aerogel samples were dried under vacuum (<1 Pa) at 383K for 12 h. The pore size distribution was obtained using the Barrett-Joyner-Halenda (BJH) method. The percent porosity, , was determined using the equation,  = (1−e /s ), where e is the envelope density and s is the skeletal density of cellulose. The thermal conductivity of aerogels was measured by Hot Disk thermal constant analyzer (TPS2500) by using transient plane source technique at room temperature. The microstructure of aerogels was observed by scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (SEM: Merlin-Carl Zeiss Microscope; gold sputtered samples; SEM-EDX: Leo 1530 VP with EDX-Analyzer; EDX detector: Oxford; the samples were not sputtered). The fractured surfaces were sputtered with thin gold layer prior to SEM imaging in order to reduce the effect of charging. For the same reason a low accelerating voltage (2–3 kV) was used. The compression tests were performed using a table top testing machine from Chatillon (TCD200-SS Ametek). The specimens were cut into cubes for analysis. A load cell of 1 kN was used and a compression rate of 1 mm/min. The compressive modulus was measured from the slope of the stress-strain curve. Measurement of tensile strength and elongation of the produced cellulose aerogel fibers was carried out with Favimat+ single fiber testing machine from Textechno (Herbert Stein GmbH & Co., KG, Mönchengladbach, Germany). The cellulose aerogel fibers were tested in the tensile mode and clamped between spring-loaded grips. The initial gauge length was set to 20 mm and initial load at 0.5 cN/tex, and crosshead speed at 2 mm/min. 10 tests were carried out for each fiber, and the average load-elongation curve was considered. Maximum elongation (%) and maximum tensile force (cN) of fibers were determined on the basis of DIN EN ISO 5079 and DIN EN ISO 1973.

3. Results and discussion Cellulose aerogel fibers and monoliths with various cellulose contents after supercritical drying are shown in Fig. 3. The shrinkage is calculated by volume change of wet gels before and after SCD drying and the shrinkage was found to be below 15%. The materials are white and opaque. In contrast to silica aerogels and other particulate aerogels [2] no visible dust particles are released on touching these materials and they are not fragile. They give the tactile impression of a dry soft sponge. A comparison of the envelope densities of cellulose aerogel monoliths and fibers with various cellulose contents is given in Fig. 4. The density values vary in the range of 0.009–0.137 g/cm3 with cellulose content varying between 0.5 and 6 wt.% for aerogel monoliths. Cellulose aerogel fibers have a similar range of envelope densities compared to cellulose aerogel monoliths. There is a linear relation between density and cellulose content as established in our

Please cite this article in press as: I. Karadagli, et al., Production of porous cellulose aerogel fibers by an extrusion process, J. Supercrit. Fluids (2015), http://dx.doi.org/10.1016/j.supflu.2015.06.011

G Model SUPFLU-3359; No. of Pages 10 4

ARTICLE IN PRESS I. Karadagli et al. / J. of Supercritical Fluids xxx (2015) xxx–xxx

Fig. 3. Cellulose aerogels (a) fibers and (b) monoliths obtained by ethanol regeneration and supercritical CO2 drying.

previous work [12]. The slope of Fig. 4 should agree with the density of the salt hydrate melt, it means the density depends linearly on the concentration of the cellulose and then gives an intercept value of zero. It should be noticed that Jin and co-workers [11] did not observe the linear relation between density of the aerogels and cellulose concentration. Innerlohinger et al. for cellulose aerogels prepared via NMMO-route [13], observed a similar relation as we did. First their density data at zero cellulose content has no intercept value of zero but a finite value and second they observe a saturation at higher contents, which shows that their dissolution medium is not effective enough to separate all nanofibrils. Our finding in Fig. 4 also shows that our aerogels are free from salt traces. It is also confirmed by EDX analysis and there are no traces of the salt found. The specific surface area (SSABET ) of the cellulose aerogel monoliths was measured with nitrogen gas adsorption analysis and the results are shown in Fig. 5 and Table 2. The BET specific surface areas of cellulose aerogel monoliths are in the range of 150–220 m2 /g similar to observations of other authors [9,11,13,19,25,26]. BET surface area was calculated over the relative pressure (P/P0 ) range of 0–0.3. The isotherms can be categorized as an IUPAC type IV with a hysteresis loop in the range of 0.8–1.0, indicating the presence of meso and macro-porous structures [27,28]. Furthermore, our research group published a mathematical model on the BET surface

Fig. 5. Adsorption-desorption isotherms of supercritically dried cellulose aerogels monoliths.

Table 2 Textural properties of cellulose aerogels monoliths at different cellulose contents.

Fig. 4. Envelope density values of cellulose aerogels regenerated in ethanol and dried with supercritical CO2 .

Cellulose content

Specific surface areas

Porosity

(wt.%)

SSABET (m2 /g)

(%)

Apparent average fibril diameter dave,F (nm)

0.5 1 2 3 3.5 4 5 6

206 ± 0.7 222 ± 0.9 208 ± 0.6 186 ± 1.0 195 ± 0.8 163 ± 1.3 156 ± 0.8 159 ± 0.7

99 99 98 96 95 94 93 91

13 12 13 14 14 16 17 17

Please cite this article in press as: I. Karadagli, et al., Production of porous cellulose aerogel fibers by an extrusion process, J. Supercrit. Fluids (2015), http://dx.doi.org/10.1016/j.supflu.2015.06.011

G Model SUPFLU-3359; No. of Pages 10

ARTICLE IN PRESS I. Karadagli et al. / J. of Supercritical Fluids xxx (2015) xxx–xxx

5

area of cellulose aerogels (0.5–3 wt.%). This model gives the correlation of the specific surface area and the fibril diameter (dave,F ) in cellulose aerogels from Ca(SCN)2 salt hydrate melt solution. The mass specific surface area should be independent of the cellulose concentration. In this study we observed that for higher than 4 wt.% cellulose there is a decrease in specific surface area up to 20%. In the model of Hoepfner et al. [12] this would indicate that the number density of nano-fibril network nodes increases leading to a decrease in specific surface area. Another explanation could be an increase in mean fibril diameter, since the specific surface area per volume is inversely proportional to the fibril diameter. The surface area per mass is just that per volume divided by the envelope density. We can use this relation to estimate the apparent average fibril diameter (dave,F ) from the skeletal density data and the specific surface area per mass (SSABET ). The average fibril diameter, dave,F , was calculated by using the relation dave,F = 4/(s ∗SSABET ) [12,28,29], where ␳s is the skeletal density of cellulose (s = 1.501 g/cm3 ). Using this relation we calculate fibril diameters in cellulose aerogels in the range of 12–17 nm. The estimated values are also shown in Table 2. The specific surface area is not constant for cellulose aerogel fibers but depends on the extrusion temperature as shown in Fig. 6. There seems to be a linear relation between specific surface area and extrusion temperature. We assume that two effects are responsible for this dependency. It has to be considered that in all cases the cellulose was dissolved at 110 ◦ C followed by cooling down to the given extrusion temperature. This means that the cellulose was completely dissolved as in the comparative studies on monolithic material, but while cooling the cellulose polymer strands already aggregate into fibrils. The higher the viscosity the lower the tendency to dissolve the starting material, and the smaller the displacements of the stiff cellulose molecules enabling them to meet each other and form fibrils. Therefore one would expect: low spinning temperature thinner fibrils and thus higher surface area. One could also argue in a different way: There is a temperature range between fully dissolved cellulose above 110 ◦ C and the gelled material below 80 ◦ C. If a spin dope is extruded from a temperature above the upper limit, the cooling of the filament after leaving the extruder nozzle will induce aggregation of the polymer strands and fibrils form to a certain radius given by shear flow assisted aggregation while the cellulose dope is mixed in the extruder [30,31]. If the spinning is performed from a temperature close to the gel-point, fibrils already exist in the spin dope, since the cellulose polymers are no longer completely dissolvable at that temperature. Thus a

pre-fibrillated cellulose material is spun into a filament and the free single polymers aggregate and connects with the already existing fibrils. The amount and the thickness of the fibrils are determined by the temperature between these two limits and the time spend in the extruder. The higher the temperature the smaller the amount and the thinner the fibrils formed. Here it has to be taken into account, that cooling to the extrusion temperature takes time and that during cooling down the cellulose containing salt hydrate melt is running in a circle (Fig. 2b). Then the linear relation could simply mean that the fibrils will become thicker and directly reducing the specific surface area, which is inversely proportional to the fibril diameter. The thermal conductivity of cellulose aerogel monoliths varies linearly with the cellulose content from 0.04 to 0.075 W/m.K at atmospheric pressure and room temperature (Fig. 7). Extrapolating to zero in cellulose content the conductivity of air (0.26 W/m.K) is almost achieved with a value of 0.032 W/m.K. Our thermal conductivity value for 3 wt.% cellulose aerogel monolith is well agree with the work done by Demilecamps et al. [32]. In most cases, heat transfer within aerogels is based on three mechanisms: heat conduction via solid state, heat transfer from gaseous phase presents in the open-porous aerogel structure and radiative heat transfer. The conductivity of a random 3D cellulose fiber network is therefore determined mainly by gas phase via heat conduction, here being of diffusive nature, since the pores are smaller than a micron and additionally conduction through the solid skeleton. Therefore the overall conductivity can be described by the weighted addition of heat conduction through the gas phase and solid state conduction through the cellulose fibers [32–35]. Therefore the thermal conductivity should vary linearly with the cellulose content in the range of low cellulose concentrations. The high absolute value of the thermal conductivity, compared to other nanostructured aerogels, is attributed to two issues. First from our experience we suppose that the Hot Disk-measurement technique always gives values larger than conventional stationary guarded hot plate methods (probably by around 20% larger values). Second, the aerogels contain large pores up to a micrometer and thus a Knudsen effect on the gas phase contribution to heat conduction cannot be expected or at least is small. Fig. 8 shows scanning electron micrographs of cellulose aerogel monoliths for 1 and 3 wt.% cellulose content. The microstructure is typical for cellulose aerogels: a spongy mesh of randomly oriented cellulose nano-fibrils connected in 3D with pore sizes ranging from

Fig. 6. Specific surface areas of cellulose aerogel fibers (3 wt.%) with respect to extrusion temperature.

Fig. 7. Thermal conductivity values of cellulose aerogel monoliths at room temperature. Solid line is linear approximation (±0.03 W/m.K).

Please cite this article in press as: I. Karadagli, et al., Production of porous cellulose aerogel fibers by an extrusion process, J. Supercrit. Fluids (2015), http://dx.doi.org/10.1016/j.supflu.2015.06.011

G Model SUPFLU-3359; No. of Pages 10 6

ARTICLE IN PRESS I. Karadagli et al. / J. of Supercritical Fluids xxx (2015) xxx–xxx

Fig. 8. Scanning electron micrographs of cellulose aerogel monoliths obtained by ethanol regeneration and supercritical CO2 drying: (a) 1 wt.% and (b) 3 wt.% cellulose content.

10 to 100 nm and cellulose fibril diameters of around 10–20 nm. This was reported in the literature several times [2,11,12,19]. The extrusion by a twin screw extruder produced fibers with a different type of morphology is observed, we call it mantle-core structure; a macro-porous outer ring or shell with an inner core of a nano-porous cellulose network as known from monoliths. Selected SEM images are shown in Figs. 9 and 10. Here we used a special preparation technique for SEM analysis. Since we observed that the fibers do not break in a brittle manner, but are rather soft and deform during breaking, we dipped the fibers into liquid nitrogen and broke them while submerged. This yields fracture surfaces looking less deformed and modified. In that way we could preserve the porous structure for SEM analysis. Results are shown in Figs. 9 and 10 . The macrostructure of the fibers can be explained by evaporation/condensation phenomena, while the hot salt melt solution

enters the room temperature ethanol coagulant. After leaving the extrusion nozzle a gel fiber passes an air gap of about l = 5 cm length before entering the ethanol coagulation bath. The time the filaments are in this air gap is calculated as ta = l/v, with v the extrusion speed. The extrusion speed can be varied in the Xplore machine between 100 and 5000 mm/min. At the different rotation speeds of the twin-screw barrels we had extrusion speeds ranging from 16 to 66 mm/s. Thus the wet fibers are exposed to air between 0.75 and 3.1 s. During this time a thin film of rather dense cellulose at the surface of the extruded fiber is formed due to the evaporation of the solvent liquid salt-melt hydrate. It is present in all fibers produced and visible in all SEM images. The hot gel then enters the ethanol bath which vaporizes immediately. The ethanol vapor enters the surface or mantle region via gas diffusion through the film, where it forms bubbles mainly aligned radially in the mantle region. The cooling of the fiber in the ethanol bath stops this pro-

Fig. 9. SEM images of 3 wt.% cellulose aerogel fibers produced by twin screw extrusion exhibiting a mantle-core structure with different extrusion conditions: (a, b) 105 ◦ C, 100 rpm, Ø 1.0 mm and (c, d) 100 ◦ C, 200 rpm, Ø 0.5 mm.

Please cite this article in press as: I. Karadagli, et al., Production of porous cellulose aerogel fibers by an extrusion process, J. Supercrit. Fluids (2015), http://dx.doi.org/10.1016/j.supflu.2015.06.011

G Model SUPFLU-3359; No. of Pages 10

ARTICLE IN PRESS I. Karadagli et al. / J. of Supercritical Fluids xxx (2015) xxx–xxx

7

Fig. 10. SEM images of the core region of the cellulose aerogel fibers shown in Fig. 9a–d in higher magnifications: (a, b) the core of the fibers shown in Fig. 9a-b, (c, d) the core of the fibers shown in Fig. 9c-d.

Fig. 11. Typical uniaxial compression stress-strain curves for cellulose aerogel monoliths.

cess in as much as the temperature in the fiber decreases. Since the characteristic time for cooling is proportional to the square of the fiber radius, thicker fibers exhibit a larger macro-pore shell than thin fibers (Fig. 9). Similar observations also made by Ratke and coworkers [2,20]. Fibers with thicknesses below 50 ␮m in diameter do not show such a core to mantle structure [21]. The large pores at the mantle shown in Fig. 9 have pore walls which also consist of nano-fibrillar structures as found in monoliths (see Fig. 10). Longitudinal sections in Fig. 10a and b reveal that the cellulose fibrils are somehow oriented in the extrusion direction or parallel to the fiber axis. In the core part of the fibers a spatially homogeneous porous network is revealed as shown in Fig. 10c and d with a fibril diameters of around 10–25 nm. The mechanical behavior of cellulose aerogel monoliths were determined by compression deformation tests. Typical stressstrain curves of porous materials (Fig. 11) exhibit three regimes:

(I) the linear elastic range followed by a plateau, where pores collapse (II) and densification (III). This behavior is in full agreement with the general statement on the compression behavior of porous bodies as given by common literature [1,36,37]. The compressive modulus was determined from the slope of the linear region of the curves in Fig. 11. Pekala et al. [38] called this value a “compressive modulus” instead of Young’s modulus because the latter connotes a property of a linear-elastic material with low total deformation strains. The compressive modulus varies between 1.4 and 16.2 MPa for cellulose aerogel monoliths with different cellulose contents. The compressive modulus of cellulose aerogels produced in this study is 5–10 times higher than that of aerogels prepared from bacterial cellulose [39], and 3 wt.% cellulose aerogels prepared in our work is almost 2 times higher than that of cellulose aerogel prepared by Demilecamps et al. [32]. For the calculation of the yield strength we take the classical definition in standard textbooks on mechanics [40]. According to that the yield strength is defined as the cross point of the deformation curve with a line drawn parallel to the linear-elastic line at 0.2% irreversible deformation. Fig. 12 shows this yield strength as function of cellulose content. In the literature [36,41] on aerogels Young’s modulus scales with the density in form of a simple power law E ∼ n , with a power n = 1.8 – 3.6 for different aerogels like silica, carbon and resorcinolformaldehyde. The 0.2% offset compressive yield strength behaves in the same way with a power law  ∼ m and m = 2.3 – 3.1 [38,42,43]. Our findings for those values are n = 1.9 and m = 2.2, respectively and are thus at the lower end of the observations for aerogels. Taking a closer look into the scaling models of Gibson and Ashby [36] one notices, that the pre-factor of these power laws is always the strength of the bulk material. This is in our case that of the cellulose fibrils. Currently the strength of the nano-fibrils inside the aerogels is unknown. If it depends on the density itself, the scaling law exponent would change. Also Poisson’s ratio of cellulose aerogels was investigated. Poisson’s ratio is a measure of the change in lateral dimensions on elastically loading a material. The

Please cite this article in press as: I. Karadagli, et al., Production of porous cellulose aerogel fibers by an extrusion process, J. Supercrit. Fluids (2015), http://dx.doi.org/10.1016/j.supflu.2015.06.011

G Model SUPFLU-3359; No. of Pages 10 8

ARTICLE IN PRESS I. Karadagli et al. / J. of Supercritical Fluids xxx (2015) xxx–xxx

Fig. 12. Compressive Modulus (a) and Compressive yield strength at 0.2% offset (b) as a function of cellulose wt.% content for cellulose aerogel monoliths.

cellulose aerogels exhibit a Poisson ratio of zero. This is in good agreement with Sescousse et al. [3] and Liebner et al. [1] who also reported a zero Poisson ratio for cellulosic aerogels. The mechanical behavior of fiber or filament form of materials cannot be determined by compression tests. It was investigated by using standard tensile test measurements. Tensile force versus maximum elongation (%) values was reported for each test. Typical tensile test curves of 3 wt.% cellulose aerogel fiber are given in Fig. 13. The behavior of the filaments in tensile is similar to compressive loading: a linear regime is followed by an irreversible plastic deformation until final fracture occurs (the densification regime does naturally not exist). The curves in Fig. 13 show that there is a large spread of data. The maximum force at fracture varies from 42 to 62 cN and the maximum elongation from around 7–11%. The spread of the filament diameter is not sufficient to explain these strong deviations. We think that defects in the filaments, observable in the microscope, like uneven, non-smooth surfaces, kinks and slight misalignments of filament segments favor the spread in the data. Averaging data from many tests for different cellulose contents leads to the values given in Table 3. These data show that with increasing cellulose content the maximum elongation increases as well as the filament strength, which is in the range of a few MPa. In order to describe the filaments strength a scaling law similar to the yield strength can be used. Assuming in accordance with Gibson

Table 3 Properties of selected cellulose aerogel fibers for extrusion parameters: 95 ◦ C, 200 rpm and Ø 0.5 mm. Sample

A

B

C

Cellulose content (wt.%) Max. Elongation (%) Max. Tensile force (cN) Fiber strength (MPa)a SSABET (m2 /g)

3 9.8 ± 1.01 55.0 ± 4.94 4.51 ± 0.41 123 ± 0.94

4 10.1 ± 2.17 75.8 ± 15.84 5.06 ± 0.89 111 ± 0.71

5 11.3 ± 2.61 62.4 ± 9.02 6.42 ± 0.93 119 ± 0.58

a

Calculated by using max. tensile force obtained from tensile test.

and Ashby, a relation  m =  m0 (1 − )3/2 we can estimate the fibril strength  m0 . We obtain a value of approx. 650 MPa for the tensile strength of a fibril. Independent of its accuracy and in spite of the remarks made above on a possible density dependence of the fibril yield strength, the value itself shows that the cellulose nano-fibrils in an aerogel are rather strong. The value itself is not astonishing, since we have a fibrous material with a porosity of 93–99% and a macroscopic tensile strength of a few MPa. This easily means that the fibrils must have strength of at least 15 times higher the value of the filaments. This strong fibril strength opens further applications for cellulose aerogel fibers. The absolute strengths of monofilament fibers are still weak compared to commercial technical fibers, but as mentioned with respect to their huge porosity they are rather strong. One should notice here: these fibers are the first open porous fibers produced. All technical fibers are either compact massive fibers or are hollow ones. The fibrous morphology and all well-known techniques of fibre processing into woven and non-woven materials are now in principle possible. Further research along this line is in progress. 4. Summary and conclusions

Fig. 13. Typical tensile force vs maximum elongation curves for 3 wt.% cellulose aerogel fiber.

Porous cellulose aerogels in monolithic and fiber form were prepared successfully by dissolving cellulose in Ca(SCN)2 salt hydrate melt followed by gelation, coagulation in ethanol, and drying in supercritical CO2 . In this work we could examine higher cellulose concentrations of up to 6 wt.%. The fiber properties were compared with cellulose aerogel monoliths. The resulting aerogels had porous 3D structures composed of nanometer sized cellulose fibrils arranged in a network. The envelope density of cellulose aerogels obtained from 0.5 to 6 wt.% cellulose content varied between 0.009 and 0.137 g/cm3 , respectively, and cellulose aerogel fibers also had envelope densities in this range. The BET specific surface area of the supercritically dried aerogels are found to be

Please cite this article in press as: I. Karadagli, et al., Production of porous cellulose aerogel fibers by an extrusion process, J. Supercrit. Fluids (2015), http://dx.doi.org/10.1016/j.supflu.2015.06.011

G Model SUPFLU-3359; No. of Pages 10

ARTICLE IN PRESS I. Karadagli et al. / J. of Supercritical Fluids xxx (2015) xxx–xxx

ranged 120–230 m2 /g. Cellulose aerogels show promising mechanical strength and thermal conductivity at atmospheric pressure and room temperature. The compressive modulus of 1.4–16.2 MPa indicates a high mechanical strength. All samples exhibited an elastic-plastic behavior and a Poisson ratio close to zero. The strength of fiber aerogel samples was high (4.5–6.4 MPa) with respect to their porosity. Producing monolithic cellulose aerogels is time-consuming regarding coagulation and washing steps until the gel is salt-free. First systematic series of extrusion experiments have been carried out using twin screw extrusion for aerogel production within this study. We found that extrusion can be carried out successfully in ranges of 95–110 ◦ C. Extrusion spinning was practical way for understand the behavior and thermal stability of cellulose in Ca(SCN)2 salt hydrate melt as a spin dope. With respect to processing washing and coagulation of filaments and fibers is advantageous compared to monoliths, since their small cross-sectional area allows faster washing and drying. SEM images of monoliths show two hierarchical pore structures with pore sizes ranging from 10 to 100 nm and fibril diameters of around 10–25 nm. The morphology of cellulose aerogels prepared as fibers made by micro-extrusion process is similar to the monolith form. The only difference of the fibers is their mantlecore structure, which originates from the evaporation of the coagulation liquid and its permeation into the fiber surface while the hot salt melt solutions is submerged into the regeneration bath. There is a linear correlation between extrusion temperature and the surface area of the supercritically dried cellulose fibers. The special morphology of the fibers with a meso and macro porous structure built upon a nano-porous network opens new fields of applications. One drawback of the cellulose aerogel fibers is their low tensile strength compared to compact technical fibers, but our preliminary test showed that using another type of salt hydrate melts such as ZnCl2 and different methods of filament production can solve the problems [21,44–46]. Acknowledgements This work is supported by Deutsche Forschungsgemeinschaft (DFG), performed by ITA RWTH Aachen and DLR Cologne, Germany (Project ID: RA 537/13-1). References [1] L. Falk, A. Nikita, S. Christian, P. Antje, R. Thomas, Bacterial cellulose aerogels: from lightweight dietary food to functional materials, in: functional materials from renewable sources, Am. Chem. Soc. (2012) 57–74. [2] L. Ratke, Monoliths and fibrous cellulose aerogels, in: M.A. Aegerter, N. Leventis, M.M. Koebel (Eds.), Aerogels Handbook, Springer New York, 2011, pp. 173–190. [3] R. Sescousse, R. Gavillon, T. Budtova, Aerocellulose from cellulose-ionic liquid solutions: preparation, properties and comparison with cellulose-NaOH and cellulose-NMMO routes, Carbohydr. Polym. 83 (2011) 1766–1774. [4] R. Gavillon, T. Budtova, Aerocellulose: new highly porous cellulose prepared from cellulose-NaOH aqueous solutions, Biomacromolecules 9 (2008) 269–277. [5] R. Sescousse, R. Gavillon, T. Budtova, Wet and dry highly porous cellulose beads from cellulose-NaOH-water solutions: influence of the preparation conditions on beads shape and encapsulation of inorganic particles, J. Mater. Sci. 46 (2011) 759–765. [6] J. Cai, S. Kimura, M. Wada, S. Kuga, L. Zhang, Cellulose aerogels from aqueous alkali hydroxide-urea solution, Chemsuschem 1 (2008) 149–154. [7] M.C.V. Nagel, A. Koschella, K. Voiges, P. Mischnick, T. Heinze, Homogeneous methylation of wood pulp cellulose dissolved in LiOH/urea/H2O, Eur. Polym. J. 46 (2010) 1726–1735. [8] J. Cai, Y. Liu, L. Zhang, Dilute solution properties of cellulose in LiOH/urea aqueous system, J. Polym. Sci. B: Polym. Phys. 44 (2006) 3093–3101. [9] M. Hattori, Y. Shimaya, M. Saito, Solubility and dissolved cellulose in aqueous calcium- and sodium-thiocyanate solution, Polym. J. 30 (1998) 49–55. [10] S. Fischer, H. Leipner, K. Thummler, E. Brendler, J. Peters, Inorganic molten salts as solvents for cellulose, Cellulose 10 (2003) 227–236. [11] H. Jin, Y. Nishiyama, M. Wada, S. Kuga, Nanofibrillar cellulose aerogels, Colloid Surf. A 240 (2004) 63–67.

9

[12] S. Hoepfner, L. Ratke, B. Milow, Synthesis and characterisation of nanofibrillar cellulose aerogels, Cellulose 15 (2008) 121–129. [13] J. Innerlohinger, H.K. Weber, G. Kraft, Aerocellulose: aerogels and aerogel-like materials made from cellulose, Macromol. Symp. 244 (2006) 126–135. [14] F. Liebner, A. Potthast, T. Rosenau, E. Haimer, M. Wendland, Cellulose aerogels: highly porous, ultra-lightweight materials, Holzforschung 62 (2008) 129–135. [15] F. Liebner, E. Haimer, A. Potthast, D. Loidl, S. Tschegg, M.A. Neouze, M. Wendland, T. Rosenau, Cellulosic aerogels as ultra-lightweight materials. Part 2: synthesis and properties, Holzforschung 63 (2009) 3–11. [16] M. Deng, Q. Zhou, A. Du, J. van Kasteren, Y. Wang, Preparation of nanoporous cellulose foams from cellulose-ionic liquid solutions, Mater. Lett. 63 (2009) 1851–1854. [17] O. Aaltonen, O. Jauhiainen, The preparation of lignocellulosic aerogels from ionic liquid solutions, Carbohydr. Polym. 75 (2009) 125–129. [18] Z. Liu, H. Wang, Z.X. Li, X.M. Lu, X.P. Zhang, S.J. Zhang, K.B. Zhou, Characterization of the regenerated cellulose films in ionic liquids and rheological properties of the solutions, Mater. Chem. Phys. 128 (2011) 220–227. [19] C.B. Tan, B.M. Fung, J.K. Newman, C. Vu, Organic aerogels with very high impact strength, Adv. Mater. 13 (2001) 644–646. [20] C. Hacker, T. Gries, C. Popescu, L. Ratke, Solution spinning process for highly porous, nanostructured cellulose fibers, Chem. Fibers Int. 59 (2009) 85–87. [21] B. Schulz, M. Schestakow, I. Karadagli, B. Milow, G. Seide, T. Gries, L. Ratke, High strength cellulose aerogel filaments, in: XII International Conference on Nanostructured Materials (NANO 2014), Section: Polymer, Organic and Other Soft Matter Materials, Moscow, Russia, July 13–18, 2014. [22] G. Xplore Instruments, The Netherlands, Micro compounders, Micro compounder 15ml, in, 2015. [23] F. Fischer, A. Rigacci, R. Pirard, S. Berthon-Fabry, P. Achard, Cellulose-based aerogels, Polymer 47 (2006) 7636–7645. [24] J. Cai, S.L. Liu, J. Feng, S. Kimura, M. Wada, S. Kuga, L.N. Zhang, Cellulose-Silica nanocomposite aerogels by In Situ formation of silica in cellulose gel, Angew. Chem. Int. Ed. 51 (2012) 2076–2079. [25] F. Liebner, E. Haimer, M. Wendland, M.A. Neouze, K. Schlufter, P. Miethe, T. Heinze, A. Potthast, T. Rosenau, Aerogels from unaltered bacterial cellulose: application of scCO(2) drying for the preparation of shaped, ultra-lightweight cellulosic aerogels, Macromol. Biosci. 10 (2010) 349–352. [26] Z.G. Wang, S.L. Liu, Y. Matsumoto, S. Kuga, Cellulose gel and aerogel from LiCl/DMSO solution, Cellulose 19 (2012) 393–399. [27] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T. Siemieniewska, Reporting physisorption data for gas solid systems with special reference to the determination of surface-area and porosity (Recommendations 1984), Pure Appl. Chem. 57 (1985) 603–619. [28] G. Reichenauer, Structural characterization of aerogels, in: M.A. Aegerter, N. Leventis, M.M. Koebel (Eds.), Aerogels Handbook, Springer, New York, 2011, pp. 449–498. [29] M. Wiener, G. Reichenauer, S. Braxmeier, F. Hemberger, H.P. Ebert, Carbon aerogel-based high-temperature thermal insulation, Int. J. Thermophys. 30 (2009) 1372–1385. [30] E.M. Hendriks, M.H. Ernst, R.M. Ziff, Coagulation equations with gelation, J. Stat. Phys. 31 (1983) 519–563. [31] R.L. Drake, A general mathematical survey of the coagulation equation, Topics in Current Aerosol Research (Part 2), (1972) 201-376. [32] A. Demilecamps, C. Beauger, C. Hildenbrand, A. Rigacci, T. Budtova, Cellulose–silica aerogels, Carbohydr. Polym. 122 (2015) 293–300. [33] M. Faessel, C. Delisee, F. Bos, P. Castera, 3D Modelling of random cellulosic fibrous networks based on X-ray tomography and image analysis, Compos. Sci. Technol. 65 (2005) 1931–1940. [34] H.-P. Ebert, Thermal properties of aerogels, in: M.A. Aegerter, N. Leventis, M.M. Koebel (Eds.), Aerogels Handbook, Springer, New York, 2011, pp. 537–564. [35] M. Koebel, A. Rigacci, P. Achard, Aerogel-based thermal superinsulation: an overview, J. Sol–Gel Sci. Technol. 63 (2012) 315–339. [36] L.J. Gibson, M.F. Ashby, The mechanics of foams: basic results, in: Cellular Solids, Cambridge University Press, 1997, pp. 175–234. [37] H. Lu, H. Luo, N. Leventis, Mechanical characterization of aerogels, in: M.A. Aegerter, N. Leventis, M.M. Koebel (Eds.), Aerogels Handbook, Springer, New York, 2011, pp. 499–535. [38] R.W. Pekala, C.T. Alviso, J.D. LeMay, Organic aerogels: microstructural dependence of mechanical properties in compression, J. Non-Cryst. Solids 125 (1990) 67–75. [39] N. Pircher, S. Veigel, N. Aigner, J.M. Nedelec, T. Rosenau, F. Liebner, Reinforcement of bacterial cellulose aerogels with biocompatible polymers, Carbohydr. Polym. 111 (2014) 505–513. [40] G.E. Dieter, Part 1: Mechanical fundamentals, in: Mechanical Metallurgy, McGraw-Hill, New York, 1986, pp. 3–100. [41] H.S. Ma, A.P. Roberts, J.H. Prevost, R. Jullien, G.W. Scherer, Mechanical structure-property relationship of aerogels, J. Non-Cryst. Solids 277 (2000) 127–141. [42] G.W. Scherer, J. Gross, R.W. Pekala, C.T. Alviso, Elastic properties of crosslinked Resorcinol-Formaldehyde gels and aerogels, J. Non-Cryst. Solids 211 (1997) 132–142. [43] J. Yang, S. Li, L. Yan, J. Liu, F. Wang, Compressive behaviors and morphological changes of resorcinol–formaldehyde aerogel at high strain rates, Micropor. Mesopor. Mater. 133 (2010) 134–140. [44] M. Schestakow, I. Karadagli, L. Ratke, Influence of regeneration fluids on structure and properties of cellulose aerogels, in: XII International Conference

Please cite this article in press as: I. Karadagli, et al., Production of porous cellulose aerogel fibers by an extrusion process, J. Supercrit. Fluids (2015), http://dx.doi.org/10.1016/j.supflu.2015.06.011

G Model SUPFLU-3359; No. of Pages 10 10

ARTICLE IN PRESS I. Karadagli et al. / J. of Supercritical Fluids xxx (2015) xxx–xxx

on Nanostructured Materials (NANO 2014), Section: Polymer, Organic and Other Soft Matter Materials, Moscow, Russia, July 13–18, 2014. [45] M. Schestakow, B. Schulz, I. Karadagli, L. Ratke, Production of cellulose aerogel fibres from industrial pulp, in: 11th Seminar on Aerogels, Properties-Manufacture-Applications, International Society for Advancement of Supercritical Fluids (ISASF) and Hamburg University of Technology (TUHH), Hamburg, Germany, 2014, p. 22.

[46] I. Karadagli, M., Schestakow, B., Schulz, L. Ratke, Production of porous cellulose aerogel fibres by an extrusion process, in: Cellular Materials - CellMat 2014, Deutsche Gesellschaft für Materialkunde e.V., DGM, Dresden, Germany., 2014.

Please cite this article in press as: I. Karadagli, et al., Production of porous cellulose aerogel fibers by an extrusion process, J. Supercrit. Fluids (2015), http://dx.doi.org/10.1016/j.supflu.2015.06.011