1. Introduction
Sugar beet is one of the most productive crops of the world. According to the data of the Food and Agriculture Organization of the United Nations, the harvested amount adds up to approximately 308 million tons in 2019, of which about half comes from Europe and one-third from the former Russian Federation [
1]. The water insoluble residue after extraction of the sugar, the sugar beet pulp (SBP), consists mainly of approximately equal parts of cellulose, hemicellulose and pectins (in sum 75–85%) and smaller amounts of lignin (<9%), proteins (<7%), lipids (<2%), saponins (<2%) and ash. Approximately 45–50 kg of sugar beet pulp (dry matter) can be obtained from 1 ton of fresh sugar beet [
2,
3,
4].
In our recent paper [
5] we summarized the literature of SBP containing bio-based and/or biodegradable composites, mainly from the last two decades and reported our work on material characterization of PLA–PBS–SBP composites. According to this, SBP containing composites represent a material with expectable mechanical properties. Due to the different polarities of the polymeric components and the highly polar SBP, compatibilizers have to be used to enable higher moduli and strengths. However, the maximum achievable properties are limited due to the low tensile modulus of approx. 5000 to 7000 MPa [
5,
6] acting in the composites. Particularly in the case of coarse SBP particles, which may consist of cell agglomerates due to the cell structure of the sugar beet plant and therefore have a low inherent strength, the strength of such composites are limited. Even high amounts of compatibilizers cannot further increase their strength. In addition, SBP containing composites often exhibit small elongation at break of only a few percent and impact strengths in the range of 1 to 3 kilojoules per square meter. We therefore concluded that biodegradable and full or partially biobased PLA–PBS–SBP composites are, from a material mechanics point of view, just one of many composites with agricultural residues as filler, and therefore do not represent anything unique to pay special attention to.
However, in contrast to classical natural fibers, SBP has the peculiarity of high content of pectin. SBP, such as thermoplasticized starch, can be processed with plasticizer in an extrusion cooking process to produce a pure SBP plastic material [
2,
7,
8,
9]. The plasticized pectinic phase therein acts as the matrix for the non-soluble components of SBP. From a materials point of view, however, such pure thermoplasticized SBP materials suffer from their high sensitivity to water [
9].
Even though the quality of sugar beet pectins does not approach that of apple and citrus pectins concerning water uptake and gelation [
10,
11], they are produced in low amounts with some niche applications [
12,
13]. Nevertheless, SBP pectins can form gels if the disturbance of the gelling ability, which is caused by the acetylation [
10,
14,
15,
16] of the secondary hydroxyl groups, is forced by deacetylation or chemical crosslinking [
17]. Due to the extraction and purification processes, pure pectins are expensive chemicals. However, even without any chemical conversion, composites with SBP as filler will be highly sensitive to water [
18].
Fully biodegradable composites may play a role in end-of-life options for application in agriculture and horticulture beside the well-known mulch films. Due to their thinness, biodegradation will take place in shorter time periods than for products with dimensions in the millimeter range or beyond. Such thick-walled products often suffer from their low surface to volume ratio and the limited attack of the microbiome to the surface. On the other hand, the possible degradation of polyester by hydrolysis is very slow under environmental conditions due to low water absorption and lack of catalytic support in the bulk material if proton and water concentrations are low [
19]. Composites containing a strongly water-attracting and water-retaining material can therefore support the internal degradation processes. Additionally, they may have a destructive effect through relaxation processes as a result of water absorption. In this article, we report our phenomenological investigation on the water absorption of PLA–PBS–SBP composites.
2. Materials and Methods
2.1. Materials
Polymers: Poly (lactic acid) (Ingeo 3251D, NatureWorks LLC, Minnetonka, MN, USA) was purchased from Resinex Germany GmbH (Zwingenberg, Germany). Poly (butylene succinate) (injection molding grade) was a donation from MCPP Germany GmbH (Düsseldorf, Germany).
Fillers: Chalk was purchased from Omya GmbH (Cologne, Germany). Talc was purchased from Mondo Minerals B.V. (Amsterdam, The Netherlands).
Additives: Maleic anhydride modified PLA (PLA-g-MAH, development type) was a donation of BYK-Chemie GmbH (Schkopau, Germany). All chemicals were used as received.
Ground SBP types were provided by Pfeifer & Langen (Cologne/Elsdorf, Germany) in paper bags. The grinding of the SBP had been accomplished before by Jäckering Mühlen- und Nährmittelwerke GmbH (Hamm, Germany) using an air turbulence mill of type “Ultra-Rotor”. Two ground types were used. A fine type (D50 = 23.4 µm), in which the cell structure is destroyed and only cell wall remnants are present, and a coarse type (D50 = 500/600 µm, two different lots), in which intact cells still exist. Pictures of the cell structure and ground SBP are presented in our previous publication [
5].
We analyzed the water content of the SBP prior to use and noticed that it changed from about 8 to 9 percent after receiving the sample to about 10 to 11 percent before processing depending on humidity conditions. Compound compositions were calculated on a dry mass basis. Due to the natural origin of SBP, deviations in the content of its constituents may occur using SBP from different years and/or different cultivation areas. No analysis of the SBP components was performed.
2.2. Compounding
An intermeshing co-rotating twin-screw extruder ZSK 25 from Coperion (Stuttgart, Germany) with the screw diameter D being 25 mm and the screw length L = 40 D was used for compounding. The polymers were dry blended and fed by a gravimetrical dosing feeder into the hopper. SBP was dry blended with coupling agent and/or fillers and fed by a gravimetrical dosing feeder into a side feeder.
The extruder consists of eight zones, which can be tempered individually. In the area of the second zone, a liquid dosing connection is installed. A side feeder for the addition of further materials (SBP, filler, coupling agents) follows this. This zone is also equipped with an atmospheric degassing system. At the beginning of the seventh zone, volatile components can be removed from the melt with the aid of a vacuum suction port (~200 mbar). A nozzle plate with two 3 mm diameter nozzle holes completed the process. Details of screw design are described in our previous work [
5] (screw design 2).
Zone temperatures of the extruder were set to 60, 170, 170, 170, 170, 160, 160 and 165 °C for the production of SBP composites with PLA/PBS. The measured temperature of the materials at the die was 176 ± 3 °C (1σ). The strands were water cooled, granulated by an SGS 50-E granulator (Reduction Engineering GmbH Korntal-Münchingen, Germany) and dried in a dryer (dry air generator, model: LUXOR 80; drying bins: 15 L, Motan-Colortronic GmbH, Kirchlengern, Germany) at 60 °C for several hours until the humidity was below 0.02% (moisture balance MA 30, Sartorius AG, Göttingen, Germany).
Compounds with PLA, PBS and SBP were produced in different grades. As starting formulations, we selected compounds based on three different PLA:PBS ratios with mineral fillers and replaced the mineral fillers with SBP in three or two steps [
5]. For improving the adhesion of the SBP to the matrix, PLA-g-MAH was used as coupling agent. The compositions and the alphanumeric codes for sample designation are shown in
Table 1:
2.3. Injection Molding
Test specimens were injection molded on a Battenfeld 600 injection molding machine equipped with a standard injection molding tool according to ISO 20753:2017 Type 1A (ISO 527 dog-bone-shaped, 2 nests). Die temperature: 170 °C; shot volume: 36–37 cm3; injection speed: PLA rich blends: 10 cm3 s−1, for PBS rich blends a 4-step profile with higher velocities was used; holding pressure: 700–800 bar depending on composition; holding time 20 s (30 s PBS rich); cooling temperature and time: 30 °C/30 s. The compounds were pre-dried in a dryer (dry air generator, model: LUXOR 80; drying bins: 15 L, Motan GmbH, Germany) at 60 °C for 1.5 h immediately before injection molding.
2.4. Microscopy
Optical microscopy was conducted with a Keyence VHX 6000 System (Neu-Isenburg, Germany) equipped with a VH-Z20T objective (20×–200×) and a VHX-S660E specimen stage.
2.5. Water Absorption
The specimens for water adsorption tests were equilibrated in air with ambient humidity and not dried prior to measuring. The dog-bone test specimens were immersed in distilled water and stored at 23 ± 3 °C in 2 L boxes. About 15–24 specimens, three of each material composition, were put in one box. We removed the test bars individually, immediately freed them from adhering water with a linen-cotton cloth and weighted them on a 0.1 mg balance. Due to the evaporation of molecular films, which cannot be wiped off from the surface, the balance showed a drift which slowed down. We took the first stable value (~10–30 s). Calculation of water absorption was completed according to Equation (1):
with
wt indicating the weight of the specimen after time t and
w0 indicating its weight prior to water immersion. Arithmetic mean values of the relative absorption values of the three specimens were used.
3. Background: Penetrant Sorption in Polymers
Due to better understanding of the sorption curves a short introduction into some fundamentals of sorption into polymers will be given here. The text is based on the review from Bond and Smith [
20] as well on the paper of Mensisteri et al. [
21] and Petropoulus et al. [
22]. The fundamental mathematics of diffusion can be found e.g., in the classical textbook from Cranck [
23].
The description of sorption processes of penetrating substances (penetrants) in polymers can ideally be described by assuming a Fickian diffusion profile [
23]. The mathematical treatment of the underlying second law of Fick (Equation (2)):
provides in the integrated form a method to determine the diffusion coefficient (Equation (3);
Mt,
M∞: mass uptake at time
t or
∞, respectively;
l: half of the thickness of a film or plate of thickness;
D: diffusion coefficient). At the beginning of the sorption process, this type of diffusion is characterized by a dependence of the absorbed mass
Mt on the square root of time (Equation (4), concentration of penetrant at
c(
x = ±
l,
t = 0) =
csurface;
c(
x,
t = 0) =
c0) [
22] and
Mt/M∞ < 0.6 [
23]). Additionally, an independence of the specific absorption from sample thickness follows.
The characteristic diffusion time of a process is defined by [
20,
22] (Equation (5)):
However, water absorption cannot always be adequately described by Fick’s law of diffusion with constant boundary conditions. Ideal Fickian type sorption occurs when the diffusion rate is much faster than the relaxation rate of the polymer. This often occurs with polymers above the glass point. It is generally assumed that deviations from Fick’s behavior correlate with relaxation processes at limited velocity. Absorption processes are therefore divided into three cases [
20,
21]:
Case 1 (diffusion-controlled adsorption, low uptake, in general T < Tg, and/or low activity of penetrant) or Fickian diffusion, in which the transport of the penetrant is essentially a stochastic diffusion process driven by the presence of a chemical potential (e.g., concentration gradient). Such transports generally take place in polymer penetration systems in which the penetration agent has a negligible hygroelastic effect on the polymer, i.e., the rate of diffusion is significantly higher than the rate of relaxation processes. Case I adsorption is characterized by the dependence of the absorption from the square root of the exposure time from the beginning (and the specific absorption is also independent of the sample thickness, Equation (4)). Diffusion and Relaxation are decoupled.
Case 2 (relaxation-controlled adsorption, high uptake, in general
T >
Tg,
T <
Tg and high activity of penetrant) in which the relaxation of the polymer in response to the osmotic pressure of the penetrant is faster than any internal diffusion process. As a result, the penetrant tends to move into the polymer with an unsteady concentration profile; areas of the swollen, saturated polymer are separated clearly from the unswollen dry polymer. The hygroelastic swelling is the result of the reorganization of the macromolecules of the polymer to reduce the osmotic pressure of the penetrant. Case 2 sorption characteristics is reported for polymers below their glass transition temperature include (1) a stepwise discontinuity of the penetration concentration within the polymer from a glassy region to a plasticized swollen region with a relatively high penetration concentration, and (2) an initial linear progress of the discontinuity with time in plate-shaped samples and thus a linear mass increase with time (see also [
24]). The linearity of the mass increase ends with a change of the migration velocity of the swollen polymer-penetration medium front. Diffusion and Relaxation are strongly coupled.
Case 3 non-Fickian or anomalous adsorption as a mixture of cases 1 and 2 in which two sub cases can additionally be distinguished [
20,
21]: 3a, the relaxation controlled adsorption (
Tg >
T moderate uptake, separate processes of water uptake may be clearly be seen; partial decoupling) and 3b, the diffusion controlled relaxation (
Tg <
T, high uptake, no separate processes, ~linear uptake with the square root of time,
t >
t0,
Mt/M∞ > 0.6, see Equation (4)).
Real absorption curves often show characteristics of both limit cases or subcases. The sorption mechanism is controlled by the physico-chemical behavior occurring within the polymer or polymer composite system and may include aspects of dissolution processes, interactions between polymer/composite and penetration agent (e.g., hydrogen bonds), diffusion, relaxation, swelling and stress build-up. Thus, unusual diffusion is not necessarily a separate process from the two limiting cases, but a process controlled to varying degrees by the limiting mechanism (diffusion and relaxation) and influenced by the interaction between the penetrant and the absorbent. External factors such as the morphology of the composite, temperature, external loadings and the penetrant activity also influence the absorption mechanism.
A qualitative number for describing the coupling of diffusion and relaxation is the Deborah number “Deb” (Equation (6)) which is the ratio of the characteristic relaxation time
λ to the characteristic diffusion time of the penetrant molecule [
20,
22]:
The relaxation caused by the penetrant can be described by an exponential approach (Equations (7) and (8)). The actual change in concentration per time of the penetrant is proportional to the difference of finite concentration and actual concentration. The proportional factor is the frequency factor ß, which is the inverse of the characteristic time λ of the relaxation (M∞,R: water uptake due to relaxation on infinite time).
Below the glass transition temperature chain mobility is low and the completion of the penetrant induced local relaxation may be sufficiently slow to exercise a substantial, or even controlling, influence on the course of the sorption process [
22].
If Deb is much greater than 1 (being D/l2 or D high in relation to β), the characteristic time of the relaxation process is much greater than the time associated with the diffusion time and the processes are decoupled. Dependent of the specific times, at first a purely Fickian sorption will be present until a (Fickian type) equilibrium is reached which is then followed by a second sorption process due to relaxation of the polymer. Conversely if Deb is much less than 1 the relaxation process is much faster and diffusion and relaxation act simultaneously and case 2 behavior will be seen (normally if test temperature is higher than glass transition temperature of the polymer). If Deb is around 1 anomalous case 3 diffusion characteristics occur.
The distinction of a coupled diffusion relaxation two-stage sorption, which will be modelled by Equation (9) (Equation (2) extended by an exponential term [
20,
21];
M∞,F,r/M∞ = mass uptake at infinite time belonging to Fickian (F) or the relaxation (R) fraction, resp.):
from a system with a polymer penetrant interaction mechanism [
20] (also named two stage sorption or pseudo-Fickian uptake) may be not obvious at once: This pseudo-Fickian uptake arises from a second diffusion mode covering special sites in the polymer with other binding properties (Equation (10)).
Especially, polar penetrants like water may cover sites resulting from the (apolar) free volume plus special polar sites (polar chain ends, hydrogen bonding capable groups, voids, molecular sized gaps between the phases in composites). The resultant sorption may then be modelled by two Fickian terms with different diffusion coefficients (see [
25,
26] for example).
In general, the superposition of the separated processes in Equations (9) and (10) may be extended to a generalized Equations (11) and (12):
It should be kept in mind that the above listed equations only represent model equations, which may not explain all processes on a molecular level in full detail. Especially, no models for explaining case 2 sorption are given here (see [
24,
27]) and variations of dimension were not taken into account. The mentioned literature provides further insights in this highly complex theme.
On long time scales we noticed an additional effect in the experiments (see
Section 4.1 and
Section 4.2). It is an accelerated water absorption with time, which is particularly pronounced in SBP free composites. This accelerated water absorption is caused by the onset of hydrolysis of the polyesters, described by the following kinetic description (Equation (13)):
[
H2O]
free,a is the free water in the composite with activity a, which is not bound to strong sorption sites (e.g., the pectinic acid groups or terminal acid groups of the polyesters) and which can react with the ester functions. Assuming that this amount of reactive water is constant, also assuming that the amount of ester bonds is nearly constant in the time range in which the composites do not show mechanical failure, and taking in mind that the change in water uptake is proportional to the change in newly generated acid groups by hydrolysis, Equation (13) can be rewritten:
Equation (11) must therefore be extended by a further exponential element and partial weight fraction:
As this exponential element has no finite value, the partial weight fractions (Equation (12)) are dependent from the time (saturation level
M∞) and the partial weight fractions for different compositions of the composites are not comparable directly. Absolute values (see Equation (12):
w∞ Fi,Rj × M∞ = M∞,Fi,Rj) for each characteristic step of adsorption should be chosen. When comparing composites with different amounts of SBP, the absolutes values per percent of SBP give comparable values (see
Section 4.2).
5. Conclusions
Water uptake in composites of SBP and the bio-based polyesters PLA and PBS can be described using a model with one diffusion term and two or three relaxation terms, or two relaxation terms and one exponential term, respectively. The analysis of the time dependent sorption at the beginning shows clearly the presence of a relaxation mode at early stages of the sorption. In the case of the polar 9:1-f/c-X composites, the analysis of early sorption clearly indicates the coupling of diffusion and relaxation processes.
The diffusion coefficients are in the expected range of 10−8 cm−2s−1 and they do not significantly change in the presence of the SBP. Since the first relaxation β1 also occurs in SBP-free composites and generally starts after about 20–30 days, we interpret it as the softening effect of the diffusing water followed by the morphological rearrangement of the polymeric chains on a molecular level (sub-nano-scale to nano-scale) with the inclusion of additional water at then accessible sites. The presence of coarse SBP or larger amounts of fine SBP accelerates this process. With increasing amounts of SPB, the relative proportion of adsorbed water also increases, especially in the less polar PBS-rich composites. We interpret this with the provision of transport routes for water uptake by the SBP. Additionally, some SBP constituents itself rearrange and absorb greater amounts of water due to their high content on polar binding sites, especially the highly polar carboxyl groups in the pectinic moieties.
We have no clear understanding of the second relaxation on a medium time scale (~300–500 days) since we also detected this phenomenon in SBP free, only mineral filled composites. For the SBP containing composites, the assumption of rearrangement of the amorphous OH group-containing domains of the carbohydrates in SBP under water uptake is reasonable, but it does not explain the different relative water uptakes per percent SBP in the fine and coarse SBP type composites. We assume that osmotic processes may play a role here. This is supported by the relative water absorption of the third relaxation of composites with coarse SBP particles, containing whole cells, compared to composites with fine particles containing only fragments of cells. For the SBP-free composites a rearrangement of the morphology in chain dimensions (nano-scale) is assumed. The type of the “third” relaxation in the 4:3/3:4-c-X composites as well as the influence of SBP on the disintegration and biodegradation of biodegradable polyesters will have to be analyzed in more detailed future research.