Barrier and sorption properties of porous poly(vinyl alcohol)–cellulose fibre composites

Various composites were prepared by impregnation of cellulose fibre nonwovens (CFNs) with poly(vinyl alcohol) using two methods. The composites obtained contained up to 0.4 volume fraction of polymer, and the volume fraction of voids was not less than 0.3. Scanning electron microscope images show no great changes in CFN frame structure compared to untreated CFN. Air and water vapour permeability decreases with growth of the volume fraction of polymer as a result of filling of the voids and formation of closed voids. As determined by sorption characteristics, the composite and its base absorb water vapour differently from CFNs. Biodegradation of samples in soil was inspected. Test data show a faster degradation for samples with a small polymer amount, where less than 10% of initial weight is left after two weeks exposure in soil.


*
Large amounts of plastics are used to package various products.Since packaging is a short-term product, its proportion in waste damps is remarkable.At the same time public concern over dwindling landfill space and accumulation of surface litter increases.Biodegradable materials can notably contribute to reducing the growing waste problem caused by different petrochemical materials [1].
Cellulose-based materials are one of the main raw materials used in packaging.They are widespread, renewable, and with great strength at a relatively low cost [2,3].Cellulose fibre nonwovens (CFNs) as natural polymers are environmentally friendly and biodegradable materials.
Poly(vinyl alcohol) (PVA) is a widely produced water soluble and completely biodegradable synthetic polymer.It is easy to process and has good mechanical characteristics.Thus PVA is an excellent constituent for green composites.Degradation of PVA is quite slow and largely depends upon such structural properties as the degree of polymerization and hydrolysis, distribution of hydroxyl groups, stereoregularity and crystallinity [4].However, composites with cellulose materials can accelerate the process of degradation [2,5] and offer new industrial applications in many fields [6].
Many packaging applications demand certain barrier properties as well as mechanical resistance.It is generally recognized that loss or gain of moisture plays a critical role in assessing the quality and shelf life of packaged food.Due to their hydrophilic nature, cellulose fibres tend to absorb water vapour from the environment, especially, under high-humidity conditions, as well as grease and oil.In addition, resistance to grease is also a property which should be improved, and biopolymers can provide it.Moisture sorption induces alterations in mechanical properties of CFNs, which has led to the need for combining cellulose packaging materials with polymers.Cellulose fibre nonwovens are a perspective base for targeted further treatment to create a packaging with a better barrier or other functional properties [3,7,8].
Barrier properties of plastic packaging play a great role in designing suitable materials.The vapour transmission rate has significance in considering the selection of barrier materials for hygroscopic foods.The gas transmission rate helps in material selection for, e.g., oxygen-sensitive foods, and for some products resistance to grease is important [1].
The main objective of this work was to study the barrier and biodegradation characteristics of CFN-PVA composites and to assess correlation of these characteristics with their microstructural features.Mechanical properties of these composites were investigated in [9][10][11] in particular.The development of nature-friendly biodegradable packaging materials is a topical problem.Conjunction of such widespread materials as cellulose and PVA can be both practical and economically advantageous.

MATERIALS AND METHODS
On the basis of previous studies [12] three dissimilar CFNs with different voids content (Table 1) were chosen for impregnation with PVA water solutions.Partly crystalline water-soluble PVA was used (molecular weight 133 kDa, density 1.28 g/cm 3 , content of acetate groups 0.98 wt%) and PVA solutions with various polymer concentrations (7-15%) were prepared.
The composites were prepared by two different methods: the applicator method (AM) and the calender method (CM).In accordance with the AM, polymer solution was laid on the nonwoven by casting and subsequent drawing of the applicator along the sample.In case of the CM, the impregnated sample was squeezed once between elastomer-coated calender rolls.
The amount of polymer in the composite was controlled by use of solutions with different polymer contents as well as by applying different gap heights to both the applicator and calender rolls.The prepared samples were stored at 17 °C till complete drying.
The volume fraction of voids in a CFN, ϕ V M , was estimated as follows: where The amount of polymer in the obtained composites was calculated by the formula where ϕ POL C -volume fraction of polymer in the composite; V POL C -volume of polymer in the composite, The volume fraction of cellulose fibres in the composite, ϕ CEL C , was evaluated as follows: where V CEL М -volume of cellulose in the CFN.
The volume fraction of voids in the composite, ϕ V C , was determined as follows: where The structure of a CFN and its composite were studied on the Tescan, Mira/LMU Shottky scanning electron microscope (SEM) at the × 600 magnification and voltage of 15 kV.Samples for tests were cooled in liquid nitrogen and fractured by bending.Brittle fracture cross sections were covered with gold (Emitech, K550X) [2].
Air permeability q was estimated as a mean value of five parallel results according to the standard EN ISO 9237 [13] by measuring the air flow (L/min) through a 5 cm 2 area under 2 kPa pressure differences on the air permeability device "SDL Atlas".
Water vapour transmission (WVT) was evaluated by the desiccant method according to the standard ASTM E 96 [14].Test specimens were sealed to a test dish containing anhydrous calcium chloride (desiccant).The sets were stored for four weeks in relative humidity 50 ± 2% at 32 °C and periodically weighed.Water vapour transmission (g/h × m 2 ) was calculated as follows: where G -weight change, g; t -time during which G occurred, h; A -test area (dish mouth), m 2 .Hygroscopic properties of composites were determined following the requirements of the standard ASTM D 3201 [15].Samples were exposed in 97% relative humidity (maintained by 3% calcium chloride solvent) at a temperature of 25 °C for seven days and dried in an oven at 105 °C until constant weight was attained."Apparent" moisture content w (%) prior (6) and after (7) high-humidity exposure was calculated to find out the ratio w A /w P using the formulas where w P -moisture content prior to high-humidity exposure, w A -moisture content after high-humidity exposure, m -sample weight prior to high-humidity exposure, m D ovendry sample weight, m W -sample weight after high-humidity exposure.
Biodegradation of CFNs and their composites was determined by the soil burial test.A sample of 3 × 3 cm size was inserted in the Petri dish with humidified soil to imitate conditions in the environment and exposed to 25 °C temperature.After the defined time periods samples were removed from soil, cleansed, and weighed.
where z -weight loss, %; m 0 -initial sample weight, g; m E -sample weight after exposing in soil, g.

RESULTS AND DISCUSSION
The main components building a CFN are cellulose fibres and voids.The length and width distribution of cellulose fibres was discussed in [10].Schematic depiction of composite structure is shown in Fig. 1.The CFN samples contain a small amount of adhesives and fillers.Composites with the volume fraction of polymer up to 0.4 and the volume fraction of voids up to 0.3 were prepared.The SEM images of the composite do not reveal significant changes in CFN structure in comparison with untreated CFN.Although slight compacting of CFN structure and a thin polymer layer around and among the fibres for all composites can be observed, no remarkable alteration of CFN structure occurred (Fig. 2) [10].
The ternary phase diagram (Fig. 3) shows the relationship between the contents of all three components of the composite, i.e., volume fractions of cellulose, voids, and polymer.To ensure certain permeability of gas and/or moisture, which is essential for potential use of composite as packaging material, our aim was not to reach the volume fraction of voids less than 0. As expected, the air permeability q of composites declines considerably with increase in polymer content (Fig. 4a).Values of air permeability q drop about 2-3 times for composites with a small polymer content (ϕ POL < 0.10), in comparison with untreated CFN (Fig. 4b).The q values decrease by about one order for composites made by the CM technique and even more than 50 times for CFN I samples with a high polymer content (ϕ POL ∼ 0.4) prepared by the AM technique.
A general trend can be noted: while the volume fraction of voids in the composite reduces almost linearly (as a consequence of structural permanence of CFN [10]), air permeability drops sharply (Fig. 5).Even small amounts of polymer cause multiple decay of permeability, most probably due to the formation of closed voids.
Another important characteristic of packaging materials, WVT, was evaluated for the period of four weeks.Figure 6 represents the principle of the WVT process, i.e., steady mass increase in test sets which indicates continuous permeability.The WVT values are specified in Table 2. Since the WVT of PVA is comparatively low -0.15 g/(h m 2 ), the WVT values of composites are considerably lower relating to base material.
The higher ϕ POL value, the lower is transmission ability (Fig. 7).
The amount of absorbed water vapour at the equilibrium state of untreated CFN is in the range 12-18%.Figure 8 indicates the change in the moisture content of the same sample after and before the experiment.Quite a great amount of moisture has absorbed all composites.Composites CFN I and CFN II absorbed less moisture than an untreated nonwoven, while composite CFN III absorbed more.It seems that CFN I and CFN II are more hydrophilic than polymer, and treatment with polymer makes water vapour sorption  more limited, whereas CFN III appears more hydrophobic and PVA accelerates water vapour absorption.
Biodegradation tests confirm that composites with a small polymer content ϕ POL lose their integrity and weight more rapidly than those containing more polymer, even faster than untreated CFNs (Fig. 9).Already after two weeks of experiment less than 10% of sample initial mass was left for all three untreated nonwovens with ϕ POL ≤ 0.6 and no more than 40% for base material and composites with a higher ϕ POL .
Composites with little ϕ POL were completely degraded when four weeks had passed, whereas base material and composites with higher ϕ POL had lost at least 80% of their initial mass.It should be noted that further degradation occurs slowly, particularly disintegration of composites.However, after six weeks no CFN I and CFN III samples were left in the test dish.
The results derived show that changes in structure after incorporation of polymer are of great importance for microorganisms consuming it.As predicted, air and water vapour transmission declines with increase in the volume fraction of polymer, most probably due to filling of voids with polymer and formation of closed voids.
The water vapour sorption character observed suggests that absorption is determined by the hydrophobic/hydrophilic nature of CFNs.
Fast biodegradation of composites takes place in the first two weeks when more than 90% of the initial weight of composite with ϕ POL ≤ 0.06 disintegrates.
After this period degradation continues much slower, except CFNs with a small polymer amount that disintegrate completely.This indicates that the process of degradation depends on the availability of both cellulose and polymer.
The tests performed show that barrier and sorption properties as well as biodegradability of the obtained composites meet the packaging requirements.
thickness of the CFN sample, S -area of the sample (100 cm 2 ); V CEL М -volume of cellulose fibres in the CFN, V CEL М = m М /ρ CEL ; m М -weight of the CFN sample; ρ CEL -density of cellulose (1.52 g/cm 3 ).
3. Therefore the values of the components contents on the diagram are rather closely spaced.The more porous CFN, the more polymer solution it can accumulate.The volume fraction ϕ POL up to 0.4 in composites can be reached on the basis of the most porous CFN I.The diagram proves that the AM technique allows greater impregnation of nonwovens than the CM technique.

Fig. 4 .
Fig. 4. Air permeability values q C (a) and ratio q M /q C (b) of composites as function of polymer content.

Fig. 5 .
Fig. 5. Air permeability q and volume fraction of voids ϕ V C versus volume fraction of polymer ϕ POL in the composite.

Fig. 6 .
Fig. 6.Dynamics of mass increase during the water vapour transmission test for CFN I composites (figures at the symbols -volume fraction of polymer).

Fig. 8 .
Fig. 8.Comparison of sample moisture content after (w A ) and prior to (w P ) water vapour sorption by the applicator method.

Fig. 9 .
Fig. 9. Course of biodegradation of CFN II and its composites with different polymer contents studied by the calender method (figures at the symbols -volume fraction of polymer).