Article pubs.acs.org/Biomac
Retrostructural Model To Predict Biomass Formulations for Barrier Performance Y. Z. Zhu Ryberg, U. Edlund, and A.-C. Albertsson* Fibre and Polymer Technology, Royal Institute of Technology (KTH), Teknikringen 56-58, SE-100 44 Stockholm, Sweden S Supporting Information *
ABSTRACT: Barrier performance and retrostructural modeling of the macromolecular components demonstrate new design principles for film formulations based on renewable wood hydrolysates. Hardwood hydrolysates, which contain a fair share of lignin coexisting with poly- and oligosaccharides, offer excellent oxygen-barrier performance. A Hansen solubility parameter (HSP) model has been developed to convert the complex hydrolysate structural compositions into relevant matrix oxygenpermeability data allowing a systematic prediction of how the biomass should be formulated to generate an efficient barrier. HSP modeling suggests that the molecular packing ability plays a key role in the barrier performance. The actual size and distribution of free volume holes in the matrices were quantified in the subnanometer scale with Positron annihilation lifetime spectroscopy (PALS) verifying the affinity-driven assembly of macromolecular segments in a densely packed morphology and regulating the diffusion of small permeants through the matrix. The model is general and can be adapted to determine the macromolecular affinities of any hydrolysate biomass based on chemical composition.
■
INTRODUCTION Wood hydrolysates containing a fair share of oligo- and polysaccharides are derived from aqueous process liquors generated in the hydrothermal treatment of wood and they are attractive material resources, especially in the packaging industry, because they are renewable, abundant, and easily processed in aqueous solution. Extensive attempts have been made to extract and isolate pure hemicelluloses, such as xylan,1 from wood hydrolysates, followed by extensive purification2,3 to produce well-defined matrices and for applications in fields such as biofuel,4 food additives,5 and gas barrier films.6,7 However, extensive post-treatments and purification operations © 2012 American Chemical Society
consume money, energy, and time to an extent not acceptable for industries producing high-volume consumer products. To realize the forest as a biorefinery8 and make biorefinery fractions commercially interesting resources, we have to find ways of utilizing them in a nonpure state. We have shown that films based on a softwood-based hydrolysate that is rich in galactoglucomannan provide a better oxygen-barrier performance than analogous films based on Received: May 25, 2012 Revised: July 14, 2012 Published: July 17, 2012 2570
dx.doi.org/10.1021/bm300821d | Biomacromolecules 2012, 13, 2570−2577
Biomacromolecules
■
highly purified galactoglucomannan.9,10 Wood hydrolysates are, nevertheless, complicated and delicate to work with. One of the obstacles to their facile implementation in the scaling-up and commercial production of hydrolysate-based products is the lack of qualitative and quantitative tools to access the structure−property relationships. These tools are necessary for the reproducibility and optimization of the formulations in industrial products from wood hydrolysates necessitated by the structural and compositional complexity of wood hydrolysates. The complexity is related to the variety of hemicellulose structures in different wood types as well as to the variety of wood components, more or less degraded, being released to the process liquor in processes run under different conditions. Furthermore, wood hydrolysates have different compositional profiles and vary with respect to lignin content, molecular weight, and degree of acetylation, branching and so on, depending on the post-treatment methods that may aim at more or less purified products.11 To facilitate formulation and design of a wood-hydrolysate-based industrial product, a comprehensive and straightforward model is urgently needed to translate the compositions of such complicated matrices to performance and to covert this information into data that are related to a desired property. Our hypothesis is that a calculation model based on Hansen Solubility Parameters (HSP) is a viable method to investigate the macromolecular interactions between components in a hydrolysate-based system and relate them to the oxygen-barrier performance. HSP models are well established in industry, for instance, for predicting solubility in pigment, petroleum, and polymer blends.12,13 We have previously shown that HSP may be an adequate method to investigate the complicated hydrolysate matrices.10,11 HSP modeling employs the concept of Hansen’s space, which is a three-dimensional coordinate system with the three Hansen’s solubility parameters on the x, y, and z axes. In Hansen’s space, each component in a mixed system can be quantified and visualized with a specific coordinate. The distance between different coordinates (components) estimates the interactions and affinities between them in the system, and these are in turn related to the properties of the entire system. The ability of a matrix to function as a good barrier is related to the molecular ability to pack densely, and this is in turn related to the strength of the inter- and intramolecular interactions in the matrix. In this way, a formulation recipe of wood hydrolysates giving a good oxygen-barrier performance can be derived, based simply on their chemical compositions and the structural information of any cocomponents. HSP modeling thus becomes a key which opens a door for the industry to utilize wood hydrolysates in a cost-efficient and reproducible way. Our aim is to use HSP parameters to create a retrostructural model to predict the barrier performance of a biomass formulation. The model shall deal with the structural and compositional complicities by quantifying the molecular interactions and compatibilities of the different components of hydrolysate-based films/coatings, and furthermore, to relate and predict the oxygen-barrier properties of formulations based on the model. In addition, Positron annihilation lifetime spectroscopy (PALS)14−16 is used as an analytical tool to detect the actual macromolecular packing, the free volume size, in the hardwood hydrolysate-based coating matrices to support the model predictions.
Article
EXPERIMENTAL SECTION
Materials. Carboxymethyl cellulose sodium salt (CMC) was purchased from Sigma-Aldrich (CAS No. 9004−32−4). The degree of substitution of CMC was 0.6−0.9 and the viscosity 400−1000 mPa·s in a 2% H2O solution at 25 °C. Highly purified xylan (>90% xylose) from birch, herein denoted XylBi, was purchased from SigmaAldrich (CAS No. 9014−63−5, X0502). Polyethylene terephthalate (PET) films with a thickness of 38 μm were kindly provided by Tetra Pak Packaging Solutions AB. Hydrolysates. A hardwood hydrolysate (HWBi) was obtained from birch after alkaline cooking and dewatering in a sodium sulfate pulping process. The birch chips were first steamed at 100 °C for 15 min and then cooked at 160 °C with sodium sulphite cooking chemicals. The liquid phase was then collected under alkaline conditions (pH = 8− 11) during a dewatering process. This liquid phase was filtered through a ceramic membrane with a cutoff of 5 kDa and the retentate was precipitated in methanol. The precipitate obtained was dissolved in water and lyophilized (LyoLab 300) into a beige-yellow dry powder. Another hardwood hydrolysate (HWBiA) was kindly supplied by Södra Cell AB. The hydrolysate was obtained from the process liquor used in the hydrothermal treatment of wood chips from birch (with 90% xylose, being a purified commercially available hemicellulose. Both hardwood hydrolysates HWBi and HWBiA contained a small fraction of lignin, ranging from 7 to 10%. The molecular weights (Mw) and PDIs of the hydrolysates were determined by size exclusion chromatography (SEC). XylBi had the highest molecular weight (∼10300 g/mol) and was essentially pure with >90% xylose residues, indicating that XylBi had a comparatively long backbone with a low degree of branching. HWBi also had a comparatively high Mw (∼9800 g/mol) and was more heterogeneous in terms of repeating unit structure than XylBi, indicating also a more branched structure. HWBiA had a low Mw (∼2100 g/mol) and was also acetylated. Hansen Solubility Parameters (HSP). The aim of the HSP calculations was to quantitatively represent each hydrolysate as a set of coordinates in the Hansen’s space, reflecting the structural information expressed as dispersion energy (δD), dipole−dipole effects (δP), and hydrogen bonding (δH). From the distance between hydrolysates and other cocomponents in the system, the compatibility of that system, that is, the intermolecular affinity can be predicted. A strong intercomponent affinity is directly related to an energetically favored immobilization of the component chain segments in a dense structure. A dense structure characterized by a low free volume provides, in turn, an effective barrier toward the permeation of small molecules like oxygen. Hence, careful calculations of the Table 1. Properties of XylBi, HWBi, and HWBiA origin type carbohydrate composition [% (w/w) of total anhydro sugars]
arabinan rhamnan galactan glucan xylan mannan
Klason lignin [% (w/w)] degree of acetylation Mwd PDId a
XylBi
HWBi
HWBiA
birch purified hardwood hemicellulose n.d.a n.d. n.d. n.d. ≥90b n.d. n.d. 0 10300 3.7
birch hydrolysate 2.48 1.28 9.11 10.55 75.53 1.04 9.2 0 9800 1.6
birch/aspen hydrolysate 4.07 2.79 4.37 5.96 77.59 4.87 7.0 0.26c 2100 3.3
Not determined. bInformation offered by Sigma Aldrich for chemical X0502. cReference 11. dObtained from SEC. 2573
dx.doi.org/10.1021/bm300821d | Biomacromolecules 2012, 13, 2570−2577
Biomacromolecules
Article
Table 2. Hansen Solubility Parameters δD, δP, δH, and δ for Pentose and Hexose Units, HWBi and HWBiAa pure hemicellulose acetylated hemicellulose hardwood hydrolysate other components
pentose unit hexose unit pentose unit (DSAcb = 0.26) hexose unit (DSAcb = 0.26) HWBiA HWBi ligninc CMC
δD (MPa1/2)
δP (MPa1/2)
δH (MPa1/2)
δ (MPa1/2)
((Ri‑lignin)/(Rolignin))
Ri‑lignin
16.9 18.9 16.1 16.3 17.2 19.11 21.9c 18.7
17.8 15.5 15.2 12.7 15.4 18.39 14.1c 13.0
23.1 29.9 25.6 21.2 25.4 25.76 16.9c 24.0
33.7 38.6 33.9 29.6 34.3 37.14 31.0c 33.1
0.9 1.0 1.0 0.9
6.07 6.42 5.87 5.57 4.09 5.73
1.0 0
a
The component interactions with lignin ((Ri‑lignin)/(Rolignin)) and with CMC (Ri‑CMC) are also shown. bInformation obtained from supplier. cFrom Hansen solubility handbook, ref 12.
groups into the matrix leading to an even more compact molecular packing in the system. As can be seen in Table 2, the additional lignin in the hydrolysates shortens the distance of the hydrolysate from CMC from a value of about 6 to about 5. This indicates that the presence of lignin also aids the formation of a compatible system with hydrolysate and CMC. In a sense, this system mimics what Nature has always been creating: wood as a ternary composite of macromolecules. In the wood fiber cell, lignin, hemicelluloses, and cellulose interact to a perfect extent through extensive hydrogen bonding.13 These interactions and the component compatibility result in a compact structure balancing hydrophilicity/hydrophobicity, mechanical performance, and permeability. A film based on either of the hydrolysates HWBi or HWBiA is therefore expected to have a better oxygen-barrier performance than a film based on the pure xylan XylBi. HWBiA was calculated taking into consideration the degree of acetylation of the polysaccharide. It is evident from the HSP calculations, the substitution of acetyl groups improves the compatibility of pentose and hexose units with CMC from 6.07 and 4.42 to 5.87 and 5.57. Thus, a shorter distance of HWBiA to CMC is obtained (4.07). In summary, the calculations predict an increasing ability to provide a good oxygen-barrier coating/film in the order HWBiA > HWBi > XylBi. The presence of lignin leads to an even more compact system with extra interactions stemming from hydrogen bonding. The oxygen-barrier ability of a matrix such as a hydrolysate may also be assessed by comparing its solubility parameters to those of oxygen. According to König and Schuch,18 a polymer tends to be a better barrier for oxygen, the more the solubility parameters differ from the solubility parameters of oxygen. The solubility (S) of oxygen directly influences the permeability (P) through the classic relationship P = S × D, where D is the diffusion coefficient. HSP is widely used to investigate both S and D. The HSP parameters of O2 have been determined to be δH = 14.7, δP = 0, and δD = 0.12 Based on these figures, the distance of hydrolysates and pure xylan-based films to oxygen in the Hansen’s space were calculated to be 29.5 (XylBi), 30.1 (HWBiA), and 32.9 (HWBi). Pure xylan is more closely located to oxygen than the hydrolysates, indicating a higher tendency to solubilize O2 and, according to the correlation mentioned above, anticipated not to perform as well as the hydrolysates as an oxygen barrier. The hydrolysates and xylan were formulated as barrier coatings on PET. Considering the HSP of PET,12 the distance from PET to O2 in the Hansen’s space was quantified as 11.6. PET is considered to be a moderate oxygen barrier, and based on the much larger distances from hydrolysates to O2
than from PET to O2, HSP modeling predicts that the hydrolysates will act as very efficient barriers. Oxygen-Barrier Performance of Films. We produced films based on XylBi, HWBi, and HWBiA with CMC and determined their OPs to verify the predictions of HSP modeling. The films were thin, on the order of 100 μm, and semitransparent. The films based on XylBi and HWBi were offwhite, while HWBiA films were light brown. Coatings were transparent, with a thickness on the order of 10 μm. They adhered well to the surface of the PET films. The OP was determined for all the coatings (Table 3). As expected, all the coatings improved the oxygen-barrier Table 3. Oxygen Transmission Rate (OTR) and Oxygen Permeability (OP) of Coatings on PET Based on XylBi, HWBi, and HWBiA at 50% RH coating on PETa PETo
none
XylBi
highly purified xylan hydrolysate birch hydrolysate birch /aspen
HWBi HWBiA
OTR at 50% RH (cm3 m−2 day−1)
OP at 50% RH (cm3 μm m−2 day−1 kPa−1)
38.0 ± 0.4 41.5 ± 2.1
38.9 ± 5.0
14.6
12.3 ± 1.2
5.0
43.4 ± 1.7 49.0 ± 6.3
8.9 ± 0.9
3.8
5.2 ± 0.9
2.5
thickness (μm)
a
All formulations are based on a CMC amount of 50% (w/w) and aqueous solution with a concentration of 0.4 g/14 mL
performance of PET dramatically. At 50% relative humidity (RH), the OP was reduced from 14.6 to less than 5.0 cm3 μm m−2 day−1 kPa−1. Compared with the definition of a good oxygen barrier (OP < 39 cm3 μm m−2 day−1 kPa−1) for food packaging,23 all coatings showed excellent barrier properties. The OP of HWBiA-based coating was the lowest (2.5 cm3 μm m−2 day−1 kPa−1). This is in agreement with the predictions of the HSP calculations and the hypothesis that lignin works as a “glue” between the hemicellulose chains to improve the compatibility of the components and aiding the formation of a densely packed system. The substitution of acetyl groups in wood hydrolysate HWBiA further improved the compatibility of the system according to the HSP calculation, and coatings based on this hydrolysate had, as expected, a lower OP than HWBi-based coatings. The OP measurements give the trend HWBiA > HWBi > XylBi, which was the same as that given by the HSP calculations. 2574
dx.doi.org/10.1021/bm300821d | Biomacromolecules 2012, 13, 2570−2577
Biomacromolecules
Article
Positron Annihilation Lifetime Spectroscopy (PALS). To further explore the macromolecular packing and to confirm our predictions of compact molecular packing, PALS was applied on the hydrolysate- and xylan-based films based on the coating formulation. During PALS measurements, the ortho-positronium (o-Ps), which is the size of a hydrogen atom (1.06 Å), is applied as a probe to investigate the sizes of the free volumes (holes) in the matrix, an amorphous glassy matrix in this work. The “pick-off” annihilation happens when o-Ps meets an electron on the wall of the free volume (hole), and in this way, the lifetime of o-Ps becomes a measure of the size of the hole. The holes are in the theoretical framework used assumed to be spherical, as in the classic Tao and Eldrup model in the field of PALS.18,20 In reality, voids in a macromolecular structure have irregular geometries that are constantly changing somewhat because main chain bond rotations of the macromolecules make them rearrange. Hence, it is impossible to decide an exact shape for free volume holes. The geometry of xylan has been described as comparable with the Flory definition of a “semi-flexible chain”,24 and that would suggest that the chains adopt a somewhat elongated structure with some main chain sugar units oriented parallel to each other, giving rise to somewhat tubular rather that perfectly spherical holes. However, the holes are imagined as spherical to facilitate the calculations. For a spherical geometry, only one parameter, the radius, has to be determined, while other geometries are associated with three individual length-scale parameters in the three-dimensional space. There are a few studies and models where the classic Tao Eldrup model is modified to account for different shapes of holes, such as cylindrical plates, for designed structure-known polymers.25 The assumptions do not, however, influence the overall trends of the estimated free volume size. Through PALS, the relative intensities I3, lifetime τ3, and distribution σ3 of pick-off annihilation of o-Ps were detected (Table 4). The hole size and hole volumes were calculated according to eqs 1 and 2 (Figure 1).
Figure 1. Distribution n(rh) (probability density function, pdf) of free volume hole radius (filled symbols) and volume (open symbols) of films based on HWBi (blue), HWBiA (red), XylBi (gray), and SW (green). The latter data are reproduced with permission from ref 28. Copyright American Chemical Society. The distributions are normalized to unit area below the curve.
Regarding the free volume hole and volume distribution (Figure 1), HWBi has a wider distribution than XylBi and HWBiA. HWBi has the smallest average hole radius, 1.972 Å, and thus the smallest volume 37.8 Å3. This value is smaller than that of highly purified xylan XylBi, which has an average hole radius of 2.050 Å and a hole volume of 40.8 Å3. Again, this is in direct agreement with the HSP prediction and OP results. As expected from the HSP calculations, lignin plays an important role, interacting with both hemicellulose (pentose units and hexose units) and cellulose (CMC). This leads to a close macromolecular packing and a compact system with strong interactions such as hydrogen bonding and further to a good oxygen-barrier performance, with an OP as low as 3.8 cm3 μm m−2 day−1 kPa−1. HWBiA, on the other hand, shows the largest hole size (2.103 Å) of the three samples in the PALS measurements, although it is supposed to have the smallest hole size according to HSP prediction. This can be explained by the sensitivity of PALS measurements to the molecular weight. Hole size is calculated mostly from the lifetime of pick-off annihilation of o-Ps, τ3, but τ3 is independent of molecular weight in a glassy polymer only when the molecular weight is over a threshold of 103 g/mol (Mw).26 However, in the low molecular weight range, the smaller the Mw, the longer is the lifetime it shows.27 It is noticed that XylBi has a Mw of ∼10300 g/mol and HWBi ∼9800 g/mol, while HWBiA has a considerably lower Mw (∼2100 g/ mol). This explains the highest τ3 obtained and thus the largest hole size calculated for the HWBiA sample. The actual hole size of HWBiA may not be as large as this experiment suggested, especially according to the observation of the relative intensities o-Ps, I3. The generality of this HSP structural modeling concept, as well as its agreement to PALS derived hole sizes, is demonstrated by considering the corresponding modeling concept applied to pure acetylated galactoglucomannan (AcGGM) and an AcGGM-rich softwood hydrolysate biomass (SW) released by the hydrothermal treatment of spruce wood.3,28 HSP calculations show a close interaction between lignin and AcGGM in the spruce wood hydrolysate and suggest
Table 4. ortho-Ps (o-Ps) Lifetime Parameters; the o-Ps Intensity I3, the Mean o-Ps Lifetime τ3, and Standard Deviation of o-Ps Lifetime Distribution σ3 for Films Based on XylBi, HWBi, and HWBiA film samplesa
I3 (±1.0%)
τ3 (±0.04 ns)
σ3 (±0.04 ns)
XylBi HWBi HWBiA
19.6 17.7 18.9
1.302 1.260 1.340
0.345 0.408 0.340
a
All formulations are based on a CMC amount of 50% (w/w) and aqueous solution with a concentration of 0.4 g/14 mL
The results show that all the films have a very short o-Ps lifetime, τ3 (1.30 ns), compared to that of common glassy amorphous polymers (in the range of 1 to 8 ns).14 Polysaccharide films based on softwood hydrolysates have a similar lifetime range (1.20−1.40 ns).10 HWBi had the shortest lifetime among all the studied films (1.260 ns). These short lifetimes indicate that the matrix is characterized by small holes and a dense system, which is consistent with HSP calculations and OP results. The relatively high hydrogen parameters in HSP calculations, which indicate abundant hydroxyl groups in the matrix, contribute to a strong affinity between components, close macromolecular packing, and thus a compact structure with low OP. 2575
dx.doi.org/10.1021/bm300821d | Biomacromolecules 2012, 13, 2570−2577
Biomacromolecules
Article
HSP modeling seems, therefore, to be a powerful tool to predict in a retrostructural fashion how a film formulation based on wood hydrolysates should be composed to offer an efficient oxygen barrier. The industrial community is anxiously awaiting strategies for biomass conversion into commercially valuable products and new tools to achieve the biomass utilization in a cost-efficient way need to be discovered, elaborated, and integrated into existing value chains to increase both sustainability and profit.
that their mutual affinity results in an immobilization of the chains in a dense disordered structure. This explains why the SW-derived film formulation is a better oxygen barrier (1.6 cm3 μm m−2 day−1 kPa−1 at 50% RH) than the analogous film formulation based on highly purified AcGGM (8.7 cm3 μm m−2 day−1 kPa−1 at 50% RH). The OP of the SW based films is lower than the OPs of analogous films based on XylBi and HWBiA, and HWBi in agreement with PALS data showing that the SW film28 is characterized by a smaller hole size (Figure 1). The model also shows a stronger compatibility between the spruce wood hydrolysate and CMC than between purified AcGGM and CMC, again in agreement with the actual oxygen permeability performance. It is noteworthy that these results show that the less refined biomass, with a more heterogeneous composition, packs more densely than the corresponding highly purified counterpart, and leads to better barriers, an insight that may have important implications for the economic feasibility and commercial interest of these matrices. The direct correlations found between a smaller distance in the Hansen’s space, a smaller average free volume hole size, and a lower OP of the resulting films give useful information on how to compose a hemicellulose or hydrolysate mixture to provide an efficient oxygen-barrier film. For instance, the HSP modeling accurately predicts that CMC would be a better cocomponent to AcGGM and AcGGM-rich softwood hydrolysates than chitosan if a film formulation with low OP were targeted. The relative intensities I3 describe the interaction of o-PS wave overlaps with the electrons in the matrix materials. The interaction is dependent on the electron environment of o-Ps in the hole. Either a small hole or a dense hole wall (i.e., a higher molecular weight) gives rise to the observed relative intensity I3. Even with its low molecular weight, HWBiA gives an intensity higher than HWBi, and this indicates a small hole size in the matrix of HWBiA. If we gather together the predictions of interactions and compatibilities by HSP calculations, actual oxygen permeability, and the free volume obtained by PALS regarding the films and coatings based on XylBi, HWBi, and HWBiA, a clear and very good agreement is evident (Figure 2).
■
CONCLUSIONS
■
ASSOCIATED CONTENT
A retrostructural model, based on Hansen solubility parameter (HSP) calculations, was developed that allows us to choose among available biomasses and to predict feasible industrial formulation recipes for desired packaging purposes, simply by knowing the structural composition of the biomass and of any cocomponents. The HSP calculation method allowed for a simplification of the polysaccharides present in liquors generated in hydrothermal treatment of wood and revealed strong interactions, not only between hemicellulose and lignin in these hydrolysates, but also between hydrolysates and a cocomponent, carboxymethyl cellulose (CMC). All these strong inter- and intramolecular interactions led to a stronger compatibility between hydrolysates with CMC and drive the formation of a densely packed matrix that finally leads to a lower oxygen permeability. A series of free-standing thin films and coatings was prepared from different hardwood hydrolysates from birch (HWBi), birch/aspen (HWBiA), and a pure hemicellulose (xylan from birch (XylBi)) to demonstrate the viability of the model. Positron annihilation lifetime spectroscopy was used to quantify the free volume hole sizes in the matrices of the films and this confirmed the predicted trend of HSP calculations. The model predicts the matrix ability to serve as a good oxygen barrier as increasing in the order HWBiA > HWBi > XylBi, which was in agreement with the results of oxygen permeability analyses.
S Supporting Information *
Details showing how the Hansen solubility parameters (HSP) of hydrolysates HWBi, HWBiA, and xylan XylBi were calculated by a group contribution method, including the consideration of hexose and pentose compositions, lignin contents, and acetylation degree. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors gratefully thank VINNOVA (Project No. 200904311) for financial support. The authors gratefully thank Dr. Margaretha Söderqvist-Lindblad from Södra Innovation for kindly providing the HWBiA hydrolysate and Tetra Pak Packaging solutions for the PET films. The authors also thank Prof. Reinhard Krause-Rehberg from the University of Halle in Germany for PALS technique support.
Figure 2. Quantifications of oxygen permeability, free volume as measured by PALS, predicted distance in Hansen’s space, and molecular weight for XylBi, HWBi, and HWBiA and their corresponding films and coatings. *Indicates that the value is sensitive to the molecular weight of HWBiA. 2576
dx.doi.org/10.1021/bm300821d | Biomacromolecules 2012, 13, 2570−2577
Biomacromolecules
■
Article
REFERENCES
(1) Teleman, A.; Tenkanen, M.; Jacobs, A.; Dahlman, O. Carbohydr. Res. 2002, 337 (4), 373−377. (2) Fundador, N. G. V.; Enomoto-Rogers, Y.; Takemura, A.; Iwata, T. Carbohydr. Polym. 2012, 87 (1), 170−176. (3) Pinto, P. C.; Evtuguin, D. V.; Neto, C. P. Carbohydr. Polym. 2005, 60 (4), 489−497. (4) Steen, E. J.; Kang, Y.; Bokinsky, G.; Hu, Z.; Schirmer, A.; McClure, A.; del Cardayre, S. B.; Keasling, J. D. Nature 2010, 463 (7280), 559−562. (5) Sedlmeyer, F. B. Food Hydrocolloids; American Chemical Society (ACS): Washington, DC, 2011; Vol. 25, pp 1891−1898. (6) Hansen, N. M. L.; Plackett, D. Biomacromolecules 2008, 9 (6), 1493−1505. (7) Goksu, E. I.; Karamanlioglu, M.; Bakir, U.; Yilmaz, L.; Yilmazer, U. J. Agric. Food Chem. 2007, 55 (26), 10685−10691. (8) Ragauskas, A. J.; Williams, C. K.; Davison, B. H.; Britovsek, G.; Cairney, J.; Eckert, C. A.; Frederick, W. J.; Hallett, J. P.; Leak, D. J.; Liotta, C. L.; Mielenz, J. R.; Murphy, R.; Templer, R.; Tschaplinski, T. Science 2006, 311 (5760), 484−489. (9) Edlund, U.; Ryberg, Y. Z.; Albertsson, A.-C. Biomacromolecules 2010, 11 (9), 2532−2538. (10) Zhu Ryberg, Y. Z.; Edlund, U.; Albertsson, A. C. Biomacromolecules 2011, 12 (4), 1355−1362. (11) Ibn Yaich, A.; Edlund, U.; Albertsson, A.-C. Biomacromolecules 2011, 13 (2), 466−473. (12) Hansen, C. M. Hansen Solubility Parameters: A User’s Handbook, 2nd ed.; CRC Press: Boca Raton, FL, 2007. (13) Hansen, C. M.; Bjorkman, A. Holzforschung 1998, 52 (4), 335− 344. (14) Dlubek, G.; Kilburn, D.; Bondarenko, V.; Pionteck, J.; KrauseRehberg, R.; Alam, M. A. Macromol. Symp. 2004, 210, 11−20. (15) Dlubek, G.; Pionteck, J.; Bondarenko, V.; Pompe, G.; Taesler, C.; Petters, K.; Krause-Rehberg, R. Macromolecules 2002, 35, 6313− 6323. (16) Townrow, S.; Kilburn, D.; Alam, A.; Ubbink, J. J. Phys. Chem. B 2007, 111 (44), 12643−12648. (17) ASTM D3985−05 Standard test method for oxygen gas transmission rate through plastic film and sheeting using a coulometric sensor. (18) Tao, S. J. J. Chem. Phys. 1972, 56 (11), 5499−5510. (19) Nakanishi, H.; Wang, S. J.; Jean, Y. C. Positron Annihilation Studies of Fluids; World Scientific: Singapore, 1988; pp 292−298. (20) Eldrup, M.; Lightbody, D.; Sherwood, J. N. Chem. Phys. 1981, 63 (1−2), 51−58. (21) Persson, J.; Dahlman, O.; Albertsson, A.-C.; Edlund, U., Nordic Pulp Paper J. 2012, 27, in press. (22) Henriksson, G.; Lawoko, M.; Martin, M. E. E.; Gellerstedt, G. Holzforschung 2007, 61 (6), 668−674. (23) Krochta, J. M.; De-Mulder, C. L. C. Food Technol. 1997, 51 (2), 61−74. (24) Dervilly-Pinel, G.; Thibault, J.-F.; Saulnier, L. Carbohydr. Res. 2001, 330 (3), 365−372. (25) Jasińska, B.; Kozioł, A.; Goworek, T. J. Radioanal. Nucl. Chem. 1996, 210 (2), 617−623. (26) West, D. H. D.; McBrierty, V. J.; Delaney, C. F. G. Appl. Phys. 1975, 7 (3), 171−174. (27) Yu, Z.; Yahsi, U.; McGervey, J. D.; Jamieson, A. M.; Simha, R. J. Polym. Sci., Part B: Polym. Phys. 1994, 32 (16), 2637−2644. (28) Edlund, U.; Yu, Y.; Zhu Ryberg, Y.; Krause-Rehberg, R.; Albertsson, A.-C. Anal. Chem. 2012, 84 (8), 3676−3681.
2577
dx.doi.org/10.1021/bm300821d | Biomacromolecules 2012, 13, 2570−2577
