Journal of Materials Science

, Volume 47, Issue 3, pp 1187–1195 | Cite as

Moisture sorption and plasticization of bloodmeal-based thermoplastics

Materials in New Zealand

Abstract

Sorption characteristics, thermo-mechanical and mechanical properties of bloodmeal-based thermoplastics have been investigated between water activities (aw) of 0.2 and 0.8, using water and tri-ethylene glycol (TEG) as plasticizers. Three different mass ratios of TEG to water were used, 1:1, 1:2 and 5:6 with a total plasticizer content of 60 parts per hundred parts bloodmeal. It was found that the equilibrium moisture content and mechanical properties were highly dependent on relative humidity suggesting that material properties may vary during use. The BET and Flory–Huggins equations gave the best fit for desorption and adsorption, respectively, but a significant difference was observed between adsorption and desorption behaviour below a water activity of 0.6, which was thought to be due to changes in intermolecular interactions. The monolayer adsorption capacity (0.05 g/g) was unaffected by the TEG content, using the BET sorption isotherm. The water activity required to form a monolayer (awl) was also independent of the amount of TEG, but was different for adsorption and desorption (about 0.5 and 0.2, respectively). Increasing TEG did not have a strong influence on the equilibrium moisture content, especially at low water activity. Dynamic mechanical analysis revealed that the glass transition temperature decreased almost linearly with increasing water activity, ranging between 3 and 85 °C, however, above a water activity of 0.6 a second transition was observed, most likely due to phase separation. Depending of TEG content, tensile strength increased from about 10 to 15 MPa at a water activity of 0.4, where after a drastic decrease was observed. A similar trend was observed for elongation at break. At low water activity (below 0.4) elongation was less than 3%, increasing between 30 and 50% at higher water activities. It was concluded that 10–15 wt% represented a critical point above which mechanical properties becomes very sensitive to the relative humidity of the environment.

Introduction

Over the past 20 years emerging environmental and economic concerns have driven research into sustainable and biodegradable alternatives away from petroleum-based plastics. A number of different protein sources such as wheat gluten, corn, sun flower, keratin, casein, soy, gelatine and whey have been successfully processed into biopolymers [1]. More recently, it has been shown that bloodmeal can also be converted into a thermoplastic under certain conditions [2, 3], but very little is still understood regarding its behaviour, especially regarding plasticization and water sorption properties.

The mechanical properties of protein-based plastics are largely associated with the distribution and concentration of inter- and intra-molecular forces. Plasticizers can reduce intermolecular interactions between polymer chains and increase product flexibility. The desired effect is to decrease the glass transition (Tg) with a minimal decrease in modulus or tensile strength [4]. Hydrogen bonding, van der Waals forces, hydrophobic interactions and ionic bonding are altered upon addition of plasticizers, leading to altered thermal and mechanical properties [5, 6, 7]. Hydrophilic hydroxyl groups are thought to be the active sites for plasticizers, creating hydrogen bonding between the polymer–plasticizer–water or plasticizer–polymer, interfering with protein–protein interactions and allowing chain mobility [8, 9].

Water is also a good plasticizer for proteins thereby directly influencing mechanical properties typically by increasing elongation and reducing strength [1, 5, 10, 11, 12, 13, 14, 15, 16, 17], making it important to know water sorption behaviour before practical application. Water diffusion increases in the rubbery state as water and plasticizers reduce the amount of protein–protein interactions, creating more free volume between chains [5, 9, 17, 18, 19, 20]. This accounts for the rapid increase in moisture content at high humidity [19].

It has also been shown that increasing temperature or the addition of plasticizers (in addition to water) affects equilibrium moisture content. Increasing plasticizer content causes an increase in equilibrium moisture content, whilst temperature has the opposite effect [8, 19, 21, 22]. Plasticizers suitable for proteins are often hydrophilic, leading to increased water binding, leading to higher equilibrium moisture contents at any given water activity [6, 8, 17, 18, 19, 20, 23]. The requirement of using a plasticizer in protein-based thermoplastics therefore leads to varying amounts of moisture in the product, which in turns influences mechanical properties.

The objective of this article was to assess moisture adsorption and desorption of bloodmeal-based thermoplastics (BMT) using different amounts of plasticizers. In addition, the influence of equilibrium moisture content and plasticizer content on mechanical and thermo-mechanical properties will be assessed.

Water adsorption models

Water sorption of biopolymers is characterised by absorption, adsorption and desorption. Absorption refers the condensation of water molecules into capillaries or pores of the absorbent. Adsorption is the formation of a monolayer or multilayer on the materials surface [24]. Desorption is the loss of water; this can occur when materials are initially processed with large amounts of water. Type II isotherms are usually observed for protein-based films [5]. Type II isotherms represent the formation of a monolayer followed by unlimited multilayer condensation and are typical for nonporous or macroporous adsorbents.

There are a number of isotherms that have been used for biopolymer sorption characterisation: BET, Guggenheim, Anderson, and de Boer (GAB), Caurie, Halsey, Smith, Oswin, Bradley, Harkins–Jura, Peleg, Iglesias, Henderson, Darcy Watt, Flory–Huggins, Ferro-Fontan, Park and modified GAB [5, 8, 17, 18, 21, 22, 23, 24, 25, 26].

The sorption theory of non-porous solids, such as proteins, is based on partial chemisorption, surface impurities or phase changes [27]. Iglesias and Chirife recognised that it is not possible to give a single relationship for water sorption over the entire water activity range in foods, due to their complex nature [27]. The same problem exists with protein-based polymers. Even with the large number of mathematical models proposed to fit hydrophilic materials, none give accurate results through all water activities.

Some of the more common models used in protein-based plastics are the GAB, BET, Flory–Huggins and Oswin models. Guggenheim, Anderson and de Boer simultaneously developed the semi-theoretical three parameter GAB equation, which is an extension of the BET equation. It can be applied over a wide range of water activities and is accepted as the most versatile model in literature. The GAB equation is a three parameter semi-theoretical, multi-molecular, localized and homogeneous sorption model [27]. The parameter Xm represents the moisture content to form a monolayer (g/g) and C and K are related to sorption enthalpies.
$$ {\text{BET}}:\;X_{\text{eq}} = \frac{{X_{\text{m}} Ca_{\text{w}} }}{{\left( {1 - a_{\text{w}} } \right)\left( {1 + \left( {C - 1} \right)a_{\text{w}} } \right)}} $$
(1)
$$ {\text{GAB}}:\;X_{\text{eq}} = \frac{{X_{\text{m}} CKa_{\text{w}} }}{{\left( {1 - Ka_{\text{w}} } \right)\left( {1 - Ka_{\text{w}} - Ka_{\text{w}} CKa_{\text{w}} } \right)}} $$
(2)
$$ {\text{Flory}}{-}{\text{Huggins}}:\;X_{\text{eq}} = A{ \exp }\left( {Ba_{\text{w}} } \right) $$
(3)
$$ {\text{Oswin}}:\;X_{\text{eq}} = A\left( {\frac{{a_{\text{w}} }}{{1 - a_{\text{w}} }}} \right)^{B} $$
(4)
The BET sorption equation provides an estimation of the monolayer moisture content; it assumes that the rate of adsorption of the first later is equal to the evaporation of the second, the binding energy of all the adsorbate is equal and the binding energy of the other layers is equal to that of pure adsorbate. The constant C can be calculated using Eq. 5, where awl is the water activity necessary to form a monolayer [5, 27]
$$ C = \left( {\frac{1}{{a_{\text{wl}} }} - 1} \right)^{2} . $$
(5)

The Flory–Huggins equation is commonly used to model sorption data of hydrophilic materials and is based on the Flory–Huggins interaction parameter, describing the polymer/solvent affinity for one another [28, 29]. The Owsin equation is a fully empirical model fitted to the data and is widely used in moisture sorption studies concerning biopolymers and meat products; it is based on a series of sigmoid expansions to form a mathematical equation representing the typical shape of sorption [25, 27].

Materials and methods

Materials

Bloodmeal was obtained in powder form from Wallace Corporation, Hamilton, New Zealand and sieved to an average particle size of 700 μm. Technical grade sodium dodecyl sulphate (SDS) was obtained from Biolab NZ, analytical grade sodium sulphite from BDH Lab supplies and agricultural grade urea from Balance Agri-nutrients (NZ).

Method

BMT has been developed earlier and has been patented by Novatein Ltd., New Zealand [3]. Thermoplastic protein was prepared by blending 100 parts by mass bloodmeal with 3 parts SDS, 3 parts sodium sulphite and 10 parts urea dissolved in water. Samples were prepared by dissolving all additives in the appropriate amount of water followed by blending with bloodmeal powder in a high speed mixer after which the required amount of plasticizer was added. The mixtures were stored over night before extrusion.

The total amount of plasticizer (water plus triethylene glycol) was kept constant at 60 parts per 100 parts bloodmeal (pphBM). Three different mass ratios of TEG to water were used, 1:1 (30 parts TEG:30 parts water), 1:2 (25 parts TEG:35 parts water) and 5:6 (20 parts TEG:40 parts water).

Extrusion was performed using a ThermoPrism TSE-16-TC twin-screw extruder at a screw speed of 150 rpm using a temperature profile and screw configuration shown in Fig. 1. Actual melt temperatures were within 2–5 °C of the set temperatures. The extruder had a screw diameter of 16 mm, an L/D ratio of 25 and was fitted with a single 10 mm circular die. A relative torque of 50–60% of the maximum allowed in the extruder was maintained (12 Nm per screw maximum), by adjusting the mass flow rate of the feed. The extruder was fed by an oscillating trough and the extruded material was granulated using a tri-blade granulator from Castin Machinery Manufacturer Ltd., New Zealand.
Fig. 1

Extruder screw configuration and corresponding temperature profile

Test specimens were produced using a 22 mm screw diameter BOY 15 S Injection Moulding Machine. Specimens were injected through a cold runner into a water heated mould. The shape of the tensile test specimens was in accordance with ASTM D638. A temperature profile of 70 (feed zone), 115 and 120 °C (die zone) was used employing 1200 bar injection pressure and 400 bar back pressure at screw speed of 150 min−1. A 20-s cooling time was allowed in a mould locked with 30 kN locking force. Samples were injection moulded directly after extrusion and granulation, without further conditioning.

Sorption isotherms

Eight saturated salt solutions were prepared and placed into separate air tight containers at 18 °C to yield the required relative humidity, as outlined in Table 1.
Table 1

Saturated salt solutions used

Salt

 

Relative humidity (%)

Lithium chloride

LiCl

11.1–12.6

Potassium acetate

CH3COOK

23.1 ± 0.3

Magnesium chloride

MgCl2

33.1 ± 0.2

Potassium carbonate

K2CO3

43.2 ± 0.4

Sodium bromide

NaBr

59.1 ± 0.5

Potassium iodide

KI

69.9 ± 0.3

Sodium chloride

NaCl

75.5 ± 0.2

Potassium chloride

KCl

85.1 ± 0.3

Relative humidity was measured using a Lutron HT-3005 hygrometer. Samples for adsorption were pre-dried at 70 °C for 3 days before being placed in humidity chambers. Desorption samples were placed directly into chambers after injection moulding. Moisture content was monitored gravimetrically over time; it was assumed that when three consecutive masses were observed equilibrium had been reached. Samples were left for 30–37 days to equilibrate. Final moisture content was determined gravimetrically by oven drying at 103 °C for 24 h. Tests were conducted in triplicate and the averages taken.

Mechanical properties

Tensile strength (TS), elongation (E) and modulus of elasticity (EM) of each specimen have been determined according to ASTM standard D638-03. Samples were injection moulded into a standard dog bone shape, 12 mm wide, 3 mm thick with a 150-mm gauge length. After conditioning, tensile properties were determined using an Instron model 33R4204. An extension rate of 5 mm/min and an extensometer gauge length of 50 mm were used for testing. Samples were tested in replicas of six directly after removal from the humidity chambers.

Dynamic mechanical analysis (DMA)

DMA was used to investigate effect of moisture content and storage relative humidity on the glass transition temperature. The experiments were carried out on a PerkinElmer DMA8000 instrument. Rectangular samples were tested in a single cantilever setup, using a displacement of 0.02 mm. The sample size was 30 mm long, 7 mm wide and 4 mm thick. For sub-zero cooling, liquid nitrogen was administered by the PerkinElmer cooling system. A temperature scan was done at three frequencies; 1, 10 and 30 Hz, over a temperature range at least 30 °C below and 20 °C above the Tg according to the ASTM D4064-01 standard. Tg was evaluated at a frequency of 1 Hz and taken to be the highest tan δ point at the peak as outlined in the ASTM standard (E1640-04).

Results and discussion

Sorption isotherms

Moisture sorption isotherms were constructed for each different TEG:water ratio. A characteristic sigmoid shape, typical of amorphous, hydrophilic polymers was obtained [17]. As expected, each experimental isotherm (Fig. 2) showed that equilibrium moisture content increased with water activity. Samples that were not dried before testing (raw samples) gained moisture above water activities of aw = 0.6, i.e., adsorption occurred rather than desorption. This is simply due to the initial moisture content being below the equilibrium moisture content at those water activities. At low water activities a gradual increase in moisture content was observed, followed by a rapid increase above aw = 0.6. This is likely caused by exposure of more polar hydroxyl groups as molecular mobility and free volume increase with increasing moisture content. The relationship to molecular mobility is further explored in terms of the glass transition later in the paper.
Fig. 2

Experimental data for adsorption (open symbols) and desorption (filled symbols). Filled square 30 parts TEG:40 parts water; filled triangle 25 TEG:35 Water; filled circle 20 TEG:40 water. Horizontal lines represent initial moisture content for each formulation

TEG content did not significantly influence equilibrium moisture content; the only noticeable difference was observed above aw = 0.6 (Fig. 2). Only raw samples (i.e., not dried) showed a slightly higher equilibrium moisture content with increasing TEG. One would normally expect that increasing hydrophilic plasticizer would lead to an increase in equilibrium moisture content [8]; however, this was not the case for BMT.

This difference is further evident from the fact that one would expect equilibrium moisture content for dried and raw samples to be equal over the entire range of water activity. However, desorption and adsorption were significantly different below aw = 0.6 (Fig. 2). This could be as a result of drying at 70 °C; it is well known that heating proteins effects their conformation and structure. Complete depletion of water and breaking of hydrogen bonds between water and protein side chains could lead to the formation of more hydrogen and hydrophobic interactions between protein chains. Proteins chains would then be bound more tightly, as the number of protein–protein interactions increase, leading to higher water resistivity and thus lower moisture uptake. If no change in protein–protein interactions occurred, then in theory, equilibrium moisture content should be equal regardless if it was measured from adsorption or desorption. Once the material is properly plasticised (at aw > 0.6), the previous argument is no longer valid and a less pronounced difference is observed between adsorption and desorption.

The Flory–Huggins, GAB, BET and Owsin models have been fitted to equilibrium data. Each model has been statistically analyzed to determine the goodness of fit by calculating the root mean square error (RMSE) and the coefficient of determination (R2). Error function values and constants for each model are shown in Table 2. All models showed a very good fit, as indicated by low RSME and high R2 values. All four models tested fitted desorption data better than adsorption, suggesting that none of these models can accurately account for processes occurring during protein rehydration, or adsorption. Selecting the most appropriate model almost becomes arbitrary, although statistically, the Flory–Huggins and BET equations gave the best fits to the adsorption and desorption processes, respectively. To avoid comparison based purely on empirical constants, the BET isotherm was selected for further comparison as it allows estimation of the monolayer adsorption capacity as well the water activity to form a monolayer.
Table 2

Model constants and error values for different models used

 

Xm

C

K

A

B

awl

RMSE

R2

GAB

 Adsorption

  30 TEG:30 water

0.090

0.882

0.817

   

0.021

0.975

  25 TEG:35 water

0.096

0.871

0.807

   

0.023

0.970

  20 TEG:40 water

0.066

0.911

0.855

   

0.020

0.976

 Desorption

  30 TEG:30 water

0.015

156.5

0.073

   

0.015

0.984

  25 TEG:35 water

0.013

156.5

0.075

   

0.016

0.982

  20 TEG:40 water

0.013

156.5

0.073

   

0.017

0.980

BET

 Adsorption

  30 TEG:30 water

0.055

1.290

   

0.468

0.025

0.956

  25 TEG:35 water

0.054

1.396

   

0.458

0.027

0.947

  20 TEG:40 water

0.057

0.853

   

0.520

0.021

0.967

 Desorption

  30 TEG:30 water

0.059

12.37

   

0.221

0.016

0.979

  25 TEG:35 water

0.059

9.181

   

0.248

0.011

0.989

  20 TEG:40 water

0.053

21.96

   

0.176

0.013

0.986

FH

 Adsorption

  30 TEG:30 water

   

0.005

4.882

 

0.014

0.986

  25 TEG:35 water

   

0.005

4.762

 

0.016

0.982

  20 TEG:40 water

   

0.003

5.444

 

0.012

0.989

 Desorption

  30 TEG:30 water

   

0.018

3.554

 

0.015

0.982

  25 TEG:35 water

   

0.015

3.754

 

0.018

0.971

  20 TEG:40 water

   

0.017

3.491

 

0.017

0.976

Osw

 Adsorption

  30 TEG:30 water

   

0.065

0.917

 

0.024

0.959

  25 TEG:35 water

   

0.067

0.898

 

0.026

0.952

  20 TEG:40 water

   

0.054

1.011

 

0.022

0.967

 Desorption

  30 TEG:30 water

   

0.113

0.700

 

0.013

0.986

  25 TEG:35 water

   

0.106

0.736

 

0.013

0.987

  20 TEG:40 water

   

0.103

0.690

 

0.013

0.986

In Fig. 3, the sorption data fitted using the BET model is shown for the three compositions tested. The monolayer capacity (Xm) was independent of the amount of TEG added as well as between adsorption and desorption. The water activity required to form a monolayer (awl) was about 0.5 for adsorption and 0.2 for desorption. This difference highlights the earlier observed difference between adsorption and desorption. At a water activity of about 0.5, there seems to be enough water present in the material to alter protein–protein and protein–water interactions. These observations are further supported by observations regarding mechanical properties.
Fig. 3

Sorption isotherms for different ratios of TEG:water, fitted using the BET isotherm

Thermo-mechanical properties

Figure 4 shows the DMA results at 1 Hz for each composition at various water activities. At water activities above about 0.6, two peaks can be seen in the tan δ curves. The second less-defined peak could be caused by loss of moisture due to heating or the presence of other protein, such as blood serum albumin (BSA). Mo and Sun [30] studied urea-modified soy protein isolates and identified multiple peaks corresponding to endothermic transitions of 7S and 11S globulin proteins. Alternatively, it could also be indicative of some phase separation, leading to plasticizer rich phases. However, at low water activities, only one peak was observed indicating a single phase or complete compatibility. It was thought that a fair amount of chain mobility is required for sufficient phase separation to occur, and TEG alone was not enough to facilitate this process. The point above which phase separation started also corresponds to the point where significant differences were observed between adsorption and desorption, as discussed earlier.
Fig. 4

DMA scans at different water activities and plasticizer content

The glass transition temperatures for each composition are shown in Fig. 5. Increasing water content lead to a decrease in Tg and decreased almost linearly with moisture content up to about 15 wt% (Fig. 5a). Above 15 wt% one of the observed Tgs showed almost no dependence on water content, likely caused by saturation of the protein chains from water adsorption; however, the Tgs corresponding to the second peak continued the linear trend. In Fig. 5b, the glass transition temperature is shown as a function of water activity in which an almost linear decrease in Tg is observed with an increase in water activity up to aw = 0.6. At this point, a sharp transition is observed where the Tg drops markedly. The relative magnitude of the second Tg peak (lower temperature peak) increased with increasing moisture content (or water activity) indicating that phase separation is more severe at higher water content.
Fig. 5

Glass transition temperatures (Tg) for materials containing different amounts of TEG. (aTg vs. moisture content; bTg vs. water activity)

It was thought that the rapid drop in Tg (reflected by the second Tg peak) above 15 wt% moisture is due to a transition from a glassy to a more rubbery material. This transition is brought about by the plasticization effect of water, reducing protein–protein interactions and allowing molecular motion. The material exposes more hydrophilic groups due to increased free volume allowing more water uptake, but only up to 15%. It is interesting to note that this point corresponds to the water activity at which the difference between adsorption and desorption becomes less pronounced (Fig. 2) as well as the water activity required to form a monolayer (awl) according to the BET model. It can therefore be concluded that a moisture content of about 15% represents a critical point in BMT, below which very little plasticization is observed and protein–protein interactions are dominant.

From Fig. 5a, one can observe that the amount of TEG had almost no influence on the Tg. However, Fig. 5b would suggest a small decrease in Tg with increasing TEG. The decrease in Tg with increase in TEG content is expected and in agreement with others who have investigated the effect of polyol addition [4]. Hydrophilic plasticizers such as TEG reduce hydrogen bonding between protein chains by binding to hydrophilic amino acid side groups, disrupting protein–protein interactions. However, the transition at 15 wt% moisture was unaffected by TEG content.

Mechanical properties

Mechanical properties have been assessed for samples conditioned to equilibrium at different water activities. Injection-moulded samples were used without pre-drying, i.e., corresponding to the desorption curves shown earlier. The effect of moisture content and water activity on tensile strength (TS), elastic modulus and elongation at break (EB) are shown in Fig. 6.
Fig. 6

Mechanical properties of materials containing different quantities of TEG as a function of moisture content and water activity

TS decreased with increasing moisture content caused by the plasticizing effect of water. At moisture contents above about 10 wt% there was little further change in TS, probably due to the loss of chain interactions. This is in agreement with others who have studied the effect of moisture content on mechanical properties [5, 9, 18]. When considering the TS as a function of water activity, a slight increase in TS was evident at low water activity (most prominently when using 25 and 30 parts TEG) before a rapid decrease occurred. This would only indicate that tensile strength is not sensitive to moisture below about 10 wt%.

Above a water activity of about 0.4, slight differences in TS between various compositions can be observed. As expected, the decrease in TS was as a result of an increase in TEG. The decrease in TS is most probably due to a reduction of chain interaction due to side chains interacting with TEG. Many researchers have reported the same with increasing hydrophilic plasticizer content [9].

Modulus of elasticity decreased as moisture content or water activity increased (Fig. 6), as would most polymeric materials with an increase in plasticizer. Below a water activity of about 0.6 samples with higher TEG contents also had a lower modulus, although the standard deviation of the data were very large and makes a definitive conclusion difficult. Above the critical water activity, all the materials tested were much more ductile and the scatter in the data was also less pronounced.

From Fig. 6, it can be seen that elongation hardly changed below a water activity of about 0.4, but then changed significantly, reaching an optimum at aw = 0.6. This suggests a water content or storage water activity of about 15 wt% and 0.6, respectively, would correspond to a water content required for a transition from a glassy to rubbery state brought about by increased plasticization due to absorbed water. The decrease in elongation at higher moisture content and water activity is likely due to the protein chains being held so loosely that failure occurs before any significant elongation can occur. It was noted that at a storage water activity of 0.85 (not shown on this graph) the plastic was saturated with water and wet to touch. This is an indication that excessive plasticization and loss of chain interactions leads to premature material failure.

Conclusions

Bloodmeal-based bioplastics have been shown to be very susceptible to environmental water activity or relative humidity. A large amount of water is adsorbed and desorbed depending on storage conditions. This has a significant impact on material properties. The implication is that the material cannot safely be implemented in environments where humidity is constantly changing.

Significant differences between adsorption and desorption was observed below a water activity of 0.6 and it was concluded that this was most likely due to changes in intermolecular interactions. The BET equation was used to model equilibrium moisture content and it was found that the monolayer adsorption capacity was unaffected by the TEG content. The water activity required to form a monolayer (awl) was also independent of the amount of TEG, but was different for adsorption and desorption.

The glass transition temperatures also showed a drastic change around a moisture content of 15 wt%. A near linear relationship was observed below this point, with almost no dependence on moisture content above 15 wt%. The difference was attributed to the material becoming more ductile at the moisture content associated with a water activity of 0.6. This was further supported by observations regarding the mechanical properties of the material. Tensile strength and elongation at break both showed a significant change at a water activity of approximately 0.6.

One can conclude that the application of BMT would be very sensitive to atmospheric conditions, a common problem for biopolymers based on biomass. However, TEG did not exacerbate moisture sorption, which is often the case with biopolymers. The results also suggested that conditioning the material to around the critical point (±15 wt% water) would make it least sensitive to small changes in relative humidity.

Notes

Acknowledgements

The authors would like to thank Wallace Corporation for their support in supplying bloodmeal.

References

  1. 1.
    Verbeek CJR, van den Berg LE (2009) Macromol Mater Eng 295(1):10CrossRefGoogle Scholar
  2. 2.
    Verbeek C, van den Berg LE (2011) J Polym Environ 19:1. doi:10.1007/s10924-010-0232-x CrossRefGoogle Scholar
  3. 3.
    Verbeek CJR, Viljoen C, Pickering KL, van den Berg LE (2009) NZ Patent NZ551531Google Scholar
  4. 4.
    Vanin FM, Sobral PJA, Menegalli FC, Carvalho RA, Habitante AMQB (2005) Food Hydrocolloid 19(5):899CrossRefGoogle Scholar
  5. 5.
    Zhang Y, Han J (2008) J Food Sci 73(7):E313. doi:10.1111/j.1750-3841.2008.00867.x CrossRefGoogle Scholar
  6. 6.
    Su J-F, Huang Z, Zhao Y-H, Yuan X-Y, Wang X-Y, Li M (2009) Ind Crops Prod 31(2):266CrossRefGoogle Scholar
  7. 7.
    Hernandez-Izquierdo VM, Krochta JM (2008) J Food Sci 73(2):30CrossRefGoogle Scholar
  8. 8.
    Maria Martelli S, Moore G, Silva Paes S, Gandolfo C, Laurindo JB (2006) Lebensm Wiss Technol 39(3):292CrossRefGoogle Scholar
  9. 9.
    Cho SY, Rhee C (2002) Lebensm Wiss Technol 35(2):151CrossRefGoogle Scholar
  10. 10.
    Verbeek CJR, van den Berg LE (2009) Recent Pat Mater Sci 2(3):171CrossRefGoogle Scholar
  11. 11.
    Sharma S, Hodges JN, Luzinov I (2008) J Appl Polym Sci 110(1):459. doi:10.1002/app.28601 CrossRefGoogle Scholar
  12. 12.
    Swain SN, Rao KK, Nayak PL (2004) J Appl Polym Sci 93(6):2590CrossRefGoogle Scholar
  13. 13.
    Zhang J, Mungara P, Jane J (2001) Polymer 42(6):2569CrossRefGoogle Scholar
  14. 14.
    Pommet M, Redl A, Guilbert S, Morel M-H (2005) J Cereal Sci 42(1):81CrossRefGoogle Scholar
  15. 15.
    Ortiz MER, San Martin-Martinez E, Padilla LPM (2008) Starch-Starke 60(10):577. doi:10.1002/star.200800212 CrossRefGoogle Scholar
  16. 16.
    Hochstetter A, Talja RA, Helén HJ, Hyvönen L, Jouppila K (2006) Lebensm Wiss Technol 39(8):893CrossRefGoogle Scholar
  17. 17.
    Kristo E, Biliaderis CG (2006) Food Hydrocolloid 20(7):1057CrossRefGoogle Scholar
  18. 18.
    Mali S, Sakanaka LS, Yamashita F, Grossmann MVE (2005) Carbohydr Polym 60(3):283CrossRefGoogle Scholar
  19. 19.
    Perdomo J, Cova A, Sandoval AJ, García L, Laredo E, Müller AJ (2009) Carbohydr Polym 76(2):305CrossRefGoogle Scholar
  20. 20.
    Fabra MJ, Talens P, Chiralt A (2009) Food Hydrocolloid 24(4):384CrossRefGoogle Scholar
  21. 21.
    Cassini AS, Marczak LDF, Noreña CPZ (2006) J Food Eng 77(1):194CrossRefGoogle Scholar
  22. 22.
    Brett B, Figueroa M, Sandoval A, Barreiro J, Müller A (2009) Food Biophys 4(3):151. doi:10.1007/s11483-009-9112-0 CrossRefGoogle Scholar
  23. 23.
    Jangchud A, Chinnan MS (1999) Lebensm Wiss Technol 32(2):89CrossRefGoogle Scholar
  24. 24.
    Srinivasa PC, Ramesh MN, Kumar KR, Tharanathan RN (2003) Carbohydr Polym 53(4):431CrossRefGoogle Scholar
  25. 25.
    Alix S, Philippe E, Bessadok A, Lebrun L, Morvan C, Marais S (2009) Bioresour Technol 100(20):4742CrossRefGoogle Scholar
  26. 26.
    Sopade PA, Ajisegiri ESA, Chukwu O, Abass AB (2010) J Food Process Eng 33:385CrossRefGoogle Scholar
  27. 27.
    Al-Muhtaseb AH, McMinn WAM, Magee TRA (2002) Food Bioprod Process 80(2):118CrossRefGoogle Scholar
  28. 28.
    Kim SD, Chakravarti S, Tian J, Bell P (2010) Polymer 51(10):2199CrossRefGoogle Scholar
  29. 29.
    Jonquières A, Perrin L, Arnold S, Lochon P (1998) J Membr Sci 150(1):125CrossRefGoogle Scholar
  30. 30.
    Mo XQ, Sun XZ (2001) J Am Oil Chem Soc 78(8):867CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC 2011

Authors and Affiliations

  1. 1.University of WaikatoHamiltonNew Zealand

Personalised recommendations