Journal of Thermal Analysis and Calorimetry

, Volume 112, Issue 3, pp 1303–1315 | Cite as

Identifying transition temperatures in bloodmeal-based thermoplastics using material pocket DMTA

Article

Abstract

Bloodmeal can be used to manufacture thermoplastics, but requires water, urea, sodium sulphite, and sodium dodecyl sulphate to modify chain mobility. Transition temperatures of bloodmeal, modified bloodmeal, and processed bloodmeal-based thermoplastics were compared using material pocket dynamic mechanical thermal analysis. The glass transition temperature (Tg) of bloodmeal dropped from 493 to 263 K using only water as a plasticizer but was restored when freeze dried. Modifying bloodmeal lowered Tg to 193 K. This was raised by drying, but not to that of unmodified bloodmeal indicating a permanent change. Three additional transitions were identified above Tg, for modified bloodmeal between 300 and 480 K. These were thought to be transitions of dehydrated bulk amorphous regions, amorphous regions between crystallites and chains segments in crystallites and were also seen at lower temperatures when replacing some water with tri-ethylene glycol (TEG). Material pockets increased resolution in processed samples. One broad Tg was observed in consolidated bars, at 335 or 350 K with or without TEG. In material pockets, these resolved into three transitions, similar to those observed before processing. Changes in relative magnitudes suggested some chain rearrangement leading to more bulk amorphous regions. Differences were detected between onset of drop in storage modulus and peaks in loss modulus and tan δ in pockets or bars, but generally led to the same conclusions. For bar samples, it was helpful to compare natural and log modulus scales. Good practice would use all these techniques in parallel to correctly identify relaxation temperatures.

Keywords

Relaxation Dynamic mechanical thermal analysis (DMTA) Thermoplastic protein Bioplastics Glass transition 

Introduction

Environmental concerns regarding conventional plastics have led to the development of bio-based alternatives using proteins or polysaccharides. Thermoplastics have been produced from plant and animal proteins, including wheat gluten, soy, sunflower, corn, gelatine, keratin, casein, whey and bloodmeal [1]. Bloodmeal is highly denatured protein that is readily available as a by-product of the meat processing industry. Protein crosslinking induced by heat treatment prevents the formation of an extrudable melt from bloodmeal and water alone [2]. Sodium sulfite (SS), sodium dodecyl sulfate (SDS) and urea are also added. SS disrupts chemical crosslinking (by reducing disulfide bonds) and SDS disrupts hydrophobic interactions. Urea, a protein denaturant, disrupts secondary structures such as beta sheets [3]. The combination of these modifies bloodmeal into a thermoplastic which can be extruded and injection moulded [2, 4, 5]. Extrusion relies on the material having a low enough glass transition temperature (Tg) for large scale chain rearrangements without protein degradation at the processing temperature [1]. The Tg can be modified further by adding additional plasticisers such as polyols.

Dynamic mechanical thermal analysis (DMTA) is a well-known technique for detecting thermal transitions in polymers by measuring the response of a sample to an oscillating force while varying frequency or temperature. For viscoelastic materials, the response is described by the complex modulus (E*) which is made up of two components, the storage modulus (E′) and the loss modulus (E″). The ratio of E″/E′ is called tan δ and used as a measure of mechanical damping. Typically, for an amorphous polymer, E′ drops rapidly by two orders of magnitude, while E″ and tan δ show maxima at the onset of large scale co-operative chain movements seen as the Tg. At lower temperatures, smaller drops in E′ and accompanying peaks or shoulders in E″/tan δ relate to the onset of shorter range chain movements. A key advantage of DMTA is sensitivity to relaxations near or below the glass transition [6] and the ability to identify these additional transitions that are not apparent in DSC or thermo-mechanical analysis (TMA) [7].

A newer application of DMTA is the use of material pockets which allow powders and other non-self-supporting samples to be analysed [8]. The technique relies on the metal pocket behaving elastically within the temperature and frequency range studied and that any tan δ peak observed is due to relaxation in the pocket’s contents [9]. This is a reasonable assumption as the pocket has a modulus several orders of magnitude higher than the contents. Such pockets are now commercially available from Tritech and Perkin Elmer, and have been used to successfully identify glass transitions and study relaxation behaviour of freeze-dried proteins [10], carbohydrates [8, 11], milk powders [12], solutions containing bio-materials [13] and thermosetting resins [14]. The technique has been mostly applied to pharmaceutical or food-based materials, with only a few studies concerned with plastics [9, 14], none of which considering protein-based thermoplastics.

The objective of this paper was to investigate the use of material pocket DMTA to determine transition temperatures in thermoplastic proteins derived from bloodmeal. In particular, the relaxation behaviour of bloodmeal, modified bloodmeal (pre-processed thermoplastic) and the resultant thermoplastic were compared using powdered samples. Analysing bloodmeal and modified bloodmeal required the material pocket as these materials are only available in powder form. As a comparison, solid bars (i.e. injection moulded thermoplastics) of the resultant product were also analysed.

A further objective was assessing the validity of various methods of identifying relaxation temperatures, specifically relating to the material pocket. Different techniques used were the peak in tan δ, onset of drop in storage modulus (E′) or peak in loss modulus (E″). These can typically vary by 10–25 °C [7]. One of the challenges of interpreting DMTA results is the inconsistent use of these identification techniques. In Table 1, a summary is given of the techniques used to identify transition temperatures in studies where the material pocket was used [8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20].
Table 1

Published studies making significant use of the material pocket DMTA technique

Year/citation

Materials analysed

Method of determining Tg from material pocket DMTA

Other techniques used to validate Tg valuesa

2005 [8]

Lactose

Peak in tan δ

MDSC

2005 [15]

Celocoxib

Peak in tan δ

DSC

2007 [14]

Benzoxazine monomers

Peak in tan δ

DEA

2008[11]

Xanthan powder

Peak in tan δ

PTA DSC/Tg not seen

2009 [10]

Hen egg white lysozyme, bovine pancreatic ribonuclease A, ovalbumin, bovine serum albumin, sucrose, trehalose and hydroxyethyl starch

Peak in tan δ and onset of E′ drop both shown. Onset of drop used for fitting

DSC/HDSC, TMA

2008 [16]

Poly(d,l-lactic acid), neat and blended with cyclodextrin

Peak in tan δ

DSC

2009 [9]

Poly(d,l-lactic acid), neat and blended with cyclodextrin

Peak in tan δ

None

2010 [17]

Ball-milled cellulose

Peak in tan δ

DSC

2010 [18]

Pharmaceuticals and polymer blends

Not stated

DSC

2010 [13]

Human serum albumin, porcine heparin, trehalose, lactose, dextran, influzena antigen, and neat human plasma analysed in solution

Peak in tan δ

MDSC

2010 [6]

Milk powders with varying lactose and protein contents

Onset of E′ drop

DEA

2011 [12]

Milk powders withvarying lactose and protein contents

Peak in E

DEA

2011[19]

Skim milk/maltodextrin systems

Peak in E

DSC

2011 [20]

Phosphate based 2-hydroxyethyl methacrylate hydrogels

Peak in tan δ

DSC

aTechnique abbreviations: DSC differential scanning calorimetry, MDSC modulated DSC, HDSC hyper DSC, DEA dielectric analysis, TMA thermomechanical analysis, PTA phase transition analysis

Several earlier studies have demonstrated that Tg values obtained by material pocket DMTA show agreement with other mechanical and calorimetric techniques (Table 1) and validation using other thermal techniques has not been repeated in this work.

Methods

Materials and sample preparation

Thermoplastic protein was produced from bloodmeal using a patented process [4], with and without tri-ethylene glycol (TEG) as plasticizer. 20 g urea (Ballance Agri-nutrients), 6 g sodium dodecyl sulphate (SDS) (Merck) and 6 g sodium sulphite (SS) (Ajax Finechem) were dissolved in 120 g distilled water at 50 °C. The solution was mixed with 200 g sieved bloodmeal (Wallace Corporation) in a high speed mixer for 10 min. To prepare the material with TEG, the amount of water was reduced to 80 g, and 40 g of TEG (Merck) was added after the first 6 min of blending. The resultant mixtures were stored below 4 °C overnight in sealed plastic bags. Sample descriptions, compositions in parts per hundred bloodmeal (pphBM) and abbreviations are shown in Table 2. WBM was prepared in the same manner as PPM and PPM–TEG, but with 60 pphBM distilled water as the only additive.
Table 2

Sample formulations used and their abbreviations

Sample description

Sample formulation/pphBM

Abbreviation

Moisture contenta

Moisture content/freeze driedb

Bloodmeal

Water

Urea

SS

SDS

TEG

Sieved bloodmeal

100

     

SBM

7.2 %/TG

N/A

Wet bloodmeal

100

60

    

WBM

40 %/TG

3 %/TG

Pre-processed thermoplastic protein without TEG

100

60

10

3

3

0

PPM

35 %/TG

5.5 %/TG

Pre-processed thermoplastic protein with TEG

100

40

10

3

3

20

PPM–TEG

28 %/TG

5.2 %/TG

Moulded thermoplastic protein

PPM extruded, injection moulded and conditioned at 23 °C and 50 % relative humidity.

MTP

7.9 %/oven drying

6.5 %/TG

Moulded thermoplastic protein with TEG

PPM–TEG extruded, injection moulded and conditioned at 23 °C and 50 % relative humidity

MTP–TEG

8.4 %/oven drying

3.1 %/TG

aMethod used to determine moisture content shown in brackets

bMoulded bar samples were ground prior to freeze drying

The mixtures were extruded in a Thermoprism TSE-16-TC twin screw extruder at 150 RPM. The temperature settings were 70, 100, 100, 100, and 120 °C from feed zone to die. Granulated extrudate was injection moulded into Type 1 tensile test specimens [21] using a BOY-35A injection moulder with a temperature profile of 100, 115, 120, 120, 120 °C from feed zone to die and a screw speed of 200 RPM. The injection pressure was set to 110 bar in profile sections 0–4 and 30 bar in sections 5–7. The mould was heated to 70 °C using water. Injection moulded specimens were conditioned in a humidity chamber set to 23 °C and 50 % relative humidity for a minimum of 7 days prior to testing.

All samples were tested at their respective moisture contents after conditioning or as freeze-dried powders (Table 2). Samples were dried over two nights in a Freezone® 2.5 Litre Benchtop freeze dryer (Labconco Corporation, Kansas City) set to auto mode (Collector temperature −50 °C, vacuum <11 Pa).

Dynamic mechanical thermal analysis

Samples were tested in single cantilever mode using a DMA 8000 (Perkin Elmer) fitted with a high temperature furnace and controlled with DMA software version 14306. The instrument was cooled to below 173 K by evaporation of liquid nitrogen after mounting each sample. Evaporated nitrogen was then fed back into the chamber as a purge gas. Experiments were performed at a programmed heating rate of 2 K min−1 up to 523.15 K for powders. For self-supporting bars, the stiffness drops before this and samples were stopped at 393.15 K. Self-supporting samples were cut from larger injection moulded pieces with approximate geometry of 3.5 × 6.5 × 30 mm. Powder samples were analysed by mounting ~50 mg powder in ~1.0 × 7.4 × 28 mm (folded dimensions) material pockets (Perkin Elmer) [8] which were then crimped with pliers. Material pockets and self-supporting samples were tested using a free length of 12.5 mm and a dynamic displacement of 0.05 mm at 0.1, 0.3, 1, 3 10, and 30 Hz.

Multi strain scans of sieved blood meal (SBM) were also performed at 1 Hz with displacements from 0.01 to 0.1 mm confirming that at the displacement of 0.05 mm was within the linear elastic region across the temperatures tested. All samples were weighed before and after testing to determine any mass loss.

Transition temperatures determined at different frequencies were examined for their correlation to the modified Arrhenius equation [6, 8]:
$$ { \ln }\,f = { \ln }\,A - \left[ {\frac{{E_{\text{A}} }}{RT}} \right] ,$$
(1)
where f is the test frequency, A is the pre-exponential factor, EA is the Arrhenius activation energy, R is the universal gas constant and T transition temperature determined from peak in tan δ, onset of drop in E′ or peak in E″ at test frequency (K).

Moisture content

Moisture content of powdered samples was assessed using thermogravimetic analysis (TG) in a Texas Instruments SDT 2960 analyser. Approximately, 10 mg powdered sample was placed in the sample crucible and heated at 10 K/min from room temperature to 873 K under constant air flow. Moisture content was determined from the cumulative mass loss up to 393 K.

The moisture content of injection moulded samples was determined gravimetrically by oven drying for 24 h at 376 K and recording the mass loss before and after drying.

Results and discussion

Thermoplastic processing techniques such as extrusion and injection moulding require a polymer to be in a molten state, or several Kelvin above its softening point. This would imply a temperature above the material’s Tg if it is amorphous or the melting point for crystalline or semi-crystalline materials. To evaluate the additives efficiency at reducing the processing temperature, thermal transitions in bloodmeal were compared to those in modified bloodmeal (pre-extruded thermoplastic protein) as well as to bloodmeal with additional water added.

Bloodmeal and modified bloodmeal

From Fig. 1a, three peaks in tan δ were apparent for bloodmeal (labelled Ti − Tiii from left to right). Although E′ and E″ represent composite moduli of the pocket and the polymer, peaks in E″ and the onset of the drop in E′ were consistent with the tan δ peaks. Ti, at 182.4 K, was thought to be a gamma relaxation, involving side chains or very small portions of main chains.
Fig. 1

Representative 1 Hz DMTA data for: a sieved bloodmeal, b sieved bloodmeal plus 60 pphBM water, c pre-processed material without TEG and d pre-processed material with TEG

The temperature of Tii at 336 K is similar to the denaturation temperature of haemoglobin or bovine serum albumin in solution [22, 23]. However, bloodmeal is already denatured and is a dehydrated protein with only 7 % moisture. It was thought that this was a β-relaxation, relating to onset of similar, but shorter range motions to the Tg. β-relaxations typically obey an Arrhenius relationship over a broad frequency range [24]; however, Tii in bloodmeal did not (Table 3). It was thought that the loss of residual water during testing up to this temperature obscured the frequency dependence of this transition.
Table 3

Transition temperatures identified from peaks in tan δ in bloodmeal and pre-processed material

Sample

Figure label

Temperature/K for each frequency/Hz

Fits to Arrhenius model

0.1

0.3

1

3

10

30

 

EA/kJ/mol

r2

SBM

Ti

n/a

176

182

189

195

197

Yes

59

0.964

 

Tii

338

337

336

336

340

342

No

n/a

n/a

 

Tiii

488

490

493

496

499

504

Yes

754

0.983

WBM

Ti

257

261

263

264

265

265

No

n/a

n/a

 

Tii

304

310

315

315

317

318

No

n/a

n/a

 

Tiii

388

388

388

388

n/a

386

No

n/a

n/a

 

Tiv

494

497

500

502

506

508

Yes

842

1.000

PPM

Ti

203

195

193

192

189

190

No

n/a

n/a

 

Tii

252

255

257

259

259

260

Yes

352

0.905

 

Tiii

397

397

397

392

398

395

No

n/a

n/a

 

Tiv

425

434

443

449

453

457

Yes

279

0.962

 

Tv

479

479

482

486

490

497

Yes

587

0.914

PPM–TEG

Ti

191

190

191

191

189

187

No

n/a

n/a

 

Tii

240

241

241

241

239

237

No

n/a

n/a

 

Tiii

400

400

399

399

414

424

No

n/a

n/a

 

Tiv

497

510

510

509

508

501

No

n/a

n/a

FD WBM

Ti

195

199

205

210

228

227

Yes

56

0.940

 

Tii

346

350

355

361

355

353

No

n/a

n/a

 

Tiii

490

493

496

500

504

508

Yes

688

0.996

1 Hz data is in bold. Transitions were deemed to fit to the Arrhenius model over the frequency range tested if r2 > 0.90

Bloodmeal’s Tiii, at 493 K, was thought to be the Tg, associated with the α-relaxation, indicating the onset of large range chain motions. Good correlation with the Arrhenius equation (r2 > 0.99) indicated a frequency dependence characteristic of long-range chain relaxation (Table 3). However, TG revealed the onset of significant mass loss immediately beyond this temperature (Fig. 2). Degradation would prevent processing at temperatures near this Tg, requiring plasticisers to lower the Tg to a temperature well below 393 K.
Fig. 2

Mass loss versus temperature for SBM determined by TgA

Using 60 pphBM water as plasticiser, a peak in tan δ was found at Ti = 263 K (Fig. 1b). Although the temperature of this transition decreased at lower frequencies, it did not fit an Arrhenius frequency relationship (Table 3). This peak was thought to be a hybrid of the Tg of water plasticised protein, and the melting of any bulk or free water in the sample. Melting of bulk water has previously been shown to exhibit a peak just below 273 K at 1 Hz in material pocket DMTA [11]. In DSC studies of sunflower proteins and extruded soy proteins, no peak for freezable water was seen below about 26 % moisture [25, 26]. Above this moisture content, it appears the Tg is obscured [25], or merges with the peak for water melting [26]. Wet bloodmeal contained about 40 wt% water so this is likely also the case here. The significant drop in storage modulus in the same temperature region further supported the conclusion that this peak was indeed the plasticised Tg.

At higher temperature, Tiv for WBM was thought to be the dry Tg of bloodmeal after evaporation of water during testing. A small minima in tan δ was also seen around 350–360 K with an associated peak in E′ at Tiii = 366 K. This dip gave rise to what appeared to be two peaks in tan δ (Tii and Tiii). In gluten-based plastics, a drop in tan δ and increase in modulus were seen as evidence of denaturation by means of crosslinking [27]. Although bloodmeal is already denatured and crosslinked, this behaviour is reminiscent of cured thermosetting resins, which exhibit additional crosslinking on heating [7]. If this is the case, it would be expected that the storage modulus show little drop at the end of the rubbery plateau, similar to synthetic crosslinked polymers. However, such effects are obscured as the pocket supporting the sample also restricts the drop in modulus.

It has previously been demonstrated that the plasticising effect of water alone was not enough to allow bloodmeal to be processed via extrusion [2]. Although a transition region is seen below 273 K in wet bloodmeal, excessive crosslinking will prevent thermoplastic flow. This required the use of additional additives for processing.

Figure 1c shows the DMTA trace for modified bloodmeal (PPM). As with wet bloodmeal, there is a peak in tan δ, a little below the freezing point of water (Tii = 257 K). There was also an additional peak (Ti) at 193 K detected also as a peak in E″ and a drop in E′. It was thought that Tii represents melting of unbound water, depressed due to colligative interactions with protein chains, with Ti the glass transition of PPM, lowered by the addition of urea, SDS and sodium sulphite. These additives reduce chain interactions and would increase the free volume by increased chain mobility. The behaviour was similar to native BSA, which is one of the components of bloodmeal. The Tg of BSA determined by DSC as been shown to be 193 K at a water content allowing free water to crystallise [28].

At high temperatures, behaviour is more complex. Three peaks were seen above 390 K, by which stage water was expected to have evaporated, compared to only one peak in SBM and WBM. The central peak (Tiv in Fig. 1c) showed very good agreement with the Arrhenius equation over a temperature range of >30 K with the frequencies tested (Table 3). It was thought that this was the main dry amorphous glass transition, after the plasticization effect of urea, SDS and SS. The broadness of this peak and the appearance of the two smaller peaks were attributed to the blend proteins in bloodmeal (mainly bovine serum albumin and haemoglobin), each of which is a complex heteropolymer. In both gluten- and soy protein-based plastics with more than one protein subunit, multiple Tgs have been detected [27, 29, 30]. Furthermore, as well as random coils, proteins fold into ordered regions (α-helices and β-sheets) suggesting a semi-crystalline nature, detectable by wide-angle X-ray scattering (WAXS) [31]. Crystallinity both broadens the glass transition [32] and sometimes causes multiple glass transitions, arising from a purely amorphous phase, amorphous material restrained by crystallites, crystal slippage and hindered rotation of chains inside folded crystals [33]. Determination of which transitions correspond to each of these possible relaxations is beyond the scope of this paper, but it is important to note that behaviour is not characteristic of one simple amorphous phase.

When TEG was included (Fig. 1d), two peaks in tan δ were again observed below 273 K. Ti, at 191 K, was comparable that seen without TEG and again thought to be the Tg of plasticised proteins in the material. Similarly, it was again thought that Tii could correspond to a bulk freezable plasticiser phase, as the same mass of total plasticiser (water plus TEG) was present. In PPM–TEG, however, Tii dropped further to 241 K, 32 K below the freezing point of water alone. TEG is a known freezing point depressor in aqueous solutions [34], so a plasticiser phase consisting of both water and TEG should have a lower transition temperature than a phase of water alone.

While replacing some water with TEG made only small difference to the low temperature relaxation behaviour, it made a profound difference to above ambient properties. The broad transition, previously centred at 428 K without TEG, was not apparent with TEG. It was thought that TEG increased mobility such that the material was in a rubbery state at this temperature, even after some moisture evaporated. Since TEG is present at a relatively low proportion, a dehydrated Tg was still observed at Tiv = 505 K.

Using material pockets has enabled chain relaxation behaviour of both bloodmeal and pre-processed modified bloodmeal to be investigated. Large differences in Tg were seen between SBM (493 K) and WBM (263 K) due to the plasticising effect of water. Both PPM and PPM–TEG showed transitions at ~191 K, believed to be the plasticised Tg after the effect of processing additives. Altering the plasticiser composition to replace some water with TEG does not affect the wet Tg, but reduces the dry Tg such that it is not seen after water evaporation during scanning. While water alone is not enough to ensure processibility of bloodmeal, the alpha relaxation of pre-processed thermoplastic protein is heavily influenced by interactions with water. Because of large differences in moisture content (Table 2), samples were freeze dried in subsequent experiments to minimise the effect of water on relaxation behaviour. Further investigation into compositional effects by varying TEG or moisture content are beyond the scope of this paper, but these effects have previously been explored for processed bar samples conditioned at different relative humidity [35].

Freeze-dried pre-processed material

DMTA plots of freeze-dried WBM (Fig. 3a, b) were very similar to SBM (Fig. 1a) suggesting addition of water did not cause any irreversible change in its relaxation behaviour. The shift of Ti and Tii to the right, when compared with SBM is to be expected as moisture content has been reduced from 7 to 3 %, effectively a decrease in plasticisation. This shift confirms that Tii is affected by water, supporting the earlier hypothesis that evaporation could obscure the frequency dependence in this region.
Fig. 3

a Multifrequency DMTA data for freeze-dried bloodmeal plus 60 pphBM water and b representative 1 Hz DMTA data. c Modified bloodmeal without TEG after freeze drying and d modified bloodmeal with TEG after freeze drying

FD PPM (Fig. 3c) showed similar three peak behaviour to its non-freeze-dried counterpart at elevated temperatures, confirming the hypothesis that in PPM these were dry relaxations seen after water evaporation. Again, the middle of these three showed good correlation with the Arrhenius equation over a wide temperature range (Table 4). Below ambient, the freeze-dried sample differed from the wet sample. One low temperature tan δ peak was seen at 215 K. Given the amount of water that has been removed, this will be only a local motion, rather than the long range motions responsible for the peak at 193 K without freeze drying. Minimal drop in modulus in the freeze-dried sample confirmed this. It was, therefore, concluded that Ti and Tii seen earlier in the wet samples (Fig. 2) were glass transitions in water-plasticized protein.
Table 4

Transition temperatures identified from peaks in tan δ in freeze-dried bloodmeal and pre-processed material

Sample

Figure label

Temperature/K for each frequency/Hz

Fits to Arrhenius model

0.1

0.3

1

3

10

30

 

EA/kJ/mol

r2

FD WBM

Ti

195

199

205

210

228

227

Yes

56

0.940

 

Tii

346

350

355

361

355

353

No

n/a

n/a

 

Tiii

490

493

496

500

504

508

Yes

688

0.996

FD PPM

Ti

n/a

207

215

217

227

240

Yes

58

0.956

 

Tii

361

366

360

363

365

364

No

n/a

n/a

 

Tiii

419

422

427

431

437

443

yes

372

0.994

 

Tiv

472

472

477

477

473

471

no

n/a

n/a

FD PPM–TEG

Ti

199

204

206

211

225

229

yes

66

0.948

 

Tii

266

266

269

275

282

282

yes

176

0.916

 

Tiii

326

330

337

342

347

354

yes

194

0.996

 

Tiv

506

503

504

504

504

504

no

n/a

n/a

1 Hz data is in bold

Including TEG further plasticised the freeze-dried material (Fig. 3D), even in the absence of water. The tan δ peak at 337 K (labelled Tiii) was considered to be the glass transition, corroborated by a large drop in modulus and good frequency dependence (Table 4). The peak at 269 K (Tii) would then be a β-transition and also showed frequency dependence. The peak labelled Ti at 206 K was thought to represent the same short-range motions responsible for Ti in FD PPM without TEG and FD WBM. This is justified as all these peaks showed similar Arrhenius activation energy (Table 4). The high temperature peak at Tiv = 504 K in FD PPM–TEG was thought to be an unplasticised dry-Tg similar to Tiv for FD PPM (Fig. 3C).

The results for freeze-dried material showed that urea, SDS and sodium sulphite had a plasticising effect in dry protein and that TEG caused additional plasticisation even after drying.

Processed material

Consolidated bars

Figure 4 shows the results for processed material as self-supporting bar samples, (a) without and (b) with TEG. In both the cases, a large peak in tan δ is apparent (labelled Tii), with a second peak at Ti = 0.74Tii, Tii and Ti were thought to be the glass transition and a beta relaxation, respectively, as it has long been recognised that due to similar, but more localised motions, Tβ ≅ 0.75Tg in amorphous and semi-crystalline polymers [33, 36].
Fig. 4

Representative 1 Hz DMTA plots of a processed bar sample with no TEG, b processed bar sample with 20 pphBM TEG. The grey shading represents ±20 K from the peaks identified in tan δ

If the onset of drop in E′ had been used to determine Tg, it could mistakenly have been identified 100 K lower than the temperature from the large tan δ peak, as the drops in modulus for both transitions appeared convoluted. Variation of up to 20 K (indicated by the grey shading in Fig. 4) is normal and to be expected [7], but differences of 100 K indicate different transitions are being identified by the two techniques. To clarify this further, the modulus curves were also plotted on a logarithmic scale (Fig. 5). Natural and logarithmic E′ scales give different onset values, but are both accepted for assigning Tg values [37]. Natural scales have been used for the bulk of this work as the pocket prevents sample stiffness dropping more than an order of magnitude. In the case of the bar samples; however, the logarithmic scale clarifies that the drop of two orders of magnitude, typical of Tg, is associated with the upper tan δ peak (Tii). The onset of this drop is still ~50 K lower than the peak in tan δ, (308 K compared with 350 K in PPM and 287 K compared with 335 K in PPM–TEG) and the peaks themselves, especially with TEG, are rather broad. This is typical of low crystallinity polymers, in which crystals restrict the general long range segmental motions responsible for the amorphous Tg, leading to a broad transition [38].
Fig. 5

Representative log scale 1 Hz DMTA E′ plots of a processed bar sample with no TEG and b processed bar sample with 20 pphBM TEG

Comparison between bars and ground processed material

A direct comparison between pre-processed material and processed bar samples is restricted because of the absence of the stiff pocket and extensive moisture loss during processing. To overcome these, bar samples were ground and analysed in material pockets before and after freeze drying (Fig. 6). DMTA results of bar and ground-processed samples indicated that material pockets can successfully be used to detect transitions present in the bulk material. In addition, as the pocket maintains stiffness, it enables analysis at higher temperatures at which stiffness drops in self-supporting samples.
Fig. 6

Representative scans of a MTP in material pocket, b freeze-dried MTP in a material pocket, c MTP–TEG in a material pocket and d freeze-dried MTP–TEG in a material pocket

In ground MTP, peaks in tan δ were seen at Ti = 257 and Tii = 344 K (Fig. 6a), corresponding to peaks at Ti and Tii seen in the bar (Fig. 4a). The higher temperature transition seen in Fig. 4a (Tii), resolved into two peaks, Tii = 344 K and Tiii = 380 K (Fig. 6a) using the material pocket. This would suggest the material pocket is more sensitive to determine multiple transitions, typically appearing convoluted in a bar sample.

In ground MTP–TEG, a transition was seen at Ti = 236 K, thought to correspond to Ti in the bar sample. The bar sample had a broad transition centred at Tii = 335 K, while the ground sample showed peaks at Tii = 308 K at Tiii = 375 K (Fig. 6b). Previous studies of protein/sugar mixtures found that material pocket DMTA was more sensitive to phase transitions than other techniques [8, 10]. The difference in MTP–TEG could be due to resolution of multiple Tgs being enhanced in the material pocket. Two glass transitions are sometimes distinguished in semi-crystalline polymers, Tg(L), arising from the purely amorphous phase and Tg(U), arising from amorphous material restrained by crystallites. A general rule when this is the case is that Tg(U) ≅ (1.2 ± 0.1)Tg(L) [33]. In MTP–TEG, Tiii/Tii = 1.2, suggesting that the stiffness of the pocket has resolved the broad Tg seen in the MTP–TEG bar into these two components.

While the material pocket technique offered better resolution of phases as detected by peaks in tan δ, it does not mean that the transitions detected in bar samples were wrong. The peak in tan δ seen in bars at 354 K (without TEG) or 339 K (with TEG) is the behaviour of the bulk material when not constrained by a non-dampening metal. Where self-supporting samples can be prepared, the material pocket may be useful for resolving multiple transitions, but should not replace self-supporting samples to determine average behaviour of bulk material.

With samples that cannot be self supporting, it should be remembered that results are for the hybrid system. From the results presented here, the shape of the E″ curves were quite different to the shape of the tan δ curves for bar samples (Fig. 5), while they were similar in material pockets (Fig. 6). In DMTA, E″ is not directly measured, but calculated from the phase lag and sample stiffness, which is a combined property of the material pocket and the sample. The material pocket stiffness is much higher than the polymer which may distort E″ resulting in a response that appears different, but still gives similar transition temperatures.

Because of this composite effect between the material pocket and the material, differences between transition temperature assignment were compared when using E′, E″ and tan δ (Table 5). Low temperature transitions (Ti) were very similar between all methods used for assigning the transition temperature, regardless of using bar or material pocket samples. However, the second transition seen in bar samples (Tii) resolved into two peaks when using the material pocket technique (Tii and Tiii). Tii detected in bar samples varied with about 40 K between the methods used to determine the transition. When this peak was resolved into two peaks in the material pocket, the variation between the temperature assignments using the different techniques was also about 40 K for Tii, but much less for Tiii.
Table 5

Transitions identified in processed material for bar and material pocket samples

  

Transition temperatures/Ka

Bar:

Ti

Tii

 

Pocket:

 

Tii

 

Tiii

Tiv

MTP

 tan δ

Bar

257

350

64/0.945

Pocket

257

344

986/0.978

380

−453/0.980

426

410/0.987

 E

Bar

239

195/0.916

310

Pocket

255

339

673/0.972

 

378

−336/0.943

421

395/0.98

 E

Bar

210

186/0.960

Not detected

Pocket

225

301

384/0.992

399

408

 LogE

Bar

218

308

Pocket

222

304

Not detected

402

308/0.936

MTP–TEG

 tan δ

Bar

246

154/0.922

 

335

  

Pocket

236

308

287/0.993

375

455

 E

Bar

240

202/0.987

272

126/0.986

Pocket

232

296

141/0.934

373

446

E

Bar

218

245/0.959

Not detected

Pocket

228

86/0.979

Not detected

Not detected

448

 LogE

Bar

223

231/0.950

285

395/0.959

Pocket

226

285

Not detected

450

aIf the transition has fit the Arrhenius equation, activation energy (EA/kJ/mol) and goodness of fit (r2) are shown after the temperature. Data for other frequencies not shown (0.1, 0.3, 3, 10 and 30 Hz)

One could, therefore, conclude that between the methods used to define transition temperatures, the variation is reasonable for material pockets, but more severe in bar samples. It would be good practise to use E′, log E′, E″ and tan δ to assign transitions and material pockets may also offer greater resolution

Comparison of freeze-dried material before and after processing

For a comparison at similar moisture content, pre-processed and processed powders were compared after freeze drying. The results are shown in Figs. 3c and 6b for materials without TEG and Figs. 3d and 6d for materials with TEG. The general shape of the tan δ curves was very similar before and after processing. However, the relative magnitude of the triple peaks between 300 and 500 K was different before and after processing for both cases. It was thought that assessing the materials as semi-crystalline (α-helix and β-sheets), Tii would correspond to a glass transition of bulk amorphous material, Tiii to chains trapped between crystallites and Tiv to a relaxation temperature of chains within crystallites. On this premise, it would appear that extrusion and injection moulding resulted in chain rearrangement leading to more bulk amorphous chains evident from a greater drop in E′ at lower temperatures. This was less evident in samples containing TEG as these would be significantly more plasticised leading to much lower transition temperatures.

Conclusions

Bloodmeal had a Tg at 493 K which is too high for thermoplastic processing. Adding water lowered the Tg to 263 K, although a low Tg alone is not enough to ensure processing.

Modified bloodmeal had a much lower Tg of 193 K due to urea, SS and SDS reducing chain interactions. Modified bloodmeal also showed complex transitions at temperatures above the Tg. These transitions were thought to be associated with dehydrated bulk amorphous chains, amorphous sections between crystallites and short chain segments within crystalline regions. Including TEG as plasticiser lowered these transitions considerably.

It was concluded that bloodmeal’s Tg was largely affected by water even after modification in pre-processed material. Analysis before and after freeze drying revealed that without protein modification the material resorted to the relaxation behaviour of standard bloodmeal, while modification resulted in a permanently lowered Tg.

The material pocket technique gave similar results compared to injection-moulded bar samples at low temperatures, but had increased resolution at higher temperatures. For processed materials, higher temperature transitions seen in bar samples, resolved into three peaks using the material pocket due to the semi-crystalline nature of the material. The relative proportion of these transitions appeared different in processed material compared to pre-processed materials suggestion that some chain rearrangement occurred resulting in more bulk amorphous regions.

Using the material pocket did not lead to additional difficulties in identifying transitions. However, it was found that the technique used for identifying a transition lead to slightly different transition temperatures. Using the onset in E′, peak in E″ or tan δ should be used in combination, preferably checking for transitions on a natural and logarithmic scale. This phenomenon was even more pronounced in consolidated bar samples.

References

  1. 1.
    Verbeek CJR, van den Berg LE. Extrusion processing and properties of protein-based thermoplastics. Macromol Mater Eng. 2010;295(1):10–21. doi:10.1002/mame.200900167.CrossRefGoogle Scholar
  2. 2.
    Verbeek CJR, van den Berg LE. Development of proteinous bioplastics using bloodmeal. J Polym Environ. 2010;1:1–10. doi:10.1007/s10924-010-0232-x.Google Scholar
  3. 3.
    Bennion BJ, Daggett V. The molecular basis for the chemical denaturation of proteins by urea. Proc Natl Acad Sci USA. 2003;100(9):5142–7. doi:10.1073/pnas.0930122100.CrossRefGoogle Scholar
  4. 4.
    Verbeek CJR, Viljoen C, Pickering KL, van den Berg LE. Plastics material. NZ Patent NZ551531. Waikatolink Limited, Hamilton (2009).Google Scholar
  5. 5.
    Verbeek CJR, van den Berg LE. Structural changes as a result of processing in thermoplastic bloodmeal. J Appl Polym Sci. 2012. doi:10.1002/app.36964.
  6. 6.
    Silalai N, Roos YH. Dielectric and mechanical properties around glass transition of milk powders. Dry Technol. 2010;28(9):1044–54. doi:10.1080/07373937.2010.505520.CrossRefGoogle Scholar
  7. 7.
    Menard KP. Dynamic mechanical analysis: a practical introduction. 2nd ed. Boca Raton, FL: CRC Press; 2008.CrossRefGoogle Scholar
  8. 8.
    Royall PG, Huang CY, Tang SWJ, Duncan J, Van-de-Velde G, Brown MB. The development of DMA for the detection of amorphous content in pharmaceutical powdered materials. Int J Pharm. 2005;301(1–2):181–91. doi:10.1016/j.ijpharm.2005.05.015.CrossRefGoogle Scholar
  9. 9.
    Pinheiro A, Mano JF. Study of the glass transition on viscous-forming and powder materials using dynamic mechanical analysis. Polym Test. 2009;28(1):89–95. doi:10.1016/j.polymertesting.2008.11.008.CrossRefGoogle Scholar
  10. 10.
    Carpenter J, Katayama D, Liu L, Chonkaew W, Menard K. Measurement of t-g in lyophilized protein and protein excipient mixtures by dynamic mechanical analysis. J Therm Anal Calorim. 2009;95(3):881–4. doi:10.1007/s10973-007-8986-7.CrossRefGoogle Scholar
  11. 11.
    Raschip IE, Yakimets I, Martin CP, Paes SS, Vasile C, Mitchell JR. Effect of water content on thermal and dynamic mechanical properties of xanthan powder: a comparison between standard and novel techniques. Powder Technol. 2008;182(3):436–43.CrossRefGoogle Scholar
  12. 12.
    Silalai N, Roos YH. Coupling of dielectric and mechanical relaxations with glass transition and stickiness of milk solids. J Food Eng. 2011;104(3):445–54.CrossRefGoogle Scholar
  13. 13.
    Gearing J, Malik KP, Matejtschuk P. Use of dynamic mechanical analysis (DMA) to determine critical transition temperatures in frozen biomaterials intended for lyophilization. Cryobiology. 2010;61(1):27–32.CrossRefGoogle Scholar
  14. 14.
    Gârea S-A, Iovu H, Nicolescu A, Deleanu C. Thermal polymerization of benzoxazine monomers followed by GPC, FTIR and DETA. Polym Test. 2007;26(2):162–71.CrossRefGoogle Scholar
  15. 15.
    Gupta P, Bansal AK. Devitrification of amorphous celecoxib. AAPS PharmSciTech. 2005;6(2):E223–30. doi:10.1208/pt060232.CrossRefGoogle Scholar
  16. 16.
    Mano JF. Thermal behaviour and glass transition dynamics of inclusion complexes of alpha-cyclodextrin with poly(d,l-lactic acid). Macromol Rapid Commun. 2008;29(15):1341–5. doi:10.1002/marc.200800180.CrossRefGoogle Scholar
  17. 17.
    Paes SS, Sun SM, MacNaughtan W, Ibbett R, Ganster J, Foster TJ, et al. The glass transition and crystallization of ball milled cellulose. Cellulose. 2010;17(4):693–709. doi:10.1007/s10570-010-9425-7.CrossRefGoogle Scholar
  18. 18.
    Williams MA, Jones DS, Andrews GP. A study of drug-polymer miscibility using dynamic mechanical thermal analysis. J Pharm Pharmacol. 2010;62(10):1400.Google Scholar
  19. 19.
    Silalai N, Roos YH. Mechanical relaxation times as indicators of stickiness in skim milk-maltodextrin solids systems. J Food Eng. 2011;106(4):306–17.CrossRefGoogle Scholar
  20. 20.
    Kemal E, Adesanya KO, Deb S. Phosphate based 2-hydroxyethyl methacrylate hydrogels for biomedical applications. J Mater Chem. 2011;21(7):2237–45. doi:10.1039/c0jm02984j.CrossRefGoogle Scholar
  21. 21.
    ASTM International. D638-03 standard test method for tensile properties of plastics. PA 19428-2959. ASTM International, West Conshohocken, PA; 2004.Google Scholar
  22. 22.
    Guo JX, Harn N, Robbins A, Dougherty R, Middaugh CR. Stability of helix-rich proteins at high concentrations. Biochemistry. 2006;45(28):8686–96. doi:10.1021/bi060525p.CrossRefGoogle Scholar
  23. 23.
    Michnik A. Thermal stability of bovine serum albumin DSC study. J Therm Anal Calorim. 2003;71(2):509–19. doi:10.1023/a:1022851809481.CrossRefGoogle Scholar
  24. 24.
    Heijboer J, Secondary loss peaks in glassy amorphous polymers. In: Boyer RF, Meier DJ. Midland macromolecular institute. In Dow Chemical Company, editors. Molecular basis of transitions and relaxations: papers. Midland macromolecular monographs. London: Gordon and Breach Science Publishers; 1978. p. xii, 429.Google Scholar
  25. 25.
    Rouilly A, Orliac O, Silvestre F, Rigal L. DSC study on the thermal properties of sunflower proteins according to their water content. Polymer. 2001;42(26):10111–7.CrossRefGoogle Scholar
  26. 26.
    Zhang J, Mungara P, Jane J. Mechanical and thermal properties of extruded soy protein sheets. Polymer. 2001;42(6):2569–78.CrossRefGoogle Scholar
  27. 27.
    Jerez A, Partal P, Martinez I, Gallegos C, Guerrero A. Rheology and processing of gluten based bioplastics. Biochem Eng J. 2005;26(2–3):131–8. doi:10.1016/j.bej.2005.04.010.CrossRefGoogle Scholar
  28. 28.
    Panagopoulou A, Kyritsis A, Serra RSI, Ribelles JLG, Shinyashiki N, Pissis P. Glass transition and dynamics in BSA-water mixtures over wide ranges of composition studied by thermal and dielectric techniques. BBA-Proteins Proteomics. 2011;1814(12):1984–96. doi:10.1016/j.bbapap.2011.07.014.CrossRefGoogle Scholar
  29. 29.
    Mo X, Sun X. Effects of storage time on properties of soybean protein-based plastics. J Polym Environ. 2003;11(1):15–22.CrossRefGoogle Scholar
  30. 30.
    Mo XQ, Sun XZ. Thermal and mechanical properties of plastics molded from urea-modified soy protein isolates. J Am Oil Chem Soc. 2001;78(8):867–72.CrossRefGoogle Scholar
  31. 31.
    Oliviero M, Maio ED, Iannace S. Effect of molecular structure on film blowing ability of thermoplastic zein. J Appl Polym Sci. 2010;115(1):277–87.CrossRefGoogle Scholar
  32. 32.
    Menczel JD, Prime RB. Thermal analysis of polymers: fundamentals and applications. Hoboken: Wiley; 2009.CrossRefGoogle Scholar
  33. 33.
    van Krevelen DW, Nijenhuis Kt. Properties of polymers :their correlation with chemical structure : their numerical estimation and prediction from additive group contributions, 4th edn. Amsterdam: Elsevier; 2009.Google Scholar
  34. 34.
    Dow Chemical Company. Triethylene glycol. 2007. http://msdssearch.dow.com/PublishedLiteratureDOWCOM/dh_004d/0901b8038004d042.pdf. Accessed 8 March 2012.
  35. 35.
    Verbeek C, Koppel N. Moisture sorption and plasticization of bloodmeal-based thermoplastics. J Mater Sci. 2011;47:1187–95. doi:10.1007/s10853-011-5770-7.CrossRefGoogle Scholar
  36. 36.
    Boyer RF, Turley SG. Molecular Motion in polystyrene. In: Boyer RF, Meier DJ, Midland Macromolecular Institute, Dow Chemical Company, editors. Molecular basis of transitions and relaxations: papers. Midland macromolecular monographs, vol. 4. London: Gordon and Breach Science Publishers; 1978. p. xii, 429.Google Scholar
  37. 37.
    ASTM International. E1640-04 standard test method for assignment of the glass transition temperature by dynamic mechanical analysis. West Conshohocken, PA: ASTM International; 2004.Google Scholar
  38. 38.
    Boyd RH. Relaxation processes in crystalline polymers: experimental behavior—a review. Polymer. 1985;26(3):323–47. doi:10.1016/0032-3861(85)90192-2.CrossRefGoogle Scholar

Copyright information

© Akadémiai Kiadó, Budapest, Hungary 2012

Authors and Affiliations

  1. 1.School of Engineering University of WaikatoHamiltonNew Zealand

Personalised recommendations