Full Paper

Preparation and Recycling of Plasticized PLA

Roberto Scaffaro,* Marco Morreale, Filippo Mirabella,Francesco Paolo La Mantia

PLA is one of the most interesting new polymeric materials, mainly because of its favorableenvironmental features. On the other hand, one of its drawbacks is represented by itsbrittleness and low impact resistance. A possible solution may be the use of proper impactmodifiers. Furthermore, limited knowledge existson the behavior of PLA upon recycling. In thispaper, a detailed investigation has been carriedout on the behavior of PLA systems upon addingsmall amounts of commercial impact modifiers,and the effects of multiple recycling steps havebeen studied. The use of impact modifiers caneffectively improve mechanical properties. Recy-cling led to a significant reduction in the impactstrength; however a relatively high fraction of theother mechanical properties was still retained.

Introduction

Poly(lactic acid) (PLA)hasbeengaininga rising interest over

the last years as a biodegradable and environmental-

friendly substitute of traditional non-biodegradable poly-

mers.

Several ways exist to synthesize PLA, however, poly-

merization through lactide formation is the most used on

industrial scale, allowing to obtain optimum results in

terms of mechanical and thermal properties.[1,2]

PLAmainproperties can significantly change, depending

on the distribution and the ratio of the two D and L

enantiomers in the final polymer.

R. Scaffaro, M. Morreale, F. Mirabella, F. P. La MantiaDipartimento di Ingegneria Chimica dei Processi e dei Materiali,Universita di Palermo, Viale delle Scienze, 90128 Palermo, ItalyFax: þ39 91 702 5020; E-mail: scaffaro@dicpm.unipa.itM. MorrealeLibera Universita Kore Enna, Facolta di Ingegneria e Architettura,Cittadella Universitaria, 94100 Enna, Italy

Macromol. Mater. Eng. 2011, 296, 141–150

� 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim wileyonlin

PropertiesofPLAhavebeendirectly comparedwith those

of other common packaging polymers,[1] finding that the

mechanical performance is similar to that of PET and

globally higher than that of PS. On the other side, resistance

against UV radiations and barrier properties are the worst

of the set, although barrier properties against CO2, O2 and

water are intermediate.

Lima et al.[3] studied the behavior of crystallinity,

rheological properties and thermal degradation uponusing

different processing techniques. They found that low

amounts of the D(þ) enantiomer (below 1.5 wt.-%) can

accelerate the crystallization process: spherulite growth

speed depends on molecular weight and decreases upon

increasing the latter. A remarkable level of brittleness was

observed, especially uponprocessing via injectionmolding.

Brittleness is an important limit of PLA, therefore several

ways have been investigated in order to improve its

mechanical properties.

Some researchers prepared nanocomposites based on

PLA and nanofillers such as montmorillonite or calcium

carbonate.[1,4–6] It was found that the nanocomposites had

elibrary.com DOI: 10.1002/mame.201000221 141

Table 1. Main physical properties of the materials used asreported by the manufacturers.

Material Tg Tm Density MFIa)

-C -C g � cm�3 g � (10 min)�1

PLA 55–60 160–170 1.25 10–30

Polyone 55 n/a 1.2 n/a

Sukano 55 n/a 1.25 n/a

a)At 210 8C and 2.16 kg load.

142

www.mme-journal.de

R. Scaffaro, M. Morreale, F. Mirabella, F. P. La Mantia

higher thermal stability and the nanofiller acted as

nucleating agent; the use of poly(ethylene glycol) (PEG)

significantly facilitated intercalation. Calcium carbonate

and organomodified clay nanocomposites showed differ-

ent impact breaking mechanisms: the former broke by

crazing (due to microvoids), while the latter by shear

yielding.

Some papers are also available on blends based on PLA

and other biodegradable polymers. Lopez-Rodriguez et al.[7]

studied crystallization behavior, morphology andmechan-

ical properties of PLA/polycaprolactone blends. PCL showed

poor compatibility with PLA throughout all the composi-

tion range investigated and influenced also PLA crystal-

lization. Shibata et al.[8] investigated mechanical proper-

ties, morphology and crystallization of poly(L-lactic acid)

(PLLA)/poly(butylene succinate) and PLLA/poly[(butylene

succinate)-co-lactide] blends. None of the blends showed

miscibility but, on the contrary, clear phase separation as

testified by formation of microdomains. Other researchers

prepared blends of PLA and other biodegradable polymers

such as poly(butylene adipate terephtalate) (PBAT) and

polyhydroxyalkanoate (PHA).[9,10] In the former, reductions

of elastic modulus and of tensile strength were observed,

while impact strength significantly increased; in the latter,

significant enhancements of the globalmechanical proper-

ties were found.

Interesting and cheap ways to improve PLA properties

are based on elastomeric polymers addition.

The validity of this way is justified also by considering

that the addition of appropriate elastomers to non-

biodegradable polymers and polymer blends proved to

increase properties such as, for instance, impact strength

and elongation at break.[11–13]

Nijenhuis et al.[14] investigated the miscibility and the

mechanical properties of PLA/poly(ethylene oxide) (PEO)

blends, finding significant increases of elongation at break

upon increasing PEO amount, but also some reduction of

the tensile strength. Anderson et al.[15] prepared PLA/linear

low-density polyethylene (LLDPE) blends, finding that

LLDPE enhanced impact strength, even though semicrystal-

line PLA should be used instead of amorphous one.

Piorkowskaetal.[16] studied the influenceofpoly(propylene

glycol) (PPG) on PLA. This polymer had plasticizing effects

and significantly reduced Tg and elastic modulus, while

elongation at break was greatly increased. On the other

hand, it showed a strong tendency to phase separation

(higher upon increasing the molecular weight).

Hiljanen-Vainio et al.[17] studied the modifications

caused by the addition of elastomeric urethane polyesters

(PEU) to PLA. The former were obtained by modifying

urethane polyester with biodegradable elastomers. Sig-

nificant enhancements of the toughness were observed.

Similarly, Ryynanenetal.[18] synthesizedmultiblockpoly(e-caprolactone)-block-poly(D,L-lactide)-block-poly(e-caprolac-

Macromol. Mater. Eng. 2

� 2011 WILEY-VCH Verlag Gmb

tone) copolymers to be used with urethane polyesters in

order obtaining PEUs.

Byrne et al.[19] investigated on the properties of PLA

blends containing four different commercial master-

batches, finding impact strength or clarity improvements

depending on the specific additive used, without suffering

significant worsening of tensile properties or thermal

stability.

The increase of the industrial interest on PLAmakes it of

primary importance to assess its behavior upon recycling,

especially if additivesareused,making theevaluationmore

difficult. To our best knowledge, there is not much data in

the literature on the recyclability of PLA. Pillin et al.[20]

investigated the evolution of mechanical and rheological

properties ofneat PLAupon recycling, finding that themain

tensile and rheological properties worsened with the

thermomechanical cycles, except for the tensile modulus

which kept substantially constant.

In this work, the effect of two different impactmodifiers

on PLA has been studied, by characterizing themechanical,

thermomechanical, rheological and morphological beha-

vior. The recyclability of the obtained materials has been

thus investigated.

Experimental Part

Materials and Preparation

The PLA used is commercialized as ‘Natureworks 3001D’ by

Natureworks LLC (USA). Itsmainphysical properties are reported in

Table 1.

Twodifferent impactmodifierswereused.Thefirstone isknown

as ‘PolyoneOnCapBIOImpact’ (herebyindicatedas ‘Polyone’) and is

commercialized by Polyone Corporation (USA). It is a masterbatch

with proprietary composition, containing a specific organic

elastomer which improves the impact strength of PLA. The second

one is commercialized by Sukano (Switzerland) as ‘IMS550’ (hereby

indicated as ‘Sukano’) and is a proprietary composition master-

batch too, based on an elastomeric organic impact modifier. The

main properties of both the impact modifiers are reported in

Table 1.

All the materials, prior to processing, were pre-treated in order

eliminating their humidity, by keeping them in a vacuum oven at

011, 296, 141–150

H & Co. KGaA, Weinheim www.MaterialsViews.com

Table 2. Labels and compositions of the investigated blends.

Blend PLA Polyone Sukano

wt.-% wt.-% wt.-%

PLA10P 90 10 0

PLA12.5P 87.5 12.5 0

PLA15P 85 15 0

PLA4S 96 0 4

PLA6S 94 0 6

PLA8S 92 0 8

Preparation and Recycling of Plasticized PLA

www.mme-journal.de

110 8C for 4 h. This choice was based on direct desorption

measurements performed at that temperature (which showed

4h to be sufficient to achieve a constant weight).

Several systems were prepared by varying the amounts of PLA,

Polyone and Sukano, as reported in Table 2.

Eachsystemwasprocessedvia twodifferentways: thefirstwas a

direct feeding of the materials to a Sandrettom30 (Italy) injection

molding press (temperature profile: 180–200–200 8C, injection

pressure¼ 900 bar, injection speed¼ 14cm � s�1) in order to obtain

samples for further characterizations; the second was similar, but

the injection molding was preceded by an extrusion step in a

corotating twin-screw extruder (OMC, Italy; L/D¼ 25, D¼ 19mm,

T¼170–170–180–180–190–190–190 8C, speed�270 rpm)andgran-

ulation in an Accrapak (UK) BM15 grinder. In both cases, it was

necessary to adjust the feeding rate to the injection molding

machinebetween62and64.5 cm3 � s�1dependingonthesystemfed.

Recycling was performed by reprocessing up to three cycles the

obtained granules in the injectionmoldingmachine, following the

same preparation steps and processing parameters indicated

above.

Characterization

All the samples, before characterization, were annealed in oven at

110 8C for 90min.[6] This method was adopted in order to mitigate

and wipe out the influence of different crystallinity degrees of the

materials, causedbythedifferent thermalhistoriesassociatedwith

the different processing techniques, on the mechanical properties

of the samples.

Rheological characterization was performed in two ways. The

first one included melt flow index (MFI) measurements on the

samples (load¼2.16 kg, T¼190 8C), while the second one included

flow curves determination by using a Rheometric Scientific RDAII

rotational rheometer (T¼190 8C, strain¼10%, frequency range

0.1–500 rad � s�1). In order to determine the changes in the

molecular weight upon processing, percent variations of the

dimensionlessNewtonianviscosity,h0i/h0P,werecalculated,where

h0i¼Newtonian viscosity of the ‘i’ system, h0P¼Newtonian

viscosity of PLA as delivered. The direct relationship between

Newtonian viscosity and weight-averaged molecular weight is

given by the well-known power law

www.M

h0 ¼ kM3:4w (1)

aterialsViews.com

Macromol. Mater. Eng. 2

� 2011 WILEY-VCH Verlag Gmb

where h0 is the Newtonian viscosity, k is a constant andMw is the

weight-average molecular weight.

Mechanical characterization included tensile, flexural and

impact and tests. Tensile tests were performed by using an Instron

3365 (USA) universal testing machine according to ASTM D882 on

specimens 100mm long, 5mm wide and 2.1–2.3mm thick.

Crosshead speed was set at 5mm �min�1 with a 50mm grips

distance. At least seven specimens were tested for each material.

Flexural testswereperformedbymeansofan Instron3365 (USA)

universal testing machine according to ASTM D790 on specimens

108mm long, 35mmwide and 2.1–2.3mm thick. Crosshead speed

was set at 0.9mm �min�1 and at least eight specimenswere tested

for each material.

Impact tests were performed according to ASTM D256 in Izod

mode, by using a Ceast (Italy) 6545/000 universal testingmachine.

At least six specimens were tested for each material.

Thermomechanical resistance tests were based on heat-deflec-

tion temperature (HDT) measurements, according to ASTM

D2990 (flexural load¼ 1.8MPa, temperature increase rate -

¼120 8C �min�1) on 120mm long, 15mm wide and 3mm thick

specimens, by using a Ceast (Italy) 6505/000 equipment.

Hygroscopic sensitivity of the recycledmaterialswasperformed

according to ASTM D570 which was followed in order to perform

the characterization. Humidity absorption testswere performed in

a controlled environment, by putting the specimens in a Binder

(Germany) climatic chamber at T¼ 23 8C and 50% relative

humidity. Weight variations of the samples were recorded daily

and reported as wt.-% of absorbed water, i.e., according to the

formula

011, 29

H & Co

DW ¼ Wf�W0ð Þ=W0�100 (2)

where Wf is the measured weight at t time and W0 is the initial

weight.

Some materials were analyzed by differential scanning calori-

metry in order to explain the results of mechanical tests. The

samples were tested in a Perkin-Elmer (USA) DSC7 differential

scanning calorimeter. They were subjected to a first heating to

210 8C in order to erase the previous thermal history, then cooling

(at 10 8C �min�1) and a second heating (to 210 8C at 10 8C �min�1).

Crystallization, cold crystallization and melting temperatures (Tc,

Tcc and Tm, respectively) were then determined, as well as the

related enthalpies.

Finally, morphological characterization was based on scanning

electronmicroscopy (SEM)analysis (FEIQuanta200ESEM,USA). All

the samples were broken in liquid nitrogen and covered with gold

in order to make them electrically conductive.

Results and Discussion

Rheology

The values of melt flow rate of the virgin materials were

found tobe 11.9, 13.1, 69.5 and�0.1 dg �min�1, for PLA, PLA-

E, Sukano and Polyone, respectively. The values for the

extruded blends are reported in Figure 1.

6, 141–150

. KGaA, Weinheim143

Figure 2. Rheological curves of neat PLA and blends obtained by extrusion plus injectionmolding (E-in) or direct injection molding (in).

0

5

10

15

20

25

0 2 4 6 8 10 12 14 16Impact modifier, wt%

MFI

[dg/

min

]

PLA+Sukano, theoreticalPLA+SukanoPLA+Polyone, theoreticalPLA+Polyone

Figure 1. MFI as a function of the impact modifier content for: Sukano, actual trend;Sukano, theoretical trend according to linear composition law; Polyone, actual trend;Polyone, theoretical trend according to linear composition law (assuming a MFI of0.1 dg �min�1 for Polyone).

144

www.mme-journal.de

R. Scaffaro, M. Morreale, F. Mirabella, F. P. La Mantia

It canbeobserved that PLAshows the lowest valueofMFI

(and thus the highest viscosity) if compared to the blends.

With concern to the impact modifiers, Sukano has a

relatively high MFI, while Polyone shows a very low MFI

and thus high viscosity. The behavior of the blends is

significantly different upon changing the amount of the

impactmodifier. The increase in Sukano concentration (4, 6,

8wt.-%), in fact, leads to a gradual increase ofMFI (thus to a

reduction of viscosity), with a roughly linear trend, while

the increase in Polyone concentration (10, 12.5, 15 wt.-%)

seems to initially increase theMFI and then stabilize it at a

nearly steady value.

Macromol. Mater. Eng. 2011, 296, 141–150

� 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinhe

Forbothof the impactmodifiers, apoor

compatibility with the PLA primary

phase can be invoked.

In fact, if the blends obeyed a simple

additive rule, the MFI trend would have

coincided with the theoretical line in

Figure 1. The actual lines, instead, are

significantly higher than the theoretical

ones, thus indicating incompatibility in

terms of viscosity.

Further considerationsmaybedoneon

the compatibility of the two impact

modifiers. Figure 1 shows that the actual

and the ideal behavior of the PLA/Sukano

blends are practically parallel, while the

behavior of PLA/Polyone blends is sig-

nificantly different, with the two curves

diverging. This seems to indicate an

increased incompatibility, for PLA/Poly-

one systems, upon increasing Polyone

level.

Figure 2 reports the rheological curves

for neat PLA, processed via injection

molding (PLA-in) or by extrusion fol-

lowed by injection molding (PLA-E-in)

and some blends with the impact modi-

fiers, processed only by injection mold-

ing.

It can be observed that PLA-E-in

viscosity is lower than that of PLA-in in

the whole frequency range. This proves

that the additional processing step

through the corotating twin-screwextru-

der induces a significant, additional

degradation. A quantitative assessment

on the reduction of themolecular weight

upon processing can be obtained by

considering the reduction in the dimen-

sionless Newtonian viscosity, h0i/h0P. In

particular, itwas found thath0i/h0P is 0.86

for PLA-in and 0.74 for PLA-E-in, and thus,

accordingtoEquation (1), thereduction in

molecular weight is of �4 and 8%, respectively, in total

agreement with the above qualitative considerations.

The blends containing the impact modifiers generally

show lower viscosities than thoseof neat PLA. Furthermore,

Sukano containing blends show a slightly more non-

Newtonian behavior in comparison to Polyone containing

blends.

The effect of the impact modifiers content is reported in

Figure 3. It can be observed that all the blends show

viscosity valueswhich are globally lower than those of PLA-

E-in. This confirms the above considerations about the

degrading effect associated with the double processing

im www.MaterialsViews.com

Figure 3. Flow curves of PLA and blends obtained through extrusion and injectionmolding (E-in) upon changing the modifiers content.

Preparation and Recycling of Plasticized PLA

www.mme-journal.de

technique and the reduction of viscosity caused by the

additionof the impactmodifiers. Sukanoshowsamorenon-

Newtonian behavior than Polyone and these effects are

more remarkable upon increasing the modifiers content.

The crossover point between Sukano and Polyone related

curves occurs approximately between 10 and 100 rad � s�1,

which is the typical range for industrial processing.

Theflowcurvesofneat SukanoandPolyone (not reported

in the figure for sake of clarity) are, respectively, signifi-

cantlyhigherand lower thanthose reported inFigure3.This

means that the behavior of the blends is remarkably

different from what predictable on the basis of a simple

additive rule and, therefore, there is incompatibility

betweenPLAand the impactmodifiers (especially Polyone).

Table 3. Flexural properties of the investigated materials.

Material Flexural modulus Flexural strength

GPa MPa

PLA-E-in 3.8� 0.08 85.7� 5.4

PLA10P-E-in 3.73� 0.05 88.8� 1.4

PLA12.5P-E-in 3.59� 0.13 91.7� 2

PLA15P-E-in 3.63� 0.11 94� 2.4

PLA4S-E-in 3.67� 0.08 89.2� 2.3

PLA6S-E-in 3.55� 0.11 94.6� 3.3

PLA8S-E-in 3.42� 0.1 99.3� 2.3

PLA-in 3.77� 0.18 90� 4.1

PLA12.5P-in 3.37� 0.3 91.2� 3.2

PLA6S-in 3.45� 0.34 93.5� 3.0

PLA8S-in 3.15� 0.03 97.1� 1.1

Sukano 1.36� 0.12 28.3� 1.6

Polyone 1.6� 0.14 37.6� 3.2

Mechanical Properties

The results of the flexural tests are reported in Table 3. As

predictable, the values of the neat impact modifiers are

much lower than those of the blends. It can be stated that

the addition of the impact modifiers generally reduces the

flexural modulus and increases the flexural strength.

Further considerations may be done with concern to the

effect of the impactmodifierused, theprocessing technique

and the amount of modifier.

The addition of the impactmodifiers led, in both cases, to

a slight reduction of the modulus and an increase of the

strength.

The ‘E-in’ samples showed slightly higher moduli than

‘in’ ones and also the flexural strength was slightly higher.

This can be explained considering two separate effects,

namely the degradation due to processing and the

dispersion of the modifier-related phase inside the matrix

phase. The double processing caused higher degradation,

but also better dispersion of themodifier’s phase inside the

www.MaterialsViews.com

Macromol. Mater. Eng. 2011, 296, 141–150

� 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinhe

matrix phase and the benefits related to

this better dispersion overcame the

worsening of the properties due to

degradation. This can explain the slightly

higher values of the flexural strength.

With regard to the moduli, the observed

trends can be explained considering

recrystallization effects subsequent to

the degradation.[20] This can explain also

the behavior observed in the neat PLA

samples, where themodifier’s dispersion

factor does not apply.

Regarding the amount of modifier

used, the modulus increases upon

increasing Sukano wt.-%, while in the

case of Polyone it decreases upon switch-

ing from 10 to 12.5 wt.-% and then keeps

almost constant on switching from 12.5

to 15 wt.-%. The flexural strength

increases in both cases upon increasing the modifier

content.

Indeed, the results foundwere somewhatunexpected, on

the basis of the compatibility issues previously reported

and discussed. A possible explanation may be found

hypothesizing that the impact modifiers exert some

impediment to crack propagation.

Table 4 reports the results of impact tests. First, it can be

observed that significant differences occur between E-in

and in materials. In particular, E-in samples show higher

impact strength values, even up to 50%. This can be

explained considering that, for the impactmodifiers to fully

exert their primary function (absorbing and dissipating

im145

Table 4. Impact properties and HDT values of the investigatedmaterials.

Material Impact Strength HDT

J�m�1 -C

E-in in E-in in

PLA 87� 11 52� 16 62� 1.2 63.5� 0.3

PLA10P 91� 22 – 61.5� 1 –

PLA12.5P 120� 5 107� 5 62.4� 1.4 63.7� 0.8

PLA15P 124� 32 – 58.1� 0.6 –

PLA4S 93� 21 – 61.7� 1.7 –

PLA6S 121� 15 102� 10 62.7� 0.4 62.3� 1.8

PLA8S 141� 20 131� 45 60.7� 1.2 58.7� 1.3

Sukano – 295� 28 – 50.2� 0.1

Polyone – 368� 12 – 52.8� 0.2

146

www.mme-journal.de

R. Scaffaro, M. Morreale, F. Mirabella, F. P. La Mantia

impact energy), it is of fundamental importance to

minimize the presence of defects and lack of homogeneity.

Therefore, the double processing of E-in systems, which

necessarily improves homogeneity, leads also to the higher

impact resistance detected.

Another interesting result is the effect of Sukano and

Polyone, respectively. The neat impact modifiers have

relatively high impact resistance and, inparticular, Polyone

shows higher values than Sukano. However, the blends

containing the latter have higher impact strength than the

ones containing the former. The explanationmay be found

considering that, inagreementwith thepreviously exposed

considerations, Polyone seems to have a slightly poorer

interaction with the matrix, thus it is likely that the

Table 5. Tensile properties of the investigated materials.

Material E

MPa

in E-in

PLA 2 338� 118 2 329� 81 41

PLA4S – 2 229.9� 27

PLA6S 2 237� 19 2 227� 21 47

PLA8S 2 215� 41 2 200� 27 46

PLA10P – 2 194� 35

PLA12.5P 2 294� 110 2 270� 26 45

PLA15P – 2 103� 46

Sukano 1 122� 9 – 14

Polyone 1 109� 5 – >

Macromol. Mater. Eng. 2

� 2011 WILEY-VCH Verlag Gmb

dispersion and the hindrance to crack propagation are

worse than in the case of Sukano.

Some considerations should also be done regarding the

effect ofmodifier’s content on the impact strength. It canbe

observed that Sukano delivers the best increment of the

properties: at the lowest modifier content (4 wt.-% Sukano

vs.10wt.-%Polyone) impactstrengthvaluesaresimilar;but

upon increasing their content, Sukano still delivers

improvements while Polyone substantially leads to a

constant value, which is significantly lower than the one

obtained by Sukano. The measured values of HDT are

reported in Table 4 as well.

The results are not particularly surprising. The differ-

ences between the materials are quite small (confined in a

6 8C range) and, beyond the experimental error, it can be

only stated that the lowest HDT values belong to the

materials containing the highest amounts of impact

modifiers. This is in agreement with the HDT of the neat

impact modifiers, which are the lowest among the whole

set of materials here investigated.

The values of the tensile properties (elastic modulus,

tensile strength, elongationatbreak)are reported inTable5.

With concern to the elastic modulus, there are not

significant differences between the samples obtained by

injection molding and those obtained by extrusion

followed by injection molding. Also, the differences

between the two impact modifiers are small: they reduce

the elastic modulus of the blends reasonably because of

their elastomeric nature.

A few considerations can be done on the results obtained

for tensile strength. With concern to the processing

technique, there are not great variations between the

two. As regards the impact modifiers, Sukano gives, on

average, slightly better results than Polyone; average

amounts seem to be optimal in both cases.

TS EB

MPa %

in E-in in E-in

.2� 0.9 40.7� 0.4 3.2� 0.7 2.9� 0.4

– 47.3� 1.8 – 7.5� 1.5

.6� 2.6 46.2� 2.1 8.4� 2 7.6� 1.4

.3� 1 46.5� 1.1 8.5� 3.4 7.9� 0.8

– 44.6� 0.6 – 7.1� 1.6

.4� 1.4 45� 1 7.8� 1.1 7.7� 0.6

– 43.9� 1.2 – 7.9� 0.7

.3� 0.5 – 14.5� 2.9 –

20.8 – >227 –

011, 296, 141–150

H & Co. KGaA, Weinheim www.MaterialsViews.com

Figure 4. SEM micrographs of (a) PLA-in, (b) PLA-E-in, (c) PLA12.5P-E-in, (d) PLA15P-E-in,(e) PLA6S-E-in, (f) PLA8S-E-in, (g) PLA12.5P-in, (h) PLA6S-in.

www.MaterialsViews.com

Macromol. Mater. Eng. 2011, 296, 141–150

� 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Preparation and Recycling of Plasticized PLA

www.mme-journal.de

Morphological Characterization

Figure 4a–h reports some SEM micro-

graphs of some representative sam-

ples. The first two show, respectively,

PLA-in and PLA-E-in. As predictable,

typical morphologies of glassy poly-

mers can be observed. Micrographs (c)

and (d) (respectively, PLA12.5P-E-in

and PLA15P-E-in) show a biphasic

morphology, with probably elasto-

meric microdomains dispersed in the

glassy matrix. In particular, micro-

graph (c) shows a non-homogeneous

dispersion and a poor interfacial adhe-

sion between the microdomains and

the glassy matrix; however, microdo-

mains size is relatively small.

Micrograph (d) allows observing

that, upon increasing Polyone’s

amount, the average size of the

microdomains increases, probably

because of coalescence phenomena.

This can, in turn explain the behavior

both of the mechanical properties and

the rheological properties (such as the

behavior detected by MFI measure-

ments) observed and therefore, based

on the mechanical properties, the

concentration which should corre-

spond to the best compatibility

between matrix and modifier should

be 12.5 wt.-%.

Micrographs (e) and (f) refer to

Sukano and show, respectively,

PLA6S-E-in and PLA8S-E-in. It can be

observed that no microdomains exist

but, instead, microholes where the

impact modifier accumulates during

phase separation, as already observed

for a similar system.[16] An increase in

themodifier’s amount does not lead to

an increase in the size of these micro-

holes, but just in their number, with-

out significant coalescence phenom-

ena and thus justifying the better

mechanical performance observed.

Finally, micrographs (g) and (h)

(reporting, respectively, PLA12.5P-in

and PLA6S-in) show some materials

which underwent only the injection

molding stage. It can be observed that

thephaseswhicharenotmisciblewith

the PLA matrix are less homoge-

147

Table 6. Flexural and tensile properties of recycled and reference materials.

Cycle Flexural modulus Flexural strength E TS EB

MPa MPa MPa MPa %

PLA8S-E-In 3 420� 10 99.3� 2.3 2 200� 27.5 46.5� 1.1 7.9� 0.8

first recycling 3 971� 11 78� 0.5 2 369� 60 45� 1.8 3.1� 0.2

second recycling 3 920� 14 67.3� 1.5 2 310� 55 47.3� 0.3 3.4� 0.2

third recycling 3 799� 45 61.2� 1.6 2 149� 28 46.2� 0.5 3.3� 0.1

148

www.mme-journal.de

R. Scaffaro, M. Morreale, F. Mirabella, F. P. La Mantia

neously dispersed than in the analogous samples which

underwent also the extrusion stage. This can explain why,

even though extrusion causes some degradation of the

material, extruded and then injection molded samples

showed, on average, the best mechanical properties.

Recycling

Based on of the above reported results, the study of the

effect of recycling on the properties of plasticized PLA was

investigated by means of choosing a reference blend,

‘PLA8S-E-in’.

Table 6 reports the flexural properties (flexural modulus

and flexural strength) of the reference material, recycled

from one to three times, as well as the tensile properties

(elastic modulus, tensile strength and elongation at break).

It can be observed that the reference material shows a

slightly lower modulus than the recycled materials. On the

other hand, flexural strength is significantly higher. This

behavior can be easily explained considering that the

reprocessing led to a partial degradation of the material.

This in turn, caused an increase of the crystallinity, in the

10

100

1000

0001001011Freq [rad/sec]

Visc

osity

[Pa

sec]

0 recycling

1st recycling

2ndrecycling

3rdrecycling

Figure 5. Shear viscosity of reference and recycled materials.

recycled/degraded materials, which

induces stiffening and, therefore, an

increase of the modulus.

Upon increasing the number of

reprocessing, the decrease of the

molecular weight becomes more

and more significant, to such an

extent that the possible positive

effects due to the increase in crystal-

linity are overbalanced by the nega-

tive ones, due to the reduction of

molecular weight. This causes a

module decrease, initially slight after

the second recycling and then quite

significant after the third recycling.

Furthermore, the molecular

weight reduction and the increased

rigidity caused a progressive

decrease of flexural strength upon

reprocessing.

Macromol. Mater. Eng. 2

� 2011 WILEY-VCH Verlag Gmb

The above mentioned reduction was proved and

quantified by rheological tests. Figure 5 reports the flow

curves of the reference and of the recycled materials.

The flow curves unequivocally show a gradual reduction

of the molecular weight, which becomes quite significant

for the last recycling.

As regards the tensile properties, also in this case the

recycled materials prove to be slightly stiffer and, on the

whole, less ductile (in agreement with the increase in

crystallinity, andwith other papers on recycling of polymer

systems[21–23]), confirming the greatly reduced deform-

ability of these recycled materials. The second and third

recyclingdidnot affect theproperties as thefirst didand the

explanation can be easily found considering the develop-

ment of contrasting phenomena (degradation and recrys-

tallization), similarly to flexural properties.

These results become evenmore striking upon analyzing

the impact properties, Table 7. It is quickly evident that the

recycling leads to a suddendrop in the impact strength, as if

the impactmodifierunderwent anearly complete loss of its

properties. This probably happened during the first

recycling operation, since the second recycling caused only

011, 296, 141–150

H & Co. KGaA, Weinheim www.MaterialsViews.com

Table 7. Izod impact strength and HDT values of reference andrecycled materials.

Cycle IS HDT

J �m�1 -C

PLA8S-E-in 141� 20 60.7� 1.2

first recycling 71� 7 68� 1

second recycling 64� 3 69� 1

third recycling 63� 10 66� 1

Table 8. Thermal properties of recycled PLA, as a function ofrecycling number.

Cycle Cooling Second heating

Tc DHc Tcc DHcc Tm DHm

-C J � g�1 -C J � g�1 -C J � g�1

first recycling n/a n/a 102.9 29.7 164.5 40.5

second recycling n/a n/a 99.9 27.2 163.4 44.9

third recycling 115.6 1.4 97.9 25.5 162.9 46.1

Preparation and Recycling of Plasticized PLA

www.mme-journal.de

a slight, further decrease of the impact strength, and the

third recycling almost did not have direct effects on it.

Finally, the values of HDT (Table 7) for the investigated

materials showoncemore that significantdifferencesoccur

between reference and recycled ones. In particular, the

latterproveagain tobestiffer, inagreementwith theresults

of flexural and tensile tests. The HDTs are, in fact, 10–15%

higher than that of the referencematerial. Even in this case,

recrystallization during recycling can easily explain the

observed results. Similar behavior, in fact, was found in

other polyesters [poly(ethylene terephthalate), polyamide-

6] subjected to recycling.[21,22]

In order to demonstrate this hypothesis, which was

proposed as an explanation for several of the results

discussed, DSC analysis was performed on the recycled

materials. Crystallization (Tc) and melting (Tm) tempera-

tures, as well the related enthalpies (DHc and DHm,

respectively) are listed in Table 8.

It can be observed that, upon increasing the number of

recycling operations, metastable crystallization tempera-

ture (aswell as the enthalpy) decreases. As regardsmelting,

0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

0 1 2 3 4 5 6 7 8 9Exposure time [days]

Wat

er a

bsor

ptio

n [w

t%]

1st recycling2nd recycling3rd recycling

Figure 6.Weight percent variation of recycled samples exposed to 50% relative humidityenvironment.

the enthalpy increases, while the tem-

perature decreases. Furthermore, no

crystallization is detected during cooling

for the materials subjected up to two

recycling steps, while a small crystal-

lization peak begins to appear at the

temperature of 115.6 8C. This trends are

in complete agreement with the results

of Pillin et al. on thermomechanical

degradation of neat PLA and, in conjunc-

tion with other studies on the thermal

behaviorofPLA, allowstating thathigher

mobility of the macromolecules occurs,

resulting also in recrystallization phe-

nomena as proposed above.[20,24–27]

Hygroscopic characterization, per-

formed according to the previously dis-

cussedmethodon the recycledmaterials,

led to the absorption curves shown in

Figure 6.

www.MaterialsViews.com

Macromol. Mater. Eng. 2

� 2011 WILEY-VCH Verlag Gmb

Water absorption is slightly lower upon increasing the

recycling number mainly because of the slightly higher

crystallinity achieved by the materials (even though the

difference in the behavior of first and second recycling is

practically negligible). However, the overall weight varia-

tions are quite small.

Conclusion

Several PLA blends with addition of two different impact

modifiers were prepared, in order to find the best

compromise in terms of overall properties and cost.

Rheological, mechanical and morphological characteriza-

tion was carried out. Systems were prepared by using

different amounts of impact modifier, and the influence of

the processing conditions was investigated as well.

Rheological characterization showed that the addition of

impact modifiers reduces the overall viscosity of the

material. Both modifiers used had a certain degree of

incompatibility, PLA matrix and Polyone, in particular.

Polyone-containingblendshaveamoremarkedNewtonian

011, 296, 141–150

H & Co. KGaA, Weinheim149

150

www.mme-journal.de

R. Scaffaro, M. Morreale, F. Mirabella, F. P. La Mantia

behavior at low frequencies compared with Sukano.

However, viscosity decreased upon increasing impact

modifier’s amount in both cases and the processability

was rather similar.

Flexural characterization showed that the addition of

any of the modifiers results in a decrease of the flexural

modulus and an increase of the flexural strength, however

the differences between the different modifiers were not

particularly significant.

Impact tests pointed out that significant higher strength

can be achieved by using 8 wt.-% Sukano and performing a

separate extrusion step before injection molding; HDT and

tensile tests confirmed the general trends.

Morphological characterization showed a biphasic nat-

ure of the blend in both cases, with different appearance of

the domains; Sukano underwent a more homogeneous

dispersion in the PLA matrix.

As regards the recyclability, the results clearly pointed

out that the properties are affected by the change of

crystallinity correlated with the decreased molecular

weight, due to reprocessing. As a consequence, the recycled

materials were stiffer, with reduced deformability, impact

resistance and water absorption if compared to the

reference one.

Received: June 8, 2010; Revised: September 22, 2010; Publishedonline: November 22, 2010; DOI: 10.1002/mame.201000221

Keywords: differential scanning calorimetry (DSC); mechanicalproperties; PLA; recycling; rheology

[1] R. Auras, B. Harte, S. Selke, Macromol. Biosci. 2004, 4, 835.[2] A. Sodergard, M. Stolt, Prog. Polym. Sci. 2002, 27, 1123.

Macromol. Mater. Eng. 2

� 2011 WILEY-VCH Verlag Gmb

[3] L. T. Lim, R. Auras, M. Rubino, Prog. Polym. Sci. 2008, 33, 820.[4] L. Jiang, J. Zhang, M. P. Wolcott, Polymer 2007, 48, 7632.[5] D. Garlotta, J. Polym. Environ. 2001, 9, 63.[6] J. Y. Nam, S. S. Ray, M. Okamoto, Macromolecules 2003, 36,

7126.[7] N. Lopez-Rodrıguez, A. Lopez-Arraiza, E. Meaurio, J. R. Sarasua,

Polym. Eng. Sci. 2006, 46, 1299.[8] M. Shibata, Y. Inoue, M. Miyoshi, Polymer 2006, 47, 3557.[9] L. Jiang, M. P. Wolcott, J. Zhang, Biomacromolecules. 2006, 7,

199.[10] J. Zhang, L. Jiang, M. P. Wolcott, Int. SAMPE Technical Con-

ference, Seattle, WA 2005.[11] R. Scaffaro, F. P. La Mantia, R. Bertani, A. Sassi, Macromol.

Symp. 2003, 202, 67.[12] T. L. Dimitrova, F. P. La Mantia, F. Pilati, M. Toselli, A. Valenza,

A. Visco, Polymer 2000, 41, 4817.[13] F. P. La Mantia, Polym. Degrad. Stab. 1993, 42, 213.[14] A. J. Nijenhuis, E. Colstee, D. W. Grijpma, A. J. Pennings,

Polymer 1996, 37, 5849.[15] K. S. Anderson, S. H. Lim, M. A. Hillmyer, J. Appl. Polym. Sci.

2003, 89, 3757.[16] E. Piorkowska, Z. Kulinski, A. Galeski, R. Masirek, Polymer

2006, 47, 7178.[17] M. Hiljanen-Vainio, J. Kilma, K. Hiltunen, J. V. Seppala, J. Appl.

Polym. Sci. 1997, 63, 1335.[18] T. Ryynanen, A. Nykanen, J. V. Seppala, eXPR. Polym. Lett.

2008, 2, 184.[19] F. Byrne, P. G. Ward, J. Kennedy, N. Imaz, D. Hughes, D. P.

Dowling, J. Polym. Environ. 2009, 17, 28.[20] I. Pillin, N. Montrelay, A. Bourmaud, Y. Grohens, Polym.

Degrad. Stab. 2008, 93, 321.[21] F. P. La Mantia, D. Curto, R. Scaffaro, J. Appl. Polym. Sci. 2002,

86, 1899.[22] F. P. La Mantia, M. Vinci, Polym. Degrad. Stab. 1994, 45,

121.[23] M. Marrone, F. P. La Mantia, Polym. Rec. 1996, 2, 17.[24] Z. Kulinski, E. Piorkowska, Polymer 2005, 46, 10290.[25] T. Miyata, T. Masuko, Polymer 1998, 39, 5515.[26] M. L. Di Lorenzo, Eur. Polym. J. 2005, 41, 569.[27] M. Salmeron Sanchez, J. L. Gomez Ribelles, F. Hernandez

Sanchez, J. F. Mano, Termochim. Acta 2005, 430, 201.

011, 296, 141–150

H & Co. KGaA, Weinheim www.MaterialsViews.com