Introduction

Polylactide or polylactic acid (PLA) is currently receiving considerable attention for conventional utilization such as packaging materials as well as production of fibres, and more recently, as (nano)composites for technical applications. In competition with petroleum-based polymers and showing a principal position on the market of bio-based polymers, PLA is undoubtedly one of the most promising candidates for further developments owing of its additional biodegradability [1–6].

Furthermore, very recent trends highlight the huge potential of PLA-based products with higher added value for engineering applications, i.e., in transportation, electronic and electrical devices, mechanical and automotive parts, etc. Consequently and to reach the end-user demands, the profile of PLA properties can be tuned up by combining PLA with different dispersed phases: micro- and nano-fillers, flame retardants, impact modifiers, plasticizers and other additives [3, 4, 7–15].

Polymer nanocomposites (PNC) with improved properties (stiffness, thermal stability, fire retardancy (FR), antistatic to conductive electrical characteristics, lower permeability, crystallization enhancement, anti-UV and antibacterial action, etc.) have been produced by combining PLA with various nanofillers such as organo-modified layered silicates (OMLS), carbon nanotubes (CNT), graphite derivatives, polyhedral oligomeric silsesquioxanes (POSS), zinc oxide, etc. [7, 8, 16–24].

Halloysite nanotubes (HNT) have recently received considerable attention as a new type of nanofiller for enhancing the mechanical, thermal, crystallization, fire and other specific properties of different polymers like PA6, PA12, PP, PS and epoxy resins, etc. [25–34].

As reported elsewhere [25–27, 31], the composition of HNT is similar to kaolin, whereas taking into account the state of hydration, the ideal chemical formula can be expressed as Al2Si2O5(OH)4 × nH2O, where n (representing the hydrated and dehydrated forms) equals 2 and 0, respectively for HNT and kaolin. The tubular structure of HNT is believed to be the result of hydrothermal alteration or surface weathering of aluminosilicate minerals. Due to several characteristic features such as nanoscale lumen, relatively high length-to-diameter ratio, and low hydroxyl group density on their surface, HNT are considered as promising competitors and cheaper alternatives to both CNT and OMLS. Conventional nanoclays require exfoliation to separate the layers and to obtain the adequate dispersion needed for uniform properties in PNC whereas HNT do not require exfoliation, making possible the production of stronger, lighter materials without the complexity and processing cost associated with intercalation and exfoliation [34, 35].

Unfortunately and compared to the extensive evaluation of OMLS and CNT as nanofillers for PLA, investigations on PLA/HNT nanocomposites appear to be missing or have not been conducted so far to evidence the overall effect of HNT addition on main PLA properties. In our opinion, the interest for HNT is rather new, and more likely represents the main reason explaining the lack of studies concerning the production and characterization of PLA-HNT nanocomposites. Generally, the addition of nanofillers can provide PLA with relevant specific properties but sometimes also triggers problems such as loss of mechanical and thermal characteristics, degradation of this water-sensitive polyester matrix, which are important aspects to be considered when targeting a potential application.

In response to the demand for new PLA grades characterized by improved characteristic features, HNT were evaluated as potential new nanofiller for PLA. Commercially available HNT (or additionally silane treated) were added into a PLA matrix (extrusion grade) via melt-compounding technology and the resulting nanocomposites were characterized in detail for highlighting their performances. Surface treatment of HNT was also studied via silanization reaction in order to improve the nanofiller dispersion within PLA.

Experimental

Materials

Poly(l,l-lactide)—hereafter called PLA, was kindly supplied by NatureWorks LLC. Characteristics of PLA are as follows: number average molecular weight, Mn(PLA) = 88,500, index of polydispersity, Mw/Mn = 1.8, d isomer content <2 %.

Halloysite nanotubes (referred below as “HNT”) were kindly supplied by NaturalNano, Inc. (USA) as HNT(QM) (HNT milled and treated). Following the information presented elsewhere [36, 37], the organic modifier used to surface-treat these HNT(QM) is a quaternary ammonium chloride salt [31], in amount less than 5 % according to the technical sheet of the product [38].

A selected silane, i.e., 3-(Trimethoxysilyl) propyl methacrylate (98 %), supplier Acros Organics (called below TMSPM), was used for the additional treatment of the nanofiller. The silane-treated nanofiller will be reported as HNT(s).

Production of PLA Nanocomposites

Before processing by melt-blending, PLA was dried overnight at 80 °C under vacuum. To minimize the water content for melt-blending with PLA, the nanofiller was dried at 105 °C for 48 h and then used directly for melt-compounding. The nanofiller was mixed together with PLA at 200 °C under moderate mixing (cam blades) by using a Brabender bench scale kneader (50EHT model) following a specific procedure: 3 min premixing at 30 rpm- speed in order to avoid an excessive increase of the torque during melting of PLA, followed by 7 min mixing at 70 rpm. For the sake of comparison, pristine PLA was processed under similar conditions of melt-compounding. Throughout this report, all percentages are given as wt %.

In a second step, PLA-HNT(s) nanocomposites containing up to 12 % HNT(s) were obtained after the previous treatment of the nanofiller (HNT(QM)) with a selected silane (TMSPM). Firstly, HNT were treated with 3 % silane (from a solution of 1 % TMSPM in methanol/demineralized water (9/1)) following the slurry method reported by Sadler and Vecere [39]. The chemical treatment of the nanofiller surface was followed by evaporation of the solvents (i.e., water and alcohol) and finally drying under vacuum at 80 °C.

Plates (~3 mm thick) for mechanical characterizations were produced by compression molding at 190 °C by using an Agila PE20 hydraulic press. More specifically, the material was first pressed at low pressure for 240 s (3 degassing cycles), followed by a high-pressure cycle at 150 bars for 150 s. The resulted samples were then cooled under pressure (50 bars). Specimens for tensile and Izod impact tests were obtained from plates by using a milling-machine in accordance with ASTM D 638-02a (specimens type V) and ASTM D 256-A norm (specimens 60 × 10 × ~3 mm3), respectively.

Characterization

Thermogravimetric Analyses (TGA)

TGA were performed by using a TGA Q50 (TA Instruments) with a heating ramp of 20 °C/min under air flow, from room temperature up to 600 °C (platinum pan, 60 cm3/min air flow rate).

Differential Scanning Calorimetry (DSC)

DSC measurements were performed by using a DSC Q200 from TA Instruments under nitrogen flow. The samples obtained by compression molding were investigated. The procedure was as follows: first heating scan at 10 °C/min from 0 up to 200 °C, isotherm at this temperature for 2 min, then cooling down to −20 °C at 10 °C/min and finally, second heating scan from −20 to 200 °C at 10 °C/min. The first scan was meant to erase the anterior thermal history of the samples. The events of interest, i.e., the glass transition temperature (Tg), cold crystallization temperature (Tc), enthalpy of cold crystallization (ΔHc), temperature and enthalpy of polymer chain rearrangement (Tr and ΔHr, respectively), melting temperature (Tm) and melting enthalpy (ΔHm) were evaluated. The degree of crystallinity (χc) was determined by subtracting ΔHc and ΔHr (if it was evidenced) from ΔHm, and by considering a melting enthalpy of 93 J/g for 100 % crystalline PLA.

Screening analyses to have information about the effect of HNT addition on the crystallization kinetics of PLA were performed through determination of crystallization half-time (t1/2) during isothermal crystallization. Using DSC, the PLA samples were heated to 200 °C at a rate of 10 °C/min, held 2 min at this temperature to erase their thermal history, step followed by high speed-cooling (rate of about 40 °C/min) to the iso-crystallization temperature of interest (110 or 120 °C) and maintained under isothermal conditions for up to 60 min.

Mechanical Testing Measurements

Tensile testing measurements were performed by using a Lloyd LR 10 K tensile bench in accordance to the ASTM D 638-02a norm at a speed rate of 1 mm/min using a distance of 25.4 mm between grips. Notched impact strength (Izod) measurements were performed by using a Ray-Ran 2500 pendulum impact tester and a Ray-Ran 1900 notching apparatus, in accordance to the ASTM D 256 norm (method A, 3.46 m/s impact speed, 0.668 kg hammer).

All mechanical tests were carried out by using specimens previously conditioned for at least 48 h at 20 ± 2 °C under a relative humidity of 50 ± 3 % and the values were averaged out over five measurements.

Dynamic Mechanical Analysis (DMA)

DMA on selected samples (specimens of 60 × 12 × 2 mm3 performed by injection molding at 200 °C, DSM micro-injection molding machine) were carried out using DMA 2980 apparatus (TA Instruments) in a dual-cantilever bending mode. To avoid a different thermal history after the processing by injection molding or an additional crystallization process during testing, the specimens were annealed under similar conditions (at 110 °C under vacuum for 90 min). The dynamic storage and loss moduli (E′ and E″, respectively) were determined at a constant frequency of 1 Hz as a function of temperature from 0 to 120 °C, at a heating rate of 3 °C/min.

Transmission Electron Microscopy (TEM)

Transmission electron micrographs of PLA/HNT nanocomposites were obtained with a Philips CM200 apparatus using an accelerator voltage of up to 120 kV. The nanocomposite samples (70–80 nm thick) were prepared with a Leica Ultracut UCT ultracryomicrotome by cutting at −100 °C. Reported microphotographs represent typical morphologies as observed at, at least, three different locations of the sample.

Results and Discussion

HNT as Nanofiller for PLA

Firstly, it is worth mentioning that HNT from different sources can vary in the level of hydration, morphology (dimension of inner and outer diameter, length, length-to-diameter ratio) and colour, depending on the substitutional metals and on mineral origin [25, 30, 31]. However, TEM images at different magnifications (Fig. 1a, b) of as received HNT(QM) assess the tubular morphology of HNT (outer tube diameter of 50–60 nm and lumen size of 20–30 nm, Fig. 1b) as well as the presence of nanotubes of diverse length, ranging typically from hundreds of nanometres to above 1.5 μm. Noteworthy, the presence of nanotubes of various lengths in the TEM images (Fig. 1a) is mainly ascribed by us to the milling process performed by the industrial supplier, which can affect, more or less, the length-to-diameter ratio of HNT.

Fig. 1
figure 1

a, b TEM images at different magnifications of as received nanotubes (HNT(QM))

Full size image

On the other hand, in the perspective of PNC preparation, it is worth recalling that PLA is very sensitive to hydrolysis, temperature, shear, etc. during melt-processing, so all precautions known in the state of the art should be applied to limit polyester chains degradation.

From TGA of as received HNT(QM) (Fig. 2) it comes out that the weight loss during thermal heating (20 °C/min) of the as received nanofiller occurs in three distinguishable regions:

Fig. 2
figure 2

TGA of as received HNT(QM) (under air, 20 °C/min)

Full size image
  1. 1.

    Below 100 °C, indicating the presence of absorbed free water (about 1.3 %).

This water will be removed by drying before mixing with PLA, but further precaution it is required to avoid the contact of the nanofiller with the atmospheric moisture.

  1. 2.

    A weight loss of ~2.4 % in the interval of temperature between 200 and 400 °C (d-TG shows a maximum above 260 °C), ascribed to the loss of hydration water molecules that can be present between the layers of HNT and to degradation of the organomodifier product.

Based on the different sources of information, i.e., publications and patents, it is reported that the HNT can be surface treated with quaternary ammonium chloride salts [31, 37, 38]. However, from TGA results it comes out that the nanofiller is containing a relatively low amount of hydrated water (a previously drying process carried out by the supplier is assumed), whereas the dispersing agent used as organomodifier is in evident lower amount by comparing to the commercial OMLS that are often modified with 20–40 % quaternary ammonium salts [7, 8, 40].

  1. 3.

    Dehydroxylation at higher temperature than 400 °C (weight loss above 12 %, with a maximum on d-TG curve at about 497 °C).

Furthermore, these data are consistent with the mechanisms and results of different studies reported elsewhere connected to HNT and their utilization [26, 34, 41].

To minimize PLA degradation by hydrolysis, the nanofiller was dried at 105 °C for 48 h. Besides, it was presumed by us that a higher temperature of drying, e.g., 200 °C, could lead to thermal degradation of the organomodifier of HNT(QM), as reported in the case of OMLS [40].

As far as the additional treatment by silanization is concerned, various silane agents have been reported for modifying the surface of HNT: γ-methacryloxypropyltrimethoxysilane [29], aminoethylaminopropyltrimethoxysilane [34], γ-aminopropyltriethoxysilane [42].

HNT contain two types of hydroxyl groups, inner and outer hydroxyl groups, which are located between the layers and on the surface of the nanotubes, respectively [25]. Following the chemical mechanism of silanization, hydrolysis of the alkoxy groups into silanols first occurs, which can react with the hydroxyl groups of the inorganic HNT and eliminate water. Furthermore, it is assumed that Si–O–Si layers can be formed by reaction of silane molecules with each other to give a multimolecular structure of bound silane coupling agents [43] on the surface of filler (here, HNT) that will behave as interfacial zone, effectively playing a compatibilizing role. Accordingly, in our experiments, the HNT were additionally treated with 3 % silane agents functionalized with a methacryloxy moiety, i.e., TMSPM, which proved high efficiency in terms of HNT(s) dispersion in PLA (vide infra).

Properties of PLA-HNT(QM) Nanocomposites

Morphology

Information concerning the morphology of nanocomposites was mainly obtained through TEM analyses. Figure 3a, b and c, d show selected pictures of the nanocomposites produced by addition into PLA of 6 and 12 % HNT(QM), respectively. It is obvious from the TEM images recorded at lower magnification (Fig. 3a, c) that a good quality of the distribution/dispersion of the HNT(QM) is reached throughout the polyester matrix even though the melt-mixing was performed under moderate shear (internal kneader). However, some small aggregates (identified by white circles in Fig. 3c) were evidenced in the TEM pictures containing the higher amount of HNT(QM), i.e., 12 % HNT.

Fig. 3
figure 3

a–d TEM pictures at different magnifications of PLA-HNT(QM) nanocomposites containing 6 % (a, b) and 12 % (c, d) halloysite nanotubes

Full size image

Besides, the TEM images at higher magnification (Fig. 3b, d) attest for the good dispersion of HNT(QM) through PLA where quite well-individualized nanotubes are typically observed. This may be explained as in the case of PP-HNT(QM) nanocomposites [31] by the presence of the organomodifier at the nanotube surface, which decreases the surface free energy and nanotube/nanotube interactions, thereby breaking up the formation of aggregates during the melt-blending process.

Following the TEM observations, HNT(QM) thus appear easily dispersible into PLA, and as reported elsewhere [25, 31], the limited intertubular contact area and rod-like geometry, associated with the surface treatment, can be considered as key-factors that explain their good dispersion. In the same context, it is also worth mentioning that a good dispersion of HNT was also identified in some other polar polymers such as PA6 or PA12 [25, 44].

Thermal Properties

To highlight the effect of HNT(QM) on PLA thermal stability, TG measurements performed on the nanocomposites have been compared to those of pristine PLA. For simplification and easier interpretation, Table 1 summarizes the values of the temperature for 5 % weight loss (T5%) and of the temperatures corresponding to the maximum rate of thermal degradation (TD—from d-TG). It is of interest to note that T5% is often considered as the initial decomposition temperature. From the data presented in Table 1, it comes out that the nanocomposites show comparable thermal stability (T5% as parameter) and only a slight decrease in TD is recorded with respect to the processed unfilled PLA. The decrease of TD at higher loading of nanotubes is mainly assigned to the residual water of hydration and to the degradation of ammonium organomodifier (see “HNT as Nanofiller for PLA” section), without totally excluding some additional effect of metallic impurities naturally contained in clays, which can also trigger faster degradation of the polyester matrix. Additionally, as suggested in previous studies, since HNT display lower aspect ratio and specific surface area than most of the OMLS, it is reasonable to assume that the barrier effect of HNT with tubular structure may somewhat be inferior to those of layered silicates [29, 44].

Table 1 TGA data of pristine PLA and PLA-HNT(QM) nanocomposites (under air flow, 20 °C/min)
Full size table

Figure 4 shows representative DSC curves recorded during the second heating scan of PLA-HNT(QM) nanocomposites compared to those of pristine PLA. Firstly, from these data it comes out that the addition of nanofiller does not modify significantly the Tg values, which remain in the range 61–63 °C, and Tm values of PLA matrix (Tm ≈ 163–170 °C). The multiple Tm peaks or presence of shoulders on DSC curves, can be ascribed to the melting of crystalline regions of various size and perfection formed during cooling and crystallization processes, without excluding the influence of additional factors that can affect the melting of PLA (different crystalline structures, presence of fractions of low molecular weight, the chemical treatment and nanofiller loading, the effect of thermo-mechanical processing and time–temperature history, etc.) [7]. However, the recrystallization and the melting can be competitive in the heating process, whereas the double-melting behaviour is attributed to the slow rates of PLA crystallization and recrystallization [45]. Furthermore, the lower temperature peak is generally associated to the melting of the small crystals produced by the secondary crystallization, and the peak recorded at higher temperature, to the melting of the major crystals formed in the primary crystallization process [46].

Fig. 4
figure 4

DSC traces obtained during second heating scan (10 °C/min) of PLA and PLA-HNT(QM) nanocomposites

Full size image

DSC also reveals that the polymer matrix undergoes cold crystallization, while Tc in nanocomposites decreases in correlation with the percentage of HNT. Furthermore, there is no clear changes of χc recorded in non-isothermal crystallization (rate of cooling of 10 °C/min) meaning that HNT(QM) do not seem to behave as efficient nucleating agents of PLA crystallization like in the case of other PNC containing the same nanofiller [25, 30, 31, 44]. However, it is assumed that due to a relatively high cooling rate (10 °C/min) for a polymer characterized by low crystallization ability, the χc of PLA in nanocomposites is remaining limited, to values in the range of 3–4 %, somewhat comparable to those recorded for the pristine PLA (2.7 %).

Thermo-Mechanical Properties

In general, addition of nanofillers into a polymer matrix is expected to result in improved properties, such as better tensile and flexural strength, higher modulus, dimensional stability, higher heat distortion temperature, etc. According to previous investigations, by adding HNT in different polymers (PA6, PA12, PBT, PP, etc.) the flexural strength, modulus, tensile and impact strength can be improved [25, 27, 28, 31, 44]. The reinforcing effect of HNT on different polymers, frequently observed at low loadings, is generally ascribed to the rod-like and high aspect ratio structure, whereas their interfacial properties can play a key-role as well.

Figure 5 shows the evolution of both tensile strength and rigidity (Young’s modulus) of the nanocomposites in function of the nanofiller content. Comparison to pristine PLA is also given. On one hand, by comparison to neat PLA, both rigidity and stress (given as maximum tensile strength) are gradually improved in nanocomposites by increasing the percentage of nanofiller up to 6 %. However, at higher loading, the rigidity is maintained at high level while no further increase in tensile strength is observed. Such an observation is generally attributed to somewhat inevitable aggregation of the nanofiller at higher loading. In fact this assumption was obviously confirmed by additional experiments to assess the production of PLA composites filled with 20–30 % HNT(QM), which revealed a dramatic decrease in tensile strength (results not discussed here).

Fig. 5
figure 5

Evolution of tensile strength and Young’s modulus of PLA and PLA–HNT(QM) nanocomposites with different content of nanofiller

Full size image

The PLA nanocomposites containing up to 12 % HNT(QM) revealed similar nominal strain at break (εb) as the pristine polyester matrix (values of about 5 %), whereas the Izod impact strength (see Fig. 9) is not decreased, with values comparable to neat PLA (2.8 kJ/m2), at least for the investigated contents in nanofiller (from 3 to 12 %). By comparing to other nanofillers that lead to PLA nanocomposites with improved rigidity, mostly in the detriment of the decrease of εb and impact strength (e.g., expanded graphite [21]), this behaviour is quite remarkable and is commonly ascribed to the intrinsic stiffness of the nanotubes- resulting from their tubular structure, their good dispersion, as well as to the good interfacial properties between filler and matrix [25, 27, 28, 44]. The excellent tensile strength (about 70 MPa for a nanocomposite containing 6 % HNT(QM)), corroborated with the increase of rigidity and good level of impact properties, suggests that the PLA-HNT composites can be potentially interesting in engineering applications.

Figure 6a, b show the temperature dependencies of storage (E′) and loss (E″) moduli of pristine PLA and nanocomposites obtained by addition of 3–12 % HNT(QM). For the sake of clarity, only one representative curve as the average of minimum two measurements is shown for each sample.

Fig. 6
figure 6

a, b Dependencies of E′ a and E″ b versus temperature: pristine PLA with respect to PLA-HNT(QM) nanocomposites

Full size image

At low temperature the values of E′ decreased gradually in a similar way (Fig. 6a) for all the samples. However, in correlation to nanofiller content, E′ increases distinctly for the nanocomposite samples, e.g., at 20 °C the value of E′ is 3,400 MPa for neat PLA, 4,100 and 4,600 MPa for PLA containing 6 and 12 %, HNT(QM), respectively. In addition, in relation to the potential interest for the utilization of these nanocomposites in mechanical applications and at higher temperature, it is worth mentioning that the value of E′ of 3,000 MPa is obtained at 50.2 °C for pristine PLA with respect to 65.9 and 68.9 °C for nanocomposites with 6 and 12 % HNT(QM), respectively (Table 2).

Table 2 Comparative DMA data of PLA and PLA nanocomposites containing HNT(QM) and HNT(s)
Full size table

DMA was also used for investigation of the loss modulus (E″) as function of temperature (Fig. 6b).

The increase of E″ with the filler content can be attributed mainly to the contribution of the mechanical loss generated at the interface regions between nanotubes and PLA matrix [47]. However, in agreement with the DSC results, the data reveal that there are no significant differences between the temperatures corresponding to the maximum of E″ (~73 °C) of PLA and those of PLA-HNT(QM) nanocomposites. Furthermore, it should be also noted that DMA results from the viscoelastic behaviour are correlated with the mechanical properties, suggesting the reinforcing effect of nanofiller. Additionally, the reinforcement factor calculated following the procedure reported elsewhere [48] is improved by addition of nanofiller, noteworthy not only in the glassy state, but also above the Tg of PLA.

As preliminary conclusion, interesting mechanical properties have been recorded at relatively low amount of nanofiller (3–6 % HNT(QM)), which can be ascribed to the reinforcing effect of the halloysite nanotubes that proved to be finely distributed and dispersed within the PLA matrix, at least at relative content up to 6 %. As demonstrated by DMA (dynamic solicitation), melt-blending PLA with HNT leads to the enhancement of the storage modulus (E′) and offers the possibility to use PLA in applications requiring rigidity at higher temperatures of utilization.

PLA-HNT(s) Nanocomposites

In order to further increase the dispersion ability of the inorganic nanotubes in PLA and increase the affinity between the dispersed HNT and the polyester chains, additional treatment of HNT(QM) with silane (TMSPM) has been performed (see experimental). The so-prepared silanized halloysite nanotubes, coined as HNT(s), were then dispersed in PLA by melt-processing for producing PLA nanocomposites containing up to 12 % HNT(s). The main objective of this section is to discuss the most important aspects that can be associated to the effectiveness of silane treatment and to the key-performances of the related PLA-HNT(s) nanocomposites.

Thermal Properties and Crystallinity

First of all, Fig. 7 shows the comparative TG curves of HNT(s), those of PLA and the corresponding nanocomposites. On one hand, HNT(s) are characterized by a weight lost above 16 % up to a temperature of 550 °C, whereas the maximum on d-TG curve is slightly shifted to 504 °C. On the other hand, the addition of HNT(s) into PLA is leading once more to nanocomposites characterized by good thermal stability (T5% in the range 334–337 °C, TD above 370 °C) quite similar with the neat PLA.

Fig. 7
figure 7

TGA of HNT(s), PLA and PLA-HNT(s) nanocomposites (under air, 20 °C/min)

Full size image

Moreover, the second DSC heating scan (non-isothermal crystallization, after previously cooling with a rate of 10 °C/min) of PLA-HNT(s) samples did not attest significant modifications (increases of χc, changes of Tg or Tm, etc.- results not shown here), with respect to the pristine PLA and PNCs containing HNT(QM). Interestingly enough, the most important differences concern the values of χc as they are revealed by the first DSC heating scan, which typically assesses the thermal characteristics of the specimens used for mechanical testing. For illustration (Table 3), the PNC filled with 3–12 % HNT(s) are typically characterized after the compression molding process by a higher χc, i.e., values in the range 16–24 % by comparing to less than 15 % in the case of nanocomposites containing similar loadings of HNT(QM), or to pristine PLA, that practically shows a χc of about 10 %.

Table 3 Comparative DSC data (first heating, 10 °C/min) of PLA and PLA nanocomposites containing HNT(QM) and HNT(s) (samples after the compression molding process)
Full size table

It is important to remind that the use of PLA in technical applications (e.g., injection molded products) remains limited because PLA is known for a relatively too slow crystallization rate when it is compared with many other semi-crystalline thermoplastics. A measure of the effectiveness of different nanofillers as nucleating agents is the crystallization half-time (t1/2) as recorded in isothermal crystallization experiments. The effects of HNT(s) with respect to HNT(QM) on the crystallization kinetics of PLA have been compared through t1/2 recorded during isothermal crystallization experiments at 110 or 120 °C [21]. The relative crystallinity was calculated by integrating for each exotherm the total area under the curve. The t1/2 was taken as the time at which the relative crystallinity (area) was equal to 50 %. Table 4 shows the comparative crystallization half-time of PLA nanocomposites containing HNT(QM) and HNT(s).

Table 4 Crystallization half-time (t1/2, expressed in minutes) of PLA (processed) and PLA nanocomposites containing HNT(QM) and HNT(s)
Full size table

In relation to the addition of HNT(QM), at the exception of the decrease of t1/2 specifically recorded for the nanocomposite containing 3 % nanofiller, e.g., at 110 °C from 6.6 to 3.8 min (processed PLA compared to nanocomposite, respectively), the presence of HNT(QM) did not allow for reducing the t1/2 values. Noteworthy, these results are supported by the experimental evidence and direct measurements of crystallinity on the specimens obtained by compression molding (the first DSC heating scan, Table 3), the nanocomposite containing 3 % HNT(QM) being characterized by a χc of 15.1 %, somewhat higher than those recorded for pristine PLA (10.2 %).

Furthermore, the improved crystallinity of PLA-HNT(s) nanocomposites (Table 3) is in agreement with the determination of t1/2 during isothermal crystallization tests at 110 or 120 °C (Table 4) that suggest powerful nucleation ability and higher crystallization kinetics for HNT(s). To explain the higher crystallization (nucleating) effect recorded in presence of HNT(s), several factors should be taken into account such as a better dispersion of nanofiller and more nucleating sites, modification of the interfacial properties by presence of the organomodifier at the nanotube surface. However, it is believed that other techniques, e.g., the polarized optical microscopy (POM) and X-ray diffraction (XRD) can provide additional answers and supplementary information [30].

Morphology and Mechanical Properties

TEM images of PLA-HNT(s) samples at different magnifications reveal that, following the treatment with the silane agent, the dispersion of the nanofiller within PLA matrix appears to be improved (Fig. 8a, b) even at high loadings, i.e., 12 % HNT(s). Overall microscopic analysis indicates well distributed and homogenously dispersed HNT(s) throughout PLA matrix, by considering that the number of clusters/aggregates that can be evidenced in TEM images is lower.

Fig. 8
figure 8

a, b TEM pictures at different magnifications of PLA-12 % HNT(s) nanocomposites

Full size image

Concerning the mechanical properties, as already specified, on one hand, DMA does not confirm additional improvements using HNT(s) (see Table 2). It is important to remind that for these evaluations the possible PLA crystallization during DMA testing, at a relatively low heating rate (3 °C/min), and the differences of crystallinity between the injection-moulded samples were purposely decreased by a previous thermal treatment (at 110 °C, during 90 min, see experimental section).

On the other hand, the values of tensile strength and rigidity are comparable to those obtained using HNT(QM). For instance, it is worth mentioning that addition of 3–6 % HNT(s) yields nanocomposites characterized by a tensile strength of 65 MPa, whereas the modulus of elasticity (2,750 MPa at 6 % HNT(s)) is equivalent to those of PNCs containing HNT(QM). In the same context, it is also important to precise that similarly to HNT(QM), addition of HNT(s) is leading to nanocomposites that do not show the decrease of εb, e.g., the εb of PLA- 12 % HNT(s) nanocomposite is 6.6 % by comparing to 4.9 %, value obtained for the pristine PLA.

Moreover, as illustrated in Fig. 9, the mechanical testing highlights some improvement for the Izod impact strength (3.5 kJ/m2 for a nanocomposite containing 12 % HNT(s), only 2.8 kJ/m2 for pristine PLA), which can be once again connected to the effectiveness of HNT(s). The enhancement in impact strength is generally ascribed to the fact that the nanotubes play an important role in hindering the cracks induced by impact solicitation or can provoke a bridging, to the absorption of energy by nanotube breakage, to a better interfacial interaction leading to higher absorption of energy during impact deformation and to plastic deformation around nanotubes [25, 27, 31, 49]. In the same context, it is believed that additional factors can explain these improvements: a relatively better dispersion and the decrease of number of agglomerates, assumption which is consistent with TEM investigations, the surface modification and compatibilization with a functionalized silane, without excluding that the crystallization of PLA matrix can play a significant role in toughening as well [50].

Fig. 9
figure 9

Impact strength (Izod) of PLA and PLA nanocomposites with different content of HNT(QM) and HNT(s)

Full size image

Finally, from these results it comes out that addition of HNT(s) is leading to the preservation of PLA thermal stability and evident improvements in the crystallization rate, better impact strength, good rigidity and tensile strength (but not additional increase), correlated to optimistic dispersion through PLA matrix. By comparing to both OMLS and CNT, HNT can represent a good alternative as nanofillers for PLA. Overall testing data suggest that a combination of mechanical strength, rigidity and toughness, associated to a better processing due to a higher speed of crystallization can be obtained by addition of HNT. Furthermore, improvements in melt-mixing (e.g., using twin-screw extruders), compatibilizing and the utilization of co-additives, are expected to lead to nanocomposites with improved properties.

Conclusion

HNT were recently considered as nanofillers for the production of PNCs as possible alternatives to both OMLS and CNT. This first study demonstrates that it is possible to obtain competitive nanocomposites using HNT and PLA via melt-blending technology. Halloysite, a specific grade (HNT(QM)), has been investigated as potential nanofiller for PLA and the resulting nanocomposites were characterized for highlighting their performances. Following TEM observations, the nanofiller is easily dispersible into PLA matrix, whereas the thermal analyses did not evidence any significant decrease in PLA stability (TGA) or modification of the main thermal parameters (DSC). Their mechanical characterization revealed high tensile strength and Young’s modulus (e.g., at 6 % HNT(QM), 70 and 2,800 MPa, respectively), no decrease of εb, and good impact properties, at least for a content of (3–12) % nanofiller. As demonstrated by DMA, melt-blending PLA with HNT leads to the enhancement of the storage modulus (E′) and offers the possibility to use PLA in applications requiring higher temperature of utilization.

Moreover, the additional nanofiller surface-treatment using 3-methacryloxypropyltrimethoxy silane leads to nanocomposites showing better morphology at higher nanofiller contents, thermal stability and good tensile stress characteristics, associated to higher impact strength (3.5 kJ/m2 by addition of 12 % HNT(s), with respect to 2.8 kJ/m2 for pristine PLA). Furthermore, by using HNT(s) the degree of crystallinity (χc) of test specimens was found higher, results consistent with the improved PLA nucleating ability and higher rate of crystallization as evidenced by determining the half-crystallization time. By considering the good dispersion into PLA matrix, the panel of mechanical performances due to a significant reinforcing effect of HNT, it is assumed that for targeted uses these nanocomposites may constitute interesting materials for engineering applications.