Elsevier

Polymer

Volume 207, 20 October 2020, 122882
Polymer

On the use of solubility parameters to investigate phase separation-morphology-mechanical behavior relationships of TPU

https://doi.org/10.1016/j.polymer.2020.122882Get rights and content

Highlights

  • Segmented thermoplastic polyurethanes are synthesized from a biobased polyol.

  • Hansen Solubility Parameters (HSP) can be used to predict phase compatibility.

  • Microscopic observations and thermomechanical behavior confirm material morphology.

  • The soft continuous phase contains polyol chains and short miscible hard segments.

  • Dispersed or continuous hard domains provide a thermoplastic elastomer behavior.

Abstract

This study proposes a thorough investigation, especially based on thermodynamics, to predict phase separation in a linear thermoplastic polyurethane, denoted TPU, prepared from fatty acid-based soft segments and MDI (4,4′-methylene bis(phenyl isocyanate))/BDO (1,4-butanediol) hard segments and specially designed for bitumen modification. Hansen’ solubility parameters (HSP) of both segments are evaluated to predict their compatibility. The later ones are evaluated either individually from the corresponding segment synthesized separately or from a decomposition of the solubility diagram of the TPUs into two distinct spheres. In a second step, phase separation is experimentally analyzed by combining differential scanning calorimetry, microscopy techniques, and small angle X-ray scattering (SAXS). The microstructure of the TPUs is described considering one soft phase made of polyol chains and short miscible hard segments and a hard phase organized as semi-crystalline nanodomains either dispersed or assembled as ramified (nano)objects within the soft phase. The dynamic mechanical properties of the TPUs can be explained by the presence of such well-defined hard domains in the structure of the TPU, acting as reinforcing fillers while maintaining a thermoplastic elastomer mechanical behavior to the TPU above the glass transition of the continuous soft phase.

Introduction

Polyurethane polymers are widely used in consumers daily life products and find applications in a wide range of comfort and insulation products [1], thanks to the versatility of their macromolecular architecture allowing them to be categorized as thermosets, thermoplastics, or elastomers. Thermoplastic polyurethanes, TPUs, usually come from the combination of three components, i.e. a polyol or macrodiol having a high molar mass (up to 3000 g mol−1), a diisocyanate, and a diol or chain extender of low molar mass [1,2]. In order to reduce petrochemical polymers production, TPUs from biobased materials have known a high development in the last years. The biobased building blocks could be provided by the polyol, e.g. derived from dimer fatty acids [1], but also from the chain extender, such as isosorbide which has known a high development in polyurethanes in recent years [3,4], and even by the diisocyanate itself, as an example with the fatty acid-derived diisocyanate dimeryl diisocyanate (DDI) [3,5]. With the rise of the biobased raw materials available for the synthesis of TPUs, polymers of high biobased content can now be produced.

TPUs exhibit soft and hard blocks in their structure, provided by the reaction of the polyol with the diisocyanate and of the diisocyanate with a chain extender, respectively [1,2]. Depending on their macromolecular architecture, in relation with their synthesis procedure, thermoplastic polyurethanes can organize as segmented pseudo-block copolymers, sometimes with (micro)phase separation between soft and hard segments. This property can first be revealed by means of thermal analysis like differential scanning calorimetry (DSC), which can evidence the glass transitions and possible meltings relative to each phase, and can give a first idea of their purity from the values of the associated temperatures [6,7]. In addition, Camberlin and Pascault reported that it was possible to precisely estimate from DSC the segregation degree between both type of segments, based on the change of heat capacity at Tg [8].

As thermoplastic elastomers (TPEs), above the α main mechanical relaxation, associated to the glass transition of the soft phase, TPUs show a rubbery plateau before the flow of the polymer occurring as the hard segments aggregates collapse. As a consequence, the mechanical properties depend on the initial components used which control the ‘quality’ of the phase separation between blocks, as shown for example with difference of thermomechanical properties observed for TPUs synthesized either with MDI diisocyanate or toluene diisocyanate (TDI) [9]. The hard segments content has also a large influence on the phase separation, and particularly on the modulus at the rubbery plateau [10].

Other characterization techniques are commonly used to reveal phase separation such as transmission electron microscopy, TEM, atomic force microscopy, AFM, or X-ray scattering [11]. Based on microstructure analyses, the phase separation relates with the presence of hard domains, either within spherulite-like structures [12], globular structures [4,13,14], or combinations of both [15]. At nano-scale, X-ray analysis provides information about the organization of the hard domains in the structure of the soft phase, such as the crystalline form [16] or interdomain spacing as highlighted by Aneja and Wilkes [17]. These authors reported interdomain spacing of 10.6 nm for a polyurethane prepared from poly(tetramethylene oxide) soft segments and 1,4-butanediol extended piperazine-based hard segments.

The specific characteristic of thermoplastic polyurethanes to organize as segmented block copolymers allows them to behave as TPEs, i.e. to display a constant rubbery modulus above the Tg of the soft continuous phase as crosslinked rubbers.

The ability of TPEs to phase segregate results from the thermodynamic incompatibility between the two blocks. Indeed, considering the block copolymers as a binary blend of soft and hard blocks, the system can be described based on Flory-Huggins’ theory [18,19] by the change of free enthalpy,ΔGM, expressed in the following equation by:ΔGMkT=ΦAlnΦAvANA+ΦBlnΦBvBNB+χ'ABΦAΦBwith Ni being the number of monomers in the chain i, vi the volume of each monomer on chain i, Φi the volume fraction of block i, and χAB the Flory interaction parameter, defined as follows:χAB=Vref×(δAδB)²RTwhere Vref is the reference volume, and δA, δB the Hildebrand solubility parameters [18] of blocks A and B, respectively. The solubility parameter is defined by:δ=(EV)12where V is the molar volume and E the energy of vaporization of the component. The required condition to obtain a miscible blend, meaning no phase separation, implies ΔGM < 0. Phase separation is expected well-defined in block copolymers such as styrene-butadiene-styrene (SBS), a common polymer modifier for bitumen, where both blocks show rather different solubility parameters [20].

The interaction parameter, χ, (calculated from the solubility parameters) allows to describe the phase diagram in block copolymers that usually show a LCST-type (lower critical solution temperature) behavior. The diagram describes the transition of a homogeneous disorganized system to a mesophase structure, where various types of nanostructures, such as spheres, cylinders, or lamellae coexist, depending on the block compositions [21]. By extension, this can also apply to the TPU morphology based on hard and soft segments.

As an example, for polyurethanes based on poly(propylene oxide) (PPO) and various hard segments (MDI/BDO, MDI/3,5-diethyltoluene diamine), a high solubility parameter of the hard segments leads to a high interaction parameter and an organized structure is observed for the hard domains even for a lower hard segment content [22]. This comes from the high difference in solubility parameters between the soft PPO segment and the hard segments (17 MPa1/2 vs. more than 22 MPa1/2, respectively), leading to a high degree of phase separation which can be also favored by the ability of the hard chains to self-assemble and organize. Hard segments having urea functions are even more capable of self-organizing, due to the occurrence of additional hydrogen bonds [23], and display even higher solubility parameters. As a consequence, a highly phase separated polymer containing well organized hard domains relates with a higher modulus in the rubbery state [22].

Camberlin and Pascault [24] pointed out that the quality of phase separation is also dependent on the architecture of the polyol (precursor of the soft segments) which is considered by estimating interaction parameters in polyurethanes having MDI/BDO hard segments. With those type of HS, they reported an interaction parameter for a PPO-based soft segment χ = 2.8, whereas with the same hard segments the corresponding hydrogenated (1,2)-polybutadiene soft segment displayed an interaction parameter χ = 4.7. Although the authors could calculate a critical value of χcr = 0.34 (that takes into account the low polymerization degree of the considered segments) above which no miscibility should theoretically be possible, the system based on PPO yet displayed high miscibility presumably forced by the existence of covalent bonds between both types of segments, in contrast to that with polybutadiene soft segments that was definitely immiscible.

In the late 60s, Hansen has described the global solubility parameter, δ, as the combination of three components reflecting dispersive (London) (δD), polar (δP), and hydrogen bond (δH) interactions [[25], [26], [27]]:δ2=δTOT2=δD²+δP²+δH²

The components δD, δP, and δH are named Hansen’ Solubility Parameters, HSP. Hansen also defined a 3D solubility diagram (δD,δP,δH) where a solubility sphere having a radius R0 can be defined for large molecules. The analysis is based on solubilization tests in solvents of known HSP, with a sphere that encompasses the good solvents of the component in the 3D space [28]. Compatibility between two compounds can be estimated based on their HSP values: compounds with close HSP and with a low distance between centers of their respective solubility spheres are expected to display a good mutual affinity. This implies that the solubility spheres of the components are overlapping. The distance Ra between centers of the solubility spheres of two components named as 1 and 2, is given by:Ra2=4(δD1δD2)2+(δP1δP2)2+(δH1δH2)2

For a given component/solvent pair, the considered solvent would exhibit strong affinity with the component for a very low value of Ra (inferior to R0, the radius of the solubility sphere). In this case, the solvent coordinates are expected to be inside the component solubility sphere. However, when considering two different components, a value of Ra = 8 MPa1/2 is usually considered as the upper limit for compatibility [25].

In the last decades, literature has reported several works dealing with the solubility of polymers and other compounds [25]. First records were on the use of turbidimetric titration to estimate HSP of polymers such as polystyrene [29] (PS), polyurethanes (combined with intrinsic viscosity) [30] or polyols [31]. Turbidimetric titration has also been used to calculate bitumen solubility parameters [32]. In the last years, inverse gas chromatography proved to be an accurate method to determine HSP of block copolymers such as poly(ethylene-co-vinyl acetate) [33], or SBS [34].

Others reported the use of solvent mixtures and a viscosimetric method to determine HSP in order to calculate the interaction parameter in block copolymers [35]. Although these authors found discrepancies between theoretical and experimental values, the tendency of the polymer to phase separate is definitely higher if the interaction parameter is high.

Another method consisting in simple solubility tests in various solvents combined with calculations by a dedicated software was used to estimate HSP on polystyrene and polybutadiene [20] or bitumen [36]. In recent years, Bouteiller et al. reported the determination of HSP and solubility spheres with the new HSPiP software [28] (allowing to calculate or predict HSP based on a solvent database and algorithm) for low molar mass organogelator molecules and went even further with the determination of a gelation sphere [37,38].

HSP have also been used to predict miscibility in polymer blends, as reported by David and Sincock [39]. Furtwengler et al., as well as Zhang and Kessler [40,41] used Hansen’ solubility spheres diagram obtained with HSPiP software to estimate compatibility between polyols to make stable emulsions with a view to manufacturing foams. This method has also been reported by Redelius to study compatibility of bitumen with different polymers, such as SBS, polyethylene sulfide or polyether sulfone [36].

Mieczkowski showed that the solubility parameters of the polyurethane polymers are very dependent on the soft segment molecular architecture [30]. In fact, a poly (ethylene adipate) (PEA)-based polyurethane proved to be more polar than the corresponding PPO-based polyurethane.

Conventional soft blocks considered for the synthesis of TPUs usually show total solubility parameters lower than 20 MPa1/2, while hard blocks display higher solubility values. As an example, Table 1 reports the solubility parameter of a range of soft and hard segments often considered for the synthesis of polyurethanes.

The present study reviews the synthesis of a TPU with a soft segment derived from fatty acids. TPUs derived from such polyols and MDI-BDO hard segments indeed already proved to be particularly well suited to the manufacture of bituminous binders [44]. The solubility parameters of the soft segments are evaluated and compared to those of the hard segment in order to predict phase separation in the TPU based on thermodynamic incompatibility of the segments. Phase separation is also evaluated by using conventional methods such as differential scanning calorimetry and morphology characterization techniques (AFM, TEM, and SAXS). From these results, relationships between thermomechanical state and physical behaviors, i.e. rheology and dynamic mechanical properties (DMA), of the investigated TPU are established. This work is the first step of a more global study aiming to optimize the blends of the present TPU with bitumen, by establishing the interactions occurring between the polymer's blocks and the bitumen's fractions, and consequently better understanding the resulting mechanical and morphological properties of the material.

Section snippets

Materials

The polyester polyol Radia 7285 based on fatty acid was provided by Oleon Co., the diisocyanate 4,4′-methylene bis(phenyl isocyanate) (MDI), and the chain extender, 1,4-butanediol (BDO), were provided by Sigma Aldrich Co. The solvents used were also provided by Sigma Aldrich Co.

The polyol was used to synthesize TPUs along with MDI diisocyanate and BDO as chain extender. A two-step synthesis was used consisting in the formation of a pre-polymer in the first step, made with excess of MDI and the

Physico-chemical characterization of initial building blocks of TPU

The polyester polyol Radia 7285 supplied by Oleon Co. obtained from dimerization and esterification of rapeseed oil is 80 wt% biobased. This polyol displays a very high molar mass dispersity and the calculation of the average molar mass by 1H NMR gave a lower value of 3200 g mol−1. This latter value is in a good agreement with the bi-functionality of the polyol and the hydroxyl number of 35.8 mg KOH. g−1 provided by the supplier. It is well known that dimerization of fatty acids can lead to

Conclusions

In this work, we intended to highlight the importance to consider the thermodynamics to predict and tailor (nano/micro)phase separation phenomena in thermoplastic polyurethanes (TPU) by using Hansen’ solubility parameters. Therefore HSP have been accurately measured for MDI/BDO hard segments and the fatty acid derived soft segment which were considered for design a series of TPUs having various hard segment contents. By using the HSPiP software, the construction of the solubility spheres for

Author contributions

All authors have contributed to the writing of the manuscript and have given approval for its final version.

CRediT authorship contribution statement

Raïssa Gallu: Conceptualization, Methodology, Investigation, Writing - original draft. Françoise Méchin: Conceptualization, Writing - review & editing, Supervision. Florent Dalmas: Conceptualization, Formal analysis, Investigation, Writing - review & editing, Supervision. Jean-François Gérard: Conceptualization, Writing - review & editing, Supervision. Rémi Perrin: Conceptualization, Resources. Frédéric Loup: Conceptualization, Resources.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors express their thanks to Soprema Co. and Eiffage Infrastructures Co. for their financial support and to Professor Jean-Pierre Pascault for helpful discussions and advice. Dr Catherine Ladavière (IMP) is gratefully acknowledged for the MALDI-TOF analysis.

The authors would also like to thank the NMR Polymer Center of the “Institut de Chimie de Lyon” (FR3023), for assistance and access to the NMR facilities. The “Centre Technologique des Microstructures, plateforme de l’Université

References (70)

  • E. Díez et al.

    Polymer-solvent interaction parameters of SBS rubbers by inverse gas chromatography measurements

    Fluid Phase Equil.

    (2011)
  • D.J. David et al.

    Estimation of miscibility of polymer blends using the solubility parameter concept

    Polymer

    (1992)
  • L. Cuvé et al.

    Synthesis and properties of polyurethanes based on polyolefine : 1 . Rigid polyurethanes and amorphous segmented polyurethanes prepared in polar solvents under homogeneous conditions

    Polymer

    (1991)
  • B. Bengtson et al.

    Thermal and mechanical properties of solution polymerized segmented polyurethanes with butadiene soft segments

    Polymer

    (1985)
  • B. Fu et al.

    Synthesis and properties of monodisperse hydroxy-terminated oligomers of 1,4-butanediol and 2,4-toluene diisocyanate

    Polymer

    (1985)
  • J. Blackwell et al.

    Structure of the hard segments in polyurethane elastomers

    Polymer

    (1979)
  • J.T. Koberstein et al.

    Small-angle light scattering studies of macrophase separation in segmented polyurethane block copolymers

    Polymer

    (1984)
  • Y. Swolfs et al.

    Linking the morphology of a high hard segment content polyurethane to its thermal behaviour and mechanical properties

    Polymer

    (2015)
  • C. Check et al.

    Kinetic aspects of formation and processing of polycaprolactone polyurethanes in situ from a blocked isocyanate

    Polym. Chem.

    (2018)
  • S. Abouzahr et al.

    Structure-property behaviour of segmented polyether-MDI-butanediol based urethanes: effect of composition ratio

    Polymer

    (1982)
  • J. Blackwell et al.

    Structure of polyurethane elastomers : effect of chain extender length on the structure of MDI/diol hard segments

    Polymer

    (1982)
  • G.P. Baeza et al.

    Multiscale organization of thermoplastic elastomers with varying content of hard segments

    Polymer

    (2016)
  • J. Blackwell et al.

    Structure of polyurethane elastomers. X-ray diffraction and conformational analysis of MDI-propandiol and MDI-ethylene glycol hard segments

    Polymer

    (1981)
  • M.F. Sonnenschein

    Polyurethanes: Science, Technology, Markets, and Trends

    (2015)
  • E. Delebecq et al.

    On the versatility of urethane/urea bonds : reversibility, blocked isocyanate, and non-isocyanate polyurethane

    Chem. Rev.

    (2013)
  • T. Calvo-Correas et al.

    Synthesis and characterization of polyurethanes with high renewable carbon content and tailored properties

    ACS Sustain. Chem. Eng.

    (2016)
  • A. Saiani et al.

    Origin of multiple melting endotherms in a high hard block content polyurethane . 2 . Structural investigation

    Macromolecules

    (2004)
  • A. Saiani et al.

    Origin of multiple melting endotherms in a high hard block content polyurethane. 1. Thermodynamic investigation

    Macromolecules

    (2001)
  • Y. Camberlin et al.

    Quantitative DSC evaluation of phase segregation rate in linear segmented polyurethanes and polyurethaneureas

    J. Polym. Sci. Polym. Chem. Ed.

    (1983)
  • C.G. Seefried et al.

    Thermoplastic urethane elastomers. III. Effects of variations in hard‐segment concentration

    J. Appl. Polym. Sci.

    (1975)
  • B. Fernández-D’Arlas et al.

    Tailoring the morphology and melting points of segmented thermoplastic polyurethanes by self-nucleation

    Macromolecules

    (2016)
  • C. Li et al.

    Morphology of segmented polybutadiene-polyurethane elastomers

    Macromolecules

    (1988)
  • E. Tocha et al.

    Morphology of polyurethane revisited by complementary AFM and TEM

    J. Macromol. Sci. Part B.

    (2002)
  • J. Blackwell et al.

    Hard-segment polymorphism in MDI/Diol-based polyurethane elastomers

    J. Polym. Sci., Polym. Phys. Ed.

    (1984)
  • L.A. Utracki et al.

    Polymer Blends Handbook

    (2014)
  • Cited by (41)

    View all citing articles on Scopus
    View full text