Contents lists available at ScienceDirect

European Polymer Journal

journal homepage: www.elsevier.com/locate/europolj

Importance of film morphology on the performance of thermo-responsivewaterborne pressure sensitive adhesives

Ehsan Mehravara, Michael A. Grossb, Amaia Agirrea, Bernd Reckb, Jose R. Leizaa, José M. Asuaa,⁎

a POLYMAT and Kimika Aplikatua Saila, Kimika Zientzien Fakultatea, University of the Basque Country UPV/EHU, Joxe Mari Korta Zentroa, Tolosa Hiribidea 72,Donostia-San Sebastian 20018, Spainb BASF SE, RAD Dispersions & Colloidal Materials Research – B001, Carl-Bosch-Strasse 38, 67056 Ludwigshafen, Germany

A R T I C L E I N F O

Keywords:Miniemulsion polymerizationPressure-sensitive adhesivesCrystallinityCore-shellBlendAdhesive propertiesMechanical propertiesTemperature-responsive behavior

A B S T R A C T

Temperature responsive pressure sensitive adhesives (PSAs) are of interests in many applications due to thepossibility of removing strong adhesives by moderate heating. In this work, semicrystalline poly(octadecyl ac-rylate) obtained by miniemulsion polymerization is used as a key component of the thermal responsiveness.Comparison of the adhesive performance of crystalline hard core-soft shell PSAs with other PSAs of the sameoverall composition and similar molar mass distributions but prepared by blending soft and hard latexes showsthat the arrangement of the phases in the film play a key role in performance of the PSAs. Good adhesiveperformance and thermal responsiveness is only achieved when the crystalline domains are dispersed in acontinuous soft phase containing entangled nanogels. Interestingly, this morphology is achieved with blends, butno with core-shell dispersions. The reasons for this difference are investigated.

1. Introduction

Pressure sensitive adhesives (PSAs) are a special type of adhesivesthat firmly adhere into a wide range of surfaces by the application of alight pressure (1–10 Pa) for a short contact time (1–5 s) [1,2]. Water-borne (meth)acrylic PSAs have the fastest growth in commercial ap-plication because of the low environment impact, balanced end-productproperties, compatibility with additives and processability, competitivecost and oxidative ultraviolet resistance [3]. Other advantages of(meth)acrylic PSAs dispersions are their high solids, their ease appli-cation, and the fact that they may be formulated, in many instances,without the need for addition of tackifiers [4].

The major challenge in the production of PSAs is to meet the con-flicting properties that are required for application (tack, peel strengthand shear resistance). As the pure (meth)acrylic formulations presentsome intrinsic limitations as low cohesive strength [5], attempts toachieve this goal by using polymer-polymer and polymer-inorganichybrids have been reported. Thus, polyurethane has been incorporatedinto the (meth)acrylic PSAs to improve cohesive strength [6–12]. Core-shell particles that upon water evaporation yield a percolating cross-linked structure that contains a less crosslinked polymer have reportedto increase the adhesive energy on polyethylene [13]. Several types ofnanoparticles (clays, carbon nanotubes and graphene, polymers, mo-lybdenum disulfide, silica) [14–22] have also been used to improve the

performance of the PSAs.Polymer nanoparticles are particularly interesting because in addi-

tion of improving the PSA performance, they can have temperaturetriggered functionality. Thus, the tackiness of the PSAs containing poly(methyl methacrylate) nanoparticles was switched by infrared sintering[18] and PSAs containing semicrystalline domains that strongly ad-hered to substrates could be easily and abruptly removed by moderateheating [22].

In most of the cases intimate contact between the different phases ispreferred as it provides better properties [19,21,22]. This is oftenachieved by synthesizing composite particles, i.e., particles that containall the components.

However, Agirre et al. reported an interesting case in which blendsallowed the unusual simultaneous increase of both the peel strengthand the shear resistance, whereas core-shell particles of the sameoverall composition exhibited the classical decrease of the peel re-sistance as shear resistance increased [22]. Although it may be of im-portance for the design of optimized adhesives, no attempts to under-stand the fundamental reasons for these findings were reported.

This work is an attempt to understand the fundamental differencesbetween hard core-soft shell PSAs and those prepared by blending hardand soft latexes. A semicrystalline poly(octadecyl acrylate) (also knownas poly(stearyl acrylate)) that can provide thermal-responsiveness tothe PSA was used as a hard phase and a copolymer containing 2-

https://doi.org/10.1016/j.eurpolymj.2017.11.004Received 19 June 2017; Received in revised form 30 October 2017; Accepted 2 November 2017

⁎ Corresponding author.E-mail address: jm.asua@ehu.es (J.M. Asua).

European Polymer Journal 98 (2018) 63–71

Available online 04 November 20170014-3057/ © 2017 Elsevier Ltd. All rights reserved.

MARK

ethylhexyl acrylate (2EHA), methyl methacrylate (MMA) and me-thacrylic acid (MAA) as soft phase. The adhesive performance of thecore-shell and blends was compared showing that the blends had betteradhesive properties. A deep characterization of the adhesives was car-ried out finding that the arrangement of the nanogels in the adhesivefilm is the most likely reason for the different behavior. In addition, itwas observed that these adhesives were temperature responsive.

2. Experimental section

2.1. Materials

Technical grade monomers, methyl methacrylate (MMA, BASF), 2-ethylhexyl acrylate (2EHA, BASF) and methacrylic acid (MAA, BASF),and stearyl acrylate (SA, Aldrich) were used without purification.Potassium persulfate (KPS, Fluka) as water soluble radical initiator,sodium bicarbonate (NaHCO3, BASF) to control the miniemulsionviscosity by reducing the electrostatic interactions among droplets andDisponil® FES 32 (Fatty alcohol ether sulfate+ 4 ethylene oxide, so-dium salt) (BASF) as anionic surfactant were used as received.Deionized water was used as polymerization media.

2.2. Experimental design

Tables 1 and 2 summarize the experimental design in which thecomonomer composition was varied from SA/SC(M)A=0/100 to SA/SC(M)A=40/60 wt/wt (SC(M)A= short chain (meth)acrylates). TheSA amount in copolymers was limited in order to be able to form tackyfilms. The SC(M)A compositions were 2EHA/MAA=99/1wt%(Table 1) and 2EHA/MMA/MAA=84/15/1 wt% (Table 2). SA is awater insoluble monomer, and therefore miniemulsion polymerizationthat does not require monomer transport through the aqueous phase[23,24] was used. 1A is a SA homopolymer synthesized by miniemul-sion polymerization [25,26]. The semicrystalline copolymers wereprepared by seeded semibatch emulsion polymerization by using Latex1A as seeds. Thus, the poly(SA) latex was charged into the reactor, andthen the pre-emulsion of SC(M)A monomers and the initiator solutionwere fed separately during 3.5 h under starved conditions (Runs 1D–4Dand Runs 1D′–4D′). The reference latexes, 1E and 1E′, that were devoidof SA, were synthesized by seeded semibatch emulsion polymerizationby using polystyrene seeds (0.1 wt% based on monomer, particle dia-meter= 31 nm). The SA miniemulsion preparation and the synthesis ofthe latexes are described in Supporting Information.

Blending was used to prepare semicrystalline adhesives as well. Theamorphous SC(M)A latexes (1E and 1E′) were mixed with the poly(SA)Latex 1A to obtain blends with different copolymer composition andcrystallinity (Blends 1–4, Table 1; Blends 1′–4′, Table 2).

2.3. Characterization

PSAs latexes and films. Monomer droplet and particle sizes of thelatexes were measured by hydrodynamic chromatography (HDC); thegel fraction and swelling degree were measured by Soxhlet extraction,using THF as the solvent; the molar mass distribution of the solublefraction of polymers was determined by size exclusion chromatography(SEC/RI); the molar mass distribution of the whole polymer was de-termined by asymmetric-flow field-flow fractionation (AF4/MALS/RI);the thermal characterization of the PSAs films was carried out by dif-ferential scanning calorimetry (DSC); the morphology of latex particlesand adhesive films were studied by means of transmission electronmicroscopy (TEM) and atomic force microscopy (AFM) respectively; theadhesive property studies were 180° peel, shear resistance, loop andprobe tack tests; the mechanical properties of the adhesives were stu-died in terms of linear viscoelastic (characterized with rheometer) andnonlinear elastic (tensile test) measurements. The detailed descriptionof the characterization methods and adhesive and mechanical tests isprovided in Supporting Information.

3. Results and discussions

The volume average particle diameter of the Latex 1A used as seedin Series D and D′ and those of the final latexes are given in Tables 1 and2. It can be seen that the final particle size decreased with increasing SAcontent. The reason was the increasing number of seed particles usedfor higher SA contents. The TEM micrograph of latex 2D′ is shown inFig. 1. It can be seen that the semicrystalline particles present a core-shell morphology with the dark poly(SA) domains in the core. More-over, the small amorphous particles observed in the TEM micrographsshow that some secondary nucleation occurred during the second stageof polymerization, most likely by homogeneous nucleation (some of thesmall amorphous particles are indicated by arrows in the TEM image).

Tables 1 and 2 showed that there were no substantial differences ingel content and crystallinity between the latexes obtained by two stagepolymerization (Series D and D′) and the blends. The gel fraction andcrystallinity values measured for blends agreed well with the averagevalue calculated taking into account the fractions of poly(SA) and poly

Table 1The summary and characteristic of the PSAs latexes synthesized with SC(M)A: 2EHA/MAA=99/1wt%.

Latex SA/SC(M)Aa (wt/wt) Particle size (nm) Gel content (%) Swelling degree Mw, Soluble part (g/mol) Đ Tm (°C) Xcb (%) Tgc (°C)

Series A Batch miniemulsion1A 100/0 170 79.2 ± 0.2 14.1 ± 0.5 392000 3.3 51 41.8 17.0

Series D Latex 1A+3.5 h addition of SC(M)A pre-emulsion and KPS solution4D 40/60 220 71.0 ± 0.1 10.3 ± 0.1 83000 3.3 51 15.9 −62.33D 30/70 236 69.4 ± 0.1 10.2 ± 0.1 86000 2.8 50 11.5 −63.02D 20/80 270 66.3 ± 0.0 10.5 ± 0.2 93000 2.8 50 6.9 −63.01D 10/90 317 66.5 ± 0.0 9.7 ± 0.0 100000 2.8 48 3.4 −65.8

Series E Seeded emulsion copolymerization1E 0/100 269 63.8 ± 0.1 10.8 ± 0.3 123000 2.7 – – −67

Blends Blending 1A and 1E latexesBlend 4 40/60 – 71.0 ± 1.7 12.5 ± 1.5 184000 3.8 49 17.0 −67.7Blend 3 30/70 – 67.5 ± 0.4 12.1 ± 0.4 162000 3.3 49 10.9 −67.2Blend 2 20/80 – 67.0 ± 0.5 11.7 ± 1.1 139000 3.4 48 6.6 −67.8Blend 1 10/90 – 64.2 ± 0.6 12.3 ± 2.0 129000 3.1 48 3.3 −67.0

a SC(M)A: 2EHA/MAA=99/1wt%.b Referred to the whole polymer.c For Run 1A, the onset temperature was taken as Tg, whereas for the rest, the inflection point temperature was considered. Poly(SA)(Run 1A) shows weak transitions at low

temperature ranges corresponding to the transitions of amorphous part (α and β transitions) [27].

E. Mehravar et al. European Polymer Journal 98 (2018) 63–71

64

(SC(M)A). However, the swelling degrees (the swelling degree is in-versely related with the crosslinking density of the gel) of the blendswere higher than the Series D and D′ latexes. This indicates that thestructure of the gel formed by polymerization of the short chain (meth)acrylates in Series D and D′ was affected by the fact that it was formedin the shell of the poly(SA) seed particles. In addition, it can be ob-served that the gel fractions of the latexes containing MMA comonomer(Runs in Table 2) were lower and the swelling degrees were higher thanin the latexes devoid of MMA (Runs in Table 1), because the presence ofMMA units in the copolymer chains reduces the possibility of hydrogenabstraction [28–30]. Furthermore, the reactivity ratio of MMA is higherthan that of the acrylates, and hence, the radicals are predominantly inMMA units, whose activity for hydrogen abstraction is significantlylower than that of the acrylate radicals. In the same vein, MMA radicalsterminate by disproportionation, which reduces the probability of gelformation. The lower sol molar mass of the latexes prepared withoutMMA (Table 1) was a consequence of the higher gel content because thelong polymer chains are preferently incorporated to the gel. Tables 1

and 2 also show that the Tg of the latexes containing MMA was higher.Fig. 2 presents the shear resistance, peel and loop tack strengths

measured at 23 °C for the films cast from latexes in Series D (Table 1),and D′ (Table 2) as well as the amorphous latexes 1E and 1E′. It can beseen that the shear resistance of the adhesives increased with SA con-tent, and hence with crystallinity. Therefore, the presence of crystallinedomains made the material more cohesive and increased the creep re-sistance. On the other hand, the adhesives containing MMA (even withlower gel fraction, see Tables 1 and 2) showed substantially highershear resistance than the softer adhesives devoid of MMA, likely due tothe higher rigidity of the amorphous copolymer.

Fig. 2 also shows that the peel strength of the adhesives containingMMA slightly increased with crystallinity and was higher than thesofter adhesives of Series D that had a lower Tg. Conversely, the looptack strength of the adhesives containing MMA decreased with in-creasing crystallinity as it may be expected because stiffer polymersincrease the adhesive shear and often peel resistance, but decreasetackiness [31]. On the other hand, it is remarkable that in Series D(2EHA/MAA) the tackiness increased with SA content (crystallinity).The properties of adhesives containing MMA are closer to commercialformulations and are studied comprehensively in this article.

Fig. 3 presents the comparison of the adhesive properties of thelatexes synthesized in Series D′ (core-shell particles, latexes 1D′–4D′)with those of the Blends 1′–4′ and the latex 1E′. It can be seen that theshear resistance of the blends also increased with crystallinity and itwas substantially higher than for core-shell (Series D′). The peelstrength of the blends was similar to the core-shell latexes. In addition,it was observed that the loop tack strength and work of adhesion Wa

(measured by the probe-tack test) of both core-shells and blends de-creased with increasing crystallinity and the loop tack of the blends washigher. Moreover, up to 20 wt% SA contents, the Wa of the blends washigher or at least equal than for the core-shell particles (Series D′). Thisis remarkable because adhesives showing a high shear resistanceusually exhibit low loop tack strength and Wa. Whereas a high shearresistance requires a solid-like response (high creep resistance), looptack strength and work of adhesion require polymer flow and energydissipation during deformation, which is the response of the highlyviscous liquid [22,32]. These conflicting requirements were met by thesemicrystalline blends.

The probe tack measurements of Series D′, the blends and 1E′ latexare shown in Fig. 4. In these experiments, first the probe comes intocontact with the adhesive film under a controlled load and setting time.

Table 2The summary and characteristic of the PSAs latexes synthesized with SC(M)A: 2EHA/MMA/MAA=84/15/1 wt%.

Latex SA/SC(M)Aa (wt/wt)

Particle size(nm)

Gel content (%) Swelling degree Mw, Soluble part (g/mol)

Đ Tm (°C) Xcb (%) Tgc (°C)

Series A Batch miniemulsion1A 100/0 170 79.2 ± 0.2 14.1 ± 0.5 392000 3.3 51 41.8 17.0

Series D′ Latex 1A+3.5 h addition of SC(M)A pre-emulsion and KPS solution4D′ 40/60 209 65.4 ± 0.1 14.1 ± 0.3 170000 3.7 51 16.1 −46.03D′ 30/70 232 60.7 ± 0.1 14.6 ± 0.2 176000 3.7 50 11.2 −46.32D′ 20/80 256 57.0 ± 0.0 14.1 ± 0.2 209000 3.1 49 7.1 −46.01D′ 10/90 311 52.9 ± 0.1 13.9 ± 0.6 228000 4.0 48 3.7 −46.4

Series E′ Seeded emulsioncopolymerization1E′ 0/100 294 52.1 ± 0.1 18.3 ± 0.1 178000 4.1 – – −49

Blends Blending 1A and 1E′ latexesBlend 4′ 40/60 – 63.0 ± 0.7 16.8 ± 0.8 235000 3.6 51 16.5 −46.3Blend 3′ 30/70 – 60.0 ± 1.0 17.2 ± 0.6 234000 3.2 50 11.0 −46.4Blend 2′ 20/80 – 57.4 ± 0.2 17.2 ± 0.0 233000 3.1 50 7.1 −47.0Blend 1′ 10/90 – 54.0 ± 0.1 16.7 ± 1.0 236000 3.1 49 3.7 −45.7

a SC(M)A: 2EHA/MMA/MAA=84/15/1 wt%.b Referred to the whole polymer.c For Run 1A, the onset temperature was taken as Tg, whereas for the rest, the inflection point temperature was considered. Poly(SA)(Run 1A) shows weak transitions at low

temperature ranges corresponding to the transitions of amorphous part (α and β transitions) [27].

Fig. 1. The TEM micrograph of synthesized latex in Run 2D′ (scale bar is 1 µm). Thearrows show some small amorphous particles.

E. Mehravar et al. European Polymer Journal 98 (2018) 63–71

65

When the probe is elevated from the adhesive, a negative pressure iscreated, which induces the formation of cavities (in a high-performingadhesives). The number of cavities reaches a maximum at the max-imum stress, σmax. Later, the stress falls to a plateau because of thelateral propagation of the cavities and the vertical deformation as thewalls between the cavities that become fibrils [33–36]. The force that isneeded to draw the fibrils determines the plateau stress, σp, at higherstrains. Finally, the stress drops to zero when the fibrils break or detach

from the probe at maximum strain, εf (strain at failure). The area underthe stress-strain curve represents the work of adhesion, Wa. The probe-tack curve of Run 1E′ (Fig. 4) shows a sharp maximum stress at ratherlow strains followed by fibrillation plateau. The curve finally ends up bya decrease in the force to zero, which was the result of the detachmentat the interface between the probe and the adhesive layer, via an ad-hesive (interfacial) debonding mechanism. Adhesive 1D′ had also anadhesive failure, but adhesives 2D′, 3D′ and 4D′ presented a cohesive

0

200

400

600

800

1000

0 10 20 30 40 50

Shea

r re

sist

ance

(min

)

SA (wt%)

2EHA/MMA/MAA=84/15/1 wt%

2EHA/MAA=99/1 wt%

0123456789

10

0 10 20 30 40 50

Peel

adh

esio

n (N

/25m

m)

SA (wt%)

2EHA/MMA/MAA=84/15/1 wt%

2EHA/MAA=99/1 wt%

0

1

2

3

4

5

6

7

0 10 20 30 40 50

Loo

p ta

ck (N

/25m

m)

SA (wt%)

2EHA/MMA/MAA=84/15/1 wt% 2EHA/MAA= 99/1 wt%

Fig. 2. The effect of SC(M)A composition and SAcontent on the shear resistance, peel and loop tackstrengths of semicrystalline PSAs synthesized in SeriesD and D′. The properties of 1E and 1E′ adhesives arealso included. The tests were carried out at 23 °C and55wt% relative humidity.

0

1000

2000

3000

4000

5000

6000

0 10 20 30 40 50

Shea

r re

sist

ance

(min

)*

SA (wt%)

(a) Blend

Core-shell

0123456789

10

0 10 20 30 40 50

Peel

adh

esio

n (N

/25m

m)

SA (wt%)

(b)

Blend Core-shell

0

1

2

3

4

5

0 10 20 30 40 50

Loo

p ta

ck (N

/25m

m)

SA (wt%)

(c)

Blend Core-shell

0

10

20

30

40

50

60

70

80

90

0 10 20 30 40 50

Wa

(J/m

2 )

SA (wt%)

(d)

Blend Core-shell

Fig. 3. (a) Shear resistance, (b) peel adhesion, (c) looptack strengths and (d) Wa (measures from probe-tacktest) of core-shell latexes synthesized in Series D′ (Runs1D′–4D′), Blend samples (Blend 1′–4′) and Run 1E′.The tests were carried out at 23 °C and 55wt% relativehumidity. *The shear resistance test of Blend 4′ samplewas without failure after one week.

E. Mehravar et al. European Polymer Journal 98 (2018) 63–71

66

failure.The comparison of Run lE′ and the blends (Blends 1′- 3′) showed

that blends presented shorter plateaus and that the length of plateaudecreased with increasing crystallinity most probably because the fi-brils were too stiff and they detached under smaller strain. It means thatthe stress needed to deform fibrils is greater than the adhesive force tothe substrate. The probe-tack curve of Blend 4′ is characterized by asharp decrease of the stress without a fibrillation plateau and a very lowadhesive energy.

To shed light on the differences observed in the tack adhesion of theSeries D′ (core-shell) and the blends both the linear viscoelasticity,which determines whether there is fibrillation or interfacial crackpropagation during debonding, and the nonlinear viscoelasticity at highstrains, which influences the fibril detachment, were studied.

Figs. S1 and S2 in Supporting Information show the variation ofstorage and loss modulus (G′ and G″, respectively) and tanδ with fre-quency at 23 °C for the core-shell (Series D′) and blend films, respec-tively. It can be seen that there is an increase of the storage and lossmodulus with crystallinity content, indicating that the introduction ofhard crystalline domains enhances the stiffness of the PSAs. Moreover,the comparison of core-shell and blend adhesives (Figs. S1 and S2 inSupporting Information) shows that the dynamic moduli of the blendadhesives are substantially higher than those of core-shell adhesives.The Dahlquist criterion establishes that in order to have a high tackadhesion G′ should be smaller than 0.1MPa [37]. The rationale behindthis criterion is that a certain softness of the adhesive is needed to forma good contact in a short contact time. At the relevant debonding fre-quency (1 Hz) the Run 1É, Runs 1-3D′ in Series D′ and the blend 1′adhesives fulfill the criterion. However, this criterion is not fulfilled byRun 4D′, blends 3′ and 4′, leading to lower work of adhesion (Wa).Interestingly, no effect was observed in neither the loop tack nor thepeel adhesion.

On the other hand, the ratio tanδ/G′ has been related to the energydissipation at the interface of adhesive-substrate to avoid detachment ofthe adhesive and a value of tanδ/G′>5MPa−1 has been recommendedfor steel substrates [36]. Fig. 5 presents the effect of the SA content on

tanδ/G′ for both Series D′ and blends. It can be seen that this ratiodecreased with the SA content (the adhesive became stiffer) and thatthe core-shell samples had higher tanδ/G′ values. Fig. 5 shows that nodetachment at small strains was expected for Series D′ up to 30wt% ofSA and for the blend with 10wt% of SA. On the other hand, early de-tachment was expected for the blends with ≥20wt% of SA, but Fig. 4bshows that this was not the case with 20 and 30wt% of SA becauseclear plateaus that extended to relatively long strains were observed.This indicates that the detachment occurred during fibril elongation,i.e., in the non-linear rheology regime. Fig. 6 presents the tensile ex-periments carried out with the films prepared with Series D′ and blends.Run 1E′ is included as a reference. It can be seen that 1E′ shows bothshear softening and hardening behavior typical of (lightly) crosslinkedadhesives. The behavior of Series D′ is interesting. Whereas the Young’smodulus, which represents the unrelaxed modulus and is directlycomparable with G′, increased with SA content, the strain hardeningdisappeared. This agrees with the cohesive failure observed in the probetack experiments (Fig. 4a). In addition, this indicates that for the core-shell, the interconnection between nanogels is lost.

On the other hand, in the case of the blends, Fig. 6b shows that boththe Young’s modulus and the strain hardening increased with the SAcontent. The increase in the stiffness of the adhesive caused the earlierdetachment of the adhesive from the substrate observed in probe tackexperiments of Fig. 4b. The high Young’s modulus and the strong strainhardening shown by the blends explain the good shear resistance of theblends with ≥20wt% of SA.

The result presented above showed that the blends were more co-hesive than the core-shell adhesives, which led to higher shear re-sistance without sacrifying peel adhesion, loop tack and Wa (in this caseup to 30wt% of SA). In an attempt to determine the origin of the dif-ferences between the core-shells and the blends, the microstructure ofthe films was determined by AFM and the MWD of the polymers wasmeasured by asymmetric flow field flow fractionation (AF4).

Fig. 7 presents the AFM phase images of the cross-section of ad-hesives 2D′, 4D′, Blend 2′ and Blend 4′ (the AFM phase and heightimages of the air-film interface of the films are shown in SupportingInformation (Fig. S3)). It can be seen that the films cast from blendspresent a distribution of hard phase (bright areas that correspond to theSA homopolymer, Latex 1A) in a soft matrix (dark areas correspondingto latex 1E′). On the other hand, the films cast from the core-shell la-texes present a homogenous distribution of hard phase (the poly(SA)core) separated by a thin film of soft polymer, which was thinner as theSA content increased.

The thermal characterization of the core-shell and blend adhesivesshowed that both the crystallinity and the Tg of the amorphous SC(M)A

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 2 4 6 8 10

Stre

ss (M

Pa)

Strain

(a)

Series D´ (core-shell)

Run 4D´ Run 3D´ Run 2D´ Run 1D´ Run 1E´

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 2 4 6 8 10

Stre

ss (M

Pa)

Strain

(b) Blend 4´ Blend 3´ Blend 2´ Blend 1´ Run 1E´

Fig. 4. Representative stress-strain curves in probe-tack tests on (a) core-shell (Series D′)and (b) blend (Blends 1′–4′) PSAs. The stress-strain behavior of amorphous PSA (Run 1E′)is also included. The tests were carried out at 23 °C and 55% relative humidity.

Fig. 5. tanδ/G′ as a function of the poly(SA) concentration for core-shell (Series D′) andblend adhesives. Measurements were made at a temperature of 23 °C and a frequency of1 Hz.

E. Mehravar et al. European Polymer Journal 98 (2018) 63–71

67

copolymer for the core-shell and blend adhesives were the same (thefirst heating DSC scan of Run 2D′ and Blend 2′ are shown in Fig. S4 inSupporting Information) and no signal of polymer interpenetration wasapparent. Therefore, the copolymerization of the SC(M)A monomers inthe thin shells did not influence the Tg of the amorphous copolymermatrix nor led to any significant interpenetration between the polymerof the shell and that of the core.

In the case of the core-shell latexes, the soft shell forms a continuousmedium in the adhesive film where the hard cores are dispersed. Fig. 6shows that this led to the loss of the strain hardening capability be-coming liquid-like as the thickness was reduced. The fraction of theshell polymer that is not soluble in solvents (gel fraction in Tables 1 and2) is expected to be formed by nanogels [11] that when entangled yieldthe strain hardening. The fact that thin shells did not have any hard-ening indicates that the nanogels are not entangled. The reason mightbe that the nanogels are smaller or that because of the geometric re-strictions they did not interconnect.

The molar mass distributions were measured by AF4 using re-fractive index (RI) and multi-angle light scattering (MALS) detectors. RIchromatograms clearly show that the presence of two populations in allthe samples (Fig. 8(a)). The area of the first peak agrees with the solublefraction and that of the second peak with the non-soluble polymer (The

gel fraction values measured by both soxhlet extraction and AF4 tech-niques are compared in Table S1 in Supporting Information). When theRI and light scattering signals were translated into molar mass, it wasfound that a constant Mw was provided by the MALS detector above acertain elution time (Fig. 8(b)). This effect has been observed by otherauthors [38,39] but no explanation was given. Likely the reasons for theobservation was that the hydrodynamic radii of the nanogels were thatbig that the diffusion coefficients were much lower than the minimumcross-flow velocity of the equipment used. Under these conditions allthe nanogels above a certain size (which for the samples analyzed herewere all of them) remained very close to the AF4 membrane and therewas no separation between them. Therefore, the plateau represents theweight average molar mass of the nanogels. Comparison of the plateauvalues shows that the molar mass of the nanogels for the core-shelladhesives is lower than the blends. The lower molar mass may be theresult of the lower probability of bimolecular termination betweennanogels placed at opposite location in the shell. However, the molarmasses of the nanogels still were very high and hence they were ex-pected to form entanglements.

Geometrical restrictions may be the reason for the lack of en-tanglements between the nanogels. Taking into account that the coreswere homogeneously distributed, the average interparticle space (IPS)

0

0.1

0.2

0.3

0.4

0.5

0 5 10 15 20 25

Stre

ss (M

Pa)

Strain

(a)

Series D´ (core-shell)

Run 4D´ Run 3D´ Run 2D´ Run 1D´ Run 1E´

0

0.1

0.2

0.3

0.4

0.5

0 5 10 15 20 25

Stre

ss (M

Pa)

Strain

(b) Blend 4´ Blend 3´ Blend 2´ Blend 1´ Run 1E´

Fig. 6. The tensile experiment results of (a) core-shell(Series D′) and (b) blend (Blends 1′-4′) PSAs. Thestress-strain behavior of amorphous PSA (Run 1E′) isalso included. The tests were carried out at 23 °C and55% relative humidity.

Run 1E´ Run 2D´ Blend 2´

Run 4D´ Blend 4´

Fig. 7. AFM phase images of the cross-section of the films cast from Runs 1E′, 2D′, Blend 2′, Run 4′ and Blend 4′.

E. Mehravar et al. European Polymer Journal 98 (2018) 63–71

68

between the cores in the adhesive film can be estimated as follows [40]:

⎜ ⎟= ⎛⎝

⎛⎝

∅∅

⎞⎠

− ⎞⎠

IPS d 1cm

1/3

where dc is the diameter of the core, ∅m the maximum packing whichfor monodispersed cores can be taken as 0.64 (assuming randompacking), and ∅ the volume fraction of the cores. For the core-shells2D′ and 4D′, the estimated distances between cores are 80 nm and29 nm, respectively. The molar masses of the nanogels that fit inspheres of these diameters diameters (assuming single polymer chain(polymer network) in the polymer particle and polymer density 1.1 g/cm3) are 1.8×108 and 8.5× 106 g/mol, respectively. These molarmasses are much smaller than those measured for the nanogels in theAF4. Therefore, it looks like there was no room to accommodate thenanogels between the cores and they migrate to the larger spacesamong cores as illustrated in Fig. 9. This broke the interconnectivitybetween nanogels leading to a liquid like behavior.

The AFM micrographs (Fig. S3 in Supporting Information) alsoshows that the air-film interface was relatively free of hard domains andthis may explain the relative insensitiveness of the peel adhesion andloop tack to the SA content.

The adhesives contained crystalline domains that had a meltingpoint of 50 °C. Therefore, they may show a temperature responsivebehavior in the vicinity of the melting point. Fig. 10 presents, the effectof the temperature used in the test on the shear resistance for Latex 4D′and Blend 4′. Latex 1E′ is included as a reference. It can be seen that theshear resistance of the amorphous adhesive (1E′) decreased about 10times when the test temperature increased from 23 to 51 °C. On theother hand, in the same temperature interval, adhesive 4D′ decreasedabout 40 times with the decrease accelerated in the melting region.However, neither 1E′ nor 4D′ showed any accelerated decrease in themelting region (42–52 °C). The most striking result was that of Blend 4′that evolved from no failure in more than one week at both 23 °C and

30 °C to a decrease of more than two orders of magnitude in the meltingregion leading to a completely liquid-like behavior above this region.

Loop tack and peel adhesion of the adhesives containing crystallinedomains were also very sensitive to the temperature used in the test(Fig. 11), and in general the effect increased with the SA content andwas stronger for blends than for core-shell (the effect on the amorphous1E′ was weak).

These results show that the presence of crystalline domains of poly(SA) does not guarantee the temperature-responsiveness of the ad-hesives. In addition, these domains should be dispersed in a continuoussoft phase containing entangled nanogels. The lack of entanglementsbetween nanogels leads to liquid like behavior and poorer adhesiveproperties. Temperature responsiveness is of practical importance be-cause strong adhesives can easily be removed by moderate heatingthem above melting point of crystalline domains. It is worth mentioningthat temperature-responsiveness of the adhesives can be modulateswith a variety of crystallizable long side chain polymers with meltingtemperature between −35 °C (poly(lauryl methacrylate)) to 60 °C (poly(behenyl acrylate)) which open this possibility to be used in differentapplications.

4. Conclusions

In this work, hard core-soft shell pressure sensitive adhesives (PSAs)were synthesized by miniemulsion polymerization and compared withPSAs of the same overall concentration prepared by blending hard andsoft latexes. A semicrystalline hard phase (poly(octadecyl acrylate))was used aiming at thermal responsiveness.

Although both types of adhesives had better adhesive properties

0

0.5

1

1.5

20 30 40 50 60 70

Run 1A

Run 1E'

Run 2D'

Blend 2'

Run 4D'

Blend 4'

Rel

ativ

e Sc

ale

Elution Time (min)

104

105

106

107

108

109

1010

1011

1012

20 30 40 50 60 70

Run 1A

Run 1E'

Run 2D'

Blend 2'

Run 4D'

Blend 4'

Mol

ar M

ass

(g m

ol-1

)Elution Time (min)

(a) (b)

Fig. 8. (a) RI chromatogram and (b) the molar mass vselution time plots for Run1A, Run 1E′, Run 2D′ Run4D′and Blend 2′.

2EHA/MMA/MAA nanogel

poly(SA)

2EHA/MMA/MAA smaller chains

Fig. 9. Possible accommodation of the second stage nanogels in the voids among poly(SA)core particles in Series D′ PSA films.

1

10

100

1000

10000

20 30 40 50 60 70

Shea

r re

sist

ance

(min

)*

Temperature (°C)

Blend 4´ Latex 4D´ Latex 1E´

melting region

Fig. 10. Effect of the test temperature on the shear resistance of the latexes 4D′, Blend4′and 1E′. *The shear resistance tests of Blend 4′ adhesive at 23 and 30 °C were withoutfailure after one week. The Y axes are in Log scale.

E. Mehravar et al. European Polymer Journal 98 (2018) 63–71

69

than the soft polymer, there were strong differences between them. Itwas found that the blends have much better shear resistance withoutsignificant reduction in peel adhesion, loop tack and work of adhesion.The differences were not due to variations in the molar mass distribu-tion and were attributed to the film morphology that in the case ofblends was a distribution of hard domains in a continuous soft matrix,whereas for core-shells was a homogenous distribution of hard spheresseparated by a thin honeycomb structure of soft polymer. The distancebetween cores, which determined the thickness of the homogenousfilm, was that small that there was no room to accommodate the na-nogels between the cores. Consequently, they migrated to the largerspaces among particles breaking the interconnectivity between nano-gels and leading to a liquid like behavior and poorer adhesive proper-ties.

The crystalline domains provided strong temperature-responsive-ness to the blend adhesives, which is of practical importance as strongadhesives can be easily removed by moderate heating. The removaltemperature can be fine tuned modifying the melting temperature ofthe crystalline domains, which for the case of long side chain (meth)acrylates is easily done by varying the length of the side chain.

Acknowledgements

The financial support of the Industrial Liaison Program onPolymerization in Dispersed Media (Akzo Nobel, Allnex, Arkema,Asianpaint, BASF, DSM, Inovyn, Solvay, Stahl, Synthomer, Vinavil,Wacker) is gratefully acknowledged. The authors would like to grate-fully acknowledge Dr. Frank Pirrung for technical support in mini-emulsification process, Ms. Christina Laue for assistance in synthesizingthe latexes, Ms. Melanie Hook for assistance in studying the adhesiveproperties and Dr. Loli Martin and Ms. Elodie Limousin for AFM study.

Appendix A. Supplementary material

The SA miniemulsion preparation and synthesis of latexes; the de-tailed description of the characterization methods and adhesive and

mechanical tests; AFM phase and height images of the air-film interfaceof the films cast from Runs 1E′, 2D′, Blend 2′, Run 4′ and Blend 4′; firstheating DSC scans and the first derivative of heat capacity for Runs 4D′and Blend 4′; The comparison of gel fraction measured using both AF4and soxhlet extraction techniques.. Supplementary data associated withthis article can be found, in the online version, at http://dx.doi.org/10.1016/j.eurpolymj.2017.11.004.

References

[1] I. Benedek, M.M. Feldstein, Handbook of Pressure-Sensitive Adhesives and ProductsFundamentals of Pressure Sensitivity, CRC, Taylor & Francis, Boca Raton (FL),London (MN), New York (NY), 2009.

[2] S. Ebnesajjad, A.H. Landrock, Adhesives Technology Handbook, William Andrew,2014.

[3] S.D. Tobing, A. Klein, Molecular parameters and their relation to the adhesiveperformance of acrylic pressure-sensitive adhesives, J. Appl. Polym. Sci. 79 (2001)2230–2244.

[4] M.D. Gower, R.A. Shanks, The effect of varied monomer composition on adhesiveperformance and peeling master curves for acrylic pressure-sensitive adhesives, J.Appl. Polym. Sci. 93 (2004) 2909–2917.

[5] R. Jovanović, M.A. Dubé, Emulsion-based pressure-sensitive adhesives: a review, J.Macromol. Sci. Polym. Rev. 44 (2004) 1–51.

[6] A. Lopez, E. Degrandi-Contraires, E. Canetta, C. Creton, J.L. Keddie, J.M. Asua,Waterborne polyurethane−acrylic hybrid nanoparticles by miniemulsion poly-merization: applications in pressure-sensitive adhesives, Langmuir 27 (2011)3878–3888.

[7] A. Lopez, E. Degrandi, E. Canetta, J.L. Keddie, C. Creton, J.M. Asua, Simultaneousfree radical and addition miniemulsion polymerization: effect of the diol on themicrostructure of polyurethane-acrylic pressure-sensitive adhesives, Polymer 52(2011) 3021–3030.

[8] R. Udagama, E. Degrandi-Contraires, C. Creton, C. Graillat, T.F. McKenna,E. Bourgeat-Lami, Synthesis of acrylic− polyurethane hybrid latexes by mini-emulsion polymerization and their pressure-sensitive adhesive applications,Macromolecules 44 (2011) 2632–2642.

[9] E. Degrandi-Contraires, R. Udagama, E. Bourgeat-Lami, T.F. McKenna, K. Ouzineb,C. Creton, Mechanical properties of adhesive films obtained from PU−acrylic hy-brid particles, Macromolecules 44 (2011) 2643–2652.

[10] A. Lopez, Y. Reyes, E. Degrandi-Contraires, E. Canetta, C. Creton, J.L. Keddie,J.M. Asua, Simultaneous free-radical and addition miniemulsion polymerization:effect of the chain transfer agent on the microstructure of polyurethane-acrylicpressure-sensitive adhesives, Macromol. Mater. Eng. 298 (2013) 53–66.

[11] V. Daniloska, P. Carretero, R. Tomovska, M. Paulis, J.M. Asua, High-performanceadhesives resulting from spontaneous formation of nanogels within miniemulsion

0

1

2

3

4

5

0 10 20 30 40 50

Loo

p ta

ck (N

/25m

m)

SA (wt%)

Series D´ 23 ºC 53 ºC

0

1

2

3

4

5

0 10 20 30 40 50

Loo

p ta

ck (N

/25m

m)

SA (wt%)

Blends 23 ºC 53 ºC

0123456789

10

0 10 20 30 40 50

Peel

res

ista

nce

(N/2

5mm

)

SA (wt%)

Series D 23 ºC 53 ºC

0123456789

10

0 10 20 30 40 50

Peel

res

ista

nce

(N/2

5mm

)

SA (wt%)

Blends

23 ºC

53 ºC´

Fig. 11. Loop tack strength and Peel adhesion at 23and 53 °C of the latexes synthesized in Series D′ (la-texes 1D′–4D′), blends (Blends 1′–4′) and latex 1E′.

E. Mehravar et al. European Polymer Journal 98 (2018) 63–71

70

particles, ACS Appl. Mater. Interfaces. 6 (2014) 3559–3567.[12] V. Daniloska, P. Carretero, R. Tomovska, J.M. Asua, High performance pressure

sensitive adhesives by miniemulsion photopolymerization in a continuous tubularreactor, Polymer 55 (2014) 5050–5056.

[13] F. Deplace, M.A. Rabjohns, T. Yamaguchi, A.B. Foster, C. Carelli, C.H. Lei,K. Ouzineb, J.L. Keddie, P.A. Lovell, C. Creton, Deformation and adhesion of aperiodic soft–soft nanocomposite designed with structured polymer colloid parti-cles, Soft Matter 5 (2009) 1440–1447.

[14] T. Wang, C.H. Lei, D. Liu, M. Manea, J.M. Asua, C. Creton, A.B. Dalton, J.L. Keddie,A molecular mechanism for toughening and strengthening waterborne nano-composites, Adv. Mater. 20 (2008) 90–94.

[15] G. Diaconu, M. Micusik, A. Bonnefond, M. Paulis, J.R. Leiza, Macroinitiator andmacromonomer modified montmorillonite for the synthesis of acrylic/MMT nano-composite latexes, Macromolecules 42 (2009) 3316–3325.

[16] T. Wang, A.B. Dalton, J.L. Keddie, Importance of molecular friction in a softpolymer−nanotube nanocomposite, Macromolecules 41 (2008) 7656–7661.

[17] T. Wang, P.J. Colver, S.A. Bon, J.L. Keddie, Soft polymer and nano-clay supra-colloidal particles in adhesives: synergistic effects on mechanical properties, SoftMatter 5 (2009) 3842–3849.

[18] R.S. Gurney, D. Dupin, J.S. Nunes, K. Ouzineb, E. Siband, J.M. Asua, S.P. Armes,J.L. Keddie, Switching off the tackiness of a nanocomposite adhesive in 30 s viainfrared sintering, ACS Appl. Mater. Interfaces. 4 (2012) 5442–5452.

[19] A. Bonnefond, M. Mičušík, M. Paulis, J.R. Leiza, R.F. Teixeira, S.A. Bon,Morphology and properties of waterborne adhesives made from hybrid polyacrylic/montmorillonite clay colloidal dispersions showing improved tack and shear re-sistance, Colloid. Polym. Sci. 291 (2013) 167–180.

[20] V. Daniloska, J.L. Keddie, J.M. Asua, R. Tomovska, MoS2 nanoplatelet fillers forenhancement of the properties of waterborne pressure-sensitive adhesives, ACSAppl. Mater. Interfaces. 6 (2014) 22640–22648.

[21] Y. Yamamoto, S. Fujii, K. Shitajima, K. Fujiwara, S. Hikasa, Y. Nakamura, Softpolymer-silica nanocomposite particles as filler for pressure-sensitive adhesives,Polymer 70 (2015) 77–87.

[22] A. Agirre, C.D.L. Heras-Alarcón, T. Wang, J.L. Keddie, J.M. Asua, Waterborne,semicrystalline, pressure-sensitive adhesives with temperature-responsiveness andoptimum properties, ACS Appl. Mater. Interfaces. 2 (2010) 443–451.

[23] J.M. Asua, Miniemulsion polymerization, Prog. Polym. Sci. 27 (2002) 1283–1346.[24] J.M. Asua, Challenges for industrialization of miniemulsion polymerization, Prog.

Polym. Sci. 39 (2014) 1797–1826.[25] E. Mehravar, J.R. Leiza, J.M. Asua, Synthesis and characterization of comb-like

acrylic-based polymer latexes containing nano-sized crystallizable domains,

Polymer 84 (2016) 167–177.[26] E. Mehravar, J. Leswin, B. Reck, J.R. Leiza, J.M. Asua, Waterborne paints con-

taining nano-sized crystalline domains formed by comb-like polymers, Prog. Org.Coat. 106 (2017) 11–19.

[27] E. Mehravar, J.R. Leiza, J.M. Asua, Performance of latexes containing nano-sizedcrystalline domains formed by comb-like polymers, Polymer 96 (2016) 121–129.

[28] C. Plessis, G. Arzamendi, J.R. Leiza, H.A. Schoonbrood, D. Charmot, J.M. Asua,Modeling of seeded semibatch emulsion polymerization of n-BA, Ind. Eng. Chem.Res. 40 (2001) 3883–3894.

[29] C. Plessis, G. Arzamendi, J.R. Leiza, H.A. Schoonbrood, D. Charmot, J.M. Asua,Kinetics and polymer microstructure of the seeded semibatch emulsion copoly-merization of n-butyl acrylate and styrene, Macromolecules 34 (2001) 5147–5157.

[30] I. González, J.M. Asua, J.R. Leiza, The role of methyl methacrylate on branchingand gel formation in the emulsion copolymerization of BA/MMA, Polymer 48(2007) 2542–2547.

[31] D.L. Esmay, Minnesota Mining and Manufacturing Company. Removable pressure-sensitive adhesive tape. U.S. Patent 4599265, 1986.

[32] Z. Czech, Crosslinking of pressure sensitive adhesive based on water-borne acrylate,Polym. Int. 52 (2003) 347–357.

[33] A. Zosel, The effect of fibrillation on the tack of pressure sensitive adhesives, Int. J.Adhes. Adhes. 18 (1998) 265–271.

[34] H. Lakrout, P. Sergot, C. Creton, Direct observation of cavitation and fibrillation in aprobe tack experiment on model acrylic pressure-sensitive-adhesives, J. Adhes. 69(1999) 307–359.

[35] H. Lakrout, C. Creton, D. Ahn, K.R. Shull, Influence of molecular features on thetackiness of acrylic polymer melts, Macromolecules 34 (2001) 7448–7458.

[36] F. Deplace, C. Carelli, S. Mariot, H. Retsos, A. Chateauminois, K. Ouzineb, C. Creton,Fine tuning the adhesive properties of a soft nanostructured adhesive with rheo-logical measurements, J. Adhes. 85 (2009) 18–54.

[37] C.A. Dahlquist, Pressure-sensitive adhesives, in: L.R. Patrick (Ed.), Treatise onAdhesion and Adhesives, vol. 2, Dekker, New York, 1969, pp. 219−260.

[38] A.C. Makan, T. Otte, H. Pasch, Analysis of high molar mass branched poly-butadienes by SEC-MALLS and AF4-MALLS, Macromolecules 45 (2012) 5247–5259.

[39] S. Podzimek, J. Machotova, J. Snuparek, M. Vecera, L. Prokupek, Characterizationof molecular structure of acrylic copolymers prepared via emulsion polymerizationusing A4F-MALS technique, J. Appl. Polym. Sci. 131 (2014), http://dx.doi.org/10.1002/APP.40995.

[40] T. Hao, R.E. Riman, Calculation of interparticle spacing in colloidal systems, J.Colloid Interface Sci. 297 (2006) 374–377.

E. Mehravar et al. European Polymer Journal 98 (2018) 63–71

71