Taguchi-based optimization of adhesion of polyurethane …insite.cavs.msstate.edu/publications/docs/2007/04/2389JAST_vol21... · Taguchi-based optimization of adhesion of polyurethane

J. Adhesion Sci. Technol., Vol. 21, No. 8, pp. 705–724 (2007) VSP 2007.Also available online - www.brill.nl/jast

Taguchi-based optimization of adhesion of polyurethane toplasticized poly(vinyl chloride) in synthetic leather

S. NOURANIAN, H. GARMABI ∗ and N. MOHAMMADIPolymer Engineering Department, Amir Kabir University of Technology,P.O. Box: 15875-4413, Tehran, Iran

Received in final form 25 February 2007

Abstract—In this investigation, the effects of formulation and processing factors on the adhesionbetween polyurethane (PU) and plasticized poly(vinyl chloride) (pPVC) layers was studied using theTaguchi method for experimental design. Among the many factors, nine parameters were screenedand tested at two or three levels, taking advantage of the Taguchi L27 orthogonal array. The factorsstudied were PVC type, PVC K-value, plasticizer type and content, filler type and content, fusiontemperature and fusion time of PVC, and PU type. Using the results of T-peel adhesion test at 60◦Cas a response, the data were analyzed by appropriate software based on the ANOVA technique. Theeffect of the various factors on the adhesion was found to be in the following descending order:PU type, PVC fusion temperature, PVC type, plasticizer content, PVC K-value, PVC fusion time,interaction between plasticizer type and PVC fusion temperature, plasticizer type, filler type, fillercontent, interaction between filler type and PVC fusion temperature, and interaction between PVCK-value and filler type.

Keywords: Adhesion; polyurethane; poly(vinyl chloride); Taguchi method; synthetic leather.

1. INTRODUCTION

Adhesion between dissimilar polymers and relevant phenomena have challengedtoday’s technologists and researchers. The complexity of the mechanisms involvedin the achievement of a satisfactory bond has added to the research efforts neededto solve the problems faced in industrial applications. From the theoretical point ofview, two main mechanisms of adhesion are responsible for the interfacial adhesionbetween polymeric materials: adsorption and diffusion [1]. Though different innature, neither of them can be accepted as the sole mechanism of adhesion inpolymers. Although somewhat controversial, the diffusion mechanism introduced

∗To whom correspondence should be addressed. Tel.: (98-21) 6454-2428; e-mail:[email protected]

706 S. Nouranian et al.

by Voyutskii in 1949 [2] explains with good accuracy the phenomena relevant to theinterfacial adhesion of high polymers with similar solubility parameter (δ).

The thermodynamic work of adhesion (WA) is another concept based on adsorp-tion theory of adhesion [1], which relates the work needed to separate two adheredsurfaces to their surface free energies (equation (1)).

WA = 2(γ Da γ D

s )1/2 + 2(γ Pa γ P

s )1/2, (1)

where γ is the surface free energy, subscripts a and s stand for adhesive andsubstrate, and superscripts D and P stand for dispersion and polar components ofsurface free energy, respectively.

In the industrial practice, adhesion plays an important role. In the syntheticleather industry, where a lack of a systematic study of the phenomena involvedin different applications has been felt for a long time, so much concern has beenfocused on the problems relative to adhesion, especially as more new productsare being manufactured. Among these products is a hybrid leather composed of apolyurethane topcoat and plasticized poly(vinyl chloride) (pPVC) skin layer. Here,two dissimilar polymers with a close solubility parameter (δ = 9.7 cal1/2/cm3/2

for PVC and 9.5–10 cal1/2/cm3/2 for PU) are brought in contact and processed tothe final product. Experience has revealed that the selection of inappropriate layermaterials and formulations and the use of inappropriate process conditions leads toweakness in the final product in many cases, showing this weakness as a diversity offailure phenomena such as delamination (debonding), non-uniform properties in thetensile and flexure tests, etc. In this study, we conducted a thorough investigationof the parameters affecting the adhesion between polyurethane (PU) and pPVClayers in synthetic leather in order to reach some resolutions on the aforementionedproblems. Due to the complexity of physical and chemical factors involved in theachievement of a satisfactory bond between these layers, an initial differentiationwas made between formulation and process-related parameters.

The Taguchi method for the design of experiments (DOE) [3] was chosen as atool for the organization of the experiments, evaluation of the effects of variousparameters and interpretation of the results. It has to be mentioned that so far noreports have been published on the use of the Taguchi method in the transfer coatingindustry. Transfer or spread coating is the main method for the manufacturing ofsynthetic leather.

Problems with the adhesion of polyurethane to PVC were recognized previouslyin many applications. Several patents have been granted to inventors of adhesionpromotion methods between PU and PVC [4–6].

In the work of Leriche and Turin [4], a non-pretreated PVC layer was covered witha two-component polyurethane film in the presence of a plasticizer compatible withboth polymers (e.g., diisooctyl phthalate or DIOP) with a content of 2–12 wt% basedon PU resin. Curing of the PU film was conducted while pressure was applied to thesurface. According to the inventors, the adhesion improved through the diffusion-promoting effect of the plasticizer in the interfacial region.

Study of the parameters affecting adhesion of PU to pPVC 707

In an attempt to introduce chemical bonds of the covalent type at the interfaceof a PU/PVC joint, Boba and co-workers [5] used a completely different approach.They added a chemical containing active hydrogen groups (e.g., hydroxy, carboxy,amino, etc.) to the PVC layer, while making available free isocyanate groups in thepolyurethane layer. Of said chemical 0.5–20 wt% was added. The amount of freeisocyanate was in the range of 0.5–10 wt%. After completion of chemical reactionbetween active hydrogens and isocyanate groups in the two layers (during the curingprocess), a strong chemical bond was achieved, resulting in high adhesion strength.

Fogle and Cooley [6] used N-substituted 2-pyrrolidone and/or ethoxylated alkylphenol as an effective adhesion-promoting agent for PU and PVC. Based ontheir reports, an amount of 1–15 vol% of the said agents would suffice for theachievement of acceptable adhesion strength. After an adequate contact timebetween the layers, these agents were removed from the interface through reactionwith an organic isocyanate.

In the present study, our main objective was to investigate the parameters affectingadhesion between PU and plasticized PVC at elevated processing temperatures.Both formulation and processing conditions were taken into account.

2. EXPERIMENTAL

2.1. Selection of parameters

In order to conduct a systematic assessment of interfacial adhesion between pPVCand PU layers, an initial study was made on the physical and chemical factors likelyto have influence on the adhesion. In this connection, the role of processing andformulation conditions in achieving a satisfactory bond between the said layers wastaken into account. Nine factors were screened for the study as shown in Table 1.

2.2. Materials

The materials and processing conditions chosen for the sample preparation wereselected based on the most common practice in the synthetic leather industry.

Table 1.Factors screened for the study of adhesion

Factor No. Factor Relationship1 PVC type PVC formulation2 PVC K-value PVC formulation3 Plasticizer type PVC formulation4 Plasticizer content PVC formulation5 Filler type PVC formulation6 Filler content PVC formulation7 PVC fusion temperature Process condition8 PVC fusion time Process condition9 PU type PU formulation


2.2.1. Ingredients of PVC and PU layers. Two main PVC types were used inthe formulation of PVC plastisols, namely emulsion and microsuspension. Othernormal components of the PVC formulation were heat stabilizers, co-stabilizers,plasticizers and fillers.

For the PU layer, polyester-type aromatic and aliphatic polyurethane solutionswere used as the pre-skin to be coated on top of PVC layer. Dimethyl formamide(DMF) was used as the solvent for aromatic-type, while a 1:1 ratio of toluene(TOL) to isopropyl alcohol (IPA) was used for aliphatic-type PU. The solvent isa regular component of any PU formulation used in the transfer coating processand is completely removed during processing leading to a thin film. In order toobtain a better contrast between PU and PVC layers, the PU layer was prepared inblack color, while PVC layers were colorless. The material grades and suppliers areshown in Table 2. In addition, some general characteristics of PU and PVC gradesare summarized in Table 3.

2.2.2. Release paper. A semi-matt plain release paper, Alfakote CR (CartieraDi Crusinallo Favini, Italy), was used for the coating of PVC paste and samplepreparation.

2.3. Equipment

2.3.1. Sample preparation equipment. For the preparation of PVC pastes and PUsolutions, a laboratory-scale mixer with variable speed (200–2000 rpm) was used.Samples were coated on the release paper using a laboratory coating blade and fused(or dried) in the oven of a laboratory-scale, two-head coating machine (Colombo,Italy).

2.3.2. Sample testing equipment. A tensile testing machine with a load capacityof 5 metric tons (Santam, Iran) equipped with a heating chamber with temperatureregulation in the range of 50–450◦C was used for the acquisition of peel test data.For the calculation of sample surface energy, a contact angle measuring instrumentwas used (Krüss Model G10, Germany).

2.4. Design of experiments

The Taguchi method for the design of experiments [3] was chosen for the organiza-tion of the experiments and analysis of the results. Using the L27 orthogonal array,a mixed-level design (six factors at three levels and three factors at two levels) wasmade. The use of more than two levels for factors makes it possible to study someeventual non-linear effects. The interactions between the factors were considerednegligible. The factor levels were chosen in accordance with the common industrialpractice and in a range suitable for preparing the maximum number of samples inall 27 “treatments”. It has to be mentioned that each experimental setup is called atreatment. The factor designations and levels are shown in Table 4.


Table 2.Material trade names and specifications

Ingredient Trade name Specification SupplierPVC layer

PVC powder E 7012 Emulsion Vestolit, GermanyPVC powder P 1345 K 80 Emulsion Vestolit, GermanyPVC powder P 1430 K 90 Emulsion Vestolit, GermanyPVC powder B 7021 Microsuspension Vestolit, GermanyHeat stabilizer CBZ 277 Ba/Cd/Zn stabilizer Dahin Group, TaiwanCo-stabilizer ESBO-132 Epoxidized soybean oil Dahin Group, TaiwanPlasticizer DOP Dioctyl phthalate Farabi Petrochem., IranPlasticizer DOA Dioctyl adipate Dahin Group, TaiwanPlasticizer (Unimoll BB) (BBP) Butylbenzyl phthalate Bayer, GermanyFiller – Calcium carbonate Alborz Carbonate, IranFiller – Talc Towlipoudr, Iran

PU layerPU pre-skin Larithane AL 233N Aliphatic Novotex Italiana, ItalyPU pre-skin Larithane MS 128 Aromatic Novotex Italiana, ItalySolvent DMF Dimethyl formamide Caldic, The NetherlandsSolvent TOL Toluene Isfahan Petrochem., IranSolvent IPA Isopropyl alcohol –Leveling agent Noresil S 900 Liquid Novotex Italiana, ItalyPigment Tecnopur TP 3035 Paste Rohm & Haas, Italy

Table 3.Some general characteristics of PVC and PU grades

Specification Grade

E 7012 P 1345 K 80 P 1430 K 90 B 7021PVC

K-value 67 80 90 70Viscosity number (cm3/g) 112 170 227 125Particle dimensions (Sieve analysis –

retained on 0.063 mm sieve) (%) <1 <1 <3 <1

AL 233 N MS 128

PUType Aliphatic, polyester Aromatic, polyesterMolecular weight (g/mol) 50 000 50 000Isocyanate structure Blend of IPDIa and HMDIb MDIc

a Isophorone diisocyanate.b Dicyclohexyl-4,4′-methylene diisocyanate (hydrogenated MDI).c 4,4′-methylene diphenyl diisocyanate.

The factors were assigned to the columns of an array as designated in Table 4.Note that factor A has three K-values (67, 80 and 90) corresponding to its firstlevel and only one K-value (70) corresponding to its second level. Therefore,


Table 4.Designations and levels of factors

Factor Factor No. of Level 1 Level 2 Level 3designation levelsA PVC type 2 Emulsion Microsuspensiona –B PVC K-value 3 67 80 90C Plasticizer type 3 DOP DOA BBPD Plasticizer content (phr) 3 50 53 57E Filler type 2 Calcium carbonate Talc –F Filler content (phr) 3 7 10 13G PVC fusion temperature (◦C) 3 170 175 180H PVC fusion time (min) 3 3 3.5 4I PU type 2 Aliphatic Aromatic –

a The microsuspension type of PVC has a K-value of 70.

a branched design had to be chosen as follows: (1) factors A and B were designatedto columns 1 and 2 of the orthogonal array (Table 5), (2) the interaction columns[3] for columns 1 and 2 (i.e., columns 3 and 4) were left blank and (3) level 1 offactor B (K-value = 67), corresponding to level 2 of factor A (microsuspensiontype PVC), was replaced by the correct K-value of 70. Other factors were assignedto columns 5–11 (Table 5).

The mixed-level nature of the orthogonal array required “dummy treatment” forall two-level factors, i.e., A, E and I [3]. In dummy treatment, a third level isassumed for two-level factors using either the value for level 1 or 2. Based onthis approach, the value of level 1 for factor A was chosen for the third level of thisfactor. In the same manner, the third level of factor E acquired the value of level 1for this factor, and the third level of factor I acquired the value of level 2 for thisfactor. The final design is shown in Table 5.

2.5. Sample preparation

Table 6 shows the formulations of PVC and PU layers used for the preparation ofsamples.

2.5.1. Preparation of the pastes. For the PVC part, solid ingredients (PVC andfiller) were added gradually to the liquids (plasticizer, heat stabilizer, and epoxidizedsoybean oil) while the paste was mixed with an initial speed of 600 rpm. The speedwas then increased to 1600 rpm. After 5 min, a homogeneous paste was obtained,which was kept unstirred for 24 h in order to eliminate air bubbles introduced duringthe mixing process.

PU solutions were mixed gently using a steel bar for approximately five minutes.The total rpm for the mixing was around 300. The order of addition of ingredientswas PU solution, solvent, leveling agent, and black pigment. The deaeration(degassing) of PU solutions was complete in 2 h.


Table 5.The final design of experiments and the corresponding factor levels

Treatment A B C D E F G H I1 2 3 4 5 6 7 8 9 10 11 12 13

1 1 1 1 1 1 1 1 1 12 1 1 2 2 2 2 2 2 23 1 1 3 3 1 3 3 3 24 1 2 1 1 1 2 2 2 25 1 2 2 2 2 3 3 3 16 1 2 3 3 1 1 1 1 27 1 3 1 1 1 3 3 3 28 1 3 2 2 2 1 1 1 29 1 3 3 3 1 2 2 2 1

10 2 (1) 1 2 1 1 2 3 111 2 (1) 2 3 1 2 3 1 212 2 (1) 3 1 2 3 1 2 213 2 (1) 1 2 1 2 3 1 214 2 (1) 2 3 1 3 1 2 115 2 (1) 3 1 2 1 2 3 216 2 (1) 1 2 1 3 1 2 217 2 (1) 2 3 1 1 2 3 218 2 (1) 3 1 2 2 3 1 119 1 1 1 3 2 1 3 2 120 1 1 2 1 1 2 1 3 221 1 1 3 2 1 3 2 1 222 1 2 1 3 2 2 1 3 223 1 2 2 1 1 3 2 1 124 1 2 3 2 1 1 3 2 225 1 3 1 3 2 3 2 1 226 1 3 2 1 1 1 3 2 227 1 3 3 2 1 2 1 3 1

The Taguchi L27 orthogonal array can accommodate 13 different factors, indicated in this tableby 13 columns. The total number of treatments (experimental setups) is 27. In this study ninefactors were investigated, indicated by letters A to I (Table 4). These factors were designated ascolumns 1–11, intentionally leaving columns 3 and 4 blank. Numbers 1, 2 and 3 indicate levels ofthe factors as described in Table 4. Level 1 in parentheses (column B) indicates K-value 70 replacingK-value 67 as per the branched design described in the text.

2.5.2. Preparation of the test samples. Samples were prepared in a sandwichform on the release paper support, where a thin PU layer was sandwiched betweentwo thicker PVC layers using a laboratory-scale coating knife. This was donein order to minimize the unwanted effects of non-uniform elongation of PU andPVC layers during the peel test. PVC fusion time and temperature were set inaccordance with the factor levels for each treatment in the design. First, a 1.00-mm-thick layer of PVC plastisol was coated on the release paper with the aid ofcoating knife and heated in an oven at the temperature specified by the treatmentsetup (Table 4). Coating was obtained by pouring the plastisol on the release paperbehind the coating knife and pulling the release paper with a constant speed through


Table 6.PVC and PU formulations used for the preparation of the samples

Ingredient phra

Formulation of PVC layersPVC powder 100Plasticizer (level 1/2/3) 50/53/57Filler (level 1/2/3) 7/10/13Heat stabilizer 2Epoxidized soybean oil 2

Formulation of PU layersPU solution 100Solvent for aliphatic PU (1:1 ratio of TOL to IPA) 20Solvent for aromatic PU (DMF) 25Leveling agent 0.6Black pigment paste 3a part per hundred resin.

the knife gap. Then, a thin PU layer was coated on top of the fused PVC layerin the same manner. For the PU layer, drying of the coated film was conductedin an oven in two successive heating cycles: (1) 2 min at 80◦C and subsequently2 min at 100◦C for aliphatic PU and (2) 2 min at 90◦C and subsequently 2 minat 130◦C for aromatic PU. After completion of film drying, another layer of PVCplastisol with a thickness of 1.00 mm was coated over the dried PU layer and fusedin the oven, again at the temperature specified by the treatment setup (Table 4). Thefinal samples were cut into standard dimensions (2 × 200 mm) as specified in theASTM D 1876-95 standard.

2.6. Sample testing

The T-peel test (ASTM D 1876-95) was used for the acquisition of peel forcedata. The peeling was conducted at 60◦C, because at room temperature, severalsamples experienced cohesive failure in the PVC layer resulting in no data. Peelforce measurements were conducted on three separate samples for each treatment.The peel rate was 200 mm/min and the average peel force was calculated using theobtained peel force curves for these samples. Some samples did not yield peel forcecurves because of the early cohesive failure in the PVC layer even at the elevatedtemperature.

For eight selected samples, the thermodynamic work of adhesion (WA) [1] wascalculated using the Kaelble–Rabel–Wendt–Owens and Wu methods [9–11].

3. RESULTS AND DISCUSSION

As mentioned before, peel force was chosen as the response (y), in order to analyzethe interfacial adhesion between PU and pPVC layers. Among the 27 samples,


Table 7.Response values (average peel forces) for the treatments

Treatment Peel force (kN/m)1 –2 2.53 2.04 2.35 3.46 0.97 2.28 1.69 –

10 –11 2.912 3.113 3.314 –15 1.916 2.017 2.018 –19 –20 1.921 2.122 1.523 –24 1.925 1.926 2.427 –

19 tests gave a set of data adequate for the calculation of average peel force. Dueto the strong adhesion between the layers, eight samples could not be peeled and,therefore, no experimental results were obtained for these samples. As mentionedpreviously, this was due to the cohesive failure in the PVC layers during peeling.Table 7 shows the response values measured for the treatments.

3.1. Linear regression modeling and data analysis

As the Taguchi method for experimental design is a fully saturated design, treat-ments without response affect negatively the determination of the significance offactors. Therefore, a semi-empirical model was derived using the response valuesof 19 successful samples. A linear regression method [7, 8] was used for this pur-pose. The derived equation was:

y = 2.6382 + 0.1509XA − 0.2932XB − 0.1648XC − 0.3033XD

+ 0.1456XE + 0.1216XF + 0.4072XG − 0.2289XH − 0.5071XI, (2)


Table 8.Calculated response values (average peel forces) for thetreatments with no actual peel force data

Treatment Peel force (kN/m)1 3.39 2.1

10 3.314 2.918 4.519 3.623 3.527 1.8

Table 9.Results of data analysis using the ANOVA technique

Factor Factor Degrees of Sum of Mean of %designation freedom (DOF) squares squares ContributionI PU type 1 6.15 6.15 35.65G PVC fusion temperature 2 3.08 1.54 17.88A PVC type 1 2.12 2.12 12.27D Plasticizer content 2 1.68 0.84 9.73B PVC K-value 2 1.04 0.52 6.01H PVC fusion time 2 0.97 0.49 5.65C × G Interaction between C and G 4 0.62 0.15 3.57C Plasticizer type 2 0.61 0.30 3.52E Filler type 1 0.51 0.51 2.94F Filler content 2 0.30 0.15 1.74E × G Interaction between E and G 2 0.12 0.032 0.72B × E Interaction between B and E 2 0.022 0.011 0.13

where y is the response and XA, XB . . . XI are variables indicating the respectivelevels of factors A, B . . . I.

Using the above equation, the response values for eight samples without peel forcedata were calculated (Table 8).

Then, through implementation of analysis of variance (ANOVA) technique withthe aid of DesignExpert software version 6.0.6 trial edition, final analysis was madeon peel force data. The results are shown in Table 9.

3.2. Results of thermodynamic work of adhesion (WA)

The thermodynamic work of adhesion is calculated based on the adsorption theoryof adhesion [1]. In this theory, no consideration is given to the diffusion phenom-enon, and only the secondary bonds are taken into account. Two main methods fordetermination of surface energy of solids are harmonic-mean and geometric-meanmethods [11]. The former is proposed by Wu [9] and the latter by Owens, Wendt,Kaelble and Rabel [10]. Both methods use data from contact angle measurements


to calculate the dispersion and polar components of solid surface energy. The ex-pressions for these methods are shown in equations (3)–(6).

Harmonic-mean method:

(1 + cos θ1)γ1 = 4

(γ d

1 γ dS

γ d1 + γ d

S

+ γ P1 γ P

S

γ P1 + γ P

S

), (3)

(1 + cos θ2)γ2 = 4

(γ d

2 γ dS

γ d2 + γ d

S

+ γ P2 γ P

S

γ P2 + γ P

S

). (4)

Geometric-mean method:

(1 + cos θ1)γ1 = 2(√

γ d1 γ d

S +√

γ P1 γ P

S

), (5)

(1 + cos θ2)γ2 = 2(√

γ d2 γ d

S +√

γ P2 γ P

S

), (6)

where θ is contact angle and γ is surface energy. Subscripts 1 and 2 refer totest liquids 1 and 2 and subscript S refers to solid. Superscripts d and P refer todispersion and polar components, respectively.

The harmonic-mean method is mainly used for low-energy surfaces, while thegeometric-mean method is preferred for high-energy surfaces [11]. The results,however, are quite close to each other. In this study, we used both methodsfor the determination of thermodynamic work of adhesion. This was done forcomparison purposes only. The test liquids were water and diiodomethane, and theircontact angles on the surfaces of PU and PVC were measured using an appropriateapparatus described in Section 2.3.2. Using the bundled software, the dispersion andpolar components of the surface energy for PU and PVC were calculated and inputinto equation (1), whereby the thermodynamic work of adhesion was calculated.These results for eight selected samples are shown in Table 10. By determiningthe thermodynamic work of adhesion for selected samples, a comparison betweenpeel force results and the results obtained through harmonic-mean or geometric-mean methods was performed. The peel force measurement includes a significant

Table 10.Thermodynamic work of adhesion (WA) results for eight selected samples usingthe (a) Kaelble, Rabel, Wendt and Owens method and (b) Wu method

Treatment WA (mJ/m2)

(a) (b)1 839 9042 966 10506 753 8287 761 8319 891 95411 1070 115014 968 105023 757 822


viscoelastic loss in the polymeric layers, which will result in a major differencebetween the thermodynamic work of adhesion and the peel force results. Thisrelationship has been discussed previously by several researchers [12–16].

3.3. Main effects of factors

The main effects of nine factors and three interactions were determined by theANOVA technique, which are given as follows.

3.3.1. PVC type. As indicated in Fig. 1, a stronger interfacial adhesion resultedfor microsuspension-type PVC compared to emulsion-type PVC. Contribution ofthis factor to the adhesion between layers is 12.27%, which ranks it as the thirdimportant factor among others (Table 9).

The observed behavior could be attributed to the lack of emulsifier in micro-suspension-type PVC, which results in better and faster fusion of PVC at the sametemperature and time conditions compared to emulsion-type PVC. This results ina better adhesion between the layers. It should be mentioned that a higher andfaster fusion of PVC provides better macromolecular diffusion into the substrateand increases the degree of polymer chain entanglements. This is in accordancewith the diffusion theory of adhesion proposed by Voyutskii [2].

Figure 1. Main effects of PVC type (factor A) and PVC K-value (factor B) on peel force (adhesion).The numbers on the x-axis correspond to factor levels. Note that PVC type is a two-level factor; thus,for this factor, number 3 on the x-axis is an indication of level 2.


3.3.2. PVC K-value (molecular weight). The K-value is a measure of PVCmolecular weight. Figure 1 shows that peel force decreases with increasing K-valueand then levels off. The contribution of this factor to the adhesion is 6.01%, whichmakes it as the fifth important parameter.

Voyutskii [2] has described a similar trend for cellophane tapes bonded topoly(isobutylene) with different molecular weights. Likewise, he observed the lev-eling off of the adhesion strength at higher molecular weights. Based on the dif-fusion theory, with increasing molecular weight, the number of polymer chain freeends at the interface capable of penetrating into the substrate is reduced. There-fore, at a relatively high molecular weight, the middle segments of macromolecularchains (normally with limited diffusion due to steric hindrance) are the only sourcefor diffusion. This results in reduced peel force. Above a certain molecular weight,the diffusion of middle chain segments is the only dominant diffusion mechanismand this will be independent of chain length; therefore, the adhesion strength willnot be affected noticeably by increasing the molecular weight.

3.3.3. Plasticizer type. The contribution of this factor to the adhesion strengthof PVC and PU layers is 3.52%, meaning it is the sixth important parameterinfluencing the interfacial adhesion properties.

As shown in Fig. 2, butylbenzyl phthalate (BBP) resulted in a higher peel forcethan dioctyl phthalate (DOP) and dioctyl adipate (DOA). This could be attributedto the high compatibility of this low-molecular-weight plasticizer (BBP) with

Figure 2. Main effects of plasticizer type (factor C) and plasticizer content (factor D) on peel force(adhesion). Numbers 1, 2 and 3 indicate factor levels as designated in Table 4.


Table 11.Clear-point (solid–gel) temperatures of plasticizers [17]

Plasticizer Clear-point temperature (◦C)Butylbenzyl phthalate (BBP) 102Dioctyl phthalate (DOP) 117Dioctyl adipate (DOA) 138

PVC or in other words, its strong solvating power compared to other plasticizers.BBP results in faster fusion of PVC at lower temperature and, thus, providesa higher diffusion rate. Based on the diffusion theory, a higher intermoleculardiffusion between polymeric layers provides a stronger adhesion. Clear-point(solid–gel transformation) temperature is a good indication of fusion characteristicsof plasticizers. Table 11 shows the clear-point temperatures for the three plasticizersused in this study [17]. As shown in Table 11, BBP has the lowest clear-pointtemperature and, thus, provides the fastest fusion of PVC.

3.3.4. Plasticizer content. This factor has an average contribution of 9.73%,which makes it as the fourth important parameter. Figure 2 shows a reduction ofpeel force with increasing plasticizer content, which was expected.

Voyutskii [2] observed the same trend for the effect of plasticizer contenton adhesion. It has been reported that at low plasticizer content in a highlyviscous adhesive, the addition of plasticizer would lead to a relative increasein adhesion due to the facilitation of molecular motion and thereby interlayermolecular diffusion [2]; however, at higher plasticizer levels, “molecular extraction”or removal of adhesive macromolecule segments from the substrate at the interfaceincreases and, therefore, adhesion strength is reduced sharply.

3.3.5. Filler type and content. The filler type has a minor effect on the adhesionstrength (2.94%) of pPVC and PU layers. Talc powder results in a slightly strongeradhesion compared to calcium carbonate powder.

According to Voyutskii [2], addition of filler to the interface of bonded layersgenerally decreases the adhesion strength. This is mainly due to the attachmentof macromolecules on the surface of filler particles and the resulting decrease inmotion, which in a way leads to decreased diffusion. Furthermore, the presence offiller particles at the interface reduces the effective contact area of the molecules ofthe two layers, leading again to decreased adhesion. However, this is the case whencohesive strength of the polymer is higher than its adhesion strength. Otherwise, theaddition of a surface-active filler, like carbon black, may actually lead to a strongerinterfacial adhesion.

Filler content showed a minor effect on the adhesion, with a contribution of1.74%. It seems that, because the three levels chosen for filler content are so closeto each other, the effect of this factor is not noticeable in this study.


3.3.6. PVC fusion temperature. The sharp increase in peel force with increasingfusion temperature (Fig. 3) is a strong indication of the significance of PVC fusiontemperature on the final adhesion strength. The contribution of this factor is 17.88%,which makes it the second most important parameter.

With reference to the diffusion theory, the increase in adhesion strength withincreasing fusion temperature is mainly due to the facilitation of macromolecularmotion and, more specifically, polymer chain segmental motion. At highertemperatures, the steric hindrance for macromolecular motion is minimized.

3.3.7. PVC fusion time. As indicated in Fig. 3, adhesion strength is reduced withincreasing fusion time. The contribution of this factor is 5.65%, which means it is amoderately important parameter in the overall adhesion property.

The reason for this behavior is thought to be the possible thermal deteriorationof polyurethane or PVC surface during long exposure times at high temperatures.The loss of plasticizer in PVC layer and possible polyurethane degradation at theinterface deteriorates the intermolecular diffusion between the layers and leads to aweaker interfacial adhesion. In order to confirm this hypothesis, thermogravimetricanalysis was conducted on a PVC sheet sample with 57 phr DOP as the plasticizerand an aromatic polyurethane film (Larithane MS 128). The test was conductedisothermally at temperature of 180◦C corresponding to level 3 of PVC fusion

Figure 3. Main effects of PVC fusion temperature (factor G) and PVC fusion time (factor H) on thepeel force (adhesion). Numbers on the x-axis indicate factor levels.


Figure 4. Isothermal TGA graphs for PVC sheet with DOP as plasticizer and aromatic PU(Larithane MS 128) film. The temperature was 180◦C.

temperature as per the Taguchi design described earlier. As shown in Fig. 4, theweight loss initiation was observed after 3 min of exposure time for PVC and 4 minof exposure time for polyurethane. This is in accordance with the fusion time ofPVC during sample preparation step in this study.

3.3.8. Polyurethane type. Among all the factors studied, polyurethane type hasthe highest percentage contribution, namely 35.65%, on the adhesion strength. Asshown in Fig. 5, aliphatic-type PU results in considerably higher adhesion strengthin comparison to the aromatic type.

The polyol part of both PU grades is a polyester type and, therefore, molecularstructure, and intermolecular forces could not be responsible for the enormousdifference in adhesion characteristics. The reason should be attributed to thedifference in diisocyanate groups attached to the polymer chains in the PU resins.

The aromatic diisocyanate possesses a much lesser diffusion rate compared tothe aliphatic diisocyanate, because of the steric hindrance. This may lead tosignificantly lower adhesion strength. The conclusion can be confirmed by acomparison between Tg values of the two polymers. As shown in Figs 6 and 7, theTg for aliphatic-type PU (Larithane AL 233N) is around −50◦C, while it is −20◦Cfor aromatic-type PU (Larithane MS 128). Thus, the macromolecule chain mobilityof aliphatic PU is greater than that of aromatic one.

3.3.9. Interactions of factors. The contributions of the interactions to the totaladhesion strength are generally low and negligible. However, one noticeableinteraction was discovered in this analysis, which is C × G interaction, i.e., aninteraction between plasticizer type and PVC fusion temperature.


Figure 5. Main effect of polyurethane type (factor I) on the peel force (adhesion). Numbers on thex-axis indicate factor levels.

Figure 6. DMTA diagram for aromatic polyester-type PU (Larithane MS 128). The onset of decreasein modulus is considered as the Tg for the polymer.

This interaction has a contribution of 3.57% to the adhesion strength. As shownin Fig. 8, BBP results in the highest adhesion at the lowest temperature (170◦C)among the three plasticizers used for the study. However, at higher temperatures,


Figure 7. DMTA diagram for aliphatic polyester-type PU (Larithane AL 233N).

Figure 8. Main effects of C × G interactions (interaction between plasticizer type (C) and PVCfusion temperature (G)) on the peel force (adhesion). Numbers on the x-axis correspond to the levelsof factor C and numbers on the graphs correspond to the levels of factor G.


this behavior is reversed. The reason probably lies in the fact that BBP has ahigh volatility rate at elevated temperatures. Thus, a high proportion of BBP isevaporated leading to a considerably weaker adhesion.

The other two interactions, i.e., B × E interaction (PVC K-value and filler type)and E × G interaction (filler type and PVC fusion temperature), showed only minorinfluences on the interfacial adhesion of pPVC and PU layers.

4. CONCLUSIONS

The purpose of this study was an in-depth evaluation of the parameters affecting ad-hesion of two dissimilar polymers, namely polyurethane and plasticized poly(vinylchloride). The results were best interpreted using the diffusion theory of adhesionproposed by Voyutskii [2]. This study is in a way a reconfirmation of the basics ofVoyutskii’s diffusion theory for high polymers using the Taguchi method for exper-imental design. Comparing the results with the thermodynamic work of adhesiondata based on adsorption theory of adhesion suggests that the main difference is at-tributed to the viscoelastic loss during sample peeling [12–16], as discussed earlier.

Based on the ANOVA results in the Taguchi method for the design of experiments,the optimum level of factors, i.e., levels corresponding to the highest achievedadhesion strength (peel force), was determined as follows:

A2, B1, C3, D1, E2, F2, G3, H2 and I1,

where the capital letters indicate factors and subscripts point to the factor levels.This study reveals that four main factors influencing the adhesion strength

between PU and pPVC layers are PU type, PVC fusion temperature, PVC type andplasticizer content. The combined contribution of these four factors to the adhesionstrength is more than 70%.

REFERENCES

1. A. J. Kinloch, Adhesion and Adhesives: Science and Technology. Chapman & Hall, London(1987).

2. S. S. Voyutskii, Autohesion and Adhesion of High Polymers. Wiley Interscience, New York, NY(1963).

3. R. Ranjit, A Primer on The Taguchi Method. Van Nostrand Reinhold, New York, NY (1990).4. C. G. A. Leriche and J. A. J. Turin, US Patent No. 6,045,918 (2000).5. J. Boba, S. N. Varadbachary and V. F. Pogozelski, US Patent No. 4,361,626 (1982).6. O. Fogle and J. Cooley, US Patent No. 4,175,161 (1979).7. R. H. Myers and D. C. Montgomery, Response Surface Methodology – Process and Product

Optimization Using Designed Experiments. Wiley, New York, NY (1995).8. C. F. J. Wu and M. Hamada, Experiments – Planning, Analysis, and Parameter Optimization.

Wiley, New York, NY (2000).9. S. Wu, J. Polym. Sci. C34, 19 (1971).

10. D. K. Owens and R. C. Wendt, J. Appl. Polym. Sci. 13, 1741 (1969).


11. S. Wu, Polymer Interface and Adhesion. Marcel Dekker, New York, NY (1982).12. A. Sharif and N. Mohammadi, J. Adhesion Sci. Technol. 16, 33–45 (2002).13. A. N. Gent and J. Schultz, J. Adhesion 3, 281 (1972).14. N. Mohammadi and R. A. Pearson, Proceedings of the 25th International SAMPE Technical

Conference, Philadelphia, PA, p. 655 (1993).15. K. L. Mittal, in: Adhesion Measurement of Films and Coatings, K. L. Mittal (Ed.), pp. 1–13.

VSP, Utrecht (1995).16. A. Sharif, N. Mohammadi and N. Taheri Qazvini, in: Polymer Surface Modification: Relevance

to Adhesion, K. L. Mittal (Ed.), Vol. 3, pp. 477–487. VSP, Utrecht (2004).17. W. V. Titow, PVC Technology, 4th edition, p. 129. Elsevier Applied Science, London (1984).

Documents

Taguchi-based optimization of adhesion of polyurethane …insite.cavs.msstate.edu/publications/docs/2007/04/2389JAST_vol21... · Taguchi-based optimization of adhesion of polyurethane