Kuram2013 Reciclaje Pa 6

Materials and Design 49 (2013) 139–150

Contents lists available at SciVerse ScienceDi rect

Materia ls and Design

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

Investigating the effects of recycling number and injection parameters on the mechanical properties of glass-fibre reinforced nylon 6 using Taguchi method

Emel Kuram a, Enis Tasci b, Ali Ihsan Altan b, Mehmet Metin Medar a, Faruk Yilmaz b, Babur Ozcelik a,⇑a Department of Mechanical Engineering, Gebze Institute of Technology, 41400 Gebze, Kocaeli, Turkey b Department of Chemistry, Gebze Institute of Technology, 41400 Gebze, Kocaeli, Turkey

a r t i c l e i n f o a b s t r a c t

Article history:Received 9 November 2012 Accepted 3 February 2013 Available online 19 February 2013

Keywords:Glass fibre-reinforced polyamide Injection molding RecyclingTaguchi method Regression analyses

0261-3069/$ - see front matter � 2013 Elsevier Ltd. Ahttp://dx.doi.org/10.1016/j.matdes.2013.02.027

⇑ Corresponding author. Tel.: +90 262 605 27 79.E-mail address: [email protected] (B. Ozcelik).

In this study, the glass-fibre reinforced nylon 6 (PA6-GF) was reprocessed in the five processing cycles.The recycl ed PA6-GF samples were characterized by its chemical, thermal, and mechanical properties as a function of the number of processing cycles. It was also investigated how the controlled factors affect the output factors and what the optimal injection settings of the controlled factors can be employed toobtain the best mechanical properties. To achieve these aims Taguchi’s mixed level parameter design (L18) was employed for the experimental design. Number of recycling, melt temperature, mold temperature, injection pressure , and holding pressure were considered as the control factors. Regression analyses were applied to predict tensile strength, yield strength, impact energy and impact strength. Analysis of variance (ANOVA) was used to determine the effects of the control paramete rs on tensile strength,yield strength, impact energy, and impact strength. In the plastic injection of PA6-GF, the numbe r of recycling was found to be the most effective factor on mechanical properties. From the experimental results itwas concluded that there was a decrement in mechanical properties after each reprocessing cycle.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

As plastic materials are widely used in recent years, the increasing consumptio n of plastic material as a consequence of market demand contributes to the large volume of plastic disposal which affects the environment negatively. Since plastic materials are non-degrad able, it takes a long time (up to hundreds of years) tobreak down and disposal of plastics creates space problems . The recycling is one of the most effective methods for diminishing the negative effects of waste plastics on environment. Recycling process reduces both the quantities of plastics sent to landfillsand raw material extraction. Recycling process can also contribute to the economy of all countries. Nowadays, recycling process has received immense interest in academic fields. Recycling of poly- mers has been paid more and more attention, especially for polypropylene (PP), poly(ethylene terephthalate) (PET), polyethylene (PE), polystyrene (PS), acrylonitrile–butadiene–styrene (ABS), and polycarbona te (PC) due to the environm ental concerns. However ,the recycling process can change mechanical, physical, chemical,and visual properties of plastic.

Nylon (Polyamide-PA) contains the amide repeat linkage in the polymer backbone. PA is a tough, semi-crystal line polymer with alow glass transition and is extensively used in the manufactur e of

ll rights reserved.

automob ile parts and textile fibres due to its high mechanical and impact strength and good processability . There are many types ofPA such as PA6, PA12, and PA66. Among them PA6 has a molecular chain made of a base unit with 6 carbon atoms and it is prepared by ring-openi ng polymerizat ion of e-caprolactam monomer . PA6 is a high-streng th engineering thermoplasti c. In general, glass fibreis added to PA6 for improving the stiffness and strength. Glass fibrealso has low cost, high chemical resistance and excellent insulating propertie s. But, excessive wear on molding dies takes place as the hardness of glass fibre is high. The effects of glass fibre on the mechanical propertie s of virgin PA have been investigated extensively [1–8] and an increase in the mechanical properties with the presence of glass fibre is well known from the literature.

Some studies in the area of recycling of PA are available [9–18].Eriksson et al. [9] reported an experimental and theoretical study of the mechanical performance of the recycled glass-fibre reinforced polyamide 66. Mechanical properties in an accelerated service-related environment of recycled glass-fibre-reinforced polyamide 66 were reported by Eriksson et al. [1] and polymer reinforce d with 30 wt% of short fibres was remolded up to seven times. Eriksson et al. [10] studied the effect of thermal aging onunreinfor ced and glass-reinfo rced recycled polyamide 66. Eriksson et al. [11] also studied the effect of in-plant recycling of glass fibrereinforce d polyamide 66 and injection molded bars were exposed to thermal aging, coolant aging, and creep testing. In another work Eriksson et al. [12] investigated the effects of impurities on

http://dx.doi.org/10.1016/j.matdes.2013.02.027

mailto:[email protected]

http://dx.doi.org/10.1016/j.matdes.2013.02.027

http://www.sciencedirect.com/science/journal/02613069

http://www.elsevier.com/locate/matdes

Fig. 1. Chemical structure of polyamide (nylon 6).

140 E. Kuram et al. / Materials and Design 49 (2013) 139–150

mechanical properties of recycled glass fibre reinforce d polyamide 66. Lozano-González et al. [13] made a study about multiple recycling of nylon-6 by injection molding on its physical–mechanical properties and morphology in order to understand how many times it is possible to recycle the nylon-6 without significant loss of the properties. Scaffaro and La Mantia [14] evaluated the rheo- logical and the mechanical properties of polymer blends of virgin and recycled polyamide 6 by changing the amount of recycled polymer. Maspoch et al. [15] investiga ted the effect of the number of reprocessin g operations (3 times) and of the fraction of recycled material (15%, 30% and 50%) added to the virgin polymer. Thermal,mechanical (tensile, flexural and impact) and rheologic al properties of a product of recycled and filled PA6 were reported [15].The fatigue behaviou r of the reprocessed glass fibre reinforced polyamide 6,6 had been studied by Bernasconi et al. [16]. To our knowledge the biggest number of recycling of PA by injection molding was complete d by Su et al. [17]. In this study, virgin PA6 was repetitively processed until the 16th cycle. They evaluated mechanical and rheologic al properties as a function of number ofrecycling process. After each cycle, they also analyzed the changes in chemical structure, molecular weight, molecular weight distribution, and crystalline behaviour. Goitisolo et al. [18] investigatedthe effect of reprocessin g on the structure and mechanical properties of PA6 based nanocompo sites by means of repeated injection molding cycles.

In plastic injection molding, the process parameters such asmelt temperature , mold temperature, injection pressure, holding pressure, injection speed, holding time, and cooling time should be optimized to produce plastic products with good mechanical properties. One of the optimizati on methods is Taguchi method and this method uses orthogonal array, signal-to-noi se (S/N) ratio and analysis of variance (ANOVA). By using Taguchi’s orthogonal array, time and cost required to carry out the experime nts can bereduced. Then the experimental results are transformed into the S/N ratio in order to measure the quality characteristics deviating from the desired values. A greater S/N ratio shows better quality characterist ics (optimal level of the process paramete rs). ANOVA is conducte d to understand the significant process parameters.Taguchi method has been extensive ly utilized in engineering area.However, few papers investiga te the effect of processing condi- tions on mechanical propertie s of recycled plastics using Taguchi method [19–21]. Mehat and Kamaruddin [19] investigated the flexural testing results of recycled PP by adopting the three levels L9 Taguchi orthogonal arrays. Four processin g parameters namely melt temperature, packing pressure, injection time, and packing time were considered and number of recycling was not taken into consideration. Mehat and Kamaruddin [20] used Moldflow Plastic Insight (MPI) integrated with L18 Taguchi orthogonal array to sim- ulate the significant processing parameters (mold temperat ure,melt temperature, injection time, packing pressure, packing time,and cooling time) affecting the residual stress of the recycled PP.By incorporating the significant parameters (melt temperat ure,packing pressure, injection time, and packing time) obtained from the preliminary simulation, L9 Taguchi orthogon al array was utilized to investigate the flexural modulus and stress at yield ofthe recycled PP. In another study, Mehat and Kamaruddin [21]studied the effects of processing parameters (melt temperature ,packing pressure, injection time, and packing time) on the flexuralproperties of the samples produced from the recycled plastics invarious compositions using L9 orthogonal array.

It is necessary to understand the relationship among the various controllable parameters and to identify the important parameters that influence the mechanical propertie s of recycled polymer.Although there are several studies on recycled PA in the literature,the optimizati on of injection parameters using recycled glass-fibrereinforced PA is not investiga ted. Therefore, we report herein how

the input paramete rs and number of recycling influence the output in recycling of glass-fibre reinforce d nylon 6 (PA6-GF) by using Taguchi experimental design method, which is the novelty of this study. The possible changes in the chemical structure were tested by fourier transform infrared spectroscopy (FT-IR) measureme nt.Different ial scanning calorimetry (DSC) and thermal gravimetric analysis (TGA) were used to analyse the thermal properties ofPA6-GF with reprocessing. The impact fracture surfaces and fibrelength of recycled PA6-GF were investigated by scanning electron microscop y (SEM) and polarized optical microscopy (POM), respectively. The effects of structura l changes on the mechanical properties were determined by tensile and impact tests.

2. Experimen tal details

2.1. Materials

30% Glass-fibre reinforced nylon 6 (Akulon� K224-G6) with adensity of 1350 kg/m 3 used in this study was obtained from com- mercial sources, provided by DSM company. Chemical structure ofnylon 6 is shown in Fig. 1. 30% glass-fibre reinforced polyamide was coded as PA6-GF through this study. Prior to molding, PA6- GF was dried, for at least 3.5 h, at 80 �C to eliminate air bubbles forming during the injection molding process. Test parts were in- jected by a plastic injection machine (YIZUMI-UN90A2) which had a maximum clamping force of 900 kN and an injection pressure of 222 MPa. A four-cavi ty mold (double cavity for tensile specimens and double cavity for impact specimens) was manufactur edin the computer numerica l control (CNC) machine accordin g to ISO 527 [22] for tensile specimens and ISO 180 [23] for impact specimens standards .

PA6-GF material was recycled five times: 0, 1st, 2nd, 3rd, 4th and 5th. Here 0 refers to virgin PA6-GF. A lot of samples in the form of dumbbell shaped (dog bone) were produced in the injection molding machine. From these, 5 samples were used for the tensile tests of virgin PA6-GF. The remaining plastic was shredded and the regrind material was molded again to produce 1st recycle sample.These procedures were repeated to obtain 2nd, 3rd, 4th and 5th recycled samples.

2.2. Taguchi method

Taguchi method uses specially constructed tables known as‘‘orthogo nal array’’ to design experiments . Using of these orthogonal arrays makes the design of experiments very easy and it re- quires a lesser number of experiments . As a result experimental cost, time, and effort reduce. Using Taguchi’s orthogon al array, atotal of 486 (6 � 3 � 3 � 3 � 3 for the full factorial design) sets ofexperime nts are reduced to only 18 sets of experiments , thus this means less experimental cost, time and effort. In the current study,Taguchi’s mixed level parameter design (L18) was used for the experime ntal design. Number of recycling, melt temperature, mold temperat ure, injection pressure, and holding pressure were considered as the control parameters. The control factors and their levels are shown in Table 1. The experimental plan for four injection

Table 1The Taguchi L18 (61 � 34) orthogonal array.

Experiment number Number of recycling Melt temperature Mold temperature Injection pressure Holding pressure

1 1 1 1 1 12 1 2 2 2 23 1 3 3 3 34 2 1 1 2 25 2 2 2 3 36 2 3 3 1 17 3 1 2 1 38 3 2 3 2 19 3 3 1 3 210 4 1 3 3 211 4 2 1 1 312 4 3 2 2 113 5 1 2 3 114 5 2 3 1 215 5 3 1 2 316 6 1 3 2 317 6 2 1 3 118 6 3 2 1 2

Table 2Different contr ol variable s and their levels.

Code Control parameters Unit Level 1 Level 2 Level 3 Level 4 Level 5 Level 6

A Number of recycling – Virgin 1st recycle 2nd recycle 3rd recycle 4th recycle 5th recycle B Melt temperature �C 260 270 280 – – –C Mold temperature �C 70 80 90 – – –D Injection pressure Bar 80 90 100 – – –E Holding pressure Bar 35 40 45 – – –

E. Kuram et al. / Materials and Design 49 (2013) 139–150 141

parameters (melt temperature , mold temperature , injection pressure, and holding pressure) with three levels (34) and one parameter (number of recycling) with six levels (61) (Table 2) were organized by the Taguchi method (L18 orthogonal array). Other injection paramete rs such as injection speed (65%), holding time (2.2 s), and cooling time (25 s), were kept constant during the experiment.

Then the experimental results are transformed into a signal-to- noise (S/N) ratio. In Taguchi method, the term ‘‘signal’’ represents the desirable effect for the output characteristic and the term ‘‘noise’’ stands for the undesirabl e effect for the output characteristic. S/N ratio measures the quality characterist ics deviating from the desired values. There are three S/N ratios; smaller the better ,nominal the best, and larger the better . A high value of S/N impliesthat the signal is much higher than the noise factors. A greater S/N ratio corresponds to better quality characteristics (optimal level of the process paramete rs). Larger the better S/N was used in this study because a higher mechanical propertie s were desirable.Quality characteri stic of the larger the better is calculated in the following equation :

S=N ¼ �10log101n

Xn

i¼1

1y2

i

!" #ð1Þ

yi is the ith measured experimen tal results in a run/row and n ex- plains the number of measureme nts in each test trial/row .

The final step of Taguchi method is to verify the results using confirmation experiments. The confirmation experime nt is conducted with a new set of factors and result is compared with that achieved from the regression model.

In this paper, we also modelled the relation between input parameters of plastic injection molding (number of recycling, melt temperature , mold temperature , injection pressure, and holding pressure) and output parameters using regression analysis. Regres- sion analysis is used to predict unknown dependent variables as a

function of independent variables. The general form of the first-or-der regressio n model is given in the following equation :

y ¼ b0 þXk

i¼1

bixi þ e ð2Þ

where b is a coefficient of each term, k is a number of independen tvariable s and e is an error. The statistical significances of the developed mathem atical models were evaluated by the correlation coef- ficients (R2). High R2 means good relations hip between the proces sparamete rs and the experimenta l results and a higher R2 value is al- ways desirable.

2.3. Measurem ents

The infrared spectra of the samples were taken on a Perkin–El-mer Spectrum 100 spectrophot ometer with an attenuated total reflectance (ATR) objective. The samples were obtained from the injection molded tensile specimens.

The melting behaviour was studied by DSC using a Mettler-To- ledo DSC 822 calorimeter with a constant nitrogen flow rate of10 mL min �1. All the samples (5–7 mg) were first heated from 25to 250 �C with a heating rate of 10 �C min �1 and held for 5 min to erase the thermal history, then cooled to 25 �C at 10 �C min �1,and finally heated to 250 �C at 10 �C min �1. The melting enthalpy (DHm) was determined in the heating scan from the melting peak area. The percentage crystallinity (Xc) of 30% reinforced PA6-GF composites was calculated accordin g to formula in the literature [24] and the melting enthalpy value of 100% crystalline PA6,190.6 J/g, was taken from [18]. TGA was carried out on a Mettler- Toledo TGA/SDTA 851 thermogravim etric analyzer with a heating rate of 20 �C min �1 from room temperature to 700 �C under nitrogen ambient.

The impact fracture surfaces of PA6-GF were directly observed with SEM (Philips XL30 SFEG). The lengths of the glass fibres in the injection molded specimens were analyzed by using Leitz Wetzler

Fig. 2. Specimen dimensions for tensile strength (ISO 527).

Fig. 3. Specimen dimensions for izod impact test (ISO 180).

Fig. 4. FTIR spectra of the reprocessed test specimens.


Orthoplan-Pol POM. The photographs of each sample were taken manually using a digital camera.

Experiment number 1 (virgin), 2 (virgin), 3 (virgin), 5 (1st recycle), 8 (2nd recycle), 11 (3rd recycle), 14 (4th recycle) and 17(5th recycle) were chosen for FTIR, DSC, TGA, SEM, and POM from orthogonal array in order to reduce labour and were named as T1,T2, T3, T5, T8, T11, T14 and T17, respectivel y in the figures of FTIR,DSC, TGA, SEM and POM.

For the tensile tests, PA6-GF was molded into the shape of dog bone specimens using the injection molding machine. The specimens and dimensions for the tensile and izod impact test are shown in Figs. 2 and 3. The standard dimension of tensile specimens was 175 mm length, 20 mm width and 4 mm thickness asshown in Fig. 2 [22]. The standard dimension of impact specimens in accordance with ISO 180 [23] was 80 mm � 10 mm � 4 mm asshown in Fig. 3. Impact specimens were molded at these dimensions, then notch was produced at impact test samples. The tensile strength and impact tests were determined by using INSTRON 5569 and INSTRON Ceast 9050, respectively . Tensile test was carried out according to ISO 527 standard. A dog bone specimen,which was clamped vertically in the grips of the testing machine,was held constant at one end and was pulled at a constant rate of elongation at the other end. The tensile test speeds were 50 mm/min (ISO 527). 5.5 J of the impact hammer was employed for the impact test measure ments. All mechanical tests were repeated five times and the mean values were presented.

Fig. 5. DSC curves for the reprocessed test specimens (second heating and cooling).

3. Results and discussions

3.1. Physical and chemical structures

FTIR was used to identify whether chemical structural changes occurred or functional groups formed else than present in virgin nylon 6 during the recycling processes. Fig. 4 shows the FTIR spectra of the chosen test specimens. As can be seen from the spectra,the characteristic bands of nylon 6 [17,18] appeared for the samples without any significant change, indicating that the chemical nature of nylon 6 in PA6-GF remained unchanged after five reprocessing cycles.

The melting and crystallization behaviou r of the recycled samples were examine d by DSC. Fig. 5 shows the DSC curves of the samples in the second heating and the cooling run. These curves clearly indicate the presence of characteri stic melting, which isthe typical of nylon 6 polymer. From Fig. 5, it was concluded that the endotherm that peaked at about 219 �C for all samples was

due to the crystalline melting [1]. No decreasing trend of melting point (Tm) was observed with increasing recycling number, but the highest Tm values were obtained for virgin samples (T1, T2,and T3) as shown in Table 3. It was also seen in this table that the percentage crystallinity did not significantly change with

Table 3Melting enthalpy (DHm), melting temperature (Tm) and percentage crystallinity (Xc)results.

Material DHm (J/g) (2nd heating) Tm (�C) Xc (%)

PAGF-T1 46.08 219.90 16.92 PAGF-T2 42.68 219.13 15.67 PAGF-T3 41.67 219.13 15.30 PAGF-T5 37.30 218.57 13.70 PAGF-T8 39.11 218.98 14.36 PAGF-T11 42.15 219.02 15.48 PAGF-T14 49.56 218.96 18.20 PAGF-T17 42.46 218.38 15.59

Fig. 6. TGA curves of the test specimens.


increasing recycling number. The reason of no significant change for Tm can be attributed to the fact that the reprocessing does not modify the main characteri stics of the crystalline phase [18].However, in some studies an incremen t in the Tm and crystallinity was observed when PA6 was reprocessed [17].

TGA is a useful analysis method for not only quantitative determination of the degradation behaviour but also the composition ofthe glass fibre and the matrix in a composite material [25]. In TGA analysis, the thermal stability of samples was obtained in terms ofthe weight loss (%) as a function of temperature (Fig. 6). The firstweight loss was observed between 50 and 150 �C due to the loss of water (moisture) in the composites. From TGA curves in Fig. 6,it was concluded that the main degradation started at about 350 �C for all samples and major degradation occurred in the range of �350–480 �C. The degradation delay of the matrix with the glass fibre content was well known. The residue at 700 �C was due to the amount of glass fibre content in PA6-GF that was left behind after the burning of the matrix (30% glass-fibre reinforced nylon 6 was used in this study, see the Materials part). It was also found that the curves of all recycled samples were not significantly different to those of virgin samples.

SEM was used to examine the impact fracture surfaces of the reprocessed PA6-GF materials. As it is known the interface between reinforcing fibre and matrix resin is the key for the thermo- plastic composites. Fig. 7 shows the SEM images of impact fracture surfaces of PA6-GF materials. It could be clearly seen that the fracture mostly occurred in the matrix phase (nylon 6) rather than inthe interface between the glass fibre and PA6. This result implies the good adhesion between the glass fibre and nylon 6.

POM was employed to examine the change in the lengths ofglass-fibre after each recycling process. First, the small pieces obtained from the injection molded specimens of virgin and

recycled materials (the chosen ones) were dissolved in concen- trated nitric acid. Then, a few drops of each mixture were taken onto a glass slide and put on a hot plate to get rid of the acid. Later,the fibres on each glass microscope slide were observed and measured after 16� magnification by using Leitz Wetzler Orthoplan- Pol Polarized Optical Microscopy. The photograp hs were taken manually using a digital camera. As it appears clearly in Fig. 8,the length of the fibres was shortened after injection molding processes as the glass fibres were exposed to breaking in the injection molding machine.

3.2. Mechanica l properties

In tensile tests, the specimen is pulled at constant rate. The specimen elongate s during the test and the increment resistance of the specimen is measured by a load cell. This load value (force)and also maximum force are recorded during the test. The elongation of the specimen is lasted until a rupture and force value atbreak is also saved. The tensile strength is calculated according to the Eq. (3). Then the stress–strain curve is plotted. Typical stress–strain curves of plastics shows the information about mechanical properties such as elastic modulus, yield strength, tensile strength, and % elongation. Sample stress–strain curve for PA6- GF is depicted in Fig. 9. From this curve it was concluded that virgin PA6-GF broke after the peak of maximum stress. Maximum stress values were obtained with virgin samples. 4th (T14) and 5th (T17)recycled samples had minimum stress and strain values. Maximum strain values occurred for the samples of 1st (T5) and 3rd (T11)recycled material. It was concluded that 1st and 3rd recycled samples were more ductile than other samples.

Tensile strength ¼ Maximum load Cross section area

ð3Þ

3.2.1. S/N analysis S/N ratio measure s the quality characterist ics deviating from

the desired values and a high value of S/N means that the signal is much higher than the noise factors. A greater S/N ratio shows the optimal level of the process parameters. There are three S/N ra-tios; smaller the better , nominal the best and larger the better . In this study, the mechanical properties were analysed by larger the better S/N quality characteristic due to higher mechanical properties were desirable . The experimental results and the S/N ratio values calculated by taking Eq. (1) into considerati on are listed in Table 4.

The level of a parameter with the highest S/N ratio is the optimal level. All the optimal processing parameters (the highest S/Nratios) were highlight ed in circles in Figs. 10–13. Hence the optimal process parameter setting for the tensile strength was A1B3C3D3E 3 (Fig. 10). Thus, the best combination values for max- imizing tensile strength were virgin PA6-GF, melt temperature of280 �C, mold temperature of 90 �C, injection pressure of 100 bar and holding pressure of 45 bar. The tensile strength of PA6-GF increased with increasing all parameters (melt temperature, mold temperat ure, injection pressure and holding pressure) used in the current study.

The analysis of the results showed that the optimum combination of process paramete rs for high yield strength was A1B1C3D3E 1, namely, virgin PA6-GF, melt temperat ure of 260 �C,mold temperature of 90 �C, injection pressure of 100 bar and holding pressure of 35 bar (Fig. 11). The yield strength of PA6-GF decreased with increments of melt temperature and holding pressure.

The optimum combinati on of process parameters for high impact energy was A1B1C2D1E 3, namely, virgin PA6-GF, melt temperature of 260 �C, mold temperature of 80 �C, injection pressure of 80 bar and holding pressure of 45 bar (Fig. 12). The impact energy of PA6-GF decreased with incremen ts of melt temperat ure

Fig. 7. SEM images of impact fracture surfaces of test specimens (a) 1000 �; (b) 250 �.

Fig. 8. Optical microscope images of the glass fibres extracted from the test specimens.


and injection pressure and increased with an incremen t of holding pressure. The increment of the mold temperature initially increased the impact energy; however , as this parameter was further increased, impact energy eventual ly decreased.

A1B1C2D1E 2 was the optimum combination of process parameters for high impact strength (Fig. 13). For impact strength, the optimum parameters were as follows: virgin PA6-GF, melt temperature of 260 �C, mold temperature of 80 �C, injection pressure of

80 bar and holding pressure of 40 bar. The impact strength ofPA6-GF decreased with increasing melt temperature. The increments of the mold temperature and holding pressure initially increased the impact strength; however , as these both parameters were further increased, impact strength eventually decreased.The decrement of the injection pressure initially decreased the impact strength; however , as injection pressure was further increased, impact strength eventually increased.

Fig. 9. Sample stress–strain curve for PA6-GF.


From the main effects plot it could be concluded that the number of recycling was more significant paramete r because the slope gradient was very big and this result was similar to those obtained with variance analysis, which was explained in Section 3.2.2.

Better mechanical properties of virgin PA than recycled PA were reported in the literature. Licea-Claver íe et al. [2] used nylon 6,6 reinforced with mixed glass and carbon fibres. The tensile strength,young modulus and impact energy of the virgin nylon were found to be higher than that of the recycled nylon [2]. Lozano-Gonzálezet al. [13] recycled the nylon-6 in order to know physical–mechan-ical properties. The nylon-6 was recycled 10 times and it was found that the properties of samples did not suffer any change until the 8th recycling cycle. Changes of 10–15% in the properties between virgin nylon and nylon with 10 times recycling materials were observed, except for percentage of elongation which decrease d 70%until the 10th recycling cycle [13].

Even though the results in this study showed that the mechanical properties of virgin PA6-GF were better compared to recycled PA6-GF, the results of mechanical properties were still acceptab leuntil 3rd recycling. Especially the mechanical properties of the recycled PA6-GF obtained from 1st recycling were not largely different from those of virgin PA6-GF and thus it could be used

Table 4Experimental and S/N results.

Experiment number

Tensile strength 2% Yield strength

Experimental result (MPa)

S/N(dB)

Experimental result (MPa)

1 150.83 43.57 99.73 2 153.28 43.71 100.16 3 167.66 44.49 97.64 4 143.91 43.16 96.91 5 149.23 43.48 93.50 6 142.41 43.07 94.33 7 134.95 42.60 89.06 8 128.09 42.15 90.23 9 136.15 42.68 89.25 10 122.21 41.74 87.31 11 124.73 41.92 87.82 12 121.46 41.69 83.69 13 108.12 40.68 75.37 14 111.80 40.97 74.61 15 110.54 40.87 67.77 16 115.27 41.23 70.51 17 103.39 40.29 71.38 18 109.72 40.81 66.72

for different applications in industry. Increasing the number ofrecycling up to four to five times resulted in a faster deterioration rate in mechanical propertie s of recycled samples. In the current study molding process repeated five times but further molding operation s (up to seven times) investiga ted for glass-fibre-rein-forced polyamide 66 by Eriksson et al. [1] and they reported that further remolding (after 5 times) had less effect on the deterioration rate of mechanical property (elongation at break) [1]. The length distribut ion of the glass fibres strongly influences the mechanical propertie s of samples molded from glass fibre reinforced thermo- plastics [9]. To this end, Eriksson et al. [1] measured fibre lengths of samples and found that this value was 0.341 mm for virgin polymer and decrease d to 0.150 mm after seven times recycling. They also found that the tensile strength and modulus of samples decreased with the increasing number of recycling process due tothe lower degree of reinforcement imposed by shorter fibre lengths [1], in agreement with the results presented in [16]. Tensile strength was about 145 MPa for virgin PA, about 105 MPa after fivetimes recycling and this value decreased to about 95 MPa after seven times recycling. In our paper, the effect of injection parameters was small as compared to number of recycling thus it is possible toconsider these parameters constant . Tensile strengths of virgin sample, 1st, 2nd, 3rd, 4th and 5th recycled samples computed the average of experime nts 1 to 3, 4 to 6, 7 to 9, 10 to 12, 13 to 15and 16 to 18, respectively . Similar approach explained above for tensile strength was also used for yield strength, impact energy and impact strength analysis. From this analysis it was concluded that tensile strength was about 157 MPa for virgin PA6-GF, about 109 MPa after five times recycling and these values were close tostudy reported by Eriksson et al. [1]. Tensile strengths of 1st, 2nd,3rd and 4th recycled samples computed in a similar manner and were found as 145, 133, 122 and 110 MPa, respectively . For tensile strength, 30.6% decrement from virgin to the 5th cycle was observed. For tensile strength Lozano-González et al. [13] reportedcontrary result to our study and the other studies in the literature.They observed that the tensile strength of nylon-6 increased when the number of recycling processes increased and 11% increment from virgin to the 9th cycle was reported. The increase of tensile strength during reprocessing was due to the increase of molecular weight [13].

Different results for yield strength were observed in the literature. Yield strength did not significantly change upon reprocessin g

Impact energy Impact strength

S/N(dB)

Experimental result (J)

S/N(dB)

Experimental result (kJ/m2)

S/N(dB)

39.98 2.97 9.46 67.83 36.63 40.01 2.96 9.43 66.78 36.49 39.79 2.74 8.76 61.83 35.82 39.73 2.82 9.01 63.66 36.08 39.42 2.61 8.33 58.64 35.36 39.49 2.47 7.85 56.01 34.97 38.99 2.69 8.60 60.60 35.65 39.11 2.03 6.15 46.10 33.27 39.01 2.06 6.28 46.10 33.27 38.82 2.17 6.73 48.44 33.70 38.87 1.97 5.89 43.56 32.78 38.45 1.93 5.71 43.07 32.68 37.54 1.74 4.81 38.85 31.79 37.46 1.63 4.24 36.00 31.13 36.62 1.52 3.64 33.22 30.43 36.97 1.58 3.97 34.59 30.78 37.07 1.51 3.58 33.39 30.47 36.49 1.48 3.41 32.53 30.25

Fig. 10. Main effects plot of S/N ratios for tensile strength.

Fig. 11. Main effects plot of S/N ratios for yield strength.


[18] and an increment in the yield strength after each reprocessin goperation was reported [17]. Literature results were contrary toour present study. Su et al. found the overall increase in the yield strength from the 1st to the 16th cycle as approximat ely 25%[17]. In our study yield strength decreased from 99 MPa for virgin PA6-GF to 69.5 MPa for the 5th cycle, which meant 29.8%reduction.

Impact energies of virgin sample, 1st, 2nd, 3rd, 4th and 5th recycled samples were found as 2.89, 2.63, 2.26, 2.02, 1.63 and

1.52 J, respectively . In our study impact energy decreased 47.4%until the 5th reprocessing cycle.

Impact strengths of virgin sample, 1st, 2nd, 3rd, 4th and 5th recycled samples were found as 65.5, 59.4, 51.0, 45.0, 36.0 and 33.5 kJ/m 2, respectively, which meant 48.8% decreme nt. Eriksson et al. [9] found the impact strength 60.6 kJ/m 2 for virgin PA, about 41.9 kJ/m 2 after five times recycling and this value decreased to36.9 kJ/m 2 after seven times recycling. The reasons of decrease inimpact strength with increasing number of recycling can be ex-

Fig. 12. Main effects plot of S/N ratios for impact energy.

Fig. 13. Main effects plot of S/N ratios for impact strength.


plained by the increase in the number of fibre ends as a result ofprocess induced fibre breakage, at the fibre ends an applied load leads to stress concentratio n, which may initiate matrix crack for- mation as the first stage of failure before catastrophic crack propagation occurs and by the lower energy absorption during crack propagation for shorter fibres [9]. Lozano-González et al. stated in their study that the impact properties of nylon-6 had no significant change till the 7th recycling cycle; after this cycle, a decrement in the impact properties was observed [13]. In another study, the impact strength of PA6 decreased from 327 J/m for 1st

processed to 128 J/m for 16th processed [17]. The reduction of impact strength with recycling was attributed to the behaviou r ofmolecula r chain scission. Shorter molecular chain and broader chain length distribution result in poor chain entangle ments, thus a reduction in toughnes s for the multiple processed PA6 [17].

3.2.2. ANOVA Analysis of variance (ANOVA) was used to determine the statis-

tically significant paramete rs influencing the mechanical properties of PA6-GF and to determine the percentage contribution of


each control factor during injection molding. ANOVA results are shown in Tables 5–8. The analyses were carried out for the level of confidence 95% (the level significance is 5%). The number ofrecycling, melt temperature, mold temperature, injection pressure and holding pressure influenced the tensile strength values by95.09%, 0.39%, 0.44%, 0.10% and 3.20%, respectively (Table 5). From the analysis of this table, since P value was lower than 0.05 one can observe that the number of recycling and holding pressure have statistical and physical significance on the tensile strength, especially the number of recycling factor.

The most influential factor for yield strength was number ofrecycling with a percentage contribution of 96.31% and followed by melt temperature (Table 6). Other factors, mold temperature ,injection pressure and holding pressure contributed to yield strength 0.18%, 0.20% and 0.39%, respectivel y, which can be ne- glected. Only the number of recycling had statistical and physical significance on the yield strength.

The number of recycling, melt temperature , mold temperature ,injection pressure and holding pressure affected the impact energy

Table 5ANOVA for tensile strengt h.

Factors Degree of freedom, DF Sum of squares, SS

Number of recycling 5 24.5034 Melt temperature 2 0.0995 Mold temperature 2 0.1141 Injection pressure 2 0.0267 Holding pressure 2 0.8257 Error 4 0.2000 Total 17 25.7695

Table 6ANOVA for yield strength.



Table 7ANOVA for impact energy.



Table 8ANOVA for impact strength.



values by 92.22%, 5.31%, 0.88%, 0.25% and 0.35%, respectively (Ta-ble 7). The number of recycling and melt temperature had statistical and physical significance on the impact energy, especially the number of recycling factor.

For the impact strength, the most significant factor was the number of recycling with a percentage contribution of 92.60%.The second most influential factor was melt temperat ure (5.24%contributi on). The mold temperature , injection pressure and holding pressure factors gave 0.83%, 0.27% and 0.14% contribution,respectivel y (Table 8). The number of recycling and melt temperature had statistical and physical significance on the impact strength, especiall y the number of recycling factor.

It was concluded from the ANOVA tables that the number ofrecycling was the most significant parameter influencing the mechanical properties investigated in this study.

3.2.3. Regression analysis Regressio n analysis is used to determine the relationship be-

tween independen t variables and dependent variables, and to pre-

Mean squares, MS F ratio, a = 5% P Contribution %

4.9007 98.02 0.000 95.09 0.0498 1.00 0.446 0.39 0.0571 1.14 0.405 0.44 0.0134 0.27 0.778 0.10 0.4128 8.26 0.038 3.20 0.0500 0.78

100.00


4.7285 87.90 0.000 96.31 0.2510 4.67 0.090 2.04 0.0222 0.41 0.687 0.18 0.0247 0.46 0.661 0.20 0.0478 0.89 0.479 0.39 0.0538 0.88

100.00


14.7580 73.92 0.000 92.22 2.1226 10.63 0.025 5.31 0.3500 1.75 0.284 0.88 0.1008 0.51 0.637 0.25 0.1380 0.69 0.552 0.35 0.1996 0.99

100.00


16.1873 80.49 0.000 92.60 2.2905 11.39 0.022 5.24 0.3630 1.80 0.276 0.83 0.1162 0.58 0.602 0.27 0.0630 0.31 0.748 0.14 0.2011 0.92

100.00

Fig. 14. Comparison of measured–predicted values for (a) tensile results and (b)impact results.


dict depende nt variables as a function of independent variables. Inmultiple regression analysis, R2 (R-squared) is value of the correlation coefficient and it is used to justify the validity of developed regression model. R2 can be defined as:

R2 ¼ Sum squared residual ðSSRÞSum squared total ðSSTÞ ð4Þ

Table 9Results of confirmation experiments.

Code Output responses Exper

A1 B3 C3 D3 E3 Tensile strength (MPa) 143.1Yield strength (MPa) 97.29Impact energy (J) 2.68 Impact strength (kJ/m2) 59.22




R2 ¼ 1� Sum squared error ðSSEÞSum squared total ðSSTÞ ð5Þ

Tensile strength, yield strength, impact energy and impact strength were dependent variables and number of recycling, melt temperat ure, mold temperat ure, injection pressure and holding pressure were independen t variables. In this study, multiple linear regressio n models were developed for prediction of output responses. The model equations were as follows:

Tensile strength ¼ 150� 10:1 � Aþ 1:05 � Bþ 1:49 � C

þ 1:03 � Dþ 4:01 � E ð6Þ

R2 ¼ 96:0% R2ðadjÞ ¼ 94:3%

Yield strength ¼ 111� 6:24 � A� 1:62 � Bþ 0:148 � C

þ 0:181 � D� 0:703 � E ð7Þ

R2 ¼ 94:3% R2ðadjÞ ¼ 91:9%

Impact energy ¼ 3:49� 0:288 � A� 0:148 � B� 0:0192 � C

� 0:0317 � Dþ 0:0383 � E ð8Þ

R2 ¼ 95:8% R2ðadjÞ ¼ 94:1%

Impact strength ¼ 80� 6:74 � A� 3:43 � B� 0:399 � C

� 0:773 � Dþ 0:599 � E ð9Þ

R2 ¼ 96:1% R2ðadjÞ ¼ 94:5%

where A, B, C, D and E were coded values pertaining to the number of recycling, melt temperat ure, mold temperat ure, injection pressure and holding pressure, respective ly.

In multiple regressio n analysis, R2 should be between 0.8 and 1.In this study, regression models were consistent with the experimental data (R2 > 90%). Thus, these mathematical equations could be utilized to predict mechanical propertie s of PA6-GF in injection molding.

To measure the validity of the developed regressio n models, the predicted values obtained from regression models were compared with the experimental results. The relationships between the experime ntal values and predicted values were plotted and were illustrate d in Fig. 14. In general, the plots for experimental and predicted values were identical and the variations between experimental and predicted values were minimal. The average absolute

imental values Predicted values Error (%)

1 162.64 13.65 7 98.778 1.52

2.72 1.49 0 61.251 3.43

6 152.52 7.51 1 103.424 9.01

2.94 12.64 0 66.913 14.40

6 156.99 1.49 51 101.508 0.75

3.10 16.98 70.06 17.85

1 152.98 3.37 8 102.211 2.96

3.06 18.60 69.46 19.29


errors for tensile strength, yield strength, impact energy and impact strength were found to be 2.41%, 2.37%, 4.70% and 4.77%,respectively .

3.2.4. Confirmation experiments The aim of confirmation experiment is to validate the conclu-

sions drawn during the analysis. In order to verify the develope dmodels, confirmation experiments should be conducted. In this study, the confirmation experiments were carried out by using the optimum levels for tensile strength, yield strength, impact energy and impact strength. The results of confirmation experiments in Table 9 showed the comparis on of the foreseen Eqs. (6)–(9) and experimental values. To measure the accuracy of predictio n models for confirmation experiments, error for models is estimated asfollows and given in Table 9:

Error ¼ Measured Values � Predicted Values Measured Values

�� 100 ð10Þ

As error values must be smaller than 20% for reliable statistical analyses, error values below 20% was accepted in the literature [26–28]. The predicted results had very close values with the experimental results, thus mathematical equations developed inthis study could be used to predict mechanical properties of PA6- GF in injection molding.

4. Conclusions

Current study focused on Taguchi experimental design for investigatin g the effect of the number of recycling and process parameters on the chemical or physical structures and mechanical properties of glass-fibre reinforce d nylon 6 during the injection molding. In the experiments , different recycling samples, mold temperature , melt temperature , injection pressure and holding pressure were utilized as processing parameters.

FTIR results indicated that chemical nature of PA6-GF remained unchanged as a consequence of multiple reprocessing. From DSC analysis it was found that Tm and the percentage crystallinity did not significantly change with increasing number of recycling. All TGA curves of the recycled and virgin samples were close to each other. The residual weight of all samples at 700 �C was in good agreement with glass fibre content in PA6-GF.

From the experimental results it was concluded that the virgin PA6-GF gave better mechanical properties than recycled PA6-GF.Even though the results showed that the mechanical properties of virgin PA6-GF were better compare d to recycled PA6-GF, the results of mechanical properties were still acceptab le until 3rd recycling. Especially, the mechanical properties of the recycled PA6-GF obtained from 1st recycling were not largely different from those of virgin PA6-GF and thus could be used for different applicati ons in industry.

ANOVA was used to determine effects of the control paramete rson tensile strength, yield strength, impact energy, and impact strength. Among the paramete rs investigated in this study, number of recycling had the greatest effect on mechanical properties.

Regression analyses were applied to predict tensile strength,yield strength, impact energy and impact strength. The comparison of measured–predicted results proved that predicted values for each response were close to experimentally measured values.The average absolute errors for tensile strength, yield strength, impact energy, and impact strength were found to be 2.41%, 2.37%,4.70% and 4.77%, respectively. Hence mathematical equations developed in this study could be utilized to predict mechanical properties of PA6-GF in injection molding.

Acknowled gements

The authors thank to TUBITAK for supporting of this study (Project no. 110M245).

References

[1] Eriksson P-A, Boydell P, Månson J-AE, Albertsson A-C. Durability study ofrecycled glass-fibre-reinforced polyamide 66 in a service-related environment.J Appl Polym Sci 1997;65:1631–41.

[2] Licea-Claveríe A, Carrillo FJU, Alvarez-Castillo A, Castano VM. Characterization of mixed fibre nylon composites incorporating composite scrap. Polym Compos 1998;20(2):314–20.

[3] Thomason JL. The influence of fibre properties of the performance of glass- fibre-reinforced polyamide 6,6. Compos Sci Technol 1999;59:2315–28.

[4] Laura DM, Keskkula H, Barlow JW, Paul DR. Effect of glass fibre and maleated ethylene–propylene rubber content on tensile and impact properties of nylon 6. Polymer 2000;41:7165–74.

[5] Thomason JL. Micromechanical parameters from macromechanical measurements on glass reinforced polyamide 6,6. Compos Sci Technol 2001;61:2007–16.

[6] Laura DM, Keskkula H, Barlow JW, Paul DR. Effect of glass fibre surface chemistry on the mechanical properties of glass fibre reinforced, rubber- toughened nylon 6. Polymer 2002;43:4673–87.

[7] García M, González N, Eguiazábal JI, Nazábal J. Structure and mechanical properties of new hybrid composites based on polyamide 6,6 reinforced with both glass fibres and a semiaromatic liquid crystalline polyester. Polym Compos 2004;25(6):601–8.

[8] Thomason JL. The influence of fibre length, diameter and concentration on the impact performance of long glass-fibre reinforced polyamide 6,6. Compos Part A – Appl Sci 2009;40:114–24.

[9] Eriksson P-A, Albertsson A-C, Boydell P, Prautzsch G, Månson J-AE. Prediction of mechanical properties of recycled fibreglass reinforced polyamide 66. Polym Compos 1996;17(6):830–9.

[10] Eriksson P-A, Boydell P, Eriksson K, Månson J-AE, Albertsson A-C. Effect ofthermal-oxidative aging on mechanical, chemical, and thermal properties ofrecycled polyamide 66. J Appl Polym Sci 1997;65:1619–30.

[11] Eriksson P-A, Albertsson A-C, Boydell P, Månson J-AE. Durability of in-plant recycled glass fibre reinforced polyamide 66. Polym Eng Sci 1998;38(2):348–56.

[12] Eriksson P-A, Albertsson A-C, Boydell P, Månson J-AE. Influence of impurities on mechanical properties of recycled glass fibre reinforced polyamide 66.Polym Eng Sci 1998;38(5):749–56.

[13] Lozano-González MAJ, Rodriguez-Hernandez MAT, Gonzalez-De Los Santos EA,Villalpando-Olmos J. Physical–mechanical properties and morphological study on nylon-6 recycling by injection molding. J Appl Polym Sci 2000;76:851–8.

[14] Scaffaro R, La Mantia FP. Characterization of monopolymer blend of virgin and recycled polyamide 6. Polym Eng Sci 2002;42(12):2412–7.

[15] Maspoch ML, Ferrando HE, Velasco JI. Characterisation of filled and recycled PA6. Macromol Symp 2003;194:295–303.

[16] Bernasconi A, Davoli P, Rossin D, Armanni C. Effect of reprocessing on the fatigue strength of a fibre glass reinforced polyamide. Compos Part A – Appl Sci 2007;38:710–8.

[17] Su K-H, Lin J-H, Lin C-C. Influence of reprocessing on the mechanical properties and structure of polyamide 6. J Mater Process Tech 2007;192–193:532–8.

[18] Goitisolo I, Eguiazábal JI, Nazábal J. Effects of reprocessing on the structure and properties of polyamide 6 nanocomposites. Polym Degrad Stab 2008;93:1747–52.

[19] Mehat NM, Kamaruddin S. Optimization of mechanical properties of recycled polypropylene products using Taguchi method. Malaysia Polymer International Conference (MPIC) 2009:125–35.

[20] Mehat NM, Kamaruddin S. Investigating the effects of injection molding parameters on the mechanical properties of recycled plastic parts using the Taguchi method. Mater Manuf Process 2011;26(2):202–9.

[21] Mehat NM, Kamaruddin S. Optimization of mechanical properties of recycled plastic products via optimal processing parameters using the Taguchi method.J Mater Process Tech 2011;211:1989–94.

[22] ISO 527, Plastics – determination of tensile properties, International Organization for Standardization, Switzerland, 1997.

[23] ISO 180:2000(E), Plastics – determination of izod impact strength,International Organization for Standardization, Switzerland, 2000.

[24] Feng W, Ait-Kadi A, Brisson J, Riedl B. Polymerization compounding composites of nylon-6,6/short glass fibre. Polym Compos 2003;24:512–24.

[25] Lee SM, Cho D, Park WH, Lee SG, Han SO, Drzal LT. Novel silk/poly(butylenesuccinate) biocomposites: the effect of short fibre content on their mechanical and thermal properties. Compos Sci Technol 2005;65:647–57.

[26] Taguchi G. Introduction to quality engineering. New York: McGraw-Hill; 1990.[27] Ozcelik B, Sonat I. Warpage and structural analysis of thin shell plastic in the

plastic injection molding. Mater Des 2009;30:367–75.[28] Ozcelik B. Optimization of injection parameters for mechanical properties of

specimens with weld line of polypropylene using Taguchi method. Int Commun Heat Mass Transfer 2011;38:1067–72.

Documents

Kuram2013 Reciclaje Pa 6