Monitoring of Rheological Indicators of LDPE Per-Åke Clevenhag and Claes Oveby Tetra Pak Carton Ambient AB ABSTRACT LDPE,s from high-pressure autoclave reactors for extrusion coating with Melt Flow Rates (MFR,s) from 6 to 9g/10 min., and with densities of 917-921 kg/m 3, have been considered uniform commodities for a long time. However, as coating lines became faster, differences in processing performance among LDPE,s have been observed even though they met MFR and density specifications. These problems included web break and edge instabilities causing PE missing. Various LDPE,s behaved differently in the coating process due to differing rheological properties. This paper presents results of our investigation comparing several LDPE,s. Results showed significant differences in their rheological properties. These differences significantly affected performance of the LDPE,s in the extrusion coating process. A further finding was significant variation in rheological properties and resulting performance within the same grade from the same supplier. INTRODUCTION In extrusion coating a thin molten polymer film is coated on some kind of substrate. At high extrusion coating speed, even a minor disturbance on the melt web causes major quality problems, which very rapidly lead to large quantities of waste. Therefore, higher coating speeds require polymers with high and even quality in order to avoid waste due to polymer edge instability and web break. In optimising the extrusion coating process it is of utmost importance to balance the rheology of the polymer. This paper describes how oscillatory measurements can be used to characterize LDPE. The G and η 0 have been found to be useful parameters to predict the extrusion coating performance and to monitor the product consistency of LDPE. MATERIALS AND METHODS Rheological Measurements The rheological measurements were performed on a StressTech controlled stress melt rheometer. Polymer pellets were melted, compressed and stamped into discs with a diameter of 25 mm and a thickness of 1 mm. Parallel plates were used and measurements were done under nitrogen atmosphere. Oscillatory measurements were performed at 170 C in the linear viscoelastic region with a frequency sweep between 20 and 0.01 Hz. Extrusion Coating Trials Thirteen grades of autoclave LDPE from different suppliers were tested in two pilot extrusion coating lines at three different occasions. All grades had MFR between 7 and 9 g/10 min. and densities of 918-920 kg/m 3. The melt temperature was monitored to 295-305 C, by means of an infrared camera and air gap was between 180-300 mm. The methodology was in principle the following. Extruder screw rpm was set to give 10 g/m 2 at 150 m/min. in order to obtain stable processing conditions. The line speed was then increased in steps of 50 m/min. until the web broke. The line speed at which the web broke (draw-down speed) and the neck-in were reported. This was made in duplicate for each grade of LDPE. G and η 0 Monitoring of LDPE Grades In order to determine the product consistency, in terms of G and η 0, LDPE suppliers were selected to send samples from at least 50 lots to Tetra Pak during a period of 4 to 8 months for G and η 0 measurement.
RESULTS AND DISCUSSION Rheological indicators When a material undergoes oscillatory stress with frequency, the response can be expressed in terms of a storage modulus, G a loss modulus, G and a complex viscosity, η*. From the frequency sweep two rheological indicators are found, G at G = 500 Pa and the zero shear viscosity. G is determined in the following way. The loss modulus is plotted versus the storage modulus (log-log plot) in the range of 200-700 Pa of the loss modulus 1. A linear relation is obtained and the storage modulus can be determined at a loss modulus of 500 Pa (log 500 = 2.7). An example is shown in Figure 1 where a G of 107.5 is found. 3.1 2.9 y = 0.64x + 1.40 R 2 = 1.00 LogG'' (Pa) 2.7 2.5 Fig. 1. The log loss modulus (G ) versus log storage modulus (G ) for the determination of the G at G equal to 500 Pa. Modulus measured at 170 C in a frequency sweep between 20 and 0.01 Hz. log 500 l G = 10 m (1) where 2.3 1.5 1.7 1.9 2.1 2.3 2.5 LogG' (Pa) l = intercept of the linear regression line m = slope of the linear regression Zero shear viscosity is determined by extrapolation by the use of the 3 parameters Cross equation: η0 η* = 1+ n τω (2) where η* = complex viscosity [Pas] η 0 = zero shear viscosity [Pas] τ = characteristic relaxation time [s] ω = frequency [rad/s]
n = power law index [-] The curve fit calculation is preferably done with the Microsoft Excel Solver 2 but any other software giving comparable data could be used. Prediction model The results from the extrusion coating trials were analysed by the use of a multivariate data analysis software 3. G and η 0 were set as variables and observed draw-down as the response (17 observations). The model diagnostics evaluation gave the following results: R 2 = 0.76 (estimates goodness of fit) Q 2 = 0.72 (estimates goodness of prediction) From the multivariate data analysis the following relationship was found: DD = 1311 7.55*G 0.022*η 0 (m/min.) (3) A relationship between draw-down and neck-in, with an R 2 of 0.93, was also found: NI = 0.13*DD + 44 (mm) (4) MFR showed no correlation to observed draw-down. An R 2 of 0.37 was achieved when observed draw-down was plotted versus measured MFR. The results from the oscillatory measurements, the pilot extrusion coating trials and the results from the predictions are given in Table 1. Table 1. Summary of the data measured and generated at the pilot extrusion coating trials. LDPE MFR G η 0 Predicted DD Observed DD and NI - g/10 min. Pa Pas m/min. m/min. mm 1 7.1 105.1 4220 414 400 99 2 7.0 132.7 4640 194 200 71 3 8.0 108.2 3580 405 365 91 4 6.8 123.4 5000 257 300 87 5 7.2 120.9 4280 292 300 84 6 7.1 117.2 4420 317 325 85 7 7.5 118.5 4280 311 290 80 8 7.2 108.5 4230 388 350 88 9 8.9 105.2 3220 435 550-10 7.0 132.7 4640 194 170-11 8.0 108.2 3580 405 430-12 6.8 116.3 4900 314 390-13 7.2 120.9 4280 292 322-14 7.1 117.2 4420 317 352-15 7.0 106.5 4575 396 400-16 7.5 118.3 5565 284 290-17 8.3 105.4 3635 425 375 -
Consistency of LDPE Process capability is defined as the 6σ interval that defines a statistically controlled process. An index of capability is C p where: C p USL LSL 6σ = (5) where USL and LSL are the upper and lower specification limits, respectively. This is an indicator of what the process could do if properly centered. A similar index, C pk, indicates the capability of the process as it is currently centered. C pk is the smallest of: C pu USL X 3σ = (6) or C pl X LSL 3σ = (7) where X is the average value of the population. The process capability, P pk, is defined in the same way as C pk but takes also the process variation into consideration. P pk is calculated using the overall variation. Both the between-subgroup and within-subgroup contribute to the overall variation. C pk is calculated using the within-subgroup variation, but not the shift and drift between subgroups. In order to control the repeatability of the test method and the equipment a reference sample of LDPE was measured every 10 th measurement during the whole monitoring period. The results are found in Table 2. Table 2. Repeatability according to TAPPI method TM1200 of test method and equipment during the monitoring. Mean Standard deviation Repeatability Repeatability ratio (%) G (Pa) 112 0.8 2.2 2.0 η 0 (Pas) 3670 47 130 3.5 The mean value and consistency of G and η 0 for the different LDPE,s are shown in Table 3. The tolerance (USL LSL) for G is set to 10 Pa and for η 0 to 800 Pas.
Table 3. Results from monitoring of LDPE. LDPE G (Pa) P pk (G ) η 0 (Pas) P pk (η 0 ) A 116 1.11 4300 0.86 B 108 0.73 4350 0.69 C 122 0.60 4650 0.80 D 107 0.81 3500 1.02 E 106 0.81 3500 0.77 F 120 0.66 4350 0.72 G 113 0.61 3650 1.21 H 109 0.61 3800 0.77 I 115 0.74 4450 0.67 J 131 0.90 4300 1.06 K 113 0.62 4850 0.49 L 113 1.11 4650 0.90 M 120 0.65 5350 0.40 N 113 1.42 4200 0.77 O 112 0.72 4850 0.29 There are considerably differences in both level of G and η 0 and also consistency among the different LDPE,s. These differences will have high impact on the extrusion coating performance of the different LDPE grades. CONCLUSIONS LDPE,s from high-pressure autoclave reactors for extrusion coating with Melt Flow Rates (MFR,s) from 6 to 9g/10 min., and with densities of 917-921 kg/m 3, have been considered uniform commodities for a long time. However, as coating lines became faster, differences in processing performance among LDPE,s have been observed even though they met MFR and density specifications. The parameters MFR and density gives very limited information on the processing performance of the product. Very often a resource demanding test run of the LDPE material in a pilot or production coating line is necessary in order to determine the processability. Extrusion coating performance can be conveniently, and considerably less resource demanding, determined and predicted by the rheological method described in this paper. The advantages with this method are as follow: Very good correlation with processing behaviour. Reduced costs in the evaluation process. Improved quality of the LDPE, which reduces the LDPE edge trim and risk for manufacturing defects. REFERENCES 1. Shroff, R and Mavridis, H., New Measurs of Polydispersity from Rheological Data on Polymer Melts, Journal of Applied Polymer Science, vol. 57, 1605-1626 (1995). 2. Roberts, G.P., Barnes, H.A., Mackie, C., "Using the Microsoft Excel solver tool to perform non-linear curve fitting, using a range of non-newtonian flow curves as examples", Applied Rheology 11, 2, 271-276 (2001). 3. SIMCA-P10, Umetrics AB, Umeå, Sweden.
2005 PLACE Conference September 27-29 Las Vegas, Nevada Monitoring of Rheological Indicators of LDPE Presented by: Per-Åke Clevenhag Claes Oveby Tetra Pak Carton Ambient AB Contents Introduction of process and material Rheological method Pilot extrusion coating trials to find correlations Results from rheological monitoring of LDPE A case study Summary The process: Extrusion coating Neck-in Draw-down 1
The process: Extrusion coating 2
LDPE- a uniform commodity product? A new extrusion coating LDPE grade has an unknown processing behaviour. LDPE is specified by density and melt flow rate. Pilot extrusion coating trials are necessary to judge the performance. A processing trial in pilot plant costs USD 15 000. Question Isn t there a simple low cost test method to predict the processing behaviour of LDPE? Methods Melt Flow Rate Rheology Temperature: 190 C Measuring time: 6+3 min. Poor correlation with processability Temperature: 170 C Measuring time: 4+6 min. Correlates well with processability 3
Melt Flow Rate at 190 C ASTM D1238 Gives g/10 minutes for a given temperature and load. A high MFR corresponds to a low viscosity. A one point measurement. 1.E+04 Pas 1.E+03 1.E+02 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 rad/s Rheology at 170 C 1.E+05 1.E+04 Pas, Pa 1.E+03 1.E+02 Viscosity* η* G' G'' 1.E+01 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 Hz G (storage modulus) A Cole Cole plot 3.1 2.9 y = 0.64x + 1.40 R 2 = 1.00 G at G = 500 Pa (log500 = 2.7) LogG'' (Pa) 2.7 2.5 2.3 1.5 1.7 1.9 2.1 2.3 2.5 LogG' (Pa) 4
η 0 (zero shear viscosity) A Cross model: ηo η* = 1+ ( τω ) n Pas 1.E+04 1.E+03 η o 1.E+02 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 rad/s Rheological indicators G = f(mwd, LCB) e.g.: G' M M z w a Amount M w MWD M z η 0 = f(m w, LCB) a η 0 = KM w LCB Chain length (K and a depend on LCB) Pilot extrusion coating trials LDPE: High pressure autoclave reactor MFR*: 6.8-8.9 g/10 min. Density: 918-920 kg/m 3 Film temperature: 295-305 C *(190 ; 2.16 kg): 5
Multivariate Data Analysis X variables: G and η 0 G (Pa) 105 133 X η 0 (Pas) 4220 4640 Y DD (m/min.) 400 200 Response (=Y): Draw-down (DD) 108 123 105 121 3580 5000 3220 4280 365 300 550 300 Quality factors obtained: R2 = 0.76 Q2 = 0.72 117 119 109 133 108 4420 4280 4230 4640 3580 325 290 350 170 430 116 4900 390 121 4280 322 117 4420 352 107 4575 400 118 5565 290 105 3635 375 Observed versus predicted DD (from G and η 0 at 170 C) 600 500 R 2 = 0.76 Observed draw-down (m/min.) 400 300 200 100 G η 0 0 0 100 200 300 400 500 600 Predicted draw-down (m/min.) Draw-down versus MFR at 190 C 600 500 R 2 = 0.37 400 m/min. 300 200 100 0 6 6.5 7 7.5 8 8.5 9 9.5 g/10 min. 6
Monitoring of LDPE LDPE suppliers were selected to send samples from produced lots during a period of 4-8 months for G and η 0 measurement. 15 grades. > 50 samples of each grade. 128 124 LDPE1_G' Monitoring LDPE Example Individual Value 120 116 112 UCL=116.15 _ X=112.79 LCL=109.43 108 1 6 11 16 21 26 31 36 Observation 41 46 51 LDPE2_G' LDPE1: MFR = 7.5 g/10 min. Density = 918 kg/m 3 LDPE2: MFR = 7.1 g/10 min. Density = 917 kg/m 3 Individual Value 128 124 120 116 112 108 1 6 11 16 21 26 31 Observation 36 41 46 UCL=127.32 _ X=119.93 LCL=112.54 6500 6000 LDPE1_eta0 Monitoring LDPE Example Individual Value 5500 5000 4500 UCL=4974 _ X=4633 LCL=4291 4000 1 6 11 16 21 26 31 36 Observation 41 46 51 LDPE2_eta0 LDPE1: MFR = 7.5 g/10 min. Density = 918 kg/m 3 LDPE2: MFR = 7.1 g/10 min. Density = 917 kg/m 3 Individual Value 6500 6000 5500 5000 4500 4000 1 6 11 16 21 26 31 Observation 36 41 46 UCL=6255 _ X=5361 LCL=4467 7
Capability Capability index = what we want/what we get P p USL LSL = 6σ (P pk takes accuracy to a target into consideration) Process Data LSL 108 Target 113 LDPE1_G' LSL Target USL O v erall C apability Pp 1.16 PPL 1.11 Monitoring LDPE Example USL 118 PPU 1.21 Sample Mean 112.789 Ppk 1.11 Sample N 55 Cpm 1.16 StDev (O v erall) 1.43265 100 104 108 112 116 120 124 128 132 136 140 O bserved Performance Exp. Overall Performance % < LSL 0.00 % > USL 0.00 % Total 0.00 % < LSL 0.04 % > USL 0.01 % Total 0.06 LDPE2_G' LDPE1: MFR = 7.5 g/10 min. Density = 918 kg/m 3 Process Data LSL 115 Target 120 USL 125 Sample Mean 119.932 Sample N 50 StDev (O v erall) 2.54661 LSL Target USL O v erall C apability Pp 0.65 PPL 0.65 PPU 0.66 Ppk 0.65 Cpm 0.66 LDPE2: MFR = 7.1 g/10 min. Density = 917 kg/m 3 O bserved Performance 100 104 108 112 116 120 124 128 132 136 140 Exp. Overall Performance % < LSL 6.00 % < LSL 2.64 % > USL 2.00 % > USL 2.33 % Total 8.00 % Total 4.97 Process Data LSL 4250 Target 4650 LDPE1_eta0 LSL Target USL O v erall C apability Pp 0.95 PPL 0.91 Monitoring LDPE Example USL 5050 PPU 0.99 Sample Mean 4632.6 Ppk 0.91 Sample N 55 Cpm 0.94 StDev (O v erall) 140.689 3000 3500 4000 4500 5000 5500 6000 O bserved Performance % < LSL 0.00 % > USL 0.00 % Total 0.00 Exp. Overall Performance % < LSL 0.33 % > USL 0.15 % Total 0.48 LDPE2_eta0 LDPE1: MFR = 7.5 g/10 min. Density = 918 kg/m 3 Process Data LSL 4950 Target 5350 USL 5750 Sample Mean 5360.88 Sample N 50 StDev (O v erall) 323.511 LSL Target USL O v erall C apability Pp 0.41 PPL 0.42 PPU 0.40 Ppk 0.40 Cpm 0.41 LDPE2: MFR = 7.1 g/10 min. Density = 917 kg/m 3 O bserved Performance 3000 3500 4000 Exp. Overall Performance 4500 5000 5500 6000 % < LSL 8.00 % < LSL 10.20 % > USL 12.00 % > USL 11.45 % Total 20.00 % Total 21.66 8
Summary of monitoring 6000 600 m/min. 500 m/min. 400 m/min. 5500 LDPE2 5000 0 (Pas) 4500 LDPE1 4000 3500 Ref. 600 m/min. 500 m/min. 400 m/min. 3000 100 105 110 115 120 125 130 135 G' (Pa) 1.6 Capability for LDPE 1.4 1.2 Ppk for eta0 1.0 0.8 0.6 LDPE1 0.4 LDPE2 0.2 0.0 0.0 0.2 0.4 0.6 0.8 Ppk for G' 1.0 1.2 1.4 1.6 Case study: missing PE 6000 600 m/min. 500 m/min. 400 m/min. η 0 (Pas) 5500 5000 4500 538 m/min 423 m/min 431 m/min 339 m/min 4000 3500 604 m/min 519 m/min 600 m/min. 500 m/min. 400 m/min. 3000 100 105 110 115 120 125 130 135 G' (Pa) 9
Why a Rheology Method? Very good correlation with processing behaviour Saves money in the LDPE evaluation process Improves the quality of the LDPE, which reduces the PE edge trim and the risk for missing PE Differences in both level of G G and η 0 and also consistency among different LDPE s has been monitored. We have submitted this rheology method for a TAPPI method standard Thank You PRESENTED BY Per-Åke Clevenhag Claes Oveby Tetra Pak Carton Ambient AB Please remember to turn in your evaluation sheet... 10