Real-time wide-angle X-ray diffraction during polyethylene blown film extrusion.
Films produced by blown film extrusion are widely used in packaging applications (1). The effect of blown film process parameters on the microstructure and the mechanical properties, viz. tensile and tear strength, have been discussed in literature (2-4). Over the years, X-ray diffraction has been a powerful and reliable technique to obtain information about the microstructure, such as crystallinity, crystalline orientation, and crystalline interplanar spacing, in the films (5-7). Most of microstructural measurements on films have been carried out offline on the processed blown films as a postmortem analysis. How- ever, online microstructural measurements offer the advantage of monitoring the quality of a film in real-time and adjusting the process parameters to obtain desired properties.
Real-time measurements of microstructure during film-blowing have gained significant attention from both modeling (8-10) and experimental (11-13) perspectives. Bull-winkel et al. (11) used simultaneous online small angle light scattering (SALS) and infrared temperature measurements to study the microstructure evolution during linear low-density polyethylene (LLDPE) film extrusion. They related the change in average scattered intensity of the light to the crystallization process. Nagasawa et al. (12) reported the first online measurements of orientation development during blown film extrusion by using birefringence. Additional studies were conducted by Ghaneh-Fard et al. (13) during the film-blowing of a LLDPE.
Recently, we have successfully used real-time Raman spectroscopy to estimate crystallinity development during the blown film extrusion of polyolefins (14), (15). Our study demonstrated Raman spectroscopy as a powerful nondestructive technique to monitor microstructural evolution in industrial settings, as it offers remote sampling capabilities when coupled via fiber optics. However, like any vibrational spectroscopic techniques, it is not a primary measurement technique. The integral intensities from Raman spectra are generally calibrated using primary measurement techniques such as differential scanning calorimetry (DSC), density measurements, or X-ray diffraction (16).
Although X-ray diffraction measurements during fiber-spinning of LLDPE, isotactic polypropylene (i-PP), Nylon 66, and polyethylene terephthalate (PET) have been reported in literature studies (17-19), real-time X-ray diffraction measurements during the film-blowing process have not been reported in the literature. Unlike fiber-spinning, where the wide-angle X-ray diffraction (WAXD) is performed on a single or bundle of filaments (20), bubble geometry (hollow cylinder) adds to the complexity of real-time WAXD measurements during film-blowing. The use of high intensity X-ray sources, such as the synchrotron radiation, can reduce data collection time (21), (22) during continuous processing, but the equipment is typically not easily accessible. The purpose of the present study was to explore the use of WAXD technique using a conventional laboratory X-ray source for measurement of crystallinity in real-time during blown film extrusion of a low-density polyethylene (LDPE).
Materials and Processing
Low-density polyethylene (LDPE, Dow 6401, 0.92 g/cc) blown films with a melt flow index (MFI) of 2 g/10 min were formed using a lab-scale extruder (19 mm, 24:1 L/D, Alex James and Associates, Greenville, SC) equipped with a die of 25.4 mm diameter and 0.25 mm gap (Haake, AZ). A die temperature of 210[degrees]C and a throughput of 20 g/min were maintained during the experiment. The films were fabricated using a single-lip air ring adjusted to supply air at a velocity of 12 m/s, as measured with an air velocity transducer (TSI model 8455, St. Paul, MN). Of the various combinations of processing parameters, a blow-up ratio (BUR) of 0.6 and a take-up ratio (TUR) of 5.5 were chosen for all real-time measurements in this study. In prior studies, we have investigated a number of other combinations of BUR and TUR (14). However, in this study, a low BUR was chosen not only to keep the bubble very stable, but more importantly to generate uniaxial extensional flow field to keep the X-ray analysis tractable. The stretch-rates were measured before and after each WAXD and Raman spectroscopy measurements. In addition, control blown film samples were also formed, one with a diameter of 2.2 cm (small) and another with a diameter of 8 cm (large), for offline X-ray diffraction experiments.
A custom-built X-ray diffraction system (Rigaku/MSC), shown in Fig. 1, was used throughout this study. It consisted of an X-ray generator, a beam-collimator, and an image-plate detector stationed on a motor-driven Z-platform. The platform can be moved axially from a distance of 1 m up to a distance of 5 m above the ground with a resolution of 1 mm.
[FIGURE 1 OMITTED]
WAXD patterns were obtained from a Rigaku 2-D diffractometer (Rigaku/MSC) using Cu-K[alpha] radiation with conditions of 45 kV and 0.67 mA. The incident X-ray beam was collimated to a beam size of 0.5 mm diameter and was focused on the bubble surface using a videocamera arrangement. All the WAXD data were obtained in transmission mode with an exposure time of 30 min and with two repetitions. During the experiment, the lateral movement of the bubble was constrained so that it was stable, and the X-ray beam was focused on the longitudinal axis of the bubble. The patterns collected at different axial distances during the film-blowing were corrected for air-scattering background using POLAR software (STAR, SUNY, NY) and analyzed using GRAMS/32 software (Galactic, Salem, NH).
Further, to compare the real-time X-ray diffraction results with those obtained by a different technique, simultaneous real-time Raman spectroscopic measurements were conducted following a protocol that was established in earlier studies (14), (15). Raman spectroscopy is a rapid measurement technique that can identify variations in the film line over a time period of 1 or 2 min. The data-collection time used during the experiment was 2 min. The online Raman spectroscopy system consisted of Raman probe connected to a charged-coupled device (CCD) de tector (Renishaw Raman system 100, Glocestershire, UK). Figure 1 displays the overall location of the Raman probe along with the X-ray source and image-plate.
RESULTS AND DISCUSSION
Offline X-Ray Diffraction Measurements
Prior to online X-ray diffraction measurements, offline control studies were conducted on the processed blown films to explore the nature of the X-ray diffraction pattern for different geometries of the film. Figure 2 shows a schematic of three different methods for obtaining an X-ray diffraction pattern on a blown film: (a) Transmittance of X-ray beam through a lay-flat blown film, (b) Grazing incidence of X-ray beam on a hollow cylindrical blown film (simulating the film bubble), and [c] Transmittance of X-ray beam through the hollow cylindrical blown film, where the beam is transmitted through the front-face (the face immediately encountered by the X-ray beam) and the back-face (the face encountered by the X-ray beam before reaching the image-plate detector) with air between them. The normalized 20 (theta)-intensity spectra obtained from the diffraction pattern for a grazing-incidence on a hollow cylindrical blown film and a lay-flat film (flattened after grazing-incidence experiment) are presented in Fig. 3. The prominent crystallographic planes of interest in polyethylene are the (110) and (200) peaks at diffraction angles of 21.6[degrees] and 23.8[degrees], respectively, and are used to calculate the orthorhombic crystalline content (6). The WAXD intensity spectrum obtained from the grazing incidence of an X-ray beam on the hollow cylindrical geometry is not different from the spectrum obtained for the lay-flat film. However, the use of grazing incidence X-ray diffraction during the real-time measurements is impractical because of the lateral movement of the bubble during continuous processing.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
Figure 4 presents the X-ray diffraction pattern along with the 2[theta] spectrum generated by the transmittance of the X-ray beam through the cylindrical blown film. A composite diffraction pattern is formed on the image-plate detector that consists of the response from the front and back faces, at a high and low angle, respectively, caused by the differences in distance between a given face and the image-plate detector. The spectrum obtained using the lay-flat film is overlaid with the spectrum from the cylindrical blown film for comparison. As expected, the intensity of diffracted X-ray beam was slightly lower for the front-face compared to that of the back-face. This is attributed to the scattering of diffracted X-ray beam by the back-face in its path from the front-face to the image-plate detector. However, the secondary scattering of the scattered X-ray from the front-face was not detected, because it is expected at an angle that is beyond the range of the image-plate detector.
[FIGURE 4 OMITTED]
With the appropriate choice of the face-to-detector distance, the spectrum from a given face of the cylindrical blown film can be matched with its spectrum from the lay-flat geometry. For the X-ray system used in this study, the sample-to-detector distance was 11.3 cm. The hollow cylindrical blown film, 2.2 cm diameter, was mounted such that the front-face was at a sample-to-detector distance of 11.3 cm. Figure 5 presents a comparison of the X-ray spectrum obtained on the lay-flat film to the spectrum obtained on the hollow cylindrical film. The known sample-to-detector distance of 11.3 cm leads to a match of the diffraction peaks from the front-face (Fig. 5b) to the diffraction peaks from the lay-flat film (Fig. 5a). On the other hand, a sample-to-detector distance of 9.1 cm, calculated from the diameter (2.2 cm) of the cylindrical blown film, leads to a match between the diffraction peaks from the back-face (Fig. 5c) and the peaks from the lay-flat film.
[FIGURE 5 OMITTED]
Real-Time X-Ray Diffraction Measurements
Two-dimensional WAXD images for LDPE bubble obtained at various axial distances along the line are presented in Fig. 6. The pattern formed in the image-plate is a composite pattern due to scattering from front- and back-faces of the bubble. Since the bubble diameter during the experiment was smaller compared to the controlled samples for the processing condition studied, the (200) peak of the back-face of the bubble overlaps the (110) peak of the front-face of the bubble.
[FIGURE 6 OMITTED]
Below the frost-line height (FLH) (i.e., axial distance < 50 cm), the WAXD pattern exhibits an isotropic amorphous halo as expected for the melt state of the polymer. At a distance of 51 cm, close to the FLH, the crystalline (110) and (200) reflections start to appear in the pattern, indicating the onset of crystallization. At higher axial distances ( > 76 cm), the crystalline reflections become stronger, indicating an increase in the crystal population. The patterns also show the existence of preferred orientation of the crystalline structure in the bubble. The (110) reflections intensifies in the equatorial direction, and the (200) reflections appear on the meridian. Simultaneously, the amorphous halo decreases.
The 2[theta]-intensity spectra from the WAXD patterns obtained at various axial locations in the blown film line are presented in Fig. 7. As expected, the WAXD spectrum for the melt exhibits a diffuse amorphous halo and lacks any defined peaks. As the ploymer travels past the freeze line, the (110) and (200) crystalline peaks increase in intensity at the expense of the amorphous halo due to crystallization.
[FIGURE 7 OMITTED]
From the composite WAXD spectra, the peak position, peak height, peak width, and integrated intensity for each diffraction are and amorphous background can be extracted. The deconvolution of the peaks was carried out using a mixed Gaussian-Lorentzian (G-L) function. The center, width of the peak, and the percentage Lorentzian function in the mixed G-L fit that were estimated on the amorphous spectrum (from the melt) were later used to calculate the amorphous content of the composite intensity spectrum obtained at other locations in the film line.
The mixed G-L fit performed on the composite X-ray diffraction spectra of amorphous (molten) and semicrystalline states of the LDPE bubble is presented in Fig. 8a and b, respectively. The crystalline content was determined by fitting all of the crystalline peaks obtained for the front- and back-faces of the bubble. The crystalline fraction in the bubble was estimated using [X.sub.c] = [I.sub.c]/([I.sub.c] + [I.sub.a]), where [I.sub.c] is the area under crystalline peaks and [I.sub.a] is the area under the amorphous halo (6). The crystallinity value of 39 [+ or -] 2 wt% measured for the processed blown film using cylindrical geometry of the bubble was consistent with a value of 40 [+ or -] 2 wt% measured for the same sample with a lay-flat geometry. These results indicate that a suitable deconvolution procedure may be used to analyze the composite spectra.
[FIGURE 8 OMITTED]
The development of crystallinity as a function of axial distance is presented in Fig. 9. First, the crystallinity profile from online WAXD measurements is presented in Fig. 9a, which displays a sigmoidal shape. The crystalline growth starts near the FLH, steeply increases at lower axial distances, and then plateaus at higher axial distancesnear the nip-rolls. Crystallinity content as low as 2 wt% was successfully detected by real-time WAXD technique. Real-time temperature measurements were also performed using an IRCON Infrared pyrometer (Modline 340, Niles, IL), and the profile is shown on the secondary axis in Fig. 9a. The temperature profile shows a plateau at [approximately equal to]98[degrees]C as a consequence of the exothermic heat of crystallization.
[FIGURE 9 OMITTED]
Next, in Fig. 9b, a comparison of real-time crystallinity measurements from WAXD and Raman spectroscopy is presented. The crystalline content from the Raman spectrum was calculated by calibrating the ratio of integral intensities from 1418 and 1300 c[m.sup.-1] bands with the crystallinity obtained from DSC (16). The crystallinity profiles from the two different techniques are consistent (within [+ or -]3 wt%). The crystallinity curves from the present study also show good agreement with trends reported in previous experimental studies (11), (14), (23).
Finally, Fig. 10 presents a comparison of an offline spectrum obtained from a large-diameter cylindrical film to that obtained from a small-diameter blown film shown earlier (see Fig. 5). It is evident that the diffraction peaks generated by the two faces of the large cylindrical blown film do not overlap. Therefore, for a large-diameter cylindrical bubble, typically obtained from a pilot-scale blown film extrusion, the response from one of the faces of the bubble can be used to analyze for crystallinity without the need to analyze a composite spectrum. While we see no fundamental limitations of the technique, a few practical limitations will need to be addressed. For instance, the X-ray based system will need to be installed in a secluded section of the production floor and then a more powerful source can be used to reduce data-collection time. Thus, in combination with vibrational techniques such as realtime Raman spectroscopy, X-ray diffraction offers a robust solution for real-time microstructural measurements during film-blowing.
[FIGURE 10 OMITTED]
From the above discussion, it is evident that the X-ray diffraction technique discussed in this study offers a novel route for real-time process monitoring in terms of the microstructure (rather than process variables such as T and [delta]P).
The feasibility of using WAXD for real-time crystallinity measurements during the blown film extrusion of LDPE was established. From the evolution of (110) and (200) peaks, it was evident that the crystallization process starts near the FLH, shows a steep growth immediately past the FLH, and then plateaus at higher axial distances near the nip-rolls. The real-time crystallinity profiles obtained from WAXD were consistent with those measured using real-time Raman spectroscopy.
Any opinions, findings, conclusions, or recommendations expressed in this material are those of the authors and do not necessarily reflect those of the National Science Foundation, which supported this work.
(1.) T. Kanai and G.A. Campbell, Film Processing, Hanser/ Gardner Publications, Clincinnati, Ohio (1999).
(2.) R.M. Patel. T.I. Butler, K.L. Walton, and G.W. Knight, Polym. Eng. Sci., 34, 1506 (1994).
(3.) A. Gupta, D.M. Simpson, and I.R. Harrison, J. Appl. Polym. Sci., 50, 2085 (1993).
(4.) X.M. Zhang, S. Elkoun, A. Ajji, and M.A. Huneault, Polymer, 45, 217 (2004).
(5.) M.C. Branciforti, L.M. Guerrini, R. Machado, and R.E.S. Bretas, J. Appl. Polym. Sci., 102, 2760 (2006).
(6.) G.P. Tiley and S.L. Aggarwal, J. Polym. Sci., 18, 17 (1955).
(7.) P. Maiti, P.H. Nam, M. Okamoto, T. Kotaka, N. Hasegawa, and A. Usuki, Polym. Eng. Sci., 42, 1864 (2002).
(8.) S. Muke, H. Connell, I. Sbarski, and S.N. Bhattacharya, J. Non-Newt. Fluid Mech., 116, 113 (2003).
(9.) L.K. Henrichsen, A.J. McHugh, S.S. Cherukupalli, and A.A. Ogale, Plast. Rubber Compos., 33, 383 (2004).
(10.) C. Bangshu and G. Campbell, AIChE J., 36, 420 (1990).
(11.) M.D. Bullwinkel, G.A. Campbell, D.H. Rasmussen, J. Krexa, and C.J. Brancewitz, Int. Polym. Process., 16, 39 (2001).
(12.) T. Nagasawa, T. Matsumura, and S. Hoshino, Appl. Polym. Symp., 20, 275 (1973).
(13.) A. Ghaneh-Fard, P.J. Carreau, and P.G. Lafleur, Int. Polym. Process., 12, 136 (1997).
(14.) S.S. Cherukupalli and A.A. Ogale, Polym. Eng. Sci., 44, 1484 (2004).
(15.) S.S. Cherukupalli and A.A. Ogale, Plast. Rubber Compos., 33, 367 (2004).
(16.) G.R. Strobl and W. Hagedorn, J. Polym. Sci. Polym. Phys. Ed., 16, 1181 (1978).
(17.) K. Katayama, T. Amano, and K. Nakamura, Kolloid Z. Z. Polym., 226, 125 (1968).
(18.) J. Dees and J.E. Spruiell, J. Appl. Polym. Sci., 18, 1053 (1974).
(19.) B.S. Hsiao, R. Barton, and J. Quintana, J. Appl. Polym. Sci., 62, 2061 (1996).
(20.) P.E. Lopes, M.S. Ellison, and W.T. Pennington, Plast. Rubber Compos., 35, 294 (2006).
(21.) M. Cakmak, A. Teitge, H.G. Zachmann, and J.L. White, J. Polym. Sci. Part B: Polym. Phys., 31, 371 (1993).
(22.) R. Kolb, S. Seifert, N. Stribeck, and H.G. Zachmann, Polymer, 41, 1497 (2000).
(23.) T. Kanai and J.L. White, Polym. Eng. Sci., 24, 1185 (1984).
Correspondence to: A.A. Ogale; e-mail: email@example.com
Contract grant sponsor: National Science Foundation (Engineering Research Centers Program); contract grant number: EEC 9731680.
Published online in Wiley InterScience (www.interscience.wiley.com). [C] 2008 Society of Plastics Engineers
Giri Gururajan, (1), (2) H. Shan, (2) G. Lickfield, (2), (3) A.A. Ogale (1), (2)
(1) Department of Chemical and Biomolecular Engineering, Clemson University, Clemson, South Carolina 29634
(2) Center for Advanced Engineering Fibers and Films, Clemson University, Clemson, South Carolina 29634
(3) Department of Materials Science and Engineering, Clemson University, Clemson, South Carolina 29634
|Printer friendly Cite/link Email Feedback|
|Author:||Gururajan, Giri; Shan, H.; Lickfield, G.; Ogale, A.A.|
|Publication:||Polymer Engineering and Science|
|Article Type:||Technical report|
|Date:||Aug 1, 2008|
|Previous Article:||Tribological behaviors of aligned carbon nanotube/fullerene-epoxy nanocomposites.|
|Next Article:||Interfacial modifications in PS/PMMA and PVC/EVA blends by E-beam and microwave irradiation: a free volume study.|