ON-LINE X-RAY DIFFRACTION INSTRUMENTATION
Summary
The principal goal of the Center for Advanced Engineering Fibers and Films (CAEFF) is to develop computer-based process/product models for fiber and film industries. The Center researchers from eight different academic departments at Clemson University, together with researchers at the Massachusetts Institute of Technology and the University of Illinois, and sixteen partner companies are working together to incorporate the necessary molecular information for these models to accurately predict final fiber and film properties. A team of faculty and students from Clemson, the University of Illinois, and DuPont has been charged with providing the accurate real-time data on the development of structure during fiber and film processing needed to verify these models. Raman spectroscopy instruments are already in place for use in real-time measurement of structure.
X-ray measurements have been employed in the past to monitor crystalline growth as well as crystalline and amorphous orientation. However, long exposure times were needed because of sample vibration and size as well as x-ray intensity and detector sensitivity. The x-ray instrumentation described in this document will be designed to overcome these problems in the measurement of crystalline growth as well as crystalline and amorphous orientation during both fiber and film formation. The x-ray instrument will give Center faculty and students the unique capability of measuring structural changes in polymeric materials during processing by both wide-angle and small-angle x-ray scattering.
The equipment will establish the critical links between our present and projected online equipment, our existing off-line analysis capabilities, and the data requirements of the research program of the CAEFF. The two components of the CAEFF research program, empirical and computational, which span all of the research thrusts, will benefit from the availability of this equipment. While the principal utilization of the equipment described in this document will be in Thrust 1.2 of the Center for Advanced Engineering Fibers and Films, it will also find application in our education mission and in other research programs on the campus. The system will provide data on the morphological structure of polymeric materials which develops during actual processing, on-line and in real time.
Project Description
Introduction
The essential elements of polymer structure are conventionally modeled as idealized crystalline and amorphous phases. Polymer crystalline material, characterized by long-range order, is amenable to measurement by wide-angle x-ray scattering (WAXS). The d-spacing, crystal lattice and size distribution can be determined from these data (1). X-ray data analysis techniques have been developed which can elucidate some aspects of the amorphous domains (20). Small-angle x-ray scattering (SAXS) data can illuminate the larger scale relative organization of these elements (texture), and can detect microscopic void spaces. Incorporated into both of these elements is the net orientation of the polymer chains, within both the crystalline and the amorphous regions. The degree of this orientation is also amenable to measurement by x-ray diffraction (XRD), although there are several optical techniques, including birefringence and optical dichroism, which are well suited for measuring molecular orientation.
In fiber and film production, draw-induced orientation is the dominant factor that influences the properties of the final end-use article. The ultimate structure of the fiber or film is a consequence of the processing conditions; in particular, the rate of temperature change, which affects crystallinity, and the stress applied during the drawing process which effects orientation and may also facilitate crystallization. In order truly to understand the nature of the effects of these, and other, process variables on the structure, it is necessary to. One method for in-situ determination of structure development during processing is on-line, real time XRD during the fiber and/or film formation process.
Off-Line XRD
The published literature that addresses the development of structure during production of fibers and films is dominated by discussion of off-line measurements of structure and properties, accompanied by inferential statements of the effects of processing parameters in structure development. Real-time fiber and film structure development data is most often inferred from measurements of draw ratios, throughput rates and processing temperatures and pressures, and occasionally filament tension. These data are then related to off-line measurements of crystallinity and orientation. There are several reports of on-line determination of birefringence and filament temperature, most of which bear testament to the limitations of the extant technology for obtaining these data. It has been amply demonstrated that XRD on static samples is unsurpassed in the amount and quality of structural data it affords, and so motivates us to devise methods for its application on-line.
The literature covering XRD of stationary (off-line) samples is more prevalent than that for XRD of moving (on-line) samples (1). The data from the off-line studies have been rich in their exposition of film and fiber structural forms. In conjunction with associated physical property testing, significant advances have been made in determining the relationships between structure and resultant properties. These advances have covered the conventional polymers, as well as new, high-performance materials.
Evidence for spatial correlation of the polymer chains in the amorphous phase of crystallizable polymers was reported using the X-ray diffraction patterns of poly(ethylene terephthalate) resins, films, and fibers (20). Their data suggests that there may be two average interchain distances in the amorphous phase. The two distances might arise from the rotational correlation of the aromatic groups on adjacent chains.
SAXS and WAXS were used by Murthy to enable quantitative prediction of the properties in fibers of semicrystalline polymers with data from drawn and annealed nylon 6 fibers. The length, diameter, and orientation of the fibrils and the lamellae, and the spacing between the fibrils and the lamellae, were determined using SAXS. Changes in the amorphous orientation were studied by analytically separating the amorphous scattering in WAXS patterns into isotropic and anisotropic components (21). Their data suggest that glass transition behavior is due to the superposition of relaxations in the non-oriented and the oriented amorphous components.
Another use of XRD is in the study of blend morphology, an application central to the research conducted in Thrust 3 (below). Characterization of the morphology of blends of poly(butylene terephthalate) and a thermotropic liquid crystalline polymer (TLCP), the co-polyester HX-8000 series from DuPont, by WAXS found that the blends show stable fiber formation at shear rates dependent on the TLCP content (28). Those authors found that the highly fibrillated TLCP phase is coupled with an increase in the TLCP molecular orientation. A recent investigation of blend spinning was also carried out in our laboratories (23). Models were developed, and confirmed using some on-line data, that successfully described the morphology of the fibrillar structure formed when spinning blends of PP and PS.
Miyata, et al., conducted melt spinning of poly(ethylene 2,6-naphthalene dicarboxylate) (PEN) up to a take-up velocity of 9 km/min. (17). From their investigations, spinning of PEN was divided into three regions in terms of the mechanism of fiber structure formation. At a take-up velocity of up to 2.5 km/min, the fibers registered a birefringence of less than, 0.08 implying only a slight increase in molecular orientation. At a take-up velocity of 2.5-4.5 km/min, the birefringence was 0.08-0.25, and, although some experimental evidence indicated that orientation-induced crystallization did not occur, there was an increase in the fiber density that suggested the formation of an ordered structure. At a take-up velocity greater than 4.5 km/min, the birefringence exceeded 0.25, and they found evidence of crystallization. Presence of a neck-like deformation in the spinning line was also observed. The solidification temperature of the spinning line inferred from the diameter profile suggested that crystallization occurred at relatively low temperatures as compared to that under quiescent conditions. The authors therefore concluded that the presence of elongational stress in the spinning line promoted crystallization.
On-Line XRD
Chappel reported the first on-line measurements of structure by XRD in 1964 (4). In a work published by Katayama in 1967 (11), WAXS and SAXS, and several other parameters including birefringence, temperature and diameter, were measured on a running, large diameter (ca. 200 micron) monofilament. Those authors used a stationary x-ray source and a moveable spinneret, as have several authors since that time. Their data supported the generally held and reasonable belief that crystallization of the polymer occurs in the threadline, although additional crystallization occurs as the material ages.
There then followed a hiatus until Joe Spruiell published his landmark research series. His prolific works addressed the on-line structure development of polyolefins, and the polyamides. Jim White collaborated with Spruiell, and then published a review with Cakmak in 1986 (30). In 1993, Cakmak, White and others reported work on poly(vinylidene fluoride) (3). The most recent work comes from the group of Jerold Schultz, some of which is not yet published (26).
The scattering cross-section for organic polymers is relatively small, on the order of 100A (16). Thus, in order to obtain intensity of the diffracted beam sufficient for assigning reciprocal lattice points to crystal planes, and to determine the other parameters of interest, most researchers have used relatively long exposure times and high intensity incident beams. Using conventional sources, with accelerating voltages of 45kV and beam currents of 40mA, exposure times have ranged up to two hours (8). The high intensity x-ray source of choice for many has been synchrotron radiation (3, 26). These sources are, of course, associated with rather large particle accelerators, such as the Advanced Photon Source (APS) at Argonne National Laboratory, which limits their utility. A rotating anode can be used to obtain a higher power output (30kV, 300mA) without causing overheating of the x-ray source. By use of this technique, Katayama (11) was able to reduce exposure times to 3 min for WAXS and 15 min for SAXS. As noted above, however, his sample was a rather large diameter monofilament, which would also increase the diffraction intensity.
The results of on-line measurements have normally addressed specific aspects of the fiber and film forming process. For example, the results of Haberkorn, et al., suggest that the occurrence of neck like deformation in high-speed spinning is related to the crystallization process (8). They used on-line measurement of diameter and temperature as well as WAXS to obtain their data. Filament diameter was measured using laser light scattering methods and filament temperature was measured with an infrared camera. The WAXS measurements taken online were done along the spinline using a Laue-type camera and a mobile x-ray generator. The film exposure time was 1 - 2 hours, and a guide stabilized the filament. Azimuthal and equatorial scans were taken from the patterns by densitometer measurements. Diameter profiles showed a definite necking phenomena for speeds above 4200 m/min, the solidification point moved up the spinline with increasing speed. The temperature profile above the necking range remained similar at all speeds indicating no effect on the cooling process. A maxima in temperature was, however, observed at all speeds, and was easily related to the enthalpy of crystallization that is released during the solidification process. Increasing mass throughput rate moved the solidification point away from the spinneret suggesting that necking occurs at the same diameter for any throughput. The corresponding temperature profiles also indicated that necking occurred at around the same temperature (± 2 C) regardless of throughput, or even the cooling conditions and orifice diameter. WAXS patterns showed an amorphous region above the solidification point. Weak reflection points were seen at the solidification point whose intensity rapidly increased and leveled off with position down the spinline. The estimated oriented crystallization also increased rapidly and leveled off within the first 10 cm from the solidification point. Again, the throughput rate did not seem to have any measurable effect on the crystallinity and crystallization rate. The length of the crystallization region along with the spinning speed gave as the half-time of crystallization t1/2 = D H/2v, which for PA66 is about 0.3ms at 5500m/min, and about 0.5 ms at about 4500m/min.
The work of Haberkorn cited above was performed on a 7-filament thread, whereas the majority of the published reports on XRD from moving filaments have used monofilament as the sample configuration. Ben Hsiao, et al., (9a) have described a set-up for monitoring structural changes during drawing and spinning. The sample they used was a 34 filament yarn. In order to meet the stated goal of this research, we eventually must be able to access single filaments of smaller diameter from a bundle of filaments.
NSF Research Support in Relation to Equipment
EEC-9731680 Center for Advanced Engineering Fibers and Films -- August 1, 1998 through July 31, 2003 - $12,000,000
This award established an Engineering Research Center (ERC) for Advanced Engineering Fibers and Films at Clemson University with the Massachusetts Institute of Technology as a partner institution. The vision of the CAEFF is to provide an integrated environment for the systems-oriented study of next generation fibers and films. The research program emphasizes the use of computation/visualization tools to overcome the barriers of experimental development, structure/property relationships, and control of structures. The investigations will target conventional polymers and processes as well as liquid-crystalline systems, metallo-organic systems, and intractable polymer systems. The new processes will include supercritical solution processing, in-situ processing, and self-assembly processing. The focus of each thrust is to develop integrated models that can predict the development of structure during fiber and film formation, enabling final properties to be estimated. Thus, molecular detail must be included into these process models. Verifying these models will require complete and precise measurements of the development of structure and orientation during fiber and film formation for a variety of precursors and processes. This equipment, along with existing complementary instrumentation for monitoring the development of structure, will provide these critical data. Funds for purchase of this equipment are available through an equipment grant from the NSF, the State of South Carolina and CAEFF.
Center Research Activities
The following paragraphs detail the three areas of application for the XRD equipment planned by Center researchers. These three topics can provide valuable data for testing of our initial models of structure development during fiber and film formation.
CAEFF Thrust 1
The primary CAEFF application area utilizing the XRD equipment is the determination of fiber and film structure in Thrust 1, Topic 2: Experimental Verification (M. Ellison, Topic Leader). In this topic, the XRD equipment is a critical part of the process monitoring system. Process melt temperature and pressure, and filament tension and speed are currently monitored on one of our fiber melt spinning extrusion systems; these data will soon be available on our other systems. We have used these systems in defining the relationships between pressure fluctuations in the spin pack, and filament uniformity (27). In addition, we used data from this on-line monitoring system to substantiate a model we developed of droplet deformation in blend spinning (23). Threadline temperature and speed are measured by calibrated IR camera techniques and by laser Döppler velocimetry, respectively; in the future, additional structural information will be garnered via on-line Raman spectroscopy (below). Using data from off-line measurements and the literature, our on-line results will enable the development of interrelationships between these process parameters and fiber and film structure. Subsequent testing of the physical properties of the materials will be used to establish the resulting structure/property relationships.
We are currently installing a Raman spectrometer for use in our on-line studies. We have a laboratory system in place. The on-line equipment will enable us to obtain data on the relative degree of orientation and crystallinity in the spinline. The addition of the x-ray equipment will allow us to "train" the Raman spectrometer in the identification of structural features so that we can more accurately assign quantitative values to the orientation and crystallinity data from the Raman (Alan Kennedy, DuPont Co., pers. comm.).
The kinetics of stress-induced crystallization is the central polymer physics issue to be explored in this Thrust (1.2) by application of this on-line XRD equipment. The study of this phenomena has been an important barrier to developing a thorough understanding of structure development in fibers and films, in particular at processing speeds which exceed 4,500 m/m. The extrusion facilitates in place in the CAEFF include winding capabilities of up to 7,000 m/m, enabling us to address these very important questions.
The data that we obtain from our on-line measurements will be crucial for the development and verification of models of fiber and film formation. In Thrust 1.1, these models are the basis for the visualization efforts that are central to the overall mission of the center. Additional information is available at the Index of CAEFF research groups.
CAEFF Thrust 2
The second Thrust area deals with alternate precursors, such as liquid-crystalline materials, sol-gel forming polymers, and solution-processed polymers. Data from real-time measurements by the XRD equipment will be used to verify models for the development of structure and orientation during fiber and film formation for each of these precursors. Additional information is available at the CAEFF website (above).
CAEFF Thrust 3
The third Thrust addresses issues dealing with so-called in situ processing. New processing methods are being developed to construct abundant fiber and film structures directly within melts. Whereas present-day polymer blending is characterized by breaking down structure to obtain a well-mixed state, it has been recently demonstrated that chaotic motion can be induced in molten minor phase bodies to cause multilayer films and fibers to evolve within major phase melts. In this process, characteristic dimensions continuously decrease sometimes exponentially with time. Initially large minor phase bodies refine in response to complex interactions between phase interfaces, induced fluid motion, and additive mass transfer. Upon capture by solidification, such structured blends can have significantly enhanced physical properties. Additional information is available at the CAEFF website (above).
General Description of Research Instrumentation and Needs
Although fiber spinning and film formation are similar in many respects, the physical processes are sufficiently different as to obviate the use of a single XRD system applied to both. Thus, this document applies to two XRD systems, one for a fiber line and one for a blown film line. Much of what follows are generic to both; differences, where critical, will be noted. The requisite engineering, however, may be significantly different for the two processes. Interested parties are urged to visit our laboratories to gain an understanding of the physical settings in which the XRD equipment will be housed. Contact either of the Co-PIs listed below.
Recognizing the difficulties inherent in on-line XRD, our first real time data will be obtained limited to the region of the threadline wherein the fibers and/or film are in the solid state and therefore have a stable position. We will determine the effects of the process variables on the morphology of the material in this region. Those data should fairly well agree with off-line data. We will begin with a simple polymer such as polyolefins, and then study polyamides and polyesters. When our expertise in the solidified region is confirmed, we will extend our research up the line toward the spinneret or die face. Thus, the engineering design must not contain any elements that preclude conducting XRD as near to the spinneret face as practical.
The limited literature resources notwithstanding, the efficacy of XRD for real-time determination of the development of structure during polymer fiber and film formation has been well documented, and the primary barriers identified. There are three, interrelated main areas of concern: the translation of the filament lateral to the beam direction, attaining sufficient incident beam intensity, and having sufficient detector sensitivity (the intensity of the diffracted beam).
Although little reference has been made in the cited literature to the lateral motion of the filament, that motion is presumably a significant source of signal deterioration in these measurements. As the filament moves, the diffracted signal must vary in strength and in sharpness. This motion will, of course, not directly influence the actual Bragg angle. However, this sample motion can severely affect accurate determination of degree of crystalline content (a function of peak intensity). Crystal size determination from information contained in the peak width, and estimation of d-spacing from data contained in the position of the peak, can both be compromised if the data is blurred. Finally, owing to the relatively small sample (fibers) combined with the small scattering cross-section will result in diminished scattering intensity.
This problem has been addressed by several methods, none of which are totally satisfactory. If one obtains data only in regions where the material is sufficiently solid (i.e., cool) so as to survive touching with a guide, the sample can be held in the beam. This clearly sets unacceptable limits on the regions of examination; in addition, the presence of a mechanical restraint modifies the process. Alternatively, very long exposure times can, when combined with careful analysis (8), also overcome these difficulties to a certain extent.
We suggest two approaches to this problem. The first is to design a means through which to know when the fiber is in the x-ray beam, and thus, through coincidence-type measurements, that the data is valid. As part of the larger program, we are designing a filament position instrument that works on the principles of laser diffraction and uses a high-speed CCD imaging device. This instrument will also provide diameter data. For certain polymers, or by the addition of fluorescing additives, x-ray fluorescence may also be used to determine when the filament XRD data is valid. Our second approach is to use the filament position as determined by the laser diffraction instrument to cause the x-ray beam to track the filament through manipulation of the beam focusing optics. Other approaches are invited.
Beam intensity has been improved of late by several technological advances. The initial intensity and the efficiency of the subsequent collimation of the incident beam have been improved remarkably. The use of higher-power sealed tubes for the x-ray source, and diffraction-based collimation and focussing have combined to result in significant increases in intensity. The concomitant improvement in frequency selection has resulted in high quality monochromatic x-ray beams.
Finally, with the advent and application of next-generation multi-wire, or gas-filled detector technology, and of sensitive imaging plate components, together with development of extremely fast hardware and software for precise decoding of photon position, significant improvements have been obtained in the resolution of diffraction patterns. The exposure times for film-type detectors are too long, even with the increased intensity of the diffracted beams. Photodetector systems (diffractometers), which are more sensitive, necessitate the acquisition of data from several scans at different angles in order to map the entire reciprocal space and determine the crystal structure parameters. Space limitations preclude the use of these linear detectors. Having the capability to obtain data from the entire space rapidly without scanning, significantly reduces data collection time.
Of paramount importance is the ability to collect both WAXS and SAXS on the running filament or film simultaneously. This will necessitate having an aperture in the detector nearest to the source (for wide-angle) through which the diffracted x-ray beam can pass to the small-angle detector.
The filament or the film must of course be processed under ambient conditions in air. Therefore, special engineering consideration must be given to designing evacuated tubes with windows transparent to the x-rays at the end that is positioned near the sample, with the detector or source at the other. Nevertheless, scattering of the x-rays by the air prior to detection is a barrier to success. One suggested approach to this problem is to establish the baseline which results from the air scatter, and attempt to remove it through software manipulation of the collected data.
The temperature of the melt can exceed 300C, and there are cooling air flows. This relatively high temperature present in the melt processing environment demands insulation and cooling of the instrumentation components near the high temperatures be designed into the system. Finally, there is an absolute need for radiation safety.
Instrument Summary
The instruments must have sufficiently high-powered x-ray sources and associated beam focussing elements to provide the required beam intensity. The detector system must provide excellent sensitivity and stability to be successful in this work. SAXS and WAXS must be obtained simultaneously. An acquisition time as short as possible must be achieved. We recognize that there are several legitimate approaches to meeting the challenges inherent in this project.
Project Management
Interested parties are encouraged to contact directly the following PIs for this project:
Dr. Michael S. Ellison, Professor
School of Textiles, Fiber and Polymer Science
864-656-5956
Dr. William Pennington, Associate Professor
Department of Chemistry
864-656-4200
References Cited
1. Alexander, L. E., "X-ray Diffraction Methods in Polymer Science", Krieger Pub. Co., Huntington, N.Y. 1979.
2. Anderson, D. P., "Carbon Fiber Morphology, Wide-Angle X-Ray Studies of Pitch and PAN-Based Carbon Fibers," Interim Report December 1988-April 1990, No. WRDC-TR-90-4137, University of Dayton Research Institute under Air Force Contract No. F33615-87-C5239, 1991.
3. Cakmak, M., Teitge, A., Zachman, H. G., and White, J. L., J Polym. Sci: Part B: Polym. Phys. 31, 371 (1993).
4. Chappel, F. P., Culpin, M.F., Gosden, R. G., and Trantor, T. C., J. Appl. Chem., 14, 12 (1964).
5. Chen, S., Yu, W. and Spruiell, J. E., J. Appl. Polym. Sci. 34, 1477 (1987).
6. Dauche, F. M., "High Performance Carbon Fibers from Mesophase Produced by Supercritical Fluid Extraction," Ph.D. Dissertation, Clemson University, Clemson, SC, 1997.
7. Folkes, M. J. and Holloway, M. J. Engr. Plast., 5, 157 (1992).
8.Haberkorn, H., Hahn, K., Breuer, H., Dorrer, H. D., and Matthies, P., J. Appl. Polym. Sci., 47, 1551-1579 (1993).
9a.Hsiao, B.S., Barton, Jr., R., Quintana, J., Simple On-Line X-Ray Set-Up tp Monitor Structural Changes During Fiber Processing, J. Appl. Polym. Sci., 62, 2061 (1996).
9.Hosangadi, G. S. M. S. Ellison, and J. N. Gowdy, "Signal Processing Methods for Melt Pump Outlet Pressure Analysis," SE-CON, IEEE, Raleigh, NC, March 1995.
10. Iannelli, P., Macromolecules 26, 2303 (1993).
11. Katayama, K., Amaon, T., and Nakamura, K., Kolloid Z., Z. Polym., 226, 125 (1968).
12. Kikutani, T., Kazuhito, N., Takarada, W and Ito, H., Int. Symp. Orient. Polym, Montreal, Sept., 1998.
13. Kudo, K., Mochizuki, M., Kiriyama, S., Watanabe, M., Hirami, M., J. . Appl. Polym. Sci. 52, 861 (1994).
14. Lu, F. M. and Spruiell, J. E., J. Appl. Polym. Sci. 34, 1541 (1987).
15. Lu, F. M. and Spruiell, J. E., J. Appl. Polym. Sci. 40, 623 (1993).
16. Mark, J. E., ed. "Physical Properties of Polymers Handbook," AIP, Woodbury, NY, 1996, p. 302.
17. Miyata, K., Kikutani, T., Okui, N., J. Appl. Polym. Sci. 65, 1415 (1997).
18. Mortimer, S. A., and Peguy, A. A., Text. Res. J. 64, 544 (1994).
19. Murthy, N. Sanjeeva, Text. Res. J. 67, 511 (1997).
20. Murthy, N. S.; Correale, S.T.; Minor, H., Macromol. 24, 1185 (1991).
21. Murthy, N. S.; Correale, S. T.; Moore, R. A. F., J. Appl. Polym. Sci: Applied Polymer Symposium, "Science and Technology of Fibers and Related Materials: Proceedings of the Fiber Society 50th Anniversary Technical Conference Aug 19-23 1990" (1991).
22. Nothwang, W., "Bench-Scale Optimization of Ribbon-Shaped Carbon Fiber Production", Masters Thesis, Clemson University, Clemson, SC, 1997.
23. Padsalgikar A. D. and Ellison, M. S., , Polym. Engr. and Sci. 37: 994, 1997.
24. Rabiej, S., Wlochowicz, A., J. Appl. Polym. Sci. 46, 1205 (1992)
25. Ramesh, C., Gupta, V. B., and Radhakrishnan, J., J. Macromol. Sci. - Phys., B36, 281 (1997).
26. Samon, J. M., Schultz, J. M., and Wu, J., Personal Comm. (Presented at APS March, 1998).
27. Shumpert, B. B., Padsalgikar, A. D., Hosangadi, G. S., and Ellison, M. S., Int. Polym. Process. XI: 347, 1996.
28. Van Eijndhoven-Rivera, M. J.; Wagner, N. J.; Hsiao, B., J. Polym. Sci., Part B: Polymer Physics 36, 1769 (1998).
29. Wang, Y., Cakmak, M., and White, J. L., J. Appl. Polym. Sci., 30, 2615 (1985).
30. White, J. L. and Cakmak, M., Adv. Polym. Tech., 6, 295 (1986).
31. Ziemenski, K. F. and Spruiell, J. E., J. Appl. Polym. Sci. 35, 2223 (1988).