Home

Polyimide-Organosilicate Hybrid Materials

1. eeesesssssss 40 Table 3 2 Summary swelling measurements in NMP for GFDA 6FpDA DABA 25 and 6FDA 6FpDA DABA 12 5 hybrid materials Surface area to volume ratio is 7128 Amount of NMP is 200 grams for every gram of sample Error 10 49 Table 3 3 Summary of Permeability Diffusivity and Solubility coefficients for 6FDA 6FpDA DABA 25 hybrid materials Evaluated at 35 C and 4 atm absolute Overall error 5 for permeability ssssss 52 Table 3 4 Summary of Permeability Diffusivity and Solubility coefficients for 6FDA 6FpDA DABA 12 5 hybrid materials Evaluated at 35 C and 4 atm absolute Overall error 5 for permeability essssssssssss 53 Table 3 5 Summary of density measurements before and after annealing All measurements were evaluated at 25 C Relative error 1 0956 63 Table 4 1 Swelling data for the polyimides and polyimide organosilicate hybrid materials Both hybrid materials consist of an organosilicate covalently bonded to a 6FDA 6FpDA D7ABA polyimide matrix With the exception of the MTMOS based hybrid all materials swelled to such a degree that measurements were not DOSSO DID rr E COEM emer vom T E t 74 xi Chapter 1 Literature Review 1 1 Introduction During the past two decades gas separations using polymeric membranes have gained ground in the chemical process industry Separations
2. DIG nb exe PIOOram is is d is diea ssa uas rH EHE Re AREA Eia AREN 91 AO Man Menisiaastdatetceuitait en bn dne deu 91 A 6 Description of Buttons Left Side essssssssssss 93 A 7 Description of Blocks Bottom of Screen ssssssuss 94 A 8 batchrun exe Program Batchrun sse 103 A 9 Main Menu ES HIPH CR 103 A 10 1210141010 RPR ATE ATETA TTE AA 103 A 11 Testing Consecutive Gases sssssssssssssesse eene 104 E E E A A A R 106 vii List of Figures Figure 1 1 Sample results from a gas permeation test using the Time Lag uie PEE 6 Figure 1 2 A plot illustrating the permeability as a function of pressure for a material that has been plasticized adapted from Sanders et al For comparison typical results from the Dual Mode Sorption Theory are shown The pressure at which the upswing occurs varies widely from polymer to polymer 17 Figure 2 1 Chemical structures of TMOS MTMOS and PTMOS 27 Figure 2 2 Chemical structure of the 6FDA 6FpDA DABA polyimide functionalized with APTEOS blue The DABA group is highlighted in red The 6FDA 6FpDA repeat unit is highlighted in green sssesesssssss 28 Figure 2 3 Schematic diagram of the permeation system All valves and pressure transmitters are interfaced with a computer sssesssssss 31 Figure 2 4 Cross sectional schematic diagram of
3. diameter and slightly increased in thickness upon annealing 62 Table 3 5 Summary of density measurements before and after annealing All measurements were evaluated at 25 C Relative error 1 0 Alkoxide Before Anneal After Anneal Change DABA 25 g cm g cm Pure Polymer 1 472 1 468 0 3 22 5 TMOS 1 508 1 510 0 2 15 0 TMOS 22 596 MTMOS 1 427 1 434 40 596 15 0 MTMOS 1 451 1 449 0 296 22 5 PTMOS 15 0 PTMOS 1 443 1 444 40 196 Alkoxide Before Anneal After Anneal 96 Change DABA 12 g cm g cm Pure Polymer 1 468 1 469 0 0 22 5 TMOS 1 511 1 518 40 496 15 0 TMOS 22 596 MTMOS 1 435 1 435 0 0 15 0 MTMOS 1 439 1 442 0 2 22 5 PTMOS 1 435 1 420 1 1 15 0 PTMOS 1 442 1 457 1 1 63 However a more likely explanation for the large increases in permeability and diffusivity may be that the molecular chain relaxation is enhanced during the annealing process Unfortunately the effects of chain relaxations on gas diffusion are not well characterized in the literature We believe that heating the polymer above the T caused a relaxation of the polymer chains and subsequently improved the molecular packing and enhanced the chain mobility 59 Physical changes to the polymer such as free volume and chain mobility would affect the diffusivity more than the solubility which would be consistent with our permeation data Additionally this heat treatment may have induced a more uniform red
4. 2 3 The two stage vane vacuum pump is an Alcatel 2010 SD model with a capacity of 10 L min The temperature box consists of 3 4 plywood lined with 34 Styrofoam insulation and a small fan to circulate the air An Omegalux silicone rubber flexible heater is coupled with an electronic digital thermometer to provide heat and maintain the desired temperature An MKS Baratron 722A absolute pressure transducer with a range of 0 to 100 Torr and a resolution of 0 196 of full scale is used to measure the pressure of the permeate vacuum side An Omega PX 621 pressure transducer with a range of 0 to 300 psig is used to measure the feed pressure An Omega resistance thermal detector RTD with a resistance of 100 ohms is used to measure the temperature of the feed gas An Accel 486 computer using LABTECH Notebook graphical interface programming software controls a series of Nupro solenoid valves The computer uses a Keithley Systems DAS 802 input board to control the valves and a Keithley Systems PIO 24 input output board to control the measurement devices All valves and measurements devices are connected using 1 8 or 1 16 OD 316 stainless steel tubing and appropriate SwagelokQ fittings 30 Pressure Pressure Control Valves Gas Supply Vacuum Pump Figure 2 3 Schematic diagram of the permeation system All valves and pressure transmitters are interfaced with a computer 31 2 5 Gas Permeation Cell A 316 stainless s
5. A Psolid unknown density of solid sample P sia ASH P jiquia P liquid density of liquid A mass of sample in air B mass of sample in liquid Iso octane Piiquia 0 688 g ml at 25 C was employed as the liquid for these measurements because it will not dissolve the samples vapor pressure is low toxicity is low and low surface tension aids in wetting the samples Sample films of the pure polyimide and hybrid materials were first subjected to 180 C for 24 hours under vacuum and then measured using this technique The average density of three different samples ranging 40 to 120 mg in mass was recorded Relative error for this process was 0 5 or 0 004 g cm It is important to note that buoyancy in air is not taken into account and may be corrected by adding 0 001 g cm The change in liquid volume when the unknown sample is submerged is assumed to be negligible 37 Chapter 3 Annealing Polyimide Organosilicate Hybrid Membranes 3 1 Abstract Polyimide organosilicate hybrid membranes were subjected to annealing to enhance gas separation performance These membranes consisted of organosilicate domains covalently bonded to a 6FDA 6FpDA DABA polyimide using partially hydrolyzed tetramethoxysilane TMOS methyltrimethoxysilane MTMOS or phenyltrimethoxysilane PTMOS The transport properties of the hybrid membranes were evaluated using pure gases He O2 N CH4 CO at 35 C and a feed pressure of 4 atm The permeability
6. C J Cornelius E Marand Hybrid Inorganic Organic Materials Bases on 6FDA 6FpDA DABA Polyimide and Silica Physical Characterization Studies Polymer 2002 43 2385 C J Cornelius E Marand Hybrid Inorganic Organic Materials Based on a Series of Silica and Polyimide Composites Gas Transport Properties Journal of Membrane Science submitted May 2001 C J Cornelius C Hibshman E Marand Hybrid Organic Inorganic Membranes Separation and Purification Technology 2001 25 181 C J Cornelius Hybrid Inorganic Organic Materials Physical and Gas Permeation Properties for a Series of Fluorinated Polyimide Composites Ph D Dissertation Virginia Polytechnic Institute and State University July 2000 K Ghosal B D Freeman Gas Separation Using Polymer Membranes An Overview Polymer for Advanced Technologies 1994 5 673 W J Koros G K Fleming Membrane based gas separation Journal of Membrane Science 1993 83 1 L M Robeson Correlation of Separation Factor Versus Permeability for Polymeric Membranes Journal of Membrane Science 1991 62 165 M Smaihi J C Schrotter C Lesimple I Prevost C Guizard Gas Separation Properties of Hybrid Imide Siloxane Copolymers with Various Silica Contents Journal of Membrane Science 1999 161 157 R Tamaki Y Chujo K Kuraoka T Yazawa Application of Organic Inorganic Polymer Hybrids as Selective Gas Permeation Membranes Journal of Materials Chemistr
7. As a result relatively few polymers have the necessary combination of favorable mechanical chemical and gas separation properties 21 Fluorinated polyimides are particularly appealing materials for use as membranes because they have high temperature stability and have favorable transport properties Recent efforts to combine polymers with sol gel derived organosilicates have surfaced in the literature attempting to arrive at materials with enhanced separation characteristics Smaihi et al synthesized a polyimide organosilicate hybrid system using polycondensation imidization and sol gel chemistry processes A polyimide PMDA was coupled to the organosilicate TMOS via aminopropyltrimethoxysilane APrTMOS and aminopropylmethyldiethoxysilane APrMDEOS producing a homogeneous microstructure The organosilicate network was formed during sol gel co reaction of the hybrid polyamic acid siloxane solution The authors reported a decrease in gas permeability with increasing TMOS content Some gas pair selectivities increased with increasing organosilicate content while other selectivities peaked at moderate concentrations of TMOS It was suggested that the methyl side groups of the TMOS did not interact with the permeating gas but contributed to the modification of the polymeric network namely influencing the degree of cross linking This degree of cross linking was determined to be the limiting factor to the gas transport Tamaki et al
8. DABA content in the polyimide matrix In each of the figures the pure polyimide refers to 6FDA 6FpDA DABA with no organosilicate content which is used to qualify the properties of the hybrid systems For the DABA 12 polyimide based systems shown in Figures 3 20 and 3 22 the normalized permeability tends to increase with increasing molecule size and the normalized selectivity generally decreases with increasing molecule size This can be interpreted as the annealing process favors the permeability of larger molecules but does not favor the selectivity of gas pairs with large diameter differences These results may be a consequence of the polymer chains experiencing a relaxation mechanism during the annealing process which allows for the redistribution of free volume while maintaining localized cooperative motions Different observations were noted for the DABA 25 polyimide based systems in Figures 3 19 and 3 21 First for the MTMOS and TMOS based hybrids the generalization that normalized permeability increases with increasing molecule size was broken at the nitrogen permeation Second and most notable 65 is the increasing selectivity with annealing of the CO2 CH gas pair for the MTMOS and TMOS based membranes Undoubtedly these observations correspond to the promising results presented in the boundary diagrams in section 3 7 It also suggests that the MTMOS and TMOS based hybrids have similar mechanisms of diffusion whereas the PTMOS ba
9. Transform Infrared Spectrometry Attenuated Total Reflectance was employed to detect chemical bonding changes in the materials after the annealing process with BIO RAD FTS 40A Spectra were obtained sampling 64 scans at a wave number resolution of 2 cm and aperture opening of 2 cm The angle of incidence was 45 and a KRA 5 SPP crystal Thallium Bromoiodine Si and Single Pass Parallelepiped with dimensions of 50mm x 10mm x 3mm was used Based on these dimensions the IR beam is reflected approximately 18 times 2 8 Thermogravimetric Analysis Mass Spectrometry TGA MS was performed in a High Resolution TA 2950 TGA instrument was coupled to a Pfeiser Thermostar mass spectrometer Sample sizes ranging from 6 5 to 7 5 mg were subjected to a temperature sweep of 20 to 960 C ramped at 10 C per minute under nitrogen flow of 90 uL min Mass was measured every 2 seconds and the effluent gas was simultaneously scanned every 100 seconds 2 9 Density Measurements The densities of the samples were obtained using a Mettler AJ100 analytical balance fitted with a Mettler ME 33360 density determination kit This method is based on the Archimedean Principle Simply stated a solid immersed in a liquid loses as much of its own weight as the weight of liquid it has displaced 36 Knowing the density of the liquid the mass of the sample in air and the mass of the sample in the liquid the unknown density can be determined using the equation
10. V and the known volume V are filled with an ideal gas at a known pressure P Note V also includes the small volume of tubing used to connect the known volume to the system This volume is calculated and added to the known volume Valve 1 is closed so we have a known volume V and pressure P of an ideal gas The permeate side of the system is exposed to vacuum by opening Valve 2 When the pressure of the permeate side P2 is lowered Valve 2 is closed and P2is recorded Valve 1 is opened and an equilibrium pressure Peg is recorded The unknown volume V is calculated using Equation 2 1 The temperature and amount of gas are assumed to be constant for the calculation V P P V MO E P 2 1 P Pa For this experiment we used aluminum foil as the nonpermeable membrane helium as the ideal gas and a known volume of 40 cm Of course more accurate volume calculations result from larger differences in pressures P and Ps The volume of the permeate side for this system including the tubing to 34 the pressure transducer is 0 900 0 004 cu in 14 74 0 06 cm This volume is the average calculation of 14 trials with pressure differences ranging from 2 to 7 atm Pressure Permeation Valve Known To permeation system i Unknown I EM Valve 2 To gas supply Vacuum Pump Figure 2 5 Schmeatic diagram of measuring permeate volume using an ideal gas 35 2 7 FTIR ATR Fourier
11. Vu OA KERN R RECO LHRRMRRN UE cR EYE VM N VERMES iv Acknowledgements aine nnnc ran aai bu n a ra Fa Ran dE ara ra cR DA FR RA dag Rea V INA C viii BILSELTS xi Chapter 1 Literature Review eeeeeeeeeeeeeeeeeeeneeeee nennen nennen 1 1T IDOQUCHO sce dne tee tu rdv av aa i eiu eos 1 1 2 Gas Transport through Polymeric Membranes suse 2 LS Solution Difusion Model ane na m ras 3 1 4 TROT AG Method os cere dete cas doque dedu asia cadeau vn meinen E irae dena ed Sd 4 1 5 Dual Mode Sorption Theory entrenar ad rua areae 9 1 6 Gas Transport through Hybrid Membranes sueesssusss 12 1 7 GOsPIASucizanon Theory ied ededs dot ids icd ti tui d east ded edbto ded saves ded edd 14 1 8 Organic Inorganic Membranes using Sol Gel Chemistry 18 1 9 Polymer Organosilicate Hybrid Materials sssssssssss 21 1 10 Annealing Polyimide Organosilicate Hybrid Membranes 24 Ghapter 2 Experimental oon di dde bini dad s ea CE epa Sae cients 26 2 1 Membrane Materials alice heeds vee aaa bea ee Da Via ORA MEE Gea APT 26 22 Annealimng Proce dU a a a eaa a a aaa aa aa iita i ea iaaa 29 2 3 Gas Permeation General Information esssesessessss 29 2 4 Gas Permeation Equipment zc cu IR ONERE I EDH ERU 30 Zor Gas P tmeat
12. although at temperatures above 400 C they begin to degrade as has been well documented Although carbon molecular sieve CMS membranes typically use a polyimide as starting material the pyrolyzing temperature gt 500 C is much higher than the annealing temperature in this study 400 C 5 4 25 Chapter 2 Experimental 2 1 Membrane Materials All membranes were previously synthesized and characterized by Dr Chris Cornelius A detailed description of the synthesis can be found elsewhere To summarize a series of polyimide organosilicate hybrid materials were synthesized from 80K My 6FDA 6FpDA DABA polyimides containing 12 5 and 25 0 mole percent DABA with respect to the total diamine content during synthesis The nomenclature 6FDA 6FpDA DABA refers to 4 4 Hexfluoroisopropylidene diphthalic Anhydride 6FDA 4 4 Hexfluoroisopropylidene dianiline 6FpDA and 3 5 Diaminobenzoic Acid DABA The notation DABA 12 and DABA 25 is used to reflect the respective DABA contents of the two different polymers studied Three different types of alkoxide precursors were considered in the sol gel chemistry and are illustrated in Figure 2 1 tetramethoxysilane TMOS methyltrimethoxysilane MTMOS and phenyltrimethoxysilane PTMOS The type of alkoxide precursor in the organosilicate domain had an effect on the physical and gas permeation 26 characteristics These effects were due to the functionality and steric hindranc
13. annealed a non crosslinked aromatic polyimide Matrimid 5218 below 15 Tg to form charge transfer complexes which restricted chain mobility and the effects of plasticization Crosslinking the polymer is another technique for suppressing CO2 plasticization Unfortunately a decrease in permeability usually accompanies increases in crosslinking Bos et al have blended a polyimide Matrimid and an oligomer Thermid and heated the film at 265 C to effectively crosslink the polymer blend The crosslinked Matrimid Thermid film suppressed CO plasticization whereas the non crosslinked films did not In a separate study Staudt Bickel and Koros have shown that increasing the crosslink density in fluorinated polymer membrane will deter plasticization effects up to 20 atm Both studies effectively reduced plasticization via crosslinking and without significantly reducing the permeability Dual mode Plasticized e Permeability Barrer Pressure atm Figure 1 2 A plot illustrating the permeability as a function of pressure for a material that has been plasticized adapted from Sanders et al For comparison typical results from the Dual Mode Sorption Theory are shown The pressure at which the upswing occurs varies widely from polymer to polymer 1 8 Organic Inorganic Membranes using Sol Gel Chemistry First the term sol gel needs to be defined A sol is a colloidal suspension of solid particles in a liquid
14. for most of the membranes increased 200 500 after the annealing process while the permselectivity dropped anywhere from 0 to 50 The exceptions were the 6FDA 6FpDA DABA 25 22 596 TMOS and MTMOS hybrid membranes both of which exhibited increases in the CO permeability and CO2z CH permselectivity The transport data was compared to Robeson s 1991 upper bound and exceeded the boundary in some cases The increase in permeation was attributed to increases 38 in the free volume and enhanced segmental mobility of the chain ends resulting from the removal of sol gel condensation and polymer degradation byproducts 3 2 Visual Observations Before annealing the membranes were optically transparent and flexible After annealing some membranes were still transparent while others were opaque Although annealing caused the membranes to be somewhat brittle they were still flexible and durable 3 3 TGA MS The four samples shown in Table 3 1 were analyzed using TGA MS to determine if the organic ligands methyl and phenyl were being decomposed and to qualitatively measure the amount of polyimide degradation The 400 C annealing temperature was higher than the T for all samples and the total weight loss did not exceed 3 096 when the sample was heated to the annealing temperature The percent weight loss in the hybrid materials was only slightly higher than that observed for the pure polyimide which would be consistent with the loss
15. gas permeation cell 33 Figure 2 5 Schmeatic diagram of measuring permeate volume using an ideal gas uote tait a HL ALL Le t ee Oe ge es A ee Eee eae 35 Figure 3 1 TGA MS spectra for GFDA 6FpDA DABA 25 pure polyimide degradation Intensity units are arbitrary sss 41 Figure 3 2 TGA MS spectra for GFDA 6FpDA DABA 25 22 5wt PTMOS hybrid material Intensity units are arbitrary essssseeeeeeee 41 Figure 3 3 FTIR ATR spectra of 6FDA 6FpDA DABA 25 pure polyimide Shading highlights areas of change after the annealing process 44 Figure 3 4 FTIR ATR spectra of 6FDA 6FpDA DABA 25 22 5 MTMOS based hybrid Shading highlights areas of change after the annealing process 44 Figure 3 5 FTIR ATR spectra of 6FDA 6FpDA DABA 25 22 5 TMOS based hybrid Shading highlights areas of change after the annealing process 45 Figure 3 6 FTIR ATR spectra of the 6FDA 6FpDA DABA 25 22 5 PTMOS Shading highlights areas of change after the annealing process 45 Figure 3 7 Thermal hydrolytic degradation of anhydrides 46 Figure 3 8 TEM images of 22 5 MTMOS based hybrid material before and after the 400 C anneal for 30 minutes esses 50 Figure 3 9 Boundary diagrams of 6FDA G6FpDA DABA 25 pure polyimide and hybrid materials for He Oz The line represents Robeson s 1991 Upper Bound The blu
16. into the sol gel process that are later removed by calcination leaving a continuous network of mesopores and micropores Membranes created by this method exhibit an increase in gas flux proportional to the organic ligand volume The gas selectivity of these membranes is dependent upon the resulting pore size and shape left by the templating group Raman et al demonstrated this technique by using a methyltriethoxysilane MTMOS and tetraethoxysilane TMOS system to show that the organic components can be removed by calcining the membranes at 550 C At a calcination temperature of 400 C permeance increased by a factor of 10 while CO2 CHs selectivity dropped by about 20 At a calcination temperature of 550 C the CH permeance decreased dramatically while the CO2 CH selectivity increased The authors concluded that the organic ligands were removed from the gel at 550 C when the network collapsed to create a fully densified silica network and suggested that the increased selectivity was a result of a molecular sieving mechanism They also noted that the constraint imposed by the underlying alumina support may have affected their results by limiting the densification of the inorganic network to temperatures above 400 C In a separate study Lu et al pyrolyzed methacryloxypropyl ligands from a methacryloxypropylsilane MPS TEOS sol gel system at 350 C The permeation data was consistent with molecular simulations and verified that a second
17. of byproducts from the sol gel condensation reactions 39 Table 3 1 Summary of Tg and 5 weight loss temperatures for GFDA 6FpDA DABA 25 pure polyimide and hybrid materials T CC 59 WiLoss c e ib Loss al Pure Polyimide 303 432 2 496 22 5 wt PTMOS 328 448 2 5 22 5 wt TMOS 340 438 2 8 22 5 wt MTMOS 340 459 2 6 DSC data collected at 10 C min in air and 2 ramp error 0 7 TGA data collected at 10 C min in nitrogen error 0 3 Carbon dioxide m z 44 28 12 water m z 18 HF m z 220 and CF m z 69 were each detected for all samples during a temperature scan from 20 to 960 C 9 As shown in Figure 3 1 carbon dioxide m z 44 28 12 removal occurred at two different temperatures 400 C and 520 C for all samples which is consistent with the literature However no peaks were observed at m z 28 and 14 because nitrogen was used as the carrier gas and thus dominated the signal at these values Since carbon monoxide peaks m z 28 12 coincide with the nitrogen and the carbon dioxide peaks it was not possible to verify the presence of carbon monoxide but several groups did report it as a common degradation product for polyimides using other techniques 40 1 0E 08 1 0E 09 4 Intensity lieth Te ect hec 1 0E 10 100 200 300 400 500 600 700 800 900 1000 Temperature C Figure 3 1 TGA MS spectra for GFDA 6FpDA DABA 25 pure polyimide degra
18. of feed pressure for various polyimide and hybrid systems at 35 C The permeabilities were calculated using the appropriate fugacity Error 2 ssuuuuss 73 Figure 4 4 CO CH ideal selectivity plotted as a function of feed pressure for various polyimide and hybrid systems at 35 C Error 496 74 Figure 4 5 Plot of pure CH diffusivity as a function of feed pressure for various polyimide and hybrid systems at 35 C The values were calculated using the appropriate feed pressure sssssssssssssssessessssessssseessessseeesesseeeseesesseeeeeee 78 Figure 4 6 Plot of pure CO diffusivity as a function of feed pressure for various polyimide and hybrid systems at 35 C The values were calculated using the appropriate fugacity accetta tacite aha aee bcn cen eres ee pete Re etd ee Drs et E DET EVE ea dE 78 Figure 4 7 Plot of pure CH solubility as a function of feed pressure for various polyimide and hybrid systems at 35 C The values were calculated using the appropriate feed pressure 2 5 2 5 2 02 27 1301 cR S rh dara Da Erik revu nnna nnmnnn nanana 79 Figure 4 8 Plot of pure CO solubility as a function of feed pressure for various polyimide and hybrid systems at 35 C The values were calculated using the appropriate Mille lo i oce Ime 79 List of Tables Table 3 1 Summary of T and 5 weight loss temperatures for GFDA 6FpDA DABA 25 pure polyimide and hybrid materials
19. related to the gas concentration gradient within the membrane Transient diffusion through a uniform slab can be described using Fick s First Law J D C EJ 1 1 Ox where Jis the flux cm STP cm s Dis the diffusion coefficient cm s C is the concentration of the gas in the polymer cm GSTP cm and xis the thickness of the membrane cm Diffusion is only in the x direction assuming the concentration gradients in the other directions are negligible Assuming D is independent of temperature and that the concentration obeys Henry s Law Equation 1 1 simplifies to J DS x 1 2 where S is the solubility coefficient At steady state Equation 1 2 simplifies can be written as e pxs Pte 1 3 where J is the flux at steady state pris the feed pressure pp is the permeate downstream pressure and is the thickness of the membrane Thus the permeability P defined as the product D and S can be substituted into Equation 1 3 resulting in J P dr e 1 4 To model real gases the fugacity of the gas f may replace the pressures as shown in Equation 1 4 This would be particularly important at pressures approaching non ideal conditions 1 4 Time Lag Method The Time Lag Method allows one to determine the permeability P solubility S and diffusivity D coefficients The time lag is simply the amount of time required for a gas to permeate through a membrane The integral technique was used in thi
20. that if total immobilization occured the permeability would be constant with increasing pressure If the gas penetrant molecules were not completely immobilized the permeability would decrease with pressure increases Another deviation stems from the assumption of equilibrium testing using glassy polymers that are inherently in a state of nonequilibrium The Free Volume Theory has been employed to predict the diffusion coefficient of polymers Pace and Datyner developed a simple model using four parallel chains to form a cage around a gas molecule The gas molecule will jump from cage to cage provided a sufficient activation energy is available to make this diffusive jump The theoretical relationship for this model is shown in Equation 1 20 where Ep is the activation energy required for the diffusive jump o is the diameter of the gas molecule and CED is the cohesive energy density of the polymer A is the jump length which needs to be estimated since there is no empirical methods for determining its magnitude 2 ae ATO CED 1 20 11 The limitation of this model is the parameters are difficult to determine experimentally and do not apply to heterogeneous hybrid systems Other models such as Non Equilibrium Lattice Theory and Activated Complex Theory attempt to describe gas transport through glassy polymers using statistical mechanics non equilibrium thermodynamics and structural arguments These models each have their ad
21. the gas molecules adsorbing into free volume holes where the gas penetrants are assumed immobilized These two contributions can be added to form equation 1 18 C a bp 1 bp C C C k p 1 18 where kpis Henry s law constant b is the hole affinity constant C y is the concentration in the holes at saturation and p is the partial pressure of penetrant in the gas phase These constants are determined empirically to represent data collected from sorption isotherms To employ the Dual Mode Sorption Theory experimentally the parameters C 4 and b are calculated using equilibrium sorption isotherms for a specific glassy polymer Using equation 1 18 the solubility coefficient S kp can be determined The permeability coefficient P is determined via a steady state permeation test see Solution Diffusion Model and the diffusivity coefficient is defined as D P S see Equation 1 12 This model has been fairly effective at 10 describing the transport of gases through glassy polymers However this model does have its limitations In particular real polymer systems are dynamic which leads to variations that cannot be explained with the Dual Mode Sorption Theory One of these variations is the partial or incomplete immobilization of gas penetrant molecules D R Paul et al investigated the validity of assuming that the gas penetrant molecules are immobilized in the holes of a glassy polymer 9 9 They predicted
22. wt PTMOS 1 25 0 55 22 5 wt MTMOS 1 01 0 19 22 5 wt TMOS 1 18 0 04 DABA 12 Before Anneal After Anneal Pure Polyimide 0 17 22 5 wt PTMOS 0 49 22 5 wt MTMOS 1 13 0 60 22 5 wt TMOS 2 37 0 76 Indicates sample swelled to such a degree that measurements was not possible 49 3 6 TEM Studies TEM images of the 22 5 MTMOS based hybrid material before and after annealing are shown in Figure 3 8 The TMOS and PTMOS based materials are not shown because the size of the silica domains on the order of nanometers was too small to discern However the MTMOS based hybrid had larger micron sized silica domains in addition to smaller nano sized domains The larger silica domains that were visible in the MTMOS based hybrid demonstrated a significant change with the annealing process The wavy striations that accompany the annealed sample are an artifact of sample microtoming These artifacts indicate that the silica domains become much harder after the annealing process thereby supporting the conclusion that the condensation reactions were driven to a higher degree of completion resulting in more Si O Si bonding in the silica structures Before Anneal After Anneal Figure 3 8 TEM images of 22 5 MTMOS based hybrid material before and after the 400 C anneal for 30 minutes 50 3 7 Gas Transport Gas permeation experiments were performed to determine the effects of the annealing process on gas transport properties of th
23. 090 and 1180 4 All three hybrid samples TMOS PTMOS and MTMOS had significant changes between 960 1280 cm evidence that the silica bonding was changing during the annealing process For example in Figure 3 4 the peak at 1020 cm asymmetric stretching vibration Si O Si nearly doubles when compared to the benzene peak at 716 cm This indicates that more Si O bonds were formed and 43 that the gel condensed as a result of annealing The peak at 716 cm remained constant because the number of benzene rings did not change Benzene Ring Polyimide a C O Cr ma f Before F FH C N Van After AUVMA n a IN 2000 1800 1600 1400 800 600 400 1200 1000 Wave Number cm Figure 3 3 FTIR ATR spectra of 6FDA 6FpDA DABA 25 pure polyimide Shading highlights areas of change after the annealing process Si Bondin dis Benzene Ring Si CH sic ps 4 Before Anneal j C 0 i Wu C N NEED EET imi Polyimide After Anneal o E 0 2000 1800 1600 1400 1200 1000 800 600 400 Wave Number cm Figure 3 4 FTIR ATR spectra of 6FDA 6FpDA DABA 25 22 5 MTMOS based hybrid Shading highlights areas of change after the annealing process 44 Benzene Ring Si Bonding Before Anneal M NU ws rors we After Anneal aaa s 1 2000 1800 1600 1400 1200 1000 800 600 400 Wave Number cm Figure 3 5 F
24. 1 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed membranes Upper Bound Pure Polymer 22 5wt MTMOS 15 0wt MTMOS 22 5wt PTMOS 15 0wt PTMOS 22 5wt TMOS Pure Polymer 22 5wt MTMOS 15 0wt MTMOS 22 5wt PTMOS 15 0wt PTMOS 22 5wt TMOS o b m o9tb n O N Selectivity a 5 6 7 8 9 10 20 30 40 O Permeability barrers Figure 3 12 Boundary diagrams of 6FDA 6FpDA DABA 25 pure polyimide and hybrid materials for Oz Nz The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed membranes 55 Upper Bound Pure Polymer 22 5wt MTMOS 15 0wt MTMOS 22 5wt PTMOS 15 0wt PTMOS 22 5wt TMOS Pure Polymer 22 5wt MTMOS 15 0wt MTMOS 22 5wt PTMOS 15 0wt PTMOS 22 5wt TMOS 70 60 50 40 gt DrPee DOE CO CH Selectivity o 30 20 30 40 50 60 70 8090100 200 CO Permeability barrers Figure 3 13 Boundary diagrams of 6FDA 6FpDA DABA 25 pure polyimide and hybrid materials for CO2 CH 4 The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed membranes Upper Bound Pure Polymer 22 5wt MTMOS 15 0wt MTMOS 7 5wt MTMOS 22 5wt PTMOS 15 0wt PTMOS 22 5wt TMOS Pure Polymer 22 5wt MTMOS 15 0wt MTMOS 7 5wt MTMOS 22 5wt PTMOS 15 0wt PTMOS 22 5wt
25. 2 ff D 3 5 1 5 4 L o j f 5 Unsteady state f Steady state P4 1 ou 0 5 0 N T 0 2000 4000 6000 8000 Time seconds Figure 1 1 Sample results from a gas permeation test using the Time Lag Method The diffusion coefficient can be calculated from the time lag using a derivation of Fick s Second Law which is shown in Equation 1 8 2 C_p C ot ox 1 8 The following boundary condition are imposed to solve Equation 1 8 t 0 O0 lt x lt l C 0 t20 x 0 C Sp constant t gt 0 x C 0 Applying these boundary conditions results in Equation 1 9 where Qis the total amount of gas permeating through the membrane C is the concentration at the feed side and tis time This equation models one dimensional flow through a membrane assuming that D is independent of temperature and concentration o f 1p 2x 2 Q Dt 2 l e 4 9 IC D Ob n When tis allowed to go to infinity thereby representing steady state diffusion the summation in Equation 1 9 approaches 1 6 and is simplified to Q Dt 1 ic are 1 10 Rearrangement of Equation 1 10 results in es EE E J 2e r E 1 11 I 6D i where 0 F 6D This provides a direct route for calculating the diffusion coefficient Knowing P and D and that P D x S the solubility coefficient S can be calculated using Equation 1 12 S 1 12 d D However due to the nonequilibrium properties associated with all glass
26. 7 1811 20 D R Paul W J Koros Effect of Partially Immobilizing Sorption on Permeability and the Diffusion Time Lag Journal of Polymer Science Polymer Physics Ed 1976 14 675 21 R J Pace A Datyner Statistical Mechanical Model for Diffusion of Simiple Penetrants in Polymers Theory Journal of Polymer Science Polymer Physics Ed 1979 17 437 22 R E Kesting A K Fritzsche Polymeric Gas Separation Membranes John Wiley and Sons Inc New York United States 1993 23 C M Zimmerman A Singh W J Koros Diffusion in Gas Separation Membrane Materials A Comparison and Analysis of Experimental Characterization Techniques Journal of Polymer Science Part B Polymer Physics 1998 36 1747 24 B D Freeman Basis of Permeability Selectivity Tradeoff Relations in Polymeric Gas Separation Membranes Macromolecules 1999 32 375 25 J H Petropoulos Plasticization Effects on the Gas Permeability and Permselectivity of Polymer Membranes Journal of Membrane Science 1992 75 47 84 26 C K Yeom S H Lee J M Lee Study of Transport of Pure and Mixed 27 28 29 30 3 32 33 34 35 36 37 CO N Gases Through Polymeric Membranes Journal of Applied Polymer Science 2000 78 179 X G Li l Kresse Z K Xu J Springer Effect of Temperature and Pressure on Gas Transport in Ethyl Cellulose Membrane Polymer 2001 42 6801 E D Sanders S M Jordan R Subra
27. 97 130 41 47 J Y Ying J B Benzinger Structure Tailoring of Alkoxide Silica Journal of Non Crystalline Solids 1992 147 222 48 J A Cella Degradation and Stability of Polyimides Polymer Degradation and Stability 1992 36 99 49 T Ozawa T Arii A Kishi Thermogravimetry and Evolved Gas Analysis of Polyimide Thermochimica Acta 2000 352 177 50 E Jakab F Till T Szekely S S Kozhabekov B A Zhubanov Thermal Decomposition of Aryl Alicyclic Polyimides Studied by 86 Thermogravimetry Mass Spectrometry and Pyrolysis Gas Chromatography Mass Spectrometry Journal of Analytical and Applied Pyrolysis 1992 23 229 51 B Crossland G J Knight W W Wright Thermal Degradation of Some Polyimides British Polymer Journal 1987 19 291 52 M J Turk A S Ansari W B Alston G S Gahn A A Frimer D A Scheiman Evaluation of the Thermal Oxidative Stability of Polyimides via TGA Techniques Journal of Polymer Science Part A Polymer Chemistry 1999 37 3943 53 C W Jones W J Koros Carbon Molecular Sieve Gas Separation Membranes l Preparation and Characterization Based on Polyimide Precursors Carbon 1994 32 1419 54 V C Geiszler W J Koros Effects of Polyimide Pyrolysis Conditions on Carbon Molecular Sieve Membrane Properties Industrial amp Engineering Chemistry Research 1996 35 2999 55 1 Howe D H Williams R D Bowen Mass Spectrometry Principles and Applications 2 ed McGraw
28. A 25 polyimide based membranes Pure polymer refers to pure 6FDA 6FpDA DABA 25 polyimide Panneai is the measurement of an annealed membranes Pois the measurement of an unannealed membrane 67 Figure 3 20 Normalized permeability as a function of molecule size for the 6FDA 6FpDA DABA 12 polyimide based membranes Pure polymer refers to pure 6FDA 6FpDA DABA 12 polyimide Pana is the measurement of an annealed membranes Pois the measurement of an unannealed membrane 67 Figure 3 21 Normalized ideal selectivity as a function of molecule size difference for the 6FDA 6FpDA DABA 25 polyimide based membranes 68 Figure 3 22 Normalized ideal selectivity as a function of molecule size difference for the 6FDA 6FpDA DABA 12 polyimide based membranes P1 P2 anneai is the measurement for annealed membranes P P5 is the measurement for unannealed membranes Subscripts 1 and 2 refer to different gases The molecule size difference is measured as the difference in kinetic diameters for the selected gas pair The dashed line represents no change with annealing 68 Figure 4 1 Chemical structure of GFDA 6FpDA and 6FDA 6FpDA DABA eis uec 72 Figure 4 2 Plot of pure CH4 permeability as a function of feed pressure for various polyimide and hybrid systems at 35 C The permeabilities were calculated using the appropriate feed pressure Error 295 73 Figure 4 3 Plot of pure CO permeability as a function
29. A gel is a substance that contains a continuous solid structure Therefore the definition of sol gel chemistry involves the conversion of a sol into gel network The gel network usually involves metal or silicon based oxides and may contain organic groups Second the term organic inorganic must be defined In the literature organic inorganic may refer to any number of different materials including CERAMERS ceramic polymers ORMOSILS organically modified silicates and ORMOCERS organically modified ceramics In sol gel chemistry the inorganic part is a silica or metalloid and the organic part is typically an alkyl ligand attached to the silica or metalloid To clarify these confusing terms will refer to the sol gel network as an organosilicate Therefore the membranes in this work will be called polyimide organosilicate hybrid materials since they are comprised of organosilicates covalently bonded to a polyimide matrix Sol gel chemistry is very complex so this review will be limited to the information needed to explain this project In particular this work utilized silica based alkoxide precursors hence this discussion will not consider other metalloid based alkoxide precursors Important variables that affect the chemistry and resulting gel network include pH stoichiometry solvent temperature and pressure In general acid catalyzed systems generate a more linear network than base catalyzed systems 18 Gel
30. C The values were calculated using the appropriate fugacity 79 4 3 Conclusions The permeability was plotted as a function of methane and carbon dioxide feed pressure for two fluorinated polyimides and two polyimide organosilicate hybrid membranes For methane the permeability coefficient decreased with increasing feed pressure which is consistent with dual mode sorption theory For carbon dioxide all membranes exhibited dual mode sorption behavior up to 17 atm at which point the onset of plasticization becomes evident Incidentally all membranes exhibited plasticization behavior at a similar feed pressures which may be due to CO interacting with the 6FDA 6FpDA segments 80 Chapter 5 Recommendations 5 1 Future Work To my knowledge this is the first study that attempted to improve gas separation performance by annealing hybrid materials Obviously the effects of annealing on these systems is complicated and not completely understood Further examination is required to determine the exact mechanisms of gas transport through these annealed hybrid membranes In particular there is a need for a gas transport model which could describe the diffusion of gas through heterogeneous systems and relate this to the Time Lag Method Further studies could examine the effects of heating and cooling rates on these systems Presumably the polyimide in this study will experience time dependent relaxations Additionally since two mate
31. DA repeat unit is highlighted in green 28 2 2 Annealing Procedure A Lindberg Blue box furnace with a temperature range of 25 1100 C was used to anneal the membranes individually under air The built in programming was set to 400 C and allowed to preheat The sample was placed in the oven on a glass plate and annealed for 30 minutes The sample was removed an immediately place between two Teflon sheets to cool the sample as quickly as possible All samples received the same heating and cooling protocol 2 3 Gas Permeation General Information Gas permeation data was collected for all materials using the Time Lag Method This method utilizes the increase of flux or pressure as a function of time to determine the quantities for permeability solubility and diffusivity Each test was started after the sample was degassed to a pressure of 1 to 10 mTorr and the system reached thermal stability at 35 C This study focused entirely on the pure gas separation of He O2 N2 CH and CO all at 99 999 purity as received from the supplier The feed pressure of these gases was 4 atm and the temperature was 35 C for all trials Each membrane was tested three times for each gas and the average results were recorded to ensure reproducibility The total error for permeability is 596 based upon testing in a different gas permeation set up 29 2 4 Gas Permeation Equipment A schematic diagram of the gas permeation system is shown in Figure
32. GC 6000 oven to control the temperature will be referred to as System CJC named after Chris J Cornelius who constructed the system The other system will be referred to as System CLH after Chris L Hibshman who constructed that system Each 89 system has its own computer and both systems have similar LABTECH NOTEBOOK software programs There are several differences between the two systems First System CJC can be operated at higher temperatures at least up to 125 C whereas System CLH is limited to 50 C In addition System CJC has five feed valves one for each of five different gases This will allow for up to five different gases to be run consecutively by the batchrun process which is described later System CLH has only one feed valve which means that the feed gas must be changed prior to each testing sequence Finally the systems have different input output I O boards that interface with the computer This will be evident in the relevant programming of each system A 3 Computer Architecture There are two main executable exe files in the NOTEBOOK software package bld nb exe and batchrun exe that are used for programming bld nb exe is where the programs are visually constructed batchrun exe is used to join several bld nb exe programs together to run consecutively In order to run the batchrun exe program each bld nb exe program must be located in an individual folder Therefore each computer has a direc
33. Hill Inc 1981 56 R M Silverstein G C Bassler T C Morrill Spectrometric Identification of Organic Compounds 4 ed John Wiley amp Sons Inc 1981 57 J Gallardo P Galliano A Duran Thermal Evolution of Hybrid Sol Gel Silica Coatings A Structural Analysis Journal of Sol Gel Science and Technology 2000 19 393 58 A Venkateswara Rao G M Pajonk Effect of Methyltrimethoxysilane as a co precursor on the optical properties of silica aerogels Journal of Non Crystalline Solids 2001 285 202 59 G Zhang Y Chen H Li Y Xie Preparation of Silica Based Inorganic Organic Hybrid Membranes via the Sol Gel Route Journal of Sol Gel Science and Technology 2000 19 425 60 P D Maniar A Navrotsky E M Rabinovich J Y Ying J B Benzinger Energetics and Structure of Sol Gel Silicas Journal of Non Crystalline Solids 1990 124 101 61 K A Mauritz J T Payne Perfluorosulfonate ionomer Silicate Hybrid Membranes via Base Catalyzed in situ Sol Gel Processes for Tetraethylorthosilicate Journal of Membrane Science 2000 168 39 87 62 Andre P Legrand The Surface Properties of Silicas John Wiley and Sons New York 1998 63 Y Yan Y Hashino Z Duan S R Chaudhuri A Sarkar Design and Characterization of Interconnected Microporous Hybrid Thin Films by a Sol Gel Process Chemical Materials 1997 9 2583 64 G Qian Z Yang C Yung Matrix effects and mechanisms of the spectral shifts of coum
34. Polyimide Organosilicate Hybrid Materials Part Effects of Annealing on Gas Transport Properties Part Il Effects of CO Plasticization Christopher L Hibshman Department of Chemical Engineering Virginia Polytechnic Institute and State University Blacksburg VA 24060 021 1 A Thesis submitted to the faculty of Virginia Polytechnic Institute and State University in partial fulfillment of the requirements for the degree of Master of Science In Chemical Engineering APPROVED Dr Eva Marand Chair Dr Richey Davis Dr Ravi Saraf May 3 2002 Blacksburg VA 24060 Keywords Gas Separations Inorganic Membranes Composite Membranes Organosilicate Polyimide O 2002 Christopher Hibshman Abstract The objective of this study was to examine the effects of annealing polyimide organosilicate hybrid membranes on gas transport In addition the effects of carbon dioxide pressure on the gas transport of unannealed polyimide organosilicate hybrid membranes were evaluated The membranes in both studies consisted of sol gel derived organosilicate domains covalently bonded to a 6FDA 6FpDA DABA polyimide using partially hydrolyzed tetramethoxysilane TMOS methyltrimethoxysilane MTMOS or phenyltrimethoxysilane PTMOS The first study subjected the hybrid membranes to a 400 C annealing process to enhance gas separation performance by altering the organosilicate structures The hybrid membranes were evaluated before and after annealin
35. TIR ATR spectra of GFDA 6FpDA DABA 25 22 5 TMOS based hybrid Shading highlights areas of change after the annealing process Si Bonding Benzene Ring CF Si O Before Anneal r 3 C O C N p Polyimide LN a UN After Anneal 2000 1800 1600 1400 1200 1000 800 600 400 Wave Number cm Figure 3 6 FTIR ATR spectra of the 6FDA 6FpDA DABA 25 22 5 PTMOS Shading highlights areas of change after the annealing process 45 In conjunction with the TGA MS and FTIR ATR results we postulate that the anhydride end groups thermally hydrolytically degrade in the pure polyimide as reported by Cella and illustrated in Figure 3 7 This mechanism occurs in the presence of water and results in the loss of carbon dioxide The presence of water and carbon dioxide at 400 C was confirmed by TGA MS data Water can come from several sources namely the condensation of the carboxylic acid groups in the DABA and the condensation reactions of the sol gel network in the hybrid materials Cella also pointed out that at elevated temperatures polyimides undergo rapid hydrolysis and these hydrolysis products suffer thermal degradation which leads to chain scission weight loss and crosslinking O CO H COH O gt A CO H O Figure 3 7 Thermal hydrolytic degradation of anhydrides The degradation mechanism shown in Figure 3 7 is also supported by the FTIR ATR spectra The reduction of intensity for peaks assigned to the ca
36. TMOS xoebpbbmx xoetb5pbbm He O Selectivity a 4 70 80 90 100 200 300 400 500 Helium Permeability barrers Figure 3 14 Boundary diagrams of 6FDA 6FpDA DABA 12 pure polyimide and hybrid materials for He O The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed membranes 56 100 90 80 70 60 50 40 30 He N Selectivity o 20 70 80 90 100 200 OOD RP EHO ORR KP E Upper Bound Pure Polymer 22 5wt MTMOS 15 0wt MTMOS 7 5wt MTMOS 22 5wt PTMOS 15 0wt PTMOS 22 5wt TMOS Pure Polymer 22 5wt MTMOS 15 0wt MTMOS 7 5wt MTMOS 22 5wt PTMOS 15 0wt PTMOS 22 5wt TMOS 300 400 500 Helium Permeability barrers Figure 3 15 Boundary diagrams of 6FDA 6FpDA DABA 12 pure polyimide and hybrid materials for He N The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed membranes 200 Upper Bound m Pure Polymer A 22 5wt MTMOS 15 0wt MTMOS A 7 5wt MTMOS 22 5wt PTMOS S 100 15 0wt PTMOS gt 22 5wt TMOS 5 90 m Pure Polymer 3 80 A 22 5wt MTMOS 70 15 0wt MTMOS 9 4 7 5wt MTMOS 5 60 22 5wt PTMOS 15 0wt PTMOS 50 22 5wt TMOS 40 30 60 70 80 90100 200 300 400 500 Helium Permeability barrers Figure 3 16 Boundary diagrams of 6FDA 6FpDA DABA 12 pure polyimide and hybrid mate
37. To follow along open an existing batchrun exe from one of the gas tests listed in Table A 1 by double clicking on the filename A 9 Main Menu As is common with most Windows based software there is a menu at the top of the page The File button gives the user the option of exiting the batchrun exe program or saving the current settings for later use A 10 Buttons A Clicking on the Start button will start the batchrun program A new window will appear which indicates that the first step in the Files to Run list has started B Clicking on the Stop button will stop the batchrun program but will not stop the bld nb exe program that is currently running The batchrun program will stop when current bld nb exe program is finished C If the Run on Startup box is checked the batchrun program will automatically start when opened In other words the Start button will be automatically clicked for the user when this screen is opened recommend this not be clicked unless running a series of gases 103 Number of Files to Run corresponds to the number of steps in the Files to Run list Number of Iterations determines how many times the Files to Run list repeats itself The white box at the bottom of the screen displays the current status of the batchrun program Typical displays are running stopping and completed Files to Run is always run nb exe for these programs The Working Directory corresponds to w
38. aling 68 3 11 Conclusions TGA MS and FTIR ATR analysis of the hybrid membranes showed that the annealing process is driving the condensation reactions to a greater extent creating a denser SiO network Neither the methyl nor phenyl ligands were decomposed from their respective hybrid samples during the 400 C annealing process It was shown that the phenyl ligands decomposed from the PTMOS hybrid at 520 C which is well above the annealing temperature Degradation of the pure polyimide was attributed to two primary mechanisms the condensation of the carboxylic groups along the backbone and thermal hydrolytic degradation of the anhydride endgroups In general the annealing process increased permeation by about 200 to 500 and decreased the ideal permselectivity anywhere from 0 to 50 The increase in permeability for these membranes was surprising considering that they became more crosslinked with the annealing process For the hybrid materials the changes in permselectivities after the annealing process vary with the type of alkoxide in the silica network The permselectivity of the PTMOS hybrid membranes decreased significantly because the bulky phenyl group prevented effective crosslinking with the polyimide Both the CO CH4 permselectivity and permeability increased for the 6FDA 6FpDA DABA 25 22 5 wt TMOS and MTMOS based hybrid materials The increases in permselectivity were attributed to effective crosslinking between the polyimid
39. amino benzoic acid DABA in the pure polymer greatly influenced the gas separation properties First the condensation of the carboxylic groups during the annealing process is a source of water which promotes the degradation of the anhydride end groups Because the DABA 25 systems had double the concentration of carboxylic acid groups than the DABA 12 these samples were more likely to condense during the annealing process thus increasing cross linking and producing more water to promote the degradation mechanism Second before annealing the polyimide chains in the DABA 25 membranes were more likely to hydrogen bond along the polymer chain backbone due to higher concentrations of carboxylic acid groups This was evident in the permeation data of the unannealed pure polyimide membranes as DABA 12 membranes had higher permeabilities and lower permselectivities than the DABA 25 membranes During the annealing process these hydrogen bonds were easily broken and the carboxylic groups condensed to form cross links along the polymer chain backbone After annealing the permeabilities were nearly the same for both DABA systems although the permselectivities were slightly higher in the DABA 25 membranes which may be due to the higher degree of cross linking in the DABA 25 Therefore the annealing process had a greater influence on the membrane with the higher DABA content 59 3 9 Gas Transport Hybrid Materials In a previous paper Cornelius
40. arable feed pressures considering the diverse chemical structures of the materials Apparently all the materials are affected by CO at some critical concentration in this case a feed pressure of 17 atm Similarly Staudt Bickel and Koros reported the onset of plasticization at 17 atm for a non crosslinked 6FDA mPD polyimide This observation leads us to hypothesize that CO is interacting with the 6FDA 6FpDA segments which is common to all the materials causing CO plasticization at a common feed pressure Since the hybrid materials in this study are crosslinked only at the polyimide organosilicate interface the bulk of the polymer matrix remains susceptible to plasticization Another interesting observation is the pressure dependence of ideal selectivity is linear for the MTMOS based hybrid membrane but nonlinear for the polyimide membranes In fact the polyimide membranes appear to be linear up to 14 atm at which point the permeability begins to increase dramatically The MTMOS based hybrid membrane maintains a linear relationship between ideal selectivity and pressure even at high pressures The PTMOS based hybrid membrane exhibits some unusual and unexpected behavior with higher selectivities at both low and high pressures but is linear between these points Plotting ideal selectivity as a function of pressure effectively cancels out the dual mode sorption effects and isolates the plasticization phenomenon In other 76 words th
41. arin 440 doped in sol gel derived gel glass Journal of Applied Physics 2000 88 2503 65 M H Cohen D Turnbull Molecular Transport in Liquids and Glasses Journal of Chemical Physics 1959 31 1164 66 H H Hoehn Heat Treatment of Membranes of Selected Polyimides Polyesters and Polyamides U S Patent 3 822 202 1974 67 J S McHattie W J Koros D R Paul Gas transport properties of polysulfones 1 Role of symmetry of methyl group placement of bisphenol rings Polymer 1991 32 840 68 M Moe W J Koros H H Hoehn G R Husk Effects of Film History on Gas Transport in a Fluorinated Aromatic Polyimide Journal of Applied Polymer Science 1988 36 1833 69 R Wang C Cao T S Chung A Critical Review on Diffusivity and the Characterization of Diffusivity of 6FDA 6FpDA Polyimide Membranes for Gas Separation Journal of Membrane Science 2002 198 259 88 Appendix LABTECH NOTEBOOK Software A 1 General Information LABTECH NOTEBOOK Version 7 3 0 software was employed to assist with the data acquisition from the two gas permeation systems The master diskettes and manuals for NOTEBOOK can be found in the glass cabinet above the IR equipment The purpose of this appendix is to assist graduate students in understanding the NOTEBOOK programs we developed to acquire data A 2 Systems We currently have two separate yet very similar gas permeation systems in our laboratory The gas permeation setup using the
42. ary pore size was created by the removal of the ligands It was also 20 noted that above 25 vol MPS a partial collapse of the silica network occurred Similarly Kim et al pyrolyzed methacryloxypropyl ligands from a MPS TEOS system at 400 500 600 and 700 C TGA data showed a significant weight loss at 320 C which the authors attributed to the pyrolysis of the methacryloxypropyl ligand Kusakabe et al pyrolyzed octane dodecane and octadecane ligands from the respective alkyl TEOS TEOS systems at 600 C From their TGA results they concluded that organic weight loss started at 250 C and continued until 600 C Sea et al pyrolyzed the phenyl ligands from phenyltriethoxysilane PTEOS and diphenyltriethoxysilane DPTEOS at 6009C 9 They concluded that the phenyl group was removed between 400 700 C In summary these results suggest that pore size in the silica network can be controlled by selective pyrolysis of organic templates This can lead to an effective control of transport properties 1 9 Polymer Organosilicate Hybrid Materials Much effort has been invested into developing new polymeric materials for gas separations Polymers are relatively inexpensive and easier to fabricate than inorganic membranes however their gas transport properties are limited particularly at elevated temperatures or under environmentally harsh conditions For example polymers can plasticize or degrade in the presence of certain organic compounds
43. at m z 20 at the same temperature range that was about four times larger than the peak at m z 69 This peak was attributed to the presence of HF This data is consistent with Turk et al who used TGA coupled with FTIR to analyze the decomposition gases For fluorinated polyimides CF3 and HF were present in the decomposition gas from 470 C to 580 C 42 3 4 FTIR ATR In addition to TGA MS analysis FTIR ATR was employed to correlate the changes in the chemical bonding structure during the annealing process Although the FTIR ATR technique probes approximately the first 2 microns of the film we assume that this is representative of the bulk of the sample The average thickness of the membranes is 76 microns The apparent changes in the hybrid systems upon annealing were compared to those of the pure polyimide as shown in Figures 3 3 through 3 6 Figure 3 3 highlights the subtle changes in the bonding structure for the pure polyimide upon annealing The peaks associated with carbonyl groups at 1720 cm and 1395 cm become weaker with annealing An aromatic peak appears at 1500 cm after annealing evidence of an aromatic ring structure Some peaks in the 680 800 cm range assigned to the aromatic rings in the polyimide also change slightly For the hybrid membranes shown in Figures 3 4 through 3 6 anti symmetric Si O Si stretching encompasses the range from 960 1280 cm including strong absorption bands of amorphous SiO at 1
44. ation corresponds well with the permeability data presented in Figure 4 3 These results are typical for the permeation of carbon dioxide through glassy polymers 77 Fugacity psia 0 5 9 9 14 7 19 4 24 4 28 6 6FDA 6FpDA 6FDA 6FpDA DABA MTMOS as PTMOS Diffusivity o o 0 3 0 0 0 5 10 15 20 25 30 CH Pressure atm Figure 4 5 Plot of pure CH diffusivity as a function of feed pressure for various polyimide and hybrid systems at 35 C The values were calculated using the appropriate feed pressure Fugacity psia 0 5 9 9 14 7 19 4 24 1 28 6 6FDA 6FpDA 6FDA 6FpDA DABA MTMOS PTMOS Diffusivity CO Pressure atm Figure 4 6 Plot of pure CO diffusivity as a function of feed pressure for various polyimide and hybrid systems at 35 C The values were calculated using the appropriate fugacity 78 Fugacity psia 0 5 9 9 14 7 19 4 24 1 28 6 6FDA 6FpDA 6FDA 6FpDA DABA MTMOS PTMOS Solubility CH Pressure atm Figure 4 7 Plot of pure CH solubility as a function of feed pressure for various polyimide and hybrid systems at 35 C The values were calculated using the appropriate feed pressure Fugacity psia 0 5 9 9 14 7 19 4 24 1 28 6 6FDA 6FpDA 6FDA 6FpDA DABA MTMOS PTMOS Solubility CO Pressure atm Figure 4 8 Plot of pure CO solubility as a function of feed pressure for various polyimide and hybrid systems at 35
45. ation is an irreversible process involving several interrelated reactions Three chemical equations are identified in this complex process OR acid OR catalysis 1 RO Si OR H20 RO Si OH ROH OR OR OH OH OH OH 2 Ho i on Ho Si oH HO SI 0 Si oH HO OH OH OH OH OR OH OR OH 3 RO Si OR HO Si OH RO Si O 0H ROH OR OH OR OH where R is an alkyl group CHs C2Hs etc Equation 1 is a hydrolysis reaction and equations 2 and 3 are condensation reactions All three reactions occur simultaneously The time of gelation is dependent upon pH stoichiometry solvent temperature pressure and evaporation rate It usually takes many hours to dry After the porous dry gel is obtained it is sintered at a very high temperature 600 1400 C to produce a fully densified and homogenous glass The dry gel non sintered material is very brittle and fragile but selectively porous The size of these pores is dependent upon the many variables mentioned earlier As a result of controlling pore size favorable gas permeation qualities can be obtained for desire gas separations In general the gas transport properties are superior to polymeric membranes However it can be difficult to produce defect free inorganic membranes 19 To control the pore size distribution in the gel network several research groups have created a template process 9 This process incorporates organic ligands
46. based upon its sound fundamental basis and ease of data collection believe that the values of permeability are accurate for these hybrid membranes although the diffusivity and solubility coefficients are not exact in absolute terms This assumption stems from a study done by Paul and Kemp who concluded that increasing filler content will cause very large increases in the time lag but will exhibit only minor effects on the steady state permeability for a rubbery polymer containing adsorptive filler materials In a separate study Zimmerman et al compared the diffusion coefficients determined by various experimental techniques and concluded that the Time Lag Method did not accurately represent the true diffusion coefficient However the ratios giving diffusivity selectivity were comparable with those measured by other methods believe that the solubility and diffusion coefficients from permeability measurements are useful for making qualitative comparisons between the materials before and after the annealing process To compare the performance of these materials the permeability and selectivity will be plotted on a boundary diagram as founded by L M Robeson In 1991 Robeson conducted an extensive literature search and plotted the tradeoff between ideal gas selectivity and gas permeability of polymers for 13 several gas pairs He H O2 No CH4 CO2 He concluded there was an empirical upper bound to the trade off and that m
47. dation Intensity units are arbitrary 3 0E 10 4 2 5E 10 4 2 0E 10 4 Intensity 1 5E 10 4 1 0E 10 4 5 1E 11 4 Mass to charge ratio m z 50 k 51 9 52 m 78 y e l l U u 500 600 700 Temperature C Figure 3 2 TGA MS spectra for GFDA 6FpDA DABA 25 22 5wt PTMOS hybrid material Intensity units are arbitrary 41 TGA results were inconclusive in determining if the methyl ligand was being decomposed from the MTMOS sample Several sources reported that the methyl ligands are stable up to 450 C or 500 C for similar materials 75 The PTMOS hybrid sample was the only material to show peaks at m z 78 and m z 50 at 520 C which correspond to the phenyl ligand CeHe m w 78 and is shown in Figure 3 2 The peak at m z 50 was due to the opening of the aromatic ring and loss of CH m w 28 In addition there were similar peaks at m z 51 and 52 that were due to aromatic ring openings Zhang et al have shown the phenyl ligand can be removed from a PTMOS derived gel at 520 580 C which is consistent with our observations Thus we conclude that the phenyl ligands were not removed during the 400 C annealing process Interestingly a distinct peak at m z 69 was observed for all samples between 470 and 570 C which was well above the annealing temperature We attribute this peak to the removal of the CF3 ion In addition there was a distinct peak
48. e CO permeability increases with pressure more than the CH permeability for the polyimide membranes but not the MTMOS based hybrid membrane Therefore it appears that the CO plasticization has more influence on the pure polyimides than the MTMOS based hybrid To be certain mixed gas permeation studies need to be performed to evaluate changes in selectivity with increasing feed pressure For example polymers that are plasticized by CO2 during a CO2 CH mixed gas experiment will experience a decrease in selectivity with increasing feed pressure because the polymer matrix will swell with increasing CO concentration thereby permitting accelerated diffusion of both CO and CH gases Figures 4 5 through 4 8 present the diffusion and solubility coefficients as a function of feed pressure These values were determined using the Time Lag Method and are presented for qualitative analysis only Again remind the reader that using the Time Lag Method in the analysis of glassy polymers does not always result in accurate absolute values for D and S As illustrated in Figures 4 5 and 4 6 the diffusivity of all samples increases linearly with increasing feed pressure for both gases This dependence is greater for CO than CH4 On the other hand the solubility coefficient decreases with increasing feed pressure for both gases For COs the solubility decreases exponentially which will result in higher permeabilities at higher feed pressures This observ
49. e and organosilicate materials The increases in permeability were attributed to relaxations of the polymer chains during the 69 annealing process which increased localized segmental mobility and therefore the flux of gas penetrant molecules 70 Chapter 4 Effects of Feed Pressure on Gas Transport 4 1 Abstract The transport properties of fluorinated polyimides and fluorinated polyimide organosilicate hybrid membranes were characterized as a function of feed pressure Steady state gas permeation experiments were performed at 35 C using pure CO and CH gases at feed pressures ranging from 4 to 30 atm The non crosslinked polyimide membranes consisted of a 6FDA 6FpDA polyimide and a 6FDA 6FpDA DABA polyimide The hybrid membranes consisted of phenyltrimethoxysilane PTMOS and methyltrimethoxysilane MTMOS derived organosilicates covalently bonded to the GFDA 6FpDA DABA polyimide All four materials exhibited dual mode sorption up to feed pressures of 17 atm at which point the effects of CO plasticization were observed to varying degrees 71 4 2 Results and Discussion The chemical structures of the two polyimides used in this study are illustrated in Figure 4 1 The permeabilities of pure CH4 and CO gases were measured at a temperature of 35 C and are plotted as a function of feed pressure in Figures 4 2 and 4 3 Gas permeation measurements of the non crosslinked 6FDA 6FpDA polyimide corresponded well with data publis
50. e hybrid membranes The permeability diffusivity and solubility coefficients for the materials before and after annealing are reported in Tables 3 3 and 4 4 The permeability for all gases through all membranes increased dramatically after the annealing process These increases were mostly attributed to large increases in the diffusion coefficient The increases in solubility coefficients were relatively small compared to increases in the diffusion coefficients The permeabilities and ideal permselectivities are plotted in Figures 3 9 through 3 18 and are compared to Robeson s 1991 upper bound The applicability of the Time Lag Method originally developed for homogeneous systems needs to be discussed To validate its use for these hybrid systems we point to a study by Paul and Kemp who concluded that using the Time Lag Method to measure the diffusivity of adsorptive but nonpermeable fillers in a rubbery polymer results in large increases in the diffusion time lag but only minor effects on the steady state permeability 9 It is unclear whether these large increases in diffusion time lag should be attributed to the gas penetrants taking a more tortuous path In our hybrid membranes we think that the filler materials organosilicate domains are permeable unlike the study by Paul and Kemp In fact the diffusion coefficient for the pure polyimide is similar to that of the hybrid materials suggesting that the organosilicates and polyimide have simi
51. e is observed to decrease during plasticization CO plasticization can usually be identified by plotting the CO2 permeability as a function of feed pressure for a given membrane As illustrated in Figure 1 2 at low pressures an initial decrease in gas permeability with increasing feed pressure was observed which would be consistent with dual mode transport theory At higher pressures if the CO permeability increases dramatically with increasing feed pressure then the polymer is considered plasticized The minimum pressure on the curve is considered the plasticization pressure Wessling et al noted another interesting phenomenon of CO plasticization namely that the permeability of a plasticized glassy polymer is not constant but increases with time However eventually the system will reach an equilibrium state Below the plasticization pressure the penetrant molecules loosened the short chain segments but above the plasticization pressure long chain rearrangements were enabled by the plasticizing agent loosening more dense entanglements Typically the presence of plasticization during gas transport will decrease the selectivity of the gas separation Therefore to optimize gas separation performance in the presence of a plasticizing agent it is imperative to minimize the degree of plasticization in the membranes To accomplish this task typically requires modification of the polymer membrane For example Krol et al thermally
52. e symbols represent unannealed membranes and red symbols represent annealed membranes esses 54 Figure 3 10 Boundary diagrams of 6FDA 6FpDA DABA 25 pure polyimide and hybrid materials for He N The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed mermbraries ccce ee atas badge ahaha ake ce SANE le annan nnmnnn nnna 54 Figure 3 11 Boundary diagrams of 6FDA 6FpDA DABA 25 pure polyimide and hybrid materials for He CH The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed membranes ceeceeeecccceeeeeeeeeeeeeeeeeeeeeeeeeeeeeeennaaeeeeeeees 55 viii Figure 3 12 Boundary diagrams of 6FDA 6FpDA DABA 25 pure polyimide and hybrid materials for Oz Nz The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent aririeatled miembFPalles 3a cccoc qe dea Centri oa edem ri dea depu eod eo n keen ene ep keen 55 Figure 3 13 Boundary diagrams of 6FDA 6FpDA DABA 25 pure polyimide and hybrid materials for CO2 CH 4 The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed membranes eeeeesssssssssssseeseeee nennen 56 Figure 3 14 Boundary diagrams of 6FDA 6FpDA DABA 12 pure polyimide and hybrid materials f
53. ed additional free volume thereby increasing the permeability This would be consistent with the Cohen Turnbull model which suggests that diffusion coefficients should increase exponentially with fractional free volume 9 While the free volume in the pure polyimide may be created by the degradation of the anhydride end groups as mentioned earlier the increase of free volume in 61 the hybrid samples could be attributed to the removal of the sol gel condensation products On the other hand if the free volume of the hybrid materials is indeed changing with the annealing process we hoped it would be evident in the density measurements The density of the hybrid materials presented in Table 5 was measured to the nearest 0 001 g cm before and after exposure to the 400 C annealing process Based on the TGA results one would expect a 2 3 percent decrease in density assuming the volume of the sample did not change However the results indicate no significant changes in density While in the pure polyimide the weight loss was attributed to the loss of water and carbon dioxide in the hybrid materials the majority of weight loss was attributed to water expelled during the condensation of the organosilicate network These observations suggest both the mass and volume of the samples decreased to maintain constant densities Although it was difficult to accurately measure the absolute volume of the samples we did observe that the samples shrunk in
54. employing membranes include natural gas sweetening oxygen enrichment and hydrogen recovery from ammonia purge gases Typically smaller gas molecules will diffuse through a polymer membrane selectively leaving the larger molecules behind Polymeric membranes are favored over inorganic membranes due to relative ease and low cost of processing In addition polymers can be spun into hollow fibers which maximize the surface area to volume ratio On the down side polymers exhibit a trade off between permeability and selectivity They also loose their performance at high temperatures high pressures and harsh chemical environments In constrast inorganic membranes such as zeolites or molecular sieves have excellent gas separation properties and durability However manufacturing these materials into high surface area membranes is very difficult and expensive Recently several groups have combined inorganic domains into a polymer matrix to form hybrid materials in attempts to combine the excellent gas separation properties of the inorganic materials with the processing properties of the polymeric materials C J Cornelius et al recently synthesized and characterized a series of fluorinated polyimide organosilicate hybrid membranes This project uses those same materials but focuses on annealing the hybrid materials at 400 C to enhance the gas transport properties of the organosilicate domains which need to be calcined at very high temp
55. eratures to maximize their performance In addition the effects of CO plasticization on these hybrid materials are unknown and need to be investigated 1 2 Gas Transport through Polymeric Membranes The Solution Diffusion Model is the principal model for describing gas diffusion through polymeric materials The Time Lag Method can be used to calculate the permeability diffusivity and solubility coefficients through rubbery polymers For glassy polymers describing the diffusion of gas is more complex because the molecule chains are not at a state of equilibrium Some common models for diffusion in glassy polymers include Dual Mode Sorption Theory and Free Volume Theory All of these models will be discussed in this review 1 3 Solution Diffusion Model The Solution Diffusion Model is a widely accepted concept used to calculate the flux through a membrane This model recognizes three stages to the diffusion of gas molecules First the penetrant molecule dissolves in the polymer on the upstream side then it diffuses across the membrane via a concentration gradient and finally it desorbs from the downstream side Dialysis reverse osmosis pervaporation and gas permeation can all be explained by this model The Solution Diffusion Model is expressed in terms of the chemical potential gradient which can be thermodynamically related to temperature pressure and concentration For the case of gas permeation the chemical potential is
56. es of the alkoxide precursors which influenced the degree of crosslinking between the polyimide and organosilicate domains This study focused on the silica contents of 15 0 and 22 5 weight percent sol in the original polyimide solution The organosilicate networks were bonded to the matrix via APTEOS as shown in Figure 2 2 The films were cast using THF as the common solvent as dried for 4 days Afterwards they were dried at 220 C for 12 hours under vacuum to remove residual water and solvent This process produced optically transparent membranes 2 4 mils thick The end product consisted of organosilicate domains covalently bonded to a polyimide matrix creating a crosslinked network After the drying process these films contained approximately 7 5 wt and 11 0 wt silica respectively The pure polyimide with the corresponding DABA content was used as a control in the experiments OCH CH 0 Si OCH OCH Tetramethoxysilane hee Methyltrimethoxysilane CH3 Si OCH OCH OCH Phenyltrimethoxysilane O Si OCH OCH Figure 2 1 Chemical structures of TMOS MTMOS and PTMOS 27 Oo CE Oo O CE Q EO Si N NCA SIBI N Pi N HO CF CR OH 0 0 n o o HN O SXEtO s Oo Oo CF3 en Pi OO TOT SOC CF3 CF3 Oo Oo Figure 2 2 Chemical structure of the 6FDA 6FpDA DABA polyimide functionalized with APTEOS blue The DABA group is highlighted in red The 6FDA 6Fp
57. evice Channel Permeate Cell Pressure 0 Transducer Feed Pressure Transducer 2 Cell Temperature RTD 4 Ambient Temperature RTD 5 Table A 6 Interface Channel and Bit Numbers for each measurement device of System CJC The Interface Device is 1 DAS 8PGA Device Channel Permeate Pressure 0 Feed Cell Pressure 1 Vacuum Pressure analog 3 C Thermocouple TC A thermocouple block is very similar to the analog input block but usually has a different input range This type of block is only used in System CJC System CLH used an analog input block to measure the temperature a Interface Device and Interface Channel correspond to the appropriate board name and channel number as shown in Table A 7 Table A 7 Interface Channel and Bit Numbers for each thermocouple of System CJC The interface device is 3 5508TC Device Channel Cell Temperature 81 Oven Temperature 83 99 b Temperature Scale describes the units of measurement c Input Range corresponds to the output voltage of the measurement device which can usually be found on the device or the operator s manual for the specific device d Thermocouple Type is set by the manufacture and is available in the operator s manual e Scale Factor and Offset Constant are used to calibrate the thermocouple Typically one would compare the measured value with a thermometer using a beaker of water at different temperatures Buffer Size Number of Stages Samplin
58. f Records to Close File should be equal to the buffer size in other blocks 101 f Number of Columns in File will determine how many columns are in the data file This number should equal the number of measurement devices in the program g The last 5 lines will format the columns with labels units width and decimal places G Display This block will format the display when performing a run This will not affect the data collection but is a nice visual to see during a run a The arrows on the upper right of the block if you do not see arrows in the upper right double click on the block will minimize and maximize the block To add or delete windows in the display you must first maximize the block To add a window click and drag in the black area Click exit to leave this mode To change the position or size of a window click on the window of choice to adjust b Go back to the previous Display block Click in the black area not a window and a new screen will appear This is how one adjusts the formatting of the windows such as scale and units C Go back to the previous Display block Click on a window not the black area and a new screen will appear This is used to customize the formatting or how the data is displayed 102 A 8 batchrun exe Program Batchrun Sections A 8 A 11 concern only the batchrun exe program The purpose of the batchrun program is to consecutively run programs without user interruption
59. fabrication facility This facility produced radiation hardened computer chips for space and military applications Chris was in need of a job during the summer of 1999 This is when he approached Dr Eva Marand and started working as an undergraduate research assistant His initial research project did not work as anticipated so instead he focused on annealing these hybrid materials in an attempt to improve the gas transport properties Eventually this project evolved from some unexpected observations to a Master s thesis For the future Chris has accepted a job at Eastern Research Group in Chantilly VA where he will use his chemical engineering skills as an environmental consultant 106
60. g using pure gases He O2 N2 CH4 COz at 35 C and a feed pressure of 4 atm The permeability for most of the membranes increased 200 500 after the annealing process while the permselectivity dropped anywhere from 0 to 50 The exceptions were the 6FDA 6FpDA DABA 25 22 5 wt TMOS and MTMOS hybrid membranes both of which exhibited increases in the CO permeability and CO2 CH permselectivity The increase in permeation was attributed to increases in the free volume and enhanced segmental mobility of the chain ends resulting from the removal of sol gel condensation and polymer degradation byproducts For the second study the transport properties of four membranes 6FDA 6FpDA polyimide 6FDA 6FpDA DABA polyimide MTMOS and PTMOS based hybrid materials were characterized as a function of feed pressure to evaluate how the hybrid materials reacted to CO plasticization Steady state gas permeation experiments were performed at 35 C using pure CO and CH gases at feed pressures ranging from 4 to 30 atm All four materials exhibited dual mode sorption up to feed pressures of 17 atm at which point the effects of CO plasticization were observed Format of Thesis The format of Chapters 3 and 4 is in the form of two separate technical papers Both chapters focus on characterizing the polyimide organosilicate hybrid materials synthesized by Dr Chris J Cornelius Chapter 3 corresponds to the effects of annealing a series of polyimide o
61. g Period and Stage Duration are the same values as the digital output block see Section A D Time This block measure the time of the run a Format allows one to determine how the units are displayed b Mode allows one to dictate how the measurements are taken c Buffer Size Number of Stages Sampling Period and Stage Duration are the same values as the digital output block see Section A E Calculated Off All Allows the programmer to choose from a list of many different operations 100 Operation is chosen as Off All to close all the valves at the end of the run Buffer Size Number of Stages Sampling Period and Stage Duration are the same values as the digital output block see Section A Analog Trigger Value is the same as the upper limit for the digital output blocks see Section A F File Determines the file name where the data can be written a Data File Name is the directory and given name of the data file Typically the file is listed as a data dat file If several runs are made consecutively the last digit is an ampersand which will prevent the previous file from being overwritten Storage Mode is the type of file We use ASCII Real to collect the data Header Lines allows on to add text to the top of the file such as the date of the run Data File Opening Closing Mode determines when the data file will open and when to close Number o
62. h block can have its own name and units as chosen by the programmer The first five lines are common for most blocks The following lines are used to dictate the programming of each block The list that follows is a description of how each block is used for our purposes This list certainly does not include every function of every type of block See the user s manual for more details A Digital Output Block ON OFF Each of these blocks corresponds to one of the solenoid valves in the permeation system The block name usually describes each block with its corresponding valve a Interface Device Interface Channel and Bit Number correspond to the appropriate board name channel number and bit number as shown in Tables A 2 and A 3 b Upper and Lower Limits set the values at which the valve will perform the appropriate operation The Output Polarity and Loop Type determine if the operation is to occur inside or outside the limits The values are fail close meaning they will automatically close unless instructed otherwise or if power pressure is lost to the system c The Input Block Number is automatically determined by connecting an arrow to a particular block For the permeation setup our upper and lower limits correspond to the vacuum or 95 permeate pressure so the input block number corresponds to the permeate pressure block For example the lower limit of the feed valve is about 0 200 cmHg which means the pressure must be
63. hed by Wang et al For both gases the 6FDA 6FpDA DABA polyimide membrane exhibited significantly lower permeabilities than the 6FDA 6FpDA polyimide due to hydrogen bonding interactions from the carboxylic groups on the DABA unit This phenomenon has been well documented in previous studies We believe these interactions were strong enough to decrease the chain mobility and therefore the permeability of the 6FDA 6FpDA DABA polyimide As shown in Figure 4 4 the ideal selectivity of the 6FDA 6FpDA DABA polyimide is higher than 6FDA 6FpDA polyimide which is consistent with this explanation o CF3 o CF N CR 9 Oo 6FDA 6FpDA 9 CF3 9 9 CF3 A CF3 SOHOG OFOH OOK O O i 0 0 COOH 7 1 6FDA 6FpDA DABA Figure 4 1 Chemical structure of 6FDA 6FpDA and 6FDA 6FpDA DABA polyimides 72 Fugacity psia 0 5 9 9 14 7 19 4 24 1 28 6 w 6FDA 6FpDA 1 8 4 6FDA 6FpDA DABA a MTMOS 1 6 e PTMOS CH Permeability barrers 5 v 5 o e o CH Pressure atm Figure 4 2 Plot of pure CH permeability as a function of feed pressure for various polyimide and hybrid systems at 35 C The permeabilities were calculated using the appropriate feed pressure Error 2 Fugacity psia 0 5 9 9 14 7 19 4 24 1 28 6 vw 6FDA 6FpDA 6FDA 6FpDA DABA m MTMOS e PTMOS LM To f C CO Permeability barrers CO Pressure atm Figure 4 3 Pl
64. here the run nb exe program is found and subsequently what program is run Typically a run nb exe program from a degassing program i e nb 120m is run first to degas the system This is followed by a testing program i e nb he to start the test and collect the permeation data Testing Consecutive Gases The batchrun program can be set to run a series of gases without user interruption For example if set properly the permeation system will perform 3 runs of helium 3 runs of oxygen 3 runs of nitrogen 3 runs of methane and 3 runs of carbon dioxide at one click of the Start button To do this the last filename in the Files to Run column should be batchrun exe in the directory corresponding to the next gas in line See Table A 8 for an example In addition the Run on Startup box must be checked in each batchrun program 104 Table A 8 Example of how to run consecutive batchrun programs Files to Run Working Directory 1 run nb exe nb 120 2 run nb exe nb n2 3 run nb exe nb 120 4 run nb exe nb n2 5 run nb exe nb 120 6 run nb exe nb n2 7 batchrun exe nb ch4 105 Vitae Christopher L Hibshman was born in Lancaster County Pennsylvania in 1977 After graduating from Garden Spot High School in 1996 Chris attended Virginia Tech to study chemical engineering After his freshman year at college he was employed by Lockheed Martin Federal Systems as a co op for three semesters to work in a semiconductor
65. homogenously incorporated poly N vinylpyrrolidone with a silica gel also using a sol gel process Methytrimethoxysilane MTMOS was used as the organosilicate precursor because the methyl group prevented complete hydrolysis and provided a less cross linked and more flexible gel Gas 22 permeation results showed higher CO2 N and He N selectivities than predicted by Knudsen flow However this composite was only thermally stable to 150 C Joly et al synthesized polyimide organosilicate composite membranes by the addition of TMOS to a polyamic acid PAA in dimethylacetamid DMAc solution a so called site isolation method The solution was heated to 300 C to thermally imidize the polyimide PDMA and drive the sol gel reactions to higher degrees of completion The permeation results were promising in that the composite membrane demonstrated higher permeability selectivity and solubility coefficients than the pure polyimide However IR spectroscopy data showed that the polyimide was not completely imidized in the hybrid samples Sysel et al also synthesized polyimide organosilicate hybrid materials using the sol gel process TMOS was covalently bonded to a p aminophenyltrimethoxysilane APTMOS terminated ODPA ODA poly amic acid of controlled molecular weight The films were dried by following a temperature ramping scheme up to 300 C and drying for 5 hours The size of the silica domains did not exceed 100nm and the films we
66. istribution of the free volume within the polymer making the diffusion process more sensitive to molecular size of the penetrant 9999 Moe et al reported that heating a fluorinated polyimide at 240 C for 24 hours favored the rapid transport of small gas molecules H compared to larger gas molecules CH4 9 However our permeation data implies a different result The 400 C annealing process appears to favor the diffusion of the larger gas penetrants Larger differences in gas penetrant size for a particular gas pair resulted in larger decreases in permselectivity For example the permselectivity of similar sized oxygen and nitrogen O2 N2 molecules decreased only by 5 to 10 whereas the permselectivity of helium and methane He CH 4 which had a greater difference in molecule size resulted in a decrease of 30 to 6096 This observation may be due to a more uniform free volume distribution or local chain motions favoring a specific gas penetrant size 64 3 10 Effects of Annealing To illustrate the effects of annealing the relative changes of permeability and ideal selectivity were plotted as a function of molecule size Figures 3 19 and 3 20 show the permeability after annealing Panneal normalized to the permeability before annealing Po for the two DABA contents DABA 12 and DABA 25 in the polyimide matrix Figures 3 21 and 3 22 show the ideal selectivity after annealing normalized to the ideal selectivity before annealing for each
67. lar transport properties 51 Table 3 3 Summary of Permeability Diffusivity and Solubility coefficients for GFDA 6FpDA DABA 25 hybrid materials Evaluated at 35 C and 4 atm absolute Overall error 5 for permeability He O2 N2 CH CO P D S P D S P D S P D S P D S Before anneal Pure polyimide 83 1 709 0 00 629 5 51 0 87 1 20 137 0 66 0 46 024 1 48 203 209 7 36 22 5 wt TMOS 71 5 632 0 09 5 99 492 093 1 06 1 11 0 73 O48 0 18 203 15 7 1 45 3827 15 0 wt TMOS 22 5 wt MTMOS 69 3 709 0 07 569 504 0 86 1 07 1 16 0 70 0 52 023 1 75 166 1 32 9 57 15 0 wt MTMOS 82 3 521 0 13 6 75 629 082 1 32 1 53 065 0 63 027 1 76 228 2 26 7 66 22 5 wt PTMOS 55 7 649 0 07 5 08 468 082 0 98 1 18 064 054 0 24 41 71 19 1 1 90 7 66 15 0 wt PTMOS 60 2 747 0 06 496 463 081 094 1 10 065 052 0 22 1 81 18 4 1 78 7 87 After anneal Pure polyimide 184 885 0 16 22 7 147 1 18 485 3 79 0 97 245 0 76 295 773 583 10 1 22 5 wt TMOS 196 218 0 69 229 122 142 487 333 1 11 215 045 3 71 79 8 508 11 9 15 0 wt TMOS 22 5 wt MTMOS 169 754 0 17 18 7 10 7 1 33 383 268 1 09 1 68 0 43 295 60 1 4 01 11 4 15 0 wt MTMOS 205 952 0 17 24 1 15 1 1 21 5 07 3 99 0 97 1 93 0 57 259 81 1 617 10 0 22 5 wt PTMOS 149 342 0 33 224 12 1 1 40 521 3 50 1 13 3 79 0 95 304 944 5 71 12 6 15 0 wt PTMOS 177 1100 0 12 27 7 17 2 1 22 625 494 0 96 3 71 1 00 2 83 104 7 72 10 3 p p eo d D 10 cm s zem cm s cmHg S cm atm 52 Tab
68. le 3 4 Summary of Permeability Diffusivity and Solubility coefficients for GFDA 6FpDA DABA 12 5 hybrid materials Evaluated at 35 C and 4 atm absolute Overall error 5 for permeability 2 cm atm He Oz N2 CH CO P D S P D S P D S P D S P D S Before anneal Pure polyimide 107 966 0 08 100 7 73 098 201 1 89 081 1 00 035 215 340 314 824 22 5 wt TMOS 119 297 0 31 942 556 1 29 1 70 115 112 0 75 0 22 262 30 9 15 0 wt TMOS 22 5 wt MTMOS 117 451 013 11 8 860 1 05 262 235 084 1 22 043 2 29 15 0 wt MTMOS 120 750 0 12 122 811 1 14 253 207 093 1 24 036 259 440 321 104 22 5 wt PTMOS 73 7 591 0 09 942 898 086 1 88 2 05 0 70 090 037 1 87 30 7 275 851 15 0 wt PTMOS 936 650 0 11 9 08 699 099 180 1 71 080 093 030 233 323 2 73 9 02 After anneal Pure polyimide 176 1185 0 11 214 136 119 450 358 096 234 0 72 246 70 8 555 9 68 22 5 wt TMOS 154 1015 0 12 174 121 1 09 316 203 118 130 036 2 72 476 3 26 1141 15 0 wt TMOS 22 5 wt MTMOS 233 218 0 23 290 200 1 44 7 07 576 094 357 1 23 2 23 15 0 wt MTMOS 225 1132 0 15 31 2 19 7 120 7 07 580 093 383 1 11 263 110 7 95 10 5 22 5 wt PTMOS 137 899 012 276 193 110 587 512 087 352 116 230 2909 7 39 9 35 15 0 wt PTMOS 164 1001 012 247 168 1 12 559 510 083 3 30 097 258 918 694 10 0 Tm o erm d page 2 cm s cmHg 53 E E STP 20 Upper Bound Pure Polymer 22 5wt MTMOS 15 0wt MTMOS 22 5wt PTMOS 15 0wt PTMOS 22 5wt TMOS Pure Polymer 22 5wt MTMOS 15 0wt MTMOS 22 5w
69. less than 0 200 cmHg in order to open that valve and start the test The lower limits of other valves are 1 000 cmHg a value that cannot be obtained which means the valves will always be open during the program until hitting the upper limit Buffer Size Number of Stages Sampling Period and Stage Duration are all related to how often data is measured The buffer size is simply the number of times data is measured The number of stages is determined by how many different sampling periods are desired The stage duration determines the length of each sampling period Typically these programs follow the regime described in Table A 4 which has 3 sampling stages for different sampling periods and durations The total buffer size in Table A 4 is 3720 so the number in buffer size must be 3720 or the program will not run This sampling regime is designed to collect many points in the beginning of the permeation test to accurately define the time lag 96 Table A 2 Interface Channel and Bit Numbers for each valve of System CLH The Interface Device is 0 PIO 12 Valve Interface Channel Bit Number A 2 4 B 2 5 C 2 6 D 2 7 E 1 0 F 1 1 G 1 2 H 1 3 l 1 4 J 1 5 K 1 6 L 1 7 M 0 1 N 0 2 Table A 3 Interface Channel and Bit Numbers for each valve of System CJC The Interface Device is 2 5632TTL Valve Interface Channel Bit Number 1 CO 2 CHa 3 Ne 4 O2 5 He 6 7 O On On On One 1
70. ligands is altered 47 3 5 Swelling Studies Swelling measurements tabulated in Table 3 2 show that the degree of crosslinking increases for all samples with annealing Before being subjected to annealing the pure polyimide readily dissolved in NMP but after annealing the pure polyimide registered very little NMP uptake In fact the pure polyimide swelled less than any of the hybrid samples after annealing This surprising observation indicated significant crosslinking of the polyimide during the annealing process Since the polyimide in the hybrid materials was functionalized with APTEOS the hybrid materials were effectively crosslinked without the annealing process This was evident by the swelling measurements All but one of the hybrid samples DABA 12 PTMOS had relatively low NMP uptake After the annealing process the NMP uptake for all samples decreased suggesting that the degree of crosslinking increased in the polyimide organosilicate hybrids This was attributed to the hydroxy terminated groups in the organosilicate domains condensing with the APTEOS of the functionalized polyimide 48 Table 3 2 Summary swelling measurements in NMP for GFDA 6FpDA DABA 25 and 6FDA 6FpDA DABA 12 5 hybrid materials Surface area to volume ratio is 7128 Amount of NMP is 200 grams for every gram of sample Error 10 NMP Uptake NMP Uptake JNMP JPolymer JNMP JPolymer DABA 25 Before Anneal After Anneal Pure Polyimide 0 17 22 5
71. manian Penetrant plasticized permeation in polymethylmethacrylate Journal of Membrane Science 1992 74 29 C Staudt Bickel W J Koros Improvement of CO2 CH Separation Characteristics of Polyimides by Chemical Crosslinking Journal of Membrane Science 1999 155 145 A Bos I G M Punt M Wessling H Strathman CO2 induced plasticization in glassy polymers Journal of Membrane Science 1999 155 67 M Wessling S Schoeman Th van der Boomgaard C A Smolders Plasticization of Gas Separation Membranes Gas Separation and Purification 1991 5 222 J J Krol M Boerrigter G H Koops Polyimide Hollow Fiber Gas Separation Membranes Preparation and the Suppression of Plasticization in Propane Propylene Environments Journal of Membrane Science 2001 184 275 A Bos I G M P nt M Wessling H Strathman Suppression of CO2 Plasticization by Semiinterpentrating Polymer Network Formation Journal of Polymer Science Part B Polymer Physics 1998 36 1547 B Wang G L Wilkes Journal of Polymer Science Part A Polymer Chemistry 1991 29 905 H Schmidt Non crystalline Solids 1985 73 681 A B Brennan Ph D Dissertation Virginia Polytechnic Institute and State University 1990 F Orgaz Orgaz Gel to Glass Conversion Densification Kinetics and Controlling Mechanisms Journal of Non Crystalline Solids 1988 100 115 85 38 C J Brinker G W Scherer Sol Gel Science The Physics and Chemist
72. mbranes Figures 3 9 through 3 18 show the trade off curves of a number of gas pairs for the various hybrid and polymer systems before and after annealing As can be inferred from the 60 figures before annealing the DABA 25 pure polyimide had higher permselectivities than the hybrid materials After annealing the permselectivities of the MTMOS and TMOS based hybrid materials particularly for the He CH and CO CH gas pairs shown in Figures 8 and 9 were better than that of the pure polyimides suggesting that the presence of the silica structures enhanced the integrity of the films In fact both the permeability and the permselectivity for the COz CH gas pair increased in the DABA 25 22 5wt TMOS and MTMOS based hybrid materials On the other hand the PTMOS hybrid materials exhibited unusually large increases in permeability and large decreases in permselectivity with annealing which could be attributed to the poor polymer silica interface as discussed previously The dramatic increase in the permeability of all gases through all the membranes is indeed surprising in view of the apparent increases in the degree of cross linking upon annealing as was shown in the swelling studies Typically one would expect that an increase in cross linking would increase the permselectivity and decrease the permeability and diffusivity in glassy polymers One possible explanation for these results is that the loss of the degradation byproducts creat
73. n Cell eresien edipdir aneii tog si P I pM DS 32 2 6 Gas Permeation Volume of permeate side ssssss 34 Qu FHRA see eae Taaa aaa 36 2 8 Thermogravimetric Analysis Mass Spectrometry 36 29 Density Measurements soooensssnnenneoeeeennrnnnrrneerreenrrnnnrneeerennnnnnnnneeene 36 3 1 Decir 38 3 2 Visual OBSCIVAHONS s seien odio redo odo ae e is 39 33 ICD n M 39 CMS LO DAP a a E 43 3 5 Swelling boil o T CRTETEEUE 48 3 6 Husum m 50 Dur OOSS d THMSDOFTS cites ido on ot etu E ORIS Lab nds 51 3 8 Gas Transport Pure Polyimide ssssssssseseeeeeees 59 3 9 Gas Transport Hybrid Materials sssssssssssseseeeeeees 60 3 10 Effets OF Anne IQ icio oss oti esed eate tot dete redd eese dt ctt eod ead esa oed 65 Chapter 4 Effects of Feed Pressure on Gas Transport 71 4 1 Pls MCI Iw 71 42 Hesults and DISCUSSIOD ada mailed iiini iiai iii iaia 72 Chapter 5 Recommendations eeeeeeeeeeeeeeeeeeee nennen nnn 81 5 1 IST cd WORK MEM HR 81 5 2 FACTOT CIC OS c osse t dd aaa EEES 83 Appendix LABTECH NOTEBOOK Software esses 89 A 1 General InfortmallQttcs sce a EE ER rere re rere re EE 89 A2 SSVSIBIS cas o one t eae uet 89 A 3 Computer Architecture seio eot e ER ERR 90 Ad
74. nealed membrane 67 140 Pure Polymer 4 22 5wt MTMOS A amp 15 0wt MTMOS 22 5wt TMOS 22 5wt PTMOS 100 T7777 77 77 7Z S 15 0wt PTMOS 60 Change in Ideal Selectivity P4 P2 anneatea P4 P2 40 20 ON CO CH He CO He O He N He CH T T T T 0 0 00 0 20 0 40 0 60 0 80 1 00 1 20 1 40 Kinetic Diameter Difference Angstroms Figure 3 21 Normalized ideal selectivity as a function of molecule size difference for the 6FDA 6FpDA DABA 25 polyimide based membranes 110 Pure amp 22 5wt MTMOS Of a whe Beek Geter a MET 100 A amp 15 0wt MTMOS 22 5wt PTMOS 90 4 9 15 0wt PTMOS 9 22 5wt TMOS 80 7 70 4 Change in Ideal Selectivity P P2 annealed P P2 60 50 7 O N CO CH He CO He O He N He CH 40 T T T T T T 0 00 0 20 0 40 0 60 0 80 1 00 1 20 1 40 Kinetic Diameter Difference A Figure 3 22 Normalized ideal selectivity as a function of molecule size difference for the 6FDA 6FpDA DABA 12 polyimide based membranes P1 P2 anneai is the measurement for annealed membranes P P5 is the measurement for unannealed membranes Subscripts 1 and 2 refer to different gases The molecule size difference is measured as the difference in kinetic diameters for the selected gas pair The dashed line represents no change with anne
75. o ci 2 C0 PD Table A 4 Description of typical sampling regime for permeation tests Stage 1 2 3 Sampling Period 0 1 seconds 1 0 seconds 10 seconds Stage Duration 60 seconds 600 seconds 25200 seconds Buffer Size 600 600 2520 97 B Analog Input Block Al Each of these blocks corresponds to a measurement device which is usually a pressure transducer The block name usually describes the device used in the permeation system a Interface Device and Interface Channel correspond to the appropriate board name and channel number as shown in Tables A 5 and A 6 b Input Range corresponds to the output voltage of the measurement device which can usually be found on the device or the operator s manual for the specific device C Scale Factor and Offset Constant are used to convert the output voltage of the measurement device to appropriate measurement units For example the MKS 722A series pressure transducer has a voltage output of 10 volts The range of pressure transducer is 0 100 Torr but we want the measurement in cmHg Therefore the scale factor is 1 0 and the offset constant is 0 0 as long as the pressure transducer is zeroed properly d Buffer Size Number of Stages Sampling Period and Stage Duration are the same values as the digital output block see Section A 98 Table A 5 Interface Channel and Bit Numbers for each measurement device of System CLH The Interface Device is 1 DAS 8PGA D
76. of adsorptive but non permeable zeolites in a silicone rubber The diffusion time lag increased with increasing zeolite concentration as expected These observations were mainly attributed to the adsorptive capabilities of the zeolite filler materials which correlated well with theoretical estimates More importantly the filler materials had only minor effects on the steady state permeability measurements To date no theoretical models have been developed for predicting the time lag in heterogeneous systems consisting of permeable filler materials in a permeable matrix 1 5 Dual Mode Sorption Theory The equations presented so far are based on the assumption that concentration is independent of temperature and pressure If this were true a plot of permeability versus feed pressure should be constant However experimental observations have shown that permeability decreases with increasing feed pressure for most glassy polymers Although no theoretical model has been developed to fully explain this behavior an empirical model known as Dual Mode Sorption Theory was developed to fit the data 9 Dual Mode Sorption Theory describes the heterogeneity of glassy polymers by accounting for two contributions of sorption in a membrane One contribution follows Henry s Law which accounts for the mobile gas molecules dissolved into the amorphous polymer matrix The other contribution is associated with the Langmuir isotherm which accounts for
77. on Dioxide Test nb co2 nb co2 92 A 6 Description of Buttons Left Side As you may notice there are several buttons on the left side of the screen This section will describe the purpose of these buttons A Save Recall allows one to save recall and delete any program stored in this file Single click on the button and a new screen will appear Use the arrow keys on the keyboard to move the highlighted cursor The mouse does not work in this screen Hit enter to select save recall or delete If any programs are present they will be listed on the screen If you wish to save the current program type a new name If you wish to recall or delete a program type the exact program name where the cursor is blinking and hit enter Many different programs can be saved under this section In fact if you wish to alter a program suggest saving it as a separate name This will allow you to return to the original settings by simply recalling the original program You are permitted to look at only one program at a time If you cannot open a program be sure that other NOTEBOOK programs are closed WARNING Be careful The program will not ask if you are sure you want to save recall or delete any programs It is very easy to accidentally overwrite or delete an existing program B Zoom allows one to zoom in and zoom out of the screen by clicking the arrows 93 C button will show the block number associated with each block in the blue
78. onally it may be possible to decompose some of the organic groups in the organosilicate materials thereby making the membranes more nanoporous and potentially more selective The optimum annealing temperature should be high enough to gain advantages in the gas transport properties of the organosilicate materials without degrading the polyimide matrix As mentioned earlier in the sol gel chemistry section the dried gel is sintered at very high temperatures 600 to 1400 C to produce a fully densified and homogeneous glass However these temperatures cannot be attained 24 for the polyimide organosilicate hybrid membranes because the polyimide will decompose at elevated temperatures A different approach to organic inorganic hybrid membranes is the template approach where organic templates are incorporated into the inorganic matrix and then removed without collapsing the matrix creating a continuous network of micropores The organic matrix can be removed by heating the membranes at 400 to 550 C under air for several hours As Raman and Brinker have demonstrated a calcination temperature of 400 to 550 C may effectively pyrolyze the alkyl groups trapped in the inorganic domains resulting in nanoporous structures The size of the nanopores can be controlled by the size of the alkyl group Degradation of the polyimide during the annealing process is a valid concern Most polyimides are thermally stable at high temperatures
79. or He Oz The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed MEMDFANES ss ioi E AER O EK qd sad eus eL A pd E D Sr EM IDEH dY 56 Figure 3 15 Boundary diagrams of 6FDA 6FpDA DABA 12 pure polyimide and hybrid materials for He N2 The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed MEMDFANCS aset vertunt oon aba quaint pa Eod napa aad vu Rn ru ra Fat re ERR 57 Figure 3 16 Boundary diagrams of 6FDA 6FpDA DABA 12 pure polyimide and hybrid materials for He CH 4 The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed membranes esses enne 57 Figure 3 17 Boundary diagrams of 6FDA 6FpDA DABA 12 pure polyimide and hybrid materials for Oz Nz The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed membranes eese eene 58 Figure 3 18 Boundary diagrams of 6FDA 6FpDA DABA 12 pure polyimide and hybrid materials for CO2 CH 4 The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed membranes essssssssssessseeeeeeen enne 58 Figure 3 19 Normalized permeability as a function of molecule size for the 6FDA 6FpDA DAB
80. ost commercially viable membranes need to surpass this bound in performance In 1998 B Freeman substantiated Robeson s upper bound with a theoretical explanation relating the slope of the line to molecular parameters such as molecule size interchain spacing and polymer backbone stiffness The theoretical model was not a perfect fit with the empirical upper bound but the fundamental and simple theory did merit an upper bound for simple gas pairs 1 7 CO Plasticization Theory The concept of CO plasticization is widely used to explain experimental observations of CO transport through polymeric membranes Such experimental observations tend to vary depending upon the properties of the polymer such as morphology backbone rigidity chemical structure and degree of crosslinking For example empirical results show that the effects of plasticization on gas transport tend to be weak for flexible chain rubbery polymers but stronger for rigid chain glassy polymers It is believed that the microheterogeneity of glassy polymers accounts for the stronger interactions with the plasticizing agent Another common effect is the swelling of polymer matrix which simultaneously increases free volume of the polymer matrix and segmental mobility Both of these inter related parameters strongly influence gas transport by increasing the diffusivity and therefore the permeability of the 14 membrane Additionally the glass transition temperatur
81. ot of pure CO permeability as a function of feed pressure for various polyimide and hybrid systems at 35 C The permeabilities were calculated using the appropriate fugacity Error 2 73 56 w 6FDA 6FpDA 4 6FDA 6FpDA DABA m MTMOS e PTMOS oa ine AB CO CH Ideal Selectivity 8 wo o Feed Pressure psia Figure 4 4 CO CH ideal selectivity plotted as a function of feed pressure for various polyimide and hybrid systems at 35 C Error 4 NMP Uptake Qnup QPolymer 6FDA 6FpDA 6FDA 6FpDA DABA MTMOS hybrid 1 13 PTMOS hybrid Table 4 1 Swelling data for the polyimides and polyimide organosilicate hybrid materials Both hybrid materials consist of an organosilicate covalently bonded to a 6FDA 6FpDA D7ABA polyimide matrix With the exception of the MTMOS based hybrid all materials swelled to such a degree that measurements were not possible 74 Swelling measurements for each material are listed in Table 4 1 The amount of solvent uptake by a polymer was representative of the crosslink density As expected the two polyimides completely dissolved in NMP In addition the PTMOS base hybrid swelled to such an extent that measurement was not possible As reported elsewhere this observation can be explained by poor bonding between the polyimide matrix and organosilicate domains which resulted in low crosslink density The gas permeation resul
82. pointed out that the unannealed polyimide organosilicate hybrid materials exhibit an increase in free volume and a decrease in chain mobility with increasing silica content The authors concluded that these two phenomena are due to the formation of cross links between the polymer backbone and the silica structures These cross links are primarily limited to the interface between these two components inhibiting chain packing In addition we postulated that in the PTMOS hybrid system the interface was non selective as a result of the steric hindrance introduced by the bulky phenyl groups These groups prevented efficient cross linking with the polymer matrix and deactivated the hydrolysis and condensation reactions On the other hand silica structures generated from the TMOS and MTMOS alkoxides were efficiently incorporated into the polymer matrix because both MTMOS and TMOS have fast hydrolysis rates and higher concentration of silanol groups that can undergo further condensation with the functionalized polyimides As can be inferred from Tables 3 3 and 3 4 the permeability of the various gases increased more in the hybrid materials with annealing than it did in the pure polyimide Most of this increase can be attributed to an increase in the diffusion coefficients Thus the inclusion of organosilicate domains even at 7 5 and 1196 by weight was contributing to the increase in permeability and diffusivity of the various gases in the annealed me
83. rbonyl groups at 1720 cm and 1395 cm was consistent with the thermal hydrolytic degradation of the anhydride end groups The formation of a peak at 1500 cm and the slight changes in the 680 800 cm range were assigned to the in plane vibration of a benzene ring The environment around the benzene ring was altered during degradation of the anhydride end groups as shown in Figure 3 7 46 It is important to understand that the degradation of the anhydride end groups only applies to the pure polyimide The hybrid materials are fully functionalized with APTEOS and therefore do not have any anhydride end groups This is evident by the FTIR ATR spectra in Figure 3 4 The peaks assigned to the carbonyl groups at 1720 cm and 1395 cm do not change with annealing for any of the hybrid samples Finally it is important to compare TGA MS data with FTIR ATR observations concerning the methyl groups in the MTMOS sample Peaks at 1260 cm and 768 cm were assigned to Si CHs stretching and rocking respectively 9 9 Referring to the FTIR ATR spectra in Figure 3 4 for MTMOS both of these peaks increase with the annealing process Therefore we conclude that the methyl ligands are not being decomposed during the 400 C anneal We attribute the higher intensities to changes in the bonding structure of the silicon atom attached to the methyl ligand 9 As the silica network cross links during the annealing process the environment around the methyl
84. re optically transparent Using TGA density measurements and IR spectroscopy the authors concluded that the sol gel reactions namely the condensation reactions were not completed Cornelius recently synthesized and characterized a series of hybrid materials consisting of organosilicate domains covalently bonded to a fluorinated polyimide matrix FDA 6FpDA DABA C NMR was used to confirm complete imidization of the polyamic acid Sol gel chemistry was employed to produce the organosilicate domains These hybrid materials were cast into thin 23 films for use as gas separation membranes The membranes were heat treated at 220 C to remove any excess solvents a temperature which was not high enough to complete the condensation reactions in the sol gel process Furthermore Cornelius concluded that the gas transport properties were influenced by the amount and type of alkoxide precursor used in the sol gel chemistry For example the MTMOS based hybrid samples had the largest increases in permeability and the PTMOS based samples had the largest decreases in permeability compared to the pure polyimide The optimum amount of organosilicate in the polyimide was 15 0 wt solution during synthesis 1 10 Annealing Polyimide Organosilicate Hybrid Membranes The purpose for annealing these hybrid membranes is to maximize their gas transport properties namely that of the organosilicate domains by further driving the sol gel reactions Additi
85. rganosilicate hybrid membranes on gas transport properties Chapter 4 studies the effects of feed pressure on gas transport properties of unannealed hybrid membranes These particular hybrid systems were chosen due to their optimum gas transport properties based on previous characterization Acknowledgements This project would not have been feasible without the loving support of my wife Michelle She always seemed to know the right advice and words of encouragement at the right time and was always there when need her most am especially grateful for the patience understanding and extensive efforts of my advisor Dr Eva Marand She always seemed to offer me assistance when needed it most Without her encouragement and support would not have gotten my Master s degree from Virginia Tech also want to express thanks to Chris Cornelius for his patience in answering my many questions Thanks also to Todd Pechar for being a sounding board and helpful labmate would also like to thank the North American Membrane Society for financial support via an undergraduate research fellowship am also deeply indebted to Steve McCartney for his assistance with the TEM images In addition many thanks are extended to David Williamson and Koji Yamauchi from the Department of Chemistry at Virginia Tech for assistance with the TGA MS Table of Contents ADSIIGCE i toten M MEME E EN ME ME ii Format of Thesis soda dass Ecux M TERRI SE SUE
86. rials are present in these systems rapid heating and cooling may affect the polyimide organosilicate interface where the expansion coefficients may differ In general it would be good to study this interface as it may be an important contribution to the gas transport For example preliminary small angle x ray scattering SAXS results 81 indicate that upon annealing large spherical domains are formed Specifically a TMOS based hybrid sample has very small domains 46 before annealing and very large domains 2250 after annealing One possible explanation for this is the larger domains actually contain polyimide chains constrained in an organosilicate network This study also concluded that the organic ligands were not removed during the 400 C annealing process The use of a ligand which decomposes at lower temperatures such as methacryloxypropyl which decomposes at 350 C may be advantageous in this case Although methacryloxypropyl is a relatively large organic group removing it from the organosilicate network may greatly improve the gas transport properties for olefins and paraffin separations Finally the annealing environment in the study was limited to nitrogen flowing over the membrane It would interesting to examine the effects of annealing the hybrid materials under vacuum as it may assist in removing the organic ligands or preventing some degradation to the polyimide 82 5 2 References
87. rials for He CH The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed membranes 57 Upper Bound Pure Polymer 22 5wt MTMOS 15 0wt MTMOS 7 5wt MTMOS 22 5wt PTMOS 15 0wt PTMOS 22 5wt TMOS Pure Polymer 22 5wt MTMOS 15 0wt MTMOS 7 5wt MTMOS 22 5wt PTMOS 15 0wt PTMOS 22 5wt TMOS tO OPP PrP eeOCOODPE O N Selectivity o 7 8 910 20 30 40 50 60 70 80 90100 Helium Permeability barrers Figure 3 17 Boundary diagrams of 6FDA 6FpDA DABA 12 pure polyimide and hybrid materials for Oz Na The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed membranes 60 Upper Bound Pure Polymer 22 5wt MTMOS 15 0wt MTMOS 7 5wt MTMOS 22 5wt PTMOS 15 0wt PTMOS 22 5wt TMOS Pure Polymer 22 5wt MTMOS 15 0wt MTMOS 7 5wt MTMOS 22 5wt PTMOS 15 0wt PTMOS 22 5wt TMOS 50 40 30 O PPrPEeeFO lL PDP Ee CO CH Selectivity c 20 30 40 50 60 70 80 90100 200 300 400 Helium Permeability barrers Figure 3 18 Boundary diagrams of 6FDA 6FpDA DABA 12 pure polyimide and hybrid materials for CO2 CH 4 The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed membranes 58 3 8 Gas Transport Pure Polyimide The amount of di
88. ry of Sol Gel Processing Academic Press Inc San Diego 1990 39 P F James The Gel to Glass Transition Chemical and Microstructural Evolution Journal of Non Crystalline Solids 1988 100 93 40 M M Collinson Analytical Applications of Organically Modified Silicates Mikrochimica Acta 1998 129 149 41 N K Raman C J Brinker Organic Template Approach to Molecular Sieving Silica Membranes Journal of Membrane Science 1995 105 273 42 N K Raman M T Anderson C J Brinker Template Based Approach to the Preparation of Amorphous Nanoporous Silicas Chemical Materials 1996 8 1682 43 Y Lu C Guozhong R P Kale S Prabakar G Lopez C J Brinker Microporous Silica Prepared by Organic Templating Relationship between the Molecular Template and Pore Structure Chemical Materials 1999 11 1223 44 Y S Kim K Kusakabe S Morooka S M Yang Preparation of Microporous Silica Membranes for Gas Separation Korean Journal of Chemical Engineering 2001 18 106 45 K Kusakabe S Sakamoto T Saie S Morooka Pore Structure of Silica Membranes Formed by a Sol Gel Technique using Tetraethoxysilane and Alkyltriethoxysilanes Separation and Purification Technology 1999 16 139 46 B K Sea K Kusakabe S Morooka Pore Size Control and Gas Permeation Kinetics of Silica Membranes by Pyrolysis of Phenyl Substituted Ethoxysilanes with Cross Flow Through a Porous Support Wall Journal of Membrane Science 19
89. s research which measures the accumulation of the permeate gas pressure as a function of time Typical results for this method are illustrated in Figure 1 1 The data is separated into two sections the first region representing transient diffusion and the second steady state diffusion The time lag is defined as the extrapolation of the steady state region to the x intercept as indicated by the 0 in Figure 1 1 The flux can be written in terms of the slope of the steady state region which is represented in Equation 1 5 3 y Slope V cm m 1 5 i RT A mol R is the gas constant Tis the temperature Vis the permeate volume and A is the active area of diffusion Equations 1 4 and 1 5 can be combined to express P as a function of the slope of steady state diffusion as shown in Equation 1 6 Notice that P is normalized to the differential pressure and membrane thickness 3 p Slope V 22414 cm STP l l 1 6 RT A mol PiP One assumption for this method is that the permeate pressure py remains negligible throughout the entire permeation process If permeate pressure is not negligible the flux will begin to level off as a function of time and a different boundary condition will need to be specified resulting in a complex relationship Typical units to represent the permeability coefficient are barrers which are defined in Equation 1 7 10 cm STP cm cm s cmHg barrer 2 5
90. screen D Trash Can looks like a disappearing square hole in the bottom left of screen Click and drag a block from the blue screen onto this button and the block will be deleted from the program Be careful you will not be able to retrieve it A 7 Description of Blocks Bottom of Screen At the bottom of the screen there are twenty eight blocks representing the different types used for programming To add any of these blocks to the program click and drag the block onto the blue area of the screen To change the placement of any block simply click and drag To delete a block refer to Section A 6 D To start double click on an existing block A new screen will appear To exit this screen click on the Done button or hit the Esc key The first two lines of this screen are information about the total Number of Blocks in the program and the Current Block number you are currently viewing The blocks are numbered by the order in which they are added to the programming screen The next line is labeled as the Block Type and should be highlighted in blue This blue highlight is the cursor This cursor will move by using the arrow keys on the keyboard or clicking on a different space with the mouse The block 94 type can be changed at this line by hitting enter A list will appear and the desired block type can be chosen form the list by scrolling and hitting enter The next two lines correspond to the Block Name and Units Eac
91. sed hybrid has a mechanism of diffusion similar to the pure polyimide hypothesize that the polyimide phase separates from the organosilicate domains in PTMOS based hybrid membranes during annealing thus creating a path of lesser resistance for the gas molecules to diffuse 66 800 700 600 Panneal P 300 200 100 500 400 Pure Polymer k 22 5wt MTMOS amp 15 0wt MTMOS 9 22 5wt TMOS 22 5wt PTMOS 6 15 0wt PTMOS He co 07 N CH 2 50 3 00 3 50 Kinetic Diameter Angstroms 4 00 Figure 3 19 Normalized permeability as a function of molecule size for the 6FDA 6FpDA DABA 25 polyimide based membranes Pure polymer refers to pure 6FDA 6FpDA DABA 25 polyimide Panneai is the measurement of an annealed membranes P is the measurement of an unannealed membrane 400 350 4 300 4 Panneal P 200 150 4 100 250 4 Pure Polymer k 22 5wt MTMOS amp 15 0wt MTMOS 22 5wt PTMOS 6 15 0wt PTMOS 22 5wt TMOS He co 0 N2 CH 2 50 3 00 3 50 Kinetic Diameter Angstroms 4 00 Figure 3 20 Normalized permeability as a function of molecule size for the 6FDA 6FpDA DABA 12 polyimide based membranes Pure polymer refers to pure 6FDA 6FpDA DABA 12 polyimide Panneai is the measurement of an annealed membranes Pois the measurement of an unan
92. t PTMOS 15 0wt PTMOS 22 5wt TMOS OOPPEXOC ODP SE He O Selectivity 50 60 70 80 90 100 200 300 Helium Permeability barrers Figure 3 9 Boundary diagrams of 6FDA G6FpDA DABA 25 pure polyimide and hybrid materials for He O The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed membranes 200 Upper Bound m Pure Polymer A 22 5wt MTMOS 15 0wt MTMOS 22 5wt PTMOS 15 0wt PTMOS 3 100 22 5wt TMOS D 90 W Pure Polymer e0 A 22 5wt MTMOS 15 0wt MTMOS 9 7 22 5wt PTMOS D 60 15 0wt PTMOS e 22 5wt TMOS z 50 o 40 30 50 60 70 80 90 100 200 300 Helium Permeability barrers Figure 3 10 Boundary diagrams of 6FDA 6FpDA DABA 25 pure polyimide and hybrid materials for He Nz The line represents Robeson s 1991 Upper Bound The blue symbols represent unannealed membranes and red symbols represent annealed membranes 300 Upper Bound m Pure Polymer A 22 5wt MTMOS 200 15 0wt MTMOS 22 5wt PTMOS 15 0wt PTMOS 3 22 5wt TMOS D m Pure Polymer S A 22 5wt MTMOS gt 100 15 0wt MTMOS 90 22 5wt PTMOS Q 15 0wt PTMOS 80 o 22 5wt TMOS x 70 Q 60 T 50 40 30 50 60 70 80 90100 200 300 400 500 Helium Permeability barrers Figure 3 11 Boundary diagrams of 6FDA 6FpDA DABA 25 pure polyimide and hybrid materials for He CH 4 The line represents Robeson s 199
93. teel gas permeation cell was custom designed and constructed for this study and is illustrated in Figure 2 4 This cell is designed for thin membranes measuring two inches 5 08 cm in diameter The cell consists of two parts a feed half and a permeate half each four inches 10 16 cm in diameter Two porous 100um stainless steel sintered disks manufactured by M tt Corporation are inserted to provide mechanical support for the membrane and prevent cracking or fractures Three radial static seal fluorocarbon Viton Static Seal O rings are used to seal the membranes and prevent leaking The two halves of the cell are fastened using six 7 bolts In order to accurately measure the temperature of the feed gas a RTD probe is inserted into the cavity of the feed side via 1 8 Swagelok fitting The active permeation area is 1 78 sq in 11 51 cm for gas diffusion 32 Viton O rings Dense Feed Permeat i Porous disks Figure 2 4 Cross sectional schematic diagram of gas permeation cell 33 2 6 Gas Permeation Volume of permeate side The volume on the permeate side was minimized to increase the accuracy of the data and reduce the experimental time This volume was obtained by using the ideal gas law and adding a fixture of known volume and a valve near the permeate pressure transducer as shown in Figure 2 5 A nonpermeable membrane is placed in the cell and Valve 1 is opened The permeate side of the system
94. tory that contains a list of folders to support the bld nb exe programs contained in the respective folder This list is tabulated in Table A 1 Each bld nb exe program is created to satisfy the specific testing conditions of each gas namely the 90 sampling protocol length of test and to open and or close specified valves The only difference for the bld nb exe file in the degassing programs is the length of time the system degasses A 4 bld nb exe Program Sections A 4 A 7 concern only the bld nb exe program To follow along open an existing bld nb exe program listed in Table A 1 by double clicking on the filename A 5 Main Menu As is common with most Windows based software there is a menu at the top of the page including File and Run buttons Clicking on File allows one to exit the bld_nb exe program Clicking on Run allows one to run the program currently shown on the computer screen 91 Table A 1 List of program folders in their respective directories Name in System CLH Name in System CJC Working directory C batch C cjc Degasses for 3 minutes nb_003m Degasses for 30 minutes nb_030m Degasses for 45 minutes nb_045m nb 45m Degasses for 60 minutes nb_ 60m Degasses for 90 minutes nb_090m nb 90m M EE for 120 nb_120m nb_120m os Jor tap nb_180m nb_180m xui E Tore40 nb 240m nb 240m Helium Test nb he nb he Oxygen Test nb o2 nb o2 Nitrogen Test nb n2 nb n2 Methane Test nb ch4 nb ch4 Carb
95. ts of the PTMOS based hybrid are also consistent with this explanation The PTMOS based hybrid exhibited slightly higher permeability than the 6FDA 6FpDA DABA polyimide but the poorest selectivity of all the materials The lack of polyimide organosilicate interactions allows gases to diffuse faster but with less discrimination On the contrary swelling of the MTMOS based hybrid was limited an indication that the MTMOS alkoxide is effectively incorporated into the polyimide matrix However the permeabilities of the MTMOS based hybrids were significantly higher than the 6FDA 6FpDA DABA polyimide Characterization results reported elsewhere suggest that the sol gel processing of MTMOS leads to a loosely crosslinked structure which contributes to higher diffusion rates In the case of CH diffusion all samples exhibited a decrease in permeability with increasing feed pressure up to 29 atm This observation is consistent with dual mode sorption theory thus we conclude that methane did not plasticize any of the membranes studied However different results were observed for the diffusion of CO as shown in Figure 4 3 All of the membranes 75 exhibited an initial decrease in permeability at feed pressures up to about 17 atm Between 17 and 20 atm each of the membranes began experiencing a slight increase in permeability which was an indication of plasticization It was surprising to see all the membranes respond in a similar manner at comp
96. vantages but are difficult to apply to heterogeneous systems such as the hybrid materials used for this study 1 6 Gas Transport through Hybrid Membranes For our hybrid materials we have inherent heterogeneity in the glassy polyimide in addition to the heterogeneity due to the organosilicate networks Needless to say both the Time Lag Method and the Dual Mode Sorption Theory have to be applied with caution when describing the gas transport through these hybrid membranes For example the Time Lag Method was originally developed for rubbery polymers Dual Mode Sorption Theory describes glassy polymers based on only two methods of sorption whereas have identified at least four methods of sorption for these hybrid systems i gas molecules dissolved into the amorphous polymer matrix Henry s Law ii gas molecules adsorbing in to holes in the glassy polymer Langmuir iii gas molecules adsorbing in to holes in the organosilicate network and iv and gas molecules adsorbing in to holes at the polyimide organosilicate interface With so many unknown variables sorption isotherms will provide little information pertaining to the mode of 12 transport in these membranes Therefore we have focused on transient permeation testing whose analysis assumes that sorption simply follows Henry s Law with a solubility coefficient averaged over the matrix and the organosilicate network The Time Lag Method was chosen to for this project
97. y polymers and heterogeneous systems it is necessary to designate these transport properties as effective properties In summary 3 p Slope V 22414 cm STP Y l 1 13 PrP RT A mol 1 eff 6 E P So 1 15 eff D at 1 16 Perry et al have explored the theoretical effects of incorporating non permeable inclusions into a polyimide matrix on the diffusive time lag by incorporating mica flakes into a polycarbonate film They demonstrated that the mica flakes must be oriented perpendicular to the gas flow in order to create a tortuous path for the diffusion of the gas penetrants In addition they presented a theoretical model relating the mica loading and size of flakes to increases in diffusion time lag which correlated well with experimental data The model is shown in Equation 1 17 0 with flakes 0 without flakes x 0 x 0 1 17 0 is the time lag is the aspect ratio of the flakes and 9 is the volume of flakes in the polymer For example using Equation 1 17 with a mica loading of 3096 and size aspect ratio of 20 the diffusion time lag can be increased nearly 36 times These results were attributed mainly to increasing the tortuousity of the path of diffusion and illustrate that the diffusion time lag can be altered by the presence of inclusions in the polymer These particular results are promising for improving barrier membranes In a separate study Paul and Kemp examined the effects
98. y 1999 9 1741 C Joly S Goizet J C Schrotter J Sanchez M Escoubes Sol Gel Polyimide Silica Composite Membrane Gas Transport Properties Journal of Membrane Science 1997 130 63 P Sysel R Pulec M Maryska Polyimide Silica Hybrid Materials Based on a p Aminophenyltrimethoxysilane Terminated Poly amic acid s Polymer Journal Toyko 1997 29 607 12 J B Alexopoulos J A Barrie and D Machin The Time Lag for the Diffusion of Gas Mixtures Polymer 1969 10 265 83 13 H L Frisch Fundamentals of Membrane Transport Polymer Journal 1991 23 445 14 F Vasak Z Broz A Method for Determination of Gas Diffusion and Solubility Coefficients in Poly Vinyltrimethysilane Using a Personal Computer Journal of Membrane Science 1993 82 265 15 D Perry W J Ward E L Cussler Unsteady Diffusion in Barrier Membranes Journal of Membrane Science 1989 44 305 16 D R Paul D R Kemp The Diffusion Time Lag in Polymer Membranes Containing Adsorptive Fillers Journal Polymer Science Symposium No 41 1973 79 17 A S Michaels W R Vieth J A Barrie Diffusion of Gases in Polyethylene Terephthalate Journal of Applied Physics 1963 34 13 18 J H Petropolous Quantitative Analysis of Gaseous Diffusion in Glassy Polymers Journal of Polymer Science Part A 2 1970 8 1797 19 D R Paul Effect of Immobilizing Adsorption on the Diffusion Time Lag Journal of Polymer Science Part A 2 1969

Polyimide-Organosilicate Hybrid Materials

Contents

Download Pdf Manuals

Related Search

Related Contents