Deep desulfurization of full range and low boiling diesel streams from Kuwait Lower Fars heavy crude

Information on feed quality and, in particular, various types of sulfur compounds present in the diesel (gas oil) fractions produced form different crudes and their HDS reactivities under different operating conditions are of a great value for the optimization and economics of the deep HDS process. This paper deals with deep desulfurization of gas oils obtained from a new heavy Kuwaiti crude, namely, Lower Fars (LF) which will be processed in the future at Kuwaiti refineries. Comparative studies were carried out to examine the extent of deep HDS, and the quality of diesel product using two gas oil feeds with different boiling ranges. The results revealed that the full range diesel feed stream produced from the LF crude was very difficult to desulfurize due to its low quality caused by high aromatics content (low feed saturation) together with the presence of high concentrations of organic nitrogen compounds and sterically hindered alkyl DBTs. The low-boiling range gas oil showed better desulfurization compared with the full range gas oil, however, deep desulfurization to 50 ppm sulfur was not achieved even at a temperature as high as 380 °C for both feeds. The desulfurized diesel product from the low-boiling gas-oil feed was better in quality with respect to the S, N and PNA contents and cetane index than the full-range gas-oil feed.     

Two-step adsorption process for deep desulfurization of diesel oil

An integrated adsorption process for deep desulfurization of diesel fuel was proposed and examined. Conventionally hydrodesulfurized straight run gas oil (HDS-SRGO) having less than 50 ppm sulfur was also adsorptively treated with activated carbon fiber (ACF) to attain the ultra low sulfur gas oil having less than 10 ppm sulfur. The ACF, used in cleaning-up HDS-SRGO, was successively examined in straight run gas oil (SRGO) treatment to enhance its hydrodesulfurization (HDS) reactivity over conventional CoMo catalyst by removing the nitrogen and refractory sulfur species contained in SRGO. Such integrated adsorption–reaction process makes it possible to utilize the maximum adsorption capacity of ACF and achieve ultra deep desulfurization og SRGO. Regeneration of used ACF with a conventional solvent was proved very effective in restoring its adsorption capacity.     

Thermal, spectroscopy, and morphological studies on polymorphic crystals in poly(heptamethylene terephthalate)

Polymorphism and its influential factors in poly(heptamethylene terephthalate) (PHepT) were probed using differential scanning calorimetry (DSC), Fourier-transform infrared (FTIR) spectroscopy, and wide angle X-ray diffraction (WAXD). PHepT exhibits two crystal types (α and β) upon crystallization at various isothermal melt-crystallization temperatures (T_cs) by quenching from different T_maxs (maximum temperature above T_m for melting the original crystals). Melt-crystallized PHepT with either initial α- or β-crystal by quenching from T_max lower than 110 °C leads to higher fractions of α-crystal, but crystallization from T_max higher than 140 °C leads to higher fractions of β-crystal. In addition to T_max, polymorphism in PHepT is also influenced by crystallization temperature (T_c = 25-75 °C). When PHepT is melt-crystallized from a high T_max = 150 °C (completely isotropic melt), it shows solely β crystal for higher T_c, and solely the α-crystal for T_c < 25 °C; in-between T_c = 25 and 35 °C, mixed fractions of both α- and β-crystals. However, by contrast, when PHepT is melt-crystallized from a lower T_max = 110 °C, it shows α-crystal only at all T_cs, high or low.     

Transformer oils prepared from the vacuum distillates of Egyptian crude paraffinic petroleum

Vacuum gas oil and spindle oil obtained from the vacuum distillation of paraffinic crude oil from the Western Desert region of Egypt were utilized for transformer oils production. The vacuum gas oil, spindle oil and blends therefrom containing different percentages of these oils and were subjected to aromatic extraction, dewaxing catalytic hydrogenation processes. The aromatic extraction process used N-methyl-2-pyrrolidone as solvent, for a feed ratio of 0.7:1 (wt/wt) at an extraction temperature of 55 °C, while in the dewaxing process the wax is removed by chilling the raffinate at − 42 °C using MEK-toluene (60:40,vol/vol). The hydrofinishing process was achieved in a pilot plant using NiO–MoO₃/Al₂O₃ catalyst. These refining processes improved the viscosity indices and the pour points of the resulting oils and removed most of the polar impurities. It was found that, the transformer oil formulated from 7:3 b.w. vacuum gas oil:spindle oil respectively meets the IEC 60296-2003 standard specifications requirements. Its sulfur content is very low and non-corrosive. The gassing tendency is also low due to its low aromatic content. Further, the electrical properties which are significant for transformer oils fulfill the standard specifications as a result of the low water content and polar contaminants. Moreover, addition of 2,4-dimethyl-6-tertiary butyl phenol increases the oxidation stability of the produced oil, by acting as a free radical inhibitor.     

Compositional analysis of deasphalted oil before and after hydrocracking over zeolite catalyst by Fourier transform ion cyclotron resonance mass spectrometry

Elemental compositions of components in feed and catalytically processed deasphalted oils were characterized by Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS). The processed oils which were hydrocracked over a zeolite catalyst at three different reaction temperatures (370, 380, and 390 °C) were analyzed. Species of the deasphalted oils were ionized either by electrospray ionization (ESI) or by in-beam electron ionization (EI). The ESI mass spectra were obtained from every feed and processed deasphalted oil. Over 550 chemically different compounds were observed in the feed oil mass spectra. Molecular formulas for the detected peaks were calculated by using accurate mass. The compounds with one N atom as well as one N and S atoms were detected as major and minor component, respectively, in every mass spectrum. The number of the detected species in processed deasphalted oil decreases as the reaction temperature increases. However, the carbon distribution of NS-containing species shifts to high number as the reaction temperature increases. Molecular formulas distribution against Z-value (Z-value is defined as C_nH_2n + ZN_mS_sO_o) and C-number were investigated for the ESI mass spectra. Z-value distribution of the peaks assigned to N-compounds was convergent in its compounds with Z = − 25 as increasing the reaction temperature. Detailed mass spectrum analysis reveals that compounds which were not detected in the feed oil were observed in the mass spectra of processed oils; N, S, and O-containing compounds. For the in-beam EI only the processed oil at 390 °C yields approximately 700 resolved peaks at adopted probe temperature (300 °C) of EI. Molecular formula analysis for the observed peaks was conducted as well as ESI. It reveals that the molecular formulas having Z-value (− 30 < Z < 2) and carbon number ranged from 8 to 31 except for (− 18 < Z < − 12, 15 < C-number < 22) were contained in the processed deasphalted oil. Using complementary ionization techniques to characterize the feed and catalytic reacted deasphalted oils provide better understanding of fuel processing conditions.     

High quality diesel by hydrotreating of atmospheric gas oil/light cycle oil blends

An improved process for high-quality diesel fuel production by hydrotreating atmospheric gas oil (SRGO) and light cycle oil (LCO) blends is presented in this paper. For this purpose, a set of blends of 5, 10 and 15% by volume of LCO with final boiling points of 300, 325 and 350 °C with a full range gas oil (FBP 350 °C) was hydrotreated in a pilot plant at 340-380 °C, 5.4 MPa, 2.5 h⁻¹ LHSV using a commercial Co-Mo catalyst. A relationship between the concentration of refractory sulfur compounds (those boiling above 316 °C) and aromatics content in the feedstock with the hydrotreating temperature required for meeting a 0.05% sulfur specification was found.The experimental data obtained during the desulfurization was quantitatively represented by a 1.50 to 1.56 order rate equation, with activation energies between 18.9 and 34.1 kcal/mol, depending on the feedstock.     

A model for initial-stage sintering thermodynamics of an alumina powder

A model was proposed to calculate several thermodynamic parameters for the initial-stage sintering of an alumina powder obtained after calcinations at 900 °C for 2 h of a precursor. The precursor was synthesized by an alumina sulphate-excess urea reaction in boiling aqueous solution. The cylindrical compacts of the powder with a diameter of 14 mm were prepared under 32 MPa by uniaxial pressing using oleic acid (12% by mass) as binder. The compacts were fired at various temperatures between 900 and 1400 °C for 2 h. The diameter (D) of the compacts before and after firing was measured by a micrometer. The D value after firing was taken as a sintering equilibrium parameter. An arbitrary sintering equilibrium constant (K_a) was calculated for each firing temperature by assuming K_a = (D_i − D) / (D − D_f), where D_i is the largest value before sintering and D_f is the smallest value after firing at 1400 °C. Also, an arbitrary change in Gibbs energy (ΔG _a°) was calculated for each temperature using the K_a value. The graphs of ln K_avs. 1 / T and ΔG _a° vs. T were plotted, and the real change in enthalpy (ΔH°) and the real change in entropy (ΔS°) were calculated from the slopes of the obtained straight lines, respectively. Inversely, real ΔG° and K values were calculated using the real ΔH° and ΔS° values in the ΔG° = − RT ln K = ΔH° − TΔS° relation. The best fitting ΔH° and ΔS° values satisfying this relation were found to be 157,301 J mol^− 1 and 107.6 J K^− 1 mol^− 1, respectively.     

Electrochemical behaviour of propane-fed solid oxide fuel cells based on low Ni content anode catalysts

Two different low Ni content (10 wt.%) anode catalysts were investigated for intermediate temperature (800 °C) operation in solid oxide fuel cells fed with dry propane. Both catalysts were prepared by the impregnation of a Ni-precursor on different oxide supports, i.e. gadolinia doped ceria (CGO) and La_0.6Sr_0.4Fe_0.8Co_0.2O₃ perovskite, and thermal treated at 1100 °C for 2 h. The Ni-modified perovskite catalyst was mixed with a CGO powder and deposited on a CGO electrolyte to form a composite catalytic layer with a proper triple-phase boundary. Anode reduction was carried out in-situ in H₂ at 800 °C for 2 h during cell conditioning. Electrochemical performance was recorded at different times during 100 h operation in dry propane. The Ni-modified perovskite showed significantly better performance than the Ni/CGO anode. A power density of about 300 mW cm⁻² was obtained for the electrolyte supported SOFC in dry propane at 800 °C. Structural investigation of the composite anode layer after SOFC operation indicated a modification of the perovskite structure and the occurrence of a La₂NiO₄ phase. The occurrence of metallic Ni in the Ni/CGO system caused catalyst deactivation due to the formation of carbon deposits.     

The interaction effects of dehydration function on catalytic performance and properties of hybrid catalysts upon LPDME process

Three bi-functional catalysts have been prepared by physical mixing of a commercial methanol synthesis catalyst (CuO–ZnO–Al₂O₃) with three different methanol dehydration catalysts including: H-MFI90, γ-Al₂O₃ and H-Mordenite in order to investigate the role of interaction effects of dehydration component on characteristic properties and performance of these admixed catalysts. The bi-functional catalysts have been characterized by XRD, N₂ adsorption, H₂-TPR, NH₃-TPD and XRF techniques and tested in a mixed slurry bed reactor at the same operating conditions (T = 240 °C, P = 50 bar, H₂/CO = 2, SV = 1100 ml g-cat^− 1 h^− 1) for 60 h time on stream. Among the examined bi-functional catalysts, the physical mixture of KMT + HMFI-90, which had lower reducing peak temperature (T = 200 °C), higher S_Cu (39.1 m² g-cat^− 1) and Cu Dispersion (11.6%), showed higher X_CO (84 mol%), yield of DME (Y_DME = 55.5 mol%), DME selectivity (Select_DME = 66.7 mol%) and also good stability over 60 h time on stream as compared to the other catalysts. This could be assigned, from NH₃-TPD results, to more middle strength acidic sites of H-MFI90 zeolite (SiO₂/Al₂O₃ = 90, total acid site density = 476 µmol/g-cat) which inhibits detrimental interactions with methanol synthesis catalyst and deep dehydration of methanol.     

Aromatic gas oil cracking under realistic FCC conditions in a microriser reactor

A paraffinic hydrowax feed spiked with naphthalene, anthracene, and phenanthrene was cracked in a once-through microriser reactor at 575 °C and with a catalyst-to-oil (CTO) ratio of 4.8 g_cat g_oil⁻¹. The conversion by cracking reactions is limited to the paraffinic fraction of the feed and the alkyl groups associated with the benzene ring in aromatic compounds; the aromatic probes did not crack under the applied conditions, and in fact an additional amount of naphthalene was formed by complex dealkylation and hydrogen transfer reactions. The ‘uncrackabilty’ of aromatics was directly demonstrated by processing an aromatic gas oil, containing 33.3 wt% aromatics. Experiments were performed with residence times between 0.05 and 8.2 s, keeping the temperature (525 °C) and CTO ratio (5.5 g_cat g_oil⁻¹) constant. The data was interpreted with a simplified first-order five-lump kinetic model, where approximately 19 wt% of the feed was found to be uncrackable. HCO (feedstock) conversion took mainly place during the first two seconds and coke was only formed during the first 50 ms of catalyst-oil contact. Gasoline was not overcracked to gas. Approximately 50 wt% of the LCO fraction was formed during this 50 ms and did not change thereafter.     

Activity and sulfidation behavior of the CoMo/Al2O3 hydrotreating catalyst: The effect of drying conditions

The influence of drying condition of the CoMo/Al₂O₃ catalyst prepared using citric acid as a chelating agent on the sulfidation behavior and on the catalytic activity to hydrodesulfurization of straight-run gas oil (SRGO) was investigated. The catalysts dried at 110, 220, 300 and 400 °C were studied using Raman, IRS and DTG techniques. The sulfidation behavior with straight-run gas oil spiked with dimethyldisulfide (DMDS) was investigated using gas chromatography (GC) with a thermal conductivity detector and GC with an atomic-emission detector for analysis of gas and the liquid phases, respectively. It was shown that the sulfidation behavior of the catalysts prepared using the chelating agents depends on the drying condition: the lower drying temperature, the later DMDS conversion and oxide precursor sulfidation starts. A higher activity in SRGO desulfurization was obtained with catalysts dried at 110 and 220 °C. This phenomenon was accounted for by the stabilization effect of chelating agent that retards precursor sulfiding but provides favorably the formation of active CoMoS phase and achieving the highest activity.     

Preparation and characterization of reticular SSC cathode films by electrostatic spray deposition

Sm_0.5Sr_0.5CoO_3−δ (SSC) cathode films were deposited on CGO (Gd_0.1Ce_0.9O_1.95) electrolyte substrates by electrostatic spray deposition to prepare SSC/CGO/SSC symmetrical cells. Deposition parameters were changed systematically to examine their effects on film microstructure and electrode performance. A set of deposition parameters including a 0.01 M precursor solution containing metal nitrates in a mixture solvent of de-ionized water (0.6 vol%), ethanol (1.5 vol%) and diethyl butyl carbitol (97.9 vol%), a flow rate of 6 ml/h for precursor solution, a deposition temperature of 350 °C and an imposed electric field of 10 kV/3 cm was identified for preparation of films with a highly porous reticular structure. The superior performance of a reticular SSC electrode was evidenced by its low interfacial resistances of 0.275 and 0.018 Ω cm² measured in 500 and 700 °C, respectively. These values were one-half to one order of magnitude smaller than that of the screen-printed or slurry-painted electrodes.     

Analysis and optimization of two-column cryogenic process for argon recovery from hydrogen-depleted ammonia purge gas

Traditionally, high-purity argon recovery from air is considerably difficult owing to the boiling point of argon close to that of oxygen. Recently with the increasing demands for argon, another attractive source of ammonia purge gas has been paid more attention. In this paper with an objective of minimizing energy consumption per argon product, the two-column process for recovering argon from hydrogen-depleted ammonia purge gas is analyzed and optimized in detail on the ASPEN PLUS platform. Firstly, the model of two-column process is set up using the standard unit operation blocks and PENG-ROB property method of ASPEN PLUS, in which validation of PENG-ROB property method is carried out by comparison with a total 623 experimental data from three aspects: vapor-liquid equilibrium, liquid phase density, and enthalpy. It is followed by the thermodynamic and simulation and sensitivity analysis, which on the one hand can reduce the number of decision variables related to optimization problem, and on the other hand can obtain reasonable parameter specification, variables initial values and ranges, thus effectively ensuring the later optimization algorithm converges quickly and accurately. Finally the built-in sequential quadratic programming (SQP) solver of ASPEN PLUS is adopted to solve the minimum energy consumption optimization problem of two-column process. On the processor of 2.66 GHz Intel(R) Core (TM)2 Duo CPU with 4 GB RAM, the whole optimization only takes CPU times 10 s or so to accomplish. The optimal results show that thermal state of feed to demethanizer is a very efficient and valuable means to reduce system energy consumption which at T_C05 = 103 K is only 87.4% of that at T_C05 = 109 K where T_C05 is the temperature of feed to demethanizer directly reflecting its thermal state. The condensing pressure of hydrogen-depleted ammonia purge gas also plays a vital role in reducing system energy consumption which is less at higher condensing pressure, whereas it almost has no influence on the yield and purity of argon recovery. The optimal operating pressure of flash separator used to remove the residual hydrogen in the feed hydrogen-depleted ammonia purge gas is 0.4-0.6 MPa (A); the most economical reflux ratio of argon distillation column is 1.15, and that of demethanizer varies from 0.33 to 0.45 depending on thermal state of feed to demethanizer.     

Recrystallization studies on isotropic cold-crystallized PET: Influence of heating rate

The nanostructural changes associated to the multiple melting behaviour of isotropic cold-crystallized poly(ethylene terephthalate) (PET) have been investigated by means of simultaneous wide- and small-angle X-ray scattering, using a synchrotron radiation source. Variations in the degree of crystallinity, coherent lateral crystal size and long period values, as a function of temperature, for two different heating rates are reported for cold-crystallized samples in the 100-190 °C range. The Interface Distribution Function analysis is also employed to provide the crystalline and amorphous layer thickness values at various temperatures of interest. Results suggest that samples crystallized at both low (T_a = 100-120 °C) and high (T_a = 160-190 °C) temperatures are subjected to a nearly continuous nanostructural reorganization process upon heating, starting immediately above T_g (≈80 °C) and giving rise to complete melting at ≈260 °C. For all the T_a investigated, a melting-recrystallization mechanism seems to take place once T_a is exceeded, concurrently to the low-temperature endotherm observed in the DSC scans. For low-T_a and slow heating rates (2 °C/min), a conspicuous recrystallization process is predominant within T_a + 30 °C ≤ T ≤ 200 °C. In contrast, for high-T_a, an increasingly strong melting process is observed. For both, high- and low-T_a, an extensive structural reorganization takes place above 200 °C, involving the appearance of new lamellar stacks simultaneously to the final melting process. The two mechanisms should contribute to the high-temperature endotherm in the DSC scan. Finally, the use of a high heating rate is found to hinder the material's overall recrystallization process during the heating run and suggests that the high-temperature endotherm is ascribed to the melting of lamellae generated or thickened during the heating run.     

Investigation of the intra-particle sintering kinetics of a mainly agglomerated alumina powder by using surface area reduction

An alumina precursor was prepared by the aluminium sulphate (0.20 M) and excess urea reaction in boiling aqueous solution. The precursor was calcined at 900 °C for 2 h and then δ-Al₂O₃ powder having volumetric agglomeration degree of 80% was obtained. Cylindrical compacts having diameter of 14 mm were prepared under 32 MPa by axial pressing using oleic acid as binder. Each compact was fired isothermally at various temperatures between 950 and 1400 °C. The firing time was changed from zero to 2 h. The fired compacts were examined by scanning electron microscopy (SEM) and nitrogen adsorption techniques. The specific surface areas (S/m² g^− 1) of the samples were calculated using the Brunauer, Emmett, and Teller (BET) procedure. The rate constant (k) and mechanism-characteristic parameter (n) were obtained for different temperatures between 950 °C and 1150 °C from the application of the neck-growth sintering rate (NGRM) model on the surface area reduction data. An Arrhenius equation and the parameter n for the sintering were found in the forms of k = (7.648 × 10⁶ h^− 1) exp (− 186,234 J mol^− 1 / RT) and n = 4.0 × 10^− 7 T³-1.7 × 10^− 3 T² + 2.3 T − 1030.8 respectively. The parameter n changes in the interval 0.61 <  n < 1.34 with rising temperature having maximum at about 1025 °C. Based on the SEM images and NGRM data, the intra-particle sintering was discussed.     

Oxidative desulfurization of simulated light fuel oil and untreated kerosene

An experimental investigation was conducted on the oxidative desulfurization of model sulfur compounds such as dibenzothiophene and benzothiophene in toluene as a simulated light fuel oil with a mixture of hydrogen peroxide as the oxidant and various acids as the catalyst. The influences of various parameters including reaction temperature (T), acid to sulfur molar ratio (Acid/S), oxidant to sulfur molar ratio (O/S), type of acid, and the presence of sodium tungstate and commercial activated carbon as a co-catalyst on the fractional conversion of the model sulfur compounds were investigated. The experimental data obtained were used to determine the reaction rate constant of the model sulfur compounds and the corresponding activation energy. Moreover, the adsorption of model sulfur compounds on the commercial activated carbons supplied by Jacobi Co. (Sweden, AquaSorb 101) was studied and the effects of different parameters such as temperature, and various chemical treatments on the adsorption of the sulfur compounds were investigated. Furthermore, the oxidative desulfurization of untreated kerosene with the total sulfur content of 1700 ppmw produced by an Iranian refining company (Isfahan refinery) was successfully investigated. These experiments were performed using formic acid as the catalyst and hydrogen peroxide as the oxidant at the mild operating conditions of T = 50 °C, O/S = 5, and the Acid/S = 10. It was realized that about 87% of the total sulfur content of untreated kerosene could be removed after 30 min oxidation followed by liquid–liquid extraction.     

Preparation, characterization and mechanical properties of epoxidized soybean oil/clay nanocomposites

New epoxidized soybean oil (ESO)/clay nanocomposites have been prepared with triethylenetetramine (TETA) as a curing agent. The dispersion of the clay layers is investigated by X-ray diffraction (XRD) and transmission electron microscopy (TEM). XRD and TEM data reveal the intercalated structure of ESO/clay nanocomposites has been developed. The thermogravimetric analysis exhibits that the ESO/clay nanocomposites are thermally stable at temperatures lower than 180 °C, with the maximum weight loss rate after 325 °C. The glass transition temperature, T_g, about 7.5 °C measured by differential scanning calorimetry (DSC) and T_g about 20 °C measured by dynamic mechanical study have been obtained. The difference in the T_g between DSC and dynamic measurements may be caused by different heating rate. The nanocomposites with 5-10 wt% clay content possess storage modulus ranging from 2.0×10⁶ to 2.70×10⁶ Pa at 30 °C. The Young's modulus (E) of these materials varies from 1.20 to 3.64 MPa with clay content ranging from 0 to 10 wt%. The ratio of epoxy (ESO) to hydrogen (amino group of TETA) greatly affects dynamic and tensile mechanical properties. At higher amount of TETA, the nanocomposites exhibit stronger tensile and dynamic properties.     

Polycarbonate/liquid crystalline polymer blend: Crystallization of polycarbonate

The enhanced crystallization of polycarbonate in the blend of liquid crystalline polymer/polycarbonate/(ethylene-methyl acrylate-glycidyl methacrylate) copolymer (LCP/PC/E-MA-GMA) was studied by wide angle X-ray diffraction (WAXD) and differential scanning calorimetry (DSC). The LCP/PC/E-MA-GMA 5/95/5 blends annealed at 200 °C, for 2, 4, 6, and 10 h, present an obvious crystalline structure corresponding to PC crystallization. The PC crystal obtained shows two melting temperature, T_m1 of about 214 °C and T_m2 of about 231 °C, with a total heat of fusion of 29 J/g (annealing time = 10 h). The preliminary results indicate that amorphous PC can be induced to crystallization by the synergistic action of LCP dispersed phase and reactive compatibilizer.     

Conductivity of CGO and CSO ceramics obtained from freeze-dried precursors

Ce_0.8Gd_0.2O_1.9 (CGO) and Ce_0.8Sm_0.2O_1.9 (CSO) have been prepared as polycrystalline materials using a freeze-dried precursor. This method yields amorphous nanometric powders. Crystallization of the fluorite phase occurred on heating at 600 °C or higher temperatures. The grain size of freeze-dried powders increases to about 100 nm after calcination at 800 °C, or about 200 nm after firing at 1000 °C. Freeze-dried powders were used to prepare dense ceramic disks by sintering at 1400 °C. Some disks were sintered at 1000 °C by adding small amounts of cobalt nitrate solution to assist the densification. The electrical conductivity results obtained for these gadolinia-doped ceria and samaria-doped ceria ceramics are similar to those obtained for CGO pellets obtained from commercial nanopowders (Rhodia). Though the bulk conductivity of CSO is probably higher than that of CGO, its grain boundary conductivity is inferior, and tends to control the overall behaviour, at least at relatively low temperatures.     

Study of thermal annealing on Matrimid fiber performance in pervaporation of acetic acid and water mixtures

Unlike the case for low-sorbing gas molecules such as N₂ and O₂, Matrimid^® hollow fibers swell significantly in acetic acid (HAc) and water mixtures. Acetic acid was found to plasticize the polymer and facilitate the passage of the ‘slow’ molecule, acetic acid, while not affecting the ‘fast’ molecule, water, as significantly. The membrane selectivity of water over HAc inevitably decreased with increasing HAc content in the feed for the standard membrane. Sub-T_g (glass transition temperature) thermal annealing was used to stabilize large bore size defect-free Matrimid^® hollow fibers by suppressing the HAc-induced plasticization. The resulting membrane selectivity was increased up to 95 using a 520 μm outer diameter/325 μm inner diameter and 20 cm long defect-free annealed fiber with acceptable water flux (1.6 kg/m² h) for 20 wt% HAc concentration feed streams at 101.5 °C. Micro-fluorescence spectra provide strong support for the notion that thermal annealing can reduce the free volume and promote the formation of charge transfer complexes. As a consequence, the improved rigidity of the polymer chains increases the polymer discrimination of shape and size for penetrants, although no chemical reaction occurs with thermal annealing.