R&D on adsorption processing technology using pitch activated carbon fiber

Transcription

1 1999D R&D on adsorption processing technology using pitch activated carbon fiber 1. Contents of empirical research With respect to waste water, exhausts and other emissions in the petroleum refining industry, present-day regulations are being cleared, but requirements are becoming stricter each year for the sake of environmental conservation. These regulations are expected to become stricter in the future, and in addition to restrictions on concentrations, the weight given to restrictions on gross volumes will increase. What is more, among present-day processing technologies, few are suited for the removal of rare pollutants, and costs must be extensive. Against this backdrop, it is vital that we take positive action for even cleaner refineries, not being satisfied with present conditions, and work to improve waste processing technologies. Of the methods for treating wastes by adsorbents, one of the most promising at present is a method of processing which uses activated carbon fiber as an adsorbent. The reason is that activated carbon fiber far outstrips conventional activated carbon in performance. This is because it has very sharp pore diameter distribution; its pore diameters are small, and it offers the advantage of rapid adsorption/desorption speed thanks to its unique structure. At present, however, the mechanism of its adsorption/desorption has not been adequately clarified, and pore diameter control technology in line with the ingredients targeted for and adsorption removal has not been established. For this reason the performance of activated carbon fiber is not being used to its full. The present R&D effort has been aimed at resolving these issues. It is based on the premise that if the functional characteristics of chemical reactivity, for instance, on the pore surface of activated carbon fiber can be conferred, adsorption characteristics can be improved (including adsorption of ingredients that cannot be removed by physical adsorption), the applications for activated carbon fiber as a catalyst can be expanded, and the high-performance, high functionality and expanding applicability of activated carbon fiber can be advanced dramatically. Accordingly, the present R&D effort has been aimed at the following objectives. (1) Establishment of technology for control of average pore diameter and pore diameter distribution, and for measurement and analysis of the same, so that targeted substances can be removed efficiently by adsorption. Completion of technology for manufacture of high-performance activated carbon fiber. (2) Development of the various elemental technologies such as adsorption and regeneration technologies using high-performance activated carbon fiber. Construction of a system comprised of these elemental technologies, and evaluation/verification of its performance. 1

2 This year process technology for practical application of activation reactions will be studied; attention will focus on manufacturing technology for mass production; conditions for manufacture of activated carbon fiber and correlations among basic properties will be studied; research on technology for high functionality will be completed; and an activated carbon fiber exhibiting the performance ultimately targeted will be developed. What is more, research will be completed for development of activated carbon fiber regeneration technology and forming technology and for systemization of elemental technologies aimed at rendering practical equipment in modular form so that this equipment can be used for processing of wastes from oil refineries. Details on research covering each topic are presented below. 1.1 Research on technology for activation of isotropic petroleum pitch fiber In an effort to manufacture activated carbon fiber exhibiting high adsorption performance, technology will be established for control of pore diameter and pore distribution using an alkali activation method introduced as a new activation method. Process technology for practical application will be studied and efforts will be made to develop manufacturing technology for mass production. 1.2 Research on correlations between basic properties and the conditions for manufacture of activated carbon fiber By evaluating the correlations among basic properties such as specific surface area of activated carbon fiber produced under each condition of alkali activation, carbon skeletal structure and adsorption performance, the various factors governing the adsorption performance of activated carbon fiber will be elucidated and manufacturing conditions will be optimized. 1.3 Research on technology for highly functional activated carbon fiber Using feedstock pitch adjusted by the method of catalyst ingredient addition, the impact of catalyst ingredients on alkali activation will be evaluated in detail, and the possibilities for manufacturing, through alkali activation, activated carbon fiber scattered to a high level with support catalyst, and for practical application of such fiber, will be evaluated. 1.4 Research on technology for regeneration of activated carbon fiber The relationships among adsorption/desorption volume, adsorption/desorption speed and processing conditions will be evaluated together with efficiency in regeneration, and regeneration technology of high processing efficiency will be established, on the assumption of practical application. 1.5 Research on technology for forming activated carbon fiber The changes in bulk density, pore structure and specific surface area that accompany formation will be evaluated along with their effects on adsorption performance, and a forming technology adapted to the alkali activation method will be optimized. 1.6 Research for systemization of elemental technologies including adsorption and regeneration. High-performance adsorption processing equipment will be systematized on the basis of elemental technologies covering adsorption, reproduction, forming, etc. 2

3 2. Empirical research results and analysis thereof 2.1 Research on technology for activation of isotropic petroleum pitch fiber In order to optimize the conditions for activation of isotropic petroleum pitch, the impact of activation time period and of activation temperature in CO 2 gas activation was investigated. Infusible mat of 12m in fiber diameter was treated within an activation temperature range of 973K ~ 1273K (700 C ~ 1000 C) and an activation time period range of 1 ~ 96 hours, and the basic properties of activated carbon fiber obtained by using the nitrogen adsorption method and the steam adsorption method were evaluated. The results are presented in Table 2-1. Table 2-1 conditions of isotropic petroleum pitch Specimen Activ ation temperature Activ ation time period yield Specif ic surf ace area Pore volume Median pore diameter N 2 method H 2 O method At an activation temperature of 1037K (800 C) or below, the reaction speed was extremely slow, and the specific surface area after twenty-four hours of activation was no more than 1000 m 2 /g. To obtain a specific surface area of 1500 m 2 /g or more, at least 96 hours of activation was required. In contrast, at an activation temperature of 1223K (950 C) or above, the reaction progressed dramatically and activated carbon fiber of 3000 m 2 /g or greater specific surface area could be obtained from activation at 1234K (970 C) for about 8 hours and from activation at 1273K (1000 C) for about 6 hours. In the relationship between specific surface area and median pore diameter, it was found that a classification by degree of activation could be made into the following three domains. 3

4 In the domain where specific surface area is 1500 m 2 /g or less, with a median pore diameter on the order of 0.6 nm there is almost no change, but in the domain of 1500 ~ 3000 m 2 /g a median pore diameter expands greatly from 0.7 to 1.3 nm as the specific surface area is increased. In the domain where specific surface area exceeds 3000 m 2 /g, the area remains almost totally unchanged even if activation is excessive, and only pore diameter increased sharply, to 1.6 nm or more. This tendency is totally independent of activation conditions. The relationship between specific surface area and central pore diameter is determined by activation yield, and the aforesaid results suggest that pore diameter cannot be controlled through activation conditions. Based on the above results, a study was done on a two-stage activation method in which activation takes place two times at different temperatures so as to alter the relationship between specific surface area and pore diameter. The results are shown in Table 2-2. When the specific surface area after two-stage activation was 2000 m 2 /g or less, the relationship between specific surface area and pore diameter remained almost totally unchanged as compared to the single stage activation method. At 2000 m 2 /g or more, an activated carbon fiber of extremely small pore diameter as opposed to specific surface area could be obtained, revealing that pore diameter can be controlled with the two-stage activation method. Table 2-2 Conditions for two-stage activation of isotropic petroleum pitch Specimen Activ ation temperature Activ ation time period yield Specif ic surf ace area Pore volume Median pore diameter N 2 method H 2 O method *1 Under activation temperature, activation time period and activation yield, the upper column denotes the first activation and the lower column, the second activation. *2 The activation yield denotes the weighted percentage of infusible fiber after the first and second stages of activation. Next, a new activation method employing alkali reagent was investigated for the manufacture of activated carbon fiber of higher adsorption performance. The activation process is portrayed in Figure

5 In this activation method, potassium hydroxide (KOH) of 2 ~ 8 times greater fiber weight is added to mat-shaped carbonized fiber manufactured by the melt-blow spinning method. After this fiber has been heated in melted KOH for two hours at 873 ~ 1073K (600 ~ 800 C) in an inert atmosphere, the alkali ingredients are cleaned with diluted hydrochloric acid and distilled water, and dried at ordinary and reduced pressures. Infusible fiber Carbonized fiber KOH addition (heating in melted alkali) Washing with water (KOH removal) Drying (heating under reduced pressure) The conditions of alkali activation and the basic properties of the activated carbon fiber obtained are presented in Table 2-3. In comparison to conventional activation with carbon dioxide for instance, activated carbon fiber of higher specific surface area could be obtained at lower temperature and in a shorter time period through activation employing alkali reagent. The activation yield is three times higher than with the conventional method, and the activation was found to progress very efficiently. Nevertheless, it was also noted that the pores formed through alkali activation tend to be smaller in diameter than pores from gas activation, and changes due to the conditions of activation are slight. In alkali activation, the mechanism of the activation reaction is unclear; detailed studies must be done on such things as the structural differences with pores formed by the conventional method. Table 2-3 Results of alkali activation conditions ACF basic properties Starting material agent Amount of addition Temperature Time Yield Specific surface area Pore volume Median pore diameter Carbonized fiber Twof old weight Threef old weight Carbonized f iber Carbonized f iber Fourfold weight Inf usible f iber Inf usible f iber Inf usible f iber * 1 Amount of addition: Denotes KOH weight ratio for starting material at KOH activation, and CO 2 concentration in nitrogen gas at CO 2 activation. * 2 Carbonized fiber: Product calcinated in nitrogen at 973K for 1 h. Next, the relationship between activation reaction and the amount of potassium hydroxide (KOH) added was studied for the purpose of reducing potassium hydroxide additions. It was discovered that the activation reaction progresses as the amount of KOH additive is increased, and there are increases in specific surface area, median pore diameter and pore volume. At eightfold KOH, activated carbon fiber that far exceeds 3000 m 2 /g in specific surface area can be obtained, but with increases in specific surface area, activation yield drops sharply. It is necessary, therefore, to make a general evaluation of adsorption performance and productivity (activation yield) and to determine the optimum conditions of manufacture. 5

6 2.2 Correlations between basic properties and conditions for manufacture of activated carbon fiber In activation employing alkali reagent, washing takes place adequately by using distilled water after the activation reaction, but it is suspected that potassium remains in the activated carbon fiber obtained because the alkali ingredients cannot be completely removed. Given this situation, a quantitative analysis was conducted of the compositional ingredients of alkali-activated products and gas-activated products through x-ray photoelectron spectroscopy analysis and elemental analysis. The results of elemental analysis are given in Table 2-4. With alkali activated products, roughly one percent potassium remains, and it was found that the oxygen content is 2 to 7 times greater than in gas-activated product (CO 2 gas-and steam-activated products). The amounts of residual potassium and oxygen tend to decrease as the specific surface area becomes greater. The same results were obtained with XPS analysis. It is suspected that the increase in oxygen is due to the introduction of oxygen atoms in KOH during the activation reaction, but it has not yet been clarified why the residual amounts decreases together with an increase in specific surface area. Table 2-4 Results of elemental analysis of activated carbon fiber Specimen name Specific surface area Elemental analysis Carbon (wt%) Hy drogen (wt%) Oxygen (wt%) Potassium (wt%) Alkali-activated product CO2 gas-activated product Steam-activated product And the wastewater treatment performance of the alkali activated active carbon fiber was evaluated and compared with the conventional product. The basic surface basic properties of each activated carbon fiber and adsorption performance with respect to phenols, etc. are shown in Table 2-5. As a result, in comparison with activated product obtained by the conventional method and with activated carbon fiber sold on the market, alkali activated product has an exceptionally high equilibrium amount of adsorption in relation to phenols and total organic carbons, and with activated carbon fiber of 2500 m 2 /g or greater specific surface area, in particular, adsorption performance (250 mg/g) greater than 200 mg/g, the final target value of equilibrium amount of absorption for phenol was obtained. Nevertheless, at a specific surface area of 2500 m 2 /g or more, adsorption performance hardly changes at all even if the specific surface area is increased. Thus it appears that a specific surface area in the range of 2300 ~ 2500 m 2 /g is most ideal in consideration of the balance between adsorption performance and activation yield. 6

7 Shown in Table 2-6 are the results of measurements of hydrogen sulfide equilibrium amount of adsorption. For these measurements, activated carbon fiber was left at 298K (25 C) adsorption temperature for 24 hours in a fixed volume atmosphere of hydrogen sulfide at 100 mg/1 (100 ppm) concentration, and after the adsorption equilibrium was reached, the hydrogen sulfide concentration of the atmosphere was measured. It is known that in the adsorption of hydrogen sulfide by activated carbon fiber, amount of adsorption is altered by the steam concentration in the atmosphere. Adsorption was thus measured at steam concentrations of 0%, 40% and 80%. The measurements indicated that in comparison to conventional gas activated product, the volume of hydrogen sulfide adsorption with alkali activated carbon fiber is exceptionally high, largely surpassing the target value of 60 mg/g. The hydrogen sulfide equilibrium adsorption volume at 0% steam exhibits a high value as the specific surface area of the activated carbon fiber becomes smaller. With alkali activated carbon fiber, the change in pore diameter due to specific surface area is small, and the lighter the degree of activation, the larger is the volume of residual oxygen. It appears, therefore, that the above results are not due to differences in pore diameter, but rather because the residual oxygen in the activated carbon fiber contributes to adsorption of hydrogen sulfide. 7

8 Table 2-5 Evaluations of wastewater treatment with each type of activated carbon fiber Adsorbent starting material Specimen name method ACF basic properties BET method specific surface area t - plot pore diameter Phenols Equilibrium amount of adsorption Adsorption performance Total organic carbon Equilibrium amount of adsorption Total nitrogen Equilibrium amount of adsorption Petroleum pitch Phenol PAN Coal pitch Table 2-6 Evaluations of hydrogen sulfide treatment with each type of activated carbon fiber Adsorbent starting material Specimen name method ACF basic properties Specific surface area Pore diameter Hydrogen sulfide adsorption performance Steam Steam Steam Equilibrium amount of adsorption Equilibrium amount of adsorption Equilibrium amount of adsorption Petroleum pitch 8

9 2.3 Research on technology for highly functional activated carbon fiber The adsorption performance of activated carbon fiber is governed by physical characteristics such as specific surface area, average pore diameter and pore diameter distribution, but the chemical characteristics of pore interior surface make up another important factor. As part of an effort to achieve highly functional activated carbon fiber, and in order to investigate methods for contributing catalytic characteristics to activated carbon fiber, a compound (silver nitrate), which becomes a catalytic ingredient in the process of pitch production by the nitrate acid oxidation method, was added; and evaluations were made of the impact on pitch reaction, the impact on activation reaction, the residual amount of additive, additive residual format, degree of additive scattering, and the pore structure of the activated carbon fiber obtained with catalyst added. In pitch production, 40% nitrate acid is added to the raw material and after nitration, an optional volume of silver nitrite is added and the moisture content and unreacted nitrate acid are removed by distillation. Thereafter, an isotropic pitch containing 0.04 ~ 13 wt% silver of prescribed softening point is obtained through thermal polymerization and decompressed distillation. Next, the pitch obtained undergoes spinning, infusibilization and activation under normal methods and conditions, resulting in the production of activated carbon fiber containing silver. In measuring the specific surface area and pore diameter of this activated carbon fiber, employing the nitrogen adsorption method, it was found that results are about the same as for regular activated carbon fiber not containing silver. From x-ray diffraction measurements and observations with the transmission type electron microscope, it was determined that metallic silver measuring 0.1 ~ 0.5m in granular diameter scatters uniformly inside the fiber. In evaluating the catalytic characteristics of activated carbon fiber containing silver, no elevation of reactivity was noted, nor any major effect on adsorption performance. Studies covering other applications will have to be made in the future. 2.4 Technology for regenerating activated carbon fiber In order to investigate the conditions for separation of hydrogen sulfide adsorbed into activated carbon fiber, experiments were conducted on the equilibrium adsorption of hydrogen sulfide in the gaseous phase, and thereafter, reproduction tests were conducted using activated carbon fiber specimen adhered to saturation with hydrogen sulfide. Tests were conducted by the following method. Hydrogen sulfide adsorption specimen was washed with a fixed volume of distilled water and allowed to dry adequately; then the equilibrium adsorption volume of hydrogen sulfide was measured again. The equilibrium adsorption volumes of hydrogen sulfide before and after reprocessing are presented in Table 2-7. As a result, it became clear that almost the entire volume of adsorbed hydrogen sulfide was separated as sulfuric acid by washing with distilled water, and the original adsorption performance is recovered. 9

10 Table 2-7 Reproduction tests on hydrogen sulfide adsorption specimen First amount of adsorption (fresh sample) Second amount of adsorption (after regenerating) Equilibrium concentration Equilibrium amount of adsorption (mg/g) 2.5 Research on activated carbon fiber forming technology In using activated carbon fiber as adsorbent, various usage formats, such as the method of filling mat-shaped fiber into a porous pack, are conceivable, but various problems arise. The circulation speed of wastes is limited because of pressure loss by the pack; the fiber in this format is not suitable for regeneration; and filling may be difficult depending on the shape of the pack. Accordingly, an investigation was made of forming method with the aim of forming a shape for filling in the course of activated carbon fiber production. The following three procedures, shown in Figure 2-2, are conceivable as methods of forming activated carbon fiber. (1) Formation of mat-shaped pitch fiber immediately after spinning, then infusibilization and activation (case 1) (2) Formation and activation after infusibilization of mat-shaped pitch fiber (case 2) (3) Formation after completing the final stage of activation in mat shape (case 3) Of these procedures, in case 3 a binder (e.g., pitch, resin) must be added in forming because a thermal history of 1173K (900 C) is received through activation (same carbon structure as in carbonization at 1173K or above). As a result, the surface of the activated carbon fiber becomes covered with binder and a drop in adsorption capacity can be anticipated due to closure of the pores. For this reason, case 3 was dropped from the study and simplified forming equipment was used to explore the remaining two procedures. In case 1, it was found that mat-shaped pitch fiber can be formed by controlling the formation temperature but problems arise thereafter in the infusibilization procedure. On the other hand, in formation using infusible fiber (case 2), formation was possible without using binder, and in the activation process thereafter, activated carbon fiber possessing the same basic properties (e.g., specific surface area, pore diameter) could be obtained under exactly the same conditions of activation as in the case of regular mat product. Isotropic pitch Spinning Mat-shaped pitch fiber Infusibilization Mat-shaped infusible fiber (alkali method) Mat-shaped activated carbon fiber Case 1 formation Case 2 formation Case 3 formation Formation pitch fiber Infusibilization Formation infusible fiber (gas method) Formation activated carbon fiber Figure 2-2 Formation processes 10

11 Next, the suitability of infusible fiber compact produced by the aforesaid method to alkali activation was examined. After mat-shaped, infusible fiber was compacted to a size of 100 mm x 100 mm x 20 mm by adding heat and pressure, it was filled into a crucible made of nickel; potassium oxide equivalent to four times the weight of the compact was added, and activation took place for two hours at 973K (700 C). Upon completion of the activation reaction, the nickel crucible was removed from the activation tower, and observation of its interior revealed that the compact had been broken down into powder form. The reason carbon fiber compact breaks down under alkali activation is perhaps that the fiber is greatly enlarged during such activation. In order to curtail this enlargement, the compact was activated after being carbonized at high temperature. It was thus confirmed that breakdown of compact due to fiber enlargement becomes difficult when the temperature of carbonization before activation is elevated to 1073K (800 C). But under this condition the activation reactivity drops low and adequate adsorption performance cannot be obtained. In using alkali activation, it will be necessary to implement case 3 in which binder is used for forming after activation treatment. 2.6 Systemization of elemental technologies involving adsorption, reproduction, etc. Based on the phenol adsorption performance of high-performance activated carbon fiber test manufactured, the quantity of activated carbon fiber used was calculated from assumed values for phenol concentration in oil refinery wastewater, for waste flow volume and for total waste volume. A study was also done for optimization of the size and configuration of activated carbon fiber compact for the purpose of improving efficiency in reproduction, exchange and filling of activated carbon fiber to adsorption treatment equipment. Nevertheless, an evaluation of formats for practical application has not yet been made because a process for practical application of alkali activation has not been established and problems remain with forming and reproduction technologies. 3. Results of empirical research (1) By studying in detail the conditions for processing by the alkali activation method, equilibrium amount of adsorption could be further added in relation to phenols in wastewater and hydrogen sulfide in exhaust gasses. In contrast to initial adsorption performance after development, performance could be improved about 2.5 times with phenols and over four times with hydrogen sulfide so that the final targeted values of performance were cleared. (2) From a comprehensive evaluation of production efficiency and of adsorption performance from the standpoint of alkali activation processing conditions and reaction yield, it was determined that activated carbon fiber of 2300 ~ 2500 m 2 /g in specific surface area is most ideal. 11

12 4. Summary It was discovered that the alkali activation method is outstanding in terms of both manufacturing efficiency and adsorption performance of activated carbon fiber. The final targeted value for adsorption performance was cleared by a wide margin but because the mechanism of activation reaction has not been elucidated adequately, a more detailed analysis of the alkali activation reaction will be required for establishing a practical process in which this activation method is used effectively. What is more, in the development of forming and reproduction technology, applicability to alkali activation will be absolutely essential. In the future, therefore, the relationships among compact bulk density, activation and adsorption/desorption characteristics will have to be determined and process technology for practical application will also have to be established. Copyright 1999 Petroleum Energy Center all rights reserved. 12