Removal of tar base from coal tar aromatics employing
solid acid adsorbents
Jeffrey Chi-Sheng Wu
a,*, Hsueh-Chang Sung
a, Yu-Fu Lin
a, Shi-Long Lin
baDepartment of Chemical Engineering, National Taiwan Uni6ersity, Taipei, Taiwan
10617, ROC
bNew Material Research and De6elopment Department, China Steel Corporation, Kaohsiung
81233, Taiwan, ROC Received 17 November 1999; received in revised form 30 March 2000; accepted 17 July 2000
Abstract
Aromatic compounds from coal tar generally contain a small amount of tar bases, such as quinoline and isoquinoline. These nitrogen-containing compounds can poison the acid-type catalysts and downgrade the aromatic products because of stinking odor. Four solid acid catalysts, silica-alumina, HY, NH4-mordenite, andg-alumina are
used to remove tar bases by adsorption. Wash oil (WO), refined naphthalene (RN), and an intermediate distillate from the China Steel coke plant (Taiwan) contain quinoline ranged from 0.03 to 8.9%. Quinoline and isoquinoline can be selectively removed from a mixture due to their strong chemisorption on acidic sites, thus the remaining compounds are not disturbed following adsorption. Naphthalene, a neutral compound, is physically adsorbed on solid acids, and is desorbed near its boiling point. Silica – alumina gives the best adsorption results because its wide-ranging pore sizes are accessible to the bulky quinoline molecule. The adsorption of HY and mordenite are significantly decreased because of the extremely diffusion limitation of quinoline in pore channels. Solid acids can be completely regenerated in air at 500 – 600°C. The adsorption ability of silica – alumina can be completely restored even after three cycles of regeneration. The adsorption rate increases with temperature although the capacity decreases. This work demonstrates that the adsorption through the use of solid acids is an effective method that can be used to reduce the amount of tar bases in coal tar aromatics. © 2000 Elsevier Science B.V. All rights reserved.
Keywords:Quinoline; Coal tar; Adsorption; Silica – alumina; HY zeolite; Mordenite
www.elsevier.com/locate/seppur
1. Introduction
Coal is converted into a variety of solid, liquid, and gaseous products when thermally pyrolyzed or distilled by heating without contact with air.
Liquid or coal tar contains many chemicals, mostly benzene, toluene, xylene, naphthalene, and methylnaphthalene [2]. These chemicals are used as raw materials for producing many fine chemi-cals used in medicines, dyes, and pigments. Coal tar aromatics may still contain undesirable tar bases, even after purification from distillation or crystallization. Tar bases are compounds of nitro-gen-containing aromatics. Quinoline is one of the * Corresponding author. Tel.: + 2-3631944; fax: +
886-2-3623040.
E-mail address:[email protected] (J.C.-S. Wu).
1383-5866/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 3 - 5 8 6 6 ( 0 0 ) 0 0 1 9 8 - 2
major compounds in the tar base family. The odor of quinoline can make coal tar products unacceptable to customers. Furthermore, the oxi-dation of contaminated quinoline in aromatics during storage can downgrade the products. Acid-type catalysts were found to be poisoned if reac-tant containing quinoline or other tar bases [4,6,11]. Removal of quinoline can be difficult because its molecular structure is similar to naph-thalene. A commercial product from the coke plant of China Steel Chemical Co., refined naph-thalene (RN), contains several hundred ppm of quinoline even following two-stages of recrystallization.
Industrial recovery or removal of tar-bases from wash oil (WO) is traditionally carried out by extraction, using mineral acids such as sulfuric acid. The extracted tar-bases in an aqueous solu-tion are then neutralized and recovered from the mixture using liquid – liquid separation. Kawasaki Steel Co. developed a novel process that could recover indole from coal tar using oligomerization technique [10]. However, if the production of coal tar aromatics was on a small scale, using such a process would not be economical because of the substantial equipment investment required. Even if the tar bases are removed only by acid washing without recovery equipment, the direct discharge of used acidic sludge can upset wastewater treat-ment plant because of high BOD and low pH.
Audeh (1979) [1] reported that nitrogen com-pounds of Arab light oil can be effectively re-moved by contacting with HCl adsorbed silica – alumina and X zeolite. The regeneration of adsorbent was considerably complicated using NH3 to remove Cl followed by heat treatment
under Ar. The basic asphaltenes could be selec-tively removed from asphaltene-containing hydro-carbon feed through the adsorption using transition-metal-oxide acids catalysts. Regenera-tion was carried out by steaming and calcinaRegenera-tion in air [7,9,12]. Nitrogen compounds were removed from a hydrotreated shale oil by adsorption on US-Y zeolites [5]. Sakanishi et al. (1995) [13] reported that quinoline was adsorbed on sup-ported alumina sulfate and was recovered using supercritical CO2 from methylnaphthalene oil.
Solid acids such as silica – alumina and zeolites are well known for their acidic properties. Acidity is due to unbalanced charges within their oxide structure. These acids comprise Bro¨nsted and Lewis acid sites, which are used as cracking cata-lysts in the petroleum industry. Both acidic sites offer a high affinity for basic compounds, like pyridine and quinoline [5] Oxide solid acid is oil resistant and thermally stable making it suitable for processing coal tar aromatics. This study fo-cuses on the adsorption of tar bases using solid acids to remove undesired quinoline and iso-quinoline from coal tar aromatics produced by the China Steel coke plant.
2. Experimental
The following four solid acid catalysts were chosen for adsorption, amorphous silica – alumina (Strem chemicals), g-alumina (Strem chemicals), HY zeolite (PQ zeolites B.V., CBV 740), and NH4-mordenite (PQ zeolites B.V., LZ-M8). The
SiO2/Al2O3 ratios of silica – alumina, HY zeolite,
and NH4-mordenite provided by manufacturer
are 7.25, 42, and 10, respectively. The specific surface area of silica – alumina, g-alumina, HY zeolite, and NH4-mordenite were measured by N2
adsorption. These solid acids were thermally acti-vated at 400 – 600°C in air to remove water and hydrocarbon impurity, and then were stored in a desiccant for later use. Three coal tar aromatics, WO, RN, and an intermediate distillate (MNO) were obtained from the coke plant of China Steel Chemical Co., Taiwan. WO is an aromatic mix-ture with boiling points ranging from 200 – 300°C. RN is a commercial product of white crystalline that is purified through the use of two-stage crys-tallization. MNO is one of the streams taken from the middle of the distillations. These coal tar aromatics were used directly without any further treatment.
Equilibrium (saturated) adsorption was per-formed in a covered glass beaker to prevent the evaporation of hydrocarbons. The glass beaker typically contained 1 – 5 g of solid acid with coal tar aromatics. The adsorption of coal tar aromat-ics was performed at temperatures between 25 –
100°C. The adsorption of RN was at 90°C, that is, slightly higher than its melting points. The weight ratio of coal tar aromatics to solid acid ranged from 1 to 20. Equilibrium was attained by checking the composition of aromatics when it did not change. Normally it took 24 h to attain equilibrium. Steady flow adsorption was carried out in a packed-bed column. Approximately 50 g of solid acids were charged in the column, which had an inside diameter of 1.5 and length of 10 cm. The solid acid is extruded pellet with 2-mm di-ameter and 4 – 5 mm in length. The residency time was calculated from the flow rates and the packed-bed volume. The column was maintained at adsorption temperature by heating tape. The desorption of used solid acids was performed in a quartz tube inside an oven as illustrated in Fig. 1. The desorbed components were carried by N2and were collected in an acetone-filled bottle. Solid acid, under N2 flow, was heated from room tem-perature to 200°C within 10 min and stayed at that temperature for 1 h, then acetone was sucked from bottle and fresh acetone was refilled. The same procedure was repeated from 200 to 600 with 100°C step. Therefore, the desorbed compo-nents in acetone can be analyzed by gas chro-matography (GC) at each temperature. Air regeneration of used solid acids was in a furnace
at 500 and 600°C. Solid acids turned black or brown after adsorption. The regeneration was completed when the solid acid reverted to its original white color because the original adsorp-tion capacity was restored. All coal tar aromatics were dissolved and diluted in acetone before GC analysis. A GC equipped with mass spectroscope was used to identify the components initially and another GC equipped with FID was used to measure the concentration of components. An HP-5 capillary column was used to analyze WO and MNO. The analysis of RN used a CP-Wax 52CB capillary column.
The weight loss of adsorbed solid acids in mal desorption was performed employing ther-mogravimetric analysis (TGA). The strength of solid acids was characterized using various indica-tors in benzene solvent. The quantity of acid sites was titrated by base, which contained 0.1 N n-butylamine dissolved in benzene. The color change of the indicators normally took several hours to 2 days during titration.
3. Results and discussion
The amount of acidity of the three solid acids is listed in Table 1. The acidic strength is indicated
Table 1
The specific area and acidity of solid acidsa
SiO2–Al2O3 HY NH4-mordenite g-Alumina
750 Specific area (m2/g) 425 480 200 0.605 0.628 H0= −3.3 (mmole/g) 1.189 – 0.533 – 0.421 – H0= −5.6 H0= −8.2 – – – –
aTitrated by n-butylamine in benzene solvent, indicators, dicinnamalacetone H
0, −3.3; Benzalacetophenone H0, −5.6;
an-thraquinone H0, −8.2.
by Hammett acidity function H0, which was
mea-sured when the indicator changed color. The amount of acidity is given by the amount of n-butylamine titrated. For silica – alumina, the amount of acidity stronger than H0= − 3.3 is
1.189 mmole/g, and that stronger than H0= −
5.6 is 0.421 mmole/g. The low acidity of morden-ite could be caused by incomplete decomposition of NH4-form under our heat treatment condition. The acidity of g-alumina was nearly undetected during measurement. The original compositions of WO, RN and MNO are listed in Table 2. They represent the coal tar aromatics with various com-positions. Quinoline is a weak heterocyclic base with basic ionization constant Ka = 8.9 × 10− 10
[8]. Quinoline and isoquinoline are the major base components in these samples, which ranged from 0.003 (RN) to 8.9% (MNO).
Table 3 lists the equilibrium adsorption results of WO mixed with various solid acids at room temperature for 24 h. Quinoline and isoquinoline are effectively reduced in WO. The remaining components stay at about the same level, indicat-ing that quinoline and isoquinoline are selectively removed by solid acids. Silica – alumina leads the highest adsorption performance. The concentra-tions of quinoline and isoquinoline are reduced from 2.64 to 1.21 and 0.87 to 0.29%, respectively. The quinoline reduction ong-alumina is the low-est one because it contains the least acidity. Gen-erally the uptakes of quinoline and isoquinoline follow the order of the amount of acidity (Table 1). The uptake of quinoline is not only related to the amount of acidity, but also the acidic strength of solid acids. It would expect that HY zeolite should have better adsorption capability because it has the higher acidic strength and largest
spe-cific area (Table 1). However, the internal acidic site is not accessible for quinoline because it is extremely diffusion-limited caused by the penetra-tion of channels in Y zeolite [3]. Fig. 2 presents the WO adsorptions on silica – alumina, HY and mordenite. The concentrations of quinoline and isoquinoline gradually decline on silica – alumina and reach a constant value after 24 h. The con-centrations of quinoline and isoquinoline on HY and mordenite quickly decline to low levels within 1 h and only a small amount more quinoline and isoquinoline can be removed within 20 h. Obvi-ously, quinoline and isoquinoline adsorptions only occur on the outer shell of crystals of HY and mordenite, thus they are saturated in a short period of time. Most acidic sites of silica – alumina are available for quinoline/isoquinoline because silica – alumina contains a wide-range of pore Table 2
Compositions of coal tar aromatics
WO wt.% RN MNO Naphthalene 3.64 99.32 28.81 Quinoline 2.64 0.025 8.92 1.81 0.005 Isoquinoline 0.87 a-Methyl-naphthalene 7.80 0.006 11.30 14.71 b-Methyl-naphthalene 0.034 33.19 – – Biphenyl 7.61 18.07 Acenaphthene – – – 14.11 – Dibenzofuran 7.26 Flourene – – Thianaphthene – 0.574 – Dimethyl-naphthalene – – 6.94 Indole – 0.028 – 23.29 0.008 Othersa 9.03
aOthers include anthracene, phenanthrene, H
2O, etc.
Table 3
The reduction of bases of WO by solid acidsa
WO
Component (wt.%) After adsorption by
Al2O3–SiO2 g-Al2O3 HY zeolite Mordenite
Naphthalene 3.64 3.79 3.01 3.78 3.75 Quinoline 2.64 1.21 1.71 1.45 1.51 0.29 0.32 0.87 0.31 Iso-quinoline 0.38 14.71 b-Methyl-naphthalene 15.64 14.39 15.53 15.36 18.07 Acenaphthalene 18.85 19.71 18.55 18.70 14.26 14.86 14.11 14.40 Dibenzofuran 14.26 Fluorene 7.26 7.44 7.88 7.60 7.36 38.52 38.12 38.38 38.68 38.70 Othersb
aSolid acids pre-calcined at 400°C, 4 ml WO+2 g solid acid, adsorption 24 h at room temperature.
bOthers includea-methyl-naphthalene, biphenyl, thianaphthene, dimethyl-naphthalene, indole, anthracene, phenanthrene, H 2O,
etc.
sizes. It also takes longer to reach adsorption equilibrium. The opening aperture of mordenite is even smaller than that of Y-zeolite, thus its ad-sorption capacity of quinoline/isoquinoline is even lower than that of HY. Air regeneration also indicated that HY and mordenite required a higher temperature and a much longer time than silica – alumina. The limited diffusion is the major hurdle using HY and mordenite. Therefore, ad-sorption experiments focused on silica – alumina.
Table 4 displays the concentration changes of quinoline and isoquinoline on RN in the 15 and 60 min of adsorption at 90°C, using silica – alu-mina. Except quinoline and isoquinoline, the rest of the compounds are almost unchanged after adsorption. The adsorption was performed at the weight ratio of RN to silica – alumina of 5. The solid acid is not saturated because the original concentration of quinoline and isoquinoline are extremely low in RN. Isoquinoline can be com-pletely removed in 15 min, but it takes 60 min to completely remove quinoline. This is attributed to that isoquinoline adsorption is stronger (see later discussion) and, thus, takes less time. Further-more, the concentration of isoquinoline is lower than that of quinoline.
The equilibrium adsorption capacity of silica – alumina was given by isotherms displayed in Fig. 3 at the temperature ranged from 25 to 90°C. The isotherms were derived by varying the weight ratio of MNO/silica – alumina. The amount of
adsorbed quinoline and isoquinoline on silica – alumina was calculated from the difference of initial and final concentrations of MNO in ad-sorption. The adsorption capacity of silica –
alu-mina depends on temperature and
quinoline/isoquinoline concentration. The higher temperature gives a lower adsorption capacity. The uptake increases with an increasing concen-tration at a low temperature. However, at 90°C, the uptake increases only slightly with increasing concentration. Possibly the adsorption at higher
Fig. 2. The concentration changes of WO adsorption on solid acids.
Table 4
The removal of tar bases in RN (5 g RN+1 g silica–alumina)a
Component RN Adsorbed at Adsorded at 60 min 15 min (wt.%) 99.32 Naphthalene 99.26 99.28 0.597 0.582 Thianaphthene 0.574 0.005 0.025 0.000 Quinoline 0.005 Iso-quinoline 0.000 0.000 0.006 0.004 a-Methyl-naphth 0.005 alene 0.034 b-Methyl-napht 0.041 0.032 halene Indole 0.028 0.030 0.023 Othersb 0.008 0.063 0.078 aSilica–alumina pre-calcined at 500°C.
bOthers include dimethyl-naphthalene, phenol, etc.
sorption of WO on silica – alumina. The tempera-ture varies from room temperatempera-ture to 100°C. The adsorption of quinoline can be observed from the decline curve of concentration within the first 2 h. The rate of adsorption increases with increasing temperature. Low temperature is favorable for higher uptake, however, low temperature de-creases the adsorption rate. For practical applica-tions, there will be an optimum temperature for obtaining an economical adsorption capacity while removing quinoline at a reasonable rate.
The TGA result of RN adsorbed silica – alu-mina is displayed in Fig. 5. The weight of RN adsorbed silica – alumina is significantly decreased near 200°C, then it is almost unchanged while the temperature rises and reaches 700°C. The des-orbed components at 200 and 600°C are listed in Table 5. Naphthalene is attributed to the 40% weight loss of RN adsorbed silica – alumina. Most of the naphthalene is desorbed near its boiling point, 218°C. This finding implies that the tion of naphthalene is merely a physical adsorp-tion on solid acid because naphthalene is a neutral compound. Naphthalene is loosely stuck in the pores and surface of silica – alumina during ad-sorption. The substantial amount of desorbed naphthalene contains no quinoline and isoquino-line. Therefore, adsorbed naphthalene on silica –
Fig. 3. Adsorption isotherms of SiO2– Al2O3 with MNO.
Fig. 4. The effect of adsorption temperatures of WO on SiO2– Al2O3.
temperature requires stronger acidic sites (i.e. stronger chemisorption), thus uptake is imposed by the quantity of stronger acidic sites available. These isotherms can be used to estimate the mini-mum amount of silica – alumina required to re-move quinoline/isoquinoline for a given concentration. Beside the equilibrium adsorption capacity of solid acids, the rate of adsorption is also crucial for practical application. The adsorp-tion rate relates to adsorpadsorp-tion temperature. Fig. 4 illustrates the temperature effect of quinoline
ad-Fig. 5. TGA of RN adsorbed SiO2– Al2O3under N2
environ-ment.
0.1282 and 0.0048 to 0.0311%, respectively. Both increments are about five times, which coincide with the RN/solid acid ratio (Table 5). That is, quinoline and isoquinoline will not be desorbed until near 600°C. The concentrations of quino-line and isoquinoquino-line are extremely low thus they will be adsorbed on the most available, that is the strongest, acidic sites. Therefore, un-der this circumstance, strong chemisorption oc-curs. This is the reason why silica – alumina can selectively remove quinoline and isoquinoline from a coal tar mixture even at a very low con-centration. Indole does not appear at 200°C. It is desorbed at 600°C and the concentration is only slightly higher than the original. Indole ad-sorption was found to be weaker than that of quinoline [5], and might be gradually desorbed between 200 and 600°C.
The desorption components of MNO ad-sorbed silica – alumina are displayed in Fig. 6. Most of naphthalene was desorbed below 220°C, which is close to its boiling point (218°C). Quinoline and isoquinoline become the major desorbed components starting at 300°C, that is still higher than their boiling points (238 and 242°C, respectively). The maximum desorp-tion of quinoline occurs at 450°C while the des-Table 5
The components of desorption in RN adsorbed silica– aluminaaunder N
2flow
Components RN SiO2–Al2O3desorption
(wt.%) component at 600°C 200°C 99.32 99.12 Naphthalene 99.37 0.1282 0 Quinoline 0.0249 0 0.0331 Isoquinoline 0.0048 0.006 a-Methyl-naphthalen 0.0035 0.0051 e 0.034 0.0328 0.0373 b-Methyl-naphthalen e 0.574 0.556 Thianaphthene 0.565 0.0277 Indole 0 0.0368 Others 0.0084 0.0287 0.0083
aRN adsorbed silica–alumina was obtained the adsorption
of 5 g RN, 1 g silica–alumina.
Fig. 6. The desorption component of MNO adsorbed SiO2–
Al2O3. alumina should be recovered near 200°C before
air regeneration to prevent excess loss of RN. The major component of 600°C desorption is still naphthalene because it is the major compo-nent of RN. The amount of desorption is very small at 600°C (Fig. 5). Quinoline and isoquino-line are not desorbed at 200 until 600°C. The original and desorbed concentrations of quino-line and isoquinoquino-line change from 0.0249 to
Table 6
The adsorption ability after three regenrations in air at 500 and 600°Ca
Removal quinoline %
Temperature (°C) Number of regeneration Removal isoquinoline (%)
97 500 (60 min) 1 100 2 94 100 100 100 3 96 1 100 500 (120 min) 93 100 2 100 3 100 1 600 (60 min) 100 100 98 2 100 3 100 100 100 600 (120 min) 1 100 100 2 100 100 3 100 a
Ratio of RN/silica–alumina, 5; adsorption at 90°C; adsorption time, 45 min; removal percentage = (original concentration−con-centration after adsorption)/original conconcentration−con-centration×100%.
orption of isoquinoline is required even above 600°C. Compared with the desorption of RN on silica – alumina, quinoline and isoquinoline are gradually desorbed as the temperature rises. The amounts of quinoline and isoquinoline in MNO are substantially higher than those in RN (Table 2). The adsorption of such an amount will occur on the acidic sites with wide ranges of acidic strength. Thus, the wide desorption temperature reflects the chemisorption strength between quin-oline/isoquinoline and silica – alumina. The shape of molecule is also an important factor. The chemisorbed isoquinoline is stronger than chemisorbed quinoline, resulting in a higher des-orption temperature [5].
Table 6 summarizes the adsorption ability of silica – alumina after three regenerations in air, at 500 and 600°C for 1 and 2 h, respectively. The removal of quinoline and isoquinoline in RN is complete or nearly complete. The adsorption ability following three regenerations shows no difference from that of a fresh one, indicating that the acidity is completely restored. Regener-ation below 400°C was not completed because the solid acids were still brown or black in color, even with a longer heating time. Thus, regeneration should be at least 500°C for 1 h.
For industrial applications, a packed-bed is a convenient way to operate in a continuous
pro-cess. Fig. 7 displays the break-through curve of the quinoline and isoquinoline removal of RN in the silica – alumina packed-bed column by flow-through adsorption at 90°C. The residency time is required approximately 1 h (see Table 4). The effluent of RN shows no quinoline and iso-quinoline until after the 140 min operation. Des-orption and regeneration can be thus performed after 140 min, depending on the maximum al-lowance of quinoline in RN product. Unlike a
Fig. 7. The adsorption of RN in packed-bed column at 90°C (residency time, 1 h).
batch process, the capacity of silica – alumina is not yet saturated for adsorbing quinoline of such amounts found in RN. The maximum capacity of silica – alumina cannot be not fully utilized if a continuous process with a limited residency time is required.
4. Conclusion
This study presents a feasible and effective method for removing tar bases such as quinoline from coal tar aromatics. Strong chemisorption of quinoline and isoquinoline occurs so that solid acids can selectively remove undesirable quinoline without disturbing the major compounds of WO, RN, and MNO. Silica – alumina shows the best adsorption results among four solid acids. The pore-diffusion limitation of HY and mordenite markedly downgrades their capacity for removing quinoline and isoquinoline. The regeneration can be achieved by simply burning off the adsorbed compounds. Solid acid catalysts are commonly and widely available from catalyst suppliers. The removal of tar base using solid acids is a simple and economical process needing no major capital investment.
Acknowledgements
The authors would like to thank the China Steel Corporation for financially supporting this research under the project number RE87605.
References
[1] C.A. Audeh, US Patent 4,137,154 (1979).
[2] G.T. Austin, Chreve’s Chemical Progress Industries, fifth ed., MaGraw-Hill, New York, 1984.
[3] A. Corma, V. Forne´s, F. Rey, Quinoline as probe molecule for determination of external bro¨nsted and lewis acidity in zeolites, Zeolites 13 (1993) 56 – 59.
[4] A. Corma, V. Forne´s, J.B. Monto´n, A.V. Orchille´s, Cataltyic cracking of alkanes on large pore, high SiO2/
Al2O3 zeolites in the presence of basic nitrogen
com-pounds. Influence of catalyst structure and composition in the activity and selectivity, Ind. Eng. Chem. Res. 26 (1987) 882 – 886.
[5] J. Ellis, J. Korth, Removal of nitrogen compounds from hydrotreated shale oil by adsorption on zeolite, J. Fuel 73 (10) (1994) 1569 – 1573.
[6] C.M. Fu, A.M. Schaffer, Effect of nitrogen compounds on cracking catalysts, Ind. Eng. Chem. Prod. Res. Dev. 24 (1985) 68 – 75.
[7] K.A. Gould, D.C. Grenoble, L.L. Murrell, W.J.M. Pi-eters, US Patent 4,422,926 (1983).
[8] S.N. Holter, Kirk-Othmer Encyclopedia of Chemical Technology, vol. 19, third ed., Wiley, New York, 1982, pp. 532 – 533.
[9] R.B. Long, F.A. Caruso, US Patent 4,419,219 (1983). [10] A. Matsuura, S. Tanaka, T. Horita, N. Sato, Y.
Na-gaoka, O. Nishimura, A Novel Process for Tar-Base and Indole Separation from Coal Tar, Kawasaki Technical Report No. 23, October 1990, pp. 61 – 64.
[11] G.A. Mills, E.R. Boedeker, A.G. Oblad, Chemical char-acterization of catalyst. I. Poisoning of cracking catalysts by nitrogen compounds and potassium ion, J. Am. Chem. Soc. 72 (1950) 1554.
[12] Murrell L.L., Grenoble D.C., Long R.B., US Patent 4,424,114 (1984).
[13] K. Sakanishi, H. Obata, I. Mochida, T. Sakaki, Removal and recovery of quinoline bases from methylnaphthalene oil in a semicontinuous supercritical CO2separation
ap-paratus with a fixed bed of supoorted aluminum, Ind. Eng. Chem. Res. 34 (1995) 4118 – 4121.
