n-dodecane mixtures on modified X- and Y-zeolites

n-dodecane mixtures on modified X- and Y-zeolites

Adsorption Studies of n-Olefin/n-Paraffin Mixtures on X- and Y-Zeolites II. Adsorption of Tetradecene-1/n-Dodecane Mixtures on Modified X- and Y-Ze...

459KB Sizes 2 Downloads 27 Views

Adsorption Studies of

n-Olefin/n-Paraffin

Mixtures on X- and Y-Zeolites

II. Adsorption of Tetradecene-1/n-Dodecane Mixtures on Modified X- and Y-Zeolites

H. H E R D E N , W-D. EINICKE, U. MESSOW, K. Q U I T Z S C H , AND R. S C H ( ) L L N E R Department of Chemistry, Karl Marx University Leipzig, 7010 Leipzig, Liebigstr. 18, German Democratic Republic

Received February 2, 1983; accepted July 26, 1983 The adsorption of tetradecene-l/n-dodecane mixtures from liquid solution on X- and Y-zeoliteswas studied. The obtained surface excess isotherms and the heats of immersion depend very strongly on the nature of the cations contained (Li, Na, K, Rb, Cs, and the divalent cations Ca, Sr, and Ba) and the degree of the ion exchange. The influence of the silicon/aluminum ratio on the shape of the excess isotherms and the heat of immersion was also studied. It was shown that the separation factors, obtained by evaluating the excessisotherms, and the heat of immersion can be interpreted in terms of the cation distribution and the charge to radius ratio. INTRODUCTION

other important raw materials for the chemical industry. In our recent papers (5, 6) we have shown that the equilibrium properties of olefin/paraffin mixtures depend on the nature of compensation cations in the zeolites and the silicon/aluminum ratio. For a better understanding of this p h e n o m e n o n we report here on the systematic studies of the adsorption of tetradecene- 1/n-dodecane mixtures on a large n u m b e r of cation-exchanged zeolites with various methods.

The adsorption from solution on A-coal, silica gel, zeolites, and other solids is of great importance in the chemical industry. The use of adsorption processes and in the purification of food and drugs should be mentioned in this context. Much attention is paid to such physicochemical problems in high-pressure liquid chromatography. In recent years the zeolites have been developed as very important solids in the selective separation by means of adsorption from gases and also from liquid phase mixtures. The investigation of the equilibrium data and the molecular transport p h e n o m e n a is therefore of great theoretical and practical interest (1). The typical molecular sieve effects of the zeolites studied from the gas phase also occurred in the adsorption from solution (2). Is is well known that the U.O.P. (3) and Asahi (4) processes are employed for the adsorption from the liquid phase on zeolites in order to produce n-paraffins, p-xylene, n-olefins, and

EXPERIMENTAL The zeolites used were prepared by means of the ion exchange of N a X and N a Y (VEB Chemiekombinat Bitterfeld, Wolfen, G D R ) from aqueous solution. Na0.alKo.59X (1.4) is a sample of an X-zeolite, where 59% of the sodium ions have been exchanged by potassium ions and the silicon/aluminum ratio is 1.4. The zeolites were activated for 20 hr at 673°K and a pressure of less than 10 m P a in 565 0021-9797/84 $3.00

Journal of Colloid and Interface Science, Vol. 97, No, 2, February 1984

Copyright © 1984 by Academic Press, Inc. All rights of reproduction in any form reserved.

566

HERDEN~ET AL.

a special tube (diameter 25 m m , height 10 m m ) under so-called "deep-bed" conditions. Three m i lli liters of a tetradecene- 1/n-dodecane mixture ofweU-known composition was added to about 1 g of activated zeolite. The system was stirred at 293°K and after 48 hr the equilibrium was reached. The composition of the bulk phase was analyzed by means of chrom a t o g r a p h y a n d r e f r a c t o m e t r y . T h e last method is only applicable if no chemical reaction (i.e., double bond isomerization) occurs. In the case of zeolites containing Na, K, Rb, Cs, and Ba we could show that these conditions exist. For Li, Ca, and Sr-containing zeolites the adsorption data are related to the sum of the olefins. We used the surface excess according to the following definition: no n~(") = - - (x ° - x]), m

[1]

where n o is the total n u m b e r of moles in the original solution brought into contact with m grams of adsorbent, and x ° and X~l are the mole fraction of c o m p o n e n t 1 (olefin) in the bulk phase before and after adsorption, respectively. From the excess isotherms we calculated the separation factors (5) according to

R T fx~=l n~(n) d(x]~/]). In K = nS---~ axe=0 xlzx]3,]

[2]

As shown by us (5) for the used systems, the activity coefficients of the bulk phase (3']) and also of the adsorbed phase are unity. Therefore the separation factor and the equilibrium constant are equal From the plot ofx]x~/n~ (n) vs x] we obtained the limiting adsorption as shown in (7). For the capacity ratio/3 we used the value 1.15. The individual isotherms can be calculated according to ns =

S 1 nml/3Kxl

1 + (/3K- 1 ) x ]

[3]

In order to determine the limiting adsorption values independently from a model isotherm equation we used a pycnometric method described by Dubinin et al. (8): Journal of Colloid and Interface Science, Vol. 97, No. 2, February 1984

n~ o = m ; 1 {ma - [Vo + Irr2h - mslps]pa

- ~rr2(H- h)pv},

[4]

where ms = mass of the zeolites. m~ = mass of the adsorbate v0 = reference m a r k on the capillary r = radius of the capillary h = height of the meniscus of the liquid adsorptive ps = density of the zeolite crystals together with the micropores p~ = density of the liquid adsorptive pv = density of the adsorptive vapor H = total length of the capillary. The lattice constants a0 of the dehydrated zeolites were taken from the literature or determined experimentally by an X-ray method. The change of the lattice constant during the adsorption of the hydrocarbons can be neglected. The heat of immersion oftetradecene-1 and n-dodecane were measured by means of an LKB 2107 sorption microcalorimeter at 303°K. About 120 mg of zeolite were weighted in the Batch cell and activated outside the calorimeter for 20 hr at 673°K and a pressure of less than 10 mPa. The Batch sorption vessel (18) comprises a glass tube with an upper and lower chamber. The adsorbent was added to the lower chamber and after evacuation it was closed by screwing down the pin to the o-ring seals. After the sample was heated in the external oven, the closed Batch vessel was introduced into the calorimeter. Here the immersion was started by unscrewing the pin. Using the known weight of the adsorbent and the results of the calibration run, the heat of immersion weight per gram was calculated. Amounts of the heat of vaporization as well as those of enthalpy effects that concern wetting the glass tube by investigated liquids have reached one-tenth of the total heat of immersion of zeolites. The olefins and paraffins from Fluka AG, Buchs (Switzerland) were dried with 3-]k molecular sieve and distilled very carefully before use.

ADSORPTION. OF

n-OLEFIN/n-PARAFFIN M I X T U R E S O N Z E O L I T E S , II

RESULTS AND DISCUSSION

In the whole concentration range the tetradecene-1 is the preferentially adsorbed component which is due to the specific interaction energy of the olefin with the zeolites. The excess isotherms are all of type II of the Schay classification (9). The obtained values of the separation factors, the limiting adsorption (nsl, n~l SD ) and the heat of immersion (Aih) are summarized in Table I. The limiting adsorption determined according to the absolute method of Dubinin is about 10% larger than the appropriated value c a l c u l a t e d from isotherm equation due to the assumptions made. In Figs. 1, 2 the typical shapes of the excess isotherms on some zeolites are demonstrated.

567

The individual isotherms on NaX (1.4) and K X (1.4) at 293°K are shown in Fig. 3, which are in agreement with the shapes of isotherms of hydrocarbons on zeolites in the gas phase (1). A systematic analysis of the excess isotherms in dependence of the silicon to aluminum ratio of the zeolites makes it possible to draw some interesting conclusions with respect to the nature of the active sites of adsorption. In Fig. 4 the dependence of the separation factors and the heats of immersion at a number of cations per unit cell are shown. As known from X-ray single crystal data by Hseu (10), the cation distribution for dehydrated NaX and NaY are: NaX-on S I 3.8 Na, 32.3 on two S I'-positions, 30.8 on S II and 7.9 on S III. NaY-on S I 9.3 Na, on S I'

TABLE I Results O b t a i n e d for the L i q u i d Phase A d s o r p t i o n of Tetradecene- I a n d n - D o d e c a n e o n Various F o r m s o f Zeolites Limiting adsorption (mmole/g)

Heat of immersion (J/g)

Zeolite

Separation factor

r~,

n~s'~

Cl4

Cl2

NaX(1.4) Nao.t7Lio.a3X( 1.4) Nao.9oKo.loX( 1.4) Nao.a3Ko.17X(1.4) Nao.74Ko.26X( 1.4) Nao.67Ko.33X(1.4) Nao.41Ko.s9X(1.4) KX(1.4)

102 91 100 100 91 87 82 82

1.05 1.09 1.05 1.04 0.90 0.80 0.73 0.73

1.076 1.108 -----0.776

115.3 118.8 114.5 108.6 104.2 103.9 96.5 98.9

95.8 91.5 -96.3 92.9 -81.5 79.9

Nao.35Rbo.tsX( 1.4) Nao.50Cso.5oX( 1.4)

72 65

0.67 0.63

---

84.5 72.9

68.5 49.3

Nao.s0Cao.2oX( 1.4) Nao.41Cao.59X( 1.4) CaX( 1.4) SrX( 1.4)

135 458 467 382

1.03 1.05 1.00 0.93

-----

108.5 -98.2 119.5

83.4 77.7 68.1 86.7

Nao.9oBao.loX(1.4) Nao.soBao.2oX(1.4) Nao.tsBao.32X(1.4) Nao.41Bao.s9X(1.4) BaX(1.4)

170 245 285 330 333

0.97 0.89 0.88 0.86 0.83

------

-103.6 102.4 102.5 100.5

-94.7 85.0 84.1 83.1

NaX(1.19) NAY(2.08) NAY(2.55) NAY(2.8) KY(2.08)

110 80 35 21 81

1.08 0.97 0.94 0.93 0.73

--1.057 ---

-103.6 95.7 90.3 91.1

103.6 80.3 76.1 72.8 76.3

Lattice parameter ao (/~)

Re£

24.920 24.714

(10)

25.076

(14)

24.790

(I0)

Journal of Colloid and Interface Science, Vol. 97, No. 2, February 1984

568

HERDEN ET AL. 1.4

1.2 G(n) nl 1.0 mmole g

0.8 /

~ ""o~ o

0.6 ~

0.4

~

o

~"



0.2 i 0.1

. 0.2

0.3

0.4

0.5 e X1

0.6

0.7

0.8

0.9

i 1.0

FIG. 1. The excess isotherms of tetradecene-1/n-dodecane mixtures for (O) NaX (1.4), (0) KX (1.4), (v) Nao.5oCso.5oX (1.4), and (A) Nao.35Rbo.65X (1.4) at 293°K.

1.4 1.2 G(n) nl 1.0 mmole



0.8

0.6

0.4

0,2

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

I

1.0

l X1 FIG. 2. The excess isotherms of zetradecene-l/n-dodecane mixtures for (0) CaX (1.4); :(O) SrX (1.4), and (A) BaX (1.4) at 293°K. Journal of Colloid and Interface Science. Vol.97, No. 2, February 1984

ADSORPTION OF n-OLEFIN/n-PARAFFIN MIXTURES ON ZEOLITES, II

569

1"114

mmole

0.1

g



¢

//'

0.2

0.1

I

0.1

J

0.2

I

0.3

i

0.4

I

I

0.5

0.6

i

0.7

I

O~

I

0.9

I

1,0

e X1

FIG. 3. The individual isotherms for tetradecene-1 on (O) NaX (1.4), and (e) KX (1.4) at 293°K. The solid lines are calculated accordingto Eq. [3].

16.7 Na and 31.2 Na in two S II-positions, S Ill are not occupied. The difference between N a X and NaY appears to be related to the S III occupation. Because of their-size the admolecules (critical dimension 4.9 ~) the cations on S I and S I' should not act as adsorption centers. If we assume that the cations in S II and S III have a strong influence on the adsorption properties, a decrease of S III should be connected with the decrease of our data, The fraction of S Ill-type cations is S III/S II + S III = 20.4% in good agreement with the decrease of the heat of immersion (17.0% for

tetradecene-1, 20.1% for dodecane) and the separation factors (25.5%). The different shape of the curve for the separation factors can be due to the influence of the entropic terms as shown in (22). Similarly results were obtained by Schrllner (19) from the gas phase adsorption measurements of n-butenes on the same zeolites. Now let us consider the behaviour of the tetradecene-1/n-dodecane solution and the heats of immersion onto zeolites which are exchanged by alkaline ions. In Fig. 5 is demonstrated the dependence of these values vs Journal of Colloid and Interface Science, Vol. 97, No. 2. February 1984

570

HERDEN

E T AL.

130

In K

120 - - A i h

110

100

90 f 80

70



40

50

,

|

60 cations

70 per unit

!

80

90

cell

FIG. 4. The dependenceof the separation factors and the heats of immersion of the number of cations per unit cell.

the formal charge to radius of the ions e/r. The first observation is that the NaLiX zeolite has a smaller value than expected. This is in agreement with a lot of other investigations (11) concerning the studies of adsorption on such zeolites. Jaroniec et al. (20) made a detailed theoretical analysis of the excess isotherms on the basis of a heterogeneous model of the adsorption from solution. For the Li-containing sample he found a significant change of the heterogeneity parameter and the energetic parameter. The location of the Liions in synthetic zeolites X and Y of dehydrated LiX and LiY are made by Herden et al. (12). The reason for the unexpected behavior should be that the small Li-ions distorted the TO4-tetrahedrons and therefore cannot interact directly with the adsorbed molecules as the other cations in position S II. Furthermore we have residual sodium ions in the large cage and also hydroxyl groups which are formed during the ion-exchange Journal of Colloid and Interface Science, Vol. 97, No. 2, February 1984

process. From NaX to CsX we found a nearly linear relationship of the separation factors with e/r. The cation distribution of the mixed ionic forms NaRbX and NaCsX are unknown but by means of the results and the fact that the S I' position can be occupied by Cs-ions (21) we suggest that the sodium ions should be in position S I and S I', e.g., all sites in the large cage S II and S III are occupied by Rband Cs-ions. The same tendency as the separation factors show the limiting adsorption. Analogous results were obtained by adsorption studies of octene from the gas phase (13). Furthermore we studied the influence of a change in the potassium content onto zeolites. The dependence of the separation factors and the heat of immersion of the potassium content are given in Fig. 6. From NaX (1.4) to Na0.83Ko.17 (1.4) we found a plateau. With a larger degree of ion exchange, the values decreased linearly and from 50% up to 100% we found another plateau. It seems that initially

ADSORPTION OF n-OLEFIN/n-PARAFFIN

571

M I X T U R E S O N ZEOL1TES, II

4.6

tetradecene-I

4.5

~

1 120 --Aih

InK 100

dodecane 4.4

80

60

4.3

40 4.2

Cs

Rb

K

0.7

Na

0.8

0.9

1.0

Li

1.1

1.2

1.3

1.4

1 20

1.5

~ [AI'] PiG. 5. The e/r dependence of the separation factors and the heats of immersion.

120

tetradecene-1

4.7

dodecane

In K

--Aih 4.5

4.4

40

4.3

20

~o

2'o

~o

4b

;o

do

7'0

;o

;o

K-content [ equ. %]

FIG. 6. The dependence of the separation factors a n d the heats of immersion o f the potassium content. The solid lines show the experimental and the interrupted line shows the theoretical line after case (i) of

Eq. [5]. Journal of CoUoidand InterfaceScience, Vol. 97, No. 2, February 1984

572

HERDEN T A B L E 11 C a t i o n D i s t r i b u t i o n of N a K X ( 1 . 4 ) Zeolites Zeolite

S I + S I'

S II

Na Na, 1 K Na, 1.7 K Na, 1.7 K

Unlocated

NaX(I.4) Nao.9oKo.loX(l.4) Nao.s3Ko.17X(l.4) Nao.74Ko.26X(l.4)

2.85 1.85 1.15 1.15

Nao.41Ko.59X(l.4)

1.15 Na, 1.7 K

3.2 Na 3.2 Na 3.2 Na 2.3 Na, 0.9 K 3.2 K

KX(I.4) ÷

2.85 K

3.2 K

3.95 3,95 3,95 3,95

Na Na Na Na

2,95 Na, 1K 3,95 K

+ The occupation factors for S I, S I', and S II are taken from Ref. (14) for sample KF 86.5.

the Na-ions in the small cage are replaced by K-ions and thus they cannot act as adsorption sites. Hseu (10) suggested for dehydrated NaX a relative preference of the cations in various sites. For univalent cations the following se-

E T AL.

quence was given S II --~ S I' > S III > S I. The ions with lower ionic potential should therefore occupy S I, in total agreement with our results. Furthermore we found no difference between K X (1.4) and K Y (2.8). The Xray analysis of a series of dehydrated KX- and KY-zeolites (14) showed that the difference is the occupation of S III, that means that the potassium ions in S II should be the most preferred adsorption sites or the difference between both kinds of potassium is very small. Adsorption measurements of n-butenes on similar zeolites from the gas phase (15) showed an analogous behavior. The curve in Fig. 6 can be explained on the basis of three different cases: (i) only the potassium ions on Site S II are the preferred adsorption sites, (ii) the potassium ions in position S III play the dominant role, and (iii) both the cations on S II and S III play the same role. We calculated

6.2

5.E

5.4 In K

5.0

4.6

iO

1

i

20

~

30

i

i

--

I

i

40

50

60

70

80

Me (II) c o n t e n t [equ. %]

FIG. 7. The d e p e n d e n c e o f the separation factors o f the Me(II)-content. Journal of Colloid and Interface Science. Vol. 97, No. 2, February 1984

i 9O

I 100

ADSORPTION OF n-OLEFIN/n-PARAFFIN MIXTURES ON ZEOLITES, II the separation factors and the heat o f immersion o f the mixed ionic forms according to

573

gree. This is in good agreement with results obtained by other m e t h o d s (17). REFERENCES

K(AIh ) = PNaKNa(AIh)Na + PKKK(AIh)K, [5] where K(A~h) is the separation factor (heat o f immersion) o f the mixed ionic form, Ki(Aih)i are the values for the pure N a X and K X , and Pi is the relative fraction o f the s o d i u m and potassium ions inside the large cage. With respect to the assumptions m a d e case (i) shows the best agreement with the experimental points (interrupted line in Fig. 6) which gives us a hint o f the cation distribution, are summarized in Table II. If divalent cations are present in the zeolites, then the m a x i m u m o f the excess isotherms shifted toward a lower concentration o f the olefin in the bulk phase, which is due to the stronger specific interaction energy. Accordingly, the separation factors increased considerably in the sequence BaX < SrX < CaX. The isomerization activity o f these zeolites should be to the f o r m a t i o n o f O H - g r o u p s during the activation process. The dependence o f the separation factors with the ion-exchange degree are d e m o n strated in Fig. 7. A n interpretation o f these curves can be given on the basis o f cation distribution o f the divalent cations in the zeolitic lattice. T h e sample Na0.8oCao.2oX (1.4) shows that this a m o u n t o f calcium has only a small influence. It is well k n o w n f r o m the literature (16) that the calcium ions are first located in S I-position, which allows an octahedron coordination and is therefore very favorable f r o m the energetical point o f view. But in such position the influence o n the adsorption properties is very low. As can be seen from Fig. 7 the b a r i u m ions m u s t be present in the large cage at low cation content (i.e., Na0.90Bao.loX (1.4)), which is demonstrated in the large increase at such ion-exchange de-

1. Breck, D. W., "Zeolite Molecular Sieves." Wiley, New York, 1974. 2. Eltekov, Yu. A., and Kiselev, A. V., "Molecular Sieves," p. 267. Society of the Chemical Industry, London, 1968. 3. Broughton, D. B., Chem. Eng. Prog. lO, 49 (1977). 4. Seko, M., Miyake, T., and Indana, K., Hydrocarbon Process. 59, 133 (1980). 5. Herden, H., Einicke, W-D., Messow, U., Qutzsch, K., and Schrllner, R., Chem. Technik. 34, 189 (1982). 6. Everett, D. H., Trans. Faraday Soc. 60, 1803 (1964). 7. Dubinin, M. M., Fomkin, A. A., Seliverstova, I. I., and Serpinsky, V. V., "Adsorption of Hydrocarbon in Zeolites" (Suppl. Vol.), p. 1. Berlin, 1979. 8. Nagy, L. G., and Schay, G., Magy. Kern. Foly. 66, 31 (1960). 9. Foti, G., Nagy, L. G., and Schay, G., Acta Chim. Acad. Hung. 80 25 (1974). 10. Hseu, T., PhD thesis, University of Washington, 1972. 11. Haflfinger, R., PhD thesis. Karl Marx University, Leipzig, 1980. 12. Herden, H., Einicke, W-D., Schrllner, R., Mortier, W. J., Gellens, L., and Uytterhoeven, J. B., Zeolites 2, 131 (1982). 13. Einicke, W-D., PhD thesis. Karl Marx University, Leipzig, 1982. 14. Mortier, W. J., Bosmans, H. J., and Uytterhoeven, J. B., J. Phys. Chem. 76, 650 (1972). 15. Harlfinger,R., Hoppach, D., Hofmann, H. P., Quitsch, K., and Schrllner, R. Z. Phys. Chem. (Leipzig.) 261, 65 (1980). 16. Angell, C. L., and Schaffer, P. C., J. Phys. Chem. 70, 1413 (1966). 17. Herden, H., Einicke, W-D., Schrllner, R., and Dyer, A., Inorg. Nucl. Chem. 43, 2533 (1981). 18. Harlfinger, R., Hoppach, D., and Hofmann, H. P., Z. Phys. Chem. (Leipzig) 261, 33 (1980). 19. Schrllner, R., Hoppach, D., and Harlfinger, R., Chem. Technik. 32, 138 (1980). 20. Jaroniec, M., Einicke, W-D., Herden, H., and Schrllher, R., Monatsh. Chem., in press. 21. Pluth, J. J., PhD thesis. University of Washington, 1971. 22. Herden, H., Einicke, W-D., Messow, U., Quitzsch, K., and Schrllner, R., J. Colloid Interface Sci., 97, 559 (1984).

Journalof Colloidand InterfaceScience,Vol.97, No. 2, February1984