croatica chemica act a ccacaa (4) 653-670 (1990)

CROATICA CHEMICA ACT A CCACAA 63 (4) 653-670 (1990)

CCA-1968 YU ISSN 0011-1643UDC 547.652.1

Original Scientific Paper

Naphthalene Adsorption on 13X Molecular Sieve

1Centre de Genie et Technologie Alimentaires, U.S.T.L. - CC 23 - Place E. Bataillon,34095 Montpellier Cčdex 5- France .

and

2Chaire de Gčnie des Pro~edes Agro-Alimentaires, INA.P.G. - 16, Rue Claude Bernard,75231 Paris Cčdex 05

Received December 21, 1989

In this paper, naphthalene adsorption on 13X molecular sieve has been in-vestigated. The isotherms and the net heat of adsorption have been determinedin the range between 40 °C and 380 °C. Analysis of the results clearly indicatesthat the Dubinin-Radushkevitch model provides the best fitting equation for datapoints, with a specific limitation due to some steric hindrance effect.

INTRODUCTlON

In a previous work, simultaneous heat and mass transfer between a freely mo-ving object and a gas fluidized bed of adsorbent particles were investigated. Subli-mation of naphthalene spheres into a 13X molecular sieve bed was used as amodelsystem.P In an attempt to give an interpretation of these experimental data, the iso-therms and the adsorption heat of this adsorbent/adsorbate pair have been deter-mined and are report ed in the present paper.

Two well-known gas-solid chromatography techniques, thoroughly describedelsewhere.' " were used in a slightly modified from: the so-called "continuous in-jection" and "pulse injection" methods. With the first, a high concentration of ad-sorbate could be reached between 40 °C and 200 "C, The second method, used athigher temperatures (up to 400 °C), gave access to the region of isotherms corre-sponding to low adsorbed quantity.

Corresponding author

654 R. JOULIE ET AL.

EXPERIMENT AL

Adsorbate: Naphthalene (ClOH8)

Plane molecule, 0.7 nm long and 0.5 nm wide.

Pressure vapour (Pa) versus temperature (K):

10glO ps = 13.575

10gLO ps = 10.538

10gLO ps = 9.561

372B.75([

2690.08([

2210.38([

0-80 "C range"

80 - 218 °C range"

218 - 477 °C range[best fitting equation of datapoints in the literature]

Adsorbent: molecular Sieve - 13X type

The cavit~ area is in the order of 760 m2/g and the corresponding volume of the poresis 0.32 cm3/g.1 .n The product was pulverized and sieved to get a 200-400 flm particle size.

Apparatus

To determine the naphthalene adsorption isotherms on the 13X molecular sieve, a GIR-DEL chromatograph (Model 30) equipped with a flarne ionisation detector was used.

The adsorbent is placed in a pyrex tube having an internal diameter of 3 mm so as toform a 10 cm long short packed bed. Such a column reduces the influence diffusion of, andoperates near the equilibrium condiuons.F

Before each run, the adsorbent was activated in anitrogen flow by heating to 390 °C forabout 14 hours. Then, the column was isolated from the outside medium and weighed to de-termine the mass of pure adsorbent.

Continuous Injection Method

The experimental apparatus is shown in Figure 1. The columns - Cl, C2, C3 - are sweptwith a constant flow rate of nitrogen.

Cl is a thermostatically controlled column, 30 cm long, filled with naphthalene crystals, 0.5to 1 mm in size. This bed thickness offers a large enough exchange area with a negligible pres-sure drop. The naphthalene vapour resulting from the sublimation of crystals is carried awayby nitrogen at a 2 mm/s mean velocity. The partial presure of the vapour, p, is a function oftemperature TI. To keep it at a constant value, column Cl is jacketed and thermostatically COI1-trolled water is circulated around it.

A temperature T2, 10 °C lower than TI, brings about a partial condensation of the na-phthalene and fixes the saturated vapour pressure at the end of column C2.

The gaseous mixture then enters the third column, C3, containing the adsorbent. The tem-perature of this column is maintained at a constant value T (T> T2) in the chromatograph fur-nace. The gaseous flow rate is measured under outside conditions (pressure and temperature)with a soap film burette. Due to the law pressure drop, the total pressure in C3 only slightlydiffers from the atmospheric pressure.

The response of the flame ionisation detector was found to be linear up to a partial pres-sure of 400 Pa naphthalene vapour.

NAPHTHALENE ADSORPTION

---------=---- ---- ---.--- ':::'--

.. :

---- -= ---------.: ---- - ---T2

T

Adsorbent

Naphthalene

Manometer

C~ Sublimation column

C2 Condensation column

C3 Adsorption column

T~, T2, T Temperature of C~, C2 and C3columns, measured by thermocouples

F.I.D. Flame ionisation detector

Figure 1. Diagram of the continuous injection system

655

Experiments proceed as follows: when the condensation and sublimation columns (Cl andC2) are at their respective equilibrium temperatures, C3 is placed into the chromatograph. LetTA be the initial value of T. Adsorption takes place until saturation of the adsorbent is reached.This state corresponds to the plateau region of the curve in Figure 2. Then C3 is weighed againand replaced into the system. Return to the equilibrium level occurs a few minutes. The valueof ml, the mass of naphthalene that has been adsorbed, is a specific characteristic of the tern-perature and vapour pressure, TA and p. At times tI, tu ts the adsorbent temperature is raisedto present values TB, Te, TD which correspond to the desired isotherms. The areas under thepeaks Al, A2, A3 are proportional to the desorbed quantities. When the final temperature TDis reached and the last equilibrium established, C3 is weighed again; thus, the residual massadsorbed at temperature TD and pressure p is obtained (m4).


TQJL:::l••tn TA~f---------_....JneQJI-

Te

To

Time

.-<rtl Equilibr-iumc:,~ p Late.au\Il

u,~I;ntnLcl'O••tneOLI;U

t·I

Time

Figure 2. Chromatogram aspect in the modified continuous injection technique

Assuming that the sum of the areasa, +A2 +A3, is proportional to the difference ml-m4,the intermediate adsorbed masses ma and m, may be easily calculated by the following equa-tions:

(1)

NAPI-ITHALENE ADSORPTION 657Several runs have thus been carried out at different vapour pressures.

Pulse Injection Method:

The basic principles of the method are straightforward. Practically, the procedure is quitedifficult to perform because it requires injection of a very small quantity of solid compoundinto the column.

The injection of given amount of a gaseaus mixture with a syringe was not very repro-ducible. The injection of a solutian of naphthalene in a solvent was not satisfactory either be-cause the peak associated with the naphthalene was masked by the peak corresponding to thesolvent. For this reason, an injection method of solid naphthalene was chosen.

The classical chromatograph injector is replaced by a "ROSS" injector. This device (Figure3) consists of a pyrex tube passing through a heating element and connected with the adsorbentcolumn inside the chromatograph furnace. Outside the furnace, a part of the tube not sweptby the carrier gas cantains a sliding pyrex rod, one end of which is shaped like a small spatulaand may be loaded with naphthalene. The other end of the rod is fixed to a piece of ferro-magnetic material so that, by means of a magnet, the spatula may be quickly introduced intothe heating element.

Adsorption column

~IIII

Stop-valve

Tight connectian

Manometer

ferromagnetic piece

slide rod

- Heating elementof the injector

Figure 3. Diagram of the injector in the pulse injection method

At the start of the experiments, a given volume of a titrated solutionof naphthalene inhexane is evaporated and naphthalene crystals appear. The slide rod is set into the injectortube. When the pressure of the carrier gas returns to its initial value, naphthalene is quicklyintroduced into the heating part of the injector. A few seconds later, a very small peak appearsdue to the trace of hexane injected with the naphthalene. This time is chosen as the origin.The naphthalene peak appears a short time later.

In order to get consistenly reproducible resu Its, the mass of the injected naphthalene waskept constant for all experiments. This mass equals 0.477 Ilg, corresponding to alO,tl sampleof a 47.7 mg/l standard naphthalene/hexane solutian. Injections were repeated at various de-creased temperatures. As a general rule, an increase in both the retention time and the widthof the peaks was observed.


A preliminary standardisation of the detection system is needed for a correct use of chro-matographic recordings. The sensitivity of the detector was determined by successively injectingincreascd amounts of naphthalene and plotting the area of the peak, Ap, versus the injectedmass mj. In the working range, the curve is linear and fits the classical relationship:

V2

mj = J :;dV=KlAp

VI

(2)

where VI and V2 are, respectively, gas volumes passed through the column at the beginningand the end of elution. Constant K, depends both on the recorder velocity and nitrogen flowrate. The variation of the retention time, tCR, as a function of the injected mass of naphthalene,mi, was thus determined (Figure 4). (It is worth not ing that, in order to relate the measuredretention times, (MR, to the arnbient temperature To and to the mean atmospheric pressurepo, the following correction is carried out:

TTo

(3)

withPl and po being the pressures at the inlet and the outlet of the packed column, respectively).

The adsorption isotherms are deduced from the chromatographic peaks following the well-known principles of elution analysis. According to this rnethod, the adsorbate to adsorbent massratio at T and p is given by:

m*

p

madsorbale J MVR d= -- tpmadsorbenl RTo

(4)

In this equation, VR represents the specific retention volume, i.e. the volume correspondingto the unit of mass of adsorbent when the temperature is T and the mean pressure in thecolumn pm (calculated from the James and Martin tormota."

The correct interpretation of experimental recordings involves classicaIcorrection steps toaccount for both the temperature/pressure of gas flow rate and diffusion phenomena."

RESULTS

Isoterms Obtained by Continuous Injection:In a first series of experiments the isotherms between 40 and 80 "C were ob-

tained. The resuIts clearly show that at any temperature the adsorbed quantities ofnaphthalene are relatively important and largely independent of the partial press ure.

NAPHTHALENE ADSORPTION 659

1\ '

+\+\

\\

+

Operating conditions

Temperatures: - column

\\\

380·C- injector 390·C- detector 390·C

Nitrogen flow rate: 30 cm3/min

\\+\

\+ -,

OL- ~ -L L_ ~~

O 2 3 ~mj (p9)

••O

30

20

10

Figure 4. Corrected retention time versus the injected mass

Then, other measurements were carried out to get isotherms in the 120 °C_ 200"C, Adsorption was less tha n expected, but the shape of the curves remained thesame leading to the same conclusion: almost idependent of the partial pressure (Fig-ure 5). Thus, it is worth noting that little information is available concerning thelow pressure zone on the curves m*(p), which is detrimental to modelling.


1 + T= 40,7 °C-+m~ -----++

/ + - 46,9°C++---- +

+/+/I

+- S6,9"C+

+- ++-----

+_67,O·C/ -+/ ---+

0,1 /++ + -77,O·C+/ +

/ ----+.....---+

+/+/

°

To interpret the adsorption isotherms, we compared the experimental resultsto the following classical models':

Langmuir (localised monolayer adsorption);- B.E.T. (localised adsorption with multimolecular layers);

_ + _ 122,2·C----+-++-+,+/' _ + _ 142,2·C

+' -+-, +- + + _ 162,2·C+,+-+' _+-+ -+ + - 181,7·C'+'+ -+-+-------ti +--+ + - 203,7°C~.,+' -+-+-+1+0++',

';0° 20p (Pa)

Figure 5. Adsorption isotherms in m* vs p coordinates (40 °C/200 "C)


Harkins and Jura (non-localised monolayer adsorption);Dubinin, first and second structural types (adsorption involving the volume

filling of micropores).The examination of these models led us to the conclusion that the Dubinin mo-

dels were those giving the best-fitting curves. Assuming that the ratio W/Wo betweenthe volume W filled und er the relative pressure p!ps and the total volume Wo ofthe micropore is a function of the adsorption free energy: ead = RT Ln (ps!p) asdefined by Polanyi15 (the standard state corresponds to the equilibrium between thecondensed phase and its saturation vapour), and also assuming that W is propor-tional to the mass adsorbed per unit-mass of adsorbent, the Dubinin - Radushke-vitch model leads to the equation:

(5)

As Shown in Figures 6 -A and 6-B, experimental Ln(m*) versus eid curvesare linear.

From a phenomenological point of view, the steep increase of m * at low partialpressures might be explained by the Polanyi - Dubinin hypothesis, namely the in-stantaneous filling of micropores by the adsorbate.

Isothenns Obtained by Pulse Injection:To get valuable information on the low adsorbed quantities, the isotherms were

determined in a range of high temperatures between 330 °C and 380 °C (at lowertemperatures, the retention time quickly increases and the concentration too slowlyfalls to zero).

A reversible adsorption of naphthalene may be observed again. The curvesm*(p) thus obtained are presented in Figure 7. As before, the Dubinin model ap-pears to be the one that best agrees with data points (Figure 8).

Isosteric Heat of Adsorption:According to the Clausius Clapeyron equation, the net isosteric heat of adsor-

ption, Q* st, may be calculated from the slope of Ln(p!ps) versus l/T curves accordingto:

* _ _. [aLn(p!ps)]Q st - Qst - QL - R a(l/T) m" (6)

with Qst, the absolute isosteric heat of adsorption and QL, the talent heat of na-phthalene condensation.


0,5

T (·C)

329,5

339,5

349,0

359,0

369,0

378,0

oo 10 20 30

P x 103 (Pa)

Figure 7. Adsorption isotherms in m" vs p coordinates (300 °C/400 "C)

40

DISCUSSION

It is interesting to find that the maximum adsorbed volume at the lower tem-perature (40.7 "C) is only half of the pore volume of the 13X molecular sieve: 320cm3!kg offered and 153 cm3!kg filled under these conditions. This behaviour can beascribed to steric hidrance within the pores, due to the form and size of the na-phthalene molecule. This hypothesis is in agreement with the results obtained by

1-" n(m*) -15

Naph

thal

ene

0'1 ~

13X

mole

cula

rsi

eve

T(Oe)

<c-,<,<,<,<,

I?'1 '-

~I

0-16

L+

3~9,5

c::: t:339,5

trI0

t:l0

349,0

""-""-<,<,<

.""'-

I> r-'

6359,0

-17L

IJ369,0

03780

I

89

1011

1213

c2-9

22

c..adx10

(Jmol-

)---

Figure

8.Adsorption

isothermsin

Dubinin

coordinates(300

°C/400

"C)


10

8

6<;

122,2 oe:>--- T (OC)1<; 2,2162,2181,7

CD-20

2,0 2,5

T (OC) --e- 378 369 359 339,5 329, 5

Ln(p/pJ

-18

-19

-20

-21

0,8

0,6

0,2

0,1

0,05

®

1,5 1,6 ...

Figure 9. Isosteric curves

666 R. JOUUE ET AL.

Laurent and Bonnetain 17 for the adsorption of organic moleculcs of various sizeby 5 A molecular sieve. These authors clearly demonstrate that the adsorbed volumebecomes smaller tha n the total offered volume if cavities cannot contain more tha n4 or 5 molecules. In our case, the mean maximum number of naphihalene moleculeswhich can pack into asingle cavity is approximately 5 to 8. Consequently, at 40.7°C, the number of naphthalene molecules adsorbed does not exceed 3 to 4. Thus,the steric hindrance hypothesis seems to be realistic.

The hypothesis of the heterogeneous distribution of pores was shown not tohold because the crystalline greating of zeolithes BX is a regular arrangement oftruncated octahedrals, constituting a stable, rigid and regular structural framework."

Similarly, it is worth recalling that adsorbed quantities of benzene 25 times hi-gher than the ones obtained in this work have been reported in literature for similarworking conditions (at 427 °C with 13X molecular sieve"). This result corroboratesthe assumption of a steric hindrace in the pores. Indeed, the two molecules (C6H6and ClOH8)do exhibit similar planar structures; their properties are very similar inmany respects but they notably differ in dimensions, the first with one aromatic nu-cleus and the second with two of them joined together.

As previously mentioned, with the continous injection method the mass of na-phthalene adsorbed cannot be directly obtained from experiments in the range ofvery lowp-values. To circumvent this problem, the isotherms have been calculatedfrom the Dubinin model in that range and then the isosteric curves drawn from theextrapolated curves. The straight-lines thus obtained are shown in Figure 9-A: weonly draw them in a narrow range of the adsorbed amount between 0.4 and 1%.From them, Q;, values have been estimated by the exposed method and reportedin Figure 10-A as a function of the adsorption rate. The heat of adsorption, whichcould be expected to be independent of the mass adsorbed, there appears to slighlyincrease with it. This is probably due to some inaccuracy inherent in the method.A mean value around 540 kl/kg (70 kl/mol) may be roughly proposed for Q:t within20%.

The isotherms obtained by the elution method give much more reliable valuesfor isosteric diagram drawing (Figure 9-B). The correlation coefficients of linearregression straight-lines are all in the 0.990-0.996 range. As it appears from FigurelO-B, the variation of Q;, as a function of m* is negligible and a value of 450kl/kg for the net isosteric heat of adsorption is found. It agrees with the previousvalue (540 kl/kg) within the confidence interval already indicated (20%).

CONCLUSIONS

From this study, three main points become clear:- the heat of adsorption of naphthalene on BX molecular sieve, around 450

kl/kg (60 kl/mol), is of about the same magnitude as those traditionaly associatedwith reversible physical adsorption;

- reversible adsorption occurs even at temperatures as high as 400 °C;- the best fitting curve for data points is obtained with the Dubinin model.However, when compared to the adsorption of benzene under the same working

conditions (temperature, pressure, adsorbent), the adsorption of naphthalene ap-pears to involve much lower quantities of product. In an attempt to explain thisgap, the idea of steric hindrance in the pores has been hinted at. This assumptionagrees well with the existence of an unoccupied pore volume even at fairly low tem-peratures (in Figure 11, where the reduced mass, m*, at saturation conditions, p

500

)

500

":00

200

100


/.,-/ 'Dl.

Qst+ (kJ !!'noi)

Average value ~ 70obtained by the pulse ~injection met~~+

~ ~/+

60

o 50

o 5

60)50

-+-++----+ ---- + ---- +-----

30

® 20

10

OL.... -"L...-__ ~ ~~OO 0,5 1,0

••Figure 10. Isosteric heat of adsorption vs recovery rate

= ps, has been plotted against temperature, the difference from the theoretical massthat could entirely filI the pores, 0,32 kg/kg, is clearly shown). From these remarksit follows that the isotherms may not be entirely adequate to give a complete de-scription of the phenomena occurring when naphthalene vapour and 13X molecularsieve are put into contact, due to distortions induced by the hindrance effect.

Acknowledgements. - The authors acknowledge the kind assistance of CNRS (Centre Na-tional de la Recherche Scientifique / National Center for Scientific Research) in the financialsupport which permitted to bring out this work.


m*o(kg/kg)

-- I I IMaximum capacity of adsorption

-

0,2 - -

\+\+\+\

0,1 - \ -+

\+~

+'-.++

O I I

O 100 200 300 400T(·C)-

Figure 11. Adsorbed masses under saturated vapour pressure vs. temperature

Ap A, A2, A3Kl, K2mj

ml··m4m*

SYMBOLS

Areas of chromatographic peaksConstantsMass injected into the chromatographAdsorbed masses at TA...TDAdsorbed mass per unit mass of adsorbent

MM

M

PpmpspoPlQLQStQ*St

Rtl ... 13

rcatMR

TTA ...TDToTt,T2VRVl,V2WWoEad

NAPHTHALENE ADSORPTION

Value of m: corresponding to filled up microporesMolecular weightPartial pressure of the vapour ,Mean pressure in the adsorption columnSaturation vapour pressureAtmospheric pressureInlet chromatograph pressureLatent heat of condensationAbsolute isosteric heat of adsorptionNet isosteric heat of adsorptionUniversal gas constant (per mol)Intermediate timesCorrected retention timeMeasured retention timeAdsorption temperature (chromatograph furnace)Particular of TOutside temperatureCl and C2 column temperaturesSpecific retention volumeGas volumes passed through the adsorption columnVolume of adsorbate filling the microporeTotal volume of the microporeAdsorption free energy (per mol)

REFERENCES

669

MM.L -1.T-2M.L -1.T-2

M.L -1.T-2

M.L -1.T-2

M.L -1.T-2L2.T-2

L2.T-2

L2.T-2

M.L2.T-2.e-1

TTTeeeeL3

L3L3

L3M.L2.T-2

1. R. J o u I i e, G. M. R i o s, and H. G i b e r t, Drying Techno. 4 1 (1986) 111.2. R. J o il I ie, Ph. D. Thesis - INP Toulouse (1988).3. L. Dem o u r g u e s, Chim. Anal. 45 3 (1963) 103.4. A. B. Li t tle w o o d, Gas Chromatography, Academic Press., New York, London (1962).5. v. P o nec, Z. K n o r, and C. S lavoj, Adsorption on solids, Butterworths, London (1974).6. A. S a i n t Y rie i x, Bul. Soc. Chim. de France (1965) 3407.7. M. Van S w a a y, Gas Chromatography, Butterworths, London (1962).8. S. B r u n aue r s, The adsorption of gases and vapors, Oxford Univ. Press (1945).9. R. C. W e a s t, Handbook of chemistry and physics, 64th Ed. CRC Press inc., Florida, USA (1983).

10. R. H. Per r y, and C. H. C h i I t o n, Chemical Engineer's Handbook, 5'h Ed. Mac Graw Hill, Ko-gakusha (1973.)

11. P. E. E b erI y, 1. Phys. Chem. 65 (1961) 68.12. L. D. B elj a k ova, A. V. K i seI ev, and N. N. Kov a I eva, Bul. Soc. Chim. de France. 1 (1967)

285.13. A. J. Mar t i n, and A. T. Ja m e s, Biochem. 1. 50 (1952) 679.14. E. B e eh t o I d, Discussion (Author's additional comments) in ref. 7 (1962) 49.15. M. P o I a n y, Trans. Faraday Soc. 28 (1932) 316.16. M. M. D u b i n i n, and L. V. Rad il S h k e v i t eh, Dokl. Akad. Nank SSSR, 55 (1947) 33l.17. A. La ure n t, and L. B o n net a i n, C. R Acad. Sc. Paris. 258 (1964) 180.18. D. W. Bre c k, Zeolite molecular sieves, Wiley, Interscience Publication - New York (1974).


SAŽETAK

Adsorpcija nafta lena na molekulskim sitima 13X

R. Ioulie, G. M Rios i H Gibert

Ispitivana je adsorpcija nafta lena na molekulskim sitima 13X. Adsorpcijske izoterme i to-pline adsorpcije odrede ne su u području izmedu 40 i 380 "C, Rezultati analize jasno pokazujuda Dubinin-Radushkevitsch-ev model daje jednadžbu koja se najbolje slaže s eksperimental-nim rezultatom, uz specifično ograničenje zbog steričkih smetnji.

croatica chemica act a ccacaa (4) 653-670 (1990)

Documents