Adsorption characteristics of n-propanol on silica gel by gas-chromatography

ADSORPTION ON SILICA

CHARACTERISTICS OF n-PROPANOL GEL BY GAS-CHROMATOGRAPHY

M MARTIN, Departamento

de Qulmlca Tecmca,

J CUELLAR

and M A GALAN

Facultad de Clencias.

(Received

for publrcatron

Umversldad 5 October

de Salamanca.

Salamanca,

Spam

1978)

Abstract-A pulse-response chromatographlc method was used to measure a reversible equrhbnum adsorptlon constant and a rate constant for lrreverslble adsorptlon of n-ptopanol on silica gel particles of various sizes, at 1 atm m the 17%250°C range For the first time revetslble and ureverslble adsorptlon was observed usmg chromatographlc techmques after passmg through the column for 48 hr a flow of n-propanol with mtrogen as carrier gas Values of KA and k, were calculated at 175, 200 and 250°C and from these values the heat of adsorptlon for reversible sites and achvatlon energy for ureverslble sites were calculated

When a gaseous pulse contammg an adsorbable component IS passed through a bed of adsorbent partlcles, the retention time, pl. depends on the charactenshcs of the adsorption process as well as upon the gas velocity, the bed void fraction and the bed length If the adsorptlon and desorptlon processes are hnear and rapld with respect to the retention time pI. it IS directly proportlonal to the equllrbnum constant KA and mvolves no mdlvldual rate constants In these condltlons, equations for the retention time are denved by solvmg, in the Laplace domam, the differential mass conservation equation wlthm the adsorbent particles and wlthm the bed of length z EL, IS then related to the first dertvatwe of the concentration c(z, p) m the Laplace domam The procedure IS described m detail by Schneider and Snuth[l] When rapld, reversible adsorption IS accompanied by completely irreversible adsorption Gahin et al 121 found the followmg expression for reversible adsorptlon RA =

$

[k,C,

-

kdn + k,C,]

irreversible adsorption rate parameters by applymg chromatographlc technique m a fixed bed of s&a particles METaoD

Each expenment consisted of injecting a pulse of gas contauung n-propanol (using nitrogen as carrier gas) into the bed and analyzing the measured concentration vs time peak m the bed effluent Values of the first-absolute moment were calculated from the measured efauent peaks by the expression

I* I.L1= O_ I0

(1)

(B

+

iC(z, I) dr (4) C(z, t) dt

where C(z, t) IS the concentration peak at the exit of the bed The data for drfferent velocltles and particle sizes K.., and k, as described were sufficient to determme below The results at ddferent temperatures were used to evaluate the heat and actlvatlon energy of adsorption

where k, IS the first-order, u-reversible rate constant The first absolute moment can be shown to be a function of k,, K,, and of the mass transport rate coefficients EA, km and D,, according to the followmg equation

_

the gel

-AL

PROCEDURE

The apparatus conslsted of a gas chromatograph, with a flame lonrzatlon detector, m which the separation

ppKA,(k$z

(coth ho/h,) - cosech’ ho

where ho

=

R

(t?$ >I” c

column contamed a bed of silica gel The apparatus was completed with four scrubbers contammng n-propanol mmersed m a thermostat The lines between the scrubbers and the packed bed were short and were insulated to avold condensation A flow of mtrogen was bubbled through these scrubbers after which the mtrogen was saturated to 97-98% with n-propanol Experiments were done by bubbling mtrogen through

(3)

and Ka IS the equlhbrmm adsorption constant Equation (2), can be used to evaluate the n-reversible rate constant The present study was undertaken m order to test experimentally eqn (2) and to determme equllbrmm and 691

692

M

MARTIN er al

scrubbers at room temperature, but when necessary, the n-propanol concentratton m the flow of nrtrogen was moddied by changmg the thermostat temperature This change of n-propanol concentration was only done m one experlmental run because eqn (2) 1s based upon a first order rate equation for reversible and Kreversible adsorption To test this assumption first moments were measured at 175°C for mlet pulses contammg 0 6 and 1 mole% n-propanol in nitrogen The 175°C results showed that the first absolute moment was mdependent of the concentration of n-propanol m the inlet pulse This shows that the adsorption isotherm IS a straight lme m this range of concentration As the first moment IS the same m the range of concentration between 0 and 1 mole%, one may work with any concentration m this range This means that even when nitrogen IS not completely saturated with n-propanol, the first moment will be the same, as 1s shown by the experimental work (Fig 3) A flow diagram and details of the apparatus used are gven m Fig 1 The properties of the slhca gel (Hewlett-Packard chromatographuz, 30-60 mesh) obtamed by using an hehum-mercury poroslmeter are gtven m Table 1 and the properties of the columns used are given m Table 2 Pnor to the runs, each bed was pretreated with a flow (20-30cm’/mm) of pure nitrogen at 300°C for 48 hr Prehmmary measurements were made to determine the adsorption characterlstlcs of the system Pulses of 0 25 cm3 (25”C, 1 atm) contammg approximately 1 0 mole% n-propanol m nitrogen were mtroduced mto the column at increasing temperature, starting at 25°C The temperature was increased up to 200°C but no response peak was obtained, this means that some lrreversrble adsorption occurs To saturate these n-reversible sites and following the technique developed by Chm et al [IO], a flow of n-propanol, with nitrogen as

Table 1 PropertIes of sllrca gel parttcles

the

surface area (cm*/& Void volume (cmYg) Partxle density, (g/cm’) Internal void fraction Average pore radms (A)

630 0 30 1 30 0 38 I?

Specific

Data given by Hewlett Packard supplters Table 2 Characterlstlcs

of packed-bed

Column No 1 Packed length (cm) 18 0 94 Column I d (mm) Parhcles 1 mesh Size range 60-70 2 averaee radms (mm) 0 230 Bed voldlfractlon 0 385 Temp “C 175,200,250 Gas flow rate l&150 (cm’lmm 25’C, 1 atm)

carrier that,

gas, peaks

was

passed

through

of reasonable

2 18 0 94

the bed

magnitude

MOMENT

3 18 094

30-60 0 420 0444 175,200,250 IO-50

were

same process of saturation was repeated for every column and temperature FIRST

columns

for 48 hr

After

The before each run obtamed

ANALYSIS

In order to use the first moment m eqn (1) It IS necessary to determine (CL&. the retention trme in the dead volume present m the lme between the pulse mJectlon and entrance to the bed and between the bed exit and the detector The dead volumes were obtamed by mtroducmg 0 25 cm3 pulses of n-propanol m nitrogen and replacing the bed by a short capillary tube at 200°C The first moment 1s gven by the followmg expression deduced by Schneider and Smith [I J

NITROGEN NEEOLE

CYLINDER VALVE

THERMOSTAT MERCURY

WITH

SCRUBBERS

MANOMETER

THERMOMETER SIX-WAYS PACKED FLAME

VALVE BED

AIR

CYLINDER

diagram of the apparatus

WITH

IONIZATION

HYDROGEN

Rg 1 Schematic

80-100 0 163 0 417 175 200,250 10-150

CYLINDER

ADSORBENT DETECTOR

MATERIAI

AdsorptIon

charactenstlcs

(P& can then be obtamed from eqn (5) Data for first moments vs z/v are plotted m Figs 3-5 for 175, 200 and 250°C It can be seen from these figures that first moments are sensltlve to particle size besides, these first moments are not m a straight lme This lmphes the existence of some pure u-reversible adsorptlon together w&h reversible adsorptlon Data must be correlated by eqn (2) with which the theoretical model developed m (2) IS expenmentally checked To separate the contrlbutlon of reversible and u-reversible adsorption the followmg method was tned First of all, the values of KA were obtsuned n-propanol was passed through a column packed with R = 0 40mm parucles for three days to be sure that all ureverslble srtes were saturated After that, chromatographic pulses were Introduced mto the column at 175°C The first moments of the eflluent peaks so obtamed were plotted vs z/u (Ftg 2) These data show a straight hne for very low retention times This means that only reversible adsorptlon exists From the slope of this hne the value for KA can be deduced by the Schneider and Smith method [ l] The K,+ value so obtamed was 1015 cm”/g Other temperatures were not tied because it was very d&cult to saturate all lrreverslble sites and lrreverslble adsorptlon was present m these cases However, Rgs 3-5 show that data for low z/u values are not sensttlve to particle size and It can be deduced

t93

of n-propanol

Fig 3 First moments as a fun&on of retention times for reversable and lrreverstble adsorptlon at 175°C and different particle sues

I

0

I I / 0

0

oc

200 0

cl

I

R042mm

0

R023mm

0

R016mm

Rg 4 First moments as a function of retentron times for reversable and ureverslble adsorptlon at 2CWC and different partxle sues

that the tangents at the ongm for these first moment lmes are the first moments for pure reversible adsorptlon given by the followmg equation Fig

2 First moments as a function of low retenbon reversible adsorption at 175°C

times for

IL*-wf-$=;t

[

1+ +

(P +

PPL)]

(6)

M

694

MARTIN ef al

For low Reynolds and becomes 200

number

-

thus equation

can be slmphfied

$!&)O AB

0)

D IS aven by the well known Hlrschfelder eqztion [4] If It IS assumed that there IS equlmolecular-counter dlffuslon and no surface dtiuslon, the effective mtraparticle dtffuslvlty, D,, can be calculated[5] from the followmg equation 1 1 z De=DK++DAEMf 250

-c

0

R 042mm

0

R 023mm

A

R 016

025

effective

Knudsen

D Kcff= 19,400~

$

dlffuslvlty[6],

mm

050 z/v

where DKe~ IS the given by equation

(9)

app71 J(

075

(II

Fig 5 Fust moments as a functton of retention times for reversable and ureverslble adsorptlon at 250°C and different particle sizes From the slope of these tangents the KA value IS obtamed for each temperature aven m Table 3 It IS remarkable that the KA value obtamed at 175°C was 1064 cm31g, which LSm good agreement Hrlth the K.., value for pure reversible adsorption at the same temperature (1015 cm’/g) Once KA values are known, it IS necessary to calculate the values of the mass transfer coefficient k,,,, the effective mtrapmcle dlffuslvlty 0, and the ax& dlsperslon coefficient (&) The values of the mass transfer coeficlent were ca.lculated by the correlation of Wakao ef nl [3] gven by the followmg equation

where “a” 1s the specific surface Internal tortuoslty factor which following expression

)

of the particle and 7, IS given[7] by the

ti3 7, = i-@

(11)

and DABe~ IS the effective molecular dlffuslvlty calculated [8] from the followmg expresslon

D ABelT=-3AB 7, The axial dispersion coefficient IS calculated from [ 11 E.A=y

(10)

which IS

(12)

for n-propanol

ff9.433

m the bed

(13)

(7) Table 3 Reversible adsorptlon equdlbnum constant Temperature (“0

id/T

PP&

cm&3

(“K-l)

-AH (kcal/mol)

1384 E

1064 661 338

2 23 2 11 191

648

KA

175 200 250

Table 4

Intrapartxle

um2 = 19 10

ddhslon

where 7, IS the external tortuoslty factor, which, as IS known from the hterature[ 11, has a value between 1 3 and 1 7 In the present case, followmg Schneider and Smith[73, a value of 1 43 was taken The different values used for k,, 0, and E, are riven m Table 4 Once all these parameters are known, values of k, can be obtamed from eqn (2) for each temperature Values of &, are qven m Table 5 of n-propanol

on silica gel

Tr= 143

7, = 3 03

k,,, Gxnls)

~.a 175 200 250

0,

(cm%)

(cm2/s)

0 15 0 17 0 21

0 019 0021 0026

R=OO42 (cm) 3 57 4 05 500

EA

(Cm%)

R=O023 (cm)

R =0016 (cm)

a=0444

a=0417

6 52 7 39 9 13

9 37 10 62 13 12

0046 0053 0065

0044 0049 0046

(I = 0 385 0065 0061 0 056

695

AdsorptIon characterlsucs of n-propanol

passmg n-propanol for 2 days through the column after which almost all u-reversible sites are occupied by n -propanol In a simdar way In KA vs l/T, “K, was plotted (FIN 7) From the slope of this other straight lme the heat of reversible adsorption was calculated Its value was, AH = -6 48 kcai/mol This value seems to represent physical adsorption and It IS slmdar to the heat of adsorption of SO, on s&a ge1[9]

Table 5 Irreversible adsorptIon constant values and actrvatmn energy T” (“C)

k, (cm%g)

E

WI

13 16 21

175 200 250

(kcallmol) 3 15

Then, Ink, was plotted vs l/T, “K (Fig 6) From the slope of this &aught lme, the actlvatlon energy for Irreversible adsorption was obtamed, Its value was equal to 3 15 kcal/mol This value seems to be low, but rt IS necessary to take mto account that u-reversible adsorptlon IS found after

CONCLUSIOU Reversible and irreversible adsorption IS demonstrated for the first time for the n-propanol-slhca gel system by the combmed use of a chromatographlc technique and

-O-O-O-

3

I 19

I 21

I 20 103/

T

I 22

I 23

IOK-‘)

Fig 6 Rate constant for irreversible adsorption as a function of temperature

t

11 3

I 19

I

I 21

20 Id/

T

I 22

I 23

(OK-‘1

Fig 7 AdsorptIon equihbnum constant for reversible adsorpuon as a funcuon of temperature

M

6%

MARTIN el al

adsorbed concentration of n-propanot, mole/g total adsorptlon rate, mol/cm3 (s) R part&e radms, cm E time, s to pulse mlection time, s lJ gas velocity in the void space of the bed, cm/s 2 bed length, cm

the momentum theory, after saturatmg the column for 48hr Values for eqtuhbrwm adsorption KA and lrreverslble adsorptlon constant, k,, have been obtamed for different temperatures From these values, heat of adsorptlon for reversible sites and actlvatlon energy for Irreversible adsorptIon were obtamed Acknowledgements-The authors want to express their thanks to Prof J R Alvarez and to Dr Raquel Rodrrguez for theu stlmulatutg

dtscusstons

about the adsorptron

problems

n

RA

Greek s ytn bols bed void fractton mtrapartlcle void frachon retention time of adsorbable

component

m bed,

S

particle density, g/cm3 external and Internal tortuoslty

NOTATEON

concentration of n-propanol m the gas, mole/cm3 concentration of n-propanol m the mtrapartlcle pores, mole/cm3 effective mtrapartlcle ddfuslvlty. cm’/s effective molecular d&Tuslvlty, cm% effectwe Knudsen ddTuslvlty, cm’/s molecular ddTuslvlty, cm*/s Knudsen dtiusmlty, cn?/s actlvatlon energy, kcat/mole axial dlsperslon coefficient, based on crosssectronal area of column heat of adsorption, k&mole adsorption rate constant, cm’/(g) (s) desorptlon rate constant, s-’ rate constant for u-reversible adsorptlon, cm”/(g) (s) fluld to particle mass transport coefficient, cm/s adsorption equthbnum constant for reversible adsorptlon, cm3/g

fractors

JUWERFZNCES

Schneider P and Smith J M , A ZCh E / 1968 14(5) 762 [2] Galan M A, Suzulu M and Smltb J M , Ind Engng Chem Fundls 1975 M(3) 273 131 Wakao N , Oshlma T 0 and Yagl S , Chem Engng Japan [l]

1958 22 780 [4] Huschfelder

J 0 , Curtts C F and Bud R R , Molecular Whey, New York

Theory of Gases and Ltqurds, p 539

1954 [5J Sattertield

C N and Sherwood T K , The Role of ~tffus~on WICatalysts, p 18 Addlslon Wesley Reading, Mass 1%3 [6] SatterIieid C N and Sherwood T K , The Role of D&won m Catalysts, p 17 Ad&son Wesley, Readmg,

Mass 1963 [7J Schneider P and Smith J M , A ZCh E J 1968 1416) [8] Sattertield C N and Sherwood J K , The Role of Drffusron in Catalysrs, p 15 Addison Wesley, Readmg, M&s i%3 [9] Galan M A and Smith J M , J Cat 1975 38 206 [lo] Chm H M , Hashunoto N and Smith 3 M , Znd Engng Chem Fundls 1974 13 282