Gas Flow Through Porous Barriers

8/6/2019 Gas Flow Through Porous Barriers

1/35

eScholarship provides open access, scholarly publishing

services to the University of California and delivers a dynamic

research platform to scholars worldwide.

Lawrence Berkeley National Laboratory


Title:

GAS FLOW THROUGH POROUS BARRiERS

Author:

Jacobson, N.S.

Publication Date:

05-19-2011

Publication Info:


Permalink:

http://escholarship.org/uc/item/77t6t8c8

Local Identifier:

LBNL Paper LBL-8283
http://escholarship.org/uc/item/77t6t8c8http://uc/search?creator=Jacobson,%20N.S.http://escholarship.org/uc/http://escholarship.org/uc/lbnlhttp://escholarship.org/uc/lbnlhttp://escholarship.org/http://escholarship.org/http://escholarship.org/http://escholarship.org/


2/35

yTh 'pywh , weeks.Fo, =all


3/35


4/35

i i

GAS FLOW THROUGH POROUS BARRIERSContents

Abstract . . . . Part I . CO 2 and He through Porous BaO . .

Introduction .Experimental .Results and Discussion .Conclusions . .References . . .

Part I I . 503 Through Porous Alumina.Introduction .ExperimentalResults and Discussion .Conclusions: .Acknowledgment .References . .

iii

11

. 2 5

9.11.12,12.12

. .19,23.25.26


5/35


6/35

i i i

GAS FLOW THROUGH POROUS BARRIERSNathan S. Jacobson

Materials and Molecular Research Division, Lawrence Berkeley Laboratoryand Department of Materials Science and MineralEngineering. University of CaliforniaBerkeley, CaliforniaABSTRACT

In Part I th e possibi l i ty of CO 2 surface diffusion through theporous BaO that resul ts from BaC0 3 decomposition is examined. CO 2 andHe flow rates through a BaO barr ier are compared and both are found toexhibit similar behavior. Because He is known not to undergo surfacediffusion. i t is concluded that CO 2 goes through BaO by ordinary Knudsenflow.

In Part II the decomposition of S03 to S02 and 02 in a porousalumina barrier is studied. The goal is to determine i f this reactionwil l equilibrate in the barr ier . A stream of S03 is run through thebarrier and the exit gas compositions are determined as a function oftemperature with a mass spectrometer . These compositions are found todiffer considerably from the calculated equilibrium values. indicatingthe reaction does not equi l ibrate in the barr ie r .


7/35


8/35

1

Part I , CO 2 and He through Porous BaOINTRODUCTION

The products of a carbonate decomposition are CO 2 and a porousoxide, In a single crystal decomposition, the porous oxide forms anouter layer and the CO 2 must escape through i t . One important questioni s how the CO 2 passes through this porous layer. In the pressure rangeof in te res t th e rat io of th e mean free path of CO 2 to the oxide porediameter is large and hence Knudsen flow is the expected mechanism ofCO 2 transport, However, in a recent theoret ical paper, Searcy and1Beruto suggest th e possibi l i ty that a chemisorbed layer of CO 2 forms(as CO) ions) and that surface diffusion may be the major mechanism ofCO 2 transport through the oxide pores,

Differentia t ing volume flow from surface flow is a problemaddressed by Barrer et al ,2 and others 3, The behavior of a non-adsorbedgas, such as He, is compared to the gas in question. Any additionalflow components are assumed to be a consequence of surface diffusion.Such comparisons had apparently never been made with th e porous oxidesproduced in carbonate decompositions.

In these laborator ies , Roberts 4 studied CO 2 and He diffusionthrough porous CaO. Because the pores of CaO are very small , too smallto resolve with the scanning electron microscope (SEM), leakage aroundthe edges of the barrier was feared to be the primary path for gasescape. 5BaO from the decomposition of BaC0 3 has larger pores, so thatflo\\I through th e barrier \Vas expected to make a larger contributionrelat ive to leaks. Therefore, the flow of He and CO 2 through porous BaOwere compared in this study.


9/35

2

In addition, pore diameters were estimated from these flowexperiments on the assumption that only Knudsen flow was important.Comparison of these estimated diameters with estimates from BET surfacearea measurements and from direct observations of pore dimensions withthe SEM corroborated the assumptions made in the flow calcula t ions.

EXPERIMENTALThe experimental apparatus consisted of a Nuclide model HT-12-60

mass spectrometer connected to a gas in le t system, The gas in le t systemwas essent ial ly the same as that described in Part II of this thesis , butwas constructed of stainless stee l instead of glass and tef lon, A BaC03disk was mounted on the end of a smooth ground alumina tube in theKnudsen cel l chamber of the mass spectrometer. This is shown in Fig, I ,The disks were cut from single crystals of natural barium carbonate(witherite). The impurities, revealed by spectrographic analysis , arel isted in Table 1. The porous BaO barrier was formed in s i by decom-posing these disks of BaC0 3 overnight at about 1200 oK, CO 2 evolved wasmonitored with the mass spectrometer, This in i tu decomposition avoid-ed any problems with hydration of the oxide by exposure to the atmosphere,

The gas in le t apparatus allowed select ion of ei ther He or C0 0 , whichwas used to f i l l a bal las t tank. The pressure was measured with aDatametrics model 1173 capacitance manometer. A typical experiment wasin i t ia ted by f i l l ing a bal las t tank with the desired gas to a pressureof 100 microns. Next, the bal las t tank was opened to the porous barrierand the pressure drop from 100 to 1 0m was recorded as a function of timeon a str ip chart recorder, These experiments were performed a t constanttemperatures, Temperatures were measured with a c h r o m e l ~ a l u m e l


10/35

Thermocouple f o r - - + - . . . j j ~ ~ temperaturemeasurements

Tubewith hi

I IIII

cement

nectionsystem

Fig. I . Porous oxide ba r r i e r mount,

Pt top inner shieldAlumina cap

- Barrier holder- + J - - - + - - Corbonate disk

AI

porous oxi barrierThermocouple forcontrolling furnaceTo heat shi d

no tube

f lange mounted onKnudsen cell vacuum chamber

t

X B L 5810


11/35

4

Table 1. Spectrographic Analysis of Barium Carbonate.

CaSrMgAlCu

,015%1%

< .001%< .001%< .001%


12/35

5

thermocouple located on top of the alumina barr ier holder.An in i t i a l experiment was done a t room temperature using He and an

undecomposed disk. This procedure was assumed to yield the leak rate .Then the carbonate disk was heated to about l200 0 K and decomposed over-night. High temperature flow rates were measured for both CO 2 and He.The leak rate was subtracted from these results . When th e sample hadcooled to room temperature, the He flow rates were again measured.

Pore dimensions were estimated from BET surface area measurementsand were also observed and measured with the SEM. Problems were en-countered with the barium oxide sample hydrating before i t could beexamined. After some experimentation, an aluminum microscope stage wasnotched and th e decomposed disk was se t vertically in the notch.Immediately before placing the disk in the SEM vacuum chamber, i t wasscraped to expose a fresh surface. Insulators generally must be platedwith a conductor before SEM examination, but the BaO could be examinedwithout f i r s t being plated.

RESULTS AND DISCUSSIONThe flux density j of molecules str iking a surface is given by the

Hertz-Knudsen-Langmuir equation 6 P- - - - ~ where P is th e pressure, M/2/TMRTj :=is the molecular weight of the gas, R is the gas constant, and T is thetemperature. The product jA , where A is the area of the part icular sur-face, gives the number of molecules which str ike that surface per unittime: dn PAdt - I2/TMRT However, in this experiment one has a porousbarrier and not a l l molecules which str ike the barrier wil l pass through.So th e area becomes an ' e ffect ive area ' and the number of molecules perunit time passing through the barr ie r is given by: dndt


13/35

6

Now consider the flow of gas from the bal las t tank. This is givendn Vb dPby the time derivative of the ideal gas law: = --- - - where Vb anddt RTb dtTb are th e volume and temperature of the bal las t tank. One now has two

expressions for the number of molecules passing through the barr ier perunit time. and they can be equated: dP Vb PAeffd t RTb "" I2'ITMRT Rearranging andintegrating gives the f inal expression for the effective area: Aeff =V /2 MRT ~ l n P As shown by the previous equations. the effect ive areaRTbis direct ly proportional to the number of molecules which pass throughth e barr ier . Furthermore. i t contains a correction for mass and temp-erature and thus provides a good basis for comparing He and CO 2 flow.

Effective areas are l isted in Table 2. A leak ru n was done with aBaC0 3 disk. The effective area for this run was subtracted from th eeffective areas for each Baa run. yielding th e data in th e table. Theessential feature of these data is that both He and CO 2 exhibit similarbehavior. CO 2 shows no short ci rcui t path via surface diffusion.Because He is known not to undergo surface diffusion. i t can be concludedthat both gases go through the barr ier by Knudsen flow.

The primary experimental problem was the leak rate. The porousbarrier was press f i t . since no adequate high temperature sealent wasfound. However. using an oxide with large pores and subtracting out theleak rate minimized this problem. Another p r o b l e m ~ a s that th e aluminabarrier holder reacted sl ightly with the Baa barrier . This occurred onlyon the edges and appeared to have no effect on flow through the majorportion of the barrier .

Pore diameters were estimated from th e flow experiments. If th eporous barr ier is approximated as a collection of parallel capil lar ies ,


14/35

Table 2. Transmission Properties of BaO from Decomposition.it Material Temp. oK Gas Effective Area Calculated Diameter

cm2 tim1 BaC0 3-1eak ra te 101 He 1.0

BaO 1300 He 1. 1 1.7BaO 1300 1.4 1.4BaO 295 He 2, 8 2.8

2 BaC0 3-1eak ra te 293 He 2. 5BaO 1316 He L8 2.9BaO 1316 1.5 2, 4BaO 299 He 2. 2 3. 6

3 BaC03-1eak rate He 1.1 "- lBaO He 2.3 2. 6BaO 1294 CO2 3. 4 3. 8BaO 295 He 4.4 4. 9

4 BaC0 3-1eak rate 339 He 1.1BaO 1332 He 1.1 1.2BaO 1332 CO2 2.2 2.4BaO 310 He 1.6 1.8


15/35

8

th e effect ive area is given by: /Aeff = n Ap T where n is th e number ofcapi l lar ies , Ap is the cross sect ional area of a capil lary, and { isthe clausing factor with d the capi l lary diameter and t the capi l larylength. Porosity is defined as the ratio of pore volume to to ta l volumeor S "" DApt where A is th e to ta l area of the barr ier ' s end. PuttingAtth is expression for effect ive area gives A d From th isnto the "" As T'f texpression pore diameters ca n be estimated. Values are l isted in Table2.

These pore diameters were compared to pore diameters estimated fromsurface areas. . 8This method, described by Satterf ield, is also basedon the assumption of cylinders in the porous medium, The rat io of cylin-

2V 2 (nTIr 2t)der volume to cylinder surface area gives the pore radius: :s = ~ 2 T I r t ) = r where n i s the number of cylinders, t is th e cylinder length, and ris the cylinder radius. The volume used in the above equation is th evolume of the pores. This can be calculated from porosity and density.

VPorosity is defined as s = P where V is th e pore volume and V i sV +V P spSIthe solid volume, Fo r a one gram sample, V = where p is the density,s pMaking th is substi tution and rearranging gives Vp So the com-1t::plete expression for pore radi i is r = S ( l ~ s ) P

The surface area for a solid with large pores is rather small andBET surface area measurements are not very accurate in this range:

Measured BET Surface Area Calculated Diameter2

il l /gr )Jill.08 ILl,2 5 3.5


16/35

9

However, the l . results are consistant with pore sizes calculated fromflow rates.

An SEM photograph of the pores is shown in Fig. 2. These poresranged from 2 to 10 in diameter, in reasonable agreement with theother data. Because the pores are i rregular in shape and th e threemethods of estimating average pore diameters have different dependenceson pore dimensions, closer agreement is not expected.

CONCLUSIONSThe major result of this study is the demonstration that CO 2 does

not go through BaO by chemisorbed surface diffusion, but rather byordinary Knudsen flow, much in the same way He goes through BaO. Inaddition. the pores of BaO were estimated by each of three independentmethods to average several microns in diameter.

The possibi l i ty s t i l l remains that CO 2 strongly adsorbs on theporous oxide decomposition product. I t is possible that CO 2 forms at ightly held chemisorbed monolayer on th e oxide and that additional CO 2gas, because i t is not strongly bound to the monolayer, goes through thebarr ie r by Knudsen flow. Routh calculations indicate that i f a stronglyadsorbed monolayer is formed there should be a noticable time delaybefore CO 2 introduced from the high pressure side of th e barrier f i r s tescapes a t th e other side, Delay should be especially apparent with anoxide which has very small pores, sich as CaO, Further experiments ,vitha pressure gauge on the exit side of the barr ie r would test thishypothesis.


17/35

-10-

Fig. 2. SEM picture of BaG. XBB781 367


18/35

11REFERENCES

1. A. W. Searcy and D, Beruto, J. Phys. Chem., 82, 163 (1978).2. R. Ash, R. M, Barrer, J, H. Clint, R. J. Dolphin, and C. L. Murray,

R. Soc. (London) Phil. Trans., 275, 255 (1973).3. C. N. Satterfield. Mass Transfer in Heterogeneous Catalysis, MIT

Press, Cambridge, 1970, pp. 4 7 ~ 5 5 . 4. J. A. Roberts, Jr . , unpublished work, Lawrence Berkeley Laboratory

(1977).5. T. K. Basu and A. W. Searcy, J. Chern. Soc., Faraday Trans. I , 72,

1889 (1976).6. S. Dushman and J. M. Lafferty, Scientif ic Foundations of Vacuum

Technique, 2nd edition, John Wiley & Sons, Inc. , New York, 1962.7. O p ~ c i t . , Ash et a l .8. Op. c it . Satterfield.


19/35

12

Part I I . S03 Through Porous AluminaINTRODUCTION

The effect of a porous barr ie r on a reactive gas mixture has beenstudied by Searcy and Mohazzabi. 1 They examined the dimerization ofNaC vapor, a reaction which has a low act ivat ion energy. Their resul tsshowed the reaction readily reached equilibrium in the barr ie r ,

But suppose a reactive gas mixture with a high activation energy ispassed through a porous barr ie r , The question to be answered then be-comes th e extent to which the barrier wil l act as a catalyst and shi f tthe reaction to equilibrium. For this study the decomposition of sulfurtrioxide was chosen: 1S03 = S02 + 2 2 , This reaction has a high act i -vation energy and there is currently considerable interest in the sulfuroxides.

The experiment consisted of passing S03 a t a known in le t pressurethrough a porous alumina barrier and measuring the exi t gas compositionwith a mass spectrometer. The barrier was varied in temperature fromroom temperature to about lOOOoK. S03 to S02 rat ios were measured asa function of temperature and they were compared to the calculatedequilibrium S03 to S02 rat ios, In addition, th e experimental S03 pressure drop was compared to the calculated equil ibrium pressure drop.

EXPERIMENTALThe apparatus consisted of a Nuclide Model HT-12-60 mass

spectrometer and a glass and teflon gas inlet system. A porous aluminabarrier was mounted on the end of a mullite tube and heated from roomtemperature to lOOOoK in the Knudsen cel l chamber of th e mass spectro-meter. This apparatus is shown in Figs. 1 and 2.


20/35

H vacuum TeViton 0 rings

Thermo-gouge

To Her 'Ie To Tod 0pump

. 1. Gas i n l e t system.

pump Gloss tuconnection tomoss spectrometer"7

rethane ~ 1 ?

r i

5

manometer

l Iost tonk

L Ii

iI- 'W!


21/35

Moss spectrometer inlet slit

Furnace -alumino form withKanthal windingson the outside Alumino cap

~ ~ , - - - - t + - - - T h e r m o c o u p l e Porous b a r r i e r + t - - ~ ~ r : f J : 2 ~ for temperaturemeasurement

U ro n ium glass - - - # - -__Pyrex - ~ .....

Fig. 2. Porous barr ier mount.

--Thermocouple forcontroliing furnace

._ - Mullite tube

Springs to securealumina cap

Large flo mountedon Knudsen cell vacuumchamber

- Smaller flange

o jaint- connects to gas inlet lineXBL 5812


22/35

15

The source of S03 was th e vapor above fuming sulfuric acid.vapor is almost entirely S03 and contains negligible water, sincereacts violently with S03' Perry 's Chemical Engineer's Handbook 2

Thiswatergives

th e vapor pressures of S03 above fuming sulfuric acid of differentstrengths, Fuming sulfuric acid is available in three strengths fromMCB; the solut ion 12-18% free S03 was found to give the desired in le tpressures of 50-100 S03'

S03 is a very corrosive gas. In i t i a l ly , the stainless s tee l gasin le t apparatus described in Part I was used. However, when th e bal las ttank was f i l led , a steady pressure of S03 could not be maintained, presumably because S03 continuously adsorbed on the metal. In additionsome S03 was reduced to S02; a lower l imit on the S03 to S02 rat io wasse t a t about 3 with th e mass spectrometer. In order to overcome theseadsorption and reaction problems, an iner t in le t system was constructed,

3somewhat similar to that of Krishan Lal Luthra. The system containedno stopcock grease with which S03 might react . Instead high vacuumstopcocks with a-r ings and teflon bodies were used. These are availablefrom Kontes of California. In addit ion a-ring seals were used for a l lglass to glass connections.

Obtaining sat isfactory a-rings was the major problem in the designof the in le t system. Teflon a-rings wil l not react signif icant ly withdry S03 at room temperature, but they are not compressible enough forhigh vacuum applications. 4The Parker a-Ring Handbook states thatethylene-propylene a-rings are usually sat isfactory for use with S03'however, these were found to deteriorate rapidly in our S03 system.Viton a-r ings seemed to react somewhat less with S03' so they were used


23/35

16

and replaced frequently. I t was also found that i f the inner O ~ r i n g s onthe stopcocks (which were exposed to the highest concentration of S02)were removed. adequate glass to tef lon seals formed.

S03 is readily adsorbed a t room temperature. Therefore. the entiregas in le t system was wrapped with heating tape and kept a t 120C. I twas found that when the bal las t tank was f i l led with S03' a fair lyconstant pressure could be maintained. With this in le t system a lowerl imit of about 5 could be se t on the S03 to S02 ra t io . This result wasa defini te improvement over that obtained with the stainless s tee l s y s ~ tern, but the in le t gas was s t i l l not pure S03'

Selection of a pressure gauge for use with S03 posed a minordiff iculty. Nearly every type of pressure gauge contains exposed partsmade of metals. which would be corroded by th e S03' After some i n v e s t i ~ gation. i t was decided to simply use thermocouple gauges and replacethem frequently, Oddly enough Hastings gauges (which have base metalcases) were found to l a s t longer than Varian gauges (which have stainlesss tee l cases). At the time of writing this thesis , an a l l glass spiralgauge was located. This gauge, which would be ideal for this study,is available from Electronic Space Products, Inc.

Porous alumina barr iers l ike those of Mohazzabi5 were used, Theywere mounted on the end of a smooth ground mullite tube, as shown inFig, 2, Two steps were taken to minimize leakage: 1) A thick walledtube was used to lenghtenthe path through which gas molecules mustt ravel to escape bet\veen the porous barr ier and the supporting mulli tesurface. 2) The cap which held the barr ier on the tube was springloaded to maintain a t ight f i t ,


24/35

17

The end of the tube was positioned in the center of a four inchlong furnace. The furnace was arranged to bring the molecules totemperature before they entered the barr ier and to be certain thebarr ier was in the hottest part of the furnace. However, this put theexit side of the barrier several inches from the mass spectrometer in le ts l i t . This si tuat ion probably contributed to the large background.

This remark concerns the major experimental problem in this study.The information of in te res t is the flux of S03 and S02 that leaves thebarr ie r and travels through a s l i t to the mass spectrometer ionizationchamber. Immediately behind the s l i t is a shut ter . This is a movablemetal plate , which can interrupt the molecular beam from the barr ie r .Molecules which go around th e plate are background. However, the differ-ence between th e unshuttered and shuttered beam is the S03 and S02 fluxwhich leaves the barrier . Generally this should be greater than 50% forreliable measurements. However, for S02 th e difference was sometimesas low as 5%. This indicates that a large background pressure of S02accumulated in our vacuum chamber.

Fortunately, background S03 readily condensed on th e walls of thevacuum chamber. At room temperature the difference between the unshut-tered and the shuttered S03 beam was about 85%. When th e walls werewarmer, the S03 condensed less and the difference between the unshutteredand shuttered beam was about 45%. These differences are large enough togive rel iable S03 measurements.

The problem then is th e large background pressure of S02' Minormodifications to reduce this pressure were no t successful. A cap with asmall opening was placed over th e porous barr ie r . I t was thought this


25/35

18

would direct the beam into the s l i t , but i t made no signif icantdifference. A large tantalum heat shie ld was placed around the barrier .I t was hoped that excess S02 might be pumped off by forming tantalumoxides and sulf ides, but this did not occur,

The best solution to the S02 background problem would be toincrease the pumping speed between the barrier and the shut ter . Thismight be accomplished by building a vacuum chamber with walls which arealways a t l iquid nitrogen temperatures, despite th e temperature of thebarr ier . This approach would require major equipment modificat ions,However, without these modifications, S02 measurements are of question-able re l iabi l i ty .

A typical experiment consisted of carefully pumping down and bakingout both the gas in le t system and the mass spectrometer. The gas in le tsystem was maintained a t 120C throughout the experiment, Approximately5 ml of fuming sulfur ic acid were added to a tube, which was connectedto th e gas in le t system. The a ir above the tube was slowly pumped offand the ent i re system was f i l led with S03 vapor, The valve to thefuming sulfur ic acid tube was l e f t open throughout the experiment. Thisarrangement provided a constant pressure of between 50 and 100 forseveral hours. Temperature was measured with a c h r o m e 1 ~ a l u m e l thermo-couple mounted on the cap which held the barr ier in place. Measurementswere taken every 50 from room temperature to about 10000K. Thesemeasurements were compared to the calculated equilibrium values todetermine i f equilibrium had been reached.

The equilibrium constant for S03 = S02 + % 2 6 is K1/2PSO PO2

PSO3


26/35

19

log K = 8,8557 - 5 4 ~ 5 . 5 1,21572 log T. Suppose one ha s a known in le t

2 . 23 in lnor PSO - 2KPso + 4KPso PSO - 2KPso "" 0

This polynomial wa solved

2 2 3 2 3

infor PSO using an in le t pressure of PSO =2 370 ~ m , every 10 from 300 0 to 1300 oK. The polynomial root finder MULLERCin the C E f u ~ computer l ibrary was used. In addit ion the equilibriumpressures of S03 and the S03 to S02 rat ios were calculated. These valuesare only approximations. The in le t stream was not 100% S03 and we havenot considered the pressure gradient through th e barr ier , However, weare making a rough comparison to major compositional changes over a widerange of temperatures, so approximate equilibrium values are a l l that arenecessary.

RESULTS AND DISCUSSIONIn i t ia l ly we to compare experimental S03 to S02 rat ios

to calculated equilibrium S03 to 502 ra t ios , In order to obtain theexperimental ra t ios , we needed to correctly in terpre t th e mass spectraof S03 and S02' The dif f icul ty here is that S02+ is formed both as theparent io n of S02 and a fragment ion of S03' There are two approachesto the problem that resul ts . One approach is to establ ish a mass spec-trometer fragmentation pattern for S03' but th is was not possible because we did not have a source of pure S03' Another approach, used byLau, Cubicciotti and Hildenbrand,7 is to turn down the energy of theionizing electrons to 17 so the S03 does not fragment. However, th eresul tant signals were very weak and our instrument had too much noiseto give rel iable readings. Therefore, we decided to use 70 eV ionizing


27/35

20

electrons and simply regard the S03 to 502 ratio as a lower l imit . Thiswas acceptable because we were looking for large changes in th e S03 toSOz ra t ios .

The reaction of S03 and th e alumina barr ier to form A12(S04)3needs to be considered: A1203 (a ) + 3 S03(g) = A12(S04)3 (s) . Atlower temperatures the production of AlZ(S04)3 is quite strongly favored,but S03 pressures were l i t t l e changed on passing through the porousalumina barrier . Thus th e 503 probably reacts with A1 20 3 to form a protect ive coating of A1 2 (S04)3' Above 780 0 K the equilibrium pressure of8S03 from A12(S04)3 decomposition drops well below 100 and the de-composition of S03 can proceed unhindered. Thus the data from 780 0 K tolOOOoK is of primary interest .

Table 1 shows the experimental and calculated equilibrium S03 to802 rat ios for a porous barrier . These data are plotted in Fig. 3. Theessen t ia l feature is that th e experimental rat ios show relat ively l i t t l echange compared to the equilibrium rat ios. Thus i t is quite probablethat the reaction is not equil ibrating in the barrier . Note that th eabsolute 502 peak intensi ty increases and the difference between theunshuttered and shuttered S02 beam becomes small at higher temperatures.This is due to the accumulation of S02 in th e ion chamber, a problemdiscussed previously. Fortunately, our conclusion can be verified byanother approach.

The difference between the unshuttered and shuttered S03 beam was85 to 45% throughout the temperature range and therefore should yieldrel iable measurements. A good test for equilibrium is to see i f theexperimental S03 pressure drops as rapidly as the calculated equilibrium


28/35

Table L SO a . 8 n ~ Thick Porous Bar r i e r .

In le t pressure 70 micronsTOK ISO+ SO3 2Shut te r Shut te r Closed Shut te r Shut te r Closed ' I294 .049 .007 .125 .105 2. 1 6.430 .039 .009 .235 1.4 1.473 .039 .009 .225 .205 1.6 3.1xl02540 .038 .012 .410 .380 .87 4.1xlO581 .043 .017 .370 .330 .65 1.625 .025 .225 .205 1.0 5.9688 .037 .019 .420 .6 0 1.4730 .040 .020 .610 .580 .66 6.8xlO-1

N781 .026 .013 1.020 .970 .26 2.7x10- I- '826 .018 .010 1.230 1.150 .1 0 1.873 .024 .011 1. 080 1.020 .22 6.5x10-292!1 .025 .012 1.100 1.050 .26 3.3xl0- 2968 .023 .013 1.250 1.200 .20 1.7xlO- 2


29/35

-I10

700

Fig. 3. S03 to S02 ra t ios .

~ 2 2 -

" Experimental'----. - - . ~ ~ - - - - ~

800 T (K) 900 1000XBL 789- 13


30/35

23

S03 pressure on temperature. Pressure is related to mass9spectrometer intensi ty by P kIT, where P is the pressure, k is the

machine constant, I is the intensi ty, and T is the temperature. Figure4 shows a plot of IT vs. T and a plot of calculated equilibrium pressurevs. T, The experimental S03 pressure changes only sl ight ly as comparedto the equilibrium S03 pressure. This verif ies the conclusion reachedfrom S03 to SO rat ios ~ ~ th e reaction does not equlibrate in a porous2alumina barr ier .

CONCLUSIONSI t has been shown that the decomposition of S03 does not

equi l ibrate in a porous alumina bar r ier . This is expected, since aluminais generally not a catalyt ical ly active materia l .

Before continuing this study, i t would be best to modify theapparatus so more rel iable 802 measurements could be made. These m o d i ~ fications involve increasing the pumping speed between the porous barrierand th e mass spectrometer in le t s l i t . The next step would be to examinecatalytically active porous bar r iers , such as FeZ03 or Pt. These barriersshould push the reaction closer to equilibrium.


31/35

Pso :3(atm)

-510 "' .......Experimental

Fig. 4. S03 pressure drop.

-24-

T (K)XBL - 5814


32/35

25

ACKNOWLEDGMENTI am very grateful to Professor Alan W. Searcy for his guidance and

kind encouragement throughout this study. A very special ' thank you'is due to Dr. James A. Roberts, J r . for instruction in the use of themass spectrometer and his frequent advice throughout this study. Dr.David Meschi also provided much useful advice. Professor Searcy'slaboratories have a part icularly friendly and pleasant atmosphere towork in - a l l of the members of his group are to be thanked for this .

The technical s taff at LBL provided invaluable aid, Thanks arepart icularly due to Emory Kozak and the LBL glass and ceramic shops, Iam grateful to Gay Brazil for typing this material.

This work was supported by the Division of Materials Sciences. Officeof Basic Energy Sciences, U,S. Department of Energy,


33/35

26

REFERENCESL Po Mohazzabi and A, W Searcy, J. Chern. Phys" 5037 (1976).2. J . H. P e r r y ~ R, H. Perry, C, H, Chilton and S, D, Kirkpatrick, eds.

Fourth edit ion, McGraw Hill ,New York, 1963, p, 3-26,

3, Krishan La1 Luthra, Ph,D, Thesis, University of Pennsylvania,4, Parker Seal Co., Parker O-Ring Handbook, 1977, p , B26.5. Ope cit" Mohazzabi.6. R. D, Pehlke, Unit Processes of Extractive Metallurgy, American

Elsevier, New York, 1973, p, 288,7. K. H, Lau, D. Cubicciott i , and D. L, Hildenbrand, J , Chern, Phys.,

4532 (1977),8, K. H, Stern and E, L, Weise, National Bureau of Standards Report

NSRDA-NBS 7, U.S. Government Printing Office, Washington, DC. 1966,9. R. T, Grimley, in The Characterization of High Temperature Vapors,

John Wiley & Sons, New York, 1967, p. 225,


34/35


35/35

Gas Flow Through Porous Barriers

Documents

Transcript of Gas Flow Through Porous Barriers