Mineralogy of Clay

8/13/2019 Mineralogy of Clay

1/13

USE OF CL YS S PETROLEUM CR CKING C T LYSTSB i- T. H . MiLLIKEN ,* A. G. OBLAD,** AND G. A. M ILL S** *

Introduction. The use of catalysts in the petroleumindus t ry in recent years has undergone a remarkableexpansion. The most important of these catalysts areemployed in cracliing processes, first introduced in 1936(Houdry et al . , 1938). As a consequence of the growthof catalyt ic cracking, the manufacture of cracking catalysts has in itself become a major industry with an estimated sales value of $61,000,000 in 1952. Principal emphasis has been placed on two tj^pes of catalysts bothessentially composed of silica and alumina, one derivedfrom clay and the other synthesized from aluminum andsilicate solutions. A total of approximately 470 tons perday of cracking catalysts are manufactured and used toprocess about 2,000,000 barrels of crude daily. Clay catalysts account for about 40 percent of the total manufact u re . Such a special use of clays has placed requirementson these materials significantly different than those metin other uses of clays. Moreover, the very part icularphysical and chemical propert ies required for good commercial hydrocarbon-cracking catalysts l imit the economic use of clays to specific types. Until recently, onlya few deposits of subbentonites were of commercial interest for catalyst manufacture.

As in certain other uses of the surface propert ies ofclays, i t i s the pecul iar s t ructural features of clays whichenables them to play a special role for catalyst manufacture.At present only two clays are used for the preparat ionof commercial cracking catalyst : montmori l loni te andhalloysi te. The patent l i terature ci tes many other natural ly occurring materials as sources of such catalysts;kaol ini te, vermicul i te, and bauxi te, for example. In thispape r, only montm ori l loni te, halloysi te, and ka ol ini te wi l lbe discussed. The general propert ies of cracking cata-* Ass is tant D irector of Researcl i , Hou dry P rocess Corp. , Researcl iand Development Laboratories , Marcus Hook. Penn.** Associa te Mana ger of Researcl i and Development, H oudry Pro cess Corp. , Res earch and Developm ent Laborato ries , M arcusHook, Penn.** Director of Resea rch, Hou dry Pro cess Corp. , Rese arch and Development Laboratories , Marcus Hook, Penn.

lysts and raw clays from which they are manufacturedare shown in table 1, which prese nts compa rat ive datafor chemical , physical , and catalyt ic propert ies.The catalyst is employed in the cracking of heavypetroleum fractions. In such a process, oil, more or lessin the vapor state, is passed over the catalyst at 425-500 C, atmospheric pressure, and contact times of 6-20seconds. Under the directive influence of the catalyst, aseries of complicated reactions (Greensfelder, 1951) takesplace with the consequent produ cts much different fromthose obtained in thermal cracking. Catalyt ic cracking isemployed in the petroleum industry because of the highquality of gasoline produced and because of the desirabledistribut ion of products. Typical ly, some 6 percent of amethane to propane fract ion, 10 percent butanes, 45percent gasoline, and 40 percent recycle oil are produced. At the same time 1 to 3 percent of the chargestock is deposited on the catalyst as a nonvolatile hydro-carbonaceous residue commonly cal led co ke . After a10 to 20 minute cracking period the coke is removedfrom the catalyst by control led combust ion. In thisburning process, the temperature is l imited to perhaps600C, al though a higher local surface temperature mayprevai l . After such a regenerat ion the catalyst i s readyfor reuse. During use at elevated temperatures, the catalyst i s exposed to various organic compounds includingthose containing oxygen, sulfur, nitrogen, as well asheavy metals contained in the charge stock. In addi t ion,the catalyst must wi thstand water over a wide range ofvapor pressures. Whereas early commercial catalyt iccracking apparatuses ut i l ized a fixed bed, present dayuni ts ut i l ize moving beds with t ransfer of the catalystfrom a cracking zone to a regenerat ion zone. (Ardernet al . , 1951; Murphree, 1951). A typical process is i l lus

trated by the flow sheet shown in figure 1. In order tomainta in a he at bal anc e between the endothermicvaporizat ion and cracking of the hydrocarbons and theexothermic regenerat ion, a large quanti ty of catalystmust be ci rculated continuously. Depending upon re-

Chemical\l2O3 - - -M gO - - - - - - . _ _ . - -F e20 3 - - - - - -C aO - - - ->fa20

PhysicalCa ta lyt ic

Table 1. P roperties of new) cracking catalysts after calcination at 55V C .

Synthetic

SiOj-AljOj

87.512.500.10.00.1300-600

45-50

Si02-MgO

66.31.631 . 80.20.2

65045-50

ClayMontmorillonite

R a w

63.724 . 35.52.52.02.0

5

Act iva ted

71 . 822.13.91.50.30.2300

40

Houdrytype 1

76.718.24.40.11.0

23040

R a w

49.149.70.10.60.20.2

12-20

Halloysite

Activated

57.140 .6

1 . 70 . 40 . 00 . 1

160

35 40

Physical and catalytic properties on fresh catalysts after calcination at about 550* C.( 314 )
http://l2o3/http://l2o3/http://l2o3/


2/13

Part VII] CLAY TECHNOLOGY IN THE PETROLEUM INDUSTRY 315

HOUDRIFLOW

FEED

HEATER

EXCESSFLUE GAS

FIGURE 1

finery requirements, 0.5 to 1.5 tons of catal3'st are circulated per barrel of crude charged. This tremendouscirculation requires a high degree of physical stability.In short, the catalyst must maintain cracking activitywhen svibjected to high temperatures, chemical attackfrom steam and contaminants in the charge, and havegood resistance to abrasion.

Cracking Catalysts From Montmorillonite. There isno known chemical or physical test which will enable oneto predict whether or not a particular montmorilloniteclay will respond to acid treatment and produce a catalyst of high activity. Very few montmorillonite claysare known which yield satisfactory catalysts. The situa

tion is similar to that encountered in the bleaching clayindustry where it is also necessary to treat the claysample with acid and test for decolorization activity inorder to evaluate the potentialities of a particular clay.

Three deposits of montmorillonite clay suitable for themanufacture of cracking catalysts have been exploitedon a commercial scale. Two are in Arizona and one inMississippi. Recently the entire supply of such clay hascome first from Chito and then from Chambers, Arizona.Some catalyst has been manufactured from clay minedat Jackson, Mississippi, but this material has proved tobe less desirable than that from the Arizona depositsand its manufacture has been discontinued. Certainother American deposits have been reported by Mills,Holmes, and Cornelius (1950) to be satisfactory for theproduction of cracking catalysts. They found, upon examining a variety of montmoriUonite clays, a wide rangeof susceptibility to activation and that a few claysyielded catalysts of maximum activity. These highlyactive catalj^sts proved to be comparable with syntheticsilica-alumina catalysts. Suehiro (1949) has also reported on the activation of certain clays for the production of cracking catalysts.

Acid Treatment. Acid treatment of bentonites forthe removal of ' ' basic' ' constituents has been studied bya number of investigators. Nutting (1933; 1935; 1937;

8 iO "1 2"R lO a" REMOVAL .GMS/ IOO GMS CLAY

FIGURE 2. Chemical compos ition ofclay remaining after acid treatment (ordinate), calculated on the basis of constantSiO2=100. Solid circles show Mississippiclay, open circles Nevada clay. (AfterMills et al., 1950, p. 1177.)

1943) has studied the changes in chemical compositionbelieved to be related to decolorizing ability. Other authors (Burghardt, 1931; Ilagner, 1939; Schroter, 1940;Hofmann, et al., 1935; Lopez-Gonzalez, et al., 1952;Escard, et al., 1950) have reported the chemical composition of many clays acid treated to give material withoptimum bleaching properties. Figure 2 shows the changein composition of Ash Meadows, Nevada, clay as a function of acid treat ment (Mills, et al., 1950). In this figure"R2O3" represents the "basic" constituents of the rawclay.

In order to simplify the discussion on the catalyticproperties of the treated clays, only the initial activities

0 2 4 6 8 10 12 14

" A " REMOVAL, GMS. /IOO GMS. CLAY

FIGURE 3. Cata lyti c cracki ng activ ity ofacid-treated bentonite as a function of severityof acid treatment. Solid circles show Mississippi clay, open circles Nevad a clay. (AfterMills et al., 1950, p. 1181.)


3/13

316 CLAYS AND CLAY TECHNOLOGY [Bull. 169

300

3 260__o -"^ o

"^220

:IBO^ ^

t;""! / " SURFACE AREA^ 100

o /

S 60

20,

o

60

r.G-^

o,o-o-^

POROSITY

S 30/o

O 20

' : ^'

^ 16"OR. PELLET DENSITY

"=1'2 ^ - o - o o ^ ^

0 2 4 6 9 10 12 14

"RjOj" REMOVAL, CMS./ 10 0 GMS. CLAY

FIGURE 4. Effect of sever ity of acidtreatme nt on physical properties (measured on calcined pellets) for Nevada clay.(After Mills et al., 1950, p. 1177.)

as measured by the empirical Catalytic Activity Test A(Alexander, 1947) will be used. This test, referred to asCAT-A, utilizes the conversion of a light East Texas gasoil to gasoline as a measiire of activity.

In figure 3 the effect of "R2O3" removal on activityof two montmorillonitic clays is shown. The activity risesas the "R2O3" is removed until a maximum is reached.On further removal of basic constituents, the activitybegins to decline. It would appear that the two clays,aside from having different activity maxima versus R2O3removal, activate to different magnitudes. The activityand product distribution of these particular clay catalysts are similar to those obtained on testing active synthetic silica-alumina catalysts.

The physical changes taking place during removal oftlie basic constituents of the clay are of interest. Thechange in surface area and porosity with increasingremoval of the basic components of the clay is shown infigure 4 (Mills, et al., 1950). The area increases continuously unti l 14 percent R2O3 is removed. On fur the r acidtreatment, the area decreases. Lopez-Gonzalez (1952),Esca rd (et al., 1950), and others have also reportedvariation of surface area with acid treatment.

The physical properties of acid-treated clay catalystshave also been extensively studied by Ries (1952) andby Oulton (1948). These authors reported that the grossphysical form, pellet or powder, has no effect on surfacearea. It is fur ther reported that a 0 to 40 micron fractionof powdered clay had the same surface area as a 40 to100 micron fraction. This finding indicates that the frac

tion of the total surface area contributed by the mega-surface is negligible compared with the micro-surface.A typical nitrogen adsorption isotherm may be utilizedas a measure of the surface as shown in figure 5.

The pore-size distribution of clav catalysts has beenrepor ted by Oulton (1948), Ries (1952), Drake (1949),and Mills, Holmes, and Cornelius (1950). Pore size distribution reported by Oulton (1948) for a commercialclay catalyst (Filtrol) is shown in figure 6.

The great similarity between the nitrogen-adsorptionisotherms of the acid-activated montmorillonite clay and

o. t 0.6 0.3 1.0

P/Po

HITBOGEtl PRESSURE

FIGURE 5. Nitr ogen isotherm for acid-act ivated Ash Meadows,Nevada, montmorillonite. Surface area is 340 square meters pergram.

\\\\\

1

i11

1

1

i ^

\

0

1

i11

1

1

i ^

\

1:11r-20 21 22 23 24 25 2 6 27 2B 29 30 31 32 33 34 35 36 37 38 39 40

RADIUS OF CAPILLARIES, ANGSTROMS

FIGURE 6. Dist ribu tio n of pore size versus surface area foracid-treated montmorillonite clay.


4/13

Part VIII Ci.AY T E C H N O L O G Y I N T H E P E T R O L E U M I N D U S T R Y 317

those of synthetic silica-alumiiia gel-type catalysts or ofsilica gel itself is indicative of a common physical structure. It is logical to assume that the siliceous residue leftafter removal of part of the basic components is responsible in large part for this similarity.

X-Ray Diffraction. Davidson (1947), Grenall (1948;

1949), and Holmes and Mills (1951) have reported thattypical montmorillonite X-ray diffraction patterns havebeen obtained on montmorillonites after acid treatmentand calcination (550C) . The patt ern is more diffusethan that of raw montmorillonite, which indicates eitherthe presence of smaller particles or dilution with anamorphous material. A cursory inspection of the diffraction patterns shows no major shifts in crystal structureto have occurred. This would indicate, in support of thesurface-area data, that a portion of the clay is destroyedby acid treatment to form an amorphous or gel-liliefraction and that the remaining material is virtuallyunaffected by acid except for the removal of the exchangeable cations. The 9.8 A spacing of the layers inthe calcined-montmorillonite structure leaves only a 1 to1.5 A distance between layers. As this distance is too

small for the entry of a nitrogen molecule, it seems probable that the relative surface area measured by theB.E.T. method includes only the gel surface created bythe acid treatment.

Differential Thermal Analysis. Davidson (1947) hasreported the results of analyses on a commercial claycatalyst (Filtrol) before and after calcination at aseries of temperatures. The differential analysis curve(fig. 7) of the acid treated clay before calcination showsan endothermic peak at 120-150C (250-350P) and asecond endothermic peak at 480-650C (900-1200F).The first peak is associated with the removal of reversibleinterlayer water contained in the clay structure. Onheating the same material to 480C (900F) the first

sharp peak disappears and is replaced by a broad band.Davidson interpreted this band to be associated with thereversible water held by the gel structure created by acidtreatment and calcination. The endothermic peak occurring at 480C (900P) to 650C (1200F) and foundfor both materials is attributed to the loss of hydroxylwater associated with the aluminum ions of the clay

I800*F

FIGURE

. 2 .


5/13

318 CLAYS AND CLAY TECHNOLOGY [Bull . 169

for activated montmorillonites (table 1).This is evidentfrom a comparison of the nitrogen isotherm for such acatalyst, illustrated in figure 8,with a similar isothermfor a catalyst derived from montmorillonite, figure 5.

Because of thetemperature range in which they areused, all cracking catalysts are, of necessity, heated totemperatures above 550C. Kaolin, when subjected to

such heat treatment, no longer produces an X-ray diffraction pattern. Thus no significant information concerning structure can beobtained by X-ray analysisofcatalysts prepared from kaolins. The latest literaturesurveyofthe structure of calcined kaolinsby Richardson(1951) indicates that such materials have no detectablestructure asrevealedbyX-ray diffraction. It ispertinentthat failure of active synthetic silica-alumina catalyststo give an X-ray patt ern indicates they also do notpossessastructure of long-range order. In a truly amorphous structure thesilica andalumina should behaveaschemical entities. However, thesilicain a calcined kaolinis not removed with sodium carbonate (Richardson)while thealumina is readily removed byacid (Walthall,et al.,1945). This apparent discrepancy between the reactivityof thesilica andthatof thealumina incalcinedkaolins issignificant andwillbediscussed in more detaillater.

The Chemical Composition and Crystal Structure ofClays in Relation to Their Behavior asCracking Catalysts. While the foregoing data areinformative as tothe physical structure as well as the overall chemicalcomposition of the catalysts prepared from clays, theydo not describe thecritically important chemical propertiesof the surface of the catalyst. The reversible watercontent of these catalystsin therange 450-650 C throwssome light on the chemical character of their surfaces.The method used in the Houdry Laboratory to determine thereversible water content issimilar to that usedin thermal-weight-loss experiments. However, the fur

nace is so designed that a constant flow of air of controlled water content and larger samples can be used.The sample isdried to constant weight in the apparatusat 650C in flowing air. The air isdried tocontainnotmore than 10 ' atmospheres par tial pressure of watervapor as determined by dewpoint. The furnace andsample are then cooled to the desired temperature andair with various water contents passed over thesample.The gain in weight of thesample is followed untilnotmore than 1 mg ispickedup in an hour. This maytake

as long as 24 hours at a given set of conditions. Anumber of materials were investigated in this manner.All materials, except pure silica gel, show a definitereversible water content. Thewater content has beenexpressed as the gain in weight of the dried samplewhen exposed to one atmosphere of water vapor (100percent steam) at 450C. These data aretabulated in

table2.Under the experimental conditions, pure silica gel

shows little or no reversible water content. However,alumina, silica-alumina synthetic catalyst, and theclaycatalysts show thepresence of reversible water contentin varying degree. If the following assumptions aremade: one, that thewater held in this fashion is presentonlyon the alumina, and, two, that thealumina presenthas the same area pergram as the total material, thenit is possible to calculate for the alumina the ratioofreversible water content to thesurface area. This valueis shownin thelast column oftable2 andindicates thatwithin limits a close relation exists between the amountof alumina present and the reversible water content,except in thecaseof theacid activated montmorillonite.

As pointed out byDavidson (1947), there is areversible water content of approximately 2percent in freshlyactivated montmorillonite-clay catalyst. In this case itis difficult todistinguish,at theconditionsof theexperiment, that part of thereversible water content associated with theresidual clay stru cture. An inspectionofthe water adsorption isotherms at 510 Cfor thealuminaand themontmorillonite-clay samples respectively (fig.9) shows a distinct difference between the twosamples.

FIGURE 9. Open circles show acid-t reated montmorillo nite (Ne

vada, Ash Meadows) ; hachured circles show activa ted alumi na

(Harshaw Chemical Co.). Water sorption: 5100 isotherms.

Tahle 2.

Sample description

Surface

area,sqm/g

PercentAI2O3(calculated)

Percentreversible

H2O piclc-up

450-650C

Gramsreversible

HJOper sqmof AI2O3

Ash Meadows, Nevada bentoni teacid activated andcalcined

Silica 87.5 percent, alumina12.5percent, synthetic gelcatalyst

Halloysite (Eureka), calcined and

330

250

164

29078

22.1(11.5)

12.5

40.6

0.099.8

2.4(0.4)

0.23

0.8

0.00.66

3.1 XIO-"1.1X10-1

0.8X10-*

0.9X10-1Silica gel ex ethyl orthosiiicate,

330

250

164

29078

22.1(11.5)

12.5

40.6

0.099.8

2.4(0.4)

0.23

0.8

0.00.66

330

250

164

29078

22.1(11.5)

12.5

40.6

0.099.8

2.4(0.4)

0.23

0.8

0.00.66 0.8X10"'

330

250

164

29078

22.1(11.5)

12.5

40.6

0.099.8

2.4(0.4)

0.23

0.8

0.00.66

For thealumina a linear relationship exists between itsequilibrium water vapor pressure and itsspecific wateradsorption when thedata per taini ng thereto areplottedon log-log coordinate paper. The same plot of the data

for the clay parallels that for the alumina for a goodportion of the curve. However, the parallel ism breaksdown between 10"^and10-^ atmospheres water pressure.Isotherms at other temperatures show the same difference between thealumina and theclay, indicating thatthe clay contains twotypes of water. If isochores areplotted, using smoothed isotherms as the source of thedata, theheats of sorption can be calculated from theresulting slopes (fig. 10) . It is interesting to comparethe distribution of energy of water adsorption for thetwo materials. The graph of these relationships shows


6/13


LEGEND

GAMMA ALUMI NA CHorshow Cfiem, Co.)

' ACID TREAT ED MONTMOR lLLO N ITE (Nevada Ash Meodows)

?,i WATER CONTENT % DRY BASIS

2 2 h EATOFSOR BTl ON K E A L / M O L E H^O

Cy ^ , ^

V / -// ^^' ^ : : >

/ ' " -

^-'' ,,-&'o..o.-' '

:o--''

1 1 1 i [ 1 1 1 i

FIGURE 12. Density of silioa alumina synthetic catalysts.

arrangement of silica and alumina found in naturalzeolites andclays.

Nature of the Catalyst. In the foregoing discussionit was pointed out that clays in their natural state

are not catalysts and that theprocess of acid treatmentand calcination form a new composition, at least inpart, having many characteristics of silica-alumina gelcatalysts.

At this point it is pertinent to discuss some experimental work that has been done on gel catalysts andwhich has a bearing on the relationship between thechemical and physical properties and the structure ofthe material. Oblad, Milliken. and Mills (1951) haveshown by various physical and chemical tests that thestructure of synthet ic silica-alumina gels after calcination is a mixture of extremely small silica and aluminaparticles. The alumina present has the specific gravityof gamma-alumina (fig. 12) . The authors have proposed that aluminum and silicon share oxygens atthe linear interfaces between the alumina and silicaparticles. The degree of oxygen sharing is enhancedby the method of preparation, wherein the hydroxides

are copreeipitated or treated after precipitation with astrong base so that all or nearly all the alumina is inthe four-coordinate state or "acid" form according toPauling (1930). If the mixed gels are put in the ammonium "zeolite" form the material may be calcined,at which time it losesall theammonia (viz. fig. 11), andyields a material with a high surface area (200 to 300

sq m/g) and cracking activity. It retains none of itsoriginal base-exchange capacity but, nevertheless, has abase-exchange potential.

Utilizing strong inorganic or organic bases it wasfound that the activity of the catalyst formed by thismethod couldbe almost completely poisoned with aslittleas0.06milliequivalents ofbase pe r gram of catalyst.Anaccurate correlation was found between the amount ofquinoline chemisorbed by a catalyst and its abilitj^ focrack hydrocarbons (fig. 13 ).

On the basis of these and other data it wasproposedthat the alumina is present as gamma-alumina, which islargely six-coordinated aluminum, and, asshown in figure10 andtable 2, contains reversible water, probably asOH" groups. The low acidity, 0.06 me./g, is associated

with thealumina at ornear the silica-alumina interface,where the strain set up by the bond sharing of the twodissimilar materials causes the normally six-coordinatedalumina totend tobecome four-coordinated or more likesilica in structure. However, the influence of silica alonewas believed to be insufficient to bring about the coordination shift for the following reasons: first, zeoliticallyheld ammonium ions ar e completely lost from the catalyst by heat treatment andunder the experimental conditions are not reversibly adsorbed; second, under theproper conditions calcined silica-alumina catalyst canbe partially reconverted to a silica-alumina "zeolite."

The calcined catalyst on treatment with solutions ofammonium acetate at various pH, said treatment beingcontinued to approximately equilibrium by continuously

LEGEND

O DIRECT EXCHANGEONCATALRST

ORE-EXCHANGE FROMPHINDICATED BYDOTTED LINE

20 25 ^ 0 ^ 40

% GASOLINE - CAT-A

FIGURE 13 . Quinoline chemisorption at 316C as a function of activity for cracking light east Texas gas oil. Circlewith central dot represents Si02AUOs (Houdry type S) ;circle with dash represents Si021 percent AI2O3; rectanglewith central dot represents clay catalys t (Filtr ol) ; plusshows position of SiOsMgO; V shows SiOzZrOj

pHOFAMMONIUM ACETATE SOLUTION

FIGURE 14. Base-exchange capaci ty of calcined syntheticsilica-alumina versus pH of exchange solution.


9/13


running fresh salt solution over the sample, gives therelationship shown by the curve in figure 14. In thiscurve the base-exchange capacity is shown as a functionof pH of the exchange solution. A sample was pretreatedwith ammonium acetate solution of 9.5 pll. It was foundto contain 0.96 me./g of dry calcined catalyst. Thesquares on the graph represent the base-exchange ca

pacity of the material measured subsequently at theindicated pH. This experiment appears to refute the concept that acids of varying strength exist in the catalystsince it indicates a new structure has been created bythe pretreatment at 9.5 pH capable of holding 0.81 me.ammonium ion per gram at 4.5 pH. This is to be compared with the value of 0.25 me. NH4* base-exchangecapacity per gram at 4.5 pH on the original calcinedmaterial.

From these data it appears that the presence of bothsilica and a basic ion are necessary for the alumina toshift from six coordination to four coordination at pHlower than about 10. Further, it can be postulated thatfor the bulk of the alumina particle the structuralchange taking place during either the loss or the gain of

a basic ion is not readily reversible and that the newstructure is relatively stable at the experimental conditions. This bulk quality is modified the closer the aluminum ions in the particle are to the silica-alumina interface, at which point even very weak bases, such as olefins,paraffins, or other hydrocarbons, can cause the coordination shift.

Structural reversibility requires that bonds be brokenand reformed in the alumina structure. In the presenceof water and OH", as in the ammonium acetate experiment just discussed, this would occur readily. In the caseof catalysts at normal cracking temperatures, 450-500C,the shift can be presumed to occur with equal facilityfor the following reasons: first, the 0"" and OH" ions arehighly mobile in the structure, particularly at 450-

500C ; and, second, the hydroxyl or water content ofthe alumina in this range is reversible and is held with awide range of energies (viz. fig. 10). At the temperatureof 450 C the isotope 0^* ion can be substituted for the0*^ ions in the lattice to equilibrium in less than 30 seconds by treating a sample of the silica-alumina catalystwith a known amount of H20^^. As a base approachesthe alumina-silica interface the demand for an extraoxygen or hydroxyl necessary for the shift from six tofour coordination can be supplied to the site of greateststraini.e ., those sites closest to the silicafrom thelower energy sites by means of the mobility of the oxygen ions. Actually the transfer is probably effected moreby small shifts of individual adjacent ions rather thanthe long range movement of a specific 0"^ or OH" from alow energy site to a higher energy site.

The presence of water vapor will increase the numberof low-energy sites available, make possible the formation of more acid sites, and thus change the activity ofthe catalyst. That such takes place has been found byHansford (1947) for certain hydrocarbon reactions andby Mills and Ili nd in (1950) for deuterium-hydrogen exchange between various hydrocarbons and catalyst.

Application of the Reversibility Concept to ClayStructures. It is obvious tha t the catalys t structure is avery specific one and that clays must be modified by par

ticular chemical and physical treatments to approxi-mate the synthetic silica-alumina catalyst. Moreover, theapplication of some of the above concepts to certain characteristics of clays seems possible and may offer an explanation for some of the anomalies in clay behavior.

One of the least reproducible procedures used in clayinvestigations is the preparation of the so-called '' hydro

gen" or "acid" clays. Various investigators using similar or apparently similar techniques on samples of thesame montmorillonite clay report widely varying results (Kelley, 1948). It is quite likely that the variationsdepend on the very low stabi lity of the four coordinatealuminum ions in the "acid" state. As Kelley haspointed out, the use of electrodialysis for cation removalresults in partial destruction of the clay and an appearance of free silica and alumina at the electrodes. Electro-dialysis requires very long periods of time comparedwith the mild acid treatment usually employed in preparing a hydrogen clay. The long time allows the degradation of the clay to proceed to a point where it isobvious that destruction has occurred. Likewise on prolonged mild acid treatment the destruction. of the clayis equally apparent (Gedroiz, 1924). Further evidenceas to the instability of the "acid" form of the clay isthe previously mentioned nonreversibility of the ammonium-clay, and the drop in temperature of the endo-thermic peak from 700 to 500C. In this last instance,the acid-treated clay has a differential thermal analysisendothermic peak of 500 C while the raw montmoril lonite has a 700C peak. This drop represents a verypronounced change in the energy with which the aluminum ion layer hydroxyls are held. Since the four-coordinate aluminum ions in the silica layer are alwayswithin one oxygen distance from the hydroxyl position,any change in the structure of this aluminum would affect the energy with which the hydroxyl is held.

The three important phenomena described in the previous paragraph can be explained on the basis of structural change. An "irreversible" change takes place asa consequence of removal of the cations from the base-exchange sites. The degree to which this change takesplace is a function of time and temperature and particle size. The inherent instability of extremely smallparticles leads to rapid breakdown of the structure withthe loss of the stabilizing basic ion. The larger particlestake considerably longer times or more severe treatingconditions before losing the structural characteristics ofthe alumina.

There are a number of instances where materials containing four-coordinate aluminum ions in the "acid"state do not yield active catalysts. Thomas (et al., 1950)has reported that mildly acid-treated, calcined "hydrogen" montmorillonite is not catalytically active. It has

been reported here that calcined ammonium montmorillonite and calcined ammonium natrolite are not activecracking catalysts. The surface areas of those montmorillonite samples are adequate for cracking activity. Inthe opinion of a number of workers in this field thesematerials should be catalysts, for, according to theirtheories, a four-coordinated aluminum with an associatedproton should be present in each. The present authorsthink that such is far from true. They consider that anisolated four-coordinate aluminum with an associatedproton is insufficient for activity. Furthermore, the exist-


10/13


48

44

Ui

og36

32

28

16

h:


11/13

324 CLAYS ANDCLAY TECIIXOLOGY [Bull.169

tralize" the acid cracking function which is necessaryfor activity. Theiron, and in montmorillonites themagnesia, remaining in the clay after acid treatment aregenerally isomorphously substituted for six-coordinatealumina in the clay lattice.Thepresence of free ironandmagnesia materially alters thenature of the catalyst. Ifthese ions remain in their original lattice positions, they

have only minor effects on the activity of the finishedclay catalyst. Free magnesia doesnot appear to be basicenough to act as a poison. Instead magnesia modifies thecatalysts prepared from clay containing it by shiftingthe distribution of products, giving somewhat less gasfor the same conversion than synthetic silica-aluminacatalysts (see fig. 17), thus being more like syntheticsilica-magnesia.

The iron in freshly prepared or "new" clay catalystscauses a slight increase in coke formation over that givenby synthetic catalysts. On continued use, particularlywhen the catalysts is exposed to petroleum crudes possessing a high content of organic sulfur compounds, theiron tends to leave its lattice positions and as a " f ree"or non-isomorphously substituted iron causes large

changes in the performance of the catalyst. Figure 15illustrates the effect of small amounts of iron added tosilica-alumina synthetic catalyst. Table 3shows the effectto be similar when a clay catalyst containing iron istreated with hydrogen sulfide. In both instances the decrease in gasoline yield and the increase in cokeand gasindicate thepresence of ''fre e'' iron.

At present there are three expedients for avoidingor preventing the damaging effect of iron. Two are nowused commercially, thethird ha sbeen extensively testedon pilot plant scale. The three expedients will be described in the following paragraphs.

The first expedient is represented by Filtrol Corporation's " S R " catalyst. This catalyst is prep ared froma halloysite with a fairly low iron content. The commercial results, using this catalyst with a charge stockhaving a high sulfur content, appear to be satisfactoryinsofar asresistance to theeffects of sulfur is concerned(E . J. Thomas, 1950).

The second expedient used in many commercialcracking plants is specific for catalysts prepared fromiron-containing montmorillonite. It was found that pre-treatment of the clay catalyst with steam before itenters the cracking zone will, to a large extent, preventthe harmful activation of the iron. This phenomenon isprobably related to the reversible water content of theclay in the following manner. Aspointed out earlier inthis paper, a large part of the reversible water contentcan be associated with the residual montmorillonitestructure. This water, as has been postulated by Grim

and Bradley (1948), comes from the hydroxyl groupson the alumina layer occurring in the hexagonal ringof silica tetrahedra. Bradley, on the strength of X-raydata, has proposed that on the loss of this OH' groupthe oxygen and aluminum ions in the lattice shift toa new arrangement shown in figure 16. The new arrangement requires a slightly larger oxygen-to-oxygendistance than is normal in a closely packed oxygenstructure, and results in a strain that is relieved byreadsorption of water and reformation of the originallattice with the hydroxyl groups replaced. A strained

structure is more easily destroyed than an unstrainedstructure. Thesituation issimilar in a way to the straininduced by the large surface to bulk ratio of extremelysmall particles discussed previously. Thus the reactionbetween the iron ions in the strained lattice and thesulfur ions in thehydrocarbons or hydrogen sulfide onthe surface of the lattice is enhanced and takes place

under conditions at which little or no reaction wouldoccur with an unstrained lattice. The clay catalyst during the regeneration portion of the cracking cycle losesthe clay lattice OH" groups since the temperature inregeneration is 620C and the partial pressure of watervapor seldom exceeds 0.05atmospheres (fig. 10). If thismaterial isintroduced to the reactor with a high-sulfur-eontent crude oil without presteaming, the lattice is ina strained state and the iron reacts with the sulfur inthe hydrocarbon charge. If the catalyst before reachingthe cracking zone is presteamed at the proper temperature and steam pressure to insure replacement of thehydroxyl groups, the lattice is not strained and theiron-sulfur reaction proceeds more slowly or is prevented entirely.

The third means of solving the iron problem is toremove it selectively from the clay (Shabaker et al. ).The treatment about to be described yields a catalystof low iron content. This catalyst is known as HoudryType1. For montmorillonite the technique involves acidactivation of the clay followed by treatment at 750Cwith a dry inert gas,such asnitrogen, containing from1 to 25 mole percent hydrogen sulfide. Since at theseconditions the hydroxyl water is lost from the claystructure and the lattice is in the strained state described above, the iron reacts readily with the hydrogensulfide. The color of the clay, originally light tan oroff-white, turns black on sulfidation. The sulfided clayon oxidation (500C) becomes a bright brick red, indicating the presence of free iron oxide. The clay with

the iron in the sulfide form is leached with cold, diluteacid to remove the iron and minor amounts of alumina.During the course of this treatment the clay structureis destroyed to an even greater degree than in normalacid activation. However, some residual lattice structure can be seen on X-ray diffraction of samples of thismaterial. If the clay catalyst prepared in this fashioncontains magnesia in the 1.0 to 6.0 percent range, itgives a distribution of cracked products on testing different from those obtained with the original clay catalystor synthetic silica-alumina catalyst. The relationshipbetween the gas and gasoline yields for a number of different catalysts is shown in figure 17. The two curvesfor commercial clay catalyst and "modified" or iron-free Type 1 clay catalyst were obtained by testing

Filtrol catalysts and Houdry Type 1 catalysts preparedfrom Filtrol clay. Catalysts of different activity levelswere obtained by steam treatment. It can be seen thatthe Type 1 tends to be more like a synthetic silica-magnesia catalyst than the original clay catalyst beforemodification. These data indicate that the magnesia(4.9 percent) in the clay lattice remaining after acidactivation has been activated by the destruction or partial destruction of the lattice during the iron removal.When montmorillonites containing little or nomagnesiaare activated and modified, using the procedure de-


12/13


scribed, the values obtained for the gasoline and gasyields fall on the synthetic silica-alumina catalyst curve.Thus the properties of the Type 1 catalyst can be likethose of a synthetic silica-alumina or tend towards thoseof a silica-magnesia by selection of clays of differentmagnesia contents. AVhen iron and magnesia are notpresent in the final catalyst prepared from montmoril-

lonite-type clays, the cracking' characteristics are verysimilar to synthetic silica-alumina gel-type catalyst. Thisfinding further supports the proposal that the new formcreated when clay is activated is essentially a mixtureof amorphous silica and alumina. The residual clay lattice of the modified clay catalyst disappears on heatingto 785'(J aiul the structure appears to be amorphoussince no X-ray diffraction pattern is observed. The thermal stability of iron-free Type 1 gives it an importantadvantage over ordinary acid-activated claj' catalystwhich is deactivated at about 75C lower temperature.Using the Type 1 technique, ca talysts can be pre paredfrom both montmorillonites and kaolins with excellentsulfur stabilit.y. In the case of montmorillonites containing magnesia, the new catalyst structure has specificcracking characteristics that are desirable in many

applications.

Summary. The changes taking place in a montmoril-lonite-, kaolinite-, or halloysite-clay structvire on acidtreatment and calcination can be interpreted as involving the creation of a structure with physical and chemical characteristics and catalytic activity similar to synthetic silica-alumina gel catalyst. Such clay propertiesas high temperature dehydration behavior only reflectthe degree to which a minor fraction of the clay structure has remained unattacked. When magnesia is absentand when iron is removed selectively by furthe r treatment, there results a catalyst having essentially identicalcharacteristics as synthetic silica-alumina. The originalstructure and chemical composition of the raw clay play

a large part in determining the degree and type of destruction taking place on acid and heat treatment. Inspite of the work that has been done in the field of claycatalysts, one must still resort to empirical testing ofany given clay in order to establish its suitability as araw material for catalyst manufacture.

Consideration of experimental data presented for reversible hydration at high temperatures, loss of ammoniaupon heating the ammonium zeolite form, and titrationexperiments by base leads to a concept of reversiblecoordination shift from six to four of the aluminumion. This concept has been utilized to explain behaviorof clays in electrodialysis.

Acknowledgment. Permission by the Houdry ProcessCorporation to publish this work is acknowledged withappreciation. Contributions of other members of thislaboratory are also acknowledged, particularly those ofE. B. Cornelius, J. J. Donovan, and S. G. Ilindin.

DISCUSSIONM. W. Tamele:

Milliken's idea of considering clay to be a reservoir of silica andalumina is interesting and may resolve some of the discrepanciesnoted in the past. Various ideas have been recorded in the literature as to how much R2O3 must be removed to achieve optimumactivity, and the material in this paper reconciles the situation.

Wil l Milliken exp lain t he presence of pores of 25 A radi us (fig.6) . I cannot see how such pores can be formed by the removal ofthe inner layer of the montmorillonite lattice, but rather it wouldindicate to me that another phase is being introduced.

T. H. Milliken:

The 25 to 26 A radius pore size was calculated assuming thepores to be capillaries and essentially round. This is obviously notthe case and for iuterlayer spacing, different methods of calcula

tion should be used. We have no correlation of this spacing withX-ray diffraction powder spectra. Although we cannot pick upspacings of this magnitude with our instrument; second-, third-,and fourth-order reflections should show up. There is a good chanceof a new phase being introduced. These longer spacings may bedue to the presence of particles of silica between the originallyswollen layers preventing them from coming together again. DoMacEwan's techniques show these?

D. M. C. MacEwan;

Spacings up to 120 A have been measured in o\ir lalioratovLes(by K. No rr is h) , and I do not think the re is any doubt th at spacingof the reported pore size iu the activated material could lie observed if they exis t as definite spacings. I have examined a


13/13


(Ireenall, Alexander, 1948, Montmorillonite cracking catalyst,X-ray diffraction : Ind. and Eng. Cliemistry, v. 40, pp. 2148-2151.

Greenall. Alexander, 1949. Montmorillonite cracking catalyst,demonstration of presence of hydrogen ion in heated Filtrol claycatalysts : Ind. and Eng. Chemistry, v. 41, pp. 1485-1480.

Grim, R. E., and Bradley, W. P., 1948, Rehydration and dehydration of the clay minerals: Am. Mineralogist, v. 33, pp. 50-59.

Hagne r, A. F., 1939, Adsorptive clays of the Texas Gulf coa st:Am. Mineralogist, v. 24, pp. 67-108.

Hansford, R. C , 1947, A mechanism of catalytic cra cking : Ind.and Eng. Chemistry, v. 39, pp. 849-852.

Hofmann, U., and Endell, K., 1935, Mitteilung'uer die alkivie-rung der rohen bleicherde: Angewandte Chemie, v. 48, pp. 187-191.

Holmes, J., and Mil ls, G. A., 1951, Aging of a bentoni te cracking catalyst in air or steam: Jour. Phys. Colloid Chemistry, v. 55,pp. 1302-1320.

Houdry, E., Burt, W. F., Pew, A. B. .Tr., and Peters, W. A. .Ir.,1938. Catalyt ic processing by the Houdry process : Nat. Petrole umNews, V. 30, no. 48, pp. R570-R580.

Houdry, E., Burt, W. F., Pew, A. E. .Ir., and Peters, W. A. Jr.,1938a, Catalytic processing of petroleum hydrocarbons by theHoudry process: Refiner and Natural Gasoline Manufacturer, v.17, pp. 574-582, 619.

Kelley, W. P ., 1948, Cation exchange in soils : Am. Chem. Soc.Mon. Ser., No. 109, 144 pp.. New York, Reinhold Publis hing Corp.

Lopez-Gonzalez, J . de D., and Deitz, V. R., 1952, Surface changesin an originfil and activated bentonite: Nat. Bur. Standards .Tour.Research, v. 48, pp. 325-333.

MacE wan, D. M. C, 1951, The montmorillonite minerals (mont-morillonoids), in Brindley, G. W., Editor, X-ray identification andcrystal structures of clay minerals, p. 120, London, Mineralog.Soc. Clay Minerals Group.

Mills, G. A., U. S. Patent 2,485,626 to Houdry Process Corporation (kaolin catalysts).

Mills, G. A., and Hindin, S. G., 1950, Chemical characterization of catalysts. II. Oxygen exchange between water and crackingagents : Am. Chem. Soc. .Tour., v. 72, pp. 5549-5554.

Mills, G. A., Holmes, J., and Cornelius, E. B., 1950, Acid activation of some bentonite elavs: Jour. Phys. Colloid Chemistry, v. 54,pp. 1170-1185.

Murphre c, E. V., 1951, Fluid cata lytic cracking process, inProgress in jietroleum technology, pp. 30-38, Washington, D. C,American Chemical Society.

Nut tin g, P . G.. 1933, The bleaching clays : U. S. Geol. Survey,Circ. 3, 51 pp.

Nutting, P. G., 1935, Technical basis of bleaching clay industry :Am. Assoc. Petroleum Geologists Bull., v. 19, pp. 1043-1052.

Nutting, P. G., 1937, A study of bleach clay solubility : FranklinInst. Jour., V. 224, pp. 339-362.

Nut tin g, P. G., 1943, Adsorbent clavs : U. S. Geol. Survey Bull928-C, p. 127.

Oblad, A. G., Milliken, T. H., and Mills, G. A. 1951, Chemicalcharacteristics and structure of cracking catalysts, in Frankenburg,W. G., et al.. Advances in catalysis and related subjects, v. 3, pp.

199-247, New York. Academic Press, Inc.Oulton, T. D., 1948, The pre size-surface area distribution of u

cracking catalyst: Jour. Phys. Colloid Chemistry, v. ,52, pp. 120l>1314.

Pauling, Linus, 1930, The structure of some sodium and calciumaluminosilicates: Nat. Acad. Sci. Proc, v. 16, pp. 453-459.

Richardson, H. M., 1951, Phase changes which occur on heatingkaolin clays, in Brindley, G. W., Ed itor , X-ray identification amicrystal structures of clay minerals, pp. 76-85, London, Mineralog.Soc, Clay Minerals Group.

Ries, H. E., 1952, Structure and sintering properties of cracking catalysts and related materials, in Frankenburg, W. G., et al..Advances in catalysis and related subjects, v. 4, pp. 87-149, NewYork, Academic Press, Inc.

Ritter, H. L., and Drake, L. C, 1949, Pore-size distribution inporous ma ter ial s: Ind. and Eng. Chemistry, Anal. Ed. (Anal.Chemistry), v. 17, pp. 782-786.

Schroter, G. A., and Campbell, I., 1940, Geological features of

some deposits of bleaching clay : Min. Technology, v. 4, pp. 1-31.Shabaker, H. A., Mills, G. A., and Denison, R. C, U. S. Patents

2,466,046 to 2,466,052 and 2,561,422 to Houdry Process Corporation.

Suehiro, Yoshiyuki, 1949, Preparation of activated clays andI)roperty of its t abn let t in use for ca tal ytic cracki ng : Chem. Soc.Japan Jour., Ind. Chem. Section, v. 52, p. 16-17.

Thomas, C. L., Hickey, J., and Stecker, G., 1950, Chemistry ofclav cracking catalysts: Ind. and Eng. Chemistrv. v. 42, pp.866-871.

Thomas, E. J., 1950, Fluid cataly tic cracking of high-sulfur stockwith natural catalysts: Oil and Gas Jour., v. 48, no. 46, pp.221, 224, 228.

Walthall, J. H., Miller, P., and Striplin, M. M., 1945, Development of a sulfuric acid process for production of alumina fromclay: Am. Inst. Chem. Eng. Trans., v. 41, pp. 53-140.

Date post:	04-Jun-2018
Category:	Documents
Upload:	hajinguk
View:	224 times
Download:	1 times

Mineralogy of Clay

Documents