Americar Mineralogist Vol. 57, pp. 246-252 (1972)

FLUORINE LOSS FROM SILICATES ON IGNITION H. A. Er-ArrAR, M. L. Jecrsox, ANDV. V. Vor,r, Departmentof SoiIScience, Uniu ersity o! Wi sconsin,M adison, 5S?06 Assrnacr Fluorine loss from hectorite (Z.Z% F in the (2 pm fraction) upon heating for ?4 hours at temperatures of 110. and 300.C was negligible, and at 840"C was less than S/o; however, upon ignition at SSO.Cfor 6 hours, the fluorine loss was l00Vo. Addition of Mg(OAc), failed to prevent ignition loss of fluorine from hectorite at 950"C while the addition of Ca,(OAc)" or CaCO, decreased it from l00Vo to 55%. Fluorine losses from NaF (Control) ranged firom 7/o at 110"C without SiOz to 100% at 9b0. with SiOg (ground qtafiz) addition. X-ray diffractograms showed that Catr'e had. been formed by the heating at 950.C of calcium salts with NaF. The F/Si atomic ratios in the evolved gas averaged 6.3 from NaF/SiO, samples which contained initial F/Si ratios of 1.2 to 2.6, but averaged 3.6 from samples having initial F/Si ratios of 0.8 to 0.7. Ifence, besides SiFc gas loss, other chemical forms of fluorine are lost upon ignition, sueh as NaF or MrSiF", in which M is H, Na, or other cation. INtnooucrrow

Fluorine occurs as a, constituent of many minerals and rocks (Correns,1956) as well as in soils (Robinsonand Edgington, 1946). Phosphoritesand fluorite (CaF2) occur in marine deposits (Degens, 1965) and apatite is found in neutral and alkali soils (Lindsay and Moreno, 1960). In metamorphicrocks fluorine is mainly found in fluorite and topaz (Correns,1956).Someamphibolesand micascontain F proxying for OH. Somemicas contain as much as 2 to 4 percent fluorine (Robinsonand Edgington, 1946). The chemical analysis of California hectorite (A.P.I. 34A), a magnesiumaluminosilicateclay, as reported (Bowden and Gruner, 1950, p. 55), did not include a fluorine determination. The fluorine content of this unfractionated sample was 1.7 percent, while that of the 12pm fraction of this hectoritewas 2.3 percent (Huang and Jackson,1967), basedon the Willard and Winter MCIO+ distillation combineda colorimetricdetermination of the fluorine in minerals and rocks. Use of HzSOr or H3PO4distillation and titration of the F has been employed (Ingamells,1962). Synthetic fluorosilicateshave been prepared under specifiecienvironmental and ionic conditions (Van Valkenburg and Pike, 1gb2; Van Valkenburg, 1955; Johnson and Shell, 1963) for industrial requirements.The synthesizedfluorosilicatesdid not undergo direct exchangeof F for OH in water or aqueousNaOH solutions (Johnson 46

FLT]OR'IDELOSSFROM SILICATES

247

and Shell, 1963). Unlike earlier reports of F exchangefor OH on edgesof clays (Weiss,1959;Hofmann et aL.,196l), OH from aluminosilicate clays did not exchangefor F from aqueousneutral KF solutions (Huang and Jackson,1965)I rather the releaseof hydroxyl ions from mineralsin such solutionswas accompaniedby an equivalent dissolutionof structural aluminum and iron. Thesetrivalent ions in the KF solution, as illustrated by formed solublefluorocomplexes the type equation: AI(OH). + 6KF ---+I(,AIF6 + 3KOH (soluble) (soluble) Ignition lossof fluorine from hectoritewas surmisedto be the cause of.a 5.4 percentdeficit in a system of mineralogicalanalysis which involvedthermal gravimetricanalysis gGA\ of hydroxyl loss (Alexiadesand Jackson,1966). Although the behavior of F of silicateson thermal treatment is an important considerationfor both theoretical mineralogyand quantitative clay mineralogicalanalysis,the reaction mechanismfor fluorine loss on heating is not given in standard analyticaltexts. The purposeof the presentstudy was to establishthe effectof both ignition temperatureand added materials on fluorine volatilization from hectorite,as comparedto controls (NaF with and without ground quartz as a sourceof Si) and to determinethe chemicalform, such as SiF4, HF, etc., in which fluorine is evolved during the ignition of fluorosilicates. t

aNo Mnrrrons MATEn'rALS Samples of CaJifornia hectorite (A.P.I. sample Ma', 0'l c of < 2pm size fraction) and reagent grade NaF (Control) were ignited at 110', 300", and 540"C for 24 hours and at 950"C for 6 hours to determine the total fluorine volatilized' The effects on F loss caused by 0'1 g additions of Me(oAc)u' MgCo"' ca(oAc)" CaCOa, and SiO, (finely ground, quartz) added to separate 0'1 g samples of hestorite or NaF prior to ignition were determined. After ignition, the samples were fused in platinum crucibles with a flux mixture of 2.5 e of Naoco" and 0.5 Zno (Huang and Jackson, t967). The residue was digested with distilled. wa.ter on a steamplate at 85"C for one hour and quantitatively transferred to a flask for stearn distillation. Fluorosilicic acid was distilled from the residue in the presenceof HCIO, at a temperature of 135 b r40"c. Aliquots from 250 ml of distillate were used for fluorine determination by the zirconi,um eriochrome cyanine B method (Haung and Jackson, 1967). Additional samples of NaF and SiOn were fused with ZnO and Na"CO", and treated with Hclol in preparation for the extraction and the determina.tion of silicon in the residue, in order to eslablish (by difference) the amount of Si lost' The treated residue was boiled in 5 ml of distilled water for 10 minutes, the

248

ELATTAR,

JACKSON, AND VOLR

soluble material was removed by washing with 0.2 N lrcl, and the silica remaining was transferred to a nickel beaker and dissolved by boiling in 0.b .v NaOH for 2.5 minutes (Jackson, l9b6). The silicon in solution was then deter_ mined by the heteropoly blue molybdosilicate method (Boltz and,Mellon, 1g4z).

Rnsur,rs aNDDrscussroN The fluorine volatilized from separatesamplesof hectoritedid not exceed3 percentof the original fluorinepresentupon heatingat 110.c, 300oC,and 540'C for 24 hours (Fig. l) After ignition at gb0'C for 6 hours,however,hectoritelost all of its initial fluorine.The addition of ca(oAc)2 to hectoritedecreasedF lossto bb percent,but addition of Mg(OAc)zdid not decrease lossof F (Fie. 1). Analytical grade NaF (Control) heated at 950oC gave a loss of 54 percent,of its initial fluorine content; however,the fluorine loss decreased to 4 to 12 percentwhen Mg(OAc)2,MgCO3,Ca(OAc)2,or cacog was addedto the NaF crystals previousto ignition. The in-

Hectorite

I

QHector:.te l

XHectorite

+ Mg(OAc), + Ca(OAcJZ

'tt o N

P60 o

o+u

110

950 Tenperature, oC

+ For 24 hout, *l*

I

For 6 hours +

Frc. l. Fluorine loss on ignition of hectorite, hectorite * f Ca(OAc)n.

Mg(OAc),, or hectorite

FLUOR'IDELOSSFROM SILICATES

249

creasedrecoveryof fluorine from the treated NaF samplessuggested the formation upon ignition of an alkaline earth fluoride (CaFz or MgFr) in the absenceof silicon (Fig. 2). Little differenceoccurred betweenthe use of the agetateor carbonatesalts.When the NazCOsZnO flux mixture was usedto fuse the NaF for the analyses(Huang and Jackson,1967)noneof the fluorinewas lost. The addition of SiOzto the NaF beforeheating at 950'C increased the fluoride loss to between 75 and 79 percent, thus providing evidencethat fluorine evolution is enhancedby the presenceof silica, analogousto the results on heating hectorite (Fig. 1). The fluorine could be evolved as SiFa accordingto the equation: 4 F- + SiO,9{9J:*-!::-; 2 O'- + SiFot Q) or from a hydroxy silicate systemas HF accordingto the equation: OH- + F-

e s o - cr o r 6 h o u l . > HFT + O'-

(3)

giving different 7G.4.results. In further tests in the presenceof SiO2, the fluorine loss from NaF rangedfrom 76 to 82 percentin the presenceof Mg(OAc)z or MgCO3,and decreased only lo 42lo 53 percent in the presenceof Ca(OAc)2 or CaCO3.In the presenceof Si and Ca or Mg, the fluorine is partly held by the divalent ions and partly volatilized. Further study showedthat the F/Si ratio lost depended on the initial NaF/SiOz ratio taken (Table 1). The F/Si atomic ratios of constituentsvolatilized rangedfrom 10 to 3.3, averaging6.3 (Table 1) when the initial F/Si ratio was high (1.2 to 2.6). The results suggestedthat fluorine was lost both as SiF+ and NaF from sampleshaving a F/Si atomic ratio in excessof 1. This was true even though the stoichiometricratio of SiF+ involves a F/Si ratio of 4 (equation 2). With an initial F/Si molar ratios of 0.3 to 0.7, the ratios of F/Si in the evolved gas averaged3.6 (Table 1), suggestingthat fluorine was evolvedfrom NaF in the presenceof excess SiO2approximatelystoichiometricallyas SiFa, accordingto equation (2\. For the thermal synthesisof fluoromicas,an excessof F above stoichiometrywas required (Van Valkenburgand Pike, 1952;Johrrson and Shell, 1963),possiblyto make up for F volatilization during the synthesisthrough fusion. Also in that experiment,moisture and CO2 had to be excludedto prevent,fluorine volatilization from melts containing Al2O3,MgO, SiO2,and KzSiFo (Van Valkenburg and Pike, L952\.Theseresultslike ours, indicate the easewith which F can be volatilized at high temperaturesfrom chargescontainingboth Si and F. In the presenceof a great excessof NazCOe-ZnO,as used for

250

EI-ATTAR, JACKSON,AND VOLK

MATERIALSWERE HEATEDAT 95O"C EXCEPTCOF2

STARTING MATERIALS ?

CoF2

lCoFt Lrl J (J

o

(9

o J

I

lNo2Go(CQ)2.2

F

t

o z

lrJ F

z Co(OH)2

1.64 1.932.32 3.15 s.to Frc. 2. X-ray diffractograms of materials a"fter ignition at 950'C for 6 hours and CaF, (control).

251

FLUONITDELOSSFNOM SILICATES T A B L Ei .

F A l { D S i V O L A T I L T Z EU D P 0 N T G N I T I 0 NO F i l I X T U R E SO F N a F A N D S i O , A T 9 5 0 o C F O R6 H O U R S

S a n p le T ak e n (mg) NaF Si02

Total* (mg) F

Si

F

Si

ilolar Ratio

Volatilized

Residual (ng)

(%)

(mg) F

Si

F

Si

Taken F/Si

Volatilized F/Si

First Series -- with high F/Si

I38.9 90.8 81.4 104.3 100.5 I04.0 I00.2 99.8 98.7 132.9 7 7. 3

73.7 69.1 72.5 100.0 101.3 10 3 . 0 |01.l 100.8 |00.2 1 4 4. 2 86.0

60.8 3 9- 7 35.6 45.6 43.9 45.5 43.8 43.7 43.2 58.1 33.8

34.3 32.2 33,5 46.6 47.2 48.0 47..1 47.D 46.7 67.2 40.0

s3.8 40.3 49.5 5t 4 52.7

104,0 109.6 98.9 108.9 107.4

23.5 17.6 2 1. 2 21.9 22.9

48.4 51.2 46.2 50.8 50,1

0 0 0 nd nd

nd nd nd 41.7 41.3

23 5 17.6 2 1. 2 nd nd

hd nd nd 9 2 8.8

28.1 24.7 25.8

106.2 12.3 113.5 12.6 10 9 2 il.3

49.3 52.9 5I.0

nd nd 0

44.4 47.8 nd

nd hd 11.3

4.9 5.2 nd Av.

12.5 8. I nd 10.0 6.3 nd nd nd nd 1.5 nd

nd nd 26.9 nd nd 40.0 40.0 36.6 36.6 nd 31.3

48.3 3l .6 28.4 35.6 37.7 37.3 35.9 35.8 35 4 s0.6 27.0

72 65 88 8.9 BO 7.1 10.3 l0 l r4.0 8.8

79.4 79.5 hd 78,1 8 5. 8 nd nd nd nd 8 7 .I nd 82.O

nd nd 19.8 nd nd 16.6 15.l 22.0 21.6 nd 21.9 .] 9.5

2 6 I 8

9.9 6.9

1 I I I 1 I

.5 4 4 4 .4 4

6.0 6.2 6.8 1.4 5l 5.2

1. 2

4.6 Av. 6.3

nd nd nd 18.0 l7;5 t/,U 9,9 9,8 nd T]9

0.7 0.5 0.7 06 0.7

4.0 2.9 3.8

0.4 0.4 0.3

3.1 3.6 3.3 Av. T]6

S e c o n dS e r i e s - - w i t h l o w F / 5 i 100 100 100 nd nd AV. nd nd t00 lTit

,",,""";:;,",":":::::"''' ",':':::":"ii".:::,t,":"::"':";;t,'1"::':::l:t;:':::""::" examplein the standardmethod of analysisfor fluorine,no F is lost during fusion. CoNcr,usroNs 1. Loss of fluorine from the magnesianaluminosilicate,hectorite, was completeupon ignition of sample at 950"C for six hours. 2. In the absenceof Si, either calciumor magnesiumsalts decreased the ignition loss of fiuorine from NaF (control) from 54 to lessthan 7 percent. 3. In the presenceof SiOz with NaF or in hectorite samples,calcium salts were more effectivethan magnesiumsalts in decreasing fluorine loss,becauseof the formation of CaFz (fluorite). Even with calcium presentfluorine loss was still 42 to 53 percent of the total F present. 4. On ignition of variable amounts of Natr' with SiOz, ihe F/Si atomic ratio in the evolved gas approached4 when SiOzwas in excess of NaF but increasedto 10 when NaF was in excessof SiO2,indicating gaseousloss of not only SiFe but also of NaF. Acrwowr,gtcunNrs Researeh supported in pa.rt by the College of Agricultural and Life Sciences, University of Wisconsin, Madison, under project 1123; in part by National Science Foundatiou grants GP 4144- a,nd GA ll0&Jackson; and in part by a

252

EI-ATTAR,

JACKffiN,

AND VOLR

Ford Foundation grant to the Institute of Land Reclamation, Alexandria University, Eg"ypt, U.A.R.; a contribution through an International Consortium for Interinstitutional Cooperation in the Advancement of Learning (ICICAL). RnnnnrNcns Aloxnors, C. A., eNo M. L. Jacrsox (1966) Quantitative clay mineralogical analysis of soils and sediments. Clnys Clag Mineral. t4, ZS-52. Bonz,D. F., ewo M. G. Mrr,r,oN (1g47) Determination of phosphorus, germanium, silicon, and a,rsenicby the heteropoly blue method. AnaI. Chem. tg, 873-877. Bowonx, W. C., exo J. W. GnuNsR (1gb0) Chemical analyses.In p. F. Ken et. al., editors, Analgtical Data on Reference CIag Minerals. Amer. Petrol. Irctit. Res. Proj. 49,PreI. Rep. 7,8&-58. ConneNs,C. w. (1956) Geochemistr5',of halogens. phgsi.cs and, Chemistry ol the Emth. r,181-%3. Docuxs, E. T. (1965) Geoclwrn'istry o,f Sedi,ments,Prentice-Hall, Inc., Englewood Cliffs, New Jersey, pp. 141-188. IlorlteNw, U., I[. P. Bonnlr, exo W. Gnolrns (1g61) The dimensionsof crystals of clay minerals. Z. Anorg. AIIg, Chem.3O8, 143-154. Iluawc, P. M., eNn M. L. Jecrsow (lg6b) Mechanism of reaction of ne,utral fluoride solution with layer silicates and oxides of soils. SoiI Sci. Soc. Amer. Proc. 29,661-665. (1967) Fluorine determination in minerals and rocks. Amer. AND M,ineral. s2, lmB-l1AI . fwcervrnr,r,s,C. O. (1962) The application of an improved stezLmdistillation apparatus to the determination of fluoride in rocks and minerals. Talanta 9, 507-516. Jecrsou, M. L. (1956) Soil Chemical Analysis Ad,uanced. Course,5th printing by the author, Dept. of Soil Science, Univ. of Wis., Madison 58706. JonNsoN, R. C., ewo I{. R. Srrnr,r, (1968) Wa,ter-swelling synthetic fluormicas and fluormontmorillonite. U. 8. Btn. Mines, Rep. Inuest. 6235. 'W. LrNnsev, L., eNn E. C. Monnro (1900) Phosphate phase equilibria, in soils. Soil Sci. Soc.Amer. Proc. 24, 177-782. RonrNsow, W. O. ero G. Eocrwcrox (1g46) Fluorine in soils. Soit Sc,i.61, 841353. Ver.r Ver,rnrvsunc, A. (1955) Synthesis of a fluorotalc and attempted synthesis of fluorochrysotile and fluoroanthophyllite. J. Res. Nat. Bw. Stand. SS, ZIS217. AND R. G. Pr