Neutron Diffraction Refinement of an Ordered Orthoclase Structure

American Mineralogist, Volume 58,pages 500-507, 1973

NeutronDiffraction Refinement of an orderedorthoclase Structure

Erweno PnrNcB

Inst,ttute for Materials Research, National Bureau o.f Sta:ndards, Washington, D. C. 20234

Gnenrprrp DoNNnylNo R. F. MenrrN

Department of Geological Sciences, McGiII Uniuersity, Montreal, Canada

Abstract

The crystal structure of a pegmatitic monoclinic potassium feldspar, (IG rnnonn) "Nao (Si,,*Al'*) t0"*(OH)o.1, from the Himalaya mine in the Mesa Grande pegmatitedistrict, California, has been refined with 3-dimensionalneutron-diffraction data to an unweighted R value of 0.031 for 721 symmetry-independentobserved reflections. Atomic coordinatesdiffer by no more than 3 estimated standard deviations from those of Spencer B adularia, yet the specimen does not have the adularia morphology, and no diffuse reflections with (h + k) odd have been observed.Direct refinement of the tetrahedral cation distribution shows that the Al content of the T(2) sites is not significantly different from zero (actually -0.016 with an e.s.d. of 0.029); in other words the Al-Si ordering in the tetrahedral sites is essentially complete. The mean SiO distance in the T(2) sites is 1.616A, appreciably greater than the values predicted by various regression lines relating bond distance to aluminum content. This indicates that the observed mean T,(m)-O, ?,(O)-O, and Tn(m)-O bond lengths reported for low albite and maximum microcline are consistentwith full Si occupancy. This ordered orthoclase occurs in gem pockets in a microcline-bearingpegmatite. The association suggestsstable growth of ordered orthoclase above the field of stability of microcline and metastablepersistence to lower temperatures.Perhaps because of more rapid crystal growth, the bulk of the pegmatitic K-feldspar ordered to common orthoclase, then transformed to maximum microcline.

Introduction interestingaspect became apparent: the refinement During a continuing study of the occurrence of of the structureof this feldspar, which was classified hydroxyl ions and water molecules in minerals, a as orthoclaseon the basisof chemicalcomposition, valence sum (Donnay and Allmann, apparentsymmetry from the powder pattern,angle 197A) was per- o), formed on orthoclase,using published atomic param- betweenoptic axes (2V" = 63" i.1 and mor- eters (Jo,nesand Taylor, 1960; Colville and Ribbe, phology,indicated essentially complete order of the 1968). This summation showed a low value of the aluminum and silicon cations between the T(1) sum for the oxygen atom in position O", and, there- and T (2) tetrahedral sites, although the mean fore, a neutron difiraction study of a feldspar crystal Z-O distancesfor thesesites would haveindicated was undertaken to investigate the possible substitu- lessthan completeorder accordingto the published tion of small amounts of OH- for the oxygen in this regressionlines (Jones, 1968; Ribbe and Gibbs, position. However, even before our own negative 1969). The pulposeof this paper is to describethe results were obtained, a neutron refinement of a refinementof the structure,to comparethe results sanidine structure (Brown, Hamilton, Prewitt, and with thoseof studieson similar crystals,and to dis- Sueno, 1971) ruled out any measurableamount of cussthe implicationsour findingsmay have on feld- H in this feldspar species. Infrared absorption spar mineralogy. spectra of clear hydrothermal feldspar have shown no evidence of hydroxyl-for-oxygen substitution SpecimenDescription (J. Zemann, private communication, August 10, The neutron diftraction data were collectedfrorn 1,971,). a cleavagefragment taken from an overgrowthwhich As the investigation progressed,however, another occurs on perthite individualsfrom the Himalaya 500 NEATRON DIFFRACTION OF AN ORDERED ORTHOCLASE 501 tourmalinemine (Jahns,1954,map 8, locality20) in the Mesa Grande pegmatitedistrict, San Diego County, California. As in other granite pegmatite dikesof SanDiego County,long pendantsof micro- MICROCLINEPERTHITE cline perthite having quafiz rods in graphic inter- WITH GRAPHICQUARTZ growth grow from the hanging wall (Figure 1). Towardsthe core of the pegmatite,the quartz rods die out abruptly,presumably in responseto slower ORTHOCLASEPERTHITE crystalgrowth, and massiveqvartz aggregatesform GEMORTHOCLASE separatebodies next to microclineperthite (Jahns OVERGROWTH and Wright, 1951).Nearer the gempockets, microcline perthitegives way to orthoclaseperthite; these POCKETPEGMATI perthites are coated by a single-phaseorthoclase MASSIVEQUARTZ overgrowthup to I cm thick where they project into the gem pockets.Simpson (1962) suggestsa hydrothermal origin for the microcline perthite- quartzintergrowths from this area.Jahns and Burn- ham (1969, p. 856) proposegrowth of relatively Fro. 1. Cross section of an asymmetrically-zonedpegma- Burnham few feldsparnuclei from an aqueousfluid, with the tite dike. modified after Figure 58 of lahns and (1969). Thicknessof dike is measuredin meters' The clear, transportedthrough the fluid from an constituents single-phase overgrowth (labeled "gem orthoclase over- adjacentsilicate melt. The presenceof hydrothermal growth"), which coats orthoclase perthites which project fluids is further suggestedby corrosionfeatures ob- into gem pocket, yielded the specimensused in this study' served on distal portions of blocky perthites (cl The country rock is gabbroic. Simpson,1964, p.23) andby fluid inclusionsin the feldspars.The overgrowthpostdates the main epi- 19'72).A value of.0.47 percentRb'2O sode of corrosion;it occursin pocketsin associa- munication, by wet chemical analysisof the tion with gem tourmaline, beryl, qvaftz, and fine- has been found which the neutron crystal was ob- grainedmuscovite, which alsocoats the etchfeatures. specimenfrom (S. Horska, private communication,1972), Jahnsand Wright (1951) note the occurrenceof a tiined by the value of 0.49 percent clear variety of orthoclase("dental spar") among and this is confirmed analysis(Table 1)' pocketpegmatite minerals in the adjacentPala dis- Rb2O from the electron-probe trict, and mentionthat microclineis subordinatein Procedure such pockets. Experimental Table 1 is a summaryof chemicaland crystal- The cleavage fragment used in the neutron refinement lographic data on this crystal, along with similar weighs 48.356 mg so that on the basis of an observedden- It was information, as given by Colville and Ribbe, for sity of 2.563g/cm",it has a volumeof about 19 mm'' with its D axis approximately but not exactly SpencerB adulariaand SpencerC orthoclase.X-ray mounted parallel to the d axis of a computer-controlled,4-circle, neu- rasterscan images of a portion of the neutroncrys- difiractometer at the National Bureau of Standards - tron tal, at a scaleof 100 p 3.5 cm, givenoindication Reactor (Alperin and Prince, 1970). All reflections with 20 angle of Na and K inhomogeneity(S, Horska, private x2 -5" and within a spheredefined by a limiting by a communication).No variationsin A1 to Si ratio of tOO' and a wavelength of 1.232A were measured wbich first measured the intensity at the peak were noted in additionalelectron-microprobe analy- procedure position and at background positions on either side' If ses;no correlationcan be madewith the very subtle ihe peak intensity did not exceed the mean background variationsin extinction angle,well within 1o, that by at least 2a (o - (1,, ! It)'/"), the reflection was con- are found in the untwinnedovergrowth. sideredto be "unobserved",and assignedan upper limit of greater than R. H. Jahns and A. A. Chodos (unpublished intensity based on o. If the peak intensity was measuredby a 0'20step scan' If the peak data) consistentlyfind rubidium and cesium en- 2a, the peak was intensity exceeded1000 counts, the counting time was re- riched in the pocket orthoclasecrystals of the San duced in order to avoid spendingtime on meaninglesssta- Diego County pegmatites;one specimenof Hima- tistical precision. laya orthoclaseovergrowth contains 0.11 percent For each observed reflection the integrated intensity was RbzO and 0.014 CszO (R. H. Jahns,private com- determined, and from it a value of the structure factor, F' and 502 PRINCE, DONNAY, AND MARTIN

Ta.eI.n1. Compositions and Crystal Data of PotassiumFeldspars

Chenlcal Analysis Hlnalava Hlnalaya , Spencer B Spencer C weight percent orthoclase? orthoclase- aduLaria* orthoclase*

s102 o).Jz tt tt (1 R 64.2 &zo: L8.25 18.45 18.5 18.3 Kzo L4,25 14,76 15.1 15.3 Na20 1.12 1.08 1.1 no UAU 0.'0s nll 0. 01 0.1 Sr0 nar nil 0.5 0.1 Ba0 0.25 nnt o.7 0.1 Rbr0 0,47 0.49 Hzo' o,44 H2o 0. 02

TOTAL 100.1 7 100.19 99.25 99.0 (K9. Formulabas€d on 4(st, At) (Kg.gaNa6.1gRbo.olp0.05) (K6.g6Na9.16tr6.94) (K9.9gNa6.19Ba9.61) 91Na6.66Ca6.91) per F,U. (r4artin & Donnay ' (s1:.01A19.99) (sr3.66r'11.s6) (s12 (s12 L97z):z=4 .96A11.62) .99Ar1.91) [07.96oH6.641 [07.950H6.64I [07.93(on) 6,62 I oe.oo Fomula welght: 275,80 275.27 276.72 277.05

wavelength used (l) I .54051 not glven not glven a (A) 8. s532(11) 8.554(2) L 561(2) b(i) 12.953 (14) L2,970 (4) L2.996(4) c (A) 7 ,2099(Lr) 7.2o7 (2) 7.L92(2) 90" 90" 90' B 115"073(9) 116.007 (r0) ' 115. 010(10) . Y 90" 90" 90" v(43) 718. 90(11) 7L8,7 718.9 Dga1.(8/ct3) 2,543 2.557 2.559 oo6"(g/cm3) 2. s53(1) 2.566 2,563 at 230c 2vo 63" 58,4' 43.6'

Observeddiffuse reflectlons (h + k) odd

t'llet chenlcal analysls by S. Horska, McGlll Unlverslty. HrO+ deslgnates H20 llberated between llToC and 1000.C. +Analytlcal data obtained by C. C. Hadldlacos on the electron nlcroprobe of the Geophysical Laboratory, Carnegle Instltutlon of Washlngton. *Data fron colvllle and Rrbbe (1968, Table 1). The glven forrrulae dlffer sllght1yi the conversion factor used by colvllle and Rlbbe ls based on 8 oxygen atons per ceil. optic axlal angles (i voi'fror Spencer (1937), /1A11 dlmensions for Hlnalaya orthoclase determlned by R. F. Martln,

its standarddeviation, op, basedon counting statistics.At this (1951). The scattering factor for the tetrahedral sites was stage a systematiccomparison of pairs of reflectionsof the assumedinitially to be that for Si, which is 0.42 X 10-nm type hkl and, h*l was made in order to determine whether a (Bacon, 1972). The scattering factor for Al is 0.35 y departure from monoclinic symmetry was observable.The 10-" m, so that substitutionof small amounts of Al in either overall agreementindex, definedby S : >lF(hkl) - F6kD/ point-position will reduce the scattering factor by a few >F(hkl), was about .02, and there was no discernibletrend as a percent. This will appear in the final parameters as a function of sin 0 or of reflection index /<.Therefore in the sub- slightly larger temperature factor. The scattering factors sequentstages of refinement we have assumedthat the srucrure for Na and K are both 0.35 X 10'* m, so that the Na:K is monoclinic, belongingto space$ovp Cz/m, with the cell occupancy of the alkali sites is immaterial to the refine- dimensions given in Table 1. There are 764 independent re- ment. Substitution of other alkali metals, such as Rb, or flectionswithin the limiting sphere,of which 721 wereobserved. vacanciesin amounts up to a few percent also appearsin The observed reflections were used in a least-squares the form of slightly difterent temperature factors, and has refinement,using the program RFrNE(Finger, l96ga), start- an entirely negligibleeffect on thesecalculations. ing with the position parameters of Jones and Tavlor Repeated cycles of refinement with isotropic temperature NEUTRON DIFFRACTION OF AN ORDERED ORTHOCLASE 503

Tlsl-n 2. Atomic Coordinatesand Anisotropic Temperature Factors for Himalaya Orthoclase

Atoo x y , Blt* RZZ B:l Btz 823 B3l Bi"o

R .2840(3) 0 .1388(4) 1.5(r) 2.4(1) 2.0(1) 00 .52(1) 2,0E(4) r(1) .0095(2) .1842(1) ,2241(2' 1.38(5) .99(s) .94(5) -,23(4) -,15(4) ,59(4) 1.08(3) r(2) .70E5(2) .1176(1) .3446(2\ 1.38(5) .95(5) 1.10(s) ,o2(4) .03(3) .55(4) 1.14(3) cA(l) 0 .1451(1) 0 2.13(6) 1.s4(5) 1.88(5) 00 .95(4) 1.82(3) oA(2) .6365<2) o .2853(2) 1.e4(6) .97(s) r.s3(5) 00 .33(4) 1.54(3) 0B .8256(1) .1459(1) .2279(2', 2 ,78(4, 2.45<4',) 2,23(4) -.24<3) .16(3) r.47(3) 2.11(3) 0c .03s3(1) .3118(r) .2599<2) 1.66(4) 1.40(4) r.75(3) -.30(3) -.36(3) .54(3) r,65(2) oD . I8I8(1) .1252<7) .4080(1) r.91(4) 1.60(4) 1.00(4) .14(3) .13(3) .23(3) r,65<2)

*B'".." a*p.."".d ln i.2 and are glven as th€ coefflcients of the expres6lon 2s*23.-a2hke*5*3.^+2k expt(-1/4(h2a*2r..+k2b*28"" b*c*B^.+2 hc*a*B.,)1. standard devlations {n the lest

lflcant flgure are shom ln paretrtheses. factors led to a weighted R value of 0.094 and a conven- Discussion tionat R of 0.082. Four more refinement cycles with anisotropictemperature factors and an extinction parameter Cationdistribution of the Zachariasen(1958) type led to a weighted residual R of 0.038,and a conventionalR of 0.033. Figure2 is a plot of the Al/(Al + Si) distributions, At this point the chemical analysis shown in Table I as determineddirectly by neutron diffraction, relative became available, and two further stagesof least-squares to the mean 7-O distancefor the two sites in this refinement, each with four cycles, were performed, In one, study (circles)and the two sitesin sanidine(squares) the alkali site was assumed to contain 4 atomic percent as determinedby Brown et al. (1971).The figure vacancies,and silicon and aluminum were distributed be- proposedby Jones tween the tetrahedral sites according to the regression also shows the regressionline formula of Jones (1968).1 (See Table 4.) The weighted (1968) for evaluation of aluminum content when residual R after this stagewas 0,03532. In the other refine- only the mean f-O distances are available. The ment the above content of the alkali siteswas kept fixed but points from this study are plotted at Al/(Al * Si) the contents of the tetrahedral sites were allowed to vary ratios of 0.0 and 0.5 becausethese are maximum under the constraintthat the overall ratio of Si to must Al be points, due to the impossibility of an 3.0 to 1,0. This refinement led to a weighted residual R likelihood of 0.03494.The ratio of residualsis 1.0110,compared with aluminum content being negative or greater than a ratio (Hamilton, 1965) of 1.0062 necessaryfor the result 0.5.The one-endederror barson thesepoints indicate to be significant at the .005 level of confidence. In other the confidence interval determined by a 26 nnge words there is only one chance in 200 that this amount from the value determinedin the refinement,and the of improvement could be due to chance alone. This result shaded area is a 26 confidenceband as given by is therefore highly significant. It should be noted that the constraint on the overall chemical composition removes Jones. much of the correlation between occupancy factors and Altough the confidencelimits on the direct refine- temperature factors which would otherwise make simul- ment in this study and thoseon the regressionline taneous refinement of both a more uncertain procedure do overlap slightly, the agreementmust be con- (Finger, 19'68b). The final unweighted R was 0.031. The sidered poor. There is a strong indication that a largest correlation betweer the occupancy factor and any other parameter was 0.434 with pss of T(l). The final range of possibleT-O distancesexists for a given parameters are shown in Table 2 and observed and calcu- Al/Si ratio, and that at the pure silicon end this lated structurefactors are given in Table 3. range lies mostly above the regressionline. The The fractionalAl contentof the f(l) sitesis 0.516(29), mean Tr(m)-O, f2(O)-O, and ?z(z)-O bond with a correspondingvalue for the Al content of the f(2) lengthsof 1.610,1.615, and I.612A in low albite sites of -0.016(29), where the number in parenthesisis the estimated standard deviation of the least significant (Ribbe, Megaw, and Taylor, 1969) and 1.614, figures. Because the actual fractional Al content of the 1.611,and l.612A in maximummicrocline (Brown T(2) site cannot be negative,we assumethe T(2\ site to be and Bailey, 1964) arc then consistentwith full Si occupiedonly by silicon and the Si to Al ratio in the I(l ) occupancy.These results indicate that, althoughthe site to be 1.0. mean 7-O distancecan be used as an indication of the degreeof order,the rangeof uncertaintyin any givencase is at least5 percentin aluminumcontent. 'Use of the linear relationship obtained by Ribbe and The fact that the orthoclaseand sanidinepoints hap- Gibbs (1969) doesnot changethe presentconclusions. pen to lie on a straightline may well be fortuitous. 504 PRINCE, DONNAY, AND MARTIN

Tlslp 3. Observedand Calculated Structure Factors of Himalava Orthoclase*

i'o'-O ! 22t -?09 2 !77 -565 -! 92r -9Zl -2 75c 749 H'f I 179 -159 H' l0'-l 6 517 -52A "2 I l5t2 l5A2 c llOS -ll0l -l 625 -82c 0 1566E-1576 5 691 650 -905 -6A0 -{ -7 -r2l9 -6 -665 q, 0 90 7 916 5 897 6 665 I 5c{ 5A9 2C7 ?59 469 6c7 I l2J3 67! 12,0 2 1559 1566 7 c?8 0l? E lll cl 5 1259 -122r i'5.-6 -2 9t. 2b -5 167 -ls7 7 553 555 -c 66. 5l c 206 167 9 625 65q tO 96r c9 5 ?!l -756 0 !20 300 -5 55s 55c -2 l!2! -l!01 -6 2J3 -2lt 6 {15 907 7 921 855 5 255 -260 2 556 560 -l !!5 5c! H.9,-5 0 502 -521 -c 2t6 -226 E 995 lolE H,l,-5 H'2,-5 9 26t -2a2 5 r05 -40 { r02a t025 l 169 t?6 2 ASZ 885 -2 655 66r 6 tlll lll2 I 255 2t7 -r 590 -589 c czt -{2J 0 rl I -506 H'0,-7 -5 6!c 6lt -6 576 -365 H'J'-t 6 t65 120 5 {tr 59{ -l S85 -556 5 I l2 8l -l 509 -510 -c 995 -995 t 150 -lzc I 5C7 5r9 d,l5'-5 -e f!2 r27 -l 1016 t025 -2 259 2c! -l 169 -?69 9 946 9t5 J l164 -t155 H'10'0 0 159 r55 I 870 908 0 t6c t5a -5 to3t -1029 -t lq59 -t{19 56269 t 9JJ -9!9 -1022 2 9Cr 6 5 9qJ 2 202 -ztA -5 c5? -q65 I {45 -42? -{ 552 5q{ f 293 515 -6 315 -2ct ! {!J \22 -797 4 797 5 tt25 ll05 4 506 5l! -l t0o5 -1005 5 195 lE9 -? f,65 !6c -4 295 a9C -6tt c 615 7 201 167 6 896 90! t 526 521 5 7{J -760 0 lse! tc97 -? 5lt J25 H,9,-r -2 365 lel -Jcl -cO7 0 324 9 568 8 567 580 S 9Ol -9ol 7 576 379 2 l22l l2l{ -5 409 -6?l 0 685 609 l0 !l? 5l? 5 159 r5l c 228 2Jl -J 5?5 55? _5 102 -74 -l l0l. -6t H'0'-6 H'l,-c 7 Ct? A09 6 9!9 9!9 -t 655 87q -t 609 620 H'll'-6 I 978 9E! N.2''2 9 r59 -172 6 Ct9 42f I f,ai -c00 I l?6r -127? 5 lE9 -20n -2 9{. -lc -5 lr5 ct2 -! 9ll -6ae J 605 602 5 c9Z 50? I 619 615 5 571 56t -lllr 0 lI02 -5 rl. -c? -8 l2l -3lc x,!'O 5 915 499 5 55a J69 5 cC7 -ca9 -715 -t 2 708 855 C65 -6 S8c !65 I A25 Al0 | !20 5S2 7 l0c6 -10J9 5 98. S0 H'15'-J c 265 2c3 I 654 -66a -c s5l 56i -9 t29 -tz? 3 205 -26! -6 50f, !0c 9 521 3cl -59. 6 567 3 lc2 150 -2 958 912 -7 lol5 -1016 5 6{1 6!5 -{ 362 t59 H,9,-! a,lt,-s -3 56E 585 -25 6 AO. 5 6c7 8c8 O 515 6A0 -5 IOSA -tO!9 | 261 -251 -A 75t t46 a,7. O -l 229 ?q! l0 EBJ 90c 7 l5o -572 2 555 55C -5 ll?t llcg 9 9c. 55 0 lc9 -t59 -5 29J Sl5 -l c59 45t I 600 621 9 552 510 C 1057 loj? -l 925 -952 2 50! !05 -7 2!A 2t7 -l 50{ 265 I 212 _t87 5 210 196 H'o'-t 6 5lc 129 Hr5,-5 4 961 -951 -5 8qr 850 I 256 2r9 I r5J s6S 5 28c zla H,l'-5 I 79 -48 H,{,-8 6 260 -279 -J 96 -46 5 Jl5 106 5 856 -662 -9 -c26 {fc t0 2ct -212 -5 559 550 A 93. 36 -t 762 759 5 555 -556 7 150 !?? {, 15,-2 -2 -95 105 -7 576 -595 2 52c -522 -1 l2c9 l2!2 7 t5C r05 -5 0 lll9 llc! 155 4cO H'z.-l c c?7 c?5 I c92 -{63 H'6. -2 H'll''q -5 2fO 260 -3c -t ? lcz {56 c6{ 6 5!9 -595 5 015 425 H,9. -2 -l ll8 -92 9 l6q? 16!7 -l 1282 -1506 -A 92. 5l 5 ?r0 -6AA -6 6t2 -67t -! 229 215 I 9!. -5? 6 l55c 1525 I 1025 tola -6 692 ?OA i,q,-7 7 ll2 -24 -c t69 -taz 4 55A 5?6 -5 160 102 -l c7t '{70 s 5C3 6!0 I cor cl5 3 652 685 -4 7!1 7C! 9 570 395 -2 t2. -J6 6 r060 1065 -5 565 -565 I 6l? -6e6 5 590 -?00 l0 l{60 tr50 5 710 -?2t -2 cco -{q6 0 585 5?5 0 t4Jc -lcll -l ll9 -!5 J coz -(2t 7 9{J 956 0 t26 -1 30 2 6t5 -6r r 2 601 -607 L 152 7C0 5 loJ 9 H,t5.-l aro,-q 9 96 l{ 2 90 5 c 25r -?54 I a3{ zrt ! r22o -1198 t !16 -505 | 678 -696 5 9l -77 -3 214 27O 6 746 -7C! -2 69{ 6aa 5 606 6ll -5 1007 -996 -6 690 69c H'l'-z 6 ltc6 -ll!q I 19{ -166 -J Coq {08 0 565 57s 7 l8{ -172 H, ll'-J -! 9C 50 -. 55{ 560 6 565 -57j -l 666 -680 2 A!{ 432 -l {12 cll -2 -7 l!99 lcl5 c09 rn9 H,q'-6 I lEo -lt9 4 l09E 1069 H,9,-l -J 952 -98C I 92. 75 -!15 0 ll57 lt77 -5 521 H,A'o J ?{J -7!5 6 525 50J -l 2cJ -25J 5 lA5 -169 -5 2 1895€ 2021 20c 20! -2 t7c -t55 5 L22 87 _A 580 -516 6 t?! 70 -? 575 -5ct I 877 -91! 5 lt?9 1lc5 c -l -8 !0t e96 561 169 67. 58 O 506 lOO | 292 299 -6 !t9 5t6 -5 l7a -395 I 172 -l 66 -5 6 l5c5 lJ22 I 696 6qB -6 SSf -5?8 A ?cc -t5l -c 575 -573 H'6.-5 t55 ll5 5 59t 557 {,1!r0 -l -r8t -loc? 6 596 6lC I lrl -9r -c 9Or -9a6 c 6cA -6t6 -2 96 -!9 166 7 lo52 -00 -JJI -5 l0 96. 5 568 561 -? lt{o -tt56 6 665 6?6 Hr5.-5 o 529 5C2 -2 152 -5t I 120 {66 C75 -Ja! -6 -! 7 {01 0 591 -591 I 65? -650 2 269 t92 0 c?2 c09 5 47. H'lt.-z 5CC !50 -262 -l H,o'-J 9 {ld cl2 -5 254 -2c9 c l5t5 r45{ 2 ?50 -r52 5 26t ltJ 5l{ -5 -161 H.5'-g H,{r-s -5 975 -946 6 20{ 166 c 279 -256 7 500 271 160 -6 -5 -600 166 I 5? Hr l, -l -l {25 -{57 8 756 77t 6 597 -60? 591 -c -l -{6J I76a 1755 I 995 -9a5 -q 789 7qz I C53 -C37 a l2r5 -1196 4.9.0 c8o -2 -c90 -? -556 c56 c65 c75 3 5J5 S2l -Z 2ag -?62 3 156 -105 I Aza A5J 2 !48 -5 -275 -7 -7c0 -605 o 72O ?60 ?69 5 599 620 O 1566 -156C 5 959 95! 5?. 2A 5 7!6 | 57c -! -5 -cA9 -676 2 592 560 ttz 7d9 7 647 -617 2 1056 lO55 7 9BL -976 -a 358 576 cA2 5 645 -16t8 -l g9q -3 c 16{0 66o q ?Zc -7!5 9 at? A2{ -b ll22 rl0c -c 887 -A79 !46 591 1 265 271 -l c9I -500 6 8A 5? I r8O -qgl H'!,-t 6 66t -6r? -c 55e 55C -? 55c -324 -266 6 2br 5 26t 2At I 92. -l I h' 5. -2 -2 7Cl 764 0 1559 -l!c2 {,tt,_! 0 96. t0 -l?39 -l H'9'J l0 l25J 5 5og 321 t05 -87 lO ltr -94 0 1980E 1905 2 9! -20 2 989 990 -56 -5 c 7 9c. I 25A 252 -7 e99 -505 ( 127? -125c 97. 6l 25C 225 -58C -! 4,0'-2 9 !57 ll8 5 274 26r H'c.-{ -5 zJA -25{ 6 C02 -562 I 579 lgc !96 -l -c0 5 ?25 -2?4 -J 2r7 -190 8 157 t72 7J -6 '24r -acz H. I ' 0 f tZO t7 -6 2lO 169 -l loa -7c I ra! -186 H'10,-7 I 07. I -6 -l0lt -? 25b 265 9 Bg9 -B9S -c 29J -266 I 802 778 5 l9l 202 Hr6'-5 5 !0t5 ll98 llq5 -{ -790 -?54 l!03 -l5bl -9 {9b -5?l -2 9t6 -926 | 797 5 2{7 0 c30 {{9 -2 J 190 t96 -9? 9e5 9r2 -7 590 599 H,5'-6 O 169 lc2 5 87J -066 7 t59 2 9A? t00? 0 l46lt-1665 -5 5q6 57( 2 jto -5t! 7 qoq 9t5 -r cJ7 c5! c t2!6 t220 -t -17 H, 2 lS9{E-1r57 !8. -5 tl55 -ltc5 c 555 -Jr5 9 57q -5Al I 98. 2A -a aJ3 -ar7 l l,0 -252 4 129? 12?6 -l ?12 ?21 -t t7t -t6A 6 6A8 -616 a 429 825 0 053 85t o 25f -t655 -10{2 -5 -556 b lTcI I 305 3rl I 92. ll 5 255 ?tf 2 Aco a59 2 1042 512 -98t -l -4 a rl{ el9 H,2,-A I l{55 -1450 lO t22 6 7 201 -184 c 968 875 A62 207 lql -l -105r -2 l0 580 J95 5 2Ac -279 -f 159 760 6 l69t l5C7 6 2c6 2!9 1028 t72 !?0 -ll9 2 r!2 -4t2 ' 2t! -?02 {,{,-3 -s 16l -127 6 lllA ll0! 7 102. 0 c66 {6! -Joq H,or-l C l7I -lCA 9 c8c -510 -J 75{ 765 a,l0r-5 2 407 6 721 -le6 -6 357 lqq -, zfa 2aa -t 9c. -!l 8,12,-5 c 5?l 55t -6 -2 -cl0 Z!2 257 8 lJo 47 Hr5,-5 -C ?OtO -1925 I rsc 150 I c58 -156 59{ -6 -906 a90 -8tc -675 OO! -2 155! l5c7 5 loal -1069 3 lStJ 1295 -6 lt65 I 155 0 0 6tb -c -9q ll5 ata'-7 -5 t22O -1203 O 649 -692 56t4? 5 405 -8tC -f 106 -Lz e 651 6{1 2 ?59 739 -2 q -c -606 5!{ J!! -5 6?5 -686 2 t2t? -tzOO 7 tcz 105 -? ZZcaE ?lc5 l6J 155 | 77 121 802 -q55 -2 -226 -t7l -12? -2 -5!0 0 clB 2t0 -t ?tt -717 c 558 568 0 alc 650 6 l?5 6 t02. 6t5 o{ 2 1586E lcl0 o 5c7 527 t ll99 -t?02 6 c5l -c51 2 5{n 5C6 0 ltc c ltal ll59 2 56. -50 5 585 !9? g !05 -512 { 2$55E 2?26 H,l0r-c H'lzr-q -722 6 E02 6lc 4 ?66 ?5q 5 t!67 -t555 tO 2t4 -161 -5 A6A -67t 6 ?!l -Ar9 -4 6e. -56 -2 7lL x,15.-3 6 1205 1209 5 t50 751 ? f,85 lA5 -9 795 7A5 -l 829 850 -2 slJ 5tc 0 6f,0 Ec9 I 575 58t l0 971 992 I 542 55e 9 c2O 42! H.q'-2 -7 lat -177 I C32 c37 0 160 -r77 2 95. c 5 299 5t9 -5 71. 50 J 916 -930 2 992 1005 r 3!5 -361 q,2'-6 n'0,0 H,s,-q -6 aor -Al8 -5 5At 574 5 t53! t!05 c 126 c! 6 166 rao H'15'-2 -9 1022 t0c? -l 797 -A00 7 ?96 297 -6 270 27C 6 521 -102 -6 -cJ7 -2 q29 {eJ {lO -5 605 -612 -2 I 156 -l l7t 9 20t -tt{ -c le69 l2t? I 2c5 r99 q'12'-J -l 576 585 -6 ?051 1926 o 669 675 -l lrt 50! o 2t6 -zot -2 510 501 I 405 9t7 -c l0e5 luo9 2 252 Z2r -l c29 -qq6 ? 158 t29 0 67c -69? H'10,-J -2 1096 Il0? ! laz 19 -? -I2c lc9 c cl8 4oo I 5c2 -559 c tl!! -tr5l eo6 a7l 2 465 466 0 5t0 -551 6 749 759 3 tc6 I 16 6 r!5 ?5c t59 -5 285 277 -c 5t6 -586 2 561 -5r2 {'t5'_l d'!.-a E ll5 -525 5 90 12 I 682 -69A cJl -J 90. 90 6 945 -972 -2 525 JJI c 167 l7l r0 lA9 -lto 7 520 -515 t0 cZ9 -qq9 -l 906 954 I 55r -f,st 0 259 227 6 lt2 -lA? -l zao 2f2 -576 -l I 886 666 9 563 I loc al 2 Ic79 -lr{6 I 96r It -da5 5 645 H.2,-5 H'c.-t 5 at5 5!t c {25 -cl2 H' l2r'2 3 95. -71 -rs6 5 l!6 H,5,-3 269 27t 5 t0t5 t00l 6 505 -510 t 98t 96c -4 557 570 -8 l59O -l5c? l9l -152 7t292 -a 696 -106 a lt45 -1109 -{ 6c5 oct x,l5'0 -2 162 -Bc -7 512 526 -6 9{9 9c5 557 -551 -6 {25 {15 -2 C5J -{5J -? B' l. O 642 -676 -5 5c[ 5q6 -s 35c -!i8 66c -659 -c 915 -9lr B'10'-2 0 529 -5{A -! t72 -l15 _t -tos ? lzc t0 -5 25q -235 -2 llCl -tlrz lJ9 27 Hr7'-3 -2 206aE-20?5 2 C57 45q t09 -l -671 -l -6 -5c q 157 -c55 l5S 169 { 662 at6 85c O 6JJ 656 0 lt65 llSc 96. -562 -9lO -c -a2{ r 580 6 910 I 760 -?55 A A6A -C62 -5 ll26 ll!l 8t0 6 ??o 22! -819 -CF5 5 {25 rl5 E 809 5 c96 o ztf, 175 -5 105{ t06c -5?6 0 991 -100r H' l2'-l 0 1652 r7t! 5 !54 !7S I 0 5tg 5 723 725 6 6!l -892 261 265 -l 9ct -9{5 ? -r50 -1876 e r595 -1373 clz 7 l97o 6 lll -lB 0 79C -789 r lct9 r!92 c -6ec -9 u?5 -clc 9 560 568 H'?'-r 9 !t2 132 674 -6Aa I 651 666 J 155 5t 655 {05 -9Jl -2 rg7 -r7l H'c'o eoc ta? 5 C25 -420 5 426 AoJ 6 -6 -156 -510 791 H.lr_6 2Ol H.3'-2 246 -256 | 777 77A I 516 0 ole -4 -725 771 '710 721 _6 9!5 9!C a 100. -5f 2 -J -2 -167 56J 5?5 7aa -7 25c 2!Z -6 lAqE -t?u5 c ?68 7q9 -l 6{ 5 0 lolt -t05c -5 382 !90 -c 5r9 -5rl * The column headingsare h,2CfJF" and 200F"; k and / are listed at the head ofeach group ofreflections. Unobserved reflections are indicated by asterisksand strongly extinguishedreflections by Es. NEUTRON DIFFRACTION OF AN ORDERED ORTHOCLASE 50s

TABLE 4. Bond kngths, Site Occupancies,and Inter-bond for O cannot exceed 3 atomic percent on any one Angles of Orthoclase (Himalaya and Spencer C) and site. Adularia (Spencer B). Data for Spencer Samples from Colville and Ribbe. 1968. Comparison with other leldspars

Btulaya Spmcer B Sp6c.r C A comparisonof interatomic distancesand angles adularla between the present Himalaya orthoclase and the Bond length8 snd slte occupanctes r (r) -oA(1) r.664(1)l r.oor(z)i r. esqG)l SpencerB and C data of Colville and Ribbe (1968) -08 L661(2) r.645(6) r,54r(5) -oc r.673(2) r,670(3 ) 1.66r(7) agreementwithin experimentallimits, namely -0p r.612(2\ L 671(4) 1.566(5) shows

T(1) -0 I.667+.00I I.6635 I.6555 3 sigma or better, betweenSpencer B adularia and T(I) occupancy* slo.5gdo,lt s1o.otNo. g9 slO,6OA10,g if the crystal structureis 1(I) occupancy** S10.50[0.50 our crystal. Indeed, detailed r(2) -0a(2) r.627

Tetlahedrsl occupancy frm neulrm refhe@nt.

In no way do we wish to imply that a straightline function can be extendedbeyond 50 percent Al o< occupancy(Fig.2). ur 1.66 C' z Hydrogen content v,F 61.64 The valencesummation (Donnay and Allmann, o I 1970) based on the experimentalinteratomic dis- F tances (Table 4) and an ordered occupancy does qz 1.6? not showa significantlylow valuefor (Table ITJ Oc, 5). = Experiencehas shown (Donnay and Allmann,1,972) that deviationsfrom ideal valuesof lessthan 1Oper- r.60 cent are of dubious significance,and we are therefore not surprised that a difference Fourier map .t .2 .3 .4 .5 computedwith the final set of calculatedF s con- Arl(Ar+si) tained no negativepeaks with depressionsgreater FIc. 2. Plot of the mean 7-O distance as a function of the than -168 on a scaleon which a hydrogenatom atom fraction A1l(Al + Si) as determined by neutron dif' would show up with amplitude -6000. Moreover fraction. The circles are points from the present work and the sanidineresults of Brown et al. (1971)' the - 168 peak,which lay on a mirror planebetween the squaresare The indicated errors in the present work represent 2o ftom the T(2) sites,was in quite an inappropriateplace the physically impossible values resulting from the refine- for the hydrogenatom of a hydroxyl group. We may ment. The line is the regressionline of Jones (1968) and the thus concludethat systematicsubstitution of (OH)- shaded area ls a 2o confidenceband on either side of it. 506 PRINCE. DONNAY. AND MARTIN

Tesre 5. Valence Summation for Himalaya Orthoclase,with Tetrahedral Occupancies Based on Least-SquaresRefinement of Neutron Data

--Alka1 t ) t v(v,u, ) c Ko. sio.soNo.so -^1.00 esNto.to%. o4

---oAt{{ (1 ) (0.138)0.125* #0.878 1.881

il1oo{z) (0.204)0.r85 /10.982 2.1-49

IJJ q (0.094)0.085* 0.883 1.018 1,986

tiJ uc (0,076)0.069* 0.863 1.002 1.934

lJt 5 (0,120)0,109* 0,865 1.986

L(A) 2.911 1.666 1.618

r. til 2.795 2.130

6.444 3.149 3.160

v, (v. u. ) a,9619 3,5014 4,0014 ,, (1,060)0.961 3,489 3.999

*Trc bonds fron catlon.

tfTwo bmds to anlon.

tThe 4Z tacanciee in the alkall poGitlon increase iobs and thls effect nu6t be taken

account when lnto derivlnS the hEx value. The effective lotrlc radl1 of nine-coordinated a1kal1 ton6, properly weighled on the assunptlotr of fu1I occupancy, aEe added to that

of three-coordlnared orygen 1on to glve an ior"d of 2.887i. The ratto io6"/{1e6 ls

used to multlply Lbax, 3,309i, ileternlned on the basls of full occupEncy, to give the

vecancy-corEected value of LMx of 1,412i, gem pockets. The apparent anomaly of the persist- This occurrence strongly suggeststhat ordered ortho- ence of the monoclinic phase may be due to the clase does have a field of thermodynamic stability, presence of small amounts of the large Rb and Cs close to the pegmatite solidus; the usual rates of ions that would act as stabilizers of the monoclinic regional cooling may be too rapid for ordinary ortho- structure. Indeed, Gordiyenko and Kamentsev clase to attain the ordered configuration reached in (1967) have found pegmatitic K-feldspars to be the pegmatite gem pockets discussedhere. monoclinic if they contain more than 2.5 percent Although the intensity agreement between hkl Rb2O, but this is a value much higher than that and hkl reflections indicates that the bulk crystal is recorded in this case. monoclinic, on the average, the temperature factors Rate of crystal growth is one factor which might suggest that there is appreciable structural disorder. explain the association of maximum microcline and For example, the equivalent isotropic temperature ordered orthoclase in the same dike. Where growth factors of the Z(1) and I(2) sites are 1.08(3)4' and was faster, perhaps in responseto more rapid cool- 1.14(3)A' respectively, as compared with values ing, the monoclinic K-feldspar began ordering but ranging from 0.29(3;A' to 0.38(2)A' in maximum only got as far as a "normal" orthoclase. As the microcline (Brown and Bailey, 1964). Some of the rate of growth decreased,with the temperature still disorder may be due to non-homogeneous alkali above the orthoclase-microcline transition, perfect, distributions. In addition, the I(1) sites occur in gem-quality crystals of ordered orthoclase, quartz, pairs, and aluminum atoms can substitute for silicon beryl, and tourmaline formed in pockets. When the atoms at either site of any given pair. Therefore temperature fell below the orthoclase-microcline aluminum can be assigned exclusively to the I(l) transition, the "normal" orthoclase near the hang- sites and still give only short-range order, leaving ing wall transformed to maximum microcline. In the possibility of considerable long-range disorder. contrast, the ordered orthoclase, which had presum- The incipient long-range ordering indicated by ably attained a stable balanced Al distribution in the diffuse reflectionswith h * k odd, which was reported ?1 sites, persistedmetastably to lower temperatures. for adularia (Colville and Ribbe, 1968), was not NEUTRON DIFFRACTION OF AN ORDERED ORTHOCLASE 507

observedfor the Himalaya orthoclase,nor could we Geophysical Laboratory, Carnegie Institution of Washing- (unpublished). detectany evidenceof twinning. We looked for both ton, Washington,D.C. (1968b) Determination of cation distributions by phenomena using chips from the neutron crystal least-squaresrefinement of single crystal x-ray data. and exposingthem to CuKa radiation for periods Carnegie Inst. Washington Year Book, 66,216-217. of up to 100 hours on precessionand Weissenberg GonnrvENro, V. V., lNo I. Yr.. KluENrstv (1967) On the cameras. nature of rubidium admixture in potassicfeldspar. Geo- chem.Int. 4,408-412. Acknowledgments HAMrrroN, W. C. (1965) Significancetests on the crystal- lographic R factor. A cta C ry stallo gr. 18, 502-5 10. Ms. S. Horska of McGill University carried out the Jeuws,R. H. (1954) Pegmatitesof Southern California, 1r, wet chemical analysis; Mr. C. G. Hadidiacos of the Geo- Geology o,f Southern California. Calif. Diu. Mines Bu'll. physical Laboratory, Carnegie Institution of Washington, 170,Chapter 7,37-50. performed the electron-probeanalysis, Ms. S. Horska and rNo C. WlvNr. BunNseu (1969) Experimental ProfessorW. H. Maclran, also of McGill University, per- studies of pegmatite genesis.L A model for the deriva- formed additional probe studies on a chip of the neutron tion and crystallization of granitic pegmatites. Ecorr. crystal to test its homogeneity. R.F.M. thanks Professor Geol. 64,843-864. R. H. Jahns (Stanford University) for guidance in the lNo L. A. WnrcHr (1951) Gem- and lithium-bear- field, for helpful comments on this manuscript and for ing pegmatites of the Pala district, San Diego County, permission to quote unpublished analytical results. E.P. California.Calif . Diu. Mines Spec.Rep.7 4,72 pp, thanks Dr. L. W. Finger, of the Geophysical Laboratory, JowEs,J. B. (1968) A1-O and SiO tetrahedral distancesin Carnegie Institution of Washington, for severalhelpful dis- alumino-silicate framework structures. Acta Crystallogr. cussionson the use of the least-squaresprogram. Professor B24,355-358. J. D, H. Donnay and Dr. Walter C. Hamilton read the eNo W. H. Tevlon (1961) The structure of ortho- manuscript critically. The support of the National Research clase.Acta Crystallogr. 14, 443-456. Council of Canada (grants A7721 and, A6259) is gratefully MenrrN, R. F., ervo GlsnrrrrE DoNr.Iev (1972) Hydroxyl acknowledged. in the mantle. Amer. Mineral. 57.637-653, References Rrsse, P. H., lNr G. V. Grsss (1969) Statistical analysis and discussionof mean AllSi-O bond distancesand the AI-rEmN, H. 4., eNo E. Pnrxcs (1970) A time-shared aluminum content of tetrahedra in feldspars. Amer. computer system for diffractometer control. l. Res., Nat. Mineral. 54.85-9'4. Bur. Stand.74C,89-95. H. D. Mrclw, eNo W. H. Tnvron (1969) The BacoN, G. E. (1972) Coherent neutron scattering ampli- albite structures.Acta Crystallogr.825, 1503-1518. tudes. Acta Cry stallogr. A2E, 357 -358. SnlxNoN. R. D.. eNo C. T. Pntwlrr (1969) Effective ionic BnowN, B. E., lNo S. W. Benry (1964) The structure radii in oxides and fluorides. Acta Crystallogr. B25, of maximum microcline. Acta Crystallogr. 17, 139l-1400. 925-946. BnowN, G. E'., W. C. HeunroN, C. T. Pnrwrrr, lxo SIuesow, D. L. (1962) Graphic granite from the Ramona S. SUENo (1971) Neutron diffraction study of AllSi pegmatite district, California. Amer. Mineral. 47, 1123- ordering in sanidine. Geol. Soc. Amer. Abstr. Programs, 1138. 3, 514. (1964) Geology of the central part of -.theRamona Cor-vrrr,r, A. A., lNo P. H. Rrsss (1968) The crystal pegmatite district, San Diego County, California. Short structure of an adularia and a refinementof the structure contributions to California geology, Ed, J. L. Burnett, of orthoclase.Amer. Mineral. 53,25-37, Calif . Diu. Mines Spec.Rep. 86,3-23. Dren, W. .4'.,R. A, Howrc, .lNo J. ZussrvreN(1963) Rock- Spr,Ncpn,E. (1937) The potash-sodafeldspars. L Thermal Forming Minerals. Vol. 4, Framework Si/,cdte.r. John stability. Mineral. Mag. 24, 453-49'4. Wiley and Sons,New York. Wnrcnt, T. L.,.eNo D. B. Srewenr (1968) X-ray and opti- Doxu,lv, GeenreLrr, eNo R. Auu.rNN (1970) How to cal study of alkali feldspar. I. Determination of composi- recognizeO=, OH- and H,O in crystal structuredetermined tion and structural state from refined unit-cell parameters by X-rays.Amer. Mineral. 55, 1033-1015. and2V. Amer. Mineral. 53, 38-87. AND- (1972) A survey of valence-sumap- Zlcnlnr,rsr,N, W. H. (1968) Extinction and Borrmann effect plications. (abstr.) Ninth Congress,International Union in mosaiccrystals. Acta Crystallogr.A24,421-424. of Cystallograpbers,Kyoto, Japan. FINGER,L. W. (1968a) Computer programs for least-squares Martuscript receiued,Mav 22, 1972; refinement, Fourier summation, and error calculations. acceptedlor publication,Ianuary 8, 1973.