Quick viewing(Text Mode)

A Spectroscopic Study of Irradiation Colodlg of Amazonite: Structurally Hydrous, Pbbearing Feldspar

A Spectroscopic Study of Irradiation Colodlg of Amazonite: Structurally Hydrous, Pbbearing Feldspar

AmeficanMineralogist, Volume 70, pages 794-804, 1985

A spectroscopicstudy of irradiation colodlg _ofamazonite: structurally hydrous, Pbbearing

ANI.le M. Horwlstrnr AND GEoRcE R. Rossurc'N Diuision of Geological and Planetary Sciences2 Califurnia Institute of Technology, Pasailena, California 91125

Abetract Irradiation-inducedcolor in amazonitecan developonly in potassiumfeldspar having both structurally bound HrO and Pb impurities. color is controlled by either (1) an absorption minimum in the p spectrumbetween three overlapping bands in the and a broad band at 625to 643nm, resulting in a blue color, (2) a combination in p of one UV band and a broad band at 720 nm, resulting in a greencolor, or (3) both of the above superimposed,resulting in a blue-greencolor. All optical variations correlat€dwith an EPR p"it".n indicative of Pb3+ or Pbr *. The different types of color are associatedwith a limited range in Pb content and structural stgte. For constant Pb content, the intensity of color is linearly relatedto the amount of struct-urallybound H2O, up to a limiting value.Dependence of color intensity on both Pb and HrO concentrationstrongly suggeststhat and water occur in a 1:1 ratio in the color centers.The first order reaction kinetics ofamazonite color formation by irradiation and the observation that water is not consumedin the process suggeststhai water plays a catalytic role in the irradiative transformation of Pb2+ to the amazonitechromophore.

Introduction with optical absorption spectra of amazonitesand other potassiumfeldspars in an attempt to establishthe origin of The color of the blue-greenvarieties of and color and its relationshipto water in the . (amazonite) is radiation-induced (Przibram, 1956,p. 253).Chemical data indicate lead to be the color- ExPerimental ing agent (Foord and Martin, 1979),yet Pb2* cannot be Optical absorptionspectra wcre obtained from opticallyorient' becauseall electronictransitions the sole causeof the color ed polishedslabs of feldsparusing polarizers in a Caty l7l (pers. of Pb2+ occur in the ultraviolet region. E. E. Foord spectrometer.Infrared spectra were obtained with wire grid comm.) has shown that amazoniteswith less than 10fi) polarizersin a Perkin Elmer 180 sp€ctrometer.Both visibleand ppm Pb are blue, whereasthose with a higher lead content IR spectrawere digitized and scaled,but only the visible data are ,and that the apparentintensity usually increases requireda baselinesubtraction. Data from computerpeak-fitting with lead enrichment. However, some feldspars with as did not give accurateareas for the peaksbecause the baseline much as 1fiX) ppm Pb are not colored (Foord and Martin, (whichis a combinationof a UV tail and scattering)could not be drawn in by 1979).Using electron paramagneticresonance (EPR) tech- adequatelymodelled. Therefore, the baselinewas + was estimated by multiplying niques,Marfunin and Bershov(1970) concluded that Pbr hand, and the integrated intensity the peak height by the full width at half height (Wr,r). Becauscthis centers are present in amazonites,but not in other feld- approach precludesresolution of two overlapping componentg Tarashchanet al. (1973)found that the intensity of spars. averageintensity and total areasare reported for doublets' an ultraviolet absorption band in amazonite,attributable 78 K on a Varian EJine spec- + EPR spectra were taken at to a Pb2 transition, increasedupon heating but decreased trometer at about 9.2 GIiIz' All samples werc coarsely-ground upon subsequentirradiation, and suggestedthat radiation powders weighing about 100 mg. Doubly integrated intensities converts Pb2* to Pbr*. However, Spelt and Lehmann were calculated from the empirical formula + (1982) did not find the Pbl EPR signal in Australian DII amazonite.Equally mysteriousis the correlation between loss of color and weight loss during dehydration of ama- (signalheight)' (signal width)' r/2 zonite (Plyusnin, 1969) becauseneither H2O nor OH- (g"tr). (r"-pte weigbt)'(modulation amplitude)' (power) alonecan producecolor. (1) The present paper presentsand correlatesEPR spectra (Eaton and Eaton, 1979).Also, a line integxalwas applied twice to r Present Address: GeophysicalLaboratory, CarnegieInstitu- doubly integrate digitized and baseline-subtractedfirst derivative tion of Washington,Washington, D.C. 20008. spectra. Becausethe numerically calculated value dependslinearly 2 Contribution Number 4050. on thc formula-calculated value, and the numerical method is less 0003-{xxx/8 5/0708-0794$02.00 794 HOFMEISTER AND ROSSMAN: COLORING OF AMAZONITE 795

precisedue to baselineapproximation, the data -l frorn equation (l) WAVENUMBER,cm are given. 3@OO 2SOOO t5000 Samples were irradiated at room ternperature in air with a 13?Cs source which produces0.66 MeV g:mrma rays in dosesof AMAZONTTE l.4l MRads/day.Dehydrations were done in air at 200to lm0.C. Loko Grcgc Water contents of selectedsamples were measured with a hy- UJ O f . o.3nn drogen extraction apparatus(e.g. Friedman, 1953).The measuie- 2 08 ments are precise(approximately 1%), but the amount f of water P.u evolved may involve a blank of I pmole HrO (R. E. Criss, pers. o comm., 1982) which increasesthe uncertainty a to about *5%. 6 04 Three were measured several times to determine the re- producibility. Samples5 and 25 gave consistentrcsults, but thc water content varied by a factor of two for sample#12 which was inhomogeneouslyturbid from fluid inclusions.Water contents of all other feldspars were determined by comparison of their infra- AFIAZONITE red spectra to that of a feldspar with measured H2O content and Broko Hl I I td \ #3" r.o.5nn similar spectral features. (J Most of the feldsparswere provided by E. E. Foord and are L u" \ dcscribedby Foord and Martin (1979)and Foord et al. (in prep). dt \ The feldspars embrace the known range of color and lead content I eu o (Table 1),and include the exceprions(# 14, lS, tB, 22, 9, 2O,261 to Q o. trends of hand-spec'imencolor with chemistry(E. E. Foord, pers. commun.).

Visible spoctrmcopy of amazonite AMAZONITE

General properties

Greenish-blueamazonites are the most common,but the apparentcolor encompassesvalues between this and green with a slight amount of yellow (Table 1). Occasionally superimposedon the green is a grey cast due to various -cation hole centers (Hofmeister and Rossman, 1985). AHAZONTTE Absorption spectraof all amazonitesexamined have the lo d Koivy,USSR following featuresin common: (l) The lrJ t2l rrO.5nn color is intrinsic. (2) (J 08 All bands are strongly polarizedin the beta direction (per- z pendicular o to (001)),and weakly polarizedin alpha,perpen- d 06 o dicular to (010).Very weak componentsin the gamma di- o (n g4 rection were ignored becausethese are due to scattering, being highest for the most turbid microclinesand zero for g2 the nearly gem quality .(3) The spectra are dominated by a single band or two overlapping broad bands with total width W,,, of 3000 to 4500 cm-l which 209 3gg 400 sog 69,0 700 aoo 90,0 togg ttoo }JAVELENGTH, are slighly asymmetric and centered at 625 to 720 crrn.(4) nm The combination of near-IR bands with UV absorption Fig. 1. Polarized absorption spectra of natural amazonites at 23"C. (a) producesa transmissionminimum in the blue-greenregion, Blue, type B amazonite #5 from Lake George, Col- orado. The giving amazoniteits rangeof colors. finely spaced dots indicate where the alpha spectrum was estimated from a spectrum of the same chip after irradiation. There are four distinct color types (Table 2): (l) type B, The slow rise toward the UV is due to scattering from turbid common blue, low Pb microcline-perthites possessingthree regions, and from the perthitc lamellae. (b) Green, type G amazon- broad polarized absorptionsat625,385, and 330 nm along ite #3' from Broken Hill, Australia This sampleis close to gem with a UV tail (Fig. 1a); (2) type G, end-membergreen, quality so that very little scattering contributes to the spectra. The high Pb orthoclase with Arrr,r5, - 0 (Cech et al., l97l) weak band near 360 nm was induced by baselineoorrections. The consistingof one band at 72Onm plus a UV tail (Fig. lb); UV tail is markedly polarized in the same schemeas the amazon- (3) type D: superpositionof thesetwo end-memberspectral ite peak at 720 nm. (c) Medium grecn amazonite #32 trom peatl, types gives a doublet (Fig. lc) in intermediatepb-content Routt County, Colorado. The spectraare the superpositionof the B and ; the variation in relative intensity of the 220 G types such that the relative intensity of the 625 and 72O nm bands gives one apparent very broad band. Spectra and 630 nm peaksproduces a gradual shift in the transmis- taken at 77 K resolvethe two bands.(d) Blue-greentype T amazonite sion minimum, resultingin the continuousvariation #21 of hue from near Keivy, USSR. The most intense absorption from blue-greento green; (4) type T, a color is for the type T color is shifted to 643 nm and is broader than thc produc€d peak by a at 643 nm (Fig. ld) which is slightly singlet of the blue amazonites. Broad bands are also present at broader than the 630 nm peak but is not resolvableinto a 385nm (shownhere) and revealedat 330nm by irradiation. 796 HOFMEISTER AND ROSSMAN: COLORING OF AMAZONITE

Table l. Sampledescription and chemistryof amazoniteand other feldsparsin order of increasingPb content

ity Color Description Eleclron llicroprobeAnalyses EmissionsPect.t Hz-lilanometry! Sample lluseumLocal H2 l{umber tiumber i,lunselI f O" Ab An PbO Fe01s161 Pb Fe Total

l'line, Pala, CA gemmyparts 0.l0 7 15063 Elizabeth R liR 9.0 Hhite perthite 0.005 0.001 l'line, Pala, CA gemmyparts 15 25 9370 llhite Queen NR9.0 t{hite perthite 0.13,0, I'tine, Pala, CA gemmYParts

15057 t{igwan 7.5BG9l? Pale blue' turbid (e9 I 0) 0.01510.0I Creek, C0 mottled Perthite

14 15058 Hawthorne, 2.588/4 Bright blue 0.020 0.015 t{evada perthi te

15 15105 NewYork 5\912 Grey ( 95 5 0) 0.030 0.030 llts, CA perthite

16 15055 Landsverk 7.58G7/8Blue turbid 0.030 0.015 Perthite D[{NH- l,figwam 7.58G7/7 Blue zoned (97.2 2.5 0) 0.045* 0.030 10934 Creek, C0 Perthite

23 10102 Chittaranjan, 7.58G716Blue, zoned 94.3 5.6 0.1 O.05 BLD Bahir. India coarse Perthite

5 15020 LakeGeorge 7.58G6/8 Dk. blue Perthite (95 5 0) 0.050 0.030 0.06,0.07 Colo rado with whlte overgrorths on selected faces 0.07 0.09,0.15 ,0. 17 12 15020 LakeGeorge - 7.58G718B'lue turbid 95.7 4.3 0 ND trace 0.07 Colorado Perthite (95 5 o)

18 15056 JoseBukuru IOGT/4 Grey-greenoPaque (97 3 0) o.07 0.15 Paddock,Nigeria microperthite 5'2 0 0.085BLD 0.10* 0.07 22 - Anelia, VA l0G7/8 Greenperthite 94'8

24 Amelia,vA NR9.5 l{hite section 95.2 4.8 0 0.12 BLD3 0.10* 0.03 fron #22

19' 13758 NeviYork 2,5RG716Blue-green 90.7 9.3 0 0.19 0.01 0.07 0.003 ilts, CA turbid Perthite

19 13758 NewYork 2.5RGl/2 Blue-green (93 7 o) 0.20 0.015 i{ts, CA tufbid Perthite

9 15020 Lake6eorge, 5BG6/8 Green-blue 93.7 6.3 0 0.20 0.05 0.12* 0.02 0.09 Col orado perthtte ( 95 5 o)

21 15021 Keivy, 2.58G516Dk. blue-green 95.5 4.5 0 0.20 0.03 0.10 0.03 0.l1 Kola-Penn., v./ coarsewhite (gs 5 0) USSR al bite blebs

20 15057 Monapo, 7.58G8/6Med. green (92.8 7.1 0 r.0) - 0.30 0.015 llozamblque turbid Perthite

32 15065 , Routt 58912 Greencoarse 94.2 5.7 0.1 0.81 BLD 0.70 0.07 Co., Colorado Perthite

6 15066 Keivy, 7.5G618 Malachitegreen 95.3 4.6 0.1 0.83 BLD 1.0r 0.03 Kola- Penn., w/ coaPseihi te ( 95. 5 4. 5 0 r.o) USSR blebs

26 15059 El Rancho, 5C9/2 Grey-green 93.9 6.1 0 1.10 BLD 0.705 0.015 Jefferson Co., turbid with Colorado grey inclusions

3 13756 BrokenHill, 2.5G712 Grey-green 94.2 5.7 0.1 1.79 trace 2.0 0,07 N.s.w., orthoalasewith (94.4 5.6 o 1.75) Australia grey incl usions

3' 13756 BrokenHill 2.5G7/6 Med.green 91.3 8.7 0 1.60 BLD t{.S.tl.,Austral ia orthocl ase

3' 13756 ErokenHill 2.5C518 Green 92.7 7.6 0.3 1.80 BLD 0.t2 N.S.!1.,Austral ia orthoclase

ffithat0r'Ab,Anareinmole'.P|obevalueslnParenthesesarelrom|0or0[yers.Lonm.,

JPc!!r UYI oP" '! rreu (ruuru'il f,iiiin, conklln,and simnons, in prep) ;[Ti':;ll.;3i':;::il8l:;,i"i]i::'"iIof'?["(13ff'f"3"$'ff,r'jlatAll emissi6nemisslon speitrirscopyspectroscoPy daia,data' exceptexcepf fort0r samples.with*samples.wlrn- dreare ]lJ,i)r'.8lllEllir sli-=6'f_b'=-tiiioi-iirit beiowlinit ofot detbition.Oetbttton.- ND = not determined.deterinined. tl'lultiptl,lultiple values are fromfrqm duDlicatedupl runs. HOFMEISTER AND ROSSMAN: COLORINO OF AMAZONITE 797

Table 2. Optical absorption parametersof natural amazonites

Color Sample Visible Peak UV-Vis Peak UV Peak Type Number 1t2 I ll ^ Ll?. , W cm-I iir:l cm-l cn-2 Cm-I Cm-I cm-l cin-l cm-1

B 88 16000 >2500 1.t3 >2825 -26200 -3000 0.6 -1800 NDlI q -0 -0 ND B 58 16050 3350 8.4 28140 ?5780 3500 5.5 19250 30770 ?700 o.r 16470 c 16000 t,7 5700 1.t 3850 ND B LzB 16000 3350 10,I 33840 26000 3500 6.6 23100 ND 2.2 7370 1.3 4550 ND B 98 16000 3450 8,72 30080 25800 3150 5.9 18585 30860 2850 (o 1681s l. 15 3970 1. I) JOZf, ilD T 21 9 15530 3760 13.16 49480 25700 -4000 8.16 -32600 30824 ND* q ?.7 10260 t.32 -5300 ND* D 238 15700 3760 3.50 13160 26000 3350 -2 -6700 30000 ? 4L -6 000? 15880 3400 0.7I 2430 0.4 1340 NO D 228 14840 4380 4.40 t9270 25700 3900 3.52 13730 31100 3000 3.96 -1 900 1.10 4820 -0.50 -2000 -0.50 -1500 D I9'B 14840 -4500 6.84 -30800 -26000 3350 4.2t 14100 ND c 1.68 -7600 weak ND D 69 14840 -4230 r?.1 -51180 26000 -3700 -9 -33300 ND c 2.2 -9300 -1.6 -5900 ND D 328 14180 4090 5.26 21510 -3800 -16000 . 25600 4.21 29000 3050 4.56 I 3900 -3800 1.75 -6660 1.75 -6650 1.75 5350 G ?68 13710 3100 4.20 13020 28500? v wkt 31300? vwK 0.95 2950 ND r{D G 3'B 13700 3000 11.10 33300 ? 29400? c 2.94 8820 ?

1/r is the peak position, t,l|l/zis the full width at half hetght, I is intenstty, and II is the integrated intensity * This band is |{eakly present in gamma,and strongly present in an irradiated sample in beta and moderate'lyin alpha f Very weak ll Not determined

doublet; like the type B spectra,the type T has both 385 out prior heating, amazonitecolor can increase,decrease, and 330nm bandspresent. or remain the sameafter gamma irradiation (Tables3 and The parametersof the peak(s)at 625 to 72Onm in beta 4);3 this implies that only color saturated through radi- are sufficientto quantitatively describezrmazonite color be- ation should be comparedto the chemistry. cause:(1) The integratedintensity in the alpha polarization The kineticsof amazonitecoloration weredetermined by is linearly correlatedwith that in the beta polarization for monitoring the intensity of the near-IR peak as a function eachof the 625to 720 nm peaksand the 385 nm peak (Fig. of y-ray dose (Table 4). Coloration is initially rapid but 2).3Sullicient data were not availableon the 330 nrn p""k saturation occurs at about 1fi) MRads. The data are best to establishthis correlation.(2) The integratedintensity for fit by a first order rate law: each of the 385 and 330 nm peaks is a linear function of dX that of the band(s)at 625-720nm (Fig. 3).3Table 3 sum- : -k(c - x)' (2) mariz€s the optical properties of the 630-720 absorption dt for all amazonitesexamined.3 where X is the number of color sites present at a given Responseof color to heating anil irrailiation time, and C the maximum Dumber available.Integration from 0 to t and rearranghg gives Amazonite color can be removedby heating and regen- erated by irradiation only if the heating is applied below ln (l - X/C): -kt. (3) about 500"C and for lessthan about l/2 hour. Otherwise, Representingthe number of color sitesby the intensity,and color is at best only partially restorable.After extensive estimatingC from the saturation color givesthe graphical heating the blue or green color cannot be recovered:in- representationof Figure 4; the slopesare sample-specific. stead,a smoky color similar to smoky occurs.With- Relationship of amazonite color to leail content 3 Amazonite becomesgreener and usually,but not always, Figures 2 and 3 on amazonite peak parameters, Figure 9 of more intensely near-IR spectra of water in feldspar, Table 3 of optical propertieg colored with increaslnglead content (E. E. pers. Table 5 of EPR properties,and Table 7 of the dehydration of Foord, commun.). The hue changes with increasing amazonitemay be obtained from the authors or by ordering doc- Pb because(1) type B spectra occur for amazoniteswith ument AM-85-271from the BusinessOIIice, Mineralogical Society low lead contents; (2) tyW G only occurs for samples of America, 2000 Florida Ave., N.W., Washington D.C. 200@. having greater than lo/o Pb, and (3) for type D and T, the Plcaseremit $5.00in advancefor the microfiche. average peak energy decreaseswith increasing Pb, due to HOFMEISTER AND ROSSMAN: COLORING OF AMAZONITE

Table 4. Irradiation r€sponseof 625 to 72Onm band of selected N agoo I amazonites € 0 sogg Sanpl e U1t, Dose, I, C, l- t/c k1, llumber cm-r [Rads cm-l cm-l l4Rads-l F 12 heated 3300 0 0.58 0.0277 | +oos (450.C) 4.8 0.51 v) 20 0.33 z 45 8.I 0.165 ul 70 8.9 0.082 sooo t t21 ot 0,021 I H 0 9,5 14.0 0.321 0.0230 <) 11 10.I 0.279 o 29 t1 t 0.200 u 2OOO 40 12.4 0.114 t- 67 13.1 0.064 94 i3.4 0.040 u I45 13.85 0.011 o tsoo trJ 3" heated 3000 0 6.5 0.508 0.0230 t- ( 300") 18 0.369 z 29 4.8 0.262 Hs 56 5.4 0, 169 83 5.0 0.o77 o o o.s Lo 1.5 2.O

0 6.7 9.37 0,284 0.0852 15 8.4 0.103 25 9.0 0.0039 Pb0, wt % 15 9.36 0.001 Fig. 5. Dependenccofamazonitc peak parameters on leadcon- 0 5.95 = = 10 q7q tent of natural amazonites'Triangle type B. Diamood type 35 4.58 D. Square: type G. X : type T. For integratedintensity for low 0 13.16 l€adcontents, intensity generally increases with leadcontent, but 1I.7 is statisticallyinsigrrificant and beyond0'5 wt% 9l Lr.2 this correlation no clearcut trendis visible. I 15 Ene lnlenslEy nedsureo. L 15 fne esfrndEeo ndxrmun lntensity obtalnablethrough irradlatlon. kl is the slope of the line defined by ln(l-llC) as a function of dose of l.4l Ilrads/dav. generally increaseswith lead content for the low Pb sam- ples,but there is tro c,ommontrend for all samples;because an increase in the 720 nm peak relative to the 630 nm the eye is more sensitive to Sreen than to blue, visual as- Pb content.The intensity of the componentwith increasing sessmentsuggcsts a continuoustrend. color is related to the integrated intensity of bands near 630-720 nm. Figure 5 shows that the integrated intensity EPR spoctroscopy of amrzonite colol centers EPR spectra were measuredfor each of the four different amazonite color types and for sampleswhich had been first heated to remove the amazonite color, and then irradiated to reveal centersextraneous to amazonitecoloration. For to the x Broken Hill orthoclase,no signalcould be attributed o e green color. For all microclines examined, a distinct set of ul signals(consisting ofa large first derivativenear g.11: 1.56 z with two smaller satellite absorption-emission features at ucl o : 1.83and 1.38)was observedin the blue-gr@nsam- o &nr o ples, but not in the heat-bleached and irradiated samples (Fig. 6a). The presenceof two or three of these sets in some t! o probably the principal values of z samples (Fig. 6b) reflects o the anisotropic g-factor and hyperfine interaction, A (G. u pers. colnmun., 1984). The absence of a Lehmann, c) amazonite-type EPR sigrrals in orthoclase may be at- tributed to sigrral broadening resulting from disorder (cf. Fe3* in feldspar: Marfunin et al.' 1967; Gaite 1970)in that broadnessin a low intensity 0 40 80 t20 and Michoulier, DOSE,MRods signalcould easilyrender it invisible. The sum of the DII for each of the three EPR regions Fig.4. Logarithmis plsl of 1-absorbance/absorbance*as a using equation 1 (Table 5)3depends linearly on function of cumulative radiation dose. Triangle:B #12. Dia- calculated optical intensities in the alpha mond: D #19'. Square: G #3'. The constantsdiffer for each the sum of the integrated sample (see Table 4), and the linear correspondenceimplies first- and beta polarizations(Fig. 7). The slight deviation of the order kinetics. one type D sample may be due either to uncertainty in HOFMEISTER AND ROSSMAN: COLORING OF AMAZONITE 799

s FACT0R due to the averagingof anisotropic spectraby the powder 156 t39 method. From Figure 7, the ratio R of the more accurately -/ determinedsatellite to the central component rangesfrom 06 ./ ./ 5.5 to 7.1. Theoretically, this ratio is predicted from the isotopic abundancesby F : a o4 a (2r+ r).%(r:0)/"/,$) (4) z LAKE GEORGE- I2 LJ (Goodman and Raynor, 1970,p. 219).The elementwhose z abundancesshow the H 02 closestcorrespondence is lead (22.60/o l:112,7'1.4% l:0, R:6.85). For 10eyear old Pb (ap- propriate for most amazonites),the model of lead isotopic change by Stacey and Kramers (1975) gives a value of R:7.1. The only other elementswhich give comparable 08 ratios are Sn, W, and Pt. Not only is the value of R for Pb YORKMTS - I9 closestto the experimentalratio, but Pb is strongly con- E oo centrated in amazoniteswhereas Pt, Sn, and W have not a z been detected (Foord, pers. commun.). Thus, Pb colors P oq amazonite blue to green; the charge state must be +3 or z KEIVYUSSR - 2I H + 1 in order to yield an EPR signal, but powder sp€ctra cannot be usedto differentiatebetween o? LAKE GEORGE- E Pb+ and Pb3+. The relationship of color to the concenttotions

34gO 3600 3800 4ae9 42go 44@O 46e0 4AOO SO@O of Pb and structural water MAGNETICFIELD, Gouss Heating above 500'C removesan essentialingredient of Fig. 6. First-derivativeEPR powderspectra taken at 77 K. amazonitecolor. Structural changesdo not occur at such Spectraare scaled for 100mg sampleweight and run conditionsof low temperatures,as indicated by similar IR spectralpat- 1.0mW, lG modulationand x 1000gain. Frequencies were 9.194 terns of vibrations before and after heating (exceptfor the GHz exceptfor #19 (v:9.18 GHz).The large,broad band near water bandsnear 3600cm-r: seebelow). However, volatile B"rr: 1.8is probablydue to hematiteinclusions. (a) Amazonite speciesare likely to be lost over the temperaturerange in #12 firomLake Georgg Colorado. line, natural blue color which amazonitecolor is affected. (typeB).Dashed line, the same sample heated at 900.for ll2hour andirradiated (grey). Features near g.rr: 1.83,1.56, and 1.39arc connectedwith amazonitecolor. (b) Top, #19', doubletfrom New York Mountains,28 MRad dose.Middle, #21, type T, Keivy, USSR,natural. Bottom, #9, blueamazonite from Lake Georgc, Colorado,natural. Smaller features near 3480 and 362() Gauss are | .83 / t .ss ,/t .s6 dueto theAl-O- holecenter (Hofmeister and Rossman, 1984). All N sampleshave the samefeatures as amazonite#l2,btt #9 aad t, aooo #21 showadditional similar features at slightlydifferent posi- o tions,and with muchlower intensities. H H +oog J peak areas or to differencesin optical extinction coer O ficients for the B and G peaks; becausethe deviation H qlnnot be explainedusing the intensity of only the 625 nm F 4 2OOO component,the "color" of all varieties(types B, T, D, and o thereforeG) arisefrom the samecolor centers. Marfunin and Bershov (1970)reported single g- valuesof 1.39,1.56, and 1.837for an EPR centerconnected o with amazonitecolor, and attributed this center to pbr+. o.oo o.20 a.40 o.ao o.ag Theseg-values are similar to ours, suggestingthat the cen- ters are the same.The following factors support their as- EPR DII signment of the amazonite color center to Pb: The large Fig. 7. Dependcnceof EPR doubly inrcgrated central first derivative peak with two satellites (Fig. 6a) intensitieson the sum of the integrated intensity of absorption bands in beta and suggeststhat the isotopes of the element involved are alpha near 630 to 720 nm. Trianglg type B. Damond, type D. X, dominantly of zero nuclear spin, with a small proportion Type T. The lines are labeled with g.n for th€ thrce EPR signals. having l: U2. The lineshapesof the two satellites are EPR DII scaled for 100 mg same weight, and run conditions of I absorption-emissionrather than first derivative, probably mW, lG modulation and x 1000gain. 800 HOFMEISTER AND ROSSMAN: COLORING OF AMAZONITE

Speciation of water in amazonite betweena of (010)and a of (001),urs s€en in Figure 8b. For A typical IR spectrum of microcline-amazoniteshows fluid inclusion water in microcline, an averagemolar ex- mostly broad bands (Fig. 8a). Cooling the sample with tinction coeflicient(e) of 23 liter/mole' cm ( + 5%) in a on liquid nitrogen splits the broad bands into three compo- (010) at 3440 cm-1 was determinedfrom IR spectra and nents, nearly doubles the intensity, and shifts the ccntral total water contentsof samples#5,9, and 12. formation of ice, There is a variable amount of in microcline- absorption to 3200 crn-r. This indicates 1. showing that the broad, isotropic (with respectto one face) amazoniteIR spectraat 3630 and 324f cm- IR sPectraof (Fig. bands are due to fluid inclusions (Aines and Rossman, nearly gem quality green orthoclasefrom Australia positions 1984).Band intensitieson spectrataken from the two dif- 8b) show anisotropy in intensity at four band ferent facesare not equivalentbecause a higher of (3630,3550, 3440 and 3240cm-r). Thesepositions are th€ fluid inclusions occursperpendicular to (001),as is readily sameas those of bands of structural water in white micro- observablethrough a microscope,producing the difference cline (see below), and show that a small fraction of the water in {lmazonite may be incorporated in the feldspar structure. IJAVELENGTH,urn Infrared spectra from the least turbid regions of three 2.5 3 0 4.O white microclinesfrom the Elizabeth R mine and one from the White Queenmine showstwo pairs of polarizedbands (Fig. 8c). One pair (3630 and 3550 cm-r) is polarized Lokr Grorgr o.4 O ^ strongly in alpha and weakly in beta. The other pair (3400 lrl and 325Ocm-l; is less strongly polarized. Near-infrared 4. u.. spectrataken on thicker sectionsof the samecrystal (Fig. o u 9)3show two strongly polarizedbands of HrO at 19fi) and 8 u., nm (-5250 cm-1;, but OH bandsare not detectedat co 1950 22ffi nm (4550crn- r) in agreementwith Solomon and Ro- o.l ssman(1979). Correlation of the intensity of the 19fi) nm absorptions with that of the fundamentalsalong with the of the 2200 nm X-OH band indicates that hy- AMAZONTTE absence is presentas H2O in thesemicroclines (Hofmeister, b Brckrn Hi | | drogen 9.4 BecauseIR spectra taken on white microcline at lrJ l. O.O5cm 1984). (J liquid nitrogen temperatureswere not noticeably different 7. u. water indicated by fi) from the room temperature spectra, the E. bands at 3630, 3550, 3440, and 324Ocrrn-t in the IR and 3 u, m 5130 and 5260 cm-r (1950 and 1900 nm) in the near- as isolated molecules in the feldspar struc- gl infrared o@urs ture (structural water). From sample 7 that had the least amount of fluid inclusion water, the molar absorptivity, e,

a was calculatedas 120liter mole'cm in a at 3630cm-l. A a t{ICROCLINE El lz R rlac possible site for the water is the M site, due to its size; td t =O.O5 cm charge balance could occur by substitution of a divalent (J s.e cation (Pb2* ?) for a nearby K+, or by replacingAl3* with 1 si4+. uo 8 s., 6 Dependenceof color on structural water content o.2 Infrared measurementsof the water regions in all ama- zonites were made (Table 6). Structural water con- centrationsgreater than about 7 ppm are accurateto + I M 3700 34f,€ 3t@ 2A8(, z€,g0 ppm; however,concentrations below 5 ppm are subjectto WAVENUMBER. a much larger error due to the difficulty of measuringthe "'-l 1 Fig. 8. Polarized infrared spectra of water in feldspar. (a) Ama- height of the 3620cm- peak on top of the fluid inclusion zoaite #5 from Lake George, Colorado. The lower pair was band. Uncertainties in the fluid inclusion water contents taken at room temperaturg and the upper pair was taken after thc are *l0%o, due to inhomogeneitY. samplewas cooled with liquid nitrogcn. (b) Green orthoclase #3 The variation in color, fluid inclusion water, and struc- data were scald from Broken Hill, N.S.W., Australia. Spectral tural water for the 3 samples from Broken Hill (which from 1 nm and 1.6 mm thick samples.The ya' pair is from (fi)l) otherwisehad the samechemically) suggests that color de- while cf are from (010). (c) Gemmy area of white microcline 17, and not on fluid inclusion water Elizabeth R Mine, Pala, CA. Absorption bands present at 363Q pendson structural water 3550,3,f60,3400,3250,and 3050crl-r are due to two diffcrent (Fig.10). typesof water molecules. Pieces of typc B sample #12 werc heated at various HOFMEISTER AND ROSSMAN: COLORING OF AMAZONITE 801

Table 6. Infrared water absorptivities,derived water content,and closest M site occupied by Pb ranges from 0.06 to 1%. infared order parametersof amazonite Thus, either HrO must affect the radiative transformation of Pb2* lrom afar, or Pb and HrO occur as a coupledpair Sanple I(3620)n Structural I ( 344q)S Fluid Incl usion d/mt Number cm-l H20, ppm cm-l H20, ppm in the feldspar.Insofar as irradiation does not changethe IR water spectr4 then if HrO affectsthe irradiative transi- 8 -0.2 2.47 820 3.63 tion regardlessof distance to Pb eventually all ordinary 14 0.7 42* 5.00 1670 2.72 15 -0.03 -2' -25 lead would be transformedto the chromophore.This does t6 0.23 14 7.0 2300 4 0.10 1 t1 700 4.54 not happenbecause under irradiation the intensity ofcolor 23 0. 16 9 1.76 590 3.64 in orthoclasesgrows only slightly larger than that of micro- 5 0.31 I8 900 1.58 12 0.34 19 4.4 1470 2.88 clines with 1/100 the amount of Pb. Thus, it is likely that 12 heated 0.196 11 2.82 940 18 -0 -0 -56 -18000 0.49 much of the structural water is locally coupled with Pb, 22 0.28 15 2.35 780 2,30 possiblyduring growth, by the substitution. 24 -0 -0 4.05 1350 19 0.20 12 9.1 3000 3.07 19' 0.25-0.42 t3-23 5.8 1930 Pb2* + H2O:2K+. (5) 9 0.15-0.2 8-12 2.76 920 ?.94 21 0.13-0.27 7-15 1.60 530 20 -0.15 -8 -t.7 -560 3.06 Calculation of extinction coeficients and comparison 32 -0.20 -L2 840 1.18 of amazonite color centers to Pb3+, Tl2+, and Tl" 6 -0.12 -7 l 58 530 3.62 26 0.25 4.90 1630 0.83 centers in KCI 3 0.31 2.5 840 0.21 3' 0.4I 700 0.15 Extinction coeffrcients were calculated from the two 3" 0.64 3s 2.0 670 0.57 trends in Figure 12, using effective molar lead contents tl Absorptionsdue t; fluid inclusions and subtracted; intensity measuredon alpha equal to the molar water [email protected] type B ama- $ Absorptjonsonly due to fluid inclusions; intensity takenfrom q zonite,e:4300 liter/mole.cm, and for type G, e:25(X) t This quantity was detemined from lR spectra as shownin Figure 13 and was uses as dn indication of Al/Si order/disorder liter/mole.cm. Oscillator strengthsf # calculatedfrom I# :4.6 x l0-e e W172 (6) 'Il2+ are given in Table 8, along with those of and Tl' in temperaturesfor ll2 hour, and then irradiated at a fixed dose. Monitoring the changein water presentby infrared spectroscopy showed that heating decreasedthe con- centrations of both fluid inclusion and structurally-bound water at similar rates (Table 7)3.The color developedafter FLUID INCLUSfONH20, ppm irradiation is linearly dependenton structural water con- centrationsfor less than 12 ppm HrO (Fig. l1), with satu- la6a t506 ration in color occurring above 12 ppm. This shows that most but not all of the structural water presentis involved A in amazonitecolor centers. T Euidence Pb: H : I : 3 in for 2O color centers ()E For the type B and G amazonitesin which the molar concentration of Pb exceedsthat of structurally bound HrO, the integratedintensity of the color F dependslinearly H on the concentrationof structural water (Fig. @ 12).Samples z A 14 and 15 have a lower molar concentrationof Pb than of IJ F HrO; these were included in Figure 12 by plotting thc z A molar amount of water equal to their molar {rmount of H lead. That samples #14 and 15 fall on same trend as the other samples strongly suggeststhat Pb and structural H2O occur in a one to one ratio in amazonitecolor cen- ters. The occuren@of some doublets outside the trends ta defined by the type B and type G samples may be related to the partitioning of water among the two different types STRUCTURALH20, ppm of color sites.The one type T samplealso lies outsidethe B or G trends, but this is probably due to differencesin Fig. 10. Variation of radiation-enhancedcolor with water speciesin amazonite-orthoclase.Squares (lower axis), molar absorptivity for the thres different amazonitepeaks. structurally bound water. Triangles (upper axis), water occurring as fluid in- Someof the scattermay also be attributed to not all of the clusions. Open symbols are samplesfrom Broken Hill # 3, 3', and structural water being involved in (e.9., color c€nters # l2\. 3"; the fiIled symbol represents sample #26. Because of the If Pb and HrO are statistically distributed among the "blank" in the H, manometer measurementsof a few ppnr, the M-sites of amazonite,the probability of HrO having the line doesnot intersectthe origin. 802 HOFMEISTER AND ROSSMAN: COLORING OF AMAZONITE ^12 T E 3 to zF H8 c) H L LrJ o O o 34 H o-F e. LAKE GEORGE, oz BLUE SII.IGLET U' 6

E a c @ 5 la 15 2A 25 o to (o STRUCTURAL WATER CONTENT (PPM)

Fig. 11 . Variation of intensity of color with structural water by dehydration of amazonite # 12,Type B from Lake George, Colorado. Dose,4OMRads. Saturationoccurs above 12 ppm HrO.

KCI which are isoelectronicwith Pb3+ and Pbl* respec- tively. Pb3* producesthree bands in KCI at 465, 303, and 216 nm (Schoemakerand Kolopus, 1970),but the oscillator N coo6 strengths were not measured.The amazonite absorption I E pattern resemblesthat of Tl2+ in that both have nearly (, bands (Table 8). The difference aooo equal strength absorption in the number of bandsis of unknown significancebecause F amazonite may have a fourth band at the uninvestigated H wavelengths.Amazonite spectraresemble those of a 4'aa,0 shorter z Tl" in that both have a band at about 625 nm: however, bJ unlike arnazonite, the 625 nm band in Tl' is 10 times F z 3oa0 stronger than tho higher energy bands, and additional H bands are present at lower energy. This comparison sug- a geststhat Pb3+ is the chromophore in amazonite rather ld 29,00 F thanPbl+. u o 19,a0 +, bJ Table 8. Oscillatorstrengths of amazonite,Tl2 andT" F z ffi# Reference H le Amazonite, B 625 0.063 this work 385 0.04? 330 0.036 STRUCTURALHZO, ppm Amazonite, G 720 0.036 enhanced amazonite color Fig. 12. Dependence of irradiation KCI:Tl 2+ 354 0.13 Delbecqet a]', 1966 on structural water @ntent. Filled triangles, type B blue singlets. 294 0.22 type D doublets. Filled squareg typ€ C green ?62 0.14 Open , 220 0.22 singlets.X, typ€ T blue-greensinglet. Sample #14 and 15 have higher molar conoentrations of water than lead, and werc plotted KC'l:Tl' 1500 8.1 x 10-5 r?60 9.3 x 10-5 as discussed in the text. The envelopes enclose the two end- 640 0.48 mernber color types. Both a triangle and a (not shown) 380 0.03 -340 not determined inters€ct the origin of both graphs. " HOFMEIST'ER AND ROSSMAN: COLORING OF AMAZONITE 803

Relationships between amazonite color type, feldspar structure and Pb content .7 ro{ P E BLUE-GREEN,, The degreeof AfSi order can be O I determinedby infrared E spectroscopyfrom the separationof IG BLUE | ..o :r2o-- the two overlapping t peaks near 768 and 728 cm-r (-13 pm) (Hafner o 15. ^s I \ and o o',, Laves, 1957).As a measureof this separationwe used the E ratio of the depth of the trough to the height of the over- 5 lap, d/m (Fig. 13). The maximum degree of Al/Si order g. seemsto decreasewith increasedincorporation of pb into ol! E the potassiumfeldspars (Fig. 1a),and at very high pb con- o tent (- 1.5%) only orthoclase has been observed. The v various color types are restricted to specificpb contents and structural states:(1) Type B occursat low pb contents; (2) Type G is limited to high Pb and high structural states; (3) Doublets (D) have intermediatePb content and inter- Pb, ppm mediate structural states. This information agrees with Fig. 14. Order parameter d/m as a function of lead content of analogousX-ray data: Colorado amazoniteswith lessthan amazonites.Dots, uncolored.Triangles, type B. Diamond, type D. about 2000ppm Pb have - A131,rs, 1 (Foord and Martin, Square, type G. X, type T. Dashed lines also show the boundaries 1979); Mongolian amazonite with 1950 ppm pb has of amazonitetypes. A : 0.87 (Pivec et al., 1981);Norwegian amazoniteswith 3000 ppm Pb have A - 0.6 to 0.8 (Tibbals and Olsen, 1977).The trends of Figure 14 may indicate that incorpor- bly, incorporation of a moderate amount of Pb can cause ation of Pb into locally changesthe slight distortions from maximum microcline structure. structure such that large amounts(about 2%o)are sufficient to produce an overall orthoclasestructure. The occurence A model for the mechanism of of type G and type B peaks together in samplesof inter- amazonite colorttion mediatestructural stateshows that M-sites of end-member We have shown that (1) amazonite color results from pb. microclineand orthoclasegeometries exist for electronictransitions involving Pb3+ or Pbl+; (2) the pro- For intermediate Pb content and low structural state, duction of the unusual charge state upon irradiation re- unusualamazonite coloration occurs,such as for the Keivy quires the associationof a water moleculewith the pr€cur- #21 samplewith three EPR Pb signals.The changein the sor Pb site; (3) irradiative production of color is rate limit- EPR spectra is accompaniedby a 400 cm-l shift in the ed by one first order reaction; (4) the speciationand con- energy of the optical bands, suggestingan intermediate centration of molecular water are unchangedby the pro- structure possi- for the chromophore sites in this sample. cess;(5) different color typesare determinedby the propor- tion of the Pb located in M-sites with either microcline or orthoclasegeometry. Comparison WAVELENGTH,um of the optical spectraof amazoniteto those l2 t5 20 of Tl2+ and Tl' in KCI suggestsPb3+ as the chromopbore in amazonite.Also, Pb3* has been demonstratedto form 1.4 by irradiation of natural calcite and other oxides (Rohrig l2 and Schneider,1969; Born et al., L97l; Andlauer et al., uJ to 1973; Popescu and Grecu, 1975). Conversely, previous 2 B analysis of the EPR center in amazonite suggestedPbl + uo o.8 T (Marfunin and Berhov, 1970),as does the g-value of 1.53 8 s.e for the even Pb isotopes(G. Lehmann and J. Weil, pers. d) D comm., 1984).Perhaps EPR spectra on scv- < 6.4 eral different color types of amazonite would resolve this o.2 dilliculty. Points (2) and (4) abovesuggest that water plays a cata- 9go 780 sog lytic role in amazonitecoloration. We proposethat gamma I",AVENUMBER,.'-l radiation dissociateswater molecules,forming Ho ad OH', radicals that are the primary products of Fig. f3. Infrared powder absoption spectra of amazonites.B, water radiolysis (Draganic #4 from Wigwam Creek, CO., T, #2t from Keivy, USSR. D, and Draganic, l97l). H" has been observedin #19'from New York Mts., CA., G, #3,,from BrokenHill, Aus- irradiated hydrous glassesand quartz (Van Wierengen and tralia. Spectrawere offset for clarity. The ratio d/m, where d and Kats, 1957;Weeks and Abraham, 1964;Peilson and Weil, m are the intensitiesindicated near 13 pm, was usedas a measure 1974). The radical H" probably diffuses and removes of Al/Si order/disorder. smoky color by reduction, as occurs in quartz (e.g.Kats, 804 HOFMEISTER lND ROSSM,{N: COLORING OF AMAZONITE

L962,p. 194). We suggest that as in quartz (ibid.), hydrogen Hafner. St. and Laves, F' (1957)ordnung,/unordnung und ul- einiger returns to a site similar to its original one. The radical OH' trarotabsorption II. Variation der Lage und Intensitat Zur Structur von Orthoklas und probably remains stationary, oxidizing nearby to Absorptionenvon Feldspaten. Adular, Zeitschrift fiir Kristallographie,109, 2M-225. form hole centers. With both reduction and oxidation oc- Hofmeister,A. M. (1984)A SpectroscopicStudy of FeldsparsCol- curing, it is possible to form either Pb3+ or Pbl*. Natural ored by Irradiation and Impurities, Including Water. Ph.D. (primarily from aoK) is suf- radioactivity in amazonites Thesis.California Institute of Technology,Pasadena, California. ficient to produce this color over geologic time (Hofmeister, Hofmeister, A. M. and Rossman,G. R. (1985)A rrodel for the 1984). irradiative coloring of smoky feldspar and the inhibiting influ- enceof water.Physics and Chemistryof ,in press' Kats, A. (1962)Hydrogen in alpha-quartz.Philips ResearchRe- Acknowledgmants ports,17, 13!195, 2Ol-279. Specialthanks are due to E. E. Foord (U.S.G.S.,Denver) for Marfunin, A. S., Bershov,L. V., Meilman, M' L.' and Michoulier, generouslysharing his samplesand researchideas with us. Sam- J. (1967\ Paramagneticresonance of Fe3* in some feldspars. ples were also donatedby S. Booth (Zinc Corp. Ltd., Broken Hill, SchweizerischeMineralogische und PetrographischeMitteil- Australia), G. Brown (Stanford),R. Cubba and S. Rigden (Cal- ungen,47, 13-20. tech), R. Currier (JewelTunael Imports, Arcadia, California), S. Marfunin. A. S. and Bershov,L. V. (1970)Paramagnetic centers in Ghose and A. Irving (U. of Washin4on), and R. Reynolds(San feldspar and their possible crystallochemicaland petrological Bernardino County Museirm, California). We thank S. Chan, C. significance.Doklady Akademii Nauk SSSR' 193, 412414 Martin, and D. Blair (Caltech Chemistry Division) for assistance (transl.Doklady Akademi Nauk Translations'193, 129-l3O). and advice with EPR sp€ctroscopy.Thanks are due to J. R. Perlson, B. D., and Weil, J. A. (1974) Atomic hydrogen in a- O'Neil and R. E. Criss (U.S.G.S.,Menlo Park) for assistancewith Quartz. Journal ofMagnetic Resonance,15' 59'+-595. hydrogenmanometry there. Discussions with L. T. (Caltech) Pivec, E., Sevcik, J. and Ulrych, J. (1981) Amazonite from the were very helpful. Critical review by E. E. Foord (U.S.G.S., alkali of the Avdor Massif,Mongolia. TschermaksMin- Denverl S. Ghose (U. Washington),G. Lehmann (Muenster),R. eralogischeund PetrographischeMitteilun gen,28' 277-283. F. Martin (Mccill U.), and J. Weil (U. Saskatchewan)substan' Plyusnin,G. S. (1969)On the coloration of amazonite.(in Russian) tially irnproved the manuscript.This work was supportedin part Zapiski Vsesoyuztogo MineralogischeskogoObshchestva, 98' by NSF grantsEAR-79-19987, 79-04801, and 83-13098. 3-17. Popescu,F. F., and Grecu, V. V. (1975)Temperature dependence of Pb+ 3 EPR spectrum.Phyiica StatusSolidi (b)' 68' 595-601. References Przibram, K. (1955)Irradiation Colors and Luminescence,Perga- Aines, R. D., and Rossman,G. R. (1984)Water in minerals? A mon Press,London, England. peak in the infrared.Journal of GeophysicalResearch, 89, 4059- Rohrig, R. and Schneider,J. (1969)ESR of Pb3* in ThOr. Physics n7L Letters,3OA,37l-372. * Andlauer,B., Schneider,J. and Tolksdo{ W. (1973)ESR analysis Schoemaker,D., and Kolopus, J' L. (1970)Pb* as a hole trap in of Pb+3ions in YrGarOrr. PhysicalReview B' 8, 1-5. KCI: ESR and opticalabsorption of Pb***. Solid StateCom- Born, G., Hofstaetter,A., and Scharmann,A. (1971)2Sr,r-states of munications,8, 435439. Pb+3-ions: correlations between g-values and hyperfine split- Solomon, G. C., and Rossman,G. R. (1979),The role of water in ting constantsA in the EPR Spectra.Zeitschrift fiir Physik,248, structural states of K-feldspar as studied by infrared spec' 7-12. troscopy. Geological Society of America Abstracts with Pro- Cech, F., Misar, 2., and Povondr4 P. (1971) A green lead- grams,11, 521. containing orthoclase.Tschermaks Mineralogische und Petro- Speit, B., and Lehmann, G. (1982)Radiation defectsin feldspars. graphischeMitteilungen, 15,213-231. Physicsand Chemistryof Minerals,8,77-82. Delbecq,C. J., Ghosh, A. K., and Yuster, P. K. (1966)Trapping Stacey,J. S., and Kramers, J. D. (1975)Approximation of terres- and annihilation of electronsand positive holes in KCI-T|CI. trial lead isotope evolution by a two stage model and PhysicalReview, 151, 599-609. PlanetaryScience Letters, 26, 207-221. Draganic,I. G., and DraganigZ.D.(1971) The RadiationChem- Tarashchan,A. N., Serebrennikov,A. I., and Platonov, A. N. istry of Water. AcademicPress, New York. (19?3)Peculiarities of the lead 's luminescencein amazonite. Eaton, S. S., and Eaton, G. R. (1979)Signal area measurementsin Konstitutsiya i SvoistvaMineralov, 7, 106-111. EPR. Bulletin of Magnetic Resonance,1, 130-f 38. Tibbals, J. E., and Olsen,A. (1977)An el€ctronmicroscope study Foord, E. E., and Martin, R. F. (1979)Amazonite from the Pikes of sometwinning and exsolutiontextures in microclineamazon- Peak Batholith. MineralogicalRecord lO, 37T382. ites.Physics and Chemistryof Minerals, 1,313-324. Friedman, I. (1953) Deuterium content of natural waters and Van Wieringen,J. S.,and Kats, A. (1957)Paramagnetic resonance other substances.Geochimica et CosmochimicaActa,4, 89-103. and optical investigation of silicate glassesand fused silica, col- Gaite, J.-M., and Michoulier, J. (1970)Application de la resonance oured by X-rays.Philips ResearchReports, 12,432454. paramagnetiqueelectronique de I'ion Fe3* a I'edtude de la Weeks,R. A., and Abraham M. (1964)Electron spin resonanceof structure des feldspaths.Bulletin de la Soci6t6 Francaise de irradiated quartz: atomic hydrogen.Journal of Chemical Phys- Min6ralogieet de Cristallographie,93, 341-356. ics.42.68-72. Goodman, B. A. and Raynor, J. B. (1970)Electron spin resonance of transition metal complexes.Advances in Inorganic Chernistry Manuscript receioeil, April 12, 1984; and Radiochemistry,13, 135-363. accepteilfor publication, March 14, 1985'

Top View