American Mineralogist, Volume 70, pages 106-123,I9E5

Iligh resolution solid-statesodium-23, aluminum-27, and silicon-29nuclear magnetic resonancespectroscopic reconnaissance of alkali and plagioclasefeldspars

R. JeMes Ktnrpernrcr Department of Geology University of lllinois at Urbana-Champaign l30l West Green Street. Urbana. Illinois 61801

Ronnnr A. KrNsev. KeneN ANN SuIrn School of Chemical Sciences University of lllinois at Urbana-Champaign 505 South Mathews Avenue, Urbana, Illinois 61801

Donelo M. Hr,NoBnsoN Department of Geology University of lllinois at Urbana-Champaign 1301 West Green Street. Urbana. Illinois 61801

eNo Enrc OlorrBI-o School of Chemical Sciences University of Illinois at Urbana-Champaign 505 South Mathews Avenue. Urbana, Illinois 61801

Abstract

We presentin this paperhigh-resolution solid-state magic-angle sample spinning - 23,aluminum -27 , andsilicon-29 nuclear magnetic resonance (NMR) spectraof a varietyof plagioclaseand alkali feldsparsand feldsparglasses. The resultsshow that MASS NMR spectracontain a great deal of informationabout structures. In particular,we observethe following.(l) AUSiorder/disorder and exsolution effects in alkalifeldspars are readily observable.(2) To our presentlevel of understandingthe publishedstructure refinementsof anorthiteare not consistentwith the NMR of spectraof anorthite.(3) The spectraofthe intermediateplagioclases are interpretable in termsofe- and/-, with much of the signalapparently coming from strainedsites between albiteJike and anorthite-likevolumes in the e-plagioclases.(4) The nearest-neighbor atomic arrangements in and anorthiteglasses are very similar to thosein the correspondingcrystals, althoughthe rangeof cation-oxygendistances is largerin the glasses.(5) Thereare no aluminum-27peaks in the rangeof 0 ppm that would indicateoctahedrally coordinated aluminumin eitherglass.

Introduction spectroscopicaspects of this problem' Future work will examine the mineralogical and petrological applications We present in this paper high-resolution solid-state in more detail. Both magic-anglesample spinning (MASS) sodium-23,aluminum-27, and silicon-29nuclear magnetic and variable-angle sample spinning (VASS) techniques resonance(NMR) spectroscopic results for a variety of have been used. plagioclaseand alkali feldsparsand feldspar glasses.The The samplesdiscussed here include low albite, micro- objective of this work is to explore the applicability of line perthite, cryptopenhite, sanidine, a range of plagio- high-resolutionsolid-state NMR to the examinationof the clase compositions, and albite and anorthite glasses.The structuresof complex silicate materials, such as feldspar samplesare listed in Table 1. and feldspar glasses.The emphasisin this paper is on the In this paper we briefly summarizefeldspar nomencla- m03-004)vE5/0102-0 r06$02.00 106 KIRKPATRICK ET AL.: NUCLEAR MAGNETIC RESONANCE OF FELDSPARS t07

Table l. Silicon-29 NMR chemical shifts of feldspars and spar are given in the compendia of Smith feldspar glasses (1974) and Ribbe et al. (1983). For the benefit of those unfamiliar with the feldspars, CHEMI6L ruTERIAL LOCALIN SHTF1 COffiNTS we summarize their main features (following Smith, 1974 and Ribbe et al., 1983), state some of our assumptions, Alblte &elta, Vlrglnla l2n (Ab99) -96.: T20 and define some of our nomenclature and symbolism. -104.4 Tlo

AIblte Class Synthetlc -98 l. Feldspar structures are based on a framework of corner- (Abr00) linked SiO+ and AlOa tetrahedra. The ratio of aluminum to Sanldin€ Cyonlng -97 poorly separated (or60Ab38An2) -IOI silicon rangesfrom l:3 to l:1. We use the notationQ4(nAl) to

Mlcrocllne -93 Ab T2tr describe the silicon sites, where the superscript indicates the Perthlte -96.0 T2b four bridging oxygen atoms and the integer n (0 to 4) denotesthe T2o + Ab T2o -l0l,8 Tln number of aluminum (AlOo) tetrahedra to which a silicon (SiO+) -106,0 Ab Tlo tetrahedronis connected. CryptoperthlEe Ab T2E ( Spencer R) -96-o AbDo+0rT2 We assume,as is customary in dealingwith tetrahedralsilicon ( or43,4b,53,k4) Nomay) -98.0 Or Tl and aluminum occupancies,that the "aluminum avoidancerule" - 103,5 Ab TID (Loewenstein, -47 1954; Goldsmith and Laves, 1955)holds for all 01lgocla6e Bakersvllle, Shoulder (An Fak) ( An23 ) North Carollna -93.2 feldspars.That is, no aluminum tetrahedron is linked to another -97.2 - 105.3 Ab peak aluminum tetrahedron,only to four silicon tetrahedra.A// alumi-

Labradotlte Sydney Ene6, -83 num sites, therefore, should be A4 (4Si). We will show, however, (Ans2 ) kbrado! -88 -91 that for our synthetic anorthite this is not completely true. - 101 -107 2. Feldsparframeworks have inherent monoclinic or pseudo- monoclinic symmetry. Because AlO4 tetrahedra are slightly Eytomlte Crystal kyl -84 Corlelatlon to (An70) Mlnnesota -89 larger than SiO4 tetrahedra, some patterns of AllSi distribution

- 101 give monoclinic frameworks, whereasothers causeslight distor- tions and triclinic symmetry. The pattern of tetrahedral site Pacaye Volceno -83.0 Flrst tro peaks (h92) GuatgMla -84.6 poorly reeolved -89.2 occupancy leads to the important concept of AVSi tetrahedral site order/disorder.Other things being equal, high AVSi disorder -42.9 Synthetic Cortelatlon to promotes (Anr00) -84.8 publlshed Etructure monoclinic frameworks, whereas increasing order -89'6 reflnerentB poor promotes triclinic frameworks having increasing cell obliquity Anorthlte Glass Synrherlc -87 (indicated by the (hroo) magnitude ofthe departure ofthe B-axis from B*). The large, irregular cages in the tetrahedral framework are occupiedby ,sodium, or calcium ions. When occupied ture and structure and previous NMR work on feldspars, by the larger potassium ions, the cages are stretched and the framework tends review the theory of MASS NMR (especially as it per- to retain its monoclinic symmetry if it has a proper AUSi distribution. Ifthe AUSi pattern is not appropriate, tains to the study of quadrupolar nuclei), and present and the symmetry is triclinic. When occupied by the smaller, sodium discussour results. Theseresults are, we believe,only a or calcium ions, the cages crumple a little and distort the beginning of the application of MASS NMR to feldspar framework slightly so that the symmetry is triclinic, regardlessof mineralogy, and we present our conclusions with the the AVSi order/disorder (except for the highest temperature recognition that they are in someinstances tentative. It is sodium feldspar, monalbite). clear from the work we have done, however, that MASS 3. Feldsparcompositions are conveniently describedin terms NMR spectraof alkali and plagioclasefeldspars contain a of percentagesof the endmembercomponents KAlSi3Os (Ortho- great deal of information about Al/Si orderidisorder in the clase, Or), NaAlSi3Or (Albite, Ab), and CaAl2Si2Os(Anorthite, tetrahedral sites, about chemical bonding in the sodium, An). The rock-forming feldspars occur in two series, the alkali feldspars(Or-Ab) plagioclases(ALAn). aluminum and silicon sites, and also, we believe, about and 4. The principal potassic feldspars include, from low to high distortion of these sites due to exsolution. Our results temperatureforms, microcline (triclinic, Cl), or low indicate that solid-state MASS and VASS NMR. when sanidine (macroscopicallymonoclinic; monoclinic by polarized combined with X-ray and electron difraction and trans- light microscopy but likely to be triclinic (low obliquity) by X-ray mission electron microscopy (TEM) can provide new and electron diffraction), and high sanidine (monoclinic, C2im). insights into the structures of a wide range of The principal sodic feldspars are low (temperature) albite and phases. high albite, both triclinic, Cl; high albite has the lower cell obliquity. High sanidineand high albite form a continuous solid Previous work solution seriesdespite their different symmetries. 5. In the triclinic alkali feldspars(microcline, low albite, high Feldspar structure and nomenclature albite) there are four sets of similar but not identical tetrahedral sites conventionally labeled Tlo, Tlm, T2o, T2m. At high Feldsparsare the most abundantminerals in the earth's temperatures, the aluminum and silicon ions are randomly crust and have received corresponding attention from distributed over these sites. If a high temperature feldspar is mineralogistsand petrologists. Excellent reviews of feld- quenched, this disorder is retained. In the lowest temperature KIRKPATRICK ET AL,: NUCLEAR MAGNETIC RESONANCE OF FELDSPARS forms, maximum low albite and maximum microcline, AVSi type that appears to be body centered by diffraction, even at order is, ideally, complete,with all the aluminum in the Tlo sites room temperatures.It is consideredto have a composite struc- and all the silicon in the other three sites. ture made from domains of P-anorthite. Adjacent domains are 6. In the monoclinic alkali feldspars there are two sets of "out-of-step" to one another (antiphasedomains) with respectto similar but not identical tetrahedral sites conventionally labeled AVSi ordering, which, coupled with the earlier mentioned 1- Tl and T2. High sanidineideally has one-quarterof an aluminum pseudosymmetryof each domain, causessystematic difiraction and three-quaxtersof a silicon in each site on the average absences.These diffraction absencesgive an apparent1l symme- (complete Al/Si disorder). Orthoclase (low sanidine) is more try for the averagestructure. ordered, with equal numbers of aluminum and silicon atoms randomly occupying the Tl sites and silicon only occupying the NMR of feldspars T2 sites. Feldspars have been investigated by single crystal 7. Non-endmember, phase, high temperature alkali single sodium-23and aluminum-27wide-line NMR at low mag- feldsparsexsolve into intergrowths called perthites when cooled netic field strengths since the early 1960's. Brun et al. slowly enough. Cryptoperthite is perthite in which the inter- (1963) growth occurs on a submicroscopic scale. In the finer crypto- (1960),Hafner etal. (1962),and Hafner and Laves perthites, the lamellaeof the potassium and sodium-rich phases investigated alkali feldspars and were able to resolve are thought to maintain coherency across the interface causing separatepeaks from different phases. Hafner and Hart- strain of the sites near the interface. In the coarser perthites, mann (1964) determined electric quadrupole coupling microperthites and macroperthites,the boundariesbetween the constantsof a variety of compositions. Staehli and Brink- lamellae are noncoherent, and there probably is no significant mann(1974a,b)examined the aluminum-z7NMR of anor- straining of the sites near the interfaces. thite using wide-line methods and were able to separate 8. The plagioclasesare even more complex. The albite end- the eight aluminum sites, determine that the P to I members have been described. The calcic endmembersare P- transition takes place between 220" and 250'C, and again anorthite (triclinic, Pl) and /-anorthite (triclinic, 1l), which occur values the electric quadrupole coupling below and above 250'C, respectively. On cooling, the 1 to P measure the of transition is not quenchable.P- and /-anorthite have doubled c constants. The quadrupole coupling constants of feld- cell edge lengths of about 144 comparedto the 7A of the other sparsand many other have been summarizedby feldspars.Even the P-anorthite structure has a high degreeof1- Ghoseand Tsang(1973). pseudosymmetry(Ribbe in Ribbe, 1983). Silicon-29 chemical shifts obtained by high-resolution P-anorthite has 16 sets of similar but not identical tetrahedral MASS NMR have been presentedby Lippmaa et al. sites. Eight are occupied by silicon and eight by aluminum. 1- anorthite has 8 sets, four occupied by silicon and four by aluminum. Both kinds of anorthite ideally have complete Al/Si Table 2. AverageSi-O bond distances,total cation-oxygen site order. Each silicon tetrahedron is connected to four alumi- bond strengths,and predictedsilicon-29 chemical shifts for num tetrahedra, Qo(4Al), and vice versa, Aa(4Si). anorthite 9. Standard compositional ranges of the plagioclasesof in- termediate compositions are labeled by the names oligoclase, St11cotr S1-O Bond Strength sEczsl-rsr Anro-ro,andesine, Arro-so, labradorite, Anso-ro,and bytownite, Slte A su (v.U.) AnTs-e6(see Table l). These names refer to bulk compositions, not phases.The intermediate plagioclasesare all triclinic. ksaw et al. (1962) At the highest temperaturesthere is extensive solid solution between the Ab and An components. At lower temperatures, Tr(000) I.5I3 8,058 -1.438 rr (00r) t.5 16 8.083 -I.471 intermediate composition plagioclasesexist as exceedingly fine rr(M0) L608 8.036 -r.27 | intergrowths of albite-like and anorthite-like domains on several Tl (rzt) r,626 7.990 -1.266 scales. The complexity of these plagioclasesarises from the r2(0zo) I .513 8.187 -1.441 T2(OzL) I .610 8,264 -1.361 incompatibilitiesof the AVSi site orderingsin albite and anorthite T2(n00) t.602 8.I5t -t.346 -1.469 and from the fact that subsolidusatomic rearrangementprocess- 12(n01) 1.528 7.879 es are so sluggish. We shall acceptthe e-plagioclasemodel as describedby Ribbe and by Smith in Ribbe (1983).The intergrowths in plagioclases Tl(000) I .616 8. I44 -1.444 rl ( o0r) 1.6t3 8. I23 -t,243 occur on at least two scales. The finer type, e-,is rl (Dzo) I .613 8,143 -t.451 -1.268 thought to consist of a modulated structure composedof coher- Tl (nzl ) I .613 7.926 ent albite- and anorthite-like domains of the order of a few tens 12(020) I.6r7 -l,446 T2(021) r.608 8.r83 of A across. e-plagioclaseseems to be restricted to the range T2(D00) I .515 8.000 -t.442 Anzo-ro. T2(n0r) I .614 8.086

A second category of intergrowth occurs on the micrometer klus ( 1978) scale and is associatedwith the so-calledperisterite (-Anz-ro), BQggild (-An+7-se) and Hiittenlocher (-An6s-q) gaps. Perister- Tr (000) I .618 8. 146 -r.445 rr(001) I .613 8.r30 -1.458 ite intergrowths are considered here to be composed of albite rr(Ezo) t.6t4 8.075 -t.252 -r,213 (-An2) and an e-plagioclase(-An25). Bpggild ones are believed rl(Mr) I .513 1 .939 to be composed of two e-plagioclases(-Anor and -56) and T2(020) I .617 8.r29 -r.460 T2(ozt) r .613 8.143 Hiittenlocher intergrowths are thought to be composedof an e- T2(oo0) I .614 8.009 -1.345 I .6I6 8.06l -1.445 plagioclase(-Anoa) and /-anorthite (-Anso). This /-anorthite is a T3(Eot) KIRKPATRICK ET AL.: NI]CLEAR MAGNETIC RESONANCE OF FELDSPARS l(D

(1980) for what they call anorthite, albite, sanidine, tetrahedral sites in feldspars. This is because they are orthoclase, and adularia. They presented spectra for shielded from the imposed magnetic field by slightly albite and sanidine, discussed structural interpretations, different electron distributions. Silicon-29 has a nuclear and proposed peak assignments for the albite peaks. spin I : ll2, and,as describedelsewhere (Andrew, l97l; Oldfield et al. (1983)have discussedthe effectsof iron Smith et al., 1983),spinning of a powderedsample on an impurities on the aluminum-27 and silicon-29 MASS axis inclined at 54.7" to the applied magnetic field (the NMR spectraof a variety of feldspars.K. A. Smith et al. "magic-angle") at kilohertz frequencies greatly narrows (19E3)have discussed silicon-29 MASS NMR spectrosco- the NMR peak breadth and increasesresolution. This is py of a number of silicates,including some feldspars. due to an efective averaging of the nuclear dipolar and Recently,J. V. Smith et al. (1984)have presentedhigh chemical shift anisotropy interactions. Thus, nonequiva- resolution silicon-29MASS NMR data for alkali feldspars lent silicon sites in an ordered crystal each give a single and discussed the peak assignments.Solid-state alumi- nalrow peak. Peaks for glassesand disordered crystals num-27and silicon-29MASS NMR have probably found are generally broader, because a range of magnetically their most important application to date in the study of nonequivalentsites is present. zeolitecatalyst materials (e.g., Klinowski et al., 1982). Sodium-23and aluminum-27,unlike silicon-29,have spins I : 312and 5/2 respectively and, therefore, exhibit Experimental methods quadrupolarbehavior. Although MASS averagesthe first order quadrupole interaction, for sites with large electric Spectrometers field gradients (and thus large quadrupole interactions) The spectrapresented here were recorded using three "home- second-orderterms in the nuclear quadrupolar Hamilto- built" Fourier transform NMR spectrometersusing the methods nian, Hq, becomeimportant, and may causea significant describedby Meadowset al. (1982)and Smirhet al. (1983).The linebroadening.Such efects have been discussedin detail three spectrometersare based on 3.52, 8.45. and ll.7 Tesla by Abragam(1961), Kundla et al. (1981),Meadows et al. superconducting magnetsand use Nicolet ll80 or 1280comput- (1982),Ganapathy et d. (1982),and Schrammand Old- ers for data acquisition. All are equipped with Andrew-Beams field (1982). We review here those aspects of theory type magic-angle sample-spinningrotor assemblies (Andrew, necessaryto understand the spectra presented. l97l), which can also be used to obtain static (non-spinning) (2I spectra.Chemical shifts are reported in ppm relative to 3M NaCl A nucleusof spin I has + l) different nuclear energy for sodium-23,in ppm relative to lM Al(H2O)l+ for aluminum- levels when placed in a magnetic field, as shown for 27, and in ppm relative to tetramethylsilanefor silicon-29.More silicon-29(l = ll2, two energylevels) and sodium-23(I = positive chemical shifts correspond to decreasedshielding. 312, 4 energy levels) in Figures lA and D. In NMR spectroscopythe energies(or frequencies)of the transi- Samplepreparation and characterization tions (1/2, -ll2), (112,312), etc. are measured.In the All sampleswere ground to a fine powder using an agate(SiO2) absenceofany other magneticinteractions, only a single pestle and mortar and were tightly packed into Delrin rotors. resonancefrequency would be observed.However, in Most sampleswere pure minerals and were simply ground and real samples, there are a variety of interactions which run. Pacayaanorthite was purified by hand picking and magnetic cause slight changesin these nuclear energy levels. For separation. The Wyoming sanidinewas purified by hand picking silicon-29 the main efect is due to an orientation-depen- and shaking in water in a small mill to remove the soft matrix. dent shielding of the main magnetic field at the nucleus. The anorthite glass was prepared from 99.D9Vo pure SiO2, This is called the chemical shift anisotropy (CSA) and CaCO3,and Al2O3by melting for about one hour at 1600.C,three yields (Fig. times, with grinding in an agate mortar between meltings. a broad resonancein powdered samples 1B). Syntheticanorthite crystals were grown by crystallizing the glass In MASS this interaction is effectively averagedby rapid at about 1500"Cfor about four hours and then quenching the sample rotation in a sample rotor spun by an air-bearing crystals. The albite glass was prepared by the Coming Glass at frequencies,2,, typically from about 2,000to 6,000Hz. Technical Staffs Division of Corning Glass Company. The result is a sharp-line spectrum with a centerband All sampleswere characterized by powder X-ray diffraction (CB, Fig. lC) at the samefrequency as in the absenceof and optical microscopy. the CSA, togetherwith a number of spinningsidebands (SSB),spaced at the spinningfrequency (2.). Theoretical background For quadrupolarnuclei (sodium-23,| = 312,aluminum- A useful introduction to NMR spectroscopyis given in 27, | = 5 12),there are a number of transitions. In general, the book by Farrar and Becker (1971). Magic-angle the resonancesof these quadrupolar nuclei in powders sample-spinning(MASS) in general,and silicon-29MASS are broadened by the interaction of the nuclear electric NMR spectroscopyin particular, have been discussedby quadrupolemoment (eQ) with the electric field gradientat Lippmaaet al. (1980)and by Smith et al. (19E3).We note the nucleus (eq). This interaction is describedin terms of here only that a given nuclide ordinarily has different the quadrupole coupling constant, e2qQ/h. For sodium- nuclear resonancefrequencies in different kinds of sites, 23, the first order quadrupole interaction leads to two whether in diferent materials or in the same material. broad absorptionsdue to the satellite transitions (312,ll2; even if the sites are as similar to each other as the -lD, -3/2) together with a relatively sharp central reso- 110 KIRKPATRICK ET AL.: NUCLEAR MAGNETIC RESONANCE OF FELDSPARS

A B c

-roo -r50 -200 -roo -r50 -200 PPM PPM F D E cB ( t/2, -t/21 ZEEMAN QUADRUPOLAR

t,"'t""t""t"" 200 t50 loo 50 0 200 r50 roo 50 0 KHz KHz G H

Fig. 1. Energy level diagrams,static spectra,and MASS NMR spectraof a spin I = 1/2nuclide (A,B,C), a quadrupolarnuclide (spin | : 312)with a small quadrupole coupling constant (D,E,F), and a quadrupolar nuclide with a large quadrupole coupling constant (D,G,H). (A) Energy level diagramof spin I = l/2 nuclide showingZeeman splitting into ll2, and - l/2 spin statesand small additional splitting due to chemical shift anisotropy (CSA). (B) Simulation of a static spectrumof a spin I : l/2 nuclide (silicon-29in Na2Si2O5) (C) MASS NMR spectrum of spin I : l/2 nuclide (silicon-29)in Na2Si2O5showing the center band (CB, the true resonance)and spinning sidebands(SSB) spacedat the spinning speed, zr. (D) Energy level diagramof a spin I : 3/2 quadrupolar nucleus showing Zeernan splitting and a small additional splitting due to the quadrupolar interaction. (E) Simulation of a static spectrum of a quadrupolarnuclide (I : 3/2 in this case)which has a small quadrupolecoupling constartt.(F) Simulationof a MASS spectrumof an I = 3/2 nuclide showing the centerband(CB), which is the central (+ll2 + - l/2) transition and the spinningsidebands (SSB) mapping out the shapeof the satellite spectra.(G) Simulation of the static spectrumof a quadrupolarnuclide with a large quadrupolarcoupling constant(sodium-23 in Na2MoO+).Only the central (+ 12+ -ll2)transition is shown. (H) Simulationof the VASS NMR spectrumof sodium-23in Na2MoOashowing the narrowing possible at 75'. The MASS spectrum of this compound shows less narrowing.

nance(112, -ll2), as shown in Figure lE. In general,with er than about 3 MHz), additional structure generally MASS, these interactions are again averaged, and a appearson the central line due to the higher order effects sharpercentral line (112, -ll2) together with a number of which arise when the quadrupole interaction is a signifi- sidebandsdue to the satellite transitions are observed, cant fraction of the actual resonance frequency. For 23Na2MoOa Figure lF. When the quadrupole coupling constant is example, for the quadrupole coupling con- about 1 MHz in frequency units, only the central (l/2, stant is approximately 2.6 MHz, and at all fields in the -ll2) resonanceis readily detected, since there is a large range3.52 to 11.7Tesla there is a so-calledsecond-order residual broadening of the satellite lines, due to small broadening of the central transition, Figure lG. This fluctuations around the magic-angledue to wobbling of interaction has a different angular dependencethan that the sample rotor. causing the first order chemical shift or quadrupolar When the quadrupolecoupling constant is larger (great- broadenings. In order to reduce this interaction it is KIRKPATRICK ET AL.: NUCLEAR MAGNETIC RESONANCE OF FELDSPARS llt necessaryto spin the sample at a different angle in the t'si magnetic field (variable-angle sample-spinning, VASS, 8,45T MASS Ganapathyet al., 1982). Thus, for silicon-29,or for quadrupolarnuclei in highly -96'. symmetric environments (where electric field gradients -92'" -,oc,,A are small)-such as sodium-23 in NaCl, magic-angle ALBITE (54.7")sample-spinning yields the highestresolution spec- tra, Figures lC and F. For quadrupolar nuclei in highly asymmetric sites, such as sodium-23in Na2MoO4, vari- ,rr,1/t/\"t"fdli -l able-anglespinning, at 75" in this case, yields the narrow- f - . - -L10 - - -qo -160 estline width, Figure lH. Both techniquesare usedhere. 100 LOU iJJ As discussedin detail elsewhere (Meadows et al., -ror 1982),the second-orderquadrupole interaction is a field- D dependent interaction, and results in different apparent chemical shifts for different applied magnetic freld strengths(Abragam, 1961).As a result, there may be \ significant discrepanciesin ihe chemical shifts of a qua- drupolar nucleus when results obtained at different mag- ---f-] -tlo -too -160 -+o - - netic field strengthsare compared.All data for sodium-23 10c 160 and aluminum-27 presented here were obtained at the three different magnetic field strengthsto help determine E the true chemical shifts. Because silicon-29 has spin CRYPTOPERTHITE I : ll2, its chemical shifts are not field-dependent. For the quadrupolarnuclei it is often useful to simulate (calculate)the spectrausing the methods of Ganapathyet al. (1982).The parametersneeded for the simulation are l--T-----_l -+o . the quadrupole coupling constant (e2qQ/h),the asymme- IUU IOU PPM try parameterofthe electric field gradient tensor (4), and 2eSi the spinning angle (0). The asymmetry parameter is a Fig. 2. 8.45 Tesla NMR spectra (corresponding to a resonance feldspars measureof the deviation of the electric field gradient at frequency of71.5 MHz) ofnatural alkali and a synthetic albite glass. (A) Albite (Amelia, Virginia), 2.4 kHz the nucleus from cylindrical symmetry (Abragam, 196l). MASS, 3l scansat 600 sec recycle time, 8.6 kHz spectralwidth, Such simulations are used here in conjunction with pub- 20 prsec90'pulse excitation, 4096data points, zero-filledto E192, quadrupole lished coupling constants and asymmetry 25 Hz linebroadening due to exponential multiplication. (B) pa"rametersto determine isotropic chemical shifts, and to Microcline perthite (Delaware Co., Pennsylvania), 2.3 kHz look for contributions from exsolved phases. MASS, 3000scans at l0 sec recycle time,8.6 kHz spectralwidth, For the silicon-29NMR spectrawe report the chemical 2.6 psec 23' pulse excitation, 4096 data points, zero-filled to shiftsof narrow peaksto +0.1 ppm, but subscriptthe last 8192,5 Hz linebroadeningdue to exponential multiplication. (C) place becauseaccuracy and reproducibility from labora- Synthetic albite glass, 2.4 kHz MASS, 87 scans at 600 sec tory to laboratory is probably only about 0.3 ppm. This is recycle time, 8.6 kHz spectral width, 30 psec 90" pulse points, primarily due to magnet drift over the relatively long excitation, 2048 data zero-filled to 8192, 100 Hz linebroadeningdue to exponential multiplication. (D) Sanidine times (up to 12 hours) needed to collect some silicon-29 (Wyoming), 2.4 kHz MASS, 2000 scans at l0 sec recycle time, spectra. For broader silicon-29 peaks and shoulders we -r 8.6 kHz spectral width, 2.6 p,sec23" pulse excitation, 4096data report the shifts to only I ppm. points, zero-filled to 8192, 25 Hz linebroadening due to For the quadrupolarnuclei we report the valuesofmost exponentialmultiplication. (E) Cryptoperthite ("Fredriksvarn", of the singularities or peak maxima to + 1 ppm. This is Norway),2.7 kHz MASS,65 scansat 6ff) sec recycletime, 8.6 becauseeven at ll.7 Tesla the peaksare at least 10ppm kHz spectral width,22 psec 90" pulse excitation, 204E data wide. We report some of the narrowest singularities to points, zero-filled to 8192, 25 Hz linebroadening due to -+0.1 ppm with the last place subscripted. exponential multiplication.

Results and discussion from these spectra, together with sample locality and Silicon-29 composition. We present in Figures 2 and 3 the silicon-29 MASS The spectra of all the crystalline samples contain NMR spectra of some representativefeldspars and feld- considerable structure arising from silicon sites with spar glassesobtained at a magnetic field strength of 8.45 different electronic environments. The spectra of albite Tesla. Table I summarizesthe chemical shifts obtained and anorthite glass are both broad, single peaks with t12 KIRKPATRICK ET AL.: NUCLEAR MAGNETIC RESONANCE OF FELDSPARS

ttsi discuss the alkali feldspars whose spectra are shown in 9,45 T MASS Figure 2. The MASS NMR of low albite (Fie. 2,{) was first examined by Lippmaa et al. (1980) and has also been discussedby K. A. Smith et al. (1983)and J. V. Smith et al. (1984).Because low albite has three particularly well resolved peaks, it has become a standard material in many NMR laboratories. The assignmentsof the three peaks to the three silicon sites in albite (Harlow and Brown, 1980)listed in Table I are those of Lippmaa et al.

-rlo - (1980),based on qualitativearguments, and elaboratedby tou -{o - lUU - LOU K. A. Smith et al. (1983),based on a procedurethat uses cation-oxygenbond strength.The peak at -92.2 ppm is assignedto T2m, which has two aluminum next-nearest neighbors and two silicon next-nearest neighbors, Qa (2Al), and those at -96.3 and -104.7 ppm to T2o and Tlm, which are Q4(tlt) sites.These assignments are the sameas those of J. V. Smith et al. (1984)based on the mean secant of the Si-O-T angle. -L+O - 1uU -160 -+0 -lJJ - tou The silicon-29 NMR spectrum of microcline perthite -82't I (Fig. 28) exhibits five resolved peaks. The sample is a typical microcline perthite from a pegmatite. The micro- SYNTHETIC monrxrreI ANORTHITE GLASS i cline phase has a 13l-l3l powder X-ray peak separation I of -0.7E" (CuKa) and is a maximummicrocline. Our site I ; I assignmentsare based on the NMR measurementsof I fiA*"-; cation-exchangedfeldspars by J. V. Smith et al. (1984), -f- r-- which are diferent from those of K. A. Smith et al. (1983) -t+0 -100 -160 -rto -100 -160 PPIY PPIY based on the structure refinement of Brown and Bailey (196/.).The new assignmentsof J. V. Smith et al. give the 2eSi Fig. 3. 8.45 Tesla MASS NMR spectra(corresponding to sameorder of peaksas for low albite. We assignthe peaks a resonancefrequency of 71.5 MHz) of natural and synthetic at -96.0, -98.6 and -101.8 ppm in Figure 28 to T2m, plagioclase glass. (A) Oligoclase feldspars and a synthetic T2o and Tlm, respectively,of maximummicrocline. The (Bakersviile,North Carolina),2.7 kHz MASS, E2 scansat 300 small peak at -93, the added intensity under the peaksat sec recycle time, 8.6 kHz spectral width, 30 psec 90" pulse -96.0 -98.6 ppm peak -106.0 ppm excitation, 2048 data points, zero-filled to 8192, 25 Hz and and the small at linebroadening due to exponential multiplication. (B) belong to the small amount of intergrown low albite in the Labradorite (Sydney Mines, Labrador), 2.4 kHz MASS, 52 perthites. scansat 600 sec recycle time, 8.6 kHz spectral width, 9 psec 90' The spectraof low albite and microcline perthite clear- pulse excitation, 4096 data points, zero-filled to 8192,75 Hz ly demonstrate the high degree of Si/Al tetrahedral site linebroadeningdue to exponential multiplication. (C) Bytownite ordering that has been deduced for these minerals by X- (CrystalBay, Minnesota),2.3kHz MASS, 313 scansat 8.6 sec ray difraction. The fact that there is no recognizable recycle time, l0 kHz spectral width, 2.3 psec 23' pulse silicon-29signal for the Tlo site in either spectrum is also points, zero-filled to 8192, 25 Hz excitation, 4096 data consistent with the accepted models for these minerals linebroadeningdue to exponential multiplication. (D) Anorthite which propose that Tlo is occupied entirely by alumi- (PacayaVolcano, Guatamala),2.7 kHz MASS,64 scansat 600 sec recycle time, 8.6 kHz spectral width, 22 psec 90' pulse num. excitation, 204E data points, zero-filled to 8192, 25 Hz The spectrum of cryptoperthite in Figure 2E has some linebroadeningdue to exponential multiplication. (E) Synthetic similarities to those of albite and microcline perthite. anorthite, 2.5 kH.z MASS, 3l scansat 6C)0sec recycle time, 8.6 There are four resolved peaks, but the backgroundunder kHz spectral width, 25 trr,sec90o pulse excitation, 2048 data these peaks is higher than for the other two minerals. points, zero-filled to 8192, 25 Hz linebroadening due to This cryptoperthite comes from "Fredriksvarn" (now exponentialmultiplication. (F) Synthetic anorthite glass,2.4 kHz Stavern) in the well known Larvik region in southern MASS, 9l scansat 600sec recycle time, 8.6 kHz spectralwidth, Norway. Many of the cryptoperthites from this region are g,sec data points, zero-filled to 8192, 25 90'pulse excitation, 2048 moonstones.Our sampleis the much-studied"Spencer 100Hz linebroadeningdue to exponential multiplication. R" (Spencer, 1937).The cryptoperthites from the Larvik region are highly sodic, vary from place to place and are essentially no structure. Because silicon-29 has spin I = likely to be zoned compositionally and texturally (Spen- ll2 and does not suffer from quadrupolar broadening, cer,1937; Oftedahl, 1948;Muir and Smith, 1956;Smith these broad, structureless peaks imply a continuum of and Muir, 1958).Spencer R has a bulk composition (mole electronic environments around the silicon sites. We first percent) of about Or+rAbsrAnr (Spencer, 1937). The KIRKPATRICK ET AL.: NUCLEAR MAGNETIC RESONANCE OF FELDSPARS ll3 aggregateperthite of our sample has monoclinic optics. By qualitative comparisonof the -92.3, -96.0, and Albite exsolution lamellae, though very fine, are clearly -103.5 ppm peaksof SpencerR (Fig. 2E) with thosefor visible under the polarizing microscope, even at low low albitein Figure2A,we believethat the componentof magnifications. Nonetheless, transmission electron mi- the peak at -96.0 ppm (Fig. 2E) contriburedby the T2 crographsof SpencerR (e.g.Bollmann and Nissen, 1968) site of orthoclase is appreciably larger than the peak at show albite lamellaeless than 1000Aacross. -98.9 ppm contributedby the Tl site. This is consistent Stewart and Wright (1974)report the potassic phaseof with the characterization of the potassic phase of the SpencerR to be monoclinic (orthoclase)by powder X-ray cryptoperthite as orthoclase. difraction and to have a composition (mole percent) of We chosethis cryptoperthite for study partly becauseit Ors2. They also report it to have the second largest is thought to represent a potassic feldspar having a large amount of "strain" (Aa) of any potassic feldspar for number of strained sites in the coherent boundary regions which they have data. betweenlamellae (Stewart and Wright, 1974).Weexpect- It is to be noted that even the alkali feldsparsfrom the ed to observeconsiderably broadened NMR peaksarising Larvik area have significant calcium contents ranging from ranges of distorted silicon sites in these boundary from around Ana to at least Anle (Spencer, 1937; Ofte- regions. The individual peaks in the cryptoperthite spec- dahl, 1948;Muir and Smith, 1956;Smith and Muir, 1958). trum are just as sharp and narrow as those for low albite Ribbe (1983, p. 243), for example, reports that the ex- and microcline perthite. The unresolved background in solved plagioclaselamellae in cryptoperthite from Larvik the cryptoperthite spectrum, however, is much larger. (presumed by us to be similar to Spencer R) have a We interpret this background as signal arising from compositionin the rangeAnrs_ra (oligoclase). In line with strained sites in coherent boundary regions. The signal in the bulk composition Ora3Ab53Anafor Spencer R, trans- the narrow peaks is coming from the unstrained volumes missionelectron micrographsof this and similar perthites away from the lamellae boundaries. from the Larvik area show the cross-sectionalarea ofthe The silicon-29 NMR spectrum of sanidine (Fig. 2D) plagioclaselamellae to exceed slightly that of the ortho- consists of a broad doublet. This feldspar (Or6sAb36An2) clase. occurs as clear, glassy, I to 5 mm phenocrystsin rhyolite The NMR spectrum (Fig. 2E) of the cryptoperthite is from Wyoming. Our interpretation of the sanidine spec- dominatedby sodic plagioclasepeaks at -92.3, -96.0 trum is the sameas for the orthoclase in the cryptoperth- -103.5 and ppm becauseofthe high degreeofexsolution ite. The two overlapping peaks represent the two sets of and the high proportion ofsodic feldspar. Thesepeaks are tetrahedral sites, Tl and T2, which are the only two almost identical to the ones for low albite in Figure 2,4, present in monoclinic alkali feldspars. We infer that the and 28, despite the probably higher An content and peak at - 101ppm for sanidinecorresponds to the peak at higher temperature state of the plagioclase lamellae in -98.9 ppm for orthoclaseand arisesfrom the Tl site, and SpencerR. that the peak at -97 ppm in sanidine correspondsto the Only two peaks attributable to the potassic feldspar peak at -96.0 ppm for orthoclase and arises from T2. phaseare apparent:the added height ofthe peak at -96.0 Again, this is consistent with the low albite and micro- and the smallpeak at -98.9 ppm. cline data, for which T2m and T20 are less shielded(less In monoclinic alkali feldspars(orthoclase and sanidine) negative)than Tlm. thereare two setsof tetrahedralsites, Tl andT2.In ideal The peaks for sanidine are much broader than any of orthoclase, the Tl sites are thought to be randomly those discussed so far. Because silicon-29 has spin populated by aluminum and silicon ions; the T2 sites are | = l/2, there is no broadeningfrom quadrupolarinterac- thought to contain only Si ions. Such a phase should tions. Moreover, there should be no peak broadening exhibit two silicon-29NMR peaks, the one corresponding from field inhomogeneity because our magnetic field to T2 having about twice the area ofthe one correspond- inhomogeneity is less than I ppm. Consequently, we ing to Tl. In idealsanidine, on the other hand,the Tl and interpret the peak broadeningto arise from AVSi disorder T2 sites are both thought to be randomly populated by in the occupancy ofthe two tetrahedral sites. The appar- one-quarter aluminum and three-quarters silicon, on the ently equal areas under the two overlapping peaks indi- average. In this case, one expects two NMR peaks of cates an approximately equal distribution of silicon and about equal area. aluminum between the Tl and T2 sites, as expectedfor a Unfortunately, the unavoidable interference by the high sanidine.Because ofthis disorder,for eachofthese -96.0 plagioclase peak at ppm makes it diffcult to sitesthere are probably Q4(0Al), Q4(lAl), Q4(2Al), and estimatethe relative areasof the two peaksfor orthoclase Q4(3Al)sites for both Tl and T2. Becauseof the overlap- -96.0 at and -98.9 ppm. We infer that the low peak at ping of the peaks from each of these eight sites, and -98.9 is from the Tl site becausesuch a site is more becauseof KA.{a disorder, they cannot be resolved. The likely than a T2 site to resemble the Tlm site of micro- wings in Figure 2D in the range of 105 and 88 ppm are -101.8 cline havinga peak at ppm. The addedheight of probably due to Q410Al) and Q4(3Al) sites respectively. -96.0 the peak at ppm, therefore,is assignedto the T2 A significantconcentration of Q4(4Al) sites would not be site.This is closeto the peakat -98.6 ppm (Fie. 28) from expectedfor a phase with a SiiAl ratio of 3/1 (Klinowski the T2o site of microcline. et al., 1982). ll4 KIRKPATRICK ET AL,: NUCLEAR MAGNETIC RESONANCE OF FELDSPARS

We consider now the silicon-29 spectra of our plagio- there is a small peak at -87 ppm. Becausethe chemical clases (Fig. 3). We have already dealt with the most shifts are identical to those of albite, we assignthe peaks commonNa endmember,low albite.We considerfirst the at -93.2, -97.2, -105.3 ppm to the T2m, T2o and Tlm calciumend-member, anorthite, and then the intermedi- sitesof albite.We believethe low peakat -87 ppm arises ate plagioclases. from Q4(4Al) sites like those in anorthite.W. -H. Yang The spectrum (Fig. 3E) of the synthetic anorthite has (1984personal communication) has found a very similar three well resolved peaks: at -82.9, -84.8 and -89.6 spectrum for an An23plagioclase. ppm and a small peak at about 95 ppm. It is possible to interpret the oligoclase spectrum in The spectrum of the natural anorthite (Ansz), Figure terms of a single e-plagioclaseof composition An23, a 3D, is essentially identical to that of the synthetic anor- composition between the peristerite and Bpggild gaps. thite, exceptthat the peaksat -83.0 and -84.6 ppm are Mclaren (1974)gives transmissionelectron micrographs not as well resolved in the natural sample. The intensity of an oligoclase from the same locality as ours which in the range of -95 ppm may be due to Q4(3Al) sites, show cross-hatchedlamellae having a spacing of about although the signal intensity is at about the noise level. 1004. This is about the right order of magnitudefor the Anorthite at room temperature ("P-anorthite") has spacings(204) that Smith (1974)says are to be expected eight nonequivalentSi sites (Megaw et al., 1962;Wain- for the modulated structure of Ab-rich and An-rich do- wright and Starkey, l97l; Staehli and Brinkmann, mains in this compositional range. l974a,b;Kalus, 1978).The peaksin the anorthitespectra We believe that the peak broadening and overlap and can be assignedusing a correlation between the secantof the extra intensity in the -90 to -100 ppm range come the mean Si-O-Al angle for each site and the silicon-29 from a range of silicon site environments in strained chemical shifts, similar to those for the silica polymorphs regions between the Ab-rich and An-rich domains in the and alkali feldsparsdeveloped by J. V. Smith and Black- modulated structure. The fact that the cbll volume of well (1983)and J. V. Smith et al. (1984).This correlation albite (660A3)is less than the comparablehalf-cell volume gives three clusters of points using all three available (6704) of anorthite suggeststhat the bond angles are structure refinements (Megaw et al., 1962; Wainwright distorted and that the silicon tetrahedra of the albite and Starkey, 1967; and Kalus, 1978), while the bond domains may be slightly stretched whereas those of the strength sum and mean Si-O bond distance correlations anorthite domains may be slightly compressed in the of K. A. Smith et al. (1983)and Higgins and Woessner strained volumes around the coherent boundaries. The (1982)do not. The only discrepancyis the reversal of longer Si-O bonds in the outer parts of the albite domains Tl(ooi) and Tl(mzo) in the Wainwright and Starkey are related to decreased shielding and less negative structure, probably due to a long Al-O distance for their chemical shifts. The shorter Si-O bonds in the anorthite refinement. The variations in the bond strength sum and are related to increased shielding and more negative bond distanceare small(0.2 V.U. and0.005A, respective- chemical shifts. There are probable changes in T-O-T ly), and the predicted values are in the range of Q4(4Al) bond angles also. The spectrum for this composition, sites(Lippmaa et al., 1981;K. A. Smithet al. (1983). therefore, is albiteJike becausethe "albite" domains are . The small peak for the synthetic anorthite at about -95 larger. Becauseof this the ratio of strained to unstrained ppm is probably due to Q4(3Al) sites, which indicate a volumes of the albiteJike material is relatively low. lack of perfect AVSi order in this rapidly crystallized Because the volume of An-rich domains is small, all or sample.The poorer resolution of the peaks in the natural most of this material is strained. This accounts for the anorthite (AnqJ is probably due to the albite component. lack of anorthite signal in the -83 ppm range. In addition, Many of the peaks of the spectra of the intermediate there could be a significant number of Q4(3Al) sites along plagioclasesin Figures 3,4,,38, and 3C can be assignedto the boundaries between the domains contributing signal silicon sites similar to those in albite or anorthite. For in the range -90 to -95 ppm. none of thesesamples, however, can all the peaksof both The small peak at about -108 may be due to Q4(0Al) albite and anorthite be recognized. Consequently, a sim- sites in the boundaries between domains or to quartz ple model involving only domains of albite and anorthite impurity, which was not detected optically or by X-ray. structures and compositions cannot account for the ob- The spectrum of labradorite (Fig. 3B) shows rather ill- servedspectra. We can, however, account for the spectra definedpeaks or shouldersat -94, -101, and -107 ppm in terms of the earlier-mentionedlow albite, e-plagioclase which may be assignableto Q4(zat) and Q4(lAl) albite- and /-anorthite structures. We attribute much of the like sites, but more signal than the oligoclase at less -88 signal in each spectrum to sites in strained volumes in e- negative chemical shifts, a poorly resolved peak at plagioclase. ppm, and a shoulder at -83 ppm. The significant amount The spectrum of oligoclase in Figure 3A, has many of signalnear -83 ppm indicates that the most deshielded similarities to that of albite (Fig. 2A). The principal Q4(4Al) anorthitelike sites are present in this sample. differencesare that the peaks at -93.2, -97.2, and This labradorite (AnsJ falls in the B/ggild composition- -105.3 ppm for oligoclaseare broader,overlap consider- al gap, and probably consists of two e-plagioclases.It ably more, and have diferent relative heights, and that seemslikely that the albite-like peaks arise mostly from KIRKPATRICK ET AL.: NUCLEAR MAGNETIC RESONANCE OF FELDSPARS

Q4(lAl) and Q4(2Al) sites in albite-rich volumesin the Q4(4Al). Becauseof the rangeof sites in albite glass,it more albite-rich e-plagioclase.As for the oligoclase, the seemslikely that a range of sites are present in anorthite -93 -Ifi) extrapeak height in the to ppm range probably also.This is consistentwith the presenceof Q4(3Al) sites comes from sites in the strained regions around the in crystalline anorthite. Becausethe amount of signal in marginsof the albite volumes. If the calcium-rich phasein the silicon-29 spectrum of anorthite glass at values more this material is also an e-plagioclase,some contribution to shielded (more negative) than - 100 ppm is small, the the albite-like peaks must also come from albite volumes concentration of Qa(0Al) sites must be quite small. The in this material. The intensity at less negative chemical presenceof a significant amount of signal at values less -88 shifts (the peak at ppm and the shoulder at -83 ppm) shielded (less negative) than -80 ppm must be due to comes from Q4(4Al) anorthite-like sites, which occur in distortedQ4(4Al) sites.Some of thesemay be causedby both e-plagioclases.As for the oligoclase, some signal in the presenceofAl in the secondtetrahedral coordination the central part of the spectrum could be due to e4(3Al) shell. In perfectly ordered anorthite all ofthese would be sites at the margins of albite and anorthite-rich volumes. occupied by Si, but in AVSi disordered material some of The relatively shielded value of the peak at -107 proba- these must be occupied by Al. bly indicates some contribution from e4(OAt) sites. The bytownite spectrum (Fie. 3C) shows poorly re- Aluminum-27 -84 -89 solvedpeaks at and ppm assignableto e4(4Al) Figures 4 and 5 show the ll.7 Tesla aluminum-27 -95 -l0l anorthitelike sites, shoulders at and ppm MASS NMR spectra of our samples. Table 3 lists the assignableto Q4(2Al) and e4(lAl) albitelike sites, and little signal in the range of - 105ppm, although there is a small shoulder that may be real. AI II,7T MASS This bytownite (An7s) probably consists of (volumes 57 of) I-plagioclase and e-plagioclase. The sites, e4(4Al) MICROCLINE which we believeto causethe peaksat -84 and -89 ppm, PERTHITE are in both materials.The and siresare Q4(2Al) e4(lAl) I in only the e-plagioclase.It seems likely that no - 105 ppm peak is present(or it is very small) becauseessential- ly all of the albite volume is strained, causing all of the silicon sites in the albite to becomedeshielded. The peak -84 at ppm is probably due to overlap of the -82.9 and -84.8 peak ofpure anorthite. The silicon-29NMR spectra of the albite and anorthite glasses(Figs. 2C and 3F) consistof singlebroad peaks, with no apparent structure. They are each approximately 12 to 15 ppm broader than the total peak breadth of the correspondingcrystalline phases.The peak for albite has a maximum at -98 ppm and covers the range from about -85 -ll5 ppm to ppm. The peak for anorthite has a Ioc c oor5o maximum at -87 ppm and covers the range from about 27Al -72 to - 105ppm. Fig. 4. ll.7 Tesla MASS NMR spectra(corresponding to Assuming, as discussedbelow, that all the aluminum in a resonance frequency of 130.3 MHz) of alkali feldspars and albite glass. (A) Albite (Amelia, Virginia), 3.5 kHz MASS, l0 these glasses is in tetrahedral coordination, both must scansat 25 sec recycle time, 15 kHz spectral width, 3 90' have a fully polymerized framework structure. trr,sec This is pulse excitation, 8192data points, 10 Hz linebroadeningdue to because the ratio of tetrahedral cations (Si + AI) to exponential multiplication. (B) Microcline perthite (Delaware oxygen is l/2. All the silicon sites must, therefore, be Co, Pennsylvania),3.5 kHz MASS, l0 scansat 25 sec recycle Qa(nAl).One question is what valuesdoes n have.W. -H. time, 15 kHz spectralwidth, 3.5 psec 90'pulse excitation,8192 Yang (1984,personal communication) has examined the data points, l0 Hz linebroadening due to exponential silicon-29MASS NMR behavior of analbite produced by multiplication. (C) Albite glass (synthetic), 5 kHz MASS, 200 heating Amelia albite at 1073'C for 120 days and scans at I sec recycle time, 15 kHz spectral width, 3 psec 90' pulse points, found that detectable amounts of e4(3Al), e4(2Al), excitation, Sl92 data 50 Hz linebroadeningdue to (D) (Wyoming), Q4(lAl), and Q4(0Al) are present. The signal in rhis exponential multiplication. Sanidine 3.2 kHz MASS, 158scans at 25 secrecycle time, 15kHz spectralwidth, spectrum covers the range from -85 to -115 ppm. 3.5 90o pulse excitation, 4O96 data points, 50 Hz Becausethe spectrum of albite glassextends ;r.sec to even less linebroadening due to exponential multiplication. (E) shieldedvalues, it seems likely that detectableamounts of Cryptoperthite("Fredriksvarn", Norway), 3.2kHz MASS, 76 all five possible types of silicon sires et4Al) through scans at 5 sec recycle time, 15 kHz spectral width, 4 psec 90' Qo(Oetl are present in albite glass. pulse excitation, 4096 data points, 50 Hz linebroadeningdue to For anorthite glass the average silicon site must be exponential multiplication. 116 KIRKPATRICK ET AL.: NUCLEAR MAGNETIC RESONANCE OF FELDSPARS t'At lt,7T MASS at 59.3ppm and a full-width at half-heightof about 8 ppm. We assignthis peak to the Tlo site. 160 A /\.o B Figure shows observed and simulated aluminum-27 l, 6 r, OLIGoCLASE , LABRADORITEI and static and MASS NMR spectra for albite at 11.7, r'l' \,./\ 8.45, and 3.5 Tesla. The great narrowing of the spectra I t f'-' with increasingmagnetic field strength demonstratesthat \----..- d \n --'^-j the peak shapeis dominatedby the second-orderquadru- pole interaction. The resolution increasesas the squareof the field strength (Meadows et al., 1982).The simulations in Figure 6 are in very good agreementwith the experi- mental spectra. These simulations assume one kind of aluminum tetrahedral site with a quadrupole coupling constant of 3.29 MHz and an asymmetry parameter of 0.62. Thesevalues are basedupon and in good agreement with the earlier single crystal NMR measurementsfor albite by Hafner and Hartman (1964). The simulations yield an isotropicchemical shift of 62.710.3ppm. This is in the range of chemical shifts expected for tetrahedral lOC ;c 0 toc 50 PPtl 116 AlCl3 PPlil frm AlCl3 aluminum(Mtiller et al., l98l). The spectrum of microcline perthite (Fig. 48) consists Fig. 5. I 1.7Tesla 27Al MASS NMR spectra(corresponding to of a doublet having maxima at 54.3 and 57.3 ppm. Single a resonancefrequency of 130.3MHz) of natural and synthetic crystal aluminum-27NMR studies by Hafner and Hart- plagioclase feldspars and anorthite glass. (A) Oligoclase mann (1964)have shown that microcline has one type of (Bakersville, 36scans at l0 sec North Carolina),36kH2 MASS, aluminum with a quadrupole coupling constant of 3.27 recycletime 15kHz spectralwidth, 5 psec 90'pulse excitation, MHz and an asymmetryparameter of 0.21. Figures 7C 4096 data points, 50 Hz linebroadening due to exponential MASS NMR spec- multiplication. (B) Labradorite (Sydney Mines, Labrador), 3.3 and E show the observed aluminum kHz MASS, 1065scans at 5 sec recycle time, 15 kHz spectral trum of microcline perthite and a simulation assumingthe width, 4 psec 90' pulse excitation, 4096 data points, 50 Hz Hafner and Hartmann values. The agreementis not very linebroadeningdue to exponential multiplication. (C) Bytownite good. There is additional intensity in the experimental (Crystal Bay, Minnesota), 3.4 k}lz MASS, 549 scans at 5 sec spectrum around 60 ppm. We believe that this is due to recycletime, 15kHz spectralwidth, 4 p.sec90'pulse excitation, aluminum in the exsolved albite. Figure 7D shows a two 4096 data points, 50 Hz linebroadening due to exponential multiplication. (D) Anorthite (PacayaVolcano, Guatamala),3.3 shifts of feldspars and kHz MASS, 500 scans at 5 sec recycle time, 15 kHz spectral Table 3. Aluminum-27 NMR chemical glasses width, 4 pcsec90'pulse excitation, 4096 data points, 50 Hz feldspar linebroadeningdue to exponential multiplication. (E) Synthetic FIILD lp"mutl anorthite,5.1 kHz MASS, 1300scans at 10sec recycletime, 15 UTERIAL STRENGTH CHEMICAL kHz spectralwidth, 4 psec 90'pulse excitation, 4096data points, ( TESI,A) STIIFT (F) -24,2,ZO,t3 50 Hz linebroadening due to exponential multiplication. AIbtte (Ab99) I sec a.45 53,56 Syntheticanorthite glass, 4.9 kHz MASS, 1889scans at 59 recycletime, 15kHz spectralwidth, 4 psec90'pulse excitation, Albtte Class (AblOO) 3,5 4096 data points, 50 Hz linebroadening due to exponential I .45 54 multiplication. II.7 55 -0160) Sanldine ( TI.7 58

ilicrocllne Perthlte 8.45 50,55 54,57

ite 1t .7 58 observed chemical shifts. Aluminum-Z7 has spin | : 512 Cryptoperth It.7 60 and, therefore, the aluminum-27NMR spectra are affect- 0ligoclase (An23) ed by peak broadening and frequency shifts due to Labradortle (An52) lI.7 60 quadrupoleinteractions as discussedabove. This quadru- Bytocntte (An70) lI.7 62,56 polar broadening prevents us from obtaining as much Anorthite (An92) Ll.7 61,55 as with silicon-29. AII of Anorthtte (Anl00) 45 detailed structural information 8.45 58 the observed aluminum-27chemical shifts are consistent tL.7 62,55 with aluminum in four-fold coordination (Miiller et al., Anorthite Class (Anlo0) 55 8,45 1981;Fyfe, et al.,1982). 11.7 60 We consider first the aluminum-27NMR spectra of the (Fig. 'lFor quadrrpolar nuclel appareot chemlcal shift changes cith fleld strength. alkali feldspars(Fig. 4). The spectrumfor low albite Value at 11.7 Tesla ls essenttally tle true lsocropic chentcal shift. 4A) consistsof a single asymmetricpeak with a maximum KIRKPATRICK ET AL.: NUCLEAR MAGNETIC RESONANCE OF FELDSPARS |7 'hr aLBrtE maximum is at 58.5 ppm, intermediatebetween low albite mss and cryptoperthite. Single crystal NMR studies of sani- dines (Brun et al., 1960)have not yielded discrete rota- tional patterns for different types of aluminum. The high resolution aluminum-27 MASS NMR lineshape of sani- dine (Fig. 4D) agrees with this and is consistent with a range of quadrupole coupling constants, asymmetry pa- rameters and/or isotropic chemical shifts. The Ho field dependenceofthe lineshape(data not shown) shows that --r[. the average quadrupole coupling constant is about 3 MHz, becausethe location and breadth of the centerband is similar to those of albite and microcline. We consider now the aluminum-27 spectra of the 20a o a -zao plagioclases.The spectrum of synthetic anorthite (Fig. 5E) is more complicated that those of the alkali feldspars. 27Al Fig. 6. static and MASS NMR specrra of albite along It consistsofa peak at 61.5ppm and a broad shoulderat with the MASS NMR computersimularions at I1.7, g.45and, 3.5 about 55 ppm, and is strongly asymmetric. Tesla.(A) ll.7 Tesla,static, 30 scansat 25 sec recycletime, 7g Anorthite at room temperaturehas eight slightly differ- kHz spectralwidth, 3 psec 90" pulse excitation, 4096data points, 50 Hz linebroadeningdue to exponentialmultiplication. (B) I 1.7 Tesla,4.0 jg kHz MASS, l0 scansat 25 sec recycletime, kHz 'ht spectral width, 3 psec 90opulse excitation, 8196data points, l0 n,7T MASS Hz linebroadeningdue to exponential multiplication. (C) ll.7 Tesla, MASS simulation, 65 Hz linebroadening.(D) g.45 Tesla, SPECTRUM SIMULATION static, 50 scansat 25 sec recycle time, 56 kHz spectral width, 3 psec 90opulse excitation, 4096data points, 50 Hz linebroadening due to exponential multiplication. (E) 8.45 Tesla, 4 kHz MASS, 50 scansat 25 sec recycle time, 56 kHz spectralwidth, 3 !.sec 90o pulse excitation, 8196data points, l0 Hz linebroadeningdue to exponentialmultiplication. (F) 8.45Tesla, MASS simulation. 150 Hz linebroadening. (G) 3.5 Tesla, static, 300 scans at 5 sec recycle time, 24 kHz spectral width, 3 psec 90. pulse excitation, {D6 data points, 100 Hz linebroadening due to exponential multiplication. (H) 3.5 Tesla, 4 kHz MASS, 300 scans at 5 sec recycle time, 24 kHz spectral width, 3 psec 90opulse excitation, zOlX data points, 100 Hz linebroadening due to exponential multiplication. O) 3.5 Tesla, MASS simulation, 100 Hz linebroadening.

component simulation assuming 75Vo microcline and25Vo albite. Figures 7A and B show the observed albite spec- 'io trum and a simulation at ll.7 T for comparison. The PPM agreement between the two component simulations and Fig. 7. ll.7 Tesla, 27Al MASS spectra and computer the observed spectmm is excellent. The simulation (Fig. simulationsof albite and microcline perthite. (A) Albite, 4.0 kHz 7E) yields an isotropic chemical shift for microcline of MASS, l0 scansat 25 sec recycle time, 4 kHz spectral width, 3 58.5+0.5 ppm. As for the albite, the aluminum-27 NMR g,sec90" pulse excitation, E196data points, l0 Hz linebroadening data for microcline clearly indicate that essentially all the due to exponential multiplication. (B) Albite, computer aluminum is in one tetrahedral site. simulation using quadrupolecoupling constant of 3.29 MHz and The spectrum for the cryptoperthite (Spencer R, Fig. asymmetry parameter of. 0.62, 65 Hz linebroadening. (C) 4E) consists of a single symmetric peak having a maxi- Microcline perthite, 4.0 kHz MASS, 20 scans at 15 sec recycle mum at 57.9 ppm. As mentioned earlier, the bulk compo- time, 4 kHz spectralwidth, 3 g,sec90o pulse excitation, El96 data points, sition of this sample is about Ora6: there is a dominant l0 Hz linebroadeningdue to exponential multiplication. (D) Microcline perthite computer simulation using sum of albite oligoclase and a less abundant orthoclase. The quadrupo- simulation (B) and a second component having a 3.22 MHz lar broadening masks the splitting of the peaks from the quadrupolecoupling constant and asymmetry parameterof 0.21, Al sites in the orthoclase and oligoclase. 75 Hz linebroadening due to exponential multiplication. (E) The spectrum (Fig. of sanidine aD) is also a single Computer simulation using 3.22 MHz quadrupole coupling symmetric peak, but is slightly broader than those of the constant and asymmetry parameter of 0.21 and 75 Hz other alkali feldspars (12 ppm wide at half-height). The linebroadeningdue to exponential multiplication. llE KIRKPATRICK ET AL.: NIJCLEAR MAGNETIC RESONANCE OF FELDSPARS ent aluminumtetrahedral sites. Single crystal NMR stud- 'uNo ies yield eight different electric field gradient tensors with n,7T MASS quadrupolecoupling constants ranging from 2.66 to 8.42 MHz (Staehli and Brinktnann, l974a,b). The broadened peak in the MASS spectrum is due to overlap of the MICROCLINE signalsfrom the different Al sites. PERTHITE The aluminum-27spectra of the four natural, intermedi- ate plagioclases(Figs. 5A, 58, 5C, and 5D) all have peak maximabetween 60.2 and 61.5 ppm. The peaksbecome broader and more like those of the synthetic anorthite with increasing An-content. Even for the Pacaya anor- thite, the large peak and shoulder are less resolved than for the synthetic anorthite. The quadrupole broadening preventsresolution of the many closely spacedpeaks' The aluminum-27NMR spectra of albite and anorthite glass (Figs. 4C and 5F) consist of broad, structureless peaks (about 18 and'28 ppm wide, respectively)with maxima at 55.0 and 59.6 ppm, respectively.Quantifica- tion of the total aluminum content in albite glassusing the heightof the free-inductiondecay yields about 75 percent of the expected intensity. Once again we interpret these spectra as indicating that the aluminum sites in the glass have a wide range of quadrupole coupling constants, -50 asymmetry parameters, and/or isotropic chemical shifts. The anorthite glass spectrum is skewed to less positive Fig. 8. 23NaMASS NMR spectra of natural alkali feldspars chemical shifts, with a very broad underlying component. and synthetic albite glass at ll.7 Tesla (coresponding to a (Amelia, Quantification of the aluminum indicates that only about resonance frequency of 132.3 MHz)' (A) Albite 50Voof the expectedintensity is observed. As for the Virginia),4.0 kHz MASS, 30scans at 25sec recycle time' 15kHz points, l0 albite glass, we interpret this low intensity as indicating spectral width, 4 (psec 90opulse excitation, 8192data multiplication. (B) that the aluminum in the glass is in a wide range of Hz linebroadening due to exponential Microcline perthite (Delaware Co., Pennsylvania), 4.0 kHz electronic environments, many of which have large quad- MASS, 100scans at 25 secrecycle time, 15kHz spectralwidth, 3 rupole coupling constants. ;.r.sec90o pulse excitation, 8192data points, l0 Hz linebroadening Little is known about the effects of Al-O bond dis- due to exponential multiplication. (C) Albite glass (synthetic)' tances or angles, total cation-oxygen bond strength, or 4.7 kHz MASS, 1ffi3scans at I secrecycle time, 15kHz spectral other structural parameters on aluminum-27 chemical width, 4 psec 90'pulse excitation, 4096 data points, 150 Hz shifts and quadrupole coupling constants, so it is not linebroadeningdue to exponential multiplication. (D) Sanidine possible at the present time to give a more detailed (Wyoming),4 kHz MASS, 159scans at 5 sec recycle time, 15 analysisof our resultsin structuralterms. It seemslikely, kHz spectralwidth, 3.5 psec 90'pulse excitation' 4096 data multiplication. however,that at leastpart ofthe peak broadeningis due points, 50 Hz linebroadeningdue to exponential (E) ("Fredriksvarn", Norway), 4 kHz MASS' to a wide range of electric field gradients around alumi- Cryptoperthite 708scans at 5 sec recycle time, 15kHz spectralwidth, 4 psec 90o num in the glasses.There is no signalin the chemical shift pulse excitation,4096 data points, 50 Hz linebroadeningdue to ppm, indicate octahedrally- range near 0 which would exponentialmultiPlication. coordinated aluminum, although becausewe do not ob- serve the full intensity expectedon the basis of aluminum content, we cannot rule out the presence of highly distorted, six-coordinated sites. X-ray radial distribution studies(Taylor and Brown, 1979)and Raman spectrosco- has a smallernuclear spin than aluminum-27,for the same py (Mysen, et al., 1982)do, however, suggestthat 6- quadrupole coupling constant, sodium has a larger sec- coordinated aluminum is unlikely to be present in these ond-order quadrupole breadth and hence broader lines glasses. (Meadowset al., 1982). The sodium-23 spectrum of low albite (Fig. 8A) con- -21.9 Sodium-23 sists of a doublet with peaks at -13.1 and ppm' We show in Figures 8 and 9 the ll.7 T sodium-23 The split peak is not due to the presenceof two sodium quadru- MASS NMR spectra for all of our sodium-containing sites in albite, but is causedby the secondorder samples.Table 4lists the chemical shifts at the singulari- pole interaction, which is only partially averaged by ties. Sodium-23 has spin I = 312 and like aluminum-27 MASS. Single crystal sodium-23 NMR studies of low suffersfrom quadrupolarbroadening. Because sodium-23 albite have yielded a quadrupolecoupling constantof 2'62 KIRKPATRICK ET AL.: NT]CLEAR MAGNETIC RES0NAN1E oF FELDSPARS ll9 tuNo n,7T MASS Figure I I, which generatesdifferent lineshapesdepending on the spinning angle, when the secondorder quadrupole interactiondominates (Oldfield et al., 1983;Ganapathyet al.,1982). As expected, better averagingis obtained near 40o or 75". Note that at 75' the first order spinning sidebandsare absent (Ganapathyet al., 1982). The sodium-23spectrum of microcline perthite at ll.7 T (Fig. EB)shows peaks at -13.1, -21.9 and -26.5 ppm. From the Ho magnetic field dependenceof the lineshape (data not shown) there appearto be two types of sodium, one in exsolvedlow albite (-13.1 and -21.9 ppm) and onein microcline(-26.5 ppm). The -26.5 ppm line is not appreciably shifted in going from 8.45 T to ll.7 T, suggestingthat the sodium site in microcline has a smaller electric field gradient and, hence, is in a more symmetric environment than in albite. Spectra of microcline ob-

n -cn o _En tainedat ll.7 T simulatewell with an albite component PPMfrom NoCl PPM frodt NoCt and an additional llVo contibution from a line with a 23Na quadrupolecoupling constant of about Fig. 9. MASS NMR spectraof natural plagioclase 0.5 MHz, this additional feldsparsat IL7 Tesla(corresponding to a resonancefrequency intensity being from sodium in the microcline. of 132.3MHz). (A) Oligoclase(Bakersville, North Carolina). 3.6 The low quadrupole coupling constant is in agreement kHz MASS,ll8 scansat l0 secrecycle time, 15kHz spectral with the small magnetic field dependenceof the chemical width, 5 psec 90' pulse excitation,4096 data points, 50 Hz shift. However, because equally good spectral simula- linebroadeningdue to exponentialmultiplication. (B) tions can be obtained using two powder patterns having Labradorite(Sydney Mines, Labrador),3.3 kHz MASS, 725 e2qQ/hof about 3 MHz, further work is neededin order to scansat 5 secrecycle time, 15 kHz spectralwidth, 4 psec90. substantiatethis very tentative assignment. pulseexcitation, 4(D6 data points, 50 Hz linebroadeningdue to The sodium-23 MASS NMR spectrum of the crypto- exponentialmultiplication. (C) Bytownite(Crystal Bay, perthite (Fig. 8E) shows a doublet due to the second- Minnesota),3.4kHz MASS, 314 scans at 5 secrecycle time, 15 order quadrupoleinteraction. These peaks are very simi- kHz spectralwidth, 4 ,rsec90o pulse excitation, 4096 data points, lar to those 50 Hz linebroadeningdue to exponentialmultiplication. (D) of albite, except that they are less well Anorthite(Pacaya Volcano, Guatamala), 3.4kHz MASS,1000 resolved. This spectrum is in agreementwith the silicon- scansat 5 secrecycle time, 15kHz spectralwidth, 4 psec90. 29 and aluminum-Z7data, which indicate that some Si/Al pulseexcitation, 21096 data points, 100 Hz linebroadeningdue to ordering has occurred in this material, but that neither is exponentialmultiplication. as complete as in low albite or microcline perthite.

Table 4. Sodium-23NMR chemical shifts of feldspars and MHz and an asymmetryparameter of 0.25 (Hafner and feldsparglasses Hartmann, 1964).The static and MASS lineshapesas a function of magneticfield strength(Fig. FrdLD p"anettt l0) are consistent STERIAL STRENGTH frE{I(;AL MMNTS with the lineshape being dominated by the quadrupole ( tEsLA) SqISl interaction. The 3.52T MASS spectrumis very complex, (Ab99) Albtte Quad(upole Spttt containing many spinning sidebands,due to the fact that -I1.1 the spinning r t.7 Qudrupore Spllr speedis much less than the spectralbreadth, -21.9 when expressedin frequency units. AlblEe clasB (Abroo) rl.7 -13 We haveobtained an excellentsimulation of the I1.7 T sanrdtne ( "or60) -21 MASS spectrum of albite (Fie. il) assuminga single Ilcrocltne Perrhtre -Il sodium site with a quadrupole coupling constant of 2.5g -26 MHz and an asymmetryparameter of 0.25, and yielding -r3 Crtptoperthtre Quadrupole Spllr an isotropic chemical shift of -6.8-{-0.3 ppm, in good -22 agreement with the value of -7.3 ppm obtained by olt8oclase (An23) lI.7 -15 Lippmaaet al. (1980).The agreementbetween the experi- Labradortt€ (An52) 11.7 -I6 mental and simulated spectra clearly indicates that there Bytohlte (An70) ll.7 -14 is only one sodiumsite in albite. Ano(Ehlte (e92) lr.7

The values of the quadrupole coupling constant and rFor quadrupolar nuclet apparedr cheolcat shlrr chanS€s di!h fteld srreErh. If asymmetry parameter of sodium-23 in albite are peak le oot spllt by second-order quadrupolar effecrsr value ar tt.7 Tesla ts con- e€e€nttally th€ t3orroplc chetrtcal shtft. firmed by variable angle sample-spinning(VASS) NMR, r20 KIRKPATRICK ET AL.: NUCLEAR MAGNETIC RESONANCE OF FELDSPARS

ttNo of sites with different quadrupole coupling constants, ALBTTE asymmetry parameters, and/or isotropic chemical shifts' STATIC In the plagioclases, much of this is probably due to distortion of the sodium sites near the margins of albite and anorthite volumes. The relatively large spinning il,77 sidebandsin these spectra, like the aluminum-27 NMR spectra of these same samples(Fig. 5), are probably due to the presence of magnetic impurities (Oldfield et al.' 1983). The sodium-23NMR spectrum of albite glass (Fig' 8C) with a 8,457 is very broad and featureless,again indicating sites range of quadrupole coupling constants, asymmetry pa- rameters. and/or chemical shifts.

SummarY and conclusions The foregoing results demonstrate that MASS-VASS NMR spectroscopy complements well the other main, reasonablydirect methods of investigatingthe structures of feldspars-single-crystal X-ray and neutron diffraction, ,T electron microscopy .t 5C3 electron diffraction, transmission PPMt?6 NoCl (TEM) and single crystal NMR spectroscopy' Each of these methods has its strengths and its weaknesses'The Fig. 10. 23Nastatic and MASS NMR spectraof albiteat 11.7, particular strength of NMR spectroscopyis that in many 8.45 and 3.5 Tesla. (A) 11.7Tesla, static,30 scansat 25 sec recycle time, 132kHz spectral width, 3 psec 90'pulse excitation' 4096 data points, 50 Hz linebroadening due to exponential ttNo T VASS multiplication.(B) 1l .7 Tesla,4.0 kHz MASS, 30 scansat 25 sec 8,45 recycle time, l3zkHz spectralwidth, 3 psec 90" pulseexcitation, SPECTRUM SIMULATION El96 data points, l0 Hz linebroadening due to exponential A (C) 8.45 Tesla, static, 100scans at 25 sec recycle multiplication. 750 time, 95 kHz spectralwidth, 3.5 psec 90'pulse excitation'4096 data points, 50 Hz linebroadening due to exponential multiplication.(D) E.45Tesla, 4.8 kHz MASS, 100scans at 25 sec recycle time, 95 kHz spectral width, 3 psec 90' pulse excitation, El92 data points, 50 Hz linebroadening due to exponentialmultiplication. (E) 3.5 Tesla, static, 1000scans at 25 sec recycle time, 39 kHz spectral width, 3.5 psec 90' pulse excitation, 40!)6 data points, 100 Hz linebroadening due to exponentialmultiplication. (F) 3.5 Tesla, 3'5 kHz MASS' 1000 scansat 15 sec recycle time, 39 kHz spectral width, 3 psec 90o pulseexcitation,4096 data points, 100Hz linebroadeningdue to exponential multiplication.

The sodium-23 MASS NMR spectrum of sanidine (Fig. D 8D) consists of a single broad peak at -21'2 ppm, which STATIC is slightly broader than those of albite and microcline perthite. The lack of a well resolved quadrupolar splitting for sanidine suggests to us, once again, that Al/Si disorder of sodium sites with different quadru- -1JJ -1oc results in a range loJ J.," rco orrll pole coupling constants, asymmetry parameters' and/or chemical shifts. The presence of a split sodium-23 peak in alkali feldspars appears to be a good indicator of AUSi ordering. The sodium-23 MASS NMR spectra of the intermediate plagioclases (Fig. 9A-9D) all show single, relatively sharp -13.0 -15.9 centerband resonances in the range of to ppm, that are similar in shape and breadth to the sanidine spectrum. As for sanidine, this probably indicates a range anele),(C) 40', (D) static' KIRKPATRICK ET AL.: NLICLEAR MAGNETIC RES2NANCE oF FELDSPARS 12l

instances provides it direct evidenceabout the electronic silicon-29NMR spectrum is relatively well separatedinto environment of individualkinds of atomic sites,even for a number of peaks. Three of these can be assignedto those ditrering as little from one another as the Tlm, T2m albiteJike sites and to Tl and T2 sites in orthoclase (low and T2o silicon sites in the triclinic alkali feldspars. sanidine).The resolutionofthe peaksis, however,not as Moreover, MASS-VASS NMR allows direct and easy clear as for the microcline perthite. There is only one comparison of sites in crystalline and noncrystalline aluminum-27 resonance, again indicating that neither materials. MASS-VASS NMR spectroscopyuses easily phase has a compositionapproaching an end member. obtained powder samples whereas single-crystal NMR The sodium-23 NMR spectrum shows a somewhat spectroscopy requires difficult-to-obtain, large, high_ smoothedout doublet due to quadrupolar splitting, again quality singlecrystals. The major limitationson MASS- suggestive of at least enough Al/Si ordering to cause VASS NMR samplesare that they must be relatively free distortion of the sodium site. There is no sodium-23 paramagnetic of impurities and that at least about l-10 mg resonancespecifically assignableto the orthoclase. for sodium-23 and aluminum-27 andabout 100-500mg for Sanidine. Because sanidine is highly disordered,its silicon-29 must be available. spectra are also relatively easy to interpret. The two Although MASS-VASS NMR spectrometerscurrently broad, poorly split silicon-29 NMR peaks are due to a are quite expensiveand not numerous,this is likely to rangeof silicon sites causedby AUSi disorderin the Tl change for the better in the near future. and T2 sites and to Or-Ab solid solution. These peaks We now summarizefor each type of material we have cover the chemical shift range of the albite spectrum and examinedthe resultspresented above. Again, we empha_ indicate a continuum of electron distributions in e4(0Al), size that while our interpretations are consistent with the Q4(1Al), Q4(2Al), and e4(3Al) sires. The aluminum-27 available knowledge of and ideas about feldspar struc- NMR spectrumconsists of a singlepeak which is broader ture, they must at this time be consideredtentative. Our than that of albite, also suggestinga range of aluminum interpretations of the intermediateplagioclase spectra are environmentsin Tl and T2. The sodium-23resonance is the most uncertain. not split by quadrupolar effects, suggestinga range of sodium environments Alkali and probably a more symmetric feldspars envrronment. Albite. Becauseit has an orderedand well-understood structure, the NMR spectra of low albite are the least Plagioclases ambiguous to interpret. The silicon-29NMR spectrum Anorthite. The silicon-29NMR spectrumof synthetic hasthree well-resolved peaks of nearlythe sameintensity anorthiteconsists of three Q4t+et)peaks that are assign- due to the three differentsilicon sites.The aluminum-27 able to the eight silicon sitesin anorthite plus a small peak NMR spectrum is a single relatively narrow peak, due to that we assign to Q4(3AD sites. The silicon-29NMR the onetetrahedral aluminum site. Both the silicon-29and spectrum of the natural anorthite (Ansr, a phenocryst aluminum-27 data clearly support the high degreeof Al/Si from Pacayavolcano (Guatamala), is similarbut lesswell tetrahedral site ordering deduced by X-ray diffraction. resolved.The aluminum-27NMR spectrumof both anor- The sodium-23 spectrum consists of a sharp, well-re- thites are broad, with one well resolved peak, and a solved doublet due to the second-orderquadrupolar inter- shoulderat lesspositive (more shielded)chemical shifts. action, indicativeof a rather asymmetricsodium site. All MASS NMR is apparently not able to resolve the eight of theseconclusions are consistentwith the well-known different aluminum sites. The sodium-23 spectrum of structure of albite. natural anorthite is discussedbelow along with those of Microcline perthite. The spectra of this sample reflect the other intermediate plagioclases. presence the of both microcline and low albite. The Intermediateplagioclases. The spectraof the interme- silicon-29 NMR spectrumcontains three peaks which are diate plagioclasesseem to be consistentwith structural assignableto the threesilicon sites in microcline,and two models involving e- and /-plagioclaseswith e4(4Al) anor- small peaks due to the small amount of intergrown albite. thite-like sires,e4(lAl) and e4(2Al) albite-likesites, and The third albite peak apparently underlies two of the distortedequivalents ofall threein strainedregions where microcline peaks. The aluminum-27 NMR spectrum of the albite-likeand anorthite-likevolumes meet. e4(3Al) the perthite consists of a quadrupole split doublet from sitesmay also be presentat the boundaries. the microcline,and extra intensitydue to the albite. The The silicon-29NMR spectra of the oligoclase(An3) sodium-23 NMR spectrumconsists of a quadrupolarsplit and labradorite (An5) exhibit three peaks that can be doublet,due to low albite, and a secondcomponent due readily assignedro Q4(1Al)and e4(2Al) sites in albite. to sodium in the microcline. The sodium site in the The peaksdo not all have the sameheight, as they do in microcline is tentativelythought to be in an envlronment low-albite. We have attributed the extra height in the -93 more symmetric - than that in albite. and lfi) ppm range to deshieldingof the albite-like sites Cryptoperthite. The NMR spectraof the cryptoperthite and increasedshielding of the anorthite-like sites caused (Spencer R) are consistentwith partiat AVSi ordering and by coherencystrain of the silicon sites in e-plagioclase separationinto potassium and sodium-richvolumes. The and, perhaps, to Q4(3ADsites at the margins. In the t22 KIRKPATRICK ET AL.: NUCLEAR MAGNETIC RESONANCE OF FELDSPARS bytownite (Anro) the most shielded silicon peak is not Facility. E. O. is a USPHS Research Career Development presentor is very small,perhaps suggesting that all ofthe Awardee, 1979-1984. albite-like volumes have strained silicon sites. The natu- ral anorthite (Anqz)has no peaks attributable to albite-like References sites. Abragam, A. (1961) The Principles of Nuclear Magnetism' Peaks or shoulders in the silicon-29 NMR spectra Clarendon Press. Oxford. which are attributable to anorthite-like sites increase in Andrew, E. R. (1971)The narrowing of NMR spectraof solidsby intensity with increasing An-content. For the oligoclase, high-speedspecimen rotation and the resolution of chemical there is only one shoulder at a relatively shielded chemi- shift and spin multiplet structuresfor solids' Progressin NMR cal shift (-88 ppm), which we interpret as arising from Spectroscopy,E, 1-39. Bollmann, W. and Nissen, H. -H' (196E)A study of optimal strained Qo(4Al) anorthite-like sites. The peaks assign- phaseboundaries: the case of exsolved alkali feldspars. Acta able to sites are larger in labradorite. We have Qo(4Al) Crystallographica, A24, 546-557 . peaks sites in e-plagio- attributed these to anorthite-like Brown, B. E. and Bailey, S. W. (1964)The structureof maxi- clases.The peaksattributable to anorthite-like sites in the mum microcline. Acta Crystallographica, 17, 1391-1400' bytownite are very large. We attribute these peaks pri- Brun, E., Hartmann,P., Staub,H. H., Hafner,S., andLaves, F' marily to sites in /-anorthite, although some of the signal (1960)Magnetische kernresonanzz:ur beobachtung des Al' Si- must come from sites in the e-plagioclase.As noted ordnungs/unordnungsgradesin einigen feldspiiten. Zeitschrift above, the silicon-29NMR spectrum of the natural anor- fiir Kristallographie, I 13, 65-76. thite is similar to the synthetic anorthite, but less well Farrar, T. C. and Becker, E. D. (1971)Pulse and Fourier Aca- resolved. Transform NMR: introduction to theory and methods. The aluminum-27 NMR spectra of the intermediate demic Press,New York. Fyfe, C. A., Gobbi, G. C., Klinowski, J., Thomas,J. M., and plagioclasesconsist of broad peaksthat becomemore like Ramdas, S. (1982)Resolving crystallographicallydistinct tet- An content. that of anorthite with increasing rahedral sites in silicalite and ZSM-5 by solid-state NMR. peaks The sodium-23NMR spectrashow single,unsplit Nature, 296,530-516. of about the same breadth as in sanidine.Because the Ganapathy,S., Schramm,S., and Oldfield' E. (1982)Variable- silicon-29NMR peaksindicate significant Al/Si order, we angle samplespinning high resolution NMR of solids. Journal tentatively attribute the non-split sodium-23resonances of Chemical Physics, 77, 43ffi-4365. to variable amounts of strain of the sodium sites, which Ghose,S. and Tsang,T. (1973)Structural dependenceof quadru- 27Al param- increases the range of electronic environments around pole coupling constant e2qQ/hfor and crystal field Mineralogist' sodium. eter D for Fe*3 in aluminosilicates. American 58, 748-755. Glasses Goldsmith,J. R. and Laves, F. (1955)Cation-order in anorthite (CaAlzSizOs)as revealed by gallium and germanium substitu- glasses The NMR spectra of the albite and anorthite tions. Zeitschrift fiir Kristallographie, 106, 213-226. consist of single, essentially structurelesspeaks that are Hafner, S. and Hartmann, P. (1964)Elektrische Feldgradienten broaderthan the spectra ofcrystalline albite or anorthite. und Sauerstof-Polarisierbarkeitin Alkali-Feldspiiten (NaAl- The silicon-29 resonancesare broader by about 20 ppm' SirOs)und KAlSi3Os).Helvetica Physica Acta, 37, 348-360' The aluminum-27 resonancefor albite glass is about 20 Hafner, S., Hartmann, P., and Laves, F. (1962)Magnetische ppm wider than that of low albite, whereasthe aluminum- Kernresonanz von Al27 in Adular. Zur Deutung der Adular- und Petrographische 27 resonanceof anorthite glass is only a few ppm wider struktur. SchweizerischeMineralogische -294. than that of synthetic crystalline anorthite. This would Mitteilungen, 42, 277 Hafner, S. and Laves, F. (1963)Magnetische Kernresonanz von seem to indicate a much larger increase in the range of AI27 in einigen Orthoklasen. SchweizerischeMineralogische glass.The sodium-23resonance aluminumsites in albite und PetrographischeMitteilun gen, 43, 65-69- of albite glass is about 20 ppm broader than that of low Harlow, G. E. and Brown, G. E. (1980)Low albite:an X-ray and albite, again indicating a somewhat larger range of elec- neutron diffraction study. American Mineralogist' 65' 986- tric field gradientsat the sodium sites in the glassyphase. 99s. 2esi,27A1, 23Na Higgins,J. 8., and Woessner,D. E. (19E2) and Acknowledgments spectra of framework silicates. (abstr.) Transactions of the We wish to thankDrs. GlennBuckley, David Stewart, and AmericanGeophysical Union (EOS)' 63,1139. (Title JohnHiggins, who suppliedus with someof the samplesand Kalus, C. K. (1978)Ph.D. Thesis,Universitv of Munich. compositions,W. -H. Yang and RichardOestrike for useful unavailable.) discussions,and an anonymous reviewer for manyuseful sugges- Klinowski, J., Ramdas,S., Thomas, J. M., Fyfe' C. A. and tions. This work was supportedin part by the Solid-State Hartman, J. S. (1982)A re-examinationof Si, Al ordering in Chemistryand Experimental and Theoretical Geochemistry Pro- zeolites NaX and NaY. Journal of the Chemical Society, gramsof the U.S. NationalScience Foundation (Grant DMR FaradayTransactions II, 78, 1025-1050. 8311339and EAR 8207260and 8408421) and in partby theU.S' Kundla, E., Samoson, A', and Lippmaa, E.' (1981) High- NationalInstitutes of Health,(Grant HL-19481) and has benefit- resolution NMR of quadrupolar nuclei in rotating solids. ted from facilities provided by the University of lllinois- ChemicalPhysics Letters, 83, 229-232. NationalScience Foundation Regional NMR Instrumentation Lippmaa, E., Magi, M., Samoson,A., Engelhardt,G., and KIRKPATRICK ET AL.: NUCLEAR MAGNETIC RES2NANCE oF FELDSPARS 123

Grimmer, -R. (1980) A. Structural studiesof silicatesbv solid Schramm, S. E. and Oldfield, E. (19E2)High resolution solid state high-resolution 2eSi NMR. Journal of the American state NMR of quadrupolarnuclei: quadrupbleinduced shifts in ChemicalSociety, 102,48E9-4893. variable-anglesample spinning of a borosilicate glass.Journal Lippmaa,E., Magi, M., Samoson,A., Tarmak, M., and Engel- of the Chemical Society, Chemical Communications,980-81. hardt, (1981) G. Investigation of the structure of zeolites bv Smith, J. V. (1974)Feldspar Minerals, volume 2esi l. Springer- solid-statehigh-resolution NMR spectroscopy.Journal of Verlag, New York. the American Chemical Society, l03,49924996. Smith, J. V. and Blackwell C. S. (1983) Nuclear magnetic Loewenstein, W. (1954) The distribution of aluminum in the resonanceof silica polymorphs. Nature, 223-225. tetrahedra of silicatesand aluminates.American Mineralogist, Smith,J. V., Blackwell,C. S., and Hovis, G. L. (1984)Nuclear 39,92-96. magnetic resonanceof albite-microcline series. Nature. 309. Mclaren, A. C. (1974)Transmission electron microscopy of the r$-t42. feldspars. In W. S. MacKenzie and J. Zussman,The Feld_ Smith, I. V. and Muir, L D. (1958)The reaction sequencein spars, p. press. 378-423. Manchester University larvikite feldspars.Zeitschrift f0r Kristallographie, I 10, I l-20. Meadows, M. D., Smith,K. A., Kinsey,R. A., Rothgeb,T. M., Smith, K. A., Kirkpatrick, R. J., Oldfield,E., and Henderson, Skarjune, R. P., and Oldfield,E. (19S2)High-resolution solid_ D. M. (1983)High-resolution silicon-29nuclear magneticreso- state proceedings NMR of quadrupolarnuclei. of the National nancespectroscopic study of rock forming silicates.American Academyof Scienceof the United States,79, 135l_1355. Minerafogist, 68, l20Ll2l5. Megaw, H. D., Kempster, C. J. E., and Radoslovich,E. W. Spencer,E. (1937)The potash-sodafeldspars L Thermal stabil- (1962) The structure of anorthite, CaAl2Si2O8.II. Description ity. Mineralogical Magazine 24, 453494. and discussion. Acta Crystallographica15, 1017-1035. Staehli, J. L. and Brinkmann, D. (1974a)A nuclear magnetic Muir, I. D. and J. V. Smith (1956)Crystallisation of feldsparsin resonance study of the phase transition in anorthite, CaAl2 larvikites. Zeitschrift fiir Kristallographie, 107, l g2_195. Si2Os.Zeitschrift fiir Kristallographie, 140, 3601373. Mtiller, D., Gessner,W., Behrens,H. J., and SchelerG. 09gl) Staehli, J. L. and Brinkmann, D. (1974b) Assignmentand Determination of the aluminum coordination in aluminum- structural dependenceof electric field gradients in anorthite oxygen 27Al compounds by solid-state high-resolution NMR. and simplefield gradient calculationsin somealuminosilicates. j9, Chemical Physics Letters, 59-62. Zeitschrift ftir Kristallographie, 140, 374-392. Mysen,B. O., Virgo, D., and Seifert,F. A. (19g2)The structure Stewart, D. B. and Wright, T. L. Onq AUSi order and of silicate melts: implications of chemicaland physical proper_ symmetry of natural alkali feldspars and the relationship of ties of natural magma. Reviews of Geophysicsand the Space strained cell parameters to bulk composition. Bulletin de la Physics, 20, 353-383. Socidte Frangaisede Mineralogie et de Cristallographie, gT, Oftedahl, C. (1948)Studies on the igneous rock complex of the 356-377. Oslo region. IX. The feldspars. Skrifter utgit av Det Norske Taylor, M., and Brown, G. E. (1979) Srructure of mineral Videnskaps-Akademi i Oslo I. Mathematisk-Naturidenskapel_ glasses.I. the feldsparglasses NaAlSijO8, KAlSi3Os,CaAl2 lig Klasse.1948. No. 3. NaturvidenskKl. 3. Si2Os.Geochimica et CosmochimicaActa, 43,61J7. Oldfield, E., Kinsey, R. A., Smith, K. A., Nichols, J. A., and Wainwright,J. E., and Starkey,J. (1971)A refinementof the Kirkpatrick, (1983) R. J. High-resolution NMR of inorganic structure of anorthite. Zeitschrift fiir Kristallographie, 133, solids. Influence of magnetic centers on magic-anglesample- 75-E4. spinninglineshapes in some natural alumino silicates. Journal of MagneticResonance, Sl, 325-327. Ribbe, P. H., (1983)Feldspar Mineralogy, Vol. 2. Reviewsin Mineralogy, 2nd Edition. Mineralogical Society of America, Manuscript received, October 26, 1983; Washington, D.C. acceptedfor publication, September4, 1984.