Octahedral Site Fe2* Quadrupole Splitting Distributions From
Total Page:16
File Type:pdf, Size:1020Kb
American Mineralogist, Volume 83, pages 1316-1322, 1998 Octahedral site Fe2*quadrupole splitting distributions from the Miissbauerspectra of arrojadite Isa.rru SHINNoI'* AND ZHE Lr2 'GraduateSchool of Social and Cultural Study, Kyushu University, Fukuoka 810, Japan ,Institute of Geology, ChineseAcademy of Science,Beijing 100029,China ABSTRACT The Mijssbauerspectra of arrojadite,(K,Ba)(Na,Ca),(Fe,*,Mn,Mg),oAl(pO*),, (OH,F), at 298 and'95 K were investigatedfor the first time. The spectraat both temperatureswere analyzedin terms of their Fe'?*quadrupole splitting distributions (QSDs). The overall QSDs at both temperaturescan be interpreted in terms of five octahedral site Fer* QSD contri- butions. The quadraticelongation, (r), and the variation ofbond angles,o,, for the different sites were calculated on the basis of the structural data obtained by Moore et al. (1981). The five QSD contributions are tentatively assignedto Fe2* in the M3, M4, M5, M6, and M7 sites, basedon the structural determination and the relation of the quadrupolesplitting to the distortion of the octahedra,respectively. The Fe,* ions are randomly distributed over the M3, M4, M5, M6, and M7 sites. In addition, Mcissbauerdata from arrojadite and related phosphateminerals indicate that the mean value of the isomer shift of Fer* in the octahedralsites in phosphateminerals is -0.07 mm/s larger than that in silicate minerals. This difference is explained in terms of electron affinity. INrnooucuoN transition metal phosphategroup. Several mineralogical To date, many Mcissbauerspectra of silicate and oxide studieson arrojadite-dickinsonitewere carried out (Head- minerals have been published, but relatively few Mciss- den 1891;Ziegler 1914;Quensel1937; Mason 1941;Gui- bauer studies concern phosphate minerals. Gonser and maraes t942;Lindberg 1950). Moore and Ito (1979) an- Grant (1967) first examined single and polycrystalline alyzed 12 samplesof the family and proposedthe general formula : samplesof naturally occurring vivianite, Fe.(PO.)r.SH,O ; XY.M?;AI(PO"),,(OH,F), where X large cat- : guz+) : subsequently,the oxidation mechanism of vivianite has ions suchas Kr*, Ba2*,Y (Na'", and l\r[ (psz+, been investigated many times (Takashima and Ohashi Mn2* and Mg'*). Moore et al. (1981) determinedthe 1968; Tricker and Ash 1979; Vochten et al. 1979; Dor- crystal structure for the seriesusing single-crystal X-ray mann and Poullen 1980; McCammon and Burns 1980; diffraction. In this study, intrinsic Fe2* QSDs were ex- Burns 1981). Chandra and Hoy (1967) reported that the tracted from Mtjssbauer spectra of arrojadite at 298 and ordered phase of ludlamite, Fe.(PO*)r.4HrO,has a Mciss- 95 K, and the iron distributions over the different sites bauer spectrum with two hyperfine fields below 15 K. are determined. Lithium orthophosphate,LiFeorMnorPOo, orders antifer- romagnetically at low temperature(Schideler and Terry ExpnnrunNTAL METHoDS 1969). Kostiner (1972) reported the Mtjssbauer parame- The sample occurs as large cleavable massesin a gra- ters for the phosphateminerals triplite, zwieselite, trip- nitic pegmatite from India associated with feldspar, loidite, and wolfeite. The cation distributions of the ter- quartz, muscovite, beryl, and spodumene.The purity of nary orthophosphates,(Zn, Fe, Me).(PO*),, (Me : Ni, the anojadite was checkedusing X-ray diffraction and no Mg, Co), and (Co, Fe).(PO"),, having the farringtonite other phases were found. Its chemical composition was structure were investigatedby Nord and Ericsson (1985) analyzedusing a CamecaSX5l electron microprobe. The and Nord et al. (1985). The cation partitioning in hydro- analyseswere carried out on six arrojadite crystals with thermally prepared sarcopsites(Fe,Mn,Co,Mg),(PO.), colors ranging from olive to grass green. The average was analyzedby means of X-ray powder diffraction and compositionis Na,O 6.83 (6.13-7.85),K,O 1.84(1.70- 57Fe Mcissbauerspectroscopy (Ericsson et al. 1986; Char- 1.88),CaO 2.59 (2.43-2.68),MnO 14.16(14.16-15.50), alampides et al. 1988). All of the above studies provide FeO 28.33(27.60-28.79), MgO 1.18(1.01-1.24), Al,O3 useful information on the crystal chemistry and bonding 2.45 (2.35-2.51),P,O, 41.71(40.32-43.10), total99.69 in phosphateminerals. wt%o(97 .05-102.28), where the bracketsdenote the rang- The anojadite-dickinsonite family constitutesan alkali es. The chemical formula can be written as I! r, (Nao uoCaonr)r r, (MDo.oFer rrMgo u,),. rsAl, o0 (Pr2 25oos) * E-mail:Shinno@rc kyushu-u ac jp (OH,F), based on 48 O atoms in the formula unit. 0003-004x/98/1I 12-13 16$05.00 1316 SHINNO AND LI: MOSSBAUER OF ARROJADITE t3t7 r.02 r.20 1.00 a u) 0.80 H 0.98 F tr k (!) 1.00 a 0.40 a q F 0.95 0.00 0.90 0.00 1.00 2.00 3.00 4.00 -4.00 -2.00 0.00 2.00 4.00 Velociry(mm/s) A(mmis) Ftcunn 1. Mrissbauerspectra of arrojaditeat 298 K (a) and FIcuru 2. The octahedral-siteFe2* QSDs of anojaditeat 298 K. 95 K (b). Dots arethe data Solid line : the fit for one sener- alizedsite with fiveQSD Gaussian components. trum-specific parameters (BG, background, f, half The Mrissbauer spectrum of arrojadite was measured width), two site-specific parameters (60, 6,), and three at 298 K using a computer-controlledconstant accelera- component-specificparameters (h, the height of one line tion Mtjssbauer spectrometer (PH-805) with 512 chan- in the symmetric elemental Lorentzian doublet; A, the nels, whereasthe spectrum at 95 K was obtained with a center of the GaussianQSD component; o^, the width of constant acceleration spectrometer(Austin Science As- the GaussianQSD component). sociates)in conjunction with a 1024 multichannel ana- The Voigt-based QSD method is used to describe the lyzer (ORTEC MCA 7700) and an OXFORD 4l0j4 in- intrinsic Mijssbauer line shape (the thin-limit spectrum). strument cryostat with a temperaturerange of 77-300 K In this case, the only correct Lorentzian line width, f, is and a variation of 0.1 K. An approximately5 mCurie 57Co the natural one, i.e., f :0.194 mm/s. Becausefinite ab- source in a palladium matrix was used in the measure- sorber thickness can cause spectral broadening, the raw ments. The detector used is a xenon (methane) propor- spectra should be corrected for thickness before the tional counter. The velocity scaleswere normalized with Voigt-basedQSD method is used.Ping et al. (1991) in- respectto the center of the spectrumof metallic iron foil vestigatedquasi-crystals by using the Voigt-basedmeth- at 298 K. od, with the absorberthicknesses in the range of 0.005- 57Fe/cm2. A Voigt-based method for arbitrary shape eSDs and 0.087 mg In their study, the full thickness cor- hyperfine field distributions (F{FDs) has been developed rection was not performed and the I obtained is in the (Rancourt and Ping l99la, l99lb; Ping et al. 1991; Ran- range of 0.214-0.219 mm/s. In this study, the absorber cowt 1994a, I994b; Rancourt et al. 1994 Rancourt et al. thickness is small, 0.006 mg slFelcm2,and hence the 1996). This method was used to fit the raw spectrain this Voigt-based method was used directly. The resulting I study. The method assumesa certain number m of gen- values equal 0.219 mrn/s at 298 K and 0.289 mm/s at 95 eralized sites each having their own continuous eSD. K, respectively. Each normalized site-specificQSD is composedof a cer- tain number (n, for site i ) of Gaussiancomponents being Rnsur,rs the sum of more than one Voigt line. The corresponding The Mcissbauerspectra of arrojadite at 298 and 95 K - - - fit can be expressedas nt n2 . n^V. The center each consists of two broad peaks, and the spectrum at shift 6 of each site's distribution component is related to 298 K shows very small shoulders,indicating ferrous ions its quadrupole splitting, A, as 6 : Do+ 6,4, where Enis in several crystallographic sites in the crystal structure the value of 6 when the distributed hyperfine parameter (Fig. 1). The octahedral-siteFe," QSDs obtained from has a value of zero, and D, is the coupling of 6 to the fitting the spectraare shown in Figures 2 and 3. The cal- distributed hyperfine parameter.Therefore, at most 2 + culated Mcissbauerparameters for the QSDs are sum- 2m + 3 X3, n, fitting parametersare required: two spec- marized in Table l. 1318 SHINNO AND LI: MOSSBAUER OF ARROJADITE 1.60 This model has been successfullyused to interpret most Mcissbauerspectra of minerals. Howeve! when tetrahe- dral and octahedrally coordinated sites accommodatea large variety of cations, minerals often display a wide t.20 range of local environments.This leads to a continuous distribution of quadrupole splitting. In this case,the QSD method should be used to analyzeMtissbauer spectra. Ar- rojadite falls in the later case consisting of two broad b 0.80 peaks. Probably, several QSDs may be contained in the Mcissbauerspectra of arrojadite. To date, a synthetic annite-oxyannite series and syn- 0.40 thetic and natural Al-deficient membersof the phlogopite- annite serieshave been analyzedby QSDs method (Ran- court I994a, 1994b: Rancourt et al. 1994; Rancourt et al. 1996). There are cis and trans octahedraoccupied by cat- 0.00 ions in mica structure.According to Fe2* QSDs analysis 0.00 1.00 2.00 3.00 4.00 of synthetic annites,the overall QSDs can be interpreted - in terms of four QSD contributions centeredat L' 2.55 A(mm/s) mm/s for Fe'z*O4(OH),octahedra (cis and trans not re- - Frcunn 3. The octahedral-siteFe2* QSDs of arrojadite at 95 K. solved), L 2.35 mm/s for Fe'?*Oo(OH)Foctahedra (cis and trans not resolved), L - 2.15 mm/s for cis-Fe'*OoF, octahedra,and A - 1.5 mm/s for trans-Fe2*OoF,octahe- Drscussron dra (Rancourt 1996).