American Mineralogist, Volume 88, pages 955961, 2003

NaBLi solid-solution in A-site-vacant amphiboles: synthesis and cation ordering along the ferri-clinoferroholmquistiteriebeckite join

GIANLUCA IEZZI,1, GIANCARLO DELLA VENTURA,2,* FERNANDO CMARA,3 GIUSEPPE PEDRAZZI,4 AND JEAN-LOUIS ROBERT1
1 ISTO, UMR 6113, 1A, Rue de la Frollerie, F-45071 Orlans Cedex 2, France Dipartimento di Scienze Geologiche, Universit di Roma Tre, Largo S. Leonardo Murialdo 1, I-00146, Italy 3 CNR-Istituto di Geoscienze e Georisorse, sezione di Pavia, via Ferrata 1, I-27100 Pavia, Italy 4 Dipartimento di Sanit Pubblica, Sezione di Fisica, INFM Universit di Parma, Via Volturno, 39, I-43100 Parma, Italy 2

ABSTRACT
Amphiboles were hydrothermally synthesized at 500 C and 4 kbar in the system Li2O-Na2O2+ FeO-Fe 2 O 3 -SiO 2 -H 2 O, with nominal compositions along the riebeckite [ Na 2 Fe 3 3+ 2+ 3+ Fe2 Si8O22(OH)2]-ferri-clinoferroholmquistite [ Li2Fe3 Fe2 Si8O22(OH)2] join, where the exchange vector is NaLi1 at the B-site. Experimental products were characterized by powder XRD and SEM-EDAX, confirming very high amphibole yield along the join (>95%, plus minor quartz). The XRD patterns can be indexed in C2/m, and the refined cell parameters show linear variation as a function of composition. For the BLi end-member, the IR spectrum shows a single sharp main band centered at 3614 cm1, which is assigned to the FeFeFe-OH-A configuration. With increasing B Na in the mineral, this band broadens and shifts 4 cm1 to higher frequencies. This effect can be attributed to the change in M4 site occupancy. Minor ANa (partial solid-solution toward arfvedsonite) is also observed with increasing sodium in the system. Mssbauer spectroscopy confirms the cation distribution provided by IR data, and shows that a small, but significant amount of Fe2+ occurs at M4 along the join. Infrared spectroscopy shows that the Li end-member has a very ordered structure, whereas intermediate compositions show local heterogeneities associated with the presence of two different B sites occupied by Na or Li.

INTRODUCTION
Li-bearing amphibole minerals belong to two main groups: Fe-Mg-Mn-Li (holmquistite) and sodic amphiboles. Holmquistite has been described from some thirty localities around the world (London 1986), whereas Li-rich sodic amphiboles have been described from igneous rocks (Questa Caldera, Hawthorne et al. 1996a; Hurricane Mountain, Hawthorne et al. 1996b; Strange Lake, Hawthorne et al. 2001), and from metamorphic rocks (Kajlidongri, India, Hawthorne et al. 1992). Both amphibole types are typical of highly fractionated syenites and alkali syenites, and granitic pegmatites (Deer et al. 1999; London 1986). Before the 1990s, the distribution and the crystal-chemical behavior of Li in amphiboles was poorly characterized, primarily because of the difficulties in analyzing Li at the micrometer scale. These problems have been resolved with the application of micro-analytical techniques and detailed crystal-structure refinements of suites of samples from petrologically well-characterized environments. In holmquistite, Li2(Mg,Fe2+)3(Al,Fe3+)2Si8O22(OH)2, Li is a B-group cation

* E-mail: [email protected] Presently at Dipartimento di Scienze della Terra, Universit G. DAnnunzio, I-66013 Chieti Scalo, Italy E-mail: [email protected]
0003-004X/03/0007955$05.00 955

and is completely ordered at M4. In sodic amphiboles such as leakeite, NaNa2[(Mg,Fe2+)2(Al,Fe3+)2Li]Si8O22(OH)2, Li is a Cgroup cation and is strongly ordered at M3 (Hawthorne et al. 1993, 1994). Sodic ferripedrizite, Na(Na,Li) 2[(Mg,Fe 2+) 2 (Al,Fe 3+) 2 Li]Si8O22(OH)2 (Oberti et al. 2000; Caballero et al. 2002), has Li at both the M4 and M3 sites, thus bridging the Mg-FeMn-Li and the Na amphibole groups. The available data show that Li can enter the amphibole structure through two main exchange reactions: (1) 2 [M2,3](Fe2+, Mg) = [M3]Li + [M2](Fe3+, Al), connecting arfvedsonite (eckermannite) with leakeite (Hawthorne et al. 1994); and (2) [M4]Na = [M4]Li, connecting riebeckite with clinoholmquistite and leakeite with pedrizite (Oberti et al. 2000). The apparent exchange vector A Na + [M3]Li = A + [M3]Mg is the sum of different mechanisms and has no crystal-chemical significance, as it does not allow for local charge balance (Caballero et al. 2002; Oberti et al. 2003). Holmquistite and riebeckite are classified in two different groups (Leake et al. 1997) and it has been assumed that there is a miscibility gap between these two amphiboles (Deer et al. 1999). However, recent investigation of amphiboles from the Pedriza Massif, Sierra de Guadarrama, Spain (Caballero et al. 1998; Oberti et al. 2000; Caballero et al. 2001) showed complete miscibility between Na and Li at the B-group sites (Caballero et al. 2002; Oberti et al. 2003). There are no experimental studies concerning the relation

956

IEZZI ET AL.: BNa BLi SOLID-SOLUTION IN AMPHIBOLES

between Na-rich and Li-rich amphiboles. End-member riebeckite [ Na2Fe2+Fe3+Si8O22(OH)2] was synthesized by Ernst 3 2 (1962) who also studied its stability as a function of T, P, and fO2. Here, we present a crystal-chemical characterization of synthetic riebeckite, and show that complete exchange between Na and Li at the B-site in the amphibole structure can occur at relatively low T and P.

EXPERIMENTAL METHODS
Starting materials were prepared as silicate gels according to the method of Hamilton and Henderson (1968). Intermediate compositions were prepared by mixing end-member gels in stoichiometric proportions at intervals of 20 mol%. The starting materials were placed in gold capsules, together with 15 wt% bidistilled water, and the capsules were then arc-welded. Hydrothermal syntheses were done using internally heated pressure vessels at T = 500 C and PH2O = 4 kbars for a duration of 14 days. The fO2 is considered as internally buffered for all experiments, as the permeability of gold is very low at this temperature (Chou 1987; Iezzi et al. 2003). The temperature was monitored by three sheathed typeK thermocouples closely adjoining the sample position (accuracy 5 C). Pressure was continuously recorded by a strain gauge, calibrated against a Heise-Bourdon tube gauge (error = 20 MPa, Scaillet and Evans 1999). The experiments were quenched by switching off the current supply, thus causing a temperature drop of about 40 C/min down to ambient temperature. Step-scan X-ray powder-patterns were collected on a Scintag X1 diffractometer, operating in the vertical q-q configuration, with Ni-filtered CuKa radiation and a Si(Li) solid-state detector. Cell dimensions were refined by wholepowder-pattern refinement (Rietveld method) using DBW3.2 (Wiles and Young 1981). Samples were prepared for FTIR spectroscopy as KBr pellets using the procedure of Robert et al. (1989a). FTIR spectra in the OH-stretching region (40003000 cm1) were collected at room temperature with a Nicolet 760 spectrophotometer equipped with a DTGS detector and a KBr beamsplitter. Nominal resolution was 2 cm1, and the spectra are the average of 64 scans. The relative changes in line width of the absorption profiles were quantified by the autocorrelation function, following the method described in Salje et al. (2000). Mssbauer spectra were recorded in transmission mode at 297 K, with a conventional constant-acceleration Mssbauer spectrometer. The source was 57Co/ Rh with an activity of about 370 MBq (10 mCi). Calibration was done using either an a-Fe foil or an SNP absorber. Isomer shifts are expressed relative to aFe; doublets were fitted with unconstrained parameters. Additional details of sample preparation and experimental conditions can be found in Iezzi et al. (2003).

FIGURE 1. SEM image of synthetic end-member riebeckite (sample 237). The scale bar is 2 mm. The amphibole crystals are acicular and between 0.10.2 mm across and 0.5 to 2 mm long.

TABLE 1. Refined cell parameters along the ferro-ferriclinoholmquistiteriebeckite join

Sample Li/(Li + Na) a () 231 100 9.512(4) 232 80 9.555(3) 233 60 9.612(1) 234 50 9.645(1) 235 40 9.678(1) 236 20 9.726(1) 237 0 9.773(1)

b () 18.064(9) 18.089(7) 18.090(2) 18.090(2) 18.076(2) 18.074(2) 18.078(2)

c () 5.323(2) 5.325(1) 5.3266(7) 5.3267(6) 5.3273(7) 5.3286(6) 5.3319(7)

b () 101.88(5) 101.80(3) 102.28(1) 102.50(1) 102.76(1) 103.14(1) 103.40(1)

V (3) 895.22 901.77 904.83 906.85 908.41 911.89 916.51

EXPERIMENTAL RESULTS
The synthesis products were characterized by powder XRD and by optical and electron microscopy. The powders consisted primarily of amphibole, with minor quartz; for all experiments the amphibole yield is considered to be >95%. The amphibole crystals (Fig. 1) are acicular, with average dimensions of 1 to 5 mm length and 0.21 mm width. The small size of the amphibole crystals precluded EMPA analysis. Cell parameters All XRD patterns were indexed in C2/m. The a edge, the b angle, and V decrease linearly as Li increases (Na decreases) at the B-site (Table 1, Fig. 2), whereas the b and c dimensions remain virtually constant. The variation in b agrees very well with the data of Oberti and Ghose (1993) and Robert et al. (1993) on the relation between the aggregate cation radius at the B-site and b angle. In particular, the data presented here confirm the different trends for monovalent and divalent substitutions at M4 predicted by Oberti and Ghose (1993, see Fig. 3). Figure 3 shows that cation charge is the principal factor in controlling the slope of the trend. However, other factors also play a role; for example, the nature of the A-cation significantly affects the position of the regression line, as indicated by the parallel trends observed for BSr substituted richterites (Fig. 3). Mssbauer spectroscopy Mssbauer spectra were recorded from the Li and Na endmember experimental products and from the intermediate B (LiNa) composition. The Mssbauer parameters are listed in Table 2, and the resolved spectra are stacked in Figure 4. The spectrum of end-member ferri-clinoferroholmquistite was described by Iezzi et al. (2003). It shows three well-defined doublets, two assigned to Fe2+ and one assigned to Fe3+. On the basis of known literature data on amphiboles (Hawthorne 1983a, 1983b; Redhammer and Roth 2002), the Fe2+ doublet with the largest quadrupole splitting (2.82 mm/s, Table 2) was assigned to Fe2+ at M1, and the doublet with QS = 2.25 mm/s was assigned to Fe2+ at M3. The doublet with the smaller isomer shift and quadrupole splitting was assigned to Fe3+ at M2 (Table 2 and Fig. 4). The Fe2+/Fe3+ ratio, calculated from the band areas for this sample, is 1.5, in agreement with the nominal compo-

IEZZI ET AL.: BNa BLi SOLID-SOLUTION IN AMPHIBOLES

957

FIGURE 3. Relation between b and the mean ionic radius at M4 for some synthetic amphiboles. Circles = this work; squares = BSr substituted potassium-richterite, diamonds = BSr substituted sodiumrichterite; data from Robert et al. (1993).

FIGURE 2. Variation in cell parameters and volume for synthetic amphiboles along the riebeckite-ferri-clinoferroholmquistite join. TABLE 2. Mssbauer parameters for analyzed amphiboles
Sample Li/(Li + Na) Site Isomer Quadrupole G I Relative c2 shift splitting FWHH site d (mm/s) D (mm/s) (mm/s) distribution (%) 1.13(1) 2.82(1) 0.29(1) 38 1.15 1.11(1) 2.25(1) 0.33(1) 21 0.39(1) 0.28(1) 0.32(1) 41 1.15(1) 2.82(1) 0.32(1) 38 1.91 1.08(1) 2.43(1) 0.45(1) 20 0.39(1) 0.34(1) 0.37(1) 38 1.10(1) 1.80(1) 0.37(1) 4 1.14(1) 2.85(1) 0.33(1) 41 1.07 1.10(1) 2.40(1) 0.30(1) 15 0.44(1) 0.39(1) 0.37(1) 40 1.80(1) 1.10(1) 0.32(1) 4

231 Li 100 M1 Fe2+ M3 Fe2+ M2 Fe3+ 234 Li 50 M1 Fe2+ M3 Fe2+ M2 Fe3+ M4 Fe2+ 237 Li 0 M1 Fe2+ M3 Fe2+ M2 Fe3+ M4 Fe2+

sition (Iezzi et al. 2003). In the Mssbauer spectra of Na-bearing amphiboles (samples 234 and 237), there is a fourth lessintense doublet (Table 2 and Fig. 4) that can be assigned to Fe2+ at M4 (Hawthorne 1983a; Redhammer and Roth 2002), suggesting a slight departure from nominal stoichiometry (this was also confirmed by IR spectroscopy, see below). For Nabearing samples, Fe2+/Fe3+ is reasonably close to the nominal value (1.5). The resolved doublets in end-member ferriclinoferroholmquistite are significantly sharper (except that of sample 237, assigned to M3) than those in the Na-rich compositions (Table 2 and Fig. 4). According to Hawthorne (1983a,

FIGURE 4. Mssbauer spectra of the two end-members and the 50:50 composition along the ferri-clinoferroholmquistite-riebeckite join.

958

IEZZI ET AL.: BNa BLi SOLID-SOLUTION IN AMPHIBOLES

(marked by an arrow in Fig. 5). The IR spectra in the OHstretching region also show other minor features that represent less than 5% of the total absorbance. The most prominent feature is a broad band centered at 3672 cm1 in end-member riebeckite. This band appears with introduction of Na into the system; its intensity increases slightly with increasing Na, up to XNa (where XNa = Na/Na + Li) = 0.40, and then remains roughly constant up to XNa = 1 (Fig. 5). In addition, the position of this band shifts toward higher frequency with increasing Na. A second minor feature occurs at 3646 cm1, but is not present in end-member riebeckite (Fig. 5). A third minor feature occurs at 3636 cm1; it is barely discernable in the spectrum of end-member Na-free ferri-clinoferroholmquistite, and increases slightly in intensity with increasing Na. Assignment of the FTIR bands Assignment of the several components in the spectra of Figure 5 is based on previous work on amphiboles (e.g., Della Ventura 1992; Della Ventura et al. 1997, 1999). The basic assumption is that any charge-balance perturbation at both the NN (nearest-neighbor) and NNN local environments around the O3 anion causes a wavenumber shift of the OH-stretching band with respect to a reference band. The reference band in this system is obviously the main peak at 36143618 cm1; this [M1] Fe 2+[M1] Fe 2+[M3] band is assigned to the Fe2+-OH- A -BLi configuration (Iezzi et al. 2003). Note that in all samples the intensity of this band is >95% of the total absorbance, indicating that Fe2+ is almost the only occupant of the M1 and M3 sites that bonds to OH at O3; this observation, coupled with the Mssbauer data for the same samples, indicates that the long-range composition of these amphiboles is very close to the nominal value. The occupancy of the A-site in amphiboles has a major effect on the OH-spectrum; this feature was firstly recognized by Rowbotham and Farmer (1973), and was later characterized by Robert et al. (1989a), Della Ventura (1992), and Hawthorne et al. (1997), among others. In an A-site occupied amphibole, the OH-band is ~ 60 cm1 higher than in an A-site vacant amphibole with an identical NN local environment around O3. The frequency of the broad band at ~3672 cm1 is 60 cm1 higher than that of the main peak at 36143618 cm1, and hence this band can be confidently assigned to the configuration [M1]Fe2+[M1]Fe2+[M3]Fe2+-OH-ANa-BLiNa; it indicates the presence of a small amount of Na at the A-site. This assignment is also consistent with the fact that this band appears in the IR spectrum as Na is added to the system. The first spectroscopic (IR) evidence of CLi in amphiboles was given by Addison and White (1968) who reported the presence of a band due to the MgMgLi-OH configuration in the spectrum of a natural Li-bearing riebeckite. This band was shifted to higher frequency with respect to the MgMgMgOH band; however, the exact frequency was not given. From the work of Robert et al. (1989b) on synthetic micas, we know that the presence of a monovalent cation (Li) at the octahedral sites causes a wavenumber shift of +30 cm1 of the OHband, due to the +5 aggregate cation charge (MgMgLi) at the NN octahedral environment around the OH group. The minor band at 3646 cm1 is shifted 30 cm1 toward higher frequency

FIGURE 5. Infrared spectra in the OH-stretching region for synthetic amphiboles along the ferri-clinoferroholmquistite-riebeckite join. For the unlabeled arrow see text.

1983b), the broadening of the Mssbauer doublets is consistent with increasing disorder at the NNN (next-nearest-neighbor) sites around Fe. FTIR OH-stretching spectra All IR spectra in the OH-stretching region are very similar, and are dominated by one main band (Fig. 5). In end-member ferri-clinoferroholmquistite (sample 231, Fig. 5), this band is very sharp (FWHM 4 cm1), roughly symmetric, and is centered at 3614 cm1. With increasing BNa, this band broadens and shifts linearly toward higher frequency. In end-member riebeckite (sample 237, Fig. 5), it is centered at 3618 cm1 and is asymmetric toward the high-frequency side, suggesting the presence of a second overlapping component at 3620 cm1

IEZZI ET AL.: BNa BLi SOLID-SOLUTION IN AMPHIBOLES

959

with respect to the main band at 3614 cm1; therefore, it can be assigned to the configuration [M1]Fe2+[M1]Fe2+[M3]Li-OH-A . Note that this band in not present in the spectrum of end-member riebeckite. Its presence in all spectra along the join suggests that a very small amount of Li enters the octahedral strip in all of the amphibole samples. Single-crystal structural work on amphiboles shows that C Li is strongly ordered at M3 (Hawthorne et al. 1994; Oberti et al. 2003). Finally, assignment of the minor band at 3633 cm1 is difficult. On the basis of the data discussed above, it must be assigned to a configuration with a vacant A-site. Given the chemical system, we are left with the configuration Fe2+Fe3+LiOH-A . This inference is supported by the work of Della Ventura et al. (1999) on synthetic pargasite-richterite solidsolutions, which shows that the presence of a trivalent cation (Al) at the NN OH-coordinated sites produces a negative shift of ~15 cm1; actually, the 3633 cm1 band is shifted 13 cm1 with respect to the Fe2+Fe2+Li-OH-A band at 3646 cm1. Note that this 3633 cm1 component is not present in the spectrum of end-member riebeckite (sample 237, Fig. 5), where it is replaced by a new weak band at ~3636 cm1 (indicated by an unlabeled arrow in Fig. 5) whose assignment is uncertain, but is probably related to Fe2+ at M2. Ordering of cations along the join Infrared spectroscopy is extremely sensitive to both LRO and SRO, and two different methods have been developed in the last few years to characterize cation order in mineral structures. The first is based on the measure of frequencies and relative intensities of OH-stretching bands in well-characterized solid-solutions (e.g., Hawthorne et al. 1997, 2000; Della Ventura et al. 1997, 1999). However, quantitative use of this method is greatly inhibited by the ignorance of possible variation in absorption coefficients with frequency (e.g., Skogby and Rossman 1991), as discussed by Hawthorne et al. (1997). The second approach is based on the fact that vibrational spectra can show substantial line broadening and shifting as cations are substituted in a solid-solution. The line broadening is due to the local heterogeneity of a structure, which relaxes differently around cations with different size or charge (Salje et al. 2000). The resulting variation in line width is analyzed by the autocorrelation method (e.g., Salje et al. 2000; Carpenter and Boffa Ballaran 2001) which is a fast and easy way to quantify line-width variations in powder absorption spectra from sequences of samples of varying composition. The variation in calculated line-width, expressed as Dcorr, scales with some local-order parameter, qod. The interesting thing about this method is that one can ignore the assignment of absorption bands to specific structural environments, and this is certainly a great advantage when dealing with complex spectra. For this reason, it has been used in lattice (<1200 cm1) vibrational regions of different silicate minerals (e.g., Boffa Ballaran et al. 1998, 1999, 2000; Atkinson et al. 1999, and references therein). The systematic presence of a single main band in the spectra (Fig. 5), shows that the amphiboles produced along the studied join are very close to the nominal compositions (see Iezzi et al. 2003). Minor departures can be ascribed to ANa (partial solid-solution toward arfvedsonite) and CLi. Infrared spec-

tra show that CLi is not coupled with ANa, as both 3646 and 3633 cm1 bands are locally connected with vacant A-sites (see preceding paragraph). This short-range-order information is in apparent contrast to the bond-valence calculations of Hawthorne et al. (1992) and to the single-crystal work of Oberti et al. (2003) which shows that [M3]Li is locally strongly associated with Na at the Am site. However, it must be emphasized that we are dealing here with very small amounts of CLi in the structure, and perhaps the [M3]Li ANa relation, which is of general validity in controlling the bulk charge-balance in CLi-bearing amphiboles (Oberti et al. 2003), can be locally violated. As previously stated, the main band in the spectrum of endmembers ferri-clinoferroholmquistite and riebeckite are very sharp (Fig. 6); this suggests that these amphiboles have a strongly ordered structure (Iezzi et al. 2003). For intermediate B NaBLi compositions, this band broadens considerably (Fig. 6). This behavior is evidently due to Na-Li disorder at M4, and can be treated by means of the autocorrelation method. Autocorrelation spectra were calculated for the segment of the OH-spectra of Figure 5 which contains only the main OHstretching band (36003625 cm1). The values of Dcorr (Fig. 7) at intermediate compositions have a positive deviation from ideality and reflect local heterogeneities in the Na-Li disordered samples. Recent single-crystal X-ray work shows that,

FIGURE 6. Enlargement in the 36303600 cm1 region for selected amphiboles along the ferri-clinoferroholmquistite-riebeckite join.

960

IEZZI ET AL.: BNa BLi SOLID-SOLUTION IN AMPHIBOLES

Caballero, J.-M., Monge, A., La Iglesia, A., and Tornos, F. (1998) Ferriclinoholmquistite, Li2(Fe2+,Mg)Fe3+Si8O22(OH)2, a new BLi clinoamphibole from 2 the Pedriza Massif, Sierra de Guadarrama, Spanish Central System. American Mineralogist, 83, 167171. Caballero, J.M., Oberti, R., Ottolini, L., and Cannillo, E. (2001) Complete solidsolution between amphiboles with Li and Na at the B-group sites. EUG XI, Strasbourg, Journal of Conference Abstracts, 6, 1, 541. Caballero, J.M., Oberti, R., and Ottolini, L. (2002) Ferripedrizite: a new monoclinic B Li amphibole end-member from the Eastern Pedriza Massif, Sierra de Guadarrama, Spain, and a restatement of the nomenclature of Mg-Fe-Mn-Li amphiboles. American Mineralogist, 87, 976982. Carpenter, M.A. and Boffa Ballaran, T. (2001) The influence of elastic strain heterogeneities in silicate solid solutions. EMU Notes in Mineralogy, 3, 155.178. Chou, I.M. (1987) Oxygen buffer and hydrogen sensor technique at elevated pressures and temperatures. In H.L. Barnes and G.C. Ulmer, Eds., Hydrothermal experimental techniques, p. 6199. Wiley, New York. Deer, W.A., Howie, R.A., and Zussman, J. (1999) Double chain-silicate, Rock Forming Minerals. Ed. The Geological Society, V.2B, second edition, 764 p. Della Ventura, G. (1992). Recent developments in the synthesis and characterization of amphiboles. Synthesis and crystal-chemistry of richterites. Trends in Mineralogy, 1, 153191. Della Ventura, G., Robert, J.-L., Raudsepp, M, Hawthorne, F.C., and Welch, M.D. (1997). Site occupancies in synthetic monoclinic amphiboles: Rietvield structure refinement and infrared spectroscopy of (nickel, magnesium, cobalt)richterite. American Mineralogist, 82, 291301. Della Ventura, G., Hawthorne, F.C., Robert, J.-L., Delbove, F., Welch, M.D., and Raudsepp, M. (1999) Short-range order of cations in synthetic amphiboles along the richterite-pargasite join. European Journal Mineralogy, 11, 7994. Ernst, W.G. (1962) Synthesis, stability relations, and occurrence of riebeckite and riebeckite-arfvedsonite solid-solutions. Journal of Geology, 70, 689736. Hamilton, D.L. and Henderson, C.M.B. (1968). The preparation of silicate compositions by a gelling method. Mineralogical Magazine, 36, 832838. Hawthorne, F.C. (1983a). The crystal-chemistry of amphibole. Canadian Mineralogist, 21, 174480. (1983b). Quantitative characterisation of site-occupancies in minerals. American Mineralogist, 68, 287306. Hawthorne, F.C., Oberti, R., Ungaretti, L., and Grice, J.D. (1992). Leakeite, NaNa2(Mg2Fe3+Li)Si8O22(OH)2 a new alkali amphibole from the Kajlidongri 2 manganese mine, Jhabua district, Madhya Pradesh, India. American Mineralogist, 77, 11121115. Hawthorne, F.C., Ungaretti, L., Oberti, R., and Bottazzi, P. (1993) Li: an important component in igneous alkali amphiboles. American Mineralogist, 78, 733745. Hawthorne, F.C., Ungaretti, L., Oberti, R., and Cannillo, E. (1994). The mechanism of [6]Li incorporation in amphiboles. American Mineralogist, 79, 443451. Hawthorne, F.C., Oberti, R., Ungaretti, L., Ottolini, L., Grice, J.D., and Czamanske, G.K. (1996a). Fluor-ferro-leakeite, NaNa2(Fe2+Fe3+Li)Si8O22F2, a new alkali 2 2 amphibole from the Canada Pinabete pluton, Questa, New Mexico, U.S.A. American Mineralogist, 81, 226228. Hawthorne, F.C., Oberti, R., and L., Ottolini (1996b). Lithium-bearing fluorarfedsonite from Hurricane mountain, New Hampshire: a crystal-chemistry study. Canadian Mineralogist, 34, 10151019. Hawthorne, F.C., Della Ventura, G., Robert, J.-L. Welch, M. D., Raudsepp, M., and Jenkins, D.M. (1997). A Rietvield and infrared study of synthetic amphiboles along the potassium-richterite-tremolite join. American Mineralogist, 82, 708716. Hawthorne, F.C., Welch, M. D., Della Ventura, G., Liu, S., Robert, J.-L., and Jenkins, D.M. (2000). Short-range order in synthetic aluminous tremolites: An infrared and triple-quantum MAS NMR study. American Mineralogist, 85, 17161724. Hawthorne, F.C., Oberti, R., Cannillo, E., and Ottolini, L. (2001). Li-bearing arfvedsonitic amphiboles from the Strange lake peralkaline granite, Quebec. Canadian Mineralogist, 39, 11611170. Iezzi, G., Della Ventura, G., Pedrazzi, G., Robert, J.-L., and Oberti, R. (2003) Synthesis and crystal-chemistry of ferri-clinoferroholmquistite, Li2Fe2+Fe3+Si8O22(OH)2. European Journal of Mineralogy, 15, 321327. 3 2 Leake, B.E., Woolley, A.R., Arps, C.E.S, Birch,W.D, Gilbert, M.C., Grice, J.D., Hawthorne, F.C., Kato, A., Kisch, H.J., Krivovichev, V.G., Linthout, K. Laird, J., Mandarino, J.A., Maresch, W.V., Nickel, E.H., Rock, N.M.S., Schumacher, J.C., Smith, D.C., Ungaretti, L., Whittacker, E.J.W., and Youzhi, G. (1997). Nomenclature of amphiboles: Report of the subcommittee on amphiboles of the International Mineralogical Association, commission on new minerals and minerals name. European Journal of Mineralogy, 9, 623651. London, D. (1986). Holmquistite as a guide to pegmatitic rare metal deposits. Economic Geology, 81, 704712. Oberti, R. and Ghose, S. (1993). Crystal-chemistry of a complex Mn-bearing alkali amphibole (tiroidite) on the verge of exsolution. European Journal of Mineralogy, 5, 11531160. Oberti, R., Caballero, J.-M., Ottolini, L., Andres, S.L., and Herreros, V. (2000) Sodicferripedrizite, a new monoclinic amphibole bridging the magnesium-iron-manganese-lithium and the sodium-calcium group. American Mineralogist, 85, 578585. Oberti, R., Cmara, F., Ottolini, L., and Caballero, J.-M. (2003) Lithium in amphib-

FIGURE 7. Variation in line width (Dcorr) measured for the 3600 3625 cm1 stretching region for synthetic amphiboles along the ferriclinoferroholmquistite-riebeckite join.

in monoclinic amphiboles, Li occurs at a site (M4') closer to the strip of octahedra than that occupied by Na and Ca (M4), and adopts sixfold coordination (Oberti et al. 2003 and references therein). Therefore, local heterogeneities in intermediate Na-Li samples can be associated with the presence of two different B sites with slightly different geometries and distortions. The consequent slightly different charge contribution to O3 gives rise to slightly different O-H vibrational frequencies, which combine to give a broader absorption. The range in charge contribution can be quantified by the variation in the O-H band widths which gives the measure of the microscopic spontaneous strain induced by the distribution of Na and Li (and minor Fe2+) at the B-site. As already observed by Carpenter and Boffa Ballaran (2001), Dcorr can be compared directly to the enthalpy of mixing. Therefore, given the frequency range considered here (3600 3625 cm1), the characteristic length scale of phonons (which is known to scale with 1/w, where w = frequency: Salje 1992) must be of the order of magnitude of a single B site, i.e., ~45 .

ACKNOWLEDGMENTS
Sincere thanks are due to B. Di Sabatino for constructive comments at the beginning of this work. S. Lomastro assisted with XRD data collection, and J. Roux and B. Scaillet assisted during the synthesis work. Thanks are due to R. Oberti for suggestions and discussions, which helped in clarifying some aspects of the present work. Part of this work was done during the stay of G.I. at I.S.T.O.C.N.R.S. (Orlans), which was financed by the University of Chieti and an EGIDE-French Foreign Affairs Ministry fellowship.

REFERENCES CITED
Addison, W.E. and White, A.D. (1968) Spectroscopic evidence for the siting of lithium ions in a riebeckite. Mineralogical Magazine, 36, 743745. Atkinson, A.J., Carpenter, M.A., and Salje, E.K.H. (1999). Hard mode infrared spectroscopy of plagioclase feldspars. European Journal Mineralogy, 11, 721. Boffa Ballaran, T., Carpenter, M.A., Domeneghetti, M.C., Salje, E.K.H., and Tazzoli, V. (1998) Structural mechanisms of solid solutions and cation ordering in augite-jadeite pyroxenes: II. A microscopic perspective. American Mineralogist, 83, 434443. Boffa Ballaran, T., Carpenter, M.A., Geiger, C.A., and Koizol, A.M. (1999) Local structural heterogeneity in garnet solid solutions. Physics and Chemistry of Minerals, 26, 554569. Boffa Ballaran, T., Angel, R.J., and Carpenter, M.A. (2000) High pressure transformation behaviour of the cummingtonite-grunerite solid solution. European Journal of Mineralogy, 12, 11951213.

IEZZI ET AL.: BNa BLi SOLID-SOLUTION IN AMPHIBOLES

oles: detection, quantification, incorporation and nomenclature in the compositional space bridging sodic and BLi-amphiboles. European Journal of Mineralogy, 15, 295307. Redhammer, G.J. and Roth, G. (2002). Crystal structure and Mssbauer spectroscopy of the synthetic amphibole potassic-ferri-ferrorichterite at 298 K and low temperatures (80-110 K). European Journal of Mineralogy, 14, 105114. Robert, J.-L., Della Ventura, G., and Thauvin, J.-L. (1989a). The infrared OH-stretching region of synthetic richterites in the system Na2O-K2O-CaO-MgO-SiO2H2O-HF. European Journal of Mineralogy, 1, 203211. Robert, J.-L., Bny, J.-M., Bny, C., and Volfinger, M. (1989b). Characterization of lepidolite by Raman and infrared spectrometries. I. Relationships between OHstretching wavenumbers and composition. Canadian Mineralogist, 27, 225235. Robert, J.-L., Della Ventura, G., Raudsepp, M., and Hawthorne, F.C. (1993). Rietvield structure refinement of synthetic strontium-rich potassium-richterites. European Journal of Mineralogy, 5, 199206. Rowbotharm, G. and Farmer, V.C. (1973). The effect of A site occupancy on the hydroxyl stretching frequency in clinoamphiboles. Contributions to Mineralogy and Petrology, 38, 147149. Salje, E.K.H. (1992) Hard mode spectroscopy: Experimental studies of structural

961

phase transitions. Phase transitions, 37, 83110. Salje, E.K.H., Carpenter, M.A., Malcherek, T., and Boffa Ballaran, T. (2000) Autocorrelation analysis of infrared spectra from minerals. European Journal of Mineralogy, 12, 503519. Scaillet, B. and Evans, B.W. (1999). The June 15, 1991 eruption of Mount Pinatubo: I. Phase equilibria and pre-eruption P-T-fO2-fH2O conditions of the dacite magma. Journal of Petrology, 40, 381411. Skogby, H. and Rossman, G.R. (1991). The intensity of amphibole OH bands in the infrared absorption spectrum. Physics and Chemistry of Minerals, 18, 6468. Wiles, D.B. and Young, P.A. (1981). A new computer program for Rietvield analysis of X-ray powder diffraction patterns. Journal of Applied Crystallography, 14, 149151.

MANUSCRIPT RECEIVED JUNE 28, 2002 MANUSCRIPT ACCEPTED MARCH 1, 2003 MANUSCRIPT HANDLED BY PETER BURNS