Quick
Search: 		  
advanced search
 GSW Home   GeoRef Home   My GSW Alerts   Contact GSW   About GSW   Journals List   Help 
European Journal of Mineralogy Email Content Delivery
JOURNAL HOME HELP FEEDBACK/COMMNET SUBSCRIBE ARCHIVE SEARCH TABLE OF CONTENTS
European Journal of Mineralogy; June 2001; v. 13; no. 3; p. 521-532; DOI: 10.1127/0935-1221/2001/0013-0521
© 2001 E. Schweizerbart'sche Verlagsbuchhandlung Science Publishers
This Article
Right arrow Abstract
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (12)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by WODARA, U.
Right arrow Articles by SCHREYER, W.
Right arrow Search for Related Content
GeoRef
Right arrow GeoRef Citation

Articles

X-site vacant Al-tourmaline

: a new synthetic end-member Ulrich WODARA and Werner SCHREYER

Institut für Geologie, Mineralogie und Geophysik, Ruhr-Universität, D-44780 Bochum, Germany

* e-mail: werner.schreyer{at}ruhr-uni-bochum.de

This paper was presented at the EMPG VIII meeting in Bergamo, Italy (April 2000)


   Abstract
Top
 Abstract
Introduction
 Syntheses
 Physical properties
 Chemical composition
 Discussion
 Stability problem
 Potential environments for the...
 Acknowledgements
 References
 
As a missing link of the tourmaline classification, an X-site vacant analogue of olenite was synthesized from three different starting compositions in the system Al2O3-B2O3-SiO2-H2O (ABSH) over the PT-range 4–40 kbar, 450–700 °C, however never as a single phase product; dumortierite with either quartz or with jeremejevite-OH were the coexisting solids. The starting materials containing 100% excess B2O3 produced an Al-tourmaline with exceptionally low cell parameters similar to those of synthetic excess-boron olenite. Electron microprobe analyses indeed show a surplus of boron over the classical 3.0 per formula unit, which is linked with a deficiency in silicon. There is a negative linear correlation between the cell volume a of Al-rich tourmalines and their tetrahedral boron contents. A preliminary structural formula of the synthetic Al-tourmaline obtained at 20 kbar, 600 °C is: ({square}0.96 Na0.004) Al3 (Al5.83 Si 0.19) [Si4.49 B 1.51 O18] (BO3)3 (OH)3.21 O0.79.

The small Na impurity is probably due to contamination from the NaCl pressure cell, octahedral Si is hypothetical. If the excess Si is considered to be due to analytical error, other hypothetical formulae exhibiting octahedral vacancies result. As none of the cations in the various structural sites of the tourmaline exceed 50% of the theoretical value, the following formula can be defined as idealized Al-tourmaline end-member: {square} Al3 Al6 [Si6O18] (BO3)3 O2 (OH)2. In nature, such X-site vacant Al-rich tourmalines would be expected to occur in Al-rich and (Mg, Fe)-, Ca- and alkali-free products of hydrothermal systems or of wall-rock alterations, or in their metamorphic equivalents such as dumortierite quartzites. An unusually Al-rich, complex tourmaline from a hydrothermal system in Nevada contains as much as 56 mol % of the above end-member together with schörl, foitite and rossmanite components.

Key-words: Al-tourmaline with tetrahedral boron, new tourmaline end-member, hydrothermal synthesis, system Al2O3-B2O3-SiO2-H2O.


   Introduction
Top
 Abstract
 Introduction
Syntheses
 Physical properties
 Chemical composition
 Discussion
 Stability problem
 Potential environments for the...
 Acknowledgements
 References
 
Following the recent proposal for the classification of minerals of the tourmaline group by Hawthorne & Henry (1999), its general crystal-chemical formula may be written as

Formula
In all tourmalines known up to the nineteen-seventies the X-site was found to be occupied by large cations such as Na and Ca. It was surprising, therefore, that tourmalines could also be prepared in synthetic systems not containing any such large cations. Rosenberg & Foit (1975, 1979, 1985) as well as Werding & Schreyer (1978, 1984) synthesized tourmaline in the system MgO-Al2O3-B2O3-SiO2-H2O (MABSH). Whereas Werding & Schreyer (1984) presented evidence that their synthetic tourmalines consistently had the formula {square}(Mg2Al)Al6[Si6O18](BO3)3(OH)4, Rosenberg & Foit (1985) also obtained tourmalines in which Mg partially to completely occupies the X-site. At any rate, both studies proved the existence of the first X-site vacant tourmaline, then named "alkali-free tourmaline". Because this is related to end-member dravite by the vector {square}AlNa-1Mg-1, a more specific name, "alkali-free dravite", was subsequently used by Werding & Schreyer (1996).

It was not until 1993 that a ferrous analogue of the above X-site vacant synthetic MgAl-tourmaline was defined from nature and named foitite (MacDonald et al., 1993). The fact that this natural mineral contains as much as 0.25 Na per formula unit (p.f.u.) in X is immaterial from the viewpoint of mineralogical classification, which defines endmembers and allows up to 50 mol % of a second component in binary, and up to 33 mol % of other components in ternary solid solutions. Thus, this natural foitite may still be classified in the "X-site vacancy group", which is one of the three major compositional groups of tourmaline after Hawthorne & Henry (1999). Consequently, natural Mgdominated X-site vacant tourmalines are called magnesiofoitites according to these authors. A natural example for this from the literature (Kuyunko et al., 1985), then not recognized as a new endmember, had been re-introduced by Werding & Schreyer (1996, p. 149). The new species magnesiofoitite was defined by Hawthorne et al. (1999). An X-site vacant tourmaline related to the Li-bearing tourmaline elbaite was decribed by Selway et al. (1998) and named rossmanite. It has the end-member formula {square}(LiAl2)Al6[Si6O18](BO3)3 (OH)4. Thus, the "elbaite" described by El-Hinnawi & Hofmann (1966) with a total X-site occupancy of 0.22 is actually a rossmanite. In summary, when ignoring the variable occupancy of the V and W positions, there are so far three end-members of the X-site vacancy group of tourmalines (Table 1).


View this table:
[in this window]
[in a new window]

  		Table 1. Comparison of common end-member formulae of Na-bearing tourmalines with the requisite X-site-vacant analogues.

 
Whereas all tourmaline minerals discussed before contain mainly Mg, Fe2+, or appreciable amounts of Li, in the Y-position, there is also a purely Al-bearing natural Na-tourmaline, which was first described and named olenite by Sokolov et al. (1986). It has the end-member formula NaAl3Al6[Si6O18](BO3)3O3(OH). Olenite is very rare in nature: only two localities are known thus far. Moreover, in at least one of these occurrences, olenite contains a considerable amount of excess boron replacing silicon in the tetrahedral T-site of the structure (Ertl et al., 1997), a fact ignored by Hawthorne & Henry (1999) in their crystal-chemical review. In an early experimental study of the relevant synthetic system Na2O-Al2O3-B2O3-SiO2 -H2O (NABSH) at 1 kbar, Rosenberg et al. (1986) had obtained only small amounts of minute tourmaline crystals that did not allow complete chemical analysis. More recent experimentation at high pressures, however, resulted in high yields of an excess-boron olenite having as much as 38 % of its tetrahedral silicon replaced by boron (Wodara & Schreyer, 1998; Schreyer et al., 2000). The existence of natural and synthetic olenites, no matter whether with or without excess boron, raised the question as to whether or not there is also a related X-site vacant tourmaline like for the Mg, Fe2+, Li-bearing ones.

Experiments aimed at tourmaline synthesis in the system Al2O3-B2O3-SiO2-H2O (ABSH) reported by Werding & Schreyer (1984) and Rosenberg et al. (1986) had remained without success. In the former case, it is now clear that the temperatures applied were too high. The first successful attempts to synthesize pure Al-tourmaline were mentioned by Werding & Schreyer (1996), referring to early stages of the work by Wodara that finally led to his thesis (Wodara, 1996). The tentative formula proposed by Werding & Schreyer (1996) for the product, however, was based on preliminary partial analyses and turned out to be wrong. In the present paper, the main results of Wodara's (1996) experiments in the ABSH-system are reported, which show remarkable similarities to those on synthetic olenite in the NABSH-system (Schreyer et al., 2000).

Table 1 summarizes the end-member formulae of Na-bearing tourmalines and compares them to the related known X-site vacant end-members. The analogue of olenite is hypothetical at this stage.


   Syntheses
Top
 Abstract
 Introduction
 Syntheses
Physical properties
 Chemical composition
 Discussion
 Stability problem
 Potential environments for the...
 Acknowledgements
 References
 
In Fig. 1 relevant crystalline phases of the system ABSH are plotted together with two hypothetical compositions of Al-tourmaline: T1 represents the formula Al3Al6[Si6O18](BO3)3O2(OH)2 as given in Table 1. It is derived from the olenite end-member formula NaAl3Al6[Si6O18](BO3)3O3(OH) by the vector {square}HNa–1 which was already implied in the original olenite formula Na1-xAl3Al6B3Si6O27(O, OH)4 as given by Sokolov et al. (1986). T2 stands for the formula {square}Al3Al6[Al2Si4O18](BO3)3(OH)4 which is related to T1 by the substitution 2(AlHSi-1). It was considered as a possible candidate for the composition of Al-tourmaline by Werding & Schreyer (1996) and also served this purpose in the experiments to be described here.


Figure 1
View larger version (11K):
[in this window]
[in a new window]

  		Fig. 1. Ternary projection of the system Al2O3-B2O3-SiO2-H2O (ABSH) from water showing relevant known crystalline phases as well as theoretical compositions and starting materials for tourmaline synthesis. T1 represents the theoretical formula {square}Al3Al6 [Si6O18](BO3)3O2(OH)2 as in Table 1, T2 = {square}Al3Al6[Si4Al2O18](BO3)3 (OH)4. SM1 and SM2 are the starting materials for the compositions T1 and T2, however each with a 100 mol % excess of B2O3. Synth. Al-T is the projection point of the synthetic Al-tourmaline as analyzed here (Table 4). Dumortierite solid solution after Werding & Schreyer (1990). The mineral boralsilite (Grew et al., 1998) was not encountered. Hydrous solids are underlined.

 

View this table:
[in this window]
[in a new window]

  		Table 4. Mean values of twenty microprobe analyses of synthetic Al-tourmaline (20 kbar, 600 °C) compared to theoretical composition {square}Al3Al6[Si6O18](BO3)3O2(OH)2 of Table 1.

 
AlSi-gels prepared according to the method of Hamilton & Henderson (1968) and AlSi-oxide mixes (from {gamma}-Al2O3 and aerosil) with the Al/Si-ratios appropriate for T1 and T2 were used as starting materials for the synthesis runs. The component B2O3 was added in nearly all cases in excess of 100 mol % relative to stoichiometry, in order to aid tourmaline crystallization (Werding & Schreyer, 1984). These excess-boron starting materials are indicated as SM1 and SM2 in Fig. 1. Crystalline H3BO3 or B2O3 were the sources of boron. In one set of experiments water was added together with H3BO3, in others the H2O-contents of the boron starting materials plus those of the gels (about 12 wt.%) or oxide mixes (about 4 wt.%) were considered sufficient to provide a hydrous fluid under run conditions.

Standard piston-cylinder equipment available at Ruhr-Universität Bochum was used for runs above 10 kbar. The low-pressure runs were performed in hydrothermal bombs. After the runs the products were washed in boiling distilled water in order to dissolve any excess boric acid present from quenching the coexisting fluids.

Run conditions and results of selected experiments are listed in Table 2 for the various starting materials employed. As shown by powder X-ray diffraction work, Al-tourmaline could be synthesized in varying quantities between 4 and 40 kbar at generally low temperatures (450–700 °C), but single-phase products were never obtained from the three different starting compositions used (SM1, SM2, T1; see Fig. 1). The highest yields (85 %) resulted from the SM1 gel with only excess B2O3 (Table 2, C). Higher amounts of water, either added or present in H3BO3, had a negative effect on tourmaline growth. Except for composition SM2, gels proved to be more productive than oxide mixes. Best conditions of synthesis are in the range 20–30 kbar and 600 °C; both higher pressure (40 kbar) and higher temperature (700 °C) are less favourable for tourmaline growth. Additional crystalline phases obtained are invariably dumortierite and quartz. In the products from oxide mixes, jeremejevite-OH, Al6[BO3]5 (OH)3 (Stachowiak & Schreyer, 1998), appeared as well. In one unproductive run at low pressure (Table 2, F), the mullite-type phase Al4B2O9 (see Werding & Schreyer, 1996, p. 128) formed. The gel run with the aluminous composition SM2 (Table 2, E) did not lead to tourmaline formation, even under the best conditions of synthesis, but to single-phase dumortierite. Perhaps importantly, the only run with stoichiometric boron starting material T1 (Table 2, G) also produced tourmaline, but in smaller amount than with excess boron under identical run conditions.


View this table:
[in this window]
[in a new window]

  		Table 2. Results of selected synthesis experiments on Al-tourmaline (for explanation of starting materials see Fig. 1 and text).

 
As a whole, the experimental results of Table 2 define a large growth field for the Al-tourmaline phase in pressure - temperature space. However, they do not lend themselves to any conclusions as to its chemical composition because of the presence of other solids and of fluids with unknown compositions. Therefore, in-situ chemical analyses of the tourmaline phase were vital.


   Physical properties
Top
 Abstract
 Introduction
 Syntheses
 Physical properties
Chemical composition
 Discussion
 Stability problem
 Potential environments for the...
 Acknowledgements
 References
 
Under the optical microscope, the reaction products obtained from the gel starting materials invariably appear as undiagnostic, extremely fine-grained felts exhibiting low birefringence due to small grain size. Early SEM studies revealed aggregates of very small prisms with lengths generally below 3 micrometres. However, with better instrumentation and under higher magnifications (Fig. 2a, b), tourmaline crystals with typical striations along the prism faces, trigonal morphology and length/width ratios of about 6 can be distinguished from extremely thin needles of dumortierite with length/width ratios of up to 60. During the course of this study it turned out to be of vital importance that a run with the oxide mix SM2 (Fig. 1) at 20 kbar, 600 °C (Table 2, F) yielded - in addition to fine-grained phase mixtures - larger homogeneous volumes with occasional rectangular cross sections and dimensions of up to 20 by 50 micrometres (Fig. 3), which were initially regarded as single crystals of tourmaline (Wodara, 1996). Recent re-inspection revealed, however, that they are actually poly crystalline with minute grain sizes. Nevertheless, the microprobe analyses performed on these volumes (see later) gave very consistent values with only small scatter. Therefore, they represent singlephase felts of tourmaline crystals that had formed in particular portions of the capsule due to a zonation of the reaction products.


Figure 2
View larger version (122K):
[in this window]
[in a new window]

  		Fig. 2. a, b. SEM images of portions of the run product for tourmaline synthesis at 30 kbar, 600 °C (see Table 2, C). In both pictures relatively stout prisms of the tourmaline phase can readily be distinguished from extremely thin needles of dumortierite. a) Note typical striations along a prism face of tourmaline on the right. Massive portion on the far right is the embedding medium. b) Largest crystal shown is tourmaline exhibiting trigonal morphology.

 

Figure 3
View larger version (143K):
[in this window]
[in a new window]

  		Fig. 3. Thin-section photograph of a rectangular fragment of a polycrystalline single-phase aggregate of tourmaline (Tu) embedded in a matrix of a fine-grained phase mixture. Product of synthesis run at 20 kbar, 600 °C (see Table 2, F).

 
In Table 3 those X-ray powder diffraction peaks of the synthetic Al-tourmaline are listed, which can be separated in the pattern from peaks of the coexisting phases dumortierite and quartz. Nevertheless, there are overlaps so that intensities can only be given in a qualitative way. The refined lattice constants of this Al-tourmaline are: a = 15.690(6) Å; c = 7.039(5) Å; V= 1500(1) Å3. They are larger than for the synthetic excess-boron olenite (Schreyer et al., 2000), but smaller than for elbaite and other tourmaline end-members (Fig. 4). Compared to the two known occurrences of natural olenites (see Schreyer et al., 2000, Fig. 3), the synthetic Al-tourmaline described here has smaller cell parameters as well.


View this table:
[in this window]
[in a new window]

  		Table 3. Selected X-ray powder diffraction data of synthetic Al-tourmaline prepared at 30 kbar, 600 °C. In the last column the requisite peaks of synthetic olenite (Schreyer et al., 2000) are shown for comparison.

 

Figure 4
View larger version (8K):
[in this window]
[in a new window]

  		Fig. 4. Plot of tourmaline cell dimensions c versus a for common end-members (see Table 1) as taken from Deer et al. (1986), compared to those of the new Al-tourmaline synthesized here as well as of the synthetic excess-boron olenite of Schreyer et al. (2000).

 

   Chemical composition
Top
 Abstract
 Introduction
 Syntheses
 Physical properties
 Chemical composition
Discussion
 Stability problem
 Potential environments for the...
 Acknowledgements
 References
 
Twenty electron-microprobe analyses were performed on the single-phase tourmaline felts of the type shown in Fig. 3 from the run product obtained from SM2 starting material at 20 kbar, 600 °C (Table 2, F). The CAMECA SX 50 electron microprobe of the Ruhr-Universität was employed using pyrope, jadeite and dravite as standards for Si and Al, Na, and B, respectively. Acceleration voltage was 15 kV and beam current 15 nA, except for the boron measurement (6 kV, 50 nA, using PC 1). Relatively small standard deviations even for the boron values suggest fairly high reliability for the latter element as well. Water could, of course, not be determined, but was calculated by difference, which undoubtedly introduces uncertainties.

The analytical results are summarized in Table 4 and compared to the theoretical values relating to the formula {square}Al3Al6[Si6O18](BO3)3O2(OH)2 as derived in Table 1 (= T1 in Fig. 1). The synthetic product clearly shows a strong deficiency in silicon and an excess of boron. The small amounts of Na detected are perhaps due to a contamination of the run product from the NaCl pressure cell of the pistoncylinder apparatus, but they could also be derived from unavoidable impurities in the chemicals used, if sodium fractionated into the tourmaline. Although the synthetic Al-tourmaline is thus not exactly an "X-site vacant tourmaline" phase of the ABSH-system, this feature is ignored here, not the least because mineralogical classification allows it. The amount of water (hydrogen) calculated is also higher than for the theoretical end-member, with the uncertainty discussed before. The composition found by analysis is plotted in Fig. 1, where it lies closest to the SM1 starting material.

Based on the experiences with synthetic olenite and the analytical and spectroscopic evidence cited there (Schreyer et al., 2000), the following hypothetical structural formula can be derived for the synthetic Al-tourmaline (cf. Table 4):

Formula
The excess boron is allocated to the tetrahedral ring site, where it substitutes for silicon. However, because the sum (Si + B) exceeds 9.00 p.f.u., a small amount of additional Si is provisionally accomodated in the octahedral Z site of the structure which —due to a lack of Al — would otherwise show vacancies. The problem of silicon in octahedral coordination was extensively discussed for the synthetic olenite (Schreyer et al., 2000), which displays this same feature even more strongly. However, no unambiguous evidence for octahedral silicon in tourmalines could be brought forward thus far. Because, in the present case, the excess silicon only slightly exceeds the analytical uncertainty, attempts were made at calculating structural formulae with different assumptions. One of these is to use (B+Si) = 9 as a basis, which indeed results in octahedral vacancies of 0.36 p.f.u. and H = 4.47 p.f.u. The latter value seems impossibly high for a tourmaline. If the excess charges of the hydrogen beyond 4.0 are redistributed, the following formula is obtained:

Formula
Again, octahedral vacancies appear due to the low Al, but B and Si, although in excess of 9.0 p.f.u., are now well within their analytical uncertainties. As all the alternative formulae show high hydrogen contents, it is clear that the synthetic Al-tourmaline differs from the theoretical end-member formula mainly by virtue of the substitution B3+ + H+ = Si4+. This was also found to be the main substitution linking ideal olenite (see Table 1) with the synthetic excess-boron olenite of Schreyer et al. (2000). The possible significance of octahedral vacancies in the formulae of Al-rich tourmalines will be discussed later.

With this compositional knowledge it is now possible to evaluate the significance of the cell parameters of the Al-tourmaline synthesized here (see before and Fig. 4), which — together with those of the synthetic olenites (Schreyer et al., 2000; Marler et al., 2000) — are the lowest ones ever found for any tourmaline. Because both the Altourmaline and these olenites carry substantial amounts of excess boron replacing silicon in the tetrahedral site, this substitution is most likely to have an important influence: according to Shannon (1976) boron in a tetrahedral site has only an effective ionic radius of 0.11 Å compared to silicon with 0.26 Å. In Fig. 5 the cell volumes V of the synthetic Al-tourmaline and of all natural and synthetic olenites known thus far are plotted against the contents of tetrahedral boron. Except for the type specimen of olenite from the Kola Peninsula (Sokolov et al., 1986), there is a negative linear trend between the two variables. Interestingly, this trend ends at zero tetrahedral boron very close to a value, which was reported by Rosenberg et al. (1986) for a chemically undefined Na-Al tourmaline synthesized at 1 kbar. This suggests that perhaps a low-pressure olenite with stoichiometric boron had been obtained by these authors. On the other hand, the original natural olenite (Sokolov et al., 1986) lies clearly off this trend. Note, however, that boron was not analyzed in the type-specimen, but assumed to be stoichiometric. Fig. 5 would suggest, therefore, that even this original olenite sample may contain a small amount of excess boron in the tetrahedral site. On the other hand, it seems that other chemical variables, especially the occupancy of the X-site, but also the presence of some Ca and Li in the olenite from the Koralpe, Austria (Ertl et al., 1997), have only a minor influence on the cell volume.


Figure 5
View larger version (15K):
[in this window]
[in a new window]

  		Fig. 5. Plot of cell volumes V of Alrich tourmalines taken from the literature as well as of the new Altourmaline synthesized here against their contents of tetrahedral boron replacing silicon as derived from chemical analyses. For the natural olenite type-specimen of Sokolov et al. (1986) as well as for the synthetic NaAl-tourmaline of Rosenberg et al. (1986) boron was assumed to be present in stoichiometric amounts only, without any [4]B.

 

   Discussion
Top
 Abstract
 Introduction
 Syntheses
 Physical properties
 Chemical composition
 Discussion
Stability problem
 Potential environments for the...
 Acknowledgements
 References
 

End-member definition

It is evident from Table 4 and the structural formulae derived for the synthetic Al-tourmaline that none of the crystallographic sites of the general tourmaline formula, that is X, Y, Z, T, [3]B and (V+W), shows occupancies that deviate by more than 50% from those of the initial theoretical formula given for the X-site vacant analogue of olenite in Table 1. Therefore, despite the excess-boron nature of the Al-tourmaline and its potential silicon in an octahedral site, its end-member formula remains {square} Al3 Al6 [Si6O18] (BO3)3 (OH)2 O2. This conforms to the rules of the IMA for defining and classifying minerals and may become important for the case that this "Al-tourmaline" be discovered in nature and requires a new name.

It should be mentioned, in this connection, that it is by no means certain that the composition found here for the synthetic Al-tourmaline (Table 4, Fig. 1) is unique in the ABSH-system. Based on the indications for solid solution in the related synthetic olenite system (Schreyer et al., 2000; Marler et al., 2000), it may even be probable that the Al-tourmaline analyzed is part of a more extended tourmaline miscibility range within this ABSH-system. Thus, the theoretical end-member itself, as defined before, may be part of such range, but special conditions of water fugacity and boron activity may be necessary to obtain it by experiment. The one run with the stoichiometric starting material T1 (Table 2, G) had yielded a tourmaline phase, but in smaller amount than with the excess-boron mixture SM1 under identical run conditions (Fig. 1). This may be due to the fact that the T1 composition is farther away from the tourmaline composition analyzed, but also that an Al-tourmaline with stoichiometric boron is harder to crystallize. Further syntheses on the T1 composition and detailed analytical and X-ray work on the products are necessary to clarify these compositional uncertainties.


   Stability problem
Top
 Abstract
 Introduction
 Syntheses
 Physical properties
 Chemical composition
 Discussion
 Stability problem
Potential environments for the...
 Acknowledgements
 References
 
All experiments conducted in the present study were synthesis runs and, thus, cannot be used to define a true thermodynamic stability field of the Al-tourmaline phase. In the absence of any run reversals to test equilibrium, the new tourmaline endmember could even be regarded as a metastable phase. On the other hand, with run durations of up to 720 hours at low, and up to 288 hours at high pressures (Table 2), it seems somewhat unlikely that only a transitional phase with a tourmaline structure was obtained. Moreover, the SEM pictures of one run product (Fig. 2a, b) do not suggest any reaction relations between tourmalines and the other Al-borosilicate phase dumortierite. On the contrary, they seem to imply two coexisting euhedral crystal species. This is not the case with increasing temperatures: Werding & Schreyer (1984, Table 2) had already shown that at 800 °C, 20 kbar, no tourmaline phase is present, but only dumortierite + quartz. Therefore, the upper stability limit of the Al-tourmaline phase synthesized here is expected to lie not far above 700 °C at 20 kbar (see Table 2).

Additional insight into the stability problem may be gained from the phase assemblages obtained in the synthesis runs (Table 2). With water having been added in excess, a fluid phase must always have coexisted with the solids under run conditions. Most probably, these hydrous fluids contained some dissolved boron and silica, but little alumina. They may thus project along the B2O3SiO2 side of the compositional triangle (Fig. 6). The most common four-phase assemblage obtained from the SM1 starting material (tourmaline + dumortierite + quartz + fluid) may — in theory — represent equilibrium in a four-component system (Fig. 6). In the two runs at 20 kbar, 550 °C with different durations (Table 2, A), the longer one produced more tourmaline, which might indicate a reaction relationship with tourmaline growing at the expense of the other phases at this low temperature. The relations are different for the products obtained from the SM1 and SM2 oxide starting materials (Table 2, D and F), which contain additional jeremejevite-OH. These five phases including fluid clearly indicate disequilibrium. However, it is also possible that quartz — here as well as partly in the SM1 gel runs (?) — is actually a quench product from the fluid. Thus, for the SM2 oxide run, the stable assemblage could be tourmaline + dumortierite + jeremejevite-OH + fluid (Fig. 6). Indeed, in this purely hydrous system ABSH, an assemblage jeremejevite + quartz may be unstable, contrary to fluorinebearing systems (Stachowiak & Schreyer, 1998, p. 886). The failure to grow any tourmaline in the single-phase dumortierite product of the SM2 gel run (Table 2, E) is more difficult to explain: the coexisting hydrous fluid virtually contained only the dissolved component B2O3 (Fig. 6), and there could be a reaction relationship between this assemblage dumortierite + B-rich fluid and the other pair Al-tourmaline + jeremejevite-OH. In fact, the latter assemblage could have a lower thermal stability than Al-tourmaline itself, which was just surpassed in this run.


Figure 6
View larger version (12K):
[in this window]
[in a new window]

  		Fig. 6. Projection of the ABSH-system as in Fig. 1 showing possible phase relations of the new Al-tourmaline (Al-T) and other phases as indicated for the condition 20 kbar, 600 °C (see Table 2) derived from the behaviour of the two different starting materials SM1 and SM2. For further discussion see text. Fluid compositions and dashed tie lines are hypothetical and do — for reasons of clarity — not depict the coexistences of fluids with dumortierite solid solutions (Du).

 
In summary, it seems that Al-tourmaline, perhaps with compositional variations, is indeed a stable phase of the ABSH-system.


   Potential environments for the occurrence of Al-tourmaline in nature and solid solubility toward this end-member
Top
 Abstract
 Introduction
 Syntheses
 Physical properties
 Chemical composition
 Discussion
 Stability problem
 Potential environments for the...
Acknowledgements
 References
 
The new synthetic Al-tourmaline is undoubtedly the chemically simplest one of all tourmaline phases known. Its composition is in obvious contrast to the extreme chemical complexity of most natural environments. Nevertheless, virtually pure Al2SiO5 polymorphs occur in such multicomponent environments; why not also the Al-tourmaline synthesized here? The answer can only lie in preferential element distribution mechanisms in tourmaline-bearing rocks and multicomponent chemical systems. Schreyer (2000) suggests that the tourmaline structure preferentially incorporates divalent cations such as Mg and Fe, but also Li, into its Y-sites to form the requisite Mg, Fe and Li endmembers as listed in Table 1, and solid solutions thereof. Tourmaline is indeed a sink for these elements, whereas any additional Al present that exceeds 6.0 p.f.u., that is the number of atoms in the Z-site of tourmaline, forms additional minerals such as white micas, Al-silicates, or even corundum. An impressive example for this distribution is provided by an extremely Al-rich corundum-rock from an Archaean Greenstone Belt in Zimbabwe (Schreyer et al., 1981, Fig. 8), which contains tourmaline coexisting exclusively with corundum and rutile. Nevertheless, this tourmaline is essentially a dravite-magnesiofoitite solid solution, Mg having completely fractionated into the tourmaline phase instead of forming some spinel. Thus, Al-richer tourmalines can only be expected to form in boron-bearing chemical environments that are virtually free of Mg, Fe, Li etc. This is the case for the two occurrences of natural olenite (Sokolov et al., 1986; Ertl et al., 1997). However, the apparent rarity of olenite in nature already suggests that such environments are not common. For the appearance of the Al-tourmaline presented here, the additional chemical constraint would have to be the lack of Na and of Ca, so that its X-site can remain more or less vacant.

Based on the still few natural occurrences of X-site vacant tourmalines, it seems that these minerals are most common in late magmatic, hydrothermal environments characterized by low-pH solutions (Fuchs & Maury, 1995). Enrichment of Al and removal of Mg, Fe and alkali, especially of Na, occurs as late wall-rock alteration and leads to "advanced argillic assemblages" (Meyer & Hemley, 1967). Similar chemical environments exist within late- to postmagmatic hydrothermal systems (Schmidt, 1985; Fuchs & Maury, 1995). X-site vacant to strongly X-site deficient tourmalines coexisting with dumortierite were first described from hydrothermally altered tuffs by Foit et al. (1989). Perhaps the type specimen of foitite (MacDonald et al., 1993), for which the exact locality is not known, is also from such environment. Klawa (1996) described foitites as the latest tourmaline generation in hydrothermally altered granites from the Erongo, Namibia. Interestingly, these foitites are partly replaced by dumortierite. In their structural investigation on a synthetic foitite prepared from a starting material of the ideal composition, Kahlenberg & Velickov (2000) report a total content of 7.42 octahedral Al p.f.u. This indicates about 21 mole % solid solution between foitite and ideal Al-tourmaline. A similar excess of Al was found by Fuchs & Maury (1995; Table 3) in a natural foitite occurring in the central zone of a hydrothermal field.

For the present discussion of particular relevance is the discovery by Fuchs (unpublished data; pers. comm. 2000) of an abnormally Al-rich and Fe-poor foititic natural tourmaline which occurs together with dumortierite in the centre of a hydrothermal system in the Humboldt Range, Nevada, USA (Fuchs & Maury, 1995). This tourmaline contains the following cations p.f.u. (31 oxygens): 0.18 Na; 0.01 Mg; 0.01 Cr; 0.91 Fe; 0.01 Ti; 7.74 Al; 5.97 Si; 0.01 F. On the basis of Mössbauer spectra, all Fe is divalent. There are no analytical values for H and B, but Li was found to be only 170 ppm. With this provision, but ignoring Li and the possible presence of Fe in both Y- and Z-sites (Fuchs & Maury, 1995), the following hypothetical cationic portion of a tourmaline formula is obtained:

Formula
There is thus an excess of 0.71 Al relative to the ideal foitite end-member formula of Table 1, but also a surprizingly high number of octahedral vacancies. If the small X-site occupancy of this foitite (0.18 Na p.f.u.) is taken to be due to a solid solution toward schörl following the vector NaFe2+{square}-1Al-1 (compare Table 1), and the minor elements as well as the octahedral vacancies are ignored, this tourmaline can actually be regarded as a ternary solid solution between the end-members schörl, foitite and Al-tourmaline (idealized formula as in Table 1). The following mole proportions can be calculated: Al-tourmaline 58 — foitite 21 — schörl 21. Thus, surprizingly, the Al-tourmaline component dominates by far, so that — following the rules of mineralogical classification — this tourmaline could actually be regarded as an impure representative of a new natural end-member. Note, however, that the hydrogen and boron contents of this tourmaline are not known for a complete evaluation of its substitutions. The small amount of lithium mentioned leads to an additional octahedral occupancy of Li 0.03, which reduces the number of vacancies to 0.32. This Li is best accounted for in terms of the new end-member rossmanite (Table 1) leading to the proportions: Al-tourmaline 56 — foitite 21 — schorl 21 — rossmanite 3.

An intriguing problem of the above Al-rich natural tourmaline from Nevada is its high number of octahedral vacancies (0.32 p.f.u.), which may also hold for the synthetic Al-tourmaline described here (see alternative formula given before). It should be emphasized, in this connection, that such vacancies are not new for Al-rich tourmalines: The Li-bearing natural olenite from the Koralpe, Austria (Ertl et al., 1997), exhibits 0.28 vacancy p.f.u., and the synthetic excess-boron olenite (Schreyer et al., 2000) shows that same number despite the allocation of excess Si to octahedral sites. In a redetermination of the crystal structure of the Koralpe olenite including new chemical analyses by Hughes et al. 1999), 0.17 octahedral vacancy p.f.u. was found as well. In order to explain this feature, Schreyer et al. 2000) had suggested a hypothetical "dioctahedral olenite" end-member with the formula Na (Al2 {square}) Al6 [Si6O18] (BO3)3 (OH)4. For the present case of X-site vacant Al-tourmalines, an end-member formula could be {square} (Al2.33 {square}0.67) Al6 [Si6O18] (BO3)3 (OH)4. In further experimental studies of tourmalines in the ABSH-system this composition should be investigated. If octahedral vacancies in Al-rich tourmalines can be confirmed by future crystallographic work, the last named end-member could be preferable in accounting for the constituing mole proportions in Fuchs' Al-rich tourmaline from Nevada discussed before.

Regional metamorphism affecting the Al-rich, Mg-Fe-Na poor rocks of hydrothermal origin at a later stage leads to Al2SiO5-bearing quartzites such as those of the Carolina Slate Belt, USA (Schmidt, 1985; Schreyer, 1987), which may even conserve typical minerals of the former hydrothermal systems such as alunite. If boron happens to be present, this could be the environment needed for the formation of Al-tourmaline with, or without, the additional presence of dumortierite. The fact that dumortierite quartzites are known to appear in metamorphic rock sequences indicates that, in principle, such environments existed in the geologic past. Unfortunately, in such rock samples from Brazil, we have not been able to find any tourmalines thus far.

Nevertheless, it is hoped that the surprizing synthesis of a pure Al-tourmaline end-member reported here, together with the above discussed tendency of natural tourmaline to form solid solutions toward this pure Al end-member, and the suggestions offered as to where to search for such tourmalines in nature will finally result in new discoveries and lead to the definition of a new end-member name within the tourmaline group of minerals.


   Acknowledgements
Top
 Abstract
 Introduction
 Syntheses
 Physical properties
 Chemical composition
 Discussion
 Stability problem
 Potential environments for the...
 Acknowledgements
References
 
The major part of this paper is taken from the Diploma Thesis in Mineralogy submitted by U.W. at Ruhr-Universität Bochum. He thanks the staff of the Institut für Mineralogie for manifold help extended. We appreciate the active involvement of H.-J. Bernhardt, Bochum, to provide the best possible electron microprobe data. Yves Fuchs, Paris, generously provided unpublished data on an unusual natural tourmaline. Insightful reviews of the manuscript were provided by C. Chopin, Y. Fuchs, J.-L. Robert and P.E. Rosenberg.

Received 19 May 2000
Modified version received 6 December 2000
Accepted 19 December 2000


   References
Top
 Abstract
 Introduction
 Syntheses
 Physical properties
 Chemical composition
 Discussion
 Stability problem
 Potential environments for the...
 Acknowledgements
 References
 
Deer, W.A., Howie, R.A., Zussmann, J. (1986): Rockforming minerals. Vol., 1B, Disilicates and ring silicates. Longman, Harlow, UK, 629 p.

El-Hinnawi, E.E. & Hofmann, R. (1966): Optische und chemische Untersuchungen an neun Turmalinen (Elbaiten). N. Jahrb. Mineral. Mh., 1966, 80–89.

Ertl, A., Pertlik, F. Bernhardt, H.-J. (1997): Investigations on olenite with excess boron from the Koralpe, Styria, Austria. Öster. Akad. Wiss., Math.-Naturw. Kl. Anzeiger Abt.I, 134, 3–10.

Foit, F.F. Jr., Fuchs, Y., Myers, P.E. (1989): Chemistry of alkali-deficient schorls from two tourmaline-dumortierite deposits. Am. Mineral., 74, 1317–1324.[Abstract][ISI][GeoRef]

Fuchs, Y. & Maury, R. (1995): Borosilicate alteration associated with U-Mo-Zn and Ag-Au-Zn deposits in volcanic rocks. Mineral. Deposita, 30, 449–459.

Grew, E.S., McGee, J.J., Yates, M.G., Peacor, D.R, Rouse, R.C., Huijsmans, J.P.P., Shearer, C.K., Wiedenbeck, M., Thost, D.E., Su, S.-C. (1998): Boralsilite (Al16B6Si2O37): A new mineral related to sillimanite from pegmatites in granulite-facies rocks. Am. Mineral., 83, 638–651.[Abstract][ISI][GeoRef]

Hamilton, D.L. & Henderson, C.M.B. (1968): The preparation of silicate compositions by a gelling method. Mineral. Mag., 36, 832–838.[ISI][GeoRef]

Hawthorne, F.C. & Henry, D.J. (1999): Classification of the minerals of the tourmaline group. Eur. J. Mineral., 11, 201–215.[Abstract/Free Full Text][GeoRef]

Hawthorne, F.C., Selway, J.B., Kato, A., Matsubara, S., Shimizu, M., Grice, J.D., Vajdak, J. (1999): Magnesiofoitite, {square}(Mg2Al)Al6(Si6O18)(BO3)3(OH)4, a new alkali-deficient tourmaline. Can. Mineral., 37, 1439–1443.[Abstract/Free Full Text][ISI][GeoRef]

Hughes, J.M., Ertl, A., Dyar, M.D., Grew, E.S., Shearer, K., Yates, M.G. (1999): Boron in the tourmaline tetrahedral site: Chemistry and structure of a boron-rich olenite. Abstract 2nd European Workshop on Tourmaline and Borosilicates, Paris 21–24 September 1999, Bull. Liaison S.F.M.C., 11, 117.

Kahlenberg, V. & Velickov, B. (2000): Structural investigations on a synthetic alkali-free hydrogen-deficient Fetourmaline (foitite). Eru. J. Mineral., 12, 947–953.

Klawa, M. (1996): Petrologie der Dumortieritgesteine der Farm Etemba, Erongo, Namibia, und assoziierter saurer Magmatite. Unpublished Diploma Thesis, Ruhr-Universität Bochum, 86 p.

Kuyunko, N.S., Semenov, Y.V., Gorevich, V.M., Kuz'min, V.I., Topor, N.D., Gorburov, V.Y. (1985): Experimental determination of the thermodynamic properties of tourmaline-dravite. Geochem. International, 22, 109–116.

MacDonald, D.J, Hawthorne, F.C., Grice, J.D. (1993): Foitite, {square}[Fe2+2(Al, Fe3+)]Al6Si6O18(BO3)3(OH)4, a new alkali-deficient tourmaline: Description and crystal structure. Am. Mineral., 78, 1299–1303.[Abstract][ISI][GeoRef]

Marler, B., Borowski, M., Wodara, U., Schreyer, W. (2000): The structure of synthetic excess-boron-tourmaline. Berichte der Deutschen Mineralogischen Gesellschaft, Beih. z. Eur. J. Mineral. Vol. 12, No. 1, 120.

Meyer, C. & Hemley, J.J. (1967): Wall rock alteration. in "Geochemistry of hydrothermal ore deposits", H.L. Barnes, ed. Holt, Rinehardt & Winston, Inc., New York, 166–235.

Rosenberg, P.E. & Foit, F.F. Jr. (1975): Alkali-free tourmalines in the system MgO-Al2O3-SiO2-B2O3-H2O (abstr.) Geol. Soc. Am. Ann. Meeting Abstr. Progr., 1250–1251.

Rosenberg, P.E. & Foit, F.F. Jr. (1979): Synthesis and characterization of alkali-free tourmaline. Am. Mineral., 64, 180–186.[Abstract][ISI][GeoRef]

Rosenberg, P.E. & Foit, F.F. Jr. (1985): Tourmaline solid solutions in the system MgO-Al2O3-SiO2-B2O3-H2O. Am. Mineral., 70, 1217–1223.[Abstract][ISI][GeoRef]

Rosenberg, P.E., Foit, F.F. Jr., Ekambaram, V. (1986): Synthesis and characterization of tourmaline in the system Na2O-Al2O3-SiO2-B2O3-H2O. Am. Mineral., 71, 971–976.[Abstract][ISI][GeoRef]

Schmidt, R.G. (1985): High-alumina hydrothermal systems in volcanic rocks and their significance to mineral prospecting in the Carolina Slate Belt. U.S. Geol. Surv. Bull., 1562, 1–59.

Schreyer, W. (1987): Pre- or synmetamorphic metasomatism in peraluminous metamorphic rocks? in "Chemical transport in metasomatic processes", H.C. Helgeson, ed. D. Reidel Publishing Company, 265–296.

Schreyer, W. (2000): Is the partitioning of boron between tourmaline and muscovite dependent on the crystallization environment? J. Czech Geol. Soc., 45, 13–20.

Schreyer, W., Werding, G., Abraham, K. (1981): Corundum-fuchsite rocks in greenstone belts of Southern Africa: Petrology, geochemistry, and possible origin. J. Petrol., 22, 191–231.[Abstract/Free Full Text][CrossRef][ISI][GeoRef]

Schreyer, W., Wodara, U., Marler, B., van Aken, P.A., Seifert, F., Robert, J.-L. (2000): Synthetic tourmaline (olenite) with excess boron replacing silicon in the tetrahedral site: I. Synthesis conditions, chemical and spectroscopic evidence. Eur. J. Mineral., 12, 529–541.[Abstract/Free Full Text][CrossRef][ISI][GeoRef]

Selway, J.B., Novak, M., Hawthorne, F.C., Cemy, P., Ottolini, L., Kyser, T.K. (1998): Rossmanite, {square}[Li-Al2]Al6(Si6O18)(BO3)3(OH)4, a new alkali-deficient tourmaline: Description and crystal structure. Am. Mineral., 83, 896–900.[Abstract][ISI][GeoRef]

Shannon, R.D. (1976): Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides. Acta Cryst., A32, 751–767.[ISI]

Sokolov, P.B., Gorskaya, M.G., Gordienko, V.V., Petrova, M.G. Kretzer, Yu. L., Frank-Kamenitskii, V.A. (1986): Olenite Na1-xAl3Al6B3Si6O27(O, OH)4 — A new highalumina mineral of the tourmaline group. Zapiski Vses. Mineralog. Obshch., 115, 119–123 (in Russian).

Stachowiak, A. & Schreyer, W. (1998): Synthesis, stability and breakdown products of the hydroxyl end member of jeremejevite in the system Al2O3-B2O3-H2O. Eur. J. Mineral., 10, 875–888.[Abstract/Free Full Text][ISI][GeoRef]

Werding, G. & Schreyer, W. (1978): Synthesis and crystal chemistry of kornerupine in the system MgO-Al2O3-SiO2-B2O3-H2O. Contrib. Mineral. Petrol., 67, 247–259.[CrossRef][ISI]

Werding, G. & Schreyer, W. (1984): Alkali-free tourmaline in the system MgO-Al2O3-B2O3-SiO2-H2O. Geochim. Cosmochim. Acta, 48, 1331–1344.[CrossRef][ISI][GeoRef]

Werding, G. & Schreyer, W. (1990): Synthetic dumortierite: Its PTX-dependent compositional variations in the system Al2O3-B2O3-SiO2-H2O. Contrib. Mineral. Petrol., 105, 11–24.[CrossRef][ISI][GeoRef]

Werding, G. & Schreyer, W. (1996): Experimental studies on borosilicates and selected borates. Rev. Mineral., 33, 117–163.[Abstract][GeoRef]

Wodara, U. (1996): Synthese und Eigenschaften von Turmalinen im System Na2O-Al2O3-SiO2-B2O3-H2O. Unpublished Diploma Thesis, Ruhr-Universität Bochum, 99 p.

Wodara, U. & Schreyer, W. (1998): Tetrahedral boron in tourmalines of the system Na2O-Al2O3-B2O3-SiO2-H2O. Terra Abstracts, Abstract supplement No. 1 to Terra nova, 10, 68.




This article has been cited by other articles:


Home page
 Can MineralHome page
A. C. Zhang, R. C. Wang, S. Y. Jiang, H. Hu, and H. Zhang
CHEMICAL AND TEXTURAL FEATURES OF TOURMALINE FROM THE SPODUMENE-SUBTYPE KOKTOKAY No. 3 PEGMATITE, ALTAI, NORTHWESTERN CHINA: A RECORD OF MAGMATIC TO HYDROTHERMAL EVOLUTION
Can Mineral, February 1, 2008; 46(1): 41 - 58.
[Abstract] [Full Text] [PDF]

Home page
 American MineralogistHome page
E. S. Grew, H. A. Graetsch, B. Poter, M. G. Yates, I. Buick, H.-J. Bernhardt, W. Schreyer, G. Werding, C. J. Carson, and G. L. Clarke
Boralsilite, Al16B6Si2O37, and "boron-mullite:" Compositional variations and associated phases in experiment and nature
American Mineralogist, February 1, 2008; 93(2-3): 283 - 299.
[Abstract] [Full Text] [PDF]

Home page
 Can MineralHome page
J. Cempirek, M. Novak, A. Ertl, J. M. Hughes, G. R. Rossman, and M. D. Dyar
Fe-BEARING OLENITE WITH TETRAHEDRALLY COORDINATED Al FROM AN ABYSSAL PEGMATITE AT KUTNA HORA, CZECH REPUBLIC: STRUCTURE, CRYSTAL CHEMISTRY, OPTICAL AND XANES SPECTRA
Can Mineral, February 1, 2006; 44(1): 23 - 30.
[Abstract] [Full Text] [PDF]

Home page
 Eur J MineralHome page
H. R. MARSCHALL, A. ERTL, J. M. HUGHES, and C. McCAMMON
Metamorphic Na- and OH-rich disordered dravite with tetrahedral boron, associated with omphacite, from Syros, Greece: chemistry and structure
European Journal of Mineralogy, October 1, 2004; 16(5): 817 - 823.
[Abstract] [Full Text] [PDF]

Home page
 Can MineralHome page
A. C. Zhang, R. C. Wang, H. Hu, X. M. Chen, and H. Zhang
OCCURRENCES OF FOITITE AND ROSSMANITE FROM THE KOKTOKAY NO. 3 GRANITIC PEGMATITE DYKE, ALTAI, NORTHWESTERN CHINA: A RECORD OF HYDROTHERMAL FLUIDS
Can Mineral, June 1, 2004; 42(3): 873 - 882.
[Abstract] [Full Text] [PDF]

Home page
 American MineralogistHome page
J. M. Hughes, J. M. Hughes, A. Ertl, M. D. Dyar, E. S. Grew, M. Wieden-beck, and F. Brandstatter
Structural and chemical response to varying [4]B content in zoned Fe-bearing olenite from Koralpe, Austria
American Mineralogist, February 1, 2004; 89(2-3): 447 - 454.
[Abstract] [Full Text] [PDF]

Home page
 Mineral MagHome page
J. Torres Ruiz, J. Torres-Ruiz, A. Pesquera, and V. Lopez Sanchez-Vizcaino
Chromian tourmaline and associated Cr-bearing minerals from the Nevado-Filabride Complex (Betic Cordilleras, SE Spain)
Mineralogical Magazine, June 1, 2003; 67(3): 517 - 533.
[Abstract] [Full Text] [PDF]

Home page
 Eur J MineralHome page
W. SCHREYER, J. M. HUGHES, H.-J. BERNHARDT, A. KALT, S. PROWATKE, and A. ERTL
Reexamination of olenite from the type locality: detection of boron in tetrahedral coordination
European Journal of Mineralogy, October 1, 2002; 14(5): 935 - 942.
[Abstract] [Full Text] [PDF]

Home page
 Eur J MineralHome page
B. MARLER, M. BOROWSKI, U. WODARA, and W. SCHREYER
Synthetic tourmaline (olenite) with excess boron replacing silicon in the tetrahedral site: II. Structure analysis
European Journal of Mineralogy, August 1, 2002; 14(4): 763 - 771.
[Abstract] [Full Text] [PDF]