J. Japan. Assoc. Min. Petr. Econ. Geol. 73, 231-240, 1978

PHASE RELATIONS IN THE SYSTEM CaO-FeOX SiO2 IN H20-CO2 MIXTURES

TETSUYA SHOJI Department of Mineral Development Engineering, The University of Tokyo

The phase relations among calcite, wollastonite , ferrobustamite, andradite, heden- bergite, magnetite and quartz in H2O-CO2 mixtures have been estimated on the basis of previous work in the systems CaO-SiOa-CO, and CaO-Fe01 SiO2-O2. The upper limit of COs content of the condition favorable to the formation of skarn involving andradite and/or hedenbergite is restricted by the following reactions:

(6) andradite+CO2 =calcite+magnetite+quartz+O2,

(8) hedenbergite+CO2+O2=calcite+magnetite+quartz. From the andradite and hedenbergite fields, the skarn consisting of these minerals is inferred to be formed in H2O-CO2 fluids containing less 30 mole % CO2 at 500•Ž , and 0.5 mole % at 300•Ž, respectively. When ferrobustamite is stable, the phase relations are more complex. The above mentioned values, however, are independent on the stability of ferrobustamite.

INTRODUCTION plot regardless CO, (Fig. 1), and listed in Table 1 with their abbreviations and molar Grandite garnet and/or chnopyroxene volumes calculated from the data complied of diopside-hedenbergiteseries are the most by Strunz (1970) and Yamanaka et al. (1977). important minerals which characterize the

skarn-type ore deposits. They frequently Table 1. List of minerals. contain a considerable amount of iron. Besidesthem, the ore deposits of this type are usually composed of calcite, magnetite and quartz with sulfidesand oxides as ore minerals. For this reason, the phase rela tions among calcite, andradite, hedenbergite, * Calculated fron the data compiled by Strunz (1970). magnetite and quartz should indicate the except ferrobustamite (Yamanakaet al., 1977). favorable conditions of formation of ore bearing skarn. SOME STABLE REACTIONS IN THE SYSTEM Ca-Fe-Si-C-O AND RELA In this paper, the writer tries to estimate quantitatively the phase relations TIONSHIP AMONG THEM in the system CaO-FeOBSiO2 under the Fig. 2 shows the schematic, isobaric T fluidscomposed of water and carbon dioxide, f co, relations among calcite, andradite, on the basis of previous studies of the system hedenbergite and quartz under a controlled CaO-SiO2-CO2and CaO-FeORSiO2-O2. oxygen fugacity. In this diagram, there Compositions of phases considered in are apparently four invariant points and this paper are shown on a CaO-FeOx-SiO2 eleven univariant curves. These univariant

(Manuscript received July 3, 1978) 232 Tetsuya Shoji

curves are equilibirum boundaries for the reactions listed in Table 2. These reactions are combined one another as follows:

Reaction(2)=1/2{9•EReaction(3)

-Reaction (1)} (1)

Reaction(3)=1/9{Reaction(1)

+2• Reaction (2)} (2)

Reaction(4)=1/3{Reaction(2)

-Reaction(1)} (3)

=1/2(3• Reaction(3)

-Reaction (1)} Fig. 1. Compositions of crystalline phases in the Reaction(6)=9{Reaction(5) system Ca-Fe-Si-C-O projected on the -Reaction(1)} (4) CaO-FeOx SiO2 plane.

Reaction(7)=Reaction(6)-Reaction(2) (5) +9•EReaction(5)} (6)

Reaction(8)=1/3(Reaction(2)-Reaction (1) =1/3(Reaction(2)+Reaction(6)}

Fig 2. Schematic analysis of phase relations in the system CaO-FeO-Fe2O3- SiO2-CO2 as a function of CO3 fugacity and temperature. Numerals correspond to the number of reaction (see Table 2). Capital letters in circles show isobaric, f02-defined invariant points (see Table 4). Abbreviations are shown in Table 1 and Fig. 1. Phase Relations in the System. CaO-FeOx-SiO2 in H2O-CO2 Mixtures 233

Table 2. Volume changes, enthalpy changes and entropy changes of some stable reactions in the system CaO-FeO-Fe202-Si02-C027% (at atmospheric pressure).

* Gu=Gustafson (1974), Gw=Greenwood(1967), H-T=Harker and Tuttle (1956), K=Kurshakova (1971a,b), L=Liou (1974), S1=Shoji (1976b), S2=Shoji (1977), T-L=Taylor and Liou (1978)

Reaction(9)=Reaction(3)-Reaction(5) (7) REACTIONS (1) TO (4) =1/9 {2•EReaction(2)-Reaction(6)} At first, let us consider the stability Reaction (10) relations among wollastonite, andradite, =2{Reaction(5)-Reaction(3)} (8) hedenbergite, magnetite and quartz. Fig. = Reaction(3)+2•EReaction (9)

Reaction(11)=1/3{2•EReaction(1)

+Reaction (7)} (9)

Equations (1) to (9) mean that the number

of independent reactions is three. Accord

ingly, if three equilibrium boundaries

among them, for example Reactions (1), (2)

and (5), have been already determined,

the other reactions can be calculated by the

thermodynamic procedure which is derived

from the reaction isobar of van't Hoff (Shoji,

1976b),

Fig. 3. Schematic analysis of phase relations at •¬. On the estimation, Lewis' rule (ideal isobaric invariant point I, where the mixing of fluids) has been assumed, and assemblage wollastonite-andradite-heden- bergite-magnetite-quartz is stable. Solid fugacitydata of carbon dioxide and oxygen, lines represent the isobaric univariant except that of the Pb-PbO buffer (see phase relations, and dotted lines buffer Appendix), have been taken from Me,7ni curves. Numerals correspond to the num ber of reaction (see Table 2). Abbrevi- (1971)and Huebner (1971), respectively. ations are shown in Table 1. NNO and FMQ represent the Ni-NiO buffer and the fayalite-magnetite-quartz one, respec tively. 234 Tetsuya Shoji

3 shows the schematic, isobaric fo2-T rela This line is not necessarily inconsistent with tions in the system Ca-Fe-Si-O. In this the results of Iypmago$a (1971a, b). diagram, there are one invariant point and From this line, the equilibrium constant four univariant curves. The univariant of Reaction (2) is as follows: curves are equilibrium boundaries for

Reactions (1) to (4). Equations (1) to (4) mean that the number of independent reactions is two. Phase relations in the system have been investigated by Rypma- •¬ The equilibrium boundary for Reaction EP Ba (1970,1971a, b), „K„…„‚‡V„p„{„€„Ž„p „y „@„Ž„u„„„y„ƒ„‘„~ (3) was determined by I ypmat oBa (1970, (1974), Gustafson (1974) and Liou (1974), 1971a), and by Liou (1974). Rypmaspa using solid phase oxygen buffers. The (1971a) stated that at 1000 bars Pffuid the equilibrium boundary for Reaction (1) was assemblage andradite-quartz is stable above determined by Gustafson (1974). The ex an oxygen fugacity of 10-13 bar at 800•Ž, periments were carried out between 500 and 10-20 bar at 500•Ž. On the contrary, bars and 3000 bars. The enthalpy and Liou (1974) concluded that at 2000 bars entropy changes, and equilibrium constant of Reaction (1) were calculated by Shoji

(1977) (Table 1). The equilibrium boundary for Reaction

(2) was determined by KypmanoBa (1971a, b), and by Gustafson (1974). IfiypinanoBa

(1971b) stated that at 1000 bars Pflnid hedenbergite is stable below an oxygen fugacity of 10-14 bar at 750•Ž and 10-23 bar at 400•Ž. On the contrary, Gustafson

(1974) concluded that at 2000 bars Pflnid hedenbergite is stable below an f o, of 10-13 bar at 800•Ž and 10-28 bar at 400•Ž. Both of the conclusions do not agree with each other at the low temperature. The data of their experiments, however, are ap proximately coincident with each other, as shown in Fig. 4, which represents a {2 ln fo2+‡™Vs(pf-1)/RT}vs. 1000/T plot of their results. The low temperature experiments performed by KypmanoBa (1971a, b) seem to have not been demonstrated as equilibrium, because the assemblage an Fig. 4. A (2 ln fo2+‡™Vs (iii-1)/RT) vs. 1000/T dradite-hedenbergite-magnetite-quartz was plot of the experimental data on Reac synthesized at 400°C within a wide fo, tion (2) reported by lcypmazoaa (1971a, b) and Gustafson (1974). Open symbols range. The straight line in Fig. 4 has been indicate reaction to andradite-magnetite- drawn for the data of Gustafson (1974). quartz; closed symbols to hedenbergite. Phase Relations in the System CaO-FeOx-SiO2 in H2O-C02 Mixtures 235

Pfluid this assemblage is stable above an fo, of 10-is bar at 800•Ž, and 10-26 bar at 500•Ž. Both of the conclusions do not

agree with each other at the low tem

perature. Liou (1974) suggested that the experiments performed by I£ypma iosa

(1970, 1971a) had not been successfully demonstrated as equilibrium because of

starting materials, run duration and experi

mental methods. Fig. 5 shows a {1/2 In fo2

+‡™Vs(Pf-1)/RT}vs. 1000/T plot of their

experimental results. The solid line in Fig.

5 have been drawn for the data of Liou

(1974). From the line, the enthalpy and Fig. 5. A{1/2 ln fo2+‡™Vs(pf-1)/RT}vs. 1000/ entropy changes of Reaction (3) are calcu- T plot of the experimental data on Re lated to be 86.1 kcal and 52.5 cal.deg-1, action (3) reported by Kypmaxoea (1970, respectively. On the other hand, the 1971a) and Liou (1974). Open symbols indicate reaction to andraite-quartz; Gustafson's (1974) results suggest that closed symbols to wollastonite (or ferro- the enthalpy and entropy changes of bustamite ?)-hedenbergite. Reaction (3) are 71.5 kcal and 38.0 cal.deg-1, respectively (Table 2). These values have demonstrated as equilibrium. For this been calculated using equation (2). The reason, the writer believes that the values dashed line in Fig. 5 shows the calculated obtained from the results of Gustafson equilibrium boundary for Reaction (3). (1974)are more accurate than those on the The dashed line is not necessarily incon- results of I£ypniagosa (1970, 1971a) and sistent with the results of Kypluaaosa (1970, Liou (1974)*. 1971a), while it does not agree with the The enthalpy and entropy changes of results of Liou (1974). Since „K„…„‚‡V„p„{„€„r„p Reaction (4) can be calculated using equa

(1970, 1971a) and Liou (1974) carried out tion (3) or (4). They are 47.4 kcal and their experiments without any attention to 16.7 cal. deg-1 respectively (Table 2). the formation of ferrobustamite, Ca5FeSi6O18 The enthalpy and entropy changes of

(see the section following the next), it is Reactions (1) to (4) are listed in Table 2, inferred that their experiments were not and the f o.-T relations of the reactions

* There are two other reasons that the writer considers thus . (i) On the basis of the results of

„K„…„‚‡V„p„{„€„r„p(1970, 1971a) and Liou (1974), the lowest temperature of hedenbergite field is calculated to be 625•Ž and 540•Ž at the oxygen fugacities defined by the NNO buffer and the FMQ buffer, respectively. These values seem to be too high for the formation of hedenbergite, because it is

inferred that the oxygen fugacities favorable to the formation of ore-bearing skarn are between the MH buffer and the FMQ buffer (Shimazaki, 1974; Shoji 1978), and the temperatures are about 400•Ž or lower (Shoji, 1975, 1978) On the contrary, (ii) the thermochemical data of Reaction (1) determined from the Gustafson's (1974) results appear to have a high reliability, because the equilibrium boundary for Reaction (6) calculated from the data is nearly coincident

with the one calculated only from the thermodynamic properties in spite of the independent

procedure (Shoji, 1977). 236 Tetsuya Shoji are shown in Table 3. listed in Table 3. The log fco2-log fo2-Pf- T relations of the equilibrium boundaries REACTIONS (5) TO (11) for Reactions (5) to (11) are also listed in Table 3. T he equilibrium boundary for Reac Fig. 6 shows the phase relations among tion (5) was experimentally determined by calcite, wollastonite, andradite, hedenberg Harker and Tuttle (1956) and Greenwood ite, magnetite and quartz in H2O-CO2 (1967). On the basis of the results, the fluids under the conditions where fluid enthalpy and entropy changes were calculat ed by Shoji (1976b). pressures are 1000 bars and 2000 bars, and oxygen fugacities are defined by the NNO The equilibrium boundary for Reaction buffer and the FMQ buffer. The diagrams (6) was thermodynamically calculated by are drawn on the case S. The equilibrium Shoji (1977), and was experimentally boundary for Reaction (4) does not exist, determined by Taylor and Liou (1978). because it is calculated to pass through the According to the former, when oxygen fugacity is. represented by the MH buffer, high temperature region. the equilibrium boundary for Reaction (6) Points A, B, C and D in Fig. 2 are invariant under a specified pressure and a at 2000 bars approximately passes through the points: T=550•Ž, Xco2=44 mole %; defined oxygen fugacity. Among them, however, only points A and B are signifi and T=630•Ž°C, Xco2=100 mole %. Mean cant under high oxygen fugacities, such as while, according to the latter, Reaction (6) the NNO buffer (Table 4, Figs. 6a and c). occurs at T=550•}10•Ž°C at Xco2=22 mole Meanwhile point C is significant under low %, T=596•}8•Ž at Xco2=50 mole %, and oxygen fugacities, such as the FMQ buffer 650•}10•Ž at 100 mole %. According to

Taylor and Liou (1978), the enthalpy and (Table 4, Figs. 6b and d), too. The mineral assemblages, temperatures and C02 contents entropy changes of Reaction (6) are 139.9 at these isobaric, f o=defined invariant kcal and 316.0 cal.deg-1, respectively.

These values differ from those estimated by points are given in Table 4. According to Table 4, the Xco2 values of point A are Shoji (1977) over the limit of error. In smaller on the case T than the case S. The the following statement, the case S means difference between both cases is about 0.4 the discussion based on the values of Shoji in log Xco2 (NNO buffer) or 1.0 (FMQ buffer). (1977), and the case T means that on the On the contrary, the Xco2 values of point B results of Taylor and Liou (1978). are approximately equal to each other The entalpy and entropy changes of between the cases S and T. Consequently, Reactions (7) to (11) can be calculated on the case T, the point A sits on the CO2- using equations (6) to (10), respectively.

The calculated values are listed in Table 2. poor position, and lines 7, 8 and 9 pass through the slightly CO, poor regions, compared with the diagrams shown in Fig. PHASE RELATIONS AMONG CALCITE, 6. WOLLASTONITE, ANDRADITE, HEDEN BERGITE, MAGNETITE AND QUARTZ SYSTEM INVOLVINGFERROBUSTAMITE

The log fo2-pf-T relations of equilibrium (IRON-WOLLASTONITE) boundaries for Reactions (1) to (4) are Recently, ferrobustamite (iron-wo- Phase Relations in the System CaO-FeOR-SiO2 in H2O-CO2 Mixtures 237

Fig. 6. Calculated Xcog T relations of equilibrium boundaries for some stable reactions in the system CaO-FeO-Fe20$ Si02 CO$ H2O. Numerals correspond to the number of reaction (see Tables 2 and 3). Capital letters in circles show the isobaric, foe -defined invariant points (see Table 4). Abbreviations are shown in Table 1 and Fig. 1. Fluid pressures and oxygen fugacities are respectively as follows: (a) 1000 bars and Ni-NiO buffer, (b) 1000 bars and fayalite-magnetite-quartz buffer, (c) 2000 bars and Ni-NiO buffer, and (d) 2000 bars and, fayalite-magnetite-quartz buffer.

Ilastonite) have been reported from many 4Ca3Fe2Si3O12+8SiO2 skarn-type ore deposits of Japan (e.g. =Ca5FeSi6O18+7CaFeSi2O6+2O2, Matsueda, 1973a; Shimazaki and Yama (Al) naka, 1973). The composition of fer instead of Reaction (3). Accordingly, robustamite is Ca5FeSi6O18on the join CaSiO3- ferrobustamite may have been formed by CaFeSi2O6. For this reason, if fer the syntheses carried out by ItypmasoBa robustamite is stable, the tie-line between (1970, 1971a) and by Liou (1974). The andradite and quartz is cut by the reaction, formation of ferrobustamite is one of the 238 Tetsuya Shoji

Table 3. Equations of equilibrium boundaries for Reactions (1) to (11) and (Al) in C02-09 mixtures.

* S=Shoji (1977), T=Taylor and Lion (1978).

Table 4. List of some isobaric, fo2-defined invariant points in the system CaO-FeO-Fe2O3 SiO2-CO3-O2-H2O.

* s=Shoji (1977), T=Taylor and Liou (19781

=9Ca5FeSi6O18+7Fe3O4+2O2 probable explanations for the incoincidence between the results of Liou (1974) and those (A4)

Gustafson (1974), which is shown in Fig. 5. 15Ca3Fe2Si3O12+9SiO2=9Ca5FeSi6O13

Let us consider the case that the experi- +Fe3O4+4O2 (A5) ments performed by „K„…„‚‡V„p„{„€„r„p (1970, 3Ca5FeSi6O18+4Fe3O4+12SiO2

1971a) and by Liou (1974) indicate the =15CaFeSi2O6+O2 (A6) equilibrium boundary for Reaction (Al). These reactions are obtained by the linear

The calculated enthalpy and entropy combinations of Reactions (Al) and. (2) or changes of Reaction (Al) are 345.0 kcal and (3). In this case Reaction (4) is not stable. 209.5 cal.deg-1, respectively. When ferro- Fig. 7a shows the log fo2-1000/T rela bustamite is stable, besides the previously tions of equilibrium boundaries for Reac mentioned reactions the following five reac tions (Al) to (A6), which are calculated on tions occur in the system Ca-Fe-Si-0: the assumption that the equilibrium 4CaSiO3+CaFeSi2Oe=Ca5FeSi8O18 (A2) boundary for Reaction (Al) was demon 7CaSiO3+Ca3Fe2Si3Od2+2SiO2 strated by the experiments of „K„…„‚‡V„p„{„€„r„p =2Ca5FeSi8O18+O2 (A3) (1970, 1971a) and Liou (1974).

12Ca3Fe2Si3O12+9CaFeSi2O6 Fig. 7a is not sufficient in the Phase Relations in the System CaO-FeOx-SiO2 in H20-CO2 Mixtures 239

Fig. 7. (a upper) Calculated log fOe1000/T relations of equilibrium boundaries for Reactions (Al) to (A6) to contain ferrobustamite. The log foe 1000/T relations of Reactions (1) to (3) are shown in Table 3. Dashed lines indicate the equilibrium boundaries to be exlcuded from Fig. 3 because of the appearence of ferrobustamite. Dotted lines represent log fo2-1000/T relation of magnetite-hematite (MH) buffer, Ni-NiO (NNO) buffer, and fayalite-magnetite-quartz (FMQ) buffer (after Huebner, 1971). (b, lower) Schematic log fo2-1000/T relations of equilibrium boundaries for Reaction (Al) to (A6), in the case that ferrobustamite breaks down to the assemblage wollastonite-hedenbergite with the increase of temperature. quantitative sense. The experiments on carried out at temperatures from 550•Ž to Reaction (Al) used for the calculation were 700•Ž (Fig. 5), while the result obtained by 240 Tetsuya Shoji the calculation indicates that Reaction (Al) If the entahlpy changeof Reaction (Al) takes place between 740•Ž and 790•Ž (Fig. is smaller than the previously stated value 7a). That is,, the calculated result does in equilibrium state, the slopes of not satisfy the assumption which was taken equilibrium boundaries for Reactions (Al) to obtain it. This disagreement probably and (A3) are gentler than those shown in suggests that the experiments of I ypmalLosa Fig. 7a, and then Reaction (A2) takes place

(1970, 1971a) and Liou (1974) were not at a lower temperature. It is expected, in equilibrium because of the existence of however,that the slope of line (Al) is surely both of wollastonite and ferrobustamite. steeper than that of line (3). If the slope is Fig. 7a does not also agree with the mode gentler, ferrobustamite breaks down to the of occurrence of ferrobustamite. Accord assemblage wollastonite-hedenbergite with ing to Fig. 7a, at any oxygen fugacity the increasingtemperature, as shown in Fig. ferrobustamite is not stable below 740•Ž. 7b. This does not agree with the considera-

Ferrobustamite, however, is presumed to tions of Rutstein (1971), Shimazaki and be stable at the lower temperatures, because Yamanaka (1973) and Matsueda (1974), it is frequently found in skarn-type ore according to which ferrobustamite exists deposits, which are inferred to be formed stably between wollastonite and hedenber below the melting temperatures of granitic gite at high temperatures. rocks, and probably at temperatures of The schematic Xco,T diagrams (Fig. about 400•Ž or lower (Shoji, 1975, 8) of some equilibrium boundaries in the 1978). system CaO-FeO-Fe2O3-SiO2-CO,involving On the other hand, Fig. 7a is not ferrobustamite were drawn on the basis of contradictory to the experimental results Fig. 6d. At low temperatures, ferro of Rutstein (1971) and the field evidence bustamite breaks down to the assemblage of ferrobustamite in the qualitative sense. wollastonite-andradite-quartz or the as-

According to Rutstein (1971), hedenbergite semblagewollastonite-hedenbergite by Reac is replaced by andradite, wollastonite solid tion (A3) or (A2), respectively. Reactions solution (may be ferrobustamite) and quartz (A3)and (A2) take place respectivelyat the at values of oxygen fugacity higher than higher and loweroxygen fugacitiesthan that the NNO buffer. This result agrees with defined by the assemblage wollastonite- the fact that the equilibrium boundary for andradite-hedenbergite-quartz.Consequently,

Reaction (Al) passes between the MH two diagrams corresponding to the high buffer line and the NNO one in Fig. 7a. and low oxygen fugacities are illustrated in Any report does not describe ferrobustamite Fig. 8. coexisting with magnetite from skarn Compared with Fig. 7d, more six iso

(Matsueda, 1973a, b, 1974; Shimazaki and baric, fos defined invariant points appear Yamanaka, 1973; Shimazaki and Bunno, in Fig. 8. The mineral assemblages at the 1976; Matsuoka, 1976; Sato, 1978; Matsueda points are as follows: calcite-ferrobustamite- and Arai, 1978). This seems to correspond andradite-hedenbergite-quartzat B', calcite- to the fact that the assemblage ferro- wollastonite-ferrobustamite-andradite-ma bustamite-magnetite is stable at higher gnetite at C', calcite-wollastonite-ferro temperatures than the assemblage an- bustamite-andradite-quartz at P, calcite- dradite-hedenbergite (Fig. 7a). ferrobustamite-andradite-hedenbergite-mag- Phase Relations in the System CaO-FeOx-SiO2 in H2O-CO2 Mixtures 241

Fig. 8. Schematic analysis of phase relations in the system CaO-FeO-Fe2O3-Si02 - CO2 involving ferrobustamite as a functions of CO2 fugacity and temperature. The upper (a) and lower (b) diagrams correspond respectively to the higher and lower oxygen fugacities than that represented by the assemblage wollastonite- andradite-hedenbergite-quartz. Capital letters in circles show the isobaric, foe-defined invariant points. The mineral assemblages are as follows: calcite-ferrobustamite-andradite-hedenbergite-quartz at B', calcite-wollas- tonite-ferrobustamite-andradite-magnetite at C', calcite-wollastonite-ferrobus- tamite-andradite-quartz at P, calcite-f errobastamite-andradite-hedenbergite- magnetite at Q, calcite-ferrobustamite-hedenbergite-magnetite-quartz at R, and calcite-wollastonite-ferrobustamite-hedenbergite-quartz at S. Abbrevia tions are shown in Table I and Fig. 1. netite at Q, calcite-ferrobustamite- calcite -wollastonite -ferrobustamite -hedenber hedenbergite-magnetite-quartz at R, and gite-quartz at S (unstable at low oxygen 242 Tetsuya Shoji

fugacities, i.e. f o,

At high oxygen fugacities (Fig. 8a), how DISCUSSION ever, it is replaced by point B', where calcite, andradite, hedenbergite and quartz associat

At oxygen fugacities lower than the ed with ferrobustamite instead of wollas

value represented by the FMQ buffer, not tonite. Point C is also replaced by point C',

only fayalite but also ferrotremolite is where calcite, wollastonite, andradite and

stable (Ernst, 1966). Since these minerals magnetite coexist with ferrobustamite

are scarcely found in skarn, they do not be instead of hedenbergite. In Fig. 8, points

discussed in this paper. Accordingly, Figs. P, Q, R and S (only at low f o.) appear.

6b and d represent the phase relations at In Fig. 6, only four equilibrium bound

the value slightly higher than the oxygen aries, that is Reactions (5), (7), (8) and fugacity defined by the FMQ buffer. (10), exist between the wollastonite- Figs. 6 and 8 are significant only on andradite field and the calcite-magnetite- the condition where the oxygen fugacity in quartz one. On the contrary, nine bound- the system is controlled by the NNO or aries exist in Fig. 8, though a few of reac

FMQ buffer. Such condition seems to tions are different between Figs. 8a and b.

simulate well the formation of skarn for the The temperature of point B is about

following reasons: (i) It is expected that 610•Ž at the NNO buffer, and 470•Ž at the

the oxygen fugacity during the formation is FMQ buffer. Fig. 7 suggests that the

buffered by any of reactions, or that it is temperature of point B' is higher than that

controlled by the nature of H2O-CO, fluid. of point B. If ore-bearing skarns are

(ii) The curves of solid oxygen buffers are formed below 400•Ž (Shoji, 1975, 1978), the

approximately parallel to one another temperature of formation is inferred to be

(e.g. Fig. 7). (iii) Many of univariant as below point B or B'. On the contrary,

semblage involving skarn minerals, such as ferrobustamite occurring in mineralized

Reactions (2), (3) and (Al), show the f o, T skarn is frequently accompanied with Mg relation which is slightly steeper than the poor hedenbergite. Consequently, ferrobusta- buffer curves, but approximately parallel to mite-bearing skarn is inferred to have been

them (Fig. 7). (iv) The f o, T relation of H2O formed at the temperature above point B'

CO2 fluid with excess carbon, by which the or S. This disagreement depends mainly

fluid drops in oxygen fugacity, is ap upon the fact that the slope of equilibrium

proximately parallel to the buffer curves boundary for Reaction (Al) is too steep as (Shoji, 1978). Consequently, the following stated previously. Since the slope is

discussion is carried out for the system, whose inferred to be necessarily steeper than line oxygen fugacity changes parallel to any of (3) as mentioned previously, however, the the solid oxygen buffer curves. temperature of point B' is always higher

The difference between Figs. 6 and 8, than that of point A, which is the lower

which is caused by the existence of ferro- limit of the hedenbergite field. If the

bustamite, appears most distinctly at slope of line (Al) is parallel to that of line the isobaric, f o.-defined invariant points. (3), and the isobaric invariant point where Though point A is not drawn in Fig. 8, it line (2) intersects line (Al) in Fig. 6 Phase Relations in the System CaO-FeOx-SiO2 in H2O-CO2 Mixtures 243

scarcely move*, the temperature of point B' extraporated to low temperatures from this is calculated to be about 450•Ž at the FMQ result passes 0.07 mole % at 300•Ž . This buffer. If the limit of hedenbergite field is is about one figure lower than the estimated higher in f o, and low in temperature than value. The difference is larger than the the values determined by Gustafson (1974) , considerable amount of estimation error. it is expected that the assemblage ferro- It is not clear what causes this incoincidence. bustamite-hedenbergite is stable below The non-ideality of mixed fluid is one of the this tempeature. probable interpretations. The calcite-magnetite-quartz field In some skarns, clinopyroxene occurs contacts with the ferrobustamite one, the closer to limestone than garnet, and vice hedenbergite one, and the andradite one versa in other skarns. It is possible to from high to low temperatures (Figs. 6 explain the genesis of the normal zonal and 8). Accordingly, under the condition arrangement of garnet and clinopyroxene and where hedenbergite is stable, the ferrobusta the reverse one by the mixing of H2O and mite field is always included in the heden- CO, fluids having different,oxygen fugacities bergite one (Fig. 8). From Fig. 6, the (Shoji, 1976). According to Gustafson upper limit of CO, content of the condition (1974), andradite coexists stably with hed favorable to the formation of skarn in enbergite between the oxygen fugacities volving andradite and/or hedenbergite is approximately represented by the NNO inferred to be as follows: Xco2=30 mole % and IM buffers. Consequently, under such at 500•Ž and Xco2=0.8 mole % at 300•Ž condition, the formation of both minerals at pf=1000 bars and the NNO buffer (Fig. is independent on the oxygen fugacity. The

6a); Xco2=50 mole % at 500•Ž and Xco2 =0.7 mineral associating with calcite and quartz mole % at 300•Ž at pf=1000 bars and the changes from andradite to hedenbergite by

FMQ buffer (Fig. 6b); Xco2=15 mole % at Reaction (9) with the increase of CO2

500•Ž and Xco2=0.4 mole % at 300•Ž at fugacity. It is expected from this fact pf=2000 bars and the NNO buffer (Fig. 6c); that hedenbergite occurs in the limestone Xco2=25 mole % at 500•Ž and Xco2=0.4 side where carbon dioxide is generated by mole % at 300•Ž at pf=2000 bars and the the decomposition of calcite. The typical

FMQ buffer (Fig. 6d). Roughly speak example is found in the zoned skarn of the ing, the CO, content of HO-CO, fluid in Yaguki mine, Fukushima Prefecture. which ore-bearing skarn involving andradite According to Shimazaki (1969), the suc and/or hedenbergite was formed is less cession is as follows; slate, epidote skarn, than about 30 mole % at 500•Ž and 0.5 mole andradite skarn, clinopyroxene skarn,

% at 300•Ž. quartz-scheelite rock and limestone. The On the other hand, the experiments of andradite skarn consists chiefly of andradite

Taylor and Liou (1978) indicate that the with small amounts of caclite, quartz and CO2-richest limit of the andradite field is others. If zoned skarn was formed at T=

100 mole % at 650•Ž and 22 mole % at 300-400•Ž, pf=1000-2000 bars, and fo2=

550•Ž. The stability limit of andradite FMQ buffer, the CO2 content is inferred to

* Reaction (2) occurs only at the temperatures below the value represented by point J in Fig. 7a. Since Gustafson (1974) experimented on Reaction (2) up to 850°C, the temperature of point J is inferred not to change largely from point I. 244 Tetsuya Shoji have been 0.3-1 mole % at the place along relations of some univariant assemblages the boundary between the andradite and in the system Ca-Fe-Si-O involving ferro- hedenbergite skarns (Fig. 6). The similar bustamite. zoned skarn is observed in the Doshinkubo (3) When ferrobustamite is stable, the deposit, the Chichibu mine, Saitama Pre phase diagrams shown in Fig. 6 must be fecture, where the zones are characterized modified as shown in Fig. 8. by garnet, magnetite and hedenbergite from (4) Fig. 6 indicates that the CO2 con quartz diorite to marble (Shoji, 1975). tent of H20-CO2 fluid in which ore-bearing Each zone consists more or less garnet, skarn containing andradite and/or heden- clinopyroxene and calcite, but free from bergite was formed is less than about 30 quartz. The mineral assemblage indicates mole % at 500•Ž and 0.5 mole % at 300•Ž, that the zoned skarn was formed under the respectively. condition represented by the triangle (5) From this estimation, for example, restricted by lines 7, 9, 10 in Fig. 2. If this the zoned skarns of the Yaguki mine and the zoned skarn also was formed under the Chichibu mine are inferred to have been same condition assumed for the Yaguki formed in the fluids containing 0.3-1 mole mine, the CO2 content is inferred to have % CO2. been 0.3-1 mole % (Fig. 6). Both of the inferred values are equal to each other, ACKNOWLEDGEMENT because line 7 passes through the vicinities I wish to thank Professor S. Takenouchi of lines 9 and 10. The zoned skarns of both of the University of Tokyo for critical mines are inferred to have been under the reading of the manuscript and many similar condition, but by the different mechanism. In the Chichibu mine, the valuable suggestions. formation of zoned skarn was due to the different rate of migrating ions. On the APPENDIX contrary, the causes forming the zonal ar Using the thermochemical data com rangment in the Yaguki mine are not only piled by Robie and Waldbaum (1968), the the difference among migration rates, but log fo2-T relation of the Pb-PbO buffer has also the CO, content of skarn-forming fluid. been computed as follows:

CONCLUSION The phase relations among calcite, wollastonite, ferrobustamite, andradite, hedenbergite, magnetite and quartz have •¬ where fo2 is oxygen fugacity (bar), T is been considered on the basis of previous temperature (•‹K), and p is pressure (bar). work. The values calculated from the equation is nearly coincident with the values of (1) Table 3 and Fig. 6 shows the equilibrium boundaries for Reactions (1) Kyparraaosa (1971a). to (11) in H2O-CO, mixtures under fluid pressures of 1000 bars and 2000 bars. REFERENCES (2) Fig. 7 shows the schematic fo2-T Ernst, W. G. (1966), Synthesis and stability rela- Phase Relations in the System CaO-FeOx-SiO2 in H2O-CO2 Mixtures 245

tions of ferrotremolite. Amer . Jour. Sci., 264, Jour. Jap. Assoc. Mineral. Petrol. Econ. Geol., 37-65. 73, 185 (in Japanese). Greenwood, H. J. (1967), Wollastonite: stability in Shimazaki, H. (1969), Pyrometasomatic copper and H2O-CO2 mixtures and occurrence in a contact- iron ore deposits of the Yaguki mine metamorphic aureole near Salmo , Fuku , British shima Prefecture, Japan. Jour. Fac. Sci ., Univ. Columbia, Canada. Amer.. Mineral., 52, 1669- Tokyo, Sec. II, 17, Pt. 2, 317-350. 1680. -(1974), Characteristics of tungsten min Gustafson, Wm. I. (1974), The stability of andradite , eralization in Japanese skarn deposits. "Met- hedenbergite, and related minerals in the allization Associated with Acid Magmatism"

system Ca-Fe-Si-O-H. Jour. Petrol., 15, (ed. Stemprok, M.), 1, 312-315, Geol. Surv., 455-496. Prague. Harker, R. I. and Tuttle, O. F. (1956) , Experimental - and Bunno, M. (1976), Scheelite-bearing data on the Pcee T curve for the reaction: skarn containing "ferrobustamite" from the calcite+quartz=wollastonite+carbon dioxide . Kasugayama mine, Nagano Prefecture (abstr.).

Amer. Jour. Sci., 254, 239-256. Joint Meeting Soc. Mining Geol. Japan, Soc. Huebner, J. S. (1971), Buffering techniques for Mineral. Japan, Jap. Assoc. Mineral. Petrol. hydrostatic systems at elevated pressure "Resea . Econ. Geol., Preprint, 44 (in Japanese). rch Technique for High Pressure and - and Yamanaka, T. (1973), Iron-wollas High Temperature" (ed. Ulmer. G.C.), 123- tonite from skarns and its stability relation in 177, Springer-Verlag, New York. the CaSiO2-CaFeSi206 join. Geochem. Jour., Liou, J. G. (1974), Stability relations of andradite- 7, 67-79. quartz in the system Ca-Fe-Si-O-H. Amer. Shoji, T. (1975), Role of temperature and CO2

Mineral., 59, 1016-1025. pressure in the formation of skarn and its bear Matsueda, H. (1973a), Iron-wollastonite from the ing on mineralization. Econ. Geol., 70, 739- Sampo mine showing properties distinct from 749.

those of wollastonite. Mineral. Jour., 7, 180- -(1976a), Role of H20-C0,(-C) mixtures 201. in the formation of zonal arrangment of skarn -(1973b), On the mode of occurrence and minerals (abstr.). Joint Meeting Soc. Mining mineral paragnesis of iron-wollastonite skarn Geol. Japan, Soc. Mineral. Japan, Jap. Assoc. in the Sampo mine, Okayama Prefecture. Mineral. Petrol. Econ. Geol., Preprint, 43

Sci. Rept., Dept. Geol., Kyushu Univ., 11 (in Japanese).

265-273 (in Japanese with English abstract). -(1976b), The stability of the assemblage -(1974), Immiscibility gap in the system calcite-quartz in H20-CO2 mxitures. Jour. Jap. CaSiOs CaFeSi206 at low temperatures. Assoc. Mineral. Petrol. Econ. Geol., 71, 379-388.

Mineral. Jour., 7, 327-343. -(1977), The stability of andradite in H2 - and Arai, K. (1978), Mode of occurrence O-CO2 mxitures. Jour. Japan. Assoc. Mineral.

and compositional variation of "ferro- Petrol. Econ. Geol., 72, 399-411.

bustarnite" from the Muko-hi orebody, the -(1978), Skarn formation. "Geological Fujigatani mine, Yamaguchi Prefecture studies of the Mineral Deposits in Japan and (abstr.). Ann. Meeting of Mineral Soc. Japan, East Asia" (Imai, H.), 201-212, Univ. Tokyo Preprint, 27 (in Japanese). Press. Matsuoka, M. (1976), Iron-wollastonite from the Strunz, H. (1970), Mineralogische Tabellen (5 Auf.). Tsumo mine, Shimane Prefecture (abstr.). Akademische Verlagsgesellschaft Geest & Portig Joint Meeting Soc. Mining Geol. Japan, Soc. K. -G., Leipzig, pp. 621. Mineral. Japan, Jap. Assoc. Mineral. Petrol. Taylor, B. E. and Liou, J. G. (1978), The low- Econ. Geol., Preprint,45 (in Japanese). temperature stability of andradite in C-O-H Robie, R. A. and Waldbaum , D. R. (1969), Ther- fluids. Amer. Mineral., 63, 378-393. modynamic properties of minerals and related Yamanaka, T., Sadanaga, R and Takruchi, Y.

substances at 298.15°K (25°C) and one (1977), Structural variation in the ferrobusta- atmosphere (1.013 bars) pressure and at high mite solid solution. Amer. Mineral., 62, 1216- temperatures. U.S. Geol. Surv., Bull. 1259, 1224.

pp. 256. „K„…p‡V„{„€„r„p, „D.„D.(1970), „{„ƒ‡U„u‚•„y„}„u„~„„„p„t„Œ„~„€„u•@„y„•„…-„‰„u„y„u•@‚•„u„p„{‡U„y„y „s„u„t„u„~ƒÂ„u‚•„s„y„„+„r„€„t„t„p„ƒ„„„€„~„y„„=„p„~„t‚•„p„t„y„„•@„{„r„p‚•„ˆ•@"„O„‰„u‚•„{„‰ „U„‰„•„‰„{„€-„‡„‰m„‰ ?? „u„ƒ„{„€„‰ „P„u‚•‚•‚•‚Ž‚• ?? ‚•‚• "(‚•‚…„t•D„G„p„t„y„{„€„r•C‚a•D‚`•D„y „^ ?? ‚•„‘„~„€„r, ‚a. ‚g.), 2, 99-115, ‚g‚•‚™‚‹‚•, ‚l‚•‚ƒ‚‹„r„p.-(1971a), ‚a„t„y„‘„~„u‚•„q‡V„u„s‚• „t‚•„r„t‚…„~„y„‘ „~‚• „ˆ‚•„t„‘ Rutstein, M.S. (1971), Re-examination of the wollastonite-hedenbergite (CaSiOs CaFeSi2O6) equilibria. Amer. Mineral., 56, 2040-2052. Sato, K. (1978), Frrrobustamite from the Fujiga tani mine, Yamaguchi Prefecture (abstr.). 246 Tetsuya Shoji

„ƒ„y„~„„„u„„„y„‰„u„ƒ„{„€„s„€., „s„u„t„u„~„q„u‚•„s„y„„„p. „s„u„€‚˜ ‚ƒ„„‚•„q„y„t‚‚„~‚•‚ƒ„„„y „s‚…„t‚…„~„q‚…‚•„s„y„„‚• „y ‚•„~„t‚•‚•„t„y„„‚• „r ‚ƒ„y‚ƒ„„‚…„}‚… Ca-Fe-Si-O2. „C„u„€‚Ž ‚o‚™•Ý ‚l‚…‚ƒ‚•‚•‚•., 1971 (3), 1974, 444-453.

49-60. ‚l‚…„t‚‚„~„y„{ (1972), „_. „P. ‚s„u‚•„}„€„t„y„~„p„}„y„‰„u‚ƒ„{„y„u•@„ƒ„r„€„y„ƒ„„„r„p•@„ƒ„†‚•„„„•‚˜•@„s‚•„•„€„r „y „~„u„{„€„„„€‚•„~„u•@„€„ƒ„€„q„u„~„~„€„ƒ„„„y„}„u„s„p„}‚•‚•„†„y„‰‚…‚ƒ„{„y‚˜ ‚•‚…‚•„{„ˆ„y„y ‚ƒ ‚™„‰‚•‚ƒ„„„y‚…„} „r‚•„t„~ - (1971b), „P„€„t„u „…„ƒ„„„€„y„‰„y„r„€„ƒ„„„y „s„u„t„u„~„q„u‚•„s„y„„„p„~

a „t„y„p„s‚•„p„}„}„u Po2-T. „Ceox, 1971, 551-561. - „y ‚`„r„u„„„y„ƒ„‘„~ . (1974), ‚b„„„p„q„y„Œ„~„€„„„Œ „t„r‚™‚•„{„y‚ƒ„y ‚™„s„t‚…‚•‚•„t‚•.., 1972, „C‚…‚•‚˜ 654-662.

H20-CO2混 合 流 体 中 に お け るCaO一 欝eOx-SiO2系 の 相 関 係

正 路 徹 也

H20-CO2混 合 流 体}ζお け る方解 石,鉄 バ ス タ ム石(fembustamite),灰 バ ンザ ク ロ石,ヘ デン 輝 石,磁 鉄 鉱, 石 英 の 相 関 係 を,従 来 発表 され てい るCaO-SiO2-CO2系 お よ びCaO-FeOズS三 〇ゼ02系 の研 究 結 果 を もとに考 察 し た。灰 バ ン ザ ク ロ石 ま たは ヘ デ ン輝 石 を含 む ス カル ン の 生成 に適 し た条 件 のCO2量 の 上 限 は,次 の2っ の 反応, (6)灰 バ ン ザ ク ロ石+CO2=方 解 石+磁 鉄 鉱+石 英+O2 (8)ヘ デ ン輝石 十CO2十 〇2=方 解 石 十磁 鉄 鉱 十石英 に よ り 決 め ら れ る。 灰 パ ンザ ク ロ石 とヘ デ ン輝 石 の 安 定領 域 か ら,こ れ らの鉱 物 を含 む ス カル ン を生成 した H:20-CO2混 合 液体 中 のCO2量 は,5000Cで3(hnole%,300℃ で0.5mole%を 越 えな い と推 定 され る。 もし 鉄 バ ス タ 云石 が安 定 な場 合 に は,こ の 系 の 相 関係 は さ らに複 雑 とな る。 しか し,上 述 し た値 は その 安 定性 に無 関 係 で あ る。