FIRE and MUD Contents

Occurrence and Origin of Sulfide and Sulfate in the 1991 Mount Pinatubo Eruption Products

By Keiko Hattori1

1 Ottawa-Carleton Geoscience Centre, and Department of Geology, University of Ottawa, Ottawa, Ontario, K1N 6N5, Canada.


ABSTRACT

Sulfide phases are found in a wide variety of the 1991 eruption products from Mount Pinatubo. Sulfides in early-formed phenocrysts (olivine and augite) in basalt fragments and dome-forming andesite are globular nickel-bearing pyrrhotite, whereas some sulfides in the glass are irregularly shaped and copper rich. Sulfides in dacitic pumice are mostly copper-rich sulfides (chalcopyrite, CuFeS2, with or without an exsolution product of bornite, Cu5FeS4). These sulfides in dacitic pumice contain significant Zn (up to 1.3 weight percent), Se, Ag, As, and Cd. Sulfides in the glass of dacitic pumice exhibit desulfidation reaction rims.

Anhydrite is commonly surrounded by the matrix glass in gray, white, and banded pumice. Smooth contact between the anhydrite and glass confirms that the dacitic melt was in equilibrium with anhydrite immediately before eruption.

The occurrence of sulfide globules in the eruption products indicates that an immiscible sulfide liquid formed in silicate melts and that the melts were once reduced, with an oxygen fugacity below the redox boundary of dissolved sulfur. Later formation of anhydrite in dacitic melt requires an addition of sulfur and oxidation of the magma. It is proposed that supercritical fluid released from ascending mafic melt beneath Mount Pinatubo provided volatile elements and sulfur. Sulfur that discharged from the hot dry melt was mostly SO2. The reduction of sulfur to H2S in the cool (~800°C), wet dacite caused oxidation of this dacitic magma. H2S formed in this way was initially precipitated in the dacite as sulfide minerals together with other volatile elements. Continued influx of SO2 led oxidation of the dacite and an increase in the sulfur solubility of the melt, which caused partial resorption of sulfide minerals and led to excess sulfur, which was precipitated as anhydrite. The proposed model is consistent with compositions of iron-titanium oxides, abundant fluid inclusions in phenocrysts, high contents of volatile elements and hydrophyllic metals in sulfides, strontium isotopic compositions of anhydrite, and sulfur isotopic values of the bulk rocks.

Note to readers: Figures open in separate windows. To return to the text, close the figure's window or bring the text window to the front.

INTRODUCTION

The eruption of Mount Pinatubo in June, 1991, introduced ~20 Mt SO2 into the stratosphere and had a significant effect on the global climate (Bluth and others, 1992; Grant and others, 1992; Gleason and others, 1993). The sulfur-rich nature of the magma is reflected in the occurrence of anhydrite phenocrysts in dacitic pumices (Bernard and others, 1991). Several models have been proposed for the origin of this sulfur, including incorporation of subducted sulfide-sulfate deposits on the East China plate (Whitney, 1992), assimilation of sulfide deposits from the underlying Zambales Ophiolite Complex (Fournelle, 1991), assimilation of hydrothermal sulfate in preexisting volcanic rocks (McKibben and others, 1992; this volume), and the introduction of sulfur gases from underlying mafic magma (Pallister and others, 1992; Matthews and others, 1992). Discussions on the origin of sulfur are so far based mostly on chemical and isotopic data for bulk rocks. This paper describes the occurrence and composition of sulfur-bearing phases in various types of eruption products with an emphasis on sulfide minerals and proposes a model for the formation of high-sulfur dacitic magma at Pinatubo.

SAMPLE DESCRIPTION

The 1991 eruption products are dacitic pumice, dome-forming andesite, and basalt. The latter two are volumetrically insignificant, but they are important because they provide evidence supporting an injection of mafic melt as a trigger of the eruption of a semi-solidified dacitic magma (Pallister and others, 1992; Matthews and others, 1992). Basalt commonly occurs as angular to rounded inclusions in andesite that range in diameter from several centimeters to meters. Small inclusions of basalt (<5 cm) are rarely found in gray pumice.

Samples used for this study include six of white pumice; six of gray pumice, three of banded pumice, nine of dome andesite, and four of basalt enclosed in andesites. Three to 10 thin and thick polished sections were made from each sample. The usage of gray and white pumice follows the description by Pallister and others (1992); "gray" pumice actually varies from gray to tan. White pumice is well vesiculated, porphyritic, and contains coarse plagioclase and hornblende phenocrysts. Inclusions are rare, but angular fragments of gray pumice are observed. Gray pumice is poorly vesiculated and has phenocrysts that are much smaller than those of the white pumice (<3 mm). Gray pumices are typically heterogeneous in texture, containing different types of fragments including white pumice, andesite, basalt, and hydrothermally altered volcanic rocks. The altered volcanic fragments do not contain glass, show extensive biotitization and silicification, and appear to have been derived from old volcanic rocks. The banded pumice contains alternating layers of various thickness (1 cm to several centimeters wide) of white pumice and gray pumice; boundaries between bands may be sharp or diffused, and some show intricate mingling textures. Gray layers contain both olivine and hornblende, and they are interpreted to be gray pumice because of the presence of plagioclase phenocrysts and fine grain size of phenocrysts, similar to that of gray pumice. The lack of calcic rims and dusty zones in the plagioclase phenocrysts and the absence of plagioclase microlites argue against an alternative possibility--that the gray layers are andesite--because calcic rims, dusty zones, and microlites are common in the plagioclase population of the dome-forming andesite. Volcanic bombs are usually gray pumice, and one such bomb (~25 cm long) of gray pumice was also examined in this study.

Our samples include two white pumice and one gray pumice specimens collected shortly after the eruption in June 1991 and one specimen of 1991 andesite dome collected along the Maraunot River in the spring of 1992. The rest of the samples were collected in August 1992. Most andesites and basalts were collected along the Maraunot River, ~4 km west of the caldera wall, and one of the basalt samples was from a fragment more than 2 m in diameter, which contains angular andesite fragments. Most pumice samples were collected along the Sacobia River near the upper end of Clark Air Base. There are no apparent differences in mineralogy and textures of samples collected in 1991 and 1992. Anhydrite is well preserved in pumice and andesite samples collected in August 1992. Good preservation of anhydrite in samples collected in 1992 was unexpected, because heavy rainfall in the area could have leached anhydrite from the volcanic rocks, as observed at El Chichón (Luhr and others, 1984). Its preservation at Pinatubo may be attributed to short exposure time of samples at the surface as a result of the daily occurrences of lahars, which continuously brought buried eruption products to the surface.

The 1991 eruption also ejected older volcanic rocks. All samples, however, have been identified with confidence as products of the 1991 eruption because of their highly angular shape and large sizes and the lack of evidence of weathering and devitrification in the glass. Older eruption products are commonly semirounded, and they show various degrees of cloudiness of matrix glass due to devitrification and nucleation of fine dusts of hematite, and Fe-Ti oxide microphenocrysts usually show well-developed exsolution lamellae.

ANALYTICAL PROCEDURES

Major chemical compositions of sulfides were determined by use of a JEOL 6400 digital scanning electron microprobe (SEM), which has a 40° take-off angle for X-rays and is interfaced to a Link X-ray analyzer system (eXL LZ5). Analytical conditions were 20 kV accelerating potential, 39 mm distance between the specimen and the analyzer, 0.8 nA absorbed current on a Faraday cup, and a counting time of 140 to 200 s. Raw X-ray spectra were reduced to elemental concentrations by use of Link ZAF4-FLS analytical software. Analytical standards were natural pyrite for Fe and S, chalcopyrite for Cu, Ni metal for Ni, Co metal for Co, and synthetic ZnS for Zn. Analyses are believed to be accurate to+-2 percent of the amount present; detection limits are ~0.1 percent.

More than 200 grains of sulfides were subjected to probe analysis, but it was difficult to obtain satisfactory analytical results for all grains because of their small size (less than several micrometers in diameter) and interference from the host phases. The compositions of representative sulfides that are large enough to provide quantitative data are given in table 2.

Trace elements of sulfides in thick sections (~100 lc mum) were determined by a proton-induced X-ray emission microprobe (PIXE). The analytical technique for trace elements in sulfide minerals has been well established (Cabri and others, 1984; Campbell and others, 1989; Czamanske and others, 1992). Operating conditions are similar to those of earlier workers; 3 meV of proton energy, 45° take-off angle, ~8 nA of specimen current, 600 to 900 s counting time, and 4x5 m beam size. Aluminum absorbers 249 and 352 lc mum thick were used for pyrrhotite and copper-rich sulfides, respectively. A synthetic pyrrhotite standard (in wt%; Fe=60.93, S=38.87, Se=0.09, and Pd=0.11) was used for calibration. X-rays were detected by a Kevex silicon detector fitted with a beryllium proton shield of 50 lc mum thick, and data reduction of the raw X-ray spectra were performed by the GUPIX program of Maxwell and others (1989). Detection limits are three times the errors obtained from the background.

OCCURRENCE OF SULFUR PHASES

ANHYDRITE

The occurrence of anhydrite in dacitic pumice has been reported by Knittel and others (1991), Bernard and others (1991), Pallister and others (1991, 1992), and others. This study confirmed its common occurrence in gray and white pumice from Pinatubo. Banded pumice also contains anhydrite in both gray and white pumice layers. Anhydrite crystals are commonly coarse, <0.6 mm, and some grains enclose microphenocrysts of apatite and low-Ti (<1 wt% TiO2) magnetite (fig. 1A). In all pumice samples, the contact between the grains and glass is smooth and there is no evidence of reactions between the two phases (fig. 1A,B), confirming that anhydrite was a stable liquidus phase in the dacitic melt immediately before eruption. An anhydrite inclusion in hornblende was reported by Fournelle (1991), but none were observed during this study, so it is suggested that such occurrences are extremely rare.

Figure 1. Back-scattered electron images of anhydrite in pumice. A, Anhydrite enclosing low-titanium magnetite (Mag) in banded pumice. B, Contact between anhydrite (Anh) and glass in white pumice. Note smooth boundary between the two phases. White specks in glass are microphenocrysts.

Anhydrite grains up to 0.4 mm in length also occur in the dome andesite. These grains display reaction rims, several micrometers wide along the contact with glass, of calcite, apatite, calcic plagioclase, and other calc-silicates (fig. 2A,B). Quantitative determination of these phases was impossible because of their small sizes, <0.5 lc mum.

Figure 2. Back-scattered electron images of anhydrite (Anh) in dome andesite. A, The reaction rim, only 10 lc mum wide, is well exposed in the lower right because anhydrite was plucked out during the section preparation (sample CN1B-C). B, Boundary between anhydrite (Anh) and glass (Gl) in sample 92814-01B. Minute minerals between glass and anhydrite are apatite, calcite, calcic plagioclase and other calc-silicate minerals. Most of them are plucked out during the section preparation.

SULFIDES

All sulfide grains are fine grained and are found in all sections examined in this study, from a variety of eruption products. Occurrences are summarized in table 1. Grains enclosed in phenocrysts are usually small, rarely exceeding 40 lc mum, but one globular sulfide of cubanite composition (~120 lc mum) was found in unexsolved rhombohedral ilmenite-hematite solid solution (ilmenitess; fig. 3). Sulfides in the eruption products display four different shapes: spherical or globular sulfide enclosed in phenocrysts and glass (figs. 3, 4), symplectic minute droplets in phenocrysts (fig. 5), and irregular-shaped sulfides in glass (fig. 6). Globular sulfides and symplectic droplets in basalt and andesite are mostly Ni-pyrrhotite, whereas irregularly shaped sulfides in glass contain high copper and commonly show exsolution of bornite, Cu5FeS4. Most sulfides in dacitic pumice, independent of shapes and host phases, are Cu-rich sulfides. Several sulfide grains in phenocrysts occur together with glass inclusions (fig. 7).

Figure 3. Photomicrograph of a rounded cubanite grain (S; 120 lc mum long) enclosed in ilmenitess in dome andesite (sample 920814-10).

Figure 4. Back-scattered electron image of spherical grain of pyrrhotite (Po) in magnetitess (Mag) in dome andesite (sample CN1B-C-X).

Figure 5. Photomicrograph showing symplectic chains of Ni-bearing pyrrhotite (Po) droplets in hornblende in basalt. Olivine and hornblende commonly enclose chromite (Chr) (sample 920814-11A). Scale bar corresponds to 100 lc mum.

Figure 6. Back-scattered electron image of irregularly shaped sulfide grain containing exsolution product of bornite (bright part) in chalcopyrite (darker matrix) in the glass of dome andesite (sample CN1B-D-C).

Figure 7. Back-scattered electron image of sulfide inclusions (Sul) in hornblende (Hbl) in white pumice (sample PP1-92526-C). Dark area (Gl) surrounding sulfide contains K, Cl, Al, Si, and Fe, and it is interpreted to be a glass inclusion.

Table 1.--Summary of the occurrence of sulfides in the Pinatubo eruption products.

[Chr, chromite; Cpx, augite; Hbl, hornblende; Il, ilmenite-hematite solid solution; K-fs, K-feldspar; Mag, titanomagnetite (ulvöspinel-magnetite solid solution); Ol, olivine; Qtz, quartz. Most sulfides in the gray pumice were small and analyses were qualitative--adequate to identify the species (below) but not good enough for inclusion in table 2]


Host

Abundance of sulfides in the host

Shape of sulfides

Cu/Fe atomic ratio of bulk composition of S-phase1

Gray pumice

 

 

 

     Glass

very rare

angular

~1

     Hbl

minor

globular, <5 lc mum

~1

     Mag

less common

globular, <4 lc mum

~1

     Pl

rare

tabular, 3x8 lc mum

~1

White pumice

 

 

 

     Glass

minor

irregular, <40 lc mum with reaction rim.

>1

     Hbl, Mag.

minor

globular, <14 lc mum

~1, 0

     Qtz

rare

angular, <20 lc mum

>1

     Pl

minor

angular, <10 lc mum

>1, ~1

Banded pumice

 

 

 

     Glass

minor

irregular, <30 lc mum with reaction rim.

0, ~1

     Hbl

minor

globular, <15 lc mum

0

     Mag

minor

globular, angular

0, >1

     K-fs

minor

irregular

~1

Andesite

 

 

 

     Glass

minor

angular

~1, >1

     K-fs

minor

angular

~1

     Hbl

common

globular

0-1

     Il

minor

globular

0-1

     Mag

common

globular

0-1

     Cpx

not common

globular

0

     Ol

rare

globular

0

     Chr

none

 

 

Basalt

 

 

 

     Glass

common

globular

irregular, <40 lc mum

0

>1

     Hbl

common

globular

0-1

     Il

rare

globular

0-1

     Mag

rare

globular

~1

     Cpx

rare

globular

0

     Ol

rare

globular

0


1 Grains containing bornite exsolution obviously have Cu/Fe atomic ratios higher than 1. Grains containing cubanite exsolution have the ratio close to 1. The ratios of Fe-sulfides with no exsolution phases are shown as 0, although some contain several wt% Cu.

 

Table 2. Chemical compositions of representative sulfides in Pinatubo eruption products.

[Cpx, augite; Gl, glass; Hbl, hornblende; Il, ilmenite-hematite solid solution; K-fs, K-feldspar; Mag, ulvöspinel-magnetite solid solution; Ol, olivine; Ox, oxides; Pl, plagioclase]


Sample no.

Weight percent

 

Atomic percent

 

Host

S

Fe

Ni

Co

Cu

Sum

S

Fe

Ni

Co

Cu

Cu/Fe

Remarks

White Pumice

PP1-A1

Gl

34.83

33.26

0.00

0.00

31.41

99.50

 

49.92

27.37

0.00

0.00

22.72

0.83

 

PP1A-3

Gl

34.35

33.12

0.00

0.00

30.96

98.43

 

49.79

27.56

0.00

0.00

22.65

0.82

 

PP1-A,x

Gl

34.99

33.53

0.00

0.23

30.73

99.48

 

50.08

27.55

0.00

0.18

22.19

0.81

 

10-PP1A4

Gl

35.17

33.43

0.19

0.00

31.71

100.50

 

49.91

27.23

0.15

0.00

22.71

0.83

 

PP1A4

Hbl

36.30

56.83

0.00

0.25

0.65

94.03

 

52.32

47.02

0.00

0.20

0.47

0.91

 

32PP1-BE1

Gl

35.54

41.74

0.00

0.00

21.69

98.97

 

50.45

34.02

0.00

0.00

15.54

0.46

1.3% Zn

33PP1-BE2

Gl

34.44

40.56

0.00

0.00

19.87

94.76

 

50.83

34.37

0.00

0.00

14.80

0.43

1.0% Zn

PP2C-B

Pl

31.67

30.43

0.28

0.00

34.24

96.62

 

47.58

26.24

0.23

0.00

25.95

0.99

 

PP2D,B

Pl

31.49

29.98

0.00

0.00

33.81

95.28

 

47.89

26.16

0.00

0.00

25.95

0.99

 

4PP1-C,A

Gl

26.55

14.83

0.00

0.00

55.84

97.22

 

41.98

13.46

0.00

0.00

44.55

3.31

Cu-rich part of a grain

PP1-C,A

Gl

31.08

27.38

0.00

0.00

37.13

95.59

 

47.42

23.99

0.00

0.00

28.59

1.19

matrix of the grain

PP1-C

Hbl

35.09

39.47

0.43

0.00

22.50

97.49

 

50.55

32.65

0.00

0.00

16.36

0.50

 

PP1-AD1

Hbl

32.73

33.82

0.31

0.00

31.33

98.19

 

48.01

28.48

0.24

0.00

23.19

0.81

 

PP1-AD2

Hbl

32.72

33.74

0.00

0.00

30.85

97.31

 

48.36

28.63

0.00

0.00

23.01

0.80

 

PP1-AE

Mag

32.88

33.90

0.00

0.00

32.80

99.58

 

47.72

28.25

0.00

0.00

24.03

0.85

 

29PP1BD1

Hbl

33.93

35.56

0.00

0.00

31.74

101.23

 

48.22

29.02

0.00

0.00

22.76

0.78

 

30PP1BA1

Gl

34.30

29.92

0.00

0.00

33.82

98.04

 

50.04

25.06

0.00

0.00

24.90

0.99

 

31PP1BA2

Gl

31.98

29.26

0.00

0.00

33.34

94.58

 

48.75

25.61

0.00

0.00

25.65

1.00

 

Gray Pumice

511-812-4B

Gl

31.61

30.98

0.00

0.00

32.84

95.53

 

47.92

26.96

0.00

0.00

25.12

0.93

 

513-812-4C

Ox

31.97

35.14

0.00

0.00

30.76

97.87

 

47.25

29.81

0.00

0.00

29.81

0.77

 

Banded pumice

528-813-17C1

Gl

39.64

61.12

0.00

0.00

0.00

100.76

 

53.04

46.95

0.00

0.00

0.00

0.00

 

536-813-17A1

K-fs

34.42

30.42

0.00

0.00

34.52

99.36

 

49.67

25.20

0.00

0.00

25.13

1.00

 

549-813-17B1

Mag

32.74

33.17

0.44

0.00

32.05

98.41

 

48.00

27.93

0.35

0.00

23.72

0.85

 

529-813-17

Gl

35.68

59.31

0.00

0.00

0.65

95.64

 

50.93

48.60

0.00

0.00

0.47

0.01

 

530-813-17

Gl

34.10

36.38

0.00

0.00

26.94

97.43

 

49.72

30.45

0.00

0.00

19.82

0.65

 

Andesite

CN1-BD2

Il

27.13

60.04

0.47

0.00

0.75

88.39

 

43.60

55.39

0.42

0.00

0.59

0.01

 

CN1-BB1

Gl

36.18

51.48

1.49

0.00

6.19

95.34

 

51.93

42.44

1.17

0.00

4.48

0.11

 

CN1-BCX

Ox

35.04

55.25

0.29

0.00

3.79

94.37

 

50.91y

46.09

0.23

0.00

2.77

0.06

 

CN1-BB2

Gl

32.05

29.10

0.00

0.00

32.21

93.36

 

49.30

25.70

0.00

0.00

25.01

0.97

 

CN1B-A1

K-fs

34.45

30.35

0.00

0.00

34.48

99.28

 

49.73

25.15

0.00

0.00

25.11

0.75

 

401-814-1BB1

Ol/Hbl

39.65

54.34

4.34

0.00

0.00

98.33

 

54.16

42.61

3.23

0.00

0.00

0.00

beside chromite

402-814-1BB

Hbl

34.72

35.20

0.43

0.00

28.42

99.18

 

49.95

29.07

0.00

0.00

20.64

0.71

 

404-814-1BB2

Gl

29.03

25.33

0.00

0.00

39.63

94.00

 

45.66

22.89

0.00

0.00

31.46

1.38

center

405-814-1BB2

Gl

30.00

16.56

0.00

0.00

53.69

100.25

 

45.05

14.28

0.00

0.00

40.68

2.85

branch part

407-814-1BC

Hbl

33.38

27.83

0.00

0.00

37.27

98.47

 

48.97

23.44

0.00

0.00

27.59

1.18

 

409-814-1BA1

Mg

33.42

38.13

0.00

0.00

28.97

100.51

 

47.79

31.31

0.00

0.00

20.90

0.67

 

412-814-10

Ox

35.64

39.87

0.00

0.00

27.40

102.91

 

49.26

31.64

0.00

0.00

19.11

0.60

in fig. 3.

414-814-10

Mag

32.68

31.76

0.00

0.00

37.95

102.39

 

46.64

26.05

0.00

0.00

27.33

1.05

 

814-10

Mag

34.94

39.09

0.00

0.00

26.86

100.90

 

48.90

31.40

0.00

0.00

18.97

0.60

 

525-814-3D

Gl

34.26

35.69

0.00

0.00

27.95

97.90

 

49.76

29.75

0.00

0.00

20.49

0.69

branch part.

527-814-3B1

Hbl/Gl

34.27

30.69

0.00

0.00

33.97

98.93

 

49.65

25.52

0.00

0.00

24.83

0.97

 

Basalt fragment

417-814-11BA2

K-fs

39.44

59.69

0.46

0.00

0.74

100.33

 

53.06

46.10

0.34

0.00

0.50

0.01

 

418-814-11BA3

Gl

39.85

58.57

0.00

0.00

0.94

99.36

 

53.89

45.47

0.00

0.00

0.64

0.01

 

419-814-11CA

Hbl

39.29

60.37

1.20

0.00

0.00

100.86

 

52.67

46.46

0.88

0.00

0.00

0.00

 

420-814-11CB

Hbl

37.54

45.93

1.04

0.00

17.10

101.61

 

51.35

36.07

0.78

0.00

11.71

0.33

 

548-814-11A1

Hbl

37.23

57.08

1.62

0.00

2.52

98.45

 

51.59

45.42

1.23

0.00

1.76

0.04

beside chromite.

549-814-11AB

Cpx

38.93

58.02

1.18

0.00

0.00

98.13

 

53.41

45.71

0.88

0.00

0.00

0.00

 

437-814-2AA

Il

39.05

58.96

1.33

0.00

0.00

99.34

 

53.04

45.98

0.99

0.00

0.00

0.00

 

432-814-2AE

Mag/Gl

30.44

17.02

0.44

0.00

52.24

100.14

 

45.56

14.63

0.36

0.00

39.45

2.70

Cu-rich part.

431-814-2AE

Mag/Gl

29.54

33.92

0.00

0.00

33.63

97.09

 

44.77

29.52

0.00

0.00

25.72

0.87

Cu-poor part.

433-814-2AE

Mag/Gl

35.21

31.62

0.00

0.00

35.03

101.86

 

49.56

25.56

0.00

0.00

24.88

0.97

another part.

435-814-2AD1

Hbl

37.35

48.13

2.80

0.00

8.29

96.57

 

52.26

38.66

2.14

0.00

5.85

0.15

 

436-814-2AD2

Hbl

34.49

37.86

0.71

0.00

23.42

96.48

 

50.40

31.77

0.56

0.00

17.27

0.54

 

434-814-2AB

Ox

36.67

46.37

0.00

0.00

19.48

102.52

 

50.35

36.41

0.00

0.00

13.44

0.37

 


 

Basalt.--Sulfide grains are relatively common in basalt fragments in the dome andesite. Phenocrysts of olivine, clinopyroxene, hornblende, ilmenitess, and ulvöspinel-magnetite solid solution (magnetitess) contain sulfides; inclusions in olivine are extremely rare, but a few inclusions were observed. Most sulfides in phenocrysts are small (<10 lc mum), globular, Ni-bearing pyrrhotite with low copper (<0.1 wt%) (fig. 5). Spherical nickel-bearing pyrrhotite also occurs in the glass (fig. 8). These occurrences suggest that a sulfide liquid had the composition of Ni-Fe monosulfide solid solution. Copper-bearing pyrrhotite and an exsolution product of cubanite (CuFe2S3) occur in apparently late phenocryst phases, such as magnetitess and hornblende, which rims olivine. Copper-rich sulfides (atomic ratio of Cu/Fe >1) are not common in basalts and are only found in glass. Chalcopyrite is also found with a complex mixture of iron-rich orthopyroxene and magnetite that appears to replace Ni-pyrrhotite and olivine (fig. 9).

Figure 8. Reflected light photomicrograph of spherical grain of pyrrhotite (Po) in basalt glass (Gl) adjacent to hornblende (Hbl).

Figure 9. Reflected light photomicrograph of a symplectic mixture of very fine grained magnetite, orthopyroxene (opx) and chalcopyrite (Cp) replacing Ni-bearing pyrrhotite (Po). Quantitative determination of silicate phase (dark part in Po) was difficult because of the small sample size, but the X-ray spectrum suggests it is olivine. Sample (920814-11) is from a large basalt fragment in andesite.

Andesite.--The dome andesite contains sulfides and anhydrite. Sulfides are similar to those of basalt; these sulfides are usually small (<20 lc mum), but some are ~0.1 mm (fig. 3). Sulfides occur in olivine, augite, hornblende, and oxides (ilmenitess, magnetitess), K-feldspar (confirmed with electron microprobe), and glass. Sulfides in augite, hornblende, oxides are mostly globular Ni-bearing pyrrhotite. Some in hornblende rimming olivine and oxides are copper bearing. One grain of Ni-pyrrhotite in magnetitess is surrounded by perovskite, possibly a reaction product (fig. 10).

Figure 10. Back-scattered electron image of Ni-bearing pyrrhotite (Po) in magneitess (Mag). Dark phase between sulfide and oxides is perovskite (Per). Sample (CH1BE-W) is a dome andesite.

Several Ni-pyrrhotite grains in phenocrysts are replaced by copper-rich sulfide along cleavages and cracks (fig. 11). Sulfides in glass have angular to irregular shapes and they are all copper rich (chalcopyrite with or without exsolution of bornite) (fig. 12). It is common to find pyrrhotite only in phenocrysts and chalcopyrite only in glass (fig. 13). The occurrences and phases of sulfides demonstrate that early sulfides are Ni-pyrrhotite and late sulfides are Cu-sulfides.

Figure 11. Back-scattered electron image of globular Ni-bearing pyrrhotite (Po), partially replaced by Cu-rich sulfides (Cp) along the cracks of olivine rimmed by hornblende. Chromite (Chr) grains at the right. Note that the rim of pyrrhotite is partially corroded.

Figure 12. Back-scattered electron image of two kinds of sulfides. Sulfides enclosed in hornblende (Hbl) are rounded Ni-rich pyrrhotite (Po), and sulfides in glass (Gl) are irregularly shaped, Cu-rich sulfide (Cp). Fine grains of chromite (Chr) are common. Sample (CH1BB-A) is dome andesite.

Figure 13. Back-scattered electron image of Ni-bearing pyrrhotite (Po) in olivine (Ol) rimmed by hornblende (Hbl) and Cu-rich sulfides (Cp, Bo) in glass. Minute chromite grains (Chr) are visible in olivines. A closeup view of the Po-Cp grain is shown in figure 11. Scale bar is 200 lc mum.

Dacitic Pumices.--Most sulfide grains in white pumice are rich in copper (Cu/Fe atomic ratio >1) and a high copper phase is formed as an exsolution product (figs. 14, 15). The sulfides were originally formed as a high-temperature Cu-Fe-S intermediate solid solution. Due to their small sizes and fine lamellae width (figs. 14, 15, 16), it was impossible to determine the precise compositions of the Cu-rich phase, but it is likely bornite (Cu5FeS4), on the basis of the Cu-Fe-S phase relations and the degree of electron back-scattering from the phase. Pyrrhotite is very rare in pumice samples, but rounded grains of pyrrhotite were found within hornblende and magnetitess phenocrysts in gray pumice.

Figure 14. Back-scattered electron image showing typical occurrence of Cu-rich sulfides in white pumice. Sulfides in 2-mm-long plagioclase (Pl) phenocryst. The bright portion is the exsolution product of bornite; the dark portion is chalcopyrite (sample PKN 29-2).

Figure 15. Back-scattered electron image of sulfide in 3-mm-long plagioclase (Pl) phenocryst (sample PP1-C-A).

Figure 16. Back-scattered electron image of sulfides showing desulfidization rim around the grain in the glass of white pumice. The sulfide on the contact with magnetite (mag) and glass (Gl) displays exsolution lamellae of bornite (bright) from chalcopyrite (dark) (sample PKN29-2-A).

Sulfides in phenocrysts are usually globular to spherical (fig. 14) and many in plagioclase are angular (fig. 15). Sulfides in matrix glass generally have an angular shape (figs. 16, 17). Some are exceptionally high in copper, forming a bornite matrix with exsolved chalcopyrite inclusions. Several Cu-Fe sulfides in glass are high in zinc, up to 1.3 wt% (table 3). Ni-pyrrhotite grains are extremely rare but are found in gray pumice and in bands of gray pumice within banded pumice. Most sulfide grains enclosed in glass show reaction rims that are composed of a fine mixture of oxides and sulfides. The presence of reaction rims reflects desulfidation (figs. 16, 17), which indicates that sulfide was not stable in the dacitic magma immediately before eruption. The narrow width of rims and dispersed occurrences of remnant sulfides prevented their quantitative compositional analysis, but they appear to be fine iron oxides.

Figure 17. Back-scattered electron image of sulfide in white pumice with oxidized rim (sample PP1-A-A). Gl, glass; Hbl, hornblende.

Selenium contents of sulfides are high with Se/S ratios in a range of 20x10-5 to 100x10-5 (table 3). The ratios are slighter higher than the values of meteorites (~33x10-5; Wedephol, 1972), sulfides in mantle xenoliths (~30x10-5; Hattori and others, 1992), and sulfides in midocean ridge basalts (~20x10-5; Hamlyn and Keays, 1986; and 30x10-5 to 60x10-5; Hattori and others, 1992).

Sulfides in gray phenocryst-poor pumices are rare, but small grains (<1 lc mum) are found in phenocrysts of magnetitess, hornblende, and plagioclase. Sulfides are extremely rare in the glass of these pumices. Gray pumices with higher phenocryst contents (>10 vol%) contain more sulfide grains than do pumices of the phenocryst-poor variety.

Table 3. Trace element contents in sulfides in white pumice (in ppm).


Sample

Phase1

Host2

Zn

As

Se

Mo

Ag

Cd

Sb

Te

S/Se

Se/S x 10-5

Shape

White pumice

UK29-3

Cp

Gl

2,233

24

314

<6

36

20

<15

35

1,074

93.1

Angular

PP1BB

Cp

Gl

2,569

<8

299

<6

32

24

<16

<22

1,192

83.9

Angular

PP1BB

Cp

Gl

2,884

<9

266

8

10

14

<15

<23

1,244

80.4

Angular

PP1BA

CpB

Gl

212

<3

170

<4

242

42

<22

<34

2,031

49.2

Angular

PP1C,F3

Cp

Hbl

1,247

12

61

<5

21

<10

<17

74

2,741

36.5

Globular

PP1A1A3

CpB

Gl

2,209

8

276

44

91

22

<18

39

1,501

66.6

Angular

PPA1A3

CpB

Gl

2,184

22

333

149

173

16

36

43

1,146

87.3

Angular

Andesite

4-92814-10C

Cub

Il

919

18

26

<5

unc4

30

<25

<35

11,540

8.7

Globular, fig. 3

Basalt fragment

92814-11AK

CubP

Gl

998

unc4

107

<18

<28

<30

62

<60

2,388

41.9

Globular

92814-11AC

 

Hbl/Gl

536

20

104

<16

25

<35

<47

<70

2,921

34.2

Angular

92814-2BZ

CuPo

Gl

927

41

76

64

46

<35

62

<65

3,778

26.5

Irregular

92814-2BA

CubP

Gl

684

28

78

<16

46

<29

43

<48

4,125

24.2

Irregular

92814-11AD

Po

Hbl

148

13

57

<19

30

<35

<65

<73

5,897

17.0

Globular

92814-11AD

Po

Gl

453

unc4

20

26

<17

<14

<29

<36

17,485

5.7

Globular


1 Sulfide phase: CpB, a mixture of chalcopyrite and exsolution product of bornite; Cp, chalcopyrite; Cub, cubanite; CubP, cubanite-pyrrhotite; CuPo, Cu-bearing pyrrhotite; Po, pyrrhotite.

2 Phase containing sulfides: Gl, glass; Hbl, Hornblende; Il, unexsolved ilmenite-hematite solid solution.

3 The grains were too small and the total of Cu, Fe, and S are in a range of 50 to 90%. Trace element concentrations are accordingly normalized.

4 unc, the presence of the element is uncertain, and the measured concentration is between the detection limit and three times the measurement error.

 

DISCUSSION

SULFUR IN SILICATE MELTS

The solubility of sulfur in silicate melts is controlled by temperature, pressure, oxygen fugacity, and melt composition. The solubility of sulfur at constant temperature is higher in melts with higher iron; mafic melts have a higher solubility of sulfur than less mafic melts (Buchanan and Nolan, 1979). Sulfur in reduced melt is predominantly S2- with minor HS-, whereas sulfur in oxidized melt is SO42- (fig. 18; Katsura and Nagashima, 1974; Carroll and Rutherford, 1988). The redox boundary of dissolved sulfur species, S2- and SO42-, is one log fO2 unit above the Ni-NiO (NNO) buffer (Carroll and Rutherford, 1988). At constant temperature and pressure, the solubility of sulfur decreases steadily as fO2 increases up to the redox boundary, and then solubility increases sharply with a further increase in fO2 (fig. 18).

Figure 18. Sulfur solubility in dacitic melt at 900°C (modified after Carroll and Rutherford, 1988). Dotted line is the redox boundary of predominant sulfur species in the melt. Thick arrow shows the change of the Pinatubo dacitic melt. At stage I, an immiscible sulfide liquid forms and scavenges nickel and copper from the silicate melt. At stage II, SO2 that has been released from mafic melt is reduced to H2S as it oxidizes the dacitic melt. The added H2S forms sulfides in the dacitic melt that are high in Cu, Cd, Zn, and Se. At stage III, continuing addition of SO2 from mafic melt increases fO2 and results in partial desulfidation of previously formed sulfides and precipitation of anhydrite in the dacitic melt. ppm, parts per million; ppt, precipitate.

Sulfide liquid formed in mafic silicate melts is enriched in Ni, Co, and Cu because the partition coefficients for these elements between sulfide liquid and silicate melt are high (>250 for Ni and Cu and 50 for Co at 1200°C; Rajamani and Naldrett, 1978). Sulfide liquid that formed in evolved silicate melts tends to be low in nickel and cobalt and high in copper because of preferential incorporation of nickel and cobalt into early silicate minerals.

The evolution of sulfur in the magma may be evaluated from the occurrences of sulfides and anhydrite, assuming that the basalt samples represent mafic melt. Mafic magma was present beneath the dacitic magma chamber shortly before the 1991 eruption (Pallister and others, 1992; Matthews and others, 1992).

SULFUR IN MAFIC MELT AT PINATUBO

Occurrences of globular sulfide grains in phenocrysts and in the glass of basalt fragments suggest that they formed from sulfide liquid and that the mafic melt was saturated with sulfur. Common occurrence of sulfides in oxides is consistent with the experimental results of the solubility of sulfur. Formation of oxides, loss of iron from the melt, probably prompted the separation of sulfide liquid in the silicate melt.

The formation of liquid sulfide in silicate melt indicates that the predominant dissolved sulfur species in the melt was S2- and that the fO2 of the melt was lower than the redox boundary of sulfur species in the melt (fO2 = NNO +1; fig. 18; Nagashima and Katsura, 1973; Carroll and Rutherford, 1988). The mafic magma was not sufficiently oxidized to contain significant SO42-, so it could not form anhydrite. The lack of anhydrite in basalt is consistent with this interpretation.

The composition of globular sulfides, Ni-pyrrhotite, indicates that the sulfide liquid was Ni-Fe monosulfide solid solution, similar to most sulfides in many igneous rocks, including midoceanic ridge basalts (Mathez, 1976), oceanic island basalts (Desborough and others, 1968), back-arc basin basalts (Francis, 1990), island arc basalts and andesites (Heming and Carmichael, 1973; Ueda and Itaya, 1981), and continental arc rocks (Anderson, 1974; Whitney and Stormer, 1983).

The contents of copper in the Pinatubo monosulfide solid solution are low, <1 wt%. The low concentration of copper in iron sulfides in the Pinatubo eruption products suggests that the mafic melt was not originally high in copper, as the Cu/Fe ratios of sulfides reflect the ratios of silicate melt. This implies that high copper in later formed sulfides requires a mechanism for the enrichment of copper other than igneous processes.

The Cu/Ni ratio in melts increases during fractional crystallization. Formation of sulfide liquid containing 30 wt% Cu requires a melt with 0.15 wt% Cu, using the partition coefficient for copper between sulfide liquid and silicate melt (Rajamani and Naldrett, 1978). The concentration of copper would not increase to this high level in silicate melts by fractional crystallization alone. If fractional crystallization was the cause, the primary mafic melt should have had exceptionally high copper. As discussed, the possibility is rejected because early-formed sulfides are low in copper. Copper, therefore, must have been added to the mafic melt from an external source after crystallization of these phenocrysts.

The occurrence of highly irregular shaped Cu-rich sulfides in the glass (fig. 6) is very unusual for sulfides in volcanic rocks. The shape suggests that they were not formed as sulfide liquid in a silicate melt. Instead, they formed after the glass was already semisolidified. This interpretation implies that the formation of Cu-rich sulfides may have taken place after the entrapment of mafic melt into the andesitic or dacitic melts.

Fe-Ti oxide assemblages, however, failed to demonstrate reduced fO2 of the mafic melt. The estimated temperatures and fO2 using the solution model by Frost and Lindsley (1992) are similar to those of dacitic melt, ~800°C and two log units above NNO (table 4). The temperatures recorded in oxides are very low for temperatures of mafic magma. Obviously, oxides in basalt have undergone subsolidus equilibration after the entrapment of the magma in cooler andesitic and dacitic magmas. Subsolidus equilibration of oxide composition is known in many slowly cooled igneous rocks (Morse, 1980; Frost and Lindsley, 1991) and experiments have documented fast subsolidus equilibration (Hammond and Taylor, 1982). The once-reduced nature of the mafic melt is, however, evident from low abundances of oxides in basalt. Formation of oxides requires high Fe3+/Fe2+. It is also supported by higher Ti and lower Fe2+ of oxides enclosed in phenocrysts (table 4). Phenocryst phases surrounding the oxides prevented them from equilibrating with the matrix glass. Reduced nature of the melt at an earlier time is also supported by the occurrence of a symplectic mixture of magnetite and Fe-rich orthopyroxene, which appear to have replaced pyrrhotite and olivine (fig. 9). The assemblage is indicative of progressive subsolidus oxidation of olivine (Johnston and Stout, 1984).

Table 4. Temperature and fO2 estimates from Fe-Ti oxides.

[And, andesite; Bas, basalt; BP, banded pumice; GP, gray pumice; WP, white pumice; capdeltaQFM, difference between the fO2 and QFM buffer in log units (QFM = quartz-fayalite-magnetite). Contents of Fe2O3 and components of ulvöspinel (Usp) and ilmenite (Ilm) calculated using QUILF program by Andersen and others (1993)]


Sample

Rock

SiO2

Al2O3

TiO2

V2O3

FeO

Fe2O3

MnO

MgO

Usp

Ilm

Temp. (C)

capdeltaQFM

UK29AX

WP

0.00

1.73

4.38

0.48

32.59

52.80

0.43

0.91

0.127

 

825

+3.1

 

 

.15

.45

28.60

.34

24.32

43.51

.00

.79

 

0.557

 

 

UK29AY

WP

.09

1.86

4.68

.48

32.83

56.65

.43

1.09

.136

 

845

+3.0

 

 

.38

.40

28.82

.23

23.44

45.18

.37

1.17

 

.539

 

 

UK29AZ

WP

.22

1.80

4.58

.44

33.06

56.64

.41

.84

.133

 

826

+3.1

 

 

.15

.29

28.60

.19

23.82

43.82

.19

.96

 

.549

 

 

92812-3A

GP

.00

1.92

4.02

.56

32.81

58.16

.00

1.08

.121

 

817

+3.2

 

 

.00

.66

28.41

.00

23.99

45.08

.00

.88

 

.547

 

 

92813-01B

GP

.00

2.95

7.24

.25

33.60

50.97

.41

2.25

.215

 

931

+2.6

(in plagioclase)

 

.00

.75

29.08

.25

22.92

43.82

.28

1.65

 

.543

 

 

92813-17E

BP

.00

2.09

4.67

.35

33.42

59.91

.38

1.71

.129

 

821

+3.1

 

 

.00

.38

29.35

.42

24.43

46.32

.00

1.10

 

.542

 

 

92814-3C

And

.00

2.10

5.41

.26

33.06

56.32

.55

1.63

.153

 

856

+2.9

 

 

.00

.51

29.43

.00

24.11

44.57

.00

1.32

 

.550

 

 

92814-11AF

Bas

.22

1.71

4.00

.36

32.70

58.94

.44

1.00

.113

 

796

+3.2

 

 

.19

.32

31.36

.00

25.82

40.22

.49

1.06

 

.590

 

 

92814-11BY

Bas

.00

3.30

11.95

.33

37.63

42.17

.45

2.64

.357

 

973

+2.0

(in hornblende)

 

.00

.59

35.75

.00

27.11

33.19

.35

2.62

 

.648

 

 


 

SULFUR IN DACITIC MELT

Like the mafic melt, the dacitic melt also once had a low fO2 promoting the formation of sulfide liquid during phenocryst crystallization, as documented by globular sulfide inclusions (fig. 18). During that time, sulfur in the melt was predominantly S2-. Fe-Ti oxides again fail to reflect the reduced condition because of later reequilibration of their compositions. They only record the condition just prior to eruption, ~800°C and 3 log units above NNO (table 4). Under this condition, the predominant dissolved sulfur in the melt was SO42- (fig. 18) and anhydrite could have formed as a stable phase, as supported by the stable occurrence of anhydrite in the glass (fig. 1A,B).

ORIGIN OF SULFUR AND ANHYDRITE

The solubility of SO42- sharply increases with the increase in fO2 (fig. 18). Formation of anhydrite in dacitic melt requires higher concentration of sulfur in the melt than does the formation of sulfide. The relatively reduced nature of the original dacitic melt, indicated by the occurrence of globular sulfides, suggests that the formation of anhydrite in the Pinatubo dacite melt requires two conditions: the oxidation of the residual melt and an addition of sulfur. The addition of sulfur is essential because the silicate melt should have lost much of its sulfur during the transition from the reduced to oxidized state, as the solubility of sulfur in the silicate melt is minimal at the redox boundary (Nagashima and Katsura, 1973). The addition of external sulfur took place during and after crystallization of hornblende, plagioclase, and magnetite, but before the eruption. Most apatite apparently crystallized after oxidation of the residual dacitic melt because apatite in the dacite is high in SO3, up to 0.8 wt% (Imai and others, 1992; this volume).

Proposed sources of sulfur in high-sulfur magma include (1) subducted seafloor sulfide deposits (Whitney, 1984; 1992), (2) evaporite beds (Rye and others, 1984), (3) sulfide deposits in underlying ophiolite complexes (Fournelle, 1991), (4) sulfate minerals from preexisting volcanic rocks and hydrothermal deposits (McKibben and others, 1992), and (5) SO2 released from underlying mafic magma (Pallister and others, 1992; Hattori, 1993). High sulfur in the dacitic melt at source (model 1) is not applicable because this study shows that the dacitic magma acquired sulfur from an external source during and after crystallization of phenocrysts. In addition, solubility of sulfur as S2- is lower than the solubility of sulfur as SO42- (fig. 18). The content of sulfur in the melt remained low until the melt was oxidized.

Model 2, incorporation of evaporite, proposed for the El Chichón eruption product, is rejected because of the lack of evaporites in the Pinatubo area. There is a thin sequence of Eocene to Pliocene shallow marine sedimentary rocks between the volcanic rocks and Zambales Ophiolite Complex, and the Complex occurs below the sea level beneath the volcano (Delfin, 1984; Delfin and others, this volume). The shallow marine rocks may contain evaporite minerals, but this possibility is discounted for three reasons. First, the assimilation of an evaporite would lower Se/S because of its low Se/S ratios (<0.1x10-5; Measures and others, 1980). Second, evaporites generally contain low amounts of copper. Third, incorporation of young evaporites would raise 87Sr/86Sr ratios, given their high 87Sr/86Sr (~0.709). The ratios of anhydrite in pumices (0.70421+-0.00001; this study) and bulk pumice values (0.70422 to 0.70426; this study) are similar to other volcanic rocks in the Bataan arc (Knittel and others, 1992).

Model 3, incorporation of ophiolite-derived sulfide, is also discounted because it would not provide sulfur with high Se/S and an oxidizing agent for the magma. The two possible sources for additional sulfur that would accompany oxidation of dacitic magma are sulfate minerals from the old volcanic rocks (model 4) and sulfur released from underlying mafic magma (model 5). In either model, sulfur ultimately originated from underlying mafic magma. In the former model (model 4), sulfur was indirectly supplied from the mafic magma to the dacitic melt. This model is appealing, as intense hydrothermal activity had been noted at Pinatubo for a long time. Isotope studies and gas analyses have indicated that the hydrothermal activity was linked to the discharge of volatiles from magma at depth (Ruaya and others, 1992). The preexisting volcanic rocks were extensively altered to form secondary minerals, including anhydrite and sulfides (Delfin, 1984, Delfin and others, this volume; author's examination of Philippine National Oil Company drill chips in 1992). In addition, two porphyry copper deposits that are the product of magmatic-hydrothermal activity (the Dizon mine and Pisumpan deposit; Sillitoe and Gappe, 1984) are located within Mount Pinatubo, and they contain abundant high-temperature hydrothermal anhydrite. Minor porphyry copper deposits might have been present directly beneath the caldera. Assimilation of such rocks into the dacitic magma chamber could have enriched sulfur and oxidized the magma.

In the latter model (model 5), sulfur was directly added to the dacitic magma from the underlying mafic melt. The mafic melt at Pinatubo was saturated with sulfur, as evidenced by the occurrences of sulfide globules in phenocrysts. When mafic melt ascends, the vapor phase separates because the solubilities of CO2 and H2O decrease with pressure decrease (Holloway, 1976; Burnham and Ohmoto, 1980). Separation of any volatile gases from a melt will cause degassing of other gases, such as H2S and SO2, because they have mutual solubilities. For example, degassing of CO2 or H2O is accompanied by release of SO2 from magmas before the saturation of SO2 itself in the magma because SO2 will be partitioned to the CO2 phase from the silicate melt as CO2 exsolves. Sulfur released from magmas is mostly SO2 and H2S. The ratio of SO2/H2S of the fluids depends on temperature, fO2, and fH2O (fig. 19). That ratio in high-temperature fluids discharged from mafic melts is generally high, as is well documented from volcanic gas data (Gerlach, 1986). H2S may be predominant in cooler and more hydrated magmas, where SO2 will be converted to H2S (fig. 19).

Figure 19. Equal concentrations of SO2 and H2S at 200 atm (dashed curve) and 2 kbar (dashed-dotted curve). "Dry" and "wet" fluids have XH2O 0.01 and 1, respectively, where X is volume ratio. Thick curves are quartz-fayalite-magnetite (QFM) and hematite-magnetite (HM) buffers. Oxidation conditions of mafic melts are shown by the screened area (from Wallace and Carmichael, 1992). Note that fluids released from progressively "drier" and hotter melts have progressively higher SO2/H2S ratios.

Underlying mafic magmas at Pinatubo likely released SO2-rich supercritical fluids during their ascent. The fluids were likely absorbed by the overlying semisolidified dacitic magma. As dacitic magma is cooler and more hydrated than mafic magma, SO2 would be converted to H2S (fig. 19). The net effect is oxidation of the dacitic magma. The H2S produced in the dacitic magma would form sulfides in the dacite melt (stage II of fig. 18) together with elements transported by the fluids, such as Zn, Cd, Cu, and Se. Continual addition of SO2 would cause further oxidation of the dacitic magma, desulfidation of sulfides, and crystallization of anhydrite.

Sulfur isotopic data of the eruption products support both models. A variation in delta34S values for individual grains of anhydrite (+3 to +16permil ; McKibben and others, 1992; this volume) appears to support the incorporation of various types of sulfur from host rocks. Most delta34S values are in a range between +6 and +8 (Knittel and others, 1992; McKibben and others, 1992; this volume; Bernard and others, this volume). Sulfates formed by hydrolysis of SO2 have similar delta34S values (Hattori and Cameron, 1986). The isotopic compositions outside this range could be attributed to a fluctuation in delta34S of gases discharged from mafic magma, because delta34S values of sulfur gases released from melt vary, depending upon temperature, fO2, and fH2O of the melts (Ueda and Sakai, 1984). Apparent isotopic equilibrium between sulfide and anhydrite in the dacitic magma (McKibben and others, this volume) may be attributed to the isotopic equilibration of sulfur between H2S and SO2 in the gas.

At present, there is no evidence to reject the former model, but the latter model is favored because of abundant occurrences of fluid inclusions in the phenocryst rims of hornblende, plagioclase, and quartz in the dacitic pumice (fig. 20), and because of 87Sr/86Sr ratios of anhydrite. Phenocrysts contain numerous fluid inclusions, some of which are high in CO2. They most commonly align with growth bands of phenocrysts near rims. The occurrence supports that fluid percolation from mafic melt into dacitic melt started while phenocrysts were still being formed in the dacitic magma chamber.

Figure 20. Transmitted light photomicrograph of fluid inclusions in plagioclases (Pl) in dacitic pumice. The inclusions usually occur near the rims of phenocrysts, but some occur in the center of phenocrysts, as shown here.

As mentioned, 87Sr/86Sr ratios of anhydrite are identical to the bulk rocks (this study), whereas alteration products (epidote and anhydrite) have various87Sr/86Sr ratios, up to 0.708 (Hattori and others, 1992).

The percolation of supercritical fluid is also in accord with the occurrence of copper-rich sulfides and their trace elements in the glass. When fluid separation takes place in silicate melts, volatiles such as Se and Cl are enriched in the fluid phase (Greenland and Aruscavage, 1986; Candela and Holland, 1984). Metals that have affinity with these volatiles are also enriched in the fluid. Experimental studies document preferential incorporation of copper into a fluid (Candela and Holland, 1984; Urabe, 1985). Significant copper and zinc are also recorded in volcanic gases emanating from silicate magma (Mizutani, 1970). The enrichment of copper in the upper part of a magma chamber is attributed to vapor transport (Lowenstern and others, 1991). Other elements that have affinity with Cl, such as Cd, Zn, and Ag, are known to be incorporated into a fluid phase (Symonds and others, 1990).

ORIGIN OF SULFUR DISCHARGED TO THE STRATOSPHERE

Various models have been proposed for the origin of sulfur discharged to the stratosphere: (1) breakdown of anhydrite in the magma (Rutherford and Devine, 1991; Baker and Rutherford, 1992), (2) breakdown/absorption of sulfides in melts (Whitney, 1992), and (3) release of gas present in a separate vapor phase (Westrich and Gerlach, 1992). The lack of reaction rims around anhydrite in dacitic pumices argues against the first model. It is also difficult to envisage a conversion of sulfur from a solid phase to SO2 even at magmatic temperatures in the short period during the height of eruption. In addition, thin reaction rims of sulfides and preservation of most sulfides in the glass phase are not in accord with the second hypothesis. The occurrence of sulfur-bearing phases documented in this paper support the proposal by Westrich and Gerlach (1992) that the SO2 was most likely derived by release of sulfur present in a vapor phase in the melt.

SUMMARY

The presence of globular and irregularly shaped sulfides in all types of eruption products indicates that an immiscible sulfide liquid formed in the silicate melts during phenocryst crystallization, that sulfur in the melt at that time was predominantly S2-, and that fO2 was below the sulfide/sulfate redox boundary. Later formation of anhydrite in the dacitic melt required oxidation of the melt, as well as addition of sulfur from an external source. High concentrations of Cu, Zn, and Se in late-formed sulfides suggest addition of these elements together with sulfur. It is proposed that the dacitic melt incorporated a supercritical fluid from the underlying mafic magma. The fluid had high SO2/H2S and was enriched in volatile and chalcophile metals. Conversion of SO2 to H2S in cool, hydrated dacitic magma caused oxidation of the dacitic magma chamber. While fO2 was still below the redox boundary, Cu-rich sulfides formed in the melt. Once fO2 became higher than the redox boundary, incorporation of sulfur into the dacitic magma caused crystallization of anhydrite.

ACKNOWLEDGMENTS

The author thanks R.S. Punongbayan for providing the lodging and transportation for the author's field work, C.G. Newhall and his family for their hospitality at Clark Air Base and their assistance for sampling in the summer of 1992, Ulrich Knittel (Institut für Mineralogie und Lagerstättenlehre, Technische Hochschule Aachen) for providing samples, H.P. Ferrer (Philippine National Oil Company) for permission to examine thin sections and for provision of drill chips in the area, H.G. Aniceto-Villarosa for assisting the examination of drill chip samples, Peter Jones (Ottawa-Carleton Geoscience Centre) for his superb operation of the SEM laboratory, and J.L. Campbell and W.J. Teesdale (Guelph University) for assisting in PIXE analysis. Comments by E.M. Cameron (Geological Survey of Canada), J.A. Donaldson (Ottawa-Carleton Geoscience Centre), J.H. Fournelle (Johns Hopkins University/University of Wisconsin, Madison), Ulrich Knittel, J.F. Luhr (Smithsonian Institution), M.A. McKibben (University of California, Riverside), N. Metrich (Centre National de la Recherche Scientifique, Saclay, France), and C.G. Newhall (USGS) are greatly appreciated. The project was supported by a grant from Natural Science and Engineering Research Council of Canada.

REFERENCES CITED

Andersen, D.J., Lindsley, D.H., and Davidson, P.M., 1993, QUILF: A PASCAL program to assess equilibria among Fe-Mg-Ti oxides, pyroxenes, olivine and quartz: Computers in Geosciences, v. 19, p. 1333-1350.

Anderson, A.T., 1974, Chlorine, sulfur and water in magmas and oceans: Geological Society of America Bulletin, v. 85, p. 1485-1492.

Baker L., and Rutherford, M.J., 1992, Anhydrite breakdown as a possible source of excess Sulfur in the 1991 Mount Pinatubo eruption [abs.]: Eos, Transactions, American Geophysical Union, v. 73, no. 43, p. 625.

Bernard, A., DeMaiffe, D., Mattielli, N., and Punongbayan, R.S., 1991, Anhydrite-bearing pumice from Mount Pinatubo: Further evidence for the existence of sulfur-rich magmas: Nature, v. 354, p. 139-140.

Bernard, A., Knittel, U., Weber, B., Weis, D., Albrecht, A., Hattori, K., Klein, J., and Oles, D., this volume, Petrology and geochemistry of the 1991 eruption products of Mount Pinatubo.

Bluth, G.J.S., Doiron, S.D., Schnetzler, C.C., Krueger, A.J., and Walter, L.S., 1992, Global tracking of the SO2 clouds from the June 1991 Mount Pinatubo eruptions: Geophysical Research Letters, v. 19, p. 151-154.

Buchanan, D.L., and Nolan, J., 1979, Soluility of sulfur and sulfide immiscibility in synthetic tholeiitic melts and their relevance to Bushveld Complex rocks: Canadian Mineralogist, v. 17, p. 483-494.

Burnham, C.W., and Ohmoto, H., 1980, Late-state processes of felsic magmatism: Society of Mining Geologists of Japan Special Issue, v. 8, p. 1-11.

Cabri, L.J., Blanck, H., El Goresy, A., LaFlamme, J.H.G., Nobiling, R., Sizgoric, M.B., and Traxel, K., 1984, Quantitative trace-element analyses of sulfides from Sudbury and Stillwater by proton microprobe: Canadian Mineralogist, v. 22, p.521-542.

Campbell, J.L., Maxwell, J.A., Teesdale, W.J., Wang, J.X., and Cabri, L.J., 1989, Micro-PIXE as a complement to electron probe microanalysis in mineralogy: Nuclear Instruments and Methods in Physics Research, section B, v. 44, p. 347-356.

Candela, P.A., and Holland, H.D, 1984, The partitioning of copper and molybdenum between melts and aqueous fluids: Geochimica et Cosmochimica Acta, v. 48, p. 373-380.

Carroll, M.R., and Rutherford, M.J., 1988, Sulfur speciation in hydrous experimental glasses of varying oxidation state: Results from measured wavelength shifts of sulfur X-rays: American Mineralogist, v. 73, p. 845-849.

Czamanske, G.K., Kunilov, V.E., Zientek, M.L., Cabri, L.J., Likhachev, A.P., Calk, L.C., and Oscarson, R.L., 1992, A proton-microprobe study of magmatic sulfide ores from the Noril'sk-Talnakh district, Siberia: Canadian Mineralogist, v. 30, p. 249-287.

Delfin, F.G., Jr., 1984, Geology and geothermal potential of Mt. Pinatubo: Philippine National Oil Company, unpublished report, 36 p.

Delfin, F.G., Jr., Villarosa, H.G., Layugan, D.B., Clemente, V.C., Candelaria, M.R., Ruaya, J.R., this volume, Geothermal exploration of the pre-1991 Mount Pinatubo hydrothermal system.

Desborough, G.A., Anderson, A.T., and Wright, T.L., 1968, Mineralogy of sulfides from certain Hawaiian basalts: Economic Geology, v. 63, p. 636-644.

Fournelle, J.H., 1991, Anhydrite and sulfide in pumices from the 15 June 1991 eruption of Mount Pinatubo: Initial examination [abs.]: Eos, Transactions, American Geophysical Union, v. 72, p. 68.

Francis, R.D., 1990, Sulfide globules in mid-ocean ridge basalts (MORB) and the effect of oxygen abundance in Fe-S-O liquids on the ability of those liquids to partition metals from MORB and komatiite magmas: Chemical Geology, v. 85, p. 199-213.

Frost, B.R., and Lindsley, D.H., 1991, Occurrence of iron-titanium oxides in igneous rocks, in Lindsley, D.H., ed., Oxide minerals: Petrologic and magnetic significance: Mineralogical Society of America, Reviews in Mineralogy, v. 25, p. 433-468.

------1992, Equilibria among Fe-Ti oxides, pyroxenes olivine, and quartz, pt. II. Application: American Mineralogist, v. 77, p. 1004-1020.

Gerlach, T.M., 1986, Exsolution of H2O, CO2, and S during eruptive episodes at Kilauea volcano, Hawaii: Journal of Geophysical Research, v. 91, p. 12177-12185.

Gleason, J.F., Bhartia, P.K., Herman, J.R., McPeters, R., Newman, P., Stolarski, R.S., Flynn, L., Labow, G., Larko, D., Seftor, C., Wellemeyer, C., Komhyr, W.D., Miller, A.J., and Planet, W., 1993, Record low global ozone in 1992: Science, v. 260, p. 523-526.

Grant, W.B., Browell, E.V., Fishman, J., Brackett, V.G., Fenn, M.A., Butler, C.F., Myor, S.D., Nowicki, G.D., Veiga, R.E., Nganga, D., Minga, A., and Cros, B., 1992, Perturbation of tropical stratospheric ozone by Mt. Pinatubo aerosols-Experimental data [abs.]: Eos, Transactions, American Geophysical Union, v. 73, no. 14, p. 632.

Greenland, L.P., and Aruscavage, P., 1986, Volcanic emission of Se, Te and As from Kilauea volcano, Hawaii: Journal of Volcanology and Geothermal Research, v.27, p. 195-201.

Hamlyn, P.R., and Keays, R.R., 1986, Sulfur saturation and second-stage melts: application to the Bushveld platinum metal deposits: Economic Geology, v. 81, p. 1431-1445.

Hammond, P.A., and Taylor, L.A., 1982, The ilmenite/titanomagnetite assemblage: Kinetics of reequilibration: Earth and Planetary Sciences Letters, v. 61, p. 143-150.

Hattori, K., 1993, High-sulfur magma, a product of fluid discharge from underlying mafic magma: Evidence from Mount Pinatubo, Philippines: Geology, v. 21, p. 1083-1086.

Hattori, K., Arai, S., Francis, D.M., and Clarke, D.B., 1992, Variations of siderophile and chalcophile trace elements among mantle-derived sulfides detected by in-situ micro-PIXE analysis [abs.]: Eos, Transactions, American Geophysical Union, v. 73, no. 14, p. 344.

Hattori, K., and Cameron, E.M., 1986, Archaean magmatic sulphate: Nature, v. 319, p. 45-47.

Heming, R.F., and Carmichael, I.S.E., 1973, High-temperature pumice flows from the Rabaul caldera, Papua New Guinea: Contributions to Mineralogy and Petrology, v. 38, p. 1-20.

Holloway, J.R., 1976, Fluids in the evolution of granitic magmas: Consequences of finite CO2 solubility: Geological Society of America Bulletin, v. 87, p. 1513-1518.

Imai, A., Listanco, E.L., and Fujii, T., 1992, Petrographic and sulfur isotopic significance of highly oxidized and sulfur-rich magma of Mt. Pinatubo, Philippines [abs.]: Abstracts, International Scientific Conference on Mt. Pinatubo, Manila, 27-30 May 1992, p. 6-7.

------this volume, Highly oxidized and sulfur-rich dacitic magma of Mount Pinatubo: Implication for metallogenesis of porphyry copper mineralization in the Western Luzon arc.

Johnston, A.D., and Stout, J.H., 1984, Development of orthopyroxene-Fe/Mg ferrite symplectites by continuous olivine oxidation: Contributions to Mineralogy and Petrology, v. 88, p. 196-202.

Katsura, T., and Nagashima, S., 1974, Solubility of sulfur in some magmas at 1 atm: Geochimica et Cosmochimica Acta, v. 38, p. 517-531.

Knittel, U., Hattori, K., Hoefs, J., and Oles, D., 1992, Isotopic compositions of anhydrite phenocryst-bearing pumices from the 1991 eruption of Mount Pinatubo, Philippines [abs.]: Eos, Transactions, American Geophysical Union, v. 73, no. 14, p. 342.

Knittel, U., Oles, D., Forster, H., and Punongbayan, R.S., 1991, Pinatubo pumice contains anhydrite [abs.]: Eos, Transactions, American Geophysical Union, v. 72, p. 522.

Lowenstern, J.B., Mahood, G.A., Rivers, M.L., and Sutton, S.R., 1991, Evidence for extreme partitioning of copper into a magmatic vapor phase: Science, v. 252, p. 1405-1409.

Luhr, J.F., Carmichael, I.S.E., and Varekamp, J.C., 1984, The 1982 eruptions of El Chichón volcano, Chiapas, Mexico: Mineralogy and petrology of the anhydrite-bearing pumices: Journal of Volcanology and Geothermal Research, v. 23, p. 69-108.

Mathez, E.A., 1976, Sulfur solubility and magmatic sulfides in submarine basalt glass: Journal of Geophysical Research, v. 81, p. 4269-4276.

Matthews, S.J., Jones, A.P., and Bristow, C.S., 1992, A simple magma-mixing model for sulphur behavior in calc-alkaline volcanic rocks: Mineralogical evidence from Mount Pinatubo 1991 eruption: Journal of the Geological Society of London, v. 149, p. 863-866.

Maxwell, J.A., Campbell, J.L., and Teesdale, W.J., 1989, The Guelph PIXE software package: Nuclear Instruments and Methods in Physics Research, section B, v. 43, p. 218-230.

McKibben, M.A., Eldridge, C.S., and Reyes, A.G., 1992, Multiple origins of anhydrite in Mt. Pinatubo pumice [abs.]: Eos, Transactions, American Geophysical Union, v. 73, p. 633-634.

------this volume, Sulfur isotopic systematics of the June 1991 Mount Pinatubo eruptions: A SHRIMP ion microprobe study.

Measures, C.I., McDuff, R.E., and Edmond, J.M., 1980, Selenium redox chemistry at GEOSECS I re-occupation: Earth and Planetary Science Letters, v. 49, p. 102-108.

Mizutani, Y., 1970, Copper and zinc in fumarolic gases of Showashinzan volcano, Hokkaido: Geochemical Journal (Japan), v. 4, p. 87-91.

Morse, S.A., 1980, Kiplapait mineralogy, II: Fe-Ti oxide minerals and the activities of oxygen and silica: Journal of Petrology, v. 21, p. 685-719.

Nagashima, S., and Katsura, T., 1973, The solubility of sulfur in Na2O-SiO2 melts under various oxygen partial pressures at 1100C, 1250C and 1300C: Bulletin of Chemical Society of Japan, v. 46, p. 3099-3103.

Pallister, J.S., Hoblitt, R.P., and Reyes, A.G., 1992, A basalt trigger for the 1991 eruptions of Pinatubo volcano?: Nature, v. 356, p. 426-428.

Rajamani, V., and Naldrett, A.J, 1978, Partitioning of Fe, Co, Ni and Cu between sulfide liquid and basaltic melts and the composition of Ni-Cu sulfide deposits: Economic Geology, v. 73, p. 82-93.

Ruaya, J.R., Ramos, M.N., and Gonfiantini, R., 1992, Assessment of magmatic components of the fluids at Mt. Pinatubo volcanic-geothermal system, Philippines from chemical and isotopic data: Report of the Geological Survey of Japan, v. 279, p. 141-151.

Rutherford, M.J., and Devine, J.D., 1991, Pre-eruption conditions and volatiles in the 1991 Pinatubo magma [abs.]: Eos, Transactions, American Geophysical Union, v. 72, p. 62.

Rye, R.O., Luhr, J.F., and Wasserman, M.D., 1984, Sulfur and oxygen isotopic systematics of the 1982 eruptions of El Chichón volcano, Chiapas, Mexico: Journal of Volcanology and Geothermal Research, v. 23, p. 109-123.

Sillitoe, R.H., and Gappe, I.M., Jr., 1984, Philippine porphyry copper deposits: Geologic setting and characteristics: Committee for Coordination of Joint Prospecting for Mineral Resources in Asian Offshore Areas (CCOP) Project Office Technical Paper 14, 89 p.

Symonds, R.B., Rose, W.I., Gerlach, T.M., Briggs, P.H., and Harmon, R.S., 1990, Evaluation of gases, condensates, and SO2 emissions from Augustine volcano, Alaska: The degassing of a Cl-rich volcanic system: Bulletin of Volcanology, v. 52, p. 355-374.

Ueda, A., and Itaya, T., 1981, Microphenocrystic pyrrhotite from dacite rocks of Satsuma-Iwojima, Southwest Kyushu, Japan and the solubility of sulfur in dacite magma: Contributions to Mineralogy and Petrology, v. 78, p. 21-26.

Ueda, A., and Sakai, H., 1984, Sulfur isotope study of Quaternary volcanic rocks from the Japanese Islands Arc: Geochimica et Cosmochimica Acta, v. 48, p. 1837-1848.

Urabe, T., 1985, Aluminous granite as a source magma of hydrothermal ore deposits: An experimental study: Economic Geology, v. 80, p. 148-157.

Wallace, P., and Carmichael, I.S.E., 1992, Sulfur in basaltic magmas: Geochimica et Cosmochimica Acta, v. 56, p. 1863-1874.

Wedepohl, K.L., ed., Handbook of geochemistry, 1972, ch. 34, Selenium: New York, Springer-Verlag, v. II-3, 34 B-O.

Westrich, H.R., and Gerlach, T.M., 1992, Magmatic gas source for the stratospheric SO2 cloud from the June 15, 1991, eruption of Mount Pinatubo: Geology, v. 20, p. 867-870.

Whitney, J.A., 1984, Fugacities of sulfurous gases in pyrrhotite-bearing silicic magmas: American Mineralogist, v. 69, p. 69-78.

-----1992, Origin and evolution of sulfur in silicic magmas [abs.]: Eos, Transactions, American Geophysical Union, v. 73, no. 14, p. 366-367.

Whitney, J.A., and Stormer, J.C., Jr., 1983, Igneous sulfides in the Fish Canyon Tuff and the role of sulfur in calc-alkaline magmas: Geology, v. 11, p. 99-102.

FIRE and MUD Contents

PHIVOLCS | University of Washington Press | U.S.Geological Survey

This page is <https://pubs.usgs.gov/pinatubo/hattori/>
Contact: Chris Newhall
Last updated 06.10.99