Abstract-A Siliceous Sublimate, Collected from an Active Fumarole of Kuju Volcano, Was Analyzed for Major Constituents

Geochemical Journal. Vol. 2, pp. 1 to 9, 1968.

Silicon content of fumarolic gases and the formation of a siliceous sublimate

FUMIHIRO HONDA and YOSHIHIKO MIZUTANI

Department of Earth Sciences, Faculty of Science, Nagoya University, Chikusa, Nagoya, Japan

(Received October 10, 1967; in revised form February 26, 1968)

Abstract-A siliceous sublimate, collected from an active fumarole of Kuju Volcano, was analyzed for major constituents. The water soluble fraction of the sublimate contains large quantities of halogen acids and silicon, and the insoluble fraction consists of hydrated silica with a trace amount of native sulfur. In order to as certain the presence of gaseous silicon compounds in fumarolic gases, condensate samples were collected from high temperature fumaroles of Kuja and Nasudake Volcanoes with a silica-free sampling device. The silicon and fluorine contents of fumarolic gases range from 1 to 52 mg Si/kg H20 and 1 to 320 mg F/kg 1120, res pectively. Silicon is transported probably as fluoride in fumarolic gases, and the sublimate is formed by hydrolysis of silicon fluoride when the gas temperature falls to the boiling point of water.

INTRODUCTION Fluorosilicic compounds are believed to play an important role in the trans portation of silicon in volcanic gases. LOVERING (1957) studied -the alteration of volcanic ash by the reaction with halogen acid at Valley of Ten Thousand Smokes, Alaska, and suggested that loss or gain of silica on the altered ash possibly re flected the attack of hydrogen fluoride or the hydrolysis of silicon fluoride. NABOKO (1957) also demonstrated the effect of gaseous fluorine compounds on basaltic rocks at Klyuchevskoy Volcano, Kamchatka. YOSHIDA et al. (1966) found a siliceous vol canic sublimate, in an area of high temperature fumaroles at Satsuma-Iwojima Vol cano, Kyusyu, Japan, which contained very high concentrations of halogen acids, and suggested that the sublimate was formed by hydrolysis of silicon fluoride in the volcanic gases. In 1965, the present authors also found a siliceous sublimate in an area of high temperature fumaroles at Kuju Volcano, Kyusyu, Japan, which was similar in occurrence and composition to that reported by YOSHIDA et al. In order to demonstrate the presence of gaseous silicon compounds in fumarolic gases, con densate samples were collected in the Kuju fumarolic area and in the fumarolic area of Nasudake Volcano, Honshu, Japan, in 1966. This study is intended to de termine the chemical composition of the sublimate and the silicon and fluorine contents of the fumarolic gases, and to discuss the behavior of silicon fluoride in fumarolic gases. 2 F. HONDA and Y. MIZUTANI

OCCURRENCE OF THE SUBLIMATE In the Kuju fumarolic area, native sulfur is commercially collected by intro ducing high temperature fumarolic gases into artificial gas channels, where native sulfur is deposited as the temperature of the gases falls. A siliceous sublimate was found on the outer wall of a gas channel in the hottest fumarole-group (the highest temperature was nearly 400°C). The sublimate is white and forms a large loose crust; 20 to 30 cm in diameter and 1 to 2 cm thick. Under the sublimate, a weak emission of gases was found, and a temperature of about 80°C was recorded.

EXPERIMENTAL Chemical and X-ray diffraction analyses of the sublimate The sublimate was washed with a known volume of distilled water. The in soluble fraction was filtered and dried up to be subjected to X-ray diffraction analysis. An X-ray diffraction pattern was obtained by a Geiger-Flex X-ray dif fractometer, with nickel-filtered copper radiation. The soluble fraction was analyzed as follow: Si : Colorimetric determination with molybdate (Aluminum salt was added to prevent the interference by fluorine (TARUTANI, 1956) ). F : Thorium-alizarin titration. Cl : Colorimetric determination with mercuric thiocyanate. SW : Gravimetric determination as barium sulfate. NHs : Colorimetric determination with Nessler reagent. Br+I : Iodometry.

Collection of fumarolic gas samples Fumarolic gases were collected through a teflon tubing into a polyethylene cold trap containing 10 to 20 g of sodium carbonate powder, which prevents for mation of colloidal sulfur from the reaction of hydrogen sulfide with sulfur dioxide, and precipitation of silica in the condensate.

Chemical analysis of fumarolic gases The condensate was treated with hydrogen peroxide to oxidize the sulfur com pounds into sulfate, and then subjected to subsequent analyses: Si : Colorimetric determination with molybdate (Aluminum salt was added to prevent the interference by fluorine.). F : Colorimetric determination with Thorin after the separation by perchloric acid distillation. Cl: Colorimetric determination with mercuric thiocyanate and iron alum.

RESULTS AND DISCUSSION The chemical nature of the sublimate is similar to that of the sublimate from Satsuma-Iwojima Volcano reported by YOSHIDA et al. (1966) as shown in Table 1. Silicon content of fumarolic gases and the formation of a siliceous sublimate 3

Table 1. Chemical composition of siliceous sublimates

Siliceous sublimates Siliceous sublimates from from Kuju Volcano Satsuma-Iwozima Volcano*

Sample No. 1 2 4 5

SiO2, insoluble 28.2% 30.7% 2.9% 44.0% S, insoluble tr. tr. 0.28 7.1 F, insoluble 0.12 1.5 Si02, soluble 2.65 1.55 0.46 0.05 H2SiF6** 18.9 11.4 6.3 0.67 HC1 3.65 5.25 8.6 4.3 HI+HBr <10-s <10-s 0.015 0.006 H2S04 <0.05 <0.05 0.01 0.09 NH3 0.011 0.004 H2O 46.5 51.0 81.4 42.2

total 100.0 100.0 100.0 99.9

* Cited from YOSHIDA et al. (1966). ** Calculated allotting the soluble fluorine to this species.

0 10 20 30 40 50 2 0*

Fig. 1. X-ray diffraction pattern of a sublimate collected from Kuju-Volcano, Kyusyu, Japan. S indicates sulfur peaks.

These sublimates are characterized by a high halogen acid content. The X-ray diffraction pattern of the insoluble fraction of the sublimate is shown in Fig. 1. The insoluble fraction consists of amorphous silica with a trace amount of rhombic sulfur. The silica must have existed as hydrated silica in the sublimate. In Table 2 are given the silicon and halogen contents of the fumarolic gases from Kuju and Nasudake Volcanoes. Relatively high concentrations of silicon and fluorine are found in the gases from KH-1 series fumaroles of Kuju Volcano, where the siliceous sublimate was found. On the other hand, in fumarolic gases from Nasudake with rather low fluorine contents, low concentration of silicon was found. According to IWASAKI et al. (1966), high temperature fumaroles of Satsuma-Iwojima Volcano, where an abundance of siliceous sublimates were found by YOSHIDA et al., are characterized also by a high fluorine content of fumarolic gases. Fluorine content as high as 2,720 mg F/kg H20 was found in the fumarolic gases of Satsuma-Iwojima 4 F. HONDA and Y. MIZUTANI

Table 2. Silicon and halogen contents of fumarolic gases

Fumarole Temp. (*C) Si ppm(wt. ) F ppm(wt. ) Cl ppm(wt. )

Kuju Volcano KH-1 d-1 240 52 310 5180 KH-1 d-2 238 51 320 6150 KO-2 148 21 46 1620 KX-5 133 13 100 5660 KO-1 b 128 26 63 2270 KO-3 128 9.8 17 2920 KX-4 123 9.7 160 5060 KX-6 118 13 48 3150 KH-1 b 110 12 110 5670 KX-3 105 18 130 4580 KH-1 e 99 19 170 3780

Nasudake CH-1 308 5.5 15 170 M-4 b 267 9.5 8.5 480 UN-2 190 2.5 1.6 570 0-1 b 188 9.0 25 560 0-1 d 174 5.7 21 550 U-1 148 4.0 1.0 1240 CH-2 135 5.0 1.4 0-2 b 118 2.5 12 120 M-5 96 1.0 3.2

Volcano in 1958 by IWASAKI et al. These evidences suggest that high temperature fumarolic gases containing a large amount of fluorine may transport silicon, and that the silicon may be present mainly as fluoride in the fumarolic gases. Based on these characters of the sublimate and fumarolic gases, it can be assumed that, as pointed out by YOSHIDA et al., the siliceous sublimates were formed by the hydrolysis of silicon fluoride with liquid water:

3 SiF4 gas + n H20(11Q.)= Si02-(n 2)H20 + 2 H2SiF6faq.) (1)

The hydrochloric acid in the sublimates was derived probably by condensation of hydrogen chloride in fumarolic gases, and the native sulfur was formed by interaction of sulfur dioxide and hydrogen sulfide, air oxidation of hydrogen sulfide, or sublimation of sulfur vapor. From these considerations, thermodynamic calculations were made to examine the possibility that the silicon in the fumarolic gases is transported as gaseous silicon fluoride. The reactions considered are:

CaSiO3 + 6 HF = CaF2 + SiF4 + 3 H2O (2) MgSiO3 + 6 HF = MgF2 + SiF4 + 3 H2O (3) Na2SiO3 + 6 HF = 2 NaF + SiF4 + 3 H2O (4) Si02+4HF=SiF4+2 H2O (5) Silicon content of fumarolic gases and the formation of a siliceous sublimate 5

The equilibrium constants of these reactions, K, were calculated from the thermo dynamical data given by KUBASCHEWSKI and EVANS (1956) :

log K2= 0.345 1.651n T+ 1.06 x 10_3 T + 14850 T-'+ 1.06 x 104 T-2 (6) log K3= -6.99-0.6191nT+0.406x10_3T-+ 13950T-1-2.12x104T-2 (7) log K.1= 0.188 1.541n T+ 0.538 x 10_3 T+ 14270 T'-' + 4.92 x 104 Tie (8) log K5 = 1.073 0.635 In T + 0.146 x 10-3 T + 5735 T-1 2.33 x 104 T-2 (9) For reactions (2), (3), and (4), Kmay be expressed as follows by the first approxi mation assuming that the activities of components are unity : y •Xto K X64 HF• P22 (10)

and for reaction (5), it may be expressed as

30 15

o kl~ I 20 0 10

W 0 teaG

SiF4 Si02 10 5

0 100 200 300 400 500 600 700 0 100 200 300 400 500 600 700 Temp., ('C) Temp., CC)

Fig. 2 Fig. 3 Fig. 2. The relation between equilibrium constant and temperature in silicate-HF systems. ®i : CaSiOs-CaF2 boundary, ®: Na2SiO3-NaF boundary, and ®: MgSiO3-MgF2 boundary. •: Estimated from the observations. Fig. 3. The relation between equilibrium constant and temperature in a Si02-HF system. •: Estimated from the observations, 6 F. HONDA and Y. MIZUTANI

K = XsiFXH •X1120 (11) F•P where K is the equilibrium constant, X is the mole fraction of each gaseous com ponent in the system: XFI20+ XSiF,+ XHF =1, and P is the total pressure on the system. All the gases treated here were assumed to be ideal. The relations be tween the equilibrium constants and the temperature are shown in Figs. 2 and 3. By using these relationships, the behavior of silicon tetrafluoride in fumarolic gases may be examined. For the fumarolic gases of Kuju Volcano, the values of (XsiF,•XFI2o)/XHFand (XsiF,•XH2o) /XHF may be estimated from the result of chemical analyses assuming that the silicon in the condensate samples was present as silicon tetrafluoride in the fumarolic gases. Table 3 gives the values of XSIF,and XIIF, which were calcu lated from the data in Table 2, assuming that all the silicon found in condensate samples was present as silicon tetrafluoride and that fluorine in excess of silicon was present as hydrogen fluoride. As the proportion of fluorides to water vapor is very small, XH2O may be as sumed to be unity. A slight excess of silicon over fluorine found in a few samples

Table 3. Postulated silicon fluoride and hydrogen fluoride contents of fumarolic gases

Fumarole Temp. (°C) ppm(volSiF4 .) Excessppm (wt.) Si HF ppm ( vol. )* Kuju Volcano KH-1 d-1 240 33 0 160 KH-1 d-2 238 33 0 170 KO-2 148 11 4 0 KX-5 133 8.3 0 62 KO-1 b 128 15 3 0 KO-3 128 4.0 4.5 0 KX-4 123 6.2 0 130 KX-6 118 8.3 0 12 KH-1 b 110 7.7 0 73 KX-3 105 12 0 77 KH-1 e 99 12 0 110 Nasudake CH-1 308 3.5 0 0.10 M-4 b 267 2.0 6.4 0 UN-2 190 0.35 1.9 0 O-1 b 188 5.8 0 0.67 0-1 d 174 3.6 0 5.3 U-1 148 0.26 3.6 0 CH-2 135 0.33 4.5 0 0-2 b 118 1.6 0 4.9 M-5 96 0.64 0 0.48 * These values are practically the same as Xsiu , x 106 and XHF x 106, respectively. Silicon content of fumarolic gases and the formation of a siliceous sublimate may be due to contamination with fine rock particles picked up in the sampling procedure, or to the situation that some part of the silicon was transported as volatile compounds other than silicon tetrafluoride. For instance, according to F. G. SMITH (1958) and J. A. WOOD JR. (1958), silica can be transported by high density water vapor at high temperatures and pressures. Under the fumarolic conditions treated in this paper, it would be reasonable to consider that the silicon exists as silicon fluoride in fumarolic gases. Here, further discussion on other forms of silicon was excluded. The total pressure of the gases is assumed to be 1 atm, because at the openings of the gas channels, the static pressure of fumarolic gases is apparently close to the atmospheric pressure. When one plots the values of (XsIF,•XHzo)/(X;F•P') and (XsiF,•X,{zo)/(XP•P) with the observed gas temperatures in Figs. 2 and 3 respectively, it is seen that under the equilibrium condition with respect to silicates the fumarolic gases may have higher proportion of silicon tetrafluoride to hydrogen fluoride than postulated in Table 3, but under the equilibrium condition with respect to silica the fumarolic gases may have lower proportion. The effect of temperature and pressure on the equilibrium amount of silicon tetrafluoride in a CaSiOa-HF system is shown in Fig. 4, where the ratio of total fluorine to water vapor, (XHF+ 4 XsIF,)/XH,o, is assumed to be constant, 2 x 10-' (this is equivalent to 210 mg F/kg H20). This figure implies that temperature and pressure competitively affect the equilibrium amount of silicon tetrafluoride, and that silicon tetrafluoride in an amount of the order of 10-5 in XsiF, may not exist under a magmatic condition, i.e. 1,000°C and 1,000 atm. From

0 3 Q1 (3 X N k 2

0 100 200 300 400 500 600 700 Temp., (*C)

Fig. 4. The effects of temperature and pressure on the equilibrium amount of silicon tetrafluoride in a CaSiOa-HF system. (XHF-+ Xsir,)/XHzo is assumed to be 2 x 10-4 atm. ( P=1 atm., ®: P=10 atm., P= 100 atm., and 40: P=1,000 atm, 8 F. HONDAand Y. MIZUTANI

Fig. 4, we consider it possible that in fumarolic gases of Kuju Volcano, silicon tetrafluoride is formed mainly by the interaction of silicates and hydrogen fluoride during the effusion process of fumarolic gases with the increasing yield of silicon tetrafluoride at rather low temperatures. The presence of an appreciable amount of silicon fluoride in the fumarolic gases may be regarded as a result of a com bination of the following two competing processes: First, the formation of silicon tetrafluoride by the attack of hydrogen fluoride on wall rocks; and second, the hydrolysis of silicon tetrafluoride by water vapor and the precipitation of silica. The dominance of the formation of silicon fluoride over the hydrolysis would be reflected in the presence of excess silicon fluoride with respect to silica as shown in Fig. 3. If the dominance of the formation of silicon fluoride over the hydrolysis was maintained throughout the effusion process, the silicon fluoride content of fumarolic gases would have been increased with lowering temperature. However, there is no increase in the silicon fluoride content with lowering temperature, as shown in Table 3. This may be attributed to the increasing contribution of the hydrolysis of silicon fluoride with lowering temperature. In summary, it may be said that in the fumarolic gases an appreciable amount of silicon tetrafluoride is present which is possibly formed by the reaction of hydro gen fluoride with wall rock silicates during the effusion process of gases under rather low temperature and pressure conditions, and that the silicon tetrafluoride is subjected to hydrolysis to form siliceous sublimate with the liquid water which is formed by the condensation of water vapor i n the fumarolic gases when gas temperature falls to the boiling point of water.

ACKNOWLEDGEMENT

The authors are greatly indebted to Professor SHINYA OANA for his continuous encouragement and help throughout this work. They also wish to thank Dr. KEINOSUKE NAGASAWA, for X-ray diffraction analysis, and Dr. SADAO MATSUO, for useful discussion and criticism.

REFERENCES

IWASAKI, I., OZAWA, T., YOSHIDA, M., KATSURA, T., IWASAKI, B. and KAMADA, M. (1966) Differentiation of magmatic emanation. Bull. Tokyo Inst. Techn. 74, 1-57. KENNEDY, G. C. (1950) A portion of the system silica-water. Econ. Geol. 45, 629-653. KUBASCHEWSKI, 0. and EVANS, E. L. (1956) Metallurgical thermochemistry. John Wiley, New York. LOVERING, T. S. (1957) Halogen acid alteration of ash at fumarole No. 1, Valley of Ten Thousand Smokes, Alaska. Bull. Geol. Soc. Amer. 68, 1585-1604. NABOKO, S. I. (1957) A case of gaseous fluorine metasomatism at an active volcano. Geochem. 452-455. SMITH, F. G. (1958) Transport and deposition of the non-sulfide vein minerals. VI. Quartz. Canad. Mineral. 6, 210-221. Silicon content of fumarolic gases and the formation of a siliceous sublimate 9

TARUTANI, T. (1956) Colorimetric determination of silica in the presence of fluoride ion (Effect of masking agents for fluoride ion). J. Chem. Soc. Japan, Pure Chem. Sect. (Nippon Kagaku Zasshi) 77, 1292-1295 (in Japanese). WOOD, J. A. JR. (1958) The solubility of quartz in water at high temperatures and pressures. Am. J. Sci. 256, 40-47. YOSHIDA, M., OZAWA, T. and OSSAKA, J. (1966) A singular silica sublimate mineral found in Satuma-Iwozi.ma Volcano. Bull. Assoc. Petro. Miner. Econ. Geol. 55, 201-21.1, 262-271 (in Japanese).