- Email: [email protected]
Radon and water in volcanic gas at Surtsey, Iceland
Geochimieaet Coamochimica Acta, 1968,Vol. 32, pp. 815 to 821. PergamonPress. Printed in NorthernIreland
Radon and water in volcanic gas at Surtsey, Iceland SVEINBJ~RN Department
of
BJSRNSSON
Natural Heat, Nat,ional Energy Authority, Reykjavik, Iceland
(Received 2 January 1968; accepted in revised form 5 March 1968) Abstract-The
radon concentration in volcanic gas at the oceanic volcano Surtsey, off the southern coast of Iceland, was found to range from 120 to 170 pC/l of noncondensing gases. The release of gas from ascending magma is briefly discussed and the suggestion made that radon could be used as a tracer for calculating the amount of water released from the erupted magma. For this purpose, information on the distribution coefficient between magma and vapor bubbles for radon is required. This information might be obtained in laboratory experiments on the release of radon and water from molten lava. From the present data it may only be concluded that the basaltic magma erupted at Surtsey contained less than 0.75 wt,. ‘A of dissolved water.
the eruption of Surtsey, 1963 to 1967, many attempts were made to sample volcanic gases released from the basaltic melt. Samples for chemical analysis, for the measurement of D/H-ratio in hydrogen gas and water vapor, and for the measurement of radon were taken simultaneously, with the same sampling apparatus. The results of the chemical analysis and the deuterium measurements are reported in the accompanying papers by SIGVALDASON and EL&SON (1968) and ARNASON and SIGTJRGEIRSSON(1968). The present paper reports the results of radon measurements and discusses the possible use of radon as a natural tracer for estimating the water content of the erupted magma. DURING
SAMPLING OF THE VOLCANIC GAS A detailed descriptionof sampling locations and the apparatus and proceduresfor collecting the gas is given in the accompanying paper by SIGVALDASON and ELfssoN (1968). The gas samples used for radon measurement are listed in Table 1. The samples are numbered in accordance with the numbers of simultaneous samples taken for chemical analysis and deuterium measurements. MEASUREMENT
OF RADON
Radon is a radioactive inert gas. The most common isotope Rn222 (T,,, = 3%25 days) is the decay product of radium (Ra226, T,,, = 1622 yr). The other naturally occurring isotopes Rn220 (T,,, = 54.5 see) and Rn21s (T,,, = 3.9 set) are relatively shortlived and therefore not of interest for the present paper. Radon decays in a cascade of shortlived CL-and p-emitters, RaA, RaB, RaC and RaC’ into the relatively longlived /?-emitter RaD (Pb210, Tl12 = 22 yr). The radioactivity of radon in gas samples is conveniently measured by introducing the gas into an ionization chamber (EVANS, 1935; ISRAEL, 1934, 1937). The arrangement of the apparatus used for the present work is schematically shown in Fig. 1. The walls of the 4.I.-ionization chamber were maintained at + 900 V by a stable dry battery supply. The central electrode was mounted in two teflon insulators, which were separated by a grounded guard ring to prevent stray currents from reaching the electrode. The ionization current was measured by a vibrating reed electrometer (Victoreen model 475 A). The background current in the chamber was equivalent to a radon activity of 2 picocuries (1 pC = lo-l2 Curie = 3.7 . lop2 disintegrations per second). 815
Table 1. Results of radon measurements
Date of
sampling
Distance from crater
Sarqle So.
(ml
Nov .1‘)5 1964 Feb. 21 1965 Sept. 2 1966 March 31 1967
JO 0 6 0 150 150 20 20
in volcanic gas at Surtsey _^_~_._
Volume of Nitrogen -i_ Water non-con-
inert gases
densing gases (ml)
(Vol%)
Radon activity PC/l of non _ non-concondensing densing gases gases g/l of
“.-__..-._ ..-._ 2.8
11
“50
!#ff
20 23 24 26 28 29 30
189 206 208 166 193 383 472
0.5 O-5 05 58 35 10.0 11.4
I;! .j -5 0,; 136 120 119 30 34 155 90.4
4‘7 4.7 4.7 2.7 2,7 6.2 6.2
f 5 i_ & * & 5
-5% 50,’ 5% zoo/b’ 20% 5% 5 “/,’ -_ ._-.
6
N2 from -
gosflosk
Fig. 1. Schematic drawing of apparatus for measurement of radon in gas samples. 1. Ionization chamber; 2 Vacuum gauge; 3. Electrometer; 4. Drying agent (Mg(CIO,),); 5. Ascarite; 6. Drying agent (Mg(CIO,),); 7. Gas sampling tube. The amount of gas contained in the gas sampling tube was detested by adjusting the g:ts pressure to one atmosphere with the aid of a niveau-flask containing saline water. Tbe volumc~ of gas was then measured at this pressure and room temperature. Two Mg(ClO,), traps and one Ascarite-trap were used for drying the gas and remo~ ing practically all of the SO, and CO, before the sample was introduced into the chamber. After reading the baokground current the ionization chamber and the purifying traps werct evacuated to about 60 mm Hg pressure and the gas samples were sucked into the chamber through the drying and absorbing agents. The chamber was then slowly brought to atmospheric pressure by bubbling N, through the sampling tube for about 15 min. In this way the bubbk,s flushedoutmore than 97 per cent of the radon dissolved in the water in the tube (seeLucAs, 19634). In the first minutes after the radon was introduced into the chamber the ionization currents increased rapidly due to the growth of the shortlived decay products of radon. After three hours the radon had attained a radioactive equilibrium with its decay products and a reliable reading of the current could be obtained. The equipment was calibrated with the aid of a standard O-099 yC radium solution using the same method as above for transferring radon from the standard solution into t0he chamber.
Radon and water in volcanic gas at Surtsey, Iceland
817
RESULTS Radon The results of the radon measurements are shown in Table 1. The radon activity in pC/l of noncondensing gases is shown in column 7. The amount of condensed water accompanying each liter of noncondensing gases is listed in column 6. Based on interpretation of the deuterium measurements (ARNASON and SIGURGEIRSSON, 1968) this water has been released from the magma and is not evaporated sea water as one might suspect from an oceanic volcano. Column 5 shows volume percent of nitrogen and inert gases, which serves as an indicator of atmospheric contamination of the samples. According to SIGVALDASOX and EL~SSON (1968) the samples from February 21 1965 are practically free from atmospheric contamination. Nitrogen and inert gases constitute only 0.5 vol “/ of t’he noncondensing gases in these samples. In the last column of Table 1 the radon activity per liter of noncondensing gases has been corrected for atmospheric contamination, assuming that all nitrogen and inert gases in excess of 0.5 vol y. is of atmospheric origin. Radium In order to calculate the initial equilibrium concentration of radon in the magma prior to degassing an attempt was made to determine the concentration of its parent nuclide, radium, in a lava bomb which was ejected from the crater on February 21 1965, when the best gas samples were obtained. A 5-g split of the lava bomb was dissolved in HF and fused with Na,CO, as described by HU~GENS et aE. (1951) until the lava was completely dissolved. The radon generated in the solution was measured with the aid of the ionization chamber. The sensitivity of the method was, however, not sufficient and it could only be ascertained that the radium concentration was less than O-4 . lo-l2 g Ra/g of lava. A split of that lava bomb and another sample, taken from lava flowing on Sept. 8 1966 were then sent to Dr. K. S. Heier at the Australian National University, who kindly determined the radium concentration in these samples. Prof. Paul W. Gast, Lamont Geological Observatory, visited Surtsey in August 1966 and collected several samples for radium measurements. He also obtained a piece of the lava bomb mentioned above and has kindly allowed us to cite his unpublished results. All results of the radium measurements are summarized in Table 2. DISCUSSION OP THE RESULTS Gas samples taken close to the crater indicated a radon concentration of 120 to 170 pC/l of non-condensing gases corrected for atmospheric contamination, whereas samples 26 and 28 taken from a lava flow, which had been exposed at the surface through a distance of about 100 meters, contained only 50 to 70 pC/l. These samples were also different in chemical composition (SIGVALDASON and EL~SSON, 1968) and deuterium ratio (ARNASON and SIOURGEIRSSON, 1968) and probably represent an advanced state in the degassing of the lava. In Surtsey, most of the gas was seen to escape in the crater and during the first meters of flow of the lava. The gas bubbles bursting at the crater presumably contained most of the gas that had exsolved as bubbles during the ascent of the magma and they may be more
818
SVEINBJBRNBJ~~RNSSON Table 2. Radium in Surtsey lava Date of flow
(Su 1, not dated) (Su 2, not dated) Su 3, Feb. 21 1965
Radium lo-l2 g/g ____0.26 o-27 0.17
Author
I
CO.4
Feb. 21 1965
Feb. 21 1965
0.17 * 0.01
Sept. 8 1966
0.11 & 0.01 I
Method
P. W. CAST, 1967 S. BJ~RNSSON, 1966
K. S. HEIER, 1967
Ionization chamber. Radon generated in solution. Gamma-ra:, spectrometry 1.74; MeV Bismuth-214 pock
representative of the composition of the primary volcanic gas then the gas tkt wa,s retained longer in solution. The radium measurements indicate an activity of radium of 0.1 to 0.3 pC/g of lava. If no radon escapes from the magma, this activity of radium will generate an equal activity of radon within 30 days. We may therefore conclude that the initial equilibrium concentration of radon in the magma prior to degassing was of the order of 0.1 to O-3 PC/g of magma. In Table 3 the radon/water ratio, ,L?,of the gas samples and the initial radon concentration in the magma, R,, are tabulated from the data in Tables 1 and 2. The radon/water ratio is, in general, not affected by atmospheric contamination. Water added because of the burning of hydrogen is insignificant. As the radium concentration in the lava flowing on March 31 1967 has not been determined, the concentration in the flow on September 8 1966 was used for calculating the initiitl radon concentra.tion on March 31 1967. It is of considerable interest to estimate the quantity of radon and volcanic gita that is released per unit mass of magma. Knowing the amount of radon initi,zll,v dissolved and the radon/water ratio in the exsolved volcanic gas, we might then be able to calculate the water content of the original magma. This suggestion will now be considered further. Since water is the main component of volcanic g;ts, the following discussion is limited to the release of water and radon from basaltic magma, which becomes supersaturated with water as pressure decreases during the* ascent to the crater. Table 3. Radon/water ratio in the volcanic gas and initial radon concentration in the magma at Surtsey Distance from crater Date of sampling (m) Nov. 25 1964 Feb. 21 196.5
Sept. 2 1966 March 31 1967
30 0 0 0 150 20 20
Sample NO.
Radon/water B (PC/g)
Radon in magma R,, (PC/d
14
20 23 24 26 28 29 30
0.17 Il.li 12.6\‘1’g
0.11
(0.11)
Radon and water in volcanic gas at Surtsey, Iceland The release of water in ascending
819
magma
The release of water in ascending magma depends primarily on the solubility of water in the magma. According to experimental data obtained by HAMILTON et ccl. (1964), RUSSELL (1957) and KURKJIAN and RUSSELL (1958) the solubility of water in basalt melts is proportional to the square root of pressure. Extrapolation of the data of HAMILTON et al. (1964) down to lower pressure indicates a solubility of 0.11 wt.% in a basalt melt (50.71 wt.% SiO,, 4.68 wt.% MgO) at 1100°C and 1 atm pressure (BJ~RNSSON, 1966). This estimate is supported by the evidence on water retained in Hawaiian lavas. According to MACDONALD (1963) the average combined water content (H,O+) in 51 recent analyses of tholeiitic basalts of Kilauea and Mauna Loa of historic age, by several-different analysts, is 0.14 wt.%. In recent and more precise analyses of samples of pumice collected during the 1959-1960 eruption of Kilauea FRIEDMAN (1967) has found a water content ranging from 0.064 to 0.104 wt.%. We may thus assume that magma containing water dissolved in excess of about 0.1 wt. o/o will saturate during the ascent and the excess water will create vapor bubbles in the magma. At 1100°C and pressures below 100 atmospheres the specific volume of water vapor deviates less than 1% from that of an ideal gas (KENNEDY and HOLSER, 1966). The total volume of bubbles released from one gram of magma containing initially 0.1 wt.% of excess water will be 6.25 cm3 at 1100°C and 1 atm pressure. The volume of bubbles will be about 95 per cent of the total liquid-gas volume. This means that magma containing only 0.1 wt.% of excess water will be changed into a froth on the last tens of meters before it reaches the surface in the crater. This extensive frothing of the magma must greatly facilitate the release of other gaseous components into the vapor phase, although the process may be too rapid for an equilibrium exsolution to be attained. The release of radon from magma According to the radium measurements the radon concentration in the magma This concentration corresponds to is of the order of 0.1 to 0.3 PC/g of magma. several thousand atoms of radon per gram. In this extreme dilution the radon gas is a tracer far from saturation in the solution. Release of radon from the magma will thus not occur unless the magma becomes saturated with some other gas component, which then partly exsolves and creates gas bubbles in the magma. The amount of radon released into gas bubbles will depend on the volume percentage of bubbles and the distribution coefficient between magma and bubbles for radon. The value of the distribution coefficient is not presently known but it should be relatively easy to determine, if the necessary laboratory facilities are available. One possible method might be to melt a lava sample of known radium concentration and let it dissolve a known amount of water under high pressure. On reducing the pressure the melt would vesiculate and the coefficient could be found from the radon/water ratio in the vapor released. At present, further evaluation of the data must be based on an assumed distribution coefficient, and will therefore be of limited value. It may however be stated that if the release of radon into the vapor bubbles is an equilibrium process,
820
SVEINBJ~~RN BJ~~RNSSON
and the ratio of equilibrium volume concentrations of radon in the magma and vapor bubbles is near unity, about 95 per cent of the radon will be released into bubbles with the first 0.1 wt.% of vapor. Even if the equilibrium concentration of radon in the magma were ten times greater than in the vapor, about 65 per cent ot’ the radon would still be released with the first 0.1 wt.% of water. Calculated content of water dissolved in the magma Although the distribution coefhcient between magma and vapor bubbles for radon is unknown, the present data provides an upper limit of the amount of water dissolved in the magma erupted at Surtsey. If R, and W, denote the amount of radon and water released from a unit mass of magma, R, the amount of radon initially dissolved in that unit mass and p t,ho observed radon/water ratio in the volcanic gas, the following expression will 1~1 valid :
Thus, the upper limit of the amount of water released may be found by dividing the observed radon/water ratio into the amount of radon initially dissolved in t(he unit mass of magma. If this is done for the data in Table 3 we obtain an upper limit of O-64 wt.% of water released on February 21 1965 and 0.56 wt.‘!!,‘! OJI March 31 1967. Adding the amount of water retained in the magma, which in Hawaiian lavas was found to be 10.1 wt.q/,, we obtain an upper limit of dissolved If practically all radon is released water of 0.74 wt.% and 0.66 wt.% respectively. into the vapor bubbles the actual amount of water released will be close to the upper limit calculated from the radon/water ratio in that vapor. Further conclusions cannot be drawn from the available evidence, until the distribution coefficient for radon is determined. It is of interest to compare the results above to the findings of MOORE (1965). In deep-sea tholeiitic pillow lavas (48.9 wt.% SiO, and 13.2 wt.y” MgO) dredged from the submarine part of the east rift zone of Kilauea, Hawaii, he found an average water content of 0.45 f 0.15 wt.% H,O+. The water content W;I,S roughly constant between 500 and 5000 meters depth and was regarded to represent the amount of water dissolved in the erupting magma. In conclusion, the measurement of radon in volcanic gas at erupting craters seems to offer a method for estimating the amount of water released from the erupted magma. To apply the method rigorously information on the distribution coefficient between magma and vapor bubbles for radon is required. It is hoped that this coefficient can be obtained in laboratory experiments. dcknowledgentents-I am indebted to Dr. I. FRIED~~AN and Dr. A. R. MCBIRNEY for their valuable criticism of the manuscript. Special thanks are due to Dr. K. S. HEIER, Australian National University for determining the radium content of the lava samples and to Prof. P. W. GAST, Lamont Geological Observatory for allowing citation of his unpublished results on radium. The valuable assistance of the Surtsey Research Society and the Icelandic Coast Guard in transporting us to Surtsey is gratefully acknowledged. A part of this work was supported by the Icelandic Science Foundation.
Radon and water in volcanic gas at Surtsey, Iceland
821
REFERENCES AR,NASONB. and SIG~RGE~SSONTh. (1968) Deu~ri~ content of water vapour and hy~ogen in volcanic gas at Surtsey. Geo&im. ~og~oc~~~. Acta 32, 807-813. BJ~RWSSONS. (1966) Radon in magmatie gas at Surtsey and its possible use for determining the content of water in the magma. Surtsey Research Progress Report II, pp. 97-110. The Surtsey Research Society, Reykjavik, Iceland. EVANS R. D. (1935) Apparatus for the determination of minute quantities of radium, radon and thoron in solids, liquids and gases. Rev. Sci. ~n~t~rn. 6, 99-112. PRIEDBL~NI. (1967) ?Vater and deuterium in pumice from the 1959-60 eruption of Kilauea volcano, Hawaii. U.S. G’eol. Survey Prof. Paper 575-B, B120-8127. HAMILTOND. L., BURNHAMC. W. and OSBORNE. I?. (1964) The solubility of w&or and effects of oxygen fug&city and water content on crystallization in m&c magmas. J. Petrol. 5,21-39. HUDGENS J. E., BENZING R. O., CALI J. P., MEYER R. C. and NELSON L. C. (1951) Determination of radium or radon in gases, liquids or solids. Nz&eonics 9, 14-21. ISRAEL H. (1934) Zur Methodik der klimatologischen Emanationsmessungen. I. iiber ein neues Emanometer. Der B~~eo~oge I, 318-327. ISRAEL H. (1937) Zur Methodik der klimatologischen Emanationsmessungen. II. uber den SBtt’igungsstrom bei emanometrischen Messungen. Gerlands Beitr. Geophys. 51, 35-49. REhNN:DY G. C. and HOLSER FV. T. (1966) Pressure-volume-temperature and phase relations of water and carbon dioxide. In Handbook of Physical Constants (editor S. P. Clark, Jr.,). Geol. Sot. Amer. &Fern.97, 371-384. KURICJIAN C. R. and RUSSELL L. E. (1958) Solubility of water in molten alkali silicates. J. Sot. Gh,sq. Tech. 42, 130-144. LUCAS H. I?. (1964) A fast and accurate survey technique for both radon-222 and radium-226. In The Natural Radiation .&+vironment (editors J. A. S. Adams and W. M. Lowder), pp. 315-331. The ‘University of Chicago Press. MACDONALDG. A. (1963) Physical properties of erupting Hawaiian magmas. Geol. Sot. Amer. Bull. 74, 1071-1078. MOORNJ. G. (1965) Petrology of deep-sea basalt near Hawaii. Amer. J. Sci. 263, 40-52. RUSSELL L. E. (1957) Solubility of water in mohen glass. J. Sot. Glass. Tech. 41,304-317. SIGVALDASONG. E. and ELfSSON G. (1968) Collection and analyses of volcanic gas at Surtsey. Geochim. Cosmochim. Acta 32, 797-805.
Radon and water in volcanic gas at Surtsey, Iceland SVEINBJ~RN Department
of
BJSRNSSON
Natural Heat, Nat,ional Energy Authority, Reykjavik, Iceland
(Received 2 January 1968; accepted in revised form 5 March 1968) Abstract-The
radon concentration in volcanic gas at the oceanic volcano Surtsey, off the southern coast of Iceland, was found to range from 120 to 170 pC/l of noncondensing gases. The release of gas from ascending magma is briefly discussed and the suggestion made that radon could be used as a tracer for calculating the amount of water released from the erupted magma. For this purpose, information on the distribution coefficient between magma and vapor bubbles for radon is required. This information might be obtained in laboratory experiments on the release of radon and water from molten lava. From the present data it may only be concluded that the basaltic magma erupted at Surtsey contained less than 0.75 wt,. ‘A of dissolved water.
the eruption of Surtsey, 1963 to 1967, many attempts were made to sample volcanic gases released from the basaltic melt. Samples for chemical analysis, for the measurement of D/H-ratio in hydrogen gas and water vapor, and for the measurement of radon were taken simultaneously, with the same sampling apparatus. The results of the chemical analysis and the deuterium measurements are reported in the accompanying papers by SIGVALDASON and EL&SON (1968) and ARNASON and SIGTJRGEIRSSON(1968). The present paper reports the results of radon measurements and discusses the possible use of radon as a natural tracer for estimating the water content of the erupted magma. DURING
SAMPLING OF THE VOLCANIC GAS A detailed descriptionof sampling locations and the apparatus and proceduresfor collecting the gas is given in the accompanying paper by SIGVALDASON and ELfssoN (1968). The gas samples used for radon measurement are listed in Table 1. The samples are numbered in accordance with the numbers of simultaneous samples taken for chemical analysis and deuterium measurements. MEASUREMENT
OF RADON
Radon is a radioactive inert gas. The most common isotope Rn222 (T,,, = 3%25 days) is the decay product of radium (Ra226, T,,, = 1622 yr). The other naturally occurring isotopes Rn220 (T,,, = 54.5 see) and Rn21s (T,,, = 3.9 set) are relatively shortlived and therefore not of interest for the present paper. Radon decays in a cascade of shortlived CL-and p-emitters, RaA, RaB, RaC and RaC’ into the relatively longlived /?-emitter RaD (Pb210, Tl12 = 22 yr). The radioactivity of radon in gas samples is conveniently measured by introducing the gas into an ionization chamber (EVANS, 1935; ISRAEL, 1934, 1937). The arrangement of the apparatus used for the present work is schematically shown in Fig. 1. The walls of the 4.I.-ionization chamber were maintained at + 900 V by a stable dry battery supply. The central electrode was mounted in two teflon insulators, which were separated by a grounded guard ring to prevent stray currents from reaching the electrode. The ionization current was measured by a vibrating reed electrometer (Victoreen model 475 A). The background current in the chamber was equivalent to a radon activity of 2 picocuries (1 pC = lo-l2 Curie = 3.7 . lop2 disintegrations per second). 815
Table 1. Results of radon measurements
Date of
sampling
Distance from crater
Sarqle So.
(ml
Nov .1‘)5 1964 Feb. 21 1965 Sept. 2 1966 March 31 1967
JO 0 6 0 150 150 20 20
in volcanic gas at Surtsey _^_~_._
Volume of Nitrogen -i_ Water non-con-
inert gases
densing gases (ml)
(Vol%)
Radon activity PC/l of non _ non-concondensing densing gases gases g/l of
“.-__..-._ ..-._ 2.8
11
“50
!#ff
20 23 24 26 28 29 30
189 206 208 166 193 383 472
0.5 O-5 05 58 35 10.0 11.4
I;! .j -5 0,; 136 120 119 30 34 155 90.4
4‘7 4.7 4.7 2.7 2,7 6.2 6.2
f 5 i_ & * & 5
-5% 50,’ 5% zoo/b’ 20% 5% 5 “/,’ -_ ._-.
6
N2 from -
gosflosk
Fig. 1. Schematic drawing of apparatus for measurement of radon in gas samples. 1. Ionization chamber; 2 Vacuum gauge; 3. Electrometer; 4. Drying agent (Mg(CIO,),); 5. Ascarite; 6. Drying agent (Mg(CIO,),); 7. Gas sampling tube. The amount of gas contained in the gas sampling tube was detested by adjusting the g:ts pressure to one atmosphere with the aid of a niveau-flask containing saline water. Tbe volumc~ of gas was then measured at this pressure and room temperature. Two Mg(ClO,), traps and one Ascarite-trap were used for drying the gas and remo~ ing practically all of the SO, and CO, before the sample was introduced into the chamber. After reading the baokground current the ionization chamber and the purifying traps werct evacuated to about 60 mm Hg pressure and the gas samples were sucked into the chamber through the drying and absorbing agents. The chamber was then slowly brought to atmospheric pressure by bubbling N, through the sampling tube for about 15 min. In this way the bubbk,s flushedoutmore than 97 per cent of the radon dissolved in the water in the tube (seeLucAs, 19634). In the first minutes after the radon was introduced into the chamber the ionization currents increased rapidly due to the growth of the shortlived decay products of radon. After three hours the radon had attained a radioactive equilibrium with its decay products and a reliable reading of the current could be obtained. The equipment was calibrated with the aid of a standard O-099 yC radium solution using the same method as above for transferring radon from the standard solution into t0he chamber.
Radon and water in volcanic gas at Surtsey, Iceland
817
RESULTS Radon The results of the radon measurements are shown in Table 1. The radon activity in pC/l of noncondensing gases is shown in column 7. The amount of condensed water accompanying each liter of noncondensing gases is listed in column 6. Based on interpretation of the deuterium measurements (ARNASON and SIGURGEIRSSON, 1968) this water has been released from the magma and is not evaporated sea water as one might suspect from an oceanic volcano. Column 5 shows volume percent of nitrogen and inert gases, which serves as an indicator of atmospheric contamination of the samples. According to SIGVALDASOX and EL~SSON (1968) the samples from February 21 1965 are practically free from atmospheric contamination. Nitrogen and inert gases constitute only 0.5 vol “/ of t’he noncondensing gases in these samples. In the last column of Table 1 the radon activity per liter of noncondensing gases has been corrected for atmospheric contamination, assuming that all nitrogen and inert gases in excess of 0.5 vol y. is of atmospheric origin. Radium In order to calculate the initial equilibrium concentration of radon in the magma prior to degassing an attempt was made to determine the concentration of its parent nuclide, radium, in a lava bomb which was ejected from the crater on February 21 1965, when the best gas samples were obtained. A 5-g split of the lava bomb was dissolved in HF and fused with Na,CO, as described by HU~GENS et aE. (1951) until the lava was completely dissolved. The radon generated in the solution was measured with the aid of the ionization chamber. The sensitivity of the method was, however, not sufficient and it could only be ascertained that the radium concentration was less than O-4 . lo-l2 g Ra/g of lava. A split of that lava bomb and another sample, taken from lava flowing on Sept. 8 1966 were then sent to Dr. K. S. Heier at the Australian National University, who kindly determined the radium concentration in these samples. Prof. Paul W. Gast, Lamont Geological Observatory, visited Surtsey in August 1966 and collected several samples for radium measurements. He also obtained a piece of the lava bomb mentioned above and has kindly allowed us to cite his unpublished results. All results of the radium measurements are summarized in Table 2. DISCUSSION OP THE RESULTS Gas samples taken close to the crater indicated a radon concentration of 120 to 170 pC/l of non-condensing gases corrected for atmospheric contamination, whereas samples 26 and 28 taken from a lava flow, which had been exposed at the surface through a distance of about 100 meters, contained only 50 to 70 pC/l. These samples were also different in chemical composition (SIGVALDASON and EL~SSON, 1968) and deuterium ratio (ARNASON and SIOURGEIRSSON, 1968) and probably represent an advanced state in the degassing of the lava. In Surtsey, most of the gas was seen to escape in the crater and during the first meters of flow of the lava. The gas bubbles bursting at the crater presumably contained most of the gas that had exsolved as bubbles during the ascent of the magma and they may be more
818
SVEINBJBRNBJ~~RNSSON Table 2. Radium in Surtsey lava Date of flow
(Su 1, not dated) (Su 2, not dated) Su 3, Feb. 21 1965
Radium lo-l2 g/g ____0.26 o-27 0.17
Author
I
CO.4
Feb. 21 1965
Feb. 21 1965
0.17 * 0.01
Sept. 8 1966
0.11 & 0.01 I
Method
P. W. CAST, 1967 S. BJ~RNSSON, 1966
K. S. HEIER, 1967
Ionization chamber. Radon generated in solution. Gamma-ra:, spectrometry 1.74; MeV Bismuth-214 pock
representative of the composition of the primary volcanic gas then the gas tkt wa,s retained longer in solution. The radium measurements indicate an activity of radium of 0.1 to 0.3 pC/g of lava. If no radon escapes from the magma, this activity of radium will generate an equal activity of radon within 30 days. We may therefore conclude that the initial equilibrium concentration of radon in the magma prior to degassing was of the order of 0.1 to O-3 PC/g of magma. In Table 3 the radon/water ratio, ,L?,of the gas samples and the initial radon concentration in the magma, R,, are tabulated from the data in Tables 1 and 2. The radon/water ratio is, in general, not affected by atmospheric contamination. Water added because of the burning of hydrogen is insignificant. As the radium concentration in the lava flowing on March 31 1967 has not been determined, the concentration in the flow on September 8 1966 was used for calculating the initiitl radon concentra.tion on March 31 1967. It is of considerable interest to estimate the quantity of radon and volcanic gita that is released per unit mass of magma. Knowing the amount of radon initi,zll,v dissolved and the radon/water ratio in the exsolved volcanic gas, we might then be able to calculate the water content of the original magma. This suggestion will now be considered further. Since water is the main component of volcanic g;ts, the following discussion is limited to the release of water and radon from basaltic magma, which becomes supersaturated with water as pressure decreases during the* ascent to the crater. Table 3. Radon/water ratio in the volcanic gas and initial radon concentration in the magma at Surtsey Distance from crater Date of sampling (m) Nov. 25 1964 Feb. 21 196.5
Sept. 2 1966 March 31 1967
30 0 0 0 150 20 20
Sample NO.
Radon/water B (PC/g)
Radon in magma R,, (PC/d
14
20 23 24 26 28 29 30
0.17 Il.li 12.6\‘1’g
0.11
(0.11)
Radon and water in volcanic gas at Surtsey, Iceland The release of water in ascending
819
magma
The release of water in ascending magma depends primarily on the solubility of water in the magma. According to experimental data obtained by HAMILTON et ccl. (1964), RUSSELL (1957) and KURKJIAN and RUSSELL (1958) the solubility of water in basalt melts is proportional to the square root of pressure. Extrapolation of the data of HAMILTON et al. (1964) down to lower pressure indicates a solubility of 0.11 wt.% in a basalt melt (50.71 wt.% SiO,, 4.68 wt.% MgO) at 1100°C and 1 atm pressure (BJ~RNSSON, 1966). This estimate is supported by the evidence on water retained in Hawaiian lavas. According to MACDONALD (1963) the average combined water content (H,O+) in 51 recent analyses of tholeiitic basalts of Kilauea and Mauna Loa of historic age, by several-different analysts, is 0.14 wt.%. In recent and more precise analyses of samples of pumice collected during the 1959-1960 eruption of Kilauea FRIEDMAN (1967) has found a water content ranging from 0.064 to 0.104 wt.%. We may thus assume that magma containing water dissolved in excess of about 0.1 wt. o/o will saturate during the ascent and the excess water will create vapor bubbles in the magma. At 1100°C and pressures below 100 atmospheres the specific volume of water vapor deviates less than 1% from that of an ideal gas (KENNEDY and HOLSER, 1966). The total volume of bubbles released from one gram of magma containing initially 0.1 wt.% of excess water will be 6.25 cm3 at 1100°C and 1 atm pressure. The volume of bubbles will be about 95 per cent of the total liquid-gas volume. This means that magma containing only 0.1 wt.% of excess water will be changed into a froth on the last tens of meters before it reaches the surface in the crater. This extensive frothing of the magma must greatly facilitate the release of other gaseous components into the vapor phase, although the process may be too rapid for an equilibrium exsolution to be attained. The release of radon from magma According to the radium measurements the radon concentration in the magma This concentration corresponds to is of the order of 0.1 to 0.3 PC/g of magma. several thousand atoms of radon per gram. In this extreme dilution the radon gas is a tracer far from saturation in the solution. Release of radon from the magma will thus not occur unless the magma becomes saturated with some other gas component, which then partly exsolves and creates gas bubbles in the magma. The amount of radon released into gas bubbles will depend on the volume percentage of bubbles and the distribution coefficient between magma and bubbles for radon. The value of the distribution coefficient is not presently known but it should be relatively easy to determine, if the necessary laboratory facilities are available. One possible method might be to melt a lava sample of known radium concentration and let it dissolve a known amount of water under high pressure. On reducing the pressure the melt would vesiculate and the coefficient could be found from the radon/water ratio in the vapor released. At present, further evaluation of the data must be based on an assumed distribution coefficient, and will therefore be of limited value. It may however be stated that if the release of radon into the vapor bubbles is an equilibrium process,
820
SVEINBJ~~RN BJ~~RNSSON
and the ratio of equilibrium volume concentrations of radon in the magma and vapor bubbles is near unity, about 95 per cent of the radon will be released into bubbles with the first 0.1 wt.% of vapor. Even if the equilibrium concentration of radon in the magma were ten times greater than in the vapor, about 65 per cent ot’ the radon would still be released with the first 0.1 wt.% of water. Calculated content of water dissolved in the magma Although the distribution coefhcient between magma and vapor bubbles for radon is unknown, the present data provides an upper limit of the amount of water dissolved in the magma erupted at Surtsey. If R, and W, denote the amount of radon and water released from a unit mass of magma, R, the amount of radon initially dissolved in that unit mass and p t,ho observed radon/water ratio in the volcanic gas, the following expression will 1~1 valid :
Thus, the upper limit of the amount of water released may be found by dividing the observed radon/water ratio into the amount of radon initially dissolved in t(he unit mass of magma. If this is done for the data in Table 3 we obtain an upper limit of O-64 wt.% of water released on February 21 1965 and 0.56 wt.‘!!,‘! OJI March 31 1967. Adding the amount of water retained in the magma, which in Hawaiian lavas was found to be 10.1 wt.q/,, we obtain an upper limit of dissolved If practically all radon is released water of 0.74 wt.% and 0.66 wt.% respectively. into the vapor bubbles the actual amount of water released will be close to the upper limit calculated from the radon/water ratio in that vapor. Further conclusions cannot be drawn from the available evidence, until the distribution coefficient for radon is determined. It is of interest to compare the results above to the findings of MOORE (1965). In deep-sea tholeiitic pillow lavas (48.9 wt.% SiO, and 13.2 wt.y” MgO) dredged from the submarine part of the east rift zone of Kilauea, Hawaii, he found an average water content of 0.45 f 0.15 wt.% H,O+. The water content W;I,S roughly constant between 500 and 5000 meters depth and was regarded to represent the amount of water dissolved in the erupting magma. In conclusion, the measurement of radon in volcanic gas at erupting craters seems to offer a method for estimating the amount of water released from the erupted magma. To apply the method rigorously information on the distribution coefficient between magma and vapor bubbles for radon is required. It is hoped that this coefficient can be obtained in laboratory experiments. dcknowledgentents-I am indebted to Dr. I. FRIED~~AN and Dr. A. R. MCBIRNEY for their valuable criticism of the manuscript. Special thanks are due to Dr. K. S. HEIER, Australian National University for determining the radium content of the lava samples and to Prof. P. W. GAST, Lamont Geological Observatory for allowing citation of his unpublished results on radium. The valuable assistance of the Surtsey Research Society and the Icelandic Coast Guard in transporting us to Surtsey is gratefully acknowledged. A part of this work was supported by the Icelandic Science Foundation.
Radon and water in volcanic gas at Surtsey, Iceland
821
REFERENCES AR,NASONB. and SIG~RGE~SSONTh. (1968) Deu~ri~ content of water vapour and hy~ogen in volcanic gas at Surtsey. Geo&im. ~og~oc~~~. Acta 32, 807-813. BJ~RWSSONS. (1966) Radon in magmatie gas at Surtsey and its possible use for determining the content of water in the magma. Surtsey Research Progress Report II, pp. 97-110. The Surtsey Research Society, Reykjavik, Iceland. EVANS R. D. (1935) Apparatus for the determination of minute quantities of radium, radon and thoron in solids, liquids and gases. Rev. Sci. ~n~t~rn. 6, 99-112. PRIEDBL~NI. (1967) ?Vater and deuterium in pumice from the 1959-60 eruption of Kilauea volcano, Hawaii. U.S. G’eol. Survey Prof. Paper 575-B, B120-8127. HAMILTOND. L., BURNHAMC. W. and OSBORNE. I?. (1964) The solubility of w&or and effects of oxygen fug&city and water content on crystallization in m&c magmas. J. Petrol. 5,21-39. HUDGENS J. E., BENZING R. O., CALI J. P., MEYER R. C. and NELSON L. C. (1951) Determination of radium or radon in gases, liquids or solids. Nz&eonics 9, 14-21. ISRAEL H. (1934) Zur Methodik der klimatologischen Emanationsmessungen. I. iiber ein neues Emanometer. Der B~~eo~oge I, 318-327. ISRAEL H. (1937) Zur Methodik der klimatologischen Emanationsmessungen. II. uber den SBtt’igungsstrom bei emanometrischen Messungen. Gerlands Beitr. Geophys. 51, 35-49. REhNN:DY G. C. and HOLSER FV. T. (1966) Pressure-volume-temperature and phase relations of water and carbon dioxide. In Handbook of Physical Constants (editor S. P. Clark, Jr.,). Geol. Sot. Amer. &Fern.97, 371-384. KURICJIAN C. R. and RUSSELL L. E. (1958) Solubility of water in molten alkali silicates. J. Sot. Gh,sq. Tech. 42, 130-144. LUCAS H. I?. (1964) A fast and accurate survey technique for both radon-222 and radium-226. In The Natural Radiation .&+vironment (editors J. A. S. Adams and W. M. Lowder), pp. 315-331. The ‘University of Chicago Press. MACDONALDG. A. (1963) Physical properties of erupting Hawaiian magmas. Geol. Sot. Amer. Bull. 74, 1071-1078. MOORNJ. G. (1965) Petrology of deep-sea basalt near Hawaii. Amer. J. Sci. 263, 40-52. RUSSELL L. E. (1957) Solubility of water in mohen glass. J. Sot. Glass. Tech. 41,304-317. SIGVALDASONG. E. and ELfSSON G. (1968) Collection and analyses of volcanic gas at Surtsey. Geochim. Cosmochim. Acta 32, 797-805.