Jökull - 01.12.1975, Qupperneq 20

point rises, the latent heat of vaporization de-
creases and the heat loss through boiling de-
creases. Thus the change to the conditions be-
low the 2160 m depth limit is not sharp. These
conditions are: that real boiling does not take
place, the latent heat of vaporization is zero.
The result of a contact of magma and water is
just that the latter forms a rather thin “bound-
ary layer” of density around 0.5 g/cm3, which
has been heated rather gradually to a tempera-
ture above 375 °C. At the same time this
“boundary layer” rises slowly and makes place
for new cold water to be heated, and so on. A
glassy crust is now formed in quite another
way than in shallow water. The initial cooling
is much less and slower, but the following heat
exchange with the water is more rapid than in
shallow water, as the surface temperature re-
mains high for sorne time. For this reason, a
relatively thick crust is cooled down to medium
temperature — and it turns out that the reten-
tion of excess argon depends on the rapidity
with which the glass was cooled. For in the
studied pillow from 2590 m depth, argon corre-
sponding to an age of 42.9 My was retained
in the outermost 0—1 cni layer, 33 My in the
1—2 cm layer, but only 4.3 My in the 3.5—4.5
cm deep layer; deeper, no excess argon was re-
tainecl in the glassy pillow. The shallow water
pillows tell the same story. Excess argon may
have been considerable in the outermost 0—0.1
cm layer, but such a thin layer might be lost
by the filing off of palagonite.
We now come to the question of the reten-
tion of the excess argon. This excess character-
ized the magma, and must originally also have
been in the interior of the pillow but later lost.
This interior argon would hardly diffuse
through a crust which at the same time was
cold enough to retain its own excess argon. The
porosity in this crust is 0.5%, while it rises to
2% in the central parts of the pillow. The
corresponding expansion must have led to the
*) On May 2, 1973, Dr. Moore, coauthor of the above
quoted paper, gave a lecture at the University of Iceland
in Reykjavik, and showed a film of submarine lava flow
at a shallow depth at Hawaii. The present author had the
good luck to be present, and for the first time realized
how extremely rapidly the boiling cools the surface of the
pillows below the boiling point, as indicated by the short
time and the small amount of vapour formation upon the
contact of open lava with the water.
formation of fractures in the crust. But we do
not think it likely that the gradual formation
of these cracks, proceeding at each time into a
plastic mass, were at the same time paths for
the argon of the interior.
As a provisory idea, we postulated that the
2590 m pillow came to rest on a hot surface,
and that the argon of the interior diffused
down through this surface during the time the
pillow was being cooled. Taking the thermal
diffusivity k to be 0.007 cm2/sec, as done by
Dalrymple and Moore, we found that the dif-
fusivity D of argon in a magma of 1200 °C
had to be of the order of 100 cm2/sec. This
result demonstrates the impossibility of the
model, for D for a real magma is at least ten
orders of magnitude less than the model re-
quires. For minerals, D is in the range 10~8 to
10~14 cm2/sec at a temperature of about 1200°C
(Dalrymple and Lanphere, 1969). In minerals
there may be regular channels along which an
inert gas like argon may move relatively easily,
while such are not found in a melt. Hence, D
may be less for the magma than for the miner-
als. The escape of excess argon from the interior
of a pillow before it is cooled down within a
few hours, is thus out of the question.
We seem to be left with only one possibility:
slowly cooled volcanic glass is far less retentive
for argon than more rapidly cooled glass, and
looses its excess argon within times of decades,
centuries or millennia.
The argon of the interior of the pillows must,
it seems, have diffused through the cracks in
the crust. Whether or not absorption of sea
water played some part in that process we can-
not tell. At any rate the conclusion seems in-
escapable that the structure of the slowly form-
ed glass is different from that of the more
rapidly forrned glass.
This result suggests the important inference
that if the potassium goes into interstitial glass
by the crystallization of a basaltic lava flow, the
produced argon will leak out easily, as this glass
was formed by slow cooling.
The reason for the different retention capa-
city of slowly and rapidly cooled glass is con-
sidered in another paper, soon to be published.
Dykes and larger intrusive bodies. The above
data frorn Hawaii tell us that excess argon will
be present in magma while it is rising up along
1 8 JÖKULL 25. ÁR