Submarine volcanics have morphologies that reflect the effusion rate, the sea floor topography, viscosity and cooling rate. Since direct observation of these undersea volcanics is limited, information concerning these eruption characteristics is often inferred from their morphology or from laboratory analogues such as liquid wax. Many undersea volcanics contain glassy portions. When a liquid is cooled through the glass transition, the resulting glass retains structural information characteristic of the quench rate. In recent papers we have demonstrated that the quench rates of natural glasses can be determined by examining the heat capacities of natural glasses when they are heated through the interval of glass transition (see Annual Report 1993). Previous measurements have demonstrated a range of quench rates from 10 to 1.2 x 10-4 K sec-1. These quench rates are consistent with a variety of cooling processes, including conduction to substrates, cooling by radiation to the air and a complex annealing process involved in the rheomorphic flow of previously rapidly quenched glasses. Accurate heat capacity measurements rely on relatively large glass samples that are not subject to compositional or other changes during the course of the thermal treatments. In this discussion we will attempt to evaluate the potential for such measurements on the small glass samples recovered from undersea basaltic volcanics. In addition we aim to establish the range of cooling rates experienced by basaltic liquids and also evaluate whether the modeled quench rates are realistic when compared with simple heat flow models.
The cooling rates determined for these glass samples show a large range. Two questions are raised from this study; are the range of cooling rate consistent with those expected for lava erupted underwater. Secondly, how can the apparently slow cooling rates be rationalized for such liquids, which as noted above, can crystallize easily.
Hyaloclastite samples yield a range of cooling rates from 0.416 to 0.00024 K sec-1, a range that is smaller than the range of cooling rates seen in subaerial volcanics but which, none the less, is larger than would be expected for simple conductive cooling of hot lava in contact with sea water. Despite problems encountered with samples crystallizing and difficulties in obtaining a heat capacity signal for small glass fragments, the quench time for glassy fragments can be obtained by modeling of the relaxation of enthalpy. The submarine volcanic glasses retain quench histories that include rapidly cooled fragments within a hyaloclastite matrix that have quench rates consistent with simple thermal models of erupting magma with an origin close to the water-lava interface. These glasses are not, however, the most rapidly cooled portions of the lava (which we assume to be represented by the hyaloclastite matrix).
Simple models of the cooling can be used to evaluate the range of instantaneous cooling rates that would be expected for a submarine basalt. For the purposes of illustration, a model of conductive heat flow in two linear half spaces can be used. For an initial eruption temperature of 1050°C (1323 K) and a thermal diffusivity of 10 x 10-6 m2 sec-1 calculations for the evolution of temperature and instantaneous cooling rate (T/t) can be calculated as a function of depth from the lava-water interface as a function of time. The instantaneous cooling rates for the interface are very rapid: at the point of glass transition (which we will arbitrarily set to 950 K) the cooling rates range up to 3500 K/min in the first 2.5 mm of the interface and the point of glass transition is reached after 4 seconds. Further from this interface the estimated cooling rates approach values comparable to the estimates for the hyaloclastite samples. At depths of 25 mm and 50 mm, the instantaneous cooling rates at the point of glass transition are, respectively 35 and 10 K/min following a time interval of 400 and 1400 seconds. A simple interpretation of the more rapidly cooled glass fragments from the samples studied is that they represent fragments of glass a few cm from the interface, the more rapidly cooled portions of which form the hyaloclastite matrix.
The apparent slow cooling rate (0.015 K/min) experienced by samples, 1-4, 3-4 and in the most extreme case 2-3 do not result from this simple linear cooling. In the simple cooling models described above the cooling rate at the onset glass transition remains in the restricted range of Log10|q| = -1 to 2 K sec-1. The slow cooling rates of Log10 |q| = -3.60 seen in sample 2-3 would not be expected to result in a glass. We suggest that the lower limit to glass formation is that of the onset of crystallization. This is taken at Log10 |q| = -1.5 Ksec-1, approximately 0.5 Log10 units less rapid than the thermal treatments applied to these glasses. The heat capacity curves can, however, be modeled with a high degree of confidence, so another explanation is required to explain the slow apparent quench rates.
The more slowly cooled glasses reflect a two-stage thermal history. This involves cooling of the erupted lava to a point within the glass transition, since initial cooling rates less rapid than this would involve crystallization. Provided the glass is retained within the glass transition interval for sufficient time, the structure can relax to a lower value of fictive temperature (enthalpy). In these glasses, the thermal history reflects not only the formation of the glass by rapid quenching at the margins but also the dissipation of heat from the lava body and is consistent with observed cooling of some erupted basaltic lavas.
Enthalpic relaxation studies of glassy portions of undersea volcanics offers a tool for determining the apparent quench rates of lava surfaces and provides a means of unraveling the complexity of their thermal history. The apparent quench rates of the hyaloclastites reflect both the cooling rates experienced at the flow margins and also the evolution of the total heat budget. We have demonstrated here that careful sampling of glassy portions of undersea volcanics provides a means of determining the thermal histories of erupted volcanic products and a calibration for elaborate thermal models which consider that the influence of such factors as eruption rate, the dissipation of heat from surfaces and the evolution of heat resultant from the onset of crystallization. The thermal studies of these basaltic flows may provide constraints on the physical processes of eruption, for example the inflation of quenched glassy rinds during pillow lava eruption. The only restriction to studies of the relaxation are whether glassy products are obtained. Consequently any relatively recent volcanic sequence that contains glassy material is suitable for study with the caveat that the spatial relations of the glasses and coexisting volcanics must be well-documented in order to produce a consistent thermal model.