EXPERIMENTALLY SIMULATING AND ANALYZING BUBBLE DYNAMICS AND NUCLEATION IN MAGMA

At any given moment, an average of 40-50 volcanoes are erupting worldwide. About 1 in 20 people live within the explosive range of an active volcano, putting around 350 million lives at risk if eruptions occur. Bubble nucleation and growth is the driving force that sends magma through volcanic vents. To better understand bubble nucleation, we experimentally test natural and synthetic obsidian samples of differing water, silica, and crystal contents. In a furnace, we heat a block of inconel to 400-900 C, depending on the characteristics of the obsidian sample to be tested. We then touch one end of the obsidian sample, insulated on all other sides with Kaolin wool and ceramic, to the inconel. We constructed an experimental apparatus, including a lever arm to ensure stability of the sample and safety of the user, and the addition of a heat shield to place on the furnace when the door is open. Bubbles form at the hot end of the obsidian but not the cold end. By positioning thermocouples strategically on the obsidian, we measure a thermal gradient in the sample and track the life cycle of bubbles from their nucleation. Working at 1 atm, we are identifying the temperature at which bubble formation occurs in natural samples of rhyolite from Armenia, Iceland, and California. We are also testing synthetic samples of rhyolite with 5 wt% water and varying crystal contents fabricated at the University of Texas. We are determining the thermal conductivity for each tested sample by using the thermal time constants of the measured heat gradient histories. We will next work to understand more about nucleation in homogeneous (crystal-free) and heterogeneous (crystal-bearing) samples, and how crystal microlite counts in heterogeneous samples affects nucleation. We will contribute to a dynamic bubble growth model based on our experimental findings. This model will seek to advance understanding of the dynamics of fragmentation from an explosive eruption using bubble size distributions.