Research of Professor Youxue Zhang

As of July 2003


Earth scientists aim at understanding the processes and history of Earth and other planets. Earth processes are diverse, ranging from microscopic processes occurring at the atomic level such as diffusion and chemical reactions, to macroscopic processes such as earthquakes, volcanic eruptions, and Earth differentiation. Earth history is long, from the violent beginning at about 4.5 billion years ago, to the appearance and demise of many species (including dinosaurs), and to the rise of humankind. Earth science is becoming increasingly broad, interdisciplinary and relevant to society.

Professor Zhang has contributed to a range of disciplines of Earth sciences: experimental petrology, volcanology, and geochemistry. More specifically, his contributions focus on kinetics and dynamics of magmatic and volcanic processes, early history of Earth, as well as recent new endeavor into Earth's surface environment. As a high-temperature experimentalist and a theoretician/modeler, one characteristic of his work is the combination of careful experiments with thorough theoretical analysis and modeling. Their experiments are guided by theoretical understanding, and followed by new and thorough theoretical analysis and quantitative modeling. They have found that such a combination is ideal in gaining fundamental insights of Earth processes.

Professor Zhang and his group as well as collaborators have addressed a broad range of questions regarding high-temperature Earth processes, as well as the early history of Earth. They experimentally examined equilibrium and kinetics of chemical reactions in minerals and melts and between minerals and melts, designed new experimental methods to simulate the dynamics of explosive volcanic and lake eruptions, proposed new hypothesis to forecast dynamics of such eruptions, and developed new theories to understand diffusion of water in magma (a critical component process in explosive volcanic eruptions) and minerals, and chemical reactions in magma (and recently, in water). Their research has probed the frontiers of volcanology and geochemical kinetics. Their work also reaches the frontiers of geochemical evolution of Earth. They have applied isotopic constraints to model the accretion and degassing history of Earth, leading to the refinement of the age of Earth, as well as its early evolution history. Some of their research findings have been communicated to a broad audience through invited review papers on water in rhyolitic melt (Zhang, 1999b) and the age and accretion of Earth (Zhang, 2002b). Selected contributions are outlined below.

Diffusion of a multi-species component

        Zhang and coworkers developed the concept and theory of the diffusion of a multi-species component (Zhang et al., 1991a, 1991b; Zhang and Stolper, 1991; Jambon et al., 1992; Zhang and Behrens, 2000).  The problem is of not only theoretical interest in kinetics, but also practical importance because the most major volatile component (H2O) in silicate magmas dissolves in melt as two species (molecular H2O and hydroxyl).  Dissolved H2O plays a critical role in driving explosive volcanic eruptions: diffusion of water supplies mass into bubbles during bubble growth in magma, and powers explosive volcanic eruptions.  Hence understanding water diffusion is essential for understanding explosive volcanic eruptions.  Dissolved water also significantly affects the properties of silicate melt (including glass) such as the strength of glass and the crystallization sequence of magma, which is why glass scientists are also interested in the problem.  Extensive experimental studies by many groups (including Zhang and coworkers) have shown that H2O diffusivity increases in a complicated way with total H2O concentration: it is proportional to total H2O concentration at low H2O content (<2 wt%), and increase exponentially with total H2O content at higher H2O content.  With high quality experimental data and the measurement of concentration profiles of the two species, Zhang et al. (1991a) developed the concept of multi-species diffusion and were able to show that (1) hydroxyl diffusion is negligible, and (2) molecular H2O is the diffusing species during the diffusion of H2O component.  The hydroxyl concentration profile is not due to the diffusion of hydroxyl itselt, but due to molecular H2O diffusion together with the interconversion reaction between molecular H2O and hydroxyl.  Previous experimental results were also reconciled using the concept (Jambon et al., 1992).  Later experimental work by Zhang and Behrens (2000) to much higher H2O contents showed that molecular H2O diffusivity increases exponentially with total H2O content (meaning that molecular H2O diffusivity is roughly constant at low total H2O and increases rapidly with total H2O content at high H2O).  This inference was confirmed by the investigation of Ar diffusion, a simple molecular component with single-atom molecules (Behrens and Zhang, 2001).  In this way, the complicated diffusion behavior of H2O is understood and quantified.  The results not only deciphered the theoretical problem of multi-species diffusion, but also found many applications in modeling volcanic eruption dynamics and other magmatic processes.  Zhang et al. also applied the concept to investigate oxygen diffusion (Zhang et al., 1991b).  Zhang and coworkers are continuing the work on H2O diffusion in dacitic, andesitic, and basaltic melts to understand volcanic eruptions of such compositions. 

Dynamics of explosive volcanic eruptions and lake eruptions

Gas-driven eruptions, including explosive volcanic eruptions and the recently recognized lake eruptions, are powered by the exsolution of an initially dissolved gas from a liquid. Explosive volcanic eruptions are a main geologic hazard, and some have changed the course of human civilization. Explosive volcanic eruptions are powered by the exsolution of H2O gas initially dissolved in magma, including the 1980 eruption of Mt. St. Helens (killed 57 people), 1902 eruption of Pele (killed 28,000 people), and the ~1600 BC eruption of Santorini (which probably destroyed the Minoan civilization). A new type of eruption has been recognized only since the 1980s. A massive gas release from Lake Nyos, Cameroon, in 1986, killed ~1700 people (Kling et al., 1987). A similar though smaller event occurred at Lake Monoun in 1984, killing about 40 people. Initially there were debates about the nature of these events, and after detailed research the events are now known to be lake eruptions powered by exsolution of dissolved CO2 in water.

Zhang and coworkers began to work on gas-driven eruptions with experimental simulation of CO2-driven water eruptions (Zhang et al., 1992; Mader et al., 1994; Zhang et al., 1997a; Zhang, 1998a). The work gradually evolved to include dynamics of lake eruptions (Zhang, 1996, 2000), and investigation of component processes during explosive volcanic eruptions such as bubble growth (Liu and Zhang, 2000), magma fragmentation (Zhang, 1999c), water diffusion in silicic melt (Zhang and Behrens, 2000), the exsolution enthalpy of water from silicate liquids (Zhang, 1999a), and energetics of gas-driven eruptions (Zhang, 2000). Three fundamental contributions are described further below.

One contribution is to develop experimental techniques to simulate volcanic eruptions. Before their work, pioneers used one-component system to simulate volcanic eruptions, either the sudden decompression of a gas phase (Kieffer and Sturtevant, 1984) or the rapid evaporation of a liquid phase (Hill and Sturtevant, 1990). Because an explosive eruption involves a liquid magma and a gaseous vapor, Zhang realized that a two-component system must be used for a more realistic simulation of the eruption process. This realization led to their work using CO2-water system at room temperature to simulate volcanic eruptions, called champagne-type experiments (Zhang et al., 1992; Mader et al., 1994; Zhang et al., 1997a; Zhang, 1998a). These experiments produced visual pictures that provided insight into the fundamental dynamic behavior of gas-liquid system. Such insights are guiding new understandings of natural eruption processes.

One such example was to understand the dynamics of lake eruptions (Zhang, 1996, 2000). After its discovery in the 1980's, initial works by others showed that there is high CO2 concentration in lake bottom water. Some scientists thought it was a "lake overturn" that released the killing CO2. Some called it a lake eruption, but the dynamics was not known. The experimental work by Zhang and coworkers (Zhang et al., 1992; Mader et al., 1994; Zhang et al., 1997a; Zhang, 1998a) showed clearly that dissolved CO2 in water can drive powerful water eruptions and showed visually how eruption proceeds. From such insight, Zhang modeled the dynamics of lake eruptions (Zhang, 1996, 2000) and showed through theoretical analysis that lake eruptions can be violent. "Lake overturn" is now considered a misnomer for such lake eruptions. An understanding of lake eruptions is essential to mitigation measures (Halbwachs and Sabroux, 2001).

Another major contribution by Zhang on volcanic eruptions is to understand magma fragmentation. Fragmentation is a defining point in a volcanic eruption: before fragmentation the eruption is a bubbly magma flow that is slow; during fragmentation many bubbles break up to release the compressed gas; after fragmentation the eruption becomes a violent gas flow carrying suspended magma particles. Seemingly benign lava flows or domes may suddenly fragment into deadly pyroclastic flows (Sato et al., 1992; Fink and Kieffer, 1993). Hence understanding the conditions for fragmentation of a bubbly magma is critical to saving lives. Zhang proposed a new fragmentation criterion for bubbly magma based on brittle failure theory (Zhang, 1999c). In this work, a bubble is treated as an inclusion in a host, the mechanical aspect of which was treated earlier (Zhang, 1998b). Because gas pressure inside a bubble is greater than that in the melt, there is a stress field in the magma. Fragmentation occurs when the tensile stress on the bubble wall exceeds the tensile strength of the magma, such that many bubbles break simultaneously, releasing high-pressure gas in the bubbles and powering pyroclastic flows (Fink and Kieffer, 1993). The theory may eventually lead to forecasting of sudden collapse of lava domes and flows.


Experimental studies of equilibrium processes

Zhang and coworkers have experimentally investigated equilibrium processes, including phase equilibrium (Withers et al., 2003), calibration (Zhang et al., 1997c), solubility of water in melt (Zhang, 1999b), and equilibrium of reactions in silicate melts. One contribution is outlined here. Water dissolves in silicate melt as at least two species, molecular H2O and hydroxyl group. The two species interconvert at high temperature. The concentrations of both species can be measured by infrared spectroscopy. Zhang and coworkers have investigated the equilibrium (Zhang et al., 1995, 1997c; Ihinger et al., 1999) by carrying out the experiments at intermediate temperatures and then quenching the samples to room temperature. The range of intermediate temperatures was determined from careful kinetic experiments. After their early work on the equilibrium of the reaction, two groups of German scientists (Behrens�� group at the University of Hannover, and another group at Bayreuth) using a different measurement technique (in situ technique) disagreed with their results and hence these authors thought that results by Zhang and coworkers were wrong (Nowak and Behrens, 1995; Shen and Keppler, 1995). During his sabbatical in 1998, Zhang visited University of Hannover and collaborated with Behrens' group on this problem. Collaboratively the two groups found that results by Zhang and coworkers were correct, and the in situ results need to be corrected for the temperature dependence of calibration coefficients (molar absorptivities). After correction for the calibration coefficients on their data, the in situ results are in agreement with the quench results. Hence Behrens' group and Zhang's group published a collaborative paper (Withers, Behrens and Zhang, 1999) to reconcile the differences, after which Zhang and coworkers also published their long-held paper on speciation equilibrium (Ihinger et al., 1999). Behrens and Zhang became good friends and are still collaborating.


Experimental studies of kinetic and diffusion processes

More of the experimental effort by Zhang's group has focused on geochemical kinetics rather than equilibrium. Disequilibrium is common in Earth processes, either in magmatic systems at high temperature, or in aqueous systems at low temperature. The work by Zhang's group has contributed to theoretical quantification of these processes. They worked mostly on high-temperature kinetic problems involving silicate melt and volcanic eruptions. Recently they expanded their work to kinetic problems in low-temperature aqueous solutions (Zhang and Xu, 2003). The kinetic problems that they have unscrambled through experiments and theoretical development include: diffusive crystal dissolution in magma (Zhang et al., 1989), bubble growth in magma (Liu and Zhang, 2000), convective crystal dissolution in water (Zhang and Xu, 2003), kinetics and geospeedometry of the hydrous species reaction in magma (Zhang et al., 1995, 1997b, 2000), water diffusion in silicate melt (Zhang et al., 1991a, 1991b; Jambon et al., 1992; Zhang and Stolper, 1991; Zhang and Behrens, 2000), and other diffusion problems (Van der Laan et al., 1994; Wang et al., 1996; Behrens and Zhang, 2001). Some fundamental contributions are described below.

One of their major contributions is to understand the complicated water diffusion behavior in magma and develop the concept and theory of multi-species diffusion (see above). 

The second fundamental contribution is to understand kinetics of hydrous species reaction and development of it into a practical geospeedometer. Zhang and coworkers investigated the kinetics of the reaction of hydrous species in rhyolitic melt as a function of temperature, cooling rate, and water content (Zhang et al., 1995, 1997b). They further quantified the kinetics as a geospeedometer, i.e., a cooling rate indicator (Zhang et al., 1997b, 2000). One of the basic aims of petrology and geochemistry is to understand thermal history of rocks, including the cooling rates. Their geospeedometer has been used to infer cooling rates in various experimental apparatus (Zhang et al., 2000) and in volcanic glasses (Zhang et al., 2000; Xu and Zhang, 2002; Wallace et al., 2003). Such investigations have revealed cooling rates and temperature in volcanic eruption columns as well as pyroclastic deposits. The understanding provides constraints on eruption column dynamics and thermal evolution of volcanic deposits. They further applied the cooling-rate experiments to infer viscosity of hydrous rhyolitic melts (Zhang et al., 2003).


Methane hydrate and possible methane-driven oceanic eruptions

Most research by Zhang's group so far has been on high-temperature processes. Recently, Zhang begun to broaden his research scope to include low-temperature geochemistry, on methane hydrate in marine sediment. His interest in this direction started from a Ph.D. thesis defense of G.R. Dickens in the same Department. A huge amount of methane is present in marine sediment in three different forms: methane hydrate (an ice-like substance that can burn), methane gas bubbles, and dissolved methane in pore water. The total amount of carbon in marine methane is estimated to be more than all other fossil carbon combined. Hence, methane in marine sediment has the potential to become a major energy source. Secondly, methane could be rapidly released from marine sediment and cause rapid climate change, affecting life on Earth. Thirdly, rapid release of methane could be a geohazard. Nevertheless, the fate and dynamics of methane release from marine sediment and how it could pass through deep ocean water and enter the atmosphere were not known. There was much confusion in the literature. Zhang and coworkers applied their expertise on geochemical kinetics and on dynamics of gas-driven eruptions to investigate the fate of methane once released from marine sediment. They recognized that due to the complicated phase relations, the kinetic processes are diverse and rich in details. They started with a paper elucidating the various types of kinetic processes, hypothesizing a new type of gas-driven eruptions (methane-driven oceanic eruptions), and modeling the dynamics of such eruptions (Zhang, 2003). They then developed the theory to treat kinetics and dynamics of dissolution and dissociation of methane hydrate in seawater, which clarified previous confusions, and carried out experiments to verify the theories (Zhang and Xu, 2003). Next, they plan to develop theories for methane bubble dissolution kinetics and dynamics, as well as the behavior of bubble plumes. From there, they plan to expand into other low-temperature kinetic processes, focusing on Earth's environment.


Age, accretion and evolution of Earth

Zhang has also maintained research interest on Earth history and evolution (Zhang and Zindler, 1989, 1993; Zhang, 1997, 1998, 2002a, 2002b). His recent contribution along this line has concentrated on the age and accretion history of Earth.

The monumental work of Patterson (1956) showed that the initial formation of Earth was close to that of most meteorites at 4.55 Ga. Over the last 30 years, workers argued for a younger core formation age based on Pb-Pb model ages and tungsten isotopic data, and for a younger gas retention age based on I-Xe modeling. However, disagreement was abundant and the younger age of Earth was not widely accepted. The I-Xe age was also challenged on the basis that iodine is volatile (Azbel and Tolstikhin, 1993). Zhang (1998c) examined the combined 129I-129Xe and 244Pu-238U-136Xe-134Xe-132Xe-131Xe system, and used a model-independent approach and total inversion to show that (1) all Xe clocks can be reconciled with a single Xe-closure age of Earth: 4.45+-0.02 Ga, 109+-23 million years younger than the formation of meteorite Bjurbole (~4560 Ma) and (2) all radiogenic components of noble gases in the atmosphere can be quantitatively accounted for by production and degassing of ~60% of the bulk silicate earth (Zhang, 1998c). Because Pu and U are refractory, the agreement between the 129I-129Xe clock and 244Pu-238U-136Xe-134Xe-132Xe-131Xe clock shows that the volatility of iodine does not affect the 129I-129Xe clock.

In a review paper (Zhang, 2002b), Zhang reviewed work on age of core formation, Xe closure, and formation of the earliest crust using U-Pb, Hf-W, I-Pu-U-Xe, Sm-Nd, and Nb-Zr systems. All data available at the time of Zhang's review were consistent with either one of the following two models: (a) a single age of 4.45+-0.02 Ga of Earth (for all events), and (b) a model of continuous Earth accretion and simultaneous core formation, followed by a giant impact stripping Earth's atmosphere at 4.45 Ga. More recent data show that earlier Hf-W data measured by Lee and Halliday (1995, 2000) and reviewed in Zhang (2002b) might be erroneous and Earth's core formation age constrained from 182Hf-182W systematics (Yin et al., 2002; Kleine et al., 2002) is much older than it was thought before. Hence model (b) (continuous Earth accretion and simultaneous core growth with a two-stage core formation age of 4.533 Ga, followed by a giant impact at about 4.45 Ga stripping Earth its atmosphere) appears to be the only viable Earth accretion model to satisfy all isotopic data.


Other studies

Other investigations by Zhang's group include various theoretical investigations of diffusion (Zhang, 1993), reaction kinetics and geospeedometry (Zhang, 1994), atomic radii of noble gas elements (Zhang and Xu, 1995), mechanical and phase equilibria in inclusion-host systems (Zhang, 1998b), oxygen barometer (Zhao et al., 1999), exsolution enthalpy of water from silicate melts (Zhang, 1999a), and energetics of gas-driven eruptions (Zhang, 2000). They have also carried out several petrological studies in collaboration with other professors in the same Department, including mineral inclusions in pyrope crystals (Wang et al., 1999), immiscibility in garnet (Wang et al., 2000a), and magnetism of mid-ocean basalts (Zhou et al., 2000, 2001). Zhang's group has also discovered a new mineral with a unique oxide structure (Wang et al., 2000b).



In summary, by combining careful experiments with thorough theoretical analysis, Zhang and coworkers have gained fundamental understanding on some key geological processes. As they explore dynamics and kinetics of Earth processes at the fundamental level, they bear in mind the relevance of such understandings to society, such as forecasting volcanic eruptions. Much of their research is relevant to society, related to hazard mitigation: they not only modeled dynamics of lake and possible oceanic eruptions, not only proposed the criterion for magma fragmentation that is a defining process of explosive volcanic eruption, their experiments and models have also provided essential data (water diffusivity, water solubility, and melt viscosity) for modeling volcanic eruptions. The geospeedometer that they developed has the potential to provide vital information to verify or invalidate volcanic eruption models.