If large, persistent thermal plumes exist in Earth’s mantle, they most likely rise from a thermal boundary layer in the interior (e.g., What is a plume? page ). The steep, super-adiabatic thermal gradient within a thermal boundary layer favors the development of boundary layer instabilities, with upwelling plumes of hot boundary layer material winning out over conduction as the more efficient mechanism to transport heat upward once the boundary layer thickens to a critical level. There is, as yet, no compelling direct evidence for the existence of any mid-mantle thermal boundary layer, although this possibility has not been ruled out. On the other hand, it is generally agreed that a thermal boundary layer does exists in the lowermost mantle, the so-called D" Region, due to the requirement that heat must flux out of the core in order to sustain the geodynamo (Lay et al., 2004). It is plausible that thermal instabilities develop in this boundary layer. Thus, the source of any large mantle plumes is generally assumed to be the D" Region (Lay, 2005).

Thermal plumes will drain hot material from the boundary layer as long as they persist, and in a homogeneous medium with temperature dependent viscosity, the most mobile material will be the lowest viscosity material at the base of the hot boundary layer. Thus, one might hope to find a geochemical signature of mantle material that has been in direct contact with Earth’s core present in the erupted lavas of a hotspot volcano fed by a mantle plume (see Osmium-Tungsten page). The plume material should also bear upward the thermal contrast of the boundary layer itself. Estimates of the D" temperature contrast that assume whole mantle convection range up to as high as 1500°C (see Lay et al., 2004), much greater than the excess temperatures inferred for many hot spot melts. These considerations should motivate interest in D" structure for all parties exploring the mantle plume hypothesis.

Seismological studies reveal substantial complexity in the seismic wave velocity structure of D", and several recent reviews highlight general attributes of the region (e.g., Lay & Garnero, 2004; Lay et al., 2004). Two key attributes of the lowermost mantle are considered here. First is the fact that global seismic tomography has resolved the presence of large-scale heterogeneity in D", dominated by large regions with shear velocity lower than average under the central Pacific and under southern Africa, the southern Atlantic and southern Indian Ocean. Figure 1 is a map showing the shear velocity heterogeneity in the lowermost mantle from a representative tomography model (Grand, 2002).

Figure 1. Laterally varying S-wave velocities at ~2700 km depth in the mantle from the tomography model of Grand (2002). Blue regions: higher than average velocity, red regions: lower than average velocity. Average is defined by the PREM model (Figure 2). Labeled subregions: areas where an S-wave velocity discontinuity has been observed a few hundred kilometers above the core-mantle boundary (CMB). This usually corresponds to regions where seismic velocity is higher than average, which may be regions where temperature is lower than average, with a post-perovskite transition at shallower depths than in low-velocity/hotter regions.

After decades of puzzling over the nature of the D" discontinuity, a viable explanation is now to hand. In 2004, mineral physics experiments discovered that the predominant lower-mantle mineral, magnesium-silicate perovskite [(Mg_1-x,Fe_x)SiO₃] undergoes a transition to a post-perovskite polymorph at temperatures and pressures similar to those in D" (Murakami et al., 2004; Oganov & Ono, 2004; Iitaka et al., 2004). This phase change is expected to result in a ~2% increase in S velocity, a small decrease or increase in P velocity, and a 1% density increase. Theoretical models suggest a strong positive Clapeyron slope for the transition (e.g., Tsuchiya et al., 2004), perhaps twice as strong as that of the 410-km deep transition from olivine to β-spinel (Wadsleyite). This is relevant to the mantle plume hypothesis, because this phase change would favor generation of thermal boundary layer instabilities and vertical transport of material in the deep mantle (Lay et al., 2005). The positive Clapeyron slope and the predicted superadiabatic temperature increase in D" could lead to post-perovskite transforming back to perovskite at the very base of the mantle (Hernlund et al., 2005) if the temperature increase is sufficient. This would result in a velocity decrease there that seismologists can seek, although it is much harder to observe a decrease than the velocity increase from the shallower phase boundary (e.g., Flores & Lay, 2005).

If the post-perovskite phase does actually exist in the lowermost mantle, it will occur in the lower-temperature regions of D". This could account for the correlation between where a shear-velocity discontinuity is observed and where high velocity (lower temperature?) volumetric structure is found in the boundary layer. Helmberger et al. (2005) assume that tomographic shear-velocity variations in D" are primarily a thermal effect, and they map the variations into a topography map for the position of the phase boundary; shallower in the mantle under higher-velocity regions and deeper in the mantle under lower-velocity regions (Figure 3). In order to account for the large-scale pattern in the velocities, it is often suggested that subducting oceanic lithosphere has reached and ponded above the CMB, with sufficient thermal anomaly retained to affect the seismic velocities (e.g., Garnero & Lay, 2003; Sidorin et al., 1999). Due to the lower temperature of slab material compared to normal boundary-layer material, one would then infer that the coldest regions of slab material should have a thicker layer of post-perovskite material. If ambient mantle temperatures are too high for the post-perovskite phase to exist, it might even be that any post-perovskite phase is confined to the vicinity of recently delivered slab material that has not yet thermally equilibrated. Lateral variations in the depth of the D" discontinuity, including abrupt steps in depth, have been interpreted as the result of thermal variations within folded, ponded slab material beneath downwellings, modulating the phase boundary (e.g., Helmberger et al., 2005; Hutko et al., 2006). Pretty speculative, but not implausible.

The large low-shear-velocity provinces under the Pacific and Africa appear to have a compositional contribution to their anomalous velocities (e.g., Lay et al., 2004). The effects of iron and aluminum on the post-perovskite phase are being examined by experiments and theory, with iron likely reducing the pressure at which the transition occurs while aluminum increases the depth extent over which it goes to completion. The post-perovskite phase appears to have the capacity to take in surprisingly large amounts of iron ((Mg_1-x,Fe_x)SiO₃ where x ~ 0.4, e.g., Mao et al., 2004), and such strong iron enrichment can reduce seismic velocities dramatically, possibly even to the extent needed to account for the ultra-low velocity zones (ULVZ) detected just above the CMB without requiring any partial melting (W. Mao, personal communication, 2005). This later finding, if substantiated, may have a big impact on our thinking about the role of partial melting or melt components in the ULVZ, and implications for thermal plumes as well.

An important recent observation is that a shear-velocity D" discontinuity about 230 km above the CMB is observed in the central Pacific on the northern end of the large low-shear-velocity province (Avants et al., 2006). As this region (Figure 1) is far removed from any subduction zone and has velocities lower than average in tomographic models, attributing the discontinuity to the post-perovskite phase transition requires that there be no direct connection to deep slab material. This implies very widespread occurrence of the shear-velocity discontinuity (and the phase transition) along with possible elevation of the discontinuity by increased iron content in a region warmer than average (not allowed for in Figure 3). In this region, but not in circum-Pacific regions (e.g., Flores & Lay, 2005), a deeper comparable-sized velocity decrease is observed about 80 km above the CMB. This could be a double discontinuity as postulated by Hernlund et al. (2005). A confirmed double crossing of the phase boundary would provide the best direct measure of the superadiabatic thermal gradient in D" yet obtained.

So, what does all this mean for plumes coming from D"? As yet, the imaging of hot upwelling features from D" remains very controversial, but the possibility has not been ruled out (see Banana Doughnut page). The large-scale structure of D" leads most geodynamicists to infer a major role at least for mid-mantle convection in inducing the large-scale structure observed. If D" is indeed dynamically circulating lower-mantle material, the presence of the post-perovskite phase is expected to enhance the vertical transport of material and thermal-boundary-layer instabilities. Thus, recent seismological developments can be inferred to strengthen the hypothesis of large-scale thermal plumes rising from the D" boundary layer, likely with diameter scales of several hundred kilometers that should eventually be resolvable by seismic tomography.

The main caveat remains the uncertain nature of the chemical heterogeneity that is believed to exist at least in the low-velocity regions of Figure 1. If iron enrichment is involved, its effects may compete with those of thermal buoyancy, mitigating the tendency to develop boundary layer instabilities. Superplumes, may simply be superpiles of iron-rich materials, possibly relics of core formation or ancient subduction, and high-velocity regions may simply be relatively cool post-perovskite masses with no slabs involved. More seismological research is needed to partition the effects of thermal and chemical heterogeneity in D" and to image better the small-scale heterogeneity if we are to answer the question of whether plumes rise from the lowermost mantle.

References

• Avants, M., T. Lay, S. A. Russell, and E. J. Garnero (2006). Shear-velocity variation within the D" region beneath the Central Pacific, J. Geophys. Res., in review.
• Flores, C., and T. Lay (2005). The trouble with seeing double, Geophys. Res. Lett., Vol. 32, L24305, doi:10.1029/2005GT024366.
• Garnero, E. J., and T. Lay (2003). D" shear velocity heterogeneity, anisotropy and discontinuity structure beneath the Caribbean and Central America, Phys. Earth Planet. Inter., 140, 219-242.
• Grand, S. (2002). Mantle shear-wave tomography and the fate of subducted slabs, Phil. Trans. Roy. Soc. London (Ser. A), 3260, 2475-2491.
• Helmberger, D. V., T. Lay, S. Ni, and M. Gurnis (2005). Deep mantle structure and the post-perovskite phase transition, Proc. Nat. Acad. Sci. USA, 10.10732/pnas.05023504102.
• Hernlund, J. W., C. Thomas, and P. J. Tackley (2005), A doubling of the post-perovskite phase boundary and structure of the Earth’s lowermost mantle, Nature, 434, 882-886.
• Hutko, A., T. Lay, E. J. Garnero, and J. S. Revenaugh (2006). A folded slab at the base of the mantle imaged by migration, Nature, in review.
• Iitaka, T., K. Hirose, K. Kawamura, and M. Murakami (2004). The elasticity of the MgSiO3 post-perovskite phase in the Earth's lowermost mantle, Nature, 430, 442-445.
• Lay, T. (2005). The deep mantle thermo-chemical boundary layer: the putative mantle plume source, in Plates, Plumes, and Paradigms, G. R. Foulger, J. H. Natland, D. C. Presnall, and D. L. Anderson, editors, GSA Special Paper 388, pp. 193-205.
• Lay, T., and E. J. Garnero (2004). Core-mantle boundary structures and processes, in The State of the Planet: Frontiers and Challenges in Geophysics (R. S. J. Sparks and C. J. Hawkesworth, editors), Geophysical Monograph Series, 150, IUGG Volume 19, 25-41.
• Lay, T., E. J. Garnero and Q. Williams (2004). Partial melting in a thermo-chemical boundary layer at the base of the mantle, Phys. Earth Planet. Inter., 146, 441-467.
• Lay, T., D. Heinz, M. Ishii, S.-H. Shim, T. Tsuchiya, J. Tsuchiya, R. Wentzcovich, and D. Yuen (2005). Multidisciplinary impact of the lower mantle perovskite phase transition, EOS, 86, pp. 1, 5.
• Mao, W. L., G. Shen, V. B. Prakapenka, Y. Meng, A. J. Campbell, D. L. Heinz, J. Shu, R. J. Hemley, and H.-K.
• Mao (2004). Ferromagnesian postperovskite silicates in the D" layer, Proc. National. Acad. Sci., 101, 15,867-15,869.
• Murakami, M. K. Hirose, K. Kawamura, N. Sata, and Y. Ohishi (2004) Post-perovskite phase transition in MgSiO₃, Science, 304, 855-858.
• Oganov, A. R., and S. Ono (2004). Theoretical and experimental evidence for a post-perovskite phase of MgSiO₃ in Earth's D" layer, Nature, 430, 445-448.
• Sidorin, I., M. Gurnis, D. V. Helmberger (1999). Dynamics of a phase change at the base of the mantle consistent with seismological observations, J. Geophys. Res., 104, 15,005-15023.
• Tsuchiya, T., J. Tsuchiya, K. Umemoto, and R. M. Wentzcovitch (2004). Phase transition in MgSiO₃ perovskite in the earth's lower mantle, Earth Planet. Sci. Lett., 224, 241-248.

Reprints/pdfs of all articles by T. Lay are available upon request. Email Thorne Lay