 |
| Titel |
Nucleation and reaction rates controlled by local reaction volume and reaction-induced stress - spinel layer growth as an example |
| VerfasserIn |
Lutz C. Götze, Ralf Milke, Ralf Dohmen, Richard Wirth |
| Konferenz |
EGU General Assembly 2014
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 16 (2014) |
| Datensatznummer |
250096099
|
| Publikation (Nr.) |
 EGU/EGU2014-11585.pdf |
|
|
|
|
|
| Zusammenfassung |
| We observed the growth of spinel sensu stricto (MgAl2O4) between periclase (MgO) and
corundum (Al2O3) in thin films deposited by the pulsed laser deposition method on
crystallographically oriented single crystal substrates. The starting samples consisted of cut
and ultra polished single crystals of either corundum (parallel (0001)) or periclase (parallel
(111)) and an amorphous source layer of the respective reactant that in the very first stages
of the experiments became polycrystalline. The cutting direction in the substrate
minerals ensures that the substrate phases start to react along their close-packed
hexagonal oxygen layers which allows topotactical growth of the newly formed
spinel. The entire layer setup on the substrate crystals was only a few 100 nm thick.
The growth of these spinel product layers was monitored in-situ using a heating
attachment and synchrotron X-ray diffraction. From the reacted samples we took electron
transparent foils by the focused ion beam method and analysed them ex-situ by
TEM.
At 1000°C we found a difference in spinel growth rate between one and two orders of
magnitude between the two substrates, all other parameters held constant. At 900 and
1000 °C spinel had formed after one hour by 0.004 nm/s (900°C) and 0.034 nm/s
(1000°C) on corundum substrate, while on periclase substrate the reaction had gone
completely through the Al2O3 source layer transforming it to spinel by at least
15-30 times higher reaction rates (boundary values) and probably even faster. At
800°C no reaction occurred between periclase layers and corundum single crystals,
whereas spinel crystallized at a (linearized) rate of 46 nm/h on periclase single
crystals.
We explain our findings by the local reaction volume at the periclase-corundum interface.
Many studies (including this one) have established that spinel grows by cation exchange in a
rather immobile oxygen sublattice. This mechanism implies a negative volume change at the
Sp-Per interface (by -13 %) and a positive volume change at the Sp-Cor interface (by +17 %).
We observed in TEM images formation of porosity at the Sp-Per substrate interface
and lattice strain in the polycrystalline corundum substrate at the Sp-Cor substrate
interface, both observations being fully in line with the respective local volume
changes. We infer that the local stress field between the growing reaction layer and
the rigid single crystal substrate is responsible for the dramatic effect on both the
nucleation and reaction rates. With respect to minerals growing at dry conditions in the
Earth’s crust (granulite facies) we infer that rates of reactions might strongly depend
on local volume changes at interfaces, that reactions are impeded by compressive
reaction-induced stress applied to rigid neighboring grains, and that nucleation of
thermodynamically stable phases on grain boundaries might be suppressed by a positive
reaction volume. |
|
|
|
|
|
|