Analysis of the morphology, composition, and spacing of exsolution lamellae can provide essential information to understand geological processes related to volcanism and metamorphism. We present a novel application of APT for the characterization of such lamellae and compare this approach to the conventional use of electron microscopy and FIB-based nanotomography.
Examining variations in mineral chemistry can shed light on changes in volcanic activity, and allow researchers to obtain timescales of volcanic processes to better understand any potential hazards. Atom probe tomography allows an unprecedented 3D visualization of mineral exsolved phases while measuring their chemical composition at atomic scale. We are currently applying the technique to the characterization of pathological mineralization, for example, in the case of kidney stones.
This method can be applied to volcanic systems where eruption transitions can occur over small timescales. These transitions are recorded in the minerals over very small spatial scales. Begin by pouring one gram of the sample into a 10-centimeter glass Petri dish and wrap a three by three-centimeter sheet of weight paper around a 10-gauss magnet.
Use the magnet to pull magnetite-rich grains between 100 and 500 micrometers in diameter from the ash sample and place the grains in a 32-micrometer pore, 8-centimeter diameter stainless steel sieve. Use a squeeze bottle of deionized water to flush the smaller adhering ash particles through the sieve for 20 to 30 seconds and allow the grains to air dry for 24 hours. The next day, affix any clean and dry ash particles to sample mounts suitable for a secondary scanning electron microscope and image the particles in secondary electron mode at a 15 to 20-kilovolt accelerating voltage and at a working distance of 10 millimeters to select the five to 10 best candidates for further analysis.
The selected grains should be predominantly magnetite. Affix the selected ash grains to a piece of clear tape and surround the samples with a one-inch diameter hollow mold that has been internally coated with vacuum grease, then fill the mold with epoxy resin mold. Once the epoxy is cured, remove the sample from the mold and peel tape from the bottom.
The ash grains should be partially exposed. Polish the epoxy-cast ash grains with silicon carbide grinding paper of five different grit sizes, from the highest to lowest grit size in a figure eight motion for at least 10 minutes per grinding paper. In between grit sizes, sonicate the sample in a bath of deionized water for 10 minutes.
After each last polish, check the sample under a microscope to ensure that no polishing grit is present and that the sample surface is free from scratches. Next use polishing cloths to polish the epoxy-cast ash grains with consecutive one and 0.3-micrometer alumina polishing suspensions in a figure eight motion for at least 10 minutes, sonicating the sample in deionized water for 10 minutes between suspension sizes. After the second suspension polish, check the sample under the microscope to ensure that no suspension is present and that the sample surface is free of scratches.
At the end of the polishing procedure, the epoxy surface should be smooth and the ash grains should be flat and well exposed. Using an available sputter coating device, coat the sample surface with an approximately 10-nanometer thick carbon-conducting coat and obtain backscattered electron images of the ash grains with the electron microscope at a 15 to 20-kilovolt accelerating voltage and a working distance of 10 millimeters to determine the location of exsolution lamellae in the magnetite. Before beginning the focused ion beam procedure, sputter coat the sample surface with a 15-nanometer layer of copper to avoid electron charging and sample drifting.
Next use a focused gallium-ion beam in a dual-beam scanning electron microscope on the polished section of interest containing the lamellae over a 1.5 by 20-micrometer region at 30 kilovolts and seven pascals. Use the ion beam to mill three wedges of material below three sides of the platinum rectangle and insert the gas-injection system to weld the wedge to an in situ nanomanipulator using gas-injection system-deposited platinum before cutting the final edge free. Using the gallium-ion beam, cut 10 one to two micrometer-wide segments from the wedge and sequentially affix the wedges with platinum to the tops of silicon posts of a microtip array coupon.
Shape and sharpen each specimen tip using annular milling patterns of increasingly smaller inner and outer diameters, starting at 30 kilovolts to produce the specimen geometry necessary for atom probe tomography and finishing at an accelerating voltage of five kilovolts to reduce the gallium implantation and to obtain a consistent tip-to-tip shape. For atom probe tomography acquisition, mount the micro coupon with the sharpened tips welded to the silicon posts onto a specimen puck and load the puck into a carousel for placement inside the local electrode atom probe. Insert the carousel inside the buffer chamber of a local electrode atom probe equipped with a picosecond 355-nanometer ultraviolet laser and turn the head of the laser.
After calibration, achieve a vacuum in the analysis chamber at or below six by 10 to the negative 11 torr and use a transfer rod to insert the puck specimen into the main analysis chamber. Then move the specimen puck to align the micro coupon with the local electrode to select the tip and update the database to indicate the tip number. In this analysis, four titanomagnetite specimen tips were successfully extracted from the single crystal and analyzed by atom probe tomography.
Two of the specimens demonstrated homogenous concentrations of both iron and titanium throughout, indicating that lamellae were not intersected. The other two specimens exhibited zones with variable concentrations in iron, oxygen, and titanium. 3D reconstructions of the atom probe tomography data permit a precise measurement of the intralamellar spacing and provide length scales that average between 14 to 29 nanometers with a one-sigma value of two nanometers for both specimens.
In addition to these measurements, atom probe tomography permits the extraction of chemical information across these lamellae at a high spatial resolution through the analysis of proxigrams using the point zero as the intersection between the lamella and the host mineral. Atomic concentrations of titanium in the crystal confirmed that the fragment is indeed a titanomagnetite and are consistent with previous petrologic analyses of Soufriere Hills Volcano eruptive products. These proxigrams also confirm that the composition of the lamella matches that of ilmenite.
Preparing the wedges to capture the lamellae is critical to the FIB SEM sample preparation as well as sharpening the tips to the correct dimension. Transmission electro microscopy can also be performed to verify the lamellae dimensions and the interlamellar spacing. This technique could allow vulcanologists to calculate the timescales of eruptive activity to better understand the potential hazards of active volcanoes.
