Wednesday, October 18, 2017

Interactive elemental maps

This post features interactive elemental maps that were created for plotting electron microprobe analysis (see previous entry). Now you can visualize the output without needing to install any software. Elemental maps show distribution of concentration for each element in spacial 2D coordinates, X and Y. In this particular example a komatiitic basalt with some alteration is used. You can navigate the analysis by choosing different element and hovering the mouse over the image to see the concentration of the chosen element (Z axis). 

The colors reflect the concentration corresponding to the wt. % value on the color bar. When Al2O3 is selected (default), elongated grains of clinopyroxene are shown with blue color (relatively ow Al2O3). This can be compared to concentration of other elements: high Ca, high Mg. This is a pyroxene pigeonite-augite. If the sample studied in SiO2 mode, one can clearly see quartz - red patches denote high concentrations of silica.


The code is written in R with plotly package for plotting ans ShinyApp package  to create an interactive output. Let me know if you would like to try this code for your own elemental maps. 

Saturday, October 14, 2017

Automated petrography: electron microprobe mapping

 altered basalt from modern seafloor

In the world of geochemistry every sample requires a detailed description. To make sense out of isotopic/chemical measurements it is important to know what each sample is composed of and how much of each component is present. One of my project deals with ancient altered basalts - something not a lot of people like to look under microscope at because of absence of primary minerals and general "mess" composed of secondary minerals. Basalts are commonly fine-grained in the first place and alteration can make them particularly microscopic; they are altered to fine-grained aggregate of  secondary minerals - chlorites, serpentines, epidotes, amphiboles and other minerals of hydrothermal origin. Even under petrographic microscope these minerals are sometimes indistinguishable from each other. In my study, I used microprobe analysis to go around this problem. With help of John Donovan from the electron microprobe analysis lab at the University of Oregon, we programmed the instrument to create maps of elemental distribution. Resulted images are posted below. I selected most contrasting elements for each sample and plotted them using R package ggplot2. Each pixel in these images represent an electron microprobe analysis.
 Panel image: A - altered basalt with komatiitic texture. Image size is 500 um along the side; B - a quartz-calcite vein with epidote, amphibole, diopside and garnet; C - fragment of altered pillow that is altered to muscovite, chlorite, amphibole and albite (200 um along the side). Abbreviations: albite (ab), amph (amphibole), cc (calcite), chl (chlorite), di (diopside), gt (garnet), ep (epidote).

Now having these images and tens of thousands of analysis I can derive composition of each mineral with very high accuracy. Using some programming algorithms I extracted a composition of a given mineral from each image and processed them to compute formula units of each mineral. Below is results plotted on classification diagrams for epidote (A), pyroxenes (B), Ca-amphiboles (C) and chlorites (D). White transparent fields are plotted using literature data on modern altered basalts (drill hole ODP 504B, eastern Pacific Ocean).

 Panel image: A - analyzed epidotes plotted on classification diagrams using end-members - epidote (Ep), clinozoisite (Clz) and two times piemontite (2Pmt); B - analyzed clinopyroxenes plotted on classification diagram for calcic pyroxenes; C - analyzed calcic amphiboles plotted on classification diagram (after Leake, 1978); D - analyzed chlorites on classification diagram (after Hay, 1954). White transparent fields are using published results for secondary minerals from the drill hole ODP 504B. Analysis of my ancient basalts show that they are composed of minerals pretty similar to those formed on modern day seafloor!

Monday, April 3, 2017

The most convincing evidence for CO2 increase

Today global warming and climate change are widely discussed. Many people refer to "scientific evidence" that proves that climate change is caused by humans, by burning fossil fuel in particular. However, I rarely hear on public media discussion of the actual data and interpretations that serve as the scientific evidence for human-induced climate change. May be there would be more productive conversation between different members of our society on climate change if the public media discussions included some background on the scientific evidence. 

Human-induced climate change has expressed itself in elevated concentrations of CO2 in the atmosphere. Elevated CO2 is a consequence of releasing carbon in the atmosphere through burning fossil fuel like coal and petrol. Among multiple evidences for such change, glacial ice cores represent the most convincing indication of human-induced elevated CO2 levels. As glacial ice forms every year it trap atmospheric air in the form of bubbles. Extracted from drill holes, glacial ice cores can be accurately dated to the recent time line (back to several thousand years). Extracted air bubbles are analyzed for concentration of gases and isotopic composition of the gases. This graph below copied from "Stable Isotope Geochemistry" by Hoefs represents the most convincing evidence for human-induced climate change.


Clear increase in CO2 concentration (plot a) starting at about 1850 marks the bloom of the industrial era. Plot B shows the carbon isotopic composition (δ13C, ‰; delta carbon thirteen) of CO2 from the atmosphere. Starting from industrial era, it becomes more negative, meaning CO2 molecules in the atmosphere are increasingly depleted in the heavier isotope of carbon - 13C - with respect to the lighter carbon, 12C. Organic matter, like coal, is extremely "light" carbon, means it is depleted in heavy carbon 13C. Burning such "light" source of carbon makes CO2 in the atmosphere "light" as well. Just for clarification, I include here a diagram from the same book on isotopic composition of all carbon sources known to Earth. It show that organic matter (such as fossil fuel) is the source of isotopically "light" carbon.


It was pointed out that the isotopic composition of CO2 from volcanoes should be shown too since volcanic emanations contribute to the level of atmospheric CO2. Valid point. The δ13C from volcanoes is almost identical to the δ13C of the air. See attached diagram. 
The image is taken from a Pennsylvania State University website

Monday, March 6, 2017

How it's done: fluid inslusion measurements (VIDEO)

Imagine a crystal of quartz growing from a hot hydrothermal fluid. Since no mineral grows without imperfections, quartz will trap some of the surrounding hydrothermal fluid (see picture below). As quartz cools down, the trapped fluid inside cools down too and it shrinks. That volume change is expressed in appearance of vapor bubble. Trapped fluid now consists of two phases - vapor and liquid. Fluid inclusions like that have a bubble and liquid now and are commonly found in hydrothermal quartz. Samples of quartz in which such fluid inclusions can be found are used as thermometers. One can heated up a fluid inclusion with bubble and liquid to make liquid and vapor become one. Such transition is called homogenization. The temperature of this transition represents the temperature when the quartz was formed. This method is well established and applied by a wide range of geoscientists, mostly economic geologists and geochemists.

 Healing of a imperfection (crack) in quartz. Surrounding fluid gets trapped. Adopted from Roedder, 1984.

An example of a fluid inclusion with bubble

Moreover, information about salinity of the hydrothermal fluid can be extracted from fluid inclusions. The freezing point of pure water is ~0°C. The freezing point of saline water is lower than that. For example seawater freezes at -2°C. Fluids with higher salinity freeze at lower freezing points. Here's the set up at Mark Reed's lab, University of Oregon for conducting heating and freezing measurements. 
 Fluid inclusion thermometry set up at University of Oregon, Mark Reed's lab

For homogenization measurements a heating element is used in combination with air flow. The element warms up the sample and thermocouple is used for measuring the temperature. A researcher regulates the temperature and observes a fluid inclusion with a bubble and liquid under the microscope. When the bubble and liquid become one (homogenize), researcher presses the footswitch and the temperature reading on the screen freezes. Careful measurements and keeping good track of inclusions (logging, taking pictures, drawings) produces the best results. Here how the stage for these measurements look like.
 Amount 300 microns thick double-polished sample of quartz is held in the camera, between 6 layers of glass and pinned down by thermocouple tip. The heat blows from the left (where the paper label was burnt) 

For conducting freezing experiments, compressed nitrogen gas is used to transfer liquid nitrogen through a set of tubing into the same stage. The sample experiences liquid nitrogen temperature of -195°C. After a fluid inclusion of interest becomes frozen (which is seen under microscope), a researcher needs to warm up the stage until ice crystals start to melt away and record when the last ice crystal disappears. That temperature is the freezing point.

Here's videos showing both heating and freezing experiments. Different inclusions are used. The stage of the microscope moves slightly so the field of view merges a little bit. I tried to maintain good focus through out the videos.The field of view in these videos is about
 Heating a fluid inclusion that homogenizes via vapor+liquid->liquid at 241°C.

Freezing an inclusion that have freezing point of -8°C which corresponds to salinity of about 10 wt. % NaCl.

Tuesday, September 27, 2016

Another outcrop, another story

Northern New Mexico offers to see a beautiful story of volcanic eruptions associated with Rio Grande rifting that was active. The river of Rio Grande carved canyon in Taos plateau exposing extensive flows of 5 million year old basalts and minor andesites and dacites. Here's one particular outcrop that provides a snapshot of geological history of the region:

The story goes like this (from bottom up):

Paleosol (old soil) marks the period of surface erosion. Underlying rock is exposed to the atmosphere and is being broken down to produce soil. The paleosol contains remnants of root channels and other traces of  life on land. Based on the ages of underlying rocks the soil was developing about 5 million years ago.

Baked and oxidized paleosol was developed because hot lava was flowing on top of moist soil. At high temperature, available iron and water and oxygen was reacting resulting in oxidation of that iron.

Crumbly breccia, aa-lava newly erupted lava was flowing on top of the soil, cooling down quickly which resulted in partly solidified rock flowing in highly viscous lava. Solidification of the material produced crumbly, chunky aggregate commonly called 'a'a-lava. The rock was erupted about 4.8 million years ago.

Massive dacite is produced by massive outflow of dacitic lava that was hot enough to flow and cool down continuously.

Sheared dacite reflects interesting property of the silisic lava. Because of high SiO2 content dacite polymerizes more so than a mafic lava, resulting in its comparatively high viscosity. So when it cools down it viscosity is so high, that lava can't flow anymore, it starts to shear. 

Geologist has something to learn about Taos plateau volcanic field. He was erupted about 25 years ago.