RockJockML Sample Preparation

 

This procedure is adapted from: 

Eberl, D.D (2003) User’s Guide to RockJock—A Program for Determining Quantitative Mineralogy From Powder X-Ray Diffraction Data, U.S. Geological Survey Open File Report 03-78, November 30, 2009 Revision. 

The full-pattern fitting method used by RockJockML depends on the shapes and sizes (not just locations) of the powder XRD peaks for the sample being very similar to those in the database of mineral standards, because calculated patterns optimized to match a sample pattern are simply linear combinations of the standard patterns in the database.  Therefore, samples must as nearly as possible be prepared in the same way as the standards, so as to ensure 1) small enough powder particles, 2) a reproducible size distribution of particles, and 3) random orientation of particles.  All of these factors can strongly affect peak shapes and sizes.

Here we present a short version of the sample preparation procedure, followed by a long version in which more context is given, including where to get the necessary equipment. 

The Short Version:

1. Grind sample to pass through a 250 µm (60-mesh) sieve.

2. Spike 1.000 g sample with 0.250 g of corundum and mix.  (Don’t use just any corundum.  See below.)

3. Wet mill the spiked sample for 5 min. with 4 mL ethanol in a McCrone XRD-mill.  

4. Ensure random orientation of particles by either mixing with a non-polar solvent like Vertrel or hexane, or by spray-drying.

A. Solvent Method.  Dry the sample overnight at about 80° C.  Put the sample in a 20-25 mL plastic scintillation vial with three 10 mm diameter plastic balls, and shake it for 10 minutes in a shaker mill to disaggregate and mix.  Add the solvent to the mixture in the ratio of 0.5 mL vertrel to 1 g of pure clay, shake again for 10 minutes in the shaker mill.  Add less solvent if the sample contains less clay.  Pass the sample through a 250 µm (60-mesh) sieve and place it in a side-loading XRD mount, with the top against a frosted glass surface. Gently tap the side against a hard surface to pack it.  If the sample is too coagulated to pass through the sieve, leave the vial open under a fume hood for several minutes to evaporate the excess solvent, and then shake and sieve it again.

B. Spray Drying.  Instead of drying the sample, leave it in the ethanol used in the wet milling step.  Dilute or evaporate the ethanol to get the desired slurry consistency.  (This takes some experience, but we have had reasonably good luck just using the slurry as-is).  Spray into the top of the heated spray dryer with an air brush, and collect the sample from the bottom.  Side-load the sample into an XRD mount as in the solvent method.

5. Side-load the sample in a powder XRD holder, with a frosted glass surface against the top surface.  Gently tap the side of the holder on a hard surface to pack the sample in without compressing it. 

6. Run the sample from 5-65° 2\theta, 0.2° per step, 2 s per step.  If you want a less noisy pattern, increase the step time.

 

The Long Version:

Step 1:  Initial Grinding.  Use a mortar and pestle, if needed, to powder the sample so it can pass through a 250 µm (60-mesh) sieve.  If some material does not pass through the sieve, collect it and grind it until it does.  (Some phases in the sample might disproportionately occupy larger size fractions, so failing to further grind excluded material until it passes through the sieve might result in changing the proportions of the phases in the sample.)

Step 2:  Spike with Internal Standard.  Mix 0.250 g of corundum with 1.000 g sample.  For best results use corundum powder that is already in the target particle size range (see Step 3 below).  It is probably best to use the same stock of corundum as was used to prepare the standards.  The standards in the RockJockML database were prepared using a 99.9% pure, synthetic Al-oxide powder with particle diameters in the 3-4 µm range.  It was manufactured by American Elements (10884 Weyburn Ave., Los Angeles, CA  90024, 310-208-0551, customerservice@americanelements.com, http://americanelements.com), Product Number:  AL-OX-03M-P.3T4UM).  When visiting the website, type “aluminum oxide powder” in the search field, click on the first result, and then click on the “Pricing” button for the AL-OX-03-P product.  This will bring up a form for requesting a quote.  (Right now it's about $1,000 per kg.)  Make sure you specify the particle size range needed in the notes.  At this writing, the direct link to the order form is here.

Step 3:  Wet Milling.  Grind the corundum-spiked sample with 4 mL ethanol for 5 min. in a McCrone XRD-mill.  After pouring off the sample, rinse out the remaining sample by adding 15 mL ethanol and running in the mill for 30 seconds.  Pour off the liquid and repeat once more.  

  • The manufacturer is The McCrone Group (630-887-7100, https://www.mccrone.com/).   The mill shakes a container with a number of stacked cylindrical grinding elements.  This configuration produces a distinctive grinding action that tends to quickly produce a very narrow, reproducible range of particle sizes (typically < 5 µm), which is important for RockJockML. 
  • You can buy sets of agate (cryptocrystalline quartz), corundum, or zirconium oxide elements from McCrone, and the RockJock 11 User Manual (D. D. Eberl) recommends using the zirconium oxide.  A rationale for this recommendation is that quartz is significantly less hard (Moh’s 7) than corundum (Moh’s 9), and so can wear down and add the common mineral quartz to the sample.  The corundum elements are actually made of a corundum-based ceramic, so how well it holds together will depend on the type of ceramic.  Any wear on the corundum elements would add to the internal standard.  The zirconium oxide elements are made of a yttria-stabilized zirconium dioxide ceramic, which can have hardness approaching that of diamond (Moh’s 10).  Furthermore, even if some zirconium dioxide does contaminate the sample, it would be more recognizable as a contaminant than added quartz or corundum.  That said, the McCrone Group actually claims that in their tests, the agate grinding elements contaminated samples the least.  The reason for this may be that degradation of the grinding elements is primarily due to interactions with each other, rather than interactions with samples.  Ultra-hard ceramics, which are made of more-or-less fused particles, could degrade more when rubbing against each other than naturally formed agate.  Nevertheless, all the standards in the RockJockML database were prepared using the zirconium oxide elements.  Therefore, we recommend using the zirconium oxide grinding elements or, in a pinch, the agate—but not the corundum.  

Step 4:  Ensuring Random Orientation.  This can be done by mixing the sample with a non-polar solvent like Vertrel(TM) or hexane, or via spray drying.  

A. Solvent Method.  Dry the sample overnight at about 80° C.  Put the sample in a 20-25 mL plastic scintillation vial with three 10 mm diameter plastic balls, and shake it for 10 minutes in a shaker mill to disaggregate and mix.  Add the solvent to the mixture in the ratio of 0.5 mL solvent to 1 g of pure clay, shake again for 10 minutes in the shaker mill.  Add less solvent if the sample contains less clay.  Pass the sample through a 250 µm (60-mesh) sieve and place it in a side-loading XRD mount, with the top against a frosted glass surface. Gently tap the side against a hard surface to pack it.  If the sample is too coagulated to pass through the sieve, leave the vial open under a fume hood for several minutes to evaporate the excess solvent, and then shake and sieve it again.

  • Vertrel(TM) is a specialty solvent used for a variety of cleaning applications.  Just search for it on the internet to find the best price.  Hexane is much less expensive, but not as pleasant to work with.  
  • Shaker mills can be bought from SPEX Sample Prep (https://www.spexsampleprep.com/accessories-list/mixermill). The 8000-Series Mixer/Mills have been standard equipment in many geology labs for some time, so sometimes you can pick up older models on eBay.  
  • We bought a case of 1000 20 mL HDPE plastic scintillation vials from Stellar Scientific here.  
  • It probably doesn’t matter much which brand of scintillation vials you buy, except that we used the ones just mentioned to design a SPEX mill adapter that can be 3D printed, so you can shake three samples at a time.  Click here to download the files.    

B. Spray Drying.  Instead of drying the sample, leave it in the ethanol used in the wet milling step.  Dilute or evaporate the ethanol to get the desired slurry consistency.  (This takes some experience, but we have had reasonably good luck just using the slurry as-is).  Spray into the top of the heated spray dryer with an air brush, and collect the sample from the bottom.  Side-load the sample into an XRD mount as in the solvent method.

Step 5. Side-Load XRD Mount.  Side-load the sample in a powder XRD holder, with a frosted glass surface against the top surface.  Gently tap the side of the holder on a hard surface to pack the sample in without compressing it. 
 
Step 6. Run the Sample.  Run the sample from 5-65° 2\theta, 0.2° per step, 2 s per step.  If you want a less noisy pattern, increase the step time.