Professional Documents
Culture Documents
Preparation of Glass Beads and Powder Pills For XRF Analysis of Silicic and Calcareous Rocks
Preparation of Glass Beads and Powder Pills For XRF Analysis of Silicic and Calcareous Rocks
- Edition 2004 -
Use of the analytical facilities (including preparation labs and machines) of ETH-IMP
For a PhD or Diploma thesis a successful participation of the ETH course
“Physikalische Methoden der Mineral- und Gesteinsanalyse” is required.
ROCK SAMPLE ANALYSIS
START
field sampling:
sign-up for XRF
- representative (max. grainsize?) (ETH NO 18.5)
- homogenous
- no alteration rims, weathering,
veins
- dry transport & storage
- labeling (no ink!): max. 7 characters
e.g. ID 12345 (2 letters & 5 numbers)
preparation of
rock powders
petrography (macroscopic): (ETH NO CO 25)
- rock type
- ore content
- water & CO2 content
- mineralogy (hardest mineral?)
Loss On Ignition
thin section preparation:
use saw in lab D 14 to prepare
rock slabs for thin sections and XRF.
For XRF:
300 g if phenocrysts > 0.5 cm
100 g if phenocrysts < 0.5 cm
glass beads
powder pills
(major & minor
(trace elements)
petrography (microscopic): elements)
- rock type
- ore content
- water & CO2 content
- hydrous phases, content (e.g. amph)
- mineralogy (hardest mineral?)
Yes
XRF analysis? END
No
Table of contents
flow chart 2
table of contents 3
regulations 4
1 rock specimen preparation 5
1.1 jaw breaker (Backenbrecher) Hydraulic Press 5
1.2 Homogenize sample 5
1.3 disk mills (Scheibenschwingmühlen) 6
1.4 Contamination effects through grinding 7
1.4.1 Oxidation effect 7
1.4.2 Contamination through mortar material 7
1.4.3 Cross-contamination effects 8
1.4.4 Literature 8
2 Preparation of powder pills for trace element determination 9
3 Preparation of Glass Beads for Major Element Determination 10
3.1 Determination of the Loss on Ignition (L.O.I.) 10
3.2 casting galss beads 12
3.2.1 The Machine 12
3.2.2 Preparation of casting powders 12
3.2.3 Casting with Claisse Flux M4 13
3.2.3.1 Start-up and operation 13
3.2.3.2 The Menu and available programs 13
3.2.3.3 Activation of a casting program 14
3.2.3.4 Sequence of events during casting 14
3.3 Cleaning and proper treatment of the platinum material 15
3.4 Troubleshooting and Error messages 16
3.4.1 Casting problems 16
3.4.2 Maintenance checks 16
3.5 Casting glass beads by hand 17
4 Lab Rules 18
For publication/thesis:
Methods of preparation and calibration in:
Nisbet et al. (1979)
Dietrich et al. (1984).
Nisbet, E.G., Dietrich, V.J., and Esenwein, A. (1979). Routine trace element determination
in silicate minerals and rocks by X-Ray Fluorescence. Fortschr. Mineral., 57 (2): 264-279.
Dietrich, V.J., Carman, M.F., Wyttenbach, A. And McKee, E.H. (1984). Geochemistry of
basalts from Holes 519A, 520, 522B, and 524, Deep Sea Drilling Project Leg 73 (South
Atlantic). Init. Reports DSDP, 7: 579-601.
Reusser, E. (2000). XRF 2000. ETHZ-IMP Internal Report. [data treatment software]
3
Regulations
• Every user has to register beforehand without exception (on calendar in E 18, key for NO
CO 25 and platinum crucibles from the assistant). By signing up, full responsibility is taken
and the rules, operating instructions and safety rules of ETH are accepted.
• Each user has to fully report in the lab journal (with assistant) after the work period is
completed. A sample is given on the first page of the lab journal. The points to note are:
Always do:
• cleaned up, consumables refilled, cleaned
• platinum crucibles and keys back
• signature
No samples will be sent to EMPA for analysis, if the following conditions are not
fulfilled:
- sign-out is incomplete or missing,
- the lab is not left in good condition,
- the lists (forms) are not readable (print or fill out on computer).
• The devices and their settings must not be modified. Working on weekends and at night
requires prior approval of the assistant.
• Preconditions for use of the lab are the successful visit of the course “Physikalische
Methoden der Mineral- und Gesteinsanalyse” and a personal instruction by the assistant.
V. Dietrich/Assistances
For any questions please contact the lab assistant Adélie Delacour, NO E 22
(E-mail: [email protected] , Tel.: 27823)
4
1 rock specimen preparation
The following instructions apply to silicate and carbonate rocks. Production of pills
from non-silicate (or carbonate) rocks without prior approval and instruction of the
laboratory leader results not only in significantly more work (preparation,
measurement, interpretation) but also results into errors of the instruments and of the
final results. No observance can result into exclusion from laboratory work and possible
compensation payments. This is especially the case for rocks with an ore content >5%
(esp. sulphidic), sulphates, slag and flint.
sample selection
The sample/specimen should be fist-sized, about 200-500 g, and shouldn’t have any veins,
paint (sample numbers etc.), grease from the saw, or weathered sections.
We advise to group samples of similar composition and to treat them group wise. This
reduces the probability of contamination! Setting up a sample list with priority assignment is
strongly recommended:
- own name, exact sampling location, date, unit, etc; representative for what?
use/scientific question?
- field description (rock type), macroscopic classification
- microscopic description (texture, mineral content, rock type)
- amphibole content, ore content (sulphide, oxide), approx. content of hydrous minerals
- sample ID (alphanumeric, no symbols, max. 7 digits) and relative priority
5
1.3 disk mills (“Sacheibenschwingmühlen”, for grinding gravel to powder)
1.3.1 large thungsten carbide (WC) shatterbox (Wolframkarbid-Mahlgefäss, gross)
Caution! NEVER knock any of the pieces of the tungsten carbide (WC) shatterbox together -
they will very probably brake! The material is very brittle and expensive. Under no
circumstances attempt to grind chert (flint), the mechanic treatment pyrolizes (strongly
exothermic) the organic content, and the disk cylinder breaks to bits.
Control before grinding: are the mobile interior parts (rings and central disk cylinder) of the
shatterbox actually mobile? Are no sample residues on or beneath the rubber ring?
Is the sealing ring in the lid placed right? Is the lever handle actually in entirely closed
position?
eclogites, garnet-rich
rocks 60 6
eclogites, garnet-rich
rocks 20 5
(only for ultrapure sample preparation, as for ppt-range analyses, e.g., trace elements in
ultramafic rocks)
A maximum of 30 g sample may be ground for app. 5 minutes (depending on rock type). Use
setting 1, only! (slow velocity)
Homogenize the sample after the grinding. Fill them in sample containers, but don’t shake
them, they will dehomogenize again.
Cleaning of the shatterbox: use warm water, dry, then acetone. The outer steel rusts easily!
6
1.4 Contamination effects through grinding
The following Information should be taken serious when attempting to grind rocks in the XRF
sample preparation lab CO 25. For any questions, contact the lab assistant.
Note: The greatest systematic errors through contamination occur during rock crushing and grinding,
not in later wet chemical digestion or analytical procedures! For powder pills, grain size must
not be greater than 64 micrometers (a sieve with 250 mesh without residues). If the powders
feels and looks like flower when rubbing between your fingertips (no grains felt), then the
powder is right.
When grinding rock pieces, Fe (II) -bearing minerals heat up and react with atmospheric O2. To
critically evaluate the effect on your sample, you might consider two ideas:
- treat all your samples the same way too keep the systematic error to the same level. This makes
your results more reproducible, precise and comparable. For example, always use the same
mortar of the disk mill for your samples, use the same amount of sample, and the same
grinding time (if equality of composition allows it, see lab manual for composition-time
relation).
- Seperately determine FeO by spectrophotometry (contact V. Dietrich) on coarser powder.
J. POTTS (1987) reports up to 20% increase of Fe2O3 relative to FeO in the same sample
allready, when grinding for less than 10 minutes!
Different mortar materials cause different contaminations. the best systematic study was done
by THOMPSON & BANKSTON (1970). They compared different mortar and sieve materials by
grinding high purity quartz. Some of their results are listed below:
7
Other useful references with similar data can be found in JOHNSON & MAXWELL (1981), p. 75,
BENNET & REED (1971), p. 10ff, and in ZIEF & MITCHELL (1976), p.153.
In addition, especially Nd and Ta among other HFSE (high field strength elements) may show
contamination by using tungsten carbide.
1.4.4 Literature
POTTS, J. (1987). A handbook of silicate rock analysis. Blackie London, 622pp.
JOHNSON, W. M., & MAXWELL, J. A. (1981). Rock and Mineral analysis. Chemical analysis
series
Vol. 47. J.Wiley, New York, 489 pp. (2nd ed.)
ZIEF, M., & MITCHELL, J. W. (1976). Contamination control in trace element analysis. Chemical
analysis series Vol. 47. J.Wiley, New York, 262 pp.
BENNET, H., & REED, R. A. (1971). Chemical methods of silicate analysis. British Ceramic
Research
Association. Academic Press, London, 272 pp.
THOMPSON, G., & BANKSTON, D.C. (1970): Sampling contamination for grinding and sieving
determined by emission spectroscopy. Appl. Spectrosc. 24:210-219.
8
2. Preparation of powder pills for trace element determination
Powder pills are produced by pressing the rock powder under a maximum pressure of 450 bar.
Before pressing, an acrylic polymer resin is added in order to have the individual grains stick
together. It is dissolved in acetone (polymerizes quickly in water), which vaporizes and leaves a
grain-coated coagulate of grains behind. Under pressure, it flows into one mass and binds the
grains together. Once the pressure is released, it solidifies and the result is a solid pill that can
easily be measured in the XRF spectrometer. An earlier version used polyvinyl alcohol instead
(much more toxic). Make sure to have read the material safety data sheets (MSDS) before using
it.
Procedure:
1. Dry the powder overnight in an oven at 110°C
2. Weigh ca. 10 g dry powder in a 150 ml beaker
3. Add a 2/3 pipette of 10% Elvacite (dissolved in acetone)
4. With a thick glass rod, stir until the powder appears to dust out of the beaker
5. Repeat twice, then add 1/3 pipette and repeat twice. Total: 2.3 pipettes (2 mL)
6. Fill the dry powder into the pressing equipment, don’t stir, but distribute equally while
pouring. Make sure the piston is right side up (polished side down)
7. Insert equipment into hydraulic press and apply pressure (350 to 450 bar).
8. Leave for one minute, NOT longer!
9. Release pressure slowly by opening the release handle.
10. Securely grab the device at bottom and top, turn upside down slowly, secure piston.
11. Take bottom plate off. Careful: don’t drop piston. Cover cylinder piece bottom with plastic
cap. Turn right side up again and re-insert into hydraulic press.
12. Slowly, by hand, press the piston down into the cylinder to press the pill out into the plastic
cup (press by turning the top wheel of the press). Hold the piston with two fingers while
doing so (you feel it releasing), and stop applying pressure as soon as the pill is released,
otherwise it will break.
13. Remove assembly and carefully drop the pill onto a clean paper towel.
CRASH-TEST: The pill gets dropped onto a stainless steel surface inside the x-ray spectrometer
and may brake if it was not made well. To simulate and test its behaviour, drop it from 8 cm high
on a double paper towel. It should neither break nor have cracks. If so, grind it in the agate
mortar/pestle and add another 1/3 mL Elvacite, try again.
The assistants perform crash tests with some of the pills to ensure that they were tested. If a pill
breaks in the XRF spectrometer, it has to be taken apart and cleaned which takes ½ day.
9
3 Preparations of Glass Beads for Major Element Determination
3.1 Determination of the Loss on Ignition (L.O.I.)
Material:
weighing:
ceramic crucible (glass cupboard)
powder spatula (drawer)
dried sample powders in jars
acetone, Kimwipes
form „L.O.I.“
ignition:
- heated oven (Muffelofen)
- long tongs
- large leather gloves
- small Nomex® gloves (densely woven cotton)
- protection glasses with light filter
- fibber ceramic plate
- desiccator with blue drying agent (if it appears pink, it has to be placed into the hot air cabinet in the large
ceramic bowl for dehydration)
Weigh the empty ceramic crucible exactly and note the mass (n1) on the L.O.I. form. Slowly add
the powder with the spatula and note the total mass (n2 = ceramic crucible + sample powder). No
exactly mass of sample is necessary but to determine it with an accuracy of 0.1 mg (0.0001g).
3.1.3 burning-off
Place the crucibles (numbered at bottom) into the oven and burn them red-hot at least 1 hour at
1050°C (1070°C if amphiboles >5%vol). Take the crucibles out and place them onto the fibre
ceramic plate (at the Bunsen burner). After 5 minutes it is cold enough to handle it with a
Nomex®-glove and place it into the desiccator. This protects your fingers and ensures that no skin
fat or tissue is stuck to the crucible which would contribute to errors of the weighing result.
10
3.1.4 Calculate L.O.I.
Weigh the crucible cold, always 10 minutes after removing it from the oven. Room temperature
is reached in the crucible after ca. 5 minutes. If the crucible is still too hot, the air in the closed
weighing chamber would be heated up, expanding and press on the scale. In experiments
performed in this lab, it became obvious that the error due to adsorption of air moisture becomes
negligible after 10 minutes. After 15 to 20 minutes though, recarbonatisation plays an
increasingly significant role. Note its total mass (n3) and calculate the L.O.I. using this formula:
3.1.4 Rehomogenization
The burned off sample has to be rehomogenized in an agate mortar until the compound has a
uniform colour. Depending on rock chemistry, at 1050° C and following cooling, a sintering,
partial glass formation, or partial crystallization might occur. The grinding of the sinter optimizes
the homogeneity and reduces the analytical errors.
11
3.2 Casting glass beads
Instruction manual Claisse Flux M4 - Automated glass bead casting machine -
The powder is transferred quantitatively into an agate mortar and homogenized with the pestle.
Coarser grains of the ignition-dehydrated powder will also be grounded in this process.
Homogenisation is necessary because the agitation in the M4 wouldn’t homogenise enough. As
long as the powder is not absolutely homogenous, any splitting of powder over the mortar rim
will alter the analytical result.
The scale: Never switch the scale off, even overnight or on weekends. The weighing will be
much more stable. If the indicated masses seem unstable, recalibrate the scale.
Operational steps:
1. Close the sliding doors; check the bubble in the level ring and adjust level on the scale feet, if
necessary.
2. Hold black key down until the display shows 'CAL ----', then release. The scale begins auto
calibration. After a while the figure 100.000 is displayed (blinking).
3. Slowly push the black lever on the right hand lower rim of the back away from you. This sets
the calibration standard mass onto the weighing lever. After a while, the figure 100.000 is
displayed constantly (not blinking), and then the figure 0.0000 blinks.
4. Slowly pull black lever back to front. The scale will return to its original display.
After homogenisation the powder mixture has to be transferred quantitatively into the titanium
crucible. Please ensure that no powder is left on the flat rim of the crucible. Remove it with
rubber spatula or a towel, because the remaining powder can interfere with the casting.
12
3.2.3 Casting with Claisse Flux M4
3.2.3.1 Start-up and operation
Important:
- The machine must never be loaded with more than 11 g total mass (powder + flux).
(Otherwise the melt will overflow upon casting and the Bunsen burners will be damaged.)
- Work with M4 only with safety goggles (the pill can burst at all the time)
- Crucible and casing dish should NEVER be touched on the inner side!!! (even small scratches
are leading to contamination of the following samples)
The operation of the machine is done through a panel of the operation microprocessor.
a. open the gas valve (do not change anything on the in-between red valve!!)
b. Make sure that the ultrasonic bath is filled with water (level up to 3 cm below rim, see mark)
and plug it in.
c. press the ‘on’ button
d. Switch the used burners on.
e. Employ the casting dish
f. employ the crucible with the weighted powder
g. put down the pivot arm (position 20°above horizontal)
h. Choose the suitable program
P8 for silica and alumina
P3 for rocks
i. close the security shutter
j. start the machine through pressing the ‘start’- button
k. after the program is finished, pick up the pill using the suction cup.
In emergency case: press the ‘STOP’-button (reset the machine by pressing the ‘reset’-button)
The M4 conducts a self-test of the electronics and the supply lines (gas) after the start-up. If the
settings do not meet the requirements, the start of the machine will be interrupted automatically.
The panel shows an error message (how to solve problems see below).
In such a case, please try first to fix it yourself with the helping manual further on. If it is an
extraordinary fault, contact immediately the responsible lab assistant.
13
3.2.3.3 Activation of a casting program
Place the crucible containing the powder softly into its holder. Insert the crucible from the top.
With one hand, open the holder und put the crucible softly in between the triangle carrier. Check
that the crucible is sitting correctly, wrong placed crucible could be scratched or fall out while
rotating. Be very careful during insertion!!! If you have cast pills before, take care when placing
the crucible, because the holder is extremely hot. Always use the special crucible tongs but never
let the tongs touch the inside or upper rim of the crucible! Only touch the outside. Melt drops
will be almost impossible to remove from scratches and will contaminate further samples.
With the same tongs, touch only the casting dish on the rim, but never on the casting surface.
Any scratches will be on the glass bead measurement surface, and will cause the incident X-Ray
beam to be out of focus. Open the holder with one hand (can be hot) and place the dish from
above softly into position. Make sure the holder encloses the dish safely and check the correct
placement. Wrong placed casting dishes can be scratched or lost while rotating.
Topple over the pivot arm (position 20°above horizontal) into the right position. The machine
will rotate the dish into casting position after start of a program. Choose the desired program
(‘P+number’) and then press the green 'Start' Button on the panel. From here on, everything is
automated. During operation, the remaining time can be read from the display:
14
3.3 Cleaning and proper treatment of the platinum material
For use of the machine, 3 casting dishes and 3 crucibles are handed out to the user. During the in
between program time, the last used crucible is placed in warm (not boiling, ca. 80°C)
hydrochloric acid (HCl 5%). 5 minutes are enough to clean the crucible if the antecedent program
has been finished properly (this means, that the crucible has been shocked) and shortly placed in
the ultrasonic bath. A hot plate (max. set dial to 6), glass beaker and acid are in the lab.
Under normal conditions of use, the casting dish doesn’t have to be cleaned by this procedure. It
is only placed into the acid after the last sample for a few minutes. This is mainly to keep the
lower side of the dish clean and reflective.
Never place the dish into the acid bath together with the crucibles! Their rims would scratch the
inner casting surface of the dish.
If the room is left for any short time, switch the hot plate off and leave the platinum in the acid
bath. As well switch the main gas supply off if you leave the room (in door frame, red button =
OFF and green = ON)
Waste acid has to be poured into the white 10 L waste acid container next to the sink. Place
container into the sink, unscrew top, pour waste acid in (GLOVES, SAFETY GOGGLES, LAB
COAT!!!). If the container filling reaches the marked top level (ca. 10 cm below top), inform lab
assistant for proper disposal. Do not pour waste acid into the sink. NEVER pour water into the
waste acid container, explosion hazard!
Under no circumstances are the inner parts of the crucible or the casting dish ever to be
touched by the tongs!! Even slightest contacts scratch the platinum surface. On these scratches,
small melt drops remain, leading to lesser quality melts and contaminated glass beads. The
platinum crucibles and/or dish would have to be replaced which takes a long time (>1 month) and
is extremely expensive (depending on platinum price ca. 5000, - CHF per item!). After extensive
use, the casting dish will deform to leave a convex surface on the analytical side of the glass
bead. Scratches and uneven surface can be flattened and polished by about 4 hours work. Uneven
surface leads to a defocused incident beam during analysis.
15
3.4 Troubleshooting and Error messages
3.4.1 Casting problems
Because the M4 is new, we do not know all the established problems. Please tell us what kind of
problems you had, so we can improve the programs or manuals.
a. The bead does not cover the entire floor of the casting dish:
a1: The bead was caste a few degrees too cold. Check if the valve on the gas bottle is open, make
new powder mixture and cast again.
a2. The whole mass is too small (6g instead of 9g flux and sample). Cast the pill again with the
exact amount of flux and sample (1.5g sample/7.5g flux)
a3. The melt is too viscous. Add a few drops of a non-wetting agent to a new weighted solution
and cast again. (Still in experimental mode)
b. Small melt remnants stay behind in the crucible:
The melt is too viscous. Add to the beginning powder a non-wetting agent (2 drops of LiBr) and
cast the pill again. (Still in experimental mode)
c. The bead has a small gas bubble trapped in its rim:
The bubble won’t change the quality of the analysis, because the outer millimetres of the bead are
not incorporated into the measurement grid. Treat the casting dish for 5-10 minutes in warm acid.
d. the pill splits
d1. The casting dish is too cold. Open the gas valve until the maximum is reached (do not touch
the in-between valve!!) and cast the pill again.
d2. There are external particles in the casting dish or the melt. Clean the casting dish before the
next cast by putting it for 5 min in the warm hydrochloric acid bath.
d3. The measurement of the powder was too improper (not 1:5 rate sample-flux). Cast the pill
again with exacter measurements.
16
3.5. Casting Glass beads by hand
If the machine is down or in repair, glass beads have to be casted by hand. The Cons are the
sweaty character of the work, the Pros the quick procedure, the fast cleaning of the thinner Pt
crucibles, and the almost unlimited number of pills possible every day.
3.5.1 Preparations
The sample powders are to be prepared exactly the same way as in the procedure for the
automated casting. The thinner crucibles are numbered. Make sure that the ultrasonic bath is
filled with water (level up to 3 cm below rim, see mark) and plug it in.
3.5.2 Melting
Place the powders in Pt/Au (95/5) crucibles into the oven and melt them during about 30 minutes
at 1150°C. Li2B4O7 lowers the melting point to ca. 650-700°C, the melt is of low viscosity and
can easily be homogenized by pioving. Sometimes, solid to viscous powder remnants stick to the
inner walls. They have to be reunified with the melt. Tilt the crucible until the remnants are
covered by melt and place it into the oven. Repeat this procedure until the inner walls are entirely
free of remnants. If the melt still shows bubbles after 10 minutes, keep shaking the crucibles and
leave them in the oven a little longer.
3.5.4 Casting
Remove the melt quickly from the oven and empty the crucible onto the casting dish.
Immediately put the crucible down, and adjust the casting dish to horizontal (note symmetric
cooling patterns). Immediately switch off the gas (not the air) to let the melt quench.
17
4. Lab Rules
The following violations and non-observance of the lab rules may lead to
payment of compensation to the IMP and to exclusion from any further lab
work:
18