Microscopy Microanalysis
Microsc. Microanal. Page 1 of 8 doi:10.1017/S1431927609991322
AND
© MICROSCOPY SOCIETY OF AMERICA 2010
Tephra from Ice—A Simple Method to Routinely Mount, Polish, and Quantitatively Analyze Sparse Fine Particles Stephen C. Kuehn* and Duane G. Froese Department of Earth and Atmospheric Sciences, University of Alberta, 1-26 Earth Sciences Building, Edmonton, AB T6G 2E3, Canada
Abstract: A method involving a graphite substrate has been developed for the mounting and analysis of sparse, fine particles from a liquid suspension to enable improved study of volcanic ash ~tephra! and atmospheric dust preserved in glacial ice. Unpolished grains may be studied by scanning electron microscope–energy dispersive spectrometry ~SEM-EDS! at full vacuum without the need for a conductive coating due to the close proximity of the underlying graphite. The same grains in the same relative positions may be subsequently examined in polished mounts by a variety of methods including SEM-EDS, electron probe microanalysis, laser ablation– inductively coupled plasma–mass spectroscopy, secondary ion mass spectrometry, and optical microscopy. Particles as small as 3–5 mm may be routinely and easily prepared for analysis as polished grains, and particles of significantly different sizes may be exposed simultaneously. The general approach also offers significant flexibility, including both single- and multiple-sample mounts, and may be adjusted to suit a variety of samples and analytical methods. Key words: specimen preparation, fine particle analysis, tephra, ice core tephrochronology, major-element analysis, scanning electron microscopy, electron probe microanalysis
I NTR ODUCTION Sparse, fine particles prepared in aqueous suspension, including atmospheric dust and volcanic ash ~tephra! particles from glacial ice cores, frequently have been studied using microfiltration on track-etched polycarbonate membrane filters followed by scanning electron microscope–energy dispersive spectrometer ~SEM-EDS! analysis ~e.g., Germani & Buseck, 1991; Zielinski et al., 1997; Zdanowicz et al., 1999; Yalcin et al., 2003; Kekonen et al., 2005!. Using wavelength-dispersive X-ray analysis of polished grains by electron probe microanalysis ~EPMA!, it is possible to produce data of greater precision that are also free of distortions due to surface irregularities. In the case of tephrochronology, the resulting higher quality data can provide for more robust identification and correlation of individual tephra beds. Preparation of fine particles from glacial ice cores and trace amounts of tephra extracted from sediment cores as polished grains has been typically achieved by centrifuging the sample, transferring the particulate material to a glass slide by pipette, embedding in epoxy resin, and polishing to expose the grains ~e.g., Dugmore et al., 1995; Davies et al., 2008!. This process, however, requires very careful and exacting polishing to expose but not remove the fine grains, and to simultaneously expose grains of different sizes. An alternative approach emReceived August 21, 2009; accepted November 30, 2009 *Corresponding author. E-mail:
[email protected]
ployed by Dunbar et al. ~2003! involves transferring particles from a filter to a drilled epoxy block using double-sided tape, followed by filling the drilled hole with resin to encase the particles, removing the tape, and polishing the surface. This method simplifies polishing and is better for exposing all of the grains on a single surface. To prepare a large number of samples taken from an ice core recovered from Prospector Russell Col on Mt. Logan ~Fisher et al., 2008!, a new procedure has been developed for fine particle mounting that involves the use of a graphite substrate. This approach is in some respects a hybrid of the latter two methods described above, but it offers several advantages and significant flexibility. Particles as small as 3–5 mm may be routinely and easily mounted in epoxy resin and polished, and particles of significantly different sizes may be exposed simultaneously during polishing. Single- and multiple-sample mounts may be prepared, and it is possible to analyze the very same grains in unpolished mounts and later in polished mounts. Unpolished grains also may be studied at full vacuum without the need for a conductive coating due to the close proximity of the underlying graphite.
S AMPLE P R EPARATION Concentration of the Particles Samples of ice obtained from the Geological Survey of Canada’s core repository in Ottawa were placed in 500 mL
2
Stephen C. Kuehn and Duane G. Froese
high-density polyethylene bottles for transport and storage and were allowed to melt. After their arrival in Edmonton, sample bottles were placed in a HEPA-filtered drying oven at ;908C in a clean laboratory to reduce the water volume to 10–15 mL. To help prevent contamination, the exterior of each bottle was rinsed to remove any adhered particles prior to evaporation. Next, the particles were concentrated by centrifuge. Just before the samples were centrifuged, each bottle was placed briefly into an ultrasonic bath to ensure that all particles were suspended. Each sample was subsequently decanted into a 15 mL centrifuge tube and centrifuged at 3000 rpm for 5 min to concentrate the particles at the bottom of the tube. Next all but 1 mL of water was removed from the tube by pipette. Each sample bottle was then rinsed repeatedly using a total of 10–15 mL of deionized water to transfer any remaining particles to the centrifuge tube. Next, the samples were centrifuged a second time, and all but 0.5 to 1 mL of water was removed by pipette. The concentrated samples were stored in centrifuge tubes until mounting. Single- and multiple-sample mounts ~Fig. 1! were prepared for analysis by EPMA using a graphite substrate coated with a thin adhesive film and later covered by an acrylic disk ~Fig. 2!. Graphite planchets and larger graphite plates were obtained from Ted Pella, Inc. ~Redding, CA! and GraphiteStore.com, Inc. ~Buffalo Grove, IL!, respectively.
Preparation of Mounts for the Study of Loose or Polished Grains (Procedure A)
4.
5.
6.
7.
To prepare a single-sample mount, the following procedure was followed: 1. Using 2000 grit ~about 10 mm! sandpaper on a flat lap wheel, gently polish one side of a 12.5 mm graphite planchet to obtain a flat surface. Larger planchets or graphite blocks may also be used. It is important to hold the planchet steady during polishing as any rocking will cause increased wear near the edges and produce a convex surface. Multiple planchets or blocks may be polished simultaneously by attaching them to a large glass slide or other solid, flat surface using thin double-sided tape and leaving a 3–4 mm space between the planchets or blocks ~Fig. 1A,B!. If the blocks or planchets are unequal in thickness, it may be necessary grind them with a coarser grit after they have been attached so that all surfaces will lie in the same plane. 2. Apply a small amount of low-viscosity epoxy to the polished surface to fill any exposed pore space in the graphite and then allow the epoxy to cure. Vacuum impregnation may be used to improve pore filling. 3. Polish the planchet again to expose the graphite and produce a flat, nonporous surface. It is important that the planchet not be ground too far at this stage to avoid
8.
exposing unfilled pore space ~Fig. 2!. Inspect each polished planchet on a binocular microscope and set aside any that have too many open pores. Affix the planchet to the center of a 47 mm diameter petri dish using a piece of double-sided tape. Stackable petri dishes are suitable for storage and transport of sample mounts. The cover also helps prevent contamination by foreign particles during sample processing. If the sample is to be studied first as unpolished grains and later mounted in epoxy and polished, use a scribe to add a label near the edge of the planchet or block. Also add at least three grooves that will accept short pieces of fine wire. Remove any graphite debris from the surface. The wire will provide transferrable reference coordinates. For this study, pieces of fine wire about 200 mm in diameter and 1–2 mm long were used. These were placed in grooves with a depth approximating one-third to one-half of the wire diameter. Working in a fume hood, dissolve about 10 Avery Spot-O-Glue TM tabs in 10–20 mL of chloroform ~trichloromethane!. If sufficiently diluted, commercially available liquid SEM adhesives that remain sticky when dry are likely to work equally well. Apply one or more coats of the adhesive solution to the surface of the planchet using a strip of filter paper ~Fig. 1C!. The objective is to obtain a very thin and uniform coating. In general, less adhesive is better than too much. If planning to add wire, apply additional adhesive to the grooves to ensure that the wire will be held firmly. If desired, place pieces of fine wire in the grooves made earlier to provide transferrable reference positions ~Fig. 1F,K!. For the best results, ensure that the end faces are vertical when the wire pieces are placed in the grooves. A minimum of three pieces of wire are needed, but it is best to include at least two extra pieces in case any prove to be unsuitable after polishing. Working in a laminar flow hood, transfer several drops of water containing suspended particles onto the planchet or block using a pipette and allow the water to evaporate completely. Keep the mount partially covered using the top of the petri dish to provide an additional guard against contamination. Evaporation may be accelerated using a hot plate set at 50 to 608C. Repeat the sample application if necessary ~e.g., for a very dilute suspension!. If particles are abundant in the suspension, be careful not to apply too much sample. Once the mount is dry, the unpolished grains may be studied on the SEM ~Fig. 1E,F!. A conductive coating is unnecessary even at full vacuum as the close proximity of the graphite substrate to the fine grains results in minimal charging. If the grains are to be later mounted in epoxy and polished, a conductive coating is best avoided as it may prevent the epoxy from adhering to the grains. Continue to the next step if study of polished grains is desired. Note that when the grains have been polished, their positions will be inverted ~Fig. 2!.
Method for Fine Particle Mounting
Figure 1. Photographs of key stages during sample preparation.
3
4
Stephen C. Kuehn and Duane G. Froese
epoxy ample time to cure fully to maximize the strength of the bond between the epoxy and the particles. 10. Using a flat lap covered with 600 grit sandpaper, remove most of the graphite. Using 2000 grit paper, gently remove the last of the graphite to expose the adhesive layer. When the graphite is nearly removed, it should be possible to see light pass through the epoxy-filled pores at the surface of the graphite. 11. Using 0.3 or 1.0 mm polishing powder on a no-nap, pellon polishing pad, remove the adhesive coating to expose and polish the grains. As the adhesive layer is typically not completely uniform and because the larger grains may sink slightly into the adhesive layer, traces of adhesive may remain when the grains are sufficiently polished. If necessary, a small portion of the mount surface may be spot polished using a small piece of polishing felt attached to a 2–4 mm diameter stub that has been affixed to the center of a polishing wheel.
Preparation of Multiple-Sample Mounts Using a Drilled Disk (Procedure B) To prepare multiple-sample mounts, the following procedure was followed:
Figure 2. Schematic illustrations of sample mounts.
9. Cut a 5 mm thick disk from 25.4 mm ~1 in.! diameter cast acrylic rod and grind both sides so that they are flat and parallel. Leave the surfaces somewhat rough to improve the bond with the epoxy ~i.e., 600 or coarser grit!. Apply one or more drops of low viscosity epoxy to the planchet and place the acrylic disk on top. Allow the
1. Cut plates of high-density, isomolded graphite into 2.5 cm square blocks. Prepare them in the same manner as the planchets above to produce a flat, nonporous surface and then coat them with a thin layer of adhesive. For this study, plates with dimensions of 10.2 cm ~4 in.! square by 3 mm thick were obtained and then cut into the smaller blocks. 2. Place an acrylic disk containing three 9 mm diameter holes onto a graphite block. 3. Add a small drop of low viscosity epoxy at each corner. This will penetrate between the graphite and acrylic. When cured, it will bond them together and provide a seal between the three sample holes. Be careful not to apply too much epoxy as some of the excess will produce a ring of epoxy within each hole in the disk. Allow the epoxy to cure fully. 4. Transfer several drops of water containing suspended particles into each hole using a pipette ~Fig. 1D! and allow the water to evaporate completely. Repeat if necessary. 5. Add a drop of epoxy and an 8 mm diameter plug of acrylic to each hole ~Fig. 2!. Add additional epoxy as needed to completely fill each hole. Allow the epoxy ample time to cure completely. Including an acrylic plug helps to ensure a flat surface for polishing. Filling each hole entirely with epoxy may result in a surface that is slightly concave as the epoxy may shrink slightly as it cures. Alternatively, the three-hole disk may be removed once the evaporation is complete and then replaced by a solid disk, using epoxy to hold the disk in place. This latter approach is better for samples that are to be examined by transmitted light microscopy. 6. Separate the acrylic disk from the graphite block using a razor blade ~Fig. 1H!. Press the blade against the
Method for Fine Particle Mounting
7. 8.
9.
10.
surface of the graphite, and then drive it between the graphite and the epoxy until they begin to separate. Work the blade around the perimeter of the disk until complete separation is achieved. The graphite block may be saved and used again. Remove any excess epoxy attached to the perimeter of the disk using a razor blade or grinding wheel. Label the surface of the mount with a scribe. The ends of the scribe marks can serve as reference positions. If desired, several holes may be made using a fine needle to provide reference positions ~Fig. 1J!. Polish the mount to remove the adhesive coating and expose the grains. If the sample holes were filled entirely with epoxy, it may be necessary to spot polish the center of each sample hole. Apply a conductive coating ~e.g., carbon! to the polished surface.
Variations Several variations of this procedure may be used to adapt it to different samples and study objectives. For example, where only a single drop of water needs to be evaporated to apply sufficient particles, it is possible to place more than one sample on a single block by placing several small drops side-byside on the same surface and allowing them to dry ~Fig. 1F!. In cases where significantly more water must be evaporated, it is better to prepare a multisample mount using a drilled disk ~procedure B, Fig. 1D!. If a drilled disk is used but the examination of unpolished grains is desired, the disk can be removed immediately after evaporation. The mount can then be further processed as in procedure A. It is also possible to substitute several short pieces of glass or plastic tubing for the drilled acrylic disk, remove these when the sample is dry, and again process as in procedure A. If separate study of particles smaller than ;5 mm is desired, these may be separated from the larger particles using a track-etched polycarbonate membrane filter. The finer particles may be captured on a second membrane filter for study, and the coarser particles may be resuspended and mounted as above.
M ICR OANALYTICAL P R OCEDUR ES Semiquantitative, energy-dispersive X-ray analyses of unpolished grains were obtained on a JEOL 6301F field emission SEM equipped with a PGT-IMIX EDS system using a 20 kV accelerating voltage, focused beam, and an analysis time of 30 s. Manual stage coordinates were recorded for each grain and for at least three reference marks on each mount so that the individual grains could be relocated later. Secondary electron images were obtained for each analyzed grain. Lower magnification images were also obtained to record the positions of the analyzed grains relative to other nearby particles. After the single-sample test mounts studied above were polished and carbon-coated, the same grains were relocated using the reference coordinates and images collected earlier.
5
EDS analyses and secondary electron images were collected for the polished grains in the same manner described above. The new positions of the previously analyzed particles were calculated using a spreadsheet ~included as an online supplement to this article! that implements the matrix algebra described by Admon et al. ~2005!. This spreadsheet takes as input the x-y stage coordinates of three reference positions measured on the source and target instruments and the particle coordinates from the source instrument. It is also possible to transfer coordinates to and from an optical ~e.g., petrographic! microscope if an x-y or digitizing stage has been installed.
Supplementary Materials Accompanying the online version of this article is a spreadsheet designed to facilitate the relocation of particles present on mounts that have been transferred from one instrument to another or that have been removed from and then returned to the same instrument. The coordinate transformation calculations require that the mounts include a minimum of three reference positions. Required input for the spreadsheet includes ~1! x-y stage coordinates for the three reference positions recorded on the source instrument, ~2! x-y coordinates for the same three reference positions recorded on the destination instrument, and ~3! x-y coordinates for the particles of interest recorded on the source instrument. From these data, the spreadsheet calculates stage coordinates for the particles on the destination instrument. Both rotation and tilting are accommodated by the calculations. Wavelength-dispersive analyses of the polished grains were obtained by EPMA using a 5-channel JEOL 8900 microprobe ~15 kV accelerating voltage, 10 mm beam diameter, and 6 nA current!. Peak and background count times for all elements were 20 and 10 s, respectively, and total analysis time was just under 2 min. Silicate materials were used for the calibration of most elements. The four most abundant elements–Na, K, Al, and Si–were calibrated using a natural rhyolite obsidian, UA5831, which originates from Lipari Island, Italy. Mineral standards used for the remaining elements were pyrope for Fe and Mg, diopside for Ca, ilmenite for Ti, willemite for Mn, and tugtupite for Cl. The UA5831 obsidian and a sample of Old Crow tephra ~UT1434! were also analyzed repeatedly as secondary standards to provide a check on calibration quality and to detect instrument drift.
R ESULTS
AND
D ISCUSSION
The mounting procedure described above has been used to prepare more than 65 samples from the Mt. Logan ice core, six samples extracted from Arctic lake sediments, and numerous test mounts containing Old Crow tephra. Two glass shards from the Mt. Logan study are shown in Figure 3 and
6
Stephen C. Kuehn and Duane G. Froese
Figure 3. Sample shards from the Mt. Logan study.
Figure 4. Unpolished and polished images of two sample shards. Note the significantly smaller particle that is exposed together with shard B in the polished mount.
illustrate the size range of the tephra particles analyzed. Four of the Old Crow test mounts were examined in both unpolished and polished form on the SEM-EDS as well as in polished form by EPMA. Compositional data were obtained at each stage, and images were obtained on the SEM. Most of the 10–30 mm size glass shards that were imaged on the SEM and analyzed by EDS as unpolished grains were successfully relocated and analyzed again as polished grains
on the SEM and by EPMA. Images of selected Old Crow shards and the resulting compositional data are included in Figures 4 and 5 and Table 1. As Old Crow tephra is secondarily hydrated, it typically produces analytical totals of 94– 96%, although totals may be lower on very fine-grained material such as that used in this study. Analysis of polished instead of unpolished grains significantly reduces the scatter in the resulting SEM-EDS data ~Fig. 5! and reduces the
0.4 0.2 0.28 0.03 0.29 0.05 0.31 10% 83%
75.2 0.8 75.34 0.47 75.55 0.25 75.06 31% 45%
57%
40%
13.24
13.01 0.13
13.08 0.19
13.6 0.4
13.7 0.7
Al2O3
81%
62%
1.85
1.73 0.07
1.73 0.07
1.9 0.4
2.4 1.0
FeO
0.06
0.06 0.04
0.06 0.03
0.0 0.1
0.1 0.2
MnO
55%
49%
0.31
0.29 0.02
0.30 0.04
1.5 0.2
1.5 0.3
MgO
69%
29%
1.50
1.46 0.06
1.48 0.06
3.5 0.6
3.2 0.7
CaO
53%
12%
3.84
3.69 0.19
3.63 0.30
3.7 0.3
3.6 0.5
Na2O
62%
37%
3.82
3.67 0.07
3.84 0.12
0.2 0.1
0.1 0.0
K2O
63%
0.26 0.02
0.26 0.03
Cl
*Each unpolished shard was typically analyzed twice by SEM-EDS and twice again when polished. Each polished shard was typically analyzed only once by EPMA.
0.4 0.2
TiO2
75.1 1.2
SiO2
Summary of the SEM-EDS and EPMA Analyses Reported as Normalized Values ~100%, water-free!.*
SEM-EDS, unpolished Mean Standard deviation SEM-EDS, polished Mean Standard deviation EPMA, polished Mean Standard deviation EPMA, reference ~Ward et al., 2008! Mean Standard deviation X-ray fluorescence, reference analysis on glass separate Mean Reduction in standard deviation, unpolished to polished, both SEM-EDS Reduction in standard deviation, SEM-EDS to EPMA, both polished
Table 1.
100.00
100.00
100.00
100.0
100.0
Total
Mean:
Mean:
95.46
95.61 1.59
92.47 1.53
n/a
n/a
Analytical Total
63%
34%
2
15
25
66
70
Number of Analyses
Method for Fine Particle Mounting
7
8
Stephen C. Kuehn and Duane G. Froese
Figure 5. Bivariate plots of the SEM-EDS and EPMA analyses. Note the much broader distribution of the SEM-EDS analyses obtained from unpolished grains as compared to the polished grains and the much tighter distribution of the EPMA data compared to both sets of SEM-EDS data.
standard deviations typically by 30–40% ~Table 1!. Moving from SEM-EDS to EPMA reduces the variability even further with standard deviations falling by another 50–60% for most elements ~Fig. 5, Table 1!.
C ONCLUSIONS A highly flexible method for the mounting of fine particles from liquid suspension is described. This approach offers the ability to study the same grains in both unpolished and polished form and works well for polishing grains as small as 3–5 mm. The method is also sufficiently simple and rapid to allow the preparation of large numbers of polished samples, making it much more convenient to obtain the higher quality analytical data afforded by polished mounts.
A CKNOWLEDGMENTS This work was supported by a Natural Sciences and Engineering Research Council of Canada Discovery grant and an Alberta Ingenuity New Faculty Award to D.G.F. Numerous discussions with George Braybrook were very helpful during procedure development and included the initial suggestion to try an adhesive film on graphite. Christian Zdanowicz and David Fisher provided helpful discussions on extraction of glass from glacial ice and provided access to the Mt. Logan ice cores.
R EFER ENCES Admon, U., Donohue, D., Aigner, H., Tamborini, G., Bildstein, O. & Betti, M. ~2005!. Multiple-instrument analyses of single
micron-size particles. Microsc Microanal 11, 354–362; doi: 10.1017/S1431927605050312. Davies, S.M., Wastegard, S., Rasmussen, T.L., Svensson, A., Johnsen, S.J., Steffensen, J.P. & Andersen, K.K. ~2008!. Identification of the Fugloyarbanki tephra in the NGRIP ice core: A key tie-point for marine and ice-core sequences during the last glacial period. J Quaternary Sci 23, 409–414. Dugmore, A.J., Larsen, G. & Newton, A.J. ~1995!. Seven tephra isochrones in Scotland. Holocene 5, 257–266. Dunbar, N.W., Zielinski, G.A. & Voisins, D.T. ~2003!. Tephra layers in the Siple Dome and Taylor Dome ice cores, Antarctica: Sources and correlations. J Geophys Res 108, 2374–2384; doi: 10.1029/2002JB002056. Fisher, D., Osterberg, E., Dyke, A., Dahl-Jensen, D., Demuth, M., Zdanowicz, C., Bourgeois, J., Koerner, R.M., Mayewski, P., Wake, C., Kreutz, K., Steig, E., Zheng, J., Yalcin, K., Goto-Azuma, K., Luckman, B. & Rupper, S. ~2008!. The Mount Logan Holocene-late Wisconsinan isotope record; tropical Pacific-Yukon connections. Holocene 18, 667–677. Germani, M.S. & Buseck, P.R. ~1991!. Automated scanning electron microscopy for atmospheric particle analysis. Anal Chem 63, 2232–2237. Kekonen, T., Moore, J., Perämäki, P. & Martma, T. ~2005!. The Icelandic Laki volcanic tephra layer in the Lomonosovfonna ice core, Svalbard. Polar Res 24, 33–40. Ward, B.C., Bond, J.D., Froese, D.G. & Jensen, B. ~2008!. Old Crow tephra ~140 6 10 ka! constrains penultimate Reid glaciation in central Yukon Territory. Quaternary Sci Rev 27, 1909–1915. Yalcin, K., Wake, C.P. & Germani, M.S. ~2003!. A 100-year record of North Pacific volcanism in an ice core from Eclipse Icefield, Yukon Territory, Canada. J Geophys Res 108, 4012; doi: 10.1029/2002JD002449. Zdanowicz, C.M., Zielinski, G.A. & Germani, M.S. ~1999!. Mount Mazama eruption: Calendrical age verified and atmospheric impact assessed. Geology 27, 621–624. Zielinski, G.A., Mayewski, P.A., Meeker, L.D., Grönvold, K., Germani, M.S., Whitlow, S., Twickler, M.S. & Taylor, K. ~1997!. Volcanic aerosol records and tephrochronology of the Summit, Greenland, ice cores. J Geophys Res 102, 26625–26640.
