|
|
|
| QUESTION 1. How is the growth structure digitized for the MicroMilling system? |
| First, consider a curved line (Line I) as a series of straight line “segments” (Example 1). This series of connected segments comprise a single line (visible line on object). The points (1,2,3 etc), where segments join, reflect the inflection in slope between segments. For example, a curved line can be represented as series of straight segments (1-2, 2-3, 3-4 etc) of changing slopes; this in effect a “spline” fit. |
| First, consider a curved line (Line I) as a series of straight line
“segments”. This series of connected segments comprise a single line
(visible line on object). The points (1,2,3 etc), where segments
join, reflects the inflection in slope between segments. For example,
a curved line can be represented as series of straight segments (1-2, 2-3,
3-4 etc) of changing slopes; this in effect a “spline” fit.
Now, given a second line (a second visible growth structure – Line II), we can do a similar spline fit, with the only requirement that the number of segments must remain constant. This is required because we need to imagine the sample path as a series of areas (A1, B1, C1, etc), and when considering depth this becomes a volume which is used to calculate total mass of material to be sampled. Note that for complex structures one can conceive of a line segment with zero length. This simply means that the two points of that segment are the same. In order to construct sampling paths between “visible” structures, we interpolate within an area (A) to produce areas (A1, A2, and A3). A sample path then becomes the series A1, B1, C1 etc. |
| QUESTION 2: How is this digitizing actually done for the MicroMilling system? |
| There are several ways that this can be done, but here are two typical
methods.
A photograph or image of the sample can be digitized “off-line” using a digitizing tablet or image analysis system. Once the visible structures are digitized (as above), then the subsample paths must be calculated (interpolated). You can do this off-line using any method you wish with numerical methods you design (eg. a polynomial fit to each curve followed by a polynomial interpolation between Lines). Alternatively, one can use MicroMill’s built-in linear interpolation as explained above. This will give you a data file that can be used to directly control the motion of the stepper motor stages. This file is read into the stage control program, reference points are located for calibration of your image to the real co-ordinate system of the sample and stages (adding the Z-axis component), and you are ready to go. *Note, this off-line method is necessary when you are working with material that has growth features that are not visible under normal conditions. For example, fluorescence or cathodoluminescence petrography is sometimes employed to see growth structures in some materials. The images obtained using these methods can then be digitized from a photo, and again, using reference points it is possible to transfer this virtual image to the sampling system. Example: Mississippian cements using cathodoluminescence petrographic mages. The on-line method uses the image that is present on the video screen. In this case, you do the digitizing of LINE segments directly. Again, you must follow the rule that each LINE must have the same number of segments. On the present video system, one can draw lines guides on the screen to help define the LINE segments. This is really just an aid to help you with digitizing. Using the pointer, you then enter the points (coordinates) for each line segment. When done, you then perform the interpolation, examine a real-time plot of the interpolated paths to verify that they properly follow the growth structures. This interpolation also generates a record file that contains important information about the mass of sample that will be removed during each traverse (this is based the volume of material calculated from the drilling depth and area of each path segment). |
| QUESTION 3. How do you collect the sample
powders during milling? How does one clean the
sampling system between microsamples? How long does this process take? |
| We generally work from polished thick sections (approx 100-200 microns
thick) so that we can use transmitted light if necessary to view internal
growth structures. After a sampling pass of the mill, a trail of
powder remains on the surface. We originally used a vacuum aspirating
system to suck up this powder onto a glass fiber filter (4mm diameter disk
and 2mm internal diameter pipet). We found that static caused by
the moving air made a large quanity of the powder to stick within short
pipet (2cm long) pipet. So we decided to take advantage of the static.
We now just use a very narrow tipped scalpel (#11) which we carefully move
along the surface that has been milled. This requires some patience, but
the powder will cling to the tip of the scalpel such that it can be transferred
to a small (in our case) metal capsule. If we do not have enough
powder from one traverse (path) we simply mill out another path and combine
the powders.
After we have collected the sample, we use a low pressure compressed air stream to clean the surface and the milling point. In extreme cases, we have sometimes dipped the tip of the milling bur into dilute HCl and then ethanol, then used air to dry the tip. This generally is not necessary. The drilling speed (both rotation rate for milling tool, and the rate of motion along a path) for a sample is completely adjustable. You want to keep the rotation rate high enough to promote cutting, but not too fast or else the sample disperses on the surface. Similarly, for harder materials, it is necessary to slow the traverse rate down to low speed. Too fast, and the milling tool does cut fast enough. A typical time for a single traverse may be 2 minutes. The sample recovery time (picking up the sample from the surface) another 2 minutes. Say 5 minutes per traverse. So, for 50 samples about 6 hours. Note that this is after the sample has been digitized. This generally takes a 1 to 2 hours depending upon the complexity of the sample and how many coffee breaks you have. |
| USEFUL LINKS |
|