Hi, new user to Matlab and I'm in Materials Engineering. I need to determine the diffusion coefficient of Ni in the Ni-Cu alloy in American Dimes.  I heard from a TA that Matlab can be utilized to determine concentrations of particles in a picture using the difference in pixel appearance.

I am wondering if this is possible.

Thanks,

Dylan

Yes it is possible

"Dylan Gibson" wrote in message <igicip$51d$1@fred.mathworks.com>...
> Hi, new user to Matlab and I'm in Materials Engineering. I need to determine the diffusion coefficient of Ni in the Ni-Cu alloy in American Dimes.  I heard from a TA that Matlab can be utilized to determine concentrations of particles in a picture using the difference in pixel appearance.
> 
> I am wondering if this is possible.
> 
> Thanks,
> 
> Dylan

On 11-01-11 02:31 PM, Joseph wrote:
> Yes it is possible
>
> "Dylan Gibson" wrote in message <igicip$51d$1@fred.mathworks.com>...
>> Hi, new user to Matlab and I'm in Materials Engineering. I need to determine
>> the diffusion coefficient of Ni in the Ni-Cu alloy in American Dimes. I
>> heard from a TA that Matlab can be utilized to determine concentrations of
>> particles in a picture using the difference in pixel appearance.
>>
>> I am wondering if this is possible.

It sounds plausible to me, but I would wonder about the sensitivity of this 
approach, and about the resolution you would need. A glance around suggests 
that unless you were just looking for a very rough approximation, that you 
would need a pixel resolution down to (estimating) about 5 nm, as the grain 
sizes would be expected to be only 20-30 nm.

http://dx.doi.org/10.1016/0040-6090(88)90203-9

How would go about writing code for such a task?

I need to grey scale the picture, that much i know.

On 11-01-13 03:57 PM, Dylan Gibson wrote:
> How would go about writing code for such a task?
>
> I need to grey scale the picture, that much i know.

I am not convinced of that at the moment. Grey scale of an image converts to 
luminance, but fresh copper reflecting a decent light source is going to have 
a high luminance and nickel reflecting a decent light source is also going to 
have a high luminance. If your light source is not uniform and pretty much 
parallel to the direction of the camera axis (and yet is not being directly 
reflected to the lens), then luminance due to the nickel could easily be 
confused with luminance due to the fresh copper.

In such a case, color is going to give you a substantial clue: fresh copper 
will have very different RGB characteristics than fresh nickel. Oxided copper 
will be relatively dull but will be a quite different color than nickel. I 
have rarely seen pure nickel, but my understanding is that it too oxides 
without difficulty into a duller grayish color.

I am not familiar with the oxidation characteristics of Ni-Cu . I see that the 
overall composition of USA dimes is 91.67% Cu and 8.33% Ni (wikipedia), 
whereas Canadian dimes are 92.0% steel (unspecified alloy), 5.5% copper, 2.5% 
nickel plating, which is different enough (and constructed differently enough) 
that I do not have practical experience with the dimes you are likely to be 
testing against.

I thus suspect that color cues will may prove more important to you than 
luminance cues -- unless, that is, that you know you are testing new dimes, 
and that the illumination is at a frequency preferentially reflected by one or 
the other material even after oxidization.

Using plain light or tuned monochrome (or narrow-band) light would probably be 
less expensive, but I wonder if for the purpose of determining diffusion, it 
might not be better to use focused X-rays -- e.g., dual source beams whose 
peak intensity depth could be adjusted. That would allow you to peak below the 
surface, and the luminance of the x-ray images would, I think, be more 
characteristic of the atomic mass of the grain rather than of the surface 
oxidization.

If you do happen to be working with laser illumination, then is it fair game 
to do some spot ablation to probe under the surface?

Another potential test device: a grid of inductance sensors, or possibly a 
grid of magnetic permeability sensors (very small magnets with measurements of 
the returned field strength?)

Thanks a lot for your input. Unfortunately I am only in 2nd year Undergraduate Studies and I have access to very little in the way of expensive machines.  I don't even know if we get new dimes, although I'm sure it would allow for more accurate diffusion results. (Assume we do get new dimes)

We (our group) have to determine the diffusion coefficient of Ni in Cu.  We were told nothing more than this, nor told any methods in which this can be achieved.  We can only do what we can afford, and after doing some research we've been informed SEM is out of the question.

We are going to analyze to new dimes under high magnification before a heat treatment, and then compare those to dimes that have under gone a heat treatment, all at varying times.

If you know any ways to determine concentration by a picture, I'm all ears.  We also are curious whether or not running current through various dimes to compare resistivity would be beneficial.  We arent even told the original concentration of the dimes (post manufacturing).

We are just kids and I could use some good advice.

Thanks a lot.

Dylan