There is a theory by the (Dutch) scientist Miedema that can be used to predict if a (dssimilar) metal combination will form intermetallic phases or not.

Using this theory W (Tungsten) is not reactive to Cu and is also less susceptible for Al en Fe.... It has a melting point (way) over 2000 °C as well.....

So, W is one candidate metal for your problem.

There is a theory by the (Dutch) scientist Miedema that can be used to predict if a (dssimilar) metal combination will form intermetallic phases or not.

Using this theory W (Tungsten) is not reactive to Cu and is also less susceptible for Al en Fe.... It has a melting point (way) over 2000 °C as well.....

So, W is one candidate metal for your problem.

Great, thank you so much Erik.

You might want to be more specific: does the material need to be a metal? DIffusion of what? self diffusion, of other substitutional elements or for interstitial elements or for heat? If it does not need to be a metal alumina would be perfectly fine.

The question itself as well as the topics below the question specifically state metal/alloy

The diffusion of what is unclear, until the elements where stated: Fe, Al, Cu

Dear Ludger,

any material can be use for sputering, it's fine. I need to have one layer with less diffusion or no diffusion to Al,Cu,Fe. I have to annealed these layer on 750 to 800 'C for 15 hours. Then diffusion barrior layer has to be stable in this temperature. Also, i don't want to use any kind of oxide. Alumina is oxide.

Dear Erik,

the alloy is mix of Al,Cu and fe. When I want to fabricate this alloy on steel, It has diffiusion from steel to these layers. I want to find a interlayer between steel and alloy and it can't be a oxide. Also I need to mention that the annealing temperature is 800'C and time annealing is 15 hours.

Best,

 Parsa

Erik, also I need to tell you, I checked out the tungsten has a interaction with Cu and it can made WCu, that this composition has so many application on resistivity and thermal conductivity.  

Cu and Tungsten are mutually virtually insoluble. Tungsten might not be such a good choice for Al. 

See the attached article in which Figure A14 shows the interfacial formation energy of W-W (dashed line) with several metallic elements (Al, Fe, Cu).

All elements above the dashed line will increase the interfacial energy and are therefore thermodynamically not stable (low affinity of W to these elements). Elements below the dashed line are decreasing the interfacial energy which is thermodynamically favorable, hence W has affinity to dissolve or even form intermetallic phases with the elements below the dashed line. The larger the distance to the line the stronger the effect.

As you can see W has a very low affinity for Cu and a only a marginal affinity to Fe and Al. Therefore, these elements will not dissolve nor form stable intermetallic phases, which you can check if you find the phase-diagrams of W-Al, W-Fe and W-Cu

Dear Hadi, in one question you are almost covering whole materials science and as result you get lot of misinformation. With your permission I would like to summarize the basics.

any physical change involves atomic motion by convection (bulk of milliions of atoms) or by diffusion (random atomic movements). Diffusion involves breaking and rebuilding of atomic bonds  between same kind of atoms (self diffusion) or added (solute) atoms in a solution or intermetallic compound. Since atom bond breaking is also the majör mechanism of melting, diffusion rate (coefficient) is proportional to the respective melting temp. kelvin. I.e., at a temp.very close to their melting temp all solids approximately have the same diffusion rate. This is the basis of atomistic kinetics, so.a material with high melting temp. and also  high diffusivity at a given temp is not possible physically. This is kinetics dictated by mechanism of the change.

What shall happen when we try to mix different kind, species, atoms is another story. Answer is in thermodynamics. All systems seek for minimum energy condition which we call as equilubrium. In constant temp. and pressure syatems Gibb's free energy "G" has to be minimized. "G" involves atomic bond energy (which we quantize as entalphy, and probability of the predicted configuration, distribution (which we call as entropy. There are three alternatives: (1) mixing will not occur at all (oil and water, Cu and W, (2) added component will dissolve in the solvent (as interstitial or substitutional solid solution in case of soliids) (3) solute and solvent will form compounds (stochiometry is required mostly) which we call as intermetallics. Which shall ocur among these three is dictated by (1) entalphy and entropy if you are looking from classical (emprical thermodynamics, (2) bond energy and statistics if you are looking from atomistic point of view.

All the data on what shall happen when we mix different kind of atoms is very effectively summarized in "phase diagrams".

PHASE DIAGRAMS ARE INEVITABLE, VITAL TOOLS FOR ALL MATERIAL SCIENTISTS.

İn early 20th century, what is expected when you mix atoms of different kinds of atoms has been stated by Hume-Rothery rules in terms of atom properties such as electro negativity, valance, atom size etc., which are still being used for approximate guidance.

Last words: if something is thermodynamically not possible then that's it, it cannot ocur. Otherwisw whether it shall take place or not depends on the  tkinetics.

thank you so much everyone. It was really good answers. I will ket you know the results after experiment.

Best,

  Parsa

Cobalt,molybdenum like refractory metals can also be added along with tungsten.

Refractory based high entropy alloys can be one of the candidate, as they have have pretty high m.p. and they are known to have show sluggish diffusion.