Chromium is used for adhesion purpose for several reasons:

 It is a material providing high binding strength with oxygen and therefore able to depassivate many material surfaces. In such cases O atoms of oxyde material can be placed in an interconnecting bridge position between subtsrate material and coating material ( best example given with glass considering that Cr-O has binding energy similar to Si-O).

This also explains why Cr  often  present reduced surface mobility and diffusion and why some homogeneous interface can often be achieved.

However, other aspects must be associated to adhesion considering that dramatic drawbacks can be observed with adhesion with surface contamination and recontamination,  interface diffusion, internal surface passivation,  film intrinsic stress and also interface mismatch stress.

Intrinsic stress (compressive or tensile) is much depending on the process with which Cr is deposited and to be considered how it can be annealed and relaxed. To be also considered that Cr can be deposited with different states with different  themodynamical properties.  

Quality of surface cleaning and depositing vacuum cleanliness has to be considered. Should not be omitted the point that at 10-6 mbar the residual gas is generally H20 and HC and that they can condense on reactive surfaces and/or at reduced temperature up to one monolayer per second. 

Cr, because of its relatively large atomic size, will generally not diffuse in many substrate materials when heated up by  means such as friction and heating by irradiation and chemical recombination energy release.

However, The binding strength of Cr-O can be significantly  higher than the binding strength of the subtrate surface oxyde, creating thus an "internal" passivation with which the intrinsic adhesion strength will be strongly affected.

This is the reason why Cr will not always provide best adhesion interlayer on aluminium considering that the Cr-O binding energy is higher than the Al-0 binding energy and that the atomic size of Al is much lower than Cr.

Especially when density of energy of different kind of external, internal and interface stress is higher than the density of  intrinsic adhesion energy , causing then delamination of the coating.

This is then the reason why Cr is not providing best adhesion for carbon coatings if not modified in its interatomic distance,  because of quite large mismatch of atomic size between Cr and C and which needs to be attenuated with interatomic single atomic nitrogen.

In some papers it is demonstrated a direct evidence that island sizes

just prior to coalescence represent the main reason for the roughness of

polycrystalline metal films. For this purpose, chromium nucleation centers (very thin film) were evaporated onto the glass substrates just before the deposition of Au in

order to inhibit the island growth. Chromium is an oxygen active material

leading to very stable nucleation centers on glass or oxidized silicon.

(Vancea et al. Surface Science 218 (1989) 108-126)

@Giuseppe, Thank you for pointing out that Chromium is an oxygen active material and this could lead to very stable nucleation centers. This fact leads us closer to the reality as another material, Titanium, is also often used as an adhesion layer (but not with glass!).

However, I feel we can dig even deeper if we pursue more fundamental properties. I don't know which ones to suggest but I'd love to hear from someone. Terms like surface energy, binding energy, etc., come to mind.

I use either Cr, Ti or Ni (2-5nm) as adhesion layer between SiO2 (grown on Si substrate) and Au. Since Cr, Ti and Ni all work well, the more relevant question to me is why Au does not stick well onto SiO2.

Fortunately, I found that there is a very hot discussion about this here:

https://www.researchgate.net/post/Is_there_anyone_who_knows_about_the_interface_adhesion_between_evaporated_Au_and_a_SiO2_layer

One plausible explanation I also found is that the heat formation of the oxide layer (gold oxide, chromium oxide, titanium oxide...) should be higher than that of SiO2.

In Diamond Growth Cr works as a Catalyst which supports the growth of the Sp3 Carbon

For deposition of DLC on steels in general, Cr has been also used due to its high activation energy, surely due to its electronic configuration of the Cr in the ground state is d5s1, with stable semi- occupied orbitals.

Chromium is used for adhesion purpose for several reasons:

 It is a material providing high binding strength with oxygen and therefore able to depassivate many material surfaces. In such cases O atoms of oxyde material can be placed in an interconnecting bridge position between subtsrate material and coating material ( best example given with glass considering that Cr-O has binding energy similar to Si-O).

This also explains why Cr  often  present reduced surface mobility and diffusion and why some homogeneous interface can often be achieved.

However, other aspects must be associated to adhesion considering that dramatic drawbacks can be observed with adhesion with surface contamination and recontamination,  interface diffusion, internal surface passivation,  film intrinsic stress and also interface mismatch stress.

Intrinsic stress (compressive or tensile) is much depending on the process with which Cr is deposited and to be considered how it can be annealed and relaxed. To be also considered that Cr can be deposited with different states with different  themodynamical properties.  

Quality of surface cleaning and depositing vacuum cleanliness has to be considered. Should not be omitted the point that at 10-6 mbar the residual gas is generally H20 and HC and that they can condense on reactive surfaces and/or at reduced temperature up to one monolayer per second. 

Cr, because of its relatively large atomic size, will generally not diffuse in many substrate materials when heated up by  means such as friction and heating by irradiation and chemical recombination energy release.

However, The binding strength of Cr-O can be significantly  higher than the binding strength of the subtrate surface oxyde, creating thus an "internal" passivation with which the intrinsic adhesion strength will be strongly affected.

This is the reason why Cr will not always provide best adhesion interlayer on aluminium considering that the Cr-O binding energy is higher than the Al-0 binding energy and that the atomic size of Al is much lower than Cr.

Especially when density of energy of different kind of external, internal and interface stress is higher than the density of  intrinsic adhesion energy , causing then delamination of the coating.

This is then the reason why Cr is not providing best adhesion for carbon coatings if not modified in its interatomic distance,  because of quite large mismatch of atomic size between Cr and C and which needs to be attenuated with interatomic single atomic nitrogen.

Mohammad Ali Mohammad I would really appreciate if you could tell me why titanium is not used as adhesion layer on glass substrate? Cheers!!!!

Thanks for detailed reply of chromium. Can you please suggest a better adhesive inter layer for coating DLC on aluminum alloy?

@Hung Phan, i am not a material science expert, but if i have to guess, i think that chemical reactions at the surface are important for adhesion of metal films and since Au is inert and does not form strong bonds with Oxygen, it poorly sticks to oxides. 

This is just an educated guess, not an answer. 

I have some work concerning DLC adhesion on Al (and its alloys) surface. The best condition was using Si as an interlayer. By using PECVD process SiH4 can be used as a precursor gas.

For athodic arc evaporation, we would like to use Cr adhesion layer due to:

1) Both Ti and Cr are cheap, but Ti is soft.

2) Cr would like to getter the contaminations (such as O) and the oxidation product is dense and stable.

3) Normally, the working gases are Ar, N2, H2, CH4 and etc. Since the 2nd ionization energy of the Cr is higher compared to the 1st ionization energies of these gases, the Cr+ dominates the plasma. This is good for deposition of adhesion layer with low internal stress under a certainly bias. However, the Ti, Zr, Hf, Nb and V are not good.

4) It's easy to maintain the arc movement on the Cr target surface. The melting points of Zr, Hf, Nb, W and V are relatively high.

thanks for nice results, can please Dr. Stephane Neuville provide publication about your answer.