User login


You are here

Metal films on polymer substrates stretched beyond 50%

Nanshu Lu's picture

A link for the paper:

Abstract: When a freestanding film of a plastically deformable metal is stretched, the film ruptures by strain localization, and the elongation of the film is less than a few percent. When the film is deposited on a polymer substrate, however, strain localization may be retarded by the substrate. This paper reports Cu films deposited on Kapton substrates and stretched up to the rupture of the substrates (at an elongation between 50-60%). When Cr adhesion layers are introduced between Cu and Kapton, few microcracks in Cu may be found, and the measured electrical resistance agrees with a theoretical prediction. Micrographs are presented to show that strain localization and debonding co-evolve.



Aaron Goh's picture

Nanshu, in principle, sticking the metal film to rubber allows deformation to even higher strains?

Nanshu Lu's picture

Thanks for your interest.

We chose to use Kapton HN as the substrate because it is not only very stretchable but also a stiff polymer (a polyimide with modulus E=2.5GPa). However, typical Young's modulus of rubber is only between 0.01~0.1GPa.

Prof. Teng Li 's simulations (the paper is available here ) showed that if the substrate is too compliant, it would not be able to provide a sufficient geometric constraint to the top metal film, specifically to retard strain localization in the film. Hence the metal film may still behave similarly to a freestanding film which may form a single neck and rupture at small strains. Of course it is worth trying more stretchable polyimide (E~3GPa) substrate to determine the ultimate rupture strain of these metal films used in our experiments.

Teng Li's picture


Excellent experiments! This is the first direct observation of the co-evolution of film necking and interface debonding.

One question about the effect of grain size on the rupture strain.  It seems in the annealed samples with large grains (several microns), there is no appreciable grain boundary cracking even at large elongation.  Do you think the huge ductility of such films results from the stronger grain boundaries or it's just due to the larger grain size. I'm wondering if you have done the experiments of stretching the as-deposited Cu films (like the one in Fig. 1a) to a large strain. In this case, the metal films with small grains (of tens of nanometer) may show both film necking and grain boundary cracking, as observed in previous experiments.

Nanshu Lu's picture

Teng, thanks for your encouragement. Smile

It is hard to characterize the grain boundray strength at this stage and we don't know for sure whether it is enhanced or not after the heat treatment.  But you are right that transgranular cracking is dominant for our annealed films (with large grains). It appears that such failure is quite similar to single crystal cracking although our Cu film is polycrystalline.

We have run tensile tests of as-depostied films as well and we are going to complete and report it soon. One of the interesting observations is that there seems to have stress-induced grain growth for such fine-grained films. Therefore strains tend to localize at the large grain regions. We do have observed co-evolution of local thinning and debonding as well. 


Nan shu, Thanks very much for your citing my work! and sharing your excellent work the paper.

I have a very primary question to consult you, how much the inital resistance? how did you monitor the resistance, all the time?

Thanks a lot

Nanshu Lu's picture

Hi, Rongmei,

Thanks for your interests on our work. To determine the initial resistance we measured the sample with a Keithley multimeter and then subtracted the system resistance which was obtained by contacting the two probes of the multimeter. We recorded in-situ resistance during tension by connecting the multimeter to the computer with a well-known software named LabView . Of course there's a little programing to do.





Hi, Nanshu, thanks for your telling me the method. hehe, comparing with your advanced method, my method is too backward. At the beginning, I kept eye on the Keithley source meter and recorded the value at once, and when one test finish, my hand is already too tired to raise. 

FIB are eupensive, so I have no chance to do it. while other students in my group did FIB observation by the way for me. unfortunately, there are mang damage on the surface of Cu film. accordingly, could the debonding of Fig. 5 (b) in your paper result from FIB?

Xi Wang's picture

Hi Rongmei, thanks for being interested in our work. The milling process from the ion impact can cause some damages to the sample, including surface amorphization and contamination. MRS bulletin May 2007 actually has a very good review article about FIB milling process. The cut in Fig.5(b) was made by several steps, including fine polish steps with very low beam current. It is unlikely the feature is due to the FIB process.


Dear Xi , thanks. I want to make cross-section specimen of Cu films on PI all the time, but failed due to PI substrates. so I will learn from you, thanks again. 

Nanshu Lu's picture

To answer your question in another point of view we used FIB to cut undeformed specimens as well and observed no debonding at all. Hence we believe that FIB will not cause metal film debonding from the substrate.


 Nan shu, I have another question want to consult you.

How about the initial stress before your tension?


Best Regards


Nanshu Lu's picture

This is what we are working on actually. We did not deal with stress-strain in this submitted paper. You must already know that an ordinary way to determine the residual stress is to do curvature measurment of the specimen.


Hi Nan shu,

I am working in buckling analysis of thin film bonded to the elastomeric sabstrate, as part of my thesis.However, i do have some difficulties in modelling the structure in Abaqus for analyis. So i kindly ask you to help if you have some helpful materials on this regard.

thanks in advance

Subscribe to Comments for "Metal films on polymer substrates stretched beyond 50%"

Recent comments

More comments


Subscribe to Syndicate