Evaluation of Freestanding Films

Testing of freestanding films requires that the films be removed from their substrates without impairing their integrity. A relatively simple procedure can be applied in which the films are deposited on substrates or intermediate layers that will readily dissolve in water or some other medium. Alkali halides, such as sodium chloride or cesium iodide, have been used to this end. Organic substances (e.g., collodion or certain polymers) have also been used as substrates. However, inherent properties of films deposited on different substrates may vary, so it is often desirable to use films deposited on a specific substrate. This is particularly true of microelectronic circuits in which metallizations, such as aluminum, are deposited on silicon or silicon oxide. In these cases, the metal films are separated from the substrate by exposure to a stream of fluorine in an inert carrier gas. The silicon is removed in the form of gaseous silicon tetrafluoride, while the metal is covered by a very thin protective metal fluoride film. As the substrate is consumed, a freestanding film is produced.

Uniaxial Tensile Testing of Films. The most direct way to obtain the stress-strain relations of thin films in tension is by uniaxial testing. In principle, this procedure is analogous to conventional tensile testing of bulk materials, yet the fragility of the films and their extreme sensitivity to even the smallest flaws make uniaxial tensile testing a difficult and often frustrating task. In view of the high surface-to-volume ratio in thin films, specimens must be virtually free of surface flaws, because even the smallest defects will lead to premature failure. Tensile specimens usually have the form of a "dog bone." This shape, which has proved to be most advantageous, is obtained either by covering the substrate with a mask during deposition or by standard photolithographic techniques.

Problems are caused also by internal stresses that are created in the films during the deposition process. When the films are separated from the substrates, the stresses often cause films to wrinkle or to curl severely, rendering them useless for testing. To alleviate this condition, special devices have been developed that keep the films stretched during ablation (Ref 1).

Equipment used for uniaxial tensile testing of thin films is generally based on the simple principle of deforming a freestanding film by a known amount through the application of a known force to the ends of the specimen. The testing devices can be divided into two categories:

• "Soft" machines, in which the loading rate is constant and the ensuing elongation is measured

• "Hard" machines, in which the elongation rate is constant and the applied loads are measured

An example of tensile curves obtained by uniaxial testing of plasma-deposited aluminum films of 1pm thickness is given in Fig. 1. The initial portion of the curves often contains a clearly noticeable "unwrinkling" stage that can be ignored or discarded.

Fig. 1 Stress-strain curve obtained by uniaxial testing of a freestanding 1 pm thick annealed aluminum film

The absence of a well-defined yield point is characteristic for aluminum in bulk and, as evidenced by Fig. 2, for thin films.

Fig. 2 Schematic drawing of a bulge-testing device. Source: Ref 3

The effect of film thickness on film strength has been the subject of intense study. Tests indicate that the strength of metallic films rises dramatically when the thickness falls below about 0.3 pm. There is evidence that the strength of films obeys the Hall-Petch relation:

Strenght = K +-b Vt where K and b are constants and t is the film thickness. However, the details concerning the thickness effect in films are still open to question and appear to be greatly influenced by the deposition conditions.

Whereas the elastic properties of metal films closely approach those of metals in bulk form, some investigations reveal a substantially higher fracture-strength level for thin films than for the corresponding bulk materials. The fracture stress of metallic polycrystalline and single-crystal films tends to rise sharply as the film thicknesses drop below 0.2 pm.

Uniaxial Creep Testing of Films. Uniaxial testing equipment can be readily employed in creep tests of freestanding thin films by keeping the load constant. In such tests, characteristic creep curves consisting of a primary and a secondary stage are obtained. The stress and temperature dependence of the secondary creep rates is very similar to that of bulk materials. For high strain rates, secondary creep rates are modeled as:

For low strain rates, the relation is:

0 0

Post a comment