Deposition of Thin Films
Fe-Pd thin films were deposited on either thermally grown oxide/Si or SiNx/Si wafers by magnetron sputtering using elemental Fe and elemental Pd targets. Deposition parameters such as the working Ar gas pressure, applied power, deposition time and the working distance were systematically varied in order to study the deposition rates of Fe and Pd and hence the composition of the film. The film composition as a function of applied power is shown in Fig. 1 below. Film thickness and composition were measured by Rutherford Backscattering Spectrometry (RBS) available at Cambridge Accelerator for Materials Science, Harvard University.

Obtaining the Correct Crystal Structure
The as-deposited films typically consist of nanocrystalline grains in the body-centered cubic (bcc) phase. However, in order to achieve SME the film initially needs to be in the face-centered cubic (fcc) structure. As the film is cooled below the M s temperature it transforms to the face-centered tetragonal (fct martensite) structure. Therefore, the as-deposited films require a heat treatment at a temperature above ~800 oC followed by a quench in order to obtain the metastable fcc parent phase at ambient temperature. Among the various heat treatment conditions studied the heat treatment at 900 oC for 15 minutes in flowing, oxygen-gettered Ar followed by rapid cooling in Ar gas was determined to give the desired phases. The x-ray diffraction (XRD) peaks of films that have the fcc/fct structures are shown in Fig. 2 below. The surface of a heat treated film containing 28.8% Pd observed by scanning electron microscopy (SEM) is shown in Fig. 3.

SME Observed by the Substrate Curvature Technique
Occurrence of SME in Fe-Pd thin films was observed by measuring the substrate curvature under thermal cycling. Changes in the film stress during phase transformation can be expressed in terms of changes in the substrate curvature through the well-known Stoney formula. The SME is characterized by a small hysteresis in the film stress vs. temperature curve. The Fe-Pd thin film sample containing 28.3 at.% Pd showed a reversible martensitic transformation upon thermal cycling between room temperature and 200 oC (Fig. 4). All the curvature measurements were performed using the laser beam curvature system in Professor Spaepen’s group. In addition to the thermoelastic martensitic transformation, the data in Figure 4 also illustrate the Invar effect observed in these films for the first time: as the temperature increases, the residual stress in the films becomes more tensile indicating that the coefficient of thermal expansion of these films is smaller than that of Si (~3 10 -6 K -1). This phenomenon is not observed in films with the bcc structure in agreement with observations for bulk materials.

Fig. 1 Variation of Pd content as a function of power.

Fig. 2 XRD peaks of heat treated Fe-Pd films with fcc and fct structures (film thickness ~0.6 mm)

Fig. 3 SEM micrograph showing the surface of a heat treated film containing 28.8% Pd.

Fig. 4 SME observed by the substrate curvature technique. M s and A s refer to martensite start and austenite start temperatures, respectively.

Conclusions
In summary, we have successfully deposited Fe-Pd films with approximately 30 at.% Pd and demonstrated the existence of shape memory effect in these films using the substrate curvature technique. Two proposals have been submitted (to DOE and NSF) based on the results obtained in this project. The results from this research project were presented in the Smart Materials Symposium and the Thin Films Symposium at the Materials Research Society Winter Meeting in Boston in December, 2003.

Three imaging methods are widely used to obtain two-dimensional sections from three-dimensional samples: deconvolution microscopy, confocal microscopy, and

multiphoton fluorescence microscopy. Multiphoton microscopy is rapidly becoming the choice for imaging live specimen in 3-D. No multiphoton imaging facility is currently available to CNS researchers, although the NSF has provided funds to purchase a commercial two-photon microscope. We have just acquired a new femtosecond system that would allow us to build a multiphoton microscope cheaply. In contrast to the NSF-funded microscope, the proposed microscope would be more flexible and allow different types of detection schemes. The trade-off is that the system will not be as polished and “turnkey” as a commercial system. Also, the laser will be used for several projects and therefore not always available. Other CNS Associated Faculty potentially interested in the proposed microscope are Weitz, Xie, and Ingber (with whom we collaborate).

Multiphoton fluorescence microscopy is based on simultaneous absorption of two or more near infrared photons by a fluorophore introduced to the specimen. The absorption probability depends on the density of the excitation photons. By using tightly focused femtosecond laser pulses as an excitation source, multiphoton absorption is strongly enhanced, but only in the small focal volume. Fluorophores outside the focal plane do not absorb due to the rapidly dropping excitation photon intensity, yielding better sectioning and less photobleaching outside the focal plane than other microscopy techniques. In addition, the near infrared wavelength of the femtosecond laser is more compatible with biological samples and falls in the “optical window” of cells and tissues allowing for better penetration depth. We proposed to build a multiphoton laser scanning microscope using the same laser as our existing femtosecond laser micromachining setup. The combined system will give us the ability to perturb live cells with subcellular precision and study subsequent biophysical processes; many additional experiments, such as microfabrication and imaging of small devices in polymers would also become possible.