The practical utility of technologies for early detection of human exposure to a variety of toxic agents has been
limited in many cases by the absence of instruments suitable for first responders and at field hospitals. Microarrays
provide multiplexed assay of a large number of human biomarkers, including cytokines and chemokines, indicators
of immune system health. Assay of saliva is less invasive and provides quick indication of exposure especially of the
respiratory system. Our pilot clinical study has uncovered an early cytokine response in human saliva. As a model
for respiratory exposure, a cohort of 16 adult volunteers was challenged with FluMist<sup>TM</sup> vaccinations, an FDA
approved, attenuated live influenza virus. Blood and saliva cytokine levels were monitored immediately prior to and
up to 7 days afterwards. Bead assay found little change in blood cytokine levels while several of those in saliva
were frequently elevated above two standard deviations on trial days one and three. We have developed a prototype
portable saliva monitoring system consisting of microarray cytokine capture plate, luminescent reporter, and whole
plate imaging. Assay is with a commercial 96-well plate spotted with up to 16 distinct biomarkers per well and read
by chemiluminescence. A battery-powered, 16-bit, cooled-CCD camera and laptop PC provide imaging and data
reduction. Detection limits of common inflammatory cytokines were measured at about 1-5 pg/ml which is within
the clinically significant range for saliva of exposed individuals, as verified for samples from the small clinical trial.
An expanded study of cytokine response in saliva of therapeutic radiation oncology patients is being launched.
The fabrication of thick orientation-patterned GaAs (OP-GaAs) films is reported using a two-step process where an OP-GaAs template with the desired crystal domain pattern was prepared by wafer fusion bonding and then a thick film was grown over the template by low pressure hydride vapor phase epitaxy (HVPE). The OP template was fabricated using molecular beam epitaxy (MBE) followed by thermocompression wafer fusion, substrate removal, and lithographic patterning. On-axis (100) GaAs substrates were utilized for fabricating the template. An approximately 350 μm thick OP-GaAs film was grown on the template at an average rate of ~70 μm/hr by HVPE. The antiphase domain boundaries were observed to propagate vertically and with no defects visible by Nomarski microscopy in stain-etched cross sections. The optical loss at ~2 μm wavelength over an 8 mm long OP-GaAs grating was measured to be no more than that of the semi-insulating GaAs substrate. This template fabrication process can provide more flexibility in arranging the orientation of the crystal domains compared to the Ge growth process and is scalable to quasi-phase-matching (QPM) devices operating from the IR to terahertz frequencies utilizing existing industrial foundries.
Commercial apparatus has recently become available to utilize a gas-cluster ion beam (GCIB) for smoothing microelectronic and photonic surfaces to sub-nanometer residual roughness. Smoothing occurs after a high fluence to the surface. However, at very low fluence, surface features are observed that are helpful in modeling the stochastic nature of the smoothing. These low fluence features have potential for nano-scale surface texturing that may result in unique electronic and optical properties. This paper addresses the impact of individual gas clusters to yield nano-scale craters in SiO<SUB>2</SUB> and nano-scale hillocks on Si. The nature of these features results from parameters of cluster species, beam acceleration, target material and residual vacuum chamber gases, as well as chemical reactions. 20 kV argon clusters impacting a smooth SiO<SUB>2</SUB> film results in pits approximately 4 nm deep, approximately 10 nm diameter with a small rim of ejecta. Higher energy 24 kV Ar gas clusters incident on silicon with approximately 20 SiO<SUB>2</SUB> cause hillocks approximately 4 nm high (projecting above the native oxide) and a approximately 40 nm diameter. The hillocks formed from Ar-GCIB on Si are composed of SiO<SUB>x</SUB> and have been found to reflect the symmetry of the underlying (100) or (111) crystallographic Si orientations.
The surfaces of single-crystal wafers of sapphire and silicon carbide with microelectronic-grade high polish were exposed to a gas-cluster ion beam (GCIB) and significant reductions in roughness were observed. Atomic-force microscopy revealed that the typical initial surfaces consisted of a fine but small random roughness together with relatively large and sharp asperities. The latter were removed efficiently and GCIB smoothing process improvements are reported. The SiC wafers also have a high density of shallow scratch marks and these too were removed, with the average roughness R<SUB>a</SUB> falling below 4 angstrom after the best process. Analysis of the SiC by Rutherford backscattering spectroscopy in channeling mode revealed that when the GCIB process was adjusted so that asperities and scratch marks were removed, there was no increase in near- and at-surface damage. In particular, no lattice damage was observed of the sort typically caused by ion implantation prior to annealing. Significantly, it was found that oxygen gas cluster ion beams provided superior results with SiC as compared with argon GCIB. Surface smoothing mechanisms are proposed to explain these results.
A computer algorithm, which matches theoretical to measured infrared reflectance spectra, was successfully employed to determine multiple thin film properties of integrated circuits. Properties, such as film thickness, dielectric constant, and free carrier concentration were determined for a variety of important electronic films both in the laboratory and in process reactors. The latter measurements were accomplished by optically interfacing a Fourier transform infrared (FT-IR) spectrometer to several reactors. Real-time process monitoring allowed determination of deposition rate, free carrier activation temperature, and the influence of reactor conditions on film properties. Finally, these measurements were nondestructive, performed in-situ and within seconds, demonstrating the utility of this method for real-time process monitoring and control.
Silicon wafers have shown promise for the fabrication of photothermal IR detectors (i.e., bolometers) from epitaxial HTS thin films of YBa<SUB>2</SUB>Cu<SUB>3</SUB>O<SUB>(7-(delta</SUB> )) (YBCO). Conventional IC-grade wafers, ultrathin wafers, and micromachined-silicon membrane windows in conventional wafers, are all suitable, but the latter provides considerable advantage for bolometer performance. The high thermal conductivity and strength of silicon make it ideal for submicron-thick window designs. Epitaxy in the HTS film is advantageous, since it reduces granular disorder, the primary cause of dark noise (resistance-fluctuations) in the detector. Mid-to-far IR transparency of Si at 90 K is unique among those substrates that support high-quality epitaxial YBCO films. This Si transparency to IR can be used for various improvements in the optical design of these devices. We review the thermal and optical advantages of silicon substrates, device fabrication issues, and bolometer modeling. Thermal modeling of membrane bolometers indicates that the steady- state temperature-rise profile is nonuniform, but that this does not degrade the response linearity of the bolometer. Certain size limits and trade-offs in the design, will be important in the final device performance. We also discuss applications to FTIR instruments, and extensions of this technology to arrays including a novel on-chip transform spectrometer design.
Efforts to grow high quality films of YBCO on Si have been complicated by factors discussed in Ref. 1, chief among
them being the reaction between YBCO and Si, which is damaging even at 550 C. This is well below the customary
temperatures for YBCO film growth. To avoid the reaction problem, epitaxial YBCO films were grown on Si (100) using an
intermediate buffer layer of yttria-stabilized zirconia (YSZ).2 Both layers are grown via an entirely in situ process by pulsed
laser deposition (PLD). Although the buffer layer prevents reaction, another problem arises; the large difference in thermal
expansion coefficients between silicon and YBCO causes strain at room temperature. Thin (<500 A) YBCO films are unrelaxed
and under tensile strain with a distorted unit cell. Thicker films are cracked and have poorer electrical properties. The thermal
strain may be reduced by growing on silicon-on-sapphire (SOS) rather than silicon.3 This allows the growth of films of
arbitrary thickness. Ion channeling reveals a high degree of crystalline perfection with a channeling minimum yield for Ba as
low as 12% on either silicon or SOS. The normal state resistivity is 250-300 i-cm at 300 K; the critical temperature, Tc
(R=0), is 86-88 K with a transition width (ATc) of I K. Critical current densities (J)°f 2x107 A/cm2 at 4.2 K and >2x106
A/cm2 at 77 K have been achieved. In addition, the surface resistance of a YBCO film on SOS was measured against Nb at 4.2
K. At 10 GHz, a value of 45 was obtained. This compares favorably to values reported for LaAlO3.
Application of this technology to produce reaction patterned microstrip lines has been tested.4 This was done by ion
milling away portions of the YSZ buffer layer prior to the YBCO deposition. YBCO landing on regions of exposed Si reacts
to form an insulator. This technique was used to make 3 micron lines 1.5 mm long. The resulting structure had a Jc of
l.6xl06 A/cm2 at 77 K. Isolation of separate structures exceeded 20 M. Several advantages of this technique are that no
solvents, etchants or photoresist come into contact with the YBCO, hence this technique has a potential for operational-asgrown
In summary, it is now possible to produce YBCO films with structural and DC electrical properties which rival the
most optimized c-axis epitaxial YBCO films on MgO, SrTiO3 and LaAlO3. Preliminary measurements of microwave
properties appear promising.
We thank Bruce Lairson for help obtaining magnetization data and Richard Johnson, Steve Ready and Lars-Erik Swartz
for technical assistance. This work benefits from AFOSR (F49620-89-C-0017). DBF received support from NSF (DMR-
8822353). DKF acknowledges the AT&T scholarship.