Copper electrodeposition into through-silicon-vias (TSV) is a rapidly developing and instrumental technique required for efficient high-density 3-D integration of complex semiconductor devices. The ubiquity of the electrodeposition of copper (Cu) in damascene interconnects has led to natural development towards Cu electrodeposition in TSVs. Many processes currently rely on either electroless plating or sputter deposition as a seed layer prior to the electrolytic deposition of Cu. However, these techniques are not capable of achieving a conformal seed layer throughout the full depth of the vias due to the geometric complexity of these devices. This work investigates via filling of electroplated Cu on substrates that have undergone atomic layer deposition (ALD) of a conformal seed metal. The intent of this work is to achieve uniform and void-free Cu electrodeposition into extremely deep vias (625μm) at aspect ratios as high as 10:1. Platinum (Pt) as a seed layer presents unique challenges due to the high affinity of carbon for collecting on the Pt surface, which renders the surface hydrophobic. In order to adequately wet the entire depth of the vias, the carbon compounds must be removed from the Pt to ensure the surface is hydrophilic. This work will present the results of both wet chemical and dry plasma processes to achieve this full-depth Pt via wetting. Further substrate preparation examines the impact of incubation in the electroplating chemistry under vacuum with an increased concentration of an accelerator additive. This step aims to further facilitate conformal plating into the vias by removing gas pockets from the vias while saturating these areas with accelerator to increase the initial plating rate in the bottom of the vias. In addition to these surface preparation techniques, development of an optimized electroplating bath chemistry and applied current regime will be presented. DC plating as well as various pulse regimes are investigated in order to tailor the filling profile of the vias. With regard to the bath chemistry, high Cu concentrations (80g/L) in methanesulfonic acid (MSA) electrolyte chemistries are compared to conventional Cu concentrations (40-60g/L) in sulfuric acid (H2SO4) electrolyte chemistries. The higher solubility of Cu in the MSA (80g/L) versus H2SO4 electrolyte (50g/L) reduces mass transport limited depletion of Cu ions through the vias 1. The particularly large concentration of Cu ions in this electroplating bath promotes a more rapid replenishment of these ions in the deep and high aspect ratio structures throughout the deposition period. In addition to the different chemistries and plating regimes studied herein, the effects of accelerator, suppressor, and leveler additives are examined through analysis of their affects on cyclic voltammetry curves as well as fill ratios throughout the vias. These fill ratios are examined after the vias are cross-sectioned and polished through the use of chemical mechanical planarization, and subsequent characterization of the vias is achieved using optical and scanning electron microscopy techniques. Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000. S. K. Cho, M. J. Kim, and J. J. Kim, “MSA as a Supporting Electrolyte in Copper Electroplating for Filling of Damascene Trenches and Through Silicon Vias.” Electrochemical and Solid-State Letters, 14 (5) D52-D56 (2011).

In recent years microelectronic packaging development has focused on 3D integration in order to compensate for the decrease in traditional Moore’s Law scaling. 3D integration allows increased I/O counts, lower RC time constants, lower power consumption, and better thermal sinking. Furthermore, 3D integration decreases areal density through stacking chips and connecting their electrical signals with through substrate vias (TSVs) as opposed to wire bonding. A significant amount of development has been achieved integrating TSVs with standard silicon (Si) substrates; however, very little development has been made integrating this technology with silicon-on-insulator (SOI) substrates 1 . Certain MEMS applications take advantage of the use of SOI wafers with relatively thick device layer silicon, however the integration of TSVs with these unique SOI substrates (e.g., 600um handle, 1um buried oxide, and 25um thick device layer) presents distinct challenges when making electrical connections through the device layer Si, buried oxide (BOX), and the handle Si. This is particularly true for cases in which the handle Si is not thinned. In this work, we present a novel TSV integration approach in which electrochemical deposition (ECD) is used to connect copper (Cu) vias, filled in the handle of the substrate, to tungsten (W) vias filled by chemical vapor deposition (CVD) in the device layer of a SOI substrate. Deep reactive ion etching (DRIE) has become the most accepted process for creating TSVs when working with semiconductor materials. After forming the vias, chemical vapor deposition, atomic layer deposition (ALD), and/or electrochemical deposition are used to deposit an insulating material for electrical isolation, a barrier layer to prevent diffusion, and a conductive material for electrical connection. CVD is typically used for all three of these purposes if the film thicknesses and aspect ratios are small enough and geometries are not overly complex. ALD is utilized for both the insulating layer and the diffusion barrier when CVD is not capable of conformally coating intricate geometries. ECD is used to fill the vias with a conductive material when via dimensions are too large for CVD, which is common in MEMS applications. This work will utilize DRIE for etching topside vias through the device layer Si and CVD for filling these vias with W. DRIE is also used for etching the handle Si and BOX layers using a hardmask on the backside of the substrate. Two approaches are investigated and presented for utilizing ECD to form Cu TSVs and connect the backside of the wafer to the W TSVs imbedded in the device layer of the SOI substrate. In both of these approaches, ALD is used to deposit a conformal layer of Al 2 O 3 to isolate the TSVs from the handle Si. A spacer etch is performed to remove the Al2O3 from horizontal surfaces, revealing the W TSVs previously fabricated in the device layer Si. The first approach is to fill the TSVs with Cu from the bottom up, initiating plating from a seed layer deposited only at the bottom of the backside vias. This approach allows a conventional makeup chemistry to be used with sulfuric acid (H 2 SO 4 ) as the electrolyte in conjunction with a periodic reverse plating regime. The second approach is to fill the vias from the outside in, initiating plating from a conformal platinum (Pt) seed layer deposited by ALD and using methanesulfonic acid (MSA), rather than H 2 SO 4 , as the makeup chemistry electrolyte 2 . The higher solubility of Cu in MSA (80g/L) compared to H 2 SO 4 (50g/L) reduces mass transport limited depletion of Cu ions in the vias, leading to conformal Cu deposition throughout the depth of the vias. The advantages, disadvantages, and challenges associated with each of these two integration approaches will be presented. Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000. G-K. Lau, J. Soon, H-Y Li, K. Chui, and Y. Mingbin, “Process Integration and Challenges of Through Silicon Via (TSV) on Silicon-On-Insulator (SOI) Substrate for 3D Heterogeneous Applications.” 17 th Electron. Packaging Tech. Conf., (2015). S. K. Cho, M. J. Kim, and J. J. Kim, “MSA as a Supporting Electrolyte in Copper Electroplating for Filling of Damascene Trenches and Through Silicon Vias.” Electrochemical and Solid-State Letters, 14 (5) D52-D56 (2011).

We present active electrostatic MEMS gas valves for Micro-Gas-Analyzer (MGA) flow control. These unique valves enable extremely low dead volume, highly integrated flow control chips for the MGA application, and potentially others (e.g., propulsion, pneumatic, and thermodynamic microsystems). We have demonstrated low leak rates ( <; 0.025 sccm, <; 0.0025 sccm on a similar passive valve design), high operating pressures 6.9×10 5 N/m 2 (100 psig), a high-pressure record for valves of this size and type, and high flow rates (>; 25 sccm) using control voltages on the order of 100 V. The valve designs presented eliminate charge build-up issues associated with insulating materials and are closely tied to a base-lined microfabrication process (SUMMiT), allowing mass production. Using this process, which incorporates only CMOS compatible materials, eliminates outgassing and absorption problems inherent to microvalve designs that incorporate elastomers or organic bonding layers, and reduces contamination when the valve is part of the chemical analysis flowpath. The results obtained indicate that even higher performance level valves (>; 1.4 × 10 6 N/m 2 or 200 psig operating pressure, at similar control voltage, flow rates, and leak rates) are possible.

We describe a novel technique that utilizes simultaneous implementation of dielectrophoresis (DEP) and magnetophoresis (MAP) to focus magnetic particles into streams for optical analysis of biological samples. This technique does not require sheath flow and utilizes a novel interdigitated electrode array chip that yields multiple streams of flowing magnetic particles in single-file columns. The MAP force placed particles in close proximity to the microelectrodes where they were subjected to a strong DEP force that generated the particle focusing effect. Particle focusing efficiency was improved using this combination DEP–MAP technique compared to DEP alone: particle stream widths were reduced ∼47% and stream width variability was reduced 80% for focused streams of 8.5 µm diameter magnetic particles. 3 µm diameter magnetic particles were strongly focused with DEP–MAP (∼4 µm wide streams with sub-µm variability in stream width) while DEP alone provided minimal focusing. Additional components of a prototype detection system were also demonstrated including an integrated magnetic pelleting component, a hand-held MHz frequency signal generator and a bench-top near-confocal microscope for optical analysis of flowing particles. Preliminary testing of a sandwich assay performed on the surface of magnetic particles showed 50 ppb detection levels of a surrogate biotoxin (ovalbumin) in a raw milk sample.

We utilized self-assembly of cyanine chromophores to study the conformational changes in various types of nucleic acid scaffolds: single and double stranded DNA, linear or circular DNA and RNA. We identified a chromophore that became highly fluorescent after aggregating upon nucleic acids. Fluorescence from the aggregate was instantaneous after self-assembly. Temporal emission profiles displayed a biphasic trend demonstrating kinetic dependence for assembly and disassembly. Absorption spectra of the aggregate showed a red-shifted "shoulder" peak indicative of J-aggregate. Fluorescence from J-aggregates was also red-shifted. We utilized cyanine self-assembly to quantize various nucleic acids. The limits of detection and quantization for ϕX174 DNA were 3 and 9 fmol, respectively. We similarly determined the sensitivity for various nucleic acids and established the optimum conditions for self-assembly. Collectively, the effects of methanol, salt, and full width at half maximum for cyanine fluorescence on DNA or carboxymethylamylose scaffolds, all suggested noncovalent, electrostatic, and hydrophobic forces were involved in supramolecular self-assembly. Our results facilitate a better understanding of supramolecular self-assembly.

In this paper, we present integrated disk-in-cage poppet valves with tuned spring stiffness for gas flow control of a microgas analyzer. The valves require zero power and close at preset offset pressures (0-35 psig) to switch from gas sample loading onto a preconcentrator to concentrated constituent sample injection into a microgas chromatograph. Air flow rates of 4.5 mL/min at pressures of - 2.5--5 psig (vacuum sample loading) were measured. Hydrogen leak rates of 0.1 muL/s (0.006 mL/min) were measured with valves closed at 15 psig. Analytical and numerical modeling was used to guide design of valve spring constants (ranging from 10 to 1500 N/m) that control the valve open position, flow rate, and closing pressure. The parameter design space is limited to a range of seat overlap, valve size, and spring stiffness that will allow adequate flow rate, sealing, and closing at predictable pressures. A linear curve defining closing pressure as a function of spring constant, valve gap, valve size, and seat overlap fit measured closing pressure data and can be used to predict closing pressure for future designs.

There is considerable interest in the oxidative fate of phenols such as p-cresol as environmental pollutants and uremic toxins. We supply a menu of spectroscopic options for the high throughput screening of laccase oxidation of p-cresol through multiple modes of detection. Laccase activity was monitored kinetically at pH 4.5 by absorption changes at 250 nm, 274 nm or 297 nm, and in endpoint mode by the bathochromic shift in absorption to 326 nm in 50 mM NaOH. Laccase oxidation of p-cresol was also detected by product fluorescence at 425 nm after excitation at 262 nm or 322 nm in 50 mM NaOH. We optimized the kinetic parameters for p-cresol oxidation (pH optimum 4.5-5.1; 37 degrees C; Km = 2.2 mM) resulting in laccase limits of detection and quantitation of 25 pg/microL and 75 pg/microL, respectively (approximately 360 pM; 25 ppb). The sensitivity for p-cresol was similar to previously reported values. The small (approximately 20%) decrease in signal strength after six cycles of excitation over a 3 h period was attributed to photobleaching or photodegradation of the emitter and not due to fluorescence decay (photoinstability). Halide inhibition was characteristic of laccases (IC(50) = 25 mM NaCl). A unique advantage of our assay is that laccase catalysis could be interrogated using multi-mode absorption or fluorescence under acidic or basic conditions, in real time or endpoint modes. Orthogonal interrogation facilitates ratiometric analysis enabling high specificity while minimizing interferences during compound library screening. The phenolic alcohol p-cresol may be a model for monolignol oxidation. Our studies might find applications in biofuels, to triage dialysis patients, or for the environmental bioremediation of phenols.

We present MEMS polysilicon microvalves for flow control of a rapid analytical microsystem (Micro-Gas-Analyzer, MGA). All valve components (boss, seat, springs, electrodes, and stops) are surface micromachined in the SUMMiTtrade microfabrication process. The valves have been characterized at high flow rate when open (60 ml/min air), low leak rate when closed (<0.0025 ml/min Hydrogen, H2), and tunable closing pressures (1 to 35 psig). Active electrostatic valves have been shown to hold closed (voltage on) against a high pressure (>40 psig) for sample loading, open for gas chromatograph (GC) loading (voltage off), and reclose against low pressure 2-5 psig.

We present a miniaturized high-throughput sensor array that will augment biofuel technology by facilitating in situ biochemical measurements upon micrometer-scale surfaces of leaves, stems, or petals. We used semiconductor processing to photopattern Foturan glass wafers and fabricated gold-plated microscopic electrode needles (ElectroNeedles) that pierced 125-mum-thick surfaces without deformation. The 5 x 5 or 10 x 10 arrays of ElectroNeedles can analyze 25 or 100 samples simultaneously, increasing throughput. Each microneedle in the array can also be individually addressed and selectively functionalized using diazonium electrodeposition, conferring multiplexing capability. Our microfabrication is a simple, inexpensive, and rapid alternative to the time-, cost-, and protocol-intense, deep-reactive-ion-etching Bosch process. We validated the system performance by electrochemically detecting p-cresol, a phenolic substrate for laccase, an enzyme that is implicated in lignin degradation and therefore important to biofuels. Our limits of detection (LOD) and quantization (LOQ) for p-cresol were 1.8 and 16microM, respectively, rivaling fluorescence detection (LOD and LOQ = 0.4 and 3microM, respectively). ElectroNeedles are multiplexed, high-throughput, chip-based sensor arrays designed for minimally invasive penetration of plant surfaces, enabling in situ and point-of-test analyses of biofuel-related biochemicals.

We developed prototype chemistry for nucleic acid hybridization on our bead-based diagnostics platform and we established an automatable bead handling protocol capable of 50 part-per-billion (ppb) sensitivity. We are working towards a platform capable of parallel, rapid (10 minute), raw sample testing for orthogonal (in this case nucleic acid and immunoassays) identification of biological (and other) threats in a single sensor microsystem. In this LDRD we developed the nucleic acid chemistry required for nucleic acid hybridization. Our goal is to place a non-cell associated RNA virus (Bovine Viral Diarrhea, BVD) on the beads for raw sample testing. This key pre-requisite to showing orthogonality (nucleic acid measurements can be performed in parallel with immunoassay measurements). Orthogonal detection dramatically reduces false positives. We chose BVD because our collaborators (UC-Davis) can supply samples from persistently infected animals; and because proof-of-concept field testing can be performed with modification of the current technology platform at the UC Davis research station. Since BVD is a cattle-prone disease this research dovetails with earlier immunoassay work on Botulinum toxin simulant testing in raw milk samples. Demonstration of BVD RNA detection expands the repertoire of biological macromolecules that can be adapted to our bead-based detection. The resources of this late start LDRD were adequate to partially demonstrate the conjugation of the beads to the nucleic acids. It was never expected to be adequate for a full live virus test but to motivate that additional investment. In addition, we were able to reduce the LOD (Limit of Detection) for the botulinum toxin stimulant to 50 ppb from the earlier LOD of 1 ppm. A low LOD combined with orthogonal detection provides both low false negatives and low false positives. The logical follow-on steps to this LDRD research are to perform live virus identification as well as concurrent nucleic acid and immunoassay detection.

Terrorism has been called 'war in the 21st century' (Barnett). Metrics are needed to quantify changes in threat state as well as system design trade-offs that weight the time required for accurate identification against the threat detection sensitivity. We present an adaptable agent-detection-platform capable of sensing multiple threat agents (chemical, biological and nuclear) simultaneously in diverse media and a performance model of this detection platform. A performance metric 'Time-To-Identify' (TTI); the time required to identify the presence of a given target analyte is presented. Experimental results that illustrate the new technologies required by this platform are presented. The model and metric predict the detection platform's sensitivity and speed for several CONcepts of OPeration (CONOP): (1) agent detection (chemical, viral or bacterial) in a 1 ml clinical sample, (2) botulinum toxin detection in milk and (3) pathogen detection in airplane cabin air.