Two of the greatest challenges in MEMS are those of packaging and integration with CMOS technology. Development of unreleased MEMS resonators at the transistor level of the CMOS stack will enable direct integration into front-end-of-line (FEOL) processing and minimal or no packaging, making these devices an attractive choice for on-chip signal generation.

Toward this goal, the authors have previously demonstrated the first fully unreleased MEMS resonator operating at 39 GHz with a quality factor (Q) of 129 ^[1] . The Si bulk acoustic resonator, surrounded on all sides by SiO₂, demonstrates the feasibility of unreleased resonators, providing a Q that is only 4x lower than its released counterpart ^[2] . At mm-wave frequency in the Landau-Rumer regime, resonator Q is limited primarily by anchor loss ^[3] . In the case of fully-clad resonators, the quality factor can be significantly improved by localization of acoustic energy using acoustic Bragg reflectors ^[4] .

The HybridMEMS lab has performed a study of fully unreleased resonator surrounded by lithographically defined ABRs, embedded in a homogeneous cladding layer (Figure 1). This one-mask design enables resonator banks of various frequencies on the same chip, providing multiple degrees of freedom in ABR design. With the goal of direct integration into FEOL CMOS processing, resonator performance is investigated for materials commonly found in the CMOS stack. The characteristics of these unreleased structures are compared with freely suspended resonators, released resonators isolated with lithographically defined ABRs ^[5] , and phononic crystal ^[6] based unreleased resonators (Figure 2).

1. W. Wang, L. C. Popa, R. Marathe, and D. Weinstein, “An unreleased mm-wave resonant body transistor,” IEEE MEMS Conference, 2011, pp. 1341-1344.
2. D. Weinstein and S. A. Bhave, “Acoustic resonance in an independent-gate FinFET,” Hilton Head Workshop, 2010, pp. 459-462.
3. R. Tabrizian, M. Rais-Zadeh, and F. Ayazi, “Effect of phonon interactions on limiting the f.Q product of micromechanical resonators,” IEEE Transducers Conference, 2009, pp. 2131-2134.
4. K. M. Lakin, “Thin film resonators and filters,” IEEE Ultrasonics Symposium, 1999, pp. 895-906.
5. R. H. Olsson, J. G. Fleming, and M. R. Tuck, “Contour mode resonators with acoustic reflectors,” US Patent 7385334 B1, 2008.
6. S. Mohammadi, A. A. Eftekhar, W. D. Hunt, and A. Adibi, “High-Q micromechanical resonators in a two-dimensional phononic crystal slab,” Applied Physics Letters, vol. 94, pp. 051906:1-3, Feb. 2009.

We integrated ultra-porous (99% porous) elements composed of nanoporous forests of vertically aligned carbon nanotubes (VACNTs) in MEMS, demonstrating their use in microfluidic applications for bioparticle isolation and health diagnostics.  Distinct from prior works where the effects of fluids on VACNT elements would result in either structural deformation or catastrophic forest collapse ^[1] , the approach here enables creation of high aspect ratio (~ 1-mm in height) nanoporous elements and preserves their shape even under flow-through conditions. An example device, shown in Figure 1, consists of a patterned and (wet) functionalized VACNT forest integrated into a PDMS microfluidic channel.

Compared to state-of-the-art designs that exploit solid materials (e.g., silicon, PDMS) for the structural features, our nanoporous elements enables fluid flow both around and through the VACNT elements, thus enhancing physical interaction between the particles in the flow and the functional elements. A ~7X increase in specific bioparticle capture when transitioning to VACNT porous designs was demonstrated for multiple device layouts ^[1] . The large surface-to-volume ratio of nanoporous materials yields also a significant increase in functional surface area (~250-500X for the layouts analyzed in our works ^{[2] [3]} ), thus further promoting bioparticle capture.

Specific isolation of bioparticles ranging over 4 orders of magnitude in size (from cells to viruses) was experimentally demonstrated (Figure 1), including the ability to perform simultaneous multiphysics, multiscale isolation on a single chip. Particles smaller than the average distance between single nanotubes in the VACNT elements (~80 nm) can penetrate the elements and can be isolated using chemical affinity; simultaneously, particles larger than 80 nm cannot enter the nanoporous elements and can be isolated on the elements’ outer surfaces using both mechanical filtration and biomolecular recognition. The nanoporous elements are versatile and could provide access to underexplored sub-micron particles (e.g., proteins, exosomes).

1. D. N. Futaba, K. Hata, T. Yamada, T. Hiraoka, Y. Hayamizu, Y. Kakudate, O. Tanaike, H. Hatori, M. Yumura, and S. Iijima. “Shape-engineerable and highly densely packed single-walled carbon nanotubes and their application as super-capacitor electrodes,” Nature Materials, vol. 5, pp. 987-994, 2006.
2. G. D. Chen, F. Fachin, M. Fernandez-Suarez, B. L. Wardle, and M. Toner, “Nanoporous elements in microfluidics for multiscale nanipulation of bioparticles,” Small, vol. 7, pp. 1061-1067, 2011.
3. F. Fachin, G. D. Chen, M. Toner, and B. L. Wardle, “Integration of vertically-aligned carbon nanotube forests in microfluidic devices for multiscale isolation of bioparticles,” Proc. IEEE Sensors 2010, Kona, HI, pp. 47-51.

Characterization of thin film layered materials is critical for many MEMS devices. Residual stresses from production determine both final shape and performance of microdevices and should therefore be accurately determined. Stresses are typically extracted using simple test structures (clamped beams and cantilevers, see Figures 1a-b) that allow for mean and gradient residual stress estimation ^[1] . However, current approaches to material characterization have two major limitations. First, their accuracy is directly proportional to their cost. This is especially true for mean compressive stress, where more accurate estimates require a larger number of different test structures. Second, they oversimplify test-structure boundary conditions by considering them to be ideal (e.g., perfectly clamped in the case of buckled beams for mean compressive stress determination ^[1] ). To overcome these issues, we have developed a new methodology for characterizing the complete stress state (effective mean and gradient stresses) in CMOS layered materials that also assesses non-ideality of clamped boundaries ^[2] . The approach uses a closed-form solution of the postbuckling problem of micromachined beams including non-ideal boundaries (Figure 1). In Table 1 we show the results relative to the characterization of four different CMOS material combinations. The outcomes show mean compressive stresses ranging between -15 and -105MPa, thus demonstrating the method’s capability to characterize structures subjected to both large and small compressive stresses. This capability is in contrast with traditional critical length methods that encounter difficulties in quantifying small compressive stresses due to their inability to distinguish between mean stress and gradient stress effects ^[2] . For the CMOS materials examined here, the accuracy was ± 2MPa for mean stresses and ±3MPa/µm for gradients. Boundary non-ideality is found to be 90% of perfectly clamped for the CMOS-released films, having such a significant effect on the extracted stresses that it must be considered.

1. M. J. Madou, Fundamentals of Microfabrication: The Science of Miniaturization, 2nd Edition, CRC Press, New York, 1997.
2. F. Fachin, M. Varghese, S. A. Nikles, and B. L. Wardle, “Characterization of the complete stress state in thin-film CMOS layered materials,” in Proc. Hilton Head, Hilton Head, SC, pp. 312-315.

The recent development of “low power” (10s-100s of μW) sensing and data transmission devices, as well as protocols with which to connect them efficiently into large, dispersed networks of individual wireless nodes, has created a need for a new kind of power source. Embeddable, non-life-limiting power sources are being developed to harvest ambient environmental energy available as mechanical vibrations, fluid motion, radiation, or temperature gradients. While potential applications range from building climate control to homeland security, the application pursued most recently has been that of structural health monitoring (SHM), particularly for aircraft. This SHM application and the power levels required favor the piezoelectric harvesting of ambient vibration energy. Current work focuses on harvesting this energy with MEMS resonant structures of various geometries. Coupled electromechanical models for uniform beam structures have been developed to predict the electrical and mechanical performance obtainable from ambient vibration sources. The optimized models have been verified by comparison to tests on a macro-scale device both without ^[1] and with a proof mass at the end of the structure (Figure 1) ^[2] . A non-optimized, uni-morph beam prototype (Figure 2) has been designed and fabricated ^{[3] [4]} . Design tools to allow device optimization for a given vibration environment have been under detailed investigation considering various geometries of the device structures and fabrication constraints, especially in microfabrication. Future work will focus on fabrication and testing of optimized unimorph beams for not only {3-1} mode but also {3-3} mode of operation using interdigitated electrode configuration.

1. N. E. duToit and B. L. Wardle, “Experimental verification of models for microfabricated piezoelectric vibration energy harvesters,” AIAA Journal, vol. 45, no. 5, pp. 1126-1137, May 2007.
2. M. Kim, M.  Hoegen, J. Dugundji, and B. L. Wardle^. “Modeling and experimental verification of proof mass effects on vibration energy harvester performance,” Smart Mater. Struct., vol. 19, p. 045023, 2010.
3. N. E. duToit, B. L. Wardle, and S.-G. Kim, “Design considerations for MEMS-scale piezoelectric mechanical vibration energy harvesters,” Integrated Ferroelectrics, vol. 71, pp. 121-160, 2005.
4. N. E. duToit and B. L. Wardle, “Performance of microfabricated piezoelectric vibration energy-harvesters,” Integrated Ferroelectrics, vol. 83, pp. 13-23, 2006.

Vertically-aligned carbon nanotube (CNT) arrays are grown on a moving substrate, demonstrating continuous growth of nanoscale materials with long-range order. A cold-wall chamber with an oscillating moving platform (Figure 1) is used to locally heat a silicon growth substrate coated with a Fe/Al₂O₃ catalyst film for CNT growth via chemical vapor deposition.  The reactant gases are introduced over the substrate through a directed nozzle to attain high-yield CNT growth.  Aligned multi-wall carbon nanotube (MWNT) arrays (or “forests”) with heights of »1 mm are achieved at substrate speeds up to 2.4 mm/s. Arrays grown on moving substrates at different velocities are studied to identify potential physical limitations of repeatable and fast growth on a continuous basis. No significant differences are noted between static and moving growth as characterized by SEM (Figure 2) and Raman spectroscopy, although overall growth height is marginally reduced at the highest substrate velocity. CNT arrays produced on moving substrates are also found to be comparable to those produced through well-characterized batch processes consistent with a base-growth mechanism. Growth parameters required for the moving furnace are found to differ only slightly from those used in a comparable batch process; thermal uniformity appears to be the critical parameter for achieving large-area uniform array growth.

Once the parameters have been optimized, a desktop continuous growth apparatus has been designed and implemented to grow VACNTs on silicon wafers (Figure 2), flexible sheets, and alumina fibers continuously. We have demonstrated and reported the ability to manufacture VACNT arrays in a continuous manner, significantly reducing the time spent, energy consumed, and reaction products created as compared to batch processing of these technologically valuable assemblies of nanoscale materials.

1. R. Guzman de Villoria, S. L. Figueredo, A. J. Hart. S. A. Steiner III, A. H. Slocum, and B. L. Wardle, “High-yield growth of vertically aligned carbon nanotubes on a continuously moving substrate,” Nanotechnology, vol. 20, no. 40, pp. 405611-405618, 2009.

Catastrophic structural failures are the cause of many physical and personal losses, with prevention estimated at billions of dollars in savings each year. Non-destructive evaluation (NDE) techniques have been pursued and employed for damage detection of such structures to detect cracks and other damage at pre-critical levels for remediation ^{[1] [2]} . Here, a novel multi-physics approach is reported that addresses drawbacks in existing techniques by taking advantage of the effects that damage, such as a crack, has on the electric and thermal transport in a material containing a CNT network distributed in the bulk material. When a potential is applied to a nano-engineered structure (Figure 1), electric field lines concentrate in the vicinity of cracks as electrons flow around damage, causing field concentrations and “hot spots” via Joule heating, an effect which is amplified because the heat flow is also impeded in areas of damage (e.g., across a crackface). These changes of temperature can be localized through a conventional infrared thermal camera. Low power operation (a 9-V standard battery is exemplary, providing a 15C rise at 1 Watt as in Figure 2), and high spatial resolution is demonstrated that is beyond state-of-the-art in non-destructive evaluation.

Using this technique, multiple applications have been identified such as crack detection in composite components that are joined by metallic fasteners, structures having internal nonvisible damage due to impact, and in situ progressive damage monitoring during a tensile strength test.  The thermal nano-engineered NDE technique demonstrated here can provide a new and effective inspection route for monitoring next-generations of safer infrastructure ^{[3] [4]} .

1. D. Barber, S. Wicks, A. Raghavan, C. T. Dunn, ,S. S. Kessler, and B. L. Wardle, “Health monitoring of aligned carbon nanotube (CNT) enhanced composites,” presented at 2009 SAMPE Fall Technical Conference, Wichita, KS, Oct. 2009.
2. S. Wicks, A. Raghavan,, R. Guzmán de Villoria, S. S. Kessler, and B. L. Wardle, “Tomographic electrical resistance-based damage sensing in nano-engineered composite structures,” in AIAA-2010-2871, presented at 51st AIAA Structures, Structural Dynamics, and Materials (SDM) Conference, Orlando, FL, April 12-15, 2010.
3. R. Guzman de Villoria, N. Yamamoto, A. Miravete, and B. L. Wardle, “Multi-physics damage sensing in nano-engineered structural composites,” Nanotechnology, vol. 22,  pp. 185502-185508, 2011.
4. R. Guzmán de Villoria, A. Miravete, N.Yamamoto, and B. L. Wardle, “Enhanced thermographic damage detection enabled by multifunctional nano-engineered composite laminates,” in AIAA-2011-1798, presented at 52^nd AIAA Structures, Structural Dynamics, and Materials (SDM) Conference, Denver, CO, April 4-7, 2011.

This work focuses on development of non-metallic substances that can catalyze carbon nanotube (CNT) growth while remaining in a non-metallic state.

We recently reported that nanoparticulate zirconia (ZrO₂) catalyzes both growth of single-walled CNTs and multi-walled CNTs by thermal chemical vapor deposition (CVD) and graphitization of solid amorphous carbon ^[1] .  We observe (Figure 1) that silica-, silicon nitride-, and alumina-supported zirconia on silicon nucleates single- and multiwall CNTs upon exposure to hydrocarbons at moderate temperatures (750°C). Using in situ, high-pressure, time-resolved X-ray photoelectron spectroscopy (XPS) of these substrates during CNT nucleation and growth (Figure 2), we show that the zirconia catalyst neither reduces to a metal nor forms a carbide. Point-localized energy-dispersive X-ray spectroscopy using scanning transmission electron microscopy confirms that catalyst nanoparticles attached to CNTs are zirconia. We also observe that carbon aerogels containing zirconia nanoparticles prepared through pyrolysis of a Zr(IV)-containing resorcinol-formaldehyde polymer aerogel precursor at 800°C contain fullerenic cage structures absent in undoped carbon aerogels. Zirconia nanoparticles embedded in these carbon aerogels are further observed to act as nucleation sites for multi-walled CNT growth upon exposure to hydrocarbons at CVD growth temperatures.

Our study clearly demonstrates, for the first time, that a non-metallic catalyst can catalyze CNT growth by thermal CVD while remaining in an oxidized state.  Currently, the yield of CNTs obtained from zirconia is significantly lower than from traditional transition metal catalysts such as Fe. Studies to identify how to optimize yield from oxide catalysts not reducible under CVD growth conditions, the role of defects in these catalysts, and the sensitivity of these catalysts to different gas-phase species are underway. We believe that characterization of the mechanisms underlying CNT growth from oxides will aid in the engineering of optimized non-metallic catalysts for CNT growth on historically challenging substrates such as carbon fiber and may offer a promising route towards control of CNT and other nanostructure characteristics.

1. S. A. Steiner III, T. F. Baumann, B. C. Bayer, R. Blume, M. A. Worsley, W. J. MoberlyChan, E. L. Shaw, R. Schloegl, A. John Hart, S. Hofmann, and B. L. Wardle, “Nanoscale zirconia as a nonmetallic catalyst for graphitization of carbon and growth of single- and multiwall carbon nanotubes,” Journal of the American Chemical Society, 2009 vol. 131, no. 34, pp. 12144-12154.

Carbon nanotube (CNT) composites are promising new materials for structural applications thanks to their mechanical and multifunctional properties. We have undertaken a significant experimentally-based program to understand both microstructures of aligned-CNT nanocomposites and nano-engineered advanced composite macrostructures hybridized with aligned CNTs.

Aligned nanocomposites are fabricated by mechanical densification and polymer wetting of aligned CNT forests ^[1] . Polymer wetting is driven by capillary forces that arise upon contact of the polymer with the nanostructured CNT forest ^{[2] [3]} , the rate of which depends on properties of the CNT forest (e.g., volume fraction) and the polymer (viscosity, contact angle, etc.). Here the polymer is unmodified aerospace-grade epoxy. CNT forests are grown to mm-heights on 1-cm² Si substrates using a modified chemical vapor deposition process. Following growth, the forests are released from the substrate and can be handled and infiltrated. The volume fraction of the as-grown CNT forests is about 1%; however, the distance between the CNTs (and thus the volume fraction of the forest) can be varied by applying a compressive force along the two axes of the plane of the forest to give volume fractions of CNTs exceeding 20% (Figure 1a). Variable-volume fraction-aligned CNT nanocomposites were characterized using optical, scanning electron (SEM) and transmission electron (TEM) microscopy to analyze dispersion and alignment of CNTs as well as overall morphology. Extensive physical property testing is underway.

Nano-engineered composite macrostructures hybridized with aligned CNTs are prepared by placing long (>20 μm) aligned CNTs at the interface of advanced composite plies as reinforcement in the through-thickness axis of the laminate (Figure 2). Three fabrication routes were developed: transplantation of CNT forests onto pre-impregnated plies ^[4] (the “nano-stitch” method), placement of detached CNT forests between two fabrics followed by subsequent infusion of matrix, and in situ growth of aligned CNTs onto the surface of ceramic fibers followed by infusion or hand-layup ^{[5] [6] [7]} . Aligned CNTs are observed at the composite ply interfaces and give rise to significant improvement in interlaminar strength, toughness, and electrical properties. Interestingly, toughness improvement has demonstrated a favorable nano-scale size effect ^[7] . Analysis of the multifunctional properties and nanoscale interactions between the constituents in both the nanocomposites and hybrid macrostructures is underway. A new route to fabricate these materials in a continuous way has been developed.

1. B. L. Wardle, D. S. Saito, E. J. Garcia, A. J. Hart, and R. Guzman de Villoria, “Fabrication and characterization of ultra-high volume fraction aligned carbon-nanotube-polymer composites,” Advanced Materials, vol. 20, pp. 2707-2714, 2008.
2. H. Cebeci, R. Guzman de Villoria, A. John Hart, and B. L. Wardle, “Multifunctional properties of high volume fraction aligned carbon nanotube polymer composites with controlled morphology,” Composites Science and Technology, vol. 69, nos. 15-16, pp. 2649-2656, 2009.
3. E. J. García, A. J. Hart, B. L. Wardle, and A. H. Slocum, “Fabrication and nanocompression testing of aligned carbon-nanotube-polymer nanocomposites,” Advanced Materials, vol. 19, pp. 2151-2156, 2007.
4. E. J. Garcia, B. L. Wardle, and A. J. Hart, “Joining prepreg composite interfaces with aligned carbon nanotubes,” Composites Part A, vol. 39, pp. 1065-1070, 2008.
5. N. Yamamoto, A. J. Hart, S. S. Wicks, E. J. Garcia, B. L.Wardle, and A. H. Slocum, “High-yield atmospheric-pressure growth of aligned carbon nanotubes on ceramic fibers for multifunctional enhancement of structural composites,” Carbon, vol. 47, pp. 551-560, 2009.
6. S. S. Wicks, R. Guzmán de Villoria, and B .L. Wardle, “Interlaminar and intralaminar reinforcement of composite laminates with aligned carbon nanotubes,” Composites Science and Technology, vol. 70, pp. 20–28, 2010.
7. J. Blanco, E. J. Garcia, R. Guzman de Villoria, and B. L. Wardle, “Limiting mechanisms in mode I interlaminar toughness of composites reinforced with aligned carbon nanotubes,” Journal of Composite Materials, vol. 43, no. 8, pp. 825-841, 2009.

GaN-based high electron mobility transistors (HEMTs) have great potential for high power/frequency applications due to their outstanding combination of large breakdown voltage and high electron velocity. Among the different possible nitride structures, InAlN/GaN heterostructures have attracted much attention recently because they enable an extremely high charge density with a thin barrier thickness ^[1] . With the use of these advantages, outstanding progress in the frequency performance of InAlN/GaN transistors has been recently achieved. Sun et al. reported a 55-nm gate length device with f_T of 205 GHz (f_max = 191 GHz) ^[2] and Lee et al. demonstrated a 30-nm gate length device with f_T of 245 GHz ^[3] .

In this study, we used an AlGaN back-barrier in InAlN/GaN HEMT structures for the first time and studied its impact on the DC and RF characteristics of these devices ^[4] . A maximum drain current of 1.49 A/mm is obtained at V_gs=2 V in the device with the back-barrier, about 27 % lower than that of the standard device (2.05 A/mm at V_gs=2 V). The smaller drain current in the device with the back-barrier mainly results from the lower sheet charge density and subsequent higher threshold voltage. However, the output conductance is significantly smaller in the device with the back-barrier, which shows an effective suppression of the  short-channel effects. In addition, in sub-100-nm-gate-length transistors, the back-barrier makes it possible to maintain a drain-induced barrier lowering (DIBL) near 50-60 mV/V while preventing the degradation of the subthreshold swing (SS). Thanks to the reduced short-channel effects, 65-nm-gate-length devices with a back-barrier showed an f_T of 210 GHz, which is higher than that of the standard device with the same gate length (195 GHz). Moreover, in a sub-30-nm-gate-length device with AlGaN back-barrier, an f_T of 270 GHz, the highest f_T ever reported in GaN transistors, was achieved.

1. J. Kuzmik, “Power electronics on InAlN/(In)GaN: Prospect for a record performance,” IEEE Electron Device Lett. vol. 22, no. 11, pp. 510-512, Nov. 2001.
2. H. Sun, A. R. Alt, H. Benedickter, E. Feltin, J.-F. Carlin, M. Gonschorek, N. Grandjean, and C. R. Bolognesi, “205-GHz (Al, In)N/GaN HEMTs,” IEEE Electron Device Lett., vol. 31, no. 9, pp. 957-959, Sep. 2010.
3. D. S. Lee, J. W. Chung, H. Wang, X. Gao, S. Guo, P. Fay, and T. Palacios, “245 GHz InAlN/GaN HEMTs with oxygen plasma treatment,”  IEEE Electron Device Lett., vol. 32, no.6, pp.755-757, Jun. 2011.
4. D. Lee, X. Gao, S. Guo, and T. Palacios, “InAlN/GaN HEMTs with AlGaN back-barriers,” IEEE Electron Device Lett., vol. 32, no. 5, pp.-617-619, May 2011.

Optogenetics is commonly used for precision modulation of the activity of specific neurons within neural circuits ^[1] , but assessing the impact of optogenetic neural modulation on millisecond-timescale local and global circuit neural activity remains difficult.  We have developed a novel strategy for designing and fabricating silicon-based microelectrode arrays with customizable electrode locations, targetable to defined neural substrates distributed in a 3-D pattern throughout a neural network in the mammalian brain, and compatible with simultaneous use of a diversity of existing light delivery devices.  Our design of these 3-D electrode arrays provides for both easy electrical and mechanical assembly, and provides for scaling of arrays to up to 1000 neural recording channels and beyond.

Our approach relies upon a number of innovations at the material, structural, electrical, and data acquisition levels.  First, typical silicon-based electrodes that are arranged in a 1-dimensional linear array, or 2-dimensional comb-like fashion, often use linear or tetrode-style electrode locations along the comb’s fingers, with stereotyped spacing and pad sizes.  Our software-driven approach enables variable spacing and pad sizes, so that electrode geometries can be customized to the cellular properties of the brain circuits under investigation.  Second, to support the assembly of such electrode arrays into a 3-dimensional array, we have developed novel electrical and mechanical connector strategies to make the assembly as automated and reliable as possible.  Third, we have developed strategies for amplifying and acquiring data that simplify the use of these probes in an intact, in vivo, mammalian context.  Fourth, we have implemented hybrid electrodes that contain both a low-impedance metallic pad for recording of spike activity, as well as an indium tin oxide (ITO) pad that can report local field potentials (LFPs) without the photo-electrochemical artifacts common in optogenetics.  Finally, these 3-D probes are designed to be easy to use, from design to surgery.  We have developed a user-friendly interface that enables neuroscientists to specify probe geometries based upon neural target geometries and coordinates, and are developing supporting surgical and behavioral strategies for use of such arrays in vivo.

1. X. Han and E. S. Boyden, “Multiple-color optical activation, silencing, and desynchronization of neural activity, with single-spike temporal resolution,” PLoS ONE, vol. 2, no. 3, p. e299, Mar. 2007.