of temperature using a controlled-temperature chuck to sweep from room .... [2] E. Sovero, D. Deakin, J. Higgins, J. DeNatale, and S. Pittman, âFast thin.
1
Model and Characterization of VO2 Thin-Film Switching Devices Tyler S. Jordan∗ Student Member, IEEE, Sean Scott† Member, IEEE, Darin Leonhardt∗ , Joyce O. Custer‡ , Christopher T. Rodenbeck∗ Senior Member, IEEE, Steve Wolfley∗ , Christopher D. Nordquist∗ Senior Member, IEEE
Abstract—This paper investigates and models the DC behavior of thin-film-based switching devices. The devices are based on Ω at room sputtered VO2 thin films that transition from 200 k Ω temperature to 390 at temperatures above 68◦ C, with the transition occurring over a narrow temperature range. The device resistance is characterized over temperature and under currentsourced and voltage-sourced electrical bias. The finite-element model predicts the device’s non-uniform switching behavior. Electrothermally-heated devices show the same transition ratio and switching behavior as externally heated devices suggesting a purely electrothermal switching mechanism. Index Terms—Vanadium compounds, Thin film devices, Switches, Resistive circuits.
I. I NTRODUCTION
T
HE latest microwave technology demands switches and limiters with high-power density characteristics, low loss, and small dimensions. Because of its unique properties vanadium dioxide (VO2 ) is a candidate for volatile RF limiters [1] and switches [2]. Unlike traditional semiconductors such as silicon whose properties are manipulated by electrical bias, VO2 undergoes a phase change (PC) metal insulator transition (MIT) from a monoclinic semiconductor state to a tetragonal metallic state when the temperature increases above 68◦ C. At this temperature the material experiences a sharp resistivity drop of several orders of magnitude [3] (approximately from 10 Ω·cm to 1 mΩ·cm) as well as an optical reflectance reflectance change [4], [5]. When the temperature drops below its transition point, VO2 returns to the high-resistance state. The exact nature of this transition is unclear and evidence has been presented for switching mechanisms that are thermal [3], [6], electrical [7], [8], [9], electrothermal [10], and optical [11], [12]. This work focuses on the nature and role of the switching mechanism and the performance implications in VO2 -based switching devices. 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. This work was supported by the Laboratory Directed Research and Development program. ∗ Sandia National Laboratories, † now with Purdue University, ‡ Sandia Staffing Alliance
l
Current Direction
A=l×w
IN
OUT w
Au
Ni
Si3N4
VO2 𝑡
Si
Fig. 1: Basic device geometry. The parallel arrows show the current flow from one terminal to the other through the VO2 material. l is the device dimension in the direction of current flow, w is the device dimension perpendicular to current flow, t is the thickness of the substrate.
II. T HEORY /C ALCULATION A. Behavioral Model Fig. 1 shows the basic lateral device layout. Electrical power dissipated in the film heats the film to its transition temperature which causes a resistance drop. We present a simplified equivalent behavioral model for the films below. The electrical film behavior is modeled with n parallel electrical resistors whose combined resistance is the device resistance (Fig. 2a). As current or voltage is applied to each element, the dissipated power Joule-heats the element and its neighbors by thermal conduction. The resistor heat is modeled by thermal power sources which connect to n nodes (Fig. 2b). Each resistor temperature is determined by its respective node in the thermal model. Each thermal node connects to its neighbors with mutual thermal resistors and to the substrate with a ground thermal resistor. The two outer-most nodes are additionally connected to thermal ground via silicon thermal resistors, representing lateral heat conduction into the substrate. The electrically-generated heat causes a resistance change in each element which affects electricity flow and, in turn, heat generation. The coupled electrical and thermal changes cause the VO2 switch’s unique switching behavior. The following equation governs how dissipated power affects each node’s temperature: ∆T = θP
(1)
where T is temperature, θ is thermal resistance to thermal ground, and P is power dissipated in the node. The power that enters the end nodes is:
0000–0000/00$00.00©2013 Sandia National Laboratories
2
such that the measured device resistance is proportional to ~ containing the VO2 film resistance. The resistance vector, R, the resistance at each node is a function of each node’s temperature. ~ = nRf ilm (T~ ) R (6)
Isrc R1 (T1 )
R2 (T2 )
R3 (T3 )
Rn (Tn )
For this model, the film sheet resistance as a function of temperature, Rf ilm (T ), is extracted using a look-up-table of experimental data obtained from 4-point-probe measurements. The sheet resistance must be multiplied by n because all n resistances are in parallel. The temperature of the film is a function of power, so:
(a)
P1 θM
P2 θM
P3
θM
Pn θSi
θSi θG
θG
θG
~ = nRf ilm (Tsubstrate + θP~ ) R
The power dissipated in the film depends upon the resistance of the film and the resistance as a function of current at each node is:
θG
~ = nRf ilm (Tsubstrate + θ(I~ I~ Rf ilm (T~ )) R (b) Fig. 2: (a) Electrical Circuit model: Each parallel resistor represents a VO2 element whose resistance is a function of its temperature. Each resistor temperature is determined by its respective node temperature in the thermal model. Either voltage or current can be sourced, depending on the test. (b) Thermal Model: The heat dissipated from each resistor element is modelled by corresponding power sources. Each node is connected to its neighbor through a mutual thermal resistance and to the substrate via a substrate thermal resistance. The two outer-most nodes are additionally connected to ground through a silicon thermal resistance.
P1 = Pn =
T1 T1 T1 − T2 + + θSi θG θM
(2)
Tn Tn Tn − Tn−1 + + θM θG θSi
and the power that enters the ith node is: Pi =
Ti − Ti−1 Ti Ti − Ti+1 + + θM θG θM
(7)
(3)
These equations can be simplified into the following matrix: 1 1 1 −1 0 ··· θSi + θG + θM θM −1 2 1 −1 θM θM + θG θM θ−1 = −1 2 1 −1 0 θM θM + θG θM .. .. . . (4) This inverse thermal resistance matrix allows each node’s temperature to be calculated simultaneously with the following equation: T~ = Tsubstrate + θP~ (5) Where T~ is the vector containing each node’s temperature, θ is the inverse of the matrix shown in (4), and P~ is the vector containing the amount of power from dissipated heat at each node. We assume that the device contact and interconnect resistance is negligible compared to the VO2 film resistance,
(8)
where is the Hadamard product (element-by-element multiplication). The resistance as a function of voltage at each node is: ~ = nRf ilm (Tsubstrate + θ(V ~ V ~ Rf ilm (T~ ))) R
(9)
where is the quotient equivalent of the Hadamard product ~ is a vector con(element-by-element division). In this case, V taining the voltage applied to each node. The total resistance of the device is given by: Rdevice = α(R1 ||R2 ||R3 || · · · ||Rn )
(10)
where α is a proportionality constant to allow for various device geometries. Equations (8) and (9) are recursive functions because Rf ilm is a function of itself. A MATLAB™ script was written to numerically compute this recursive function and produce a behavioral model for electrothermal switching assuming a linear current or voltage sweep. Fig. 3 shows the resistance versus dissipated power for sourcing both current and voltage. Current sourcing limits the dissipated power and exhibits multiple resistance levels as various sections of VO2 are switching, while voltage sourcing causes a current spike that could result in destructive power runaway depending on the value of any current-limiting resistance in series with the device. Fig. 4 shows the calculated temperature and resistance distributions along the width of the current-carrying channel as a function of the applied current or voltage. The top charts are modeled voltage-current curves. The four lower charts show a one-dimensional temperature or resistance distribution along the channel width for each current or voltage value. The xaxes of the four lower charts align with the x-axes of the top charts. The Temperature Distribution vs. Current chart in Fig. 4a shows uniform temperature distribution until switching where the device center significantly heats due to both the nature of thermal conduction and current crowding. When the device reaches its transition temperature, the resistance drop limits voltage so the overall power dissipated in the device drops
3
Resistance (Ω)
2
10
I−Source V−Source
1
10
0
10
0
0.1
0.2
0.3 0.4 Power (W)
0.5
0.6
Fig. 3: Resistance vs. Power for the modeled VO2 devices. The sharp resistance drop limits the voltage and power through the device and exhibits multiple resistance steps when current-sourced, but causes current and power runaway when voltage is sourced. Resistance values between the transition point and the final resistance value are unstable because this is a nearimmediate transition.
resulting in a temperature drop. The material maintains low resistance at this lower temperature due to hysteresis. Current takes the least-resistive path which causes more heating in the lower resistance areas in the center. Eventually, the entire device width heats past the transition temperature. The Normalized Resistance Distribution vs. Current plot in Fig. 4a shows uniform resistance distribution prior to switching. The device center switches first because the heat is most concentrated in the center. As more current is sourced, the transitioned area widens, completely switching the device. The effect of small VO2 sections switching at different current steps is evident in the R vs. P plot in Fig. 3 where there is a large initial drop in resistance followed by smaller drops. The Temperature Distribution vs. Voltage plot in Fig. 4b shows that when the device reaches its transition temperature, the resistance drop causes a current rise so power dissipated in the device spikes resulting in a sudden temperature rise. Unlike the current-sourcing case, the entire device transitions at the same power level under voltage-sourcing because voltage is constant across the device width, the current increases, and power spikes. The Normalized Resistance Distribution vs. Voltage plot in Fig. 4b shows that the entire device effectively transitions at the same voltage. This abrupt change can be seen in the R vs. P plot in Fig. 3.
the bottom of the substrate with constant cross-sectional area. Both are reasonable assumptions within our temperature range and wafer thickness. To find the device area required for switching at a given input power, we substitute for θ in (11) and get: t P. (12) A= ∆T k This relationship assumes that, prior to transition, the temperature and current flow is uniform over the entire film area. The film in the device must have an area A to switch at power P. The film resistance in a planar device at temperature T is given by: l (13) R(T ) = Rf ilm (T ). w Where R(T ) is the desired initial resistance of the device at a temperature T , Rf ilm is the sheet resistance of the film at the same temperature, l is the dimension parallel to current flow and w is the dimension perpendicular to current flow. Here the proportionality constant α in (10) is replaced by wl . Equation (12) provides the required film area for the device to switch at power P and (13) provides the film dimensions to obtain resistance R(T ). The substrate thickness (t) and thermal conductivity (k) can be controlled through the amount and type of materials that make up the substrate. This analysis is for gap devices but can be adapted to other geometries. C. Device Types Three device types were fabricated and tested: gap, interdigitated fingers (IDF), and metal-insulator-metal (MIM) (Fig. 5). The gap devices consist of two adjacent gold pads with a rectangular section of VO2 material connecting them. The IDFs have a gold trace of interdigitated fingers over a square of VO2 and function as the equivalent of gaps connected in parallel. The MIMs have the VO2 material sandwiched between the gold top metal and the nickel bottom metal. The MIM design allows low resistance with a relatively large area. The device design model can be modified to apply to each of the three devices. III. M ATERIALS AND M ETHODS
B. Device Design The three attributes of a VO2 switch are switching power threshold, resistance, and on/off ratio. On/off ratio is determined by the film as long as the film’s ‘on’ resistance is greater than any series resistance. Equation (1) tells us the power needed to heat the device to the transition temperature. For a planar device on a substrate such as the one shown in Fig. 1, θ is approximated as: t . (11) kA where t is the substrate thickness, k is the equivalent thermal conductivity of the substrate and A is the area over which power is dissipated (Fig. 1). For this approximation, we assume that k is constant as a function of temperature and the heat from the film flows directly down from the film to θ=
The devices were fabricated on a silicon substrate with resistivity greater than 5000 Ω·cm with a 300 nm-thick thermally grown silicon dioxide layer. A 100 nm-thick Ni bottom electrode was deposited and patterned by evaporation and lift-off. Following the bottom electrode, the VO2 film was deposited by reactive sputtering of a V target in an Ar/O2 (ratio of 125) ambient at a total system pressure of 1.3Pa, which generally resulted in a Vx Oy film with poor stoichiometry. After sputtering, the film was annealed using a tube furnace with an N2 ambient at 450◦ C for 30 minutes to achieve the desired VO2 stoichiometry. Four-point-probe testing indicates that the 200 nm-thick film has a sheet resistance of 200 Ω Ω at 40◦ C and 390 at 100◦ C, for a transition ratio of k approximately 500. After annealing, the films were coated with a 200 nm-thick PECVD-deposited Six Ny layer, which is patterned using dry
2 1 0.05
0.1 0.15 0.2 Current (A) Temperature Distribution (◦ C) Along Channel Width vs. Current 80 60 40
0
0.4 0.2 0
0
0.5 1 Voltage (V) Temperature Distribution (◦ C) Along Channel Width vs. Voltage
Channel Width
0
Channel Width
0
Current (A)
Voltage (V)
4
0.05
0.1 Current (A)
0.15
0.2
80 60 40
0
0.5 1 1.5 Voltage (V) Normalized Resistance Distribution Along Channel Width vs. Voltage
Channel Width
Normalized Resistance Distribution Along Channel Width vs. Current Channel Width
1.5
0
0.05
0.1 Current (A)
0.15
0.2
0
0.5 1 Voltage (V)
(a)
1.5
(b)
Fig. 4: Evolution of the one-dimensional resistance and temperature distributions as current or voltage is swept. The top charts are voltage-current measurements. The four lower charts show a one-dimensional temperature or resistance distribution along the channel width for each current or voltage value. The x-axes of the four lower charts align with the x-axes of the top charts. (a) shows the current-sourcing case and (b) shows the voltage-sourcing case.
MIM
IDF 150 μm
150 μm
150 μm
Gap
Au
Ni
VO2
SiO2
Fig. 5: Plan view optical micrographs and drawn cross-sections of the three types of devices: 5 µm × 150 µm gap, 10 fingers × 5 µm × 200 µm interdigitated fingers (IDF), and 50 µm × 150 µm metal insulator metal (MIM). Yellow is gold, light green is nickel, dark green is VO2 , and blue is silicon dioxide insulator.
IV. E XPERIMENTAL A. Thermal Switching The VO2 film sheet resistance was measured as a function of temperature using a controlled-temperature chuck to sweep from room temperature up to a peak temperature and back
down again. Each resistance measurement is extracted from a computer-controlled four-point current-voltage sweep using a Keithley 2400 Source Measurement Unit (SMU). During the measurement, the source current was auto-ranged to be as low as possible to ensure good measurement quality while avoiding electrothermally heating the film. Typical measurement currents were in the hundreds of microamperes. The sweeps were linear, showed no signs of recitfying behavior at low voltage regions, and showed no indication of electrical heating. The measured film experiences a sharp resistance drop of about 3 orders of magnitude near 68◦ C with increasing temperature. Upon decreasing temperature the film returns to a high-resistance state at approximately 50◦ C due to hysteresis (Fig. 6). Because of this hysteresis, the power will need to drop lower than the switching power to return to high-resistance.
Sheet Resistance Ω ( sq uare )
etching to make contact to the VO2 films. Finally, the device interconnect and pads were completed by lift-off patterning and evaporation of 500 nm of Ni/Au. This process is used to fabricate planar gap devices, planar interdigitated devices, and metal-VO2 -metal devices.
5
10
4
10
3
Cooling
2
Heating
10 10
0
20
40 60 80 ◦ Temperature ( C)
100
Fig. 6: Resistance vs. temperature for a VO2 film. The resistance drops nearly 3 orders of magnitude at 68◦ C and demonstrates hysteresis upon cooling.
5
Resistance (Ω)
A four-point measurement using a computer-controlled Keithley 2400 SMU with Kelvin probes was used to determine the current through and the voltage across the device. The power dissipated within the device was calculated by the product of the current and voltage. Fig. 7 shows the resistance vs. power curve for a 50 µm x 140 µm MIM device with the substrate held at 25◦ C. This transition occurs on the microsecond time scale. Consistent with the modeled results, the transition power is identical for the two measurement modes, with the dissipated power being limited in currentsourcing and spiking during voltage-sourcing. In the case of this MIM device, the minimum device resistance and on/off ratio are limited by the resistance of the bottom electrode. The same switching mechanism was observed in all three types of devices.
3 2
Resistance (Ω)
B. Electrical Switching
1
Thermal
2
Electrothermal
2
10
3 4
1
10
40
50
5
60 70 Temperature (◦ C)
6 80
90
(a)
(1) 0 mW
(2) 296 mW
(3) 612 mW
(4) 731 mW
(5) 894 mW
(6) 984 mW
(b)
I−Source V−Source
MIM : 50 µm × 140 µm Effective Area 1 −2 −1 10 10 Power (W)
0
10
Fig. 7: Resistance vs. power for a 50 µm x 140 µm MIM device, tested in both current-sourced and voltage-sourced mode. These results are consistent with the presented model.
Because the temperature increase is proportional to dissipated power (see (1)), a uniformly electrothermally-heated VO2 device would resemble the smooth resistance versus temperature curve if the power over the device was uniformly distributed, unaffected by the resistance drop, and maintained a consistent sweep. However, we observe that electrical switching distinctly differs from thermal switching. Fig. 8a shows a 6-finger 50 µm × 200 µm IDF device switched both electrically and thermally. The thermal curve was measured with low current, assuming no electrothermal heating. The electrical curve was taken by sourcing current and calculating an equivalent temperature rise by (1). The thermal curve is smooth while the electrical curve switches abruptly at a lower substrate temperature and exhibits what appears to be multiple transitions. The multiple transitions occur because only parts of the film (usually starting at the center) heat up and transition at a time. This behavior has been previously reported by others[1], [13], [14], is described earlier in Section II-A, and is seen in our model results in Fig. 4. While the measurement in Fig. 8a was conducted, the micrographs shown in Fig. 8b were taken at each current step. The images show that each filament that switches lowers the overall resistance of the device, lowering the applied voltage and power over the rest of the film, thus reducing heating of the non-transitioned region. As the current is increased and the power over those filaments increases, the heat spreads from the center to other sections, allowing those sections to transition, causing the multiple steps. The multiple
Fig. 8: (a) Resistance vs. film temperature for thermal and electrical switching. The electrical switching curve is normalized to temperature using (1). Sections of the film switch at a time, showing very different behavior for electricallyand thermally-switched IDF devices with five 14.7 µm long × 200 µm wide gaps. (b) Enhanced color images of the same device under current-sourced bias show the high and low resistance regions of the film which differ in relfectivity. Each numbered image correspond to the labels in (a). The heated, transitioned film is a different color because of the reflectivity change due to the structural phase change.
steps occur with each of the three device designs but are most pronounced in the IDFs because the design allows several small sections between fingers to exhibit the center-heating effect. Interestingly, a device exhibits the same steps at the same power levels each time it is switched. The MIT in VO2 has been observed at temperatures lower than its transition temperature when switched electrically, as suggested by numbered transition ”2” in Fig. 8a. One group believes this is possibly due to the transition of small filaments [15]. Other groups postulates that a lower temperature transition can be caused by an increase of electron density [8], [16]. To investigate the early switching mechanism of VO2 , we conducted resistance versus temperature sweeps at various voltage biases. When temperature is adjusted for the voltage bias electrothermal heating as in Fig. 9, all resistance versus temperature curves lie on the same thermal switching curve, indicating that the transition in VO2 is purely thermal for DC electrothermal switching. To further investigate the switching mechanism, we electrothermally switched the devices at different temperatures. Fig. 10 shows multiple current-sourced resistance versus power curves taken at various temperatures. The inset figure shows a linear decline in switching power versus temperature. An extrapolation of the linear progression intercepts the Substrate Temperature axis at the transition temperature which means that when the substrate is at the transition temperature, no power is required to heat it to transition temperature. The data presented in Fig. 7 and Fig. 8 fit our model and corroborate its validity. Fig. 9 and Fig. 10 suggest that for
Ω Sheet Resistance ( squar e)
6
5
10
Gap: 50 µm × 100 µm
4
10
1V
3
10
5V 7V
2
10
0
50 100 Film Temperature (◦ C)
150
Fig. 9: Resistance vs. temperature curves for a 50 µm long x 100 µm wide gap device: The x-axis is normalized to account for the electrothermal heating of the devices due to the voltage bias. The voltage bias electrothermally heats the device beyond the temperature of the substrate, causing the larger, abrupt transition. Though the substrate temperature sweep only went to 100◦ C, the 7 volt bias heated the device to nearly 150◦ C. All data lies on the same thermal transition curve.
300
Increasing IDF : Temperature 4 × 20 µm × 200 µm
200
24°C Switch Power (W)
Resistance (Ω)
from uniform device heating. Measurements of these devices under varying bias and substrate temperature reveal that the transition is due to thermal triggering via Joule heating. These observations have design and performance implications if metal-insulator-transition devices are used in switching and limiting applications.
100
20 0.1
56°C
0.4
Transition Temperature
0.2
θ = 92.5°C/W 0 20 40 60 Substrate Temperature (◦ C)
0.3 0.5 0.7 0.9 Power (W)
Fig. 10: Resistance vs. dissipated power for temperatures ranging from 24◦ C to 56◦ C in 2◦ C increments. As the temperature increases, the switching power decreases and the transition is eventually eliminated. The inset shows the switching power versus the temperature with the extrapolation to the transition temperature. The device is an IDF with four 20 µm wide, 200 µm long gaps (total gap length is 800 µm). The negative reciprocal slope of the inset provides the thermal resistance of the first filament that switches (see (1)).
these device types and stimulus, the MIT of VO2 is purely thermal for the electrothermal switching purposes. Fig 9 and Fig. 10 additionally support the purely-thermal mechanism that the model assumed. V. C ONCLUSION We have characterized and accurately modeled VO2 -based metal-insulator-transition devices to identify the impact of the electrothermal switching behavior on the terminal performance of these types of devices. Gap, interdigitated, and metalinsulator-metal devices have been characterized and all show similar response under current and voltage biasing. Correlation of electrical measurement and optical imaging shows that non-uniform heating and current crowding within the device are responsible for multi-step switching and the switching transitions at lower power and temperatures than predicted
ACKNOWLEDGMENT The authors acknowledge management support from C. T. Sullivan, as well as fabrication/materials support from M. Ballance, M. Cavaliere, P. Clem, J. Hunker and the MESAFab operations team. R EFERENCES [1] J. Givernaud, A. Crunteanu, J.-C. Orlianges, A. Pothier, C. Champeaux, A. Catherinot, and P. Blondy, “Microwave power limiting devices based on the semiconductor–metal transition in vanadium-dioxide thin films,” IEEE Trans. Microw. Theory Techn., vol. 58, no. 9, pp. 2352–2361, 2010. [2] E. Sovero, D. Deakin, J. Higgins, J. DeNatale, and S. Pittman, “Fast thin film vanadium dioxide microwave switches,” in GaAs IC Symp Tech Dig 1990, pp. 101–103, 1990. [3] F. Morin, “Oxides which show a metal-to-insulator transition at the neel temperature,” Phys. Rev Lett., vol. 3, no. 1, pp. 34–36, 1959. [4] M. Rini, Z. Hao, R. W. Schoenlein, C. Giannetti, F. Parmigiani, S. Fourmaux, J. C. Kieffer, A. Fujimori, M. Onoda, S. Wall, and A. Cavalleri, “Optical switching in vo2 films by below-gap excitation,” Appl. Phys. Lett., vol. 92, no. 18, p. 181904, 2008. [5] M. Soltani, M. Chaker, E. Haddad, R. Kruzelecky, and J. Margot, “Micro-optical switch device based on semiconductor-to-metallic phase transition characteristics of w-doped vo2 smart coatings,” J. Vac. Sci. Technol., A: Vacuum, Surfaces, and Films, vol. 25, no. 4, pp. 971–975, 2007. [6] C. Griffiths and H. Eastwood, “Influence of stoichiometry on the metalsemiconductor transition in vanadium dioxide,” J. Appl. Phys., vol. 45, no. 5, pp. 2201–2206, 1974. [7] C. Ko and S. Ramanathan, “Observation of electric field-assisted phase transition in thin film vanadium oxide in a metal-oxide-semiconductor device geometry,” Appl. Phys. Lett., vol. 93, no. 25, pp. 252101–252101, 2008. [8] B.-J. Kim, Y. W. Lee, B.-G. Chae, S. J. Yun, S.-Y. Oh, H.-T. Kim, and Y.-S. Lim, “Temperature dependence of the first-order metal-insulator transition in vo2 and programmable critical temperature sensor,” Appl. Phys. Lett., vol. 90, no. 2, pp. 023515–023515, 2007. [9] H.-T. Kim, B.-G. Chae, D.-H. Youn, G. Kim, K.-Y. Kang, S.-J. Lee, K. Kim, and Y.-S. Lim, “Raman study of electric-field-induced firstorder metal-insulator transition in vo2-based devices,” Appl. Phys. Lett., vol. 86, no. 24, pp. 242101–242101, 2005. [10] J. B. Goodenough, “The two components of the crystallographic transition in vo2,” J. Solid State Chem., vol. 3, no. 4, pp. 490–500, 1971. [11] M. Rini, A. Cavalleri, R. W. Schoenlein, R. L´opez, L. C. Feldman, R. F. Haglund Jr, L. A. Boatner, T. E. Haynes, et al., “Photoinduced phase transition in vo2 nanocrystals: ultrafast control of surface-plasmon resonance,” Opt. Lett., vol. 30, no. 5, pp. 558–560, 2005. [12] A. Cavalleri, C. T´oth, C. W. Siders, J. A. Squier, F. R´aksi, P. Forget, and J. C. Kieffer, “Femtosecond structural dynamics in vo2 during an ultrafast solid-solid phase transition,” Phys. Rev. Lett., vol. 87, p. 237401, Nov 2001. [13] J. Duchene, M. Terraillon, P. Pailly, and G. Adam, “Filamentary conduction in vo2 coplanar thin-film devices,” Appl. Phys. Lett., vol. 19, no. 4, pp. 115–117, 1971. [14] K. Okimura, N. Ezreena, Y. Sasakawa, and J. Sakai, “Electric-fieldinduced multistep resistance switching in planar vo2/c-al2o3 structure,” Jpn. J. Appl. Phys., vol. 48, no. 6, p. 5003, 2009. [15] A. Mansingh and R. Singh, “The mechanism of electrical threshold switching in vo2 crystals,” J. Phys. C, vol. 13, no. 31, p. 5725, 2000. [16] G. Stefanovich, A. Pergament, and D. Stefanovich, “Electrical switching and mott transition in vo2,” J. Phys.: Condens. Matter, vol. 12, no. 41, p. 8837, 2000.
7
Tyler S. Jordan (S’11) will receive his B.S. degree in electrical engineering from The University of New Mexico, USA, in 2014. He has conducted microelectronics research at Sandia National Laboratories and developed RF communication software and hardware at The Department of Defense.
Steve Wolfley received his B.S. degree from Brigham Young University in ChemEng. in 1986 and his Masters (also BYU) in 1991. He is employed at Sandia National Laboratories and has worked in different aspects of thin films for the past 27 years.
Sean Scott (S’07-M’11) received the B.S. and Ph.D. degrees in electrical and computer engineering in 2007 and 2011, from Purdue University, West Lafayette, IN. He was previously a postdoc with Sandia National Laboratories, Albuquerque, NM. He is currently a visiting assistant professor with Purdue.
Darin Leonhardt received the B.S. and Ph.D. degrees in Chemical Engineering from the University of New Mexico in 2005 and 2011. In 2011, he joined Sandia, where he works on the materials science, device physics, and modeling of semiconductor/detector systems.
Christopher D. Nordquist (M’96-SM’09) received B.S., M.S., and Ph. D. degrees in Electrical Engineering from the Pennsylvania State University and is currently in the RF/Optoelectronics Department at Sandia National Laboratories, where he is exploring integrated RF devices.
Joyce Olsen Custer has a B.S. degree in chemistry from the University of California, Berkeley in 1984. She has worked at Lawrence Berkeley Labs, Philips Semiconductor, and Sandia National Labs in the Coatings and Surface Engineering group.
Christopher T. Rodenbeck (S’97-M’04-SM’09) received the B.S. (summa cum laude), M.S., and Ph.D. degrees in electrical engineering from Texas A&M University in 1999, 2001, and 2004, respectively. He currently leads multidisciplinary advanced/exploratory technology development for microwave and sensor applications at Sandia National Laboratories.
