Metal Lines Research Articles

Electromigration Study of Plasma Etched Copper Lines with Copper Oxide Capping Layers

Copper (Cu) is a widely used interconnection material in integrated circuits (ICs) and large area thin film transistor (TFT) arrays. Chemical mechanical polishing (CMP) is the most popular production method in etching Cu fine lines [1,2]. Cu cannot be etched with a conventional plasma etching process at the room temperature because of the low volatility of copper halides [3,4]. Kuo’s group invented a plasma-based Cu etch method that includes two steps: the conversion of Cu into CuClx or CuBrx with a Cl- or Br-containing plasma and the subsequent removal of the halide by dipping in a dilute HCl solution [5-7]. The complete process, which can be carried out at room temperature, has been used to fabricate BiCMOS ICs and TFT LCDs [8,9]. The reliability of the Cu line can be evaluated using the electromigration (EM) method [10]. EM tests on Cu lines with different capping layers (e.g., TiW, Mo), which could prevent Cu oxidation, have been reported [11-13]. In this study, the influence of the copper oxide (CuOx) capping layer on the EM reliability is studied.The Mo/Cu stack was sputter deposited on a Corning glass substrate and etched into a 4-point pattern by the plasma-based etching method in a parallel-plate RIE reactor (PlasmaTherm 700C) at room temperature. The Cu chlorination reaction was conducted in CF4/HCl 5/20 sccm at 70 mTorr and 600 W for 2 minutes. Then, the CuClx layer was dissolved in a H2O:HCl (8:1 v/v) solution for 1 min. The Mo barrier layer was etched in CF4/O2 10/10 sccm at 60 mTorr and 600 W for 2 minutes. In order to form the CuOx capping layer, the Cu line was exposed to the O2 plasma at 200 mTorr and 100W for 3 minutes in the same PlasmaTherma reactor but under the plasma etching mode. For the comparison purpose, a control sample, i.e., the Mo/Cu line without the CuOx capping layer, was prepared and EM stressed. All EM tests were done at room temperature under the fixed current densities (J) condition.Figure 1 shows the resistance-time (R-t) and temperature-time (T-t) curves of Mo/Cu/CuOx lines stressed at different current densities. The resistance increases slightly with the stress time and jumps sharply at the line broken point. Also, the line broken time shortens with the increase of the current density. The line broken process starts from voids formation at grain boundaries followed by voids merging [11-13]. The T-t curves were calculated by the JEDEC standard guidelines of JESD33-B [14] modified with an iteration method. The change of the T-t curve follows that of the R-t curve since the temperature increase is caused by Joule heating. The extent of heat dissipation is dependent on the heat dissipation rate of the CuOx capping layer of which the composition changes during the EM stress.Figure 2 shows top views of a Mo/Cu/CuOx line at different stress times under J = 0.41×106 A/cm2. The original line has the aquamarine color. The color becomes lighter when stressed. Eventually, the broken region became dark. The line color change is a good indication of the change of the composition of the passivation layer, which is temperature dependent.The detailed EM failure mechanism with the CuOx composition change will be presented. Advantages of the CuOx passivation layer will also be discussed. Authors acknowledge the financial support of this work through NSF CMMI project 1633580. 1. A. Gladkikh and Y. Lereah, APL, 66(10), 1214-1215 (1995).2. C. S. Hau-Riege, et al., JAP, 96(10), 5792-5796 (2004).3. Y. Kuo and J.R. Crowe, JVST A, 8(3), 1529-1532 (1990).4. H. Miyazaki, et al., JVST B, 15(2), 237-240 (1997).5. Y. Kuo and S. Lee, JJAP, 39(3A), L188 (2000).6. S. Lee and Y. Kuo, JES, 148(9), G524-G529 (2001).7. Y. Kuo and S. Lee, APL, 78(7), 1002-1004 (2001).8. Y. Kuo and S. Lee, Vacuum, 74(3-4), 473-477 (2004).9. J. Yang, et al., ECS Trans., 16(9), 13 (2008).10. C. S. Hau-Riege, Microelectronics Reliability, 44(2), 195-205 (2004).11. M. Li, J. Q. Su and Y. Kuo, ECS Trans., 89(3), 87-92 (2019).12. J. Q. Su, M. Li and Y. Kuo, ECS Trans., 90(1), 65-72 (2019).13. J. Q. Su, M. Li, Y. Kuo and S. Hamaguchi, ECS Trans., 92(5), 39-46 (2019).14. EIA/JESD33-B, Standard Method of Measuring and Using the Temperature Coefficient of Resistance to Determine the Temperature of a Metallization Line (2004). Figure 1

(Invited) Reliability and Scaling Perspectives of HfO2-Based OxRAM

Hafnium oxide OxRAM (Oxide Random Access Memory) is a non-volatile memory technology that has been widely investigated over the last 15 years. Its working mechanism relies on the formation and rupture of a tiny conductive filament through a HfO2 layer sandwiched between an active (Ti) and inert (TiN) electrodes. This causes the electrical resistance to very over three orders of magnitudes between a low resistance state (LRS) and a high resistance state (HRS).OxRAM is a strong candidate to replace embedded Flash technology at scaled nodes. This is justified by several considerations but mainly driven by the low cost of integration and very high compatibility with standard CMOS technology. Yet OxRAM faces tremendous challenges in terms of reliability, variability and scaling.In this paper we first show our latest findings in terms of reliability on a Ti/HfO2(5nm)/TiN cell. Controlling LRS and HRS resistance distributions can be achieved with a proper write/erase algorithm design (see Fig. 1) and it is a key point to enable large array integration. A proper write/erase sequence ensures long bit endurance up to 107 cycles. Data retention is then evaluated at four different temperatures (125, 165, 200 and 260 °C) to extract a 10 years retention temperature of 85°C.Next, we discuss OxRAM scaling perspectives. In this context, it is essential to take into account both the resistive cell and its selector (most typically a MOS transistor). As recently shown [2], so far the OxRAM scaling has been mainly limited by the programming current, which determines the MOS selector size. Even if the OxRAM cell is integrated in the BEOL above its selector, the bitcell size is mostly determined by the underlying transistor. In addition, the OxRAM programming voltage dictates the transistor EOT and challenges the transistor reliability for scaled nodes. We demonstrated our OxRAM cell integration with a thin gate oxide FDSOI transistor (GO1/SG) and we show how this paves the OxRAM integration in a 28nm node.Eventually we discuss the ultimate scaling perspective of an OxRAM array in a Cross-Point architecture. To enable such architecture the development of a BEOL selector is mandatory and we focused on Ovonic Threshold Switching device (OTS) for this scope. The OTS works switching between a low-conductive state and a metastable conductive one, which is voltage-triggered. We demonstrated (see Fig. 3) a 1S1R (one selector one resistor) OxRAM cell where the HfO2 memory is co-integrated with a series Ge-Sb-Se-N (GSSN) OTS. We eventually discuss the limits of this architecture showing that a trade-off exists between the half pitch of the Cross-Point array metal lines and the size of the memory bank.REFERENCE[1] G. Sassine et al., 2018 IEEE International Reliability Physics Symposium (IRPS). [2] J. Sandrini et al., 2019 IEEE Electronic Device Meeting (IEDM). [3] D. Alfaro Robayo et al., 2019 IEEE 11th International Memory Workshop (IMW). Figure 1

Study of Dual-Frequency Polarizer for Electron Cyclotron Resonance Heating Systems of 105 and 140 GHz

In electron cyclotron resonance heating (ECRH) systems, the desired polarization is achieved by using two polarizers, a linear polarizer and an elliptical polarizer. These two polarizers serve as an important part of a 63.5-mm corrugated metallic waveguide transmission line for the ECRH. The dual-frequency polarizers (DPs) for the 105- and 140-GHz systems were designed in the present article by using the optimization of the phase difference method. The linear polarizer of 105-GHz shared an optimized polarizer with an elliptical polarizer of 140 GHz. The elliptical polarizer of 105-GHz shared another optimized polarizer with a linear polarizer of 140 GHz. The polarization characteristics of the single-frequency and DPs at the frequencies of 105 and 140 GHz were analyzed and compared. The polarization purity of the combination of the two polarizers was also calculated. The results indicated that arbitrary polarization state could be successfully achieved with the use of the designed DPs.

On Border Traps in Back-Side-Illuminated CMOS Image Sensor Oxides

CMOS image sensors (CISs) in back-side-illuminated configuration consist of photodiode arrays having metal lines and drive electronics beneath the active region with respect to the device/air interface so that the light reaches the photodiode active region directly. This enhances sensor quantum efficiency but reduces the electrical performance and reliability. The back-side configuration is realized by flipping the wafer upside down, bonding it to a handling wafer, mechanically thinning it, and opening a through-silicon via with a long plasma etch. As a result, gate oxides in back-side CISs show an increased density of donor-like border traps with respect to the conventional front-side-illuminated sensors. In this article, we try to add some information toward the comprehension of the origin and the electrical nature of those traps in back-side gate oxides. To this aim, we performed negative bias temperature instability stress on p-channel MOSFETs during which the traps were filled by holes tunneling from the substrate and then studied the relaxation transients of the drain current after the stress was removed. The characteristic emission times of a few specific levels were obtained at different temperatures. This allowed us to extract values of the trap activation energies, which resulted coherent with hole-capturing ${E}\,^{\prime }$ donor-like centers (trivalent silicon dangling bonds) commonly attributed to ionizing radiation.

CAN UVB VARIATIONS RECONCILE SIMULATED QUASAR ABSORPTION LINES AT HIGH REDSHIFT?

In this work we present new calculations of the observables associated with synthetic metal and HI absorption lines in the spectra of high redshift quasars, in-spired by questions and limitations raised in work with a uniform Haardt-Madau 2012 ultraviolet background (UVB). We introduce variations at z ≈ 6 to the UVB and HI self–shielding and explore the sensitivity of the absorption features to modifications of the hardness of the UVB. We find that observed SiIV and low ionization states (e.g. CII, SiII, OI) are well represented by a soft UV ionizing field at z = 6, but this same prescription, fails to reproduce the statistical properties of the observed CIV ion absorber population. We conclude that small variations in the UVB (not greater than a dex below Haardt-Madau 2012 emissivity at 1 Ryd) and HI SSh at z ≈ 6 play a major role in improving the estimation of metal ions and HI statistics at high z.

Absorption-line Abundances in the SMC-like Galaxy UGC 5282: Evidence of ISM Dilution from Inflows on Kiloparsec Scales*

Abstract We present a Hubble Space Telescope Cosmic Origins Spectrograph spectrum of the QSO SDSS J095109.12+330745.8 ( ) whose sightline passes through the SMC-like dwarf galaxy UGC 5282 ( , cz = 1577 km s−1), 1.2 kpc in projection from the central H ii region of the galaxy. Damped Lyα (DLA) absorption is detected at the redshift of UGC 5282 with log [N(H i) cm . Analysis of the accompanying S ii, P ii, and O i metal lines yields a neutral gas metallicity, , of [S/H] ≃ [P/H] = −0.80 ± 0.24. The metallicity of ionized gas from the central H ii region measured from its emission lines is [O/H] = −0.37 ± 0.10, a difference of +0.43 ± 0.26 from . This difference δ is consistent with that seen toward H ii regions in other star-forming galaxies and supports the idea that ionized gas near star-forming regions shows systematically higher metallicities than exist in the rest of a galaxy’s neutral interstellar medium (ISM). The positive values of δ found in UGC 5282 (and the other star-forming galaxies) is likely due to infalling low-metallicity gas from the intergalactic medium that mixes with the galaxy’s ISM on kiloparsec scales. This model is also consistent with broad Lyα emission detected at the bottom of the DLA absorption, offset by ∼125 km s−1 from the absorption velocity. Models of galaxy evolution that attempt to replicate population characteristics, such as the mass–metallicity relation, may need to start with a galaxy metallicity represented by rather than that measured traditionally from .

Switchless Multiplexing of Graphene Active Sensor Arrays for Brain Mapping.

Sensor arrays used to detect electrophysiological signals from the brain are paramount in neuroscience. However, the number of sensors that can be interfaced with macroscopic data acquisition systems currently limits their bandwidth. This bottleneck originates in the fact that, typically, sensors are addressed individually, requiring a connection for each of them. Herein, we present the concept of frequency-division multiplexing (FDM) of neural signals by graphene sensors. We demonstrate the high performance of graphene transistors as mixers to perform amplitude modulation (AM) of neural signals in situ, which is used to transmit multiple signals through a shared metal line. This technology eliminates the need for switches, remarkably simplifying the technical complexity of state-of-the-art multiplexed neural probes. Besides, the scalability of FDM graphene neural probes has been thoroughly evaluated and their sensitivity demonstrated in vivo. Using this technology, we envision a new generation of high-count conformal neural probes for high bandwidth brain machine interfaces.

First-Principles Study of 3d Transition-Metal-Atom Adsorption onto Graphene Embedded with the Extended Line Defect.

A type of line defect (LD) composed of alternate squares and octagons (4–8) as the basic unit is currently an experimentally available topological defect in the graphene lattice, which brings some interesting modifications to the magnetic and electronic properties of graphene. The transitional-metal (TM) atoms adsorb on graphene with a line defect (4–8), and they show interesting and attractive structural, magnetic, and electronic properties. For different TMs such as Fe, Co, Mn, Ni, and V, the complex systems show different magnetic and electronic properties. The TM atoms can spontaneously adsorb at quadrangular sites, forming a metallic atomic chain along LD on graphene. The most stable configuration is the hollow site of a regular tangle. The TMs (TM = Co, Fe, Mn, Ni, V) tend to form extended metal lines, showing a ferromagnetic (FM) ground state. For the Co, Fe, and V atoms, the system is half-metal. The spin-α electron is insulating, while the spin-β electron is conductive. For the Mn and Ni atoms, Mn-LD and Ni-LD present a spin-polarized metal; for the Fe atom, Fe-LD shows a semimetal with Dirac cones. For Fe and V atoms, both Fe-LD and V-LD show spin-polarized half-metallic properties. And its spin-α electron is conducting, while the spin-β electron is insulating. Different TMs adsorbing on a graphene nanoribbon forming the same stable configurations of metal lines show different electronic properties. The adsorption of TMs induces magnetism and spin polarization. These metal lines have potential applications in spintronic devices and work as a quasi-one-dimensional metallic wire, which may form building blocks for atomic-scale electrons with well-controlled contacts at the atomic level.

Physical conditions in two high-redshift H2-bearing GRB-DLAs, 120815A and 121024A

ABSTRACT The gamma-ray burst (GRB) afterglows provide a unique opportunity to study the interstellar medium (ISM) of star-forming galaxies at high-redshift. The GRB-DLAs (damped Lyman-α absorbers) contain a large neutral hydrogen column density, N(H i), and are observed against the GRB afterglow. A large fraction of GRB-DLAs show presence of molecular hydrogen (H2) which is an indicator of star-formation. Hence it is important to study those GRB-DLAs which have H2 lines to decipher and understand their physical conditions. The GRB-DLAs 121024A and 120815A, situated at redshift 2.30 and 2.36, respectively, are two such important H2-bearing GRB-DLAs. Besides H2, these two GRB-DLAs also show many metal lines. In this work we have carried out a detail numerical study on the H2 lines, as well as on those metal lines, in GRB-DLAs 121024A and 120815A self-consistently. We use the spectral synthesis code cloudy for this study. This modelling helps us to determine the underlying physical conditions which give rise to such lines and hence to understand these two GRB-DLAs in much more detail than any other previous investigation. We find that the hydrogen densities for these two H2-bearing DLAs are ≥60 cm−3. Moreover our study infers that the linear sizes are ≤17.7 pc for these two GRB-DLAs, and the mean gas temperatures averaged over the cloud thickness, are ≤140 K. Overall, we find that these two H2-bearing GRB-DLAs are denser, cooler, and smaller compared to those without H2.

The New Structure and Analytical Model of a High-Voltage Interconnection Shielding Structure With High-k Dielectric Pillar

The high-voltage interconnection (HVI) effect induces electric field crowding near the source region of power lateral double diffused metal oxide semiconductor field effect transistors (LDMOSs) and, therefore, weakens its breakdown voltage (BV) characteristic. To suppress such a detrimental effect, a novel HVI structure with a high- ${k}$ dielectric pillar beneath the HVI metal line is proposed. An analytical model is proposed to reveal the mechanism of the proposed structure and explain the effects of the parameters of the structure. The analytical model indicates that the high- ${k}$ dielectric pillar eliminates the electric field crowding near the source, promotes the depletion of the drift region, and avoids the premature breakdown. Meanwhile, due to the existence of the high- ${k}$ dielectric pillar, the surface electric field profile in the drift region can be modulated, and the influence of HVI can be effectively suppressed. Thus, a higher BV is obtained. The simulation results indicate that the proposed structure presents better BV, which is the same as that of the reduced surface field (RESURF) structure without HVI.

РОЛЬ ЖЕЛЕЗО СОДЕРЖАЩЕГО КОМПОНЕНТА В СОСТАВЕ ПЕРЛИТОВОГО ГЕОПОЛИМЕРА НА ЕГО ФАЗОВЫЕ И СТРУКТУРНЫЕ ТРАНСФОРМАЦИИ

Normally, the component composition of building materials makes a significant contribution when formation of final performance properties of them. Moreover, new types of binding systems, in spite of their poor knowledge, this effect on the structural characteristics of final composites also takes place. In this work, perlite-based geopolymer binder as an example, was studied.
 In framework of the study, it was found that the application of grinding mill with a metal lining leads to the formation of Fe- milling yield, which is included into crushed raw materials (perlite) com-position.
 It was established that the particles of the resulting Fe- component are nanoscaled.
 The X-ray phase and chemical analyzes were used to determine the mechanism of the effect of the Fe- component on the properties of perlite-based geopolymer. 
 Microstructure of the hardened geopolymer paste were studied using SEM microscopy. Phase and mineral composition of geopolymer paste obtained by XRD-analysis showed, when the hardening process, nanosized iron hydroxide or bernalite Fe(OH)3(H2O)0.25 is formed in the geopolymer system, that is 5 times more than metallic iron. This is initiates the structure destruction and confirmed by the low values compressive strength of the geopolymer stone.
 Keywords: Fe-bearing component, metallic milling yield, geopolymer binder, friable structure, the Scherrer formula

Hybrid nanoelectronic-magnetic device with magnetoresistive core–shell Fe/FeC nanoparticles

We propose a concept of hybrid nanoelectronic-magnetic device made of magnetic Fe–C core–shell nanoparticles deposited onto prepatterned Si (111) substrate with basic circuitry made of metallic conductive lines. The synthesis of magnetic material and the creation of nanoelectronic prepatterned interdigitated die are reported and to prove the effectiveness in devices, their magnetotransport properties are investigated. Magnetic Fe/FeC nanoparticles, 11 nm diameter, with a core–shell structure have been prepared by laser pyrolysis. Two different layouts of prepatterned interdigitated die, have been conceived using e-beam lithography, with various geometries. A range of microscopy techniques, transmission electron, scanning and optical, were employed for morphological characterization of the as-obtained structures. Magnetic and magnetotransport characterization using SQUID magnetometry has been performed onto both the core–shell nanoparticles and onto the hybrid device obtained by depositing centrifugated and dispersed core–shell nanoparticles from liquid carrier solutions. From magnetotransport measurements, it has been revealed that the hybrid device made of Fe/FeC nanosized materials on prepatterned interdigitated die exhibit a large giant magnetoresistive (GMR) effect of about 8% at 300 K. This result is promising in view of the use of such devices as arrays of nanosensors and in spintronic applications.

SN 2016gsd: an unusually luminous and linear Type II supernova with high velocities

ABSTRACT We present observations of the unusually luminous Type II supernova (SN) 2016gsd. With a peak absolute magnitude of V = −19.95 ± 0.08, this object is one of the brightest Type II SNe, and lies in the gap of magnitudes between the majority of Type II SNe and the superluminous SNe. Its light curve shows little evidence of the expected drop from the optically thick phase to the radioactively powered tail. The velocities derived from the absorption in H α are also unusually high with the blue edge tracing the fastest moving gas initially at 20 000 km s−1, and then declining approximately linearly to 15 000 km s−1 over ∼100 d. The dwarf host galaxy of the SN indicates a low-metallicity progenitor which may also contribute to the weakness of the metal lines in its spectra. We examine SN 2016gsd with reference to similarly luminous, linear Type II SNe such as SNe 1979C and 1998S, and discuss the interpretation of its observational characteristics. We compare the observations with a model produced by the jekyll code and find that a massive star with a depleted and inflated hydrogen envelope struggles to reproduce the high luminosity and extreme linearity of SN 2016gsd. Instead, we suggest that the influence of interaction between the SN ejecta and circumstellar material can explain the majority of the observed properties of the SN. The high velocities and strong H α absorption present throughout the evolution of the SN may imply a circumstellar medium configured in an asymmetric geometry.

Searching for z > 6.5 Analogs Near the Peak of Cosmic Star Formation

Abstract Strong [O iii]λλ4959,5007 + Hβ emission appears to be typical in star-forming galaxies at z > 6.5. As likely contributors to cosmic reionization, these galaxies and the physical conditions within them are of great interest. At z > 6.5, where Lyα is greatly attenuated by the intergalactic medium, rest-UV metal emission lines provide an alternative measure of redshift and constraints on the physical properties of star-forming regions and massive stars. We present the first statistical sample of rest-UV line measurements in z ∼ 2 galaxies selected as analogs of those in the reionization era based on [O iii]λλ4959,5007 equivalent width (EW) or rest-frame U − B color. Our sample is drawn from the 3D-HST Survey and spans the redshift range . We find that the median Lyα and C iii]λλ1907,1909 EWs of our sample are significantly greater than those of z ∼ 2 UV-continuum-selected star-forming galaxies. Measurements from both individual and composite spectra indicate a monotonic, positive correlation between C iii] and [O iii], while a lack of trend is observed between Lyα and [O iii] at EW[O iii] ≲ 1000 Å. At higher EW[O iii], extreme Lyα emission starts to emerge. Using stacked spectra, we find that Lyα and C iii] are significantly enhanced in galaxies with lower metallicity. Two objects in our sample appear comparable to z > 6.5 galaxies with exceptionally strong rest-UV metal line emission. These objects have significant C iv λλ1548,1550, He ii λ1640, and O iii]λλ1661,1665 emission in addition to intense Lyα or C iii]. Detailed characterization of these lower-redshift analogs provides unique insights into the physical conditions in z > 6.5 star-forming regions, motivating future observations of reionization-era analogs at lower redshifts.

Effect of Contact Plug Deposition Conditions on Junction Leakage and Contact Resistance in Multilevel CMOS Logic Interconnection Device.

Here, we developed the optimal conditions in terms of physical and electrical characteristics of the barrier and tungsten (W) deposition process of a contact module, which is the segment connecting the device and the multi-layer metallization (MLM) metal line in the development of 100 nm-class logic devices. To confirm its applicability to the logic contact of barrier and W films, a contact hole was formed, first to check the bottom coverage and the filling status of each film, then to check the electrical resistance and leakage characteristics to analyze the optimal conditions. At an aspect ratio of 3.89:1, ionized metal plasma (IMP) Ti had a bottom coverage of 40.9% and chemical vapor deposition (CVD) titanium nitride (TiN) of 76.2%, confirming that it was possible to apply the process to 100 nm logic contacts. W filling was confirmed, and a salicide etching rate (using Radio Frequency (RF) etch) of 13–18 Å/s at a 3.53:1 aspect ratio was applied. The etching rate on the thermal oxide plate was 9 Å/s. As the RF etch amount increased from 50–100 Å, the P active resistance increased by 0.5–1 Ω. The resistance also increased as the amount of IMP Ti deposition increased to 300 Å. A measurement of the borderless contact junction leakage current indicated that the current in the P + N well increased by more than an order of magnitude when IMP Ti 250 Å or more was deposited. The contact resistance value was 0.5 Ω. An AC bias improved the IMP Ti deposition rate by 10% in bottom coverage, but there was no significant difference in contact resistance. In the case of applying IMP TiN, the overall contact resistance decreased to 2 Ω compared to CVD TiN, but the distribution characteristics were poor. The best results were obtained under the conditions of RF etch 50 Å, IMP Ti 200 Å, and CVD TiN 2 × 50 Å.

Flexible thermoelectric generators for body heat harvesting – Enhanced device performance using high thermal conductivity elastomer encapsulation on liquid metal interconnects

This paper reports flexible thermoelectric generators (TEGs) employing eutectic gallium indium (EGaIn) liquid metal interconnects encased in a novel, high thermal conductivity (HTC) elastomer. These TEGs are part of a broader effort to harvest thermal energy from the body and convert it into electrical energy to power wearable electronics. The flexible TEGs reported in this paper employ the same thermoelectric legs' used in rigid TEGs, thus eliminating the need to develop new materials specifically for flexible TEGs that often sacrifice the so-called figure of merit' for flexibility. Flexible TEGs reported here embed rigid thermoelectric legs' in soft and flexible packaging, using stretchable EGaIn interconnects. The use of liquid metal interconnects provides ultimate stretchability and low electrical resistance between the thermoelectric legs. The liquid metal lines are encased in a new stretchable silicone elastomer doped with both graphene nano-platelets and EGaIn to increase its thermal conductivity. This high thermal conductivity elastomer not only reduces the parasitic thermal resistance of the encapsulation layer but it also serves as a heat spreader, leading to 1.7X improvement in the output power density of TEGs compared to devices fabricated with a conventional elastomer. The device performance is further improved by a thin Cu layer acting as a heat spreader providing an additional 1.3X enhancement in the output power at 1.2 m/s air velocity (typical walking speed). Worn on the wrist, our best devices achieve power levels in excess of 30 μW/cm2 at an air velocity of 1.2 m/s outperforming previously reported flexible TEGs.

Overlapping Limitations for ps-Pulsed LIFT Printing of High Viscosity Metallic Pastes

Laser-induced forward transfer (LIFT) technique has been used for printing a high viscosity (250 Pa·s) commercial silver paste with micron-size particles (1–4 µm). Volumetric pixels (voxels) transferred using single ps laser pulses are overlapped in order to obtain continuous metallic lines. However, interference problems between successive voxels is a major issue that must be solved before obtaining lines with good morphologies. The effects of the laser pulse energy, thickness of the donor paste film, and distance between successive voxels on the morphology of single voxels and lines are discussed. Due to the high viscosity of the paste, the void in the donor film after a printing event remains, and it negatively affects the physical transfer mechanism of the next laser pulses. When two laser pulses are fired at a short distance, there is no transfer at all. Only when the pulses are separated by a distance long enough to avoid interference but short enough to allow overlapping (≈100 µm), is it possible to print continuous lines in a single step. Finally, the knowledge obtained has allowed the printing of silver lines at high speeds (up to 60 m/s).

A novel airgap formation scheme by GO nanosheet gap sealing process with extreme low effective dielectric constant

In order to reduce the parasitic capacitance between the interconnect metal line, low-k material was developed and integrated into CMOS BEOL process, and airgap is the ultimate low-k material with the lowest dielectric constant close to 1. In this work, a novel airgap formation scheme was proposed to form the airgap by low-cost coating process, in which Graphene Oxide (GO) nanosheet was used as gap-filling suppressor to seal the gap between two adjacent metal line and form the airgap. Traditional non-conformal CVD gap sealing process was firstly developed and compared with the proposed scheme. It was found that the non-conformal CVD sealing process cannot achieve high performance airgap, and there will always be some material deposited into the gap between the metal interconnect line even for very narrow gap. For the proposed scheme, volume ratio of GO sealed airgap to gap region before sealed can reach to about 85% with much wider gap, and calculated effective relative dielectric constant of this region is about 1.57. When increasing the gap CD to value larger than GO nanosheet size, conformal film deposition can be got, which indicate that GO nanosheet with the size larger than gap CD can seal the gap in the very beginning and prevent further material depositing into the gap to make the large ratio airgap formation.

Characteristics of Metal-Metal Nano-Area Contact Resistance for Less Than 5 nm Technology Node.

Recently, better understanding of nano-area is required for 5 nm or less technology node. In particular, the high contact resistance generated in a nano-area significantly degrades the device performance. In this study, we propose a direct contact resistance measurement method without a test structure by separate processes to improve the nano-area contact resistance. The nano-area contact resistance of Ti-Ti and Cu-Cu decreased from 6.46 MΩ to 1.08 MΩ and from 3.78 MΩ to 1.48 MΩ, respectively, when the metal line and native layer formed on the surface were removed. In addition, it is confirmed that the contact resistance decreased with an increase in bonding strength in the case of nano-area homo-metal contact. However, the contact resistance is affected by the tunneling effect and bond energy according to the distance between the first layers of atoms in the case of nano-area hetero-metal contact.

Learning Localized Spatial Material Properties of Substrates in Ultra-Thin Packages Using Markov Chain Monte Carlo and Finite Element Analysis

Thinned silicon dies and thin substrates using thin core and coreless structures have enabled thin packages. For robust manufacturing and reliability of these parts, solving the warpage problem is key. While current finite element methodologies can provide some insights at the design stage, these simulations are only as accurate as the inputs such as the material properties and the stress-free temperatures. Electronic substrates are especially challenging to characterize and model as they are laminates consisting of a core with layers of resin and metal lines on either side. In this work, a hybrid approach using Markov Chain Monte Carlo (MCMC) and Finite Element Analysis (FEA) is used to learn the spatially varying properties of the substrate from Digital Image Correlation (DIC) measurements of the warpage. The analysis is carried out at room temperature and at an elevated temperature point. Image analysis on electrical artwork is also carried out to correlate the material properties to the substrate metal density. These results will be useful to package and substrate designers to understand how material properties vary over the substrate and how temperature and metal density affect material properties so that robust design for future packages to minimize warpage can be initiated by careful routing of metal lines depending on the locally desired properties of the stack.