1fNoise and Grain-Boundary Diffusion in Aluminum and Aluminum Alloys

Abstract

We have measured the 1/f noise in polycrystalline films of Al, Al-Si(1'/o), and Al-Cu(4'/o) in the temperature range of 300 to 600 K. The temperature dependence indicated activation energies of 0.69, 0.80, and 0.89 eV, respectively. These energies are similar to the activation energies found for Al diffusion along grain boundaries for films of the same size and composition measured in the same temperature range. Measurements of samples with identical compositions but differing widths and thicknesses revealed significant departures from the usual inverse volume dependence of the 1/f noise. PACS numbers: 72.70.+m, 73.60.Dt The origin of 1/f noise in metal films remains one of the longest-standing unanswered questions of solid-state physics. The details of the microscopic source of the noise remain elusive despite its almost universal presence in all metal films. An empirical expression for the 1/f-noise power, Sz, of a currentbiased metal film obtained by Hooge' in 1969 iswhere V is voltage across the sample, 0 is the sample volume, f is the frequency of measurement, and o. is an empirical parameter. Since the 1/f noise in metal films has been shown to be a fluctuation of the sample resistance, 2 the quantity Sz/V2 is independent of the current bias. Hooge originally postulated that the parameter o. was constant for a11 samples; however, many recent experiments have shown this parameter to be temperature and sample dependent. Detailed measurements of the frequency exponent of the noise have verified that the exact exponent is seldom exactly -1 (or 1/f) but is also a function of temperature and sample.Dutta and Horn proposed that the 1/f noise in a metal film results from thermally activated processes and demonstrated that the noise and frequency exponent when measured as a function of temperature can be used to measure the average activation energy and the width of the distribution of activation energies,The questions of what is activating and where still remain unanswered.We report in this paper the first systematic measurements of the temperature dependence of the 1/f'noise in polycrystalline thin films of Al, A1-Cu, and Al-Si alloy. Additions of Cu or Si to an Al film increase the activation energy for A1 atomic motion along grain boundaries6 in the films but not the activation energy for Al motion in the bulk of a grain. 7 We measured a corresponding increase in the activation energy of the

Low-Frequency Electronic Noise in Quasi-1D TaSe3 van der Waals Nanowires

“…Table I summarizes the electromigration activation energy estimated using the 1/f noise measurements for different interconnects. The values of the electromigration activation energy for the Cu and Al interconnects determined from the 1/f noise measurements are 0.76-1.10 eV and 0.67-1.14 eV, respectively [21,22,24,38,39]. Our data suggest that quasi-1D TaSe3 nanowires have similar EA values.…”

“…Hence, the noise can reveal information on the electromigration failure of interconnects and metal thin films. The noise activation energies extracted from the two commonly accepted physical models, the Dutta-Horn model [21,24] and the empirical noise model in metals [22,26,27], have shown an excellent agreement with the activation 3 | P a g e energies obtained from the industry standard electromigration mean-time-to-failure (MTF) tests. These considerations explain additional practical motivations for investigation of the lowfrequency noise in quasi-1D nanowires of TaSe3, which exhibit promise as ultimately downscaled local interconnects owing to their single-crystal atomic thread structure and exceptionally high breakdown current density [18].…”

“…Cu with TaNTa barrier 0.76-0.78 [24] 0.79 [38] Al 0.69 [21] Al -Si (1%) 0.80 [21] Al -Cu (4%) 0.89 [21] Al thin film 1.14 [22] Al -Si (0.8%) passivated 0.74 [39] Al -Si (1%) -Cu (0.5%) passivated 0.78 [39] …”

“…The measurements were performed in the temperature range from 300 K to 430 K in order to elucidate the physical mechanism of the low-frequency noise and determine the noise activation energies. The low-frequency noise is a ubiquitous phenomenon present in all kinds of electronic devices including field-effect transistors, diodes and interconnects [19][20][21]. The frequency dependence and temperature characteristics of the spectral density of the low-frequency noise have been used as reliability metrics for interconnects in Si complementary metal-oxide-semiconductor (CMOS) technology.…”

“…where A is a constant, j is the biasing current density and EA is the activation energy related to [21,23,24,37]. The noise measurements have specific advantages over the conventional reliability tests, which require time-dependent resistance measurements of interconnects under high temperature and high current stress for long periods of time (typically from 10 to 100 hours for each set).…”