Proc. SPIE. 5841, Fluctuations and Noise in Biological, Biophysical, and Biomedical Systems III
KEYWORDS: Temperature metrology, Liquids, Linear filtering, Liquid crystals, Platinum, Resistance, Bridges, Digital signal processing, Signal processing, Copper
In this paper we address the fundamental issue of temperature fluctuation during the thermal denaturation (or the unzipping of the two strands on heating) of double stranded (ds) DNA. From our experiments we observe the presence of extremely high thermal fluctuations during DNA denaturation. This thermal fluctuation is
several orders higher than the thermal fluctuation at temperatures away from the denaturation temperature range. This fluctuation is absent in single stranded (ss) DNA. The magnitude of fluctuation is much higher in heteropolymeric DNA and is almost absent in short homopolymeric DNA fragments. The temperature range over which the denaturation occurs (i.e., over which the thermal fluctuation is large) depends on the length of the DNA and is largest for the longest DNA.
Proc. SPIE. 5843, Fluctuations and Noise in Materials II
KEYWORDS: Resistance, Silver, Copper, Temperature metrology, Metals, Polymers, Interference (communication), Digital signal processing, Transformers, Nanowires
Nanowires with high aspect ratio can become unstable due to Rayleigh-Plateau instability. The instability sets in below a certain minimum diameter when the force due to surface tension exceeds the limit that can lead to plastic flow as determined by the yield stress of the material of the wire. This minimum diameter is given d<sub>m</sub> ≈ 2σ<sub>S</sub>/σ<sub>Y</sub> , where σ<sub>S</sub> is the surface tension and σ<sub>Y</sub> is the Yield force. For Ag and Cu we estimate that d<sub>m</sub> ≈ 15nm. The Rayleigh instability (a classical mechanism) is severely modified by electronic shell effect contributions. It has been predicted recently that quantum-size effects arising from the electron confinement within the cross section of the wire can become an important factor as the wire is scaled down to atomic dimensions, in fact the Rayleigh instability could be completely suppressed for certain values of k<sub>F</sub> r<sub>0</sub>. Even for the stable wires, there are pockets of temperature where the wires are unstable. Low-frequency resistance fluctuation (noise) measurement is a very sensitive probe of such instabilities, which often may not be seen through other measurements. We have studied the low-frequency resistance fluctuations in the temperature range 77K to 400K in Ag and Cu nanowires of average diameter ≈ 15nm to 200nm. We identify a threshold temperature T* for the nanowires, below which the power spectral density S<sub>V</sub>(f) ~1/f. As the temperature is raised beyond T* there is onset of a new contribution to the power spectra. We link this observation to onset of Rayleigh instability expected in such long nanowires. T* ~ 220K for the 15nm Ag wire and ~ 260K for the 15nm Cu wire. We compare the results with a simple estimation of the fluctuation based on Rayleigh instability and find good agreement.
We have studied the conductance fluctuation in metal film which is under electromigration stressing. The apparatus used by us allows measurement of noise with an ac 5-probe technique with a superimposed dc stressing current (typically 2MA/cm<sup>2</sup>). This allows measurement of noise in the film at different stages of the electromigration process till the film is damaged completely. We study both the spectral power S<sub>V</sub>(f) and also the probability density function (PDF) from the time series. The electromigration stressing was done to elevated temperature on Al and Cu metal lines grown by RF magnetron sputtering. Principal motivation of the investigation is to study low frequency defect relaxations in the metal film due to electromigration that give rise to conductance fluctuations with a spectral power S<sub>V</sub>(f)∝ 1/f<sup>α</sup>. SV(f) (both magnitude as well as the spectral power quantified through α) shows changes continuously and some times non-monotonically during the electromigration process and it is large just before the damage of the film. It was also observed that the PDF width increases significantly during the course of the em stressing and it changes from a Gaussian to a non-Gaussian PDF.
We have investigated the conductance fluctuation in a shape memory alloy (NiTi-Nitinol) near its martensite transition in order to study how the noise evolves near a phase transition that is driven by long range elastic interactions. The alloy has a martensite start and completion temperatures of 333 K and 325 K respectively and change of around 6%-8% in resistance at the transition. The conductance fluctuation (noise) measurements were carried out using a 5-probe ac technique in the temperature interval 200 K to 400 K in wire samples with typical thickness 5 μm and width of 120 μm. The samples were annealed at high temperature to remove any effect of previous temperature or strain cycle. We made the following observations: (1) There is a large change in the measured noise spectral power S(f) at the transition. (2) Away from transition temperature S(f) follows a 1/f-type behavior. This changes drastically near the transition, where large spectral density appears for f <1 Hz, and this strong dependence tends to saturate around 1 mHz. (3) Close to the transition, the probability density function (PDF) deviates from a Gaussian distribution, and shows a significant contribution from a power-law distribution. (4) When the transition is tuned by an applied constant stress (staying close to but below the transition temperature range), a similar strongly frequency-dependent low frequency S(f) appears, accompanied by a non-Gaussian PDF with a power-law statistics.
We have studied the conductance fluctuations in silver nanowires in the temperature range 4K to 375K. The nanowires with an average diameter of 15nm were electrochemically deposited using polycarbonate membrane as template. Principal motivation is to study low frequency defect relaxations in the nanowires that give rise to conductance fluctuations with a spectral power S(f) ∝ 1/f<sup>α</sup>. The Ag nanowires, stabilized at 400K with a current of few mA, show metallic temperature dependence. The S(f) was measured with a psuedo 4 probe ac technique with rms current of few tens of μA. We find that S<sub>V</sub>(f) (which is ∝1/f<sup>α</sup>) shows a rapid rise at around 220K as T is increased along with an enhancement in the exponent α. The exponent α≈1-1.1 for T<220 and it increases to ≈1.4 at T=375K. In the same temperature range S(f) rises by an order of magnitude. We analyze the data using a model assuming that there are two components to the 1/f<sup>α</sup> fluctuations--one arising from relaxation of local defects give α≈1. The other arises from the long-range diffusion of defects characterized by α≈3/2. It is seen that for T < 220K the noise arises mainly from local defect relaxation and the temperature dependence of a follows the Dutta-Horn model. Above this temperature the contribution from long-range diffusion dominates with the noise becoming thermally activated with an activation energy (~300meV). Interestingly the activation energy is similar to but somewhat higher than that seen in micron sized films.
In this work we review the investigations of conductance fluctuations in doped silicon at low temperatures (2K < T < 20K) as it is tuned through the metal-insulator transition by changing the carrier concentration n. Spectral power, S(f), of the conductance fluctuation retains a generic 1/f<sup>α</sup> dependence. In the metallic regime (n>n<sub>c</sub>) the doped Si is like a weakly-localized electron system and the conductance fluctuation is governed by the mechanism of Universal conductance fluctuations. The relative variance of fluctuation follows the temperature dependence ∝ T<sup>-β</sup>, where β≈1/2. However, the noise diverges by orders of magnitude as n decreases through the critical concentration n<sub>c</sub> and the fluctuation also becomes strongly temperature dependent with β>> 1. At the transition (n/n<sub>c</sub>≈1) the fluctuation becomes strongly non-Gaussian below 20K as observed through the second spectrum S<sup>(2)</sup>(f). At T=4.2K, we find that after subtracting the Gaussian background , S<sup>(2)</sup>(f)∝ 1/f<sup>p</sup> where p is small (< 0.5) for metallic samples (n/n<sub>c</sub>≥ 1.5) and it grows to > 1 for samples close to the transition n/n<sub>c</sub> ≈1. The growth of non-Gaussianity is accompanied by a growth in low frequency spectral weight as seen through a significant enhancement of α from close to 1 (n>n<sub>c</sub>) to nearly 1.4 for n/n<sub>c</sub> ≈1. The growth of non-Gaussian fluctuation of extremely large magnitude with significant low frequency component points to a correlated low frequency dynamics of charge fluctuation near the insulator-metal transition. This has been interpreted as the onset of a glassy freezing of the electronic system across the transition.
We have investigated the resistance fluctuations in Si:P as a function of doping level n, across the Metal-Insulator transition at low temperatures. The fluctuation size increases sharply with decrease in the doping level, and shows indications of correlations (presented elsewhere in this conference). The measured jumps in voltage in a current biased sample due to resistance fluctuations were stored digitally and the fluctuation size statistics were estimated in the form of a Probabilty Density Function (PDF). On the metallic side, the PDF's were found to have more or less a Gaussian shape, as expected from an ensemble of small uncorrelated fluctuators. However, we find marked deviation of the PDF from a Gaussian behavior as the system crosses into the insulating side. The deviation starts to occur at the tail of the distribution, and grows in size with decreasing doping levels. The deviating part of the tail could be fitted with a log-normal expression. On the insulating side, this growth of a log-normal tail is also seen to occur as the temperature is lowered. The observations have been analysed using existing theories.
We have investigated the dynamics of co-existing phases in the Charge Ordered (CO) manganite Pr<sub>0.63</sub>Ca<sub>0.37</sub>MnO<sub>3</sub> using the technique of conductance noise spectroscopy. We note that close to the CO transition temperature T<sub>co</sub> the spectral power of S<sub>v</sub>(f)/V<sup>2</sup> deviates significantly from the 1/f frequency dependence for f≤0.12Hz. Our analysis shows that this deviation can be described by a single frequency Lorentzian with corner frequency f<sub>c</sub> in addition to the usual broadband 1/f noise. Such a Lorentzian contribution to S<sub>v</sub>(f)/V<sup>2</sup> can come from a two level system (TLS). In the time serioues this shows up as RTN. For T≤T<sub>co</sub> the system shows the onset of a non-linear conduction close to a threshold value J<sub>dc</sub> = J<sub>th</sub> the noise spectra is mainly 1/f in nature. For J > J<sub>th</sub> a large low frequency component of noise (characterized again by a frequency f<sub>c</sub>) appears. We associate f<sub>c</sub> with the relaxation time tc of the TLS fluctuator so the tc = 1/f<sub>c</sub>. For thermal activation of the TLS the temperature dependence of f<sub>c</sub> will follow f<sub>c</sub>=f<sub>o</sub>exp(-E<sub>a</sub>/k<sub>B</sub>T) where E<sub>a</sub> is an energy barrier. The value of f<sub>c</sub> shows an increase with J<sub>dc</sub> showing that the value of the activation energy E<sub>a</sub> is being lowered by the applied bias.