Effective oxide thickness

Transient effects in naturally aging samples occur with some delay and are very slow. Nazar and Ahmad274 have observed a slow decrease of A1-A1203-A1 capacitance that was attributed to neutralization of Al3+ cations in the vicinity of the internal boundary and a corresponding increase of the effective oxide thickness. However, the same effect may be due to neutralization of negatively... 

Quantifying the effect of surface roughness or morphology is difficult, however. Surface preparations that provide different degrees of surface roughness also usually produce surfaces that have different oxide thicknesses and mechanical properties, different compositions, or different contaminant levels. The problem of separation of these variables was circumvented in a recent study [52] by using a modified microtome as a micro milling machine to produce repeatable, well-characterized micron-sized patterns on clad 2024-T3 aluminum adherends. Fig. 2 shows the sawtooth profile created by this process. 

In studies on Pt dotted silicon electrodes, PMC measurements revealed that tiny Pt dots increased the interfacial charge transfer compared with bare silicon surfaces in contact with aqueous electrolytes. However, during an aging effect, the thickness of the oxide layer between the silicon and the platinum dots gradually increased so that the kinetic advantage again decreased with time.11... 

Figure 12,4.1 The multiple reflection of light from microscopic oxide layers of different dimensions leads to constructive and destructive interference of light waves, producing a particular color effect. Different thicknesses reflect different colors.

Thinning of the anodic oxide coverage is found at sharp 90° edges of the substrate. This effect has been ascribed to oxide stress, because similar results are found for low-temperature thermal oxidation under conditions where viscous flow is not present. For oxide thicknesses in excess of about 100 nm, cracks develop in... 

The equation is then solved for the oxide thickness z under the assumption that no down area polishing occurs until the local step, Zj, has been removed, after which the local rate is equal to the blanket polish rate. This is captured by expressing the effective density as... 

The calibration phase focuses on the determination of the planarization length itself. This is a crucial characterization phase since once the planarization length is determined, the effective density, and thus the thickness evolution, can be determined for any layout of interest polished under similar process conditions. The determination of planarization length is an iterative process. First, an initial approximate length is chosen. This is used to determine the effective density as detailed in the previous subsection. The calculated effective density is then used in the model to compute predicted oxide thicknesses, which are then compared to measured thickness data. A sum of square error minimization scheme is used to determine when an acceptably small error is achieved by gradient descent on the choice of planarization length. 

We see from the thickness nonuniformity that if the TiN thickness is fixed, the oxide thickness has a lower standard deviation, indicating a tighter distribution of the measurement results. Since the measurement is on the same wafer, the difference suggests that the effect of fixed TiN thickness to help improve the repeatability of the measurement. (Of course, the TiN film must be uniform for this to be true.) In short, if the control of the TiN... 

Van Deventer and co-workers (23) recently investigated the thickness of vanadium oxide on as-received metal and heated specimens. Oxide thicknesses of up to 10 pm were observed, depending on the treatment. Hydrogen pretreatment seemed to increase the hydrogen permeability through the specimens. Similar effects of hydrogen pretreatment frequently have been seen. 

More recently, Gould and Irene (96) studied surface-cleaning effects on the oxidation of silicon wafers for oxide thicknesses up to 4300 A. They found... 

Conclusion. Existing empirical models such as the Deal-Grove model have been used for many years as a means of estimating oxide thickness as a function of extracted rate constants. The effects of processing on oxidation are included in these rate constants, but generally only by means of measurement. No fundamental basis exists yet for the direct calculation of these... 

Barrier-layer capacitors are based on the limited reoxidation of a reduced composition. This results, in the simplest case, in a surface layer of high resistivity and a central portion of conductive material so that the effective dielectric thickness is twice the thickness hQ of a single reoxidized layer and there is an apparent gain in permittivity over that of a fully oxidized unit by a factor of h/2ha, where h is the overall dielectric thickness (Fig. 5.51). Alternatively each... 

Suppose that some ionic species i is strongly rate-limiting and that the rapid flow of an electronic species produces a constant voltage VB across the oxide, independent of oxide thickness L. Furthermore, assume the electric field in the oxide to be due primarily to surface charge, such that space-charge effects are negligible. Then the electric field, E(x), in the oxide can be written as a constant E0... 

Another process of physical protection is the formation of an oxide layer that makes the metal passive. This procedure is used for aluminium. Aluminium is normally anodized in 10 per cent sulphuric acid with steel or copper cathodes until an oxide thickness of 10-100 pm is obtained. As the more superficial part of the oxide layer has a fairly open structure it is possible to deposit metals (cobalt, nickel, etc.) or organic pigments in the pores and seal with boiling water or with an alkaline solution. The colours after metallic deposition are due to interference effects. Chromic and oxalic acids are also used significantly as electrolyte. 

Mica particles coated with a metal oxide film have three layers with different refractive indices (layer 1 and 3 are identical, layer 2 is mica) and four interfaces (see chapter 16.2). Interference of light is generated by reflections of all six possible combinations of the four interfaces. Some of them lead to equal effects. The thickness of the mica platelets varies according to a statistical distribution. As a con-... 

The effect of pH on patterned TEOS oxide wafers was also investigated using the MIT CMP Characterization Mask Set pitch mask [1]. The pitch mask is a 6 x 6 array of various equal width lines and spaces. The lines and spaces have widths ranging from 2 pm to 1000 pm. The 12 mm pitch pattern was first created in photoresist across the entire wafer. The expiosed oxide was plasma etched to a depth of about 7000 A below the surface. This created an array of 36 sub-arrays with equal width lines and spaces so that half the area was lines and the other half spaces. The wafers were pwlished with silica slurry at 3 different pH values. The oxide thickness at the plateau area (starting surface of the oxide) for 10 sub-patterns of line widths 40,60,80,100,125, 150, 180, 200,250, and 500 pm were measured before and after CMP. The measurements were made in each sub-array in increasing order of line/space width. Five dies were measured along the diameter of the wafer. 

Unlike W plasma etch back process, the typical W CMP process usually removes the adhesion layer such as Ti/TiN or TiN during the primary polish. As a result, during the over polish step there is some oxide loss. Since the oxide deposition, planarization CMP (oxide CMP), and tungsten CMP steps are subsequent to each other, the oxide thickness profile could become worse further into the process flow. Therefore, the across-wafer non-uniformity of the oxide loss during W CMP process is one of the very important process parameters needs to be optimized. To determine the effect of the process and hardware parameters on the polish rate and the across-wafer uniformity, designed experiments were run and trends were determined using analysis of variance techniques. Table speed, wafer carrier speed, down force, back pressure, blocked hole pattern, and carrier types were examined for their effects on polish rate and across-wafer uniformity. The variable ranges encompassed by the experiments used in this study are summarized in Table I. 

The modeling and characterization of pattern dependent variations in oxide CMP further demonstrate the importance of these effects in the design of a viable CMP process. Given the magnitude of the oxide thickness variation present after planarization in comparison to the variation across the wafer as shown in Fig. 9, it is clear that process optimization should be driven by the die-level variation as much as if not more so than by the wafer-level variation. 

---