Low-k dielectric thin films are finding increased use in integrated circuits for the faster signal speed that they permit. These materials, however, have higher leakage currents and shorter lifetimes than SiO₂–based dielectrics. With the continued push to lower values of k, these problems are becoming more acute.

In this paper, we present results of several complementary characterization techniques for determining key physical properties, such as trap densities and barrier heights, that influence leakage and time-dependent dielectric breakdown (TBBD) phenomena. Electrical characterization techniques include impedance spectroscopy and the measurements of transient currents. In addition, we make use of distinctive optical characterization techniques to obtain specific information about the underlying material properties. Internal photoemission (or photocurrent) spectroscopy yields information on interfacial barrier heights from the photon energies required to induce a current. Optical second-harmonic generation (SHG) provides a sensitive, non-contact method for measuring both photo-driven and spontaneous charge transport. From these methods, we show that the increase in current with long time biasing, which is in turn a precursor to electrical breakdown, can be directly correlated with increased trap densities. Conduction models accounting for the early failure mechanism will be discussed. Partial support for this work from the Semiconductor Research Corporation is gratefully acknowledged

The electronic properties of thin film low-κ interlayer dielectric (ILD) and etch stop layers (ESL) are important issues in ULSI development. However as the semiconductor industry looks to transition to 16 nm and beyond technology nodes, numerous concerns with low-κ materials need to be addressed. Leakage currents, time dependent dielectric breakdown and stress induced leakage currents are critical problems that are not yet well understood in ILD. A topic of current interest is ultraviolet light (UV curing) of low-κ materials.

We have made electron spin resonance (ESR) and current density versus voltage measurements on a moderately extensive set (over 50 films) of dielectric/silicon structures involving materials of importance to low-κ interconnect systems. Most of the dielectrics studied involve various compositions of SiOC:H. In addition we have also made measurements on other dielectrics including SiO₂, SiCN:H and SiN:H, some of which are utilized as ESLs. In our study we have made ESR and current density versus voltage measurements both before and after exposing the dielectrics to UV light (hc/λ ≤ 5 eV), and films that have experienced an industrial UV curing process. We observe extremely gross differences in the ESR spectra and leakage current versus voltage response of these low-k films. We find that UV exposure consistently increases both the density of paramagnetic defects and the leakage current density at a given field. Paramagnetic point defects observed in these films include, E’ centers, silicon dangling bond defects in which the silicon is back bonded to oxygen, the 74 gauss doublet which is E’ center complexed to a hydrogen atom, the 10.4 gauss doublet which is a hydrogen coupled E’ center, the K-center which are silicon vacancies back bonded to three nitrogens, and possibly silicon and carbon dangling bond centers and likely organic radicals. We have also made electrically detected magnetic resonance (EDMR) spin dependent trap assisted tunneling measurements on some ILD films. The close correspondence between the ESR and SDT result establishes a direct link between the defects observed in ESR and the defects responsible for the increased tunneling currents. We have also observed a correspondence between ESR amplitudes and leakage currents. Our preliminary results suggest the UV curing process creates paramagnetic centers which take part in trap assisted tunneling. Our results indicate quite clearly that the processing parameters have extremely gross effects upon defect densities within these films.

Defect concentrations in low-k organosilicate glass [SiCOH] films deposited on high-resistivity silicon were measured with electron-spin resonance. Both plasma exposure and ultraviolet exposure were used. During argon electron-cyclotron resonance plasma exposure, ion and photon bombardment increased the measured defect concentrations. Ultraviolet lamp exposure was also shown to increase the defect concentrations. SiCOH samples with several dielectric constants were examined showing that as the value of the dielectric constant was lowered, the defect concentrations were shown to increase significantly.[i] In addition, the nature of the defects in SiCOH was investigated using air and nitrogen plasma exposure. The defects were found to be silicon dangling bonds. Air-plasma exposure increases the defect concentrations by breaking silicon-hydrogen bonds, measured by Fourier-transform infrared spectroscopy. Nitrogen-plasma exposure as well as free-radical exposure have only a small influence on the bond breaking. It was also shown that UV curing improves the chemical-damage resistance of the dielectric.

With the continuous downscaling of devices, interconnects get narrower and narrower, and necessitate using porous low-k as insulator. Plasma processes required for the integration of low-k may cause damage at the sidewalls of the patterns, which degrades the dielectric properties of the material. The impact of the modified layer on the low-k sidewalls is becoming more critical when interconnects dimensions are scaled down. Developing low-damage plasma processes for porous low-k materials integration is compulsory, but requires a trustworthy characterization technique. Electron microscopy which leads to material shrinkage and does not precisely reveal the damaged layers is reaching its limitations. A new characterization technique, so-called Scatterometric Porosimetry has been recently proposed (Bouyssou et al. JVSTB, 2010). In this paper, we explain the principles of the method and demonstrate it can be used on complex industrial-relevant dielectric patterns (more than 9 dielectric layers, 140nm pitch). This technique can also be used to determine fundamental mechanisms of plasma induced modification to porous low-k dielectrics.

Scatterometric Porosimetry (SP) is a combination of scatterometry and porosimetric ellipsometry. Scatterometric measurements under vacuum give access to the pattern dimensions while measurements under controlled partial pressure of solvent give access to the material porosity or permeation (with low-polar solvent) or to the thickness of the hydrophilic damaged layer when water is used as a solvent.

Porous dielectric materials offer lower capacitances that reduce RC time delays in integrated circuits. Typical low-k materials include SiOCH – silicon dioxide with carbon groups, principally CH₃, lining the pores. Fluorocarbon plasmas are often used to etch low-k materials, a process that leaves a fluorocarbon polymer on the low-k surface that must be removed. With porosities as high as 0.5, pores which are internally connected provide pathways for reactive species to enter into the porous network. During cleaning using oxygen plasmas, reactions of O atoms with the CH_x groups, can remove carbon as CO/CO₂. After the process, H₂O from air can form hydrophilic Si-OH which can further adsorb H₂O through hydrogen bonding or physisorption.[1] As such, O₂ plasmas can degrade the low k value of the porous SiCOH. Plasma cleaning with He/H₂ mixtures causes less damage to SiCOH as reaction of H atoms with –CH₃ is endothermic. These damage scenarios are complicated by the UV/VUV photons produced by the plasma. Photons produced by the plasma can break Si-C bonds and separate –CH₃ radicals from SiO₂ to enhance the C removal rate.[2] 130 nm photons of Ar/O₂ plasmas can penetrate into SiCOH ≈100 nm but photons of He/H₂ plasmas (<100 nm) penetrate ≈20 nm. As a result, VUV photons from O₂ plasmas can produce Si-C bond scission approximately 5 times deeper in the low-k material compared to the VUV photons from He/H₂ mixtures. These penetration depths are sensitive functions of porosity and interconnectivity. For example, penetration depth increases nearly linearly with increases in interconnectivity due to the alignment of pores. In this talk, we discuss results from modeling of the plasma damage of porous SiOCH in He/H₂ and Ar/O₂ plasmas. The HPEM (Hybrid Plasma Equipment Module) was employed to obtain the ion energy and angle distributions of reactive fluxes from inductively coupled plasmas. These are used as input to the MCFPM (Monte Carlo Feature Profile Module) with which profiles of the low-k materials after the plasma exposure are predicted. The role of photons in porous SiCOH damage and validation of numerical results will be discussed in terms of treatment time, interconnectivity and photon flux. Overall, we found that due to its lower photon penetration depth and less reactivity, He/H₂ plasmas cause approximately 3 times less damage than Ar/O₂ plasmas, which is in agreement with experiments.

1. J. Proost, E. Kondoh, G. Vereecke, M. Heyns, and K. Maex, J. Vac. Sci. Technol. B 16, 2091 (1998).

2. B. Jinnai, S. Fukuda, H. Ohtake, and S. Samukawa, J. Appl. Phys. 107, 43302 (2010).

* Work supported by Semiconductor Research Corp.

Due to a low dielectric constant (4-4.5) and high density (1.8-2.0 g/cm³), Plasma Enhanced Chemically Vapor Deposited (PECVD) boron nitride (BN) is an intriguing material for use in low-k/Cu interconnect structures as a Cu diffusion barrier material. However, relatively little is known about the electrical leakage performance of BN in Cu interconnects or the Schottky barrier formed at the interface between these two materials. In this regard, x-ray photoelectron spectroscopy (XPS) was utilized to determine the Schottky barrier formed at the interface between polished Cu substrates and PECVD BN thin films. Our measurements indicate a barrier height of 3.0± 0.2 eV for the BN/Cu interface. This barrier height is nearly 2X that determined for a-SiCN:H and a-SiOC:H Cu capping layers and is attributed to the significantly larger band gap of BN (> 5 eV).