Depth-profiled positronium lifetime spectroscopy is used to probe the pore characteristics ͑size, distribution, and interconnectivity͒ in porous, low-dielectric silica films. The technique is sensitive to the entire void volume, both interconnected and isolated, even if the film is buried beneath a metal or oxide layer. Our extension of a simple quantum mechanical model of Ps annihilation in a pore adequately accounts for the temperature and pore size dependence of the Ps lifetime for pore sizes in the range from 0.1 nm to 600 nm. It is applicable to any porous media. ͓S0163-1829͑99͒51932-2͔ Submicron thin films of porous silica and organosilicates are vigorously being developed as low-dielectric, interlayer insulators for use in future high-speed microelectronic devices. 1 Voids are introduced into the film to produce porosity and hence to lower the dielectric constant. Pores must be plentiful to lower the dielectric constant of solid silica from 4 to less than 2, yet they must be small relative to the device element size which is expected to approach 100 nm in the next decade. Important pore characteristics such as average size, size distribution, and degree of interconnectedness are difficult to probe with standard techniques ͑such as gas absorption͒ because of the submicron film thickness, the presence of a thick Si substrate and, in some cases, by the lack of pore interconnectivity ͑i.e., inaccessibility to gas absorption͒. A less standard technique, positronium annihilation lifetime spectroscopy ͑PALS͒, is well known as a bulk probe of subnanometer voids in polymers and insulators and has recently been extended to probe very thin polymer films using keV beams of positrons. 2 The technique looks promising for probing porous films since it is readily applicable to films less than 0.1 m thick, does not rely on any pore interconnectivity/accessibility, and is expected to be sensitive to pore sizes in the 0.3 nm to 100 nm range.In this paper we will explore the capability of PALS to probe the pores in two different types of porous silica films that are spin-cast on Si substrates. The first is a 0.5-m-thick silica-organic composite in which the organic component is removed by thermal decomposition to create pores after the silica component is fully cured and crosslinked. The second is a 0.9-m-thick film, formed using a sol-gel ͑aerogel/ xerogel͒ technique. We determined the film porosities using Rutherford backscattering spectroscopy to be 52% and 77%, respectively. Details on the methodology of depth-profiled PALS has been presented elsewhere. 2 Briefly, a focussed beam of several keV positrons forms positronium ͑Ps, the electron-positron bound state͒ throughout the film thickness. The binding energy of Ps ͑6.8 eV in vacuum͒ is reduced in the solid dielectric and thus Ps tends to localize in the pores.The natural ͑vacuum͒ lifetime of Ps ͑142 ns͒ is reduced by annihilation with molecular electrons during collisions with the pore surface and thus pore size information can be deduced from measuring this lifetime, ͑...

An existing model that relates the annihilation lifetime of positronium trapped in subnanometer pores to the average size of the pores is extended to account for positronium in any size pore and at any temperature. This extension enables the use of positronium annihilation lifetime spectroscopy in characterizing nanoporous and mesoporous materials, in particular thin insulating films where the introduction of porosity is crucial to achieving a low dielectric constant, K. Detailed results of the model calculations are presented along with extensive experimental results to systematically check the lifetime vs pore size calibration in a variety of low-K materials over a wide range of pore sizes.

Positronium annihilation lifetime spectroscopy is used to determine the pore-size distribution in low-dielectric thin films of mesoporous methylsilsesquioxane. A physical model of positronium trapping and annihilating in isolated pores is presented. The systematic dependence of the deduced pore-size distribution on pore shape/dimensionality and sample temperature is predicted using a simple quantum mechanical calculation of positronium annihilation in a rectangular pore. A comparison with an electron microscope image is presented.

The technique of positron annihilation lifetime spectroscopy (PALS) has been used to investigate the continuity and thermal stability of thin barrier layers designed to prevent Cu atom diffusion into porous silica, low-dielectric constant (k) films. Nanoglass™ K2.2-A10C (A10C), a porous organosilicate film, is determined to have interconnected pores with an average tubular-pore diameter of (6.9 ± 0.4) nm. Cu deposited directly on the A10C films is observed to diffuse into the porous structure. The minimum necessary barrier thickness for stable continuity of Ta and TaN layers deposited on A10C is determined by detecting the signal of positronium (Ps) escaping into vacuum. It is found that the 25 nm thick layers do not form continuous barriers. This is confirmed by the presence of holes observed in such films using a transmission electron microscope. Although 35 nm and 45 nm Ta and TaN layers perform effectively at room temperature as Ps barriers, only the Ta-capped samples are able to withstand heat treatments up to 500 °C without breakdown or penetration into the porous film. TaN interdiffusion into the silica pores is indicated by the reduction of the Ps lifetime after high annealing temperatures. The validity of using Ps diffusion to test barrier layers designed to inhibit Cu diffusion is discussed. The procedures to standardize the testing of barrier layer integrity and thermal stability using PALS are proposed. Extension to probing barrier layers in realistic vias and trenches should be straightforward.

No abstract