Search

Motivation: Plasma etching of low-k and ultra low-k (ULK) dielectric materials have seen a tremendous growth in deep nanoscale applications. Fluorocarbon based gases are the forerunners in etching low-k dielectrics as they are the F suppliers necessary to remove Si and C from SiCOH material. Different FC gases in combination with the additive gases contribute differently to the etch behavior of SiCOH due to differing reaction chemistry and formation of Fluorocarbon layer as an interfacial layer between plasma front and dielectric front [i]. Due to dynamic formation and etching of FC Layer happening at the dielectric surface, understanding FC layer properties in each gas plasma is the starting step to control etch rate and plasma induced damage in Low-k dielectrics. This paper aims to study and compare FC layer deposition by different gases grouped into CF4 and CF4/C4F6 family with N2, O2 and H2 as additive gases in each group. FC layer thickness, etch rate, optical properties of etched SiCOH and chemical composition of the FC layer by different plasma treatments have been presented in this work. Methodology: Dense SiCOH (k=2.75 and Open Porosity: 7%) with initial thickness of 157±3nm is deposited on 300mm Si Wafer with a thin SiO2 adhesion layer in between. These wafers are blanket etched in a commercial etch chamber with the given gas combinations (CF4, CF4/N2, CF4/O2, CF4/H2; CF4/C4F6, CF4/C4F6/N2, CF4/C4F6/O2, CF4/C4F6/H2) for same amount of time, total gas flow rate, similar power, pressure and temperature settings. Post Etch blanket wafers are subjected to various in-line and off-line metrology tools. In-line spectroscopic ellipsometer is used to model complex refractive index and thickness of individual layers. Tauc-Lorentz model best describes the FC layer whereas Cauchy model describes the dielectric [ii]. In-line XPS was carried out to probe the surface of FC layer and gain chemical information. Furthermore depth profile of etched layer is carried out by sputtering the surface with low power Ar ions to study variations in chemical composition with the layer depth. SEM is used to see the cross section profile and verify the results from above method Observation: Thickness evaluation from the dispersion models shows that the CF4 group has higher etch rate and lower FC layer thickness compared to CF4/C4F6 group. C4F6 has higher polymerization ability by the virtue of its lower F/C ratio. Amongst each group, Oxygen as additive gas shows the highest etch rate and least FC thickness followed by Nitrogen whereas Hydrogen has least etch rate but highest FC layer thickness. Under given etch conditions with CF4/C4F6/H2 plasma, deposition rate was higher than etching rate leading to stack growth (Fig.1a). Hydrogen supports polymerization by scavenging Fluorine away and aids in thicker FC films. XPS chemical composition shows Fluorine diffusion through FC layer into the dielectric layer bringing changes to its pristine structure. This affects Carbon% across the etched dielectric thickness (Fig.1b). Thicker the FC layer is, closer is the Carbon % to pristine SiCOH. Presence of surficial Nitrogen in the nitrogen containing plasmas shows nitrogen reacts with SiCOH Carbon, forms CN compounds and increases the etch rate. FC layer shows presence of CF3 radicals on the surface and decrease fast with the depth, followed by CF2 radicals. CF radicals are in larger quantity and are present deepest into the FC Layer. Presence of polymerized C-C and C-CFx bonds increases with the FC Layer depth and offset with increase of C-Si bond near the SiCOH interface. Conclusion: This paper shows comparison of Etch Rate and FC layer thickness during SiCOH etch using different gas combinations in the CCP plasma chamber. Thickness correlation between ellipsometric model data and XPS depth profile followed by SEM cross section measurement of layer thickness point towards the correctness of dispersion models for each layer. These models can be used for quick inline post etch metrology useful for process control. O2 containing plasmas are too aggressive on the dielectric with negligible FC layer formation, while H2 containing plasmas are less damaging but exhibit very small etch rate. CF4/C4F6/N2 plasma is the optimum combination for Low-k dielectric etch amongst other combinations in terms of desired etch rate and lower reduction in Carbon %. Relative concentration and penetration depth of different CFx radicals in FC film suggests its influence on film structure. References: Baklanov et.al, Journal of Applied Physics 113, 041101 (2013); doi: 10.1063/1.4765297 T. Easwarakhanthan et al., Journal of Applied Physics 101, 073102 (2007); doi: 10.1063/1.2719271 Figure 1

At different moisture levels the leakage current and the disruptive strength of SiCOH, a nonporous low-κ dielectric, has been investigated. If water is in the dielectric film, the current density increased of about 6 orders of magnitude due to the ionization of water molecules in the applied electric field. Furthermore, at a constant electric field the delay time until the highest current value is reached depends on the water concentration and the silanol amount in the dielectric film. The delay time can vary between 1 minute in an untreated sample and 39 minutes in a thinner and plasma treated sample. In addition, the disruptive strength in the dry state and in the water saturated state at 80% relative humidity remains constant at about 10 MV/cm. © 2014 The Electrochemical Society. [DOI: 10.1149/2.0161501jss] All rights reserved.

A porous ultra low-κ dielectric (pULK) and a dense SiCOH dielectric were investigated before and after a plasma treatment with argon in terms of the change in the bonding types, the relative permittivity and the water uptake. Fourier transform infrared (FTIR) spectroscopy revealed a change in the bonding types of the dielectrics in general and a significant increase in the hydroxyl band especially. The high hydroxyl amount leads to an increase in the relative permittivity of these dielectrics by up to 6.25% for SiCOH and up to 12.5% for the pULK material. Furthermore, if water diffuses into the dielectric films from the environment, the moisture uptake is up to 2.7 times higher in saturation at 80% relative humidity in comparison to the untreated samples. Due to the plasma damaged upper layer of the materials, the diffusion process of water into the bulk dielectrics is significant reduced. Overall, it has been found that the pULK material is more vulnerable to the used plasma treatment in comparison to the dense SiCOH film. © 2014 The Electrochemical Society. [DOI: 10.1149/2.0041501jss] All rights reserved.

Atomic Layer Deposition (ALD) is a particular thin film deposition technology which is based on alternating saturated surface reactions. As a result, the film growth proceeds in a self-limiting manner enabling the deposition of thin films with excellent thickness control, uniformity, and conformity. Although a large number of materials have been deposited by ALD so far for various applications, there are still some challenges in ALD. The deposition temperatures in ALD are typically lower compared to CVD due to the limited thermal stability of ALD precursors. As a consequence of the lower energy available for film formation the films may not meet the properties needed for application. In these cases a post deposition annealing is required to improve the film properties, e.g. to obtain the desired film structure, density, or purity. However, this high temperature processing is often impracticable due to a restricted thermal budget of the substrate, in particular when coating temperature sensitive substrates. Secondly, the reactants of an ALD process, e.g. oxygen, may react with the substrate itself leading to the formation of a parasitic interfacial layer. In order to avoid this issue, the proper choice of reactants or the use of an alternative deposition technique is essential. Furthermore, many ALD processes suffer from substrate inhibited film growth accompanied by inefficient precursor consumption and the formation of films with unfavorable properties. Finally, there are materials of interest, e.g. titanium, which so far can not be deposited by thermal ALD at all. These limitations may be overcome by the application of flash lamp annealing (FLA) in ALD. In FLA the substrate is exposed to a light flash with durations typically in the millisecond range. The light energy is absorbed within the top layers of the sample causing a rapid heating of the surface near region. On the contrary, the bulk material experiences no or only moderate heating. Consequently, FLA is a suitable technology to power high temperature processes even on temperature sensitive substrates. The film growth in flash lamp enhanced ALD is induced by this effect. Thereby, each process cycle consist of both a precursor pulse and the irradiation of the substrate with a light flash. During each single flash the surface temperature exceeds the threshold temperature which is required to achieve the thermal decomposition of adsorbed precursor molecules or to activate chemical reactions between the adsorbed precursor molecules and a second reactant. The film growth proceeds step-by-step and thus the film thickness can be controlled by varying the number of cycles. In addition, FLA in each cycle results in the periodical annealing of the already grown film and hence may lead to an improved film quality. Consequently, flash lamp enhanced ALD has a high potential for the realization of single-source processes, for the reduction of growth delay in the initial phase of film growth, for the deposition of high purity thin films, and for the deposition of new materials. In this work the principle of flash lamp enhanced ALD will be presented in detail, the technology will be reviewed and classified. Thereafter, we will give an overview about our studies on the flash enhanced ALD of aluminum-, ruthenium-, and tantalum-based thin films. These depositions were realized by flashing periodically on a substrate during the precursor pulses. We will show that the film growth is induced by the flash heating and the processes exhibits typical ALD characteristics. The obtained relations between flash parameters and film growth parameters will be discussed with the use of simulation results illustrating the temperature profile during the FLA treatment. Moreover, this work addresses the potentials of this technology as well as the technical challenges.

Erratum: Investigation of Argon Plasma Damage on Ultra Low-κ Dielectrics [ECS J. Solid State Sci. Technol., 4, N3023 (2015)] C. Kubasch,a T. Olawumi,a H. Ruelke,b U. Mayer,b and J. W. Barthaa aTechnische Universitat Dresden, Fakultat Elektrotechnik und Informationstechnik, Institut fur Halbleiterund Mikrosystemtechnik, 01062 Dresden, Germany bGLOBALFOUNDRIES Dresden Module One, Limited Liability Company & Co. KG, 01109 Dresden, Germany

Ta―N-based thin films were deposited by thermal atomic layer deposition. In this work, we introduced a tantalum source. The alternate supply of this halide but liquid precursor tantalum pentachloride, diethyl sulfide (TPDS), and ammonia (NH 3 ) resulted in Ta-N-based films with a saturated growth rate of approximately 0.2-0.3 A/cycle at 300-400°C and less than 1 atom % chlorine. By pulsing trimethylaluminium (TMA) as an additional reacting agent between the TPDS and NH 3 , the resistivity was improved up to 10 3 μΩ cm. These films showed a chlorine content of 10 atom % and an aluminum content of less than 1 atom %. X-ray photoelectron spectroscopy, X-ray diffraction, and a standard four-point probe method indicated a shift from tantalum nitride to tantalum-carbonitride-based films with increasing TMA and decreasing NH 3 pulse numbers.

The interaction between the additive components chloride, accelerator, and suppressor is the focus of the present paper. Based on Frumkin and Damaskin’s adsorption theory, an advanced concept for the superfilling mechanism is introduced. Cyclovoltammetry measurements show the additive impact on the current–potential behavior and their synergetic effect on the charge transfer across the electrode–electrolyte interface. The measurements are supported by partial fill experiments. Cross-section scanning electron microscopy micrographs reveal that the synergetic impact of an accelerator and a suppressor is very different from their individual contributions. Chloride ions support the adsorption of suppressor molecules, while the accelerator enhances the desorption of the suppressor molecules from the copper surface. The higher the local accelerator concentration in an electrolyte, the more suppressor molecules desorb from the surface. The synergetic behavior between the chloride, suppressor, and accelerator can be explained by coadsorption, which is an important key in the process of copper superfilling. For the application of interconnect structures, manifold investigations have been carried out regarding the role of plating bath additives in the electrochemical copper deposition process. To achieve void-free copper interconnects, superfilling is crucial. Superfilling with an increased copper growth rate at the bottom structure results from the impact of plating bath additives. At least two additives, called suppressor and accelerator, are required for the superfilling process. The active component of the suppressor often consists of polyethylene glycol, a long chained organic molecule with a molar mass of up to 10,000 g/mol. Due to its functional end groups, it is soluble in water and acids, while the ether groups possess free electron pairs that interact with copper ions. 1,2 Small amounts of chloride ions are necessary for the suppressor to function as a surfactant. 3,4 The active component of the accelerator bis3-sulfopropyldisulfide or mercaptopropylsulfonic acid adsorbs on Au and Ag surfaces by forming thiolate bonds. Controversies exist about the surface activity of the accelerator on copper especially in acidic solutions. 5 Spectroscopic measurements did not identify a specific coordination of the accelerator to a copper surface. 6