Search

Advanced semiconductor manufacturing technology has adopted an alkaline solution for copper (Cu) corrosion prevention instead of the traditional acidic solution in the post-cleaning process of copper chemical mechanical polishing. Low particle and residue removal efficiency has been an issue for this process. In this study, we investigated the formation of small residue defects and the cleaning mechanism to remove these defects. The results show it is insufficient to remove the small residue defects by using the traditional process parameter tuning. Adding some friction between the pad and the wafer surface allowed the defects to be effectively reduced.

This study proposes that the corrosion resistance of copper film correlates well with underlying barrier's orientation. To test the hypothesis, we performed X-ray diffraction, conducted copper removal rate experiments after chemical mechanical polishing, and tested static potentiodynamic polarization. The results all show that copper deposited on strongly (002) oriented β-Ta barrier layer demonstrated better chemical resistance against surface reaction with the slurry for strong copper (111) orientation. The findings were consistent with the result of the chronoamperometric test at 0.3 V in which the more passive film formed on the composite film with (002) β-Ta underlying barrier.

This study investigates spontaneous microstructural evolution in electroplated Cu films with various plating current densities involving direct current and pulse-reverse waveforms and various possible driving forces. Studies have explained the grain growth and resistivity decrease during the self-annealing of as-deposited Cu film, but the incubation behavior of self-annealing under various direct current and pulse-reverse current waveforms at a certain film thickness is unknown. In this study, it was found that pulse-reverse current retards the incubation behavior more significantly than does direct current. According to the measurements of resistivity, stress, and secondary ion mass spectrometer, the large stress difference between the initial and critical values and the low impurity content of pulse-reverse current postponed the incubation, and led to a slow self-annealing rate. The combination of the stress difference and the impurity effect explains the incubation behavior of self-annealing under various plating current densities. The resistivity and X-ray diffraction results suggest that stress is the primary driving force that dramatically speeds up grain growth above the critical stress, and that high current density with a rapid grain growth rate enhances the (200) texture for strain energy minimization in electroplated Cu film.

One of the major challenges in the integrated circuit beyond 90 nm is to fabricate low resistivity Cu diffusion barrier layer in the metal multilevel interconnect. Ta/TaN bilayer is one of the best candidates for Cu metal diffusion layer. It provides the advantage good diffusion performance between Cu and low-k dielectric layer. However, the resistivity is large deviation from tantalum phase variation. This is because of Tantalum of bilayer easily found in beta-phase, which is high resistance. This paper proposed a low resistivity alpha-Ta thin films were grown by treatment-TaN using Argon plasma treatment on TaN substrate. The argon treatment redistribute nitrogen profile between Ta and TaN and create a Ta(N) interface. X-ray diffraction analyses show that the interface film is composed of b.c.c.-Ta(N) grains. This Ta/treatment TaN provide new method to fabricated b.c.c.-Ta(N) difference from the previously report deposition by sputtering process. The resistivity will decrease nitrogen concentration to provide a favorable environment to promote low resistivity Tantalum alpha-phase formation.

Currently an additional in situ argon resputtering process after thin film deposition is regularly applied to improve step coverage of micro-trenches in devices. However, this process produces unexpected defects when applied to Ta/TaN bi-layers used in metal barriers for copper damascene processes. The balance between deposition and resputtering needs to be optimized; otherwise, this process can easily result in large thickness variation at trench corner and generate over-etching defects at the corner of trench bottom after the barrier formation. In this paper, the mechanism of defect formation is demonstrated, and we discuss the conditions in which these defects appear. A simple model to eliminate the side effect is proposed. The model examined the thickness variation at different locations as a function of TaN amount removed during the resputtering process. Besides, we have defined a factor, resputtering ratio, to optimize the deposition and resputtering processes. Through the model, the optimal resputtering ratio is determined to be 75%, which not only keeps low via-resistance, but also eliminates the over-etching defects.

This paper discusses the first evidence about the low pattern density causing abnormally high VIA resistance, which is at least more than two times the normal level. The low pattern density also induces the exposure-like shape formed in VIA bottom, where high concentration of sulfur element is detected in the copper film after electric- and thermal-torture testing. These effects, due to the low pattern density, are significantly important for nano-scale feature filling and the quality of VIA superfilling. This paper also proposes a mechanism to explain the observed results by inferring from the local resistance and chemical polarity difference with different numbers of VIA serially connected. The experiment with high accelerator concentration also supports the proposed mechanism by locating abnormal voids in the region with low VIA pattern density.

There are many kinds of commercial slurries used in Cu CMP. Major components include an oxidizing agent, complexing agents, inhibitors, and abrasives. We analyze the abrasive particle size by TEM and light scattering. Cu CMP polishing mechanism is also discussed under different particle size distribution. The complexing agent transportation will be the rate determining step when a small abrasive is insufficient, but the copper hydroxide removal rate will determine the overall polishing rate when the amount of smaller particles is enough.

As the dimensions of devices are shrunk quickly, the requirements of metallization become more critical. For VIA barrier and seeding layer filling and deposition, the process was mostly applied with the copper physical vapor deposition methodology in the back-end of line flow of the interconnection metallization. The criteria for barrier and seeding layer deposition are the metal continuity inside the VIA feature and grain size and orientation control for film diffusion barrier and qualities. Besides, while the interconnection size shrunk to nano-scale, the barrier thickness would be very thinner to maintain the VIA resistance; however, it would face the film conformity and continuity consistence within the wafer and different features. The integration solution would be developed and studied with the re-sputter process step adding into the convectional physical vapor deposition process. The resputter process step could not only improve the film conformity and continuity in the VIA's sidewall; but also reduce the resistance of VIA feature over 20%. The improvement of the resputter method adding into the deposition process would be contributed to the standard barrier deposition in the nano-scale feature of the interconnect. Besides, we also discussed the effect of the film properties after the resputter process introduced into the barrier deposition.

The Ta/TaN bilayer exhibits the best performance in the Cu metal multilevel interconnects, because it provides good coherence between Cu and dielectric layer. In the Ta/TaN bilayer, Ta has two phases: alpha-phase of body center cubic is preferred due to its lower resistivity (15–60 μΩ-cm), whereas beta-phase of tetragonal should be avoided due to high resistive (∼150–250 μΩ-cm). However, beta Ta most commonly forms on fcc TaN. Here we provide a simple scheme to bypass this high resistive phase by resputtering TaN prior to Ta deposition. We found that, with surface treatment by argon ion bombardment for enough time, alpha Ta phase can be directly formed, which is supported both by X-ray diffraction and resistivity measurement. Depth profiles of all elements from Auger electron spectroscopy reveals that the surface treatment induces a nitrogen deficient surface layer due to different sputtering yield, which causes phase changes from fcc TaN to hcp Ta2N followed by bcc Ta(N) and provide a favorable lattice constant for Ta alpha-phase formation.

The influence of Cu-ion concentration in a concentrated H 3 PO 4 electrolyte on the Cu removal rate and planarization efficiency (PE) of electrochemical mechanical planarization (ECMP) was investigated. With increasing Cu-ion concentration in the electrolyte, results show that the Cu removal rate significantly decreased because more Cu ions in the electrolyte facilitated the formation of a surface passive film. Electrochemical analysis shows that the current density decreased from 91 to 62 mA/cm 2 when the Cu-ion concentration was increased from 35.8 to 158.6 mM. In a low Cu-ion concentration electrolyte, Cu(OH) 2 is dominant, while in a high Cu-ion concentration electrolyte, CuO predominates. A high Cu-ion concentration in the electrolyte during ECMP promotes the formation of CuO, which retards Cu-ion diffusion from the Cu surface to the electrolyte solution, resulting in a decrease in the Cu removal rate. The intensity ratio of CuO/[Cu(OH) 2 + CuO] in the Cu 2p 3/2 X-ray photoelectron spectroscopy spectrum increased from 22.1 to 85.6% when the Cu-ion concentration was increased. The effect of Cu-ion concentration on the microscale PE of Cu ECMP is discussed.

An improved copper (Cu) process with a plating current strategy was used to stabilize the residual stress and suppress the void formation. The Cu grain size was inversely proportional to the amplitude of the plating current in the electroplated process, and it dominated the density of the grain boundary due to the aggregation of Cu grain. Therefore, the elimination of the grain boundaries happens in the first thermal cycle and generates a tensile residual stress. The electroplated process by the improved plating current strategy was presented to produce various grain sizes to enhance the interface packing density of the Cu film. This result leads to a higher density of the grain boundaries and is helpful to the grain growth only through the effect of Cu self-annealing. The residual stress maintains a stable value without the thermal treatment. This plating method strategy presents a reliable Cu interconnect process without considering the change in the residual stress.

The relationship between Cu dishing depth (DD) and Young's modulus of polishing pads during chemical mechanical polishing for various Cu linewidths is investigated. For the Cu width of 150 μm, the results show that increasing Young's modulus from 161 to 180 kgf/cm 2 slightly decreased Cu DD; there was a major linear decrease until 195 kgf/cm 2 , and Cu DD was a constant value up to 208 kgf/cm 2 . These three regions are characterized by the threshold, linear, and saturated zones. Scanning electron microscopy results show that the nonlinear behavior is in response to the changes in the ratio of pore areas within the polishing pads. We build a model and define an effective polishing length, L * C , to predict Cu DD. L * C is the rotation mode of L C , the static distance between two neighboring pores. The proposed model shows that the three regions are determined by competition between static chemical etching and variable mechanical wearing resulting from the pore area ratio in relation to Young's modulus. The model is useful for understanding the nonlinear behavior between Cu DD and Young's modulus of the pad.

In the semiconductor metallization process, the superior gap-fill capability of copper (Cu) electroplating is mainly due to external additives, such as bis-(3-sodiumsulfopropyl disulfide) (SPS), which is used as an accelerator. This study demonstrates that the byproducts of SPS induced Cu defects after a chemical-mechanical-polishing (CMP) process. In conventional cyclic-voltammetric-stripping analysis, the byproducts generated from organic additives are very difficult to quantify. In this study, the authors used mass-spectrum analysis to quantify SPS byproducts and found that the SPS byproduct, 1,3-propanedisulfonic acid, correlated with the formation of Cu defects because it influenced the properties of electroplated Cu films and the chemical corrosion rate, then induced defects after the CMP process.

Bis-(3-sodiumsulfopropyl disulfide) (SPS) is a well-known accelerator used to promote "bottom-up filling" in the semiconductor copper (Cu) electroplating process. However, the gap-filling capability and the defect performance of Cu electroplating are impacted as the SPS molecules are decomposed or oxidized. In this study, by-products generated from SPS decomposition and oxidation were identified and quantized using mass spectroscopy. It was found that 1,3-propanedisulfonic acid (PDS) was the most stable species among the SPS by-products. The PDS concentration increased with the increasing surface area of immersed Cu metals. The relationship between the consumption rate of SPS and the generation rate of PDS with different surface areas of Cu metals was particularly investigated. Furthermore, an equivalent circuit was developed and then examined using electrochemical impedance spectroscopy to characterize the properties of the aged bath with the decomposition of SPS.

Bis-(3-sodiumsulfopropyl disulfide) (SPS) is a well-known additive that promotes "bottom-up filling" in the copper (Cu) electroplating process. In the literature, SPS has been widely investigated and has proven to have an acceleration effect on Cu electroplating. However, our results show that low-concentration SPS (∼ 10 ppm) has a suppression effect on the Cu electroplating rate. The suppression behavior of low-concentration SPS is demonstrated using various electrochemical analyses, including the polarization curve, cyclic-voltammetry stripping, and electrochemistry-impedance spectroscopy (EIS). The equivalent circuit simulated from the EIS data shows that the mechanism of low-concentration SPS inhibiting the Cu electroplating rate is caused by an increase in the charge-transfer resistance. This suggests that when SPS concentration is too low, the formation of Cu(I)thiolate complexes is too low to accelerate the Cu electroplating rate while the added SPS still consumes energy from the system. The acceleration effect of SPS concentration on the Cu electroplating rate exists when SPS concentration is larger than ∼ 100 ppm.

The requirement for the superplanarization of interconnect nanotechnology beyond 100 nm poses an urgent need to study the complicated behavior of copper CMP. It is common practice in advanced copper chemical mechanical planarization (CMP) polishing to add the inhibitor into the slurry to keep the copper surface perfect and smooth and to protect the copper surface from corrosion. It is beneficial to have the polishing pressure cushion between abrasives and wafer under the different pattern features and is most important to overcome the planarization limitation. This study describes the behavior of non-Preston's phenomena under the passivated additives (inhibitors) and develops a model to explain the mechanism of the passivated-and-oxidated kinetics with non-Preston's polishing, which explains the mechanism of copper surface reactions during polishing. Furthermore, our model shows that the three regions are due to different relationships between removal rate and polishing pressure. Three regions are characterized as the threshold, linear, and saturated zones, which are governed by the chemical etching, the depth of abrasive particles indent into the copper oxide, and the oxidation rate, respectively. Most of all, the removal rate change can be simulated and predicted by the ratio between the inhibitor and oxidizer concentrations. Therefore, this study does not only contribute the understanding of the non-Preston's behavior but also provides the model under the assumption of the sequential stacked films of passivation and oxide films on the copper surface. The potentiodynamic methods are employed to test the assumption used in the mechanism and model.

Copper pits formed after copper chemical mechanical polishing (CMP) are the critical defects in the copper dual damascene process because the imperfect copper surface can cause the interconnection to be missed or broken. There are many parameters contributing a flat copper surface, such as the polishing down-force, slurry pH value, and galvanic effects between barrier and copper materials. One of the important factors influencing the pit formation is the friction force. The kinetics of the polishing friction can be studied to understand the mechanism of the pit-defect formation. A mechanism is proposed and can be explained by two characteristics of pit-defect formation during the polishing, i.e., the location of the edges between the metal line and wafer and the magnitude of the friction torque inducing the defect. Furthermore, three different polishing platforms are used to demonstrate this kinetic approach toward understanding the role of the threshold of the friction force in the pit-defect immunization. With this understanding, the pit formation can be reduced significantly.