(588d) Additives Applied to Cu Electrodeposition | AIChE

(588d) Additives Applied to Cu Electrodeposition

Authors 

Kim, J. J. - Presenter, Seoul National University
Copper (Cu) is widely used as a metallization material in semiconductor fabrication because of its properties like a low electrical resistivity. As the one of the methods for Cu deposition, electroplating has been extensively used for the high quality copper on a substrate. The mechanical properties of the electrodeposited Cu can be controlled with the modification of current/potential profiles such as pulse and pulse-reverse deposition, or the use of additives. Especially, additives can control the properties maintaining electroplating process, and also enable to modify the deposition rate topologically. Various combinations of additives are used to form the deposit with suitable properties and induce bottom-up filling in various scales of vias and interconnects. Conventionally the additives for Cu electrodeposition are classified into three categories based on the functionality. Accelerators increase the deposition rate of Cu where they adsorb. Instead, suppressors retard the Cu electrodeposition rate, physically blocking the access of cupric ions to the surface of the electrode. Levelers, generally slowing the deposition rate, similar to suppressors, have specific behavior dependent on the convective condition, and are introduced to fill micro-scaled feature and through-holes.

Through silicon via (TSV), the focused technologies for 3-dimensional (3D) packaging to solve the scaling limitation, is one of the fields using Cu electrodeposition. In the Cu deposition for fabrication of TSV interconnect, leveler acts as key materials, which is essential to maintain the inhibition at the top of TSV due to its large scaled structure and long filling time compared to damascene features. Therefore, a number of studies related to the levelers have been introduced to obtain bottom-up filling of micro-scaled features and reduce process time.

Our group also has been dedicated to study about additives applied to Cu electrodeposition, especially levelers.

New organic compounds have been developed as levelers for feature filling. This compound containing ammonium cation and halide ions strongly suppressed electrodeposition on the side and top of TSVs. These choline-based levelers induced selective deposition at the bottom of TSVs and established a flat growing surface, which can facilitate bottom-up filling of TSVs without any defects. The mechanism of gap filling is demonstrated based on the electrochemical analyses and filling profiles.

Furthermore, iodide ion which is the inorganic material has been introduced, which can function like organic leveler. It inhibited Cu deposition convection-dependently. The inhibition mechanism of iodide ion was identified based on chemical and electrochemical reaction in solution chemistry containing Cu ions, iodide ions, and other additives. This inorganic material formed non-conducting material, CuI, as inhibition layer on the Cu surface. Its convection-dependent behavior was suitable for bottom-up filling of TSVs.

CuI formed during electrodeposition with iodide ions could affect reliability of filling solution as byproduct. To solve this issue, another inorganic leveler was studied, bromide ion. The bromide ion cannot inhibit Cu electrodeposition as CuBr form due to its lower reduction potential. Although it showed the weaker inhibition strength than iodide ions, the synergistic effect with PEG-PPG enables to fill the TSV without any defects. Moreover, the filling rate doubled with this ion because of higher deposition efficiency than choline-based organic leveler or iodide ion. The mechanism of bromide ions was investigated by comparing to that of iodide ions.

In addition to effect of organic/inorganic additives on Cu electrodeposition, we have investigated the degradation mechanism of additives. As the Cu electrodeposition proceeds, the process accompanies the degradation of the plating solution, as like other chemical processes. Plating solution contains many kinds of components such as Cu metal salts and additives, and bath conditions can be changed with degradation of these materials. Consumption of the metal source is easily complemented by supplementing the additional metal salt. However, unlike Cu metal ions, additives are degraded with forming byproducts during electrodeposition. The formation of byproduct could adversely affect to reliability of the Cu electroplating process, causing defects such as voids and seam in interconnect, and unexpected change of Cu properties. Therefore, the degradation mechanism of each additive should be identified to understand and control the solution chemistry. The mechanism of bis(3-sulfopropyl) disulfide (SPS) and polyethylene glycol-polypropylene glycol copolymer (PEG-PPG) were investigated by electrochemical analysis and spectroscopic method.

SPS, widely used as an accelerator, is degraded easily in the electroplating process. SPS is decomposed into 3-mercapto-1-propanesulfonic acid (MPS) or intermediate oxide forms during the electrodeposition, and lastly converted into 1, 3-propane disulfonic acid (PDS). The formed MPS is reduced to SPS through interconversion reaction, and, on the other hand, PDS is irreversibly accumulated in the electrolyte.

The degradation reactions, which occur at both Cu cathode and insoluble anode, were confirmed electrochemically and chemically, using a membrane to separate each electrode. At the insoluble anode, SPS is decomposed mostly into PDS through irreversible electrochemical reaction. However, SPS is also decomposed by a chemical reaction as well as the electrochemical reaction at the Cu cathode. Cupric ions convert into cuprous ions by disproportionation reaction with Cu electrode, and these cuprous ions forms free radicals with dissolved O2 in the solution. SPS and Cu(I)-MPS react with the free radicals, and finally converted into PDS.

PEG and PEG-PPG copolymer, which are typically used suppressor, are slowly decomposed compared to the accelerator. In the case of this suppressor, it is divided into the small molecular weight of PEG. As the electrodeposition proceeds, the molecular weight of the PEG decreases to less than a certain value, not showing suppression effect. Moreover, by degradation of PEG, byproducts are produced, having a functional group such as hydroxyl (-OH), aldehyde (-COOH), formic ester (-COO), ketone (C=O), and propyl (-C3H6-).

Because of these degradation reactions, monitoring the condition of plating bath is necessary for reliability of processes, which has been optimized by our group.

It is difficult to monitor the change of additives’ chemical properties. Therefore, monitoring the additives concentration can be alternative to monitor the condition of the solution. Additives are decomposed to byproduct and incorporated into the deposit due to their chemical/electrochemical effects. Therefore, the degradation of additives decreases the additives’ concentration. Because the electrochemical behavior of the additives is critically affected by the concentration, it can be measured using cyclic voltammetry stripping (CVS).

CVS method uses the voltammetric approach in determining the concentrations, and it focuses on the stripping charge, which represents the electroplating rate influenced by the type and concentration of additives. Calibration curve of a target additive is obtained by CVS using standard solution whose concentration of the target additive is known. Using this curve, the concentration of the additive by interpolating the stripping charge of a sample solution can be determined.

For monitoring additives used in Cu electrodeposition solution, CVS methods have been optimized for each type of additives. Modified linear approximation technique-CVS (MLAT-CVS) is used for determining concentration of accelerator. However, MPS, the degradation byproduct, acts as an interference due to its different accelerating effects on Cu electroplating compared with SPS. Therefore, the function has been devised related to the conversion ratio from SPS to MPS, which enabled the determination of individual SPS/MPS concentration. Moreover, by controlling the pH of the base and target solutions, MPS was completely converted to SPS, and the total amount of accelerating agent could be determined when using MLAT-CVS method.

Dilution/titration-CVS (DT-CVS) method is generally used for determining concentration of suppressor. When the conventional DT-CVS analysis was applied to two-additive target solutions containing both accelerator and suppressor, the accelerator could affect stripping charge as an interference. Even small concentrations of accelerator in CVS bath can increase error in determined suppressor concentration. We suggested new method of determining the suppressor concentration in two-additive solution by introducing iodide ion which completely suppressed effect of the accelerator.

As mentioned above, iodide ion has been recently developed for defect-free TSV filling due to its excellent inhibition effect by deactivating SPS. Response curve-CVS (RC-CVS) method could be applied to monitor concentration of iodide ion due to its electrochemical behavior. Optimizing the determination process based on the mechanism of I- affecting Cu electrodeposition, the successful monitoring of I- concentration was performed in wide range with high resolution. This method was used to recover the performance of the electroplating solution containing three additives for TSV filling.