1.
Introduction
With the reduction of feature size and the increase of the integration level in an integrated circuit (IC), the requirement for cleanliness of the multilayer Cu wiring wafer surface becomes higher. Increasing attention has been paid to the CMP post cleaning process[1–4]. Benzotriazole (BTA), which acts as an anticorrosive agent in the polishing solution, is the main organic pollutant remaining after multilayer copper wiring CMP[5, 6]. In the process of CMP, it is difficult for Cu-BTA film to desorb. The BTA residue after CMP can cause the surface to be hydrophobic and increase the adsorption probability of the particles. Cu-BTA will release toxic gases and produce pores on the surface of the wafer in some processes of high temperature manufacturing of integrated circuit, which leads to the diffusion of copper ions to the dielectric layer, and the serious problems of time-dependent dielectric breakdown (TDDB). Furthermore, the BTA residue seriously affects the reliability of the copper interconnection and the performance of semiconductor devices[7, 8]. Therefore, the removal of BTA is the most important task for post CMP cleaning.
At present, the mechanism of Cu-BTA film formation is not clear. Based on the research around the world, the theoretical basis of BTA film formation is concluded as follows: the nitrogen atom in the upper NH group of the BTA triazole ring contains solitary pair electrons, the π bond formed on the benzene ring can be combined with the copper atom SP orbit to form a covalent bond. The copper atom is combined with a pair of nitrogen atoms in the 3 position of the other BTA molecule, forming an adsorption polymer membrane with multiple chains[9, 10]. In fact, BTAH and BTA? adsorb the positive Cu2O surface by electrostatic gravitation, and BTAH reacts with Cu2O forming chemisorbed Cu/Cu2O/Cu(I)-BTA composite passivation film. The adsorption structure is shown in Fig. 1.
onerror="this.onerror=null;this.src='http://www.jos.ac.cn/fileBDTXB/journal/article/jos/2018/12/PIC/18060024-1.jpg'" class="figure_img" id="Figure1"/>
Download
Larger image
PowerPoint slide
Figure1.
The adsorption structure of Cu(I)-BTA on Cu2O.
According to the diagram of the copper -BTA- redox potential in water system-pH obtained by Hirabayashi[11], as is shown in Fig. 2, the chemical properties of the solution system and the environment seriously affect the formation and adsorption of the BTA film. It is generally believed that when the PH value ranges from 4 to 10, Cu(I)-BTA is stable[12]. In order to remove BTA effectively, the pH of the cleaning solution should be less than 4 or more than 10. Hydroxyl ions ionized from the alkaline cleaning solution can react with CuO produced from CMP, which will remove the loose copper oxide layer and leave the Cu2O, making the copper wafer passivating. However, an environment with a high pH will increase the static corrosion rate of copper. Thus, the pH setting of the alkaline cleaning solution should be as small as possible within the range that can effectively destroy the Cu(I)-BTA structure, and the pH was suggested from 10.00 to 10.80.
onerror="this.onerror=null;this.src='http://www.jos.ac.cn/fileBDTXB/journal/article/jos/2018/12/PIC/18060024-2.jpg'" class="figure_img" id="Figure2"/>
Download
Larger image
PowerPoint slide
Figure2.
(Color online) Potential-pH diagram of copper water system[11].
Electrochemical measurement was used to quantitatively characterize the BTA residual after cleaning. KOH (PH = 13.90), TMAH (pH = 14.31), TMAH (pH = 14.98), and FA/O II chelating agent (Developed by Institute of Microelectronics, Hebei University of Technology Developed, pH = 14.29) were selected as alternative components of the alkaline cleaning solution to research the BTA removal on the copper surface.
2.
Electrochemical measurement of BTA residues
Due to the adsorption of BTA on the copper surface, the passivation and anti-corrosion characteristics of copper are obvious. In the electrochemical system, the amount of BTA adsorbed on the copper surface directly affects the dissolution rate of copper in the solution, so the electrochemical method can be used to achieve the quantitative detection analysis of BTA residues after cleaning. The magnitude of the corrosion current quantitatively reflects the severity of the anodic reaction. Therefore, the amount of BTA removal during the cleaning process can be characterized by the increase of the self-corrosion current density (Jcorr) in the same specific solution before and after cleaning.
In this paper, a three-electrode system was used in the electrochemical detection experiment. A platinum electrode was used as the counter electrode, a saturated calomel electrode was used as the reference electrode, copper wafer with a 10 × 10 mm2 window coated with the epoxy resin film was used as the working electrode, and a 5 g/L potassium chloride solution was used as the electrolyte. The parameter setting of the electrochemical workstation is shown in Table 1.
| Parameter | OCP | Tafel |
| Scanning scale (V) | ?1 to 1 | OCP ± 0.3 |
| Scanning time (s) | 600 | – |
| Scanning rate (10 mV/s) | – | 10 |
Table1.
Parameter setting of electrochemical workstation
Table options
-->
Download as CSV
| Parameter | OCP | Tafel |
| Scanning scale (V) | ?1 to 1 | OCP ± 0.3 |
| Scanning time (s) | 600 | – |
| Scanning rate (10 mV/s) | – | 10 |
After the electrochemical detection test, the Tafel curve is shown in Fig. 3, the Jcorr of the newly obtained copper wafer after polishing with BTA-free polishing liquid was 9.115 × 10?7 A/cm2, then the copper wafer was soaked in a saturated BTA solution for 10 min to introduce the BTA-stained, and the Jcorr of the BTA-stained copper wafer was only 1.066 × 10?7 A/cm2. The difference between the two values was large, and the BTA stain caused a significant decrease in Jcorr. This method can be used to analyze the residual BTA after cleaning quantitatively.
onerror="this.onerror=null;this.src='http://www.jos.ac.cn/fileBDTXB/journal/article/jos/2018/12/PIC/18060024-3.jpg'" class="figure_img" id="Figure3"/>
Download
Larger image
PowerPoint slide
Figure3.
(Color online) Change of corrosion current density caused by BTA contamination.
3.
Experiment
The FA/OII type chelating agent, TMAH, TEAH, and KOH four alkali cleaning solutions were prepared with the fractions of 100, 150, 200, 250, and 300 ppm, and the pH values of the prepared cleaning solutions for each group were tested. Copper wafer coupons of 3-inch used in the experiments were cut from 12-inch blanket copper wafers. In order to prevent the natural oxides produced from influencing the experimental results, copper wafers were polished utilizing an E460 polishing machine produced by the France Alpsitec Corporation. The prepared slurry was composed of 20 wt% colloidal silica abrasive, 0.15 wt% FA/O II chelating agent, 1.2 wt% BRJI 30 nonionic surfactant, and 1 ml/L H2O2 developed by Hebei University of Technology. The polishing time was 10 s. Newly polished copper wafers were immersed in saturated BTA solution for 10 min to introduce the BTA contamination. An HQE-48S PVA brush was used to carry out the brushing process for 1 min with a flow rate of 1 L/min. The cleaning solutions used in the experiments were the FA/O II chelating agent, TMAH, TEAH, and KOH. The contact angle of the copper surface before and after cleaning was measured. Each experiment was repeated three times to ensure the reality of the results, and the average value was used for analysis. According to the electrochemical measurement described in the previous section, SEM measurement was done to identify the type of residues and verify the cleaning results. The BTA removal effect of each group of cleaning experiments was studied.
4.
Experimental results and analysis
4.1
pH analysis
Table 2 presents the pH of four different alkalis. Except for 100 ppm TMAH (pH 9.92) and 100 ppm KOH (pH 9.98), the pH of cleaning solutions are all maintained at a suitable range from 10.0 to 10.8, as is shown in Table 1. Overall, TEAH has the greatest alkaline ability. The order of alkalinity from strong to weak is TMAH, FA/O II, and KOH. The pH values of KOH and FA/O II along with the change of the concentration are more linear. As an inorganic alkali, the ionization of KOH as an inorganic base in water is completed at once. The FA/O II chelating agent is a good sustained release agent. The pH value will not decline drastically in a certain range.
4.2
Analysis of the BTA removal effect and regularity
Figs. 4 and 5 show electrochemical measurement and contact angle results of the cleaning solutions based on four different alkalis respectively. According to Table 2 and Fig. 4(a), when the concentration of KOH was 100 ppm (pH = 9.98), it can be shown in Fig. 5, the contact angle of copper coupon was 45.84°, and the Jcorr value was 1.491 × 10?7 A/cm2. Compared with 67°, 1.066 × 10?7 A/cm2 of BTA-stained copper wafer, the result was almost unchanged. The result indicated that the BTA adsorption condition has not been improved by cleaning, Cu(I)-BTA adsorption on the copper wafer was still very reliable, showing a strong passivation of copper. With the increase of KOH concentration, the value of self-corrosion current density became larger, which illustrated that the increase of pH in the cleaning environment could actually destroy the adsorption structure of Cu(I)-BTA and facilitate the removal of BTA; when using the KOH concentration of 300 ppm (pH = 10.42), the contact angle was 45.07°, and the Jcorr value was 2.235 × 10?7 A/cm2, reaching the best level in the KOH group. However, compared with the original contact angle (28°) and Jcorr value of copper without contamination (9.115 × 10?7 A/cm2), the difference was obvious. It can be concluded that under the alkaline condition (pH ranging from 10 to 10.8), the effect of BTA removal was not efficient. This shows that the mechanical function of alkaline construction and brushing cannot remove BTA effectively. However, an excessively high pH environment will reduce the self-corrosion voltage of copper in the cleaning solution and increase the dissolution tendency of copper, which will be harmful to the maintenance of planarization of the wafer surface after cleaning.
onerror="this.onerror=null;this.src='http://www.jos.ac.cn/fileBDTXB/journal/article/jos/2018/12/PIC/18060024-5.jpg'" class="figure_img" id="Figure5"/>
Download
Larger image
PowerPoint slide
Figure5.
(Color online) Contact angles of Cu surface cleaned by different concentrations of alkalis.
According to Table 2 and Fig. 4(b), when the concentration of TMAH (pH = 9.92) was 100 ppm, the contact angle was 41.73°, and the Jcorr value was 1.890 × 10 ?7 A/cm2. The effect of BTA removal was poor, which verified that when the pH value was lower than 10, the adsorption structure of Cu-BTA was not destroyed; when the concentration of TMAH was increased, the effect of BTA removal was also increased, reaching a maximum at 300 ppm, the contact angle was 29.11°, and the Jcorr value was 5.762 × 10?7 A/cm2.
According to Table 2 and Figs. 4(c) and 4(d), the Jcorr values of TEAH and FA/O II chelating agents do not increase monotonically with the increase of concentration, indicating the effect of BTA removal is not simply the increase of concentration and pH value. The reason may be that the molecular weight of the FA/O II chelating agent and TEAH is relatively large. The increase of concentration increases facilitated its own reaction and affected its overall external reactivity. When the concentration of the FA/O II cleaning solution is too high, both the TEAH and FA/O II chelating agent reached their best BTA removal results at 200 ppm. The contact angle was 29.71°, the Jcorr value of TEAH was 5.282 × 10?7 A/cm2, while the contact angle was 28.68° and the Jcorr value of FA/O II was 7.231 × 10?7 A/cm2. Compared with the initial contact angle of 28° and the Jcorr value of 9.14 × 10 A/cm2, BTA was largely removed. Fig. 6 and Fig. 7 show that the SEM images of copper coupons after immersing in saturated BTA solution and after cleaning using 200 ppm FA/O II cleaning solution. It can be seen from the EDX analysis that BTA residues remained on the copper surface before cleaning. The wafer surface has been greatly improved compared to Fig. 6 after cleaning by FA/O II cleaning solution. The SEM image and EDX analysis verified that BTA was removed efficiently, as is shown in Fig. 7.
onerror="this.onerror=null;this.src='http://www.jos.ac.cn/fileBDTXB/journal/article/jos/2018/12/PIC/18060024-7.jpg'" class="figure_img" id="Figure7"/>
Download
Larger image
PowerPoint slide
Figure7.
(Color online) (a) The SEM picture and (b) EDX analysis of copper wafer after cleaning.
onerror="this.onerror=null;this.src='http://www.jos.ac.cn/fileBDTXB/journal/article/jos/2018/12/PIC/18060024-6.jpg'" class="figure_img" id="Figure6"/>
Download
Larger image
PowerPoint slide
Figure6.
(Color online) (a) The SEM picture and (b) EDX analysis of copper wafer before cleaning.
| Alkaline | Mass fraction (ppm) | pH | Jcorr (10?7 A/cm2) |
| KOH | 100 | 9.98 | 1.491 |
| 150 | 10.06 | 1.667 | |
| 200 | 10.24 | 1.980 | |
| 250 | 10.31 | 1.784 | |
| 300 | 10.42 | 2.235 | |
| TMAH | 100 | 9.92 | 1.890 |
| 150 | 10.36 | 4.103 | |
| 200 | 10.43 | 4.226 | |
| 250 | 10.50 | 5.422 | |
| 300 | 10.56 | 5.762 | |
| TEAH | 100 | 10.09 | 2.918 |
| 150 | 10.31 | 3.471 | |
| 200 | 10.46 | 5.282 | |
| 250 | 10.58 | 2.225 | |
| 300 | 10.69 | 2.313 | |
| FA/O II | 100 | 10.05 | 2.237 |
| 150 | 10.13 | 4.382 | |
| 200 | 10.26 | 7.231 | |
| 250 | 10.39 | 3.938 | |
| 300 | 10.51 | 4.159 |
Table2.
Effect of BTA removal on cleaning solutions based on four different alkalis.
Table options
-->
Download as CSV
| Alkaline | Mass fraction (ppm) | pH | Jcorr (10?7 A/cm2) |
| KOH | 100 | 9.98 | 1.491 |
| 150 | 10.06 | 1.667 | |
| 200 | 10.24 | 1.980 | |
| 250 | 10.31 | 1.784 | |
| 300 | 10.42 | 2.235 | |
| TMAH | 100 | 9.92 | 1.890 |
| 150 | 10.36 | 4.103 | |
| 200 | 10.43 | 4.226 | |
| 250 | 10.50 | 5.422 | |
| 300 | 10.56 | 5.762 | |
| TEAH | 100 | 10.09 | 2.918 |
| 150 | 10.31 | 3.471 | |
| 200 | 10.46 | 5.282 | |
| 250 | 10.58 | 2.225 | |
| 300 | 10.69 | 2.313 | |
| FA/O II | 100 | 10.05 | 2.237 |
| 150 | 10.13 | 4.382 | |
| 200 | 10.26 | 7.231 | |
| 250 | 10.39 | 3.938 | |
| 300 | 10.51 | 4.159 |
onerror="this.onerror=null;this.src='http://www.jos.ac.cn/fileBDTXB/journal/article/jos/2018/12/PIC/18060024-4.jpg'" class="figure_img" id="Figure4"/>
Download
Larger image
PowerPoint slide
Figure4.
(Color online) Electrochemical measurement results of the cleaning solutions based on four different alkalis. (a) KOH. (b) TMAH. (c) TEAH. (d) FA/O II.
4.3
Mechanism analysis
The BTA removal experiments revealed that when the pH is controlled in the range from 10.0 to 10.8, the efficiency of the three organic amine alkalis was significantly higher than the inorganic amine alkali. This is because the three organic alkalis all contained the same NC bonding with BTA molecules. According to the similar dissolution of organics, the BTA monolayer and some BTA molecules not forming the polymer adsorbed on the copper surface will be largely dissolved in organic alkaline agents and will leave the copper surface with the flow cleaning solution. In addition, under strong alkaline conditions with a pH greater than 10, the Cu(I)-BTA structure is destroyed. There is a weak ionization equilibrium, as shown in Eq. (1)[5], which releases the cuprous ions. The cuprous ions are not stable in the solution and can be oxidized to copper ions easily. The three organic alkalis can react with copper ions, which will accelerate the cracking of Cu(I)-BTA film. The polyhydroxylamine structure of the FA/O II chelating agent can form 13 effective chelating rings, and it reacts with copper ions to produce stable soluble copper amine ions, as is shown in Eq. (2)[5]. Thus, the effect of BTA removal by FA/O II chelating agent is the best.
$${ m {Cu(I)}} - { m BTA} leftrightarrow { m C}{ m u^ + } + { m BT}{ m A^ - },$$  | (1) |
$${ m C}{ m u^{2 + }} + { m R}{left( { m N{H_3}} ight)_4} leftrightarrow {left[ { m Cu-left( { m R{{left( { m N{ m H_3}} ight)}_4}} ight)} ight]^{2 + }}.$$  | (2) |
According to the results and the analysis of the BTA removal tests by different types of alkaline solutions: an alkaline environment with a pH greater than 10 destroyed the Cu-N bond formed by Cu atoms and unsaturated N in the BTA molecule. The adsorption structure of Cu(I)-BTA was also destroyed, forming a weak ionization equilibrium as shown in formula Eq. (1). However, the destructive effect of the alkaline environment on Cu(I)-BTA alone could not eliminate the BTA contamination adsorbed on the copper surface completely. It is necessary to use the complex reaction of copper ions with chelating agents and complexing agents to remove copper ions. According to the similar dissolution theory of organics, the ionization balance in Eq. (1) can be broken by increasing the solubility of BTA molecules in cleaning solutions. The disintegration of Cu(I)-BTA film will be accelerated. The BTA contamination will be removed completely by the mechanical function of PVA brushing.
5.
Conclusion
In this paper, the contact angle measurement and electrochemical measurement were used to quantitatively analyze the BTA residue after Cu CMP cleaning, and BTA removal effect of four organic amine alkalis and inorganic alkali on BTA was studied. The conclusions were obtained as follows: (1) The alkaline environment with a pH of less than 10 cannot destroy the structure of the Cu(I)-BTA adsorption film. To effectively remove BTA, the pH of the alkaline cleaning solution should be set larger than 10. (2) BTA cannot be effectively removed only by the alkaline environment and the mechanical action of brushing. Under the same condition of pH, the effect of organic alkali with the coordination structure is better than that of inorganic alkali. With the coordination reaction, the disintegration of Cu(I)-BTA is accelerated and the structure of Cu(I)-BTA film is destroyed under strong alkaline conditions. (3) The 200 ppm FA/O II chelating agent can effectively remove the BTA residue on the copper surface.
