A low-stress polishing slurry and method for ruthenium interconnect air gaps
By using low-stress polishing fluid and a two-step polishing process, the problems of insufficient ruthenium removal capacity and air gap residue in the integrated structure of ruthenium interconnect and air gap were solved, achieving efficient removal of ruthenium material and protection of air gap, and improving surface flatness and cleanliness.
Patent Information
- Application Number
- CN202611356689.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-03
- Publication Date
- 2026-11-17
AI Technical Summary
In the chemical mechanical polishing process of ruthenium interconnect and air gap integrated structure, there are problems such as insufficient ruthenium removal capacity and easy damage to the air gap structure. At the same time, the polishing slurry can easily enter the unclosed air gap and form residues.
A low-stress polishing slurry and method are employed, comprising colloidal silica, hydrogen peroxide, ammonium persulfate, citric acid, and polyethylene glycol 600. The pH value is adjusted to 8.0–9.0, and a two-step process of coarse polishing and fine polishing is combined. The polishing pressure is controlled at 0.3–1.0 psi, and the slurry is cleaned with dilute ammonia and dried with isopropanol vapor.
This method achieves effective removal of ruthenium material, reduces reliance on mechanical grinding, minimizes residue and structural damage within air gaps, improves surface flatness and cleanliness, and meets the low-stress flatness requirements of ruthenium interconnect and air gap integrated structures.
Smart Images

Figure SMS_67 
Figure SMS_68 
Figure SMS_69
Abstract
Description
Technical Field
[0001] This invention relates to the field of semiconductor chemical mechanical polishing technology, and particularly to a low-stress polishing slurry and method for ruthenium interconnect air gaps. Background Technology
[0002] As integrated circuit technology nodes continue to shrink, traditional copper interconnects face problems such as increased resistivity and RC delay under small-size conditions. Ruthenium, with its short electron mean free path and good electromigration resistance, can be used in interconnect structures for advanced technology nodes and can be combined with air gap structures to reduce parasitic capacitance of interconnect structures.
[0003] In the integration process of ruthenium interconnects and air gaps, a ruthenium metal thin film is first formed and ruthenium interconnect wires are formed by etching. Then, air gaps are formed between adjacent ruthenium interconnect wires, and a capping material is applied over the air gaps. After the capping layer is deposited, there may still be a certain height difference on its surface, which requires global planarization by chemical mechanical polishing to meet the surface flatness requirements of subsequent processes.
[0004] During the chemical mechanical polishing stage, the sealant at the top of the air gap has not yet completely sealed the air gap. Polishing fluid can enter the air gap through the top opening or micropores in the sealant and become trapped under capillary action. If the abrasive particles, chemical components, and moisture that have entered the air gap are not discharged in time, they can easily leave residues in the air gap during subsequent cleaning and drying processes, thus affecting the sealing of the subsequent sealant and the interface condition.
[0005] On the other hand, ruthenium metal has high hardness and strong chemical inertness, which can easily lead to insufficient material removal capacity under conditions of reduced mechanical action. If mechanical grinding is enhanced by increasing the content of abrasive particles or polishing pressure, the risk of mechanical damage to the air gap structure can be increased. Therefore, the integrated structure of ruthenium interconnects and air gaps requires chemical mechanical polishing to balance ruthenium removal capacity, low mechanical action, and control of air gap residue.
[0006] To address this, a low-stress polishing fluid and method for ruthenium interconnect air gaps are proposed. Summary of the Invention
[0007] The purpose of this invention is to provide a low-stress polishing slurry and method for ruthenium interconnect air gaps, in order to solve the problems that when performing chemical mechanical polishing on ruthenium interconnect and air gap integrated structures, reducing the polishing pressure can easily lead to insufficient ruthenium removal capacity, while increasing the mechanical action can easily cause damage to the air gap structure, and the polishing slurry can easily leave residues after entering the air gap that is not completely closed.
[0008] The technical solution of the present invention is achieved as follows: To achieve the above objective, the present invention provides a low-stress polishing slurry for ruthenium interconnect air gaps, comprising, by weight 100%, 0.2–0.5 wt% colloidal silica, 0.8–1.2 wt% hydrogen peroxide, 0.08–0.20 wt% ammonium persulfate, 0.15–0.30 wt% citric acid, 0.03–0.08 wt% polyethylene glycol 600, a pH adjuster, and deionized water. The pH adjuster adjusts the pH of the polishing slurry to 8.0–9.0, and the deionized water is the balance.
[0009] Furthermore, the average particle size of the colloidal silica is 30–50 nm.
[0010] Furthermore, the pH adjuster is tetramethylammonium hydroxide.
[0011] Furthermore, based on the total weight of the polishing slurry as 100%, the content of colloidal silica is 0.3 wt%, the content of hydrogen peroxide is 1.0 wt%, the content of ammonium persulfate is 0.10-0.15 wt%, the content of citric acid is 0.20 wt%, the content of polyethylene glycol 600 is 0.05 wt%, and the pH of the polishing slurry is 8.5.
[0012] In the polishing slurry, colloidal silica is used to provide mechanical removal; hydrogen peroxide is used to oxidize the ruthenium surface; ammonium persulfate is used to promote the oxidation process of the ruthenium surface; citric acid is used to complex the ruthenium species formed by oxidation, allowing the ruthenium species to enter the liquid phase and reducing their redeposition on the ruthenium surface; polyethylene glycol 600 is used to reduce the surface tension of the polishing slurry and reduce the retention of the polishing slurry and abrasive particles in the air gap.
[0013] The present invention also provides a low-stress polishing method for ruthenium interconnect air gaps, comprising the following steps: S1. A semiconductor substrate is provided, on which ruthenium interconnect wires and air gaps are formed, and an encapsulation layer is covered over the air gaps, wherein the encapsulation layer does not completely seal the air gaps. S2. Rough polishing is performed using a low-stress polishing slurry. The low-stress polishing slurry, by weight (100%), comprises 0.2–0.5 wt% colloidal silica, 0.8–1.2 wt% hydrogen peroxide, 0.08–0.20 wt% ammonium persulfate, 0.15–0.30 wt% citric acid, 0.03–0.08 wt% polyethylene glycol 600, a pH adjuster, and deionized water. The pH adjuster adjusts the pH of the low-stress polishing slurry to 8.0–9.0. The deionized water is the balance. The polishing pressure for rough polishing is 0.5–1.0 psi, and the polishing time is 30–60 seconds. S3. Fine polishing is performed using the aforementioned low-stress polishing fluid. The polishing pressure for fine polishing is 0.3–0.6 psi, and the polishing time is 15–30 seconds. S4. Rinse with deionized water or dilute ammonia, then dry with isopropanol steam.
[0014] Furthermore, in step S2, the polishing pad rotation speed for coarse polishing is 60–90 rpm, the wafer rotation speed is 55–85 rpm, and the polishing slurry flow rate is 150–250 mL / min.
[0015] Furthermore, in step S3, the polishing pad rotation speed is 40-60 rpm, the wafer rotation speed is 35-55 rpm, and the polishing slurry flow rate is 100-150 mL / min.
[0016] Furthermore, in step S2, the polishing pressure for rough polishing is 0.8 to 1.0 psi; in step S3, the polishing pressure for fine polishing is 0.4 to 0.6 psi.
[0017] Further, in step S4, the water is cleaned with dilute ammonia for 45 to 60 seconds, wherein the volume ratio of ammonia to water in the dilute ammonia is 1:150 to 250, and then dried with isopropanol vapor.
[0018] Further, in step S1, the material of the sealing layer is SiCN or SiCOH, and the height difference of the sealing layer surface is no greater than 50 nm.
[0019] The embodiments of the present invention have the following advantages due to the adoption of the above technical solutions: I. This invention employs a chemical-mechanical removal system for ruthenium using colloidal silica, hydrogen peroxide, ammonium persulfate, and citric acid. Ammonium persulfate promotes ruthenium surface oxidation, while citric acid induces the oxidized ruthenium species to enter the liquid phase, where they are mechanically removed in conjunction with colloidal silica. This reduces the dependence of ruthenium removal on mechanical grinding. Under polishing pressures of 0.5–1.0 psi, the Ru removal rate can reach 832–1150. / min, which helps to meet the low-stress flattening requirements of the integrated structure of ruthenium interconnect and air gap.
[0020] Second, this invention adds polyethylene glycol 600 to the polishing slurry, thereby reducing the surface tension of the polishing slurry and the retention of polishing slurry and abrasive particles in the air gaps. When the content of polyethylene glycol 600 is 0.05 wt%, the surface tension of the polishing slurry is 32 mN / m, the air gap residue rate is reduced from 38% without the addition of air gap protectant to 2%, and no voids appear at the sealing interface, thereby reducing the risk of residue formation in the air gaps during the polishing process.
[0021] III. This invention employs a two-step low-stress polishing method combining coarse and fine polishing. By controlling the polishing pressure in stages, the influence of mechanical action on the air gap structure and sealing layer is controlled while removing ruthenium material. When the coarse polishing pressure is 0.8 psi and the fine polishing pressure is 0.5 psi, the post-polishing flatness is 4.8 nm, the surface roughness Ra is 0.32 nm, and the remaining SiCN thickness is 285 nm. With a yield rate of 97.8%, it can balance the planarization of ruthenium interconnect surface and the protection of air gap structure.
[0022] IV. This invention employs dilute ammonia cleaning after two-step low-stress polishing, combined with isopropanol vapor drying, which further reduces residual air gaps and metal ions after polishing. When cleaning with dilute ammonia water (1:200 volume ratio) for 45 seconds followed by isopropanol vapor drying, no residual air gaps were observed, and the residual metal ions were 1.2 × 10⁻⁶. 10 atoms / cm 2 This improves the cleanliness of the wafer after polishing. Detailed Implementation
[0023] This invention provides a low-stress polishing slurry and method for ruthenium interconnect air gaps. By combining colloidal silica, hydrogen peroxide, ammonium persulfate, citric acid and polyethylene glycol 600, the oxidation, complexation and removal of ruthenium surface are achieved while reducing the mechanical polishing effect. Polyethylene glycol 600 (PEG600) is used to reduce the retention of polishing slurry in the air gap.
[0024] In this embodiment, the percentage content of each component is a mass percentage relative to the total weight of the polishing solution, with deionized water as the balance; pH is adjusted using tetramethylammonium hydroxide. The polishing experiment used an 8-inch wafer with a ruthenium film thickness of 2000 μm on the wafer surface. The IC1000 polishing pad was used, and the polishing equipment was a Reflexion LK.
[0025] In this embodiment, the ruthenium interconnect and air gap integrated structure uses an air gap formed between adjacent ruthenium interconnect wires, with a SiCN or SiCOH capping layer covering the top of the air gap. During chemical mechanical polishing, the capping layer does not completely seal the air gap, allowing polishing fluid to enter through the top opening, the micropores of the capping layer, and capillary action. Therefore, while planarizing the ruthenium surface, it is also necessary to reduce mechanical action and the retention of polishing fluid inside the air gap. In the patterned wafer used for air gap structure testing, the linewidth of the air gap is 28 nm, the spacing is 28 nm, and the gap height is 80 nm.
[0026] The polishing slurry in this embodiment, by weight, comprises colloidal silica, hydrogen peroxide, ammonium persulfate, citric acid, polyethylene glycol 600, tetramethylammonium hydroxide, and deionized water. Colloidal silica provides the mechanical grinding action; hydrogen peroxide oxidizes the ruthenium surface; ammonium persulfate promotes the oxidation process on the ruthenium surface; citric acid complexes the ruthenium species formed by oxidation, reducing ruthenium species redeposition; and polyethylene glycol 600 reduces the surface tension of the polishing slurry, minimizing the retention of the polishing slurry and its particles within the air gaps.
[0027] In one specific embodiment, the colloidal silica has an average particle size of 30–50 nm and a content of 0.2–0.5 wt%; hydrogen peroxide content is 0.8–1.2 wt%; ammonium persulfate content is 0.08–0.20 wt%; citric acid content is 0.15–0.30 wt%; polyethylene glycol 600 content is 0.03–0.08 wt%; the pH of the polishing solution is adjusted to 8.0–9.0 using tetramethylammonium hydroxide, with the remainder being deionized water.
[0028] When polishing a ruthenium interconnect air gap structure using the above-mentioned polishing slurry, the following steps are included: S1. Provide a semiconductor substrate, on which ruthenium interconnect wires and air gaps are formed, and a SiCN or SiCOH sealing layer is covered above the air gaps, with the height difference of the sealing layer surface not exceeding 50 nm; S2. Use the polishing slurry for rough polishing. The polishing pressure is 0.5-1.0 psi, the polishing pad speed is 60-90 rpm, the wafer speed is 55-85 rpm, the polishing slurry flow rate is 150-250 mL / min, and the polishing time is 30-60 seconds. S3. Fine polishing is performed using the polishing slurry. The polishing pressure is 0.3 to 0.6 psi, the polishing pad rotation speed is 40 to 60 rpm, the wafer rotation speed is 35 to 55 rpm, the polishing slurry flow rate is 100 to 150 mL / min, and the polishing time is 15 to 30 seconds. S4. Use deionized water and / or dilute ammonia water for circulating rinsing. The volume ratio of ammonia water to water in the dilute ammonia water is 1:100-500, and the rinsing time is 30-60 seconds. Then use isopropanol steam drying.
[0029] During the polishing process, a hard polyurethane polishing pad with a Shore hardness D of 50 to 70 can be used. The surface of the polishing pad has a groove structure.
[0030] Example 1 In this embodiment, colloidal silica is used as the abrasive particles. The silica content is 0.1 wt%, the hydrogen peroxide content is 0.5 wt%, the ammonium persulfate content is 0.05 wt%, the citric acid content is 0.20 wt%, and the PEG600 content is 0.05 wt%. Tetramethylammonium hydroxide is used to adjust the pH of the polishing solution to 8.5, and the remainder is deionized water.
[0031] Ruthenium wafers were polished for 60 seconds using the aforementioned polishing slurry at a polishing pressure of 0.8 psi. Test results showed a Ru removal rate of 623%. The surface roughness Ra after polishing is 0.45 nm, and the defect density is 12 defects / cm². 2 .
[0032] Example 2 The difference between this embodiment and Example 1 is that the hydrogen peroxide content is adjusted to 1.0 wt%, the ammonium persulfate content is adjusted to 0.10 wt%, and the rest of the formula and polishing conditions remain the same.
[0033] Test results show that the Ru removal rate is 892. The surface roughness Ra is 0.38 nm, and the defect density is 8 defects / cm². 2 .
[0034] Example 3 The difference between this embodiment and Embodiment 1 is that the hydrogen peroxide content is 1.5 wt% and the ammonium persulfate content is 0.15 wt%, while the other conditions remain the same.
[0035] Test results show that the Ru removal rate is 945. The surface roughness Ra is 0.42 nm, and the defect density is 15 defects / cm². 2 .
[0036] Example 4 In this embodiment, the content of colloidal silica is 0.3 wt%, the content of hydrogen peroxide is 0.5 wt%, the content of ammonium persulfate is 0.10 wt%, the content of citric acid is 0.20 wt%, the content of PEG600 is 0.05 wt%, the pH is 8.5, and it is polished for 60 seconds under a pressure of 0.8 psi.
[0037] The test yielded a Ru removal rate of 781. / min, surface roughness Ra is 0.35 nm, defect density is 6 / cm 2 .
[0038] Example 5 In this embodiment, the content of colloidal silica is 0.3 wt%, the content of hydrogen peroxide is 1.0 wt%, the content of ammonium persulfate is 0.15 wt%, the content of citric acid is 0.20 wt%, the content of PEG600 is 0.05 wt%, the pH is adjusted to 8.5 using tetramethylammonium hydroxide, and the remainder is deionized water; the polishing pressure is 0.8 psi, and the polishing time is 60 seconds.
[0039] The test yielded a Ru removal rate of 1150. The surface roughness Ra is 0.31 nm, and the defect density is 5 defects / cm². 2 .
[0040] Example 6 Compared with Example 5, the hydrogen peroxide content in this embodiment is adjusted to 1.5 wt%, the ammonium persulfate content is adjusted to 0.20 wt%, and the other conditions remain the same.
[0041] The test result showed a Ru removal rate of 1205. / min, surface roughness Ra 0.36 nm, defect density 9 / cm 2 .
[0042] Example 7 This embodiment uses 0.5 wt% colloidal silica, 0.5 wt% hydrogen peroxide and 0.15 wt% ammonium persulfate, with 0.20 wt% citric acid, 0.05 wt% PEG600, pH 8.5, polishing pressure 0.8 psi and polishing time 60 seconds.
[0043] The test yielded a Ru removal rate of 832. / min, surface roughness Ra is 0.32 nm, defect density is 8 / cm 2 .
[0044] Example 8 In this embodiment, 0.5 wt% colloidal silica, 1.0 wt% hydrogen peroxide and 0.20 wt% ammonium persulfate were used, and the remaining components and polishing conditions were the same as in Example 7.
[0045] The test yielded a Ru removal rate of 1188. The surface roughness Ra is 0.29 nm, and the defect density is 7 defects / cm². 2 .
[0046] Example 9 In this embodiment, 0.5 wt% colloidal silica, 1.5 wt% hydrogen peroxide and 0.25 wt% ammonium persulfate were used, while the other conditions remained the same.
[0047] The test yielded a Ru removal rate of 1240. The surface roughness Ra is 0.34 nm, and the defect density is 12 defects / cm². 2 .
[0048] As can be seen from Examples 1 to 9, Example 5 shows a relatively coordinated test result among Ru removal rate, surface roughness, and defect density; although increasing the content of grinding particles or hydrogen peroxide can improve the removal rate, the defect density increases accordingly.
[0049] Examples 10-14: Catalytic Activators Examples 10-14 all used 0.3 wt% colloidal silica, 1.0 wt% hydrogen peroxide, 0.20 wt% citric acid and 0.05 wt% PEG600, with the pH adjusted to 8.5 using tetramethylammonium hydroxide and the polishing pressure at 0.8 psi. The difference was in the type or content of the catalytic activator.
[0050] Example 10 In this embodiment, ammonium persulfate was used as the catalytic activator, with a content of 0.05 wt%. The Ru removal rate was tested to be 782. / min, SiCN removal rate 185 / min, the Ru / SiCN selectivity ratio is 4.23.
[0051] Example 11 In this embodiment, the ammonium persulfate content was 0.10 wt%. The Ru removal rate was measured to be 1050. / min, SiCN removal rate 210 / min, Ru / SiCN selectivity is 5.00.
[0052] Example 12 In this embodiment, the ammonium persulfate content was 0.20 wt%. The Ru removal rate was measured to be 1180. / min, SiCN removal rate 245 / min, the Ru / SiCN selectivity ratio is 4.82.
[0053] Example 13 In this embodiment, the catalyst activator was adjusted to 0.10 wt% potassium persulfate, while the remaining conditions were the same as in Example 11. The Ru removal rate was measured to be 968. / min, SiCN removal rate 198 / min, the Ru / SiCN selectivity ratio is 4.89.
[0054] Example 14 In this embodiment, 0.10 wt% sodium tungstate was used as the catalytic activator. The Ru removal rate was measured to be 723. / min, SiCN removal rate 175 / min, the Ru / SiCN selectivity ratio is 4.13.
[0055] In the above tests, the Ru / SiCN selectivity ratio corresponding to 0.10 wt% ammonium persulfate was 5.00; although increasing the ammonium persulfate to 0.20 wt% improved the Ru removal rate, the Ru / SiCN selectivity ratio decreased.
[0056] Examples 15-19: Ruthenium complexing agents Examples 15-19 consistently used 0.3 wt% colloidal silica, 1.0 wt% hydrogen peroxide, 0.10 wt% ammonium persulfate, and 0.05 wt% PEG600, with a pH of 8.5 and a polishing pressure of 0.8 psi. The difference lay in the type and content of the complexing agent.
[0057] Example 15 In this embodiment, 0.10 wt% citric acid was used as the ruthenium complexing agent. The Ru removal rate was tested to be 892. At a rate of 1000 m / min, slight redeposition of Ru was observed on the Ru surface; after the polishing slurry was allowed to stand for 24 hours, the Ru removal rate decreased by 8%.
[0058] Example 16 This embodiment uses 0.20 wt% citric acid. The Ru removal rate was tested to be 1050. / min, no ruthenium redeposition was observed; after the polishing slurry was left to stand for 24 hours, the Ru removal rate decreased by 5%.
[0059] Example 17 In this embodiment, the citric acid content was increased to 0.50 wt%. The Ru removal rate was tested to be 1105. / min, no ruthenium redeposition was observed; after standing for 24 hours, the Ru removal rate decreased by 3%.
[0060] Example 18 In this embodiment, 0.20 wt% ethylenediaminetetraacetic acid (EDTA) was used as the complexing agent. The Ru removal rate was tested to be 968. At a rate of 1000 m / min, no ruthenium redeposition was observed; after standing for 24 hours, the Ru removal rate decreased by 12%.
[0061] Example 19 In this embodiment, 0.20 wt% triethanolamine was used as a complexing agent. The Ru removal rate was tested to be 845. The rate was 1 / min, indicating slight ruthenium redeposition; after standing for 24 hours, the Ru removal rate decreased by 18%.
[0062] Examples 20-25: pH conditions Examples 20-25 all used 0.3 wt% colloidal silica, 1.0 wt% hydrogen peroxide, 0.10 wt% ammonium persulfate, 0.20 wt% citric acid and 0.05 wt% PEG600, with a polishing pressure of 0.8 psi, and different pH values were adjusted by tetramethylammonium hydroxide.
[0063] Example 20 The pH of the polishing solution was adjusted to 7.0. The Ru removal rate was measured to be 782. / min, low-k medium corrosion rate 18 / min, air gap protection status is good, Zeta potential is -18 mV.
[0064] Example 21 The pH of the polishing solution was adjusted to 7.5. The Ru removal rate was 912. / min, corrosion rate of low-k media 12 / min, the air gap protection status is good, and the Zeta potential is -22 mV.
[0065] Example 22 Adjust the pH to 8.0. The Ru removal rate is 1025. / min, low-k medium corrosion rate is 8 / min, air gap protection status is good, Zeta potential is -25 mV.
[0066] Example 23 Adjust the pH to 8.5. The Ru removal rate is 1050. / min, the corrosion rate in low-k media is 5 / min, the air gap protection status is good, and the Zeta potential is -28 mV.
[0067] Example 24 Adjust the pH to 9.0. The Ru removal rate is 998%. / min, the corrosion rate in low-k media is 4 / min, air gap protection status is normal, Zeta potential is -30 mV.
[0068] Example 25 Adjust the pH to 9.5. The Ru removal rate is 856. / min, the corrosion rate in low-k media is 3 / min, the air gap protection effect decreased and bubbling occurred, and the Zeta potential was -32 mV.
[0069] Examples 26-32: Air gap protectant Examples 26-32 used 0.3 wt% colloidal silica, 1.0 wt% hydrogen peroxide, 0.10 wt% ammonium persulfate and 0.20 wt% citric acid, with a pH of 8.5, a polishing pressure of 0.8 psi, and an air gap linewidth of 28 nm, a spacing of 28 nm and a height of 80 nm.
[0070] Example 26 No air gap protectant was added in this embodiment. The surface tension of the polishing slurry was 72 mN / m, the air gap residue rate was 38%, and the Ru removal rate was 1150. / min, severe voids appeared at the seal layer interface.
[0071] Example 27 In this embodiment, 0.03 wt% PEG400 was added. The surface tension of the polishing slurry was 38 mN / m, the air gap residue rate was 12%, and the Ru removal rate was 1080. / min, slight voids exist at the seal layer interface.
[0072] Example 28 In this embodiment, 0.03 wt% PEG600 was added. The surface tension of the polishing slurry was 35 mN / m, the air gap residue rate was 5%, and the Ru removal rate was 1050. / min, no voids were observed at the sealing layer interface.
[0073] Example 29 In this embodiment, 0.05 wt% PEG600 was added. The surface tension of the polishing slurry decreased to 32 mN / m, the air gap residue rate was 2%, and the Ru removal rate was 1030. / min, no voids were observed at the sealing layer interface.
[0074] Example 30 In this embodiment, 0.10 wt% PEG600 was added. The surface tension of the polishing slurry was 28 mN / m, the air gap residue rate was 1%, and the Ru removal rate was 980. / min, no voids were observed at the sealing layer interface.
[0075] Example 31 In this embodiment, 0.05 wt% PEG1000 was added. The surface tension of the polishing slurry was 34 mN / m, the air gap residue rate was 3%, and the Ru removal rate was 10¹⁰. / min, no voids were observed at the sealing layer interface.
[0076] Example 32 In this embodiment, 0.03 wt% sodium dodecylbenzenesulfonate was added. The surface tension of the polishing slurry was 31 mN / m, the air gap residue rate was 8%, and the Ru removal rate was 1060. / min, slight voids exist at the seal layer interface.
[0077] In the above results, the air gap residual rate corresponding to 0.05 wt% PEG600 was 2%, and no voids were observed at the sealing interface.
[0078] Examples 33-37: Polishing Pressure Examples 33-37 used 0.3 wt% colloidal silica, 1.0 wt% hydrogen peroxide, 0.15 wt% ammonium persulfate, 0.20 wt% citric acid and 0.05 wt% PEG600, with a pH of 8.5, and the polishing pressure was varied.
[0079] Example 33 The polishing pressure was set to 0.3 psi. The tested Ru removal rate was 562. / min, the change in air gap height was less than 1%, and no structural damage was observed.
[0080] Example 34 Polishing pressure was set to 0.5 psi. Ru removal rate was 832. / min, air gap height change is less than 1%, and structural damage rate is 0.
[0081] Example 35 Polishing pressure was set to 0.8 psi. Ru removal rate was 1050. / min, the change in air gap height was less than 1%, and no structural damage was observed.
[0082] Example 36 Polishing pressure was set to 1.0 psi. Ru removal rate was 1150. / min, the air gap height change is 2%, and the structural damage rate is 0.5%.
[0083] Example 37 Polishing pressure was set to 1.2 psi. Ru removal rate was 1220. / min, the air gap height change reached 8%, and the structural damage rate was 3.2%.
[0084] As can be seen from Examples 33-37, when the polishing pressure is not higher than 1.0 psi, the change in air gap height is less than 3% and the proportion of structural damage is less than 1%; when the pressure is increased to 1.2 psi, the change in air gap structure and the proportion of damage increase significantly.
[0085] Examples 38-42: Two-step polishing Examples 38-42 use the polishing slurry described in Example 5 to perform rough and fine polishing on wafers with an air gap structure. The initial SiCN encapsulation layer thickness on the wafer is 500 μm. The initial surface height difference is 40 nm, the coarse polishing time is 45 seconds, and the fine polishing time is 20 seconds.
[0086] Example 38 The coarse polishing pressure was 0.6 psi, and the fine polishing pressure was 0.3 psi. After polishing, the flatness of the 30 μm test area was 8.5 nm, the surface roughness Ra was 0.48 nm, and the remaining SiCN thickness was 325 nm. The yield rate was 96.2%.
[0087] Example 39 The coarse polishing pressure was 0.8 psi, and the fine polishing pressure was 0.4 psi. After polishing, the flatness was 5.2 nm, the surface roughness Ra was 0.35 nm, and the remaining SiCN thickness was 298 nm. The yield rate was 97.5%.
[0088] Example 40 The coarse polishing pressure was 0.8 psi, and the fine polishing pressure was 0.5 psi. The flatness after polishing was 4.8 nm, the surface roughness Ra was 0.32 nm, and the remaining SiCN thickness was 285 nm. The yield rate was 97.8%.
[0089] Example 41 The coarse polishing pressure was 1.0 psi, and the fine polishing pressure was 0.5 psi. After polishing, the flatness was 4.2 nm, the surface roughness Ra was 0.30 nm, and the remaining SiCN thickness was 262 nm. The yield rate was 96.5%.
[0090] Example 42 The coarse polishing pressure was 1.0 psi, and the fine polishing pressure was 0.6 psi. The flatness after polishing was 3.8 nm, the surface roughness Ra was 0.28 nm, and the remaining SiCN thickness was 248 nm. The yield rate was 95.1%.
[0091] In Example 40, relatively well-coordinated test results were obtained in terms of flatness, surface roughness, remaining SiCN thickness, and yield. When the polishing pressure was further increased, the remaining SiCN thickness continued to decrease, and the yield also decreased.
[0092] Examples 43-46: Cleaning and Drying Examples 43 to 46 all use the two-step polishing conditions of Example 40, the difference being the cleaning solution after polishing, the cleaning time, and the drying method.
[0093] Example 43 After polishing, the surface was rinsed with deionized water for 30 seconds and then dried using a rotary dryer. Slight residue remained in the air gaps after treatment, with a metal ion residue level of 8.2 × 10⁻⁶. 10 atoms / cm 2 .
[0094] Example 44 In this embodiment, deionized water was used for 60 seconds, followed by rotary drying. Slight residue remained in the air gaps after treatment, with a metal ion residue level of 5.6 × 10⁻⁶. 10 atoms / cm 2 .
[0095] Example 45 In this embodiment, a dilute ammonia solution (ammonia to water volume ratio of 1:200) was used for cleaning for 45 seconds, followed by drying with isopropanol (IPA) steam. No air gap residue was observed after treatment, and the residual metal ion content was 1.2 × 10⁻⁶. 10 atoms / cm 2 .
[0096] Example 46 In this embodiment, a dilute ammonia solution (ammonia to water volume ratio of 1:100) was used for cleaning for 60 seconds, followed by drying with isopropanol vapor. No air gap residue was observed after treatment, and the residual metal ion content was 0.8 × 10⁻⁶. 10 atoms / cm 2 .
[0097] Comparative Example 1 The difference between this comparative example and Example 5 is that no catalytic activator is added to the polishing solution, while the remaining components and polishing conditions are the same as in Example 5.
[0098] The Ru removal rate was tested to be 180. / min, which is significantly lower than that of Example 5, and is difficult to meet the material removal rate requirements during the planarization of ruthenium interconnect structure.
[0099] Comparative Example 2 The difference between this comparative example and Example 5 is that no air gap protectant is added to the polishing slurry, while the remaining components and polishing conditions are the same as in Example 5.
[0100] The Ru removal rate was tested to be 1150. The polishing fluid achieved a Ru removal rate of 1000 m / min, but the air gap residue rate reached 38%, and severe voids appeared at the seal layer interface. This indicates that, while a high Ru removal rate can be maintained without adding an air gap protectant, the polishing fluid and its particles easily leave residues within the air gaps.
[0101] Comparative Example 3 In this comparative example, the content of grinding particles was set to 2.0 wt%.
[0102] The Ru removal rate was tested to be 1320. / min, the proportion of air gap structural damage is 12%, and the scratch defect density reaches 85 per cm. 2 The results showed that while increasing the content of abrasive particles could improve the Ru removal rate, it also increased the mechanical action on the air gap structure, leading to increased structural damage and surface scratch defects.
[0103] Comparative Example 4 In this comparative example, the hydrogen peroxide content was set to 3.0 wt%.
[0104] The Ru removal rate was tested to be 1180. At a rate of 210 pitting defects per minute, the polished ruthenium surface exhibited significant pitting corrosion. 2 Simultaneously, RuO4 escape occurred. The results indicate that although increasing the hydrogen peroxide content can maintain a high Ru removal rate, it increases ruthenium surface corrosion and defects.
[0105] Comparative Example 5 In this comparative example, the pH of the polishing solution was adjusted to 10.5.
[0106] The Ru removal rate was tested to be 720. / min, the corrosion rate in low-k media is 42 The polishing solution's pH increased by 10.5, which was detrimental to achieving a balance between Ru removal, low-k medium protection, and air gap residue control. Simultaneously, the protective effect of the air gap protectant decreased. The results indicate that increasing the polishing solution pH to 10.5 is not conducive to simultaneously achieving Ru removal, low-k medium protection, and air gap residue control.
[0107] Comparative Example 6 In this comparative example, the pH of the polishing solution was adjusted to 6.0.
[0108] The Ru removal rate was tested to be 580. The ruthenium removal rate decreased by 1 / min, while the complexing effect of citric acid decreased. The results indicate that the ruthenium removal capacity decreases when the pH of the polishing solution drops to 6.0.
[0109] Comparative Example 7 In this comparative example, the polishing pressure was set to 1.5 psi.
[0110] The Ru removal rate was tested to be 1280. The polishing pressure was increased by 15% per minute, but the air gap height collapsed by 15%, and the air gap structure damage rate reached 8.5%. The results show that increasing the polishing pressure can increase the Ru removal rate, but it will significantly increase the degree of mechanical damage to the air gap structure.
[0111] Comparative Example 8 The difference between this comparative example and Example 5 is that no ruthenium complexing agent is added to the polishing solution, while the remaining components and polishing conditions are the same as in Example 5.
[0112] The Ru removal rate was tested to be 1120. The surface roughness Ra of the polished ruthenium was 0.78 nm, with significant ruthenium redeposition observed at a rate of 0.78 nm / min. These results indicate that although a certain Ru removal rate can be maintained without the addition of a ruthenium complexing agent, oxidized ruthenium species are prone to redeposition on the polished surface.
[0113] Test Example 1 To investigate the effect of the air gap protectant on the residue of polishing slurry inside the air gap, air gap residue tests were conducted on Examples 26-32.
[0114] Cross-sectional samples were prepared from the polished patterned wafer along a direction perpendicular to the ruthenium interconnect wires. Scanning electron microscopy combined with image analysis was used to determine the residue within the air gaps. At least 10 air gaps were randomly selected from each sample, and the total cross-sectional area St and the residue coverage area Sr of each air gap were measured. The air gap residue rate was calculated using the following formula: Air gap residual rate (%) = Sr / St × 100%.
[0115] In the formula, Sr is the coverage area of the residue in the air gap section, St is the total area of the air gap section, and the final test result is the arithmetic mean of no less than 10 air gap residue rates.
[0116] The test results are shown in Table 1.
[0117] Table 1 Test results of air gap protectant
[0118] As shown in Table 1, the air gap residual rate was 38% when no air gap protectant was added. After adding PEG600, the surface tension of the polishing slurry and the air gap residual rate were both reduced. In Example 29, after adding 0.05 wt% PEG600, the surface tension was 32 mN / m, the air gap residual rate was 2%, and no voids appeared at the sealing interface.
[0119] Test Example 2 To investigate the roles of ammonium persulfate and citric acid in the ruthenium removal process, X-ray photoelectron spectroscopy and ruthenium ion concentration tests were performed, respectively.
[0120] Ruthenium thin film samples were placed in test solutions of different compositions for 60 seconds. The test solutions were free of abrasive particles. After treatment, the samples were rinsed with deionized water and dried before XPS testing. The test used an Al Kα ray source with a photon energy hν of 1486.6 eV and a passing energy of 20 eV. Charge correction was performed using the 284.8 eV peak of C 1s.
[0121] The test results are shown in Table 2.
[0122] Table 2 Results of Ruthenium Surface Oxidation State Test
[0123] As shown in Table 2, when only 1.0 wt% hydrogen peroxide is used, the RuO2 / Ru peak area ratio is 0.15; after adding 0.10 wt% ammonium persulfate, the RuO2 / Ru peak area ratio increases to 1.85, indicating that ammonium persulfate can promote the oxidation of ruthenium surface; after further adding 0.20 wt% citric acid, the RuO2 / Ru peak area ratio is 1.80, which is close to that when no citric acid is added.
[0124] Further, the polished ruthenium wafers of Comparative Example 8 and Example 16 were placed in deionized water for ultrasonic cleaning for 5 minutes. The cleaning solution was collected, and the Ru ion concentration was detected by ICP-MS. The results are shown in Table 3.
[0125] Table 3. Test results of Ru ion concentration in cleaning solution under different formulation conditions
[0126] As shown in Table 3, the Ru ion concentration in the cleaning solution of Comparative Example 8 was 2.3 ppb, while the Ru ion concentration in the cleaning solution of Example 16 was 185 ppb. The higher Ru ion concentration detected in Example 16, which used a polishing system containing 0.20 wt% citric acid, indicates that more ruthenium species formed by oxidation could enter the liquid phase in this polishing system.
[0127] The XPS test results in Table 2 show that ammonium persulfate promotes the oxidation of ruthenium surfaces; the polishing system containing citric acid facilitates the entry of oxidized ruthenium species into the liquid phase. Therefore, the combination of the oxidation effect of ammonium persulfate on the ruthenium surface and the complexation effect of citric acid on ruthenium species is beneficial for ruthenium removal under relatively low mechanical action conditions.
[0128] Test Example 3 To investigate the effect of polishing pressure on the air gap structure and ruthenium removal performance, a polishing solution composed of 0.3 wt% colloidal silica, 1.0 wt% hydrogen peroxide, 0.15 wt% ammonium persulfate, 0.20 wt% citric acid and 0.05 wt% PEG600 was used. The pH was adjusted to 8.5, and tests were conducted according to the conditions described in Examples 33 to 37.
[0129] The test results are shown in Table 4.
[0130] Table 4. Effect of polishing pressure on air gap structure
[0131] As shown in Table 4, under conditions of 0.5–1.0 psi, the Ru removal rate was 832–1150. / min, the change in air gap height does not exceed 2%, and the proportion of structural damage does not exceed 0.5%; when the polishing pressure is increased to 1.2 psi, the change in air gap height increases to 8%, and the proportion of structural damage increases to 3.2%.
[0132] Further, two-step process tests, coarse polishing and fine polishing, were performed according to the conditions described in Examples 38-42. The initial SiCN encapsulation layer thickness of the wafer was 500 μm. The initial surface height difference was 40 nm, the coarse polishing time was 45 seconds, and the fine polishing time was 20 seconds. The results are shown in Table 5.
[0133] Table 5 Results of Two-Step Low-Stress Polishing Test
[0134] The flatness test area is 30 μm × 30 μm.
[0135] As shown in Table 5, in Example 40, when using 0.8 psi for rough polishing and 0.5 psi for fine polishing, the flatness was 4.8 nm, the surface roughness Ra was 0.32 nm, and the remaining SiCN thickness was 285 nm. The yield rate was 97.8%. Although further increasing the polishing pressure could further reduce the flatness and surface roughness values, the remaining SiCN thickness and yield rate decreased accordingly.
[0136] Therefore, a low-stress polishing method that combines coarse polishing and fine polishing can simultaneously achieve ruthenium material removal, surface planarization, and protection of the air gap structure.
[0137] Test Example 4 To investigate the effect of cleaning and drying methods on residues after polishing, the wafer was treated with the two-step polishing conditions of Example 40, and different cleaning and drying treatments were performed according to Examples 43-46. The test results are shown in Table 6.
[0138] Table 6. Results of Cleaning and Drying Tests
[0139] Table 6 shows that slight air gap residue remained after cleaning with deionized water and rotary drying; however, after cleaning with dilute ammonia water combined with IPA steam drying, no air gap residue was observed, and the metal ion residue was reduced. In Example 45, cleaning with dilute ammonia water (1:200 volume ratio) for 45 seconds followed by IPA steam drying resulted in a metal ion residue of 1.2 × 10⁻⁶. 10 atoms / cm 2 .
[0140] Therefore, it can be seen that cleaning with dilute ammonia water and combining it with IPA steam drying after two-step low-stress polishing is beneficial to further reduce the residual air gaps and metal ions after polishing.
[0141] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in the present invention, and these should all be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.
Claims
1. A low-stress polishing slurry for ruthenium interconnect air gaps, characterized in that, The polishing solution comprises, by weight 100%, 0.2–0.5 wt% colloidal silica, 0.8–1.2 wt% hydrogen peroxide, 0.08–0.20 wt% ammonium persulfate, 0.15–0.30 wt% citric acid, 0.03–0.08 wt% polyethylene glycol 600, a pH adjuster, and deionized water. The pH adjuster adjusts the pH of the polishing solution to 8.0–9.0, and the deionized water is the balance.
2. The low-stress polishing slurry for ruthenium interconnect air gaps according to claim 1, characterized in that, The average particle size of the colloidal silica is 30–50 nm.
3. The low-stress polishing slurry for ruthenium interconnect air gaps according to claim 1, characterized in that, The pH adjuster is tetramethylammonium hydroxide.
4. The low-stress polishing slurry for ruthenium interconnect air gaps according to claim 1, characterized in that, Based on the total weight of the polishing slurry (100%), the content of colloidal silica is 0.3 wt%, the content of hydrogen peroxide is 1.0 wt%, the content of ammonium persulfate is 0.10–0.15 wt%, the content of citric acid is 0.20 wt%, the content of polyethylene glycol 600 is 0.05 wt%, and the pH of the polishing slurry is 8.
5.
5. A low-stress polishing method for ruthenium interconnect air gaps, characterized in that, Includes the following steps: S1. A semiconductor substrate is provided, on which ruthenium interconnect wires and air gaps are formed, and an encapsulation layer is covered over the air gaps, wherein the encapsulation layer does not completely seal the air gaps. S2. Rough polishing is performed using a low-stress polishing slurry. The low-stress polishing slurry, by weight (100%), comprises 0.2–0.5 wt% colloidal silica, 0.8–1.2 wt% hydrogen peroxide, 0.08–0.20 wt% ammonium persulfate, 0.15–0.30 wt% citric acid, 0.03–0.08 wt% polyethylene glycol 600, a pH adjuster, and deionized water. The pH adjuster adjusts the pH of the low-stress polishing slurry to 8.0–9.
0. The deionized water is the balance. The polishing pressure for rough polishing is 0.5–1.0 psi, and the polishing time is 30–60 seconds. S3. Fine polishing is performed using the aforementioned low-stress polishing fluid. The polishing pressure for fine polishing is 0.3–0.6 psi, and the polishing time is 15–30 seconds. S4. Rinse with deionized water or dilute ammonia, then dry with isopropanol steam.
6. The low-stress polishing method for ruthenium interconnect air gaps according to claim 5, characterized in that, In step S2, the polishing pad rotation speed for coarse polishing is 60–90 rpm, the wafer rotation speed is 55–85 rpm, and the polishing slurry flow rate is 150–250 mL / min.
7. The low-stress polishing method for ruthenium interconnect air gaps according to claim 5, characterized in that, In step S3, the polishing pad rotation speed is 40-60 rpm, the wafer rotation speed is 35-55 rpm, and the polishing slurry flow rate is 100-150 mL / min.
8. The low-stress polishing method for ruthenium interconnect air gaps according to claim 5, characterized in that, In step S2, the polishing pressure for rough polishing is 0.8 to 1.0 psi; in step S3, the polishing pressure for fine polishing is 0.4 to 0.6 psi.
9. The low-stress polishing method for ruthenium interconnect air gaps according to claim 5, characterized in that, In step S4, the water is cleaned with dilute ammonia for 45 to 60 seconds, wherein the volume ratio of ammonia to water in the dilute ammonia is 1:150 to 250. After cleaning, the water is dried with isopropanol vapor.
10. The low-stress polishing method for ruthenium interconnect air gaps according to claim 5, characterized in that, In step S1, the sealing layer material is SiCN or SiCOH, and the height difference of the sealing layer surface is no greater than 50 nm.
