WO2008023215A1

WO2008023215A1 - Post chemical mechanical polishing rinse formulation

Info

Publication number: WO2008023215A1
Application number: PCT/IB2006/003085
Authority: WO
Inventors: Janos Farkas; Maria-Luisa Calvo-Munoz; Sebastien Petitdidier
Original assignee: STMicroelectronics Crolles 2 SAS; Freescale Semiconductor Inc
Current assignee: STMicroelectronics Crolles 2 SAS; NXP USA Inc
Priority date: 2006-08-23
Filing date: 2006-08-23
Publication date: 2008-02-28
Anticipated expiration: 2009-02-23

Abstract

A post-CMP rinse formulation for use in semiconductor processing, characterised in that the formulation is free or essentially free of triazole species and contains: (i) a surfactant comprising a polyalkylene glycol and/or polyoxymethylene (POM); and (ii) a complexing agent selected from one or more of ammonia, EDTA, EDDHA and an organic acid, including salts and derivatives thereof.

Description

Post Chemical Mechanical Polishing Rinse Formulation

Field of the Invention

The present invention relates to solutions for use in cleaning the surface of a substrate for use in a semiconductor device, in particular for cleaning the surface after it has been subjected to chemical mechanical polishing.

Background Art

Integrated circuit manufacturers are constantly striving to increase the amount of features on integrated circuits and hence the size of the features becomes smaller in each new generation of circuit. In recent years, lithographically patterned and etched aluminium interconnects have been replaced with copper interconnects. These copper interconnects are often laid in methods such as the dual damascene process. Copper offers many advantages over aluminium, including a greater conductivity and less of a tendency to electromigrate . However, copper has presented new problems to manufacturers. For instance, it has been found that copper has a tendency to diffuse into the dielectric material at the copper/dielectric material interface. In order to prevent this diffusion, manufacturers will deposit a barrier layer onto the dielectric prior to deposition of the copper. The optimum barrier layer material comprises tantalum or tantalum compounds. Following deposit of the barrier layer, a thin copper seed layer (of the order of a hundred nanometers thick) is then deposited, allowing a thicker copper layer to be electrodeposited onto the seed layer. The excess copper is then removed from the surface using chemical mechanical polishing and the surface is planarized.

However, particles of material removed from the surface and contained in the polishing chemical during chemical mechanical polishing (CMP) processes are a huge source of contamination of the substrate surface. The surface should ideally be particle-free before the substrate is subjected to any other processing steps. Chemical rinses before, during and after the CMP process are commonly used to remove contaminant particles, and therefore reduce defect formation in subsequent steps and also improve the yield of semiconductor devices.

The removal of contamination from a wafer surface during or after a CMP process involves either application of an external mechanical force to the contaminant particles in order to overcome their adhesion to the substrate surface or the use of a mild surface etching agent to release the particles from the substrate surface. Alternatively, the probability of adhesion from contaminating particles can be reduced by using a cleaning solution which generates oppositely charged surfaces between the semiconductor substrate and the contaminating particles. If additives are used in the CMP or rinse processes, they must be selected such that they are compatible with the materials on the wafer, i.e. without causing excessive removal or corrosion of the dielectric, copper or barrier materials.

Especially for the CMP treatment of a copper inlaid dielectric substrate, the chemical rinses normally include organic copper corrosion inhibitors, which form an organic passivation layer at the copper surface to prevent or significantly reduce any corrosion of the copper surface.

These corrosion inhibitors are also often present in the polishing slurries to prevent corrosion of the copper during polishing, or to adjust polishing rate of the substrate surface.

Corrosion inhibitors are well known in the art. They are incorporated to prevent the corrosion of the material while the surface is being stored or treated. Although not wishing to be limited by theory, the corrosion inhibitors are thought to absorb onto the surface of the material and form a thin protective film. This may protect the surface from corrosion. For example, it may act as a passivation layer when the surface is stored in an aqueous environment between processing steps. In addition, it may act as a passivation layer to corrosion involving oxygen or moisture

Corrosion inhibitors are often aromatic compounds. They often have polarizable pi-systems, and this perhaps favours the interaction of the corrosion inhibitor with the surface. They are sometimes substituted with substituents capable of hydrogen-bonding, perhaps enabling the self- assembly of the corrosion inhibitor on the surface. , Examples of corrosion inhibitors include substituted- benzenes, for example substituted by hydroxyl groups, such as in pyrogallol. Other corrosion inhibitors include quinones, such as hydroquinone . Finally, a widely used class of corrosion inhibitors is the triazoles. These include 1, 2 , 4-triazole, benzotriazole (BTA) and tolytriazole .

US-B-6443814 discloses the use of post-CMP cleaning formulations, which are buffered by organic acid buffer systems and have a pH of 2.5 to 7.5. The post-CMP cleaning formulations also contain a corrosion protection agent such as benzotriazole .

US-B-6444569 and US-B-6464568 disclose rinse steps after subjecting a substrate having a copper surface to CMP. These documents suggest the use of a corrosion protecting agent in a cleaning solution following CMP treatment.

US Patent Application No. 2004/0014319 discloses the use of an aqueous slurry containing various organic additives to treat the surface of a wafer during and/or after chemical mechanical polishing. The use of the disclosed slurry reduces the amount of defects, which is believed to originate from the reduction of silica and/or carbon precipitates on the surface due to the reduction of surface reactivity of copper and/or silicon dioxide by organic additives. The amount of organic additives in the slurry is from 0.1-2 wt% solution. This lower concentration is about 10⁵ times higher than the amount which would be expected to form an efficient passivating monolayer (assuming approximately 1000 molecular weight/nm²) . The high amount of organic additives used seems to indicate that they are far from optimal.

US Patent Application No. 2004/0014319 also teaches the use of an organic corrosion inhibitor in the slurry, which has several disadvantages. These will be further detailed below.

The present inventors have found that there are drawbacks with the rinse formulations of the prior art. The organic corrosion inhibitors, such as the triazoles, have been found to form insoluble globular deposits on the surface of the substrate. These insoluble globular deposits have been found to form particularly when the substrate is rinsed with water following an initial rinse with the rinse solution containing the triazole. In addition, the majority of these products are classified as carcinogenic, mutagenic, and present reproductive hazards. It would be desirable to create rinse formulations free of these products, particularly those formed from triazole-containing solutions, for environmental, safety and health reasons.

Additionally, the present inventors have found that the corrosion inhibitors of the prior art are not as effective at preventing corrosion of a copper surface as one would hope under certain conditions. For instance, electrodeposited copper features on a silicon wafer were found to corrode when the wafer is dipped into an aqueous solution of 1,2,4 triazole (3 % by weight) and polyethylene glycol (0.3 % by weight) for 1 hour under ambient light conditions .

The present inventors have found that many rinse solutions of the prior art also leave significant amounts of residual ions (including sulphur-containing, chloride, potassium, calcium, iron, copper and zinc ions) on the surface of the copper. These are believed to have — o -

detrimental effects on dielectric reliability, and surface leakage .

Recent advances in the field have led to porous dielectric materials being used in semiconductor devices. The use of porous materials reduces the dielectric constant between the metal lines, which results in smaller RC time constant allowing faster devices. However, layers of porous dielectric materials have been found to easily delaminate during and after CMP processes. It has been found that the nature of the rinse formulation used following a CMP process can have a large impact on whether or not the de-lamination occurs after the CMP process. It would therefore be desirable to produce a post-CMP rinse formulation that mitigates or prevents delamination of a semi-conductor device containing a porous dielectric material .

Summary of the Invention

The present invention provides, in a first aspect, a post-CMP rinse formulation for use in semiconductor processing, characterised in that the formulation is free or essentially free of triazole species and contains:

(i) a surfactant comprising a polyalkylene glycol and/or polyoxymethylene (POM) ; and

(ii) a complexing agent selected from one or more of ammonia, EDTA, EDDHA and an organic acid, including salts and derivatives thereof.

The present invention also provides, in a second aspect, a process for cleaning the surface of a substrate for use in semiconductor processing, the process comprising: a step (A) of contacting the surface of the substrate with the rinse formulation of the present invention.

Brief Description of the Drawings

Figure 1 shows the SPl defect density of 200 mm blanket

1.5 μm copper wafers having been first subjected to CMP treatment and then rinsed with one of the comparative examples (solutions A and B, as detailed in the Examples section below) or an embodiment of the formulations of the present invention (solutions C to E, as detailed in the

Examples section below) .

Figure 2 shows the SPl defect density of 300 mm copper wafers having surface layers of Cu/Ta/TaN, the wafers having been subjected to a CMP treatment and then rinsed with a comparative example (solution A) or an embodiment of the present invention (solution C) .

Figure 3 shows the normalised defect density of wafers having been subjected to a CMP treatment and then rinsed with a comparative example (solution A) or an embodiment of the present invention (solution C) , the defect density having been analysed by SEM.

Figure 4 shows the contamination of a 1000 A TEOS wafer with various elements and ions having been first subjected to a CMP treatment and then rinsed with a rinse solution

(either one of the comparative examples, solutions A and B, or embodiments of the present invention, solutions C to E) .

For each element/ion, the results for solutions A to E are shown from left to right . Figure 5 (a) shows the corrosion of copper on a wafer when immersed in an aqueous rinse formulation of the prior art (solution A) for period of 1 hour and exposed to ambient fluorescent light in a laboratory. The portion of the wafer within the oval shows corroded copper.

Figure 5 (b) shows the same type of wafer as used in Fig 5 (a) immersed in solution D for 1 hour and exposed to ambient fluorescent light in a laboratory. No corrosion of the copper was observed.

Figure 5 (c) shows the same type of wafer as used in Fig 5 (a) immersed in solution E for 1 hour and exposed to ambient fluorescent light in a laboratory. No corrosion of the copper was observed.

Detailed Description of the Invention

The present invention provides, in a first aspect, a post-CMP rinse formulation for use in semiconductor processing, characterised in that the formulation is free or essentially free of triazole species and contains :

The polyalkylene glycol may comprise polyethylene glycol and/or polypropylene glycol . The polyalkylene glycol may comprise a block co-polymer of polyethylene glycol and polypropylene glycol . Examples of such co-polymers include those produced by BASF under the tradename Pluronics and Tetronics. Co-polymers of polyethylene glycol and polypropylene glycol are known in the art and those used in the present invention may have the formula EO-PO-EO or PO- EO-PO, in which EO represents one or more repeating ethylene oxide units and PO represents one or more repeating polyethylene oxide units. The co-polymers are generally hydroxyl terminated. The co-polymers may have a molecular weight of from 1000 to 25000 g/mol, preferably from 2000 to 3500 g/mol, and preferably contain from 10 to 80 weight % ethylene oxide units.

The liquid may comprise water, such as deionised water. The liquid may comprise an alcohol, advantageously an alkyl alcohol, most advantageously ethanol and/or isopropanol . ,

The surfactant may be present in the liquid in a concentration of from 0.0001 wt . % to 3wt.%, advantageously of from 0.001 wt . % to 0.5 wt.%, still more advantageously 0.2 to 0.4 wt%, most advantageously about 0.3 wt%.

The complexing agent may be present in the liquid in a concentration of from 0.001wt% to 0.1wt%, advantageously of from 0.005wt% to 0.01wt%.

The ratio of the concentration of surfactant to the concentration of complexing agent in the liquid may be in the range of from 1:0.01 to 1:100.

The organic acid may be a carboxylic acid. The organic acid may comprise, for example, one or more of citric, oxalic, glycolic, malic and succinic acid. Alternatively the organic acid also could be ascorbic acid and gallic acids. It is believed that these acids can perform one or more functions in the formulation. Firstly, they can act as acidifiers for the solution. Secondly, they can act as complexing agents. For example it is well known that Fe, Cu₅Ca, Ni, K and Al ions can be effectively coraplexed by these organic acids. Thirdly, these acids can act as oxygen- scavengers or reducing agents.

The liquid may also contain tetramethyl-ammonium hydroxide .

If the formulation has a pH of more than 7, it may comprise one or more of ascorbic acid, EDTA and EDDHA, or salts and derivatives thereof ;

If the formulation has a pH of less than 7, it may comprise an organic acid such as described above.

The formulation is free or essentially free of triazole species. Such triazole species include, but are not limited to, triazole, benzotriazole, tetrazole and/or imidazole species. The formulation may be free or essentially free of 1, 2 , 4-triazole, benzotriazole and/or tolytriazole . "Free" of a triazole species indicates that no triazole species is detectable in the solution. Essentially free includes, but is not limited to, an amount of triazole less than 0.0001 % by weight, preferably less than 0.00001 % by weight in the formulation.

The present inventors have found that the rinse formulation of the present invention used after CuCMP provides similar, if not better, corrosion protection of a copper surface compared to a rinse formulation containing a triazole species. Additionally, at least some of the disadvantages mentioned above with reference to the prior art are overcome by use of the formulation of the present invention. For instance, few, if any, globular deposits are found on a copper surface following a rinse with the formulation of the present invention.

The present invention provides, in a second aspect, a process for cleaning the surface of a substrate for use in semiconductor processing, the process comprising: a step (A) of contacting the surface of the substrate with the rinse formulation of the present invention.

The substrate may comprise one or more layers of a dielectric material. The substrate may comprise silicon. The substrate may comprise a porous dielectric material. The substrate may comprise one or more layers of a dielectric material, upon at least one of which deposited a barrier layer, upon which is deposited a conducting layer comprising copper or a copper alloy. The barrier layer may contain a material such as Ta and/or TaN. The barrier ^■ layers may have a thickness of from 50 A to 200 A. The barrier layer may have a thickness of from 1000 A to 5000 A. An example of a typical conducting/barrier layer construction is: Cu/Cu/Ta/TaN with respective thicknesses of 4500 A /1500 A / 150 A /lOOA.

The process may further comprise, prior to step (A) , a step (B) comprising subjecting the substrate to chemical mechanical polishing.

The process may further comprise, between step (B) and step (A) , a step (C) comprising rinsing the surface of the substrate with deionised water. The rinse formulation may be contacted with the surface by means of spraying the formulation onto the surface or dipping the substrate into the formulation.

The process may further comprise removing the rinse formulation from the surface following step (A) .

Embodiments of the present invention will be illustrated further with reference to the following non- limiting Examples.

Examples

"Blanket Wafer Defectivity 200 nun data"

Solutions of deionised water were prepared containing the following proportions of additives:

Solution A (comparative example): PEG (0.3%), 1,2,4 triazole (3%), NH₄OH; pH 8.5 Solution B (comparative example) : 1, 2 , 3-benzotriazole (BTA) 0.002%, pluronic RPE2520 0.01%, citric acid 0.5%, NH₄OH; pH 3.8

Solution C (according to the present invention) : PEG (0.3%), NH₄OH, pH 8.5; the complexing agent in this solution is ammonium ions (NH₄ ⁺)

Solution D (according to the present invention) : as solution C and also containing EDTA (0.005%) and NH₄OH so that the pH of this solution was 8.8

Solution E (according to the present invention) : as solution C and also containing EDTA (0.01%) and NH₄OH so that the pH of this solution was 9.2

The wafers used for this experiment were blanket 1.5μm copper wafers annealed at 400⁰C. The wafers were polished on a Mirra polisher for 60s on a first polishing platen with the slurry EPL2361 from Eternal company and for 60s on a second platen with the slurry CuS1351 from Rodel company. After polishing on the second platen, the wafer was subjected to a rinse with one of the solutions A to E for 15s with a flow rate of 200mL/min, followed by a high pressure de-ionised water (DIW) rinse of 15s. The wafers were then scrubbed on a scrubber from LAM company with the chemical ESC784 from ATMI company and DIW. The defect density at the surface of the wafers was then measured by a particle counter SPl from KLA Tencor, which is a surface inspection technique based on a wide light scattering.

The results of this test are shown in Figure 1. This graph shows that the solutions C to E cause significantly less defects than solutions A and B. In the Figure, "DD" is defect density and the units are number of defects/cm². The threshold of the SPl recipe used for the measurement is

0.2μm (the minimum particle size measured is 0.2μm) . Each solution was used to rinse two identical wafers, as described above, and the result for each wafer is shown in the graph .

"Defect density Do and Scratches" Solutions A and C were prepared containing the constituents indicated above.

This test was performed on a 300 mm tool (Reflexion from AMAT) . The blanket Cu wafers were polished for 60s respectively on the first and the second platen. This was followed by a 15s rinse with the solution A or solution C and a 15s DIW rinse on platen 3. This was followed by a clean with ESC784 and DIW on a scrubber.

The copper bare wafers were Iμm Cu film, grown by electroplating and then annealed in a furnace at 400⁰C during 20 min under N₂ atmosphere at 50 mTorr . A three layer Cu/Ta/TaN structure with thickness 1500 A / 150 A / IOOA was sputtered on a reclaimed p-type Si(IOO) with pre-deposited IOOOA TEOS oxide before plating. Defectivity was measured with the SPl particle counter mentioned above in the same manner, with a threshold of 0.2μm.

The results of the defect density for wafers rinsed with solutions A and C are shown in Figure 2. The defect density for solution C is much smaller than solution A.

"SEM classification"

120 nm patterned wafers were polished and rinsed with solutions A and C. Subsequently the wafers were inspected by a brighfield KLA 2361 inspection tools. 50 of detected defects were then reviewed and classified on an SEM review station. The results in the graph of Figure 3 show that wafers rinsed with solution C have a lower defect density than those rinsed with solution A. Solution C creates no globular carbon defects and no corrosion-type defects can be seen.

"Corrosion Protection"

Wafers using 90nm technology node, having copper features on their surface were immersed in solution A, solution D and solution E for 1 hour at 25⁰C and exposed to laboratory ambient light during this period. The wafers in solution D (as shown in Figure 5 (b) ) and solution E (as shown in Figure 5 (c) ) showed little or no corrosion of the copper features on the wafers compared to significant corrosion of the copper features of the wafer in solution A (as shown in Figure 5 (a) ) . "Metallic and Non-metallic Ionic Residuals"

1000 A TEOS oxide wafers were polished on the Mirra polisher for 15s on the third platen with the slurry CUS1351 from Rodel company followed by a 15s rinse with one of solutions A to E and a 15s DIW rinse. This was followed by a clean with ESC784 and DIW on a scrubber.

The metallic contamination was measured by Total X-Ray Fluorescence (TXRF) . The measurements were performed with a tungsten X-ray source (35kV - 400 mA) . Five points were collected with an acquisition integration time of 500s. Figure 4 shows levels of metallic contamination for certain metals. The values reported on the graph are the average of five points .

Claims

CLAIMS :

1. A post-CMP rinse formulation for use in semiconductor processing, characterised in that the formulation is free or essentially free of triazole species and contains:

2. A formulation as claimed in claim 1, wherein the polyalkylene glycol comprises one or more of polyethylene glycol, polypropylene glycol and a block co-polymer of polyethylene glycol and polypropylene glycol .

3. A formulation as claimed in claim 1 or claim 2, wherein the liquid comprises water, ethanol and/or isopropanol .

4. A formulation as claimed in any one of the preceding claims, wherein the surfactant is present in the liquid in a concentration of from 0.2% to 0.4%.

5. A formulation as claimed in claim 4, wherein the surfactant is present in the liquid in a concentration of about 0.3%.

6. A formulation as claimed in any one of the preceding claims, wherein the complexing agent is present in the liquid in a concentration of from 0.001% to 0.01%.

7. A formulation as claimed in claim 6, wherein the complexing agent is present in the liquid in a concentration of from 0.005% to 0.01%.

8. A formulation as claimed in anyone of the preceding claims, wherein the organic acid is selected from one or more of citric acid, oxalic acid, glycolic acid, malic acid, succinic acid, ascorbic acid and gallic acid.

9. A formulation as claimed in any one of the preceding claims further containing tetramethyl-ammonium hydroxide.

10. A formulation as claimed in any one of the preceding claims having a pH of more than 7 and comprising one or more of EDTA and EDDHA, or salts and derivatives thereof.

11. A formulation as claimed in any one of claims 1 to 9 having a pH less than 7 and comprising the organic acid.

12. A formulation as claimed in any one of the preceding claims, wherein the formulation is free or essentially free of 1, 2 , 4-triazole, benzotriazole and tolytriazole .

13. A process for cleaning the surface of a substrate for use in semiconductor processing, the process comprising: a step (A) of contacting the surface of the substrate with the rinse formulation as defined in any one of claims 1 to 12.

14. A process as claimed in claim 13, wherein the process further comprises, prior to step (A) , a step (B) comprising subjecting the substrate to chemical mechanical polishing.

15. A process as claimed in claim 13 or 14, wherein the process further comprises, between step (B) and step (A) , a step (C) comprising rinsing the surface of the substrate with deionised water.

16. A process as claimed in any one of claims 12 to 15, wherein the rinse formulation is contacted with the surface by means of spraying the formulation onto the surface or dipping the substrate into the formulation.

17. A process as claimed in any one of claims 12 to 16, wherein the process further comprises removing the rinse formulation from the surface following step (A) .

18. A process as claimed in any one of claims 12 to 17, wherein the substrate comprises one or more layers of a dielectric material, upon at least one of which is deposited a barrier layer, upon which is deposited a conducting layer comprising copper or a copper alloy.

19. A process as claimed in anyone of claims 12 to 18, wherein the substrate comprises a porous dielectric material .