Evaluating and Treating CMP Wastewater

Josh H. Golden, Microbar Inc., Sunnyvale, Calif. Robert Small, Louis Pagan, Cass Shang, EKC Technology Inc., Hayward, Calif. Srini Ragavan, University of Arizona, Tucson, Ariz. -- Semiconductor International, 10/1/2000

		At a Glance
	With the transition to copper interconnect material comes concern over the toxicity and necessary treatment of copper in CMP wastewater. Examination of copper chemistry in wastewater reveals mostly insoluble copper oxides and hydroxides, except atpH<7 where cupric ion (Cu2+) occurs in solution. A simple mass balance on a CMP polisher revealed a copper concentration in the fab's final effluent of ~0.5 ppm. We review methods for treating and removing copper from wastewater, including coagulation and flocculation as well as different methods of filtration. Microfiltration results compare favorably with more mature processes, while offering higher throughput (5000 gal/min of effluent) in a smaller footprint.

Device manufacturing at 180 nm and beyond relies heavily on chemical-mechanical polishing (CMP) to reduce wafer topological imperfections and improve the depth of focus of lithography processes through better planarity. As a result, the CMP market is growing faster than other equipment sector, and is expected to reach $1.7 billion by the year 2003 and $5 billion by 2010.¹The increasing use of CMP slurry also drives the demand for ultrapure water (UPW) for slurry dilution, cleaning and rinsing wafers and production equipment. Until recently, the CMP process accounted for only about 5% of the total UPW used in IC manufacturing,² but estimates at leading-edge fabs put the usage in 2001 at 30-40% of the water consumed.² Consequently, CMP wastewater generation is expected to reach 450 million gallons by the year 2006,² and this effluent will contain increasing amounts of copper. However, this increased use of UPW by CMP processes conflicts with the NTRS's target of decreasing overall water consumption five-fold, from today's ~1500 gal/200 mm wafer to 300 gal/200 mm wafer in 2003, based on increased adoption of recycling and conservation.³

Before the recent introduction of copper CMP, the CMP waste stream was not a major issue. Silica and fluoride contaminants were typically diluted with other fab process water or treated by the fab acid waste neutralization system (AWNS). However, because the AWNS is not equipped to handle high levels of suspended solids or especially, heavy metals, fabs are beginning to run the risk of violating local regulations for these contaminants in discharged wastewater. To effectively meet regulatory challenges and safely treat CMP wastewater, it is useful to become familiar with the CMP wastewater chemistry, treatment options and regulatory issues regarding copper in CMP effluent.

CMP wastewater chemistry

CMP process effluent contains inorganic and organic contaminants (Table 1) derived from the slurry, the wafer, planarization and post-CMP cleaning processes. Contaminant solubility depends on several factors including pH and the oxidation-reduction potential of the solution.

Organic materials found in CMP effluent include metal complexing agents, surfactants, stabilizers and rheology control agents whose solubility is usually pH dependent. For example, benzotriazole (BTA), an aromatic amine and corrosion inhibitor used to regulate the rate of copper removal in copper CMP,^4-6 forms a soluble copper complex and acid salt at low pH, but becomes insoluble at pH ~5. In contrast, triethanolamine, an aliphatic dipersant and complexing agent, remains water-soluble over a wide pH range, making it difficult to remove by pH adjustment alone. Dispersants, stabilizers and surfactants including poly(acrylic acid) and aliphatic quaternary ammonia salts, are used to create metastable homogeneous suspensions of CMP slurry particles. Significantly, these materials can negatively affect coagulation and flocculation treatment schemes, so careful attention to treatment chemistry is required.

Table 1. Materials in CMP Wastewater

Inorganic Materials

Interconnect: Cu2+, complexed Cu2+, Cu2O, CuO, Cu(OH)2, WO3, Al2O3, Al(OH)3, Fe2+/Fe3+

Barrier/liner: Tantalum and titanium oxides and oxynitrides

Abrasives: SiO2, Al2O3, MnO2, CeO2

Oxidizers: hydroxylamine, KMnO4, KIO4, H2O2, NO3-

Strong acids and weak buffering acids: HF, HNO3, H3BO3, NH4+, citric acid

Strong bases: NH3, OH-

Organic Materials

Dispersants/surfactants: poly(acrylic acid), quaternary ammonium salts, alkyl sulfates, EDTA

Corrosion inhibitors: benzotriazole, alkyl amines

Metal complexing agents: EDTA, ethanol amines, oxalic and citric acid

Acids: poly(acrylic), oxalic, citric, acetic, peroxy acetic

In wastewater, most inorganic materials appear in oxidized form. The majority of insoluble inorganic contaminants come from the slurry abrasive, which may include suspended particles of silica (SiO₂), alumina (Al₂O₃) and ceria (CeO₂), typically in a 50-500 ppm concentration after dilution. Following oxide polishing, we determined that the particle size distribution broadened from 0.2 µm to a range of 0.01 µm to 0.5 µm. This change is a result of the CMP process, which presses particles together in the presence of corrosive agents at pH from 2 to 11. Abrasion of the wafer surface also introduces other insoluble inorganic contaminants such as metals, metal oxides and low-k dielectrics. The soluble inorganic species include oxidizers such as hydroxylamine and hydrogen peroxide. Importantly, these materials must be reduced to a more innocuous form before discharge.

1. Like other transition metals, in a saturated solution of metal hydroxide, cupric ion (Cu 2+ ) is more soluble at low pH (Source: C.N. Sawyer, et al, Chemistry for Environmental Engineering , McGraw-Hill, N.Y., 1974, 4th Ed., p. 147.)

Wastewater pH largely dictates the relative solubility of inorganic contaminants. In general, common transition metals exist as insoluble oxides or hydroxides at elevated pH and under oxidizing conditions (Fig. 1). Copper metal's interaction with water at different oxidation-reduction potentials is shown in Figure 2. At low pH, copper and other transition metals are highly soluble. At pH 3, cupric ion (Cu²⁺) can reach a concentration of 1900 ppm (mg/L),⁵ while a pH of 8 solution yields insoluble Cu₂O, CuO and Cu(OH)₂^4-6 and only 0.1 ppm Cu²⁺. The solubility properties of iron are similar to copper: at pH of 1.6, Fe³⁺ exists as the free hydrate, while at pH >4, the insoluble hydroxide form dominates.⁷ Some slurry formulations use ferric salts in conjunction with hydrogen peroxide to oxidize metals. Note that Figure 2 does not consider the effect of copper complexing agents in CMP wastewater and is for general reference only.

2. As shown in this Pourbaix diagram, 4,7 copper forms insoluble oxides and hydroxides at pH of 7-12.

Silica and alumina represent important exceptions to these solubility guidelines, displaying a maximum insolubility from pH 6 to 7.⁸ However, at pH >9, silica and alumina begin to dissolve and form soluble silicates and aluminum-hydroxo species.^7,8 The increase in silica solubility at pH values >9 is what drives the on-wafer oxide CMP process. Due to scaling concerns in reverse osmosis or gray water applications ( >10 ppm SiO₂, Al₂O₃), the pH should first be adjusted downwards for optimal pretreatment.

Many users will add ammonia to CMP formulations, for example, to complex and to remove metals and other on-wafer contaminants. In high pH copper CMP processes, the driving force for copper dissolution is the formation of [Cu(NH₃)_n]²⁺ complexes where n = 1, 2, 3, 4.⁴ High ammonia concentrations in the wastewater can continue to complex the cupric ion or may dissolve insoluble Cu(OH)₂ and related oxides according to:

Microbar's previous work in the remediation of copper CMP wastewater suggests that ammonia concentrations exceeding 100 ppm have little effect on coagulation processes.⁹ It was observed that silica and copper removal efficiency was maintained to less than 1 ppm at pH 8. This is because coagulants like aluminum and iron hydroxide compete with ammonia and other metal complexants, and will sequester transition metal ions by absorption and adsorption processes.⁹

Ultimately, it is difficult to predict or calculate the amount of soluble or complexed copper in CMP wastewater. The aqueous chemistry is in constant flux due to changing flow rates, variable tool uptime, dilution factors and flow segregation. In addition, ammonia, EDTA, organic acids, and other species used in slurry formulation compete with the thermodynamically favored oxides and hydroxides. Also, copper complexes have similar stability constants (Table 2),¹⁰ and compete with an aqueous medium high in hydroxide and oxygen.

Table 2. Forms of Copper and Formation Potential
Copper species	Formation constant (log k)
Cu(OH)+	6.3
Cu(OH)2(aq), Cu(OH)2(s)	11.8, 20.4
Cu(OH)42-	16.4
[Cu(NH3)n]2+ n = 1, 2, 3, 4	4.0, 7.5, 10.3, 11.8
Cu(EDTA)2-	20.5
CuOH(EDTA)3-	22.6
Cu(citrate)-	7.2
CuOH(citrate)2-	16.4
[Cu2(citrate)2]2-	16.3
Cu(dialkyldithiocarbamates)	21-30

In dilute systems, mass transport affects the kinetics of copper complexation. For example, we have observed "end of pipe" 10-fold dilution factors in CMP wastewater, which add an unquantifiable, but significant mass-transport barrier to the binding of slurry derived complexants to equally dilute copper species. Oxidizers such as hydrogen peroxide and hydroxylamine in wastewater may attack ligands complexed to copper, releasing the copper for oxide and hydroxide formation (see Maag¹¹ for further discussion).

Copper can also absorb on a silica surface, becoming a potential source of insoluble copper. Since many silica particles are negatively charged at pH 3 6, according to:

about half of the surface silica is negatively charged at pH 5.9 and will bind to cupric ions. Alternatively, hybrid copper oxide and hydroxide species may be covalently bound to the surface oxide or engage in hydrogen bonding.

The interaction of copper ions with SiO₂ has been simulated using a surface complexation model.¹² It simultaneously solves equations that define the mass balance of sites, mass action and charge balance, and considers the equilibria of the protonation and deprotonation of surface silanol groups and the reactions:

Using these assumptions, one can predict the adsorption of copper ions onto silica particles (surface area = 416 m²/g) as a function of copper solution concentration at different pH values (Fig. 3). Copper adsorption increases sharply with increasing pH, and, to a lesser degree, copper concentration. This is consistent with a predominance of negatively charged surface groups above pH 5.9. While the modeled concentrations are relatively high compared to real-world copper concentrations in CMP wastewater, the model is likely operative at the wafer surface during CMP and possibly in the wastewater itself. Finally, we have observed that silica particles in many real-world copper CMP wastewater samples display a pale blue coloration, strongly suggesting adsorbed copper in some form.

3. At higher pH, silica is more likely to absorb copper due to increased negative surface charge. (data for 50 ppm copper)

Quantifying copper in wastewater

Microbar has processed over one dozen copper CMP wastewater samples provided by both fabs and equipment manufacturers. The samples' physical characteristics (copper content, solids content, pH, conductivity, etc.) varied significantly depending on the place and time of collection. However, we have found that the amount of total copper in copper CMP wastewater typically ranges from 0.1 to 5 ppm. In an attempt to reconcile our empirical results with an actual copper CMP process, we performed a preliminary mass balance around an Avanti 472 polisher from IPEC/Westech (Phoenix, Ariz.) and Synergy cleaner from OnTrak (now Lam Research, Fremont, Calif.).

First, we polished a 200 mm wafer for approximately 1 minute, removing 5,533 Å of copper. The EKC CMP 9003 slurry contained approximately 5% silica solids, 3-5% hydroxylamine and 100 ppm BTA. We collected 450 grams of pH 4 polisher wastewater every 12 seconds for 3 minutes. We determined the copper concentration in each sample using inductively coupled plasma spectroscopy (ICP).

4. The area under the curve gives the total amount of copper in the polisher effluent: 70.2 mg by ICP.

Approximately 67 mg of electroplated copper was removed from the wafer during this period based on copper thickness, copper density, polishing rate and wafer size. The total amount of copper in the wastewater samples was 70.2 mg. The approximate silica concentration was 350 ppm.Factoring in the flow rate through the polisher (2.25 L/min), we calculate a peak copper concentration of 30 ppm in the polisher effluent over the three minute period and an actual peak concentration of 34 ppm (Fig. 4).

As expected, dilution dramatically affects copper concentration. A 2.2 L/min flow through the cleaner reduces the copper concentration from 34 ppm to 15 ppm. Boosting the polisher flow rate to 25 L/min, a typical real-world level, reduced the copper concentration to approximately 0.5 ppm.

From our simple single tool mass balance study, we can reasonably expect dilute levels of copper in CMP wastewater in the range of the 0.1 to 5 ppm levels we observed in real-world samples. A combination of pad conditioning, platen rinse and cooling and post-CMP cleaning can easily cause a ten-fold or greater dilution of copper depending on flow-rates, the number of polishers and cleaners tool up-time, and the wastewater segregation scheme. Our results also are consistent with a recent mass balance study that found <0.5 ppm of copper in various CMP wastewaters.¹¹

CMP wastewater treatment

There are a variety of treatment schemes available for the removal of suspended solids, heavy metals, and some organic materials in CMP wastewater. Coagulation and flocculation are effective, commonly used chemical processes that involve neutralization of ion and particle surface charge by oppositely charged inorganic and organic materials (see Letterman⁹ for a complete review). This procedure alon e may be enough to destabilize some soluble and suspended materials, so that they can be filtered or isolated by other means. Practically, this process is carried a step further by introducing an excess of coagulant, so that the particles and some ions are trapped within a gel-like matrix and agglomerate. This process is known as sweep coagulation, and the precipitate is called "floc." Typical inorganic coagulants used for this purpose are aluminum sulfate and ferric chloride, both of which form insoluble hydrated hydroxide gels at pH from ~5 to 8. Organic coagulants may also be used. The coagulated agglomerate can be further destabilized for gravity settling or active filtration by adding organic flocculants.

5. How particles are destabilized by coagulation.

Flocculation uses a high molecular weight organic polymer, such as poly(acrylamide) and it's negatively charged copolymers. Figure 5 illustrates one process by which suspended charged particles are destabilized. Polymeric coagulants and flocculants commonly display molecular weights ranging from 2,500 to many millions (g/mole), and may be negatively or positively charged, or neutral. Organic coagulants are typically lower in molecular weight.

Recently, a new approach has been investigated for particle destabilization in CMP wastewater without traditional chemical additions. Electrocoagulation and electrodecantation use electric fields to agglomerate charged silica particles.¹⁵ This technology has promise and warrants further investigation.

Click for larger image 6. Though proven and effective, the gravity settling scheme occupies a large footprint (~3000 ft 2 for a 300gal/min system).

After coagulation and/or flocculation, three physical mechanisms are commonly used to separate the floc from the clarified water: gravity settling, cross-flow filtration and single-pass low-pressure filtration. In gravity settling (Fig. 6), the wastewater is treated with an excess of calcium oxide (lime) to create insoluble materials including metal hydroxides, calcium silicate and calcium fluoride. Other coagulants may be added, followed by a high molecular weight flocculant. The mass settles over time to the bottom of the clarifier and water is separated from the sludge. These crude but relatively effective systems are also commonly used to treat municipal wastewater.

Gravity settling, though a simple and proven technology, typically has large space requirements. A clarifier and thickener can occupy 3000 ft² or more for a 300 gallons per minute system. More space is also needed for equalizing tanks and sand filters. Large capital costs are involved in controls, piping and tank construction. These systems are usually inflexible due to the gross-nature of the process. For details, we direct the reader to an excellent review of sedimentation and gravity clarification schemes.⁹

7. Cross-flow filtration provides effective filtration with multiple passes through the membrane array. Efficiency is 10-150 gal/ft 2 of membrane per day.

Cross-flow filtration systems have gained attention for effective solids removal and smaller system footprint compared to gravity clarification systems. Cross-flow filtration uses tangential water flow through an array of ceramic tube filters or multilayer polymer sandwiches to remove suspended solids and other contaminants (Fig. 7). Water enters the tube at pressures ranging from 25 to 75 psig, and is partitioned into water that permeates the tube (filtered) versus water that passes straight through the tubes unfiltered. The unfiltered water is continuously recycled and reconcentrated. The concentrate typically needs further treatment in a gravity settler to completely precipitate the mass. A high pressure back-flush cycle sloughs off the filter cake that builds on the surface of the membranes, as well as particles that become entrained in the filter. The resultant flux, or measure of filtration efficiency is measured in gallons per square foot of membrane per day (GFD). Cross-flow flux values range from 10 to 150 GFD. The use of these systems in CMP wastewater treatment has been reviewed.¹⁴ Suppliers of cross-flow filtration systems include EPOC Filtration and Separation Systems (Fresno, Calif.), Pall Corp. (E. Hills, N.Y.), and US Filter (Warrendale, Pa.).

Click for larger image 8. A single-pass low-pressure microfiltration system can generate a flux >200 gal/ft 2 of membrane per day.

Microbar recently introduced a new filtration technology, single-pass low-pressure microfiltration.¹⁵ Capable of generating flux values exceeding 200 GFD, the EnChem system operates at low-pressures (4-10 psi) and high flow rates (to 5,000 gal/min). Unlike cross-flow filtration, the microfiltration system processes wastewater in a single pass with no reconcentration mode (Fig. 8 ). First, the wastewater is treated with inorganic and organic coagulants that engulf and agglomerate the fine suspended particles. If the wastewater contains complexed heavy metals or fluoride, a metal removal agent and/or calcium chemistry may also be added. After pretreatment, the wastewater is filtered using an array of sock membranes with a pore size much smaller than the floc particle size. This enables low-pressure filtration without clogging the filters. After approximately 10 minutes, a filter cake builds on the membrane surface and the pressure rises to approximately 7 psi. To slough off the filter cake from the membrane surface and thus lower the pressure back to baseline, the flow is reversed for approximately 10 seconds(<2 psig gravity backflush). A sludge pump removes the filter cake from the bottom of the vessel and the process continues.

Copper removal options

The toxicity of copper to flora and fauna in aquatic systems is well known and beyond the scope of this paper. We direct the reader to an excellent review on the subject.¹⁶ Copper can be effectively removed from CMP wastewater using coagulation and microfiltration alone.⁹ For example, we observed a drop in copper content of effluent samples from 38 ppm to sub-0.1 ppm following coagulation and microfiltration. However, to insure copper removal to sub-0.1 ppm levels due to stringent discharge requirements, it may be necessary to use water-soluble heavy metal removal agents. These materials typically contain sulfide functionality, which bonds strongly to transition metals and forms insoluble precipitates. Typical metal precipitants include NaHS, dithiocarbonate, and dimethyldithiocarbamate.¹⁷ While effective, sulfide salts and small molecules suffer from toxicity issues and form very fine particles that are not easily agglomerated or filtered.

We also have studied the removal of copper in CMP wastewater using a metal chelating polymer, which forms a floc upon contact with dissolved heavy metals. Polymeric metal removal agents removed copper levels to below 50 ppb, even in the presence of ammonia and other competing materials. Relative to other complexants in CMP wastewater, a polymer containing sulfide functionality is expected to have a formation constant exceeding other species such as EDTA (Table 2),¹⁷ and more efficiently turns nanometer-scale particles into particles several microns in size for simple filtration. Toxicity issues with these materials are of limited concern.

Alternatively, copper may be removed from the CMP wastewater before filtration,^18,19 using acidification followed by ion exchange. The wastewater pH is then readjusted to 7 to 8 and may be treated by coagulation and microfiltration. Drawbacks of this approach include the large amounts of acid and base used in the pH adjustment steps, as well as the need for frequent ion bed regeneration. Suspended solids can also damage the ion exchange columns. Nonetheless, this approach may have some merit and requires further study.

Copper toxicity and disposal

There is some controversy regarding the toxicity or hazardous nature of copper CMP wastewater and sludge. The safe discharge of CMP wastewater depends on U.S. federal (EPA) and local regulations. For the regulation of the semiconductor industry, the EPA imposes:

.40 CFR 433 Subpart A: for regulation of metal finishing and electrodeposition processes related to final assembly and wirebonding. Sets standards on: cyanide, metals: Cd, Cr, Cu, Pb, Ni, Se, Zn; total toxic organics (TTO); total suspended solids (TSS); and pH.²⁰

.40 CFR 469 Subpart A: discharges resulting from all process operations associated with the manufacture of semiconductors, including interconnect and CMP processes. Sets standards on: pH, F-, As, and TTO.²⁰ Because no specific limits for solids or copper discharge are given, attention must be paid to local discharge limits.

Although current EPA regulations do not directly specify copper CMP processes or effluent, it is suggested that copper CMP is covered by 40 CFR 469.²⁰ At the date of this publication, the EPA has not ruled specifically on the hazardous or non-hazardous nature of copper CMP wastewater or sludge.²⁰ However, because sludge derived directly from plating operations is considered hazardous and bears the F006 designation, it would be sensible to separate plating rinsewaters and concentrates from CMP wastewater; otherwise a larger body of hazardous waste will be created.

Since the EPA has not made an absolute ruling on whether copper CMP sludge is hazardous, it is prudent to default to local regulations. California is home to perhaps the most stringent environmental regulations in the world. Examination of Title 22, Section 66261.24 of the California Code of Regulations reveals that a copper bearing sludge containing 2.5 g of copper/kg of sludge or less is not deemed hazardous. Based on this guideline, a copper-bearing sludge obtained from a typical coagulation and microfiltration process may be below this limit.

This is a direct result of the dilute nature of copper in CMP wastewater. We have calculated that a 30% solids sludge derived from a typical CMP waste stream containing 1 ppm copper and 400 ppm silica, will meet this guideline with at least a 50% margin.

Conclusions

CMP wastewater is constantly in flux: the contaminant concentration depends on the process and slurry chemistry, tool up-time, flow rates and wastewater segregation schemes. Coagulation and flocculation effectively treat CMP wastewaters containing high suspended solids, fluoride and heavy metals. We have found that copper is mostly insoluble and relatively dilute in real-world samples. This is consistent with a simple mass-balance study that showed typical peak concentration of copper at the end of one CMP/post-CMP clean process of ~0.5 ppm. Concentrations below 0.1 ppm are more common after dilution. Because copper-bearing sludge toxicity and disposal issues remain unresolved on the federal level, it is prudent to abide by local regulations, which may be more stringent. Finally, because of the dilute nature of copper in CMP wastewater, copper-bearing sludges derived from coagulation and microfiltration treatment schemes may be non-hazardous. .

Josh H. Golden , Ph.D. is a senior scientist at Microbar Inc., where he is involved in the development of CMP wastewater treatment technology and dispense solutions for low-k spin-on dielectrics. He received his Ph.D. in Chemistry from Cornell University. He was formerly a polymer chemist at Cytec Industries, and lead chemist in the low-k dielectric materials program at Watkins-Johnson. He has authored 19 papers and holds several patents.
Phone: 1-408-542-9069
e-mail: jgolden@microbar.com.

Robert J. Small , Ph.D., is research director of the R&D group at EKC Technology, where he is involved in developing new chemistries for post-etch residue removal, post-CMP cleaning and new CMP chemistries. He has a B.S. from Norwich University, an M.S. from Texas Tech University, and a Ph.D. in organic photochemistry from the University of Arizona in Tucson. He has co-authored over 45 articles, holds more than twelve U.S. and foreign patents and has six submitted patent applications.
Phone: 1-510-784-5846
e-mail: bsmall@ekctech.com.

Louis Pagan has 24 years of experience in the electronics and semiconductor industries, and is currently an applications engineer in the R&D area of EKC Technology. Having specialized in CMP process development, applications, CMP mechanical engineering and equipment engineering over the last 10 years, he also has designed CMP cleanrooms.
Phone: 1-510-780-1345
e-mail: lpagan@ekctech.com.

Cass Shang is an R&D chemist at EKC Technology, where she focuses on developing new CMP slurries and optimizing CMP processes. She has a B.S. in chemical engineering from Beijing University, and an M.S. in environmental and surface analysis from the University of Colorado.
Phone: 1-510-670-1478
e-mail: cass_shang@ekctech.com.

Srini Ragavan , Ph.D., is professor of Materials Science and Engineering at the University of Arizona. He has been at the University since 1978, performing research and teaching in the areas of surface and colloidal phenomena, wafer cleaning, CMP and corrosion. He is a principal investigator in the NSF-SRC Center for Environmentally Benign Semiconductor Manufacturing and in the Center for Microcontamination Control at the University of Arizona. He received his Ph.D. in Materials Science and Mineral Engineering from the University of California, Berkeley. He has published over 80 papers and holds three patents.
Phone: 1-520-621-6073
e-mail: srini@u.arizona.edu.

---

REFERENCES

1. W.C. O'Mara, "CMP Equipment and Materials Take Semiconductor Industry by Storm," Semiconductor Online, 10/12/99.
   
2. G.L. Corlett, "CMP Water Reduction and Waste Treatment Overview," Semiconductor Online, 3/17/99.
   
3. SIA, National Technology Roadmap for Semiconductors, 1997 Edition.
   
4. J. M. Steigerwald, S. P. Murarka, R. J. Gutmann, Chemical Mechanical Planarization of Microelectronic Materials, John Wiley & Sons, Inc., N.Y., 1997.
   
5. G. Wu, L. Cook, "Mechanism of Copper Damascene CMP," Proc. 3rd Int'l CMP-MIC, Feb. 19-20 1998, p. 150.
   
6. R.J. Gutmann, et al., "CMP of Copper-Polymer Interconnect Structures," Ibid., p.257.
   
7. M.J.N. Pourbaix, Atlas of Chemical Equilibria in Aqueous Solutions, National Assoc. of Corrosion Engineers, Houston, Texas, 1974.
   
8. I.K. Iler, The Chemistry of Silica, John Wiley and Sons Inc., N.Y., 1979.
   
9. R.D. Letterman, (ed), Water Quality and Treatment, McGraw-Hill, N.Y., 1999, 5th Ed.
   
10. F. Morel, J. Hering, Principles and Applications of Aquatic Chemistry, John Wiley and Sons Inc., N.Y., 1993.
    
11. B. Maag, "The Environmental Impact of Copper CMP," Masters Thesis, MIT, 2000.
    
12. Y. Sun, "Colloidal Chemical Investigations Pertaining to Copper CMP," Ph.D. Thesis, Univ. of Ariz., 2000.
    
13. B.M. Belongia, et al., "Treatment of Alumina and Silica CMP Waste by Electrocoagulation and Electrodecantation," J. Electrochem. Soc., 146, 1999, p.4124.
    
14. M. Reker, M. Lenart, S. Harnsberger, "Treatment and Water Recycling of Copper CMP Slurry Waste Streams to Achieve Compliance for Copper and Suspended Solids," Semiconductor Fabtech, 8th Ed., p. 141.
    
15. J.H. Golden, "CMP Wastewater Treatment," Proc. CMP-MIC Conference, 2000, p.305, (also see: US 05871648).
    
16. R. Eisler, "Copper Hazards to Fish, Wildlife, and Invertebrates: A Synoptic Review," U. S. Geological Survey, 1997-0002. (Order via phone: 301-497-5550).
    
17. A. Hulanicki, "Complexation Reactions of Dithiocarbamates," Talanta Review, V. 14, pp. 1371-1392.
    
18. L. Mendecino, et al., "Environmental Impact of Process Tools," SEMICON/West, 1999.
    
19. L. Kirman, "Copper Removal from CMP Wastewater," CMP Users Group, NCCAVS, SEMATECH, ERC, April 1998.
    
20. Personal communication, Mr. Keith Silva, EPA Region 9, Phone: 415-744-1907. Also see: www.epa.gov/docs/epacfr40/chapt-I.info/subch-N/.
    

Talkback

We would love your feedback!

Post a comment

» VIEW ALL TALKBACK THREADS

Related Content

• Topics
• Author

Related Content

 

By This Author

There are no other articles written by this author.

SPONSORED LINKS