WSRC-MS-98-00869

Available to the public from the National Technical Information Service, U.S. Department of Commerce, 5285 Port Royal Road, Springfield, VA 22161.

Abstract

A study was undertaken to evaluate the resistance of various materials and coatings to the adherence of molten borosilicate glass. Candidate materials were ranked on the ability to shed molten glass as the material surface temperature was elevated. Superior performance was achieved with materials that either had a high thermal conductivity or were non-metallic/non-wetting. Metallic materials that formed loosely adhering oxide layers also performed acceptably. Results of the test and application of the candidate material in a radioactive waste glass melter will be presented.

Introduction

Approximately 125 million liters (33 million gallons) of highly radioactive waste solutions, from the production of nuclear materials at the United States Department of Energy's Savannah River Site (SRS), are presently stored in large underground carbon steel tanks. Waste handling operations separate the waste into three parts: highly radioactive insoluble sludge, highly radioactive precipitate slurry, and decontaminated aqueous phase of dissolved salts. The decontaminated salt solution is being immobilized by incorporation into a concrete waste form. In preparation for long-term storage to allow for controlled decay of long-lived radionuclides, the highly radioactive sludge and slurry is being vitrified and encapsulated in stainless steel canisters in the Defense Waste Processing Facility (DWPF). The vitrification process was initiated approximately 3 years ago and is expected to take approximately 25 years to complete. The stabilized solid radioactive waste is being temporarily held at the Savannah River Site until emplacement in the Yucca Mountain repository.

In preparation for vitrification the radioactive feed is chemically processed using various acids. Borosilicate frit is blended with the processed feed and then continuously fed under normal operating conditions to the melter (Figure 1). The melt pool temperature is maintained at approximately 1150°C. Pouring rates are nominally 82 kg per hour (180 lbs/hr).

Figure 1. Schematic of the DWPF glass melter and pour spout.

Normally molten glass flows over the riser and down the side of the pour spout closest to the melter. Two knife edges were machined into the pour spout to provide a release point for the molten glass stream. The Inconel 690 pour spout is not removable and therefore, failure of this component would result in the removal of the entire melter. External heaters are used to keep this region of the pour spout between 1000 and 1050°C. The molten glass free falls from the knife edge into the stainless canister below. As the knife edges degraded the glass stream began to waver and wick down the side of the pour spout allowing the glass to contact the bellows. The temperature in the bellows was significantly cooler than the molten glass, approximately 500°C in the inlet to 350°C near the bottom, so the glass solidified in this region. During waste qualification (non-radioactive operations) the solidified glass was removed from the bellows by mechanical cleaning. However, this would be impossible during radioactive operations. Therefore, a Type 304 L stainless steel replaceable liner was developed to protect the bellows from the glass (Figure 2a and b). Initially the liner was successful, keeping glass from contacting the bellows and funneling it into the canister but as the liner aged a thick chromium oxide layer developed. When the pour stream contacted the 304L liner the glass would adhere to this oxide layer. If the pour stream instability continued the liner would plug. Remote mechanical cleaning techniques were developed to remove the glass, but were only partially successful. The liner was inexpensive and was replaced if it became deformed or could not be cleaned. Service life of the 304L liners ranged from 2 weeks to 3 months depending on the stability of the pour stream. In order to increase liner's service life, a program was undertaken to identify materials and coatings that would minimize or eliminate adherence of glass. This paper describes the testing and the selection of a candidate material used to fabricate a prototypic liner.

Experimental Apparatus and Test Procedures

The purpose of the screening tests was to determine the relative temperature at which molten glass will adhere to various materials and coatings. To perform this task a small melter capable of melting 300 grams of glass per batch (Figure 3) was constructed. The melter system consisted of two zones, 1) an Inconel 690 box (dimensions, 10.2 cm by 7.6 cm by 10.2 cm) and 2) an Inconel 690 drain tube (7.6 cm long with a 4.5 mm inner diameter). Each zone was surrounded by independently controlled pairs of heaters. The heaters and the area below the drain were covered with an insulating blanket and fiber board. Test specimens were placed on a hot plate that was positioned approximately 7.5 cm below the end of the drain. The hot plate was not insulated. Coupons were tested at 45 and 80 degree angles to the horizontal plane.

Most of the coupons were 0.318 cm thick; however, a thicker (0.953 cm) coupon was fabricated from each of the following materials, Type 304L stainless steel (304L), oxygen-free copper, and aluminum. Coupons were either used in the as-received condition (better than a 600 grit finish), ground to a 600 grit finish or polished to a 1 micron finish. All coupons were cleaned with ethyl alcohol prior to testing. The thickness of the silver and gold electroplating on the copper and 304L coupons was 2.54 microns (0.0001 in). Candidate materials, surface treatments, and coupon dimensions are shown in Table 1.

Testing was performed by melting borosilicate glass at approximately 1050°C. The drain was maintained at approximately 1025°C, which is a typical temperature for glass entering the pour spout. This would yield glass temperatures between 760 and 830°C on the surface of the test specimen. All tests were performed in air. Glass exited the drain and formed a bead approximately 5 mm in diameter every 7 to 10 seconds. The test specimens were heated using the hot plate from ambient to the temperature where glass would begin to adhere to the coupon. A piece of Inconel weld wire was used to push the glass bead off the coupon when it began to stick. The temperature at which the bead could not be dislodged from the coupon using minimal force was defined as the "glass adherence" temperature.

Results

General Observations

Results of the screening tests indicated that high thermal conductivity materials namely, copper, sterling silver, Consil^® 995 and aluminum had the best (highest) glass adherence temperatures of all the metallic materials. Solid graphite, boron nitride, and the carbon/carbon fiber composite performed extremely well with glass adherence temperatures ranging from 440 to 510°C (Table 1). When glass stuck to these materials it could be removed by lowering the temperature slightly. Highly polished surfaces (1 micron or better) did not provide a significant performance advantage over a machined surface (32 rms) or ground surface (600 grit / ~ 17 micron). Generally, diffusion and electroplated coatings did not beneficially affect the glass adherence characteristics. However, the copper coupon with the polished dense chromium coating did show a significant improvement with the glass beginning to adhere around 440°C .Increasing the slope of the coupons also did not appear to affect the glass adherence temperature significantly.

Materials	Material Condition	Dimensions (cm)*	Glass Adherence Temperature (C)
304L	As Received (mill finish)	15.2/10.2/0.318	290
	Polished 1 micron	15.2/10.2/0.318	289
	Polished 600 grit (17 micron)	15.2/10.2/0.318	313
	Gold Plated (0.0001 in)	15.2/10.2/0.318	284
	Silver Plated (0.0001 in)	15.2/10.2/0.318	297
	Dense Chromium (Armoloy)	15.2/10.2/0.318	285
	Dense Cr (Armoloy) polished	15.2/10.2/0.318	280
	Oxidized 400°C	15.2/10.2/0.318	300
	Siliconized Alon	7.6/7.6/0.318	313
	ChromePlex(TM) (Cr/Si) Alon	7.6/7.6/0.318	306
	Bidiffused (Cr/Al/Si) Alon	7.6/7.6/0.318	300
	Polished 600 grit (17 micron)	15.2/10.2/0.953	412
Inconel 690	Oxidized 900°C	15.2/10.2/0.318	290
Copper        (Oxygen Free)	As Received (better than 600 grit)	15.2/10.2/0.318	363
	Polished 1 micron	15.2/10.2/0.318	375
	Polished 600 grit (17 micron)	15.2/10.2/0.318	363
	Polished 600 grit (17 micron)	15.2/10.2/0.953	422
	Gold Plated (0.0001 in)	15.2/10.2/0.318	434
	Silver Plated (0.0001 in)	15.2/10.2/0.318	385
	Dense Chromium (Armoloy)	15.2/10.2/0.318	384
	Dense Cr (Armoloy) Polished	15.2/10.2/0.318	440
	Oxidized 900°C	15.2/10.2/0.318	341
Sterling Silver	Polished 600 grit (17 micron)	5.1 dia/0.635	418
Consil® 995	As Received (better than 600 grit)	10.7/6.5/0.318	400
Consil® 995	As Received (better than 600 grit)	6.4/6.4/0.030	389
Consil® 995	Oxidized (732°C, 40 minutes)	6.4/6.4/0.030	336
Aluminum (1100)	Polished 600 grit (17 micron)	15.2/10.2/0.953	403
Boron Nitride (BN)	Polished 600 grit (17 micron)	8.3 dia/0.737	440
BN (Combat®)	Spray (304L Substrate)	15.2/10.2/0.318	290
BN (Combat®)	Paste Type A (304L Substrate)	15.2/10.2/0.318	286
BN (Combat®)	Paste Type S (304L Substrate)	15.2/10.2/0.318	280
Graphite	Polished 600 grit (17 micron)	15.2/10.2/0.318	510
Carbon/Carbon	As Received	15.2/10.2/0.318	465
Nickel 200	Polished 600 grit (17 micron)	15.2/10.2/0.318	403
CDA 706 (C70610)	Polished 600 grit (17 micron)	15.2/10.2/0.318	430

* Length/Width/Thickness

Type 304L Stainless Steel and Inconel 690

Glass adhered to the 0.318 cm (0.125 in) thick, 304L stainless steel coupon at temperatures between 285 and 313°C. Surface coatings, highly polished surfaces, and electroplating with gold or silver did not significantly increase the glass adherence temperature. Temperature of the thicker, 0.953 cm (0.375 in), coupon exceeded 400°C before the glass began to adhere.

Glass stuck to the pre-oxidized (900°C) Inconel 690 coupon at 290°C. The glass did not spall off during cooling and was also difficult to remove at room temperature. The oxide scale was thick and uniform.

Polished and as-received coupons formed a thin gold colored oxide, similar to the pre-oxidized coupon, when heated on the hot plate above 400°C. A blue oxide was observed on these coupons around the glass contact area. Glass tenaciously bonded to this oxide and could not be removed at room temperature even with vigorous mechanical cleaning. Removal of glass at room temperature from the electroplated and diffusion coated coupons was significantly easier than cleaning the oxidized samples.

Copper Coupons

The glass sticking temperature for the various copper coupons ranged from 363 to 440°C. This was significantly higher, approximately 100°C, than the 304L and Inconel 690. As-received and polished oxygen-free copper coupons formed a non-uniform loosely adhering scale, which would spall off during the test. Glass that adhere to the freshly exposed copper substrate could be easily removed by mechanically cleaning after it cooled to room temperature. The force necessary to clean the glass deposits from these coupons was significantly less than that required to remove the glass from the stainless steel coupons. Increasing coupon thickness was beneficial, resulting in an increase in the glass adherence temperature.

The Copper Development Association alloy CDA 706 (Cu/Ni) formed a uniform oxide scale that would bond with the glass. However, upon cooling the glass would delaminate at the oxide metal interface and slide off. Oxide from the coupon was observed on the glass surface, which was in contact with the coupon. This oxide layer would quickly reform on the coupon so the glass never bonded with the copper substrate.

The copper coupon pre-oxidized at 900°C had the lowest glass sticking temperature, 341°C, when compared to the other copper coupons. This coupon contained a thick black uniform scale. As was observed with the CDA 706 coupon, the glass would delaminate upon cooling. The oxide did not appear to degrade to the point where the copper substrate was exposed during the thirty minute test.

Graphite, Carbon /Carbon Fiber Composite, and Solid Boron Nitride

Non-metallic materials including solid boron nitride, graphite, and the carbon/carbon fiber composite had the highest glass sticking temperatures of all materials tested, 440, 510, and 465°C respectively. Above these temperatures glass would adhere to these materials; however, the glass would slide off when the temperature was decreased slightly. These materials did not show any evidence of degradation, i.e., cracking or discoloration, as a result of exposure to molten glass at the test temperatures.

Other Alternate Materials - Sterling Silver, Consil^® 995, Aluminum, and Nickel 200

Temperatures in excess of 400°C were required for glass to stick to materials with a high thermal conductivity, (i.e., sterling silver, Consil^® 995, and 1100 aluminum). These materials shed glass readily upon cooling. The sterling silver and Consil® 995 were slightly discolored, but no degradation was noted. Scale formation on these materials was minimal. Increasing the Consil® 995 coupon thickness again increased the glass adherence temperature. A uniform layer of Al₂O₃ formed on the aluminum coupon at elevated temperatures and remained intact during the entire test. Nickel 200, which has a thermal conductivity much lower than the aluminum and sterling silver, performed satisfactorily up to 403°C. This alloy formed a blue uniform scale and shed glass much like the CDA 706. Glass that remained stuck to the coupon would delaminate upon cooling to room temperature. The oxide beneath the glass deposits was continuous and uniform in thickness.

Discussion

The glass adherence temperature is defined as the minimum temperature at which molten glass will begin to adhere to the test coupon. For the bellows liner higher sticking temperature are beneficial, as this would allow a broader operational temperature range. Metallic materials with high thermal conductivities, i.e., sterling silver, copper, and aluminum, had the highest glass adherence temperatures compared to the other metals (Table 2). The molten glass would rapidly

quench and slide off the coupon when the temperature differential between the molten glass and the coupon surface was large. However, as this temperature differential decreased, bonding (chemical and/or mechanical) resulted. This may occur because the chemical reaction is more kinetically favorable at the higher temperatures and because the molten glass is less viscous. Others have shown that decreasing viscosity increases the contact area thus, increasing the probability of a chemical bond [1,2]. Stresses developed between the glass and the substrate will also determine whether or not the glass will spall. If the glass bonds with the metal substrate or with a tenaciously adhering oxide and the thermal stresses generated upon cooling are not sufficient to break this bond, the glass will not delaminate and slide off. If however, a chemical bond does not exist or if the glass bonds to an oxide that is weakly adhering to the metal substrate, the glass will spall. In order for the latter to occur continually the oxide must reform rapidly. The oxide that forms on the CDA 706 and the Nickel 200 appear to perform in this fashion.

Glass stuck to both the 304L (reference material of construction of the current DWPF bellows liner) and the Inconel 690 coupons. These alloys contain chromium and readily form a spinel (NiCr₂0₄) or chromium oxide (Cr₂O₃) protective layer in oxidizing environments. The oxide that formed on the 304L above 400°C was thin and tenaciously bound to the substrate. The molten glass has an affinity for chromium oxide and thus chemically bounded with this layer. Pask [3] has shown this for other silicate glasses. The oxide scale that formed on the Inconel 690 coupon pre-oxidized at 900°C was not tightly bound to the substrate thus the glass did not adhere as tenaciously.

Electroplated and diffusion coatings on copper and 304L coupons did not significantly increase the glass adherence temperature. This suggests that the thermal conductivity and the mass (thicker coupons had higher glass adherence temperatures) of the substrate play a greater role in impeding glass adherence than does surface tension at the lower temperatures. The polished dense chromium coating on the copper coupon was the only coating that showed an improvement in performance. Further testing of this material would be necessary to understand this phenomena.

Non-metallic materials including boron nitride, pure graphite, and the carbon/carbon fiber composite also had high glass adherence temperatures, 440, 510, and 465°C, respectively. Above these temperatures glass would stick to the coupons. However, unlike the metallic coupons, the glass would slide off when the temperature was decreased slightly. The molten glass did not appear to wet the surface of these materials, especially at the lower temperatures. Therefore, surface tension may play an important role with the ability of these materials to shed molten glass.

Increasing the angle of the coupon plane did not significantly change the glass sticking temperature. This may be because the glass beads were small and their weight was insignificant compared to the surface area adhering to the metal surface. Allowing a large deposit of glass to adhere to the coupon surface or increasing the flow may affect this result. Future work will be required to evaluate the effects of continuous pouring at rates comparable to the DWPF melter, 82 kg per hour.

Conclusions

Based on the results of this study the conclusions are:

1. Glass adhered tenaciously to oxidized Type 304L stainless steel even after the coupon was cooled to room temperature.
2. Materials were identified that had excellent resistance to glass adherence. - Non-metallic materials - solid boron nitride, graphite, and the carbon/carbon fiber composite had the highest glass adherence temperatures, 440, 510, and 465°C, respectively. - Metallic materials with high thermal conductivities - sterling silver, Consil^® 995, copper, and aluminum had the highest glass adherence temperatures of the metallic materials tested. The temperatures were, 418, 400, 375, and 403°C, respectively.
3. Increasing coupon thickness (i.e., increasing the thermal mass) raised the glass adherence temperature.
4. Highly polished surfaces did not provide a significant performance advantage compared to a machined surface.
5. Glass spalled from the CDA 706, Nickel 200 and the oxygen-free copper pre-oxidized at 900°C by delaminating at the oxide / metal interface.
6. Coatings intended to alter the surface chemistry of the coupon (i.e., sterling silver, gold, and chromium) did not have a significant effect on the glass adherence temperature, except for the polished dense chromium coating.
7. Changing the angle of the coupon surface from 45 to 80 degrees from the horizontal did not have a significant effect on the glass adherence temperature.

Material Selection and Performance

The focus of this study was the glass adherence temperature including the ability of the material to shed glass once it had stuck. In order to make the final material selection for the new bellows liner, mechanical and physical properties, i.e., corrosion and oxidation resistance, melting point, and elevated temperature yield and tensile strengths, were also evaluated. Other factors that were considered were product form availability, fabricability, affect on glass quality, and costs due to material and fabrication. The candidate material selected based on the above factors was Nickel 201.

A prototypic liner was fabricated from the Nickel 201, a lower carbon version of Nickel 200. The finished component was heat treated to establish a uniform oxide layer. Heat treating was required to provide the non-stick surface and to improve the video image of the pour stream by minimizing reflections. After several months of service, during which time severe pour stream instability occurred, the liner successfully shed molten glass. Unlike the 304L liner, glass that adhered to the nickel liner was easily removed during remote mechanical cleaning operations. As a result down time associated with cleaning was significantly decreased and service life of the liner was increased from 3 to 9 months. A second-generation nickel liner is currently being fabricated that will contain an awning over the view port. This will keep the glass from solidifying in the view port during severe pour stream conditions. It is expected that this awning will further extend the life of the bellows.

Acknowledgments

This work was supported by the U.S. Dept. of Energy under contract DE-AC09-96SR18500. The authors would like to express their appreciation to the following individuals for their commitment to this task, Dan Iverson, Robert Hopkins and Thadeous Reown.

References

1. McGraw, D.A., Transfer of Heat in Glass During Forming, Journal of The American Ceramic Society, Vol. 44, No. 7 (1961).
2. Falipou, M. and Donnet, C., Sticking Temperature Investigation of Glass/Metal Contacts - Determination of Influencing Parameters, Glastech. Ber. Glass Science Technology, Vol. 70, No. 5, (1977).
3. Pask, J.A. and Fulrath, R.M., Fundamentals of Glass-to-Metal Bonding: VII, Nature of Wetting and Adherence, Journal of The American Ceramic Society, Vol. 45, No. 12 (1962).