A Study of Chlorine in High Temperature Corrosion of Alloys in an FBC System

FINAL TECHNICAL REPORT

November 1, 1998, through October 31, 1999

Project Title: A STUDY OF CHLORINE IN HIGH TEMPERATURE CORROSION OF ALLOYS IN AN FBC SYSTEM

ICCI Project Number: 98-1/4.2B-1

Principal Investigator: Wei-Ping Pan, Western Kentucky University

Other Investigators: John T. Riley, Western Kentucky University; Ian G. Wright, Oak Ridge National Laboratory

Project Manager: Ken K. Ho, ICCI

ABSTRACT

The major purposes of this research project are to help the Illinois coal industry and TVA's FBC operators at the Shawnee Power Station understand whether the previously experienced corrosion and ash deposit problems with evaporating tubes in the FBC system may or may not be associated with the chlorine content in coals. Seven 1,000-hour combustion tests firing seven different coals were conducted in a 0.1 MW_th FBC facility at WKU with the operating conditions simulating to Shawnee Power Station's parameters. Seven coals were selected with various sulfur (S) and chlorine (Cl) contents from high S (4.48%) and high Cl (0.47%) to low S (0.97%) and low Cl (0.012%). For comparison, the coals selected consist of high S & low Cl; high S & high Cl; low S & high Cl; and low S & low Cl. Among them, the low sulfur and low chlorine coal was used as the baseline of erosion investigation. The calcium/sulfur ratio was held at 3 for all tests except the erosion baseline test which was controlled at 9 as a rate equivalent to that used for a 3% sulfur content and assuming a Ca/S ratio was being used. Three types of uncooled coupons (304SS, 309SS, and 347SS) were prepared and installed, based on the advice of TVA, to simulating the tubes in the superheater region [1,020-1,100^oF (550-600^oC)]. An air-cooled A210-C evaporating tube was adapted to simulate conditions of a 663-697^oF (340-370^oC) surface temperature in an 1500^oF (815^oC) surrounding flue gas environment. Metal wastage of each coupon was obtained by measuring the thickness before and after each combustion test. No Cl was found in the corrosion scale or on the metal surface. Alkali metals, K and Na, were only observed on the surface of the alloys. High S contents were found in the outer part of the deposits, and appeared to be associated with calcium and magnesium, suggesting that the fly ash may react further after being deposited on the surface of the metal. Locations of high S concentrations correlated well with the high chromium content of the inner layers of the corrosion scale of the samples, indicating that corrosion involved sulfidation attack. Sulfur is the major factor causing the corrosion of the metal. However, based on the 304SS coupon testing results, there was no significant corrosion found for all seven 1,000-hour tests in the pilot FBC units by extrapolating the data in terms of metal lost per year. This indicates that the calcium sorbent in the FBC system can capture not only the sulfur but also the chloride effectively. This effect helps bring the gas phase chloride concentration in the system below the level of less than 60 ppm. This reduction in Cl species in the gas phase has possible implications for decreased corrosion problems not only in the freeboard, but also in the cold end of the boiler.

EXECUTIVE SUMMARY

At least half the known economically minable coal reserves in Illinois have a chlorine content around 0.3%, which has been a concern for potential end users of these coals. Between 1992 and 1993, TVA's Shawnee Plant observed that the boiler tubes in the primary superheater region of the atmospheric fluidized bed combustion (FBC) system had wastage/corrosion problems. The system had been firing some Illinois coal and in 1993 TVA declined to renew an Illinois Basin coal (0.4% Cl) contract for approximately 4 million tons per year. The decision was made solely on the chlorine content of the coal. This concern is reflected in the suggested coal chlorine limits used by boiler manufacturers and in the application of coal chlorine content limitations in coal contracts.

It is known that FBC systems can absorb sulfur oxides with limestone in the combustor with high sulfur retention efficiency. Limestone can also capture hydrogen chloride at relatively low temperatures in the freeboard region of the FBC system. The scope of this project is to study the behavior of chlorine during coal combustion in an FBC system. The results from this research project could help reduce the fuel costs for the 160 MWe FBC unit at the Shawnee Steam Plant near Paducah, KY, if raising the specification limit for chlorine content is possible after completion of the study. A successful project demonstrating the utilization of high chlorine coals may open new markets for Illinois high chlorine coals. It will also reduce the risk of serious damage to commercial units using high chlorine coal, if power plant operators know the limits for utilization of such coals.

A three-phase investigation was carried out in order to study the fate of chlorine during coal combustion in an FBC system and to study the susceptibility of boiler components to corrode in the combustion gases containing hydrogen chloride. In Phase 0 (1995-1997), a preliminary stage, two 1,000-hour burns were conducted with a western Kentucky #9 coal (3.2% S and 0.012% Cl) and an Illinois #6 coal (2.38% S and 0.31% Cl) which was requested by TVA. In Phase I (1997-1999), three Illinois high chlorine coals (0.21, 0.47 and 0.42%) were burned for a total of 3,000 hours to determine the threshold limitation of chlorine in the coal which will cause severe metal wastage (if any is due to the chlorine). In order to study which element (sulfur or chlorine) is the major element influencing metal corrosion and erosion blank in the FBC system, two coals [one Kentucky coal (0.97% S and 0.026% Cl) and one blend (4.48% S and 0.41% Cl) of Illinois high chlorine coal with pyrite] were chosen and tested in Phase II in 1999. The major contribution in this report was focused on the results obtained from Phase I and II.

In order to honor the requests of TVA for corrosion tests under conditions simulating those of the 160 MWe FBC system at the Paducah Shawnee Plant, three types of uncooled coupons (304SS, 309SS and 347SS) were installed at the top of the freeboard to simulate the tube performance in the superheater region [1,020-1,100^oF (550-600^oC) surface temperature] for the three phases. An air-cooling method was adapted for a A210-C test evaporating tube to match the conditions of 663-697^oF(340-370^oC ) surface temperature on the tube, in an 1500^oF (815^oC) surrounding flue gas temperature during phase 1 and 2.

During the past four years, seven 1,000-hour runs were conducted in the 0.1 MW_th FBC facility at WKU using seven different coals with chlorine contents ranging from 0.012% ppm to 0.47% and sulfur contents ranging from 0.97% to 4.4%, including one corrosion blank test firing low chlorine (0.026%) and low sulfur (0.97%) coal. During the corrosion blank test (6^th 1,000-hour run), the limestone feeding ratio was kept the same as that used for a 3% sulfur content coal, making the Ca/S ratio around 9. The thickness changes of the test boiler tube (A210-C) and the uncooled steel coupons should correspond to the erosion caused by particles in the flue gas. In the Phase I and II study, five samples of cooled-boiler tubes (A210-C) were examined after 5,000 hours of testing in the FBC system. A parallel test was run with a boiler tube in an electrically heated furnace with the intent of obtaining information on the corrosion rate caused by sulfur and/or chloride in the FBC system and metal wastage rate caused by oxidation only during Phase I. Sixty-six test coupons (22 each of 304SS, 309SS and 347SS alloys) were analyzed after corrosion tests in the FBC system. The major operating parameters were as follows: combustion temperature -- 1450-1550^oF (788-843^oC); Ca/S ratio -- about 3.5; fluidizing velocity -- ~1.25 m/s; excess air ratio -- ~1.25; bed height -- ~1.1 m; surface temperature on the cooled test tube (A210-C) -- ~700^oF (371^oC); surrounding temperature of test tube -- ~1480^oF (805^oC); temperature of uncooled test coupons -- ~1000^oF (538^oC).

A 1,000-hour corrosion blank test with a boiler tube (A210-C) heated in an electric furnace under a simulated flue gas atmosphere (without SO_x and HCl) was completed. The tube was designed for two different experimental conditions by having two separate parts divided by an inner plate. One side was heated under a nitrogen atmosphere. The other side was heated in a flue gas atmosphere (15% CO₂, 5% O₂, and 1% CO in nitrogen). The purposes of this test were to (1) study the deterioration rate (temperature effect) of the A210-C tube in an inert atmosphere (N₂) at 700^oF (371^oC), and (2) study the corrosion rate (flue gas effect) of the A210-C tube in a flue gas atmosphere (without SO_x and HCl) at 700^oF (371^oC). By comparing the data from the two parts of the tube heated in the same furnace at the same temperature, the corrosion rate of the tube in a flue gas atmosphere (without SO_x and HCl) can be determined. The temperature conditions were: 725^oF (385^oC) in the electric furnace, with a temperature of 650^oF (343^oC) to 760^oF (404^oC) at the surface of the boiler tube.

A total of 140 cross-section species were prepared from the 70 coupons, and 80 cross-section samples were prepared from the five cooled A210-C tubes. The analysis of the corrosion products included examination with a Scanning Electron Microscope (SEM) system with an Energy Dispersive X-ray (EDX) spectrometer. In the work, corrosion of the two sections from each coupon was selected for investigation. One was the bottom section facing the gas and particle flow (6 o'clock placement), and the other is the top section (12 o'clock). In Phase I and II, for the stainless steel test coupons, samples were taken after 250, 500, 750 and 1,000 hours of exposure in the FBC hot flue gas stream system resulting from the firing of coal with various sulfur and chlorine contents. The A210-C tube was only removed from the combustor after the entire 1,000-hour test.

No significant corrosion was found for all seven 1,000-hour runs in the 0.1 MW_th WKU-FBC facility. The maximum metal loss for stainless steel and A210-C was 3.62 mils/year (10.5 µm/1000hours) and 28.3 mils/year (82 µm/1000hours) at the requested surface temperature, respectively. Upon comparison of the results obtained from coals 98011 and 99426 (with almost the same sulfur content, ~1.0%), the worst metal loss only increases by 0.97 mils/year (2.8 µm/1000hours) for 304SS, 0.86 mils/year (2.5 µm/1000hours) for 347SS and 0.28 mils/year (0.8 µm/1000hours) for 309SS after tests with coals having chlorine contents from 0.026% to 0.47%. At the same time the worst metal loss increased from 0.66 mils/year (1.9 µm/1000hours) to 2.62 mils/year(7.6 µm/1000hours) for 304SS, from 0.55 mils/year (1.6 µm/1000hours) to 2.03 mils/year (5.9 µm/1000hours) for 347SS, and from 0.31 mils/year (0.90 µm/1000hours) to 0.86 mils/year (2.5 µm/1000hours) for 309SS, respectively, when the sulfur content increased from 0.97% to 3.2% by comparing results obtained from runs firing coals 95011 and 99426. No chlorine was found in the corrosion scale or on the metal surface of the alloy samples. The alkalis (K, Na) were only observed on the surface of the alloys. In the outer layer of the corrosion scale, high sulfur contents were associated with calcium, silicon and magnesium, which indicates that the fly ash must react further after being deposited on the metal surface. There was a good correlation between the high sulfur concentrations and the formation of high Cr-containing inner layers in the corrosion scale. Ash deposits and erosion were the two major factors accelerating metal corrosion in the FBC system.

In order to study the relationship between metal corrosion and ash deposits, a total of 170 ash deposits were collected from the test coupon surfaces during phase I and II and analyzed . The analysis included determination of metal concentrations by ICP-AES spectroscopy, determination of sulfur content with a LECO SC-432, and determination of chloride by bomb decomposition (LECO AC-350) and ion chromatography (Dionex DX-120). The major components in the deposits were calcium-based compounds, with sulfur content around 10-15%, and chloride content around 0.5-5%. It was concluded that the calcium sorbent in the FBC system can capture not only the sulfur but also the chloride effectively, and as a result decreases the gas phase chloride concentration in the FBC system down to a level of less than 60 ppm. The concentration is unlikely that it contributes much to corrosion of the metal in the FBC system.

As to future research on high temperature corrosion in an FBC system firing high chlorine and sulfur coal, we are proposing a project to determine what is the optimal (critical) operating conditions to not cause corrosion problems when high chlorine coals are used. We believe this can be accomplished by controlling the Ca/(S+Cl) ratio and flue gas velocity in the combustor. It is well known that the desired temperature for capturing SO₂ by calcium-based sorbents is around 1562^oF (850^oC) and less that 1202^oF (650^oC) for capturing HCl. In an FBC system, the sorbent reacts with SO₂ first in the dense fluidizing zone. If the limestone feeding rate is chosen only based on the sulfur content when firing a high chlorine coal, the amount of limestone used will be too low to capture HCl effectively in the freeboard. Consequently, high HCl concentrations will remain in the flue gas. For the proposed project, one or two high chlorine coals can be selected. The study can be carried out by making a change in test conditions such that feeding the coal and limestone with three different Ca/(S+Cl) molar ratios and with three flue gas velocities. The results from the proposed project will help us determine the critical operating conditions to minimize corrosion when high chlorine coals are used in FBC systems.