Figure 1

Tubes damaged by erosion-corrosion because of the erosive action of the sootblowers in a sulfidising environment

Figure 2

The topotactically grown corrosion scale. In the oxide layer many sulphides are found. The sulphides have also grown in the cracks of the oxide

Figure 3

At the metal-oxide interface the element Cl is sometimes found. In a few cases this resulted in severe intergranular corrosion

Figure 4

A Scanning Electron Microscope image, the places of presence of heavy metals are blue. In the deposit layer a sharp border is noticed, under which heavy metals were not found

Figure 5

Typical Alligator Skin Cracks in one of the tubes at the rear water wall

Figure 6

The high-sulphide layer on one of the corrosion probes, after 5 weeks of operation

Figure 7 The oxide layer after adjustment of the burners and air curtain introduction. The layer is more protective now

Corrosion in Low NOx Firing Conditions

W.M.M. Huijbregts (KEMA Nederland B.V), J.B. Janssen ( UNA), O.C.J. Rens (UNA). (paper 50)

Power Gen Conference, June 9-11, 1998, Milano

pdf available

Rather severe corrosion was observed in the 680 MWe coal fired power unit of Hemweg 8, installed with Low NOx burners. Microscopic examinations of the tube sections showed that corrosion over approx. 30% of the side walls occurred because of:

• sulphidation
• chloride enrichment on the metal-oxide interface
• thermal fatigue cracks in combination with sulphidation and chloride enrichment on the crack tips (Alligator Skin Cracking).

To stop the corrosion various countermeasures were taken:

• Installation of gas sniffle points in the waterwalls; the following gasses were measured: CO, CO2, O2 and occasionally H2S
• Application of curtain air and adjustments of the burners so that the gas environment at the water wall contained less than 2% CO
• Installation of corrosion probes, the probes can be removed and new ones can be mounted during operation
• From regular microscopic examination of the probes the quality of the layer could be derived judging from the amount of sulphides and porosity
• The corrosion rate constants in dependence of temperature and reducing gas composition were determined in laboratory tests and by means of these corrosion rate constants the material loss of the waterwall tubes was calculated.

The supercritical coal-fired boiler at Hemweg 8 has 18 burners on 3 rows each in the front and the rear wall. The boiler has a capacity of 680 MWe. The pulverised coal blend had a maximum of 1% S and 0.1% Cl. This is the generally used coal blend for all Dutch power stations. For a short while a small amount of sewage sludge was co-fired as a demonstration project. The amount and the composition of the sludge was such that the coal-sludge blend did not differ from the composition of the usually applied coal-blend regarding sulphur, chlorine and heavy metals.

The boiler came into operation firing coal in October 1993 and the first corrosion was noticed in December 1996, because of a leakage near a sootblower. This tube had failed because of a combination of sulphidation and the erosive action of the sootblower. See Figure 1.

Microscopic examination of the corrosion scales in detail also showed that the scale was typical of reducing atmospheres. Tube wall thicknesses were measured over the whole water wall and several areas with severe corrosion were found. The zones most attacked in the side walls (with wall thicknesses smaller than 4 mm) were provisionally repaired by surface welding in January.

The boiler was taken into operation again in February 1997. In May 1997 the boiler underwent massive repairs of the water walls and by means of 36 air nozzles, curtain air was introduced. Gas sniffle points were mounted at 40 locations in the water wall. Besides, corrosion probes (KEMCOP system: KEMa COrrosion Probe) were mounted for regular examination of the corrosion scales formed.

In August 1997 the boiler came into operation again. The wing burners were adjusted during operation to influence flame shape and air ratio, which had the most significant result in optimising the gas atmosphere at the side walls.

2. Examination Boiler Tubes

The boiler tubes were examined by means of various techniques:

• Advanced optical microscopy
• Scanning Electron Microscopy
• Radiograph Micro Analysis.

In the area of the sootblowers erosion-corrosion of the tube occurred. Because of the poor quality of the oxide layer and the high pressure of the sootblowers, the scale was blown away regularly

The scale on the most corroded tubes in the side walls consists of various layers:

• a topotactically grown oxide layer with many sulphides in it. The topotactically grown oxide layer should normally provide the protection from severe corrosion. Because of the porous structure and the large amount of sulphides, the protection effectiveness of this layer is very low. See Figure 2.
• besides, at the oxide-metal interface the element Cl was found and sometimes also intergranular corrosion See Figure 3.
• an epitactical layer with many coarse crystalline sulphides and oxides and in between fly-ash particles the sulphides were present as FeS and as FeS2, the latter species indicating a relatively severe sulphidising atmosphere
• a deposit layer of fly-ash particles; in this layer an enrichment of heavy metal compounds was found.The metals found were: Pb, Zn, Sb, Sn, Ni, Cu, Ge, Se and As. It was typical that the heavy metals were mainly present in the outside deposit layer. A very sharp border was often present under which the heavy metals were not noticed. See Figure 4.
• in some tubes on the outside of the deposit layer, round particles with dendritical structures were found. This indicates fast solidification of molten ash and pyrite particles. See Figure 4. The composition of these particles varied from iron sulphides to CaAlKFesilicates (fly-ash)
• many tubes had deposits of black carbon, originating from the flame impingement process.

Some of the tubes at the rear wall in the neighbourhood of the burners showed first indications of Alligator Skin Cracking (ASC). See Figure 5. When a sulphidising atmosphere is present under moderate thermal fatigue conditions the oxide scale is less protective and the cracks are deeper and filled with sulphide containing oxide. In this situation the phenomenon is called ASC. Even with the moderate amount of chlorine content in the coal, volatile Cl compounds, like HCl and FeCl2, will condense in reducing atmospheres at the coolest places in the oxide layers, the oxide-metal interface. The element Cl is also often found at the crack tip of the ASC cracks.

The composition of the gas at the sniffle points was measured, as soon as the boiler came into operation in August 1997. Over early the entire side walls the gas had a rather reducing composition. As long as the CO amount was more than 3 v % the gas was considered as too reductive, a properly protective oxide layer cannot be formed. This was concluded from practical experiences (DOE) and from theoretical thermodynamic calculations of the sulphur, chlorine and oxygen partial pressure and the corrosion products that could be considered as the thermodynamically stable ones. If sulphides and chlorides are the thermodynamically stable ones, then the chance of formation of properly protective oxide layers is very low. So the burner settings were regulated in such a way that the gas composition was optimal for the formation of a properly protective magnetite layer.

The first probes were already removed as soon as 5 weeks after the restart of the boiler. The burners were not yet adjusted properly. The corrosion scales of the probes were examined by means of microscopic techniques. The scale still contained many sulphides and a thick deposit layer of fly-ash and even heavy metal compounds were noticed. See Figure 6.

After a period of approx. 4 weeks under conditions with adjusted burners, the oxide layer improved. Fewer sulphides and deposits were present on the probes. See Figure 7.

Corrosion exposure tests on the 13CrMo4.4 steel were conducted in a gas mixture with the same sulphidising grade as that in the Hemweg 8 boiler. The temperature was chosen in the range of 350 to 450 °C and the exposure times were only ca. 40 hours.

Nevertheless, relatively thick oxide and sulphide layers were formed. From the thickness of the topotactical corrosion scale the corrosion rate constant was calculated, assuming a parabolic corrosion rate. Despite the poor protective oxide layer in sulphidising atmospheres, the corrosion will decrease at longer times according to a parabolic law. This was concluded by R. John (3) and later also in our own examinations on samples exposed in coal gasifier pilot plants and in laboratory exposures (2).

The corrosion rate can be coupled to the gas mixture by introducing a gas factor (G) in the corrosion rate equation.
X = Kt 0.5
logK = ao + a1 G + a2 1000/T

We define the gas factor G as being:

G = gas factor = log(pS2)+0.5log(pCl2)-0.75log(pO2)

The pS2, pCl2 and pO2 pressures are dependent on temperature and for each gas composition these pressures can be calculated thermodynamically. So for each temperature the gas factor was calculated too.

The first series of experiments were conducted in a temperature range of 400 to 450 °C and at a gas factor range from -6.5 up to 9. By means of regression analysis the constants a0, a1 and a2 could be determined for the corrosion rate constant K being 3.09, - 3.77 and 0.03, respectively. (These values and the Figures 8 and 9 differ from the original figures because of a mistake in the original text.)

In the Figure 8 below the measured and calculated K values are presented. In more recent publications such regression data were also published (4,5).

Knowing these constants and the gas compositions on the water wall, the corrosion rate of the tubes can be calculated. In Figure 9 below an example has been given of the thickness of the corrosion scale for , assuming various gas factors for a metal temperature of 400°C.

• Corrosion in the Hemweg 8 boiler was a result of low NOx corrosion. Reducing atmospheres were present at the water wall, so that a properly protective oxide layer could not be formed.
• Such a reducing atmosphere was a result of misadjustment of the burners and lack of an air curtain.
• It appeared that the shape of the flame plays an important role in creating the optimal atmosphere at the walls for prevention of low NOx corrosion.
• No clear indications were found that co-firing of sewage sludge has influenced the corrosion process.
• The amount of sulphides in the topotactically grown oxide layer is a measure for the protection effectiveness of the scale.
• The corrosion on the water wall tubes can be followed very easily by installing corrosion probes (KEMCOP's) and determining the quality of the oxide layer at regular times.
• Measurements of the gas composition by means of sniffle points in the water wall help to control the burner adjustment. The CO-content should be at least less than 2 v%.
• By means of the corrosion rate constants, determined in laboratory experiments for various gas compositions, the corrosion rates of the water wall tubes can be calculated.
• The gas measurements and the corrosion probes are considered to be good tools to follow possible corrosion and fouling of the water walls, in particular when various coal blends and co-firing are applied.

1. Bakker W.T., Perkins R.A., Liere J. van Journal of Material Performance 1985, 24, 9
2. Huijbregts W.M.M., Kokmeijer E., Zuilen H.G. van, Sulfidation, down-time corrosion and corrosion assisted cracking on high alloy materials in synthetic coal gasifier environments. Conference on materials for coal gasification power plant June 1993, Petten The Netherlands Materials at high temperatures vol. 11 numbers 1-4 1993
3. John R., Predicting corrosion mechanisms and rates for Alloys in High Temperature Complex Gasses. Corrosion, 1990, April 23-27, paper 279
4. Kung S. Prediction of Corrosion Rate for Alloys Exposed to Reducing/Sulphidising Combustion Gasses. NACE Conference Corrosion '97. March 1997, paper 97-136
5. James P.J., Pinder L.W. The impact of Coal Chlorine on the Fireside Corrosion Behaviour of Boiler Tubing: A United Kingdom Perspective. NACE conference Corrosion '97 March 1997, paper 97-133