As a rookie, the task of performing weekly burnouts of the atmosphere furnaces in our heat-treat shop fell to the Doctor. Perhaps it was the unbridled enthusiasm of youth, but the job was never boring and always filled with excitement as one tried to invent new and innovative ways to perform (and shorten) the seemingly simplest of tasks on a Friday night. It was only later that the importance of a proper burnout on the furnace’s ability to maintain precise process control and lower the frequency of furnace maintenance struck home. Over the years, many people have forgotten how or even why we perform this function in the heat-treat world. Let’s learn more.


The Why

Under a variety of operating conditions, carbon, in the form of soot, can build up inside a heat-treat furnace. The presence of soot adversely affects the life of the internal alloy components; is absorbed into the refractory lining, changing its thermal characteristics; and can interfere with effective heat and carbon transfer to the parts. Plugged gas inlets, malfunctioning oxygen probes and inaccurate temperature readings are some of the most commonly reported problems.

    Soot formation can also have devastating long-term maintenance consequences, requiring us to change out alloy fans, radiant tubes, roller rails, chain guides and protection tubes. The life of electric heating elements and radiant tubes are often shortened. Left unchecked, refractory life is dramatically reduced as well. 

    Case-hardening processes such as carburizing and carbonitriding are classic sources of excess carbon, but other processes (such as sintering of powder-metal parts) produce carbon due to the release of organic binders into the furnace atmosphere. A common source of soot is an endothermic or nitrogen/methanol atmosphere that is out of control (Fig. 1).


The How

For all of the above reasons and more, it is periodically necessary to remove this unwanted carbon from our furnaces. It is important to recognize, however, that the reaction of carbon with oxygen to form carbon dioxide is highly exothermic. For example, 96 kcal per gram mole of energy is released 925°C (1700°F). Thus, great care must be exercised to avoid overheating and/or creating localized hot spots, which will rapidly and catastrophically damage the furnace.

    It is customary to help control the carbon burnout reaction by lowering the furnace temperature in the range of 845-870°C (1550-1600°F) before commencement of the burnout process. If at any time during the burnout procedure the temperature rises 38°C (100°F) or more, the process should be temporarily halted.

    Here are the most common ways in which a furnace burnout can be accomplished:

1. With the atmosphere removed from the furnace, use a fixed volume of clean, dry and filtered air entering through a flowmeter (typically at a rate no greater than 20% of the normal atmosphere flow) for a prolonged period of time, typically 3-8 hours.

2. With the protective atmosphere still inside the chamber, raise the furnace dew point into the range of +15 to +21°C (+60 to +70°F) by introducing clean, dry and filtered air or water-saturated air for a prolonged period of time, typically 12-72 hours.

3. With the atmosphere removed from the furnace, introduce clean, dry, filtered and (flow and pressure) regulated compressed air through a steel lance or wand. Manually direct the air stream toward locations of heavy soot deposits for no more than 10-15 seconds at a time. This is then followed by the introduction of air through a flowmeter for a fixed period of time, typically 1-4 hours.


    While all methods are used in the industry, method 3 must be performed with extreme care to avoid damaging the furnace interior from overly high-pressure air, long-duration air impingement in a localized area or striking interior components such as the circulating fan (as the Doctor found out one day). Remember, soot (carbon) combusts at approximately 2480°C (4500°F), high enough to melt any of the materials present in the furnace.

    It has also been found[3] that carbon dioxide may be used either undiluted or with an inert carrier gas to accelerate the burnout process. The addition of small amounts of water (i.e. saturated nitrogen, air or carbon dioxide) can speed up the process as well. However, the use of oxygen (as opposed to air) is highly discouraged due to the rapid combustion and acceleration of temperature (as the Doctor again knows all too well).


The When

In the case of medium- to deep-case carburizing, weekly furnace burnouts are typically recommended (dependent to some degree on the dew point of the incoming endothermic gas). Monthly atmosphere burnouts should be adequate in most other cases. To an extent, the time between burnouts (and their duration) is dependent on when the last burnout was conducted.

    During normal operation, it is important to try to keep furnaces as soot-free as possible by first making sure they are running the right gas flows at the proper carbon-potential setpoint. The enriching gas additions being used should be limited over the course of the total cycle to an acceptable percentage of the total flow (typically 10-15% in the case of natural gas).

    It is noteworthy that most automatic systems add enriching gas in relatively short increments of high flow. These should be timed to determine exactly how much enrichment is actually entering the furnace, and in many cases the peak flow should be limited to prevent too much enriching gas from entering the chamber in too short a time.

    One of the many consequences of an improper or partial burnout is that carbon absorbed into the refractory continues to diffuse inward. This creates a situation in electric furnaces where the refractory becomes conductive and element terminals or alloy extending through the refractory may melt.


The Oxygen Probe

With so many furnaces using oxygen (carbon) probe control systems, it is worth briefly talking about them. The recommended probe burnout frequency, flow and duration vary by manufacturer based on sheath diameter and tip design but are also dependent on the process-carbon-level setpoint. In the case of gas carburizing, every four hours is a common practice using 0.05-0.3 m3/hour (2-10 cfh) of air piped to the burn-off fitting on the head of the probe. Pumped room air or filtered combustion air are most commonly used. Compressed air is not recommended due to the contamination of water and oil often present that can damage the oxygen probe. It is good practice never to exceed 90 seconds to avoid overheating the tip of the probe.

    A consistent way to verify a correct burnout is to monitor the oxygen millivolts of the carbon controller during the burnout phase. If a proper burnout is taking place, the oxygen will drop below 200 millivolts. This can also vary based upon the circulation in the furnace and the probe placement.

    A possible side effect of extended burnout duration is the oxidation of the tip of the sensor. The problem can manifest itself with oxygen millivolts being elevated over time, which will require a lower CO-factor setting for the same calculation of carbon.


The Generator

Today, many endothermic gas generators are run at dew points in the +1.5 to +7°C (+35 to +45°F) range. The primary reason for this is to reduce the frequency of generator maintenance. The consistency of the endothermic ga

s produced at the generator significantly affects all downstream processes. In case hardening, the volume of enrichment gas that is required to achieve a given surface carbon content is affected. The generator should be controlled by a dew-point analyzer and monitored with a three-gas analyzer (CO, CO2, CH4) so as to detect if the generator catalyst is sooted. The CH4 value rising above 0.5% carbon is an indicator that a generator burnout is needed. If the value does not fall below this threshold after burnout, a catalyst change is required.


The Conclusion

A successful atmosphere burnout will return the furnace to productive service. Once the furnace burnout is complete, it is natural to ask if one can begin running heat-treat cycles immediately afterward. While possible for many processes, it has been suggested[5] that several conditioning runs be performed under neutral atmosphere prior to the onset of carburizing or carbonitriding. The secret is to avoid the control system making unnecessarily high enriching-gas additions to compensate for the reabsorption of some carbon by the furnace after a burnout.  IH


Soot formation on the front load table of a mesh-belt furnace after one hour with an out-of-control furnace atmosphere 


1. Kelly, James, “Understanding Conditions that Affect Performance of Heat Resisting Alloys, Parts I & II,” Industrial Heating, March/April 1979.

2. Herring, Daniel H., and Gerald D. Lindell, The Need For Periodic Atmosphere Furnace Maintenance Part Two: Lessons Learned, Heat Treating Progress, November/December 2009.

3. Peartree, Robert, Method for removing carbonaceous deposits from heat treating furnace, 1982, Air Products & Chemicals Inc. Patent No. EP 0047067A1

4. Protective Atmospheres, Measurement Technologies and Troubleshooting, white paper, Super Systems Incorporated.

5. Pye, David, “Atmosphere Furnace Burnout,” The Experts Speak Blog, Industrial Heating, 2011.