Burner with catalytic cleaning on a production furnace

High-temperature industrial processes often produce excessive NOx emission levels, especially in instances where preheated air is used to increase efficiency. Different measures are used to reduce NOx formation, such as fuel and/or air staging and flue-gas recirculation. Flue-gas recirculation and mixing before the flame reduces NOx formation even further. Two examples are the Flox (flameless oxidation) and the Gaft (gas dynamic abated flame) burners. In radiant tubes, Gaft burners are reported to have NOx emissions of 25-45 ppm at a furnace temperature of 900C (1650F) using preheated air [1]. Flameless oxidation occurs at extremely high recirculation; the flame temperature is reduced and more evenly distributed resulting in NOx formation. However, the combustion process requires a furnace temperature higher than approximately 850C (1560F) [2].

Catalytic cleaning is used in processes where the NOx emissions are very high, such as power plant boilers, glass melting and in combustion engines. Catalytic cleaning has been tested in the Netherlands using an automotive three-way catalyst, but is not a common technology in radiant-tube burners [3]. NOx reductions exceeding 90% are reported using an automotive three-way catalyst, but high flue-gas temperatures and high thermal mass in the catalyst are possible reasons for catalyst degradation (failure). A NOx reduction of 40 to 50% was achieved using an alumina catalyst developed by Osaka Gas at the burner exit [4].

The FeCrAl catalytic system (patent pending) discussed in this work has been tested both in laboratory conditions and on an industrial scale and fits most gas burners on the market [5]. Test results show that the technique is reliable with good reproducibility, and no noticeable major degradation after 4000 operating hours. The catalytic cleaning is suitable for furnace temperatures up to at least 1000C (1830F) and at fuel input levels (of full burner input) of 50-100%. A catalytic gas radiant tube system (Kanthal GAS) will be introduced commercially based on the test results.

Fig 1 Schematic of the catalytic cleaning system

Technology Description

The principle of this technology is to operate the burner close to a stoichiometric air/fuel ratio to obtain a suitable environment for the catalyst. Formed NO is converted in the catalyst to N2 according to the basic reaction: NO + CO to N2 + CO2

The catalyst is placed between the recuperator and the outer radiant tube where the temperature is appropriate for the catalytic reactions. CO is oxidized in a catalyst located farther downstream. Air for the oxidation is supplied at a point before the catalyst. Figure 1 shows a schematic of the burner and catalysts.

Fig 2 Efficiency as a function of exhaust gas temperature and excess air ratio (_)

Excess air ratio (_) is referred to at two places in the burner system: after the three-way catalyst (_1) and after the oxidizing catalyst (_2). A reduced excess air ratio gives a potential for increased efficiency. The thermal efficiency for different excess air ratios and flue gas temperatures is shown in Fig. 2. Reducing the excess air ratio from 1.30 to 1.00 increases efficiency by 4 to 5%. A similar increase in efficiency also is obtained by reducing the flue gas temperature about 100?C (180?F). However, this potential efficiency increase is offset to a certain extent by the heat content of unburnt fuel. The unburnt fuel originates from the close to stoichiometric combustion. Evaluating the overall effect is difficult because heat-exchanger parameters also are affected. The catalytic wire mesh also may improve heat transfer, especially in recuperators without extended surfaces.

Fig 3 Oxidizing catalyst sheet (actual diameter = 100 mm, or 4 in.) Fig 4 Wire mesh (16 mesh) for an oxidizing catalyst, which has been used for 2200 hours.


The catalytic devices are based on a wire mesh structure developed by the Swedish company Katator {6, 7, 8]. The wire mesh is coated with palladium and rhodium on the three-way catalyst and coated with palladium on the oxidation catalyst. Both catalysts are folded to increase the overall catalyst surface. A number of catalyst sheets are put in a pile to obtain the necessary catalyst surface, and the catalyst is easily cut to fit the available space. A single sheet for the oxidizing catalyst is shown in Fig. 3, while Fig. 4 shows a magnified image of the mesh.

Advantages of the mesh design in reducing the NOx emissions of existing furnace equipment include a low pressure drop while still providing a high mass transfer rate and the ability to cut the mesh in the necessary shape to fit the actual burner.


The NOx reducing technology was tested on several commercially available radiant tube burners representing traditional designs and new low-NOx designs. All burners incorporated a recuperator, and all burners were mounted in a single ended radiant tube.

Burner A has a ceramic recuperator and a maximum fuel input of 14 kW. Gas and air are mixed and combusted in a bulb tube having a small outlet diameter. A separate inner tube allows recirculation of exhaust gases.

Burner B is an all-metallic recuperative burner. Fuel input is 10 to 12 kW (version B1) and about 16 kW (version B2). Fuel and air are mixed in a swirling device, and premixing (adding a small amount of combustion air into the fuel flow) was added to the design. The burner operates in high and low (3 kW) fuel inputs. A continuous inner tube in the original burner design was cut and a separate inner tube was used instead. The burner head was modified to increase the exhaust gas velocity and provide recirculation. The modifications also show the importance of the burner head design on the CO formation at the operating conditions for catalytic cleaning.

Burner C is an all-metallic recuperative burner having a maximum power of 16 kW. Fuel and air are partly premixed before combustion. The burner was redesigned in the same way as burner B.

Burner D has a different mixing device and uses preheated combustion air. The inner tube is divided, but the space between the burner head and inner tube is small. Maximum fuel input is 25 kW.

Burner E, the largest burner, has a maximum fuel input of 45 kW. The burner was changed in the same manner as for burner B.

Fig 5 Test furnace used for catalytic flue gas cleaning: Furnace dimensions in mm. Outer tube: 91/78 mm dia. by 880 mm long. Inner tube: 64/56 mm dia. by 700 mm long.

Test Conditions

Laboratory tests were performed in a specially built laboratory furnace designed for a single radiant tube (Fig. 5). The burner is mounted on the furnace roof. Burner B1 was used for endurance testing.

Burner B1 had the NOx reducing catalyst located at the recuperator. The oxidizing catalyst, consisting of 11, 100-mm (4 in.) sheets, was placed in a special annular box on the flue gas outlet and covered the entire cross section; total sheet area was 864 cm2 (134 in.2).

Kanthal APM material was used in the radiant tubes, having a surface heat flux of 4.5 W/cm2. Temperature distribution was uniform along the radiant tube during the tests. The temperature of materials and air and flue gases was measured using thermocouples.

The natural gas composition for the endurance test was 91.1% CH4, 5.25% C2H6, 1.4% C3H8, 1.4% C4H10, 0.5% CO2 and 0.6% N2, corresponding to a lower heating value of 39.0 MJ/m3. In the other tests, the methane content was lower (86-88%), while the other hydrocarbon content was slightly higher. The furnace temperature was kept constant at 950C (1740F).

Endurance Test

Tests were conducted basically to clarify the performance for one burner design (B1) and to obtain long-term catalyst performance. The best performance was obtained for the parameters listed in Table 1.

Fig 6 NOx and CO emissions for burner B1 in endurance test. The total operating time exceeds 4000 hours.

The redesign of burner B1 shows the importance of fuel and air mixing and the primary CO formation for overall acceptable performance. Pressure drop across the burner was not significantly increased with catalysts installed. Burner B1 was operated with air and fuel pressure of 2.6 and 2.2 kPa (26 and 22 mbar), respectively, at the burner inlet. Data for NOx and CO emissions after 4000 hours of operation plotted in Figure 6 show the cleaning performance is kept within an acceptable range.

Fig 7 NOx and CO emissions as a function of excess air ratio after the oxidizing catalyst, _2 (long time test of burner B1)

At lower combustion-airflow rates, CO emission is high and the NOx emission is low while the opposite occurs at higher airflow rates. An intermediate airflow area (a "green window") exists where both the NOx and CO contents are low (<50 ppm). The width corresponds to an airflow of I 0.5 m3/h. The corresponding _1 value is 0.96I0.03. Figure 7 shows the NOx and CO emissions as a function of the excess air ratio after the oxidation catalyst (_2).

There is a _2 range (1.12I0.03) where both the NOx and CO contents are low; the range is wide enough to allow control of the excess air ratio. Temperature uniformity was good (about I 20C, or I38F) along the outer tube. At a furnace temperature of 950C (1740F), the radiant tube surface temperature was measured at 990 to 1010C (1815 to 1850F).

Performance Tests With Other Burners

Tests using a number of commonly used burners for radiant tubes, both in laboratory and on a production site, were conducted to test catalytic cleaning in various burner designs. Table 2 summarizes results for burners in a lower fuel input range. Substantial NOx reduction was obtained in several operating conditions.

Table 3 shows the corresponding results for two burners in a higher input range. The method has been successfully verified for five different industrial burners in eight different tests, both in controlled laboratory conditions and in real industrial conditions.

On average, NOx emissions after the catalysts were 62 ppm, or approximately 30 mg/MJ. This value can be further reduced using a larger catalyst. However, the possible catalyst area is limited by the available space in the recuperator. There is a slight, but not significant, tendency toward higher NOx emissions at higher fuel inputs. CO emissions from all burners are in an acceptable range.

It is possible to apply catalytic cleaning in several existing burner designs with a significant NOx reduction potential. Together with an efficiency increase, the overall NOx reduction for these burners was 71-78%.

Operation in an industrial environment was tested where 2 burners out of 54 were equipped with catalytic cleaning in a 730-kW continuous carburizing furnace operating at a temperature of 950C (1740F).

Fig 8 NOx and CO emissions for burner B2 in production site test
Figure 8 shows the results with the NOx and CO emissions as the combustion airflow is changed.

Fig 9 NOx and CO levels as a function of fuel input for burner B2

Effects Of Varying Parameters

Satisfactory operation also is obtained when the burner is operated in a part-load mode. Figure 9 shows the results of operating Burner B2 at different fuel inputs. There is an increased emission level but still at a fairly low level.

Burner B1 was operated at furnace temperatures between 300 and 975C (570 and 1785F) to evaluate the effect from different furnace temperatures. Emission levels were stable at around 40 ppm below temperatures between 500 and 975C (930 and 1785F).

Fig 10 Air/fuel ratio control system based on _2 measurements

Control System

No air/fuel-ratio control was used during the tests. Air and fuel flow were stable using standard equipment regarding pressure reduction and flow. A dedicated control system will give the system even better stability and maintenance-free operation.

A way to measure _2 has been developed, which offers an appropriate method to control the system and to ensure that the operation will be kept within the green window [9]. Figure 10 shows a way to build such an excess air control system based on _2 measurement. The secondary airflow is fixed, and, thus, the measurement is appropriate to control the combustion airflow.

Global Reduction Potential

The global NOx reduction potential for gas fired radiant tube heaters could be estimated assuming 10,000 batch hardening furnaces having 12 burners in each furnace, and 2,500 continuous furnaces, each equipped with 50 burners [10]. This amounts to 245,000 single-ended radiant tube heaters where this catalytic method can be used. If the NOx reduction is 87% (due to both the direct reduction and the efficiency increase) and the average burner input is 20 kW, the total annual reduction becomes 10,000 tons.

However, the most common radiant tubes are U and W shaped. The catalytic cleaning method also can be used in such systems. Assuming 10,000 furnaces, each having 40 burners, gives a total of 400,000 burners. Burner input is higher in these systems (approximately 80 kW), and using the same emission levels as before gives a total annual NOx reduction of 71,000 tons.

NOx emissions from different industrial sectors in the USA are shown in Table 4 [11]. The catalytic cleaning method could be used in metals and steel sectors, and an estimated reduction of 81,000 tons annually represents a significant part of the total emissions.

For more information, contact Thomas Lewin, senior applications engineer, Kanthal AB, Box 502, S-734 27, Hallstahammar, Sweden; tel: +46 220 211 59; fax: +46 220 211 53; e-mail: thomas.lewin@kanthal.com