Changes in shape and dimension of automotive parts during heat treatment are influenced by many parameters, some of which can be related to the pre-treatment stage, part design, furnace-loading characteristics and the heat treatment itself. The extent to which heat treatment influences distortion behavior can range from 20 to 30%  and sometimes higher. It is possible to reduce the overall distortion significantly by focusing on using a homogeneous gear material having low tendency for distortion, a gear design suitable for heat treatment and reducing stress during manufacturing. Through experimentation, we found that by using a standardized pretreatment stage, the influence of the heat treatment on part distortion can be estimated and the influence of quenching parameters on changes in gear shape and dimensions can be systematically investigated.
The distortion behavior of gear wheels and the influence of different process parameters on the distortion values were previously investigated  to compare gas and oil quenching using optimized conditions and process parameters. However, new gear materials and designs make further investigation necessary to optimize heat treating production. The influence of the process method (gas carburizing with oil or gas quenching and low-pressure carburizing with gas quenching) on the distortion was evaluated using two different case-hardening steel grades and optimized quenching parameters for two different part designs. The results cannot be compared directly with previous results  because the tests in this study were conducted using different material lots and higher quenching rates by means of helium quenching as opposed to nitrogen quenching.
Test procedure and equipment
Gas-carburizing with oil-quenching tests were carried out in a continuous, single-track pusher furnace with a preheating furnace, a heating zone with three tray positions, as well as seven for carburizing and three for the diffusion phase. Case depths of 0.6 - 0.7 mm (0.023-0.027 in.) at 550 HV were achieved using cycle times of 16.3 minutes between tray pushes.
Low-pressure carburizing with high-pressure gas quenching was carried out in a two-chamber furnace, generally using convective heating. Parts were carburized at a temperature of 930C (1700F), after which the temperature was lowered and parts were quenched using a nitrogen or helium pressure of 15 bar. Helium quenching in combination with an optimized impeller significantly increased the quenching rate, leading to core hardness values similar to oil quenching. Due to the enhanced quenching rate, only single directional gas flow from top to bottom was applied.
Gas carburizing with gas quenching tests were carried out in a double-row pusher type furnace equipped with a high-pressure nitrogen-gas quenching device. Quenching cell design, smaller load dimension and optimized impeller allowed achieving nearly the same quenching speed using 20-bar nitrogen as that using helium and/or oil quenching. Process parameters were optimized to achieve a case depth of approx. 0.7 mm at 550 HV.
Previous research  involved testing both standard and reinforced gear wheels (gear type 14H) having an outer diameter of 187.35 mm (7.375 in.) and a tooth face width of 23 mm (0.9 in.). Gears were made of SAE 5115/5120 (~ 16/20MnCr5) with a hardenability range at J10 = 33-37 HRC. The starting material was CSN 41 4220 rectangular ingot from Czech suppliers.
The study reported here compared the same 14H gear geometry with a new MQ type gear design (Fig. 1). Starting material was continuously cast German steel TL4221 (corresponding to SAE5115/ 16MnCr5) and TL4521 (corresponding to 20NiMoCr65). Hardenability at J10 was 34 for TL4221 and 43 for TL4521. Both materials have good hardenability in the upper band of the range according to their respective standards.
Loads with 25 gear wheels in 5 layers were heat-treated in the pusher furnace. The parts were placed on the 500 x 500 mm (20 x 20 in.) trays in a star shape on a three-point fixture. The same load carriers were used in the low-pressure carburizing furnace, but with 8 layers and a total of 40 gear wheels. Both horizontal and vertical loads were tested.
The quenchant selected was high-speed quenching oil Durixol W25. Oil temperature was varied between 90 and 150 C (~195 and 300 F) to investigate the influence of temperature on changes in gear shape and dimensions. The best results with respect to distortion were at an oil temperature of 120 C (~250 F) for the standard geometry and 150 C for gears having a reinforced geometry. Part geometry has a stronger influence on changes in shape and dimension than oil bath temperature.
The smallest distortion values were achieved at the lowest circulation rate of 320 rpm (Fig. 2). In tests using reinforced gears, a frequency changer was used to widen the speed range of the circulation propeller from 170 to 1,500 rpm. To use the complete range under practical conditions, an initial speed of 750 rpm was selected and changed to 1,500 rpm after 8 seconds. Speeds of 320 and 170 rpm were tested as well. Oil bath temperature was 150 C (for expected best results).
As shown in Fig. 3, increasing oil bath temperature reduced out-of-roundness and out-of-flatness by approximately 10%. Despite the increased oil circulation rate, only slight differences in out-of-roundness and out-of-flatness occurred at a temperature of 150 C. Changing the gear geometry considerably reduced the influence of oil circulation rate on distortion. However, a rate of 320 rpm produced the least distortion.
Low-pressure carburizing plus HPGQ
Convection heating was used in LPC+HPGQ tests. Convection heating promotes considerably less out-of-roundness of the inner bore than does straight radiation heating (a nearly doubled spread of measured values). Out-of-flatness and concentricity have larger spreads, but a negligible difference in the mean values from both measurements. Shrinkage of the inner dimension is not affected by the heating method.
Cooling gas flow direction in the quenching chamber was possible from top to bottom and vice versa, and the flow direction could be changed at regular intervals; typically about 15 seconds. Change the flow direction always reduced distortion values compared with unidirectional flow (Fig. 4). Furthermore, gas flow from top to bottom produced better results than the opposite direction. Reversing the gas flow starting from top to bottom gave better results.
Quenching gas direction also influences dimensional changes, showing a clear correlation with quenching rate. The change in gas flow direction lowers overall quenching speed, causing a slightly higher shrinking of the gear inner diameter. Highest quenching rates were achieved by quenching from the top, leading to reduced shrinking compared with the reverse gas flow direction.
Figure 5 shows the influence of gas pressure and loading on changes in shape and dimensions. Lowering gas pressure to 10 bar slightly increased out-of-roundness and out-of-flatness together with a considerable increase in the spread of measured values. Concentricity remained largely unchanged.
It is assumed that a 30 to 40% increase in the mean heat-transfer coefficient (corresponding in an increase of pressure from 10 to 15 bar) produced no critical change in the temperature gradient within the part during cooling, especially during cooling below the martensite-start temperature. On the contrary, higher pressure results in an increased uniformity of heat transfer within the load, which influences the reduction of changes in shape.
Vertical loading of parts provides a further improvement of the distortion values including the spread (out-of-roundness) and the mean value (out-of-flatness and concentricity). In addition to reduced distortion, a further increase of quenching rate and core hardness values was achieved, and the improved gas flow contributed to increased quenching uniformity.
Gas and oil quench distortion results
The most favorable conditions for both heat-treating methods were determined. A comparison of results refers to the entire heat treatment method, not just the influence of quenching medium on distortion. However, presenting results only as a comparison between gas and oil quenching is justifiable if one considers the extent to which heating (convective in both cases) and carburizing affect the total change in shape and dimension, and relates this to the influence of the different quenching methods.
The mean values of out-of-roundness, out-of-flatness and concentricity produced under these given conditions are presented in Fig. 6, which compares optimized results for standard and reinforced gear wheels after gas and oil quenching. In the case of gas quenching, results from vertical loading are also included.
For standard geometry gears, a substantial decrease of out-of-flatness (~38%) and concentricity (~34%), as well as a definite reduction of the spread, was achieved using gas quenching. Out-of-roundness was slightly improved, but the spread was somewhat larger. Vertical loading further reduced deformation to within 5% for out-of-roundness and 20% for out-of-flatness.
Results for reinforced gears deviated somewhat, with a 10% lower out-of-roundness after oil quenching than for gas quenching. The spread of values was also smaller after oil quenching, which differs from reports in the literature showing substantial improvements in distortion behavior when changing from gas to oil quenching . Nevertheless, these results are significant since the comparison of both methods under optimized conditions shows that the influence of component geometry greater than both the method and process parameters. A detailed interpretation of results can be found .
TL-material quenching results
German TL4221 material (similar to SAE 5115) was compared with Ni-alloyed material TL4521, which has better hardenability. Tests were carried out using the three different thermochemical processes (gas carburizing with oil or gas quenching and LPC with gas quenching) to determine the influence of the process on the distortion of the gear parts.
Gas carburize + oil quench
Gas carburizing with oil quenching was conducted using optimized process parameters of earlier tests. Figure 7 shows distortion values due to oil quenching. There is a close correlation between out-of-roundness and out-of-flatness values with 14H-design and TL4221 steel and previously published results (see Fig. 3).
TL4521material generally resulted in higher distortion values due to the higher hardenability of the steel. Similar experiences are reported in . The greater extent of distortion was observed for all parameters. TL4221 material showed the expected slight growth due to heat treatment, which can be compensated for in the design of the green part. In contrast, the Ni-containing material showed a significant shrinkage of the inner dimension.
The MQ design showed a slightly higher out-of-roundness compared with the 14H-design. Out-of-flatness was not significantly influenced by part design. Dimensional changes of the MQ design were generally smaller compared with the 14H-design. Mean values of distortion were rather uniform throughout the layers of the charge (Fig. 8).
LPC with gas quenching used helium gas at 15-bar pressure. An optimized impeller increased the helium quench rate by 60% compared with nitrogen quenching, which allowed using only single-direction gas flow from top to bottom. Shape and dimensional changes are shown in Fig. 9. Helium quenching did not produce the low distortion values achieved using nitrogen. Helium's faster quenching rate had less influence on out-of-roundness, but increased out-of-flatness and concentricity.
TL4521 material showed higher changes in shape compared with TL4221, similar to the results using oil quenching. MQ gear out-of-roundness was slightly less and out-of-flatness slightly greater than those for the 14H gears using helium quenching.
The shrinkage of the gear inner dimensions typically observed using LPC and gas quenching was higher for TL4221 type 14H gears despite the higher quenching rates of helium, which could be due to different material lots. MQ gears had significantly reduced shrinking of the inner diameter compared with 14H gears, similar to the results obtained with oil quenching. TL4521 gears had more shrinkage than TL4221 gears, which can be compensated by the soft-machined measures, since the spread of the measure all over the load is small.
Gas carburize + HPGQ
Only MQ gears were gas carburized and gas quenched. The nitrogen quenching pressure was increased to 20 bar and load dimensions were reduced to achieve a quenching rate comparable to those for oil and 15-bar helium. The heat treatment was designed to achieve a case depth of approximately 0.7 mm (0.027 in.) at 550 HV.
Figure 10 shows shape and dimensional changes of MQ gears made of both materials for gas carburizing and nitrogen quenching compared with LPC and helium quenching. With the exception of TL4221 material processed by LPC and helium quenching, out-of-roundness and concentricity for both materials and processes are comparable. Uniform heating and carburizing in combination with a comparable gas quenching process could cause similar changes in shape.
Out-of-flatness of TL4521 gears was significantly higher than for TL4221 gears, which correlates with other distortion results, and is due to its higher hardenability. Gas carburizing using nitrogen quenching produced significantly better gear-flatness results compared with the LPC process. Compared with LPC, MQ gears had greater shrinkage, especially those made of TL4221 material.
Figures 11 and 12 show the changes in shape and dimension for TL4221 and TL4521 gears due to the different treatments. TL 4521 caused higher changes in shape compared with TL4221. IH