Wear Behavior of Roller‐Burnished High Interstitial Austenitic Stainless Steel Parts in Glass‐Reinforced Plastic Melt

Herein this investigation, profiled high interstitial austenitic stainless steel parts are burnished on a profile‐rolling machine, and afterward, the wear behavior is analyzed in a melt of glass‐reinforced polypropylene. Wear and corrosion resistance are significant properties of steel parts in the plastics and food industries. The high work‐hardening ability of high interstitial austenitic stainless steel enables burnishing parts with a significant local hardness to increase up to maximum values of ≈600 HV 1. Wear tests on a recently developed test stand reveal that the burnished austenitic stainless steel surface performs similarly to a nitrided surface of the standard nitriding steel 31CrMoV9 + QT with a hardness of ≈830 HV 0.5. Regarding the given advantage of corrosion resistance, it is concluded that roller burnishing supports the applicability of high interstitial austenitic stainless steel in plastics and food industries.


Introduction
In plastics processing, wear and corrosion of the machine parts, for example, extruder screws, determine product quality and machine availability. High corrosion and wear resistance at the same time require a complex manufacturing process.
However, burnishing steel grades with a significant work hardening ability enables a hardness increase that would conventionally require a heat treatment for martensitic steels.

Protection of Extruder Screws against Wear and Corrosion
For production of technical plastic components, the polymer often contains reinforcing materials such as glass fibers or spheres. Further additives like flame retardants or color pigments are mixed into the polymer. [1] The processed polymer and additives as well as the processing parameters temperature, pressure, and screw speed significantly influence wear and corrosion of the extruder screw. Manifold wear protection methods have been developed in recent years. A hard and smooth surface shows a wear-reducing effect. [1] Extruder screws are conventionally made of steel. Nitrided screws made of martensitic steels are a standard wear protection method. Moreover, hard facings on basis of Fe, Co, or Ni can be welded onto the screw tip. In the case of high wear and/ or corrosion requirements, material composites are manufactured by hot isostatic pressing (HIP). HIP enables to manufacture screws with a hard flight surface and sufficient material toughness in the screw core for torque transmission. [2] Manufacturing of screw geometry and realizing the required wear protection are conventionally separate process steps, which requires hard machining at the end of the manufacturing process. The positive effect of work hardening in austenitic steels on wear resistance against solids and slurries with mineral abrasives has been shown in literature. [3,4] However, work hardening is currently not used for wear protection of extruder screws.
Model tests with samples of a standardized shape are typically applied to evaluate the wear resistance of different screw materials. For testing against melted plastic, a small rectangular platelet made of the wear protecting metal and layer is used. [5] stainless steels can be applied, which feature a certain level of resistance to corrosion combined with favorable mechanical properties. [6] In a recent study, it has been shown that particular grades of austenitic stainless steels, which in general are characterized by even higher resistance to corrosion, lead to promising results when being used for extruder screws. [7] Pin-on-disc wear testing against solid glass-reinforced polypropylene resulted in similar wear resistance compared with martensitic nitriding steel. These particular austenitic stainless steel grades are called high interstitial austenitic stainless steels (HIS), since in contrast to conventional FeCrNi austenites, they are based on the alloying system FeCrMn and contain large amounts (around 1 mass%) of interstitial alloying elements carbon and nitrogen. Subject to a proper solution annealing treatment, both C as well as N are entirely dissolved in the austenitic matrix, which results in superior properties, including high strength and toughness, good corrosion resistance, and intense work hardening ability. The latter in combination with high austenite stability allows for further strengthening by cold deformation without virtually any straininduced phase transformation, which may result in hardness levels that are comparable to martensitic steels, still remaining nonmagnetic. [8] Resistance to corrosion, in particular to localized corrosion, is improved by addition of molybdenum, and is maintained even at high degrees of cold deformation. [9,10] While resistance to general surface corrosion of HIS, which was tested by means of potentiodynamic measurements in sulfuric acid, turned out to be similar to that of conventional FeCrNi austenites, the same experiment using sodium chloride solution revealed significantly improved resistance of HIS, particularly of those alloyed with molybdenum. [9,11] This was attributed to suppression of initiation of pitting corrosion and enhanced repassivation behavior, which is induced by the combined effect of nitrogen and molybdenum. [12,13] Decreased corrosion resistance of FeCrNi austenites after cold working is related to strain-induced phase transformation or decreased stability of passive layers. [14][15][16] In turn, the reason for the high resistance of HIS even in highly cold-worked condition can be attributed to their high austenite stability [16] as well as a very stable passive layer. [10] Consequently, the described combination of properties makes high interstitial austenitic stainless steels promising candidates for application in processing of plastics.

Roller Burnishing on Profile-Rolling Machines
Roller burnishing is applied to machined preforms to smooth the surface and improve geometric accuracy. Deep rolling is used to increase the strength of a part. Simple rolling tools with a ball can be used on lathes or machining centers for burnishing of cylindrical parts. High contact pressure between tool and workpiece causes plastic deformation of the part subsurface. Deep rolling of steel with a metastable austenitic structure shows potential to shorten process chains by avoiding a final heat treatment after machining. [17] Thereby, hardness is significantly increased by strain-induced martensitic transformation. Moreover, work hardening can cause a significant hardness increase of the workpiece material. To predict hardness after work hardening, in refs. [18,19] correlations between plastic strain or yield stress calculated in finite element analyses (FEAs) and measured hardness values were found for different steels. Burnishing profiled parts on a profile-rolling machine requires complex, profiled rollers but enables higher plastic strain than using simple ball type tools. In ref. [20] a significant hardness increase is shown for axisymmetric and helical profiles that were burnished on profile-rolling machines.

Aims of This Investigation
In this investigation, preformed axisymmetric workpieces were burnished on a profile-rolling machine to smooth the surface and work harden the subsurface while finishing the part geometry. Furthermore, a wear test stand was developed for analyzing the wear resistance of the burnished parts that have a geometry different from the standard platelet sample. In addition to wear testing against solid polymers shown in a previous study, wear was analyzed against melted polymers. Thus, conditions closer to that in extruders were simulated. The wear results shall show that burnishing of austenitic stainless steel parts may serve as an additional wear protection method for plastics processing machinery in corrosive environments. Burnishing would enable shortening of conventional process chains for wear-and corrosion-protected steel parts.

Numerical Simulation of Burnishing
An axisymmetric demonstrator geometry with an outer diameter of 34.7 mm was defined on basis of an extruder screw element. The geometry comprises two flights. The roller burnishing process was simulated with the FEA-system Ansys Academic Research Mechanical release 19.1 to predict work hardening in the subsurface of the profiled part. Using symmetry, just one half of the demonstrator was modeled.

Workpiece Material Characteristics
The material investigated in this study is a high interstitial austenitic stainless steel, named CN1.1Mo, which indicates the amount of carbon plus nitrogen in total in mass%, and the addition of molybdenum. Details on manufacturing route were described in a previous study. [7] The chemical composition is shown in Table 1. To achieve a fully austenitic microstructure Table 1. Properties of the investigated high interstitial steel CN1.1Mo.
Chemical composition in mass% [9] Hardening curve used in the material model free of undesired precipitations, samples were solution annealed (1150°C, 60 min) and subsequently quenched in water. The workpiece was modeled with an elastic-plastic material using Voce isotropic hardening shown in Table 1 (green line), while the hardening curve was obtained by a regression of tensile test results (blue asterisks).

Prediction of Local Work Hardening
During burnishing, the center distance between three driven rollers and the workpiece is synchronously decreased, which results in a roller radial infeed toward the workpiece. Thereby, the workpiece is set in rotation (Figure 1a). The number of workpiece revolutions during roller radial infeed was varied in a ratio of 1:2.5 between two settings, namely high feed and low feed. Material displacement on the preformed workpiece causes plastic strain and hence work hardening of the workpiece material. The preform is designed assuming a constant volume during burnishing. As a result of the process design, yield stress is significantly increased in the flight subsurface, as shown in Figure 1b

Roller Burnishing
The burnishing experiments were carried out on a three-roller profile-rolling machine manufactured by Profiroll Technologies GmbH. This machine is characterized by a maximum radial roller force of 400 kN. Preforms made of the austenitic steel CN1.1Mo were machined by turning and subsequently burnished with both settings, high and low feed, as investigated in the FEA. For each setting, three parts were burnished. For comparison purposes, the nitriding steel 31CrMoV9 in quenched and tempered condition (þQT), which is conventionally applied for manufacturing of extruder screws, was investigated as well. After burnishing, the nitriding steel was gas nitrided. As the profiled burnished parts cannot be tested on conventional model test stands, a dedicated test stand was developed for this investigation.

Wear Testing
The test stand, shown in Figure 2, consists of a bimetallic cylinder for plastics extrusion made of 50CrV4 steel with a centrifugally cast wear-resistant inlay of tungsten carbide and nickel. For melting plastic granules, the cylinder is filled through a pipe and heated by a heating band and two heating sleeves. When the plastic is completely melted, the linear actuators cause a translational movement of the sample in the cylinder, thereby causing a melt volume flow through the cylindrical gap. The sample is clamped between two guide shafts, which transmit the axial force of the linear actuators to the sample. The force of the linear actuators is measured at the guide shafts and controlled to prevent buckling of the shafts. Two force sensors on basis of strain gauges with a nominal force range of 20 kN were used. In the case of highly viscous polymers, the second linear actuator can assist the first and pulling actuator by pushing. The force control ensures that the tensile force is always larger than the pushing force. The force required for moving the sample through the polymer melt is recorded and the maximum force within the investigated samples and polymers was %2500 N. On the one hand, force recording allows for monitoring of the cylinder filling level. On the other hand, polymer degradation resulting from thermal and mechanical damage was recognized by decrease in axial force. The polymer was partly exchanged and new granules were added when the axial force fell below 1000 N.
Reinforced polymers (glass fiber and glass sphere) with similar properties, namely melting temperature and flow behavior (melt   Table 2). The polymer temperature was 200°C. Each sample was tested for 5000 cycles in the test stand. This corresponds to a distance of 5 km. The average translational speed of the sample was 10 mm s À1 to ensure conditions similar to that in a screw extruder.
Two samples, one burnished with high feed and one with low feed, were tested. The circular gap between sample and cylinder depends on geometric accuracy of the rolled part. Within this analysis, the gap width was between 0.14 and 0.16 mm. For model tests with rectangular platelets, a gap width of 0.2 mm is reported. [1]

Sample Preparation after Burnishing and Wear Analysis
After burnishing, the demonstrator parts shown in Figure 3b were separated in the middle resulting in two halves, each with one flight. Hardness was investigated with the hardness tester M1C 010 from EMCO-TEST Prüfmaschinen GmbH on polished axial sections of half the demonstrator part. Three polished samples were prepared of each of the three burnished parts, resulting in nine hardness measurement repetitions. Afterward, samples were etched with V2A-etchant at 50°C for 70 s to analyze the steel microstructure. The other half of the demonstrator was used as sample for wear testing. After wear testing, one-sided wear of the tip was recognized on most samples. This can be caused by a misalignment of the centers of sample and cylinder in the test stand. Such misalignment might result from transversal forces and elastic bending of the guide shafts while testing. Because of the sample size and initial mass, it was not possible to measure the mass loss in the case of very low wear. Thus, the shape of the outer diameter circle was measured on a 3D-coordinate measuring device manufactured by Carl Zeiss AG before and after wear testing. Furthermore, the axial contour was measured in two circumferential positions before wear testing. For quantitative wear evaluation, the wear depth Δx, shown in Figure 3a, was used. The maximum wear depth is the largest difference between ideal and worn contour. However, the wear depth varies along the sample circumference within the worn circular arc over the angle η. Information on the individual ideal sample contours and the wear depth was used to calculate the volume loss by Equation (1).
where x A is the half outer circle diameter, A j is the area between ideal and worn contour, and N arc is the maximum number of angle increments η j within the worn arc length x A · η. Area A j was calculated by Equation (2).
where x i are the coordinates corresponding to the axial increment z i and N z is the maximum number of increments within the worn contour width. The coordinates were measured every η j ¼ 0.05°a nd z i ¼ 0.05 mm. A similar approach was shown in literature for thermomechanical wear testing. [21] The mass loss was calculated by multiplying the steel densities, which are 7.73 kg dm À3 for 31CrMoV9 þ QT and 7.63 kg dm À3 for CN1.1Mo, with the volume loss.
For optical evaluation, a limited area of the wear mark, shown in Figure 3c, was analyzed using scanning electron microscopy (SEM). The surface profile was analyzed by means of a KEYENCE 9700VK laser scanning microscope.

Results and Discussion
Measured hardness values of the burnished parts are compared with yield stress values calculated with the FEA. In addition, wear marks and wear levels are compared between reference steel 31CrMoV9 þ QT and the austenitic stainless steel CN1.1Mo.

Work Hardening and Microstructure
For comparison of measured hardness with calculated yield stress, three paths, on tip, flank, and root, were defined in the axial section of the sample (Figure 1b). As shown in Figure 4a, yield stress in the root is approximately the initial yield stress of 600 MPa with a small increase near the surface. However, yield stress is significantly increased in the flank and the tip regions. The curves for tip and flank show a maximum yield stress below the part surface. The hardness curves shown in Figure 4b are in good correlation with the yield stress curves shown in Figure 4a. Yield stress is about three times the hardness. The initial hardness of approximately 300 HV 1 is increased to approximately 580 HV 1 in the tip. In the flank, the maximum value is approximately 520 HV 1. For both curves, hardness is increased over more than 2.4 mm surface distance.
In the root, there is a small hardness increase to 310 HV 1 near the sample surface. The influence of roller feed on the subsurface hardness is negligible within the investigated parameter range. Hardness curves for low feed are available in ref. [20]. Hardness was measured with a lower test force near the tip surface of single samples for a comparison between CN1.1Mo and the nitrided martensitic steel 31CrMoV9 þ QT. Up to a surface distance of %0.25 mm, the burnished and nitrided subsurface of 31CrMoV9 þ QT is harder than the one of the burnished CN1.1Mo with %590 HV 1. Within this range, the maximum hardness of the nitrided steel is about 40% higher than the hardness of CN1.1Mo ( Figure 5).
The intense cold-work hardening ability of HIS, which is confirmed by the hardness measurements of CN1.1Mo and numerous signs of dislocation movement shown in Figure 6, is a significant advantage compared with conventional FeCrNi austenitic stainless steels and is caused by different mechanisms of deformation. In particular, differences in electron configuration (density of free electrons) and stacking fault energy result in planar dislocation glide in the case of HIS, while FeCrNi austenites feature wavy slip. Furthermore, increased austenite stability prevents formation of alpha-martensite even upon high degrees of deformation (only very small amounts of epsilon martensite could be found in recent studies), thus maintaining nonmagnetic properties of HIS. [8]     of steel 31CrMoV9 þ QT shows pores on the compound layer, which is locally broken (Figure 7a, top row). After burnishing, the surface of the CN1.1Mo parts is smooth with a roughness depth of Rt ¼ 1.6 μm. Nitriding increases the roughness by %1 μm, which is consistent with the values given in literature. [22] Wear testing in glass fiber-reinforced polypropylene causes wear marks on the tip parallel to the movement direction of the sample in the test stand, which is parallel to the sample axis of rotation (Figure 8). The machining marks on the flank surface of 31CrMoV9 þ QT are smoothed by the polymer flow. However, the edge between the worn tip and the machined flank surface is clearly recognizable. The surface of CN1.1Mo, shown in Figure 8b, has more distinct valleys than the surface of the nitriding steel shown in Figure 8a, although the roughness depth is nearly on the same level.

Topography of the Roller-Burnished and the Worn Samples
The wear marks are more significant after testing in glass sphere-reinforced polypropylene ( Figure 9). Obviously, the glass spheres follow the translational sample movement through the circular gap between sample and cylinder. The roughness depth of Rt ¼ 2.7 μm is smaller in case of the nitrided surface than in case of CN1.1Mo with Rt ¼ 5 μm. It is assumed that the higher roughness of CN1.1Mo results from ploughing. In contrast, micro-cutting, as a form of abrasive wear without significant deformation, is assumed in case of the nitrided surface. Entering and leaving the gap results in wear marks in shape of a river delta, which is more significant for the ploughing wear marks on CN1.1Mo shown in Figure 9b. As a result, the transition between circumferential machining marks on the flank and axial wear marks on the tip is wider for glass spheres than for glass fibers (Figure 8 and 9).

Wear
In addition to topography analysis, the geometrical changes on the tip of the sample flight were measured as wear depth, which was used for calculation of volume and mass loss (Section 3.3). Given that the radial roller feed has no distinct influence on the near surface hardness of the samples, wear is evaluated as average between both samples low and high feed. Figure 10a allows for a comparison between gap width and the maximum wear depth. It is assumed that the differences in gap width of %0.02 mm between the samples tested against PP GS40 and the other samples have a minor effect on the wear results. On the harder nitrided steel surface, wear depth is smaller for glass fiber reinforcement than on the austenitic stainless steel. The wear depths for glass sphere reinforcement spread between 30 and 80 μm. Glass spheres cause smaller wear depths on the austenitic steel than glass fibers. This can be a result of significant ploughing of glass spheres, which does not cause material loss but a deformed surface.
The volume loss in Figure 11 includes the varying circumferential distribution of the wear depth on the sample tip. However, volume loss values are in the same order as wear depth results. In the case of CN1.1Mo, the difference in volume loss for both plastics is lower than the difference of maximum wear depth. Because of minor differences in material density, the mass loss shows the same order like the volume loss.

Conclusions
In this investigation, high interstitial austenitic stainless steel parts were produced by turning followed by burnishing on a profile-rolling machine. Hardness was significantly increased from an initial value of %300 HV 1 to an average value of %580 HV 1 by burnishing the austenitic stainless steel. Local work hardening was predicted in an FEA of the burnishing process. Calculated curves of yield stress correlate with the measured hardness over surface distance, whereas further investigations have to show whether the assumption of a linear coefficient holds true. A wear test stand simulating the conditions in a plastics extruder was developed and used for investigating wear of burnished steel samples in glass-reinforced polymer melt. This test stand was applied for comparative wear investigations on different steels. The comparison of wear on the nitrided steel surface and the burnished austenitic steel surface shows that the wear level is similar for both materials in glass sphere-reinforced polypropylene, although the nitriding steel is harder than the   burnished austenitic steel. While micro-cutting seems to be significant on the nitrided surface, an increased roughness indicates ploughing wear on the austenitic steel surface. The performance of burnished austenitic stainless steel parts might result in a wider application in corrosive environments with wear protection requirements. In the future, roller-burnished screw elements made of high interstitial austenitic stainless steel might be applied in sidefeeders of extrusion lines.