Surface hardening of AISI 4340 steel by laser linear oscillation scanning

Material and methods

The tested material was AISI 4340 steel, which was a quenched and tempered low alloy steel composed of tempered martensite structures.15 The composition of the material is 0·347C–0·331Si–0·007V–1·397Cr–0·697Mn–1·355Ni–0·169Mo (wt-%). The initial hardness of the base material was 329 HV. The surface roughness of the samples, R_a, was 2·5 μm.

This study used a work cell consisting of a CNC XY work table, a YLR-5000-S multimode fibre laser system and an ILV DC scanner installed as the laser head. The fibre laser produced a laser beam with a wavelength of 1070–1080 nm and a maximum nominal output power of 5 kW. An output fibre core with the diameter of 200 μm was used. The DC scanner, which was primarily designed for laser welding applications, contained a parabolic scanner mirror with focal length of 250 mm and a controller unit connected to it. The distance off the focus was 60 mm, producing a laser spot of 2·1 mm in diameter. The oscillation speed was affected by the oscillation frequency and followed the sine waveform. The parameters are given as tests A1–A6 in Table 1. For comparison, a conventional single track scanning test was done, shown as tests B1 and B2 in Table 1. To produce laser power densities of 15136 and 22629 W cm⁻², 2020 and 3020 W laser power with 75 mm off focus distance were used respectively. The feeding speed was constantly 15 mm s⁻¹ in the whole test.

OPEN IN VIEWER

Table 1

Processing parameters

Test	Laser power P/W	Oscillation frequency μ/Hz	Amplitude A/mm
A1	2020	10	11·0
A2	2020	20	8·5
A3	2020	40	4·5
A4	3020	20	10·0
A5	3020	40	7·4
A6	3020	100	5·9
B1	2020	NA	NA
B2	3020	NA	NA

Modulation of laser power profile

Figure 2a shows the laser energy distribution with oscillation frequency of 100 Hz and amplitude of 6·2mm produced by the DC scanner. Since the oscillation speed was reduced to 0 at the edge of oscillation, in order to avoid melting, the output laser power was modulated by programming via the controller unit. As shown in Fig. 3, the output laser power was programmed at 19 positions with equal spacing across the irradiated width, producing an approximated sine wave laser power profile with peak laser power of 2020 and 3020 W respectively and with 300 W laser power at both edges of oscillation. Since the oscillation speed followed the sine waveform as mentioned above, the ratio of oscillation speed to output laser power could be regarded as constant. Figure 2b shows a controlled laser beam profile featuring reduced laser power density near the edge of oscillation. The average laser power densities measured were 14 483 and 21 691 W cm⁻² under 2020 and 3020 W laser power respectively, which were close to those produced by the single track scanning mentioned above.

OPEN IN VIEWER

Figure 2

Laser energy distribution produced by DC scanner with oscillation frequency of 100 Hz and amplitude of 5·9 mm: a without power control; b with power control via controller unit

OPEN IN VIEWER

Figure 3

Programmed laser power output via controller unit of DC scanner

Scanning speed

The scanning speed is the combination of the oscillation speed and the feeding speed of the laser. Since the oscillation speed (v_y) is in sine waveform, it can be written as (1) where y is the position of the laser spot. At y  =  0 and y  =  2A, the laser spot is at the edges of oscillation.

Moreover (2) where v_ymax is the maximum oscillation speed, which is achieved at the middle of the oscillation.

The scanning speed (v) is therefore (3) where v_x is the feeding speed of the laser head. In this study, v_x = 15 mm s⁻¹.

The average oscillation speed ( ) is (4) v_ymax and are calculated for the parameters of oscillation test in Table 1 using equations (2) and (4). The oscillation speed of the test is described in Fig. 4, obtained on the basis of equation (1). It can be markedly seen that is dramatically higher than v_x (15 mm s⁻¹) in this study, and thus, v can be treated as v = v_y in equation (3), although at the edges of oscillation v_y = 0 and v = v_x.

OPEN IN VIEWER

Figure 4

Oscillation speed of test

Overlap width of adjacent tracks

Figure 5 schematically describes the overlapped neighboring tracks of LLOS. The laser spot starts from position O₁ and reaches O₂ in one period of oscillation. At the edge of oscillation, the laser spot moves along the feeding direction with a distance, which can be calculated as (5) where S_e0 is the distance at the edge of oscillation between the laser spots of adjacent tracks.

OPEN IN VIEWER

Figure 5

Schematic diagram of overlapped adjacent tracks in LLOS

Thus, the overlap width at the edge of oscillation S_eov can be expressed as (6) where d₀ is the diameter of the laser spot.

Substituting equation (5) with equation (6) gives (7) At the centre of the oscillation, the overlap width S_cov can be calculated as (8)

Number of scans spot experiences in centre of oscillation

During the oscillation scanning process, a position in the middle of the oscillation width may experience multiple times of laser irradiation. According to Fig. 5, the number of scans N_c is the integral part of the ratio of d₀/S_e0, which can be expressed as (9) Substituting equation (5) with equation (9) gives (10) Based on the test parameters in Table 1, equations (5), (8) and (10), S_e0, S_cov and N_c can be calculated. The results are shown in Table 2.

OPEN IN VIEWER

Table 2

Calculated S_cov and N_c of test

Test	Se0/mm	Scov/mm	N c
A1	0·75	1·35	4
A2	0·37	1·73	10
A3	0·19	1·91	22
A4	0·37	1·73	10
A5	0·19	1·91	22
A6	0·07	2·03	56

Hardness measurement

The hardness measurement was done on a Vickers hardness tester with a load of 100 g. Three series of positions in the tested samples were measured (Fig. 6). Series 1 was along the centre of the irradiated area toward the feeding direction. The starting point was randomly selected. Series 2 was across the whole irradiated width of the cross-section. Series 3 was along the depth below the centre of the irradiated track. Measurements of series 1 and 2 were done at 0·05 mm depth below the surface, which was assumed to have nearly identical hardness to the surface. In this study, the hardened depth is defined as the depth where the measured hardness decreases to below 400 HV.

OPEN IN VIEWER

Figure 6

Hardness measurement positions of samples

Hardness profile along centre of irradiated track

As shown in Fig. 7a, tests A1 and A3 produced surface hardness values of 460–703 and 652–831 HV respectively, which were relatively inhomogeneous. Test A2 did not cause hardening of the material. Test B1 resulted in relatively homogeneous surface hardness of ∼684 HV, which was close to the average hardness of 717 HV caused by test A3. Figure 7b shows that test A4 caused little hardening effect, while individual scans might be distinguished by the variation of hardness. The marked hardness variation caused by A1 and A4 can be explained by the tempering effect.16 According to Table 2, the distances between adjacent tracks of A1 and A4 at the edge of oscillation were 0·75 and 0·37 mm respectively, which were much larger than A6 (0·07 mm). Test A6 with oscillation frequency of 100 Hz produced an overlap width of 2·03 mm, which was 96·7% of the whole laser spot diameter (2·1 mm). This might have resulted in massive tempering covering almost the whole irradiated area and thus resulted in homogeneous surface hardness. Tests A5 and A6 caused respective average hardness of 795 and 777 HV, which were similar to the hardness of 767 HV produced by test B2. As markedly noticed, the hardness variation by test A6 was much smaller than that by test A5, which was caused by the higher scanning frequency of A6 (100 Hz), compared to A5 (40 Hz).

Hardness profile across irradiated width

Figure 8 indicates hardness across the width of the irradiated track. Noticeably, in the results of tests A1, A3, A5 and A6, the highest hardness was not achieved in the centre of the track but millimeters away from it. The maximum hardness values obtained were 764, 790, 909 and 892 HV for tests A1, A3, A5 and A6 respectively. For tests A3, A5 and A6, the maximum hardness was achieved at ∼2·5, 4 and 2 mm respectively, considerably higher than the average of the measured hardness along the centre of the track (Fig. 7). According to Fig. 5, this was possibly caused by the tempering effect, which was stronger in the centre of oscillation than elsewhere. According to equation (2), reducing the oscillation frequency and/or the amplitude can decrease the maximum oscillation speed. Compared with test B1 with 2020 W laser power, tests A1 and A3 produced similar hardness in the centre of the track, as shown in Fig. 8a. However, test A1 generated a much larger irradiated width, and test A3 markedly produced higher average hardness across the track. Similarly, as denoted in Fig. 8b, tests A5 and A6 both produced wider hardened tracks with higher average hardness than test B2. Figure 8a shows that test A2 did not cause hardening at all but retained the hardness values of the base material. Figure 8b denotes that test A4 caused very slight hardening, indicating that a small amount of martensite might have formed on the surface.

Hardness profile along depth below irradiated track

Figure 9 compares the hardness profiles along the depth beneath the irradiated track. As seen in Fig. 9a, test A2 did not cause hardening. Tests A1 and A4 produced very small hardened depth of <∼0·1 mm. As compared in Fig. 9a, tests A3 and B1 produced hardened depth of ∼0·55 and 0·9 mm respectively. However, between the surface and 0·5 mm depth, test A3 generated fairly consistent hardness of 682–751 HV, markedly higher than single track scanning test B1 that caused 524–683 HV. Similar phenomena were noticed in Fig. 9b. Regarding the hardness from the surface to 0·4 mm deep, tests A5 and A6 generated 641–803 and 693–831 HV respectively, which were higher than the hardness of 593–753 HV acquired by test B2. Tests A5 and A6 produced hardened depth of 0·4 and 0·5 mm respectively, which were much smaller than the 1·5 mm hardened layer by test B2. This is believed to be due to the high scanning speed in the middle of oscillation, which did not allow sufficient heating in a large depth of the material.

Microscopic evaluation

Microscopic observation on the microstructure of the tested samples provided a direct and qualitative way to visually evaluate the experimental effect. Samples tested with different laser power, frequency and amplitude of oscillation are compared in Fig. 10, which shows the microscopic view below the centre of the irradiated area toward the feeding direction of the laser beam.

OPEN IN VIEWER

Figure 10

Microscopic view below centre of irradiated surface of samples tested with different laser power, frequency and amplitude of oscillation respectively: samples were etched with 5% nital: a A1: 2020 W, 10 Hz, 11·0 mm; bA2: 2020 W, 20 Hz, 8·5 mm; c A3: 2020 W, 40 Hz, 4·5 mm; d A4: 3020 W, 20 Hz, 10·0 mm; e A5: 3020 W, 40 Hz, 7·4 mm; f A6: 3020 W, 100 Hz, 5·9 mm

As measured in Fig. 10, the depths of the heat affected zone (HAZ) produced by tests A1–A6 were respectively ∼0·05, 0, 0·45, 0·05, 3·8 and 4·5 mm, which were well in accordance to the hardened depth indicated in Fig. 9. Figure 10c, e and f shows that HAZ acquired with 40 and 100 Hz were much deeper. No HAZ was observed in Fig. 10b, meaning that phase transformation did not occur on the sample. According to equation (2), the oscillation speed produced with 20 Hz and 8·5 mm amplitude might be too high to allow sufficient absorption of laser energy. Tests A2 and A3 were done with both 2020 W laser power and, according to equation (2), with very close oscillation speed. However, the depths of HAZ were dramatically different by comparing Fig. 10b and c. A possible explanation is that with higher oscillation frequency, the scanned surface experienced more cycles of irradiation, which meant more absorption of laser power. Figure 10a and d with 10 and 20 Hz respectively shows very small and inhomogeneous HAZ depths, which might be caused by the distinguishable adjacent tracks. Differently, the individual scans could not be distinguished with 40 and 100 Hz in Fig. 10c, e and f. As shown in Table 2, this can be explained by the large overlap width and large number of overlapped scans under the frequencies of 40 and 100 Hz.