STUDY OF HOLE PROPERTIES IN PERCUSSION REGIME WITH A NEW ANALYSIS METHOD.

Paper M1203 M. Schneider*, M. Muller*, R. Fabbro*, L. Berthe* *Laboratoire pour l’Application des Lasers de Puissance (UPR CNRS 1578)/GIP GERAILP 16 bis avenue Prieur de la côte d’or 94114 Arcueil FRANCE

Abstract To drill sub-millimeter holes, laser percussion drilling has been a well-established industrial process for more than tens years. However, inherent factors due to laser source, material properties and dynamic of the process are still making physical approach very difficult. This paper deals with the study of the properties of holes drilled by an innovative source from TRUMPF that ensures top hat distribution in focal plane with a constant diameter, for any laser parameters. Results are based on several post-mortem metallographic analysis. First, a new and very efficient hole analysis method (called DODO for Direct Observation of Drilled hOle) is described and compared to the X-ray radiography and metallographic cutting. DODO solves most of the problems of previous methods, notably to ensure the an easy matching of the analysis plane with the drilling axe. Secondly, with the help of DODO method, we show the existence of a threshold between two characteristic drilling shapes defined as V and U profile. Using a geometrical optic approach, a simple model estimates the drilling depth corresponding to a drilling with U shape. We show the influence of the peak power distribution on hole morphology (profile, diameter and quality), when using a sequence of pulses. The repartition of peak power is adjusted in the sequence of pulses in order to eliminate the recast layer cracking. This problem results from the solidification of a melt layer located above a previous recast layer. To eliminate it, it is essential to remelt the previous recast layer with higher peak power pulse than the previous one. 1. Introduction Laser drilling in percussion regime is used extensively in aeronautical industries. This process consists in irradiating a metallic target with a laser tuned in the MW.cm-2 range (pulse duration in µs-ms range). The laser energy is absorbed by the surface for the heating, the melting and the vaporization of the target. This vapor expansion pushes the melt pool. In this case, the drilling is dominated by melt ejection induced by the pressure gradient between the irradiated area and the hole surroundings, see [1, 2]. Translating these physical processes into equations, requires to solve the transport equations (Navier-Stokes and heat transfer) with the laser dump as boundary conditions. Up to now, it is impossible to solve this system without introducing some approximations, then compromising the accuracy of the final solution. This is

why parametric and qualitative studies are essential to understand the correlation between each operating parameters and their own effect. This allows us to find out the main parameters that induce changes in the hole. In the first part, we present a new hole analysis called DODO (for Direct Observation of Drilled hOle) DODO is compared to conventional hole analysis : X-ray radiography and metallographic cutting. In the second part, we present the experimental conditions and the HL201P TRUMPF laser characteristics used for experiments. Finally we present and discuss some results with this method. Hole profile allows us to identify two drilling shapes separated by a threshold. At low intensity, holes are conical and higher intensity they are quite cylindrical. We show the influence of peak power, pulse duration and number of pulses on the hole profile and on the recast layer characteristics. 2. Hole analysis method 2.1 X ray radiography This method is the only one allowing non-destructive analysis. It is based on the shadowgraphy principle showed on Fig 1. In practice, the drilled sample is located very closely to a radiographic paper. The whole is irradiated with X rays. The paper reveals a projection of holes.

Fig 1 : Schematic representation of X ray radiography method. Fig 2 shows a typical image of a radiographic photograph performed with this method. Four holes appear as shaded columns. From this picture, global shape, depth and conicity can be determined. Typically, top diameter is 500 µm and the depth is 2,5 mm.

Fig 2 : X ray radiographic image. Details on images, are not enough accurate to describe precisely hole morphology and recast layer thickness. Moreover the hole bottom is less contrasted than the hole entrance, (see discussion Table 2)This method allows a very global analysis. 2.2 Metallographic cutting This method is destructive one. Indeed, the sample is cut and polished until the cross section is located on the drilling axe. With the help of a chemical reaction it is possible to reveal the metallurgical structure. However, it does not ensure the matching of the cross section position and the drilling plane due to the polishing procedure (particularly for deep holes). Difficulties concern their parallelism and their location, Fig 3 shows schematic views of these drawbacks. Fig 3.a presents the shift between drilling plane and cross section and on Fig 3.b the double tilt (α and β) between drilling plane and cross section. Besides, resin, used to operate, limits the control of the polishing and allows only a 2 D analysis in the cross section. Consequently, shape measurements (diameters and depths) are not accurate and in some cases not representative.

Fig 4 : hole profile for several tilt between drilling plan and metallographic cutting. If shift is on the other side of the drilling axe and the tilt is constant, we see on Fig 4.b that only the 0.1 rad curve is totally wrong. For 0.05 the mistake on the radius measurement is in the range of 10%. Here, we consider a perfect cylindrical hole, but in reality, holes are more conical so the mistake made on the hole depth is minimized here.

Fig 3 : Schematic representation of metallographic cutting and the different drawbacks

Fig 5 is an image of a cross section obtained by this method.

a)Shift between drilling plane and cross section b)Tilt between drilling plane and cross section The Fig 4 shows the normalized radius given by the metallographic cutting on a cylindrical hole of 0.2 mm radius and 3 mm depth. The shifts are equal to one third of the hole radius from drilling axe. The tilt is given for each curve in radian. The curve represent the holes profile for several position of the metallographic cutting. We see that for and 0.01 rad tilt, the mistake on the radius measurement is in the range of 10%. Above 0.01 radian the measurements are totally wrong.

Fig 5 : Microscopy image of metallographic cutting. This image proves that this method could qualify, the recast layer thickness and the metal granularity in the cross section. For deep hole analysis it is very difficult to match together the cross section and the drilling plane. This method allows a very local analysis with a high

accuracy, typically to determine the structure and the recast layer thickness. This method allows a very precise analysis but it remains very expensive in time and cost and cannot be used on production site. Typically, only one hole a day can be analyzed.

correctly coincident, the melt metal injected between the surfaces during the drilling process, will be easily detected after splitting. Secondly, if the drilling axis is not precisely located in the common cross section, the 2 hole shapes obtained on the 2 samples will be different and this can be very easily detected.

2.3 Direct Observation of Drilled hOle: DODO method

3. Experimental Set up. 3.1 Laser characterization.

DODO solves most of the problems of the previous methods. Fig 6 shows schematic views of the method. It consists in locating the drilling axis in the analysis plane that has been prepared. To do it, we used 2 samples and one surface of each sample is polished. These two planes are then tightly assembled. Holes are drilled in such manner that their axis are in the interface plane of samples as shown on Fig 6.a.

In a previous article [3], we have shown that the HL201P laser from TUMPF has very useful guarantees for a parametric study: - a circular “top-hat” intensity distribution in the focal plane, - a pulse reproducibility, - a constant focal plane position for any laser parameters. The circular “top-hat” intensity distribution is ensured by an optical principle. The laser beam output is homogenized through a small fiber length. The output of the fiber is imaged in the focal plane by two lenses. Fig 8 shows a beam shape as function of the distance to the focal plane. In the focal plane the laser diameter is 330 µm.

After drilling, the two samples can be easily separated and reveal half hole shape on each one, as shown on Fig 6.b Consequently, DODO ensures the observation analysis plane and so measurements associated to morphology analysis (depth, conicity).

Fig 6 : Schematic representation of new analysis method.

Fig 8 : HL201P beam shape. The distance to the focal plane is in millimeter. From geometric optics considerations we define :

tan(α)= RLl − RL0 f

Fig 7 : Hole shape SEM image. Fig 7 shows a SEM images of a half hole provided by DODO method. Pictures prove clearly the morphology and the recast layer thickness. Moreover, because resin is not used as with traditional metallographic analysis, the surface roughness of the wall and the 3D shape of the hole can be observed and evaluated directly. To produce a good quality image we need a depth of field deeper than the radius of the hole, which is in the range of 500 microns. So we use here a SEM but for direct observation an optical microscope is enough. Besides, this method allows some auto-control ways. Firstly, if the two surfaces of the samples are not

(1)

where α is the optical aperture, RLl is the laser radius on the focal lens and RL0 in the focal plane, see Fig 9. We assume laser radius is described by: (2) DLz = DL0 + z×tan(α) where DLz is the laser diameter at the distance z from the focal plane. Fig 9 : geometric approach Fig 10 experimental beam shape diameter and eq (2). There is a very good agreement between the two curves. So in the following we will assume that laser diameter is equal to optical diameter.

V profile hole group: The radius evolution is linear and decreases with the depth and increases with the peak power. The recast layer thickness is below 10 microns and it is constant along the hole. In this V profile hole group, holes depth, and so the drilling velocity, increase with the peak power. Fig 11 presents hole depth as a function of peak power. U profile hole group: Fig 10 : Geometric evolution of beam diameter and optical diameter as function of distance to the focal plane. 4. Hole morphology study with DODO method: single pulse In this part, we present some results obtained with the DODO method. We show the effect of peak power and pulse duration inside a blind hole.

Above 6 kW, the two morphological parts start to separate. In the hole entrance part, the radius decreases slower with the depth than it does in the V profile hole group. And the bottom of the hole start rounding. The hole profile is no more totally conical. The radius increases with the peak power.

In following, drillings are made with laser focused on the target surface for a normal incidence. An argon shield gas (Psurface<1bar) is used to only protect optics. Table 1 presents typical images of the hole from analysis. Pictures are directly obtained by a scanning of the samples. The recast layer thickness, which appears more brightness than the target material original, can be easily measured with this method. On these images, the front hole is at the bottom of the picture. Holes profiles are selected to obtain a representative profile among a range of thirty holes made with the same parameters. 4.1 Hole drilled with 0.5 ms pulse duration Table 1 shows hole profile as a function of peak power for a normal incidence. The pulse duration is 0.5 ms. Table 1 : Hole profile evolution as a function of peak power for a 0.5 ms pulse duration. Peak power (kW)

5

6

7

8

12

pulse duration (ms)

0.5

0.5

0.5

0.5

0.5

Fig 11 : Depth as a function of peak power with a single pulse. The holes depth are rather constant when the peak power increases. So, the drilling velocity appears to be constant. The recast layer thickness is below 10 microns and it is constant along the hole, as in the conical shape hole group. 4.2 Hole drilled with 1 ms pulse duration

a)

b)

c)

d)

e)

Table 2: Hole parts. Two specific geometry parts can be identified in holes. The hole entrance is the part of the hole where the diameter is constant, and the hole bottom is the complementary part, see Table 2.

Hole bottom

Hole entrance

Table 3 shows the holes profile evolution as a function of peak power for a 1 ms pulse duration. Shape evolution in two groups is confirmed, as with a 0.5 ms pulse duration. Holes can be divided into two groups. The first one for low peak power, below 6 kW (Table 3.a to 3.c), where hole shape is completely conical (V profile). There is only one part. The second group concerns peak power above 6 kW (Table 3.d to 3.n) for which there are two parts in the hole profile (U profile).

Table 3 : Hole profile evolution as function of peak power for a 1 ms pulse duration. P (kW) τ(ms)

3 1

4 1

5 1

6 1

7 1

8 1

10 1

12 1

13 1

14 1

15 1

16 1

17 1

18 1

a)

b)

c)

d)

e)

f)

g)

h)

i)

j)

k)

l)

m)

n)

V profile hole group Below 6 kW the holes profile are still conical, with the same characteristics, but they are deeper than with a 0.5 ms pulse duration, see Fig 11. U profile hole group Above 6 kW, the radius is rather constant in the holes entrance. And the holes bottom is rounder when peak power increases. Above 6 kW, the depth is also constant when increasing peak power. So, the drilling velocity is also constant.

13kW (in red), 18kW (in blue) and for a formula obtained with a simple model (in green). For 13 kW peak power holes bottom have a U profile until 4 pulses (Table 4.d). Between 5 and 7 pulses, (Table 4.e and 4.g) hole diameter is constant in hole entrance but holes bottom are pointed so holes are drilled with a mixed shape between U and V profile, we consider this profile like an U profile. Above 7 pulses, hole shape is conical, V profile. The recast layer thickness is below 10 microns until 7 pulses.

The recast layer thickness is always below 10 microns and it is constant along the hole. Fig 12 shows the evolution of front hole diameter as function of peak power. The diameter increases linearly with the peak power. At 0 kW, hole diameter is equal to the focal plane laser diameter (DL0). So hole diameter can be defined by:

DH = A×P+ DL0

(3)

P is the peak power and DH is the hole diameter. A is a constant in m.W-1.

Fig 13 : Graphic evolution of depth as function of pulses number for 13 and 18 kW. For 18 kW peak power holes have U profile until 6 pulses (Table 4.k to 4.p). Form 6 to 10 pulses (Table 4.p to 4.t), holes have a mixed shape. For this peak power holes shape do not become conical until 10 pulses. The recast layer thickness is below 10 microns until 10 pulses along the wall of the hole. So, until 7 pulses for 13kW, drilling has an U profile. Above 7 pulses it become conical. For 18kW and until 10 pulses drilling has U profile.

Fig 12 : Diameter as function of peak power. (A=0.03 mm.kW-1, DL0=0.35 mm) for 1 ms pulse duration. 5. Hole morphology study with DODO method: multi-pulses In this part, we present the influence of pulse number. Table 4 shows hole profile for 13 and 18 kW peak power. N is the number of pulse. Fig 13 gives the correspondence between pulse number and depth for

We have assumed that drilling velocity is, at first order, a linear function of absorbed intensity, and it is averaged on pulse duration. The laser intensity is approximated by the direct irradiation contribution. So, we define a depth of drilling (zn) like the sum of the depth of each pulse, so: n −1

( ) ( 2

n −1

zn = RL0 = 1+ zi ×tan(α) ∑ z1 ∑ RL0 i =1 RLz i =1

)

−2

(4)

Table 4: Hole profile evolution as function of number of pulse for 13 & 18 kW peak power. P (kW) 13 13 13 13 13 N 1 2 3 4 5

a)

b)

c)

d)

e)

13 6

13 7

13 8

13 9

13 10

18 1

18 2

18 18 3 4

18 5

18 6

18 7

18 8

18 9

18 10

f)

g)

h)

i)

j)

k)

l)

m)

o)

p)

q)

r)

s)

t)

with z1 =2.7 mm is the depth after the first pulse. N is the number of pulse, zi is the position at the beginning of the nth pulse. On Fig 13, the 18 kW curve equals to the model curve for all number of pulse, whereas the 13 kW curve is moved away from the fit curve from the 7th pulse. So the fit curve describes the drilling depth above the threshold as function of pulse count. 6. Hole morphology study with DODO method : two pulses In this part, we show the influence of peak power for two pulses on recast layer cracking. Table 5 shows hole drilled with two pulses. Holes are drilled with twin pulses in the left part (Table 5.a to 5.e) and with different first pulses peak power in the right part (Table 5.f to 5.i). The second pulses have a constant 16 kW peak power. The left part (Table 5.a to 5.e) : twin pulses.

n)

Table 6: Recast layer cracking. The right part (Table 5.f to 5.i) : pulse train. In this case, holes are produced with a first pulse peak power that is lower than the second one. The first pulse peak power increases and the peak power of the second pulse is constant 16 kW.

Recast layer cracking

Three areas characterize holes. The upper part is drilled by the first pulse with the lowest peak power of both pulses. The hole center and the hole bottom are drilled by the second pulse. The diameter increases between the entrance and the center area. The hole bottom is rounded. Recast layer cracking is missing for the holes 5.f and 5.g when the peak power of the first pulse is below 8 kW.

The first hole (Table 5.a) is produced with two pulses at 5 kW. Its shape is conical.

For holes 5.h with a 8 kW first pulse peak power, a recast layer cracking is observed. And for higher first pulse peak power, each hole has a recast layer cracking.

The others have U profile. Diameter is constant in the hole center and hole bottom start pointed to become curved when the peak power increases, as seen previously.

7. Discussions Table 7 shows hole produced with 5 kW peak power for first or second pulse. It summarizes holes shapes observed previously.

All holes produced by twin pulses show a recast layer cracking in the hole entrance (see Table 6).

Table 5: Hole profile evolution as function of peak power for two pulses. P (kW) N RLC

5-5 2 yes

7-7 2 yes

10-10 13-13 16-16 2 2 2 yes yes yes

5-16 2 no

7-16 2 no

8-16 2 yes- no

10-16 2 yes Hole Bottom Hole Center

2nd pulse

Hole Entrance a)

b)

c)

d)

e)

f)

g)

h)

i)

Table 7: Hole profile evolution for 5 kW peak power. P (kW) N

51

5-5 2

5-16 2

Concerning Recast layer cracking :

16-5 2

Hole drilled with a first pulse peak power below 8 kW and a second one equal to 16 kW have not any recast layer cracking. Each pulse forms a recast layers. The recast layer cracking comes, in case of multiple pulses drilling, from the solidification of the melt layer on the previous recast layer.

a)

b)

c)

d)

The first and second hole (Table 7.a and 7.b) have a V profile, they have been drilled respectively with 1 and 2 pulses. The 7. c hole is produced with a first pulse at 5 kW and the second at 16 kW. Three areas can also be observed. The diameter in the bottom part of the hole is larger than in the center one. There is also recast layer cracking. The last hole (7.d) drilled with a second pulse at 5 kW has an U profile. Whereas the hole drilled with twin pulses at 5 kW (Table 7.b) has conical shape. So, for a second pulse at 5 kW the drilling process is different if the first hole has a U or V profile. This could be explained by difference of the irradiated surface at the bottom of the first hole. The ration between a cone and a sphere surface is given by :

()

2 irradiated Scone h >1 and h> R ≈ 1 + irradiated Ssphere R

(5)

where h is the cone high and R is the cone and the sphere radius. The cone surface is always higher than the sphere surface. So the absorbed intensity decreases in consequence.

The only recast layer that do not show cracking is the one generated by the solidification of the melt layer on the substrate. So if a hole drilled, with two or several pulses, has no recast layer cracking, it is only because the last pulse had melted the previous recast layers. Table 8: Interpretation of two pulses drilling with different peak powers.

Table 8 shows two holes diameters averaged among a range of thirty holes made with the same parameters, obtained with the DODO method. The first hole is made with two pulses with 5 and 16 kW and the second with a 16 kW twin pulses.

-

the drilling velocity is constant at 2.5 m.s-1

For multi pulse drilling: The diameter is conserved with the depth, as long as the drilling intensity is higher than the threshold.

For the hole in Table 8.a, the diameter of the upper area is much larger than for a single pulse of 5 kW peak power. So the second pulse has widened the diameter of the first pulse. And the first recast layer has been melted. On Fig 12 The diameter of the hole entrance drilled with 2 pulses (600 µm) is larger than hole drilled with 5 kW (500 µm) and smaller than hole drilled with 16 kW (800µm).

The recast layer thickness is constant. The cracking results from the solidification of a melt layer above a previous recast layer. To eliminate this effect, it is essential to melt the previous recast layer with a higher peak power pulse than the previous one, see Table 8.

For the hole in Table 8.b, its diameter is defined by the first pulse peak power. The diameter is equal to 800 µm for a single pulse of 16 kW peak power. So the recast layer has not been melted by the last pulse, and as a result, recast layer cracking is observed in the hole.

[1] S. I. ANISIMOV AND V. A. KHOKHLOV, (1995) Instabilities in Laser-Matter Interaction, CRC Press, Boca Raton.

8. Conclusion Concerning DODO : DODO method ensures the matching of the analyzing plane and the drilling plane, and allows a 3D analysis inside the hole. DODO : - allows the characterization of the hole morphology (profile, radius, recast layer thickness, conicity, and depth). - is a global 3D analysis . - can be done on production site. - is the fastest and the cheapest method of hole analysis. Once the samples have been prepared, many holes can be analyzed in a very short time. Concerning drilling process : The hole diameter increases linearly with the peak power, see Fig 12. For our operating parameters, the recast layer thickness is below 10 microns and it is constant along the hole whatever the peak power. There is a threshold (6 kW incident peak power) below which : - the hole is constituted of only one area. - the diameter decreases linearly with the depth. The hole shape is a V profile. The conicity is maximum. - the drilling velocity increases quite linearly with peak power And above which : - the hole is constituted of two areas. - the diameter is constant in the hole entrance. The hole bottom is rounded, or pointed. The hole shape is an U profile.

Bibliography.

[2] V.V. SEMAK AND A. MATSUNAWA (1997) The role of recoil pressure in energie balance during laser materials processing, J. PHYS D: APP PHY 30, 2541-2552. [3] M. SCHNEIDER, R. FABBRO, L. BERTHE, L.LANDAIS, M. NIVARD AND P. LAURENS (2004), Parametric study of drilling with new innovative laser source: application to percussion regime, Proc. Intnl Congrs on Appl on Application of Lasers and Electro-Optics (ICALEO’04), San Francisco , 540-546. [4] M. SCHNEIDER, R. FABBRO, L. BERTHE, M.MULLER, (2006), Study of hole properties in percussion regime with a new analysis method, Proc. Intnl Congrs on Laser Advanced Materials Processing (LAMP’06), Kyoto, HPL17.