Laser cladding: an experimental and theoretical investigation

(1)

DOCTORA L T H E S I S

Laser Cladding: An Experimental and Theoretical Investigation

Hans Gedda

(2)

This thesis is dedicated to my family

Birgitta, Petrus and Emilia.

(3)

(4)

Preface

Since April 1999 I have been conducting experimental and theoretical research in the field of laser cladding at the Division of Manufacturing Systems at Luleå University of Technology.

The experimental work was mostly performed in our laser laboratory. Some work has been done at Duroc AB in Umeå, Luleå and at Nottingham University.

Several people have been important in completition of this work. I sincerely thank my supervisor John Powell who has guided and supported me throughout this research.

I would like to express my gratitude to Professors Alexander Kaplan and Claes Magnusson for discussions, suggestions through this work. I would also like to thank all my friends and colleagues at the division for all their help and fruitful discussions.

I would finally thank my family, Birgitta, Petrus and Emilia for their love, support and patience during the work.

Luleå, October 2004

Hans Gedda

(5)

(6)

Abstract

This thesis presents an investigation into the laser cladding process using CO₂ and Nd:YAG lasers. The work is divided into six chapters:

Chapter one is an introduction the subject of laser cladding. This presents a general overview of the two common laser cladding methods and some applications for the processes. This chapter concludes with abstracts, main figures and conclusions from all chapters in the thesis.

Chapter two is an investigation into the energy redistribution during CO₂ and Nd:YAG laser cladding. Experimental absorption measurements by calorimetry were carried out to analyse how much of the energy is lost by reflection etc. It was found that the Nd:YAG laser cladding process is approximately twice as energy efficient as the CO₂ laser cladding process.

Chapter three investigates the process parameters which affect the finished product when cladding into pre machined groves including; groove geometry, powder application method and laser type.

Chapter four presents preliminary experimental results from two new processes; Laser casting and Laser clad-casting. Laser casting is a process similar to blown powder laser cladding but without the final product joined to the substrate. The substrate acts as a mould and the casting retains the topological features of the substrate. Laser clad-casting involves the production of a clad layer between machined copper blocks. Clad tracks can therefore be achieved with large depth to width ratios and pre determined cross sections.

Chapter five describes a new technique for the production of solid wire or rods from powder by laser melting. Three techniques have been developed to ensure that the molten powder solidifies as a rod or wire rather than a series of droplets. The techniques can be used to produce welding rods, tensile test samples and other solid pieces from a wide range of powder mixes.

Chapter six presents experimental data in conjunction with mathematical models are used to explain various aspects of laser casting and laser cladding by the preplaced powder method.

Also the interaction of the melt pool with the powder bed is analysed to identify why laser castings have microscopically uneven surfaces.

(7)

(8)

Chapter I

Introduction to laser cladding

(11)

(12)

1 Introduction to laser cladding

Industrial applications require parts with special surface properties such as good corrosion resistance, wear resistance and hardness. Alloys with those surface properties are usually very expensive and it is of great interest to reduce the cost of parts with these surface properties [1].

This cost reduction can be achieved by applying a hard or corrosion resistant surface layer to a cheaper substrate. Laser surface treatment includes several different surfacing techniques using the heat of the laser beam to modify the structure and physical characteristics of the surface of a material [2].

Laser cladding is the fusion of a different metal to a substrate surface, with a minimum of melting of the substrate. The surface alloy composition must be well controlled with a high bond strength to the substrate [3]. Surface coating by laser is a method that has been developed over the last two decades. The lasers minimal and easily controllable energy delivery makes it possible to alloy, impregnate, clad, and harden components that are exposed to wear and corrosion. The method offers great advantages compared with traditional hardening and alloying methods. The method is used commercially in the aircraft engine industry and in the car industry (G.M, etc.).

Laser cladding can be carried out in a single or a two-stage process. In the single stage process, the powder is blown into the interaction zone between the laser beam and workpiece. In the two-stage process the cladding material is pre deposited on the substrate. Both techniques (see figure 1) have the advantage of the possible deposition of a wide range of alloys either using a chosen alloy in powder form or by a blend of powders with the required composition. Laser cladding with powder offers the possibility of the development of new material combinations for the future.

(a)

(b)

Cladding material

V

Figure 1. Schematic diagrams of laser cladding process.

a) Preplaced powde,r b) blown powder cladding.

(13)

Relative motion between the laser/powder supply and the substrate can be used to

continuously apply a surface coating. To cover larger surfaces, overlapping tracks are made (see figure 2).

Figure 2. Schematic of the overlapping cladding process [4].

1.1 Blown powder laser cladding

The first reference that describes the laser cladding process by blown powder is a patent from Rolls Royce Ltd in the early eighties [5]. Blown powder laser cladding can produce a high quality cladding layer with low dilution. The powder is transported into the melt pool by a carrier gas and directed at an angle in the range 38-45° towards the substrate (see figure 2 ).

The powder particles are heated when they pass through the laser beam. Melting starts at the interface and the molten particles are trapped in the melt pool. The energy must be high enough to melt the powder without too much substrate melting [3]. The powder striking the substrate ricochets but the powder striking the melt pool is completely melted. With side blown powder there is a directional effect on the clad bead shape [6] and the powder

utilisation efficiency is low compared with coaxial powder nozzle feed [7]. The coaxial system in figure 3 can avoid this problem in some extent.

Inner nozzle

Powder stream in

Workpiece Focal point

Figure 3. Cross section of a coaxial nozzle.

(14)

1.2 Laser cladding with preplaced powder

Cladding with preplaced powder is the simplest method provided the powder can be made to remain in place until melted, while the area is being shrouded by an inert gas. Some form of binder is usually used, often this is an alcohol [6]. The preplaced powder method involves scanning the laser beam over the powder bed. The general theory for cladding of pre-placed powder may be understood on the basis of the work of Powell et al. [8].

2 Commercial examples

Industrial applications require parts with good wear, corrosion and hardness properties and laser cladding is a process which can fulfill all these requirements. Laser cladding can be used to good effect in processes which require a high productivity combined with flexibility without compromising on quality.

A high and uniform quality with a low heat input makes this process suitable for a wide range of applications in which minimum distortion is desired.

Examples of industrial laser cladding applications are:

• Improved wear resistance of bearings, valves, axles, cutting tools and other parts where the working conditions are very severe

• Improved corrosion resistance

• Repairing turbine parts, moulds, tools etc

• Building up complex geometries

Figure 4. Laser clad parts.

Typical commercial applications of laser cladding are carried out by SIFCO in Ireland who are involved in the remanufacture of turbine engine components. In recent years they have

devoted a large amount of resources to the research and development of new repair

technologies for the gas turbine industry. The deposited layer can have different composition, and subsequently properties, to the underlying material. This potentially has a range of

applications in a number of areas in particulary the aerospace and the automotive industries.

(15)

Figure 5. Cladding on turbine blade.

Duroc AB in Sweden Umeå has developed the technology for cladding material on valves etc for the nuclear power plant industry and the wood industry. Figure 6 below shows a part from a laser clad chopping tool of which the service life has increased 5-6 times compared to an untreated tool.

Figure 6. Laser clad chopping tool for the wood industry.

(16)

3. Summary of the chapters

3.1 Chapter 1. Energy redistribution in laser cladding; A comparison of Nd:YAG and CO₂ lasers

Abstract

Blown powder laser cladding is a cost effective way of producing a surface layer to withstand wear and corrosion. However, the cladding process is slow. Therefore is it of great interest to investigate how much of the laser power is used in the cladding process and how much is reflected etc. In this investigation an Nd:YAG and a CO₂ laser have been compared as energy sources for the process. Every aspect of the energy redistribution during cladding has been analysed. The main energy loss to the process for both lasers is by reflection from the melt pool and the powder cloud. It was found that the Nd:YAG laser cladding process is approximately twice as energy efficient as the CO₂ laser cladding process.

Figure 1. The redistribution of laser power during the cladding process (see text for definition of P_A, P_B etc).

(Power lost by convection) P_F

P_F

P_A P_A

P_B P_B

P_E P_E

P_D P_D

P Laser beam

Substrate

Powder stream

Substrate P Laser beam

(Power radiated) (Power reflected off the surface of the clad) (Power reflected off

the powder particles)

(Power lost by conduction) P_D

P_A

(17)

Where:

P_A = Power reflected off the surface of the clad zone.

P_B = Power reflected off the powder particles as they approach the weld pool.

P_D= Power lost by radiation from the cladding zone.

P_E= Power lost by convection from the cladding zone.

P_F = Power lost by conduction from the clad zone to the substrate.

P_G= Power absorbed by the powder particles which do not enter the cladding melt pool.

Conclusions

1. Ignoring the trivial contributions of convective and radiative cooling etc, the laser power applied to the cladding process is redistributed in the following ways:

*This value includes powder and substrate melting.

2. Nd:YAG lasers are approximately twice as energy efficient as CO₂ lasers for cladding in the range of parameters covered in this paper ( and by implication, the higher power (5 kW) range covered in our earlier work [2]) i.e. given the same laser power, Nd:YAG lasers are capable of approximately double the cladding rates of CO₂ lasers.

Laser type

CO₂ Nd:YAG

Power reflected off the cladding melt 50% 40%

Power reflected off the powder cloud 10% 10%

Power used to heat the substrate 30% 30%

Power used to melt the clad layer^* 10% 20%

Figure 2. The experimental arrangement for the analysis of the absorption and reflection of the energy by the powder cloud.

Laser Beam Powder Particles

Insulated Calorimeter

(18)

Figure 4. Cross sections of grooves showing that even when there is sufficient melt to produce a flat surface the clad layer does not do so when preplaced

powder is used.

3.2. Chapter 3. Laser cladding into pre machined grooves Abstract

When laser cladding is used to improve the wear characteristics of a substrate it is not always necessary to clad the whole surface. Wear resistant individual tracks can be clad directly onto the substrate or into pre machined grooves. This paper investigates the process parameters which affect the finished product when cladding into groves including; groove geometry, powder application method and laser type.

Substrate 200 mm Groove

Powder depth 2 mm Wedge of powder Powder

depth 0 mm

Figure 3. Schematic preplaced powder.

4 mm

(19)

Figure 5. a) A cross section of the type of clad profile achieved for preplaced powder cladding if the groove must be completely filled. 5b) A micrograph showing the clad – substrate interface weld.

0.1 mm 4 mm

a)

b)

Figure 6. Blown powder cladding results for 46 g/min powder flow at a process speed of 0.5 m/min (CO₂laser).

4 mm

Figure 7. The concave top profile of an under filled groove clad by the blown powder method.

A B C D 4 mm

(20)

a b 4 mm

Laser beam

x x

Figure 8. A pair of preplaced powder clad tracks produced under identical conditions except for the depth of the powder used. (3.5 kW CO₂ laser, spot size 4 mm, cladding speed 0.5 m/min)

powder deep a = 0.75 mm powder deep, b = 1.75 mm of powder.

Figure 9. The change in cross section of a clad track as more powder is added (“x” remains approximately constant as its width is determined by the laser beam diameter on the melt

pool).

(21)

Conclusions

1. It is possible to produce almost flat topped filled grooves by either CO₂ and Nd:YAG laser if blown powder cladding is employed.

2. Pre placed powder cladding does not give flat typed clad filled grooves. However the process may be used to produce a clad track with shallow grooves on either side which could aid lubrication (Once the central protruding part of the clad layer has been machined away).

3. Grooves with too large an aspect ratio cannot be effectively filled with melt.

4. The contact angle of a clad melt on a substrate can be varied and is determined by the laser beam diameter and the amount of powder supplied to the melt.

Substrate Removed excess

clad material Clad layer

Lubricant supply and debris removal conduits

Figure 10. Schematic cross section.

(22)

3.3. Chapter 4. Laser casting and laser clad casting: New process for rapid prototyping and production

Abstract

This paper presents preliminary experimental results from two new processes:

1. Laser casting involves a process similar to blown powder laser cladding but the final product is not joined to the substrate. The substrate surface therefore acts as a mould in a laser casting process and the eventual casting retains the topological features of the substrate.

2. Laser clad-casting involves the production of clad tracks which are welded as usual to a substrate but which are laid down between machined copper blocks. The eventual clad track therefore has its cross sectional profile determined by the blocks which are

removed after completion of the cladding process. In this way clad tracks with large depth to width ratios can be achieved with pre determinated cross sections.

Figure 13 shows the difference between laser cladding and laser casting.

Figure 11. Comparison of laser cladding and casting.

45°

10-15 mm

5 mm

45°

Interfacial melting between the clad layer and substrate

No interfacial melting

Unmelted layer of powder particles

(23)

3 mm

Machining line

Figure 12 shows an example of laser casting.

Figure 12. Successful laser clad-casting of cross hatched grooves. a) substrate (mould), b) substrate and casting, c) casting. Process parameters: laser power 3 kW (Nd:YAG), beam diameter 5 mm, process speed 0.8 m/min., Ni based powder, powder flow 80 g/min (in Ar), inter-track distance 3mm.

Figure 13 shows the difference between standard laser cladding (a+b) and clad-casting (c).

a) Standard clad b) Maximum height c) Required clad cross section clad track (semi circular cross section cross section)

Figure 13. Standard clad track cross section (a, b) and the required cross section (c ).

Clad layer

a) b) c)

(24)

Figure 14 shows the use of moulds in clad casting.

Figure 15 shows a successful laser clad cast.

Figure 15. A cross section of the clad-cast track deposited between copper blocks. (substrate width:

3mm,clad track height: 3.5 mm). Process parameters: powder feed (Nickel alloy) 40 g/min, cladding speed 0.5 m/min, laser power 3.5 kW (CO₂), beam diameter 4 mm.

Conclusions

It has been demonstrated that two new laser cladding techniques are possible and that they may provide novel answers to future production requirements.

Laser casting can be used to produce surface castings in high strength alloys to generate tool bits or stamping dies etc.

Laser clad-casting can be employed to make clad tracks with large depth to width ratios and pre determined cross sections.

Figure 14. Cross section of the clad cast mould.

Clamping Substrate Machined

copper blocks

3 mm

(25)

3.4. Chapter 5. Laser wire casting Abstract

This paper describes a new technique for the production of solid wire or rods from powder by laser melting. Three techniques have been developed to ensure that the molten powder

solidifies as a rod or wire rather than a series of droplets. The straight rods or wires produced in this way have an almost circular cross section, are several millimetres in diameter and can be pore free. The techniques can be used to produce welding rods, tensile test samples and other solid pieces from a wide range of powder mixes. The rapid thermal cycle involved means that hitherto difficult to produce mixtures and alloys can now be produced in the solid form in seconds.

Wire Powder

Mould Mould

Substrate

Laser beam

Powder

a) Cross section of mould and powder before laser irradiation

c) The cross section after laser irradiation b) During laser

processing

Figure 16. Laser wire casting.

(26)

Laser power 3kW Speed 0.4 m/min Mould separation 3 mm

Laser power 3kW Speed 0.4 m/min Mould separation 5 mm

Laser power 3kW Speed 0.4 m/min Mould separation

≥ 6 mm

Cross section Cross section Cross section

General view

General view General view

5 mm 6 mm

3 mm

5 cm

Figure 17. A selection of results of the side contact mould laser casting process.

(27)

18 a) The powder filled mould 18b) After successful prior to laser melting production of a rod

19a) Before laser melting 19b) After laser melting Figure 18. The use of a net shape mould to

form a rod.

Figure 19. Casting with wires imbedded in powder beds.

(28)

Conclusions

• Wires or rods can be cast from metal powder using a high power laser as a heat source.

• Metal powders which have been laser melted do not readily solidify as uniform cross section rods unless the tendency to form strings of droplets is inhibited.

• The presence of side wall or net shape moulds can result in rods which are ovoid or circular in cross section and approximately 100% dense. Wires incorporated into the powder bed can have the same effect in the absence of moulds.

• The casting techniques discussed in this paper could be used to produce wires or rods of a very wide range of alloys and alloy-ceramic mixtures.

3.5 Chapter 6. Melt-Solid Interactions in laser cladding and laser casting

Abstract

Experimental data in conjunction with mathematical models are used to explain various aspects of laser casting and laser cladding by the preplaced powder method. Results include an

explanation of the large range of process parameters over which low dilution clad deposits can be produced. Also the interaction of the melt pool with the powder bed is analysed to identify why laser castings have microscopically uneven surfaces.

(29)

Figure 20. Cross sections of clad tracks made under identical conditions (laser power 3500 W, powder bed depth 1 mm) at different speeds.

a) 0,1 m/min b) 0,2 m/min

c) 0,9 m/min d) 2,1 m/min

e) 3,3 m/min f) 3,8 m/min 1 mm

a b c d e f (0,1 m/min) (0,2 m/min)(0,9 m/min)(2,1 m/min)(3,3 m/min)(3,8m/min) Figure 21. The top views of the clad tracks shown in figure 20.

(30)

The main contra- intuitive feature of figure 20 is the surprisingly low amount of substrate melting over a wide range of process speeds. This phenomenon was first discussed by Powell who postulated a three stage melting process for preplaced powder laser cladding;

1. The laser rapidly melts the powder before the melt touches the substrate because, prior to substrate contact the melt is surrounded by low conductivity powder.

2. Once the melt touches the substrate it looses a great deal of heat by conduction.

This leads to partial solidification of the melt. As a result the melt-liquid interface does not move into the body of the substrate.

3. If the laser energy continues to irradiate the top surface of the melt, the energy will eventually move the melt/solid interface back down through the clad layer and across into the body of the substrate.

Figure 22 presents a graphical description of the three stage process derived from a one dimensional mathematical model.

Figure 22. Vertical temperature distribution through the preplaced powder and substrate for different time steps [9].

Figure 23. Calculated maximum melting depth through the powder (1 mm thick) and substrate ( >>¹

(31)

0 10 20 30 40 50

0 50 100 150 200

Particle Diameter [um]

Energy, Particle Distribution [a.u.]

N E N*E

Figure 24. Melt-substrate contact history in cross section.

(Black = liquid, Grey = Powder , Shaded = Solid).

Figure 25. The particle size distribution and proportion of the incident energy needed to melt the particles of different sizes in this batch.

(32)

The calculated surface shape and motion is shown in figure 26 for four different grain sizes as a function of time.

Figure 26. Calculated heating and melting of powder grains of different diameter touched by the melting front and subsequent smoothing of the droplets.

Figure 27 is a magnified photograph of the surface of a laser casting. The part of the surface shown is that which was in contact with the substrate. This photograph supports the model results presented in Figure 10 as it demonstrates that the liquid surface was covered in partially melted particles.

0,1 mm

(33)

4 Conclusions

This analysis of melt solid interactions has helped to explain the following points about the laser cladding and casting processes;

a) There is a wide parameter range over which dilution free cladding can be achieved by the preplaced powder process. This is primarily due to the difference in thermal conductivity of the powder bed and substrate.

b) If the process parameters are set outside the range mentioned above the result will be either a dilute clad layer (see figure 3a) or a casting process (see figure 2) depending on whether or not the power input to the process is increased or decreased.

c) The physics of powder particle melting by contact with a liquid pool makes it different to achieve laser casting with a smooth surface.

5 References

1. Riabkina-Fishman, M., Zahavi, J. (1996). Laser alloying and cladding for improving surface properties. Applied Surface Science, Vol. 106, no. 1-4, pp.

263-267

2. König, W., Rozsnoki, V., Kirner, P. (1992). Laser Treatment of Materials.

Conference proceedings (ECLAT 92) ISBN 3-88355-185-6, pp. 217-221

3. Yellup, JM. (1995). Laser Cladding using the powder blowing technique. Surface Coating Technology, Vol. 71, no. 2, pp. 121-128

4. Frenk, A., Vandyoussefi, M., Wagnière, J. D., Zryd, A., Kurz, W. (1997).

Analysis of the laser-cladding [laser surfacing] process for stellite on steel.

Metallurgical and Material Transactions B, Vol. 28B, pp. 501-508

5. Hoadley, A, Rappaz, M. (1992). A thermal model of laser cladding by powder injection. Metallurgical Transactions B, Vol. 23B, pp. 631-641

6. Steen, W.M. Laser Material Processing. (1998). Second edition. Springer-Verlag London. ISBN 3-540-76174-8, pp. 199-202

7. Hu, U.P., Chen, C.W., Mukherjee, K. (1997). An analysis of powder feeding systems on the quality of laser cladding. Metal Powder Industries Federation USA, pp. 21.17-21.31

8. Powell, J., Henry, P.S., Steen, W.M. (1988) Laser cladding with preplaced powder. Analysis of thermal cycling and dilution effects. Surface engineering, Vol

(34)

Chapter II Energy Redistribution in Laser Cladding; A

comparison of Nd:YAG and CO

₂

lasers

(35)

(36)

Energy Redistribution in Laser Cladding; A comparison of Nd:YAG and CO

₂

lasers

The following chapter combines information from two published papers;

1) Energy Redistribution During CO₂ Laser Cladding

(Published in Journal of Laser Applications. Vol. 14, no. 2, pp. 78-82. May 2002)

H.Gedda*, J.Powell⁺, G.Wahlström**, W-B. Li*, H.Engström*, C.Magnusson

*.

* Luleå University of Technology, Division of System and Manufacturing Engineering, S-971 87 Luleå, Sweden Phone: +46 920 91169, E-mail: Hans.Gedda@ltu.se + Laser Expertise Ltd., Acorn Park Industrial Estate, Harrimans Lane, Nottingham NG7

2TR, U.K.

** Duroc AB, Industrivägen 8, S-90130 Umeå Sweden

2) A Process Efficiency Comparison of Nd:YAG and CO₂ Laser Cladding (Published in Welding in the World, vol.46, Special Issue. pp.75-86. July 2002)

H.Gedda*, J.Powell⁺, A.Kaplan*.

* Luleå University of Technology, Division of System and Manufacturing Engineering, S-971 87 Luleå, Sweden Phone: +46 920 91169, E-mail: Hans.Gedda@ltu.se

+ Laser Expertise Ltd., Acorn Park Industrial Estate, Harrimans Lane, Nottingham NG7 2TR, U.K.

Abstract

Blown powder laser cladding is a cost effective way of producing a surface layer to withstand wear and corrosion. However, the cladding process is slow. Therefore is it of great interest to investigate how much of the laser power is used in the cladding process and how much is reflected etc. In this investigation an Nd:YAG and a CO₂ laser have been compared as energy sources for the process. Every aspect of the energy redistribution during cladding has been analysed. The main energy loss to the process for both lasers is by reflection from the melt pool and the powder cloud. It was found that the Nd:YAG laser cladding process is approximately twice as energy efficient as the CO₂ laser cladding process.

Keywords: Laser cladding; Laser processing, Energy redistribution, Surface treatment.

(37)

1 Introduction

Blown powder laser cladding involves projecting a stream of metal powder (in an inert gas jet) into a laser generated melt pool on the surface of a metal substrate (see figure 1).

The result of this process is a clad track of the cladding metal on the substrate. Such tracks can be overlapped to cover areas of the substrate with a harder and/or more corrosion resistant surface. The process is not energy efficient as a large proportion of the incoming laser power is reflected or reradiated from the cladding zone as shown in figure 2. Figure 2 demonstrates all the different ways in which the incident laser energy is redistributed during the cladding process.

Figure 1. Blown powder laser cladding.

Powder particles Laser beam

Clad layer

(38)

A power balance for laser cladding can be expressed as follows:

P_tot= P_C+P_L (1)

Where:

P_tot= The output power of the laser.

P_C = The power utilised in melting the cladding material and welding it to the surface of the substrate.

P_L= The power lost by reflection, radiation, convection etc.

P_c in equation 1 can be expanded as follows:

P_C= P_P+P_S (2)

Where:

P_P = The power utilised in melting the cladding powder.

P_S= The power utilised in melting the surface of the substrate in order to achieve aclad/substrate weld.

P_L in equation 1 can be similarly expanded:

P_L= P_A+P_B+P_D+P_E+P_F+P_G (3) Figure 2. The redistribution of laser power during the cladding

process (see text for definition of P_A,P_B etc).

(Power lost by convection) P_F

P_F

P_A P_A

P_B P_B

P_E P_E

P_D P_D

P Laser beam

Substrate

Powder stream

P_D P_A

Substrate P Laser beam

(Power radiated) (Power reflected off the surface of the clad) (Power reflected off the powder particles)

(Power lost by conduction)

(39)

Where:

P_A = Power reflected off the surface of the clad zone.

P_B = Power reflected off the powder particles as they approach the weld pool.

P_D= Power lost by radiation from the cladding zone.

P_E= Power lost by convection from the cladding zone.

P_F = Power lost by conduction from the clad zone to the substrate.

P_G= Power absorbed by the powder particles which do not enter the cladding melt pool.

Figure 2 gives a visual representation of equation 3. Of course these “losses” are to some extent necessary to the cladding process; It is not possible to heat a metal to well above its melting point without having radiant or convective thermal losses, a liquid sitting on a comparatively cool solid will always lose heat by conduction etc. For the purpose of this discussion however, it will be taken that any influence which could minimise P_A, P_B, P_D, P_E, P_F or P_G would increase the efficiency of the cladding process. This reduction in any of the factors of equation 3 would, of course, increase the proportion of the power available to the cladding process.

The aim of commercial cladding is to cover the surface of one metal with another at the lowest cost. Clad depths are usually stipulated and the biggest cost element of the process is laser time.

Therefore the simple aim of commercial cladding can be expressed as follows:

• To cover metal A with a known thickness of metal B at the fastest possible rate with a high quality interfacial bond.

Returning to equation 1 it is clear that the process can be speeded up if there is an increase in the proportion laser power available producing the clad layer P_C. The requirement here would be to melt enough powder to achieve the correct clad thickness at a faster linear speed. Such an increase in P_C must not be employed to melt the substrate to a greater depth. The process must be accelerated to achieve the same (minimum) substrate melt depth at a higher process speed.

To summarise:

• The efficiency of laser cladding could be improved by minimising any of the losses in equation 3. This would lead to an increase in P_Cand the process could be accelerated to produce the same clad depth with a minimal depth of substrate melting.

Earlier work by the present authors [1] quantified the individual elements of equations 1,2 and 3 for CO₂ laser cladding. The results of that work concluded that the laser power was

redistributed in the following proportions:

Power reflected off the workpiece (P_A) Power reradiated from the workpiece (P_D) Power reflected off the particles (P_B) Power absorbed by the process (P_C+P_F)

= 50%

= 1%

= 9%

= 40%

100%

(40)

Of the power absorbed by the process (40%) three quarters of it was employed in simply heating the substrate and only the remaining 10% of the original laser power was used to melt material to produce a clad layer.

This present work involves repeating this quantification of the power redistribution for Nd:YAG and CO₂ laser cladding in order to compare the efficiency of the two types of laser for this process. These experimental trials were carried out at a laser power of approximately 3 kW for both types of laser. This allowed a direct comparison of the lasers and also a

confirmation of the previous published results [1] at a different power level (the earlier work was carried out at a power level of 5 kW).

2 Experimental work

2.1 General

The substrate material used in this study was (SS 2172) steel with the following composition:

Table 1. Steel composition (substrate)

C Si Mn P S V N Fe

wt % 0.16 0.22 0.94 0.014 0.022 0.06 0.009 98.6

The cladding material was cobalt based with the following composition:

Table 2. Cladding powder composition

Cr C Si Mo Ni Fe Co

wt % 27.2 0.27 1.0 5.5 2.3 0.3 63.4

The substrate specimens were grit blasted before cladding was carried out. The laser used was a Rofin Sinar RS 6000 CO₂ laser with a maximum output power of 6 kW and the Nd:YAG laser was a Haas Laser HL 3006 D 4 kW. The powder feeder was a TECFLO ^TM 5102. The shielding/carrier gas employed to propel the powder was argon.

2.2 The power absorbed by or reflected off the powder cloud above the clad zone During the cladding process the laser beam must travel through the powder cloud in order to reach the cladding zone (see figure 2). A proportion of the laser energy is reflected off the powder cloud and is lost to the cladding process. Another portion of the incident energy is absorbed by the particles but some of this energy is also lost to the process because not all the heated particles join the cladding melt pool.

A simple experiment was set up to discover what proportion of the original laser power would penetrate the powder cloud (see figure3 below). A commercially available “power probe” was used to measure the laser power with and without the powder stream turned on. The powder flow rates were typical of the cladding process as were all the other process parameters. The average results from several such tests are presented in table 3. The energy absorbed by the

(41)

after it had passed through the beam (see figure 3). The power reflected off the powder cloud could then be easily calculated as shown in table 3.

Table 3. Power absorbed and reflected by the powder cloud irradiated by the two types of laser Laser

type Laser output Power *

(Watts)

Powder flow rate

(g/min)

Post powder cloud power

(Watts)

Total power reflected and

absorbed by the powder cloud **

(Watts)

Power absorbed by powder cloud **

(Watts)

Power Reflected off powder cloud (P_B)

**

(Watts) Nd:YAG 2743

(100%) 30 2506

(91%) 237

(9%) 18

(1%) 224

(8%)

CO₂ 2695

(100%)

30 2457 (91%)

238 (9%)

22 (1%)

218 (8%)

* Measured by the power probe but with zero powder flow.

** Percentages are approximate

It is clear from table 3 that we now have an approximate value for P_B (the power reflected off the powder cloud) for the parameter range covered here:

P_B= 8 % P_totfor the Nd:YAG laser and the CO₂laser (4)

One other component of equation 3 can also be identified from table 3 after the same

parameters were used for actual cladding. This parameter is P_G, the level of power absorbed by particles which do not enter the cladding melt pool. A number of cladding trials were carried out and these showed that, over this range of parameters, the proportion of particles which formed the clad track was 60% (The range was 57%-63%). It can then be concluded that 40%

of the heat collected by the powder cloud (1% P_tot see table 3) does not contribute to the cladding process.

Figure 3. The experimental arrangement for the analysis of the absorption and reflection of the energy by the powder cloud.

Laser Beam Powder

Particles

Insulated Calorimeter

(42)

2.3 The power lost by radiation from the cladding zone (P_D).

The total energy radiated from the clad pool can be calculated from the pool temperature, surface area and emissivity. If the emissivity of the liquid metal pool is taken as equal to one then the calculation is simplified and the maximum possible radiation power can be estimated:

P_D= σ^T⁴^A ⁽⁶⁾ Where:

σ is the Stefan-Boltzman constant (5.7*10^-8Wm^-2K^-4) T is the surface temperature of the melt (K)

A is the area of the melt surface (m²)

In this case the surface temperature of the melt was approximately 2300 K [1] and its surface area was 19 mm².

This gives a maximum value for P_D of:

P_D=5.7*10⁻⁸*(2300)⁴*19*10⁻⁶= 30 Watts (7)

P_D≈ 1% of Ptot for both the Nd:YAG and the CO₂ laser (8)

2.4 Power lost by convection from the clad zone (P_E).

The cladding zone is a molten alloy with a surface temperature of ≈ 2300 K

and a known surface area. This melt is exposed to a stream of argon which carries the powder to the clad zone. The argon flow was measured and found to have an average flow velocity of 4.3 m/sec.

The rate of convective cooling of a hot body exposed to a cooler gas is given by:

Q=hA∆t Watts (9)

Where:

h = Heat transfer coefficient.

A = Surface area of the hot body.

∆t = The difference in temperature between the body and the cooling gas.

Evaluation of h from a standard text on the subject [2] gives us a value of approximately 100 W/m²K.

Q = ¹⁰⁰^*

(

^π^*⁰^.⁰⁰²⁵²

)

^*²⁰⁰⁰ (10)

P_E= 3.9 Watts (11) Or P_E= 0.1% P_tot for both Nd:YAG and CO₂ laser (12)

(43)

2.5 Power reflected off the surface of the clad zone (P_A).

Calorimetry was employed to measure the heat input to each clad sample (P_in). From this measurement it is possible to measure the power reflected off the cladding zone (P_A) in the following way:

P_A = P_tot– ( P_B+P_D+P_E+P_G+P_in) (13)

Table 4 shows the average results from the calorimetric measurements over a range of process parameters.

Table 4. Calorimetric measurements (average values)

From our earlier results:

P_A= 100 – (8+1+0.1+0.8+49) (Nd:YAG) (14) P_A= 41.1 % (Nd:YAG) (15) P_A= 100 – (9+1+0.1+0.8+39) (CO₂) (16)

P_A= 51.1 % (CO₂) (17) So far this is the first time that the measurements from the two types of laser have shown an appreciable difference. In summary it can be said that, for the CO₂ laser, approximately half of the laser power is reflected from the cladding zone. For the Nd:YAG laser this value is reduced to approximately 40%. These generally high reflectivity values confirm the work of other authors in the field [3] who suggest that the onset of melting is associated with a rise in material reflectivity. This is because a molten surface in an inert atmosphere (in this case argon) is smooth and oxide free. This smooth, oxide free surface acts as a better reflector than the solid, rough, oxidised surface which exists before melting. It is well known [4] that metals have a lower reflectivity for the 1.06 µm radiation of Nd:YAG lasers than for the 10.6 µm radiation of CO₂ lasers and this is confirmed by the above results. As we will see later in this paper, this reduction in reflectivity for the Nd:YAG laser results in a marked increase in process efficiency when cladding as compared to a CO₂ laser.

Laser

type Laser output

Power (P_tot) (Watts)

Power input to clad sample (P_in) (Watts)

Power % input to sample

Nd:YAG 2743 1367 49%

CO₂ 2695 1044 39%

(44)

2.6 The power utilised in melting the clad layer to the substrate (P_C)

Blown powder laser clad layers usually have a cross sectional geometry similar to that shown in figures 4 and 5.

Figure 4. The cross sectional geometry of a blown powder laser clad layer. Note: the melted substrate and cladding material are mixed together during the process.

As figure 4 demonstrates, the production of a clad layer usually involves melting the surface layers of the substrate. The amount of substrate melting can range from minimal to levels where the clad layer is really a dilute alloy of the substrate and cladding material.

Figure 5. Macrographs of a typical laser clad sample in cross section (see also figure 4).

P_C, the power utilised in melting the cladding material and welding it to the surface of the substrate can be calculated as follows:

P_C= Avρ(Cp∆t + ∆Hm) (18)

Where:

A = The melt cross sectional area (m²).

v = The Cladding speed (m/s).

ρ = The Density of the material melted.

Cp = Specific heat capacity of the material melted.

∆t = The difference between the melt temperature and ambient.

∆Hm= Specific heat of melting of the clad melt.

6.2 mm

Clad layer

HAZ

Melted substrate

Clad layer

Melted substrate Heat affected zone

(45)

In order to evaluate P_caccurately for both the CO₂ and the Nd:YAG laser a set of cladding trials were carried out. Laser power and focusing conditions were kept identical for the two lasers (2700 W, 5 mm beam diameter). Powder flow rates of 30, 40 and 50 g/min were employed at cladding speeds of 0.7, 0.8, 0.9, 1.0, 1.2 and 1.4 m/min.

From all these tests an average for P_c was calculated for both types of laser. Figure 6 shows cross sections of clad traces produced by both types of laser at a powder flow rate of 40 g/min and speeds of 0.7, 1.0 and 1.4 m/min.

a) CO₂laser

b) Nd:YAG laser

Figure 6. Clad cross sections at increasing process speed for both types of laser. (laser power 2700 W, laser spot diameter 5 mm, powder flow rate 40 g/min.

It is clear that a substantial amount of substrate was melted in each case. On average the melt was found to consist of a 40% substrate; 60% clad material mix for the CO₂ laser and 55%

substrate and 45% metal mix for the Nd:YAG laser. As a simplification, the material properties necessary for equation 18 were taken as being for a 50:50 mixture of cladding material and substrate.

ρ = 8020 kg/m³

Cp = 500 J/kg K

0.7 m/min 1.0 m/min 1.4 m/min

3 mm

0.7 m/min 1.0 m/min 1.4 m/min

(46)

From these values and a melt temperature of 2300 K [1], the values for P_cfor the two types of laser were:

P_C(Nd:YAG) = 506 W (19)

Which represents 18% of P_tot

P_C (CO₂) = 266 W (20)

Which represents 9.5 % of P_tot

This is a remarkable result. Here we can see that the Nd:YAG laser melts approximately twice as much material its CO₂ counterpart.

In our earlier work [1] we found that for a CO₂laser at a power of 5 kW only 10% of the laser power was used to melt metal during cladding. This result is confirmed here for a different set of process parameters. For the Nd:YAG laser however, the proportion of the laser power involved in melting is almost double the CO₂ value. Section 2.5 revealed that 50% of the CO₂ laser light was reflected from the clad zone as compared 40% for the Nd:YAG laser. It seems then, that the difference of 10% is almost exclusively given over to material melting and this represents a doubling of the energy available for melting.

2.7 Power lost by conduction from the clad zone to the substrate (P_F)

This value is easily established by subtracting (P_C) from the total power absorbed by the workpiece (P_in). Taking average values:

(Nd:YAG) P_F= P_In- P_C≈ 30%Ptot (21)

(CO₂) P_F= P_In- P_C≈ 28%Ptot (22) This is the result which would be expected given that all extra laser power which joins the

process when an Nd:YAG laser is used is involved in the melting process , (see previous section).

(47)

3 Discussion

Figure 7 presents schematics of the redistribution of energy during the laser cladding process for both types of laser.

Figure7. Schematic of the redistributions of energy during the laser cladding process (percentages are approximate).

For the sake of clarity P_E (convective losses) and P_G (lost powder losses) have been left out of figure 7 as their contribution to the energy balance is negligible.

Power absorbed by the cladding process 40% 50%

Power used to melt the Clad layer (P_C) 10% 20%

Power absorbed in heating the substrate (P_F) 30% 30%

CO₂ Nd:YAG % of power % of power Raw Power of beam (∼3 kW) (Ptot) 100% 100%

Power Reflected off the workpiece (P_A) 50% 40%

Power Re radiated from the workpiece (P_D) 1% 1%

Power Reflected off particles (P_B) 8% 8%

Power absorbed by the Process (P_C+P_F) 40% 50%

∼100%

(48)

The major mechanism of energy loss to the process is that of reflection from the melt pool and the powder cloud. Reflection off the melt pool is a function of the condition of the melt surface. As this melt is produced in an inert atmosphere it experiences no surface oxidation and thus has a high reflectivity. Dilution of the inert shroud gas with an oxidising agent would decrease this reflectivity but may have disruptive consequences on the stability of the process and the metallurgy of the clad track. Overlapping such deliberately oxidised tracks could also prove problematic.

A reduction of reflective losses from the powder cloud on the other hand would be

comparatively easy. All that is necessary is an increase in the average particle diameter. During its passage through the laser beam the particle interacts with and reflects light over an area equal to its cross sectional area (πr²) rather than half its surface area (2πr²). This because the shadow cast by any particle has an area of πr² where r is the particle radius. A particle of twice the original radius would cast a shadow four times as big but would have eight times the mass (mass ∝ r³). Thus it is clear that, for a set mass flow rate, larger particles interact with (and reflect) less of the beam. This is of course only useful within certain limits as the cladding process will break down if the particles are too large.

One very important feature of figure 7 is that, although the power absorbed by the process increases only from 40% to 50% when CO₂ and Nd:YAG lasers are compared, the power employed in melting material increases by a factor of 2 from 10% to 20 %. This doubling of the energy efficiency of the process is clearly demonstrated in figure 8 which compares a low speed (0.7 m/min) CO₂laser clad sample with an Nd:YAG laser clad sample carried out at twice that speed (1.4 m/min).

a) CO₂ laser 0,7 m/min b) Nd:YAG laser 1.4 m/min

Figure 8. A demonstration of the doubling of the process speed possible when using an Nd:YAG rather than CO₂ laser.(The powder feed rate was increased from 30 g/min for the CO₂ laser to 50 g/min for the Nd:YAG laser but the laser power (∼ 3kW) and spot size (5 mm) were kept constant.)

The doubling of the process efficiency shown in figures 7 and 8 would not be possible if the

“powder absorbed in heating the substrate” (P_F see fig 7) changed as more power was absorbed by the process. P_F remains steady (in this case at 30%) because it is determined by the amount of power the substrate needs to absorb before surface melting is initiated. This is a threshold value, which will not charge with increasing absorptivity. This being the case, any increase in absorbed power will be entirely available to the melting process.

3 mm

(49)

The CO₂ laser results given in figure 7 are almost identical to the earlier published figures from experiments carried out at 5 kW on different equipment [1]. It is therefore possible to say that these results are typical of multi kilowatt laser cladding.

4 Conclusions

1. Ignoring the trivial contributions of convective and radiative cooling etc, the laser power applied to the cladding process is redistributed in the following ways:

*This value includes powder and substrate melting.

2. Nd:YAG lasers are approximately twice as energy efficient as CO₂ lasers for cladding in the range of parameters covered in this paper ( and by implication, the higher power (5 kW) range covered in our earlier work [2]) i.e. given the same laser power, Nd:YAG lasers are capable of approximately double the cladding rates of CO₂ lasers.

3. As a large proportion (30%) of the laser power is consumed in heating the substrate it is likely that substrate pre heating by a cheaper power source^* would improve the

profitability of laser cladding. (^* flame, plasma, induction etc).

5 References

1. Gedda, H., Powell, J., Wahlström, G., W-B, Li., Engström, H., Magnusson, C.

(2002). Energy Redistribution During CO₂ Laser Cladding. Journal of Laser Applications, Vol. 14, pp. 78-82

2. Porier, D.R., Geiger, G.H. (1994). Transport Phenomena in Materials Processing. The Minerals, Metals & Materials Society, ISBN 0-87339-272-8, pp. 219-236

3. Bloehs, W., Grünenwald, B., Dausinger, F., Hügel. (1996). Recent progress in laser surface treatment. Part 1: Implications of laser wavelength. Journal of Laser Applications, Vol. 8, pp. 15-23

4. Steen, W.M. Laser Material Processing. (1998). Laser surface treatment.

Springer-Verlag London. Second edition, ISBN 3-540-76174-8, pp. 199-202

Laser type

CO₂ Nd:YAG

Power reflected off the cladding melt 50% 40%

Power reflected off the powder cloud 10% 10%

Power used to heat the substrate 30% 30%

Power used to melt the clad layer^* 10% 20%

(50)

Chapter III

Laser Cladding into pre machined grooves

(51)

(52)

Laser Cladding into pre machined grooves.

J.Powell

⁺

**, H.Gedda, A.Kaplan.**

+ Laser Expertise Ltd., Acorn Park Industrial Estate, Harrimans Lane, Nottingham NG7 2TR, U.K.

* Luleå University of Technology, Division of System and Manufacturing Engineering, S-971 87 Luleå, Sweden Phone: +46 920 91169, E-mail: Hans.Gedda@ltu.se

Abstract

When laser cladding is used to improve the wear characteristics of a substrate it is not always necessary to clad the whole surface. Wear resistant individual tracks can be clad directly onto the substrate or into pre machined grooves. This paper investigates the process parameters which affect the finished product when cladding into groves including; groove geometry, powder application method and laser type.

1 Introduction

Laser cladding is a process by which a metal powder is melted onto the surface of a metal substrate. There are two common methods of providing powder for this process;

a) Pre placed powder; where a layer of powder is applied to the surface of the substrate and subsequently melted by the laser (see figure 1a).

b) Blown powder; where powder is propelled into the cladding melt pool by means of a non oxidising gas stream (see figure 1b).

Figure1. (a) Preplaced and (b) blown powder

a)

b)

Cladding material Laser beam

(53)

Laser cladding can be used to provide a protective coating of hard or corrosion resistant metal on a weaker substrate. Tracks of the harder, powdered material are laid down next to each other to form a new surface as shown in figure 2.

The individual clad tracks which go to make up a clad layer have their cross sectional shape determined by a number of factors including laser power, laser beam width and powder

characteristics etc. Typical individual clad tracks produced by the preplaced and blown powder methods are presented in cross section in figure 3.

It is clear from figure 3 that individual tracks of laser melted powder would not generally be useful in an engineering context as the harder material forms a ridge on the substrate. This is different situation from that experienced in the field of laser surface hardening. Surface

hardening [1-3](which does not affect the substrate surface flatness) has been successfully used HAZ

Significant substrate melting

Substrate 4 mm

Substrate Heat affected zone

(HAZ)

Minimal clad layer/substrate interfacial melting

1 mm

Figure 2. A cross section of a laser clad layer of Ni-based material on a low carbon (SS 2172) steel.

Figure 3 a. A typical cross section of a single clad track produced by the preplaced powder method.

Figure 3b. A typical cross section of a clad track produced by the blown powder method.

1 mm

(54)

One early application of single track hardening was employed by the automobile industry to improve the wear characteristics of a piston and cylinder [4]. In this case a spiral track was produced on the piston and this interacted with three or four straight hardened lines down the length of the internal face of the cylinder.

In order to replicate the advantages of the single track approach for cladding it is necessary to deposit the cladding material into pre machined grooves. This paper investigates the effect of the cross sectional shape of the grooves on the eventual clad track.

2 Experimental work

If grooves are to be filled with cladding material it is important to optimise the cross sectional geometry of the groove. For this experiment “V” shaped grooves were produced with included angles of 30º, 45º, 60º, and 90º.

The gap at the top of the grooves was kept constant at 4 mm. These grooves were clad using CO₂ and Nd:YAG lasers both operating at a power of 3 kW. The substrate was mild steel and the cladding material was Nickel based super alloy (see table 1).

Both the pre-placed and blown powder techniques were investigated as follows;

a) For blown powder the mass flow of the powder stream (in Argon) was increased in five steps from 22 to 46 grams per minute.

b) For pre placed powder a wedge of powder was prepared over the groove as shown in figure 4. The depth of the powder increased from zero at one end of the groove to 2 mm at the other end (200 mm away).

A photograph of a groove clad in this way is presented in figure 5. This use of a wedge of powder is useful in demonstrating the progressive effect of an increase in powder depth. All powder wedge samples were produced at a process speed of 0.5 m/min.

Table I. Equipment and Parameters CO₂ laser, Rofin Sinar RS 6000 (6 kW) Nd:YAG, Haas Laser HL 3006 D (4 kW) Spot size at top of groove (both laser ) = 4 mm

Substrate, SS 2172 Mild steel (0.16% C) Cladding Material, Nickel based (80% Ni, 20% Cr)

Powder feeder, Sulzer Metco Single 10 C

(55)

Figure 6. Cross sections of grooves showing that even when there is sufficient melt to produce a flat surface the clad layer does not do so when preplaced

3 Results

A. CO2 laser; Preplaced Powder

Figure 6 shows cross sections of the preplaced powder cladding trials at the section in the sample where there was enough melt to fill the top of the groove. It is clear that for all these samples the melt has not assumed a flat top surface. In all these cases the melt has retained its circular curvature towards the top of its cross section. It is also noticeable that there is a pore along the bottom of the clad groove for angles less than 90º.

Substrate

200 mm Groove

Powder depth 2 mm

Wedge of powder Powder

depth 0 mm

Figure 5. A grooved powder wedge sample (see figure 4) after cladding.

Figure 4. Schematic preplaced powder.

4 mm

(56)

Figure 7 demonstrates that the addition of more powder to the melt results in a clad trace which over fills the groove. This sample also demonstrates the very low amount of substrate melting which is often typical of pre placed powder cladding [5].

B. CO₂ laser; Blown Powder

Figure 8 shows the cross sections of blown powder cladding for the four types of groove at a powder flow rate of 46 g/min and a process speed of 0.5 m/min.

Figure 8 reveals that the 30º groove is unsuitable to the process because the powder stream does not project sufficient material into the bottom of the groove. The 45º and 60º groove are successfully filled with almost flat top surface although there are small linear pores at the bottom of the grooves. The 90º grooved sample has become overfilled with melt at this powder flow rate as its cross sectional area is considerably smaller than those for the other angles.

0.1 mm 4 mm

7a)

7b)

Figure 7. A cross section of the type of clad profile achieved for preplaced powder cladding if the groove must be completely filled. 7b A micrograph showing the clad – substrate interface weld.

Figure 8. Blown powder cladding results for 46 g/min powder flow at a process speed of 0.5 m/min (CO2 laser).

A B C D 4 mm

(57)

As far as producing an overall flat surface is concerned, samples b and c in figure 8 are much more successful than the preplaced powder samples shown in figure 6. The reason why these cross section are flatter must be attributed to the action of the powder jet gas flow on the solidification dynamics of the melt. This point is supported by figure 9 which shows that under filled grooves produced by the blown powder method had concave rather than convex top profiles. Figure 9 also demonstrates the increased substrate melting common to blown powder cladding.

C. Nd:YAG laser; Preplaced Powder

Figure 10 demonstrates that a change of laser type from CO₂ to Nd:YAG does not produce a flat surface when preplaced powder is employed. The results are very similar to those given in figure 6 for the CO₂ laser.

D. Nd:YAG laser; Blown Powder

Figure 11 shows the results of blown powder cladding with the Nd:YAG laser and the

maximum flow rate. Once again the 45º groove produces an almost flat top surface and because of its smaller cross section the 90º groove is overfilled. It was noticed however that the 90º samples for both types of laser did not produce flat clad surfaces even at lower powder flows.

This retention of a curved upper melt surface is possibly related to the superior heat sink capacity of the 90º grooves.

Figure 9. The concave top profile of an under filled groove clad by the blown powder method.

4 mm

Figure 10. Tracks made by Nd:YAG laser and preplaced powder when there is enough melt to fill the groove. Once more the melt retain its curved upper surface.

4 mm

Laser cladding: an experimental and theoretical investigation

DOCTORA L T H E S I S