laser rapid mfg on vertical surfaces

Surface & Coatings Technology 224 (2013) 18–28

Contents lists available at SciVerse ScienceDirect

Surface & Coatings Technology

j ourna l homepage: www.e lsev ie r .com/ locate /sur fcoat

Laser rapid manufacturing on vertical surfaces: Analytical and experimental studies

C.P. Paul ⁎, S.K. Mishra, Atul Kumar, L.M. KukrejaLaser Materials Processing Division, Raja Ramanna Centre for Advanced Technology, PO: RRCAT, Indore, 452 013 MP, India

⁎ Corresponding author. Tel.: +91 731 248 8384; faxE-mail address: [email protected] (C.P. Paul).

0257-8972/$ – see front matter © 2013 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.surfcoat.2013.02.044

a b s t r a c t
a r t i c l e i n f o
Article history:Received 9 November 2012Accepted in revised form 21 February 2013Available online 5 March 2013

Keywords:Laser rapid manufacturing onvertical surfacesTrack geometryAnalytical modellingGravity effect

Analytical and experimental studies on geometrical aspects of the deposited tracks were carried out at differentprocessing parameters for laser rapid manufacturing (LRM) in vertical surface configuration using AISI type 304stainless steel powder on the substrate of the samematerial. The vertical downward shift of the deposited trackand its peak due to the gravity flow of the melt were found to follow quadratic dependence on the track height.The downward rounded bulging was found to be quite significant for the scan speeds lesser than 200 mm/min,while this was insignificant for the scan speeds more than 400 mm/min. A set of consolidated processing parame-ters for continuousmaterial depositionwas identified. The threshold value of laser energy and powder fed, both perunit traverse length for the continuous deposition were found to be ~96 J/mm and ~0.006 g/mm respectively. Themaximum powder catchment efficiency was ~42% for stand-off distances in the range of 15–18 mm. The surfacewaviness factor was found to decrease from ~0.95 to ~0.05 when the overlap index was increased from 30% to80%. The study provides a deeper insight into the ensuing geometrical aspects of the tracks using LRM in verticalconfiguration.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Laser rapid manufacturing (LRM) is one of the advanced additivemanufacturing processes. It is similar to laser cladding/alloying at the pro-cess endwith an extended capability of fabricating three-dimensional ob-ject directly from a solidmodel. LRM employs a high power laser beam asa heat source tomelt a thin layer on the surface of the substrate/depositedmaterial and fedmaterial to deposit a new layer as per shape and dimen-sions defined in numerical control code as per the solidmodel. A numberof such layers are deposited one over another resulting in three dimen-sional (3D) components. LRM offers many advantages over conventionalsubtractive techniques, such as reduced production time, better processcontrol and capability to form functionally graded parts [1]. It employs ahigh power laser like CO2, Nd:YAG, diode and fibre as energy source tomelt and deposit a layer of the desired material in the form of powderor wire onto the substrates/previously deposited layers forming a soundmetallurgical bond [2]. A wide variety of the deposit materials and sub-strates are reported in literature for various applications in automotive,aerospace,machinery, petrochemical, power generation and shipbuildingindustries [1–6]. Sincemost of the laser rapidmanufacturing applicationsinvolve deposition of materials on the horizontal surfaces, there havebeen theoretical and experimental studies of laser rapid manufacturingprocess in the horizontal configuration [7–13]. The technology isalso being investigated for the laser rapid manufacturing of porous

: +91 731 248 8380.

rights reserved.

structures [14]. Alimardani et al. developed a comprehensive model toevaluate the track geometry [15]. The model used the conservation ofmass within the process domain for material addition to predict localtrack height by incorporating catchment efficiency into powder feedon themolten substrate surface for each time interval. Hofman et al. de-veloped a FEMbasedmodel for the determination of the track geometryand dilution during the process [16]. An analytical approach for estimat-ing the track geometries (height and width) was presented byWang etal. based on the mass conservation of powder feeding stream [17]. Thesimulation was capable of predicting the track width and height withreasonable accuracy at medium powder feed rate. Recently, Kumar etal. used a finer modelling approach for numerically predicting singletrack geometry in two dimensions [18]. The approach involved the cal-culation of excessive enthalpies above melting point for all nodal pointsin the process domain and using those for the computation of localtrack height at every node along the track width on the substrate. LRMin vertical configuration did not find much attention except for a few ef-forts reporting the development of laser vertical cladding system [19,20].LRMon vertical surface substrate configuration is important formany en-gineering applications, such as surface cladding of turbine blade shroudand interlock, off-shore drilling heads, cylinder body, sleeve and mouldsidewalls etc. This paper presents theoretical andexperimental investiga-tions on dynamic geometrical aspects of the tracks at different processingparameters for LRM on vertical surface configuration using AISI type 304stainless steel powder on the surface of the same material. A newlydesigned LRM head was augmented to the existing laser workstationand successfully used for LRM on vertical surface substrate configuration.An analytical model incorporating gravity force was developed to under-stand its effect on molten deposits and its subsequent downward flow

Nomenclature

ai activity of species i in molten pool, weight %Aγ constant in surface tension gradient, N/(m K)Cp specific heat capacity, J/(kg K)D thermal diffusivity of the material, m2/sEl laser energy per unit traverse length, kJ/mg gravity, m/s2

hc combined heat transfer coefficient for radiative andconvective boundary conditions W/(m2 K)

ΔH0 standard heat of adsorption, J/(kg.mol)hmax maximum height of the deposited/overlapped track, mhmin minimum height of the overlapped tracks, mi overlap indexj waviness factork thermal conductivity, W/(m K)k1 constant related to entropy of segregation (3.18 × 10−3)L1 melt pool length in the forward direction, mL2 melt pool length in the rear direction, mLm latent heat of melting, kJ/kgM atomic massmd powder deposited per unit traverse length, kg/m_mp powder feed rate, kg/smp/l powder fed per unit traverse length, kg/mPL laser power, kWPp laser power loss in the powder stream, kWRg gas constant J/(kg.mol.K)rl radius of laser beam, mrp Gaussian powder stream radius, mst centre distance between the two successive overlap

track, ms line element on the top boundary of the molten de-

posit, mT temperature, KTamb ambient temperature, KTm melting temperature, KU(z) local fluid flow velocity at distance z from the substrateV molar volume of the metal, m3/molv scan speed, m/sW overall track width, mw melt pool width, mwmax maximum melt pool width, mX dimensionless linear dimension (=x/rl)Y dimensionless linear dimension (=y/rl)Z dimensionless linear dimension (=z/rl)

Greek symbolsαl laser absorptivityγ surface tension on the top boundary of the molten

deposit, N/mГ dimensionless time ¼

ffiffiffiffiffiffiffiffi2Dt

p=rl

� �Γs surface excess at saturation J/(kg mol m2)ε emissivity of the deposit surfaceηc powder catchment efficiencyμL kinematic viscosity, m2/sρ density, kg/m3

ρL density of liquid metal, kg/m3

Ω dimensionless speed (=rlv/D)

19C.P. Paul et al. / Surface & Coatings Technology 224 (2013) 18–28

tendency before solidification for prediction of track geometry. The con-solidated processing parameters for various values of laser energy perunit length of track, powder fed per unit length of track and overlappingindices for this process were experimentally identified. The effect of

powder catchment efficiency at various stand-off distances and overlapindices on waviness factor of the overlapped tracks was also experimen-tally evaluated.

2. Experimental setup

LRM head was specially designed for vertical configuration andaugmented with LRM work station consisting of a 2 kW fibre lasersystem, a 5 axis workstation in a glove box, a computerized numericalcontroller and a twin powder feeder [1]. The schematic arrangementof LRM head is presented in Fig. 1.

The external size of the nozzle is 63 mm × 63 mm × 60 mm and iscapable of depositing material on the inside diameter (ID) of a circulartube having minimum ID of 75 mm. This head has two sub-assemblies:(a) laser processing head and (b) side blown powder feeding tube.While designing the LRM head for vertical configuration, two majorobjectives were considered: first to achieve the compact-design for pro-cessing the components having narrow passage/opening and second toprovide the least perturbation in the powder-gas stream avoiding sud-den or sharp bends. These two objectives were met by providing thepowder feeding from the bottom of the LRM head. This configurationalso facilitated the same laser–powder-interaction zone and associatedphenomena on the vertical surface during the deposition of thematerialby the movement of the LRM head in either direction horizontally. How-ever, the powder catchment efficiency in the present configuration wascompromised as compared to that of conventional in-line/side-blownpowder feeding configuration [2]. The laser processing head has a Quartzlens of 200 mm focal length to focus the laser beam. The focusing laserbeam is reflected normal to the vertical plane by a water cooled Au-coated plane mirror (diameter: 25 mm and 6 mm thick) mounted at45° to the incoming laser beam axis in beam-bender. There is a provisionofmoving beam-bender up and down tomatch the laser beam size to thepowder stream size at the substrate to facilitate the maximum powdercatchment efficiency. The laser beam diameter of 2 mm was used in thepresent study. A port for inert gas is provided at the upper part of laserprocessing head to protect the Au-coatedmirror from ricocheted powderparticles whichmay enter the beam-bender and damage themirror. Thisgas also assists the shielding of molten metal from oxidation. The laserprocessing head has suitablewaterflowarrangement to cool the process-ing head during the deposition of tracks. The side blown powder feedingtube has inside diameter of 2 mm and is mounted at the bottom of thelaser head at an inclination angle of 35° to the laser beam axis. Themate-rial used for the construction of laser head is copper due to lower laser ab-sorption and higher thermal conductivity.

3. Analytical modelling

For LRMon vertical surfaces, the powder is sprayed by a lateral nozzleinto the process zone. A moving laser beamwith known intensity profilemelts the powder particles and a thin layer of the vertical substrate. Asshown in Fig. 1, the laser beam strikes the substrate through powder par-ticles cloud. A fraction of the laser power is absorbed, reflected andscattered by the powder particles and the rest reaches the substrate.Some portion of its power is reflected and the remainder is absorbedforming a molten pool on the substrate. The absorbed power is carriedaway from the melt pool surface into the feed powder and the substrateby thermal conduction and thermo-capillary (Marangoni) flow. Oncethe powder particle reaches the substrate surface, one of the followingprocesses takes place influencing the powder catchment efficiency:

a. Solid particle–solid surface impact leading to ricochet.b. Solid particle–liquid surface leading to catchment.c. Liquid particle–solid surface leading to catchment and quenching

of substrate.d. Liquid particle–liquid surface leading to catchment.

Fig. 1. Schematic arrangement of laser cladding head for processing vertical surfaces.

Fig. 2. Schematic diagram of a moving melt pool in y-direction during LRM on verticalsurface.

20 C.P. Paul et al. / Surface & Coatings Technology 224 (2013) 18–28

During material deposition using moving Gaussian heat source onsemi-infinite substrate accounting the power loss in the powder streamand vaporization heat flux, ignoring convective and radiation heat losses,an analytical model was presented by Ahsan et al. [21,22]. According tothis model, the temperature distribution is given by

T x; y; zð Þ ¼ αlPL−Pp

krlf x; y; z; vð Þ ð1Þ

where the temperature distribution function f(x, y, z, v) is

f x; y; z; vð Þ ¼ ∫∞0

Exp −Hð Þ1þ Γ2� � ffiffiffiffiffiffiffiffi

2π3p dΓ ð2Þ

H ¼Y þ ΩΓ2

2

� �2 þ X2

2 1þ Γ2� � þ Z2

2Γ2: ð3Þ

Here, Pp can be obtained from the following equation [21]

Pp ¼ _mp 1−Exp −2r2lr2p

" # !Cp Tm−Tambð Þ þ Lm� �

: ð4Þ

The melt pool boundary was estimated by calculating the temper-ature profile for T > Tmelt from Eq. (1). Thus obtained moving meltpool is divided into two regions, one representing the front part ofthe pool and another representing the rear part of the pool. As de-scribed by Ahsan and Pinkerton [22], the boundaries of these regionsof pool may be approximated as half ellipses defined by the melt poolwidth (w) and the melt pool length in the forward direction, L1, andthe melt pool length in the rear direction, L2 (refer to Fig. 2). Themelt pool has asymmetry about the x-axis that increases with thescanning speed. The maximum height of track can be deduced by fol-lowing the principle of conservation of energy and mass for powderparticle flow, as described in our previous work. It may be notedthat Marangoni flow is important for the prediction of melt poolboundaries. In the present study, the effect of fluid motion due tothe thermo-capillary (Marangoni) phenomena is incorporated bymodifying the thermal conductivity K for the estimation of track ge-ometry. Since the trackmaterial is just above themelting point duringLRM, other phenomena, like gravity, surface tension, viscous forces,predominate for the shape of the molten metal drop as the moving

laser beam passes away [18]. For better laser energy deploymentand improved powder utilization, the width of powder stream shouldnearly be laser beam diameter. For this, the maximum height of trackmay be written as

hmax ¼ ηc _mp

2ρLrlv: ð5Þ

Assuming that the track height has parabolic cross section alongx–z plane [23] and deriving w(y) from the melt-pool geometry onthe top layer of the substrate, the track height at any point (x, y) isgiven by

h x; yð Þ ¼ hmax 1− yþ L2ð Þ2L1 þ L2ð Þ2

" #1− x2

w yð Þ2" #

ð6Þ

where,

w yð Þ ¼ rl

ffiffiffiffiffiffiffiffiffiffiffiffiffi1− y2

L21

sfor y > 0

¼ rl

ffiffiffiffiffiffiffiffiffiffiffiffiffi1− y2

L22

sfor yb0:

ð7Þ

Once the powder particles reach themolten pool on the vertical sur-face, it gets melted and trapped by the surface tension and viscousforces (both temperature dependent) and counter acting gravity force.The details of temperature dependence of the material properties are


discussed in the later part of this section (in Eqs. 15–17). In the presentsection, a simplified model is presented in 2-D configuration where theshape of the molten metal drop is resulted from the downward flowunder gravity, surface tension, viscous forces as the moving laserbeam passes away. Fig. 3(a) and (b) show the schematics of the moltenmetal profiles under gravity of melt-pool at time t = 0 at a point in x–zplane and at time t = (2rl/v), i.e. at the timewhen laser passed throughthat point. As themoltenmetal is viscous and gets solidified as themov-ing laser beam passes away, the flow of molten metal is in laminar re-gime. This flow of molten metal due to gravity is resisted by theviscosity and the surface tension at the free surface of liquid metal. Fol-lowing Newtonian fluid relationship for viscosity incorporating thegravity and Kim andNa [24], the flowofmoltenmetalmay be describedas

μL∂U zð Þ∂z ¼ ρLgz−

dγdT

∂T∂s ð8Þ

where, ∂ U(z)/∂ z is the velocity gradient perpendicular to the layers offluid flow in the molten pool.

The temperature dependent surface tension gradient for alloys[24–26] is given by

dγdT

¼ γ1 ¼ −Aγ−RgΓs ln 1þ Kaið Þ− Kai1þ Kaið Þ

ΓsΔH0

Tð9Þ

where

K ¼ k1 exp −ΔH0

RgT

!ð10Þ

and

∂T∂s ¼ DTS ¼ −hc T−Tambð Þ

k: ð11Þ

Here, hc is given by [27]

hc ¼ 24:1� 10−4εT1:61: ð12Þ

A value of 0.9 was assumed for ε, as recommended for hot rolledsteel [28].

Fig. 3. Schematic diagram of deposited clad

One can substitute values of γ1 and DTS from Eqs. (9) and (11) intoEq. (8). After integrating it, one obtains

μLU zð Þ ¼ ρLgz2

2−γ1DTSz: ð13Þ

In the above equation, the constant of integration is calculated aszero by using boundary condition z = 0, U(z) = 0.

The local layer fluid displacement (x) will be function of local fluidvelocity U(z) [~xv/rl] of layers at different values of z and laser inter-action time (t = 2rl/v) due to solidification of deposit. The deposit so-lidifies as soon as the laser beam passes away. Simplification gives

x ¼ rlvμL

ρLgz2

2−γ1DTSz

" #: ð14Þ

Eq. (14) can be solved to get the loci of the top boundary of thetrack. In the above, time of solidification of deposit is the same asthat of laser interaction time. It is to be noted here that the locallayer fluid displacement (x) is less at higher scan speeds due tolower laser interaction time and subsequently lower time of local so-lidification. Material properties of AISI 304 stainless steel used in thecalculations are taken from Kim and Na [24] and they are given inTable 1.

The temperature dependence of the density [29] and viscosity [30]of liquid metal is incorporated as follows:

ρL Tð Þ ¼ ρb1þ A1 þ A3ð Þ 1−αð Þ

1−A2

� �ð15Þ

where,

α ¼ T−Tm

Tb−TmA1 ¼ 1:829A2 ¼ 0:03 M Tm=Tbð Þ½ �0:5−2:7� 10−4 M Tm=Tbð Þ½ �2 A1−1ð ÞA3 ¼ ρm

ρb1−A2ð Þ−1−A1:

ð16Þ

Here, subscripts m, b, and s in Eqs. (15) and (16) refer to the melt-ing point, boiling point, and saturation state, respectively. Further,

μL Tð Þ ¼ A:M1=2

V2=3 :T1=2: exp B:

Tm

T

ð17Þ

profile at (a) t = 0 and (b) t = (2rl/v).

Table 1Material properties of AISI 304 stainless steel.

Parameter Unit Value

ai wt.% 0.001Aγ N/(m K) 1.0 × 10−4

ρ kg/m3 7200μL kg/(m s) 0.05ε – 0.9Rg J/(kg mol K) 8314.3Tm K 1723ΔH0 J/(kg.mol) −1.88 × 108

Γs J/(kg mol m2) 1.3 × 10−8

k W/(m K) 580

Table 2Experimental trials and measured dimensions of deposits.

Laser power Scan speed Powder feed rate Height Width Remark

W m/min g/min mm mm –

800 200 6 0.815 2.375 Regular deposit800 400 6 0.325 1.75 Regular deposit800 700 6 0.225 1.61825 Non-uniform

deposit800 1000 6 0.17 1.2525 Non-uniform


deposit1200 400 6 0.5 2.195 Regular deposit1200 700 6 0.33 1.9475 Regular deposit1200 1000 6 0.285 1.96 Non-uniform


deposit1600 400 6 0.64 2.71 Regular deposit1600 700 6 0.42 2.2425 Regular deposit1600 1000 6 0.38 2.915 Regular deposit800 200 10 1.105 2.265 Regular deposit800 400 10 0.56 1.9225 Regular deposit800 700 10 0.285 1.7225 Non-uniform


deposit1200 400 10 0.655 2.5675 Regular deposit1200 700 10 0.335 2.41 Regular deposit1200 1000 10 0.27 2.245 Non-uniform

deposit1600 400 10 0.725 2.96 Regular deposit1600 700 10 0.515 2.71 Regular deposit1600 1000 10 0.425 2.5725 Regular deposit800 200 14 1.44 2.4675 Regular deposit800 400 14 0.81 1.905 Regular deposit800 700 14 0.41 1.7325 Non-uniform




deposit1200 700 14 0.5 2.395 Regular deposit1200 1000 14 0.355 2.2425 Non-uniform

deposit1600 400 14 0.955 3.1475 Regular deposit1600 700 14 0.525 2.5025 Regular deposit1600 1000 14 0.43 2.635 Regular deposit


where,

A ¼ 1:80� 0:38ð Þ � 10−8 J=Kmol1=3� �1=2

B ¼ 2:34� 0:2:

4. Experimental studies

AISI type 304 stainless steel is general purpose austenitic stainlesssteel with a face centred cubic structure. It derives its stainless charac-teristics through the formation of an invisible and adherent Chromiumrich oxide film. It is essentially non-magnetic in the annealed conditionand can only be hardened by cold working. As this material is usedwidely for various industrial applications, it has been selected as thefeed material for present study. Comprehensive experimental studywas carried out on the substrate of the same material (size: 75 mmdiameter and 12 mm thick) to study the effect of different processingparameters on the track geometry. Initially, a number of single trackswere deposited at various combination of laser power, scan speed andpowder feed rate. Table 2 summarizes the experimental trials and sub-sequent deposited track geometries. The cross sections of these trackswere examined using optical microscopy. The images of these trackswere used to measure track dimensions, i.e., width and height. The sin-gle tracks were also deposited at various stand-off distances. The pow-der catchment efficiency was evaluated experimentally by measuringthe weight of the substrate before and after single track depositionusing precision electronic balance having least count of 1 mg. The effectof overlapping parameter on the surface roughness and waviness ofthe overlapped tracks was experimentally determined by measur-ing the surface finish using Taylor–Hobson make surface roughnesstester model: Surtronic 3+ and height gauge with puppy dial and9pt?>sub-millimetre probe respectively. To demonstrate LRMon verticalsurfaces applications, the developed LRM head was used on concave andconvex surfaces of tubular geometry.

5. Results and discussion

5.1. Effect of laser processing parameters on deposited track geometry

Experiments were carried out with fibre laser of beam diameter2.0 mm. As illustrated in Eq. (14), the local fluid displacement in themolten pool due to gravity depends primarily on interaction time[=laser beam diameter (2.rl)/scan speed (v)] and consequently onthe solidification rate. The larger interaction time allows more periodfor fluid to flow downward resulting in shifted deposit and its peak.The fluid properties, i.e., density, viscosity and surface tension of mol-ten metal govern the flow tendency under gravity of the deposit. Asthese properties are the functions of the molten pool temperature,the lower interaction time results in lower bulk mean average tem-perature of molten pool and consequently, it leads to higher averagedensity and extremely high viscosity. Subsequently, it results in little

downward shift of the deposit and its peak. The height (z) of thedeposit is a function of powder feed per unit length for given laserpower and the same stand-off distance. Also, higher height of thetrack results in higher fluid velocity at farthest layer, where z = hmax.Thus, the higher powder feed per unit length leads to higher shift of de-posit and its peak for the same system configuration.

Fig. 4(a)–(d) present the analytical results showing the effect ofgravity at various scan speeds on the track geometry for a set oflaser processing parameters (laser power = 1.6 kW and powderfeed rate = 10 g/min). It is evident that the deposit shifts downwardat lower scan speeds (v = 200 mm/min). Little shift is observed forscan speed of 400 mm/min and it is insignificant for higher scanspeeds (v ≥ 700 mm/min). It is because for higher scan speeds, theinteraction time is also small and hence, the time period for depositto remain in molten state is small. As a result, higher effective viscos-ity counters the flow of molten deposit under gravity. On the con-trary, at lower scan speeds leads to higher interaction time andlonger time period for deposit to remain in molten state. The lower

Fig. 4. Vertical transverse cross section of track as per analytical model with laser power of 800 W and powder feed rate of 10 g/min at various scan speeds: (a) 200 mm/min, (b)400 mm/min, (c) 700 mm/min, and (d) 1000 mm/min.


effective viscosity offer lesser resistance to the flow of molten depositunder gravity. It is necessary to mention here that the contribution dueto temperature dependent surface tension gradient term (−γ1DTSz) isone order less than the gravity dependent term in Eq. (4). Hence, thevertical downward shift of the deposited track and its peak due to thegravity flow of the melt predominantly follows square dependence ofthe track height. Fig. 5(a)–(d) present the experimental results for theabove sets of laser processing parameters. The effect of gravity at vari-ous scan speeds on the track geometry is evident at lower scan speedin both analytical and experimental results.

Fig. 6(a)–(c) present the analytical results showing the effect ofgravity at various powder feed rate for scan speed of 200 mm/min.It is observed that the increase in powder feed rate results in increasein deposit's height [18] and subsequently higher downward shift ofdeposit's peak. There is a little shift of deposit's peak at powder feedrate of 6 g/min and it is increased to a sizable value at 14 g/min.The larger shift at higher powder feed rate for the same set of pro-cessing parameters is primarily due to increase in deposit's height(refer to Eq. (4)). This higher height gets higher localized fluid veloc-ity due to its larger distance from the substrate. This higher localized

fluid velocity results in higher shift of deposit's peak for the same in-teraction time. Fig. 7(a)–(c) presents the experimental results for theabove sets of laser processing parameters. The effect of gravity at var-ious powder feed rates on the track geometry is evident in both ana-lytical and experimental results.

The maximum displacement of deposit's peak (i.e., at z = hmax)for various scan speeds was obtained from Eq. (14) and their compar-ison with experimental data is presented in Fig. 8. It clearly indicatesthat the shift of deposit's peak is the cumulative effect of surface ten-sion, viscous force against gravity, scan speed and deposit's maximumheight. The term exhibiting the effect of surface tension, viscous forceagainst gravity is due to vertical orientation of the surface for LRM. Itmay be noted that the shift of deposit's peak increases with decreasein the scan speed for the same deposit height and it is parabolic forlower scan speeds (v b 200 mm/min). At relatively higher scanspeed (e.g. v ~ 400 mm/min), the shift of deposit's peak tends to bea straight line for various deposit heights, while it is negligible forhigher scan speeds (v ≥ 700 mm/min). The experimental results forthe shift of deposit's peak for various deposit heights are in agree-ment with that of analytical modelling.

Fig. 5. Vertical transverse cross section of track as per experimental model with laser power of 800 W and powder feed rate of 10 g/min at various scan speeds: (a) 200 mm/min,(b) 400 mm/min, (c) 700 mm/min, and (d) 1000 mm/min.


5.2. Processing parameters for material deposition on vertical surface

Itwas observed that aminimum threshold laser power, sufficient in-teraction time and optimal powder flow rate were required for success-ful deposition of material in vertical configuration (refer to Table 2).However, we find the effect of these three parameters can be accountedwith the following two parameters:

Laser Energy per unit traverse length Elð Þ ¼ Laser Power PLð ÞScan Speed vð Þ ð18Þ

Powder fed per unit traverse length mp=l

� �

¼Powder feed rate _mp

� �Scan Speed vð Þ : ð19Þ

The parameters “laser energy per unit traverse length” and “pow-der fed per unit traverse length” governs the laser energy and the ma-terial available for the single track deposition, respectively. Fig. 9presents the consolidated processing parameters for LRM on the ver-tical surface for the experimental range of parameters under investi-gation. At extremely high laser energy per unit traverse length andlower powder fed per unit traverse length, there may be vaporizationof the feed material. As a result, there may be very thin or no trackformation. On the contrary, at extremely low laser energy per unittraverse length and higher powder fed per unit traverse length, thefeed material may not fuse and form a discontinuous irregular track.Hence, there is processing zone, where a balance of both the parame-ters, results in fused continuous track. The threshold value of laser ener-gy and powder fed, both per unit traverse length for the continuousdeposition were found to be ~96 J/mm and ~0.006 g/mm respectively.

Fig. 6. Vertical transverse cross section of track as per analytical model with laser power of 800 W and scan speed of 200 mm/min at various powder feed rates: (a) 6 g/min,(b) 10 g/min, and (c) 14 g/min.


Our experimental studywith other material depicts a similar trend. Theobtained consolidated processing parameters are found to be closer interms of trend and values to the earlier published similar work forlaser rapid manufacturing in horizontal configuration [31,32].

5.3. Powder catchment efficiency

The powder catchment efficiency is one of the important processingparameters for high quality track geometry and economical viability ofthe process. It is the ratio of the powder deposited to the powder fed forthe deposition during the LRM process. Mathematically,

ηc ¼Powder deposited per unit length mdð Þ

Powder fed per unit length mp

� �:

ð20Þ

Fig. 7. Vertical transverse cross section of track as per experimental model with laser powe(b) 10 g/min, and (c) 14 g/min.

Maximum powder efficiency is obtained when there is nearly fulloverlap of powder stream diameter and laser beam diameter at thesubstrate. In side-blown powder delivery configuration, it is not pos-sible to achieve the cent–percent overlap of the diameters due to geo-metrical constraints [33]. The maximum powder catchment efficiencyderived considering the maximum overlap region of powder streamand laser beam on the substrate for our present configuration is65%. Since various nozzle stand-off distances between the nozzle tipand substrate resulted in different powder and laser beam diameters,the catchment efficiency varies accordingly. Fig. 10 shows the variationof Powder catchment efficiency at different nozzle stand-off distancebetween nozzle tip and substrate during LRM on vertical surfaces. It isobserved that catchment efficiency increases as the stand-off distanceis increased up to certain value and then it falls. The maximum

r of 800 W and scan speed of 200 mm/min at various powder feed rates: (a) 6 g/min,

Fig. 8. Comparison of theoretically calculated and experimentally observed shift ofdeposit's peak at various scan speed for different deposit heights.

Fig. 10. Powder catchment efficiency as a function of stand-off distance between noz-zle tip and substrate during LRM on vertical surfaces.


catchment efficiency is obtained where there is maximum overlappingregion of powder stream and laser beam. For our configuration, themaximum catchment efficiency is obtained at the stand-off distancebetween 15 mm to 18 mm.

5.4. Surface finish of deposited overlapped tracks

The surface characteristics of the track geometry on vertical sub-strate play an important role as it defines the manufacturing toleranceand post processing of the deposits. Generally, multi-layer overlappedtracks are used for the fabrication of any engineering components onthe surfaces. Higher waviness and poor surface roughness means lesseffective deposition to achieve on the plane surface. The LRM surfaceshave two distinct characteristics— surface roughness and surface wav-iness. The surface roughness is primarily due to semi-molten powderparticles adhered on the deposit, while overlapping parameter governsthe surface waviness to a great extent.

Fig. 9. Processing parameters fo

As shown in Fig. 11, “Overlap index” is defined as the ratio of thecentre distance between the two successive overlap tracks to the singletrackwidth, while “waviness factor” is defined as the ratio of the heightto be removed to get flat surface to total height deposited. Mathemati-cally,

overlap index ið Þ ¼ stW

� 100 ð21Þ

waviness factor jð Þ ¼ hmax−hmin

hmax: ð22Þ

Anumber of overlapped track samplesweremade at various overlapindex. Fig. 12 presents the variation inwaviness factor at different over-lap index. It can be seen that the waviness factor decreases as the over-lap index is increased. For overlap index greater than 70%, layer heightgoes on increasing in subsequent overlapping track and it leads tointer-run porosity due to a lack of fusion zone along the track. The

r LRM on vertical surfaces.

Fig. 11. Schematic diagram for various parameters for overlapped deposition.

Fig. 13. Laser vertical surface cladding on (a) internal and (b) external surface of tubulargeometry.


surface finish was also measured in the samples across and along thetrack. The surface finish was found to be 12–14 μm Ra and 9–12 μm Raacross and along the track, respectively. A similar trend is observed forsurface finish along and across the tracks for LRMon horizontal surfaces[33–35]. The surface finish during LRM on horizontal surfaces with the

Fig. 12. Variation in waviness factor at different overlap index.

same process parameters and setup configuration was experimentallyfound to be in the range of 12–25 μm Ra. Thus, the surface finish inLRM on vertical surfaces was found to be better than that of horizontalsurfaces for the same processing parameters. It is because the semimol-ten particles sprayed towards molten pool fall down due to gravity andit leads to reduced tendency of the semi-molten powder particles ad-hering to the surface.

6. Conclusions

A laser processing head for rapid manufacturing on vertical surfaceswas developed and augmented with work-station. It was used success-fully for the rapidmanufacturing on vertical surfaces on various tubulargeometries. Fig. 13(a) and (b) present the LRM carried out on the exter-nal and internal surface of tubular geometry in vertical configuration atauthors' laboratory. The processing zone for LRM on vertical surfacewith the threshold value of laser energy and powder fed, both perunit length of the deposit for the continuous tracks were found to be~96 J/mm and ~0.006 g/mm respectively. The trend and values werefound to be closer to earlier work published for laser rapidmanufactur-ing in horizontal configuration. Analytical and experimental analysis ofthe track cross-section at various processing parameters indicate thatthere is a downward shift of deposit and its peak, due to gravity duringlaser rapidmanufacturing onvertical surfaces. Therewas little downwardshift of deposit's peak for higher scan speeds (e.g. v > 400 mm/min),while it was distinctly observed with asymmetric downward bulgingof deposit at lower scan speeds (v b 200 mm/min). The maximumpowder catchment efficiency was experimentally found to be 38% for


stand-off distance between 15 mm to 18 mm with the present laserhead and configuration. The surface waviness decreases as the overlapindex increases. The surface finish during LRM in vertical configurationwas better than that of horizontal configuration for the same setup andprocessing parameters.

Acknowledgement

The authors acknowledge the technical support of Mr. P Bhargava,Mr. C H Premsingh and Mr. Deepjwalit Viashnav during the study.

References

[1] C.P. Paul, P. Bhargava, Atul Kumar, A.K. Pathak, L.M. Kukreja, in: J. Paulo Davim(Ed.), Lasers in Manufacturing, Wiley-ISTE, U.K., 2012

[2] L.M. Kukreja, R. Kaul, C.P. Paul, P. Ganesh, B.T. Rao, in: J. Dutta Majumdar, I. Manna(Eds.), Laser Assisted Fabrication of Materials, Springer Series in Materials Science,vol. 161, Springer-Verlag, 2013.

[3] C.P. Paul, H. Alemohammad, E. Toyserkani, A. Khajepour, S. Corbin, Mater. Sci.Eng. A 464 (2007) 170.

[4] J.C. Ion, Laser Processing of Engineering Materials: Principles, Procedure andIndustrial Application, Elsevier/Butterworth-Heinemann, Amsterdam; Oxford, 2005.556.

[5] J. del Val, R. Comesaña, F. Lusquiños, M. Boutinguiza, A. Riveiro, F. Quintero, J. Pou,Surf. Coat. Technol. 204 (2010) 1957.

[6] S.H. Mok, G. Bi, J. Folkes, I. Pashby, Surf. Coat. Technol. 202 (2008) 3933.[7] R. Jendrzejewski, G. Sliwinski, M. Krawczuk, W. Ostachowicz, J. Comput. Struct. 82

(2004) 653.[8] S. Kumar, S. Roy, C.P. Paul, A.K. Nath, Numer. Heat Transf. B 52 (2008) 271.[9] P. Peyre, P. Aubry, R. Fabbro, R. Neveu, A. Longuet, J. Phys. D: Appl. Phys. 41

(2008) 025403.[10] N. Ahmed, K.T. Voisey, D.G. McCartney, Opt. Lasers Eng. 48 (2010) 548.[11] J. Liu, L. Li, J. Manuf. Process. 8 (2006) 1.

[12] H. Qi, J. Mazumder, H. Ki, J. Appl. Phys. 100 (2006) 024903.[13] X. He, G. Yu, J. Mazumder, J. Phys. D: Appl. Phys. 43 (2010) 015502.[14] C.P. Paul, S.K. Mishra, C.H. Premsingh, P. Bhargava, P. Tiwari, L.M. Kukreja, Int.

J. Adv. Manuf. Technol. (2012) 757.[15] M. Alimardani, E. Toyserkani, J.P. Huissoon, C.P. Paul, Opt. Lasers Eng. 47 (2009)

1160.[16] J.T. Hofman, D. Fde Lange, B. Pathiraj, J. Meijer, J. Mater. Process. Technol. 211

(2011) 187.[17] Z. Wang, S. Dong, B. Xu, W. Xia, Adv. Mater. Res. 148–149 (2011) 628.[18] Atul Kumar, C.P. Paul, A.K. Pathak, P. Bhargava, L.M. Kukreja, Opt. Laser Technol.

44 (2012) 555.[19] P.T. Colby, D. Hotchkiss, J.W. Sears (2012) Vertical laser cladding system, US Pat-

ent US2012/0100313 A1.[20] J. He, M. Zhong, F.S. Li, W. Liu, Y. Zhang, in: S.S. Deng, A. Matsunawa, Y.L. Yao, M.

Zhong (Eds.), Lasers in Material Processing and Manufacturing II, SPIE Proceed-ings, vol. 5629, 2005, p. 418.

[21] M.N. Ahsan, C.P. Paul, L.M. Kukreja, A.J. Pinkerton, J. Mater. Process. Technol. 211(2011) 602.

[22] M.N. Ahsan, A.J. Pinkerton, Modell. Simul. Mater. Sci. Eng. 19 (2011) 055003.[23] A. Fathi, E. Toyserkani, A. Khajepour, M. Durali, J. Phys. D: Appl. Phys. 39 (2006)

2613.[24] W.H. Kim, S.J. Na, Int. J. Heat Mass Transf. 41 (1998) 3213.[25] P. Sahoo, T. Debroy, M.T. McNallan, Metall. Trans. B 19B (1988) 483.[26] T. Zacharia, S.A. David, J.M. Vitek, T. Debroy, Weld. J. Res. Suppl. (1989) 499.[27] J. Glodak, M. Bibby, J. Moore, B. Patel, Metall. Trans. 17B (1986) 587.[28] Y.S. Touloukian, Thermophysical Properties of High Temperature Solid Materials,

vol. 3, The Macmillan Co, 1967.[29] N. Mehdipour, A. Boushehri, H. Eslami, J. Non-Cryst. Solids 351 (2005) 1333.[30] G. Kaptay, Z. Metallkd. 96 (2005) 24.[31] U. de Oliveira, V. Ocelik, J.Th.M. De Hosson, Surf. Coat. Technol. 197 (2005) 127.[32] E.W. Kreutz, G. Backes, A. Gasser, K. Wissenbach, Appl. Surf. Sci. 86 (1995) 310.[33] C.P. Paul, Parametric Studies of Laser Metal Deposition for Fabrication Engineer-

ing Components (2005). PhD Thesis.[34] J. Zeng, Y.-H. Kim, Y. Chen, Adv. Mater. Res. 399–401 (2012) 1802.[35] Y. Li, J. Ma, Surf. Coat. Technol. 90 (1997) l.

laser rapid mfg on vertical surfaces

Documents

laser rapidmanufacturing

track width

track geometries height

tnotherny advahigh power

horizontal conguration

process domain

deposit materials

lrmin vertical conguration