Porous Structures Fabrication by Continuous and Pulsed Laser Metal Deposition

7/31/2019 Porous Structures Fabrication by Continuous and Pulsed Laser Metal Deposition

http://slidepdf.com/reader/full/porous-structures-fabrication-by-continuous-and-pulsed-laser-metal-deposition 1/8

Journal of Materials Processing Technology 211 (2011) 602–609

Contents lists available at ScienceDirect

Journal of Materials Processing Technology

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / j m a t p r o t e c

Porous structures fabrication by continuous and pulsed laser metal deposition

for biomedical applications; modelling and experimental investigation

M. Naveed Ahsan a,∗, Christ P. Paul b, L.M. Kukreja b, Andrew J. Pinkerton a

a Laser Processing Research Centre, School of Mechanical Aerospace and Civil Engineering, The University of Manchester, Sackville Street, Manchester M13 9PL, United Kingdomb Laser Material Processing Division, Raja Ramanna Centre for Advanced Technology, Indore 452013 (M.P.), India

a r t i c l e i n f o

Article history:

Received 13 September 2010

Received in revised form

15 November 2010

Accepted 21 November 2010

Available online 26 November 2010

Keywords:

Osseointegration

Laser metal deposition

Surface porous structures

Biomedical implants

Analytical model

a b s t r a c t

The use of porous surface structures is gaining popularity in biomedical implant manufacture due to its

ability to promote increasedosseointegrationand cellproliferation.Laser direct metal deposition(LDMD)

is a rapid manufacturing techniquecapableof producingsucha structure. Inthiswork LDMD with a diode

laser in continuous mode and with a CO2 laser in pulsed modes are used to produce multi-layer porous

structures. Gas-atomized Ti–6Al–4V and 316L stainless steel powders are used as the deposition mate-

rial. The porous structures are compared with respect to their internal geometry, pore size, and part

density using a range of techniques including micro-tomography. Results show that the two methods

produce radically different internal structures, but in both cases a range of part densities can be pro-

duced by varying process parameters such as laser power and powder mass flow rate. Prudent selection

of these parameters allows the interconnected pores that are considered most suitable for promoting

osseointegration to be obtained.Analytical modelsof theprocessesare also developedby using Wolfram

Mathematica software to solveinteracting, transient heat,temperature and massflow models. Measured

and modelled results are compared and show good agreement.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

There is growing demand for biocompatible implants in

the world, therefore the manufacturing technology required to

produce them should be developed at the same pace. The esti-

mated market for dental and orthopaedic implants is already

over 4.8billion dollars and it is expected to grow briskly in

the future. Thus, there are both humanitarian and economic

incentives to optimize implant design and manufacture. The

main requirements for implant-materials are bio-functionality

(adequate mechanical properties), bio-corrosion resistance, bio-

compatibility, bio-adhesion (bone growth), process-ability and

availability as referred by Black and Hastings (1998). The pos-

sible material options include – metals, polymers and ceramics.

Compared to other biomaterials, metals are widely used for many

biomedical applications, including – hip and knee endoprosthe-

ses, plates, screws, pins and dental implants, due to outstanding

tensile strength. The major problems concerning with metallic

implants in orthopaedic surgery is the mismatch of the mechanical

properties between bone and bulk metallic materials. Due to this

mechanical mismatch, bone is insufficiently loaded and becomes

∗ Corresponding author. Tel.: +44 161 3062365; fax: +44 161 3064646.

E-mail address: [email protected]

(M.N. Ahsan).

stress-shielded, which eventually leads to bone resorption. This

impairs the clinical performance of the prosthesis and is respon-

sible for implant migration, aseptic loosening and fractures around

the prosthesis as investigated by Kroger et al. (1998). The rela-

tionship between implant flexibility and the extent of the bone

loss has been established; it is confirmed that the changes in bone

morphology are an effect of the stress shielding and subsequent

adaptive remodelling process as observed by Sychterz et al.(2001).

Therefore, there shouldbe a suitable balance between strength and

stiffness of the implants to best match the behaviour of bone for

prolonged trouble free performance.

One consideration to achieve a strong interfacial bond and

also to reduce the modulus mismatch has been the develop-

ment of implants with engineered porous materials as investigated

by Kirshna et al. (2008). Use of engineered porous materials

effectively reduces the modulus mismatch and also provides path-

ways for bone in-growth through the pores for stable long-term

anchorage or biological fixation of the implant as explained by

Pilliar (1987). Conventional powder metallurgical processing has

been used to fabricate surface-treated or fully porous implants

for biomedical applications as shown by Ik-Hyun et al. (2003).

However, conventionally sintered metals are very brittle and

pore size, shape, volume fraction, and distribution are difficult

to control. These factors all have a major influence on mechan-

ical and biological properties. Other fabrication techniques that

use foaming agents, as elaborated by Wen et al. (2001), suffer

0924-0136/$ – see front matter © 2010 Elsevier B.V. All rights reserved.

doi:10.1016/j.jmatprotec.2010.11.014

http://dx.doi.org/10.1016/j.jmatprotec.2010.11.014

http://www.sciencedirect.com/science/journal/09240136

http://www.elsevier.com/locate/jmatprotec

mailto:[email protected]



mailto:[email protected]

http://www.elsevier.com/locate/jmatprotec

http://www.sciencedirect.com/science/journal/09240136




M.N. Ahsan et al. / Journal of Materials Processing Technology 211 (2011) 602–609 603

from limitations such as contamination, impurity and limited part

geometry.

Pore size is also very important in the porous structures to be

used for biomedical applications. Ingrowth of bone tissues into

the pores is critical for proper osseointegration. Karageorgiou and

Kaplan (2005)f oundthata100mporewastheminimumrequired

for the migration requirements while Götz et al. (2004) showed a

200m is an optimal pore size for osseointegration in laser tex-

turedTi–6Al–4V implants. Xueet al.(2007) showed thata pore size

greater than 200m was required for cell ingrowth into the pores

and for pore size less than 150m the cells span directly across the

pores in spite of going into the pores. So, proper osseointegration

and cell proliferation depends upona reasonablepore sizeselection

in biomedical implants.

Laser direct metal deposition (LDMD) is one of the few tech-

niques that has the potential to produce a structure with graded

porosity and/or composition from different biomaterials including

titanium, Ti–6Al–4V, stainless steel and shape memory alloys. It

can produce the implant with greater degree of geometrical flex-

ibility than other methods, however predicting the outcome from

a particular combination of LDMD input parameters can be diffi-

cult because of the complex nature of the process and adoption of a

‘trial-and-error’ method is both slow and expensive when medical

grade alloys are being used. There is thus a need for both a reliablemethodto produce structures with controlled porosityand a wayof

predicting their nature in advance through modelling. In this paper

LDMD techniques have been applied using a laser in both continu-

ous and pulsed mode to produce multilayer porous structures for

biomedical applications. Analytical models for both processes have

also been developed.

2. Porous structures fabrication using continuous-wave

LDMD

2.1. Experimental investigation

Forthe continuouswave laserdeposited porous structureexper-

iment, A 1.5kW Laserline LDL-160 High Power Diode Laser, withoutput wavelengths of 808 nm and 940 nm, was used as the power

source and Ti–6Al–4V was used as the substrate and deposi-

tion material. The Ti–6Al–4V powder was of size 50–120 m and

was delivered to the coaxial deposition head from a dual hopper

SIMATIC OP3 disc powder feeder. Argon, with a flow rate of 5 l per

minute, was used as the conveyance gas. Another flow of Argon

gas, with a rate of 6 l per minute, through the central passage of

the coaxial deposition head was used to shield the objective lens

and to prevent oxidation in the deposition process. The substrate

blocks were ofa nominal size 50mm × 50mm × 10mm. The blocks

Table 1

Processing parameters for deposition of porous structures.

Parameters Level 1 Level 2 Level 3

Powder flow rate (g s−1) 0.033 0.066 0.089

Track overlap ratio (%) −10 0 10

were sand blasted for laserabsorption enhancementand degreased

using ethanol prior to the experiment. Three levels of each powderflow rate and track overlap ratio at a constant laser power of 1kW

and scanning speed of 6 mm s−1 were tested, as shown in Table 1,

giving a total nine experimentalruns. Track overlap ratio is defined

as the percentage of the track width that overlaps a previous track.

Mathematically it is equal to (1 − xcentre/W ) where xcentre is the

track spacing and W is the track width.

All porous samples were initially detached approximately 1 mm

above the substrate and then cut to form a cube of approx size

12mm × 12mm × 3 mm. All surfaces were ground to 1200 grit and

then examined under a Polyvar Met optical microscope to find the

pore sizes. Due to variation in pore sizes in different planes, all

measurements were taken on y– z plane mentioned in Fig. 3. These

pores were not of circular shape so the diameter of a circle of equal

area was taken as a representative dimension of each pore. There

was some variation in pore sizes within each sample so four ran-

dom pores per sample were selected and an average value used.

The micro computed tomography (MicroCT) technique was also

employed to examine the internal morphologies of a selection of

the samples. The Metris X-Tek 225 kV MicroCT machine was used

to conduct the analysis.

The produced structures were well-bonded with the substrate

and had little visible oxidation. Fig. 1(a) shows good bonding at

clad-substrate interface while the shiny surface in Fig. 1(b) depicts

the presence of oxide.

Fig. 2 shows a 3D MicroCT image of the porous sample show-

ing the regular location of the pores. Fig. 3 show three orthogonal

images (a), (b) and (c) at y– z plane, x– y plane and x– z plane, respec-

tively. The MicroCT post-processing software can take images at

discrete slice positions through the sample in each orthogonalplane. The pictures shown are taken at the mid of the pores. The

pores are interconnected throughout the thickness of sample and

partially melted powder particles can also be seen.

The pore size is mainly dependent on overlap ratio while pore

shape changes with the powder flow rate. It is found that pores are

nearly circular at lower powder flow rate and become elongated

at higher powder flow rate. Fig. 4 shows that pore size decreases

linearly with the increases in overlap ratio at all powder flow rates.

There is a slight decrease in pore size with the increases in powder

flow rate at all overlap ratios.

Fig. 1. Continuous-wave deposition sample at 1000 W, 6 mm s−1, 0.089gs−1 and 10% overlap: (a) optical micrograph showing clad-substrate interface region; (b) clad area

showing presence of oxidation.










604 M.N. Ahsan et al. / Journal of Materials Processing Technology 211 (2011) 602–609

Fig. 2. 3D MicroCT image of the porous structure at laser power 1000 W, scanning

speed 6mm s−1, powder flow rate 0.033 g s−1 and overlap ratio 0%.

Fig.3. OrthogonalMicroCTimages ofthe sampleshownin Fig.2: (a)right, y– z plane;

(b) top, x– y plane; (c) front, x– z plane.

2.2. Analytical model

An analytical model for the interaction of multiple deposition

tracks and layers and consequential pore formation has been for-

mulated. The model assumes the initial substrate is semi-infinite

and the governing equations remain valid over the entire dimen-

sions of thesubstrate. Although thesubstrate block is actually finite

in size this is reasonable at the heat input was localised, towards

the centre and the block was clamped to the table, which acted as a

further heat sink. Convection and radiation heat losses are ignored

Fig. 4. Effect of overlap ratio and powder flow rate on pore size at laser power

1000W, scanning speed 6 mms−1

.

as justified by the previous research conducted by Pinkerton and Li

(2004a). However, the model accounts for the power losses in the

powder stream and due to the vaporization heat flux.

Cline and Anthony (1977) equation of moving Gaussian source

of heat is used tocalculate thetemperature profiles on thesubstrate

and hence melt pool size

T ( x, y, z) = ˛P l − P pkr l

f ( x, y, z, v ) (1)

where the temperature distribution function f is

f ( x, y, z, v ) = ∞

0

Exp(−H)

(23)1/2

(1 + 2)d,

H = (Y + (/2)2)2 + X 2

2(1 + 2)+ Z 2

22(2)

here k is thermal conductivity; D is thermal diffusivity of the mate-

rial; v is scanning speed and T amb is ambient temperature. r l is laser

beam radius · P l is laser power and ˛ is laser absorptivity. is

dimensionless time, is dimensionless speed and X , Y and Z are

the linear dimensions made dimensionless by dividing the laser

beam radius, i.e. = √ 2Dt/r l , = r lv /D, X = x/r l, Y = y/r l and Z = z/r l.Power loss in the powder stream, P p, can be found from the

following equation

P p = q pr

1 − Exp

−

2r 2l

r 2 p

[C (T m − T amb) + Lm] (3)

where C , T m, Lm, q pr and r p stand for specific heat capacity, melting

temperature, latent heat of melting, powder mass flow rate and

Gaussian powder stream radius, respectively.

Eq. (1) is solved using Wolfram Mathematica and the melt pool

divided into two regions, one representing the pool in front of the

heat source and one the pool behind the heat source. The bound-

aries of these regions are approximated as half ellipses defined by

the melt pool width, W , and the melt pool length in the forwarddirection, L1, and the melt pool length in the rear direction, L2, as

shown in Fig. 5. The melt pool has asymmetry about the x-axis that

increases with the scanning speed.

Forcalculations of evaporation heat fluxlossesthe overall evap-

oration model of Choi et al. (1987) has been used. The laser power

loss due to evaporation heat flux, P evap, from the melt pool can be

Fig. 5. Schematic diagram of a moving melt pool in y-direction with powder mass

addition.




calculated by the following equation.

P evap =

4W (L1 + L2)

455.32Exp[4.834 − (18836/T )]√

T

hv (4)

where hv is latent heat of vaporization.

During the deposition process, the melt pool moves in the y-

direction with uniform speed. The powder from the coaxial nozzle

strikes the melt pool and forms the clad. Considering Fig. 5, the

equations of ellipses for the regions in front and behind of the heatsource can be written as:

For y > 0 :y2

L21

+ 4 x2

W 2= 1 (5)

For y < 0 :y2

L22

+ 4 x2

W 2= 1 (6)

If y front and yrear are the front and rear limits of the melt pool at

any x position | x| ≤ W /2, as illustrated in Fig. 5, the melt pool can

assimilate powder that falls within these limits. From Eqs. (5) and

(6), we can write

y front = L1

1 − 4 x2

W 2(7)

yrear = −L2

1 − 4 x2

W 2(8)

The powder distribution in coaxial laser direct metal deposition

has been shown to have a Gaussian distribution as explained by

Pinkerton and Li (2004b). The coaxial powder stream with powder

mass flux q pf at any point ( x, y), with the laser beam centred at

origin, can be written as:

q pf = 2q pr

r 2 pExp

−2

x2 + y2

r 2 p

(9)

The deposition height, h z, canthusbe found byintegratingthe pow-

der mass flux, i.e. Eq. (9) over the distance between the front and

rear limits of the melt pool.

h z = 1

v

y front

yrear

q pf dy (10)

Eq. (10) applies to the entire range of melt pool, −W /2 ≤ x ≤ W /2.

During deposition of a parallel, overlapping track, the melt pool

will become partially inclined due to the difference in height lev-

els of the substrate and the first track on which it is overlapped.

This inclination effect will reduce the melt pool area projected

normal to the surface in the overlapped region compared to the

non-overlapped region. The melt pool for the overlapped track is

modelled to incorporate this effect. Consider Fig. 6(a), in which the

melt pool is moving in the y-direction with a uniform speed and

partially overlapping the previous track.

The track spacing is given by the variable xcentre, which rep-resents the centre to centre distance between the consecutive

overlapped tracks. The tracks thus overlap and interact if xcen-

tre < W . The area of the melt pool when over the previous track

is taken by assuming the boundaries of the melt pool are straight

lines rather than arcs of ellipses.

Using this method, a simplified diagram of the geometry of the

overlapped region is shown in Fig. 6(b), which will be used to cal-

culate the modified melt pool limits. The y front 1 and yrear 1 are the

modified meltpool limits in forward and rear directionrespectively

and can be written as:

y front 1 = L1

1 − (4( xcentre − (W/2))

2/W 2)

W − xcentre

x − xcentre − W

2

(11)

Fig. 6. (a) Schematic diagram of a moving melt pool overlapping a previous track

surface; (b) simplified diagram of the overlapped melt pool region used for calcula-

tion of the modified melt pool limits.

yrear 1 = −L2 1 −

(4( xcentre

−(W/2))

2/w2)

W − xcentre

x − xcentre −w

2(12)

The deposition height, ha, in the overlapped region can be calcu-

lated using the modified melt pool limits in Eq. (10)

ha = 1

v

y front 1

yrear 1

q pf dy (13)

In multiple track deposition, another important factor is how the

change in surface geometry affects the powder affinity. The track

corner has a highly concave surface and has high affinity for pow-

der as compared to other portions of the track. This may be related

to powder having less tendency to ricochet from the surface at this

position and the area tending to attract stray powder. The coaxial

nozzle produces a Gaussian distribution of powder, giving a max-

imum probability of powder under the central axis of the nozzle.

The effect of this high powder affinity is therefore particularly pro-

nounced at overlap ratios of about 50% ( xcentre = W /2) because the

nozzle central axis is near to the limit of the previous track. As the

overlap ratio increases the nozzle moves away from high powder

affinity zone reducing the effect. The powder affinity factor, ε, can

be written in the form:

ε = 1 + (1 − ϕ)

5Exp

−70000000

W

2− x

2

(14)

where ϕ, is the overlap ratio defined as (1 − xcentre/W ).

After incorporating the effect of the powder affinity factor, the

modified deposition height, hap, for the overlapped region, and hbp,




Fig. 7. Modelled multilayer porous structures of Ti–6Al–4V at 1000 W laser power, 6mm s−1scanning speed and 0.089 g s−1powder flow rate with: (a) −10% overlap ratio;

(b) 0% overlap; (c) 10% overlap ratio.

for non-overlapped region, can be written as:

hap =ε

v

y front 1

yrear 1q pf dy

(15)

hbp = ε

v

y front

yrear

q pf dy (16)

The final deposition height for both two regions, i.e. the total track,

h z, becomes:

h z = hap + hbp (17)

Eq. (17) definesthe deposition profile for parallel overlapping track.

A routine has been written in Wolfram Mathematica software to

realize the model andpost-process anddisplay the results. Numer-

ical integration, polynomial least-squares fit and logic selection

techniques are used. Table 2 shows the Ti–6Al–4V parameters used

for the modelling calculations.Fig. 7 shows a modelled multilayer porous structure. The results

can be interpreted from the figure; the model predicts changes in

porosity, pore size and pore shape with overlap ratio and change

Table 2

Ti–6Al–4V parameters used in the modelling calculations.

Parameter Unit Value

Laser absorptivity, ˛ – 0.5

Melting temperature, T m K 1933

Thermal conductivity, k W m−1 K−1 17

Specific heat, C J kg−1 K−1 800

Density, kg m−3 4300

Latent heat of melting, Lm kJkg−1 370

Latent heat of vaporization, hv kJkg−1 9830

in deposition parameters such as laser power, traverse speed and

powder mass flow rate.

2.3. Model verification

The model was verified by comparing the modelled pore sizes

with the experimental values, as shown in Fig. 8. The model shows

good agreementwith experimentalvalues of pore sizes in a overlap

ratio operating range −10 to 10%. The slight variations in experi-

mentaland modelleddatafor pore sizes canbe attributed tothe fact

thatduring multiplelayer depositionthe substratesurfacebecomes

rough due to the deposition of the previous layer. This can slightly

change the clad formation characteristics during the subsequent

clad.

Fig. 8. Experimental and model results comparison; effect of overlap ratio on pore

size (1000W, 6 mms−1

, 0.089gs−1

).




Table 3

Laser pulse duration and average ball dimensions.

Pulse duration (ms) Average ball

diameter (mm)

Average ball

height (mm)

50 1.25 0.56

150 1.49 0.72

250 1.75 0.87

350 1.89 0.94

450 2.12 1.02

3. Porous structures fabrication using pulsed-wave LDMD

3.1. Experimental investigation

For the pulsed wave deposition experiment a 3.5 kW CO2 laser

system, operating in pulse mode, integrated with beam delivery

system consisting of water-cooled gold coated plane mirrors (PM)

and a concave mirror (CM) of 600 mm radius of curvature, a co-

axial powder feed system and a 5-axis laser workstation, was used.

At the fabrication point, a defocused laser spot of 1.5mm diam-

eter was used to create a melt pool. Stainless steel 316L powder

with size range 45–106m was fed into the pool coaxially using a

volumetriccontrolled powder feeder. Argonwas used as the shield-

ing and powder carrier gas. A series of single layer test sampleswere first produced at a scan speed of 20mms−1 and powder mass

feed rate of 0.133g s−1. The laser power and pulse period were

1.6 kW and 500 ms, respectively. Each laser pulse was found to

produce a ‘ball’ like deposit on the plane substrate that was cir-

cular in horizontal section and elliptical in vertical section. Laser

pulses of various durations were found to give balls of different

sizes; this was primarily attributed to the gravity and surface ten-

sion effects. Table 3 presents thesize of balls created using different

pulse lengths.

Porous structures were with overlap ratios of 14%, 23% and 32%

were manufactured by depositing multiple layers of overlapping

balls. Similarly to the definition for continuous tracks in Section

2.1, overlap ratio is defined as the percentage of a ball overlapping

a previous one divided by the ball diameter and can be appliedparallel to the motion of the laser or perpendicular to it, across

tracks. Mathematically it is equal to (1 − xcentre/D) where xcen-

tre is the ball spacing and D is the ball diameter. Fig. 9, shows

a typical porous structure fabricated by pulsed laser deposition

and Fig. 10, shows the results of the density measurement for

the determination of the porosity using Archimedes principle. A

structure with porosity of more than 25% is created with an over-

lap index 14% the porosity then decreases with increasing overlap

ratio.

3.2. Analytical model

For modelling of the pulsed-wave laser deposition, the laser is

considered as a stationary source of heat during the length of apulse. Cline and Anthony (1977) equation for a moving Gaussian

source of heat is used with a scanning speed term equal to zero. Eq.

(18) is used to calculate the temperature profiles on the substrate

and hence melt pool size

T ( x, y, z) = ˛P lkr l

f ( x, y, z) (18)

where the temperature distribution function f is

f ( x, y, z) = ∞

0

Exp(−H)

(23)1/2

(1 + 2)d, H = Y 2 + X 2

2(1 + 2)+ Z 2

22

(19)

If a laser pulse begins at t = 0 and continues for a pulse duration t p,

then initially there will be heating but no melting, and then a melt

Fig. 9. A typical porous structure of 316L stainless steel fabricated by LDMD at

1.6kW laser power, 250ms pulse duration and 0.133g s−1powder flow rate: (a)

sample; (b) MicroCT image.

pool of radius R will be generated. R will then continue to increase

over the length of the pulse. If t R is the time after the pulse started

then R = f (t R). For the purposes of the model this is inverted and by

curve fitting t R expressed as a fifth order polynomial function of R.

For example if a 1.6 kW laser power and 316L stainless steel are

used as substrate material, then the polynomial becomes:

t R = −0.107035 + 416.338R − 511652R2 + 1.28132 × 103R3

+ 1.36412 × 1011R4 − 4.30024 × 1013R5 (20)

Fig. 10. Effect of overlap index on porosity of laser deposited structures (1.6kW

laser power, 250ms pulse duration, 0.133 g s−1

powder flow rate).




Table 4

316L parameters used in the modelling calculations.

Parameters Unit Value

Laser absorptivity, ˛ – 0.25

Melting temperature, T m K 1673

Thermal conductivity, k W m−1 K−1 15

Specific heat, C J kg−1 K−1 500

Density, kg m−3 8000

Thecoaxialpowder streamwithpowder mass fluxq pf with thelaser

beam centred at origin can be written as:

q pf = 2q pr

r 2 pExp

−2

R2

r 2 p

(21)

where q pr and r p stand for powder mass flow rate and Gaussian

powder stream radius, respectively. The deposition height can be

found by integrating Eq. (21) over the effective pulse duration.

h z = 1

t p

t R

q pf dt (22)

Eq. (22) defines the deposition profile in a pulse (0 < t < t p).In practice,surface tension and gravity forcesbecome dominant,

changing the shape of the profile. The shape modified in this way is

assumed to have an elliptical profile, with dimensions determined

by theextent ofthe meltedare (giving thewidth) andthe totalmass

of material in the ball.

The total deposition cross sectional area Ad, can be calculated as

Ad = R

0

h z dR (23)

The maximum ball height Hb, assuming its shape as elliptical, is

given by the following equation

Hb =4

× Ad

× R (24)

and the ball deposition profile h z, can be calculated as

hb = Hb

1 − x2

R2(25)

Eq. (25) defines the balldeposition profile, incorporating the effects

of surface tension and gravity forces.

A routine waswritten in Wolfram Mathematica software to real-

ize the model and post-process and display the results. Numerical

integration, polynomial least-squares fit and logic selection tech-

niques are used. Table 4 presents the 316L parameters used for

the modelling calculations. Fig. 11 shows that model ability to pre-

dict the deposition ball shape and size at a given set of processing

parameters.

3.3. Model verification

Experimental and model results of deposited balls diameter and

height are shown in Fig. 11. There is a good agreement between

measured andmodelled ball diameter values at allpulse durations.

The model also predicts the deposited ball heights quite well but

slightly underestimates it at lower pulse durations and overesti-

mates at higher pulse durations. This may be due to the fact that at

higher pulse durations the surface tension forces become less sig-

nificant causing the ball shape to becomemore elliptical and hence

decreasing its height. Overall the model follows the experimental

trends.

Fig. 11. Experimental and model results comparison: (a) change in ball diameter

with pulse duration; (b) change in ball height with pulse duration.

4. Conclusion

The present work shows the flexibility of laser direct metal

deposition both in continuous and pulsed mode to produce dif-

ferent levels of porosity, pore sizes and pore morphologies. The

pore size is very important for proper osseointegration in biomed-ical implants and it was shown that this can be controlled to some

extent by optimizing the deposited track geometry and offset dis-

tance. MicroCT analysis of continuous-wave deposited samples

showed pores with excellent interconnectivity were produced,

which is important for cells migration requirements. A 3D ana-

lytical model for deposition of multilayer porous structures has

been developed. The model accounts for the multiple track inter-

action in deposition of a single layer and then multiple layers to

build the porous structures. The model results are compared with

experiments and show a good agreement.

Using a pulsed beam method to create interacting balls rather

than deposition tracks allowed structures with a greater level

of controlled porosity to be created than with a continuous

beam. It is also possible to control final part characteristicssuch as pore size and overall porosity by deposition parame-

ter and overlap ratio selection using this method. A problem

with this method is that the final structures are not regular

and may consequently be liable to premature failure. An analyt-

ical model for the pulsed deposition of the individual balls has

been developed. The model accurately predicts the deposited ball

diameters and also follows the experimental trends of the ball

height.

Both methods could be able to produce the porous structures

with pore size range of 100 microns but with different sets of

processing parameters for each method. Major process variable

affecting pore sizes is laser spot diameter. With the same laser spot

diameter, pulsed-wave laser deposition can produce smaller pore

sizes compared to that of continuous-wave laser deposition.




Acknowledgements

The first author would like to acknowledge the financial assis-

tance of the Government of Pakistan in thepresent work. Thanks to

the staff of Laser Processing Research Centre (LPRC) for their help.

References

Black, J., Hastings, G., 1998. Handbookof Biomaterial Properties. Chapmanand Hall,

London, p. 135.Choi, M., Greif, R., Salcudean, M., 1987. A study of the heat transfer during arc-

welding with applications to pure metals or alloys and low or high boilingtemperature materials. Numerical Heat Transfer 11 (4), 477–489.

Cline, H.E., Anthony, T.R., 1977. Heat treating and melting material with ascanning laser or electron-beam. Journal of Applied Physics 48 (9), 3895–3900.

Götz,H.E.,et al.,2004.Effectof surfacefinishon theosseointegration of laser-treatedtitanium alloy implants. Biomaterials 25 (18), 4057–4064.

Ik-Hyun, et al., 2003. Mechanical properties of porous titanium compacts preparedby powder sintering. Scripta Materialia 49 (12), 1197–1202.

Karageorgiou, V., Kaplan, D., 2005. Porosity of 3D biomaterial scaffolds and osteo-genesis. Biomaterials 26 (27), 5474–5491.

Kirshna,B.V., et al.,2008. Engineered porous metals for implants.JOM 60 (5),45–48.Kroger, H., Venesmaa, P., Jurvelin,J., Miettinen, H., Soumalainen,O., Alhava, E., 1998.

Clinical Orthopedic Related Research 352, 66–74.Pinkerton, A.J., Li, L., 2004a. An analytical modelof energy distributionin laser direct

metal deposition. Journal of Engineering Manufacture 218 (4), 363–374.Pinkerton, A.J., Li, L., 2004b. Modelling powder concentration distribution from a

coaxialdeposition nozzle forlaser-basedrapidtooling. Journalof Manufacturing

Science and Engineering 126 (1), 33–41.Pilliar, R.M., 1987. Porous-surfaced metallic implants for orthopedic application.

Journal of Biomedical Materials Research 21, 1–33.Sychterz,C.J., Topoleski, L.D., Sacco, M., Engh, C.A., 2001. Clinical Orthopedic Related

Research, 218–227.Wen, C.E., et al., 2001. Processing of biocompatible porous Ti and Mg. Scripta Mate-

rialia 45 (10), 1147–1153.Xue, W., et al., 2007. Processing and biocompatibility evaluation of laser processed

porous titanium. Acta Biomaterialia 3 (6), 1007–1018.

Porous Structures Fabrication by Continuous and Pulsed Laser Metal Deposition

Documents

Transcript of Porous Structures Fabrication by Continuous and Pulsed Laser Metal Deposition