Ct

Na

b

a

ARRA

KPLSWD

1

mitFsmaclsmwtpftNch

p(

0d

Journal of Materials Processing Technology 210 (2010) 2261–2267

Contents lists available at ScienceDirect

Journal of Materials Processing Technology

journa l homepage: www.e lsev ier .com/ locate / jmatprotec

omparison of dry and wet fibre laser profile cutting of thin 316L stainless steelubes for medical device applications

. Muhammada,∗, D. Whiteheada, A. Boorb, L. Li a

Laser Processing Research Centre, School of Mechanical, Aerospace and Civil Engineering, University of Manchester, P.O. Box 88 Sackville Street, M60 1QD, United KingdomSwiss Tec AG, Bahnhofstrasse 7, FL-9494 Schaan, Principality of Liechtenstein, Switzerland

r t i c l e i n f o

rticle history:eceived 14 May 2010eceived in revised form 9 August 2010ccepted 21 August 2010

a b s t r a c t

In medical coronary stent fabrication, high precision profile cutting with minimum post-processing isdesirable. Existing methods of profiling thin tubular metallic materials are based mainly on the use ofNd:YAG lasers. In recent studies fibre lasers have been used for stent cutting. However, for profiling thin(<4 mm diameter, < 200 �m wall thickness) stainless steel tubes, back wall impingements often occur.

eywords:recisionaser cuttingtent

This paper presents a comparison of wet and dry pulsed fibre laser profile cutting of 316L stainless steeltubes. When water flow was introduced in the tubes, back wall damage was prevented. Meanwhile,heat affected zone (HAZ), kerf width, surface roughness and dross deposition have also been improvedcompared with the dry cutting. The scientific study on the effect of internal water flow on laser cuttingof thin tubular stainless steel material is reported for the first time.

aterross

. Introduction

One of the growing applications of laser micro-machining foredical application is the manufacturing of coronary stents. A stent

s a wire mesh tube which is deployed in a diseased coronary arteryo provide a smooth blood circulation as referred to Whittaker andillinger (2006). Coronary artery disease (CAD) reduces the bloodupply to the heart and causes angina. Stenting is a favourableinimal invasive method to open the occluded coronary artery

s surgery can be avoided. Kathuria (2005) describes the typi-al sizes of stents used in clinical practice: diameter = 2–4 mm;ength = 15–20 mm. Materials used for this application includestainless steel, chromium cobalt, nitinol (nickel–titanium shapeemory alloy) and tantalum alloys. Laser technology is a mostidely used method in processing stents as compared to elec-

ric discharge machining, water jet cutting, braiding, knitting andhotochemical etching as reported by Stoeckel et al. (2002). Stentabrications require a high precision process in order to main-

ain the slit structures and width. Started with flash-lamp pumpedd:YAG lasers, laser micro-profiling has been an established tool fororonary stent manufacture. A considerable amount of literaturesave been published on Nd:YAG laser applications in stent cutting.

∗ Corresponding author. Tel.: +44 0 1613062350; fax: +44 0 1612003803.E-mail addresses: Noorhafiza.Muhammad@postgrad.manchester.ac.uk,

jatt 07@yahoo.com (N. Muhammad), David.Whitehead@manchester.ac.ukD. Whitehead), alanb@swisstecag.com (A. Boor), lin.li@manchester.ac.uk (L. Li).

924-0136/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.jmatprotec.2010.08.015

© 2010 Elsevier B.V. All rights reserved.

Kathuria (1998) conducted a feasibility study of pulsed Nd:YAGlaser precision fabrication of metallic stents. He suggested that apulsed Nd:YAG laser is a viable tool in creating such fine and meshstructure with slit width of 100 �m. High pulse repetition rate withshort pulse duration is preferred for the high cut quality. However,the processed samples were not free from dross and spatter adher-ence. Work by Raval et al. (2004) shows that Nd:YAG laser cutting ofstents is associated with slag, oxide layer and unacceptable surfacequality.

In the last few years, the emergence of fibre laser technolo-gies has enabled their increasing applications in medical devicemicro-machining. Kleine and Watkins (2003) conducted a compar-ative study to evaluate the cutting quality between a pulsed lamppumped Nd:YAG and a fibre laser. Both systems were adjusted tohave nearly the same beam quality and beam diameter. The sameprocessing parameters were applied during cutting processes toaccomplish a valid comparison results. They have demonstratedthat cutting with an Nd:YAG laser slightly degrades the surfacequality due to wider striations zone from the top edge of the lasercut. A study by Liu et al. (2005) identified that flash-lamp pumpedNd:YAG lasers produced large kerf widths and an expansion of heataffected zones due to low facular quality. On the other hand, thepoor stability laser outputs of the Nd:YAG laser caused difficulty

in reaching small and consistent kerf width in micro-machining.Recent investigation by Meng et al. (2009) demonstrated that cut-ting quality (heat affected zone and average roughness) was betterwith a fibre laser compared to an Nd:YAG laser due single modeoutput and small focused size of the fibre laser. Miller et al. (2009)

2262 N. Muhammad et al. / Journal of Materials Processing Technology 210 (2010) 2261–2267

Table 1Chemical composition of stainless steel 316L (Kathuria, 2003).

3.0% m

emcassmc

dbtcttpnrmwiwflasdwhoifl

fiewti

2

2

mwTthlca2focitt

3.1. Effects of cutting parameters upon kerf width and surfaceroughness

Fig. 2 shows the relationship between the kerf width and lasercutting parameters. The standard deviations were taken as error

Table 2Cutting parameter ranges.

Elements C Mn Cr Ni Mo

% 0.03% max 2.0% max 17.0–19.0% max 13.0–15.5% max 2.0–

xplained that embrittlement of metal in the heat affected zoneay lead to crack formation and expansion of the stents which may

ause crack propagation and device failure. Thus significant costsre associated with post-processing of Nd:YAG laser machinedtents to produce high quality products. From the aforementionedtudies, fibre laser is seen as a potentially new technology in stenticro-machining which heat affected zone, kerf width and dross

ould be diminished to a minimum.The other issue arises in stent micro-machining is the back wall

amage either processed with Nd:YAG or fibre lasers. Preventingack wall damage is a challenge in cutting such small and thinubes. The ejected materials from a cut kerf in a form of hot parti-les adhere and create back wall damages to the opposite wall of theube. Back wall damage would cause rough surface finish and crackso the stent structures due to the dissipation of heat by the ejectedarticles. On the other hand, the laser beam transmission was alsoot prevented from reaching the opposite wall causing deterio-ation to the back wall. Raval et al. (2004) and Kathuria (2005)anaged to protect the laser beam energy transmitting to the backall by inserting the Teflon and Teflon coated brass through the

nner wall. A patent by Merdan and Shedlov (2005) discloses a backall damage prevention method in stent manufacturing. A conduituid source through the tube was proposed to wash away the debrisnd cooling any heat that generated during the cutting. One of theuggested fluids to be used is water. Meng et al. (2009) in their workesigned and utilized the tube cooling equipment that could pumpater through the tubes during the process in order to reduce theeat affected zone and prevent any damage to the opposite surfacef the tube. However, there is no scientific work published regard-ng the performance of fibre laser cutting with and without waterow.

The present work aims to investigate the basic characteristics ofbre laser cutting of stainless steel 316L tube and understand theffect of introducing water flow in the tubes on minimising backall damages and thermal effect. The influence of laser parame-

ers upon cutting quality for fixed gas type and gas pressure wasnvestigated.

. Experimental procedures

.1. The stent cutting system

The tube profiling system used was a Swisstec Micro T15achine designed for stent production. This system was integratedith a GSI JK100FL single mode fibre laser with 100 W peak power.

he system also includes a CNC motion system (3 axis: rotation,ransverse and height), a beam collimator, a cutting head withousing a focusing lens, a coaxial gas nozzle and a CCD vision system

ooking directly down to the nozzle at the cutting area. A computerontrol module was integrated for the process parameter selectionnd control. The fibre laser has a 1080 ± 5 nm output wavelength,5 �m theoretical spot size at the focal plane and a beam qualityactor, M2 < 1.1. The coaxial assist gas nozzle had an exit diameter

f 0.5 mm. The nozzle was optimized for maximum velocity andoaxial gas flow. This equipment has the ability to pump the waternside the tube as an additional means of cooling during the cut-ing process. The laser remains stationary and the tube rotates andraverses automatically during laser machining.

Si Cu N P S Fe

ax 0.75% max 0.50% max 0.10% max 0.025% 0.010% max Balance

2.2. Materials

In this work, 316L stainless steel with an outer diameter of3.175 mm and 150 �m wall thickness was used. The chemical com-position for the material is given in Table 1.

2.3. Cutting experiments

The cutting experiments were performed in two cutting condi-tions, dry and wet by using the pulsed laser mode. Nitrogen wasused as an assist gas. Dry cutting has been performed with thepresence of the N2 assist gas, and the wet cutting was performedwith the presence of an assist gas (N2) and continuous water flowthrough the inner part of the tube along the tube axis. In the wetcutting, the water pipe with the same diameter of the tube was con-nected to the tube opening and to the water supply container. In thiscase, the water flow rate was measured to be 1567 mm3/s whichwas kept constant during the experiment. Preliminary experimentswere carried out to determine the appropriate processing parame-ters to be used for the comparative study. The range of parameterschosen was based on necessary average power needed to achieve afull depth penetration for both cutting conditions at the selectedcutting speed ranges. Laser peak pulse power, frequency, pulsewidth and the cutting speed were varied in the experiments andtheir variation ranges are shown in Table 2. The range of cuttingspeed was selected from 250 mm/min to 2000 mm/min which is thelimit of the machine. The parameter variation used was the samefor dry and wet cutting condition. The gas pressure was constantat 6 bar for all the cutting experiment performed, limited by themachine. The cutting quality factors investigated were kerf width,surface roughness, dross deposition, back wall damage and heataffected zone (HAZ).

In this study, the Computer Aided Design (CAD) data of the cutprofiles was created and transferred to the Micro T15 computersystem. The data were translated into G-code programming by thecomputer system which enables one to perform the cutting processwith the desired profiling pattern. In order to study the quality char-acteristics of the fibre laser stent cutting system, a simple geometrywas designed to cut the stainless steel 316L tubes. Initially, thetubes were cut into two separated parts; Part A and Part B as shownin Fig. 1. The simple geometry was used to assess the basic charac-teristics including kerf width, surface roughness, dross deposition,back wall damages and HAZ. A more complex profile was used toassess the heat effects, particularly along the cut kerf.

3. Results

Cutting parameters Lower limit Upper limit

Peak pulse power (W) 70 100Frequency (kHz) 1.5 4Pulse width (ms) 0.1 0.2Cutting speed (mm/min) 250 2000

N. Muhammad et al. / Journal of Materials Proce

bpibrlaaltatad

Fig. 1. Scanning electron microscope (SEM) image of the laser cut geometry.

ars. The results show that the kerf width increased as the peakulse power, frequency and pulse width increased (Fig. 2a–c). Dur-

ng the cutting process, the average power supplied was controlledy these three parameters. The small variation in average poweresulted in a large variation of kerf width. Increasing the speeded to reduction of kerf width after a critical cutting speed waschieved (Fig. 2d). Less interaction time between the laser beamnd material reduced the energy supplied to the cut kerf, and thusess amount of material was melted. The kerf width size varia-

ion was investigated by varying the speed between 250 mm/minnd 2000 mm/min (maximum equipment speed limit). Critical cut-ing speed for dry cutting and wet cutting were 1250 mm/minnd 1000 mm/min, respectively, where the kerf width started toecrease after these particular points. The kerf widths obtained

Fig. 2. Kerf width as a function of laser cutting parameters in dry and wet cutting

ssing Technology 210 (2010) 2261–2267 2263

in this experiment were within the range of 28–40 �m. Minimumkerf was obtained at the highest cutting speed, 2000 mm/min. FromFig. 2, it can be clearly seen that the wet cutting produced a low kerfwidth compared to dry cut. In addition, it is seen that kerf widthincreases with the repetition rate in the fibre laser cutting as moreenergy was delivered to the process.

Fig. 3 shows the influences of laser parameters upon surfaceroughness. The surface roughness increased with the increasingpeak pulse power. However, the surface roughness reduced whenthe peak pulse power increased from 90 W to 100 W (Fig. 3a).Higher pulse frequency improves surface quality as shown inFig. 3b. Pulse width significantly affects the surface roughnesswith a rougher cut surface produced as the pulse width increases(Fig. 3c). From Fig. 3d, surface roughness decreased as the cut-ting speed increased until 1250 mm/min for the dry cut and1000 mm/min for wet cut. After these points, the surface roughnessgradually increased and were more pronounced at higher cuttingspeed for wet cutting. This result shows that more power is requiredafter certain point of speed to improve surface quality. In compar-ison with dry cutting, wet cutting led to better surface quality.

3.2. Dross formation

The dross formation after the laser cutting in dry and wet cuttingcondition was observed as shown in Fig. 4. From the observation,the processed samples were not free from dross. In both cuttingconditions, the dross deposition was heavy at the low cutting speedof 500 mm/min and slightly reduced at 1000 mm/min. The moltenmaterial was not totally ejected out from the cut kerf and wasattached to the bottom side of the cut wall. Above 1000 mm/min,the dross formation is not significant. By comparing the dross depo-

sition at cutting speed 1000 mm/min for both wet and dry cuttingconditions, dross was significantly reduced in the wet cutting con-dition at 1000 mm/min. In inert gas cutting, the energy solely camefrom the focused laser beam. The presence of inert gas does not con-tribute any additional energy to the cutting point (cooling of the cut

: (a) peak pulse power, (b) frequency, (c) pulse width and (d) cutting speed.

2264 N. Muhammad et al. / Journal of Materials Processing Technology 210 (2010) 2261–2267

t cutt

zf6dfiar

3

tpwibmw

Fig. 3. Surface roughness as a function of laser cutting parameter in dry and we

one may be resulted) thus producing high viscosity and high sur-ace tension of molten material. The obtained results showed thatbar gas pressure was not enough to act as a mechanical force torag away the molten material. The presence of water was not suf-cient to clean the dross. Water flow, however, helped to achievereduction in dross. It is shown that higher-pressure inert gas is

equired to clean the dross in achieving high cutting quality.

.3. Back wall damage

Preventing back wall damage is a challenge in cutting small andhin tubes. The ejected materials from a cut kerf in a form of hotarticles adhered and created back wall damages to the opposite

all of the tube. Back wall damage would cause rough surface fin-

sh and cracks to the stent structures due to the dissipation of heaty the ejected particles. On the other hand, the laser beam trans-ission was also not prevented and transmitted to the oppositeall causing deterioration to the back wall. In the dry cutting, it

Fig. 4. Dross deposition at different cutting

ing: (a) peak pulse power, (b) frequency, (c) pulse width and (d) cutting speed.

can be observed that the undesired particles were scattered on theopposite wall and this phenomenon was not avoidable by chang-ing cutting parameters (Fig. 5a). Heat contained in these particleswas transferred to the back wall. In the wet cutting, a clean andspatter-free back wall was obtained as can be seen in Fig. 5b. Thecontinuous water stream carries away the hot particles after theyare ejected from the cut kerf and minimizes the heat transfer to theback wall. Fig. 6a and b compares the surface profile analysis of theback wall for both wet and dry cutting conditions. The wet cuttingseems to be effective in preventing the back wall damage.

3.4. Heat effects

Heat effect on the surrounding material is a critical factor in cut-ting thin materials especially in medical device application. Smalland thin materials are very sensitive to thermal distortion. Experi-ment results show that shorter pulse width and low average powerreduced the thermal distortion. A shorter pulse width is always rec-

speed in different cutting condition.

N. Muhammad et al. / Journal of Materials Processing Technology 210 (2010) 2261–2267 2265

Fig. 5. SEM images of the back wall: (a) dry cutting and (b) wet cutting.

Fig. 6. White light inteferometer (Wyko NT1100) images of surface profiles of the back wall: (a) dry cutting and (b) wet cutting.

at 40

oattmtm

Fig. 7. Comparison of thermal effect along the cut kerf

mmended in cutting materials sensitive to distortion. At a highverage power (40 W), dry cutting (Fig. 7a) resulted in a noticeable

hermal effect and surface oxidation along the cut while wet cut-ing (Fig. 7b) reduced this effect even at high average powers. A

ore complex profile similar to a medical stent was used to assesshe heat effect particularly along the cut kerf. Fig. 8 shows the ther-

al discoloration in dry stent cutting, while cutting in the water

Fig. 8. Thermal discoloration in stent cutting before removing

W average power: (a) dry cutting and (b) wet cutting.

assistance condition gave a clean and bright surface. Wet cuttingrelatively created a controllable heating during the cutting process

and reduced the extent of heat significantly that benefits particu-larly in cutting thin and small diameter tubular materials.

The heat affected zones (HAZ) corresponding to a high aver-age power (40 W) for both cutting conditions were compared afterthe laser cut samples were undergone a series of grinding, polish-

the excess material: (a) dry cutting and (b) wet cutting.

2266 N. Muhammad et al. / Journal of Materials Processing Technology 210 (2010) 2261–2267

F assocw

iiloep3

4

qtoic

P

wq

iHawmsrpH

ig. 9. Tranverse section of the cut kerf showing the changes of the microstructureet cutting.

ng and etching processes. The grain structures were revealed bymmersing the samples in Glyceregia solution for 60 s. The averageaser power of 40 W was selected as a comparative parameter tobserve the effectiveness of water to reduce the heat affected zoneven at high power input. As expected, wet cutting resulted in lessronounced HAZ with the width <10 �m while dry cutting reached0 �m (Fig. 9).

. Discussions

In fibre laser cutting processes, many factors affect the cuttinguality. To understand the effect of each processing parameter,he approach of varying one parameter at a time and keeping thether parameters constant was used. In this experiment, the var-ed parameters were peak pulse power, pulse width, frequency andutting speed. The average power, Pave can be obtained from:

ave = Pp� f (1)

here, Pp is the peak power, � is pulse width and f is pulse fre-uency.

Increasing all these three parameters (Pp, �, f) will lead to thencrease of laser average power delivered to the cutting process.igher average power generates more energy transfer to the kerfnd increases the amount of melted material which translates to aider kerf. Increasing the pulse width causes a larger kerf due to

ore material being melted and removed by the gas. As the cutting

peed reduces, the interaction time between laser beam and mate-ial increases which creates a larger kerf. Increasing the peak pulseower and pulse width resulted in increasing surface roughness.igher frequency applied to the cutting process improves sur-

Fig. 10. Mechanism of molten material behaviour after ejected from cut k

iated with HAZ when cutting at high average power (40 W): (a) dry cutting and (b)

face quality due to high pulse overlapping. Kerf width and surfaceroughness results seem to indicate that pulse width is a dominantfactor affecting the cutting quality. The findings from this studysuggest that the introduction of water flow inside the tube reducedthe local heating during the laser cutting process thus minimizedthe amount of melted material and produced narrower kerfs. On theother hand, the presence of water improves the melt flow behaviourreduced surface roughness as compared to that in the dry cut.

In this experiment, with the flowing water, a clean back walland low thermal effect even at high average powers and high pulseenergy was demonstrated. Fig. 10 illustrates the mechanism ofmolten material behaviour after being ejected from cut kerf in dryand wet cutting. The molten metal solidifies and forms undesiredspatter on the back wall in the case of dry cutting (Fig. 10a). Thiscreated damages to the opposite wall as the wall thickness is thinand thus thermal conduction is not as efficient as in case of thickwalls. Fig. 10b shows the clean and spatter-free back wall that wasobtained by using the wet cutting. Water flow from the inner diam-eter of the tube ideally removed all the molten materials before theypermanently adhere to the back wall.

Significant improvements with wet cutting to minimize the heateffect and back wall damage were achieved in this work. As everysingle stent strut needs to be cut without any defects, the waterpreventive method shows promising results in reducing the heateffect and back wall damages. This is due to the water stream that

carries away the hot debris which minimizes the heat transfer andimpingement to the back wall.

In order to investigate effect of the water in attenuating thelaser transmission during the wet cutting, water transmission spec-trum was measured for the 900–1100 nm wavelength at every 1 nm

erf in tube cutting in Y–Z view: (a) dry cutting and (b) wet cutting.

N. Muhammad et al. / Journal of Materials Proce

Fp

bmbo

A

wfi

˛

f

z

cfawss

Bbtcocsca

ig. 11. Water transmission spectrums recorded by spectrometer for 1 cm waterath length.

y using an Ocean Optics spectrometer (Fig. 11). The measure-ent was made with a water path length of 1 cm. The relationship

etween light transmission and absorbance can be obtained basedn Beer–Lambert law as referred in Robinson (1996):

= −log10 T (2)

here A is absorbance and T is transmittance. The absorption coef-cient, ˛ is defined by:

=(

1x

)ln

(1T

)(3)

Here, x is the water path length. The absorption length, z there-ore is related to the absorption coefficient, ˛, and is given by:

= 1˛

(4)

From Eqs. (2), (3) and (4), the water absorption spectrum wasalculated and plotted based on the transmission data obtainedrom spectrometer. Through the above calculations, the lightbsorption length for water media is around 4.2 cm for fibre laseravelength (� = 1080 ± 5 nm) as shown in Fig. 12. The absorption

pectrum for this spectrum band is in agreement to the resultshown by Kruusing (2004).

The water thickness flowing inside the tube is x = 0.2875 cm.ased on the Beer–Lambert law, the fraction of the light absorbedy each layer of solution is proportional to the path length. Thus,he transmission for water path length, x = 0.2875 cm has beenomputed based on the spectrometry analysis. The transmission

btained for this particular water path length is 94%, which indi-ates that most of the beam intensity strikes the back wall andhowing that water does not absorb much laser intensity. This elu-idates that the transmission and absorption of light in water isssociated with the light path length in water. High absorptivity of

Fig. 12. Water absorption spectrums.

ssing Technology 210 (2010) 2261–2267 2267

laser energy could not be expected in a thin water path length. How-ever, the transmission is expected to be <94% due to existence ofmolten material (produced during cutting) in the water that slightlyincreases the attenuation.

The spectrum analysis shows that water does not play a signif-icant role in attenuating the laser beam in cutting small diametertube since the 100% absorption can be obtained at around 4.2 cmof water depth. In wet cutting, water played an important role incooling the work piece and carries away the debris rather thanattenuating the beam. A constant rate of water circulation removeslocal heating of water adjacent to the cutting surrounding andimproves the cooling. It is well known that water cools the cuttingzones more rapidly rather than particular gases or gas mixturessuch as air as referred to Kruusing (2004).

In the current work, the wet cutting results obtained were notfree from dross attached to the cutting edge. A conceivable reasonwas the inadequate inert gas pressure which is not sufficient toforce away the molten material from the cutting area in preventingdross.

5. Conclusions

The experimental results comparing dry and wet stainless steel316L tube cutting were reported. Wet cutting enabled significantimprovement in cutting quality. Wet cutting resulted in narrowerkerf width, lower surface roughness, less dross, absence of back walldamages and smaller HAZ which would lead to reducing the cost ofpost-processing. Laser average power and pulse width play a sig-nificant role in controlling the cutting quality. Increasing the pulsewidth increased beam/material interaction time which increasedthe kerf width and surface roughness. Wet cutting is ideally suit-able for cutting thin wall and small diameter tubes specifically instent production.

Acknowledgements

The authors are grateful to GSI group for the provision of the100 W single mode fibre laser for this work. One of the authors(first author) acknowledges the financial support from UniversitiMalaysia Perlis (UniMAP) and Ministry of Higher Education (MoHE)for her PhD scholarship.

References

Kathuria, Y.P., 1998. Laser microprocessing of stent for medical therapy. Proceed-ings of International Symposium on Micromechatronics and Human Science,111–114.

Kathuria, Y.P., 2003. Biocompatible metallic stent for medical therapy. Proceedingsof SPIE 5287, 52–61.

Kathuria, Y.P., 2005. Laser microprocessing of metallic stent for medical therapy.Journal of Materials Processing Technology 170, 545–550.

Kleine, K.F., Watkins, K.G., 2003. Fiber laser for micro cutting of metals. Proceedingsof SPIE 4974, 184–192.

Kruusing, A., 2004. Underwater and water-assisted laser processing. Part 1. Generalfeatures, steam cleaning and shock processing. Optics and Lasers in Engineering41, 307–327.

Liu, W, Du, W., Liao, J., 2005. Application of fiber laser in the field of stent cuttingand micromachining. Proceedings of SPIE 5629, 263–270.

Meng, H., Liao, J., Zhou, Y., Zhang, Q., 2009. Laser micro processing of cardiovascularstent with fiber laser cutting system. Optics and Laser Technology 41, 300–302.

Merdan, K., Shedlov, M., 2005. Inverted stent cutting process. US Patent2005/1224473 A1, United States.

Miller, P., Aggarwal, R., Doraiswamy, A., Lin, Y.J., Lee, Y.S., Narayan, R.J., 2009. Lasermicromachining for biomedical applications. Journal of the Minerals, Metals andMaterials Society 61, 35–40.

Raval, A., Choubey, A., Engineer, C., Kothwala, D., 2004. Development and assess-ment of 316LVM cardiovascular stents. Materials Science and Engineering A

386, 331–343.

Robinson, J.W., 1996. Atomic Spectroscopy. Marcel Dekker, Inc., New York.Stoeckel, D., Bonsignore, C., Duda, S., 2002. A survey of stent designs. Minimally

Invasive Therapy and Allied Technology 11, 137–147.Whittaker, D.R., Fillinger, M.F., 2006. The engineering of endovascular stent tech-

nology, a review. Vascular and Endovascular Surgery 40, 85–94.