Development of a new niobium-based alloy for vascular stent applications

J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 1 ( 2 0 0 8 ) 3 0 3 – 3 1 2

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/jmbbm

Research paper

Development of a new niobium-based alloy for vascular stentapplications

Barry O’Briena,c,∗, Jon Stinsonb, William Carrollc

a Boston Scientific Corporation, Ballybrit Business Park, Galway, Irelandb Boston Scientific Corporation, One Scimed Place, Maple Grove, MN55311, United StatescNational Centre for Biomedical Engineering Science, National University of Ireland Galway, Ireland

A R T I C L E I N F O

Article history:

Received 2 October 2007

Received in revised form

8 November 2007

Accepted 13 November 2007

Published online 17 November 2007

Keywords:

Stent

Niobium

Tubing

Magnetic resonance imaging

Elastic modulus

A B S T R A C T

This study was performed in order to develop a new stent material that would provide

reduced MR image artifact compared to current stent materials. Alloy design rationale

is initially presented and following this the development of a Nb–28Ta–3.5W–1.3Zr alloy

is described, including the manufacture of stent tubing. Tensile testing of this new alloy

showed that it had approximately twice the yield strength of current Nb–1Zr material

with a 25% higher elastic modulus. The new alloy was also confirmed to have suitably

low magnetic susceptibility. Mechanical testing of demonstration coronary stents made

from the new alloy were shown to have acceptable compression strength and elastic

recoil performance. It is concluded that this new Nb–28Ta–3.5W–1.3Zr alloy is a practical

candidate stent material for both coronary applications and peripheral uses such as carotid

or intracranial stenting, where reduced MR image artifact would be beneficial.c© 2007 Elsevier Ltd. All rights reserved.

u

d

1. Introduction

The use of metallic stents for the treatment of coronaryand peripheral vascular disease is now a standard practice.Successive iterations of coronary stents have seen restenosislevels reduced to below 10%, particularly with the relativelyrecent introduction of drug eluting stents (Ong and Serruys,2005). The majority of the commercially successful coronarystents are made frommaterials such as 316L stainless steel orcobalt–chromium-based alloys such as L605 andMP35N (Maniet al., 2007). Selection of these materials has been primarilydriven by mechanical performance requirements and to alesser extent by the availability of fabrication technologies.Whilst biocompatibility was a significant factor, prior implant

∗ Corresponding author at: Boston Scientific Corporation, Ballybrit BE-mail address: [email protected] (B. O’Brien).

1751-6161/$ - see front matter c© 2007 Elsevier Ltd. All rights reservedoi:10.1016/j.jmbbm.2007.11.003

siness Park, Galway, Ireland. Tel.: +353 87 2934292.

experience in orthopaedic applications contributed to what isnow widespread acceptance of these materials.

It was however somewhat fortunate that these materialshad reasonable compatibility with the imaging tools mostwidely used during stent implantation. During implantation,the location and deployment of the stent is typicallymonitored by x-ray fluoroscopy. The attenuation of x-raysby the stent material allows the device to be imaged,with sufficient accuracy to ensure proper placement anddeployment against the vessel wall. A number of approacheshave been used to more finely tune radiopacity, such ascoatings (Reifart et al., 2004) or marker bands made fromdense noble metals (Wiskirchen et al., 2004). These have hadvarying degrees of success but overall, the use of stainless

.

www.sciencedirect.com



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

mailto:[email protected]

http://dx.doi.org/10.1016/j.jmbbm.2007.11.003

304 J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 1 ( 2 0 0 8 ) 3 0 3 – 3 1 2

steel and cobalt–chromium alloys has been acceptable froman x-ray fluoroscopy perspective. Material such as nitinol hasbeen more challenging and the wide use of this material inperipheral stenting applications is dependent on radiopacityenhancement through markers or coatings (Stoeckel et al.,2004).

Similarly, during follow-up angiography to assess forpatency, stentsmade from stainless steel or cobalt–chromiumperform well under x-ray fluoroscopy, allowing detectionof restenotic tissue or thrombus within the stent lumen.However recent years has seen a rapid development innew imaging modalities such as computed tomography(CT) and magnetic resonance imaging (MRI). While x-ray fluoroscopy will continue to be the preferred imagingtechnique during interventional procedures, these newmethods are increasingly being used for screening and follow-up procedures. CT angiography can provide significantlymoreinformation on the configuration and condition of the vesseland stent due to the three-dimensional nature of the datacollected. This has resulted in rapid take up of CT fornon-interventional imaging, particularly for coronary vessels(Schoenhagen et al., 2004; Ehara et al., 2006). In addition,MR offers advantages in terms of eliminating both ionizing-radiation- and iodine-based contrast agents as well as beingnon-invasive for some procedures. MR angiography for thecoronary vessels is challenging due to insufficient resolutionand relatively slow image capture combined with vesselmotion artifacts (Woodward et al., 2000). These issues arebeing addressed with on-going developments by the largeimaging companies. However, MR is already seeing increaseduse for peripheral angiography where these aspects are lessof an issue (Rofsky and Adelman, 2000).

However, challenges arise when it is desired to useMR angiography for a stent follow-up procedure, or for ascreening procedure on a vessel with a previously implantedstent. The paramagnetic nature of austenitic stainless steeland cobalt–chromium stents causes a localized distortion ofthemagnetic fields resulting in loss of signal which ultimatelyshows as an artifact on the collected image. In the case ofthese paramagnetic materials the artifact usually obscuresthe stent lumen and most often extends well beyond theboundary of the device itself (Hug et al., 2000; Nitatori et al.,1999). This can prevent full or proper interpretation of thedata. This artifact problem is more critical on coronary stentsthan peripheral sizes and furthermore, many peripheralstents are made from nitinol which has lower paramagnetismand therefore less artifact (Lenhart et al., 2000). Howeverin summary, magnetic induced artifact is a problem for MRimaging of both peripheral and coronary devices and whileimaging of peripheral devices is feasible to some extent,imaging of coronary stents is not practical. A need thereforeexists to provide a stent material which meets all the regularrequirements for stent performance, but which will alsoprovide reduced image artifact during MR angiography orduring any other MR imaging procedure. This ideal materialdoes not exist and the study reported here aimed to developand evaluate a niobium-based alloy for this purpose.

Table 1 – Bulk magnetic susceptibility values for selectedmetals and alloys

Material Magneticsusceptibility, χ

Ti 182 × 10−6

Ta 178 × 10−6

Nb 237 × 10−6

W 77.2 × 10−6

Zr 109 × 10−6

Mo 123 × 10−6

Cr 320 × 10−6

Nitinol 245 × 10−6

Stainless Steel (austenitic) 3520–6700 × 10−6

Ni 600Co 250Fe 200,000

2. Alloy design rationale

One of the first selection criteria was to consider onlycompositional elements which have low bulk magneticsusceptibility but which also have appropriate mechanicalcharacteristics. Magnetic susceptibility is a measure of theextent to which amaterial becomesmagnetized in amagneticfield. Table 1 lists the magnetic susceptibility (χ) for someof the relevant elements and alloys as originally describedby Schenck (1996). Units shown are in the SI system, whichare dimensionless.

Niobium, tantalum and titanium are all immediatecandidates due to their lowmagnetic susceptibility and usefulmechanical properties. Table 2 shows some basic mechanicalproperty data for these elements as well as for some relatedcommercially available alloys. All data refer to the material inthe annealed condition, which is normally desirable from astent durability perspective.

Pure titanium and commercially available Ti alloys wereruled out due to having relatively low elastic modulus andonly moderate ductility, in comparison to stainless steel.These are important characteristics when considering theexpansion behaviour and durability of balloon expandablestents. Stent materials need to have a balance of relativelyhigh elastic modulus combined with a significant plasticdeformation capability to ensure good expansion anddurability without risk of elastic recoil or strut fractures.It is worth noting that while there are many other well-known titanium alloys being developed for orthopaedicimplant applications, the mechanical property requirementis significantly different. Specifically, the desire in theseapplications is to decrease the elastic modulus so that it moreclosely matches the behaviour of bone (Niinomi, 2008). Evenwhen titanium is alloyed with relatively high elastic modulusconstituents, the additions invariably result in a decrease inmodulus, as demonstrated in the case of tantalum additionsto titanium (Zhou et al., 2004).

Tantalum has been used in the past with some degreeof success for both the Wiktor coronary stent (Vaishnavet al., 1994) and the Strecker peripheral stent (Long et al.,1995). However, the possibility of Ta alloy development wasruled out as it was considered that the high density of thematerial would result in excessive radiopacity when imaged

J O U R N A L O F T H E M E C H A N I C A L B E H AV I O R O F B I O M E D I C A L M A T E R I A L S 1 ( 2 0 0 8 ) 3 0 3 – 3 1 2 305

Table 2 – Tensile mechanical properties for selected elements and alloys

Material Ultimate tensilestrength

0.2% Yieldstrength

Elongation(%)

Elastic modulus Reference

Ti 240–331 MPa 170–241 MPa 30 102.7 GPa ASM Handbook (1990)35–48 ksi 25–35 ksi 14.9 msi

Ti–6Al–4V 900–993 MPa 830–924 MPa 14 113.8 GPa ASM Handbook (1990)130–144 ksi 120–134 ksi 16.5 msi

Ti–3Al–2.5V 620–689 MPa 520-586 MPa 20 106.9 GPa ASM Handbook (1990)90–100 ksi 75–85 ksi 15.5 msi

Nb 195 MPa 105 MPa 25 103 GPa Poncin and Proft (2003)28 ksi 15 ksi 14.9 msi

Nb–1Zr 241 MPa 138 MPa 20 68.9 GPa ASM Handbook (1990)35 ksi 20 ksi 10.0 msi

C103 405 MPa 310 MPa 26 87 GPa ASM Handbook (1990)59 ksi 45 ksi 12.6 msi

Ta 207 MPa 138 MPa 25 185 GPa Poncin and Proft (2003)30 ksi 20 ksi 26.8 msi

316L Stainless 595 MPa 275 MPa 60 193 GPa Davies (1994)Steel 86 ksi 40 ksi 28 msi

by x-ray fluoroscopy. Furthermore, the highly refractivenature of tantalum, with a melting point of 2996 ◦C, wouldhave created significant fabrication challenges from alloydevelopment through to stent electropolishing.With titaniumand tantalum ruled out, this study focused on using niobiumas the base element for alloy development.

Having selected niobium as the platform for alloydevelopment, some commercial compositions were initiallyassessed. Pure niobium and the Nb–1Zr alloy have beenwidely used in the nuclear industry for high-temperatureapplications. However, the elastic modulus and yield strengthwere considered to be too low for achieving optimumstrength and expansion characteristics on a stent. The onlyother available commercial grade is alloy C103, which is aNb–10Hf–1Ti composition. The tensile strength, yield strengthand ductility of this material are closer to the requirementsof a stent material, however the elastic modulus was stillconsidered to be too low. Historically, the development ofniobium alloys was targeted at high-temperature applicationsin the nuclear and aerospace industries. However, most ofthe alloys with potentially useful properties for stents areno longer commercially available. In particular, alloys withTa, Hf, W and Zr additions show promising elastic modulusvalues in the range 123–140 GPa (Wojcik, 1998). Internalexploration of similar systems therefore commenced.

In addition to the magnetic susceptibility and mechanicalaspects already reviewed, the contribution which alloyingadditions make to radiopacity was also considered. Stentradiopacity is significantly influenced by geometry and strutthickness, which in turn may be controlled by materialmechanical properties. However, apart from these geometryand mechanical interactive influences, radiopacity can alsobe considered to be directly proportional to the x-ray linearattenuation coefficient. Table 3 shows linear attenuationcoefficients for some of the main elements involved in thisdevelopment and shows iron for comparison (Hubbell andSeltzer, 2004). The dependence of attenuation on x-ray energy

Table 3 – x-ray linear attenuation data for selectedalloying elements and iron

Element x-ray linear attenuationcoefficient at 100 keV

(cm−1)

Niobium 8.9Tantalum 71.4Tungsten 85.5Zirconium 6.2Iron 2.9

level should be noted, as this can alter relative positioning ofmaterials.

Tantalum was selected as the main addition in orderto provide some solid solution strengthening to theniobium, while maintaining the low magnetic susceptibility.Furthermore, the high attenuation coefficient for tantalumwould provide a significant contribution to improving theradiopacity beyond that of niobium- or iron-based alloys.Tungsten additions were also selected to provide solidsolution strengthening, though at a much lower level, astungsten has been shown to be a much more potentstrengthener than tantalum (Slining and Koss, 1973). Lowerlevels were also desirable in order to avoid excessiveradiopacity and to avoid the risk of low formability resultingfrom this highly refractive addition. Though it was alsobelieved that tungsten should contribute more significantlyto increases in modulus than tantalum.

Zirconium was retained at the low levels typicallyobserved in commercial grades not only due to itsstrengthening contribution but also due to its ability tointeract with any low levels of interstitial oxygen impurity.Like most refractory materials, niobium has a tendency toabsorb oxygen during elevated-temperature processing, withthe risk of reducing ductility. Zirconium plays an importantrole in that it ultimately leads to the formation of a ZrO2precipitate that helps maintain ductility, even in the presence


of low levels of oxygen contamination (DiStefano et al., 2000).It was therefore considered useful to retain this function inthe alloy design.

Initially the alloy selected for exploration was a Nb-28Ta-10W-1.3Zr (weight %) composition. However this initialcomposition proved to be difficult to fabricate due to lowductility. This was attributed to the tungsten content whichwas therefore reduced, leading to a Nb–28Ta–3.5W–1.3Zrcomposition which became the main composition forcontinued exploration and development. Initial trials withalloy manufacture and mechanical assessment are reportedhere. Subsequent scale up and assessment of stent tubing ispresented, as well as data on evaluation of the first prototypecoronary stents made with this material.

3. Materials and methods

3.1. Experimental Nb–28Ta–3.5W–1.3Zr ingot and stripproduction

An arc melting furnace (Materials Research Furnaces ABJ-900)was used to produce an ingot of approximately 150 g. Theraw material charge for the melt were all in powder form,obtained from CERAC Inc (Milwaukee, WI, US). Melting wasinitially performed at 250 Amp, followed by a 400 Amp step toensure melt homogeneity.

The ingot was cut and machined into three bars forannealing and cold rolling to strip. An initial vacuumhomogenization treatment was carried out at 1200 ◦C for 6h. The bars were then cold rolled to 50% reduction followedby annealing at 1200 ◦C for 1 h. After two further iterations ofcold rolling and annealing, strips suitable for the productionof tensile test samples were obtained, with a width ofapproximately 15.5 mm and thickness of 0.70 mm. The finalvacuum anneal step was performed at 1200 ◦C for 1 h.

3.2. Prototype Nb–28Ta–3.5W–1.3Zr ingot and stenttubing production

A larger prototype ingot (5 kg) was produced to the targetcomposition using vacuum arc remelting (VAR) techniques,by MetalWerks PMD Inc (Aliquippa, PA, US). The ingot wasapproximately 62 mm diameter and 150 mm long. Thiscylindrical ingot was then drilled to produce a thick walltube which was subsequently extruded to reduce diameter,increase length and to convert the microstructure from as-cast to wrought. The extrusion was then vacuum annealedat 1100 ◦C for 2 h and then pilgered to produce feedstockfor tube mandrel drawing. Standard tube drawing and inter-stage annealing processes were then used to draw downthe tubing of finished dimensions. An initial batch of tubingwas produced with outside diameter of 2.11 mm and wallthickness of 112 µm. Subsequently two further batches wereproduced, both with outside diameter of 1.83 mm but withwall thicknesses 127 µm and 152 µm. A final vacuum annealstep of 1150 ◦C for 2 h was used for all batches to produce amainly recrystallized grain structure. Tube manufacture wascarried out by Accellent Cardiology (Salem, VA, US).

Fig. 1 – Pattern of 3.0 mm × 16 mm demonstration stent.

3.3. Commercial Nb–1Zr alloy tubing

In order to establish a current baseline for the performance ofcommercially available material, a quantity of Nb–1Zr tubingwas purchased from W.C. Hereaus (Hanau, Germany). Thiswas obtained in the as-drawn condition and subsequentlyvacuum annealed at 1150 ◦C for 1 h to obtain a fullyrecrystallized grain structure. This tubing had an outsidediameter of 2.11 mm and a wall thickness of 110 µm.

3.4. Strip and tube tensile testing

Strip tensile testing was performed using samples in generalaccordance with the geometry outlined in Fig. 1 of ASTME8 [Standard Test Methods for Tension Testing of MetallicMaterials]. Specific gauge region dimensions were 3.175 mmwidth, 0.508 mm thickness and 17.018 mm length. Testingwas performed at room temperature using a 12.7 mm (0.5”)extensometer gauge length. A strain rate of 0.005/min wasused up to the 0.2% off-set yield strength and a loading speedof 0.508 mm/min was used from this point up to failure.Tensile testing of this strip material was restricted to threesamples due to the limited availability of this prototype alloy.

Tube samples were tested with the aid of internal pinsat the sample ends for gripping. An extensometer with a25.4 mm (1.0”) gauge length was used for measuring strain.A test loading speed of 0.508 mm/min was used up to the0.2% off-set yield strength and a speed of 5.08 mm/min fromyield through to failure. Twelve tensile tests were performedon the first batch of tubing, while thirteen tests were carriedout on each of the two subsequent batches. Five tensile testswere performed on the commercial Nb–1Zr tubing.

3.5. Metallographic examination

The grain structure was examined in order to confirm thata recrystallized structure had been obtained and to geta measure of the grain size. Longitudinal and transversesections were cut from the tubes, mounted in epoxy resin,ground with silicon carbide paper and polished with diamondmedia. The polished surfaces were etched in a solutionconsisting of 30 ml HCl, 15 ml HNO3 and 30 ml HF toreveal the grain structure. Samples were examined usingan Olympus IX70 microscope with digital images of themicrostructure recorded at 500x magnification. Grain sizemeasurements were made using the comparison method of


ASTM E112 [Standard Test Methods for Determining AverageGrain Size]. Three longitudinal and three transverse sampleswere taken from the initial 112 µm wall Nb–28Ta–3.5W–1.3Zrtubing. These samples were taken from one tube, at locationsmore than 300 mm apart. One longitudinal sample andone transverse sample was taken from each of the twosubsequent batches. Grain structure and size were recordedfor each.

3.6. Magnetic characterization

Magnetic property testing was carried out on a VibratingSampleMagnetometer (Lake ShoreModel 7404). Tube samplesof 5 mm length were cut, polished, cleaned and weighed.The test involves placing the sample within a uniformmagnetic field. As the sample is vibrated within the field, theinduced voltage is recorded. This voltage is proportional tothe induced magnetic moment in the sample. The magneticfield is incrementally increased from zero up to 12,000 Gauss(1.2 Tesla) and magnetic moment (emu values) recordedfor each field setting. The system records a plot of massnormalized magnetic moment versus applied magnetic fieldstrength and provides the slope of this line as an output.This slope is considered to be a measure of magneticsusceptibility and can be used for relative comparisonsbetweenmaterials. Tests were performed on one sample eachfrom Nb–28Ta–3.5W–1.3Zr tubing, Nb–1Zr tubing and also316L stainless steel tubing as a reference point.

3.7. Demonstration stent manufacture

In order to further assess the potential performanceof Nb–28Ta–3.5W–1.3Zr for stent applications, a quantityof demonstration coronary stents were manufactured.These were produced using conventional technologies oflaser cutting, dross removal and electropolishing. Processparameters were modified, in comparison to manufacture ofstainless steel devices, to account for the highly refractive andchemically resistant nature of the alloy. These stents werecut from the first batch of tubing with outside diameter of2.11 mm and wall thickness of 112 µm.

The design and dimensions of this demonstration stentwere typical for a coronary device, with a nominal expandeddiameter of 3.0 mm, length of 16 mm and a strut wallthickness of 81 µm. A schematic of the unexpanded stentpattern is shown in Fig. 1. Stents were also fabricated fromthe commercial Nb–1Zr material in order to demonstratethe benefits of the new alloy. In addition, stents weremanufactured from the cobalt–chromium alloy L605 in orderto compare performance against this relatively high strengthconventional stent material. (This particular L605 tubing washeat treated to achieve a yield strength of 448 MPa andhad an elastic modulus of 227 GPa.) The same pattern anddimensions were used for the three sets of stents.

3.8. Mechanical testing of demonstration stents

Stents were crimped onto 3.0 mm balloon catheters fortesting. Stent elastic recoil was measured on ten units. Thiswas carried out by initially inflating to nominal diameter andmeasuring the diameter at three locations along the stent.The balloon catheter was then deflated and measurementstaken at the same locations again. Using the average of

each set of three measurements, elastic recoil is thendefined as the percentage change in diameter comparedto original inflated diameter: [Inflated Diameter − DeflatedDiameter]/Inflated Diameter.

Stent compression resistance was measured on five units.Stents were initially expanded to nominal diameter and theballoon catheter removed. Each test stent was then placedinto a compression test fixture which applies a uniformlydistributed compression load to the outside diameter ofthe stent. A plot of force versus diameter was recordedand compression strength is defined as the maximum forcerecorded as the stent is reduced by 15% in diameter. This forcevalue (Newtons) is then normalized to force per unit lengthfor the stent.

4. Results

4.1. Tensile properties

The tensile properties for the initial strip and subsequentstent tubing batches are shown in Table 4. Averages andstandard deviations (SD) are presented. The strip data showsvalues in the range required for a stent material andnotable is the relatively high strength compared to Nb–1Zr.Related to this, the elongation values are correspondingly lowbut overall the properties were promising enough to meritdevelopment of the stent tubing. In particular, establishing anincreased elastic modulus for Nb–28Ta–3.5W–1.3Zr, comparedto Nb–1Zr, was an essential requirement for continueddevelopment.

It can also be seen that the properties of the tubingare in general different to the initial strip sample, with thetubing batches showing slightly lower yield strengths andhigher ductility. This is not unexpected, as the processingroute is different between strip and tubing and the levelsof cold work are not identical. Furthermore, tighter oxygencontamination control would have been possible with thevacuum arc remelted material used for making the tubes.(Oxygen contamination would lead to excessive interstitialstrengthening with a corresponding drop in ductility.) Inany event a better balance between yield strength andductility has been obtained in the tubing. There are alsosome differences evident between the three batches whichcan be attributed to improved process knowledge and controlas more experience was gained with this novel material. Insummary, the Nb–28Ta–3.5W–Zr tubing has approximatelydouble the yield strength of Nb–1Zr and has approximately25% higher elastic modulus.

4.2. Metallography

The metallographic examination revealed the Nb–28Ta–3.5W–1.3Zr material to have a recrystallized grain structureconsisting of relatively fine equiaxed grains. Representativeimages for longitudinal and transverse sections are shownin Fig. 2. The ASTM Grain Size Numbers (G) from each ofthe examined sections are shown in Table 5. In accordancewith this system, higher numbers indicate a finer grain sizeand the measured values which range from 7.6 to 9.6 confirmthat the material has an acceptable grain size range and


Fig. 2 – Metallographic photomicrographs taken at 500x showing recrystallized grain structure on (a) longitudinal and (b)transverse sections of the Nb–28Ta–3.5W–1.3Zr (1.83 mm × 127 µm).

Table 4 – Tensile properties for Nb–28Ta–3.5W–1.3Zr tube and strip compared to Nb–1Zr tube

Material Ultimate tensilestrength (MPa)

0.2% Yieldstrength (MPa)

Elongation (%) Elasticmodulus(GPa)

Nb–28Ta–3.5W–1.3Zr strip Average 476 350 16.7 129SD 16 6 0.6 14

Nb–28Ta–3.5W–1.3Zr tube Average 429 297 25.9 1212.11 mm × 112 µm SD 5 12 1.8 4

Nb–28Ta–3.5W–1.3Zr tube Average 428 342 22.8 1231.83 mm × 127 µm SD 5 3 2.6 5

Nb–28Ta–3.5–W–1.3Zr tube Average 416 339 25.2 1281.83 mm × 152 µm SD 7 9 1.7 6

Nb–1Zr tube Average 290 163 26.5 1022.11 mm × 110 µm SD 1 1 3.6 2

Table 5 – Grain size measurements in longitudinal andtransverse orientations for Nb–28Ta–3.5W–1.3Zr tubing

Material ASTM grain size number, GTube

locationLongitudinal Transverse

Nb–28Ta–3.5W–1.3Zr End 9.6 8.62.11 mm × 112 µm Middle 7.6 8.6

End 9.6 8.6

Nb–28Ta–3.5W–1.3Zr End 9.6 9.61.83 mm × 127 µm

Nb–28Ta–3.5–W–1.3Zr End 8.6 9.11.83 mm × 152 µm

uniformity for stent applications. Fine grains are desiredinherently to provide adequate strength, but also to ensuregood fatigue resistance and ensure that multiple grains arepresent across the small cross section of a stent strut. Havingseveral grains present across a strut reduces the risk of strutfracture.

4.3. Magnetic properties

Data from the magnetic measurements is presented inTable 6. The initial slope of the mass normalizedmagnetic moment (emu/g) versus applied magnetic field

strength (Gauss) plot is a relative measure of the magneticsusceptibility. The susceptibility of Nb–1Zr is lower thanstainless steel by a factor of approximately twenty. This dif-ference corresponds well with the reference data presentedearlier in Table 1 for pure niobium and stainless steel. Whilethe two niobium alloys are in the same order of magnitude,it is worth noting the lower value for the Nb–28Ta–3.5W–1.3Zrcompared to Nb–1Zr. It appears that the Ta, W and Zr alloyingadditions, all of which have low susceptibility values, have re-duced the overall magnetic susceptibility relative to pure nio-bium.

4.4. Stent mechanical performance

The demonstration stents produced had a highly polishedsurface, with well-rounded edges. Fig. 3 shows a typical

Table 6 – Magnetic susceptibility data based on slope ofmoment v field plot

Tubing material Magnetic susceptibility: Momentversus Field slope(memu/g Gauss)

Nb–28Ta–3.5W–1.3Zr 0.0008Nb–1Zr 0.0013316L Stainless steel 0.0267


Fig. 3 – Polished finish of Nb–28Ta–3.5W–1.3Zrdemonstration stent.

Fig. 4 – Elastic recoil data for Nb–28Ta–3.5W–1.3Zrcompared to Nb–1Zr and L605.

example of the stent structure. Elastic recoil data is shownin Fig. 4 for the three sets of stents. The Nb–28Ta–3.5W–1.3Zrstents had an average recoil of 2.2% which comparesfavourably with an average value of 3.8% for Nb–1Zr and 2.2%for the L605 stents. Low values of elastic recoil are desired inorder to minimize sub-optimal stent expansion in the vesseland also to reduce the need for high balloon pressures andstent over-expansions, used to compensate for these lumendiameter losses. The compression strength data for the threesets of stents is shown in Fig. 5. The Nb–28Ta–3.5W–1.3Zrstents had an average compression strength of 0.22 N/mm,compared to 0.17 N/mm and 0.32 N/mm for Nb–1Zr andL605 respectively. The trend in these values correlates wellwith the yield strengths for the materials. Good compressionstrength is required to resist the effects of negative vesselremodeling after angioplasty.

5. Discussion

The interaction of medical devices with the fields presentduring MR imaging has long been a concern, but this issuehas grown in significance in recent years as both the numberof implanted devices and clinical MR procedures has grownrapidly. These interactions may result in device movement,device heating or development of an artifact on the imagebeing collected. The critical nature of these interactions wasprobably first acknowledged in the field of neuroradiology

Fig. 5 – Compression strength data forNb–28Ta–3.5W–1.3Zr compared to Nb–1Zr and L605.

where MR examination of patients with previously implantedintracranial aneurysm clips raised concerns. Wichmann et al.(1997) have investigated how the magnetic susceptibility ofaneurysm clip materials correlated with the resulting imageartifact and demonstrated how titanium clips were saferand resulted in less artifact than cobalt alloy clips. Therapid and widespread introduction of stents has seen manystudies into the safety of these devices, in particular inrelation to movement and heating, though image artifact wastraditionally less of a concern. A study by Teitelbaum et al.(1988) on vena cava filters and early coronary stents wasone of the first to describe the ‘black hole’ artifact createdby ferromagnetism in stent materials, such as 316L stainlesssteel. Since then, the continued maturing of MR angiographyhas seen a number of systematic stent studies. Meyer et al.(2000) assessed a wide range of stents and materials andconcluded that even stents made of nitinol could not bequantifiably imaged within the lumen, due to image artifact.A similar study by Bartels et al. (2001) highlighted that whilematerial magnetic susceptibility plays the major role, eddycurrents induced in the stent by the RF signal of the scanner,also contribute to the overall artifact.

With an increasing awareness of the importance ofmagnetic susceptibility, a number of developments emergedwhere novel stent materials were explored, specifically aimedat reducing the image artifact. Most notable among thesehas been the work of Buecker, Ruebben and colleagues inAachen who developed a Cu–14Au–8Ag–2Pt–1Pd alloy stent.Their first efforts involved implantation of a basic wirewound stent in pig renal arteries with follow-up by MRangiography (Buecker et al., 2002; Spuentrup et al., 2003).The results showed full artifact-free visualization of thevessel, confirming the benefits of the exceptionally lowsusceptibility of this copper-based material. However themechanical properties of this device were not reported andsubsequent work concentrated on more practical designsincluding laser-cut stents, which were implanted in pigcoronary arteries (Buecker et al., 2004; Spuentrup et al., 2005).A number of different coronary MR angiography sequenceswere used to assess the implanted stents and all establishedfull lumen visibility with no artifact. While this approach ispromising, details of the stent mechanical performance arestill not presented, but it can be reasonably assumed that theproperties of a copper-based material would not be sufficient


to achieve practical stent dimensions and performance. Inany event, the vascular compatibility of a copper-basedalloy would also be a significant concern and though theinvestigators indicated that biocompatible coatings wouldbe needed on such devices, there would be significantpractical challenges with this approach. The benefit of apalladium–silver alloy has been explored by Van Dijk et al.(2001) using a basic wire wound stent-type structure. Theadvantage of the material from an MR imaging perspectivewas clear, with the Pd–Ag materials performing favourablycompared to stainless steel, nitinol, cobalt–chromium andtantalum devices. However, the mechanical performanceof this device has again not been addressed and while amoderate value of 110 GPa was indicated for elastic modulus,no value was given for tensile yield strength. The practicalchallenge of making stent tubing and laser-cut stents wasalso not considered. The most successful effort by far hasbeen the development of the Vistaflex platinum biliary stent.While the requirements for biliary stenting may not be aschallenging as for coronary applications, the reduced MRimage artifact has been clearly demonstrated, though notcompletely eliminated, with tantalum devices still beingsuperior (Hagspiel et al., 2005).

With most of these past efforts falling short on demon-strating mechanical performance of these low magnetic sus-ceptibility devices, this current study has placed initial em-phasis on the development of mechanical properties suit-able for coronary devices. In this regard the tensile proper-ties for Nb–28Ta–3.5W–1.3Zr tubing, presented in Table 4, aremost promising with the typical values for yield strength be-ing comparable to 316L stainless steel. Elongation is also atan acceptable level and while elastic modulus falls short of316L values, it is significantly increased compared to com-mercial Nb–1Zr. The single phase equiaxed grain structure,as shown in Fig. 2, supports the fact that the tensile proper-ties are predominantly derived from the expected solid so-lution strengthening of both tantalum and tungsten i.e. noevidence of significant second phases or large dispersoids orprecipitates. The measured ASTM grain size numbers corre-spond approximately with grain sizes in the range 13–19 µm,which is within the range suitable for the manufacture ofstent structures. The recrystallized nature of the grains alsoconfirms that the final anneal step utilized has been suffi-cient to remove cold worked structures, which could lead toreduced ductility.

Having achieved useful mechanical properties with theNb–28Ta–3.5W–1.3Zr tubing, demonstration coronary stents,as shown in Figs. 1 and 3 were successfully manufactured.The mechanical testing of these stents has now shownthat performance levels comparable with that expected ofcommercial stents has been achieved. The exceptionally lowaverage recoil value of 2.2% for the Nb–28Ta–3.5W–1.3Zrstents compares well with the range of 1.5%–16.5% measuredby Barragan et al. (2000) on a variety of commercial devices.The recoil effect in a device is partially related to the elasticmodulus of the material, and the superior recoil performanceof the Nb–28Ta–3.5W–1.3Zr in comparison to the Nb–1Zrstents, could be taken as a reflection of the measureddifferences in modulus. However recoil is also influenced byinteractions between design, work hardening rate and yield

strength and so the fact that the higher modulus of L605 doesnot result in an even lower value for recoil is not surprising.

The average compression strength of 0.22 N/mm forNb–28Ta–3.5W–1.3Zr also compares favourably with respectto other devices. Comparing against the published literatureis less useful in this instance as there is a wide variety oftest methods, so comparison with the Nb–1Zr and L605 testunits is more informative. The lower compression strengthof Nb–1Zr stents (0.17 N/mm) and the higher compressionstrength of L605 stents (0.32 N/mm) can be taken as areflection of the yield strength differences between all thesematerials. As the L605 material had been annealed toa relatively high strength of 448 MPa, the correspondingcompression strength of 0.32 N/mm can be taken as beingtowards the upper end of what would be normally expectedin clinical devices. Similarly, the lower value of 0.17 N/mmfor Nb–1Zr is likely to be at the lower end, but still veryacceptable. In this regard it is worth noting that recent workby Beier et al. (2006) used Nb–1Zr coronary stents, of similardimensions, for the first time in a human clinical study.This study showed that at 6 months follow-up, the stainlesssteels controls and the Nb–1Zr units had similar levels ofin-stent restenosis, but the Nb–1Zr stents were trendingtowards having more neointimal in-growth. Interestingly, theinvestigators have postulated that the increased neointimaformation may be attributable to the thicker struts of theNb–1Zr stents (110 µm) compared to the stainless steel units(95 µm). This points to an immediate demonstration ofone possible advantage of the Nb–28Ta–3.5W–1.3Zr material;with up to double the yield strength value of Nb–1Zr, itshould allow for easy reduction of strut thickness whilestill retaining, and increasing if necessary, the compressionstrength of the device. (For example, the strut thicknessof the demonstration stent tested here was 81 µm.) Thisreduction in thickness could then facilitate less vessel walldamage and subsequent neointima growth. It is worth notingthat this study by Beier and co-workers was focused on thepotential advantage of niobium over stainless steel in termsof reducing restenosis rate; through elimination of nickel andmolybdenum, which are present in stainless steel. The studywas not assessing MR imaging, but the Nb–1Zr should indeedalso have reduced image artifact in comparison to stainlesssteel.

The magnetic susceptibility of the Nb–28Ta–3.5W–1.3Zrmaterial, as shown in Table 6, is in the desired region andcompares even better than expected against Nb–1Zr. In lightof the successful imaging behaviour of stents made fromtantalum, platinum and palladium alloys as described earlier,the Nb–28T–3.5W–1.3Zr should be at least equivalent sinceits magnetic susceptibility is in the same range as thesematerials. This will be confirmed by carrying out an MR imageartifact study, which will be reported in a future publication.

The biocompatibility of the Nb–28Ta–3.5W–1.3Zr materialhas yet to be investigated but the constituents of the alloyhave been selected with significant consideration of thisaspect. Niobium itself is a refractory metal which very readilyforms a stable and protective oxide layer; the stability ofoxide layers has long been considered as a useful indicatorof a material’s biocompatibility. Zitter and Plenk (1987)included niobium in their study of several electrochemical


parameters and described how the low current densitymeasured for niobium, compared to gold, stainless steeland cobalt–chromium, was a strong indicator of its goodbiocompatibility. Metikos-Hukovic et al. (2003) examined howvarious oxide films performed in physiological solutions,including oxides of pure niobium and oxides on titaniumalloys with niobium additions. This work further confirmedthe stability of niobium and also indicated that niobiumadditions to Ti–Al–V alloys improved the oxide stability of thealloy. Johansson and Albrektsson (1991) have reported on astudy of orthopaedic screw implants in rabbits and concludedthat niobium performed equivalent to titanium with nohistological differences between the two materials. Matsunoet al. (2001) have implanted a number of refractory metals,including niobium and tantalum, in both subcutaneous softtissue and in bone marrow of rats. Histological studiesshowed no inflammatory response for these two metalsand no dissolution of metal into surrounding tissue wasdetected. Eisenbarth et al. (2004) also investigated thebiocompatibility of both niobium and tantalum, as wellas zirconium. This included cell viability and proliferationstudies using bovine aortic endothelial cells, which istherefore of direct relevance to the Nb–28Ta–3.5W–1.3Zrstent material. The results showed that cell growth andproliferation was better on niobium and tantalum comparedto a titanium reference surface, while zirconium did showmarginally reduced activity. Though, of note is the factthat zirconium still performed better than 316L stainlesssteel, suggesting that the low level of 1.3% Zr in this newalloy should still provide for high biocompatibility. Finallyconsidering the biocompatibility of tungsten, some of themost relevant works have been performed by Peuster andco-workers to assess the degradation and compatibility oftungsten aneurysm occlusion coils. An initial rabbit implantstudy showed that even though the pure tungsten coilscorroded, the increased levels of tungsten detected in blooddid not translate to a toxic response, either locally in theimplanted vessel or systemically in any of the major organs(Peuster et al., 2003a). A further in vitro study exposedcells from human pulmonary arteries to various levels oftungsten in the cell growth medium (Peuster et al., 2003b).Measurements of the corrosion rate of tungsten coils werealsomade and the study concluded that tungsten levels muchhigher than that caused by the corrosion would be neededto induce toxic effects in the cells. Though this study wasfor pure tungsten in a different vascular application, it doessuggest that the level of tungsten in the Nb–28Ta–3.5W–1.3Zrstent material should not cause biocompatibility problems.This will however be evaluated and reported in a futurepublication.

6. Conclusions

A new Nb–28Ta–3.5W–1.3Zr alloy has been developedwith mechanical properties superior to commercial Nb–1Zrmaterials. The improved levels of yield strength andelastic modulus, combined with low magnetic susceptibilitymake this alloy a practical candidate as a stent material,where reduced MR image artifact is critical. The promising

performance of this new alloy has been demonstratedon experimental coronary stents which showed highlyacceptable levels for critical stent performance attributessuch as compression strength and elastic recoil. Bearing inmind clinical applications where MRI is already widely used,the material may also be of significant advantage in carotidand intracranial stenting. Further evaluations will be carriedout to confirm stent MR imaging behaviour and to assesspotential long term in vivo behaviour through corrosion andbiocompatibility testing.

Acknowledgements

The authors thank Steve Larsen and Matt Shedlov of theBoston Scientific Corporation (Maple Grove, MN, USA) for theirwork on stent process development. Thanks are also due toDennis Boismier of the Boston Scientific Corporation (MapleGrove, MN, USA) for input on stent testing.

R E F E R E N C E S

ASMHandbook, Tenth ed. Vol. 2. Properties and Selection: Nonfer-rous Alloys and Special-Purpose Materials. ASM International1990. ISBN 0-87170-378-5.

Barragan, P., Rieu, R., Garitey, V., Roquebert, P., Sainsous, J.,Silvestri, M., Bayet, G., 2000. Elastic recoil of coronary stents:A comparative analysis. Catheterization and CardiovascularInterventions 50, 112–119.

Bartels, L.W., Smits, H., Bakker, C., Viergever, M., 2001. MRimaging of vascular stents: Effects of susceptibility, flowand radiofrequency eddy currents. Journal of Vascular andInterventional Radiology 12, 365–371.

Beier, F., Gyongyosi, M., Raeder, T., von Eckardstein-Thumb, E.,Sperker, W., Albrecht, P., Spes, C., Glogar, D., Mudra, H.,2006. First in-human randomized comparison of an anodizedniobium stent versus a standard stainless steel stent. ClinicalResearch in Cardiology 95, 455–460.

Buecker, B., Spuentrup, E., Ruebben, A., Guenther, R., 2002.Artifact-free in-stent lumen visualization by standard mag-netic resonance angiography using a new metallic magneticresonance imaging stent. Circulation 105, 1772–1775.

Buecker, A., Spuentrup, E., Ruebben, A., Mahnken, A., Nguyen,T.H., Kinzel, S., Guenther, R., 2004. New metallic MR stents forartifact-free coronary MR angiography: Feasibility study in aSwine model. Investigative Radiology 39, 250–253.

Davies, J.R. (Ed.), 1994. Stainless Steels — ASM SpecialityHandbook. ASM International, ISBN: 0-87170-503-6.

DiStefano, J.R., Pint, B.A., DeVan, J.H., 2000. Oxidation of refractorymetals in air and low pressure oxygen gas. InternationalJournal of Refractory Metals & Hard Materials 18, 237–243.

Eisenbarth, E., Velten, D., Mueller, M., Thull, R., Breme, J., 2004.Biocompatibility of β-stabilizing elements of titanium alloys.Biomaterials 25, 5705–5713.

Ehara, M., Surmely, J., Kawai, M., Katoh, O., Matsubara, T.,Terashima, M., Tsuchikane, E., Kinoshita, Y., Suzuki, T., Ito, T.,Takeda, Y., Nasu, K., Tanaka, N., Murata, A., Suzuki, Y., Sato, K.,Suzuki, T., 2006. Diagnostic accuracy of 64-slice computed to-mography for detecting angiographically significant coronaryartery stenosis in an unselected consecutive patient popula-tion. Circulation Journal 70, 564–571.


Hagspiel, K.D., Leung, D.A., Nandalur, K., Angle, J., Dulai, H.,Spinosa, D., Matsumoto, A., Christopher, J., Ahmed, H.,Berr, S., 2005. Contrast-enhanced MR angiography at 1.5Tafter implantation of platinum stents: In vitro and invivo comparison with conventional stent designs. AmericanJournal of Roentgenology 184, 288–294.

Hubbell, J.H., Seltzer, S.M., Tables of X-Ray Mass AttenuationCoefficients and Mass Energy-Absorption Coefficients (version1.4). [Online] Available: http://physics.nist.gov/xaamdi [2007,August 28]. National Institute of Standards and Technology,Gaithersburg, MD, 2004.

Hug, J., Nagel, E., Bornstedt, A., Schnackenburg, B., Oswald, H.,Fleck, E., 2000. Coronary arterial stents: Safety and artifactsduring MR imaging. Radiology 216, 781–787.

Johansson, C.B., Albrektsson, T.A., 1991. A removal torque andhistomorphometric study of commercially pure niobium andtitanium implants in rabbit bone. Clinical Oral ImplantsResearch 2, 24–29.

Lenhart, M., Volk, M., Manke, C., Nitz, W.R., Strotzer, M.,Feuerbach, S., Link, J., 2000. Stent appearance at contrast-enhanced MR angiography: In vitro examination with 14stents. Radiology 217, 173–178.

Long, A.L., Sapoval, M.R., Beyssen, B.M., Auguste, M.C., Le Bras,Y., Raynaud, A.C., Chatellier, G., Gaux, J.C., 1995. Strecker stentimplantation in iliac arteries: Patency and predictive factorsfor long-term success. Radiology 194, 739–744.

Mani, G., Feldman, M.D., Patel, D., Agrawal, C.M., 2007. Coronarystents: A materials perspective. Biomaterials 28, 1689–1710.

Matsuno, H., Yokoyama, A., Watari, F., Uo, M., Kawasaki, T.,2001. Biocompatibility and osteogenesis of refractory metalimplant, titanium, hafnium, niobium, tantalum and rhenium.Biomaterials 22, 1253–1262.

Metikos-Hukovic, M., Kwokal, A., Piljac, J., 2003. The influenceof niobium and vanadium on passivity of titanium-basedimplants in physiological solution. Biomaterials 24, 3756–3775.

Meyer, J.M.A., Buecker, A.S., chuermann, K., Ruebben, A.,Guenther, R., 2000. MR evaluation of stent patency: In vitrotests of 22 metallic stents and the possibility of determiningtheir patency by MR angiography. Investigative Radiology 35,739–746.

Niinomi, M., 2008. Mechanical biocompatibilities of titaniumalloys for biomedical applications. Journal of the MechanicalBehavior of Biomedical Materials 1, 30–42.

Nitatori, T., Hanaoka, H., Hachiya, J., Yokoyama, K., 1999. MRIartifacts of metallic stents derived from imaging sequencingand the ferromagnetic nature of materials. Radiation Medicine4, 329–334.

Ong, A.T.L., Serruys, P.W., 2005. Technology insight: An overviewof research in drug-eluting stents. Nature Clinical PracticeCardiovascular Medicine 12, 647–658.

Peuster, M., Fink, C., Wohlsein, P., Bruegmann, M., Guenther, A.,Kaese, V., Niemeyer, M., Haferkamp, H., von Schnakenburg,C., 2003a. Degradation of tungsten coils implanted intothe subclavian artery of New Zealand white rabbits is notassociated with local or systemic toxicity. Biomaterials 24,393–399.

Peuster, M., Fink, C., von Schnakenburg, C., 2003b. Biocompatibil-ity of corroding tungsten coils: In vitro assessment of degrada-tion kinetics and cytotoxicity on human cells. Biomaterials 24,4057–4061.

Poncin, P., Proft, J., 2003. Stent tubing: Understanding the desiredattributes. In: Proc. Materials & Processes for Medical

Devices Conference, 8–10 Sept 2003, Anaheim, CA, US. ASMInternational, pp. 253–259.

Reifart, N., Morice, M., Silber, S., Benit, E., Hauptmann, K.,de Sousa, E., Webb, J., Kaul, U., Chan, C., Thuesen, L.,Guagliumi, G., Cobaugh, M., Dawkins, K., 2004. The NUGGETstudy: NIR ultra gold-gilded equivalency trial. Catheterizationand Cardiovascular Interventions 62, 18–25.

Rofsky, N.M., Adelman, M.A., 2000. MR angiography in theevaluation of atherosclerotic peripheral vascular disease.Radiology 214, 325–338.

Schenck, J.F., 1996. The role of magnetic susceptibility in magneticresonance imaging: MRImagnetic compatibility of the first andsecond kinds. Medical Physics 6, 815–850.

Schoenhagen, P., Halliburton, S., Stillman, A., Kuzmiak, S., Nissen,S., Tuzcu, E.M., White, R., 2004. Non-invasive imaging ofcoronary arteries: Current and future role of multi-detectorrow CT. Radiology 232, 7–17.

Slining, J.R., Koss, D.A., 1973. Solid solution strengthening of highpurity niobium alloys. Metallurgical Transactions 5, 1261–1264.

Spuentrup, E., Ruebben, A., Stuber, M., Guenther, R., Buecker,A., 2003. Metallic renal artery MR imaging stent: Artifact-freeLumen visualization with projection and standard renal MRangiography. Radiology 227, 897–902.

Spuentrup, E., Ruebben, A., Mahnken, A., Stuber, M., Koelker,C., Nguyen, T.H., Guenther, R., Buecker, A., 2005. Artifact-Free coronary magnetic resonance angiography and coronaryvessel wall imaging in the presence of a new, metallic,coronary resonance imaging stent. Circulation 111, 1019–1026.

Stoeckel, D., Pelton, A., Duerig, T., 2004. Self-expanding nitinolstents: Material and design considerations. European Radiol-ogy 14, 292–301.

Teitelbaum, G.P., Bradley, W.G., Klein, B.D., 1988. MR imaging ar-tifacts, ferromagnetism, and magnetic torque of intravascularfilters, stents and coils. Radiology 166, 657–664.

Vaishnav, S., Aziz, S., Layton, C., 1994. Clinical experience with theWiktor stent in native coronary arteries and coronary bypassgrafts. British Heart Journal 72, 288–293.

Van Dijk, C., van Holten, J., van Dijk, B.P., Matheijssen, N.,Pattynama, P., 2001. A precious metal alloy for constructionof MR imaging-compatible balloon expandable vascular stents.Radiology 219, 284–287.

Wichmann, W., Von Ammon, K., Fink, U., Weik, T., Yasargil,G.M., 1997. Aneurysm clips made of titanium: Magneticcharacteristics and artifacts in MR. American Journal ofNeuroradiology 18, 939–944.

Wiskirchen, J., Kraemer, K., Koenig, C., Kramer, U., Truebenbach,J., Wersebe, A., Tepe, G., Dietz, K., Claussen, C., Duda, S.,2004. Radiopacity of current endovascular stents: Evaluationin a multiple reader phantom study. Journal of Vascular andInterventional Radiology 15, 843–852.

Wojcik, C.C., 1998. High temperature niobium alloys. AdvancedMaterials & Processes 12, 27–30.

Woodward, P.K., Li, D., Zheng, J., Haacke, E.M., Gropler, R.J., 2000.Coronary MR angiography. Applied Radiology Supplement 3,55–64.

Zhou, Y.L., Niinomi, M., Akahori, T., 2004. Effects of Ta content onYoung’s modulus and tensile properties of binary Ti–Ta alloysfor biomedical applications. Materials Science and EngineeringA 371, 283–290.

Zitter, H., Plenk, H., 1987. The electrochemical behavior of metallicimplant materials as an indicator of their biocompatibility.Journal of Biomedical Materials Research 21, 881–896.

http://physics.nist.gov/xaamdi

Development of a new niobium-based alloy for vascular stent applications

Documents

Transcript of Development of a new niobium-based alloy for vascular stent applications