An essential procedure
The implantation of stents into arteries using balloon angioplasty has become an accepted technique in the treatment of arterial disease, where constricted or partially blocked blood vessels are involved. Stents are typically laser-cut from metals such as stainless steel or nitinol. The size of a coronary stent is in the order of 1.5–3 mm in diameter and 15–30 mm in length. Brain stents are much smaller and leg stents are much larger. Once implanted at the site of an arterial lesion, a stent’s role is to keep the blood vessel open and permit the normal flow of blood through the artery.
The flexibility of a stent is essential to its performance.The most common method employed for implantation is a balloon catheter. When the target site is reached, the balloon is inflated by pressurising it through the hollow catheter tube, which, in turn causes the stent to expand so that it pushes open the artery wall at the site. The balloon is then deflated and the tube is removed. Obviously, this implantation technique requires the stent to be flexible. The operating environment is equally important. Arteries are flexible and are constantly subjected to the forces generated by the pulsating pumping action of the heart. Thus, the stent must have the capability to react dynamically in a manner similar to its host environment.
Stenting has gained considerable favour in the treatment of arterial disease because it is relatively noninvasive. However, restenosis, that is the build up of scar tissue round the stent, which causes re-blockage, has been proven to occur approximately 25% of the time with bare-metal stents. This problem has been addressed through the use of stents coated with drugs that inhibit cell formation. Usually these drugs are embedded in a polymer, either a biodegradable type such as polylactide co-glycocide or a nonbiodegradable polyurethane-based elastomer. The major characteristics that these coatings must exhibit are
The polymer/drug matrix allows the drug to be released (eluted) slowly, over a period of several months, thereby inhibiting the formation of cell masses, and hence significantly reducing the probability of restenosis. Clinical trials using drug-eluting stents have shown that the occurrence of restenosis could be reduced dramatically compared with a rate in the order of 25% for uncoated stents.
Spray coating stents
Spray nozzles are the primary method used to apply coatings to stents. Other techniques such as vacuum and super-critical CO2 deposition are methods under development, but are not in commercial use. Dip coating is not a viable alternative because of tight stent geometry. The spacing between struts, particularly where they join each other, can be as small as 50 µm. Dip coating would result in the bridging of these areas, something that is not allowed, because the bridged coating could eventually break off and enter the blood stream. Even spray-coated stents can exhibit bridging if not processed properly.
Figure 1b. Coating showing significant webbing.
Coating a stent properly is a technological challenge. Figures 1a to 1c illustrate what can occur. Figure 1a depicts a perfectly coated stent. It features a smooth, continuous coating, without any webbing or surface defects. In Figure 1b, the coating shows significant webbing. In Figure 1c, the coating exhibits an irregular, bumpy surface caused by the drops drying before they reach the target. It is important that the coating is uniform (both inside and outside), free from holes or other defects, and reproducible with a few percent weight gain from one stent to the next.
Figure 1c. Coating shows irregular, bumpy surface.
The industry has adopted two types of spray nozzles, ultrasonic and twin-fluid nozzles. The focus in this article is on ultrasonic nozzle technology, which is currently perceived to be the best method for achieving the desired results.
Figure 2. One configuration of a stent-coating system.
The operating requirements in this application call for flow rates in the order of 20–100 mL/min, spray diameters in the range of 0.5–2 mm, and small median drop diameters. Ultrasonic nozzles are capable of meeting these requirements. Figure 2 shows one coating strategy. The stent is placed on a mandrel that is attached to a rotating shaft. The nozzle is mounted above the stent. The shaft not only rotates, but also translates so that the stent is sprayed along its entire length. Typically, several traverses are required to achieve the proper coating weight.
The sprayed liquid consists of the polymer/drug system dissolved in a suitable organic solvent and diluted to approximately 0.5-2% by weight. Typically, high vapour pressure solvents are used, so that drying occurs quickly. Repeated traverses, coupled with low flow rates, produce the best coatings and maximise efficiency of material transfer. By varying rotational speed, distance between the spray and the stent and the number of traverses (and therefore flow rate), a process can be optimised. Often, stent coating is done in a nitrogen environment. Nitrogen promotes better liquid flow characteristics because it acts to lower the surface tension of the liquid as it contacts the stent surface.
Figure 3. Capillary waves.
Apart from the operational parameters described above, the key to successful coatings is specifying an appropriate type of ultrasonic nozzle. First, it is useful to briefly describe ultrasonic atomisation. When a liquid film is placed on a smooth surface and set into vibrating motion so that the direction of vibration is perpendicular to the surface, the liquid absorbs some of the vibrational energy, which is transformed into standing waves. These waves, known as capillary waves, form a rectangular grid pattern on the surface of the liquid with regularly alternating crests and troughs extending in both directions as shown in Figure 3.
When the amplitude of the underlying vibration is increased, the amplitude of the waves increases correspondingly, that is, the crests become taller and the troughs deeper. A critical amplitude is ultimately reached, at which point the height of the capillary waves exceeds that required to maintain their stability. The result is that the waves collapse and drops of liquid are ejected from the tops of the degenerating waves perpendicular to the atomising surface. The median drop diameter (DN,0.5) expected is inversely proportional to the vibration frequency (f) to the two-thirds power. This relationship was discovered in the late 19th century by Lord Rayleigh.1 Specifically,
DN,0.5 ~ (8πσ/ρf2)1/3 (1)
where σ is surface tension of the liquid and ρ is its density. Experimental studies by Lang and others2,3 further quantified this relationship. Lang established that the proportionality constant is 0.34, so that
DN,0.5 = 0.34(8πσ/ρf2)1/3 (2)
Figure 4. Cross-section of a typical ultrasonic nozzle.
A typical ultrasonic nozzle is shown in cross-section in Figure 4. Disc-shaped ceramic piezoelectric transducers convert high frequency electrical energy from an oscillator/amplifier into vibratory mechanical energy at the same frequency. The transducers are sandwiched between two titanium cylinders, which act to concentrate and amplify the vibration amplitude at the atomising surface. Titanium is used because of its good acoustical characteristics, corrosion-resistance and high strength.
One of the principal attributes of ultrasonic nozzles results from the atomisation being solely a surface phenomenon. The amount of liquid atomised depends exclusively on the rate at which liquid is introduced onto the surface.
Therefore, in principle, ultrasonic nozzles have infinite variability with respect to flow rate. Although practical considerations related to nozzle design limit this variability to a fixed amount, the ability to precisely adjust flow rates by adjusting the rate at which liquid is delivered to the nozzle is extremely useful.
The other major attribute of ultrasonic nozzles, which distinguishes these devices from all other atomising devices, is the low velocity character of the spray, typically 0.25–0.4 m/s compared with 10–20 m/s for standard pressure atomising nozzles. This 100-fold reduction in spray velocity is equivalent to a 10 000 times reduction in kinetic energy.
The liquid is delivered to the atomising surface through a feed tube that runs the length of the nozzle. The relatively large liquid feed orifice ensures freedom from clogging. Alternatively, liquid can be delivered to the atomising surface externally.
Ultrasonic nozzles for stent coating
There are two principal ultrasonic nozzle designs that are in use in coating stents. Both rely on the use of a low-pressure gas stream to shape the slow-moving drops into a narrow spray beam. The operating frequency of both designs is 120 kHz, which for water yields a value for DN,0.5 in the order of 18 μm.
Figure 5. Nozzle A — internal liquid feed.
The first design (nozzle A) is represented in Figure 5. Compressed gas, typically at 1 psi, is introduced into the diffusion chamber of the gas shroud, which produces a uniformly distributed flow of air around the nozzle stem. The ultrasonically produced spray at the tip of the stem is immediately entrained in the gas stream. An adjustable focussing mechanism on the gas shroud allows complete control of spray width.
The spray envelope is conical. Moving the focus-adjust mechanism in and out controls the width of the bow. The distance between nozzle and substrate can be varied from near contact to approximately two inches (50 mm). The narrowest beam diameter achievable at the focal plane is approximately 1.75 mm.
Figure 6. Nozzle B — external liquid feed.
The second design (nozzle B) is depicted in Figure 6. It consists of feeding liquid to the atomising surface through an externally mounted cannula. The gas is fed through the nozzle orifice. The gas stream draws the spray to it, creating a narrow spray beam of 0.5 mm. The distinguishing feature of this technique is that the liquid feed is external. This means that the liquid is completely isolated from nozzle vibrations until the time that atomisation occurs. This allows a high degree of spray stability and a greater degree of reproducibility from one spray cycle to the next.
In addition to displaying different diameter spray beams, the two nozzle designs described above produce different results in terms of drop size distribution and the properties of the liquid being sprayed. A series of experiments were conducted using a Malvern Spraytec particle size analyser (Malvern Instruments Ltd, www.malvern.co.uk) to obtain distributions. Figure 7 summarises the results. Four time-averaged distributions are shown, two for nozzle A (one for water and the other for isopropyl alcohol), and the corresponding two for nozzle B. Both the number distribution and the volume distribution are plotted for each case. The flow rate was 100 μL/min.
Figure 7. Drop-size and distribution analysis.
The first observation is that the median drop sizes agree with theoretical predictions of Equation 2: DN,0.5 = 18 µm. The values of σ and ρ for alcohol are different from those for water (σwat = 73 dynes/cm and σalc = 22 dynes/cm; ρwat = 1 and ρalc = 0.8). The resulting median drop size for alcohol is approximately 72% that for water, or approximately 13 μm. The distributions for water appear to closely follow a lognormal distribution shape, whereas for alcohol, the distribution is noticeably skewed toward larger diameters, particularly for nozzle B. Moreover, in each case, the spread of the distributions for alcohol is considerably greater than that for water. In addition, the number and volume distributions for each nozzle spraying water are similar; they nearly overlap. The same is not true when spraying alcohol.
These observations suggest that there is significant coalescence occurring, particularly for alcohol. This can be explained by the lower surface tension of alcohol compared with that of water. The lower surface energy promotes coalescence.
Because the diameter of the spray beam of nozzle B is smaller than that of nozzle A (0.5 mm versus 1.75 mm), the spatial density of spray from nozzle B is considerably greater than from nozzle A. This difference explains why, for the same liquid, nozzle B exhibits more coalescence than nozzle A.
Effects of drop size
Drop sizes have an impact on stent coatings in that size influences the solvent evaporation rate, and hence the drying time. It is not possible to predict how a particular polymer/drug/solvent system will behave during the coating process because the interaction of the three-part solution with the stent surface is relatively complex. In general terms, an abundance of large drops is acceptable if the drying rate is rapid, which occurs for solvents with high vapour pressures such a tetrahydrofuran or chloroform. This is not the case for low vapour pressure solvents such as dimethylacetamide.
The width of the distribution should be examined to determine whether it plays a role in coating behaviour. The sharpness of the distribution(ΔV) is defined as the difference between the diameter (d) for which 90% of the volume of drops in a sample have diameters smaller than that diameter (d90%), less the diameter for which 10% of the volume of drops meet that condition (d10%).
Table I. Values associated with Figure 7.
Table I shows that the values of ΔV for both nozzles are similar for a given liquid, thus this characteristic may not be significant.
Selecting the right design
Coating arterial stents is a challenging application. Ultrasonic spray nozzles are ideally suited to the application because they are capable of producing low flow rates, precisely shaped spray patterns, low-velocity delivery, and relatively small drops. The techniques used in coating may vary, as may the drug/polymer system. These variances make it important to optimise the entire system for each specific circumstance. The two nozzles described in this article display different spray properties, but both play a role in stent coating. The selection of which one to use depends on the exact nature of the process and the materials being sprayed.
1. J.W. Strutt (Lord Rayleigh), The Theory of Sound, 349, pp. 82, Dover Publications Inc. (1945).
2. W.R. Wood and A.L. Loomis,Phil. Mag., 7,417 (1927).
3. R.J. Lang, Ultrasonic Atomisation of Liquid, Jour. Acoustical Soc. of America, 34:1 (1962)
Dr Harvey L. Berger is Chief Technology Officer at Sono-Tek Corporation, 2012 Route 9W, Milton, New York 1254, USA, tel. +1 845 795 2020,