The magnet-magnet coupling system is a very important aspect of a centrifugal pump with regard to thrombus formation or hemolysis and bearing wear. 1,2 A strong magnetic force jerks the impeller down into the bottom bearing and results in thrombus formation or bearing wear. Conversely, a weak magnetic force results in decoupling of the combination of impeller and actuator and produces a decline in efficiency. Consequently, proper magnet balance is very important for a long-term, clinically applied centrifugal pump.
The impeller floating phenomenon can be observed in a pivot bearing supported pump. 3,4 The condition in which the impeller floating phenomenon occurs is the proper magnet balance. This phenomenon is affected by the magnet-magnet (impeller-actuator) coupling distance and rotational speed of the impeller. The aim of this study was to map the impeller floating phenomenon on a pump performance curve and thus achieve proper magnet balance during pumping.
Materials and Methods
A completely sealless Gyro centrifugal pump was used in this study. This pump, from the permanently implantable (PI) series, is being developed as a totally implantable artificial heart. 5,6 This pump is composed of the housing, impeller, and actuator. The pump has ultrahigh molecular weight polyethylene female bearings and alumina ceramic male bearings. Six strong neodymium-iron-boron composite magnets (diameter 10 mm) were installed inside the impeller and actuator. The direct current (DC) brushless motor (Softronics Co., Ltd., Urawa City, Saitama, Japan) was used to drive the pump. Two combinations of impellers and actuators were used for this study: an impeller with a magnetic flux of 3.4 kg at the surface of the impeller and an actuator magnetic flux of 2.36 kg at the surface of the actuator and an impeller with a magnetic flux of 3.3 kg at the surface of the impeller and an actuator with a magnetic flux of 2.12 kg at the surface of the actuator. This study used an acrylic pump, which was identical to the PI pump, to observe the impeller floating phenomenon. A schematic of the experimental setup is shown in Figure 1. These studies were performed with a glycerin–water (37% water) solution. To observe the impeller floating phenomenon, the distance between the top and bottom female bearing was lengthened by 1 mm. The minimal magnet coupling distance (MCD) (Figure 1) between the impeller magnet and actuator magnets was designed to be 9 mm. The MCD was altered from 9 mm (standard condition) to 12.5 mm by means of an acrylic spacer, which was installed between the pump and actuator. The floating phenomenon was observed in Figure 2. Figure 2a demonstrates that the impeller male bearing comes into contact with the bottom female bearing (bottom contact mode). Figure 2b demonstrates that the impeller male bearing is kept apart from the bottom female bearing and comes into contact with the top female bearing (top contact mode). We investigated the relationship between pump flow and pressure as a parameter of the impeller rotational speed (pump performance curve). The impeller floating phenomenon was mapped on the pump performance curve.
Figure 3 shows the results when using one combination of impeller (magnetic flux 3.4 kg) and actuator (magnetic flux 2.36 kg). As indicated in Figure 3, the results were plotted on a flow pressure curve. Figure 3 illustrates top contact area, top contact point, hysteresis area, bottom contact point, and bottom contact area. The MCD was altered from 9 mm (standard condition) to 12.5 mm by a spacer. At a low impeller rotational speed, the impeller position was in the bottom contact mode. When the rotational speed was increased, the impeller floating phenomenon appeared at the top contact point. At a high impeller rotational speed, the impeller position was in the top contact mode. When the rotational speed was decreased, the impeller floating phenomenon disappeared at the bottom contact point. The area surrounded by the top contact point and bottom contact point was the hysteresis area. In this area, the magnetic forces jerk the impeller down into the bottom contact mode, and hydromechanical forces cause the impeller float to balance. As indicated in Figure 3a (MCD 9 mm) through Figure 3f (MCD 12.5 mm), as the MCD increased, the impeller rotational speed at which the floating phenomenon appeared decreased. Consequently, the top contact area was extended, and the bottom contact area was reduced. However, the size of the hysteresis area remained unchanged. At an MCD of 9 mm (Figure 3a), the ordinary left ventricular assist device (LVAD) driving condition of 5 L/min against 100 mm Hg occurred in the hysteresis area. When MCD was 11 mm or higher, the LVAD driving condition occurred in the top contact area. Therefore, the proper spacers can be selected for the left pump (2.5 mm spacer [MCD 11.5 mm]) and right pump (3 mm spacer [MCD 12 mm]).
During pumping, the working forces on the impeller are magnetic and are caused by the magnetic coupling, whereas the hydromechanical forces are caused by differential pressure. The magnetic force that jerks the impeller down into the bottom bearing is calculated as EQUATION
where m1 and m2 are magnetic flux, μ0 is magnetic permeability, and r is MCD.
The attractive forces of one combination of the impeller (magnetic flux 3.4 kg) and actuator (magnetic flux 2.36 kg) and another combination of impeller (magnetic flux 3.3 kg) and actuator (magnetic flux 2.12 kg) are estimated in Figure 4. When MCD increased, the force decreased. The force of the combination of impeller (magnetic flux 3.4 kg) and actuator (magnetic flux 2.36 kg) at an MCD of 12 mm is equal to the force of another combination of impeller (magnetic flux 3.3 kg) and actuator (magnetic flux 2.12 kg) at an MCD of 11 mm. Figure 5 shows the measurement results for the impeller floating phenomenon in each condition. The top contact area, the hysteresis area, and the bottom contact area are located in almost the same position, as shown in Figure 5. The impeller position, (i.e., the impeller floating phenomenon) can be estimated using the measured magnetic forces and MCD, and this phenomenon can be regulated by altering the MCD. Because this pivot bearing supports the pump, the impeller position is very important. Durability of the ultrahigh molecular weight polyethylene female bearing 1 is affected by wear caused by an imbalance of the working forces on the impeller. To reduce wear, the working forces on the impeller should be balanced, and the friction between the male and female pivot bearing will thus decrease. When the impeller is in the top contact position, the clearance between impeller bottom and bottom housing increases. Therefore, blood flow in this clearance will be promoted, 7 the stagnant area of the pump will disappear, and a condition for antithrombogenicity will be improved. Accordingly, the impeller position has to be in the top contact mode. However, the driving condition of the left pump is different from that of the right pump. Therefore, two kinds of pumps are generally necessary for total BiVAD artificial heart implantation. However, the same pump can be useful in a variety of conditions (right and left ventricular assist) by changing the thickness of the spacer. When the pump is maintained at the proper magnet balance condition, the floating impeller phenomenon occurs automatically in response to the impeller revolution. This is called “the dynamic revolution per minute suspension.”
Maintaining the proper MCD and impeller position in the top contact mode is very important for a long-term, clinically applied centrifugal pump. The impeller position (top contact mode or bottom contact mode) is very important for a pivot bearing supported centrifugal pump in regard to thrombus formation or hemolysis and bearing wear. In this study, the impeller floating phenomenon was investigated in detail, and we demonstrated that the impeller position was regulated by changing the MCD with a spacer. It became clear that the impeller floating phenomenon could be estimated and that the same pump could cope with different conditions (right and left ventricular assist) by changing the thickness of the spacer.
1. Nakazawa T, Makinouchi K, Takami Y, et al: The effect of impeller-driver magnet coupling distance on hemolysis in a compact centrifugal pump. Artif Organs 20: 252–257, 1996.
2. Makinouchi K, Nakazawa T, Takami Y, Takatani S, Nosé; Y: Evaluation of the wear of the pivot bearing in the Gyro C1E3 pump. Artif Organs 20: 523–528, 1996.
3. Nakazawa T, Takami Y, Makinouchi K, et al: Hydraulic assessment of the floating impeller phenomena in a centrifugal pump. Artif Organs 21: 78–82, 1997.
4. Ohara Y, Takatani S, Nosé; Y: In vitro analysis of the floating mode phenomenon in the free impeller Gyro pump. Artif Organs 19: 635–638, 1995.
5. Yoshikawa Y, Nonaka K, Linneweber J, et al: Development of the NEDO implantable ventricular assist device with Gyro centrifugal pump. Artif Organs 24: 459–467, 2000.
6. Takami Y, Andrade A, Nakazawa T, et al: Eccentric inlet port of the pivot bearing supported Gyro centrifugal pump. Artif Organs 21: 312–317, 1997.
7. Ohara Y, Murase M, Nosé; Y: Hemolytic effect of the secondary vane incorporated into the back side of the impeller. Artif Organs 21: 694–699, 1997.