Secondary Logo

Journal Logo

Invited Commentaries

Will We See Nuclear-Powered Ventricular Assist Devices?

Poirier, Victor

Author Information
doi: 10.1097/MAT.0b013e31826e3ee6
  • Free

In the search for an ideal ventricular assist system, we are always led to the concept of a nuclear-powered system. Yes indeed, on the surface, the benefit of an implantable energy source with a half-life of 87.7 years is attractive, but is the technology ready today to produce such a system? In the review article of Dr. Vakhtang Tchantchaleishvili with the title “Plutonium-238: An Ideal Power Source for Intracorporeal Ventricular Assist Devices?”,1 he concludes that plutonium-238 (238Pu) is an ideal power source for such a system. I do not disagree with the choice of that isotope, but I do think that choosing the power source is just the beginning of the process. When one considers all the other needs for such a system, it becomes clear that the implantable technology needed to convert the thermal energy to blood hydraulic power at the level that is needed is just not available at this time. What we need to do is to look at the entire system in detail and not just at the energy source.

To begin with, the nuclear isotope 238Pu was first described for this application by the physicist Dr. Edward Teller, the father of the hydrogen bomb. I had the privilege of working with him when I joined the research group of Thermo Electron Corporation in 1961. My concentration at that time was to develop nuclear-fueled power sources for deep space probes that used thermionic energy conversion principles. This nuclear technology was later applied to artificial heart systems when Thermo Electron applied for and received a contract from the National Heart, Lung, and Blood Institute in 1966 to develop such a system. When we formed Thermedics Inc., as a subsidiary of Thermo Electron to concentrate on the artificial heart work, I requested that Dr. Teller become a member of my Scientific Advisory Group, where he provided valuable insight into the artificial heart technology.

Dr. Teller recommended 238Pu because of its following characteristics. First of all, the plutonium dioxide fuel is used in its heat-resistant, ceramic form which reduces its chance of vaporizing in fire. This ceramic form fuel is also highly insoluble, has a low chemical reactivity, and if fractured, tends to break into large, nonrespirable particles and chunks. These characteristics help to mitigate the potential health effects from accidents involving the release of this fuel. Also, the fuel has a high melting point of 2,150°C, has a half-life of 87.7 years, and a power density of 4 watts per cubic centimeter. With these characteristics, we would not need a large quantity of fuel, and it would not melt in the event that a recipient was cremated. Radiation shieldingwise, it is attractive because it is primarily an alpha decay system with exceptionally low gamma and neutron radiation. A disadvantage is that there is a global shortage of this isotope. It must be produced in a nuclear reactor because it is a breakdown product. If this becomes a fuel of choice, we would need to build dedicated reactors for the sole purpose of producing this fuel. We all can imagine how difficult it would be to obtain approval to build a reactor, how long it would take, and how much it would cost. Just the cost of the fuel in 1972 was $1,000/gm.2 A 50 W capsule contains approximately 120 gm and would cost $120,000 in 1972 dollars. I have no idea how much it would cost today.

The next question that must be answered is how much power is required to pump blood. A simple calculation indicates that with a left ventricular assist device (LVAD), it would take 0.9 W to produce 5 lpm of blood flow at 80 mm Hg. Total heart requirements are quoted at approximately 1.7 W. Rotary blood pumps typically operate at efficiencies of 10–20%. If we choose an efficiency of 20%, the input power requirements for the blood pump, as shown in Table 1, would be 4.5 W. This value now defines the power output requirements that the thermal to electric energy converter would need to provide. There are several types of nuclear converters that could be used to produce this power requirement, and each has its own characteristic.

Table 1
Table 1:
Power and Efficiency Requirements for a Blood Flow of 5 lpm at 80 mm Hg

Converters that operate with a nuclear source would need to use an atomic battery. There are two types of converters, nonthermal converters and thermal converters. In the nonthermal converter types, the betavoltaics use silicone to capture electrons emitted from a radioactive gas such as tritium. Of interest, the first pacemakers used betavoltaic converters based on the radioactive element promethium.3 As of 2004, about 90 of these were still in use. These systems typically produce power in the milliwatt range and operate at efficiencies of 6–8%. This technology does not produce sufficient power levels to power an LVAD.

In the thermal converter group, output power is a function of temperature differentials. This group includes the thermoelectric generators that use the Seebeck effect and operate at efficiencies of 3–7%. They are typically fueled with 238Pu.This converter type has been used on at least 41 National Aeronautics and Space Administration space missions.4 They have powered satellites, deep space probes as well as experiments on the lunar surface. I was personally involved in the Systems for Nuclear Auxiliary Power (SNAP)-27 converter that was used on the moon. I was responsible for the development of the thermal simulators that were used to test all these systems. The second type of thermal converter used thermionic energy conversion technology.5 A hot electrode (1,600°C) thermionically emits electrons over a space charge barrier to a cooler electrode producing useful power output. Cesium vapor is used to optimize the electrode work function and provide an ion supply to neutralize the electron space charge. The efficiency of these systems is in the range of 10–20%. Unfortunately, 238Pu does not operate at a high enough temperature to be used with this technology. Cobalt 60 is the fuel of choice; however, this material could not be used in the human body. The third technology to consider would be a thermophotovoltaic system, which operates at a much lower temperature. In this technology, infrared light produced from a heated surface is converted to electricity. The efficiency of these systems operates in the range of 3–7%. The final system to consider would be a hybrid system, where the waste heat from a thermoelectric system would be used to provide a heat source to the lower operating thermophotovoltaic system. Combining these two systems would increase the efficiency to 6–14%.

Unfortunately at this point, we do not have an outstanding candidate that we could use as the thermal to electric converter. For the sake of the analysis, we will evaluate the system using a thermoelectric system as these are proven systems with many years of reliable use. As shown in Table 1, with a converter efficiency of 7%, we would need to have an input power level of 64 W, which would need to be produced by the 238Pu fuel capsule. This is an excessive amount of fuel, which should not be implanted into a human body. In addition, the size of the convertor would be large. The SNAP-27 generator that was used on the moon produced 60 W. It was 18 inches in diameter and 18 inches high and weighed 30 pounds. Fortunately, we do not need to deliver 60 W, only 4.5 W. A smaller system, the SNAP-3A produced 2.7 W, was 4.75 inches in diameter and 5.5 inches high with a weight of 4.6 pounds. We would need a system twice the size and weight of the SNAP-3A to satisfy our needs if the efficiency were to remain at 7%. With this converter efficiency, we would need a fuel capsule that could produce 64 W coupled to a heat exchanger that could safely dissipate 60 W of waste heat. We could build such a system today and implant it into a large animal as we did in 1970s. But that is as far as we could go with such a system because it would be impractical for human use; too large, too heavy, too expensive, too inefficient which would result in requiring a heat exchanger and finally, an excessive amount of nuclear fuel. Fundamentally, it is an efficiency problem. We could improve the efficiency by using a hybrid system that could operate at 14%, which would reduce the fuel requirement to a more reasonable level of 32 W. If we could further improve the efficiency of the thermal to electric converter to 20% as has been demonstrated in the laboratory with advanced hybrid systems and increase the blood pump efficiency to 30%, we would only need a fuel capsule capable of producing 15 W. Of interest, the predecessor of the HeartMate vented electric blood pump system operated at an efficiency of 31%.6 This system now becomes attractive because the converters become much smaller and lighter, the fuel cost is greatly reduced, the nuclear safety issues are reduced, and a heat exchanger is no longer needed.

In conclusion, it does appear that from a technical point of view and with improvements in efficiency, nuclear-powered systems would be feasible. Whether this system would be accepted by society would depend on the cost and associated risk of nuclear fuel being put into our society in an uncontrolled manner. As we saw in the pacemaker industry, the lower cost battery technology won out. If our society does not allow us to use nuclear energy, what is our alternative? I would submit a minimally tethered system that is feasible with today’s technology. Much like an automobile is a minimally tethered system that we accept with the high risk associated with transporting 20 gallons of volatile gasoline, we could accept an artificial heart system in much the same manner. An automobile is tethered to a fuel pump to recharge its energy source, and once fully charged, it then becomes untethered. The important characteristic is a short tether time with a long untethered time. A simple analysis can be made to establish the feasibility of such a system. If we could produce a blood pump with an efficiency of 30%, we would require a power source that could deliver 3 W of power to achieve a blood flow of 5 lpm at a pressure of 80 mm Hg. With today’s technology, a Li-ion battery system containing 30 W hours of energy would be 3.5 inches × 3 inches × 1 inch thick and would weigh 0.5 pounds. This implantable battery pack would power the blood pump for 10 hours in an untethered mode. Rapid battery recharging could take place with the second generation far distant transcutaneous energy transfer systems now being developed. This advanced technology does not require close coupling and greatly reduces the need for precise alignment while eliminating the risk of infection as seen with percutaneous systems. One important concern of course is implanting an energy pack, which contains a large amount of fuel; however, it should be no different from the risk associated with a nuclear source.

The future of assisted circulation is bright. With continued development of the technology, we can look forward to smaller systems that are more reliable with longer and longer untethered operation.


1. Tchantchaleishvili V, Bush BS, Swartz MF, Day SW, Massey HT. Plutonium-238: An ideal power source for intracorporeal ventricular devices? ASAIO J. 2012;58:550–554
2. Norman JC, Molokhia FA, Harmison LT, Whalen RL, Huffman FNAn implantable nuclear-fueled circulatory support system.. I. Systems analysis of conception, design, fabrication and initial in vivo testing. Ann Surg. 1972;176:492–502
3. Bourzac K. A 25-Year Battery. MIT Technology Review. 2009 Available at:
4. Harris B. Power Sources That Last a Century. Raytheon Technology Today. 2011;1:12–15
5. Hatsopoulos G, Gyftopoulos EP. Thermionic Energy Conversion, Volume 1: Processes and Devices. 1973 Cambridge, MA MIT Press:1–19
6. Poirier V, Gernes D, Szycher M. Advances in electrical assist devices. Trans Am Soc Artif Intern Organs. 1977;23:72–79
Copyright © 2012 by the American Society for Artificial Internal Organs