Energy Storage

A battery converts chemical energy into electrical energy. The source of this energy is the electrochemical reactions that occur within the battery. The type and amount of the materials participating in these reactions determine the deliverable energy of a battery.

Chemical Reactions

During a spontaneous chemical reaction, substances react to form reaction products. For example, in combustion, a fuel combines with oxygen (O₂) from air to form water (H₂O), carbon dioxide (CO₂), and other species. The fuel is oxidized (loses electrons) during this process. Oxygen from air is reduced (gains electrons). The exact amounts of each that react are determined by the stoichiometry of the reaction. Equations 6-1 to 6-4 are examples of different types of reactions in which oxidation and reduction occur, called “redox” reactions, because electrons are transferred from one reactant to another.

(6-1)

(6-2)

(6-3)

(6-4)

Such chemical reactions occur spontaneously because the products are in a lower energy state than the reactants. The difference in energy appears as heat in the case of combustion, or rusting. A battery or galvanic cell is designed to convert much of this energy difference into electrical energy rather than heat. A battery operates because the electrons transferred during a redox reaction are channeled from one terminal of the battery through the external circuit and back to the battery through its other terminal. The maximum work these electrons can do outside the battery is related to the difference in a thermodynamic quantity called the free energy of the reaction.

Anode and Cathode

The two battery components involved in the electrochemical reaction during discharge are the anode and the cathode. The anode, lithium (Li) in most medical batteries, furnishes electrons to the external circuit, while the cathode, usually an oxide or halogen-rich compound, receives the electrons.

The anode is defined as the electrode at which electrochemical oxidation occurs, and the cathode is the electrode at which electrochemical reduction occurs. Oxidation is a process in which electrons are removed or lost, and reduction is a process in which electrons are gained. This is true in all electrochemical cells. In a spontaneous galvanic cell such as a battery, the anode is electrically negative and the cathode electrically positive. This creates a seeming inconsistency. In nonspontaneous electrochemical reactions, such as those that occur at pacing electrodes, electrochemical reactions are driven by an externally applied voltage. In this case, oxidation (loss of electrons) will occur at the positive electrode and reduction (gain of electrons) at the negative electrode. Thus, for pacing electrodes, the anode is positive and the cathode is negative. The polarities are reversed, but the underlying electrochemical processes defining anode and cathode—oxidation and reduction, respectively—are the same. The terminology is entirely consistent.

Electrolyte

The anode and the cathode must be separated so that they cannot directly react with each other, but they also need to be connected by an ionically conducting medium called the electrolyte. As the battery discharges, it furnishes electrons at one terminal, pushes them through an external circuit, and receives them back at a second terminal. The electrolyte allows electric charge, as ions, to flow within the battery, completing the circuit. The electrolyte must conduct ions but not electrons. If the electrolyte conducted electrons, the battery would be shorted internally just as if the terminals were connected by a metal wire. Batteries with lithium anodes, such as those used to power pacemakers and ICDs, must use nonaqueous electrolytes because water readily reacts with lithium. Most Li-based batteries employ a mixture of organic ethers and esters as solvents for the electrolyte. A dissolved Li salt is used to make the electrolyte conductive. For example, lithium/manganese dioxide batteries, widely used to power automatic cameras, often use electrolytes containing a lithium salt dissolved in dimethoxyethane ether and propylene carbonate ester.

Separator

The separator is a structural member of the battery that keeps the anode and cathode materials physically apart, thus preventing shorting of the battery. In batteries with liquid electrolytes, the separator is usually a porous polymer film that is immersed in and permeated by the electrolyte. In the case of lithium/iodine (Li/I₂) batteries, traditionally used for pacemakers, the separator and electrolyte are the same (i.e., growing layer of solid, ionically conductive LiI discharge product).

Current Collector

The current collector makes the connection between the positive or negative terminal of the battery and its respective active electrode material inside the cell. A current collector is usually a wire connected to a screen, or grid, which is embedded in the anode or cathode material. The current collector may also serve as a structural member of the battery to provide physical integrity and strength to that electrode. Some medical batteries use the case of the cell as the current collector for one of the electrodes. Batteries are termed “case negative” if the anode is in contact with the case and “case positive” if the cathode is in contact with the case.

Primary Batteries

Primary batteries can only be used once. They are not designed to be recharged; in fact, it is dangerous to attempt to recharge them. A familiar example of a primary battery is the alkaline zinc/manganese dioxide flashlight battery. Most batteries used to power modern implantable medical devices are primary batteries that use lithium anodes because such batteries have very high energy densities.

Secondary Batteries

Secondary batteries are designed to be repetitively discharged and recharged. Familiar examples include the lead-acid batteries used in automobiles and lithium-ion batteries, widely used in portable electronic devices, home appliances, power tools, and electric vehicles. Lithium-ion batteries are also the most relevant secondary medical battery, used as supplemental power sources in some left ventricular assist devices and widely used to power implantable neurologic stimulators for control of chronic pain. Although not currently in development for pacemakers and ICDs, features such as more frequent telemetry could lead manufacturers to reconsider lithium-ion battery use in cardiac applications in the future.

Capacity

The fundamental unit of battery capacity is the coulomb (C), or ampere-second (A-sec). This is the amount of charge delivered by one ampere (1 A) of current in 1 second. In the context of implantable devices, it is customary to express battery capacity in terms of ampere-hours (A-hr), which represents the charge carried by a current of 1 A flowing for 1 hour; 1 A-hr = 3600 C. A battery for an implantable medical device usually has a capacity rating of to 2 A-hr. The methods for determining and specifying the capacity of a medical battery produce numbers ranging from upper-limit theoretical values that can never be achieved in the field to realistic values based on detailed models or accelerated testing.^1–3 More sophisticated methods account for limited availability of the components within the battery, kinetic limitations on the chemical reaction, and losses to side-reactions over time.

Estimating the deliverable capacity of implantable medical batteries is especially difficult because of their long service life. The time frame for the operation of most implantable medical devices is so long (5-10 years) that real-time measurements of capacity are not practical. Therefore, accelerated tests and models are typically used to estimate the amount of deliverable capacity in these batteries. Technology in this area is now highly developed, and it is possible to make highly accurate projections of deliverable battery capacity under a range of usage conditions.^4–6

Energy and Energy Density

The fundamental unit of energy is the joule (J). This is the energy given to one coulomb (1 C) of charge that is accelerated by a difference in potential of one volt (1 V). One joule is also the energy transferred by one watt (1 W) of power in 1 second. Just as battery capacity is often measured in ampere-hours, battery energy is often expressed in watt-hours (W-hr) instead of joules; 1 W-hr = 3600 J.

An important battery parameter in the design of an implantable device is energy density, a quantity that can be expressed on either a mass or a volume basis. For medical applications, volume is usually more important than mass, so ratings based on volumetric energy density are most often used. The time integral of the product of voltage and current divided by the total volume of the battery is its deliverable energy density. Modern batteries for implantable devices have energy densities as high as 1 W-hr/cm³, including the case.

Stoichiometry and Cell Balance

The specific amount of anode and cathode materials that will react is determined by the stoichiometry of the battery reaction. A cell that contains exactly the required ratio of anode and cathode materials is said to be a “balanced” cell. However, most medical batteries are not designed with exactly the stoichiometric ratio of the active cathode and anode to provide predictable end-of-service characteristics. In many cases the anode (lithium) is in excess so that the voltage decrease near the end-of-service life is not too abrupt.

Cell Voltage and Current

The open-circuit voltage of a battery can be calculated from the thermodynamic free energy for the discharge reaction. This is the voltage that will be measured when there are no kinetic limitations, a condition that occurs only when an insignificant amount of current is being drawn from the battery. With the onset of current flow, the voltage at the battery terminals will be smaller than the open-circuit value. Both chemistry and battery design determine the relationship between voltage and current drawn from the battery. For example, a lead-acid battery for automotive use is constructed of very conductive materials and is designed with large-surface-area electrodes so that extremely high currents can be drawn from it to run an engine’s starter motor. On the other hand, a transistor radio battery is designed with small electrodes because relatively low currents are typically needed to power small portable electronic devices. Figure 6-2 shows a typical current-voltage relationship in which the load voltage approaches the open-circuit voltage (OCV) as the current approaches zero. At the other extreme, the maximum (short-circuit) current is observed when the load voltage approaches zero.

Figure 6-2 Typical load voltage versus current drain plot for a battery.

The load voltage is always less than the open-circuit voltage (OCV). Current drain increases toward the right. The maximum cell voltage is obtained at zero current, and the maximum current drain occurs at zero load voltage. The exact shape of the plot depends on the chemistry and design of the battery.

Internal Resistance and Impedance

Electrical impedance and resistance are important battery properties that play a significant role in the clinical performance of many implantable devices. The terms impedance and resistance are often used interchangeably but are not the same. Both are terms for the change in voltage per unit change in current in an electric circuit, but they are measured under different conditions. Impedance is the more general term, encompassing effects of resistance, capacitance, inductance, and other circuit elements on the relationship between voltage and current. The resistive component of impedance is measured using direct-current methods. Alternating-current and transient methods are used to measure the additional components of impedance besides resistance.

For simple, resistive electric circuit elements, Ohm’s law, V = IR, accurately describes a linear relationship between voltage drop, V, and the corresponding change in current, I, with resistance, R, as the proportionality constant. However, a battery is a complex electrochemical device with several nonlinear processes operating in series/parallel combinations. Different processes may dominate at different current levels, depths of discharge, and time. Consequently, the relationship between current and voltage for a battery is, in general, nonlinear, even at very low currents. Although this relationship is sometimes characterized by a quantity called R_DC, Ohm’s law should only be applied to batteries with caution.

Non-Ideal Battery Behavior

In addition to the principles and the nomenclature of battery operation, it is also important to understand factors that limit a battery’s ability to power an implantable device.

Polarization

Polarization is any process that causes the voltage at the terminals of a battery to drop below its open-circuit value when it is providing current. The internal resistance of the battery is one important cause. This is well illustrated in Figure 6-3 for the lithium/iodine battery, but the same is true for all batteries to some degree. The differences in the curves for discharge voltage versus capacity at four rates of constant current discharge are mainly caused by the voltage drop associated with internal resistance of the battery. Other contributing elements of voltage loss when a battery provides current are concentration polarization, which is associated with concentration gradients that may develop in the electrolyte or the active electrode materials, and activation polarization, which is associated with the kinetics of the electron-transfer reactions at the electrode/electrolyte interfaces.

Figure 6-3 Load voltage versus capacity plot.

For a lithium/iodine battery discharged at four magnitudes of constant current. Curve A = 40 µA/cm², curve B = 20 µA/cm², curve C = 10 µA/cm², and curve D = 2 µA/cm². The deliverable capacity decreases with increasing current in this current range because of increasing effects of polarization.

When current is drawn from a battery, all these processes occur to some extent. The net effect of these kinetic limitations is always observed as a decrease in the voltage at the terminals of the battery. In general, neither concentration polarization nor electron-transfer polarization conforms to Ohm’s law.

Battery Design Requirements

Implantable medical devices must be viewed as a system. The main requirement in battery selection for implantable devices is high reliability. Other factors include the desired longevity of the device (directly related to battery energy density, circuit design, and overall size) and an appropriate indication of impending battery depletion (end-of-service warning). The basic considerations when designing a battery for an implantable medical device include the current variations that will be imposed by the circuit as the device provides its service for the individual patient. Once the span of these application requirements is formulated, the current, voltage, and capacity requirements of the battery can be determined. An important limitation is the physical space, both volume and shape, within the device that can be allotted to the battery. Once this is determined, the required energy density can be calculated, and various battery systems and designs can be considered and evaluated. To put these decisions in perspective, one needs to realize that the battery occupies about a third of the volume of an ICD. The circuitry occupies another third, and the capacitors fill the remaining volume. In a pacemaker the battery and circuitry each occupy about half the volume.

Power Requirements

An important parameter for a device is the peak power requirement. For example, the rate of energy consumption differs greatly for pacemakers and defibrillators. Pacemakers use very small amounts of energy when they stimulate the heart, on the order of 15 µJ. Defibrillators, on the other hand, deliver as much 40 J for a defibrillation shock. A battery optimized for a pacemaker could never come close to supplying energy at the rate required to power a defibrillator. Likewise, a defibrillator battery is not an optimum choice to power a pacemaker, although it could easily supply the current needed. The high-power design of a defibrillator battery has a significantly lower energy density than that of a pacemaker battery, by a factor of as much as two. Thus, if a defibrillator battery was used primarily for pacing, and everything else were equal, it would need to be twice as large as an optimized pacemaker battery to obtain the same longevity.

Optimizing a battery for longevity and power becomes more complicated when a device performs multiple functions, such as both bradycardia pacing and defibrillation. For example, up to now, the lithium/iodine battery has been the dominant power source for implantable cardiac pacemakers, which typically have peak power demands of 100 to 200 µW. Under these conditions, the lithium/iodine battery, which is still used in many pacemakers, can maintain an adequate voltage, even when its internal resistance reaches several thousand ohms. On the other hand, an implantable cardiac defibrillator may have peak power requirements approximately 10,000 times greater than those of a pacemaker. Under such a high power demand, the voltage of an Li/I₂ battery would drop to almost zero, and the power delivered to the device would be almost nil.

In recent years the distinction between a need for high-rate and low-rate batteries has become more blurred because features such as distance (“wireless”) telemetry and multisite pacing need both more current and more capacity to operate. The result is that battery designers have been challenged to develop medium-rate batteries that can deliver more power than pacemaker batteries of the past while still having a high energy density.

Average vs. Instantaneous Current Drain

Implantable medical devices are often characterized by their average current drain, although events such as therapy delivery and telemetry may require temporary current excursions that can differ greatly from the average value. The effect of the instantaneous current demand can be mitigated by using a capacitor that buffers the battery to allow short bursts of power that may be more than two orders of magnitude higher than it could deliver directly. This allows the use of battery designs with reduced anode and cathode surface areas and improved volumetric efficiency. The same principle involving a battery plus a capacitor is used to deliver the defibrillation therapy to the heart from an ICD, even though the magnitude of the current and voltage is much greater.

Shape, Size, and Mass Constraints

All device requirements (e.g., longevity, end-of-service indication, peak power) must be balanced against device volume, shape, and mass. Volume and thickness are particularly important for safe implantation in children and for esthetic reasons in many adults.

The ratio of volume to surface area in the battery is an important parameter. The performance of a battery of fixed volume will vary substantially depending on its electrode surface area/volume ratios. The operating current and longevity demanded of the battery determine both the minimum areas and the amounts of anode and cathode needed. Increasing the electrode area/volume ratio increases current capability, but this ratio must not be too large, or the battery will be too costly and will have a diminished energy density and most likely greater complexity, raising reliability concerns.

Size, Energy Density, and Current Drain

The relationship between battery size and average current is not one of direct proportionality. For example, decreasing the average current by 50% will not permit a 50% reduction in battery size without compromising longevity because of the inactive materials in a battery (e.g., case, electrolyte, current collectors). Likewise, the usable energy density is also a function of the current demand on the cell. As the current from the cell is increased, the resulting voltage drops significantly (see Fig. 6-3), which reduces the time during which the cell can provide current at or above the minimum voltage necessary to operate the electronic circuits. Thus, usable energy density, which is directly proportional to the area under the discharge (voltage vs. capacity) curve, is also reduced. For very-high-rate batteries such as those used to power ICDs, this is not such an issue because their internal resistance is extremely low.

Effect of Pulse Width on Pacing Current

Increasing the pacing rate, pulse width, or pulse amplitude increases the average pacing current. The average pacing current, excluding static current, is directly proportional to the pacing rate. Recall that the pacing pulse results from the discharge of a capacitor through the electrode-heart interface. This capacitor produces a pulse in which the current decays exponentially with time, as shown for two different pulse widths in Figure 6-4.

Figure 6-4 Comparison of charge delivered to heart as pacing pulse width is shortened.

The two cases (A and B) assume a discharge into the same patient. The pulse width, t_w, is typically between 0.2 and 1.5 msec, and the maximum current is between 2 and 15 mA, depending on the pulse generator output voltage and the lead-heart interface resistance.

Thus, the time-dependent behavior of the current during the pacing pulse is given by the following equation:

(6-7)

where V_A is the amplitude at the beginning of the pacing pulse, R_H is the impedance of the lead plus the heart (discussed later), C is the value of the capacitor that delivers the pacing pulse, and t is the time since the beginning of the pacing pulse. In Figure 6-4, A and B, the pulse width is t_w and t_w/2, respectively. The area under each current-time curve gives the total charge delivered during the pulse. Although the width of the pulse in B is half that of the pulse in A, the charge delivered by this pulse is considerably more than half that of the longer pulse. The exact ratio of the charge delivered in the two cases depends on the values of R_H and C. Nevertheless, reducing the pulse width by a given fraction will always reduce the average pacing current by a substantially smaller fraction because of the exponentially decaying shape of the pacing stimulus current curve.

Effect of Pulse Amplitude on Pacing Current

The definition of pacing pulse amplitude may vary somewhat between manufacturers of implantable pulse generators. For our purposes, pulse amplitude is defined as the voltage delivered to the heart at the beginning of the pacing pulse (“leading edge” voltage). As stated earlier, the area under the current-time curve gives the charge delivered per pulse. Thus, doubling the amplitude doubles the current and the total charge delivered to the heart. Also, because the charge per pulse is doubled, it might seem that the average pacing current drawn from the battery would also be doubled. However, the impact on the pacing current is much larger than that, as seen from the following argument. The energy per pacing pulse is defined as follows:

(6-8)

In Equation 6-8, V_A is the average pacing stimulus output voltage of the pulse generator, I_A is the average pacing current delivered to the heart, and t_w is the pulse width. If we consider the lead-electrode-heart interface to be mainly resistive, Ohm’s law, I = V/R, can be substituted in Equation 6-8, which becomes the following:

(6-9)

In Equation 6-9, R_H is the effective ohmic load of the heart and lead. From this equation it is readily apparent that energy consumption increases with the square of the output voltage. The effect of this energy increase on the current drawn from the battery may not be intuitive. Because the battery supplies all the energy delivered to the heart at a relatively constant voltage, any increase in energy is accompanied by a proportional increase in current drawn from the battery. Thus the average pacing current is increased by a factor of four if V_A is doubled. In fact, this is the best situation; additional energy losses occur when the stimulus voltage is programmed to a higher level because the electronic circuit for increasing the stimulus voltage is not 100% efficient.

Programming Effects on Longevity of Bradycardia Pulse Generators

In summary, the wide range of available pacing parameters can have a dramatic effect on the current drain from the battery in an implanted pulse generator. For example, in the same patient, a bradycardia pulse generator with 6 years of longevity under nominal pacing parameters may reach its replacement time in 2 years at one extreme or more than 10 years at the other extreme. Therefore, it is important for clinicians to optimize parameters for the individual patient and to consider each device’s settings when predicting remaining device longevity for a specific patient.

Considerations for Longevity of ICDs

Most of the arguments concerning bradycardia pacemakers also apply to ICDs. However, ICDs are much more complicated than typical pacemakers, so it is not nearly as easy to generalize about the effect of different factors on their longevity. In general, the two main factors to consider are the expected frequency of tachyarrhythmia therapies and the percentage of time spent pacing the heart. The longevity may vary by a factor of two to three because of these issues alone, as discussed later.

Elective Replacement Indicator

All modern implantable pulse generators have an elective replacement indicator (ERI) that alerts the clinician to impending battery depletion and allows adequate time for replacement of the device. Typically, the ERI is designed to alert at least 3 months before the battery voltage drops to a level at which erratic pacing, loss of capture results, or loss of other critical features occur. Future devices will conform to CENELEC standards (European Committee for Electrotechnical Standardization). For pacemakers, the standard requires a recommended replacement time (RRT) such that at least 95% of the devices will have a prolonged service period (PSP) of at least 180 days. For ICD and CRT devices, the requirement is that at least 95% of the devices have a PSP of at least 90 days. Device manufacturers may choose to retain the ERI nomenclature and may set ERI to be coincident with RRT or between RRT and end of service.

Methods for Monitoring State of Battery Discharge

Battery Voltage

The most common method to indicate impending battery depletion is to measure the battery voltage. Most modern devices incorporate a voltage measurement circuit in the form of an analog-to-digital converter. The digitized battery voltage can be compared with a value stored in a nonvolatile memory to trigger the ERI, or it can be telemetered to the programmer where the comparison is made.

For lithium/iodine batteries, the battery voltage remains relatively constant throughout most of its discharge under low load conditions, as shown in the voltage versus capacity curve in Figure 6-5. This figure also shows the resistance of this battery as it discharges. Note that the resistance changes from a modest value at the beginning of service to a quite large value when it is almost depleted.

Figure 6-5 Relationships among voltage, resistance, and delivered capacity.

Voltage remains reasonably constant through the service life of lithium/iodine battery, whereas the resistance steadily builds up with discharge. Note the rapid falloff in voltage as the resistance rapidly increases near the end of service life.

Because the battery voltage is relatively constant for much of the pulse generator’s useful life, the telemetered voltage of this battery may not be particularly useful for estimating remaining service. Voltage characteristics during discharge differ for different battery chemistries, so clinicians should not assume familiarity with one model or that one type of battery can be applied to another. For example, the lithium/iodine battery has a fairly flat discharge curve throughout most of its life at pacing currents. However, other battery chemistries have voltage characteristics much different than Li/I₂. The lithium/silver vanadium oxide (Li/SVO) battery used in many ICDs has two distinct voltage regions before the typical ERI voltage. The region following the second plateau is sloped. The switch to the sloped discharge region can be used as the ERI trigger for this battery chemistry. This voltage is much less sensitive to current variances than the voltage chosen for an Li/I₂ battery because the internal resistance of the Li/SVO battery is much lower than the Li/I₂, so for the Li/SVO system, voltage is a good indicator of remaining service life.

False Triggering

Any large increase in average battery current, even if temporary, in pulse generators that utilize battery voltage to indicate the elective replacement point may drop the battery voltage below the ERI trigger value before the true ERI point is reached; this occurs particularly with the Li/I₂ battery because of its high resistance. For example, changes in current can result from electrophysiologic studies with noninvasively programmed stimulation (NIPS) in which the pacemaker is used to interrupt an episode of tachycardia by delivering short bursts of very-high-rate pacing, or from high pacing rates or telemetry sessions. Because ERI is usually latched (stored) when it is triggered, these temporary increases in current can cause a false, premature appearance of the ERI. In most cases, devices are designed to allow an ERI prematurely triggered to be reset. In some devices, ERI is inhibited during temporary high-current events to prevent false triggers.

In ICDs, ERI is usually based on the battery voltage or the time required to charge the capacitor. The battery voltage is usually measured during normal sensing/pacing operation and not during a defibrillation therapy, when the battery voltage is depressed. Because of the very low internal resistance of ICD batteries, pacing parameters have a much smaller effect on triggering ERI in these devices, so false triggering is less of a problem. Alternatively, some ICDs base ERI determination on the time required to charge the output capacitor. This approach is possible because most battery designs have a reduced power capability as the battery approaches depletion. This method does not usually cause false ERI triggering.

Lithium/Iodine Battery

The lithium/iodine battery is probably the most well known implantable battery because it has been used in the vast majority of cardiac pacemakers. The first implant of a pacemaker powered by an Li/I₂ occurred in 1972.^5,6,9 At least 13 million Li/I₂-powered pacemakers have now been implanted. Many factors favor the use of this battery system. When the current demand is low, it is difficult to improve on the performance of Li/I₂ batteries; their high energy density and low rate of self-discharge result in good longevity and small size. The inherently high impedance of the Li/I₂ battery has not been a major disadvantage up to now because the current required by modern pacemaker circuits is low, typically about 10 µA. Note that the much larger current delivered during a (short duration) pacing pulse is drawn from a capacitor, which can charge between pacing pulses. The voltage and impedance characteristics of the Li/I₂ cell also allow the clinician to monitor the approaching end-of-service indication. This battery system is simple, elegant in concept and is inherently resistant to many common modes of failure, as will be discussed below. As a result, Li/I₂ batteries have attained a record of reliability unsurpassed among electrochemical power sources.

Cell Structure

Figure 6-6 shows a cutaway view of a typical Li/I₂ battery. In general, Li/I₂ batteries have a single, central lithium anode that is surrounded by cathode material, which is at least 96% iodine and has been thermally reacted with a polymer material to render a conductive mixture. The central anode is seen with an embedded current collector wire, and the iodine cathode fills much of the volume inside the battery. Figure 6-6 also shows several other important structures, including the electrical feedthrough that connects the anode to the outside of the cell. The case serves as the site of the electrical connection to the cathode, which is in direct contact with the inside of the container. Another visible feature is the fill port, the means by which the cathode mixture is introduced into the cell, after which the fill port is hermetically sealed.

Figure 6-6 Cutaway view of typical lithium/iodine battery.

Easily identifiable interior parts include the anode feedthrough, anode current collector (wire), cathode, poly-2-vinylpyridine (P2VP) film on the anode, and port where the cathode material is poured into the cell. The connection to the cathode is through the battery case.

Lithium/Carbon Monofluoride Battery

Lithium/carbon monofluoride (Li/CF_x) batteries were introduced in commercial coin-type batteries in 1976. Carbon monofluoride (also known as CF_x) is a cathode material with very high capacity and moderate power capability. The CF_x material is a powder that is mixed with carbon as a conductivity enhancer plus a polymer binder and then pressed into a porous pellet. The battery comprises of a lithium anode, porous polymer separator material, and porous pressed-powder cathode pellet that are inserted into a case and hermetically sealed, similar to an Li/I₂ battery. In contrast to the Li/I₂ battery, however, the Li/CF_x battery uses a liquid electrolyte composed of a lithium salt dissolved in an organic solvent (or solvent mixture). The energy density of this battery is similar to that of the Li/I₂ battery, but it delivers significantly higher power. A typical discharge curve is shown in Figure 6-7, A.

Figure 6-7 Voltage versus percent discharged capacity.

A, Typical discharge curve of a lithium/carbon monofluoride (Li/CF_x) battery. B, Typical discharge of a lithium/silver vanadium oxide (Li/SVO) battery. C, Typical discharge curve of a CF_x-SVO hybrid cathode battery.

One feature of the Li/CF_x system is that the voltage decline near end of battery life tends to be fairly abrupt, making it challenging for device manufacturers to engineer means of maintaining adequate replacement-time warning.

Lithium/Hybrid Cathode Battery

New lithium battery chemistry with a cathode consisting of silver vanadium oxide (SVO) plus CF_x has been developed to meet the needs of implantable devices with higher rate therapies and features.^11,12 SVO has high power capability and is the same cathode material used to power ICDs. As previously described, CF_x has modest power capability and more abrupt end-of-service characteristics, but a very high capacity. The blend of the two cathode materials, called a hybrid cathode, yields a primary battery that has an energy density equal to that of a lithium/iodine battery with approximately 100 times more power capability. It can support the new device features and also has good end-of-service detection because of the SVO. Hybrid cathode batteries have been used in implantable cardiac devices since 1999. Cumulative implants through 2009 total about 300,000.

Plots of voltage versus percent of discharged capacity for SVO and CF_x batteries are shown in Figure 6-7, A and B. The voltage curve of hybrid cathode behaves like a superposition of the two in proportion to the starting composition of the mixture (Fig. 6-7, C).

The composition of the hybrid can be chosen based on the application, enhancing the capacity with a greater proportion of CF_x or enhancing the power capability and end-of-service characteristics with more SVO. For low- and medium-rate applications, the composition can be chosen such that 85% to 90% of the battery capacity comes from CF_x. In this composition range, efficiently designed medium-rate hybrid cathode batteries can match lithium/iodine in energy density at about 1 W-hr/cm³.

Two basic hybrid cathode battery designs have been used in implantable pacemakers. Both types use lithium anodes, a porous polymer separator, a porous pressed-powder cathode pellet, and a liquid electrolyte of lithium salt dissolved in an organic solvent blend. The differences in the two designs are related to the construction of the cathode pellet, either a mixture of the SVO and CF_x active materials or the CF_x and SVO in separate layers, with the SVO layer adjacent to the lithium anode. In both cases the cathodes also contain a polymer binder and carbon as a conductivity enhancer. Although the methods of construction are different, the discharge behavior of either design is similar to that in Figure 6-7, C.

Comparison of Pacemaker and Defibrillator Batteries

Implantable defibrillators deliver up to 40 J of energy to the heart within a few milliseconds. By contrast, the energy required to stimulate the heart in bradycardia pacing is 1 to 10 µJ over 1 msec or less. Clearly, the demands placed on the battery that powers an implantable defibrillator are much different from those of a battery that powers a bradycardia pacemaker. Implantable defibrillators are designed to deliver a shock within about 10 seconds after the fibrillation or tachycardia is detected. When the ICD determines that a shock may be required, it begins to charge a high-voltage capacitor (discussed later). The time needed to charge the capacitors before the shock is delivered depends on the ability of the battery to sustain a very high current during this period. Thus the implantable defibrillator requires a battery with high peak power, where power is the product of current and voltage (P = I × V).

The high power required of defibrillator batteries dictates a different design than used for pacemaker batteries. Figure 6-6 shows a cutaway view of a typical battery used to power a pacemaker, with small and rather thick electrodes. Contrast this with Figure 6-8, which shows an analogous cutaway for a typical defibrillator battery, with very different construction. The long, thin anode and cathode, with two layers of separator film in between, are rolled into a flattened coil. Alternatively, either or both of the electrodes of ICD batteries can be individual flat-plate electrodes electrically connected to each other. Either method provides the high-power electrode area necessary for ICD batteries. The electrolyte in the defibrillator battery is a highly conductive solution of a lithium salt in organic solvents. The design with large, thin electrodes and liquid electrolyte gives the defibrillator battery the high power needed for quickly charging the high-voltage capacitors in the defibrillator. As mentioned, however, these same characteristics also reduce the energy density of the battery.

Figure 6-8 Cutaway view of typical high-rate defibrillator battery.

In wound construction a sandwich of anode, separator, and cathode is rolled, flattened, and fitted inside the battery case. This battery uses a liquid electrolyte to ensure high internal conductivity. In this battery the anode is connected to the case and the cathode is connected to the feedthrough. Note the multiple connections to the cathode to ensure low internal resistance.

Lithium/Silver Vanadium Oxide

The first ICD battery chemistry put into widespread use was Li/SVO.^11–17 These batteries have a distinctive discharge curve, with a slightly sloping voltage at about 3 V in the first third of discharge, followed by a decline in voltage to a flat region with about 2.6 V, and ending in a short region of declining voltage, as seen in Figure 6-10. The uppermost curve represents the battery voltage during normal sensing and bradycardia pacing operations (also called the “background” voltage). It is this background voltage that is typically monitored by the ICD and reported via telemetry. The “load” voltage is the voltage measured during charging of the defibrillation capacitors, and it is typically much lower than the background voltage.

Figure 6-10 Typical plot of discharge characteristics of the lithium/silver vanadium oxide ICD battery.

After many years of laboratory testing and experience in the field, it is apparent that the performance of Li/SVO batteries depends on how rapidly the battery is depleted. When the battery is depleted over a longer period, its internal impedance is higher in the latter half of the discharge curve. Figure 6-11 illustrates this phenomenon, with background voltage and internal impedance of Li/SVO batteries discharged over 3, 5, and 7 years. The internal impedance is unaffected by the discharge rate until midway through discharge. After that point the internal impedance always increases, but to a larger degree for slower discharge rates (i.e., longer discharge times). The higher internal resistance results in longer charge times, especially for the longest discharge periods.

Figure 6-11 Comparison of area-normalized resistance of lithium/silver vanadium oxide battery for various battery longevities. The area-normalized resistance is the internal impedance of the battery, calculated from the drop in voltage during the pulse, the current, and the area of the cathode. The area-normalized resistance = A_cathode · (V_background − V_pulse)/I_pulse.

Charge Time–Optimized Li/SVO ICD Batteries

One manufacturer has addressed the problem of the time-dependent increase in charge time by changing the ratio of anode to cathode material in the battery. The first Li/SVO batteries were made with an excess of lithium anode material so that the voltage only dropped slowly at the end of service life. “Cathode limitation” is the most efficient battery design if the goal is to optimize energy density. However, if the goal is to maintain a low charge time, it makes sense to use only part of the cathode capacity, ending the discharge partway onto the 2.6-V plateau where the charge time starts to increase. This is accomplished by reducing the amount of anode and increasing the amount of SVO cathode. As a result, the battery always remains in the region where internal impedance is low and charge times are short.

Figure 6-12 compares the voltage and charge time of conventional SVO ICD batteries with charge time–optimized batteries that occupy the same volume in an ICD. For example, when discharged over 7 years, the charge time is considerably lower starting after 4 years for the charge time–optimized battery than the conventionally balanced battery (Fig. 6-12, A). The charge time–optimized battery maintains a much lower charge time in the latter half of ICD life, without substantially sacrificing longevity.¹⁸ The difference in charge time is much less when the battery is depleted over a shorter time, such as 3 years (Fig. 6-12, B).

Figure 6-12 Comparison of the performance of charge time–optimized batteries with conventionally balanced silver vanadium oxide (SVO) batteries as a function of time for battery longevity of 7 years (A) and 3 years (B). Both batteries occupy the same volume in the ICD.

Lithium/Manganese Dioxide

Another ICD manufacturer uses manganese dioxide batteries for use in ICDs. Lithium/manganese dioxide (Li/MnO₂) batteries have been used in consumer and military batteries since the 1980s. For example, high-power Li/MnO₂ batteries used for photoflash applications have similar electrical characteristics to those used in ICDs. Figure 6-13 provides a graph of Li/MnO₂ battery performance. During the first half of battery life, the voltage remains almost constant at 3.1 V. In the second half of battery life, the voltage gradually slopes downward. This characteristic of the battery voltage, along with charge time and the amount of battery energy consumed, allows the clinician to judge how much longer the battery may last. The internal resistance and thus the charge time of Li/MnO₂ batteries remain relatively constant until battery replacement is indicated by the ICD.¹⁹

Figure 6-13 Voltage and internal resistance with extent of discharge for a lithium/manganese dioxide ICD battery. Charge times remain fast and relatively steady as the internal resistance remains low until near the end of battery life.

(Courtesy Boston Scientific.)

Lithium/Layered Silver Vanadium Oxide–Carbon Monofluoride

Another manufacturer of ICD batteries has combined the properties of two cathode materials, silver vanadium oxide and carbon monofluoride, to create a novel ICD battery that exploits the best features of those two materials. SVO provides high power while CF_x contributes high capacity per unit volume. The construction of the cathode portion of the battery is rather complex, with a central CF_x cathode sandwiched between two metallic current collectors, with SVO electrodes on each side. This arrangement of electrodes allows the SVO, which can discharge at very high rates, to be preferentially discharged when high current is needed to charge the capacitors. Over time the extra lithium ions that accumulate in the SVO part of the electrode equalize with the CF_x part, so that the lithium ions become distributed equally within both parts of the cathode. The voltage curve starts at about 3.2 V, stays almost flat from about 20% to 50%, and then gradually declines. Figure 6-14 shows a typical discharge curve.²⁰

Figure 6-14 Discharge curve for lithium/layered silver vanadium oxide–carbon monofluoride battery as a function of percent depth of discharge (DOD). This test lasted 6 months. A 3.15-A pulse was applied about every 15 days. Upper curve, Battery voltage on a low current drain. Lower curve, Battery voltage at the end of the 3.15-A pulse. Batteries of two different sizes were compared, 6.5 cm³ (blue) and 9.0 cm² (orange), with minimal differences.

(Courtesy Dr. Esther Takeuchi, State University of New York, Buffalo.)

Principles of Operation

The defining feature of a lithium-ion battery is that it contains no metallic lithium. Instead, lithium ions (Li⁺) are shuttled back and forth between the positive and negative electrodes during charge and discharge, as shown in Figure 6-15. The most common electrode materials are lithium cobalt oxide (LiCoO₂) for the positive cathode material and a graphitic carbon to contain the intercalated lithium for the negative anode material. Many alternative materials are being developed and introduced in commercial batteries and will eventually migrate to batteries for implantable medical devices. The general cell construction is similar to that shown for the Li/SVO battery. However, lithium-ion batteries are manufactured in their discharged state (e.g., using LiCoO₂ and graphite) and then charged (or “formed”) after the cell is fully assembled.

Figure 6-15 Schematic diagram of a lithium-ion (Li⁺) battery showing lithium ions shuttling between the positive and negative electrodes as the battery is charged and discharged.

Method of Recharge

To date, all of the implantable applications for rechargeable batteries involve recharge by transcutaneous electromagnetic induction. The implanted device incorporates a wound-wire coil that acts as the secondary to an externally located primary coil that is placed adjacent to the implanted device. Typical lithium-ion batteries are capable of accepting a full charge in 2 hours or less. Recharge times in future batteries with improved chemistries could become as short as 10 to 30 minutes. In addition to battery capacity, actual recharge time in ICDs depends on the relative positions of the primary and secondary coils, the depth of the implant, and the maximum power that can be transmitted while limiting heating of the primary coil or the implanted device.

End of Service Life

Lithium-ion batteries typically exhibit a gradually declining voltage during discharge that can be used as a “gas gauge” to indicate the state of discharge at a given time (Fig. 6-16). Information about the voltage can be telemetered to a patient controller or a monitor to alert the patient of the need to recharge the battery. However, the end of service for a device powered by a lithium-ion battery is not as readily apparent as for one powered by a primary battery. Because the lithium-ion battery slowly loses capacity as a function of both time and the total number of charge/discharge cycles, the patient may eventually experience a reduced interval between recharge sessions becomes too short to be acceptable to the patient. However, because rechargeable batteries can, in principle, provide unlimited energy, the eventual end of service may be determined by the lifetime of some other component of the implanted device.

Figure 6-16 Monotonic discharge curve of the lithium-ion battery shows how it could be used as a “gas gauge” for determining the remaining capacity in a battery.

Capacitance

In its simplest form, a capacitor consists of two conductors, separated in space and electrically insulated from each other. When the conductors are charged, each stores an equal and opposite amount of charge. In an ideal capacitor, the amount of charge (Q) on each conductor is proportional to the difference in voltage (V) between the two conductors. The proportionality constant, C, is called the capacitance.

(6-10)

The simplest model of a capacitor is shown in Figure 6-17, A, in which the plates are separated in a vacuum. The conductors (or electrodes) consist of two parallel plates of area (A) separated by a distance (d). The capacitance, or ability to store charge as a function of voltage, is equal to A/d multiplied by a fundamental constant, ε₀, the permittivity of free space, which is related to the energy required to separate charge in a vacuum. The value of the capacitance can be increased by inserting insulating materials, known as dielectrics, between the electrodes. The factor by which these materials increase the capacitance is called the dielectric constant, k, which is the ratio of the permittivity of the material to the permittivity of free space, ε/ε₀. A simple capacitor containing a dielectric material is shown in Figure 6-17, B. Various materials are used as dielectrics the properties of which largely determine the properties of the capacitor.

Figure 6-17 Schematic representation of a parallel-plate capacitor with A, a vacuum, and B, a dielectric, between the capacitor plates. The capacitance is directly proportional to the area of the plates (A), and inversely proportional to the distance between the plates (d). The insertion of a dielectric material between the plates increases the capacitance by a factor of k, the dielectric constant, which is a characteristic of the particular insulating material.

Energy Storage

In an ICD, several capacitors are required to deliver a high-voltage, high-energy electric shock through the lead system to the heart. Capacitors are passive devices and therefore must store energy before it can be delivered. The process of storing energy in a capacitor is called “charging.” For a capacitor to be charged, both terminals must be connected to a charging circuit that includes a transformer and to a voltage source such as a battery. The transformer converts the low-voltage energy supplied by the battery into the high-voltage energy stored by the capacitor. The amount of energy stored on the capacitor is proportional to the square of the voltage on the capacitor, as follows:

(6-11)

Energy Delivery

During defibrillation, the energy stored in the capacitor is delivered to the heart using an electrical path between the opposing electrodes of the capacitor. This connection allows the separated charge to recombine by traveling through the heart.

When an ideal capacitor is discharged through a constant resistance, R, the voltage decays with time in an exponential manner, as seen in Figure 6-18. For an ideal capacitor discharged to 0 volts, the delivered energy would equal the stored energy. In ICDs the amount of energy delivered from the capacitor is less than the amount stored. One reason is that the therapy delivery is truncated before the capacitor is completely discharged, and the remaining energy is discarded. A truncated waveform is used because it defibrillates the heart more effectively than a waveform that is allowed to decay to 0 V, despite less energy being delivered to the heart. Other inefficiencies associated with the capacitor, device circuit, and delivery system also reduce the amount of stored energy delivered to the heart, as discussed later.

Figure 6-18 Waveform for the discharge of an ideal 100-microfarad capacitor, charged to 750 V, into a 50-ohm load. These parameters approximate those for an implantable defibrillator in use.

Energy Density

Energy density is an important parameter for capacitors as well as batteries. Capacitor energy density is typically expressed in terms of volume and has units of joules per cubic centimeter (J/cm³). The most energy-dense modern defibrillation capacitors deliver energies in excess of 5 J/cm³. As previously discussed, the highest energy density seen for implantable batteries is about 1 W-hr/cm³, which is the same as 3600 J/cm³, or more than 500 times as high as the best defibrillation capacitor. Why, then, use capacitors at all? The reasons are voltage and power. Defibrillation therapy requires the energy to be delivered rapidly at about 1000 W of power and starting at a voltage of 650 to 850 V. ICD batteries are designed to deliver the energy to the circuit at about 10 W of power with voltage between 2.5 and 3.0 V. Consequently, the ICD is designed with the battery as a large energy reservoir, analogous to the gas tank of an automobile, and the capacitor as the fuel injector to the engine. Nevertheless, it is important for the capacitors to have the highest practicable energy density in order to minimize the overall volume of the ICD.

ICD Capacitor Types

The defibrillation capacitors used in past and current ICDs are called electrolytic capacitors, most often based on aluminum or tantalum electrode materials. The metal electrode materials are processed so they have a very large surface area, which allows for a high capacitance. For aluminum electrolytic capacitors, the positive electrode (anode in a capacitor) starts as a thin foil that has been etched to create a large number of tunnels. Figure 6-19 shows a scanning electron micrograph of a cross section of a highly etched anode foil. For tantalum electrolytic capacitors, the anode is made from tantalum metal powder pressed into a porous pellet and heated to a high enough temperature to make the metal particles bond to one another (called sintering) while still remaining porous. For either material, an oxide film is electrochemically grown (called forming) on the exposed surfaces of the metal electrode. The oxide serves as the capacitor dielectric. Aluminum anodes are typically formed to support charging to about 400 V. Two capacitors are connected in series in the ICD to provide sufficient voltage for defibrillation. Tantalum anodes are typically formed to support charging to 175 to 250 V with three or four capacitors connected in series.

Figure 6-19 Photomicrograph of a replica made from the etched anode foil used in aluminum electrolytic capacitors. The long, linear structures are tunnels etched into the aluminum foil. In this example some of the tunnels are etched completely through the foil to provide an electrical connection between the two sides. Magnification is about 800×.

The high capacitance, high voltage, and relatively low resistance of aluminum and tantalum electrolytic capacitors provide their high energy density (3-5 J/cc) and quick, efficient energy delivery that make them the capacitors of choice for ICDs.

ICD Capacitor Construction

Early ICD devices used commercially produced, cylindrically wound, aluminum electrolytic capacitors developed for flash photography applications. The need for smaller, thinner capacitors and specialized shapes led ICD manufacturers to develop stacked-plate aluminum electrolytic capacitors and custom tantalum electrolytic capacitors.^21,22

Cylindrical Aluminum Electrolytic Construction

Most commercially produced aluminum electrolytic capacitors employ wound electrodes in a cylindrical shape. Strips of anode and cathode foil, spaced apart by a paper separator, are wound into a coil. Figure 6-20 shows a cutaway view of an aluminum electrolytic photoflash capacitor of the type typically used in the first few generations of ICDs.

Figure 6-20 Cutaway view of wound aluminum electrolytic capacitor used in early implantable defibrillators.

The anode and cathode connections are brought out of the capacitor using feedthroughs. The capacitor is closed with a crimp seal. Notice that two layers of anode material are used for each layer of cathode material. The entire wound capacitor coil is saturated with electrolyte (not shown).

Stacked-plate Aluminum Electrolytic Construction

ICD manufacturers have developed stacked-plate aluminum electrolytic capacitor technology to improve the energy density, minimize the thickness, and allow a variety of shapes compared with cylindrically wound capacitors. The details of construction vary among manufacturers, but several design themes are common. Foils and separators are cut to a specific geometry and layered such that multiple anode plates are sandwiched between cathodes and layers of porous separator paper. The geometry of the plates is selected so that the finished capacitor fits efficiently in the ICD. The energy density of stacked-plate capacitors can approach 4 J/cc versus the 2.5 J/cc typical of the coiled configuration. Figure 6-21 shows a cutaway view of a stacked-plate aluminum electrolytic capacitor similar to those used in current ICDs.

Figure 6-21 Cutaway view of stacked-plate aluminum electrolytic capacitor used in current implantable defibrillators.

Anode and cathode layers are cut to a specific geometry and are stacked upon one another. Anode and cathode terminals exit the welded capacitor housing through feedthroughs. The capacitor encasement is laser welded. Note that four layers of anode material are used for each layer of cathode material (some designs use up to five layers). The entire capacitor stack is saturated with electrolyte (not shown).

Tantalum Electrolytic Capacitor Construction

Tantalum electrolytic capacitors have the same basic elements as aluminum electrolytic capacitors. In tantalum capacitors the anode is made using pressed, sintered, and formed tantalum pellet. The cathode in a tantalum capacitor is typically deposited directly on the inside of the capacitor case, making it the negative terminal of the capacitor. Tantalum capacitors have similar benefits to stacked-plate aluminum capacitors but provide higher energy and simpler construction than aluminum capacitors. Rather than layered anode foils, tantalum capacitors have a single anode pellet. The simple and efficient construction minimizes the volume of materials that do not contribute to energy storage (cathode, separator, encasement), thus maximizing energy density. Tantalum capacitors used in current ICDs have an energy density of 5 J/cc or more. The primary drawback of tantalum capacitors is their mass. Because tantalum is very dense, tantalum capacitors are approximately 50% heavier than aluminum capacitors with equivalent energy. Figure 6-22 shows a cutaway view of a tantalum electrolytic capacitor similar to those used in some current ICDs.

Figure 6-22 Cutaway view of a tantalum electrolytic capacitor used in current implantable defibrillators.

Design incorporates a single tantalum anode pellet pressed from tantalum powder. A tantalum lead wire exits the welded housing by a feedthrough. High-capacitance cathode material is deposited directly on the case. The entire electrode assembly is saturated with electrolyte (not shown).

Non-Ideal Behavior in Capacitors

Although the electrolytic capacitors used in ICDs have performance characteristics well suited to the ICD application, their performance does deviate from ideal in ways that can be clinically significant.

Deformation

The non-ideal process that is most apparent is called deformation. As previously explained, the dielectric material in electrolytic capacitors is created by “forming,” immediately after which the dielectric is almost free of defects. Over time, the chemical environment within the capacitor or relaxation of residual stress will cause microscopic imperfections in the oxide film on the anode. When the capacitor is charged to a high voltage after a long period of disuse, additional energy is required to regrow oxide in these areas to heal the defects.²³ This process is known as re-formation. The additional energy needed to re-form the dielectric results in longer charge times in ICDs. In a typical aluminum capacitor, 25% or more additional energy (and time) may be required to charge the capacitor if it has been many months since the last charge. The amount of additional energy drops to about 15% for capacitors not charged for about a month. Tantalum capacitors perform similarly, although the rate of deformation may be somewhat slower. Figure 6-23 shows capacitor deformation in typical aluminum capacitors for various charging intervals.

Figure 6-23 Percent deformation.

For typical photoflash capacitor used in early ICDs and typical stacked-plate performance. Percent deformation is percentage of additional energy required to charge a capacitor after some period at open-circuit voltage.

To minimize the clinical effect of deformation, current ICDs incorporate a programmable maintenance routine that automatically charges the capacitor to high voltage to heal the dielectric at intervals of 1 month, 3 months, or 6 months.

Capacitor Failure Modes

Capacitor failure modes depend greatly on the specific conditions of device use. The failure modes most relevant to the ICD application are discussed here.

Future Developments

Electrolytic capacitors are expected to remain the technology of choice for many future generations of ICD. Research continues on materials and processing methods to continue to improve energy density and performance. There is no clear path to integration of alternate capacitor technologies, such as ceramic, polymer, or double layer, but new materials and processes are being investigated that might provide smaller size and improved performance characteristics.