Rechargeable Metal-Ion Batteries for Energy Storage

CSUN AIMS2 Program Attract, Inspire, Mentor and Support Students Department of Education Grant California State University, Northridge Kevin Miranda August 2013 CSUN Students: John Grijalva & Travis Van Leeuwen CSUN Advisers: B. Bavarian & L. Reiner

Abstract Lithium-ion batteries dominate the energy storage industry and research & development funding. Because of that the potential upside of sodium-ion batteries as a replacement is explored. Drawbacks and highlights of sodium-ion cells are briefly discussed. The current state of lithium-ion technology is explored and briefly discussed. Experimentation on a simple magnesium-copper cell is executed with note of open-cell potential, corrosion, electrolyte, and corrosion inhibitor effectiveness. Water, alum, magnesium sulfate and acetic acid are tested as potential electrolytes for the cell. Open potential is tested using VersaStat and the Ecorr vs Time test on MS-DOS. Average potential is scaled against each electrolyte. Open potential is analyzed as a function of time.

Introducion Present day rechargeable batteries have many limitations and relatively high costs. Lithiumion batteries are the most ubiquitous of the bunch with high energy density, high capacity, low memory effect, and long cycling life. However, lithium is expensive and dangerous. With its component materials becoming harder to extract or downright impossible to satisfy increasing demand, lithium needs a replacementor at least a companion to fill in the tasks ill-suited for lithium. Sodium-ion batteries present the solution to the cost problem that plagues lithium, without much loss in energy density or life cycling. The drawbacks of sodium ion batteries and its lack of popularity have caused it to be almost completely ignored by the research and development community. Although in its current state sodium does not present an immediate solution, its unequaled abundance offers a good chance that a sodium battery could change the face of energy storage. As such, a goal to build an inexpensive battery that would light up an LED bulb for an extended period of time was conceived. The battery was to be built using cheap materials that were safe, easy to acquire, and resulted in at least 2.2 volts of potential. Magnesium and copper were chosen as electrodes; many different electrolytes were tested to measure corrosion, open potential degradation over time, and effectiveness of antifreeze as a corrosion inhibitor. The effects of antifreeze on open-cell potential over time were also noted.

Sodium-ion Batteries 3.1 The case for Sodium-ion Due to the high availability of sodium on the earths crust, sodium ($135-$165 per ton of trona) is extremely cheap when compared to lithium ($5000 per ton of lithium carbonate) [Slater]. Trona (trisodium hydrogendicarbonate dihydrate) is processed and sodium carbonate can then be extracted, making trona a cheap source for sodium ions. Also, lithium reserves in the future might be limited or dwindling and thus creating unstable and highly fluctuating prices [Egbue]. When taking analyst-projected optimistic and pessimistic estimates of lithium availability and demand, lithium falls short, where as the high abundance of sodium, as trona or seawater, means that demand of any size can be

Kevin Miranda

CSUN AIMS2 2013

satisfied. Sodium-sulfur batteries are characterized by their high coloumbic efficiency, low cost, and ease of recyclability [1].

Figure 1: Projected demand of Lithium compared to projected HEV and EV demand [3].

3.2 The drawbacks of Sodium-ion batteries Sodium-ion batteries tend to require high temperatures to operate efficiently since the melting point of sodium is around 98C[1]. Sodium-sulfur and ZEBRA (Zero-Emission Battery Research Activities) cells, which are based on Na-NiCl2, both operate at temperatures in the range of 300C due to the requirement that both the anode and sodium are molten in the cell [2]. This requires that sodium-based cells employ efficient and robust battery management systems in order to maintain peak efficiency. Also, like lithium, sodium is highly reactive and explosive in water, which leads to safety concerns. Precautionary measures must be taken when choosing sodium-cell materials. Sodium creates harsh environments within cells, which can result in corrosion that in turn may lead to short circuits, unwanted conductivity, and dendrite growth in the solid electrolyte interphase (SEI) [1]. Sodium ions are nearly three times as large as those of lithium, 23g mol-1 and 6.9g mol-1 respectively [3]. A heavier atom is harder to push across the electrodes, which equates to a lower energy density. A lower energy density essentially removes sodium-based batteries from ever being applied to the transportation sector since a heavier battery would be required to hold the equivalent charge of a lithium-based one. Furthermore, sodiums redox potential is relatively high when compared to lithium, E= -2.71V and E= -3.04V respectively [2], which makes it harder to oxidize. 3.3 Why Sodium? Battery consumers, the general populace and industrial, are price sensitive. Given the many applications and roles of batteries in modern life, including: hybrid-electric vehicles (HEV), full electric vehicles (EV), mobile computing, smart grid energy storage, and many others, finding a cheap solution to the expected increase in demand in the near future creates a big opportunity for manufacturers and suppliers to gain a foothold in the multibillion dollar battery industry. Sodiums low price and high availability make it markedly important to research and develop into a contender in the fight for battery high energy density, near-infinite life cycle, decreasing weight and volume, and high recyclability and safety.

Kevin Miranda

CSUN AIMS2 2013

3.4 Where Sodium makes sense Looking at sodiums strengths and weaknesses it makes for quite an easy argument to rule out the possibility of sodium playing a huge role in the mobile or automotive sector. The chemistry behind sodium denies it this possibility due to its weight and low energy density. However, since very little research has been given to sodium when compared to lithium, it isnt impossible to say that with the right application and chemistry sodium can eventually become efficient and dense enough to be applied in the mobile and automotive industry. But, in its current state sodium is well suited for the energy storage industry. Be it the smart grid storage on a national scale or solar energy storage at the micro-consumer level, the low cost and current research dictates that sodium is one of the best players in stationary applications. Already sodium displays a higher voltage potential than the widely used nickel-metal hydride rechargeable battery [1], a sign of sodium ion batteries potential.

Figure 2: Table of reduction potential of the both the half-cells and the full cell, respectively. 25 C [1].

3.5 The chemistry of Sodium One particular system, the sodium-sulfur rechargeable battery, uses sodium and sulfur as the cathode and anode, respectively. Ceramic alumina ( -Al2O3 ) serves as both the electrolyte and the separator in this application [1]. This cell operates at temperatures between 300 and 350 C (572 and 662 F), where sodium and sulfur both remain in a liquid state [1]. As the sodium-sulfur battery discharges a solid forms, sodium polysulfide, at the sodium cathode, which increases resistance and then eventually prevents further discharge from occurring. The design requirements of the separator and electrolyte require that it be manufactured in a solid tubular form, a difficult processing technique, so that it may have high Na+ conductivity, low discharge, high energy density, strong mechanical stability and a high resistance to corrosion.

Kevin Miranda

CSUN AIMS2 2013

Figure 3: Schematic representation of Na-S cell [2].

However, at the current operating temperature range Na-S cell batteries are prone to materials degradation and corrosion. Sodium creates a very volatile and harsh environment. The use of an inhibitor can help increase the longevity of this cell and its subcomponents. Currently a Fe-75Cr alloy inhibitor is showing promising results when sprayed on the inner wall of the electrode casing [1]. The Na-S cell in particular is not ready for mass production or practical use due to its extremely high operating temperature, difficult tubular design, high rate of capacity fade, and other issues. Advancement in an effective planar design could increase energy and power densities and ease manufacturing, but this would require the right components and materials to withstand the internal environment [1]. 4 Lithium Ion Batteries 4.1 The case for Lithium-ion Lithium is the lightest metal in the periodic table and is the most easily reduced. Such a low redox potential, -3.07V, allows for lithium-ion batteries high specific energy density, 150-190 watt-hours per kilogram [4] and an extremely high capacity per gram at 3829 mAh, compared to sodiums 1165 mAh per gram [3]. Combined with lithium-ions lack of a memory effect, or capacity fade, lithium-ion cells hold the most capacity for the most cycles with little mass. This very combination is what has made lithium-ion the most popular battery in consumer electronics and increasingly in the automotive industry. 4.2 The drawbacks of Lithium-ion batteries The most significant problem in the production and operation of lithium-ion batteries is safety, say Tsivadze, Kulova and Skundin [4]. Lithiums extremely negative redox potential makes it very reactive, especially in water, and thus makes cells prone to explosions and leakage. High reactivity is the reason why nearly all lithium-ion batteries are made using non-aqueous electrolytes [6]. Cost is lithiums second biggest drawback. Not only the cost of the lithium metal itself, but also cost is highly dependent on the price of cobalt, lithium-ion batteries most popular electrode. Tin, sodium-ions most popular electrode material, on the other hand is extremely cheap and thus keeps sodium-ion relatively inexpensive [3]. As previously
Kevin Miranda CSUN AIMS2 2013

discussed in section 1.1, lithiums high cost per kilogram, about $5000, makes it very expensive in large-scale applications such as vehicles and energy storage. 4.3 Where Lithium makes sense Lithium is perfectly suited for applications where its high cost is of little to no importance, where its energy density and capacity make the only viable option, and in small-scale storage for consumer electronics and portables. Since 1991, the first year lithium-ion was commercialized by Sony, very little has changed in the electrochemical cell. Instead focus has remained on increasing structural improvements in order to increase specific energy density [4]. This means that energy density and price is slated to remain stagnant unless a new source of lithium, thereby increasing supply, or a novel way increasing capacity, thereby decreasing demand, are discovered.

Figure 4: The battery pack of the Tesla Model S contains hundreds of lithium-ion cells [teslamotors.com]

Lithium-ions application where its price matters much less than in others is evident in the luxury sedan the Model S, produced by Tesla Motors. With the price per kilowatthour near $1000 [5], only the most affluent of consumers are able to afford it. Outside of luxury cars serving a niche market, lithium-ion has been outdone by cheaper lead-acid and nickel-metal hydride batteries found in hybrids like the Toyota Prius. However, in fully electric vehicles like the Model S, Honda Fit EV, Fiat 500e, and the Nissan Leaf, lithium-ion is the only choice available that allows these cars to operate efficiently without exceeding mass or weight restrictions. Any other type of readily available battery would make the range too low and curb weight too high. For small consumer electronics the high density, long life cycles, fast charging, and high capacities highly outweigh the price. For this reason lithium-ion is nearly the sole provider of energy for smartphones, laptops, tablets and other portables worldwide.

Kevin Miranda

CSUN AIMS2 2013

Theory: 5 Designing the inexpensive battery cell 5.1 Goals The tentative goals set forth for this experiment and research were as follows: to develop a cell or combination of cells using readily available equipment and materials, low-cost and safe electrolytes, and a resulting voltage open-cell potential in the range of 2.2-2.6 volts, enough to light an LED light bulb. 5.2 Theory and calculations The electrodes chosen were a copper, as the anode, and magnesium, as the cathode. The reasoning behind this is due to the relative position of each electrode on the emf-series table [Figure 5]. Theoretically the open-cell potential would have been expected to equal the absolute difference in the potentials listed in Figure 5: |2+ 2+ | = |2.363 0.340| = 2.703 However the above calculation takes into account a few assumptions, like 1 molar ionic solution of magnesium and copper at standard temperature and pressure. These ideal requirements were forgone during experimentation, not to lack of will, but due to practicality. Instead focus was placed on electrolytes and their effects on resulting opencell potential.

Figure 5: emf-series Table

Kevin Miranda

CSUN AIMS2 2013

5.3 Materials For the experiments, available copper and magnesium samples were used [Figure 6]. Readily available salts and liquids were used, as per research goals. All electrolytes used could easily be prepared without much cost, difficulty, or danger. For electrolyte holder and electrode separator, cotton pads were used due to their high permeability, low electrolyte leakage, and extremely low cost. Ethylene glycol, or antifreeze, was used as a low-cost corrosion inhibitor.

Figure 6: Polished Copper and Magnesium electrodes prior to testing.

5.4 Procedure First, three magnesium and three copper samples are grinded and polished using 240 grit (or lower) sandpaper in order to remove passive layer. The following aqueous solutions are then prepared in a beaker: 200 mL of approximately 1 molar glacial acetic acid consisting of 100 mL of water and 6 mL of glacial acetic acid, 200 mL of saturated alum potash consisting of 14g alum potash per 100 mL of distilled water, 200 mL of saturated magnesium sulfate (Epsom salt) consisting of 25.5g per 100 mL of distilled water, and 200 mL of pure water. pH and temperature of each solution was recorded. Cotton pads are then soaked in the electrolyte currently being tested. The cotton pad is placed between one magnesium sample [if the current test requires antifreeze, the magnesium sample must be submerged entirely in antifreeze prior] and one copper sample and pressed together so as to make good contact; this is one cell. Voltage is then immediately measured using a multimeter; this is the 1-cell voltage. The previous steps were repeated in order to build a second cell but before measuring and recording voltage both cells were stacked, alternating anodes and cathodes. Voltage was recorded again as the 2-cell voltage. A third cell was built and the 3-cell voltage recorded [Figure 7].

Kevin Miranda

CSUN AIMS2 2013

Figure 7: 3-cell stack and measuring open-cell potential.

Finally the 3-cell stack is held together using rubber bands, and an LED bulb is attached to the outermost copper (positive terminal) and outermost magnesium (negative terminal) samples. In order to insure the LED glows careful attention must be paid to match the positive terminal in the LED to the positive side of the 3-cell stack, the negative terminal vis-a-vis.

Figure 8: 3-cell stack in final setup with lit LED and attached working and reference probes.

Using the VersaStat cell, the Ecorr vs Time test is run with data collecting for 1800 seconds (30 minutes) at a data collection rate of 1 point per second (1800 points total). The working electrode (green) is affixed to the LED on the positive terminal, the terminal that shares the wire affixed to the outermost copper sample. The reference electrode (white) is affixed to the LED on the negative terminal, the terminal that shares the wire affixed to the outermost magnesium sample. All other values on the test are left at their default values. The test is then allowed to run. Data is saved once the test has completed. The 3-cell stack is disassembled, the cotton pads discarded. Note was taken of any visible corrosion. The entire procedure was repeated twice for every solution without coating magnesium in antifreeze, and twice again when using the antifreeze coating for a total of four runs for any given solution.

Kevin Miranda

CSUN AIMS2 2013

Results and Conclusions 6.1 Data

Average Potential of Single Cell

2.00 1.80 1.60 1.71 1.47 1.40 1.20 1.00 0.80 0.60 0.40 0.20 0.00
Figure 9: Bar Graph of average potentials of a single cell.

1.98 1.83 Water Acetic Acid Alum MgSO4

Potential (V)

Average Potential of Antifreeze-coated Single Cell

2.00 1.87 1.87 1.85 1.50 1.00 0.50 0.00
Figure 10: Bar graph of average potential of antifreeze-coated single cell.

Potential (V)

1.50

Water +AF Acetic Acid +AF Alum +AF MgSO4 +AF

Units (Volts) Average Run #1 Run #2

Water 1.47 1.54 1.40

Water +AF 1.50 1.34 1.65

Acetic Acid 1.71 1.66 1.76

Acetic Acid +AF 1.87 1.88 1.86

Alum 1.98 1.99 1.97

Alum +AF 1.87 1.84 1.90

MgSO4 1.83 1.87 1.78

MgSO4 +AF 1.85 1.83 1.86

Table 1: 3-cell open potential and averages.

Kevin Miranda

CSUN AIMS2 2013

10

Figure 11: Average 3-stack Open-cell Potential as a function of time.

Figure 12: Average 3-Stack Open-cell Potential as a function of time.

6.2 Observations From the data gathered in Table 1, it is evident that the electrolyte which resulted in the highest voltage was Alum, with the chemical formula of KAl(SO4)2, followed by Magnesium Sulfate (Epsom salt), MgSO4, and Acetic Acid, C2H4O2, and lastly distilled

Kevin Miranda

CSUN AIMS2 2013

11

water. Alum is clearly above the rest in terms of open-cell potential, especially with no inhibitor present. For water and acetic acid, coating the magnesium electrode in ethylene glycol (antifreeze) resulted in a higher single-cell open potential. The opposite was true for the sulfates, alum and Epsom salt, which experienced a slight decrease in potential or negligible change. Corrosion-wise, water as an electrolyte resulted in very little to no visible corrosion [Figure 13]. When antifreeze was used, corrosion was nearly non-existent on some of the electrodes [Figure 14]. For acetic acid, alum and magnesium sulfate, the corrosion pattern resembled a powdery white coating on the magnesium and a very thing oxidation layer on the copper samples. [Figure 15].

Figure 13: Electrodes after 30 minutes with water electrolyte. Pitting corrosion is slightly visible.

Figure 14: Electrodes after 30 minutes with water electrolyte and antifreeze coating. Corrosion is less evident than before.

Figure 15: Corrosion layers on magnesium electrodes after 30 minutes with acetic acid electrolyte. Corrosion layer is powdery white. Corrosion is slightly visible on copper electrodes.

Kevin Miranda

CSUN AIMS2 2013

12

Every electrolyte was capable of lighting the LED when stacked to a 3-cell, and kept it lit throughout the 30 minute test on the VersaStat machine. From the graphs a familiar sharp decrease in the first few minutes of the test was common to all electrolytes, regardless of antifreeze use or not. After the point of sharp decrease the open-cell potential decreased slowly until reaching a final point for the remainder of the test [Figures 11-12]. The tangential behavior of the graph is probably due to the corrosion on the magnesium electrodes to reach a point of passive equilibrium. Once the corrosion layer has stabilized, so does the resulting voltage. The effect of the antifreeze coating on the magnesium was evident in almost every sample tested. The antifreeze, for the most part, negatively affected the open-cell potential but it decreased the amount of visible corrosion on every single run [Figures 13-15]. 6.3 Limitations Many unaccounted-for variables were present throughout the experiments and research. These include, but are not limited to: exact alloy composition of our electrode samples, surface area of copper and magnesium electrodes, volume of electrodes, quantitative measurement of contents in electrolyte solutions, constantly changing materials, and inconsistent testing methods. 6.4 Conclusions Overall the goal of creating an inexpensive cell that would light up and LED and achieve an open potential of at least 2.2 volts was clearly achieved. No components or materials tested were expensive, hard to find, or posed any immediate risk. From the limited data gathered it can be concluded that alum is the better electrolyte as it presented a 33% increase in single-cell open potential over water and nearly a 10% increase over magnesium sulfate, the next best. In terms of corrosion, water is the clear winner with very little visible corrosion present, especially when using antifreeze as a corrosion inhibitor. Antifreeze served as a well-suited corrosion inhibitor due to its commonality and effectiveness with the harsher electrolytes. Although open potential was decreased at times, or even remained unchanged, antifreeze might serve essential when considering cell life and degradation rates. Although the stated goals did not set forth the need for prolonging cell life, further testing with many different inhibitors with the most corrosive of electrolytes might help increase the understanding of antifreeze as an effective corrosion inhibitor. With the amount of data gathered, merely two testes per electrolyte, it isnt enough to conclude that antifreeze in fact played any beneficial or detrimental role on cell life or open-cell potential. In addition, a more solid conclusion as the effectiveness or detriment of each electrolyte can only be achieved with more top-down testing, which would more clearly outline patterns and behaviors that could not be concluded with two tests per.

Kevin Miranda

CSUN AIMS2 2013

13

Works Cited [1.] Cheng, Fangyi, Jing Liang, Zhanliang Tao, and Jun Chen. "Functional Materials for Rechargeable Batteries." Advanced Materials (2011): 1695-715. Wiley. Web. 14 Aug. 2013. [2.] Ellis, Brian L., and Linda F. Nazar. "Sodium and Sodium-ion Energy Storage Batteries." Current Opinion in Solid State and Materials Science 16.4 (2012): 168-77. Science Direct. Web. 14 Aug. 2013. [3.] Slater, Michael D., Donghan Kim, Eungie Lee, and Christopher S. Johnson. "Sodium-ion Batteries." Advanced Functional Materials 23.8 (2013): 947-58. Wiley. Web. 14 Aug. 2013. [4.] Tzivadze, A. Y., T. L. Kulova, and A. M. Skundin. "Fundamental Problems of Lithium Ion Rechargeable Batteries." Protection of Metals and Physical Chemistry of Surfaces 49.2 (2013): 149-54. Springer. Web. 14 Aug. 2013. [5.] Vyrynen, Antti, and Justin Salminen. "Lithium Ion Battery Production." The Journal of Chemical Thermodynamics 46 (2011): 80-85. Science Direct. Web. 14 Aug. 2013. [6.] Yoshino, Akira. "The Birth of the Lithium-Ion Battery." Angewandte Chemie International Edition 51.24 (2012): 5798-800. Wiley. Web. 14 Aug. 2013.

Kevin Miranda

CSUN AIMS2 2013

14