Thermoacoustic Refrigeration

Jonathan Newman, Bob Cariste, Alejandro Queiruga, Isaac Davis, Ben Plotnick, Michael Gordon, and Sidney San Martn We explored the basic principles of thermoacoustic refrigeration, replicating the work of Russell and Weibull to produce a small thermoacoustic refrigerator out of readily available parts. Combined with an understanding of the underlying thermodynamics, the model enables us to spread awareness of the viability of thermoacoustic devices as refrigerators and heat pumps.

Introduction
From creating comfortable home environments to manufacturing fast and efficient electronic devices, air conditioning and refrigeration remain expensive, yet essential, services for both homes and industries. However, in an age of impending energy and environmental crises, current cooling technologies continue to generate greenhouse gases with highenergy costs. Thermoacoustic refrigeration is an innovative alternative for cooling that is both clean and inexpensive. Through the construction of a functional model, we will demonstrate the effectiveness of thermoacoustics for modern cooling. Refrigeration relies on two major thermodynamic principles. First, a fluids temperature rises when compressed and falls when expanded. Second, when two substances are placed in direct contact, heat will flow from the hotter substance to the cooler one. While conventional refrigerators use pumps to transfer heat on a macroscopic scale, thermoacoustic refrigerators rely on sound to generate waves of pressure that alternately compress and relax the gas particles within the tube. The model constructed for this research project employed inexpensive, household materials. Although the model did not achieve the original goal of refrigeration, the experiment suggests that thermoacoustic refrigerators could one day be viable replacements for conventional refrigerators.

as light, sound, or water waves. In a closed tube, columns of air demonstrate these patterns as sound waves reflect back on themselves after colliding with the end of the tube. When the incident and reflected waves overlap, they interfere constructively, producing a single waveform. This wave appears to cause the medium to vibrate in isolated sections as the traveling waves are masked by the interference.1 Therefore, these standing waves seem to vibrate in constant position and orientation around stationary nodes. These nodes are located where the two component sound waves interfere to create areas of zero net displacement. The areas of maximum displacement are located halfway between two nodes and are called antinodes. The maximum compression of the air also occurs at the antinodes. Due to these node and antinode properties, standing waves are useful because only a small input of power is needed to create a large amplitude wave. This large amplitude wave then has enough energy to cause visible thermoacoustic effects. All sound waves oscillate a specific amount of times per second, called the waves frequency, and is measured in Hertz. For our thermoacoustic refrigerator we had to calculate the optimal resonant frequency in order to get the maximum heat transfer rate. The equation for the frequency of a wave traveling through a closed tube is given by:

f =

v 4L

Sound Waves and Pressure

Thermoacoustics is based on the principle that sound waves are pressure waves. These sound waves propagate through the air via molecular collisions. The molecular collisions cause a disturbance in the air, which in turn creates constructive and destructive interference. The constructive interference makes the molecules compress, and the destructive interference makes the molecules expand. This principle is the basis behind the thermoacoustic refrigerator. One method to control these pressure disturbances is with standing waves. Standing waves are natural phenomena exhibited by any wave, such

where f is frequency, v is velocity of the wave, and L is the length of the tube.

Figure 1: Shows the relationship between the phase of the wave, the pressure, and the actual arrangement of the molecules. The black line shows the phase of the sound wave, the red shows the pressure and the dots below represent the actual molecules. From Reference 2

GSET Research Journal 2006

Thermodynamics, Heat Cycles, and Heat Pumps

The second fundamental science behind thermoacoustics is thermodynamics, the study of heat transfer. The Ideal Gas Law states that the pressure on a gas is directly proportional to absolute temperature, or, as the pressure on a gas increases, the temperature increases. On a microscopic scale, the gas particles in a system will collide more frequently if the temperature is increases or if the volume is reduced. The basic thermodynamic cycles rely on this relationship between temperature and pressure. In any heat cycle, gases will expand and contract, circulating heat throughout the system. These movements of kinetic energy can be used to do work. Depending on how the heat oscillations are controlled, different heat cycles become more efficient, involving less loss of heat from the system. Thermoacoustic refrigerators use variations of these cycles to pump heat. The Carnot Cycle The most efficient cycle of thermodynamics, the Carnot cycle, takes advantage of this principle of gas expansion. The Carnot cycle uses gas in a closed chamber to extract work from the system. In engines, pistons are used to output work. The cycle begins with the piston in its rest position. Heat from an outside source is transferred to the gas via an isothermal process where the temperature does not change. By the ideal gas law, the gas expands, pushing the piston to its extended position. This is an adiabatic process where no heat is transferred into or out of the cylinder. The heated gas then transfers the heat to a low temperature container doing work on the surroundings. This is also an isothermal process.

The surroundings now do work on the system, adiabatically compressing the gas and allowing the piston to fall back to its rest position. However, because it is easier to compress the cooler gas than to add heat to the warm gas, net work is done on the surroundings. To determine the efficiency of the cycle, the total useful work done is compared to the total heat transferred. In Figure 3, the total heat transferred equals the red area plus the white area. The work extracted from the system is represented by the white area. Even the Carnot cycle, the ideal thermodynamic process where each step is reversible and involves no change in entropy,2 transfers more heat than it does work. However, the Carnot cycle has the best work output with the given temperature difference and entropy difference, so it is defined to be 100% efficient. The Stirling Cycle The Stirling cycle is a variation of the Carnot cycle, but unlike the Carnot cycle, an engine can actually be constructed that effectively utilizes the Stirling method of heat transfer. In a Stirling engine, an external heat source (often external combustion) transfers heat into the gas in the chamber. As in the Carnot cycle, the gas expands, pushing the piston to its extended position. The chamber into which the gas expands, however, has a heat sink, usually consisting of metal fins, through which the heat in the expanded gas can dissipate into a cooler chamber. The gas then compresses, and the piston returns to its rest position. A Stirling engine is useful because it can be powered by almost any external heat source, such as solar power, nuclear power, or conventional combustion.34 Both the Stirling cycle and Carnot cycle involve the following basic thermodynamic cycle: heat enters from a hot container, work comes out of the engine (i.e. moving a piston), and, as a result, the heat is dissipated into a cooler container.

Figure 2: P-V diagram of the Carnot cycle. From Reference 3.

Figure 3: T-S diagram showing the four stages in the Carnot cycle. From reference 4.

GSET Research Journal 2006

thermoacoustic refrigerator cycle. The left end is towards the closed end of the resonator tube. From Reference 5.

Figure 4: Shows the cylinder positions of the four steps of the Carnot cycle. From reference 3.

A heat pump, or refrigerator, operates on the same basic cycle as a heat engine, only in reverse. A heat pump requires an input of work to transfer heat from a cooler container to a hotter one. This heatpump cycle is the basic mechanism by which our refrigerators will work.

Thermoacoustics
Thermoacoustics combines the branches of acoustics and thermodynamics together to move heat by using sound. While acoustics is primarily concerned with the macroscopic effects of sound transfer like coupled pressure and motion oscillations, thermoacoustics focuses on the microscopic temperature oscillations that accompany these pressure changes. Thermoacoustics takes advantage of these pressure oscillations to move heat on a macroscopic level. This results in a large temperature difference between the hot and cold sides of the device and causes refrigeration. The most important piece of a thermoacoustic device is the stack. The stack consists of a large number of closely spaced surfaces that are aligned parallel to the to the resonator tube. The purpose of the stack is to provide a medium for heat transfer as the sound wave oscillates through the resonator tube. A functional cross section of the stack we used is shown in figure 6. In typical standing wave devices, the temperature differences occur over too small of an area to be noticeable. In a usual resonator tube,

heat transfer occurs between the walls of cylinder and the gas. However, since the vast majority of the molecules are far from the walls of the chamber, the gas particles cannot exchange heat with the wall and just oscillate in place, causing no net temperature difference. In a typical column, 99% of the air molecules are not near enough to the wall for the temperature effects to be noticeable. The purpose of the stack is to provide a medium where the walls are close enough so that each time a packet of gas moves, the temperature differential is transferred to the wall of the stack. Most stacks consist of honeycombed plastic spacers that do not conduct heat throughout the stack but rather absorb heat locally. With this property, the stack can temporarily absorb the heat transferred by the sound waves. The spacing of these designs is crucial: if the holes are too narrow, the stack will be difficult to fabricate, and the viscous properties of the air will make it difficult to transmit sound through the stack. If the walls are too far apart, then less air will be able to transfer heat to the walls of the stack, resulting in lower efficiency. Thermoacoustic Cycle The cycle by which heat transfer occurs is similar to the Stirling cycle. Figure 55 traces the basic thermoacoustic cycle for a packet of gas, a collection of gas molecules that act and move together. Starting from point 1, the packet of gas is compressed and moves to the left. As the packet is compressed, the sound wave does work on the packet of gas, providing the power for the refrigerator. When the gas packet is at maximum compression, the gas ejects the heat back into the stack since the temperature of the gas is now higher than the temperature of the stack. This phase is the refrigeration part of the cycle, moving the heat farther from the bottom of the tube. In the second phase of the cycle, the gas is returned to the initial state. As the gas packet moves

GSET Research Journal 2006

back towards the right, the sound wave expands the gas. Although some work is expended to return the gas to the initial state, the heat released on the top of the stack is greater than the work expended to return the gas to the initial state. This process results in a net transfer of heat to the left side of the stack. Finally, in step 4, the packets of gas reabsorb heat from the cold reservoir to repeat the heat transfer process. Penetration Depth An essential variable in building a thermoacoustic refrigerator is the spacing between the walls of the stack. If the walls of the stack are too close, the sound cannot pass through the stack efficiently since the viscous properties of air prevent the air from vibrating. If the walls are too far apart, the process described above cannot occur, since gas packets are too far away from the wall to effectively transfer heat. According to G.W. Swift, the ideal spacing in a stack is 4 thermal penetration depths.6 The thermal penetration depth is the distance heat can diffuse in a gas over a certain amount of time. For example, if a block of aluminum is at a constant low temperature and suddenly one side is exposed to a high temperature, the distance that the heat penetrates the metal in 1 second is the heat penetration. As time passes, the heat penetrates farther into the material, increasing the temperature of the interior sections. However, since sound waves are constantly oscillating between the roles of heat source and heat sink, the thermal penetration depth is roughly constant. The thermal penetration depth for an oscillating heat source is a function of the frequency of the standing wave, f , the thermal conductivity, , and density, , of the gas, as well as the isobaric specific heat per unit mass of the gas, c p , according to the equation: 5

Figure 6: The basic design of our thermoacoustic refrigerator. The design of the stack is shown on the left, which consists of closely spaced film walls. The design on the right shows the entire setup, with the loudspeaker, stack, and plug. The thermocouples (not shown) were inserted one above the stack and one below the stack.

Tcrit =

p cp

where p is the acoustic pressure and is the acoustic displacement amplitude. The variation in local wall temperature is represented by 2 Tcrit over the maximum displacement of the gas molecules. The maximum temperature variation caused by the sound waves is 2 p / c p . If these two quantities are equal, the critical temperature is reached and no heat is transferred. This temperature is important in determining the properties of a thermoacoustic device, since efficiency depends on a temperature differential caused by the sound waves that is larger than the critical temperature so that a large cooling effect is created.

k =

f c p

Procedure
To create the thermoacoustic refrigerator we followed a plan similar to one designed by Russell et al.5 We began by creating the stack, which we constructed with film and fishing wire. The design specified an optimal thermal penetration depth of 4; however, we were restricted by material constraints and achieved an acceptable penetration depth of 2.5. This was achieved with 15-lb nylon fishing wire with a diameter of 0.34 millimeters. The design called for the fishing line to be placed in 5 mm separations, so we each created looms as templates for stack. The looms were created from 5 cm wide cardboard roughly 33 cm in length with slits placed every 5 mm on each side. This allowed for a straight application of the fishing line to the film. The film was then cut to a meter in length and taped to a flat surface. The

Critical Temperature The critical temperature is the temperature at which no heat will be transferred through the stack. If the temperature difference induced by the sound wave is greater than this critical temperature, the stack will function as a refrigerator, transferring heat from the cold end of the tube to the warm end. If the temperature is less than the critical temperature then the stack will function as an acoustic engine, moving heat from the warm region to the colder region and creating sound waves. The function for the critical longitudinal temperature gradient is5

GSET Research Journal 2006

loom wound with fishing wire and the film were both sprayed with spray adhesive. The line was then applied to film with an even distribution of weight in order to insure optimal strength. After 10 minutes the glue was cured enough to cut the line from loom. However, while removing the line from the loom, we had to take care not to remove it from the film. This process was repeated approximately 3 times until the entire meter of film was covered with fishing line. We allowed the stack to cure for several more days in order to ensure the best quality. We then machined 8 by 8 inch squares of Plexiglas with a 2.5 cm whole in the center. This would later serve as the base. We then cut 2 cm diameter Plexiglas tube into lengths of roughly 25 cm. The edges of the tubes where then beveled to allow for proper sealing. The tube was then inserted and glued into the hole in the center of the Plexiglas squares. After the glue dried we then attached the Plexiglas square to a 6-inch, low-range speaker. The hole in the square was centered over the speaker and silicon caulk was used to properly seal the connection. After the glue dried we took the completed stack and coiled the film tightly enough to fit into the 2.5 cm diameter tube. We then pushed the stack down 4 cm as specified from the top of tube using a machined 2.5 cm tool designed not to destroy the stack. We then created the two thermocouples required for determining the temperature gradient. To accomplish this, we took a length of 0.015 cm chromel and a length of 0.015 cm alumel and welded them together to create each thermocouple. We then drilled holes 1 cm below and 1 cm above the stack where the thermocouples were inserted into the center of the tube. We then used the thermocouple calibration device to ensure that the thermocouples were working. We then sealed the tube by plugging the end with an aluminum cap.

Figure 7: The final modified thermoacoustic device with heat sink.

Next, we connected a frequency generator to a 40w amplifier and connected this to the speaker via a BNC to RCA connector. We then determined the impedance of the speakers and used this knowledge to properly connect the speaker to the amplifier with 8-gauge wire. After, we inserted the machined aluminum stoppers, roughly 2.5 cm in diameter, which were needed to create the closed tube necessary for standing waves. Next, we determined the proper frequency needed to achieve a standing wave. This is supposed to be at the first harmonic, or, when the wavelength is 4 times the length of the tube. We measured the length of the tube to the bottom of the aluminum cap and multiplied by 4 and divided the speed of sound at room temperatures, or roughly 349 m/s, by the length of tube. We determined the frequency to be around 340 Hz. Once we found the optimal resonant frequency for our refrigerator, we used a tone generator, which outputs a sound at a specific frequency. Basically, the generator vibrates the speaker cone at that frequency, which subsequently vibrates the air and causes the heat transfer to occur. We listened to verify for the sound of the harmonic, and when verified, we increased the intensity of the frequency and then recorded the temperatures of the two thermocouples.

GSET Research Journal 2006

Possible Modifications
One of the major problems that we had was the heat build up at the top of the tube. Most of the possible modifications we thought of involved of dissipating the heat from the top of the tube. In the original design the aluminum plug was responsible for conducting heat from of the top end of the tube into the surrounding air. However, the aluminum proved to be unable to dissipate enough heat, because as we ran the experiment the temperature of the bottom section soon reached room temperature. This is because heat will only be transferred from the cold region of the bottom end to the hot region of the top when the temperature gradient created by the sound waves is greater than the temperature difference between these two regions. When too much heat is in the system the bottom temperature stays at the surrounding temperature, while the area on the top of the tube becomes very hot. One possible way to dissipate more heat is to increase the surface area of the cap by cutting grooves into each end of the aluminum plug. The increased surface area gives air particles a larger area to collide into the aluminum plug and transfer heat, allowing for there to be more collisions at a single time, thus increasing the rate of heat conduction of the aluminum plug from the top end of the tube into the surrounding air. The grooved aluminum plug will decrease the temperature in the top end of the tube by dissipating heat faster than the flat aluminum plug could. This will decrease the temperature difference between the top end and the bottom end, allowing the bottom end to become colder than with the flat plug before the temperature difference reaches the point that it exceeds the temperature gradient created by the sound waves and heat can no longer be transferred. Unfortunately, we were unable to test this modification since we did not have the appropriate machine tools. Another possible method of dissipating the heat from the refrigerator would involve heat absorption by water. Thin pipes could be run across the top end of the stack. Liquid could flow through the stack, effectively transferring the excess heat from the system. Water, with a relatively high heat capacity, would absorb the heat quickly. The hot water could then be used for other applications, such as spinning a turbine in a generator or an engine. This would be using the device as a heat pump to power a device. Unfortunately we did not have the tools to try this modification as well. Instead, in our experiment we tried to create a heat sink around the top end of the tube, to try and create effective refrigeration on the bottom of the tube. We wrapped copper around the top of the pipe

to increase the surface area of the pipe and dissipate more heat. However, as described in the results below, this was largely ineffective, probably because the acrylic pipe was unable to effectively transfer heat. Another method for improving the heat transfer within the tube would be to use helium as the sound medium. Of all gases, helium behaves the most like an ideal gas in that diatomic helium molecules exhibit weak electrostatic attractions upon each other. Due to low dispersion forces and a low atomic mass, helium molecules have a greater thermal conductivity than any other gas except for hydrogen. Hydrogen gas, however, is reactive, whereas helium is relatively inert. Using helium as a sound medium, though beyond the reach of this study, would only require an evacuated airtight tube that would then be pressurized with the helium gas. This process is used in many professional-grade thermoacoustic refrigerators where helium gas has led to increased efficiency and heat transfer across the stack.7 Unfortunately we could not try this modification because of a lack of materials.

Results
We successfully created a thermoacoustic heat pump. The problem with this, however, is that we were attempting to create a thermoacoustic refrigerator. Our results showed that we were able to create a high temperature gradient above room temperature, but were unable to significantly cool the air. We tested three thermoacoustic refrigerators that we built. Two of these devices were created from the instructions in the procedure section without any modifications. The third had the heat sink modification that was described in the modifications section. We collected the data for these results by sampling the temperatures at the top and bottom thermocouples of the refrigerators as they ran every ten seconds, stopping when it became apparent that there would be no more significant change. In the three tests performed on the devices, the temperature difference between the hot end and cold end of the stack increased rapidly for the first thirty seconds and reached the final value after around ninety seconds of operation. The unmodified models exhibited temperature differences of 11C and 14C after thirty seconds with final temperature differences of 19C and 20C. The model with the heat sink created a larger temperature difference, reaching a temperature gradient of 22C after thirty seconds with a final temperature difference of 28C. The three tests also showed similar trends in the absolute temperatures of the top and bottom of the

GSET Research Journal 2006

Temperature of unmodified model

60 Temperature (C) 50 40 30 20 10 0 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320

allowing the excess heat to dissipate to the surroundings. However, our device did demonstrate that thermoacoustic device have the ability to create and maintain a large temperature gradient, more than 20 degrees Centigrade, which would be useful as a heat pump.

Applications
Thermal management has always been a concern for computer systems and other electronics. Computational speeds will always be limited by the amount of noise produced by computer chips. Since most noise is generated by waster heat, computer components and other semiconductor devices operate
Temperature of Second Unmodified Model
60 Temperature (C) 50 40 30 20 10 0
20 40 60 80 10 0 12 0 14 0 16 0 18 0 20 0 22 0 24 0 0

Time (s)

Figure 8: The unmodified model data. The top red bar shows the readings of the warm thermocouple. The bottom blue bar shows the readings for the cooler thermocouple.

Temperature (C)

stack. The bottom of the stack, the cold end, became colder during the initial rapid rate of temperature change. When the rate of change of the temperature difference began to decrease, and the hotter end started transferring heat back to the cooler one, the temperature of the bottom end of the stack began to increase. For the unmodified models, the maximum temperature gradient was 15C. For the modified model with the heat sink, the temperature difference was 22C. For the two unmodified models, the temperature of the bottom of the stack gradually increased until it reached room temperature. The model with the heat sink actually exceeded room temperature by 3C. Based on this data, the refrigerator was able to generate a large temperature gradient. However, as time went on, the refrigerator returned to room temperature, or even slightly above it. This can be attributed primarily to unwanted heat diffusion of the top end of the stack to the outside of the tube and to the cooler end of the tube. Heat may have also diffused through the acrylic which would have brought the cool section back up to room temperature. Additionally, the vibration of the speakers could have added heat to the cool part of the device, thus adding enough heat to result in no long term cooling effect. Therefore, over the long term, we were unable to cool the air significantly, and instead created a 20 degree centigrade heat pump.

Time (s)

Figure 9: The unmodified model data for the second refrigerator. The top red bar shows the readings of the warm thermocouple. The bottom blue bar shows the readings for the cooler thermocouple.

faster and more efficiently at lower temperatures.8 If thermoacoustic cooling devices could be scaled for computer applications, the electronic industry would realize longer lifetimes for microchips, increased speed and capacity for telecommunications, as well as reduced energy costs.9
Modified Model Temperature
70 60 50 40 30 20 10 0
20 40 60 80 0 10 0 12 0 14 0 16 0 18 0 20 0

Conclusion
Our device worked as a proof of concept device showing that a thermoacoustic device is possible and is able to cool air, abet for only a short period of time. If we were able to build the device with better materials, such has a more insulating tube, we might have been able to get better results. In order to create a working refrigerator we probably would have to attach a heat sink to the top of the device, thus,

Time (s)

Figure 10: The modified model data. The top red bar shows the readings of the warm thermocouple. The bottom blue bar shows the readings for the cooler thermocouple. As shown in the diagram, the actual temperature difference was slightly greater in this design, but not significantly different.

GSET Research Journal 2006

Although this project was specifically designed to test the effectiveness of thermoacoustic refrigeration for electronic devices, low-cost, highefficiency cooling devices have broad applications in commercial industries and households. Research conducted by Professor Steven Garrett at Pennsylvania State University has yielded reliable air conditioning devices used in submarines and space shuttles.10 However, future applications of thermoacoustic air conditioners would not be restricted to industrial uses but could offer inexpensive heating and cooling for homes. Additionally, since current air conditioners use HFCs and other potentially harmful chemicals, thermoacoustic cooling systems that employ inert gases would have long-term benefits on the environment.10 One thermoacoustic device could potentially operate an entire households air conditioner, water heater, and furnace, eliminating the need for natural gases and oils.

Ben and Jerrys Ice Cream, in collaboration with Professor Garretts research team, has begun production of thermoacoustic freezers to keep its ice cream cold. Investing over $600,000 in Garretts program, Ben and Jerrys has already placed the freezers in many of its New York stores.11 The ice cream companys experiment has successfully demonstrated the viability of thermoacoustic refrigeration.

Acknowledgments
The Governors School of Engineering and Technology for providing this opportunity to be exposed to conducting small scale research Professor Stephen Tse, Advisor Dr Galiang Sun, Graduate student Megan Smith, Graduate student Stephanie Chen, GSET counselor Mr. Anthony Welch, Program director GSET All the Governors School counselors and staff

GSET Research Journal 2006

Standing Waves. Rod Nave, Georgia State University. Available: http://hyperphysics.phyastr.gsu.edu/hbase/waves/standw.html. 17 July 2006. 2 http://hyperphysics.phy-astr.gsu.edu/hbase/thermo/carnot.html 3 http://www.howstuffworks.com/stirling-engine.htm 4 http://en.wikipedia.org/wiki/Carnot_cycle 5 Daniel A. Russell and Pontus Weibull, Tabletop thermoacoustic refrigerator for demonstrations, Am. J. Phys. 70 (12), December 2002. 6 G. W. Swift, Thermoacoustic engines and refrigerators, Phys. Today 48, 22-28 (1995) 7 http://www.rolexawards.com/laureates/laureate-36-lurie_garrett.html 8 Thermal Management of Computer Systems Using Active Cooling of Pulse Tube Refrigerators. H.H. Jung and S.W.K Yuan. Available: http://www.yutopian.net/Yuan/papers/Intel.PDF. 17 July 2006. 9 Thermoacoustic Refrigeration for Electronic Devices: Project Outline. Stephen Tse, 2006 Governors School of Engineering and Technology. 10 Frequently Asked Questions about Thermoacoustics. Penn State Graduate Program in Acoustics. Available: http://www.acs.psu.edu/users/sinclair/thermal/tafaq.html. 17 July 2006. 11 Chilling at Ben & Jerrys: Cleaner, Greener. Ken Brown. Available: http://www.thermoacousticscorp.com/news/index.cfm/ID/4.htm. 17 July 2006.