Malone Refrigeration: an Old Solution to a New Problem

Malone Refrigeration An Old Solution to a New Problem

Gregory W. Swift he refrigeration and air- Tconditioning industries are in crisis due to the discovery that chlorofluorocarbons (CFCs), the working fluid in virtually all cooling equipment, cause unacceptable environmental damage. A few of us in the Laboratory’s Condensed Matter and Thermal Physics Group are trying to help resolve the crisis by developing a new

Malone’s first liquid-based heat engine, 1925

112 Los Alamos Science Number 21 1993 Malone Refrigeration/CFCs and Cooling Equipment

CFCs and Cooling Equipment: The Size of the Problem

s recently as ten years ago, dollars per year. Kitchen refrigera- their chlorine to the stratosphere, Aonly a few experts appreciated tors alone use 8 percent of the na- and hydrofluorocarbons, which the potentially disastrous conse- tion’s electricity. Most of that contain no chlorine at all. One quences of releasing chlorofluoro- electricity is produced by burning German manufacturer is using or- carbons (CFCs) into the environ- fossil fuels. Congress is accord- dinary hydrocarbons (a mixture of ment. Today it is generally be- ingly requiring ever more efficient propane and butane) as working lieved that chlorine from CFCs is cooling equipment; for example, fluids, recognizing that their flam- destroying stratospheric ozone, kitchen refrigerators built in 1993 mability poses negligible danger which shields the earth from the must use 30 percent less electricity since only small quantities are sun’s ultraviolet radiation, and that than those built in 1990. Clearly used. the resulting increased ultraviolet the elimination of CFCs should not But these new chemicals have radiation will soon lead to millions involve making refrigerators that disadvantages. They are less com- of fatal and nonfatal skin cancers are much less efficient than present patible with lubricants than are and cataracts, to other threats to models. Therefore the use of CFCs CFCs, so they may cause present human health, and to severe dam- in cooling is the most difficult of compressors to wear out more age to agriculture and ecosystems. their common uses to eliminate. quickly. HCFCs and HFCs also To mitigate those effects, the Unit- The cooling industry is enor- significantly reduce the efficiency ed States and most other countries mous. Cooling equipment worth of cooling machinery. They are have committed themselves, $40 billion is sold in the United also greenhouse gases, with rough- through the 1987 Montreal Proto- States each year, and, since it has a ly 1000 times the global-warming col on Substances That Deplete the long useful lifetime, the total value potential of carbon dioxide per Ozone Layer and its later revi- of installed equipment is about molecule, so their use will proba- sions, to rapid elimination of CFC $200 billion. Thus the appliance bly eventually be limited by inter- production. industry and other cooling-equip- national agreements. Finally, as The rate of CFC production, ment manufacturers face a daunt- they do not occur in nature, their though being reduced, is on the ing challenge. The prospects of release into the environment in order of a million tons per year. millions of fatal cancers, tens of million-ton quantities, like the re- CFCs are used extensively as millions of cataracts, and contin- lease of CFCs, will be an experi- working fluids in refrigerators and ued enormous energy consumption ment in atmospheric chemistry air conditioners, as cleaning sol- and attendant greenhouse-gas emis- with unpredictable consequences. vents in electronics and sheet- sions have led to government regu- Completely different cooling metal fabrication, and as foaming lations that are driving a $40-bil- technologies that don’t use any of agents in foam insulation and cush- lion-per-year industry to an un- these chemicals are still needed. ions. CFCs have the advantage for precedented crisis. The situation Many are being developed, includ- many purposes of being almost dwarfs the problems of the DOE’s ing Rankine cycles using CO2, chemically inert (their behavior in nuclear-weapons complex, a mere Stirling cycles using helium, and the upper atmosphere is an excep- $12-billion-per-year industry re- cooling by the Peltier (thermoelec- tion); in particular they present no sponsible for less environmental tric) effect. Almost by accident fire or poison danger. In addition and human-health damage. three new cooling technologies— their lack of interaction with the Stopgap solutions to the CFC thermoacoustic refrigeration, the lubricants used in refrigerator com- crisis are required immediately and related Sonic Sompressor, and pressors improves the efficiency are, in fact, in progress. Industry Malone refrigeration—have been and the lifetime of the refrigerator. has begun extensive recycling of developed in part here at the Labo- Cooling-system efficiency has CFCs, especially in air-conditioner ratory; each may become part of major economic and environmental repair. Some new appliances will intermediate or long-term solutions impacts. Cooling consumes about soon use hydrochlorofluorocar- to the CFC problem. They are 20 percent of the nation’s electrici- bons, which tend to break down described in the accompanying ty, at a cost of tens of billions of and then rain out before carrying articles.

1993 Number 21 Los Alamos Science 113 Malone Refrigeration

type of cooling machinery. Its de- the mature descendants of those small thermal-expansion coefficients sign takes advantage of modern fab- nineteenth-century inventions, work (part of a larger misconception that rication techniques and environmen- unobtrusively and reliably—but not liquids resemble idealized hydraulic tally benign materials but is inspired so well that further improvement is fluid) and therefore do not couple by a turn-of-the-century invention. impossible. heat to work well enough for use in Refrigerators and air conditioners In the 1970s, John Wheatley, then engines and heat pumps. In heat are based on heat pumps, machines a physics professor at the University pumps, working fluids must cool as that exploit mechanical power to of California, San Diego, and at the the fluid is depressurized in order to pump heat from low temperature to height of a distinguished career of absorb heat from the area to be high temperature. They operate by academic research into the proper- cooled; in engines, working fluids taking a working fluid through cy- ties of liquid helium, became inter- must expand on heating in order to cles of temperature and pressure ested in improving the efficiency of do work. The cooling on depressur- changes in which the working fluid such ubiquitous heat-engine machin- ization and thermal expansion of a absorbs heat at low temperature ery. But there was trouble at UCSD. material are both proportional to one (from the air inside the refrigerator, Some faculty thought that the work thermodynamic property: its ther- say) and loses (“rejects”) heat at the was “not really physics,” and mal-expansion coefficient. The higher temperature in the room. Wheatley shared their concern that thermal-expansion coefficients of Figure 1 shows an example of such heat-engine research would not pre- gases are large, those of liquid-gas a thermodynamic cycle, the Rankine pare graduate students for a tradi- mixtures at the boiling point are es- cycle used in all present household tional career in academia. Further- sentially infinite (a fact that ex- refrigerators, in which the cooling more, the research required more plains why evaporation and conden- (absorption of heat at low tempera- sophisticated fabrication techniques sation are so useful in the Rankine ture) is produced by the evaporation than were available at universities. cycle), but those of liquids are usu- of the CFC working fluid. Revers- So (with much encouragement from ally small. However, as Figure 2 il- ing the cycle of a heat pump makes Jay Keyworth, then leader of the lustrates, near their critical points a heat engine, which converts heat Laboratory’s Physics Division) liquids do have large thermal-expan- to mechanical power as heat flows Wheatley moved to Los Alamos in sion coefficients, larger in fact than from high temperature to low. For early 1981 and assembled a team that of an ideal gas. (The critical example, the steam turbines that that included Al Migliori, Tom point is the temperature, Tcritical, drive electric generators in large Hofler, Heikki Collan, and me, to and pressure, Pcritical, above which power plants use the reverse of the begin fundamental investigations of the liquid and gas phases of a sub- Rankine cycle shown in Figure 1. old and new concepts in the fields of stance are indistinguishable.) Water is the working fluid; its ex- refrigeration and power generation. Hence liquids can indeed serve as pansion on evaporation drives the One of our research areas is de- working fluids in engines and refrig- turbines. scribed in “Thermoacoustic Engines erators. Heat pumps and engines are and Refrigerators.” Liquids also have advantages among the wonderful machines de- This article discusses another area over gases as working fluids. For veloped during the nineteenth centu- of our investigations: refrigerators instance, as shown in Figure 3, liq- ry, when pistons, crankshafts, fly- and heat engines that use liquids, uids are far less compressible than wheels, and automatically timed without change of phase to gas, as gases; that is, a given volume valves began to replace the labor of the working fluid. We called such change causes a larger pressure horses, oxen, and people. Engines devices Malone refrigerators and en- change in a liquid than in a gas. removed water from mines and later gines after the first engineer to build Large pressure changes are desirable propelled ships and trains. Refriger- such machines (see “John Malone because the heat transferred in the ators preserved beef on the two- and the Invention of Liquid-Based heat exchangers is proportional to month voyage from Argentina or Engines”). Malone’s ideas had been the pressure change in the previous New Zealand to England. Today, the ignored for fifty years, in part be- step. Low compressibility allows internal-combustion engines in cars cause of a common misconception fractional pressure changes to be and the heat pumps in refrigerators, among scientists that liquids have made large with modest fractional

114 Los Alamos Science Number 21 1993

Malone Refrigeration

RANKINE REFRIGERATOR RANKINE CYCLE

Liquid Qhot Cooling Cold heat Qcold exchanger in hot heat Condensation in exchanger hot heat exchanger 2 9 Free expansion Hot heat in flow impedance exchanger Flow 3 impedance Evaporation Compression in cold heat in compressor exchanger 1 Compressor Pressure (atmospheres) Piston 4 0.6 Win Qhot Incidental warming Qcold Vapor

230 310 W in Temperature (kelvins)

Conservation of energy: Qhot = Qcold + Win. That is, heat rejected (at high temperature) equals heat added (at cold temperature) plus work done by the compressor.

Figure 1. The Rankine Cycle in a Household Refrigerator The heat pumps in household refrigerators and nearly all other present cooling equipment use the Rankine cycle. This design has been by far the most popular for decades because it is simple and reliable: Many refrigerators run for thirty years with little or no maintenance, and the cost to buy and to run them is low. On the left is a sketch of the heat pump in a kitchen refrigerator. As the working fluid flows through the heat pump in a continuous loop, each element of the working fluid helps cool the refrigerator compartment by going through the thermodynamic cycle shown on the right, the Rankine cycle. The cycle requires work to be done on each fluid element by compressing it in order to raise its temperature above room temperature. The cycle has four steps: (1) The power piston in the compressor does work Win on the element of working fluid, which is in vapor form, by greatly increasing its pressure. As the vapor is compressed, its temperature rises above room temperature. The high pressure is what drives the working fluid around the heat pump. (2) In the hot heat exchanger the hot vapor condenses as it rejects an amount of heat Qhot to the air in the room. The pressure of the fluid remains constant. (3) The working fluid, initially all in liquid form, cools to below the temperature inside the refrigerator compartment by undergoing a free expansion in the flow impedance, a narrow tube that resists the flow of the working fluid so that it emerges at the pressure required for the next step. (4) In the cold heat exchanger the cold liquid absorbs heat from the refrigerator compartment by evaporating. The vapor absorbs a little additional heat on its way to the compressor; the total heat absorbed is Qcold. The pressure of the fluid remains constant. volume changes of the liquid. working fluids is heat capacities per for Malone refrigerators and engines Therefore the volume change creat- unit volume that are orders of mag- include those that both absorb heat ed by the power piston can be small nitude larger than those of gases at from and reject heat to water. compared to the volume available in the pressures typically reached in re- Among practical examples are the heat exchangers. Thus either the frigerators and engines. Therefore water-cooled water chillers provid- power piston and other mechanical when the working fluid is a liquid, ing air conditioning for large build- components involved in volume the volume of working fluid that ings and perhaps ocean thermal en- changes can be smaller and simpler must flow through a heat exchanger ergy conversion, in which the tem- than in a gas engine, or the heat ex- is orders of magnitude less. As a re- perature difference between the changers can be more capacious and sult, far less mechanical power is re- surface and the depths of the ocean consequently more efficient. The quired to pump a liquid through the is used to produce mechanical power low compressibility of liquids also heat exchanger, and heat exchangers and ultimately electricity. leads to low stored elastic energy for liquids can be far smaller. Be- On the other hand, the high heat per unit volume when they are pres- cause the heat exchanger can be par- capacities of liquids also present a surized, diminishing the hazard of ticularly compact if it transfers heat problem. Compressing or depressur- explosions. to or from another liquid stream, izing a liquid without transferring Another advantage of liquid particularly promising applications heat changes its temperature rela-

1993 Number 21 Los Alamos Science 115

Malone Refrigeration

tively little, but refrigeration re- later to warm the liquid flowing 1 quires that the working fluid under- from the cold heat exchanger to the Ideal gas go relatively large temperature hot heat exchanger. This process is changes. In particular, as the work- called regeneration. Critical ing fluid flows from the hot heat ex- The first thermodynamic cycle to temperature κ changer to the cold heat exchanger, use regeneration was the Stirling P its temperature must decrease from cycle, invented for engines by the Liquid CO2 room temperature to below the tem- Reverend Robert Stirling in 1816. perature inside the refrigerator. As shown in Figure 4, a Stirling heat Cooling a liquid through such a pump or engine differs from a Rank- 0 wide temperature range must be ac- ine heat pump or engine in including 220 240 260 280 300 complished by removing heat from a regenerator, which consists of Temperature (kelvins) the liquid in addition to depressuriz- walls that bound a number of narrow Figure 3. Compressibility of CO2 ing it. The most efficient way to re- channels through which the working near Tcritical move the heat is to store it and use it fluid flows from the hot heat ex- The compressibility, denoted by ∑ , is de- Refrig fig3 (@V =@P)T =V . P∑ changer to the cold heat exchanger fined as 3/5/93Here, (the di- 5 and back. (A Stirling machine also mensionless compressibility) of CO2 is differs from a Rankine machine in plotted versus temperature at the critical 4 that the working fluid flows back pressure of CO , 73 atmospheres. Plot- Liquid CO 2 2 and forth rather than around a con- ted in gray is P∑ for an ideal gas, which, 3 tinuous loop.) The temperature of like TÆ , is equal to unity at any tempera- α

T the channel walls decreases steadily ture and pressure. Since P∑ is lower for 2 Critical from the temperature of the hot heat liquid CO than for an ideal gas except at Ideal gas 2 temperature temperatures closer to the critical point 1 exchanger at one end of the regenerator to that of the cold heat exchang- than are shown, liquid CO2 has advan- er at the other end. The walls should tages as a working fluid over gases. 220 240 260 280 300 have low heat conductance along the Temperature (kelvins) channels so that they do not conduct compression in step 1. When the heat from the hot heat exchanger to cold working fluid flows back Figure 2. Thermal-Expansion Co- the cold heat exchanger. The nar- through the regenerator in step 2, it efficient of CO2 near Tcritical rowness of the channels creates ex- is warmed by the heat stored in the The thermal-expansion coefficient, de- cellent thermal contact between the channel walls. Since the fluid is al- noted by Æ , is defined as(@V =@T )P =V , working fluid and the walls. There- ways at almost the same temperature where V is volume, T is fore the temperature of each small as the nearby walls, heat transfers to temperature, and P is pressure. Here element of the working fluid is al- and from the regenerator are nearly TÆ (the dimensionless thermal-expan- ways nearly the same as that of the the reverses of each other. There- sion coefficient) of CO2 is plotted as a adjacent part of the channel walls. fore each part of the regenerator is function of temperature at the critical As the working fluid is displaced restored to its original state at the pressure of CO2, 73 atmospheres. Like through the hot heat exchanger and end of each cycle. many thermodynamic derivatives, the the regenerator (step 4 in Figure 4), In 1981 we chose the Stirling thermal-expansion coefficient goes to in- it gradually cools to the temperature cycle for our research on engines finity exactly at the critical point. Plotted of the cold heat exchanger by trans- and heat pumps that used liquids as in gray is TÆ for an ideal gas, which is ferring heat to the channel walls, working fluids. As the working unity at any temperature and pressure. which store the heat. (The heat ca- fluid in our first Stirling heat pump, Since the gases often used in heat en- pacity of the walls should be suffi- we chose liquid propylene (C3H6) gines and heat pumps are nearly ideal, ciently high that the stored heat does from more than a dozen hydrocar- the fact that TÆ of CO2 near the critical not change their temperature signifi- bon, CFC, and inorganic fluids with point exceeds TÆ of an ideal gas implies cantly.) The working fluid is then critical points just above room tem- ready to be further cooled by de- perature. The machine we built,

116 Los Alamos Science Number 21 1993

1 Malone Refrigeration/John Malone and the Invention of Liquid-Based Engines

John Malone and the Invention of Liquid-based Engines

t is hard to imagine using a liq- versatile 50-horsepower water en- infest them, and knowledge for Iuid instead of a gas in a heat engine and began an extensive pro- knowledge’s sake is better than gine, but John Malone did—per- gram of experimentation with it. their parasitical life.” After Mal- haps partly because he was not Malone claimed that his second en- one’s death in 1959, his son Ray prejudiced by a proper scientific gine was very efficient. In 1931 he wrote, “Now as patent rights have education. Malone was born in wrote, “Trials by three different in- long expired I can see no advantage England in 1880. His formal edu- dependent engineers gave 27% in- in publishing any of the informa- cation ended in his eighteenth year, dicated efficiency. Thus, after al- tion which he accumulated while when (probably in part to avoid lowing for furnace and mechanical developing his liquid engine.” some trouble with the police) he losses in a commercial engine, We can only guess why Mal- joined the merchant marine. He re- 20% overall efficiency between the one’s promising work came to an mained at sea for nearly all of the heat in the coal and the shaft end. The worldwide economic de- next fourteen years; during that horsepower can be expected.” The pression of the 1930s must have time he was wounded seventeen efficiencies of the steam engines made venture capital scarce. Some times in Arab and Latin-American that powered ships at the time were may have dismissed the idea of liq- wars. between 9 and 12 percent and those uid working fluids because it con- Leaving the merchant marine, of locomotives were between 5 and tradicted conventional wisdom. Malone founded the Sentinel In- 7 percent, much lower than the ef- Large coal-fired steam turbines strument Company and, later, the ficiency of Malone’s engine. with 20 percent efficiency were in Fox Instrument Company. He Curiously, the “27% indicated the ascendancy for applications began experimenting with liquids efficiency” quoted above is the above 10,000 horsepower. The in- as engine working fluids in the only quantitative experimental ternal-combustion engine (includ- 1920s. As part of that project, he datum in any of Malone’s publica- ing what we know today as the measured the compressibilities and tions and patent disclosures. In a diesel engine) was already more thermal-expansion coefficients of 1939 letter to Selwyn Anderson, advanced than Malone’s engine, many liquids, including hydrocar- Malone wrote about his measure- and its incomparable power-to- bons, mercury, carbon dioxide, and ments, “I refused to publish this in- weight ratio made it seem the only sulfur dioxide. In 1925 he com- formation because it cost me a lot practical choice for airplanes and pleted his first liquid-based engine, to learn it and I may yet obtain automobiles. By the time the Great shown on the title page of “Malone some reward if it is not known. Depression and then World War II Refrigeration. It burned coal, used Also because to my amazement I had ended, the steam engine was high-pressure liquid water as work- found my enemies were alleged disappearing, internal-combustion ing medium, and produced 50 centers of learning. Universities engines and turbines were becom- horsepower. Malone referred to and the like.” Later in the same ing ubiquitous, and Malone’s work that first engine as crude and cum- letter his bitterness is more evident: had been forgotten. It took another bersome, but claimed that with per- “A study of liquids as mediums in independent thinker, the late John severance it would have eventually thermodynamics will teach an engi- Wheatley, to see the promise in produced 500 horsepower. neer more about the art of thermo- Malone’s work fifty years later and Instead, in 1927 Malone com- dynamics than all the universities resume the study of liquid-based pleted a much smaller and more on earth, or the memory men who engines.

1993 Number 21 Los Alamos Science 117 Malone Refrigeration

Figure 4. A Malone Heat Pump Using the Stirling Cycle

Typically an electric motor (not shown) supplies the work required to compress the liquid working fluid by turning a crankshaft (also not shown) that operates the pistons. (Only the power piston does work; the displacer piston is operated by the same crankshaft for proper phasing.) The graph shows the thermodynamic cycle of pressure and temperature changes undergone by the working fluid; the object is the

absorption of heat Qcold at the cold heat exchanger in step 2. The graph of the Stir- ling cycle has been simplified by assuming that the four steps are separate from shown in Figure 5, could function as each other; that is, that each piston is stationary while the other one moves. In real engine or heat pump, and was heavi- machines the pistons oscillate sinusoidally in time. The simultaneous motion of the ly instrumented to allow simultane- two pistons causes consecutive steps of the cycle to overlap; the overlaps would be ous measurement of mechanical reflected in a graph of the Stirling cycle by rounded corners. In addition, because power, heat flow, and temperatures the volume of working fluid displaced by the pistons is small compared to the vol- and pressures throughout. Our ex- ume in the heat exchangers and regenerator, different elements of the working fluid tensive program of measurement on are carried through variations of the single thermodynamic cycle shown. Liquids that machine, coupled with simple are practical working fluids for the Stirling cycle and others because they possess theoretical work, taught us how liq- certain thermodynamic properties. Specifically, their thermal-expansion coefficients uid properties, machine geometry, (Æ ) are large, their compressibilities (∑ ) are small, and their heat capacities per unit and cycle thermodynamics all work volume at constant pressure (CP) are large. The heat Qcold absorbed in step 2 is (@Q =@P)T ¢P in concert to process each of the given by (@ Q/@ P)T¢P, where ¢P is the pressure difference brought about in step countless volume elements of the 1. From a thermodynamic identity and the definition of Æ , ³ ´ ³ ´ ³ ´ liquid through its own closed cycle. @Q @S @V As a result, we believe we can now = T = ¡ T = ¡ T ®V ; @P T @P T @T P predict the power and efficiency of Malone machines with reasonable where S is entropy. It follows that Qcold ' °TcoldÆ V¢P, where Tcold is the tempera- accuracy. ture of the cold heat exchanger, and V is the volume of liquid displaced through the As this fundamental phase of the heat exchangers. Thus a large thermal-expansion coefficient leads to absorption of work drew to a close in the middle a large amount of heat. On the other hand, the energy wasted in heat absorption is and late 1980s, our situation under- proportional to the square of the temperature drop ¢T caused by the depressuriza- went a number of important tion in step 1, which is given by ¢T = (TÆ=CP)¢P. Therefore a working fluid with a changes. Recognition that CFCs small value of Æ=CP is desirable. Fortunately, liquids near their critical points typi- cause unacceptable health hazards cally have thermal-expansion coefficients that are a little larger than that of an ideal and environmental damage by de- gas and heat capacities per unit volume that are orders of magnitude larger than stroying atmospheric ozone (see that of an ideal gas, so that the cooling can be both powerful and efficient. Finally, “CFCs and Cooling Equipment: The the low compressibilities of liquids are an advantage in steps 1 and 3, as can be Size of the Problem”) gave new mo- seen from the relation ¢P = ¢V/V∑ , which follows from the definition of ∑ . Thus tivation to our development of cool- large pressure changes ¢P can be achieved in the entire liquid volume V by power- ing technologies based on other piston strokes that displace small volumes ¢V. working fluids. But the pace of the research did not increase. Wheat- ley’s sudden death deprived us of pact in the marketplace within two room temperature—carbon dioxide. our leader, our keenest intellect, and or three years. A prolonged, time- Liquid carbon dioxide and dilute our most successful fundraiser. consuming attempt to get Navy sup- mixtures of methanol or ethanol in Escalating costs of experimental port for our research ended with no liquid carbon dioxide are efficient work—especially of fabrication— Navy funds coming to Los Alamos, and safe working fluids for Malone slowed progress further. The next but with researchers at a Navy labo- refrigeration. The amount of carbon step in the development of a Malone ratory beginning their own develop- dioxide used has negligible environ- refrigerator needed much more sup- ment of Malone machines. Fortu- mental impact compared even with port than our earlier work, but sup- nately, during this difficult interim the effect of the pounds per capita port was hard to find: the DOE’s phase of basic technology develop- per day we each exhale; the amount Office of Basic Energy Sciences was ment, the Laboratory’s own re- of alcohol also has negligible envi- not interested in increasing our search funds have provided a partial ronmental impact. funding for this applied work, and bridge. Our first Malone machine, though the DOE offices that support conser- For our current Malone-refrigera- designed for ease of measurement vation and renewable energy seemed tor research, we picked the most en- rather than for efficiency, was half interested in funding only those pro- vironmentally acceptable of the liq- as efficient as present-day CFC- jects that promised to make an im- uids with critical points just above based equipment. There is little

118 Los Alamos Science Number 21 1993

1 Malone Refrigeration

Qhot Step 4 The displacer piston pushes the hot, compressed liquid back through the heat exchangers and regenerator. In the hot heat exchanger, the liquid rejects heat to the Step 1 room and its temperature decreases. As it flows The cold working liquid is through the regenerator, it loses enough heat to the depressurized as it does work channel walls to lower its temperature to that of the on (pushes up) the power cold heat exchanger. The heat lost by the liquid is piston. During this step the stored in the regenerator to be returned Win liquid becomes even colder. to the liquid in Step 2. The liquid is ready to repeat Step 1. Wout

Step 4 Hot heat Q exchanger Power hot piston Step 1 Step 3 Displacer W Win piston out Pressure

Regenerator Qcold Step 2 Temperature Step 3 The power piston does work on Cold heat the hot liquid by compressing it. exchanger As the liquid is compressed, it Step 2 becomes even hotter. The displacer piston pushes the liquid through the heat exchangers and regenerator. In the cold heat exchanger, the liquid absorbs heat Hot from the refrigerated compartment (the object of the game!) and its temperature increases. Temperature As the liquid flows through the regenerator, it scale absorbs enough heat from the channel walls to raise its temperature to that of the hot heat exchanger. Cold

Qcold

doubt that a liquid-CO2 Malone re- environmentally questionable) be very thin and can be applied to frigerator can be built with an effi- “space-age” materials and tech- the sheets by electroplating before ciency higher than that of existing niques. For example, the heat-ex- punching or chemical milling, as an- CFC-based refrigerators. There are changer/regenerator assembly is a other cost-saving measure. And we two important questions, which we furnace-brazed stack of stainless hope we can eventually save still are pursuing simultaneously. Can an steel sheets; slots in some of the more money by making some of the efficient Malone refrigerator be built sheets form fluid channels when the parts from ordinary carbon steel in- inexpensively enough to enjoy wide- sheets are assembled. Although we stead of stainless steel. spread manufacture? What are the now fabricate the sheets by pho- The configuration of the pistons environmental costs of the manufac- tolithography and chemical milling and other moving parts is also influ- turing process? for speed and flexibility, we know enced by the need to reduce costs. The CO2 Malone refrigerator we that they can ultimately be mass- Our original Malone machine had are building now should provide produced very inexpensively and too many high-precision and hence partial answers to those questions. cleanly by punching. The brazing expensive moving parts, including a We are using modern fabrication metal is pure copper, which is dozen roller bearings. The cost of techniques when necessary but are cheaper than the more commonly the bearings alone was higher than avoiding expensive (and sometimes used silver alloys. The copper can the cost of the compressor in a con-

1993 Number 21 Los Alamos Science 119 Malone Refrigeration/Thermoacoustic Engines and Refrigerators

Thermoacoustic Engines and Refrigerators hermoacoustic effects, which crankshaft-coupled pistons or ro- Tconvert heat energy to sound, tary compressors, thermoacoustic have been known for over a hun- heat pumps have no moving parts dred years. They have generally or a single flexing moving part, been considered mere curiosities, such as a loudspeaker, and have no but in the early 1980s our engine- sliding seals. The lack of moving research group at Los Alamos, led parts gives thermoacoustic refrig- by John Wheatley, began to consid- erators the advantages of simplici- er thermoacoustic effects as a prac- ty, reliability, and low cost. Be- tical way to make efficient engines. cause the sound waves are confined One serious impediment to rapid in sealed cavities, the machines are progress on our experimental Mal- fairly quiet. one engines was the large number For us thermoacoustic heat of precision moving parts required. pumps had the additional advan- While looking for simpler engine tages of conceptual elegance and designs, we came across the work easy, low-cost prototype develop- of Peter Ceperley at George Mason ment. We hoped that those fea- University, who had realized that tures would lead to near-term suc- the timing between pressure cesses (which would help keep our changes and motion in Stirling en- research well funded and lend gines and heat pumps is the same credibility to our longer-term Mal- as in a traveling sound wave. In- one program). The development of spired by his work, we eventually thermoacoustic refrigerators has invented thermoacoustic heat indeed had successes, such as the pumps (and new types of thermoa- 1992 flight in a space shuttle of a coustic engines) that had at most thermoacoustic refrigerator built at one moving part. the Naval Postgraduate School and As Figure 1 shows, our thermo- a 1993 test of a thermoacoustic acoustic heat pumps use standing sonar projector (an engine rather (rather than traveling) sound waves than a heat pump) by Bill Ward in to take the working fluid (a gas) the Laboratory’s Advanced Engi- through a thermodynamic cycle. neering Technology Group. They rely on the heating and cool- After Tim Lucas, an inventor, ing that accompany the compres- noticed an article about our ther- sion and expansion of a gas in a moacoustic work in a popular sci- sound wave. Although ordinary, ence and technology magazine, he conversational-level sound pro- added yet another chapter to the duces only tiny heating and cooling story of novel refrigeration at the effects, extremely loud sound Laboratory. Lucas had invented waves produce heating and cooling the Sonic Compressor (Figure 2), a effects large enough to be useful. device for compressing conven- Whereas typical heat pumps have tional refrigerant vapors that con-

120 Los Alamos Science Number 21 1993 Malone Refrigeration/Thermoacoustic Engines and Refrigerators

tains no sliding parts. Instead a Cold heat Hot heat resonant sound wave in a cavity Loudspeaker exchanger Stack exchanger compresses the vapor and two one- way valves ensure that only low- pressure vapor enters and only high-pressure vapor leaves the Resonator W in compressor. Since the Sonic Com- pressor needs no lubricating oil, it is attractive for compressing HFC refrigerants, which do not destroy the ozone layer but have the draw-

Qhot back of being less compatible with Qcold lubricants than CFCs are (see “CFCs and Cooling Equipment: What is each gas element doing? The Size of the Problem”). The Cooler Warmer Expanding and Rejecting heat lack of sliding parts should also Plate from the stack cooling while to warmer lead to higher efficiency in small moving left part of plate systems. Furthermore, the Sonic Oscillation of gas element Compressor can replace the piston-

Pressure Contracting and driven compressor in present re- heating while frigerators without requiring any Absorbing heat moving right Plate from the stack from cooler changes in other parts. part of plate Lucas needed to suppress the Temperature production of shock waves in his compressor by the high-amplitude Figure 1. The Thermoacoustic Refrigerator sound because the shock waves An electrically driven, radically modified loudspeaker maintains a standing sound wasted energy by turning it into wave in an inert gas in a resonator. The sound wave interacts with an array of par- heat. He sought help from us be- allel solid plates called the stack. The resulting refrigeration can be understood by cause of our experience with high- examining a typical small element of gas between the plates of the stack. As the amplitude sound in thermo- gas oscillates back and forth because of the standing sound wave, it changes in acoustics. Working together we temperature. Much of the temperature change comes from compression and ex- found that the shock waves result- pansion of the gas by the sound pressure (as always in a sound wave), and the rest ed from nonlinear self-interactions is a consequence of heat transfer between the gas and the stack. In the example in the desired fundamental reso- shown the length of the resonator is one-fourth the wavelength of the sound pro- nance in the cavity and from un- duced by the speaker, so all the elements of gas are compressed and heated as wanted resonances at frequencies they move to the right and expanded and cooled as they move to the left. Thus that were exact integral multiples each element of gas goes through a thermodynamic cycle in which the element is of the fundamental frequency. compressed and heated, rejects heat at the right end of its range of oscillation, is When we changed the shape of the depressurized and cooled, and absorbs heat at the left end. Consequently each el- cavity to that shown in Figure 2, ement of gas moves a little heat from left to right, from cold to hot, during each the frequencies of the extra reso- cycle of the sound wave. The combination of the cycles of all the elements of gas nances changed so that they were transports heat from the cold heat exchanger to the hot heat exchanger much as a no longer significantly excited by bucket brigade transports water. The spacing between the plates in the stack is cru- nonlinear self-interaction in the cial to proper function: If the spacing is too narrow, the good thermal contact be- fundamental. tween the gas and the stack keeps the gas at nearly the same temperature as the Lucas’s collaboration with us stack, whereas if the spacing is too wide, much of the gas is in poor thermal contact was an example of totally suc- with the stack and does not transfer heat effectively to and from it. cessful “tech transfer.” During

1993 Number 21 Los Alamos Science 121 Malone Refrigeration/Thermoacoustic Engines and Refrigerators

ventional CFC refrigerator! A totally Elastic support for Low-pressure different design was clearly required resonator cavity vapor if Malone technology was ever to enjoy widespread use. So in our pre- Magnet Coil sent CO2 machine we are using a lin- Intake ear free-piston configuration, which valve was invented only recently and is being employed in gas-based Stirling Resonator cavity engines intended for solar power or for use in space. The pistons in a lin- Outflow ear free-piston machine are driven, valve not by a rotating motor connected to Motion of cavity a crankshaft, but by a “linear” electric motor that provides reciprocating force and motion directly in the same High-pressure way as a loudspeaker. This configu- Compressor wall vapor ration minimizes the number of moving parts and eliminates the need for Figure 2. The Sonic Compressor high-force bearing surfaces. Careful The Sonic Compressor uses electric power to compress a conventional refriger- design can even eliminate the need ant vapor by means of a high-amplitude sound wave; the model depicted can re- for any mechanical connection to the place the piston compressor in a conventional cooling system such as the displacer piston—the piston moves household refrigerator shown in Figure 1 of “Malone Refrigeration.” The electric- with the correct amplitude and phase ity drives a radically modified loudspeaker that shakes a cavity back and forth at simply in response to the fluid pres- a resonance frequency of the working-fluid vapor inside (300 hertz). In the fig- sures acting on it. ure the cavity is shown at the rightmost point of the vibration. The motion of the The present crisis in the cooling in- cavity causes the vapor to slosh back and forth—in other words, the motion gen- dustry is a unique opportunity for a erates a standing sound wave. The shape of the cavity is designed to prevent new, potentially more efficient tech- the formation of shock waves. The standing sound wave compresses and ex- nology to break the monopoly of a pands the gas; at the end of the tube farther from the loudspeaker, the range of technology that has enjoyed decades of pressure is 8 atmospheres. A pair of one-way valves at that end, which are incremental improvement. The prima- opened and closed at the operating frequency by the pressure itself, admits low- ry challenge of the next year or two is pressure vapor from the intake pipe and ejects high-pressure vapor into the out- to keep this difficult experimental pro- flow pipe. ject moving ahead, though the funding only pays for a third of the time of one the year he spent here, we solved the market until a few years later. researcher, while trying to attract the his shock problem. Of equal im- Malone refrigeration will take interest of an industrial collaborator. portance to him, we did not joint- still more time to develop but ap- At best, years of further work costing ly invent anything patentable, so pears to be the most efficient op- millions of dollars will be required to the business aspects of his project tion of the three to which we at bring Malone refrigeration to the were not complicated by the in- Los Alamos are contributing. threshold of possible widespread ap- volvement of intellectual-proper- plication. Meanwhile, as described in ty rights belonging to the Labora- “CFCs and Cooling Equipment: The tory. As Lucas’s visit was suc- Size of the Problem,” industry is pro- cessful, the Sonic Compressor ceeding promptly with more straight- could come into production in a forward interim measures. In the in- few years. Thermoacoustic re- termediate time scale, mature new frigeration will not be ready for technologies such as the Sonic Com- pressor and perhaps thermoacoustic re-

122 Los Alamos Science Number 21 1993 Malone Refrigeration

Figure 5. Chris Espinoza Adjusts Our First Malone Refrigerator (1989) frigeration may come into use. National Laboratory. If we succeed in developing a Further Reading widely used refrigerator, it would be R. E. Benedick. 1991. Ozone Diplomacy. Cam- the first Laboratory product since bridge, Massachusetts: Harvard University Press, the implementation of the Atmo- 1991. spheric Test Ban Treaty to find its Energy Information Administration. 1989. Poten- way into homes and businesses tial costs of restricting chlorofluorocarbon use. throughout the world. If we fail for EIA Service Report SR/ESD/89-01. Washington, unforeseen technical reasons, we DC: U. S. Department of Energy. will not regret having tried. But if Gregory W. Swift. 1990. Malone Refrigeration. we fail because of inadequate sup- ASHRAE Journal, November: 28Ð34. port, an opportunity to improve the Gregory W. Swift. 1989. A Stirling engine with a world environment and reduce its liquid working substance. Journal of Applied energy consumption will have been Physics 65: 4157Ð4172. lost. Steven L. Garrett and Thomas J. Hofler. 1992. It is a pleasure to acknowledge Thermoacoustic refrigeration. In Technology the contributions of Al Migliori, 2001: Proceedings of the Second National Tech- Gregory W. Swift is a staff member in the Con- densed Matter and Thermal Physics Group, where Sonia Balcer, Chris Espinoza, Frank nology Transfer Conference and Exposition. NASA Conference Publication 3136, 2: 397Ð406. he has been working on novel heat engines and re- Murray, and Alex Brown to the de- Washington, DC: NASA. frigerators since 1981. He received his B.S. in velopment of Malone refrigeration physics and mathematics from the University of Nebraska and his Ph.D. in physics from the Uni- at the Laboratory and to thank the Gregory W. Swift. 1988. Thermoacoustic engines. Journal of the Acoustical Society of America 84: versity of California, Berkeley. From 1983 to Department of Energy’s Office of 1145Ð1180. 1985 he held an Oppenheimer Fellowship at Los Basic Energy Sciences for steady Alamos. He is a fellow of the Acoustical Society of America. support of fundamental engine and Information about the Sonic Compressor can be obtained by calling Timothy S. Lucas (president of refrigerator research at Los Alamos Sonic Compressor Systems in Glen Allen, Vir- ginia) at (804) 262-3700.

1993 Number 21 Los Alamos Science 123