Solar Energy Materials 2 (1980) 381-393

~) North-Holland Publishing Company

THERMAL ENERGY STORAGE IN SALT HYDRATES

Maria TELKES
Solar-Riat, US Highway 190 West, Killeen, TX 76541, USA

Received 4 March 1980; in revised form 8 April 1980

This article reviews the material aspects of thermal energy storage in salt hydrates. In air cir-
culating systems that separate the collection and storage of solar heat (the so-called "active sys-
tems"), the backup system is an important part of the total installation. On a succession of cloudy
days, phase change storage systems can be charged by using resistance heating or heat pump. In
most locations, lower cost electric power is available during off-peak periods and the use of the
backup electric system would be economically desirable. In passive solar heating systems, a method
is described that could be incorporated into the south facing vertical wall of the building, eliminating
the cost of the wall by replacing it with polyhedral glazing and a phase change Trombe wall.

1. Introduction

The properties of thermal storage materials are reviewed. Emphasis is placed on

materials that melt at moderate temperatures around 32°C (90°F) or 21 to 24°C
(about 70 to 75° F). The storage of cooling can be accomplished in materials that melt
at lower temperatures at 13°C (55°F) or 7~C (45~F). Tests have been performed in the
University of Delaware's experimental house (1973 to 1977), and new full scale
laboratory tests have been completed with the coolness storage materials.

2. Solar heating systems

It is probable that improved solar heating systems will be developed, tested and
marketed during the next few years. Currently the following major subsystems are
identified: solar collector mounted on the roof or the south wall, circulating system
and controls, thermal storage and back-up systems.
The primary subject of this article is the thermal storage subsystem. The other
subsystems are mentioned in relation to their influence on storage and their connec-
tion with it. An important part of most solar heating installation is the back-up
system that is needed during prolonged periods of cloudy days. Electric power in the
form of resistance heating or as the energy source of heat pumps can be used as the
back-up system. It may be desirable to use electric power during off-peak periods
when electric rates may be lower. It may be possible and even economically desirable
to store electric heating effects in the thermal storage system during off-peak periods
for later use during peak periods.
381
382 M. Telkes / Thermal energy storage in salt hydrates

3. Passive solar heating

Passive solar heating systems usually involve the use of Trombe-walls [5]. The
components of such a wall are:
A solar energy transmitting wall which can act as thermal insulator when needed.
A system that stores solar energy as heat (or cooling).
Means to convey stored heat (or cooling) into the building, when needed. The
transmission and thermal insulation may be combined, provided it would be possible
to obtain structures of solar transmittance higher than 60~, having at the same time
reasonably good thermal insulation values. Studies have been made of solar energy
transmittance of a number of air-spaced transparent layers and of their thermal
resistance values R = I / U with optimum spacing. The transmitting glazing layers or
panes can be coated with low reflection coating. The optimum situation may be

Fig. I. Polyhedralstructurecoveringblackabsorber.
 M. Telkes / Thermal energy storage in salt hydrates 383

obtained by using five to seven parallel glazing layers which have a thermal resistance
R = 4 (U=0.25), or higher.
Several attempts have been made to obtain good solar transmission, at diminished
heat loss, using various combinations of glazing and transparent honeycombs [6].
I have developed a structure that resembles a "three-dimensional honeycomb",
consisting of a polyhedral structure of large-celled foam [7, 8]. This material has been
commercially produced for a few years. Unfortunately, the fabricator used poly-
styrene [9] that is attacked by solar radiation, the material turns yellow or gray-
brown, and becomes brittle. Another attempt was made to use a mixture of poly-
styrene with acrylic plastics. This was more durable but still not sufficiently sun-
resistant. Figs. 1 and 2 show a polyhedral structure and its proposed uses. Fig. 3
shows one way of fabricating panels by extruding the foamed plastic between an
external and internal transparent glazing layer. According to measurements, the

Fig. 2. Polyhedral structure in transmitted light.

~ - + 1 ~ = ~ . . _ __ ~ /-,,.~_..~CONVEYOR
PANELS

Fig. 3. Preparation of polyhedral panels.

384 M. Telkes r/ Thermal energy storage in salt hydrates

initial solar energy transmittance of panels consisting of six cell-layers was better
than 60~o while the thermal losses were as low as obtainable with one inch thick layer
of conventional foam insulation. The polyhedral structure "cell-size" reaches an
optimum at,approximately 9 mm (3/8 in.).
The polyhedral wall structure can be made of low cost plastics or other materials.
The walls of the cells can be very thin and the weight of one square meter wall-
structure, less than 50 mm (two inches) thick, can be less than 1 kg/m 2 (0.2 lb/ft2).
The emphasis is on the cell size which should not be less than 6 mm (1/4 in.) and should
not be larger than 12 mm (1/2 in.). The aim is to use highly durable materials to
develop complete wall structures in the form of modular panels, possibly about
1.2 × 2.4 x 0.05 m 3 weighing less than 10 kg (4 ft x 8ft x 2 in. thick, weighing less
than 20 lbs).

4. Thermal storage

The classical Trombe wall using concrete has an average temperature variation
AT of 10°C (about 18°F), consequently, one pound of concrete can store 4 Btu
(9.4 x 103 J/kg). The specific heat Of concrete is 0.22 Btu/lb°F (930 J/kg K). In this
case, a concrete block of one cubic foot volume cannot store more than 500 Btu
(1.9 x 107 j/m3). Computer analysis at the Los Alamos Scientific Laboratory indic-
ated the optimum thickness and thermal storage capacity of Trombe walls [10, 11].
Trombe's report [5] prefers a thickness around 1.4 ft (40-45 cm).
Phase change materials melting around 18-23°C (70-75°F) are available [12-15].
These can store up to 10000 Btu/ft 3 (3.7 × l0 s J/m 3) or approximately 20 times more
than a concrete wall of the same volume. It is immediately obvious that the wall
made of phase change materials could be twenty times thinner than the concrete
wall, to store the same thermal energy. A wall of 25 mm (one inch) thickness could
store the same amount of thermal energy as a 420 mm (1.4 ft) thick concrete wall.
In addition, the phase change material is conveniently at a constant temperature.
The "Phase Change Thermal Trombe Wall" can be built of individual components
that are stacked to form an interlocking wall. Conventional post type structure can
be used, allowing the use of 4 x 8 ft 2 (1.22 x 2.44 m 2) exterior panels formed of poly-
hedral walls and sections of the phase change thermal wall mounted on the same
upright posts.
If desired, windows can be left in the polyhedral wall that transmit solar radiation
directly into the building. Some of these windows can be openable for people who
want to look out. They can be double or triple glazed and are visually transparent.
Plans have been made to construct a building using polyhedral walls with phase
change material filled containers behind the wall.

5. Glauber's salt eutectics

Glauber's salt (sodium sulfate decahydrate) is probably the most satisfactory

material to be used in passive systems. It is available in practically unlimited quan-
 M. Telkes / Thermal energy storage in salt hydrates 385

tities, as "bulk material", in the form of anhydrous sodium sulfate, at a cost as low as
$60 per ton (1000 kg). When made into sodium sulfate decahydrate, by adding the
appropriate amount of water (56~) the cost decreases to less than 3 cent per kg
(1.5 ¢ per pound).
Several would-be experimenters and manufacturers found that repeated cycling
decreases the available heat content of Glauber's salt. The "ideal" heat of fusion
value is 60 kcal/kg. (108 Btu/lb) measured repeatedly by experts, including those of
the National Bureau of Standards [16].
The difficulties can be eliminated by using nucleating agents and thickeners.
Premature claims have been made that "no salt hydrate can survive more than 10
cycles without deterioration" [17]. Others have experimented with the use of thick-
eners that "biodegraded" in the salt hydrate itself. Still others suggested the use of
heavy walled porous structures that needlessly added weight [18]. Experiments have
been made with saturated solutions rather than salt hydrates, thereby cutting the
thermal storage capacity into half of its actual value [19]. Metal containers have been
proposed and used, regardless of their cost and corrosion. Expensive and energy
consuming rotating structures were suggested and used without regard to the energy
needed to keep them rolling [20].
After considerable experimental work and tests made with numerous thickeners,
an inorganic thixotropic material was found to be satisfactory [21]. The thickener
has the advantage that salt hydrates of low cost can be thickened most economically,
due to the low cost of the thickener itself, which is "clay-type" strip-mined in large
quantities [22]. The thixotropic mixture is an easily flowing liquid when stirred, but
it "sets" to a gel-like structure when not agitated. The difficulty is that the gelled
material in its "liquid" state, above the melting point of the salt hydrate, is quite
soft and has no "load-bearing" mechanical strength. When the material is below its
melting point, the mixture can be "hard as ice".
It would be desirable to change the material into a self-supporting form, even when
it is in the molten state. Some experimenters attempted to use porous concrete struc-
tures, and by vacuum methods forced the liquid salt hydrate into the holes of the
concrete. At best only about half of such a structure is available to be filled with salt-
hydrates [18]. Another method has been proposed: to encase thickened salt hydrate
into plastic bags and surround these with "plastic concrete", forming brick-like
blocks. It is obvious that such methods use heavy walled enclosures, occupying more
than 50~ of the volume and even more of the weight of the entire structure, decreasing
greatly its thermal storage capacity [23].

6. Recent developments

The problem has been solved by forming a crust around each crystal of the solid
salt hydrate, by chemical means. The entire mass is solidified as the entrusted crystals
form a conglomerate structure. Above the melting point, the crystals melt in their
crust-shells, but due to the small size of each encrusted crystal stratification is neg-
ligible. The entire structure is self-supporting, even when the crystals are at a tempera-
386 M. Telkes / Thermal energy storage in salt hydrates

ture above their melting point in the liquid state. The blocks must still be surrounded
by a hermetically sealed outside cover. This must be impermeable to water vapor,
otherwise the water of crystallization of salt hydrates would evaporate, as the water
content of salt hydrates can be 40 to 60~o of the total volume. The external skin or
package forms only part of the load-bearing structure [24].
The problem of crystal formation and growth is illustrated in fig. 4. Glauber's
salt crystals are placed into a rectangular tube of 20 x 20 mm 2 cross sectional area,
melted and permitted to grow at room temperature. Fig. 4 is part of a time-delay
moving picture (prepared by Dr. Harold Edgerton of MIT). The bottom layer of
anhydrous Na2SO 4 crystals look like soil, sprouting transparent crystals that "grow
like grass".
Cooling curves are shown in fig. 5, using mixtures of sodium sulfate and water,
with and without thickener. All samples contained 3~o Borax (used for crystal
nucleation). Glauber's salt Na2SO4 • 10H20 contains 44~ Na2SO 4 ; in its thickened
form, it shows a well formed plateau at its melting point. Without thickener. Glauber's
salt shows the same cooling curve as mixtures of Na2SO4 +water (with 36 and 40~,
Na2SO4, balance being water).
All materials were prepared in identically shaped containers and compared with
the cooling curve of a water-filled container. The difference between the cooling
curves and ambient air indicates the relative heat content of the mixtures in a most
conclusive way.
The results of 1000 consecutive cycling tests are shown in fig. 6. Glauber's salt was
mixed with 3~ Borax (to nucleate crystal growth) and thickened [21]. Air was cir-
culated around the tubular, horizontally placed container (To) and the temperature
of the salt hydrate (Ts) recorded continuously. One cycle required 4 h (2 h each for
heating and cooling), changes being made automatically with a sequencing device.
Each 24 h day, 6 cycles were performed and 1000 cycles required about 6 months for
completion. There was practically no difference between the first, 200th and 1000th
cycles.

Fig.4. Glauber'ssalt crystalsgrowing(actual size20 mm verticaldimension,photo by Dr. HaroldEdgerton,

M.I.T.).
 M. Telkes / Thermal energy storaoe in salt hydrates 387

90

85.
LL
O

LAJ

80
"=E

t.~ NOT"I"HICKE.I~
E
75

70 ¸

65

t I .I. I I t I I
1 2 3 4 5 6 7 8
HOURS
Fig. 5. Na2SO 4 + H20 +borax 3% +thickener 9%.

CYCLE 200--
CYCLE I 0 0 0
To
ttO ~
T o = A M B I E N T AIR
Ts = T H I C K E N E D

I00

90

80

7O To

60 I I I /
0 HOURS I 2 5 4
• HEATING -'-,~ COOLING -

Fig. 6. Sodium sulfate decahydrate, cycles 200 and 11300.

388 M. Telkes / Thermal energy storaoe in salt hydrates

7. Thermal Energy Storage (TES) system

The following is a "first-order" cost estimate, listing factors contributing to the

cost of the system :
- materials, chemicals, packaging,
- manufacturing (mixing, filling, sealing),
- sales (warehousing, transportation, profit, overhead),
- installation (local materials, labor),
- bin and support structure,
- occupied space.

7.1. Cost of materials (for Na2SO 4"10H20 )

Cost S/ton

Chemicals - Na2SO4, anhydrous 60

borax 200
thickening material 60
Shipping costs are additional

Water is added to produce Glauber's salt, changing the cost to $30/ton or 3.3 ¢/kg
(1.5 ¢/lb). The cost of 6000 lb material storing 630000 kJ (600000 Btu) is $100.

7.2. Containment

The average manufactured cost of plastics is approximately $2.20/kg ($1/lb).

Based on blow molded containers, an estimated 136 kg (300 lb) of plastic is required
to contain 2700 kg (6000 lb) of salt hydrate mixture, needed in a 630000 kJ (600000
Btu) TES at a total cost of $300. Total cost of materials is $400 (1979 dollars).
The other costs are more difficult to estimate. The factors listed above, however,
applied cumulatively to the materials cost of the prototype 630000 kJ (600000 Btu)
storage capacity, appear to be reasonable.
Total Estimated Cost of the 630000 kJ (600000 Btu) TES system, including com-
mercial installation, may be less than $2000.
There remains the consideration of occupied space. To assign a dollar value to this
is practically meaningless in a residential installation, since the TES requires only
limited volume and would inevitably be located in a space that is already available
and scarcely utilized. In other words, no new construction is necessary. A modular
phase-change TES unit may even be permanently deployed in several separate sec-
tions in a building, offering extreme flexibility.
The occupied volume of the salt hydrate in a typical 630000 kJ (600000 Btu)
heat storage unit has a net volume of 1.84 m 3 (65 ft3). The heat exchanger/containers
are designed to provide spaces for air circulation, that add not more than 1.0 m 3
 M. Telkes / Thermal eneroy storaoe in salt hydrates 389

(35 ft 3) in additional volume, leading to a total gross volume of approximately 2.84 m 3

(100 ft3).
The passive system could be based on a south wall collector of 300 ft 2 net area.
Assuming that TES is evenly distributed, the gross volume of 100 ft 3 results in a
TES wall thickness of-~ ft or 4 in. (100 mm).

8. Storage of cooling using phase-change materials

The storage of cooling combined with air conditioning systems has been tested at
the Institute of Energy Conversion of the University of Delaware and it was incorpor-
ated in the experimental building "Solar-One" (1973 to 1977).
The purpose of the storage of cooling is to lessen peaktime power demands from
many millions of air conditioners which can cause "Brownouts" during hot summer
days. The system uses standard air conditioners, which are operated during the
night, storing up "coolness" in a phase-change eutectic system by freezing it around
45°F to 55°F. During the day, when cooling is needed, the stored coolness is used by
blowing room air through heat-exchangers filled with the phase change material.
In this way, peak demands for electric power can be diminished, resulting in "peak-
shaving". The storage material is around 60° F after the stored coolness is extracted,
as night approaches and the peak electric demand is over. After 10 to 11 pm when
off-peak rates start, the air conditioner resumes operation, cooling the storage
material again, completely solidifying it by early morning. The coolness storage
material is reused every day, as this is needed.
In some respects, coolness storage materials act like ice, which absorbs heat at the
rate of 7900 Btu per cubic foot (300 kJ/m 3) of ice as it changes to water at 32°F.
Coolness storage material melting at 55°F can store nearly the same amount of
cooling as ice does.
An additional benefit is that its volume does not expand during freezing; a great
advantage as compared to ice, which involves a volume change of nearly 9% and the
danger of breaking pipes and heat exchangers. This problem is completely eliminated
when using eutectic phase change materials, which contract very slightly, if at all,
when freezing.
Phase-change materials in this case consist of salt hydrates, based on sodium sulfate
decahydrate, one of the most widely available materials, obtainable at a low cost.
This material melts at 89°F, its transition temperature, where 85% is completely
dissolving in the water of crystallization. With specified amounts of added salts,
including sodium chloride and ammonium chloride, the melting point changes to
55°F, forming an eutectic salt-hydrate system. Its water content is nearly 50% and
the cost of the ingredients is less than two dollars per hundred pounds. Other eutectics
are available, each melting at a well defined specific temperature that can be varied
between 40 and 65°F. Other materials melting in different temperature ranges are
also available.
Salt hydrates based on sodium sulfate decahydrate have been specially developed,
because they are low in cost ; in addition, they are non-toxic, non-corrosive and easily
390 M. Telkes / Thermal energy storage in salt hydrates

mixed and handled. A special crystallization promoting agent is used (a "nucleating

agent"), to provide large number of crystal seeds to start crystal growth when the
eutectic is cooled. Small quantities of a special thixotropic thickening agent are added
to suspend a large number of crystal seeds and prevent settling of the components
[21].
Coolness storage materials are contained in heat exchangers made of thin walled
plastic containers of low cost. These containers have the advantage of durability and
light weight, as compared to metal containers, which can easily corrode in this
application.
Fig. 7 shows a typical heat-curve of the 55°F material contained in a tubular
container which is 32 mm (1.25 in.) outside diameter. The wall thickness can be
around 0.8 to 1.2 mm, preferably made of polyethylene. The tubes were placed in an
insulated duct and 70-76 ° F air was blown around the tubes at the same rate as this
would occur in the bin. Starting at 40°F, the tube temperature rose to within +_2°F
of the melting point and kept at this level for 6~2h, rising slowly to a cutoff of 68°F
in eight hours. This was in accordance with the design expectations. Numerous cycles
were repeated at various air velocities and ambient temperatures.
Coolness storage systems using ice blocks have been popular for many years.
The production of ice, using commercial air conditioners, was combined with an
off-peak operation in 1977 [25]. It is described briefly in the following way. The newly
developed ice storage system has shifted the air conditioning load completely into
late night off-peak hours. The storage modules are of all plastic construction, con-
taining water that is frozen completely by the air conditioning equipment. Coolness
passes in and out of the modules through specially designed plastic heat exchanger
coils. Glycol brine is used as the heat transfer fluid. Performance of ice-type coolness
storage units can be used in residential or commercial installations [26].
During the night cycle, the system freezes the storage module, lowering its tempera-
ture to 28°F. If hot weather at night requires it, the system provides cooling to the
house as well. During the day cycle, cooling is circulated into the house from the
storage module. In the event of unusually heavy loads, the control system provides
cooling directly from the compressor-condenser unit, if the storage module is dis-
charged during the day.

70

60 ......... .- "

HOURS

Fig. 7. Coolness storage tube (55°F) in still air.

 M. Telkes / Thermal energy storage in salt hydrates 391

During 1979, experiments have been started with a low temperature phase change
material for use in bulk storage tanks, replacing water-ice. The purpose of this work
was to develop means of storing coolness economically in a liquid heat transfer
system with a phase transition point around 7° C (45°F).
As a result of experimental work, calorimetric tests have been completed on
samples subjected to 25 cycles. Cost, performance and practical application factors
have been evaluated. A complete test model (1:10 scale) has been built, filled and
cycle-tested.
A recent paper [28] predicted that: "Cooling storage is coming of age in the USA".
The Department of Energy/Argonne Conference emphasized the need and vital
interest the public utilities have in this matter. Off-peak rates exist in thirty-five states
and rising demand charges give evidence of future progress.
The need for building new generating plants, whether oil, coal or nuclear, can

BEFORECYCLING,AT 75°F

AFTER 10 CYCLES,AT 75°F

Fig. 8, 45°F material, with 0 to 8~o gel, before cycling, at 75°F and after 10 cycles, at 75°F.
392 M. Telkes / Thermal energy storage in salt hydrates

be postponed for many years, or perhaps eliminated, by making use of vast amounts of
idle capacity during nights and weekends. In Germany, France and England, over 6
million storage radiators have made use of off-peak rates in winter and done much to
flatten daily load curves. In Japan, most new commercial buildings have chilled water
storage for cooling.

Cycling tests with the coolness storage material

The coolness storage material with the transition temperature of 7°C (45° F), when
used with the thickening gel in sufficient amount, solidifies completely below that
temperature. When above 45°F, the material changes into a gel provided that more
than 6~o gelling agent is added.
Fig. 8 shows samples of the 45°F material containing no gel, 2 gel, 4 gel, 6 gel and 8
gel before and after the material has been cycled. The same samples are shown after 10
cycles. There is no change in the appearance of the material even after numerous
cycles have been performed. Adding 6 parts gel to the mixture prevents segregation of
the liquid part.

References

[1] American Association for the Advancement of Science, San Francisco annual Meeting (December
1979).
]2] M. Heylin and 3. H, Krieger. 1980's Energy Outlook; Gloom and Doom, Chem. Eng. News 121
January 1980).
[3] National Research Council's Committee on Nuclear and Alternative Energy Systems (CONAES~:
Energy in Transition, 1985--2010.
[41 Study sees Needs for Diverse Energy Sources, Chem. Eng. News 126 January 19801.
[5] F. Trombe and J. F. Robert et al., Some Performance Characteristics of the CNRS Solar House
Collectors, Passive Workshop (1976) p. 201.
[6] G. Francis, Etude d'un Collecteur de l'Energie Solaire, Comples Bull. No. 6 (May 1964) p. 56.
[7] M. Telkes, Method and Apparatus for Making Large Celled Material, US Patent 3 248 464 (1966).
[8] M. Telkes, Large Celled Material, US Patent 3 440 130 (1969).
[9] Anon., New Approach to Glazing, Mod. Plastics (December 1962).
[10] J. D. Balcomb et al., Simulation as a Design Tool, Passive Workshop (1976) p. 238.
[11] J. D. Balcomb et al., Simulation Analysis of Passive Solar Heated Buildings, Solar Energy 19 (1977)
277.
[12] M. Telkes, Thermal Storage for Solar Heating and Cooling, NSF-RANN Workshop, NSF-RANN-
75-041 (December 1975).
[ 13] M. Telkes, Solar Energy Storage, in : Critical Materials in Energy Production, ed. C. Stein (Academic
Press, New York, London, September 1976) ch. 14.
[14] M. Telkes, Solar Energy Storage, Chapter VII in :ASHRAE GRP 170 Application of Solar Energy,
eds. R.C. Jordan and Y. H. Liu (March 1977).
[15] M. Telkes, Thixotropic Mixture, US Patent 3 986 969 (October 1976).
[16] D. E. Jones and J. E. Hill, Testing of Pebble-Bed and Phase-Change Thermal Energy Storage Devices,
NBSIR 79-1737 (May 1979).
[17] U. W. Kauffman and Y. C. Pan, Thermal Energy Storage in Sodium Sulfate Decahydrate Mixtures,
Report No. NSF/RANN/SE/GI 27976/TR 72/11 (1972),
[18] D. Charoudi, Suspension Media for Heat Storage Material, Pros. Workshop on Solar Energy
Subsystems, NSF-RA-N 75-041 (1975) p. 56.
 M. Telkes / Thermal energy storage in salt hydrates 393

[19] D. R. Biswas, Thermal Energy Storage Using Sodium Sulfate Decahydrate and Water, Solar Energy
10 (1977) 99.
[20] C. S. Herrick and K. P. Zarnoch, Heat Storage Capability of a Rolling Cylinder Using Glauber's
Salt, presented at the DOE Annual Thermal Energy Storage Program Review Meeting (December
1979).
[21] M.Telkes,ThixotropicMixtureandMethod of Making Same, USPatent 3986969 (1976).
[22] Floridin Co., 3 Penn Center, Pittsburg, PA 15235.
[23] W. Worthy, Passive Solar Heating System Shows Promise, Chem. Eng. News (11 September 1978)23.
[24] M. Telkes, Phase Change Thermal Storage Materials With Crust Forming Stabilizers, US Patent
4187189 (5 February 1980).
[25] Anon., Ice storage system shifts a/c load to off-peak hours, in: Air Conditioning, Heating and
Refrigeration News (Nov. 1977).
[26] Descriptiveliterature ofCalmac Manufacturing Corporation, Box710, 150 S. Van Brunt Street, Engle-
wood, NJ.
[27] Calmac Industries, DOE contract, No. EM-78-C-02-4698.
[28] C. D. MacCracken, Refrigeration, Cooling Storage and Load Management, Am. Soc. Fteating
Refrig. Aircond. Engrs. J. (December 1979) 35.