Вы находитесь на странице: 1из 9

Applied Thermal Engineering 23 (2003) 733–741


Heat recovery from a thermal energy storage based on

the Ca(OH)2/CaO cycle
M.N. Azpiazu a, J.M. Morquillas b,*
, A. Vazquez c

Departamento de Ingenierıa Quımica y del Medio Ambiente, E.T.S. Ingenieros (UPV/EHU), Alda. Urquijo s/n,
48013 Bilbao, Spain
Departamento de M aquinas y Motores T ermicos, E.T.S. Ingenieros (UPV/EHU), Alda. Urquijo s/n,
48013 Bilbao, Spain
Departamento de Energıa y Propulsion Marıtima. E.S. da Mari~ na Civil, Campus de Riazor, La Coru~
na, Spain
Received 8 July 2002; accepted 10 January 2003

Thermal energy storage is very important in many applications related to the use of waste heat from
industrial processes, renewable energies or from other sources. Thermochemical storage is very interesting
for long-term storage as it can be carried out at room temperature with no energy losses.
Dehydration/hydration cycle of Ca(OH)2 /CaO has been applied for thermal energy storage in two types
of reactors. One of them was a prototype designed by the authors, and in the other type conventional
laboratory glassware was used. Parameters such as specific heats, reaction rate and enthalpy, mass losses
and heat release were monitored during cycles. Although in the hydration step water is normally added in
vapour phase, liquid water, at 0 C has been used in these experiences.
Results indicated that the energy storage system performance showed no significant differences, when we
compared several hydration/dehydration cycles. The selected chemical reaction did not exhibit a complete
reversibility because complete Ca(OH)2 dehydration, was not achieved. However the system could be used
satisfactorily along 20 cycles at least. Heat recovery experiments showed general system behaviour during
the hydration step in both types of reactors. The designed prototype was more efficient in this step.
Main conclusions suggested carrying out one complete cycle at a higher dehydration temperature to
recover total system reversibility. A modification of the prototype design trying to enhance heat transfer
from the Ca(OH)2 bed could also be proposed.
 2003 Elsevier Science Ltd. All rights reserved.

Keywords: Thermal energy storage; Chemical heat pump; Hydration/dehydration cycle of lime; Combustion engine

Corresponding author. Tel.: +34-946-014075.
E-mail address: iapazalm@bi.ehu.es (J.M. Morquillas).

1359-4311/03/$ - see front matter  2003 Elsevier Science Ltd. All rights reserved.
734 M.N. Azpiazu et al. / Applied Thermal Engineering 23 (2003) 733–741

1. Introduction

Research and development in high efficiency heat storage systems is very important for the
effective utilisation of thermal energy. Besides, increasing energy consumption is leading to higher
atmospheric pollution and making it necessary to find new ways to use successfully all kinds of
energy, including natural sources (solar and geothermal) and industrial waste heat.
Thermal energy produced nowadays by combustion of fossil fuels is energy chemically stored in
the past with a high storage density. Therefore thermal energy could be converted to chemical
energy and thus stored for a long period. For this reason the use of a reversible chemical reaction
with absorption of thermal energy endothermically, and its reversible release of stored heat
exothermically has been suggested. The repetition of endo/exothermic reactions in a closed cycle is
also attractive because we could employ the chemicals indefinitely, at least theoretically.
Thermochemical storage systems are more complex than others based, for example, on latent
heat because all the individual system components should be adequately assembled due to pos-
sible interactions between them. In these cases the reversibility of the selected reaction could be
disturbed by reactions involving chemical impurities [1].
Many reactions have been studied for this purpose such as the decomposition of magnesium
nickel hydride [2], the depolymerization of paraldehyde [3], the dehydrogenation of 2-propanol [4]
or carbon dioxide systems [5]. Fig. 1 shows the cycle for a thermal storage using the reaction
Ca(OH)2 fi CaO + H2 O that has also been reported [6,7].
An input energy of 148.6 kJ (¼54.0 + 94.6) can decompose 1 mol of Ca(OH)2 at 25 C to
CaO + H2 O at 510 C. The produced CaO can be easily stored as a solid at 25 C loosing 23.5 kJ.
If water is allowed to escape to the atmosphere, instead of storing it at 510 C in vapour phase,
61.5 kJ are additionally lost. Therefore, 57.2% (85.0 kJ) is the portion of input energy that results
in sensible heat of the products, and the remaining 42.8% (63.6 kJ) is actually stored in the form of
chemical energy.
When the stored high temperature steam at 510 C is used in the hydration step 125.1 kJ could
be recovered (84.2% of the initial input of energy), but the recovery of stored heat using liquid
water at 25 C can only provide the already mentioned 63.6 kJ.
This reaction cycle has been applied for several purposes such as self-heating food containers
for hot drinks like coffee, tea or milk [8], or sulphur emissions control in coal combustion [9]. The

+ 94.6 kJ
Ca(OH) 2 (s) CaO(s) + H2O(g)
510 ºC 510 ºC 510 ºC

+ 54.0 kJ - 23.5 kJ - 61.5 kJ

- 63.6 kJ
Ca(OH) 2 (s ) CaO(s) + H 2O(l)
25 ºC 25 ºC 25 ºC

Fig. 1. Ca(OH)2 /CaO dehydration/hydration cycle.

M.N. Azpiazu et al. / Applied Thermal Engineering 23 (2003) 733–741 735

authors wish to use it to preheat the combustion engine in motor vehicles [10], high pollutants
emissions and other undesirable effects of cold starting at very low temperatures could thus be
avoided, and the thermal energy of exhaust gases could be partially used.
For this purpose we used, as input energy, heat from the engine exhaust gases and water at 0 C
in the hydration step of CaO cooled down to 0 C. In this case, the exothermic reaction heat (63.6
kJ in Fig. 1), is partially used to heat CaO and water from 0 up to 25 C, and 60.6 kJ are left for
this particular application.
The previously described dehydration/hydration cycle of the Ca(OH)2 /CaO system has been
considered to analyze its heat release characteristics in a very simple static reactor and in a
prototype, designed to be used as a thermal storage device for the above mentioned application.
System performance during 20 cycles was followed studying its reversibility, chemicals degrada-
tion, side reactions due to impurities, heat release rate and mass losses. This limit of 20 cycles has
been stabilised because carbonation of CaO and Ca(OH)2 becomes a trouble in system perfor-
mance after 20 cycles of hydration–dehydration [11].

2. Experimental methods

The selected product for the experiments was a commercial Ca(OH)2 , which has the chemical
composition and characteristics shown on Table 1. It provided a very reactive CaO after dehy-
For the experimental static reactor conventional laboratory beakers of 25, 50, 100 and 250 ml
were used. Samples of 5 and 8 g of CaO were placed inside the 25 and 50 ml beakers respectively.
After cooling them down to 0 C, water at 0 C was poured over them with a selected proportion
CaO:H2 O of 1:1.5 (7.5 ml for 5 g samples and 12 ml for 8 g samples) [11].
These beakers were put into 100 ml beakers filled with an aqueous antifreeze solution at 0 C,
used to recover stored heat in the hydration step. Volumes of this heat recovery solution were
15 ml for 5 g samples and 24 ml for 8 g samples. The set was covered and surrounded by empty
250 ml beakers to avoid heat losses due to air streams around it.
The calcium hydroxide obtained in the preceding step was dehydrated in a furnace at 550 C
during 45 min, repeating these two steps up to 20 cycles.

Table 1
Chemical composition and characteristics of the commercial Ca(OH)2
Component %
SiO2 0.52
Al2 O3 0.14
Fe2 O3 0.08
P 2 O5 0.03
MnO 0.03
MgO 0.30
CaSO4 0.10
CaCO3 6.40
Ca(OH)2 92.40
Characteristics: Density: 0.35 g/cm3 . Particle size distribution: <200 lm: 95%; <90 lm: 89%. Specific heat: 1.66 kJ/kg K.
736 M.N. Azpiazu et al. / Applied Thermal Engineering 23 (2003) 733–741

Fig. 2. Scheme of designed thermal storage prototype.

During the hydration step, both temperatures of the reacting solid and of the surrounding heat
recovery solution (T1 and T2 ) were monitored with thermocouples. Samples from each cycle were
kept for further thermal analysis in a METTLER TA4000 equipment with a differential scanning
calorimetry (DSC) cell and a TC11 processor. To study the acquired data we used a computer
with the graph ware TA72, also provided by METTLER. Specific heat changes, enthalpies and
kinetic parameters were obtained in this way.
Additionally, trying to recover more efficiently the stored thermal energy, the hydration step
with 5 g samples was carried out using three 100 ml beakers filled each one with 15 ml of the heat
recovery solution, instead of only one. The beaker with the reacting CaO was successively in-
troduced in them. After the experiment all the solutions were mixed and its final temperature was
measured. The procedure was repeated up to 5 cycles.
Finally a heat exchanger prototype designed by the authors, following some ideas taken from
literature [12], was used (Fig. 2).
It consisted of a rectangular central body of 192 · 118 · 85 mm with fins of 192 · 20 · 1.5 mm to
enhance heat transfer. It was placed inside a rectangular container of 200 · 126 · 110 mm with a
lid. All was made of Cu, and covered with a 60 mm layer of polyurethane.
For the experiments 200 g of CaO and 360 ml of the heat recovery solution were used. The
hydration water was added with the same indicated proportion of 1:1.5 (300 ml). In all cases every
system component was cooled down to 0 C prior to starting the next trials. In some experiments
the heat recovery solution was renewed 90 and 240 s after the addition of the hydration water, to
see if heat recovery could reach higher efficiencies. Temperatures of the reacting solid (T3 ) and of
the aqueous antifreeze solution as it warms up (T4 ) were also monitored with copper/constantan

3. Results and discussion

Fig. 3 shows changes of heat flux received by the sample in W/g versus temperature, when it is
heated with a 10 C/min rate in the thermal analysis equipment. Data for analytical grade and for
M.N. Azpiazu et al. / Applied Thermal Engineering 23 (2003) 733–741 737

Fig. 3. DSC analysis of analytical grade (––) and commercial (- - -) Ca(OH)2 .

the commercial Ca(OH)2 , are plotted together. Dehydration temperatures (516 and 519 C) as
well as reaction enthalpies (1060 and 1040 J/g) were similar in both cases. Specific heats had
medium values of 1.66 kJ/kg K for Ca(OH)2 and 1.80 kJ/kg K for CaO.
Kinetic evaluations for this dehydration reaction were performed on DSC peaks with the
thermal analysis equipment software. Parameters were obtained according to the simple nth order
model equation combined with the Arrhenius approach of the temperature dependency of the
reaction rate constant (see Appendix A). Data derived from these calculations are presented in
Table 2.
In both cases reaction order (n) is very close to zero as it is reported by other authors [13] for
heterogeneous reactions with solids. It seems that commercial Ca(OH)2 has a slightly higher
dehydration rate that does not seem to disturb its performance as a heat storage system. Metals
additions or pressure lowering will be studied in the future in order to reduce dehydration tem-
Calcium hydroxide dehydration weight losses were of 18% for 5 g samples and 20% in 8 g
samples, evaluated with data after at least four cycles. Stoichiometric loss is a 24% indicating that
reaction is not complete with the adopted working method. Besides, heat recovery solution
temperature was lower after 15 cycles of operation as was reported in [11].
Some samples after 20 cycles were dehydrated at 1000 C and higher weight losses were ob-
tained. Reaction with CO2 , either dissolved in hydration water or present in air, could lead to
CaCO3 formation and thus higher dehydration temperature is needed. This possibility has also

Table 2
Kinetic parameters for Ca(OH)2 dehydration
Analytical grade Ca(OH)2 Commercial Ca(OH)2
ln k0 14.8 22.7
Ea (kJ mol1 ) 133 181
n 0.2 0.3
738 M.N. Azpiazu et al. / Applied Thermal Engineering 23 (2003) 733–741



Temperature (ºC) 20
Cyc 5/8g
Cyc 8/5g
10 Cyc 7/8g
Cyc 11/5g

0 100 200 300
Time (s)

Fig. 4. Heat recovery solution temperature (T2 ) evolution with time for two samples in different cycles in the static

been mentioned by other authors [14], and made us consider that the number of effective cycles
with this working method is about 20. Additionally repetitive hydration of magnesium oxide has
been studied by [15].
Hydration experiments suggested general system behaviour during this step that can be seen in
Fig. 4. Results for two samples in different cycles, carried out with the same experimental con-
ditions, are presented in it. Heat recovery solution temperature (T2 ) increased gradually with time,
reaching a basically constant value 4 min after hydration water addition. There are no significant
differences between samples and cycles reaching always an approximate final temperature of 35
C, which remained basically constant for another couple of minutes (not shown graphically).
This solution does not cool down because it is still receiving heat delivered by the reacting CaO
bed, with a lower heat transfer rate because temperature gradient is also lower.
This fact was considered as a possibility to obtain a more efficient recuperation of heat
changing periodically the heat recovery solution by a fresh one. Experiments using three 100 ml
beakers to recover stored heat made it possible to heat 45 ml of solution (3 · 15 ml in each beaker),
from 0 C up to a medium temperature of 13.5 C instead of 15 ml from 0 up to 35 C as in the
preceding experiments.
A calculation, based on the first principle of thermodynamics, shows that the 40% of the 5.43 kJ
released by the reacting CaO (5 g CaO and 1.09 kJ/g CaO) could be recovered by the heat recovery
solution (15 ml heated up to 35 C), increasing this percentage with an additional 7% when the
heat recovery solution was renewed. Heat losses were associated to heat absorption by beakers
(10%), sensible heat of the remaining Ca(OH)2 bed (5%) and, most of all, to partial evaporation of
the excess of the hydration water added.
Prototype performance was previously studied to establish the adequate position for thermo-
couples and guarantee its thermal insulation. This last point is very important to avoid heat losses
and, principally, to be sure that the heat recovery solution, initially at 0 C, is heated by the stored
heat and not by surrounding air at a medium ambient temperature of 20 C.
M.N. Azpiazu et al. / Applied Thermal Engineering 23 (2003) 733–741 739


Temperature (ºC)



0 100 200 300
Time (s)

Fig. 5. Temperature evolution with time for CaO and heat recovery solution (h.r.s.) in prototype experiments for
pouring (in black) and sudden addition (in white) of hydration water.

Fig. 5 shows CaO bed and heat recovery solution temperatures (T3 and T4 in Fig. 2) evolution
with time during the hydration step for two different experiments in the prototype. They were
performed slowly pouring the hydration water distributing it all around the CaO mass (in black)
or adding it suddenly in one system corner (in white).
Both temperatures in reacting CaO and heat recovery solution, reach a similar value about 5
min after water addition. It seems better to add hydration water slowly spreading it over the CaO
bed as heat release starts before, and heat recovery solution (360 ml) reaches a medium final
temperature of 55 C, while in the second case it warms up to 48 C.
As in the experiences with beakers, final temperatures were again constant during at least 25
min more (not graphically shown) pointing towards a more efficient heat release changing the heat
recovery solution. Fig. 6 shows data for these last experiments changing the solution by a fresh
one 90 and 240 s after hydration water addition.


Temperature (ºC)

75 h.r.s.



0 100 200 300 400
Time (s)

Fig. 6. Temperature evolution with time for CaO and heat recovery solution (h.r.s.) in prototype experiments changing
the recovery solution.
740 M.N. Azpiazu et al. / Applied Thermal Engineering 23 (2003) 733–741

In this case 360 ml of water were successively heated from 0 up to 42 C, 33 and 25 C in three
steps. There is still a temperature gradient between reacting CaO (d at 33 C) and the heat ab-
sorption media ( at 25 C) 420 s after the addition of water. Heat transfer could even go further,
in a fourth step, changing again the heat recovery solution.
Another simple calculation (again based on the first principle of thermodynamics) indicated
that the 38% of the 218 kJ released by reacting CaO (200 g) were recovered by heat recovery
solution (360 ml heated up to 55 C) and that this percentage rose up to 69% when the heat
recovery solution was changed by a new one. This 69% is significantly higher than the 40%
reached with beakers to heat water from 0 C to a similar final temperature (33.3 and 35 C

4. Conclusions

As principal conclusions of the present study we could mention the following:

1. Due to carbonation, the process does not reach a full reversibility in the dehydration step, and
therefore could be used with a limited number of cycles, that however can go up to 20. A pe-
riodic dehydration at 1000 C could be necessary to recover full system reversibility.
2. Efficiency is higher when the heat recovery solution is changed and a system with renewal of
this solution in a prototype made of Cu is a better design to heat water from 0 C up to a
medium final temperature of 33.3 C.
3. Hydration water addition should be carried out pouring it slowly with a uniform distribution
along the whole mass of CaO as in a spray.

Future experiments on prototype systems will try to minimize carbonation in order to reach a
higher number of operative cycles. For special applications in low temperature climates, reduction
of dehydration temperature of Ca(OH)2 with metals additions and working at lower pressures will
be studied, as well as hydration with antifreeze solutions at )10 C.

Appendix A

Simple nth order model equation combined with the Arrhenius approach of the temperature
dependency of the reaction rate constant. This combination results in the equation

r ¼ k0 expðEa =RT Þð1  aÞn

where r is the reaction rate in s1 ; k0 , the preexponential factor; Ea , the activation energy (J mol1 );
R, the gas constant (8.31 J mol1 K1 ); T , the sample temperature (K); a, the extent of reaction; n,
order of reaction.
Reaction rate and the extent of reaction, in the range of investigation, were calculated from
DSC peak:
M.N. Azpiazu et al. / Applied Thermal Engineering 23 (2003) 733–741 741

r ¼ DSC signal deviation from baseline ðmWÞ=peak area ðmJÞ;

a ¼ Partial area ðmJÞ=peak area ðmJÞ:


[1] M.N. Azpiazu, J.M. Morquillas, Almacenamiento Termoquımico de Energıa Termica, Centro de Publicaciones
E.T.S. Ing. Ind., Bilbao, 1988.
[2] T. Akiyama, T. Fukutani, R. Takahashi, J. Yagi, Heat storage rate of magnesium nickel hydride, Materials
Transactions, JIM 37 (5) (1996) 1014–1020.
[3] H. Kawasaki, A. Kanzawa, T. Watanabe, Characteristics or chemical heat pump through kinetic analysis of
paraldehyde depolymerization, Journal of Chemical Engineering of Japan 31 (3) (1998) 374–380.
[4] T. Doi, T. Tanaka, Fundamental study on solar chemical heat pump: basic experiment of dehydrogenation of 2-
propanol, ASME Solar Engineering G0656A (1997) 285–290.
[5] M. Kubota, K. Kyaw, F. Watanabe, H. Matsuda, M. Hasatani, Study of decarbonation of CaCO3 for high
temperature thermal energy storage, Journal of Chemical Engineering of Japan 33 (5) (2000) 797–800.
[6] I. Fujii, K. Tsuchiya, Y. Shikakura, M.S. Murthy, Consideration on thermal decomposition of calcium hydroxide
pellets for energy storage, Journal of Solar Energy Engineering 111 (1989) 245–250.
[7] M.N. Azpiazu, J.M. Morquillas, A. Vazquez, Performance of the Ca(OH)2 /CaO cycle as a long term heat storage
system, in: CALORSTOCKÕ94, 6th International Conference on Thermal Energy Storage, Spoo, Finlandia, 22–28
August 1994, pp. 469–477.
[8] S. Aitoh, Research and development of a self-heating food container using lime, in: 7th International Lime
Congress, Roma, Italy, 13–14 September 1990, pp. 353–382.
[9] P.K. Sharma, G.R. Gavalas, R.C. Flagan, Calcium pretreatment of coal for sulphur emissions control in
combustion, Fuel 66 (1987) 207–209.
[10] M.N. Azpiazu, J.M. Morquillas, A. Vazquez, Evaluation of a thermal storage system on warm-up exhaust
emissions, SAE Paper no. 921459, 1992.
[11] M.N. Azpiazu, J.M. Morquillas, A. Vazquez, H. Varona, Improvements of an energy storage system which uses a
reversible chemical reaction, in: THERMASTOCKÕ91, 5th International Conference on Thermal Energy Storage,
Scheveningen, The Netherlands, 13–16 May 1991, pp. 4.9-1–4.9-6.
[12] A. Kanzawa, Y. Arai, Thermal energy storage by the chemical reaction. Augmentation of the heat transfer and
thermal decomposition in the CaO/Ca(OH)2 powder, Solar Energy 27 (4) (1981) 289–294.
[13] J. Mu, D.D. Perlmutter, Thermal decomposition of carbonates, carboxylates, oxalates, acetates, formates and
hydroxides, Thermochimica Acta 49 (1981) 207–218.
[14] R.M. Dheilly, J. Tudo, M. Queneudec, Influence of climatic conditions on the carbonation of quicklime, Journal of
Materials Engineering and Performance 7 (1998) 789–795.
[15] Y. Kato, J. Nakahata, Y. Yoshizawa, Durability characteristics of the hydration of magnesium oxide under
repetitive reaction, Journal of Materials Science 34 (1999) 475–480.