Solar Energy 75 (2003) 395–401

www.elsevier.com/locate/solener

Experimental study of an innovative solar waterdesalination system utilizing a passive vacuum technique

S. Al-Kharabsheh, D. Yogi Goswami *

Solar Energy and Energy Conversion Laboratory, Mechanical and Aerospace Engineering Department, University of Florida,

220 MEB, P.O. Box 116300, Gainesville, FL 32611-6300, USA

Received 30 June 2003; received in revised form 1 August 2003; accepted 4 August 2003

Abstract

A solar desalination system based on an innovative passive vacuum concept, utilizing low-grade solar heat, was

studied experimentally. The system uses the natural means of gravity and atmospheric pressure to create a vacuum,

under which liquid can be evaporated at much lower temperatures and with less energy than conventional techniques. A

vacuum equivalent to 3.7 kPa (abs) or less can be created depending on the ambient temperature at which condensation

will take place. The system consists of a heat source, an evaporator, a condenser, and injection, withdrawal and dis-

charge pipes. The effect of various operating conditions (withdrawal rate, depth of water body and temperature of the

heat source) were studied experimentally and compared with theoretical results. The experimental results agreed well

with the theoretical predictions. It was found that the effects of withdrawal rate and the depth of water in the evaporator

were small while the effect of heat source temperature was significant.

� 2003 Published by Elsevier Ltd.

1. Introduction

Solar energy may be used to supply the required

energy for a desalination process either in the form of

thermal energy or electricity. The most widely used solar

desalination system is a simple solar still, where the heat

collection and distillation processes take place in the

same equipment. The main disadvantage of a simple

solar still is its low efficiency, which rarely exceeds 50%,

averaging value of 30–40% (Delyannis and Belessiotis,

2001). The daily solar still production is about 3–4 l/m2

(Kalogirou, 1997). Simple solar stills have been studied

extensively to improve their efficiency. A theoretical

analysis by Dunkle (1961), and the relations that he

derived for the heat and mass transfer within the still,

formed the basis for many research efforts since then.

Dunkle found out that the mass transfer rate depends on

the temperature difference between the water surface and

* Corresponding author. Tel.: +1-352-392-0812; fax: +1-352-

392-1071.

E-mail address: solar@mae.ufl.edu (D.Y. Goswami).

0038-092X/$ - see front matter � 2003 Published by Elsevier Ltd.

doi:10.1016/j.solener.2003.08.031

the glass cover. In order to increase this temperature

difference some researchers (Boukar and Harmim, 2001;

Kumar and Tiwari, 1996; Lawrence and Tiwari, 1990;

Tiwari et al., 1996; Yadav and Jha, 1989; Yadav,

1993; Yadav and Prasad, 1995) studied the effect of

coupling the solar still to a flat plate solar collector.

Another way to increase the temperature difference is to

reduce the temperature of the glass cover.The tempera-

ture difference between the saline water surface and the

transparent cover could be increased by adding a con-

denser to the still, thus increasing the heat sink capacity,

hence the still performance (El-Bahi and Inan, 1999;

Fath and Elsherbiny, 1993; Khalifa, 1985; Saatci, 1984).

Evaporation at a low temperature using vacuum con-

ditions, leads to a good improvement in the system

efficiency as the evaporation rate increases with the re-

duction of pressure. System productivity higher than

that from similar solar desalination systems operating

under atmospheric pressure has been reported by many

researchers (Abu-Jabal et al., 2001; Jubran et al., 2000;

Low and Tay, 1991; Tay et al., 1996; Uda et al., 1994).

The present study utilizes vacuum conditions for evap-

oration and condensation, where a vacuum is created in

396 S. Al-Kharabsheh, D.Y. Goswami / Solar Energy 75 (2003) 395–401

a unique way proposed by Sharma (1994) using natural

forces of gravity and atmospheric pressure.

This paper presents experimental test results of the

new water desalination system under passively created

vacuum conditions and compares them with the theo-

retical ones.

2. System description and operating principle

The proposed desalination system consists of a solar

heating system, and an evaporation chamber and a

condenser at a height of about 10 m above ground level,

connected via pipes to a saline water supply tank and a

fresh water tank, respectively. Fig. 1 shows a schematic

of the system. A vacuum is created by balancing the

hydrostatic and the atmospheric pressures in the supply

and discharge pipes.

The evaporation chamber has provisions to feed the

cold fluid directly to the chamber and provide solar or

other low-grade thermal energy through a closed loop

heat exchanger as well as withdrawing the concentrated

brine. The incoming cold fluid and withdrawn brine pass

through a tube-in-tube heat exchanger in order to ex-

tract the maximum possible heat from the hot brine. The

evaporation chamber is connected to a condenser, which

dissipates the heat of condensation to the environment.

It is known that the vapor pressure of seawater is

about 1.84% less than that of fresh water in the tem-

perature range of 0–100 �C. This means that if the top of

the two chambers (saline water evaporator and fresh

water condenser) are connected while maintained at the

same temperature, water will distill from the fresh water

side to the saline water side. In order to maintain the

distillation of potable water from the saline water the

vapor pressure of the saline water must be kept above

Fig. 1. Schematic of the system.

that of the fresh water, which is done by increasing the

temperature of the saline water by solar energy. To start

up the unit, it is filled completely with water initially.

The water is then allowed to drop down out of the unit

under the influence of gravity. Depending on the baro-

metric pressure, water falls to a level of about 10 m from

the ground level, leaving behind a vacuum above the

water level in the unit. Vacuum enables the distillation

of water at a lower temperature level, requiring less

thermal energy. This heat can be provided from solar

collectors, which will operate at a higher efficiency be-

cause of lower collector operating temperatures. Simple

flat plate collectors may be used to heat the saline water

in the evaporator.

As saline water in the evaporator starts evaporating,

its salinity increases which tends to decrease evaporation

rate, so it becomes necessary to withdraw the concen-

trated brine at a certain flow rate and inject saline water

at a rate equivalent to the withdrawal plus evaporation

rates. The withdrawn water will be at a temperature

equal to that of the evaporator, so it becomes necessary

to recover the energy from it. A tube-in-tube heat ex-

changer is used for this purpose, where injected water

flows inside the inner tube and withdrawn water will

flow in the annulus in a counter-current direction.

Under the influence of vacuum conditions at the sa-

line water surface in the evaporator, water can be in-

jected by the atmospheric pressure; hence no pumping

power is required. This makes the proposed system a

continuous process type, unlike a flat basin solar still

which is usually a batch process. The withdrawn con-

centrated brine can be concentrated further and used to

construct a solar pond, which may be used as a solar

energy collection and storage system. The system will

require periodic cleaning by flushing and restarting it,

so that the non-condensable gases are not allowed to

accumulate destroying the vacuum.

3. Experimental setup

A small-scale system was designed and built. A photo

of the evaporator condenser is shown in Fig. 2, and the

experimental setup is shown in Fig. 3.

The experimental system has the following specifi-

cations. The heat exchanger coil for heat input to the

saline water is a copper tube of 2.4 m length and 1.27 cm

outside diameter. The evaporator is a cylinder of 0.2 m2

cross sectional area, 0.2 m height, with a truncated cone

on top of it. The evaporator has a provision for feed

water, through a 1.27 cm diameter copper tube, enclosed

by 2.54 cm CPVC pipe that is used for withdrawing the

concentrated brine. The two pipes form a tube-in-tube

heat exchanger. The condenser is a 10.16 cm copper tube

of 0.5 m length, 0.25 cm thickness. On its lateral surface,

10 copper fins of 25.4 cm diameter and 0.0635 cm

Fig. 2. Photo of the evaporator–condenser.

Fig. 3. Photo of the experimental setup.

S. Al-Kharabsheh, D.Y. Goswami / Solar Energy 75 (2003) 395–401 397

thickness are soldered 4 cm apart. The other end of the

condenser is connected to a condensate receiver via a

1.27 cm PVC pipe.

Table 1

Measurement uncertainty limits

Quantity Uncertainty

Flow rate (kg/min) 5.3%

Output (kg/day) 6.8%

Heat input (W) 5.25%

4. Results

A number of tests were performed covering vari-

ous combinations of operating conditions. The results

are presented in this section. Theoretical results are also

presented which are obtained using a simulation model

presented in earlier papers by the authors (Al-Kharab-

sheh and Goswami, 2003a,b). The theoretical results

were based on the measured heat source and ambient

temperature profiles from the experiments. A measure-

ment uncertainty analysis was conducted using standard

deviation method to calculate the uncertainty of directly

measured values within a given level of confidence

(taken as 90%). For quantities calculated from measured

quantities, the method of propagation of errors was

used, where the total uncertainty was calculated as the

combination of uncertainties of individual components

(Wilson, 1952). The obtained uncertainty limits are

given in Table 1.

The first set of tests were performed for the heat

source temperature of 60 �C, a withdrawal rate of 0.1 kg/h, and the depth of water in the evaporator as 0.08 m.

This test was repeated six times. Fig. 4 shows the tem-

perature profiles of the saline water as a function of

time. Also included in the figure are the ambient tem-

perature profiles during those tests. As can be seen from

the figure, saline water temperature profiles show the

same trend, and at steady state conditions the difference

between the maximum and the minimum saline water

temperatures is about 2.1 �C, or about 4%. This differ-

ence can be attributed to the variation in the ambient

temperature during the tests and the measurement un-

certainty.

The water outputs during these six tests are shown in

Fig. 5. These tests were conducted under the same ex-

perimental conditions except for the ambient tempera-

ture, which was different for each day because of the

Fig. 4. Saline water and ambient temperatures for six tests

(each line style represents one test day). Source tempera-

ture¼ 60 �C; withdrawal rate¼ 0.1 kg/h; and evaporator water

depth¼ 0.08 m.

Fig. 5. Theoretical and experimental output for six tests.

Source temperature¼ 60 �C; withdrawal rate¼ 0.1 kg/h; and

evaporator water depth¼ 0.08 m.

Fig. 6. Experimental and theoretical saline water temperature

and heat input with time. Source temperature¼ 60 �C; with-drawal rate¼ 0.1 kg/h; and evaporator water depth¼ 0.08 m.

398 S. Al-Kharabsheh, D.Y. Goswami / Solar Energy 75 (2003) 395–401

outdoor tests. This explains the difference between the

results from various tests. The first day, which has the

lowest ambient temperature, has a significantly higher

output (about 0.9 kg) than the other days. It is seen from

these results that the ambient temperature affects the

system output significantly. The lower the ambient

temperature, the higher the output. A similar trend ex-

ists with respect to the fresh water temperature.

Also shown in Fig. 5 are the theoretical simulation

results based on measured temperature profiles of the

heat source fluid and the ambient. The experimental and

theoretical results agree very well. The maximum dif-

ference is about 0.049 kg, or about 5%, which is within

the measurement uncertainty. It was noticed that the

experimental values were always less than the theoretical

ones. This may be attributed partly to the fact that the

theoretical model assumes that all molecules evaporated

from the saline water inside the evaporator will reach the

condenser and condense as liquid water, whereas in real

life a number of those molecules might fall back to the

pool. Also the model assumes that the fins are an inte-

gral part of the condenser, whereas they were soldered to

the condenser surface which would reduce the rate of

heat transfer from the condenser, hence the system

output would be reduced. Another possible reason is

that the model assumes the heat loss from the system to

be by natural convection only, whereas it might be

higher because of wind. Both experimental and theo-

retical outputs are broken down into daytime output

(time during which heat is supplied to the system) and

nighttime output (time after the heat is no longer sup-

plied to the system), as shown in Fig. 5. The night time

output is a result of the heat stored in the system during

the initial hours of operation, and varies slightly for

different tests, depending on the saline water tempera-

ture at the end of the test and the ambient temperature.

The maximum and minimum experimental night time

outputs (excluding the first day) were 0.319 and 0.290

kg, respectively. The saline water temperatures at the

end of these two days were 49.7 and 49.4 �C, respec-tively, but the average ambient temperatures were 27

and 30 �C, respectively.Additional results from the previous six tests are

presented in Figs. 6 and 7. Fig. 6 shows how the saline

water temperature and the heat input vary with time

both experimentally and theoretically.

Also shown in the figure are the heat source and

ambient temperatures. The saline water temperature

increased with time and reached a steady state value of

about 50 and 48 �C, for experimental and theoretical

results, respectively. The higher value obtained from the

experiments may be due to the fact that the temperature

was measured at a distance of 10 cm from the evapo-

rator wall, where the temperature may be slightly higher

than that near the wall, where the heat loss takes place.

In the theoretical model, the saline water temperature

was assumed to be uniform throughout the evaporator.

The heat input to the system was high initially, to raise

the temperature of the saline water and the evaporator

material (stored as sensible heat). This stored heat be-

comes useful at night time. As the system reached steady

state, the experimental and theoretical values of the

Fig. 7. Variation of the output, output rate, and energy and

exergy efficiencies with time. Source temperature¼ 60 �C;withdrawal rate¼ 0.1 kg/h; and evaporator water depth¼ 0.08

m.

Fig. 8. Effect of depth of water body on the system perfor-

mance.

Fig. 9. Effect of withdrawal rate on the system performance.

S. Al-Kharabsheh, D.Y. Goswami / Solar Energy 75 (2003) 395–401 399

energy input were about 109 and 103 W, respectively.

The difference is about 5%, which is within the uncer-

tainty of measurements.

Fig. 7 shows the variation of the theoretical output

rate, accumulated output, and the energy and exergy

efficiencies with time, based on the measured tempera-

ture profiles of the heat source fluid and the ambient.

The accumulated output during the six hours test

reached a value of 0.495 kg compared to 0.462 kg ob-

tained experimentally. The difference is about 6.6%,

which is within the measurement uncertainty. At steady

state the output rate was 0.115 kg/h compared to 0.108

kg/h obtained experimentally, a difference of about 6%.

The energy and exergy efficiencies were low at the be-

ginning of the test and increased with time. When the

system reached steady state conditions the energy and

exergy efficiencies reached values of 74% and 79%, re-

spectively.

The effect of water depth in the evaporator on the

system performance is shown in Fig. 8. Three different

depths were considered: 0.06, 0.08, and 0.1 m. For these

tests, the heat source temperature was set at 60 �C and

the withdrawal rate was 0.1 kg/h. The effect of water

depth should not be significant, but there are some dif-

ferences due to the variation in the ambient tempera-

tures. The saline water temperature for the test day

corresponding to a water depth of 0.1 m was about 47

�C, as compared to about 50 �C for the other two days.

However, the test with water depth of 0.1 m gave the

highest accumulated output of about 0.908 kg, with the

output rate of about 0.124 kg/h under steady state

conditions, compared to an accumulated output of

about 0.750 kg and an output rate of about 0.118 kg/h

for the other two test days. This is because the average

ambient temperature for that day was 21 �C compared

to about 27 �C for the other two days resulting in a

temperature differential of 26 and 23 �C, respectively.The agreement between the experimental and theoretical

output results for each depth is within the experimental

uncertainty. Although no comparison of experimental

outputs and output rates can be made for the present

concept with the conventional solar still, the fact that the

experimental results agree with the theoretical predic-

tions, allows us to use a comparison based on theoretical

simulations. Theoretical simulation showed that this

new concept would give as much as two times the output

of a simple flat basin solar still for the same input and

evaporator area. (Al-Kharabsheh and Goswami, 2003b)

The effect of withdrawal rate is shown in Fig. 9.

Three different withdrawal rates were considered: 0, 0.1,

and 0.5 kg/h. For these tests, the heat source tempera-

ture was 60 �C and the depth of water body was 0.08 m.

The average ambient temperatures during these tests

were almost the same. The effect of withdrawal rate was

found to be very small. With the increase in the with-

drawal rate from 0 (batch process) to 0.5 kg/h, the

output rates and the accumulated outputs decreased

slightly from 0.119 kg/h and 0.763 kg to 0.116 kg/h and

0.755 kg, respectively.

The effect of heat source temperature is shown in Fig.

10. Three different temperatures were considered, 40, 50

and 60 �C. For these tests, the withdrawal rate was 0.1

kg/h, and the depth of water body was 0.08 m. Although

the ambient temperatures were different for the different

tests, the effect of heat source temperature is significant.

Fig. 10. Effect of heat source temperature on the system per-

formance.

400 S. Al-Kharabsheh, D.Y. Goswami / Solar Energy 75 (2003) 395–401

The output rate, accumulated output and saline water

temperature all increased with the heat source temper-

ature as shown in the figure. The theoretical results agree

very well with the experimental results, except for the

saline water temperature. As explained earlier, the dis-

agreement in these temperatures is mainly because the

measurement was made at 10 cm away from the evap-

orator wall while the theoretical analysis assumed a well

mixed uniform temperature in the evaporator.

5. Conclusions

An innovative solar water distillation system, that

uses a vacuum created by natural forces, was studied

experimentally. The system is a continuous type as op-

posed to a batch type conventional solar still. Because of

passively created vacuum conditions, it requires lower

temperatures for distillation, which can be easily pro-

vided by flat plate solar collectors. The experimental

results agree well (within the experimental uncertainty)

with the theoretical simulation results. Although it was

not possible to compare the experimental results of this

innovative system with a conventional solar still, based

on the agreement of the experimental and theoretical

results, theoretical simulations can be used for com-

parison. Based on theoretical simulations the present

system would perform much better than a simple flat

basin solar still. A multi-effect system based on the same

principle, which would utilize the latent heat of con-

densation from one stage to evaporate a part of water in

the next stage, would improve the performance even

further. The effects of water depth in the evaporator,

concentrated brine withdrawal rate, and heat source

temperature on the system performance were studied.

The effect of water depth in the evaporator and the

withdrawal rate of concentrated brine (in the range

considered) were found to be small, whereas, the effect

of heat source temperature was found to be significant.

References

Abu-Jabal, M.S., Kamiya, I., Narasaki, Y., 2001. Proving test

for a solar-powered desalination system in Gaza-Palestine.

Desalination 137, 1–6.

Al-Kharabsheh, S., Goswami, D.Y., 2003a. Theoretical anal-

ysis of a water desalination system using low-grade solar

heat. In: Proceedings of the 2003 ASME International Solar

Energy Conference, Hawaii, USA.

Al-Kharabsheh, S., Goswami, D.Y., 2003b. Analysis of an

innovative water desalination system using low-grade solar

heat. Desalination 156, 323–332.

Boukar, M., Harmim, A., 2001. Effect of climatic conditions on

the performance of a simple solar still: a comparative study.

Desalination 137, 15–22.

Delyannis, E., Belessiotis, V., 2001. Solar energy and desalina-

tion. In: Goswami, D.Y. (Ed.), Advances in Solar Energy,

An annual review of research and development, American

Solar Energy Society, Inc, Boulder, Colorado, pp. 287–

330.

Dunkle, R.V., 1961. Solar water distillation: The roof type still

and a multiple effect diffusion still. In: Proceedings of the

International Conference of Heat Transfer, University of

Colorado, Publication No. 108, pp. 895–902.

El-Bahi, A., Inan, D., 1999. A solar still with minimum

inclination, coupled to an outside condenser. Desalination

123, 79–83.

Fath, H.E.S., Elsherbiny, S.M., 1993. Effect of adding a passive

condenser on solar still performance. Energy Conversion

and Management 34 (1), 63–72.

Jubran, B.A., Ahmed, M.I., Ismail, A.F., Abakar, Y.A., 2000.

Numerical modeling of a multi-stage solar still. Energy

Conversion and Management 41, 1107–1121.

Kalogirou, S., 1997. Survey of solar desalination systems and

system selection. Energy 22 (1), 69–81.

Khalifa, A.N., 1985. Evaluation and energy balance study of a

solar still with an internal condenser. Journal of Solar

Energy Research 3 (1), 1–11.

Kumar, S., Tiwari, G.N., 1996. Performance evaluation of an

active solar distillation system. Energy 21 (9), 805–808.

Lawrence, S.A., Tiwari, G.N., 1990. Theoretical evaluation of

solar distillation under natural circulation with heat ex-

changer. Energy Conversion and Management 30 (3), 205–

213.

Low, S.C., Tay, J.H., 1991. Vacuum desalination using waste

heat from a steam turbine. Desalination 81, 321–331.

Saatci, A.M., 1984. Heat-pipe assisted solar still. In: Proceed-

ings of the 1984 ASME Annual Meeting, pp. 249–253.

Sharma, S.K., Goswami, D.Y., 1994. Low temperature energy

conversion system, unpublished invention disclosure, Office

of Technology Licensing, University of Florida, Gainesville,

FL.

Tay, J.H., Low, S.C., Jeyaseelan, S., 1996. Vacuum desalination

for water purification using waste heat. Desalination 106,

131–135.

Tiwari, G.N., Kumar, S., Sharma, P.B., Khan, M.E., 1996.

Instantaneous thermal efficiency of an active solar still.

Applied Thermal Energy 16 (2), 189–192.

Uda, K., Sato, H., Watanabe, K., 1994. Development of

advanced evacuated solar still. In: Proceedings of the 1994

ASME Joint Engineering Conference, pp. 513–519.

S. Al-Kharabsheh, D.Y. Goswami / Solar Energy 75 (2003) 395–401 401

Wilson Jr., E.B., 1952. An Introduction to Scientific Research.

McGraw-Hill Book Company Inc., New York.

Yadav, Y.P., Jha, L.K., 1989. A double-basin solar still coupled

to collector and operating in the thermosyphon mode. Solar

Energy 14 (10), 653–659.

Yadav, Y.P., Prasad, A.S., 1995. Performance analysis of a high

temperature solar distillation system. Energy Conversion

and Management 36 (5), 365–374.

Yadav, Y.P., 1993. Performance analysis of a solar still coupled

to a heat exchanger. Desalination 91, 135–144.