- Email: [email protected]
Condensation tests on glass samples for energy efficient windows
ARTICLE IN PRESS
Solar Energy Materials & Solar Cells 91 (2007) 609–615 www.elsevier.com/locate/solmat
Condensation tests on glass samples for energy efficient windows Anna Werner, Arne Roos Department of Engineering Science, Solid State Physics, The A˚ngstro¨m Laboratory, Box 534, SE-751 21 Uppsala, Sweden Received 6 September 2006; accepted 14 November 2006 Available online 22 January 2007
Abstract This article presents results from a pilot study comparing condensation patterns on small glass samples with different surface properties. Experiments were carried out on three commercial glass samples (clear float, TiO2-coated and SnO2-coated) to see how water condensed on the different surfaces. The experiments were carried out under a clear night sky in Uppsala, Sweden. It was found that the pane with the low-emittance coating stayed clean of condensation longer than the other two. In the morning, the water layer on the TiO2coated sample was smeared out so that it was possible to see through that pane, while the view through the other two was still blurred. The TiO2 coating does not prevent condensation, but makes it easier to see through the water layer. These simple tests indicate noticeable differences between different surface materials and also that these effects can be studied by exposing small samples to a clear night sky without having to perform full scale tests. r 2007 Elsevier B.V. All rights reserved. Keywords: External water condensation; Energy efficient windows
1. Introduction Most of us have experienced water condensation on windows, either between the window panes or on the inside [1]. In both cases, the condensation indicates that the window is of poor quality or has a high U-value. In the former case, humid air leaks into a space that is supposed to be closed or properly ventilated and in the latter case water condenses on the inner window surface because the window is considerably colder than the room [2]. That means that heat leaks out from the room through the window. The phenomenon of interest in this article, external condensation on windows, is experienced when the temperature of the external glass pane of a window goes below the outside dew point due to radiative cooling [3–5]. This is a relatively new phenomenon and appears during clear nights on wellinsulated windows for which the thermal losses do not balance the radiative cooling of the external glass surface. Such a decrease in pane temperature is desirable when a radiative cooling system is designed and has, therefore, been Corresponding author. Tel.: +46 18 471 77 83; fax: +46 18 50 01 31.
E-mail address: [email protected] (A. Werner). URL: http://www.angstrom.uu.se. 0927-0248/$ - see front matter r 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.solmat.2006.11.015
studied for that application [6–10], but not in relation to well-insulated windows. Condensation happens during clear nights when radiation from the window towards the sky is large. The scope of condensation depends on the environment, the building and the climate (weather). The characteristics of the outer surface also affects how fast condensation appears and how it is distributed across the surface [11,12]. Often small droplets are formed and they cause light scattering which obstructs the view through the window. A general resistance against using well-insulated windows because of this phenomenon can be noticed [13]. In order to avoid, or at least reduce, the problem with external water condensation on windows, thin film glass coating technology can be utilized. Especially two concepts are of interest here, low-emissivity coatings and hydrophilic coatings. A low-emissivity coating will decrease the occurrence of condensation, whereas a hydrophilic surface will help smearing out the water on the surface so that the view through the window is less obstructed. Many factors, both surface properties and environmental conditions, affect the process of water vapour condensation on a solid surface. There are broadly speaking two types of condensation: dropwise condensation and film
ARTICLE IN PRESS 610
A. Werner, A. Roos / Solar Energy Materials & Solar Cells 91 (2007) 609–615
condensation. In the former case, the condensate is created in the form of individual drops and, in the latter case, the condensate forms a continuous layer on the surface [14]. Mixed types of condensation also exist. Our work presented here is a first attempt to evaluate how different coatings on the outer surface affect the condensation process. The aim is to evaluate different ways of avoiding the problem of obstructed view through a window due to external water condensation and to develop a technique to monitor the occurrence of condensation. Since the perceived problem of external condensation on well-insulated windows in cold climates is relatively new, only few observations and simulations have been carried out [15–17], but no systematic study of the actual process. Experiments have not been undertaken, since the night sky is difficult to simulate in a laboratory. A more investigated problem is external condensation on poorly insulated windows in hot climates in air-conditioned buildings [18]. Then, dew on the window is a sign of poor thermal insulation, whereas in our case, dew on the window is a sign of excellent thermal quality. Wellinsulated windows constitute a way to save energy [19,20], and it is therefore, important to find a way to make the view through these windows less obstructed by possible condensation. We decided to study three glazing products (see Table 1); uncoated clear float glass (surface emissivity 0.84) as a reference since most windows of today do not have any external coating, a low-emissivity fluorine-doped tin dioxide hard coating (surface emissivity 0.15) since it is often used on internal surfaces and, therefore, commercially available, and a titanium dioxide coated self-cleaning glazing product (surface emissivity 0.84) since it has the desired hydrophilic properties and is expected to gain a larger market share. Both coated products are well known for their durability and can, therefore, be used with the coating on the external surface. Most other low-emissivity products can only be used in a sealed insulated glass unit and are, therefore, not feasible for this possible future application. The solar optical properties, such as the reduction of the solar factor (g), must be considered when external coatings are applied on windows. This was, however, beyond the scope of the present study. The object of this work was primarily to perform a feasibility study to see to which extent simple tests with only visible inspection of small
samples can provide information about the formation of condensation on different glass coatings. 2. Experimental setup Our experiments, accounted for here, were not carried out on real, installed windows, but on small test samples. The sample size was not regarded as being critical since edge effects were expected to be small also for small samples because of the low heat conductivity of glass. For the first experiment, we used smaller, 5 cm 10 cm, samples simply to test the feasibility of the methodology. In the two following experiments, larger, 30 cm 30 cm, samples were used to get a better visual impression of the condensation. The glass samples were left facing the night sky as described in Sections 2.1–2.3. In two of our experiments, the test panes were placed horizontally and in a third experiment they were put in an almost upright position to better simulate a real window situation. Since the panes were tilted instead of vertical, condensation would occur more often than on a real window. For a vertical window, the surface normal points in the direction of the horizon and the sky only covers the upper half of the hemisphere seen by the glass surface. For a tilted window, the sky covers more than half of the hemisphere (the view factor is larger than 0.5) and hence the radiative cooling from the sky is higher [21]. The inclination of the glass pane was not expected to affect the formation of condensation in any other way than changing the time of the onset of condensation, i.e., it was not expected to affect the structure of the condensation. When much condensation has formed the inclination will affect how easily the drops run down the pane, but in our experiments that latter part of the condensation process was not primarily studied. All tests were performed during clear summer nights, with almost no wind, since condensation is more likely to occur under these conditions [16]. We did not measure the relative humidity or calculate the dew point, since we were primarily interested in monitoring differences between the samples. The pane temperature was measured with thermocouples. For upcoming experiments, an IR-camera could give additional information. These initial experiments were not undertaken to verify any models when condensation occurs. We used visual inspection, but for future experiments, we have developed a testing equipment that can
Table 1 Our samples Called here
Description
Clear float glass Low-emissivity coated glass
Traditional commercial clear float glass with no coatings Commercial pane with a low-emissivity coating with a surface resistivity of about 20 OX&. This pane is today used in well-insulated windows with the coating either on the outer side of the innermost pane or on the inner side of the external pane, i.e., the coated side is not turned outside as in our experiments and measurements Commercial pane with a titanium dioxide, ‘‘titania’’, coating with self-cleaning properties. This pane is relatively new. It has been sold for only a couple of years and stands for a very small part of the world float glass pane market
Self-cleaning glass
The three glass samples tested in our experiments.
ARTICLE IN PRESS A. Werner, A. Roos / Solar Energy Materials & Solar Cells 91 (2007) 609–615
611
more accurately detect when the water film forms [22]. We did not collect the dew from the panes, but this could give an idea of the quantity of water condensed on the individual samples [23]. In all experiments described here, the samples were insulated from the back with 30 mm of polystyrene. This way there was almost no heat transport through the glass samples from behind, as is the case for the outer pane in a well-insulated window, and the equilibrium temperature was mainly determined by the surface of the sample and by the sky temperature. Since the air speed was low, convection and conduction were low and the surrounding air could not fully balance the radiative cooling of the surface. Therefore, it was easier for the glass to reach a temperature lower than the surrounding air and for condensation to occur. 2.1. High and low emissivity The aim of the first experiment was to see if there is a simple way to monitor surface temperature and to visually discern external condensation on glass samples exposed to a clear night sky. The aim was also to see if there is a simple way to compare different coated glass samples to see which one is more prone to condensation. The experiment was undertaken during night when the outdoor air temperature was about 13 1C. Two glass samples, an uncoated clear float sample and a tin dioxide-coated clear float sample, of size 5 cm 10 cm were heated in an oven to about 50 1C and placed horizontally on a piece of polystyrene (Fig. 1). A thermocouple below each sample measured the temperature. Both samples had a thickness of 4 mm. The heating to 50 1C was done to monitor the difference in cooling rate more clearly. The experiment started at midnight. 2.2. Horizontal test of three samples The aim of this second experiment was to compare condensation proneness and properties of three commer-
Fig. 2. Experimental set up at the beginning of the second experiment, about 8 o’clock in the evening, 45 min before sunset.
cial glass samples, two of which were the same as in the first experiment. Another aim was to study how the temperature of a clear float sample and a tin dioxide-coated glass sample varies with the air temperature and relative humidity. Three small (30 cm 30 cm) glass samples (clear float, tin dioxide-coated and titanium dioxide-coated) of equal thickness were placed horizontally on a piece of polystyrene (Fig. 2), i.e., insulated from below as in the previous experiment and exposed to the clear night sky. A thermocouple below the clear float sample and the tin dioxide coated glass sample measured their temperatures. The TiO2 surface has an emissivity similar to the one of uncoated glass and it can, therefore, be assumed that it had the same temperature as the uncoated sample. The samples were not pre-heated as in the first experiment described above. A lamp was installed above the samples to provide a reflected image for the photographs. It was not lit, except for when photographs were taken. There was almost no wind. 2.3. Vertical test of three samples The aim of the third experiment was to compare condensation proneness and properties of the same three commercial glass samples in an almost vertical position. Three small (30 cm 30 cm) glass samples (clear float, tin dioxide-coated and titanium dioxide-coated glass) of equal thickness were placed outdoors, almost vertically, with polystyrene as backing material. They were exposed to the clear night sky and almost no wind. Their temperatures were not measured. 3. Results
Fig. 1. The two glass samples in the morning, about 7 h after experiment begun (t ¼ 420 min). Condensation was now present on both samples. The water layer had grown differently on the two different surfaces. The clear float glass sample is to the left and the tin dioxide-coated sample to the right.
3.1. High and low emissivity In the first experiment, the clear float glass condensation occurred after about 40 min of exposure to the clear night
ARTICLE IN PRESS A. Werner, A. Roos / Solar Energy Materials & Solar Cells 91 (2007) 609–615
612
sky. After 90 min, we observed no condensation on the tin dioxide-coated glass. After 3 h also to the tin dioxidecoated glass condensation was observed, but considerably less on the clear glass. Fig. 1 shows the set up in the morning, about 7 h after the experiment begun. Condensation was then present on both surfaces, but the water had been distributed differently across the two surfaces. The tin dioxide-coated sample both had a lower cooling rate and maintained a higher temperature than the clear float sample throughout the experiment (Fig. 3) due to reduced radiant losses. As can be seen in Fig. 3, the onset of condensation in both cases took place at a surface temperature of approximately 12 1C. The experiment showed that with a very simple set up, it is possible to measure and monitor the temperatures of the glass samples, the air temperature and the occurrence of condensation on small size samples. The experiment also showed that the thermal mass of these samples is sufficiently high; it is possible to monitor differences between the samples, both surface temperature and time when condensation is present. From the pictures taken during the experiments, it is also possible to see that the condensation has different characteristics on the different surfaces. 3.2. Horizontal test of three samples In the second experiment (Fig. 2), 1 h after sunset, the clear float and the titanium dioxide-coated samples already Measurement Data from First Experiment 50
Temperature (°C)
40
T (K-glass) T (Clear glass) T (K-glass) - T(Clear glass) T (Air)
30
20
had some condensation, whereas the tin dioxide-coated glass sample had none. On the clear float sample, drops of water had formed. A paper with printed text was placed just behind the surface. This was still sharp and easy to read. After another half an hour, there was condensation on two of the samples, but on the tin dioxide-coated glass sample there was no water condensation. On titanium dioxide coated sample condensation was more than on clear float sample. The surface emissivity of the TiO2 surface is approximately the same as that of uncoated glass and the temperatures must be the same due to the radiation balance. The fact that more condensation occurs on the TiO2-coated surface could be due to surface contamination and minor differences in nucleation sites. The difference was not dramatic and this issue was not further investigated. After two more hours, the condensation layers differed much on these three surfaces. The titanium dioxide-coated sample had a lot of water on it and the tin dioxide-coated glass sample had a milky condensation layer with small, fine drops, looking more like haze. The surface was clearly wet. The uncoated sample had a condensation layer, which looked like something in between the layer on the titanium dioxide-coated sample and the layer on the tin dioxidecoated glass sample. Figs. 4–6 show what the samples looked like in the morning, about 11 h after the experiment started. The clear float looked almost like in the previous picture, the tin dioxide-coated glass was very wet and now showed big drops unevenly distributed over the surface, maybe due to some surface contamination. On the titanium dioxide-coated glass, the condensed water had formed a film rather than drops. From Figs. 4–6 it can be seen how the water drops distorted the reflection of the lamp shown in Fig. 2. The air temperature (upper curve in Fig. 7) decreased slowly during the experiment. The clear float sample started at a lower temperature, which decreased somewhat slower than the air temperature. The tin dioxide-coated glass sample started at the air temperature, but its temperature decreased faster than the air temperature. Condensation seems to have occurred at approximately the
10
0 0
50
100
150
200
Time (minutes) Fig. 3. Temperature versus time for the two tested surfaces of the first experiment: The temperature of the tin dioxide-coated glass sample (solid curve with large squares) was above the temperature of the clear glass sample (solid curve with small triangles) throughout the first experiment. The solid curve without markers shows the air temperature. The lowest curve (dashed with small squares) shows the difference between the two surface temperatures. The horizontal and two vertical dashed lines mark the approximate onset of condensation on the two samples.
Fig. 4. Clear float sample about 11 h after beginning of experiment.
ARTICLE IN PRESS A. Werner, A. Roos / Solar Energy Materials & Solar Cells 91 (2007) 609–615
613
same temperature, about 12 1C, on both the clear float and the tin dioxide-coated glass samples. This was expected since the dew point temperature varies only slightly during most nights and depends on the relative humidity which increased slowly during the experiment, as can be seen in Fig. 7. 3.3. Vertical test of three samples
Fig. 5. Tin dioxide-coated glass sample about 11 h after beginning of experiment.
The pictures (Figs. 8–10) were taken in the morning after the third experiment, when the three samples had been out all night. The samples were put one by one in a frame for the photographs to be taken. The visual appearance of the samples was clearly different. In the morning, it was not possible to see through the clear float sample or the tin dioxide-coated sample, but it was possible to see through the titanium dioxide-coated glass sample. The TiO2-coating had not prevented condensation, but made it easier to see through the water layer. In Figs. 8 and 9, it can be seen that drops of water have been formed and are sliding down the pane. The trails of these drops might be visible even after the condensed water has evaporated if the pane is dirty. This unwanted effect seems to be avoided with a titania coating (Fig. 10).
Fig. 6. Titanium dioxide-coated glass sample about 11 h after beginning of experiment.
Measurement data from Second Experiment 100
20 Relative Humidity
75
T (Air) T (Low-e Glass) T(Clear Glass)
16
50
14
RH ( %)
Temperature (°C)
18
Fig. 8. First sample in experiment, clear float glass.
12 25 10 0
8 0
1
2
3
4
5
Time (hours) Fig. 7. Temperature of clear float and tin dioxide-coated glass samples and relative humidity during a clear night with very slight wind. The relative humidity is measured on the right y-axis. The horizontal line shows the approximate glass temperature at which condensation was observed on the two surfaces. The two vertical lines point at the measurements when condensation was first detected for each glass sample, respectively. The time 0 in the graph indicates the start of the experiment, i.e., 7.45 p.m. August 19, 2005. The sunset was at around 8.30 p.m.
Fig. 9. Second sample in experiment, tin dioxide-coated glass.
ARTICLE IN PRESS 614
A. Werner, A. Roos / Solar Energy Materials & Solar Cells 91 (2007) 609–615
is formed, the emissivity increases slightly and also this surface becomes covered by water. Under more realistic conditions dew is less likely to form on the low emissivity coated glass. For a real window, the vertical position is the most common. The matter of inclination was not studied here. The tilted angles were only used to stimulate the formation of condensation. 5. Conclusion and future outlook
Fig. 10. Third sample, titanium dioxide-coated glass pane is staged like the other two panes. In the upper part of the picture, the sky is reflected in the pane.
4. Discussion For an in-depth analysis of the circumstances under which condensation is formed more detailed climatic data, such as wind speed, relative humidity and long wave incoming sky radiation, are needed [24]. The present pilot study, however, is limited to comparing condensation proneness and characteristics of different surface coatings under identical conditions and to evaluating if the differences can be detected by visual inspection. The results indicate that it is possible to study the formation of condensation on single glazed samples under a clear night sky. There are clear differences between different coatings both regarding formation of condensation and subsequent evaporation. One of the issues with external condensation mentioned by customers is the pattern appearing when the amount of condensation is sufficient for water drops to run down the usually slightly dirty glass surface, thus causing visible streaks to appear even after the condensation has evaporated. It would be of interest to evaluate the selfcleaning type of surface in this respect, but this would require more tests with larger samples. The condensation always disappears in the morning after sunrise. Even the diffuse radiation from a cloudy sky supplies sufficient energy to cause the water layer to evaporate. The rate of evaporation may differ for different surfaces depending on the wetting angle of the water but this was not further investigated. The external coating on the window inevitably leads to a reduced value of daylight and the solar energy transmitted. Thus the energy balance of the window is affected negatively and some of the energy benefit of using a wellinsulated window is lost. The overall evaluation of the benefit of these coatings must involve also the negative effects and the cost. This investigation was only focussed on the feasibility of using these coatings to prevent external condensation. For the low-emissivity coating the formation of condensation was considerably delayed, but once condensation
It was shown that experiments on small pane samples can show that the surface coating is of importance for when water condenses on the glass surface and for how it is distributed across the surface. Even though the panes in our experiments were not heated from the back as in an installed window, the SnO2-coated pane got condensation much later than the other two panes. This discrepancy is expected to be enhanced in a more realistic test. In our experiments, it was shown that although the condensation during a clear night appears relatively early on the TiO2-coated pane, it is smeared out so that in the morning the view through an uncoated or SnO2-coated pane is more obstructed. More experiments are underway using a test box with an optical monitoring system. These experiments will be possible to compare with simulation results, since weather data such as incoming long wave radiation, air temperature, relative humidity and wind speed will be collected on the site. Acknowledgements This work was sponsored by the Swedish Research Council for Environment, Agricultural Sciences and Spatial Planning (Formas). The test samples were provided through Dr. Kevin Sanderson by Pilkington European Technical Centre, UK. Dr. Torbjo¨rn Nilsson is acknowledged for coming with useful comments on the text. Anna Maria Lundin is acknowledged posthumously for covering travel expenses so that the results could be presented at Eurosun 2006 in Glasgow [25]. References [1] L.E. Nevander, B. Elmarsson, Fukthandbok, AB Svensk Byggtja¨nst, Stockholm, 1994 (in Swedish). [2] B. Jonsson, Heat transfer through windows, Swedish Council for Building Research, Stockholm, 1985 D13:1985. [3] J.A. Duffie, W.A. Beckman, Solar Engineering of Thermal Processes, Wiley, New York, 1980. [4] C.-G. Granqvist, Materials science for solar energy conversion systems, in: A.A.M. Sayigh (Ed.), Renewable Energy Series, Pergamon Press, Oxford, 1991. [5] C.-G. Ribbing, Radiative control of outdoor condensation. Kungl. Vetenskapssamha¨llets i Uppsala a˚rsbok 30. 1993–1994, Uppsala: Almqvist & Wiksell. [6] M.G. Meir, J.B. Rekstad, O.M. Løvvik, Sol. Energy 73 (6) (2002) 403.
ARTICLE IN PRESS A. Werner, A. Roos / Solar Energy Materials & Solar Cells 91 (2007) 609–615 [7] [8] [9] [10]
[11]
[12] [13] [14] [15] [16] [17] [18]
J. Khedari, et al., Renew. Energy 20 (2000) 183. E. Erell, Y. Etzion, Build. Environ. 35 (2000) 297. B. Givoni, Energy Build. 1 (1977) 141. T. Mouhib, et al., The radiative cooling of stainless steel/tin/ glass multilayer systems, in: Proceedings of Congre`s Ge´ne´ral, Lille, 2005. H.K. Pulker, Coatings on glass, in: G. Siddall (Ed.), Thin Films Science and Technology, vol. 6, Elsevier Science Publishers B.V., Amsterdam, 1999, p. 548. I. Hamberg, et al., Appl. Opt. 26 (11) (1987) 2131. A. Brovall, Elitfo¨nster, 2006, personal communication. A.J. Klemm, et al., Build. Environ. 37 (2002) 1215. J.M. Schultz, F.H. Kristiansen, Energimaessig helhedsvuderings av vinduer, 2002, Rapport BYG DTU R-035 (in Danish). B. Jonsson, Utva¨ndig kondens pa˚ fo¨nster, 1995: Sveriges Provningsoch Forskningsinstitut, SP Rapport 1995: 01 (in Swedish). M. Thyholt, S. Geving, ventilasjon og sanitaerteknikk (2000) 28. R. El Diasty, I. Budaiwi, Constr. Build. Mater. 3 (2) (1989).
615
[19] J. Karlsson, Optical Performance and Energy Efficiency, Uppsala University, Uppsala, 2001. [20] M.-L. Persson, Windows of Opportunities—The Glazed Area and its Impact on the Energy Balance of Buildings, Uppsala University, Uppsala, 2006. [21] M. Martin, Radiative cooling, in: J. Cook (Ed.), Passive Cooling, MIT Press, Cambridge, MA, 1989, p. 593. [22] P. Nilsson, Condensation Detection on Energy Efficient Windows by Light Scattering Measurement, Uppsala University, Uppsala, 2006 UPTEC F06 013. [23] T.M.J. Nilsson, Optical Scattering Properties of Pigmented Foils for Radiative Cooling and Water Condensation: Theory and Experiment, Chalmers Tekniska Ho¨gskola, Go¨teborg, 1994. [24] A. Werner, A. Roos, Influence of climate and window parameters on the occurrence of external water condensation on Swedish windows. Energy Build., submitted for publicaton. [25] A. Werner, Condensation tests on glass samples for energy efficient windows, in: Proceedings of Eurosun, Glasgow, 2006.
Solar Energy Materials & Solar Cells 91 (2007) 609–615 www.elsevier.com/locate/solmat
Condensation tests on glass samples for energy efficient windows Anna Werner, Arne Roos Department of Engineering Science, Solid State Physics, The A˚ngstro¨m Laboratory, Box 534, SE-751 21 Uppsala, Sweden Received 6 September 2006; accepted 14 November 2006 Available online 22 January 2007
Abstract This article presents results from a pilot study comparing condensation patterns on small glass samples with different surface properties. Experiments were carried out on three commercial glass samples (clear float, TiO2-coated and SnO2-coated) to see how water condensed on the different surfaces. The experiments were carried out under a clear night sky in Uppsala, Sweden. It was found that the pane with the low-emittance coating stayed clean of condensation longer than the other two. In the morning, the water layer on the TiO2coated sample was smeared out so that it was possible to see through that pane, while the view through the other two was still blurred. The TiO2 coating does not prevent condensation, but makes it easier to see through the water layer. These simple tests indicate noticeable differences between different surface materials and also that these effects can be studied by exposing small samples to a clear night sky without having to perform full scale tests. r 2007 Elsevier B.V. All rights reserved. Keywords: External water condensation; Energy efficient windows
1. Introduction Most of us have experienced water condensation on windows, either between the window panes or on the inside [1]. In both cases, the condensation indicates that the window is of poor quality or has a high U-value. In the former case, humid air leaks into a space that is supposed to be closed or properly ventilated and in the latter case water condenses on the inner window surface because the window is considerably colder than the room [2]. That means that heat leaks out from the room through the window. The phenomenon of interest in this article, external condensation on windows, is experienced when the temperature of the external glass pane of a window goes below the outside dew point due to radiative cooling [3–5]. This is a relatively new phenomenon and appears during clear nights on wellinsulated windows for which the thermal losses do not balance the radiative cooling of the external glass surface. Such a decrease in pane temperature is desirable when a radiative cooling system is designed and has, therefore, been Corresponding author. Tel.: +46 18 471 77 83; fax: +46 18 50 01 31.
E-mail address: [email protected] (A. Werner). URL: http://www.angstrom.uu.se. 0927-0248/$ - see front matter r 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.solmat.2006.11.015
studied for that application [6–10], but not in relation to well-insulated windows. Condensation happens during clear nights when radiation from the window towards the sky is large. The scope of condensation depends on the environment, the building and the climate (weather). The characteristics of the outer surface also affects how fast condensation appears and how it is distributed across the surface [11,12]. Often small droplets are formed and they cause light scattering which obstructs the view through the window. A general resistance against using well-insulated windows because of this phenomenon can be noticed [13]. In order to avoid, or at least reduce, the problem with external water condensation on windows, thin film glass coating technology can be utilized. Especially two concepts are of interest here, low-emissivity coatings and hydrophilic coatings. A low-emissivity coating will decrease the occurrence of condensation, whereas a hydrophilic surface will help smearing out the water on the surface so that the view through the window is less obstructed. Many factors, both surface properties and environmental conditions, affect the process of water vapour condensation on a solid surface. There are broadly speaking two types of condensation: dropwise condensation and film
ARTICLE IN PRESS 610
A. Werner, A. Roos / Solar Energy Materials & Solar Cells 91 (2007) 609–615
condensation. In the former case, the condensate is created in the form of individual drops and, in the latter case, the condensate forms a continuous layer on the surface [14]. Mixed types of condensation also exist. Our work presented here is a first attempt to evaluate how different coatings on the outer surface affect the condensation process. The aim is to evaluate different ways of avoiding the problem of obstructed view through a window due to external water condensation and to develop a technique to monitor the occurrence of condensation. Since the perceived problem of external condensation on well-insulated windows in cold climates is relatively new, only few observations and simulations have been carried out [15–17], but no systematic study of the actual process. Experiments have not been undertaken, since the night sky is difficult to simulate in a laboratory. A more investigated problem is external condensation on poorly insulated windows in hot climates in air-conditioned buildings [18]. Then, dew on the window is a sign of poor thermal insulation, whereas in our case, dew on the window is a sign of excellent thermal quality. Wellinsulated windows constitute a way to save energy [19,20], and it is therefore, important to find a way to make the view through these windows less obstructed by possible condensation. We decided to study three glazing products (see Table 1); uncoated clear float glass (surface emissivity 0.84) as a reference since most windows of today do not have any external coating, a low-emissivity fluorine-doped tin dioxide hard coating (surface emissivity 0.15) since it is often used on internal surfaces and, therefore, commercially available, and a titanium dioxide coated self-cleaning glazing product (surface emissivity 0.84) since it has the desired hydrophilic properties and is expected to gain a larger market share. Both coated products are well known for their durability and can, therefore, be used with the coating on the external surface. Most other low-emissivity products can only be used in a sealed insulated glass unit and are, therefore, not feasible for this possible future application. The solar optical properties, such as the reduction of the solar factor (g), must be considered when external coatings are applied on windows. This was, however, beyond the scope of the present study. The object of this work was primarily to perform a feasibility study to see to which extent simple tests with only visible inspection of small
samples can provide information about the formation of condensation on different glass coatings. 2. Experimental setup Our experiments, accounted for here, were not carried out on real, installed windows, but on small test samples. The sample size was not regarded as being critical since edge effects were expected to be small also for small samples because of the low heat conductivity of glass. For the first experiment, we used smaller, 5 cm 10 cm, samples simply to test the feasibility of the methodology. In the two following experiments, larger, 30 cm 30 cm, samples were used to get a better visual impression of the condensation. The glass samples were left facing the night sky as described in Sections 2.1–2.3. In two of our experiments, the test panes were placed horizontally and in a third experiment they were put in an almost upright position to better simulate a real window situation. Since the panes were tilted instead of vertical, condensation would occur more often than on a real window. For a vertical window, the surface normal points in the direction of the horizon and the sky only covers the upper half of the hemisphere seen by the glass surface. For a tilted window, the sky covers more than half of the hemisphere (the view factor is larger than 0.5) and hence the radiative cooling from the sky is higher [21]. The inclination of the glass pane was not expected to affect the formation of condensation in any other way than changing the time of the onset of condensation, i.e., it was not expected to affect the structure of the condensation. When much condensation has formed the inclination will affect how easily the drops run down the pane, but in our experiments that latter part of the condensation process was not primarily studied. All tests were performed during clear summer nights, with almost no wind, since condensation is more likely to occur under these conditions [16]. We did not measure the relative humidity or calculate the dew point, since we were primarily interested in monitoring differences between the samples. The pane temperature was measured with thermocouples. For upcoming experiments, an IR-camera could give additional information. These initial experiments were not undertaken to verify any models when condensation occurs. We used visual inspection, but for future experiments, we have developed a testing equipment that can
Table 1 Our samples Called here
Description
Clear float glass Low-emissivity coated glass
Traditional commercial clear float glass with no coatings Commercial pane with a low-emissivity coating with a surface resistivity of about 20 OX&. This pane is today used in well-insulated windows with the coating either on the outer side of the innermost pane or on the inner side of the external pane, i.e., the coated side is not turned outside as in our experiments and measurements Commercial pane with a titanium dioxide, ‘‘titania’’, coating with self-cleaning properties. This pane is relatively new. It has been sold for only a couple of years and stands for a very small part of the world float glass pane market
Self-cleaning glass
The three glass samples tested in our experiments.
ARTICLE IN PRESS A. Werner, A. Roos / Solar Energy Materials & Solar Cells 91 (2007) 609–615
611
more accurately detect when the water film forms [22]. We did not collect the dew from the panes, but this could give an idea of the quantity of water condensed on the individual samples [23]. In all experiments described here, the samples were insulated from the back with 30 mm of polystyrene. This way there was almost no heat transport through the glass samples from behind, as is the case for the outer pane in a well-insulated window, and the equilibrium temperature was mainly determined by the surface of the sample and by the sky temperature. Since the air speed was low, convection and conduction were low and the surrounding air could not fully balance the radiative cooling of the surface. Therefore, it was easier for the glass to reach a temperature lower than the surrounding air and for condensation to occur. 2.1. High and low emissivity The aim of the first experiment was to see if there is a simple way to monitor surface temperature and to visually discern external condensation on glass samples exposed to a clear night sky. The aim was also to see if there is a simple way to compare different coated glass samples to see which one is more prone to condensation. The experiment was undertaken during night when the outdoor air temperature was about 13 1C. Two glass samples, an uncoated clear float sample and a tin dioxide-coated clear float sample, of size 5 cm 10 cm were heated in an oven to about 50 1C and placed horizontally on a piece of polystyrene (Fig. 1). A thermocouple below each sample measured the temperature. Both samples had a thickness of 4 mm. The heating to 50 1C was done to monitor the difference in cooling rate more clearly. The experiment started at midnight. 2.2. Horizontal test of three samples The aim of this second experiment was to compare condensation proneness and properties of three commer-
Fig. 2. Experimental set up at the beginning of the second experiment, about 8 o’clock in the evening, 45 min before sunset.
cial glass samples, two of which were the same as in the first experiment. Another aim was to study how the temperature of a clear float sample and a tin dioxide-coated glass sample varies with the air temperature and relative humidity. Three small (30 cm 30 cm) glass samples (clear float, tin dioxide-coated and titanium dioxide-coated) of equal thickness were placed horizontally on a piece of polystyrene (Fig. 2), i.e., insulated from below as in the previous experiment and exposed to the clear night sky. A thermocouple below the clear float sample and the tin dioxide coated glass sample measured their temperatures. The TiO2 surface has an emissivity similar to the one of uncoated glass and it can, therefore, be assumed that it had the same temperature as the uncoated sample. The samples were not pre-heated as in the first experiment described above. A lamp was installed above the samples to provide a reflected image for the photographs. It was not lit, except for when photographs were taken. There was almost no wind. 2.3. Vertical test of three samples The aim of the third experiment was to compare condensation proneness and properties of the same three commercial glass samples in an almost vertical position. Three small (30 cm 30 cm) glass samples (clear float, tin dioxide-coated and titanium dioxide-coated glass) of equal thickness were placed outdoors, almost vertically, with polystyrene as backing material. They were exposed to the clear night sky and almost no wind. Their temperatures were not measured. 3. Results
Fig. 1. The two glass samples in the morning, about 7 h after experiment begun (t ¼ 420 min). Condensation was now present on both samples. The water layer had grown differently on the two different surfaces. The clear float glass sample is to the left and the tin dioxide-coated sample to the right.
3.1. High and low emissivity In the first experiment, the clear float glass condensation occurred after about 40 min of exposure to the clear night
ARTICLE IN PRESS A. Werner, A. Roos / Solar Energy Materials & Solar Cells 91 (2007) 609–615
612
sky. After 90 min, we observed no condensation on the tin dioxide-coated glass. After 3 h also to the tin dioxidecoated glass condensation was observed, but considerably less on the clear glass. Fig. 1 shows the set up in the morning, about 7 h after the experiment begun. Condensation was then present on both surfaces, but the water had been distributed differently across the two surfaces. The tin dioxide-coated sample both had a lower cooling rate and maintained a higher temperature than the clear float sample throughout the experiment (Fig. 3) due to reduced radiant losses. As can be seen in Fig. 3, the onset of condensation in both cases took place at a surface temperature of approximately 12 1C. The experiment showed that with a very simple set up, it is possible to measure and monitor the temperatures of the glass samples, the air temperature and the occurrence of condensation on small size samples. The experiment also showed that the thermal mass of these samples is sufficiently high; it is possible to monitor differences between the samples, both surface temperature and time when condensation is present. From the pictures taken during the experiments, it is also possible to see that the condensation has different characteristics on the different surfaces. 3.2. Horizontal test of three samples In the second experiment (Fig. 2), 1 h after sunset, the clear float and the titanium dioxide-coated samples already Measurement Data from First Experiment 50
Temperature (°C)
40
T (K-glass) T (Clear glass) T (K-glass) - T(Clear glass) T (Air)
30
20
had some condensation, whereas the tin dioxide-coated glass sample had none. On the clear float sample, drops of water had formed. A paper with printed text was placed just behind the surface. This was still sharp and easy to read. After another half an hour, there was condensation on two of the samples, but on the tin dioxide-coated glass sample there was no water condensation. On titanium dioxide coated sample condensation was more than on clear float sample. The surface emissivity of the TiO2 surface is approximately the same as that of uncoated glass and the temperatures must be the same due to the radiation balance. The fact that more condensation occurs on the TiO2-coated surface could be due to surface contamination and minor differences in nucleation sites. The difference was not dramatic and this issue was not further investigated. After two more hours, the condensation layers differed much on these three surfaces. The titanium dioxide-coated sample had a lot of water on it and the tin dioxide-coated glass sample had a milky condensation layer with small, fine drops, looking more like haze. The surface was clearly wet. The uncoated sample had a condensation layer, which looked like something in between the layer on the titanium dioxide-coated sample and the layer on the tin dioxidecoated glass sample. Figs. 4–6 show what the samples looked like in the morning, about 11 h after the experiment started. The clear float looked almost like in the previous picture, the tin dioxide-coated glass was very wet and now showed big drops unevenly distributed over the surface, maybe due to some surface contamination. On the titanium dioxide-coated glass, the condensed water had formed a film rather than drops. From Figs. 4–6 it can be seen how the water drops distorted the reflection of the lamp shown in Fig. 2. The air temperature (upper curve in Fig. 7) decreased slowly during the experiment. The clear float sample started at a lower temperature, which decreased somewhat slower than the air temperature. The tin dioxide-coated glass sample started at the air temperature, but its temperature decreased faster than the air temperature. Condensation seems to have occurred at approximately the
10
0 0
50
100
150
200
Time (minutes) Fig. 3. Temperature versus time for the two tested surfaces of the first experiment: The temperature of the tin dioxide-coated glass sample (solid curve with large squares) was above the temperature of the clear glass sample (solid curve with small triangles) throughout the first experiment. The solid curve without markers shows the air temperature. The lowest curve (dashed with small squares) shows the difference between the two surface temperatures. The horizontal and two vertical dashed lines mark the approximate onset of condensation on the two samples.
Fig. 4. Clear float sample about 11 h after beginning of experiment.
ARTICLE IN PRESS A. Werner, A. Roos / Solar Energy Materials & Solar Cells 91 (2007) 609–615
613
same temperature, about 12 1C, on both the clear float and the tin dioxide-coated glass samples. This was expected since the dew point temperature varies only slightly during most nights and depends on the relative humidity which increased slowly during the experiment, as can be seen in Fig. 7. 3.3. Vertical test of three samples
Fig. 5. Tin dioxide-coated glass sample about 11 h after beginning of experiment.
The pictures (Figs. 8–10) were taken in the morning after the third experiment, when the three samples had been out all night. The samples were put one by one in a frame for the photographs to be taken. The visual appearance of the samples was clearly different. In the morning, it was not possible to see through the clear float sample or the tin dioxide-coated sample, but it was possible to see through the titanium dioxide-coated glass sample. The TiO2-coating had not prevented condensation, but made it easier to see through the water layer. In Figs. 8 and 9, it can be seen that drops of water have been formed and are sliding down the pane. The trails of these drops might be visible even after the condensed water has evaporated if the pane is dirty. This unwanted effect seems to be avoided with a titania coating (Fig. 10).
Fig. 6. Titanium dioxide-coated glass sample about 11 h after beginning of experiment.
Measurement data from Second Experiment 100
20 Relative Humidity
75
T (Air) T (Low-e Glass) T(Clear Glass)
16
50
14
RH ( %)
Temperature (°C)
18
Fig. 8. First sample in experiment, clear float glass.
12 25 10 0
8 0
1
2
3
4
5
Time (hours) Fig. 7. Temperature of clear float and tin dioxide-coated glass samples and relative humidity during a clear night with very slight wind. The relative humidity is measured on the right y-axis. The horizontal line shows the approximate glass temperature at which condensation was observed on the two surfaces. The two vertical lines point at the measurements when condensation was first detected for each glass sample, respectively. The time 0 in the graph indicates the start of the experiment, i.e., 7.45 p.m. August 19, 2005. The sunset was at around 8.30 p.m.
Fig. 9. Second sample in experiment, tin dioxide-coated glass.
ARTICLE IN PRESS 614
A. Werner, A. Roos / Solar Energy Materials & Solar Cells 91 (2007) 609–615
is formed, the emissivity increases slightly and also this surface becomes covered by water. Under more realistic conditions dew is less likely to form on the low emissivity coated glass. For a real window, the vertical position is the most common. The matter of inclination was not studied here. The tilted angles were only used to stimulate the formation of condensation. 5. Conclusion and future outlook
Fig. 10. Third sample, titanium dioxide-coated glass pane is staged like the other two panes. In the upper part of the picture, the sky is reflected in the pane.
4. Discussion For an in-depth analysis of the circumstances under which condensation is formed more detailed climatic data, such as wind speed, relative humidity and long wave incoming sky radiation, are needed [24]. The present pilot study, however, is limited to comparing condensation proneness and characteristics of different surface coatings under identical conditions and to evaluating if the differences can be detected by visual inspection. The results indicate that it is possible to study the formation of condensation on single glazed samples under a clear night sky. There are clear differences between different coatings both regarding formation of condensation and subsequent evaporation. One of the issues with external condensation mentioned by customers is the pattern appearing when the amount of condensation is sufficient for water drops to run down the usually slightly dirty glass surface, thus causing visible streaks to appear even after the condensation has evaporated. It would be of interest to evaluate the selfcleaning type of surface in this respect, but this would require more tests with larger samples. The condensation always disappears in the morning after sunrise. Even the diffuse radiation from a cloudy sky supplies sufficient energy to cause the water layer to evaporate. The rate of evaporation may differ for different surfaces depending on the wetting angle of the water but this was not further investigated. The external coating on the window inevitably leads to a reduced value of daylight and the solar energy transmitted. Thus the energy balance of the window is affected negatively and some of the energy benefit of using a wellinsulated window is lost. The overall evaluation of the benefit of these coatings must involve also the negative effects and the cost. This investigation was only focussed on the feasibility of using these coatings to prevent external condensation. For the low-emissivity coating the formation of condensation was considerably delayed, but once condensation
It was shown that experiments on small pane samples can show that the surface coating is of importance for when water condenses on the glass surface and for how it is distributed across the surface. Even though the panes in our experiments were not heated from the back as in an installed window, the SnO2-coated pane got condensation much later than the other two panes. This discrepancy is expected to be enhanced in a more realistic test. In our experiments, it was shown that although the condensation during a clear night appears relatively early on the TiO2-coated pane, it is smeared out so that in the morning the view through an uncoated or SnO2-coated pane is more obstructed. More experiments are underway using a test box with an optical monitoring system. These experiments will be possible to compare with simulation results, since weather data such as incoming long wave radiation, air temperature, relative humidity and wind speed will be collected on the site. Acknowledgements This work was sponsored by the Swedish Research Council for Environment, Agricultural Sciences and Spatial Planning (Formas). The test samples were provided through Dr. Kevin Sanderson by Pilkington European Technical Centre, UK. Dr. Torbjo¨rn Nilsson is acknowledged for coming with useful comments on the text. Anna Maria Lundin is acknowledged posthumously for covering travel expenses so that the results could be presented at Eurosun 2006 in Glasgow [25]. References [1] L.E. Nevander, B. Elmarsson, Fukthandbok, AB Svensk Byggtja¨nst, Stockholm, 1994 (in Swedish). [2] B. Jonsson, Heat transfer through windows, Swedish Council for Building Research, Stockholm, 1985 D13:1985. [3] J.A. Duffie, W.A. Beckman, Solar Engineering of Thermal Processes, Wiley, New York, 1980. [4] C.-G. Granqvist, Materials science for solar energy conversion systems, in: A.A.M. Sayigh (Ed.), Renewable Energy Series, Pergamon Press, Oxford, 1991. [5] C.-G. Ribbing, Radiative control of outdoor condensation. Kungl. Vetenskapssamha¨llets i Uppsala a˚rsbok 30. 1993–1994, Uppsala: Almqvist & Wiksell. [6] M.G. Meir, J.B. Rekstad, O.M. Løvvik, Sol. Energy 73 (6) (2002) 403.
ARTICLE IN PRESS A. Werner, A. Roos / Solar Energy Materials & Solar Cells 91 (2007) 609–615 [7] [8] [9] [10]
[11]
[12] [13] [14] [15] [16] [17] [18]
J. Khedari, et al., Renew. Energy 20 (2000) 183. E. Erell, Y. Etzion, Build. Environ. 35 (2000) 297. B. Givoni, Energy Build. 1 (1977) 141. T. Mouhib, et al., The radiative cooling of stainless steel/tin/ glass multilayer systems, in: Proceedings of Congre`s Ge´ne´ral, Lille, 2005. H.K. Pulker, Coatings on glass, in: G. Siddall (Ed.), Thin Films Science and Technology, vol. 6, Elsevier Science Publishers B.V., Amsterdam, 1999, p. 548. I. Hamberg, et al., Appl. Opt. 26 (11) (1987) 2131. A. Brovall, Elitfo¨nster, 2006, personal communication. A.J. Klemm, et al., Build. Environ. 37 (2002) 1215. J.M. Schultz, F.H. Kristiansen, Energimaessig helhedsvuderings av vinduer, 2002, Rapport BYG DTU R-035 (in Danish). B. Jonsson, Utva¨ndig kondens pa˚ fo¨nster, 1995: Sveriges Provningsoch Forskningsinstitut, SP Rapport 1995: 01 (in Swedish). M. Thyholt, S. Geving, ventilasjon og sanitaerteknikk (2000) 28. R. El Diasty, I. Budaiwi, Constr. Build. Mater. 3 (2) (1989).
615
[19] J. Karlsson, Optical Performance and Energy Efficiency, Uppsala University, Uppsala, 2001. [20] M.-L. Persson, Windows of Opportunities—The Glazed Area and its Impact on the Energy Balance of Buildings, Uppsala University, Uppsala, 2006. [21] M. Martin, Radiative cooling, in: J. Cook (Ed.), Passive Cooling, MIT Press, Cambridge, MA, 1989, p. 593. [22] P. Nilsson, Condensation Detection on Energy Efficient Windows by Light Scattering Measurement, Uppsala University, Uppsala, 2006 UPTEC F06 013. [23] T.M.J. Nilsson, Optical Scattering Properties of Pigmented Foils for Radiative Cooling and Water Condensation: Theory and Experiment, Chalmers Tekniska Ho¨gskola, Go¨teborg, 1994. [24] A. Werner, A. Roos, Influence of climate and window parameters on the occurrence of external water condensation on Swedish windows. Energy Build., submitted for publicaton. [25] A. Werner, Condensation tests on glass samples for energy efficient windows, in: Proceedings of Eurosun, Glasgow, 2006.