Vacuum 104 (2014) 70e76
Contents lists available at ScienceDirect
Vacuum journal homepage: www.elsevier.com/locate/vacuum
Gas permeation characteristics through heat-sealed flanges of vacuum insulation panels Inseok Yeo, Haeyong Jung, Tae-Ho Song* School of Mechanical, Aerospace and System Engineering, Korea Advanced Institute of Science and Technology, Guseong-dong 373-1, Yuseong-gu, Daejeon, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history: Received 11 August 2013 Received in revised form 14 January 2014 Accepted 17 January 2014
Application of vacuum insulation panel (VIP) is becoming more practical to reduce the thermal energy loss and to reduce CO2 emission. Outstanding insulation performance and extended service life are the two critical factors in the VIP engineering. Thermal insulation performance of a VIP is degraded with time due to gradual increase of pressure in the VIP. Keeping the effectively low pressure is a key technique to improve the service life. Permeation of gases through heat-sealed flanges in the lateral direction has the greatest effect in it, together with permeation of gases through normal direction to the envelope sheets. This study is made to investigate the permeation characteristics of the envelope seal materials (typically linear low density polyethylene, LLDPE, and low density polyethylene, LDPE). A measurement apparatus using multiple radial permeation passages is designed to measure the gas permeation rate through the polymer films. The measurement shows reliable accuracy compared with other reported results. Inner pressure change in the VIP, which is enveloped with Al-foil-based film, is calculated based on the experimental results of gas permeabilties of seal materials. It is found that the permeation characteristics of heated LLDPE is the same as heated LDPE and unheated LLDPE. H2O, O2, and N2 show greatest permeations in this order. Unlike fumed-silica VIPs, glass wool VIPs may exhibit short service life due to this permeation. Ó 2014 Elsevier Ltd. All rights reserved.
Keywords: Vacuum insulation panel Gas permeability Service life
1. Introduction Recently, energy crisis and environmental issues have come to the fore as the major problems that can threaten human life. While researches for renewable energy as the solution are extensively conducted, ratio of the renewable energies to the total energy consumption is only about 8% [1]. This is far from sufficient considering the explodingly increasing energy demand and limited resources of economical energy. Alternatively, improving the energy efficiency is more practical, considering that as much as 40% of the total energy is spent in buildings to compensate the heat loss [2]. More specifically, enhancement of thermal insulation for the buildings, refrigerators, low temperature storage tanks, and so on is the thought to be a major solution, which can be made possible through the use of VIPs. Conventional insulation materials such as EPS (expanded polystyrene), PU (polyurethane) and glass wool contain gases inside the pore so that their thermal conductivities impose lower limits of thermal conductivity (kair z 0.026 W/m K).
* Corresponding author. Tel.: þ82 42 350 3032; fax: þ82 42 350 3210. E-mail address:
[email protected] (T.-H. Song). 0042-207X/$ e see front matter Ó 2014 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.vacuum.2014.01.008
On the other hand, thermal conductivity of a VIP is about 1/5 to 1/10 of kair; energy consumption can be virtually reduced to zero even with a thinner insulation [3]. Nevertheless, VIP has a few technical problems, one of which being that the thermal performance is degraded with time due to permeation of air from outside. The service life of the VIP must be guaranteed for 10e15 years for home appliances or more than 50 years for buildings, which is currently a formidable goal. Accordingly, this research is made to investigate the permeation of gases in the VIP. Here, it is appropriate to introduce the mechanism of residual gas conduction in VIPs. Thermal transport in a VIP is made by three modes [4,5]: conduction via core material and the envelope, residual gas conduction and radiation. Among these heat transfer modes, the service life of the VIP is related with the gas conduction, while it is widely conceived that the others do not change with time. (Lately, however, it is reported by S. Brunner et al. that solid conduction can be increased in a few years due to growth of the interfacial area between the nanosized SiO2 in case of pyrogenic silica core [6]). Gas conduction in rarefied state is a function of pore size and inner pressure. The resulting gas conductivity kg is derived by Kwon et al. as follows [7].
I. Yeo et al. / Vacuum 104 (2014) 70e76
Ru S t T V
Nomenclature outside radius of the annulus film, m surface area, m2 inside radius of the annulus film, m gas concentration, mol/m3 conversion factor from mole to volume diffusion coefficient, m2/s gas permeation flux, mol/m2 s thermal conductivity, W/m K gas permeability m2/s Pa total thickness of the annulus films, m pressure, Pa pressure at high pressure side, Pa pressure at high pressure side, Pa gas permeation rate, m3/s
a As b C C0 D J k K L P PH PL Q
kg ¼
kg0 1 þ 0:032 Pf
;
71
universal gas constant, J/mol K solubility, 1/Pa time, s temperature, K volume, m3
Greek symbols d permeation length, m 4 pore size, m Subscripts atm atmospheric cr critical g gaseous i gas species
(1)
where kg0 is the thermal conductivity of the gas in continuum state, P is the inner pressure in Pa (N/m2) and 4 is the pore size (m) of the core material. The relation between inner pressure, pore size and effective thermal conductivity is demonstrated in Ref. [8]. The critical pressure Pcr here is be defined as the pressure at which the effective thermal conductivity reaches half of kg0. It is smaller when the pore size is larger, which again means a shorter service life. In practice, fumed silica powder and glass wool are mostly used as the core material. Both of them have advantages and disadvantages. While the initial thermal performance at the center of a VIP with glass wool is 2.0e3.5 mW/m K, which is better than that with fumed silica powder of 3.5e5.0 mW/m K, the service life of the VIP
with glass wool is significantly shorter than that with the other one. This is because of their different pore sizes and thus, selection of the VIP should be made considering these two conflicting characteristics. Note that thermal conductivity at the center of the VIP which is mentioned earlier is lower than overall thermal conductivity because edge conduction is excluded. (This edge conduction cannot be neglected considering that metal layer of envelope is indispensable.). Consequently, more fundamental studies are necessary regarding the increase of the pressure and the change of thermal conductivity in the VIP. Pressure increase factors are [9]: a) outgassing from inner surface of envelope and the core material and, b) gas permeation through the envelope. In case of outgassing, dissolved gas in the material can be effectively eliminated by baking it in vacuum at high temperature (above 80 C). A more practical process is to add a
Fig. 1. Schematic diagram of a VIP.
72
I. Yeo et al. / Vacuum 104 (2014) 70e76
preheating before putting in vacuum furnace for a shorter time. Then the gas permeation problem only remains to be solved. This can be classified into the normal direction gas permeation and the lateral one depending on the path as shown in Fig. 1(a). The VIP envelope is mostly composed of polyethylene terephthalate (PET) as the protective layer, Al-foil or Al-metallized layer as the gas barrier, and LLDPE and LDPE as the heat-seal layers as shown in Fig. 1(b). Gas permeation across the envelope surface in the normal direction can be effectively blocked by the Al-layer. It is known that there is no appreciable pin hole if Al-foil of thickness greater than about 10 mm is used [10]. However, since the polymer side cut of the heat-sealed flange is exposed to atmospheric pressure, gas permeation through that in the lateral direction occurs without any countermeasure. Kwon et al. estimated various gas loads separately for the VIP with a polycarbonate staggered beam core and found that the gas permeation through heat-sealed flanges is mainly responsible to the pressure increase [11]. Therefore gas permeation characteristics through this passage should be investigated. The objective of this paper is thus to measure the lateral gas permeability for O2, CO2, N2 and H2O with a new measurement apparatus. Heat-sealed LLDPE and heated LDPE will be taken as the specimen and the permeability data will be used to estimate the service life. 2. Gas permeation theory and experiment 2.1. Theoretical background Gas permeation through polymer films is well described by three consecutive steps called the solution-diffusion model [12,13]. This process is based on colloidal diffusion postulated by Graham [14] and it involves sorption of the gas at the high-pressure interface, molecular diffusion of the gas through the film and finally desorption of the gas at the low-pressure interface (Fig. 2). This process contains three coefficients K, D and S [15]; K is the permeability, D is the diffusion coefficient and S is the solubility of the gas. In a steady-state, the gas permeation rate Q across a medium is expressed from Fick’s first law as follows [16,17]
Q ¼
DAs ðCH CL Þ
d
;
where As is the cross-sectional area of the medium, C is the gas concentration with subscripts H and L denoting “high” and “low” side, respectively, D is the diffusion coefficient and d is the permeation length. It is assumed that the gas does not dissociate, which is valid for diffusion through polymers. From Henry’s law, C of the gas at a surface of the medium is expressed as
C ¼ SP;
(3)
where S is the solubility and P is the partial pressure of the gas at that surface. Gas permeability K through a homogeneous polymer film is the product of D and S,
K ¼ DS:
(4)
Using K, Eq. (2) is re-written as follows:
Q ¼ K
As
d
ðPH PL Þ;
(5)
where PH and PL are the gas pressure at high pressure side and at low pressure side, respectively. The units of Q and K are taken as the volumetric gas flow rate at STP (298.15 K, 101.3 kPa). Once K is known, the pressure increase rate (dPL/dt) of the vacuum volume can be obtained mathematically, and vice versa. Suppose PL ¼ 0 initially and the volume is V. Then the molar inflow rate Q/C0 (C0 being molar volume at STP) is the molar increase rate inside the volume. Assuming ideal gas behavior,
Q V dPL ; ¼ C0 Ru T dt
(6)
where Ru and T are the universal gas constant and the absolute temperature, respectively. Thus,
dPL K As R u T ðPH PL Þ: ¼ C0 d V dt
(7)
When PH PL h DP z constant, K can be conversely expressed as
K ¼ (2)
C0 V d dP : Ru TAs DP dt
(8)
This equation is used to experimentally determine the value of K. 2.2. Measurement method of the lateral direction gas permeation
Fig. 2. Schematic diagram of the gas permeation process.
There are little reported data about permeability in the lateral direction. Many researchers investigate gas permeation for multifarious gases through various polymers using different methods. Typical examples are the volume variation method [18], the continuous flow method [19], the gravimetric method [20] and the pressure rise method [21]. The pressure rise method has been used in this paper. The experimental setup is made up of two chambers separated by the sample film as shown in Fig. 3. After both sides are evacuated by a vacuum pump, gas is introduced into one of two chambers. Thus, gas penetrates in normal direction to the film surface by the pressure difference. In the designed apparatus, sample films of LDPE or LLDPE of annulus shape together with copper gaskets are employed as shown in Fig. 4(a). The sample film combined with two copper gaskets is inserted between the vacuum flanges and then seven such units are perfectly fastened in parallel with fastening bolts. This apparatus is connected to a vacuum gauge on one side and a vacuum valve on the other as shown in Fig. 4(b). A CVM-201Ò gauge from InstruTech, Inc. is installed to measure the pressure increase
I. Yeo et al. / Vacuum 104 (2014) 70e76
73
for O2, CO2 and N2 and a CMR-361Ò gauge from Pfeiffer Vacuum is used for H2O. The former can measure from 102 Pa to 133 kPa as it is a convection-enhanced Pirani gauge and the latter is a capacitance diaphragm gauge that measures the pressure differences by the deflection of metal diaphragms, with a measurement range of 10 Pae110 kPa. The apparatus is completely sealed with copper gaskets and stainless steel vacuum flanges so that gas permeates only in the lateral direction. The experimental procedure is as follows: 1) The inside of apparatus is evacuated with a baking pre-process and then the vacuum valve is closed. 2) It is placed in a large environmental chamber. 3) The environmental chamber is evacuated. 4) Gas is introduced into the environmental chamber. 5) Pressure in the apparatus is monitored with time.
Fig. 3. Schematic diagram of a conventional gas-permeation measurement apparatus in the normal direction.
For H2O, PH is maintained as the vapor pressure at the test temperature. Fig. 4(c) shows this measurement system. Since the diffusion cell is a radial one with outer and inner radii of a and b and length L, the factor d/As in Eq. (8) is replaced by ln(a/ b)/2pL to yield [22],
K ¼
C0 V lnða=bÞ dP : Ru T DP 2pL dt
(9)
Since the films between copper gaskets are subject to compression force when tightened, the film thickness may diminish and it may have a decisive effect on the experiment of gas permeability. To confirm this uncertainty, the film thickness change is experimentally measured with the apparatus shown in Fig. 5. It is composed of a pair of vacuum flanges, a linear variable differential transformer (LVDT) and stainless steel gaskets. After repeated measurements, deformation of a 100 mm-thick LLDPE film is found to be 5.66 mm (with a standard deviation of about 0.87 mm) and it accounts for 5.7% of the film thickness. Thus, the effect of the film deformation by compressive force can be safely neglected. 2.3. Validation test The measurement apparatus is validated against available reference data. For the validation experiments, raw LLDPE films of
Fig. 4. Schematic diagram of the gas permeation measurement apparatus in the lateral direction.
Fig. 5. Experimental apparatus to measure variation of the film thickness.
74
I. Yeo et al. / Vacuum 104 (2014) 70e76
100 mm thickness are used as the sample. Outside and inside radii of the sample films are 25 mm and 18 mm, respectively, and the inner volume of the apparatus is 1.2 104 m3. For each of O2, CO2 and N2 experiments, the environmental chamber is filled with a pure gas of 80 kPa and the experiment is carried out in a constant temperature & humidity room at 23 C. Since there are no reference data of lateral direction permeation, the experimental results are compared with available data in the normal direction. Fig. 6 shows graphs of pressure increase with time for O2, CO2 and N2. Each graph is scaled differently to clearly identity the initial and the steady pressure-rise stages. Through linear fitting of the pressuretime graph after about 50 h (assuming a steady state at this time), the gas permeabilities are calculated as in Table 1. It is shown that experimental values lie within the reference values. The measurement apparatus is thus considered to be successfully validated. 2.4. Experimental specimen and condition As mentioned earlier, the polymer layers are multi-layered LLDPE and LDPE films. The heat-sealing process was performed at approximately 130 C [25]. To bind their characteristics separately, two experiments are made: 1) For the LDPE, a piece of LDPE film of 100 mm thickness is pre-heated and cut into the annulus sample film of Fig. 4(a). 2) For the LLDPE, three LLDPE films of total 300 mm thickness are heat-sealed together as the annulus sample film. (Note that LDPE sample film alone would not heat seal at the commonly employed conditions of heatsealing.)
Pressure difference of 80 kPa for O2, CO2, N2 is applied; for H2O, a pressure difference of 2.83 kPa (saturated pressure of water at 23 C) is exerted. 2.5. Results and discussion From Eq. (9), permeability K is obtained from the steady-state pressure increase rate (dP/dt) as shown in Table 2(a). The permeability of the gases increases in the following order; H2O > CO2 > O2 > N2. As mentioned earlier, gas permeability depends on solubility as well as on the diffusion coefficient. An interesting consequence is as follows. Diffusion coefficient for CO2 is smaller than that of O2 [26]. Nevertheless, the solubility of CO2 is much larger than that of O2. This eventually makes the permeability of CO2 greater than that of O2. These permeabilities are compared with non-heat-sealed LLDPE films as shown in Table 3. From the result, heat-seal effects are found to be negligible as the difference is less than 7%. Also, the gas permeability through the LDPE film is in Table 2(b). Gas permeability through the LLDPE film is found to similar to that of LDPE film. For LDPE, available data in the literature are fairly scattered, which makes it meaningless to directly compare with the current data. It is believed that the data of Table 2(b) is not very different from raw ones since LDPE films are fabricated through a thermal process already. This argument may be also applied to the foregoing discussions on LLDPE. To appreciate the applicability of current data, let’s think of the actual pressure increase in the VIP. Assuming that the VIP is placed in the atmosphere, the partial pressure Pi(t) with time t for gas species i in the VIP can be expressed by solving Eq. (7) as,
Fig. 6. Pressure increase rate of the raw LLDPE film for O2, CO2 and N2.
I. Yeo et al. / Vacuum 104 (2014) 70e76
75
Table 1 Gas permeability through the raw LLDPE films in the lateral direction for O2, CO2 and N2 (296 K, DP ¼ 80 kPa). dP/dt 105 [Pa/s]
O2 CO2 N2
Permeability 1017 [m2/s Pa] Experiment (Lateral direction)
Refs. [23,24] (Normal direction)
1.69 4.54 0.42
0.98e2.17 3.38e9.6 0.73
1.68 4.49 0.44
Table 2 Gas permeability through the heat-sealed LLDPE films and the heated LDPE films in the lateral direction for O2, CO2, N2 (296 K, DP ¼ 80 kPa), and H2O (296 K, DP ¼ 2.83 kPa). (a) heat-sealed LLDPE films; (b) heated LDPE films. (a)
O2 CO2 N2 H2O
dP/dt 105 [Pa/s]
Permeability 1017 [m2/s Pa]
4.28 11.5 1.19 3.30
1.57 4.22 0.44 34.2
dP/dt 105 [Pa/s]
Permeability 1017 [m2/s Pa]
1.69 5.61 0.53 1.27
1.70 5.66 0.53 36.2
Fig. 7. Expected pressure increase with time inside an Al-foil-based VIP of 0.3 m 0.3 m 0.01 m size (d ¼ 0.01 m).
(b)
O2 CO2 N2 H2O
As Ru T t ; Pi ðtÞ ¼ Pi;atm 1 exp Ki d C0 V
(10) 3. Conclusion
where Pi,atm is the atmospheric partial pressure of gas species i and Ki is the gas permeability. The total pressure in the VIP can be estimated by summation of the partial pressures for separate gas species. To make the problem simple, we consider only lateral direction permeation for a 0.3 m 0.3 m 0.01 m (V ¼ 0.0009 m3, d ¼ 0.01 m) VIP enveloped with Al-foil-based film. Note again that the gas permeation through the normal direction to the Al-foil is neglected. The film is multi-layered with 40 mm LDPE and 80 mm LLDPE. The partial pressure for separate gases and total pressure with time inside the VIP are calculated as in Fig. 7. This is calculated for 2% molar water vapor in the atmosphere. It is found that permeation of H2O is dominant. The water vapor permeation should be thus treated properly and fortunately, there are many H2O absorbing materials like silica and zeolite. For other gases, permeability of CO2 is relatively high but since the fraction of CO2 in the atmosphere is extremely low, its effect is insignificant. In contrast, N2 affects the pressure increase significantly due to its high composition. Using the earlier definition of Pcr, those of glass wool and fumed silica are calculated as 1000 Pa and 26,500 Pa, respectively, following the data of Ref. 8. For the VIPs of above size, the service life of glass wool VIP is 13 years in the atmosphere of 2% water vapor and that of fumed silica VIP is 620 years. Depending on the author(s), the critical pressure is taken as that of
Table 3 Comparison of permeability through heat-sealed and non-heat-sealed LLDPE films for O2, CO2 and N2.
Permeability 1017 [m2/s$Pa] Difference (%)
Heat-sealed films Non-heat-sealed films
k j 11 mW/m K. If this criterion is taken, Pcr is 250 Pa for glass wool and 26,500 Pa for fumed silica, yielding the service lives of 3 years for the former and 620 years for the latter under the same environment condition. Note that these are doubled if H2O is effectively eliminated. Yet, however, the service life of VIP with glass wool needs to be increased to be safely applied to buildings. To realize this, it is recommended to further study getters and outgassing characteristics.
O2
CO2
N2
1.57 1.69 7.1
4.22 4.54 7.0
0.44 0.42 4.8
Gas permeation through the heat-sealed flanges of VIP is investigated. Based on a theoretical solution, a measurement apparatus is devised to measure the gas permeation through using the direct method. Since gas permeability for the polymer film in the lateral direction is investigated for the first time, there are no reliable reference data to verify the results. Nonetheless, the results from the experiments are convincing when compared with existing data of normal direction gas permeability. Lateral permeation through heat-sealed LLDPE shows little difference with normal direction data (within 7%). And it behaves similarly to heated LDPE. Among H2O, N2, CO2 and O2, the gas permeability increases in the following order H2O > CO2 > O2 > N2. Under atmospheric environment, gas that largely affects pressure increase is found to be H2O and effects of N2 and O2 are fairly high. Roughly speaking, the lateral permeation of LLDPE layer significantly limits the life time of glass wool VIPs, while for fumed-silica VIPs, there is no such limitation. Acknowledgments This work was supported by the National Research Foundation of Korea (NRF) grant funded by Korea government (MSIP) (No.2013068924) and BK21 Plus Project. References [1] U.S. Energy information administration, annual energy review; 2010. [2] Department of Trade and Industry. Energy trends; 2003. pp. 4e23. London. [3] Jelle BP, Gustavsen A, Baetens R. The path to the high performance thermal building insulation materials and solutions of tomorrow. J Build Phys 2010;34(2):99e123.
76
I. Yeo et al. / Vacuum 104 (2014) 70e76
[4] Mukhopadhyaya P, Kumaran K, Normandin N, van Reenen D, Lackey J. High performance vacuum insulation panel: development of alternative core materials. J Cold Reg Eng 2008;22(4):103e23. [5] Fricke J, Schwab H, Heinemann U. Vacuum insulation panels-exciting thermal properties and most challenging applications. Int J Thermophys 2006;27: 1123e39. [6] Brunner S, Wakili KG. Hints for an additional aging factor regarding the thermal performance of vacuum insulation panels with pyrogenic silica core. Vacuum 2014;100:4e6. [7] Kwon JS, Jang CH, Jung H, Song TH. Effective thermal conductivity of various filling materials for vacuum insulation panels. Int J Heat Mass Transf 2009;52: 5525e32. [8] Fricke J, Heinemann U, Ebert HP. Vacuum insulation panel-from research to market. Vacuum 2008;82:680e90. [9] Kwon JS, Jung H, Yeo IS, Song TH. Outgassing characteristics of a polycarbonate core material for vacuum insulation panels. Vacuum 2011;85:839e46. [10] Murray L. The impact of foil pinholes and flex cracks on the moisture and oxygen barrier of flexible packaging. In: Polymers, laminations, adhesives, coatings, and extrusions conference 2005. Las Vegas, USA. [11] Kwon JS, Jang CH, Jung H, Song TH. Vacuum maintenance in vacuum insulation panels exemplified with a staggered beam VIP. Energy Build 2010;42: 590e7. [12] Tremblay P, Savard MM, Vermette J, Paquin R. Gas permeability, diffusivity and solubility of nitrogen, helium, methane, carbon dioxide and formaldehyde in dense polymeric membranes using a new on-line permeation apparatus. J Membr Sci 2006;282:245e56. [13] Flaconnèche B, Martin J, Klopffer MH. Transport properties of gases in polymers: experimental method. Oil Gas Sci Technol 2001;56:245e59. [14] Graham T. On the absorption and dialytic separation of gases by colloid septa. Philos Mag J Sci 1866;32:401e20.
[15] Stannett V. The transport of gases in synthetic polymeric membranes e an historic perspective. J Membr Sci 1978;3:97e115. [16] Comyn J. Polymer permeability. London and New York: Elsevier Applied Science Publishers; 1985. pp. 1e73. [17] Georage SC, Thomas S. Transport phenomena through polymeric systems. Prog Polym Sci 2001;26:985e1017. [18] Felder RM, Huvard GS. Methods of experimental physics. 16c. New York: Academic Press; 1980. pp. 315e77. [19] Gavarat R, Catala R, Hernandez-Munoz PM, Hernandez RJ. Evalution of permeability through permeation experiments: isostatic and quasiisostatic methods compared. Packag Technol Sci 1996;9:215e24. [20] Britton LN, Ashman RB, Aminabhavi TM, Cassidy PE. Prediction of transport properties of permeants through polymer films. J Chem Educ 1988;65:368e 70. [21] Jung H, Jang CH, Yeo IS, Song TH. Investigation of gas permeation through Almetallized film for vacuum insulation panels. Int J Heat Mass Transf 2013;56: 436e46. [22] Incropera FP, Dewitt DP. Fundamentals of heat and mass transfer. 5th ed. New York: Wiley; 2002. pp. 104e10. [23] Villaluenga JPG, Seoane B. Experimental estimation of gas-transport properties of linear low-density polyethylene membranes by an integral permeation method. J App Polym Sci 2001;82:3013e21. [24] Compañ V, Andrio A, López ML, Riande E. Effect of annealing on the permeation characteristics of gases of coextruded LLDPE films. Polymer 1996;37: 5831e7. [25] Yuan CS, Hassan A, Ghazali MIH, Ismail AF. Heat sealability of laminated films with LLDPE and LDPE as the sealant materials in bar sealing application. J App Polym Sci 2007;104:3736e45. [26] Crank J, Park GS. Diffusion in polymers. London and New York: Academic Press; 1968. pp. 1e39.