A recent study provides new clues about sorbed water in indoor surface films

Rowen and Blaine50 reported apparent surface areas on such a comparative basis. The specific surface areas for the two sorbates were comparable for titanium dioxide. However, for nylon and for wool specific surface areas were much larger for sorbing water as compared to N2. With dynamically changing indoor humidity conditions, an important consideration for sorptive partitioning of water is the accessibility of fabrics. As one example, the magnitude of clothing worn by a person is ~ 1 kg. Closets and clothing cabinets contain tens to hundreds of kg of clothing. While being worn, clothing fibers are readily accessible to moisture and to gaseousacids and bases. While stored, however, limited rates of mass transfer might extend the equilibration time scale such that gaseous interactions are not responsive to rapid dynamic changes. The influence of long sorptive time scales associated with abundant fibrous materials has not been studied even with regards to water vapor buffering.Another key consideration is the thermodynamic properties of the sorbed water. In particular, to what extent do volatile and semivolatile species partition into water that is sorbed to fibrous materials? Do the Henry’s law constants developed for partitioning to bulk water have meaning, even as estimates, for the expected partitioning between gaseous and sorbed water? Can sorbed water serve as a proton acceptor for the case of an acid or as a proton donor for the case of a base? Are the respective pKa values, e.g. for acids in bulk water, reasonably predictive of acid behavior in sorbed water? We don’t know the answers to these questions. Given the expected abundance of sorbed water indoors,ebb flow research is warranted to answer them. Water may be sorbed at any of the interfaces between solid materials and indoor air. In relation to the forms of condensed water already discussed, this surface-sorbed water has small abundance.

However, because the surface-sorbed water occurs in thin films, the equilibration time scales are rapid for species partitioning between the water film and air. We acknowledge an overlap in classifying water in aqueous surface films as compared with sorbed water in indoor materials. Compare water for an impervious surface such as window glass to a porous sorbent such as gypsum board. Water molecules are not expected to penetrate materially into window glass; instead, condensed water is present as part of an invisible surface film along with other deposited materials, such as inorganic ions and organic compounds. 56 For gypsum board, at equilibrium, sorbed water will be present throughout the bulk of the material . Mass-transfer limitations influence the degree to which water is present throughout the bulk of a porous material such as gypsum board, but would not strongly affect water’s abundance in surface films. The distinction in this comparison between surface-sorbed water and bulk-sorbed water should be clear. But what then about sorptive water partitioning to textiles indoors, such as carpet, furniture fabrics, and window coverings? In this review, we have categorized such water as bulk-sorbed, in relation to the bulk properties of the fabrics. For materials like painted gypsum board or wood furniture, contributions to total condensed-phase water abundance occur from both the bulk sorption throughout the material and from surface sorption in a film. Bulk sorption in such cases was discussed in the previous section. One should consider the possibility of additional water being present in surface films on such permeable and porous materials. To date, the relevant literature about surface films, itself limited in scope and extent, focuses on impervious surfaces.

Important lessons about the adsorption of water on glass can be found in an early investigation by Razouk and Salem. They measured the sorption of water on glass wool, glass powder, and glass microspheres. For water-washed glass, they found that the “real surface is about two to three times greater than the geometric surface.” They also found that “the amounts adsorbed at 0.20 and 0.80 relative pressures [i.e., at RH = 20% and 80%, respectively] correspond closely to one and two molecular layers .” Their experimental evidence was interpreted to support the view that, “water adsorbed by glass is made up of two parts: one part being readily removed by pumping and thus loosely held, the other part being held firmly and driven off only by heating at higher temperatures.” Water sorbed on a surface can be quantified in terms of monolayer equivalents. As suggested by the name, a monolayer corresponds to the quantity of water just sufficient to completely cover the surface at a thickness of a single molecule. A few clarifying points are needed. First, the idea of a monolayer is conceptually valuable, but not real. Sorption occurs in a patchy manner, with multiple layers of sorbed molecules occurring before a first layer is full. Second,isotherm data are often based on measurements of the mass of sorbed water per mass of sorbent. To convert to equivalent monolayers, basic information is needed about the geometry of sorbed water and about the specific surface area of sorbents. For sorbed water a nominal linear dimension can be obtained as the cube root of the effective volume occupied by a single water molecule. From this perspective, water at 18 g/mol and 1.0 g cm-3 has a molecular-specific volume of 3.0 ´ 10-23 cm3 molecule-1, and the corresponding linear scale is the cube root, i.e. 0.31 nm. Based on their literature review, McClellan and Harnsberger recommended 12.5 Å2 as the effective surface area occupied by adsorbed water molecules. The corresponding monolayer thickness to produce the proper molecular-specific volume would be 0.24 nm.

This finding is consistent with direct experimental measurements of water film thickness. We will adopt the following practice. When original sources quoted here report results in units of monolayers, then we will cite those results without amendment. . When original sources report isotherm data in other terms, such as mass of water sorbed per mass of sorbent, we then convert these data to monolayer equivalents using a monolayer thickness of 0.24 nm, based on the McClellan and Harnsberger recommendation.Another important detail is that the true surface area onto which sorption occurs often exceeds the nominal surface area, even for seemingly impervious materials such as glass. In cases in which we have computed monolayer equivalents, we use the specific surface areas measured by the BET method as reported in the original source . Studies of water sorption on the surfaces of different types of materials reveal two important points. First, at the ordinary relative humidity levels encountered indoors, it is common for the abundance of sorbed water to exceed a monolayer. Table 5 records the numbers of equivalent monolayers of sorbed water onto different types of surface materials,pot drying typically either pure minerals or mixtures of mineral origin. Among the entries, the median are as follows: at 30% RH — 1.9 monolayers; at 50% RH — 2.3 monolayers; at 70% RH — 3.8 monolayers. It is noteworthy that that the results exhibit a fair degree of homogeneity despite the diverse minerals tested. These results also are in broad agreement with the information presented by Leygraf et al. concerning metal surfaces: “a number of metals are seen to be covered by water equivalent to 2-10 monolayers for relative humidities exceeding 40% at normal room temperatures.” They also state, “The first layer of water has a high degree of ordering relative to the substrate because of its proximity to the solid surface. The second and third layers are more mobile with a higher degree of random orientation. Aqueous films thicker than three monolayers possess properties that are close to those of bulk water.” Although the properties of surface water approach those of bulk water as the film thickness grows, the water film may also have properties that differ from bulk water so as to coexist at equilibrium with water vapor at relative humidities less than 100%. A second point is that the specific surface areas of these materials commonly are much greater than the superficial or apparent surface areas. For example, a sample of sand studied by Lin et al.had a specific surface area based on the BET test with N2 of 0.4 m2 g-1 , a value 60´ as large as the nominal surface area of equivalently sized solid spheres. Most of the additional surface area was associated with internal pores, even though the within-grain porosity was determined to be only 1.4%. With the high internal surface area, at 50% RH, this sand sample had a sorbed water content of 1.1 mg per g of sand.The existence of a much higher internal surface area than the apparent or nominal surface poses difficulties for translating the isotherm-based information of the type displayed in Table 5 into information about the expected contribution of surface-sorbed water to L*. Fundamentally, we lack adequate information about the effective true surface areas of mineraltype materials exposed indoors. Another complication in applying the data reported in Table 5 is that interior surfaces are commonly coated with films of organic materials that may alter the nature of water-surface interactions . Liu et al. were among the first to characterize such films on indoor window surfaces, and many subsequent studies support an inference that organic films are ubiquitous on indoor impervious surfaces. Wu et al. demonstrated that airborne exposure to “kitchen grime” caused different surfaces to exhibit comparable thermodynamic properties with respect to sorptive partitioning of phthalates. Weschler and Nazaroff56 and Eichler et al. have modeled film formation and growth. Schwartz Narbonne and Donaldson72 exposed initially clean, gold-coated quartz crystals for periods of approximately two months in two occupied homes in Toronto.

These crystals were oriented horizontally during the exposure period, so they would accumulate both settling dust as well as organic vapors. After the exposure period, the crystals were exposed to conditions in which the relative humidity could be controlled. The humidity was systematically varied “from 5% to 85% at a rate of 1% per minute.” By measuring the change in oscillation frequency, the mass change associated with water uptake could be evaluated as a function of relative humidity. Unexposed crystals were also analyzed and these results were used for blank correction, so that only the uptake of water into the surface film was assessed. The detailed results, displayed in Figures 2- 4 of the cited reference, show an irregular feature: the water mass associated with several exposed crystals did not rise monotonically with increasing RH. Nevertheless, among 12 samples for which results were reported, at RH = 30%, 50%, and 70%, the mean ± standard deviation of water surface densities was 0.15 ± 0.09, 0.42 ± 0.26, and 0.90 ± 0.56 µg cm-2 , respectively. The corresponding effective mean monolayer thicknesses of sorbed water would be 6, 18, and 38, respectively. There was not a clear pattern of differences between pairs of rooms sampled at each site or between study sites. For an effective monolayer thickness of 0.24 nm, 18 monolayers would correspond to an equivalent thickness of 4 nm. By comparison, the experimental data on window film growth summarized in Weschler and Nazaroff show median and mean growth rates of 0.15 nm/d, so that an expected representative thickness of organic constituents after an exposure period of two months would be about 9 nm, or about 2´ the associated contribution from sorbed water at 50% RH. An important point to highlight is that the Schwartz-Narbonne and Donaldson results are based on the nominal or superficial surface area of the sorbing substrate, in this case the goldcoated quartz crystal covered with an indoor-exposure acquired surface film. If the mean water surface density of 0.42 µg cm-2 for RH = 50% applied for the average exposed surface in an indoor environment with an overall surface-to-volume ratio23-25 of 3.5 m2 m-3, then this water could contribute 1.5 ´ 10-5 L m-3 to the indoor liquid water ratio, L*. Recent studies have investigated the properties of surface water in relation to the behavior of acids. Using silica as the substrate, Fang et al.showed that a surface-bound acid can deprotonate in the presence of sorbed water and that the degree of deprotonation is greater for a strong acid than for a weak acid . Wellen et al.investigated the behavior of octanoic , nonanoic , and decanoic acids. They found a reduction in the acidity of the species at the water-air interface. In contrast to the water/air interface, at the substrate/water interface Parashar et al.used a modeling approach to “show how the acidity of pyruvic acid at the quartz/water interface is increased by almost two units” .