Energy & Buildings 202 (2019) 109394

Contents lists available at ScienceDirect

Energy & Buildings

journal homepage: www.elsevier.com/locate/enbuild

Moisture buffering in buildings: A review of experimental and

numerical methods
Brenton K. Kreiger a, Wil V. Srubar III a,b,∗
a
Department of Civil, Environmental, and Architectural Engineering, University of Colorado Boulder, ECOT 441 UCB 428, Boulder,
CO 80309-0428 United States
b
Materials Science and Engineering Program, University of Colorado Boulder, UCB 596, Boulder, CO 80309-0596 United States

a r t i c l e i n f o a b s t r a c t

Article history: Moisture buffering in buildings is well known to influence material durability, building-scale energy ef-
Received 14 June 2019 ficiency, and indoor environmental quality. In this work, we present a comprehensive meta-analysis of
Revised 20 August 2019
experimental studies and a review of numerical approaches concerning the moisture buffering capacity
Accepted 29 August 2019
of common building materials. More specifically, we synthesize and analyze reported moisture buffer-
Available online 29 August 2019
ing values (MBVs) of materials from >180 unique characterization experiments. In addition, we classify,
Keywords: compare, and critically discuss experimental methods employed to measure MBV, along with numerical
Moisture buffering methods that have been used to quantify building-scale benefits. Experimental data indicate that biotic
Building materials and chemically hydrophilic (e.g., cellulosic) materials exhibit higher MBVs than porous, abiotic (e.g., ce-
Energy efficiency mentitious) materials, which suggests new opportunities for engineering natural, synthetic, and/or hybrid
Indoor environmental quality hydrophilic materials that display hyperactive MBVs. In addition, moisture buffering effects have been
shown to yield up to 30% energy savings in certain climates. However, our analysis reveals that more con-
sistent experimental and numerical methodologies are still needed to accurately quantify building-scale
benefits. To this end, we identify specific gaps in scientific and technical knowledge and offer sugges-
tions for experimental and theoretical research that is required to advance the collective understanding
of moisture buffering and its effects on the energy consumption and indoor environmental quality of
residential and commercial buildings.
© 2019 Elsevier B.V. All rights reserved.

1. Introduction ing operation can result in up to 210% overestimation of peak heat-

ing loads and 59% underestimation of heat flux from latent heat
Active moisture management in residential and commercial and moisture effects [3]. Over- and underestimations of these mag-
buildings is an energy-intensive process. In 2017, the United States nitudes can lead to the overdesign of HVAC systems in residential
(US) Energy Information Administration reported that building en- and commercial buildings because simulated energy consumption
ergy comprised 39% of total US energy consumption [1], of which will increase due to increases in thermal conductivity of wall as-
92% was attributable to residential and commercial building heat- semblies. Accounting for hygroscopic effects of spruce plywood, for
ing, ventilation, and air-conditioning (HVAC) systems alone [1,2]. instance, has been shown to yield up to 20–30% energy savings in
To maintain indoor environmental comfort, mechanically driven terms of cooling loads in residential buildings [4–6] . HVAC control
HVAC systems stabilize indoor relative humidity (RH) within spe- strategies that consider passive moisture buffering of building ma-
cific ranges. In humid climates, for example, dehumidification re- terials have been shown to decrease energy consumption by 14–
quires both cooling and reheating air. Buildings with high in- 17% during heating periods [7]. In short, the passive ability of ma-
door humidity (e.g., natatoriums) require high degree of ventila- terials to absorb and desorb moisture is important to consider in
tion with conditioned air. Dehumidification, ventilation, and condi- estimating the peak heating and cooling demands in both residen-
tioning (i.e., heating and cooling) are all energy-intensive processes tial and commercial buildings.
that are required, in large part, to manage moisture in buildings.
Recent research shows that failing to account for passive mois-
1.1. Sources of moisture in buildings
ture absorption and desorption of building materials during build-
Moisture in buildings originates from several sources, including
∗
Corresponding author. outdoor air and indoor plants, occupants, and release from high-
E-mail address: [email protected] (W.V. Srubar III). water-content building materials (e.g., wood, concrete) soon after

https://doi.org/10.1016/j.enbuild.2019.109394
0378-7788/© 2019 Elsevier B.V. All rights reserved.
2 B.K. Kreiger and W.V. Srubar III / Energy & Buildings 202 (2019) 109394

construction. The moisture content of outdoor air depends heav- skin dryness and congestion [24]. While low RH causes adverse
ily on climate zone, which dictates average RH, types and levels health effects, excess RH, has also been linked to other respi-
of precipitation (e.g., snow, rain), and ground-level moisture. In- ratory illnesses, such as asthma, wheezing, and bronchial hyper-
door moisture conditions depend on the type and amount of flora, responsiveness (BHR) [25]. Additionally, high humidity levels can
as well as occupancy levels and associated activities. Humans can cause occupant discomfort and can alter the perception of indoor
generate 115 to 270 gs of water per hour (2.8 to 6.5 kg per day) air quality [26]. In other words, humans will perceive humid in-
through respiration and perspiration [8]. Water fixtures (e.g., show- door air as heavy, muggy, and uncomfortable and less sanitary to
ers, sinks, toilets) and the evapotranspiration of indoor plants also breath. Given that humans spend an average of 90% of their time
significantly contribute to indoor moisture levels [8]. High-water- indoors (e.g., homes, offices, schools) [27] and that high (>70%) or
content construction materials will acclimate over time and re- low (<40%) RH imparts measurable health effects [28], confining
lease moisture into the building. Overall estimates for moisture RH levels to prescribed, acceptable ranges is of critical importance
released in the first year of new construction is ∼10 kg per day in building operations.
[9]. More specifically, dimensional lumber utilized at a standard
moisture content (i.e., 15% or 19%) can release ∼200 kg of water 1.4. Impact and scope of review
in an average residential single-family home as it equilibrates [9].
Concrete can release ∼90 kg/m3 over two years post-construction Moisture buffering as a passive humidity regulation strategy
[9]. Together, these outdoor and indoor sources of moisture im- is both an economical and energy-efficient approach that is cur-
part non-trivial material, energy, and human health consequences rently not fully understood and, consequently, underexploited in
in buildings, which are further discussed in the following sections building design and construction. This review critically explores
[6]. the experimental and numerical methods of quantifying moisture
buffering effects and estimating their building-scale impact. First,
1.2. Building health experimental methods employed to measure moisture buffering
values (MBVs) of building materials are classified, compared, and
Moisture can trigger physical, mechanical, chemical, and biolog- reviewed. Empirical MBV data from >180 characterization exper-
ical deterioration mechanisms in building materials that can lead iments are synthesized and analyzed to elucidate the physical,
to building-scale damage. This has been a well-known problem for chemical, and biological characteristics of materials that display
many years, in fact, previous research from over 40 years prior has particularly high MBVs. Second, hygrothermal modeling methods
reported that up to 90% of all construction material and building for building-scale applications are classified and critically analyzed
durability issues are caused by moisture [10]. Specifically, mois- based on computational expense, convenience, and accuracy. Fi-
ture content of hygroscopic structural materials, like wood, affects nally, new, emerging research on the moisture buffering properties
size (e.g., swelling, shrinkage), strength, and stiffness [11]. Wu et al. of innovative hygroscopic materials is highlighted. We conclude
[12] found that commercially available oriented strand board (OSB) this review by discussing new opportunities and future research
exhibited 31% dimensional swelling with an increase in moisture directions that are required to advance technical understanding,
content of 24%. In terms of mechanical properties, high moisture material development, and computational modeling of moisture
levels can result in up to 70% loss in allowable strength for wood buffering effects and to leverage the benefits in the design, con-
members [13]. These physical and mechanical effects in wood can struction, and operation of residential and commercial buildings.
be exacerbated by moisture absorption and desorption cycling [14].
Moisture can also lead to other physical, mechanical, and chemi- 2. Review of experimental methods
cal deterioration effects, such as efflorescence [15], hydrolytic and
exacerbated UV degradation of polymeric materials [16], chloride- 2.1. Characterization methods
induced corrosion in reinforced concrete [17], and freeze-thaw de-
terioration [18]. Regarding biological deterioration, moisture can Moisture buffering value (MBV) is the most well accepted pa-
induce localized conditions for mold and fungal growth, which can rameter that is used to describe and compare the moisture buffer-
accelerate material aging [19]. RH is intimately related to mold ing capacities of different building materials. MBV, defined as the
growth rates, where mold growth depends on the frequency of change in mass per square meter per change in RH (g/RH/m2 ),
low and high humidity levels and time of wetness [19]. In a 2001 is most commonly characterized by a stepwise vapor sorption
study led by Klaus Sedlbauer, 250 sources were reviewed to es- process—a method that involves a measurement of mass change
tablish mold growth isopleths on common building materials [20]. with discrete increases or decreases in RH over various time steps.
The LIM (Lowest Isopleth for Mold) begins at 75% RH with up to MBV is calculated according to the following [29,30]:
2 mm/day of growth at 85% RH and 5 mm/day above 95% RH [20].
G
Biological deterioration is of critical concern regarding not only the MBV = (2.1)
structural integrity of biotic materials (e.g., wood), but also the po- RH
tential negative effects on indoor environmental quality and asso- where G is the moisture uptake in g/m2 throughout a prescribed
ciated consequences to human health and well-being. RH cycle.
Fig. 1 illustrates the historical evolution of MBV characterization
1.3. Human health methods. Seminal moisture buffering capacity research was con-
ducted in Germany at the Fraunhofer Institute of Building Physics
Both high and low levels of moisture in indoor air have been and at Lund University in Sweden in the late 1960s. By devel-
linked to health problems since the early 20th century. In seminal oping the step-response method that is still used to characterize
studies, researchers linked dry air from furnace-heated homes in moisture sorption in response to cyclic, time-dependent changes
a New England winter to skin and respiratory irritation [21] and in RH [31], these experiments set the standard for contemporary
to the diminishing health of children in grammar schools [22]. In moisture buffering experimentation. Following these studies, the
the 1910s, researchers showed that 40% RH in residential buildings Padfield method was developed as an alternative to the stepwise
of the era was enough for condensation—and the health-related model [32]. The Padfield method draws an analogy to thermal ca-
problems that follow—to occur [23]. Recent research has shown pacitance by defining the moisture capacitance of air using the
that conditioned air with RH >40% relieves nasal, pharyngeal, and unit of buf, which cannot be compared to the more standard MBV
 B.K. Kreiger and W.V. Srubar III / Energy & Buildings 202 (2019) 109394 3

Fig. 1. Historical milestones in experimental methods for MBV.

unit of g/RH/m2 . The buf is reported in meters and refers to the Table 1
MBV classification system established by the
height of 1 m2 column of air. The higher the moisture capaci-
NORDTEST method [29].
tance of a material, the higher the buf and the greater effective
volume of air that can hold moisture. The Padfield method is dis- Minimum Level Maximum Level
MBV MBV (g/RH/m2 )
tinct in that it does not define MBV as an intrinsic material prop-
Class 8 h @ 75% RH and 16 h @ 33% RH
erty. The most recent methods that have been developed, includ-
Negligible 0 0.2
ing the Japanese Industrial Standard [33], NORDTEST [29], and ISO
Limited 0.2 0.5
[34] methods, however, have all used stepwise RH variation as a Moderate 0.5 1.0
primary approach to characterize MBV. Good 1.0 2.0
Excellent 2.0 >2.0

2.2. Material- and Building-Scale experimental methods for

evaluating mbv 2.2.3. Japanese industrial standard jis a 1470–1
Introduced in 2002, the Japanese Industrial Standard (JIS A
2.2.1. The german standard din 18,947 1470–1) [33] employs a similar step-response method established
The German-developed standard DIN 18,947 [35], which was by the German Standard DIN 18,947. The experimental procedure
developed specifically for characterizing the hygric properties of involves a stepwise preconditioning method that first equilibrates
earthen plasters, is the oldest standard using methods first pro- materials at 23 °C and at either 43%, 63%, or 83% RH, prior to
posed in 1965 by Kunzel. The prescribed method applies cyclic a stepwise dynamic conditioning procedure. A minimum sample
conditioning of RH between 50% and 80% in 12 h cycles. Percent- area of 100 cm2 is required. Most studies utilize rectangular prism
mass changes over that prescribed RH range is used to define geometries in which all sides except one are sealed with alu-
moisture buffering capacity. The original German standard has not minum tape to effectively capture the effects of 1D moisture trans-
been widely applied and cited in the literature, as only two of the port. While similar to the DIN 19,847 standard and other methods
188 published experimental tests reviewed in this work employed that followed its development (e.g., NORDTEST), the JIS method is
DIN 18,947. However, additional references that were reviewed in- unique. First, there is an equal stepwise cyclic conditioning proce-
dicate that this standard has been most often used internally by dure that involves cycles of 24 h of high humidity exposure and
German industry researchers to characterize and compare hygric 24 h of low humidity exposure. Three cyclic conditioning RH levels
performances of different materials [36–38]. are specified: 33% to 53%, 53% to 75%, and 75% to 93% RH. Second,
the material surface film resistance (a function of air speed around
and geometry of sample) is prescribed as 4.80 ± 0.48 × 107 Pa/kg.
2.2.2. The padfield method Third, the material thickness must be equal to the actual, real-
The Padfield Method, developed by Tim Padfield in 1998, specif- istic thickness of the product to be tested [38,39], since, as dis-
ically focused on moisture buffering capacity of materials as it re- cussed, material thickness can make a significant impact on mois-
lates to historic preservation of buildings [32]. The method pre- ture buffering value [40]. Overall, while this is an established and
scribes a small, cyclic step-response procedure between 50% and accepted method, none of the published studies in this review uti-
60% RH. The aim of this method—and what differentiates it from lized the JIS procedure, as it is not usually referred to in English
others—is that it quantifies the effect of hygroscopic materials on publications [38].
ambient RH as opposed to the opposite, standard measurement,
which quantifies the effect of cyclic RH on changes to material
mass. The basic experimental set-up consists of a sealed chamber 2.2.4. The NORDTEST method
with a water reservoir and controlled air speed that varies between Since 2005, the NORDTEST Method has been the most pop-
0.2–1.2 m/s to ensure adequate mixing. The temperature of the wa- ular standard for hygroscopic characterization of materials. The
ter, adjusted by a thermoelectric heat pump, controls the RH of the procedure was developed to create a standardized test method
sealed chamber. The method assumes that all water lost from the that characterized a well-definable and consistent material prop-
reservoir is absorbed by the test material. While a benefit of this erty. The NORDTEST method established the conventional defini-
set-up is that any geometry of sample can be tested, one drawback tion of MBV as it is referred to herein and defined broad cate-
is the water sorption of equipment and air must be corrected in gories of MBV from negligible to excellent for commercial compar-
the quantitative analysis to ensure accurate results. As previously ison of materials (see Table 1). During its development, university
discussed, the unit to report moisture buffering was defined as the researchers carried out round-robin MBV tests to evaluate the con-
buf in the Padfield method, which cannot be directly compared to sistency of experimental protocol across material types. Because
MBV [32]. results from the round robin tests among universities showed good
4 B.K. Kreiger and W.V. Srubar III / Energy & Buildings 202 (2019) 109394

Table 2
Summary of UMBV testing conditions.

Stage Temperature (C) RH (%) Time (hr/day) Practical Application

I 23 ±0.3 50 ±0.3 12 Damp-proofing, instantaneous response capacity

II 40 ±0.3 98 ±0.3 8 Damp-proofing, adsorption
III 18 ±0.3 3 ± 0.3 4 Desorption, instantaneous response capacity

agreement, the high repeatability and consistency of results pro- 2.2.6. Ultimate moisture buffering value (UMBV)
pelled the popularity of the NORDTEST method. UMBV is an adaptation of the NORDTEST MBV [43] that ac-
Over 70% of the studies reported herein used an official or mod- counts for adverse temperature and humidity conditions that are
ified version of the NORDTEST method. The experimental method more representative of four-season climate zones. The high humid-
consists of a 24-h cyclic RH stepwise variation between 33% RH for ity level corresponds to 98% and the low corresponds to 3% with
16 h and 75% RH for 8 h. A common modification of the NORDTEST temperatures ranging from 18 to 40 °C. Samples are preconditioned
method, the two-bottle method, was used in research facilities that at a standard 23 °C and 50% RH until hygroscopic equilibrium is
could not meet the NORDTEST chamber conditioning criteria but achieved. Using these parameters, the MBV is calculated and mul-
used similar principles with slightly different RH values between tiplied by a time coefficient, ??, determined by the rough frequency
50% and 80%. Similar to the JIS method, samples tested using the of exposure to these conditions in four seasons:
NORDTEST method must be a minimum of 100 cm2 of rectangu-

III
lar geometry. However, where the JIS method specifies that the UMBV = αi MBVi (2.2)
thickness of the sample must be the thickness of the product, the
i=I
NORDTEST method only states that the sample must have a thick-
ness greater than the calculated theoretical moisture penetration where α is the time coefficient, or total hr/day each sample is
depth (TMPD). This requirement is important, because a study by subjected to the specified RH. The α parameter is simply calcu-
Roels and Janssen [39] demonstrated through a sensitivity analy- lated by t/24 where t varies based on the stage of the UMBV test
sis that the main reason for difference in MBV values between JIS (see Table 2). The resulting equation computes the comprehensive
and NORDTEST method testing of samples was the sample thick- moisture tolerance of the material over three stages that corre-
ness. For exposure to RH, samples are covered on all but one or spond to damp-proofing capability, adsorption, desorption, and in-
two sides with aluminum tape to ensure realistic conditions of ex- stantaneous response capacity, as illustrated in the table below.
posure to indoor air [29,38]. If two sides are left exposed, the cal-
culated MBV must be divided by two. Samples are generally pre- 2.2.7. Full-Scale experimentation
conditioned at 23 °C and 50% RH until hygroscopic equilibrium by Despite advances in material-scale characterization, full-scale
weight is achieved, prior to the test. As mentioned above, the ex- experimentation is arguably a more accurate method to gather
periment involves cyclic RH variation, which can be achieved us- experimental data on moisture buffering capacity or to validate
ing a climate chamber or with salt solutions in sealed chambers. predictive models that quantify energy and environmental bene-
Temperature is held isothermally at 23 °C. As a criterion to stop fits of moisture buffering. Intuitively, however, full-scale experi-
the test, the NORDTEST and most other methods require a quasi- ments are more expensive, time-consuming, and rarely reported in
steady-state equilibrium to be reached in which the change in the literature. Additionally, completely non-hygroscopic houses do
mass between absorption and desorption steps varies less than a not necessarily exist due to the hygroscopic nature of furnishings
pre-defined threshold [41], since the slope of initial moisture in- and other coating materials [44]. Some full-scale experiments were
take is shown to overestimate actual MBV [41]. Additionally, while performed in Holzkirchen, Germany, in which researchers used
the NORDTEST method does not specify any ventilation rate for a large climate chamber to show that moisture buffering capac-
testing, the standard does mention an important inverse relation- ity of gypsum board could reduce peak humidity up to 44% [31].
ship between ventilation and MBV and states that the sample sur- Fraunhofer-Institute of Building Physics in Holzkirchen conducted
face air speed should be 0.10 ± 0.05 m/s, which corresponds to another similar experiment to validate the widely used WUFI soft-
a surface film resistance of 5.0 × 107 m2 -s-Pa/kg [42]. The differ- ware, in which the moisture buffering effects of a reference room
ence between this value and the 4.8 × 107 ± 10% m2 sPa/kg value covered in aluminum was compared to that a wood-clad test room
set forth in the JIS method is important because of the established [45]. Using tracer gasses to track ventilation rates and exposed hy-
inverse relationship between surface film resistance and MBV. groscopic surface area to measure moisture transfer, a test house in
Helsinki, Finland, compared a wood-framed home with and with-
2.2.5. ISO 24,353 out a polyethylene vapor retarder to show 15% RH peak reductions
The first edition of ISO 24,353 [34] was published in 2008 and at 27 °C, also concluding that material moisture buffering is more
is based on the JIS method. Similar to the JIS method, samples are effective than increasing ventilation rates from 0.08 to 0.55 ACH
prepared with aluminum tape sealing all of the sides except one [46]. More recently, Shi et al. conducted a full-scale experiment
and must have a minimum surface area of 100 cm2 . Precondition- to test non-standard building materials by outfitting a typical civil
ing occurs at 23 °C and either 43%, 63%, or 83% RH, and step cy- defense shelter in Beijing for field measurements and numerical
cles occur over 12 h periods from 33%−53%, 53%−75%, and 75% validation [2]. Allinson and Hall [38] monitored the interior mois-
−93%. Similar to the JIS, there is no specified method for main- ture buffering effect of a rammed earth shed. Huibo Zhang et al.
taining RH values, unlike the NORDTEST method that requires a [45] conducted a moisture buffering experiment under real-world
climate chamber or salt solutions. However, these are proven to be conditions. In a study by Luyang Shi et al., two rooms were set-up
the most cost effective and simple experimental setups and used within a real home in order to compare a test room to a refer-
consistently in studies that employ the ISO standard [38]. Given its ence point to evaluate porous ceramic tile, biomass fiber wallpa-
relatively recent establishment, this method has been rarely uti- per, and vermiculite board [46]. Overall, full-scale implementation
lized and, therefore, scarcely cited in the published studies consid- is a comprehensive practice for eliminating major assumptions and
ered in this review. gathering realistic data on the applied moisture buffering effect of
 B.K. Kreiger and W.V. Srubar III / Energy & Buildings 202 (2019) 109394 5

a building assembly. However, the drawbacks for space, time, and

cost, are significant, which has motivated the industry to develop
in-situ characterization practices through numerical modeling and
material-scale experimentation to define MBV as a material prop-
erty.

2.3. Characterization methods: key assumptions

In addition to employing different experimental approaches to

characterize and report MBV, each method makes key assump-
tions that affect the general applicability of material-scale MBV to
building-scale behavior. As discussed in the next section, very few
studies go as far as to experimentally measure the moisture buffer-
ing effect of materials within building systems [31,38,45,47,48], and
no studies that were reviewed experimentally report building-scale
application of highly absorbing materials. However, many numeri-
cal studies have relied on material-scale characterization of MBV to
estimate the effect of moisture buffering on building energy con-
sumption, thus placing high dependence on the MBV experimental
methods to adequately capture moisture buffering capacity of dif-
ferent building materials.
The primary assumptions inherent to a majority of MBV ex-
perimental methods—and their potential consequences in terms of
building-scale applicability of MBVs obtained by these methods—
are as follows: Fig. 2. Frequency of methods used to experimentally determine MBV.

1. Stepwise changes in RH are representative of RH changes in

indoor environments. The potential energy savings of mois- tion, permeation) [50], which will affect the concentration-
ture buffering come from the ability of materials to pas- dependence of its diffusivity. Therefore, given the wide range
sively dampen RH peaks. Hygroscopic materials interact dy- of responses of materials to moisture, the static diffusivity
namically with indoor RH rather than discretely, as assumed assumption imparts unknown variability in the definition of
by the stepwise methodology. Therefore, given that RH will MBV as an inherent material property. Deviations beyond
vary continuously with moisture content of indoor surface the RH range of 30–70% have also been shown to exhibit sig-
materials and not remain constant for a period of time [49], nificant variations in moisture diffusivity depending on RH
providing a constantly high or low RH environment to test a [51].
material may overestimate the MBV. 5. Ventilation and airflow across the sample is negligible. Ventila-
2. The method parameters for number of cycles, temperature, and tion is mentioned in various standards (e.g. JIS, NORDTEST),
RH sufficiently provide conditions in which equilibrium can be yet not considered homogenously across many studies. For
achieved for all materials. Test methods specify isothermal example, Colinart et al. [50] and Nguyen et al. [51] do not
conditions, relevant RH steps, and a tolerance for mass varia- mention the effect of airflow in their testing of hemp con-
tions, representative of quasi-steady state equilibrium. How- crete and bamboo fibers. However, air flow will affect lo-
ever, this does not directly relate to building applications be- calizes surface RH. The MBV review of earthen materials
cause the material is not always responding with a MBV cor- by Svennberg et al. [31] states that lower air change per
responding to its quasi-equilibrium state. This assumption is hour (ACH) rates will yield higher RH levels, which can in-
important, because absorption and desorption can differ— evitably affect the definition of MBV by up to 20%. Shi et al.
and the calculation of overall MBV can differ – from equi- [47] confirmed this by measuring ACH with tracer gasses to
librium values [29]. test the effect of five different ventilation rates on MBV and
3. Sample thickness is larger than its theoretical moisture pene- confirmed that they have a significant impact on measured
tration depth (TMPD). Theoretical moisture penetration depth MBV. These findings not only illustrates the importance of
is defined as the point at which moisture content variations holding ventilation rate constant in determining MBV, but
are only 1% of those on the surface [29]. Most experiments also suggests that differing ventilation rates in real building
assume that the TMPD is within the sample thickness with- applications may account for the inconsistencies between
out substantiating it being so. A study by M. Rahim et al. material-scale MBV measurements and building-scale mois-
[40], however, illustrated that MBV continued to increase ture buffering behavior.
0.3 g/RH/m2 when varying the sample thickness from 3 to
7 cm while TMPD was previously calculated at 3.14 cm. This 2.4. Meta-Analysis of experimental data
finding indicates that MBV may continue to increase with
thickness, regardless of TMPD. 2.4.1. MBV method frequency
4. Moisture diffusivity remains constant throughout testing. Stud- Despite the wide variety of methods previously described,
ies assume moisture diffusivity remains constant within the the NORDTEST method remains the most widely applied—and,
RH range of 30% to 70%, which is approximately within therefore, the most comparable—method to characterize moisture
the bounds used in most MBV studies and realistic condi- buffering capacities of building materials. Fig. 2 illustrates the
tions for building applications. However, moisture diffusivity frequency of different methods applied in the studies reviewed
is well known to be concentration-dependent [50], and the herein. As shown in Fig. 2, the NORDTEST method was most fre-
chemistry of different materials can dictate single or multi- quently employed (>70% of studies) to measure MBV, while meth-
ple mechanisms of moisture transport (e.g., diffusion, sorp- ods “Not Specified” and the UMBV method were the second and
6 B.K. Kreiger and W.V. Srubar III / Energy & Buildings 202 (2019) 109394

Fig. 3. Reported MBV of building materials. CEB = compressed earth block, CS = clay-sand, ELS = engineered local soil, PCM = phase change material [29,30,55–64,40,65–
69,42,43,47,48,52–54].

third most commonly applied, respectively. Only five studies em- all studies was 0.47 ± 0.32 g/RH/m2 , which represented the
ployed the Padfield Method, which, as discussed, does not utilize largest deviation amongst the three error sources compared. These
the conventional step-response procedure prescribed by the other results are expected because the method error also encapsulates op-
methods. erator error because seven different studies are compared and no
data in which the lab used multiple studies to quantify MBV was
2.4.2. MBV experimental data available. However, using standard deviation for quantifying error,
A comprehensive visualization of reported MBVs is shown in method error still significantly increases the variability of the test
Fig. 3. This chart illustrates how widely MBV varies across individ- even when operator error is included.
ual studies and measurement methods, as well as variation within Fig. 4b depicts variability throughout a clay and sand plaster
identical materials and between different materials. material measured multiple times in the same study [56] using the
Table 3 reports statistical data for reported MBV, density, poros- same testing method (in-house climate chamber) to remove any
ity, and the method by which MBV was characterized per material operator or method error. Sample variability resulted in the low-
reviewed herein. We report different MBVs in Table 3 for materials est error with an average of 0.31 ± 0.05 g/RH/m2 . This finding
that were characterized by different methods. illustrates that there will likely always been an inherent material
As evidenced by the data presented in Fig. 3 and Table 3, high heterogeneity error from physical variability, which can be accentu-
variability exists for the reported values of MBV and other intrinsic ated by operator error and method error.
material properties. Previous authors have noted that results from Fig. 4c illustrates potential deviations from the mean caused by
hygric experimentation can be significantly affected by differences operator error, comparing various spruce board MBV studies from
in the experimental set-up and operator error, while the results five different universities that all employed the NORDTEST method.
can also be affected by other protocol and geometry factors, in- The mean MBV obtained was 0.57 ± 0.08 g/RH/m2 and ranged
cluding number of cycles and sample size [72]. However, some er- from 0.67 to 0.45 g/RH/m2 . In these examples, operator error is
ror in results appears inherent to various testing methods regard- less significant than method consistency, yet more important in
less of corrections made for minor differences in the experimen- terms of variability than material heterogeneity. These results agree
tal set-up. For example, findings from a round robin test with 14 with findings from previous sensitivity analyses on MBV error fac-
participating laboratories illustrate that sorption isotherm testing tors, which reported that operator error is the most significant
is consistent, yet vapor permeability procedures yield significantly source of error [72].
different results across labs [73].
Fig. 4a-c highlights the high variability that can arise due to in- 2.4.3. MBV vs. thickness, density, and porosity
consistencies between methods, material heterogeneity, and oper- The relationship between MBV and thickness, density, and
ator error, respectively. Fig. 4a depicts variability amongst MBV ex- porosity is graphically illustrated in Fig. 5a–c, respectively. As
perimental methods, in which three major methods testing 13 dif- shown in Fig. 5a, sample thickness did not correlate well with re-
ferent gypsum board samples were compared. The average across ported MBV. This result was anticipated, given that thickness is an
 B.K. Kreiger and W.V. Srubar III / Energy & Buildings 202 (2019) 109394 7

Fig. 4. Variability imparted by (a) Method: Gypsum board characterization across different study methods. Dashed lines represent within one standard deviation of the
mean (solid line); (b) Material Heterogeneity: Variability in MBV of different clay and sand plaster samples. Boundary lines represent one standard deviation of the mean
(middle line); (b) Operator Error: Spruce Plywood characterization between different testing facilities.
8 B.K. Kreiger and W.V. Srubar III / Energy & Buildings 202 (2019) 109394

Table 3
Statistical summary of reported MBV, porosity, and density by material and method. CEB = compressed earth
block, CS = clay-sand, ELS = engineered local soil, PCM = phase change material. Note that one standard deviation
is used, representing a 67% confidence interval.

MBV Density
Material n g/RH/m2 Method Porosity (kg/m3 ) References

Bamboo Fiberboard 6 3.0 ± 0.79 NORDTEST N/A 404 ± 85 [53]

Barley Straw 1 3.2 ISO 24,353 92% 108 [55]
Birch Panel 3 0.85 ± 0.22 NORDTEST N/A 600 [29]
Brick 3 0.48 ± 0.19 NORDTEST N/A 1600 [29]
Carnauba Wax 1 1.1 NORDTEST N/A N/A [58]
CEB 7 1.9 ± 0.54 NORDTEST N/A 1800 [63]
CEB + Barley Straw 4 2.6 ± 0.10 NORDTEST N/A 1735 ± 71 [55]
Cellulose 1 3.1 NORDTEST N/A N/A [59]
Cement 2 0.37 ± 0.16 UMBV N/A 1925 ± 153 [43]
Ceramic 1 0.26 Not Specified N/A 1740 [70]
1 0.95 NORDTEST N/A 1500 [42]
Clay Plaster + Fiber 4 1.7 ± 0.02 NORDTEST 41 ± 2.8% 1544 ± 144 [67]
Clay, Sand Plaster 9 0.31 ± 0.05 Not Specified N/A N/A [56]
Concrete 2 0.85 ± 0.51 UMBV N/A 1346 ± 1279 [43]
2 0.70 ± 0.42 Not Specified N/A N/A [71]
10 0.88 ± 0.56 NORDTEST 70.6% 1335 ± 762 [29,30]
Corn Pith 1 3.0 ISO 24,353 98% 48.1 [55]
Earth Plaster 1 2.9 NORDTEST N/A 1848 [55]
Earth Plaster + Fiber 6 3.0 ± 0.31 NORDTEST N/A 1362 ± 242 [55]
ELS 5 0.90 ± 0.37 NORDTEST 22.3 ± 2.2% 2076 ± 65 [69]
Fiber Wallpaper 1 0.15 Not Specified N/A 300 [70]
Gypsum 1 1.1 UMBV N/A 874 [43]
4 0.33 ± 0.18 Not Specified N/A N/A [56,61,71]
10 0.45 ± 0.30 NORDTEST N/A 977 ± 204 [29,54,59]
Gypsum, Lime Plaster 3 0.14 ± 0.06 NORDTEST N/A 900 [54]
Hemp Concrete 7 1.89 ± 0.32 NORDTEST 76.4 ± 2.7% 713 ± 645.5 [30,52,64]
Hemp Fiber 1 3.3 ISO 24,353 97% 41.1 [65]
Hemp Lime 1 2.3 ISO 24,353 83% 286 [65]
Hemp Lime Assembly 19 1.5 ± 0.70 NORDTEST N/A 610 ± 453 [62]
Laminated Wood 3 0.46 ± 0.08 NORDTEST N/A 430 [29]
MIL-100 (Fe) 1 15 NORDTEST N/A N/A [68]
Mortar + Fiber 1 3.0 NORDTEST N/A N/A [66]
Painted Gypsum 1 0.33 Not Specified N/A N/A [56]
PCM 1 0.08 NORDTEST N/A N/A [59]
9 0.81 ± 0.43 Not Specified 80% 793 ± 351 [61,71]
Pise 1 2.1 NORDTEST 24.2% 1870 [57]
Plaster 3 1.1 ± 0.40 NORDTEST N/A 670 [62,63]
Plywood 1 1.0 UMBV N/A N/A [43]
Rammed Earth 1 0.88 NORDTEST N/A 1980 [69]
Rape Straw Concrete 1 2.4 NORDTEST 75.1% 1954 [30]
Sand, Lime Plaster 1 0.90 NORDTEST N/A 1650 [52]
Sodium Polyacrylate 1 9.0 NORDTEST N/A N/A [59]
Spruce Board 4 1.2 ± 0.06 NORDTEST N/A 430 [29,58]
Spruce Board (sealed) 4 0.35 ± 0.25 NORDTEST N/A N/A [58]
Spruce Plywood 5 0.57 ± 0.08 NORDTEST N/A N/A [29]
Vermiculite Board 1 0.19 Not Specified N/A 746 [48]
WSE Mortar Assembly 2 0.92 ± 0.04 UMBV N/A 637 ± 21 [43]
Wood 1 0.4 Not Specified N/A N/A [61]
Wood 1 1.2 Not Specified N/A N/A [71]
Fiberboard 2 2.4 ± 0.47 NORDTEST N/A 458 ± 4 [62]
Wood Fiber 2 2.3 ± 0.50 ISO 24,353 91 ± 7.1% 136 ± 108 [65]

inherent physical property of each sample and directly relates to yet it is not a statistically significant predictor of MBV as shown in
the total theoretical capacity of that sample to fully absorb mois- Fig. 5a.
ture throughout the bulk, while MBV is an explicitly character- Porosity is an intrinsic material property that is well known
ized surface-dominated capability of materials to passively buffer to relate to the physical capacity of materials to absorb moisture
moisture in the air. A simple least-squares regression shows no re- [29]. As illustrated in Fig. 5b, a more significant relationship is evi-
lationship between thickness and MBV. This result is most likely dent between measured sample porosity and MBV (R2 = 0.46). This
due to the fact that most studies are based off of the NORDTEST result is expected, given that higher porosity indicates a higher
method which requires a material sample thickness greater than propensity for water vapor to not only interact with the material
its theoretical moisture penetration depth (TMPD) [29]. However, (i.e., increased surface area), but also potentially condense and re-
at least one study shows that moisture sorption capacity can dis- main in the void space. While we observed very strong correla-
play a linearly increasing relationship with thickness and MBV as tions to material densities and porosities reported in the literature
thickness is increased past its TMPD [74]. While no argument is (Fig. 5d), no statistically significant relationship between density
made for this correlation, it is hypothesized that moisture, once and MBV was observed (Fig. 5c) (R2 = 0.02). No relationship here
absorbed by the material, can move into the bulk via other trans- was expected, given that some materials, like zeolites [60] or per-
port mechanisms (i.e., capillary action, diffusion). Overall, material lite [59], may exhibit ultra-high porosities (therefore low density)
thickness is a factor for increasing moisture buffering capacity [74], and, simultaneously, a poor ability to buffer moisture. A Pearson
 B.K. Kreiger and W.V. Srubar III / Energy & Buildings 202 (2019) 109394 9

Fig. 5. Reported MBV plotted versus (a) sample thickness, (b) porosity, and (c) density. Panel (d) illustrates the inverse relationship between sample porosity and density.

Table 4
Pearson correlation values and associated statistical significance.

MBV (g/RH/m2 ) Thickness (mm) Density (kg/m3 )

Thickness Pearson 0.040 – –

P-value 0.738 – –
Density Pearson −0.203 −0.075 –
P-value 0.090 0.541 –
Porosity Pearson 0.745 0.512 −0.741
P-value 0.001 0.051 0.001

correlation confirmed that porosity is, in fact, the only physical As shown in Fig. 5, MBV is difficult to correlate to certain intrin-
material property significantly correlated with MBV at a statistical sic material properties (i.e., thickness, density) and, as elucidated
significance worth considering (see Table 4). through this meta-analysis, even more difficult to correlate to the
other RH-dependent material properties listed in Table 5. In the
2.4.4. Other reported properties reviewed studies, properties that depend on RH (e.g., vapor per-
All material characterization studies investigated in this review meability, moisture diffusivity, moisture capacity) are discussed in
report other material properties that were characterized in tan- their contribution to overall moisture buffering effect and theoret-
dem with MBV. These material properties can be grouped into two ical moisture buffering values. However, we do not attempt herein
main categories: (1) intrinsic physical properties (i.e., thickness, to relate MBV to these other properties due to the lack of report-
density, and porosity) or (2) RH-dependent material properties. A ing and, for those that are reported, the wide variation of the RH
summary of all reported material properties—and their respective conditions used for their characterization.
classification—are reported in Table 5.
10 B.K. Kreiger and W.V. Srubar III / Energy & Buildings 202 (2019) 109394

Table 5
Summary of hygrothermal material properties.

Intrinsic Material Properties

Property Units Description

Sample Thickness mm Material sample property generally defined by study

parameters.
Density kg/m3 Measure of compactness that directly relates to
porosity and therefore available space for water
sorption.
Porosity % Measure of voids in a material and directly related to
theoretical absorption capacity of a material. The
arrangement and size of these pores affect sorption
and desorption mechanisms.
W
Thermal Conductivity m·K
Ability of a material to conduct heat; changes with
moisture content and is therefore RH dependent.

RH-Dependent Material Properties

Property Units Description
kgwater
Moisture Capacity kgmaterial
Capacity of water that can be held within the pore
spaces of a material; determined from the slope of
the sorption isotherm.
kg
Water Vapor Permeability m·s·Pa
Time- and vapor-pressure dependent value that
describes the resistance of a material to transport
vapor.
m2
Moisture Diffusivity s
Rate at which moisture diffuses into a specific
material; dependent on vapor pressure from ambient
RH.
Moisture Penetration Depth mm Often defined as the thickness at which the moisture
content variation is only 1% that of the surface;
dependent on vapor pressure from ambient RH.

These results, and those from the preceding meta-analysis, and complexity. We argue that, for the field to advance, a balance
indicate that high porosity, in combination with hydrophilic must be achieved between these “white box” and “black box” mod-
chemistries of natural materials, are ideal physicochemical char- els to result in validated, productive simulation results for model-
acteristics that lead to a high MBV. The conclusion is supported ing the effects of moisture buffering in residential and commercial
by the porosity-MBV relationship illustrated in Fig. 5b, in which buildings.
natural, highly porous materials exhibit excellent moisture buffer-
ing capacities. As explicated toward the end of this review, find- 3.1.2. Empirical methods
ings such as these have accelerated interest in capitalizing on the Empirical methods employed in the literature most often take
inherent moisture buffering properties of natural materials in the simplified building models that may or may not include the mois-
development of innovative, ultra-high moisture buffering materials ture buffering effects of hygroscopic materials and use a correc-
and composites thereof. tion factor, additional equations, or experimentally derived coeffi-
cients to account for an assumption that holds true in realistic and
3. Review of computational methods universal conditions. In effect, this approach calibrates the numer-
ical model to represent practical MBV effects that are necessary
3.1. Modeling methods for model validation. For example, in one study, researchers re-
formulated the problem to find a theoretical correction factor. The
3.1.1. Method introduction correction factor was motivated by the fact that RH variations are
In addition to experimental measurements, theoretical models not idealized square wave functions in realistic conditions as they
for simulating moisture buffering and predicting their effects at are in the MBV characterization experiments [71]. The researchers
the material and building scale have been formulated and imple- model RH variation as a quasi-harmonic function and compute a
mented in the literature. These models can be classified as empir- correction factor and subsequent equations for moisture uptake
ical, semi-empirical, and physics-based methods. Empirical meth- [71]:
ods are phenomenological in their exclusive use of experimental Gin = Gout = β ∗ MBVbasic (H − L ) (3.1)
values to inform their formulation. Physics-based models, like cou-
pled heat and moisture transfer (HAMT) models, stem from funda- where MBVbasic is consistent with the NORDTEST ideal MBV defi-
mental physical equations and rely on experimental data solely for nition and H and L are the high and low RH levels, respectively.
validation purposes. Semi-empirical methods combine elements of Gin is moisture uptake and Gout as moisture release, both in kg/m2 .
both empirical and physics-based models. The classification is anal- Eq. (3.2) defines the factor β used in Eq. (3.1):
ogous to the more historical categorization of modeling methods −0.035
[α (1 − α )]
as white box (empirical), gray box (semi-empirical), and black box β = 0.888 ∗ √ √ (3.2)
(physics-based) approaches [75]. Ultimately, these categories help α+ 1−α
to better understand to what extent the model is being informed where α is a unitless time constant derived from high and low
by experimental data or fundamentally derived equations and are RH cycle times. After determining the correction factor, researchers
helpful to consider in the validation and evaluation of these mod- tested the results against a tested and validated hygrothermal
els. Due to a lack of large-scale experimental data and difficulty model and found good agreement (<3% relative error) [71].
in even obtaining assembly- or building-scale data, recent models Another strategy based solely on empirical methods is the ef-
focus on physics-based approaches—an approach that is computa- fective capacitance model in which ambient air capacity is in-
tionally limiting when the models are scaled in size, resolution, creased to account for the properties of hygroscopic building ma-
 B.K. Kreiger and W.V. Srubar III / Energy & Buildings 202 (2019) 109394 11

terials. The effective capacitance method can be within 18% er- 3.1.4. Physics-based methods
ror of a more advanced, computationally expensive software when Physics-based methods are rooted in fundamental equations
it comes to component level moisture buffering effect; accurately that are used to model physical phenomena. These models can
predicting full-building RH buffering well and less-accurately pre- quickly increase in complexity (and, therefore, computational ex-
dicting sudden moisture loadings [76]. Analysis of results displays pense) as the physical problems become more difficult (e.g.,
an influence of hygrothermal materials on zone RH, yet little influ- coupled heat and moisture transfer). However, a comprehensive
ence on overall heating and cooling demand [76]. This strategy also physics-based model can be easily modified to fit a variety of
relies completely on the assumption that the interior space con- geometries and materials, while experimentally gathered data for
tains well-mixed air with uniform properties—an assumption that empirical models are time-consuming and expensive to modify.
reduces computation time and simplifies the model significantly In formulating moisture uptake for numerical modeling pur-
because an entire air volume of a zone can be represented as a sin- poses, often, Fick’s law is used [82].
gle node where values can be extracted as averages. To improve re-  
ϑp
sults without losing simplicity and time, researchers coupled com- g = −δ p (3.3)
putational fluid dynamics (CFD) and effective penetration depth ϑx
(EPD) models based on experimental data can calculate localized
where moisture flux, g (kg/m2 -s) is related to water vapor per-
surface transfer coefficients for the hygric properties of simulated
meability, δ p [kg/m-s-Pa], and the change of water vapor pres-
building walls. Adding these coefficient values to the well-mixed
sure p [Pa] through a thickness x [mm]. As the water vapor pres-
model successfully improved the results in cases that the surface
sure changes on the surface of the material (assuming a semi-
transfer coefficient was stable and physically relevant [77].
infinite body, given that thickness is greater than theoretical mois-
In summary, the goal of the aforementioned empirical methods
ture penetration depth), a vapor flow (moisture diffusion) is in-
is to use experimental data to either create a model or improve
duced, and overall moisture content will increase (absorption) or
the accuracy of a model without having to eliminate its driving as-
decrease (desorption) over time.
sumptions. However, the primary drawback of this approach—and
For total moisture uptake, from which the theoretical MBV can
why it is limited in its application—is that empirical data are sit-
be calculated, the moisture flux is integrated, and the water vapor
uational and assembly-specific, especially with a phenomenon like
permeability is replaced with bm [kg/m2 -Pa-s0.5 ], a term that rep-
moisture buffering, which is highly driven by geometric factors.
resents surface moisture exchange:

tp
G = bm  p h ( α ) (3.4)
π
3.1.3. Semi-empirical methods
Semi-empirical models combine physics-based and empirical Here, tp is the moisture interaction time period in seconds, p is va-
methods to create a model that is rooted in derived equations por pressure in Pa, and α is the fraction of time period where hu-
and gleans additional accuracy from inputs from experimental midity is high, so for the 8/16 h scheme used in the NORDTEST
data. Ideally, these methods also combine the efficient, in situ, and method, h(α ) = h( 13 ) = 1.007 because the procedure calls for high
widespread applicability benefits of physics-based modeling with humidity one-third of the time. This calculation results in a sim-
experimental data to improve accuracy and computational speed. plified version of the equation that is only applicable in the 8/16 h
In one example of a semi-empirical model that was employed in scheme [29].
the literature reviewed herein, researchers used a lumped model Drawing an analogy to heat transfer, using the moisture flux
based on effective moisture penetration depth (EMPD) that linked and total moisture uptake over a certain time period, the moisture
practical MBV to the ideal MBV as a function of moisture effusiv- effusivity of a material can be defined using a Fourier series to
ity, RH, and air-film resistance [51]. This method uses an equation- approximate the moisture exchange for a semi-infinite body sub-
based theory in tandem with empirical data of material proper- jected to a square wave form of moisture variation at the surface.
ties and environmental conditions to inform the model. Using the The total moisture uptake G [kg/m2 ] from high to low over the
room vapor balance and a model benchmark (discussed in further time period yields the following equation that is dependent on the
sections), the model showed between 1.5% and 50% absolute er- ratio of time the material is exposed to the high humidity condi-
ror depending on material property and material type. For practical tion [41,83].
MBV, between 7.7% and 44% absolute error was observed, depend-
h(α ) ≈ 2.252[α (1 − α )]0.535 (3.5)
ing on the material. Although large, these results performed better
than the pure empirical model. Furthermore, the errors from these
types models can be explained better by the quality of model in-

G ≈ 0.568 bm  p t p (3.6)
puts.
For determining quality inputs for a more robust semi-empirical Most studies recognize that, although more simplified, isolating
model, an empirical method was developed by Woods et al. to ex- the heat transfer and moisture transfer problems does not accu-
tract data by subjecting the materials to square-wave RH profiles rately describe the behavior of either phenomena due to the sig-
and, after validation, showed that the EMPD model can predict RH nificant relationships between heat and moisture transport and
distributions [78]. In 2018, Woods and Winkler looked at the sensi- the material properties that govern the behavior [84]. There-
tivity of model inputs for a two-layer EMPD model for determining fore, the coupled heat, air, and moisture transfer (HAMT) strategy
moisture buffering [79,80]. The two-layer model focuses on short- for hygrothermal modeling remains the prevailing, most accurate
and long-term buffering layers and assumes cyclic RH variations, method to estimate building energy demands because it simulta-
resulting in a mass-based moisture buffering determination. A sen- neously solves equations for temperature, RH, and vapor pressure
sitivity analysis showed that the deep material layer, often ignored, using a purely physics-based approach. A scale analysis of govern-
is an important consideration in the calculations [81]. However, ing heat and moisture transfer mechanisms for hemp concrete, for
given the difficulty of gathering data at this material layer, using example, shows that, for up to 95% RH levels, these equations are
the combination of experimental data-based surface inputs with significantly coupled, and the main driver behind moisture move-
physics-based equations for deep-layer moisture transport can be ment in the material is the temperature gradient [85]. Above 95%
an efficient method for modeling hygrothermal behavior. RH (which is albeit unlikely to occur in a building thermal comfort
12 B.K. Kreiger and W.V. Srubar III / Energy & Buildings 202 (2019) 109394

context), liquid transfer and latent heat from phase changes be- developed to test the accuracy of emerging software. During the
come more significant [85]. Furthermore, a 10% increase in wood IEA ECBCS Annex 41, an international collaborative project to fur-
moisture content can result in 30% increases in thermal storage ther develop modeling of HAMT, a set of common exercises were
capacity due to the high specific heat of water [86]. However, it developed to test each software [92]. Additional published test
must be noted that heat and moisture transfer will occur on dif- cases are available for researchers, such as the benchmark exer-
ferent time scales, as moisture transfer mechanisms are generally cises from the European HAMSTAD project, to test the accuracy
slower than heat transfer, making the computational process even of their methods. These benchmarking exercises are presented in
more complex [86]. Overall, to be most accurate, the building en- the HAMSTAD report from 2002 [93] and are used in a study by
ergy modeling methodology must consider the significantly cou- F. Tariku et al. [87] to validate their transient model for coupled
pled effects of heat and moisture transfer through a building en- HAMT by comparing their model results to analytical results. Jud-
velope, as well as the timescales in which they will interact with koff and Neymark go further to recommend that three classifi-
mechanical equipment and building loads. cations of test cases should be used for model validation. These
While heat flux and coupled HAMT through materials have cases must include (1) an analytical verification, (2) model compar-
been more extensively studied and modeled in recent years, the ison, and (3) experimental validation [94]. In addition, Judkoff and
direction of the field continues toward eliminating simplifying as- Neymark [94,95] have developed and continuously update [95] the
sumptions to model more accurately the physical interactions in NREL BESTEST base case building developed in IEA ECBCS Annex
building-scale applications (e.g., moisture transfer through porous 21, which has been used to validate models from Rode et al. [83],
media [87]). To accurately model HAMT through a building en- Zhang et al. [71], Abadie et al. [51], Feng et al. [68], and software
velope and to capture the moisture buffering effect of different developed for the IEA Annex 41.
interior surface materials, the physical phenomena that each nu-
merical method is modeling (and associated limitations) must be 3.3. Summary of computational methods
well understood. During the moisture sorption process, for exam-
ple, water vapor can be transported into porous materials due Hygrothermal behavior is important to consider in develop-
to vapor pressure differentials and has the capability of condens- ing more accurate building energy simulation models. Numerical
ing with the pores, then moving through the material via other simulation tools enable a relatively fast and low-cost solution to
mechanisms, such as capillary action. Condensed water also has quantify building-scale benefits of materials that exhibit moisture-
a propensity to evaporate under certain conditions (e.g., air veloc- buffering effects and to predict and mitigate potential moisture ac-
ity across a surface [88]). Additionally, moisture concentration and cumulation in the building envelope. For example, the ability to
temperature do not stay uniform throughout the entire space, so predict mold growth potential is an extremely valuable tool com-
surface variation and discretized time and space must be thought- pared to costly repairs. Solutions to this particular problem are ex-
fully considered [89]. As a result, best exemplified in a numeri- plored with WUFI-Bio [96], but the concept applies further to nu-
cal study of earth-based material validated with experimental data, merically predicting condensation in structural layers and reducing
a hygrothermal model that considers coupled HAMT, pore water ventilation rates based off of feedback from RH sensors.
pressure, and water phase changes can yield accurately model hy- IEA Annex 41 explored the development of 17 different simu-
grothermal behavior [90]. Additionally, the researchers used this lation tools contributed by 39 institutions in 19 countries, and re-
validated model to evaluate the sensitivity of common modeling searchers agreed that it is necessary to model the impact of mois-
assumptions for earth-based material. The authors found it impor- ture to ensure accurate whole building energy simulation results
tant to consider the impact of temperature on moisture flux and [89]. The consensus made by top researchers provides further mo-
in-pore vapor mass condensation and evaporation. However, for tivation for the continued development of faster, more accurate
low water permeability materials, simplifying assumptions can be modeling tools.
made with high enough accuracy [90]. In summary, various modeling methods categorized as empir-
The last major consideration that was elucidated by this review ical, semi-empirical, and physics-based have been used to model
was the importance of hysteretic effects, namely the difference in moisture buffering behavior in buildings. Each method has inher-
sorption and desorption isotherms, which are have emerged as im- ent tradeoffs in accuracy, efficiency, and expense, but progress will
portant sorption characteristics to consider when capturing actual entail combining and refining these methods, such as coupling
hygrothermal behavior. However, only one study that was reviewed with CFD software to tie together material-scale and building-
included hysteresis in the modeling methodology. Hysteretic ef- scale zone interactions [70] or evolutionary strategies with multi-
fects (although computationally more expensive) yield better cor- objective searching capabilities [97]. Using advanced numerical
relation to actual building-scale behavior [91]. tools in tandem with comprehensive, coupled models of HAMT,
When modeling the moisture buffering effect of materials in and a solid understanding of the physical processes at the
buildings, 24 h cycles are generally needed before all components building-scale and a need to verify and validate computational ap-
can be assumed to have reached steady-state equilibrium [91]. proaches with benchmark standards, numerical simulation remains
Overall, hygrothermal models can become complex, computation- a powerful tool that will continue to be exploited to quantify mois-
ally expensive, and potentially impossible to converge on a solu- ture buffering effects on building energy consumption.
tion with user-defined acceptable relative error in this time span
due to a variety of inter-related properties. In order to simplify and 4. Research trends and future developments
utilize these models, future development will need to consider the
scope regarding accuracy and time-scale for the built environment, 4.1. Emerging and growing importance of MBV
in order inform change when compared to traditional construction
practices. The results from this meta-analysis indicate not only a grow-
ing interest in understanding and characterizing moisture buffer-
3.2. Numerical benchmarks ing, but also a consensus on the importance of considering mois-
ture buffering effects in building design and operation. As illus-
Given the multitude of different numerical HAMT modeling ap- trated in Fig. 6, 75% of all studies that report the MBV of conven-
proaches that have been introduced to the field in recent years, tional and innovative building materials were published after 2012.
numerical benchmarks based off experimental datasets have been The data in Fig. 6 also suggest that a wider variety of materials are
 B.K. Kreiger and W.V. Srubar III / Energy & Buildings 202 (2019) 109394 13

studs, hemp-lime) materials were explicitly studied to evaluate hy-

grothermal effects on energy consumption. Results show that these
materials can potentially reduce energy use an average of 15%
[4] and, depending on the climate, up to 30% [71].

4.2. Innovative materials

The results from this review indicate that best-performing hy-

groscopic materials for moisture buffering must exhibit high mois-
ture uptake capacities as well as low desorption temperatures.
Studies show that certain materials—especially natural materials—
can exhibit over 30 times the MBV of standard building materials
(e.g., gypsum, plywood, concrete, plasters) [68], and this increase
can impart significant savings in cost and energy to building op-
eration when passive dehumidification through moisture buffer-
ing effects are considered. We find that most natural materials
have moisture buffering effects in the excellent range (see Table 6).
However, one main drawback is that their variability is higher—
expectedly, engineered materials have a more statistically consis-
tent MBV. Earth plasters with the addition of natural fibers and
hempcrete mixes exhibit high moisture buffering values while not
deviating to far from commonly accepted construction materials.
Some more holistic, natural surface finishing products, however,
Fig. 6. Recent studies that characterize and report MBV.
such as natural wax coatings that display high MBVs, can be a
cost-effective retrofit to improve the moisture buffering effects of
existing buildings.
Novel synthetic materials, like superabsorbent hydrogels, are a
being analyzed and studied with respect to their moisture buffer- promising class of materials that may be engineered to exhibit
ing behavior. hyperactive moisture buffering behavior. Superabsorbent polymers
It is anticipated that additional MBV characterization studies— (SAPs) are crosslinked networks of ultra-hydrophilic polymers that
especially studies that investigate the MBV of non-conventional can absorb up to 10 0,0 0 0% of their dry weight in aqueous so-
building materials—will remain a primary interest (and need) of lutions [99]. The ability of the polymer to absorb fluids is at-
the field. Currently, state-of-the-art energy-efficient building de- tributed to the abundance of hydrophilic functional groups present
sign emphasizes tight, well-insulated, high-performing envelopes. on the polymer backbone, while the crosslinks in SAP networks
The importance of ensuring indoor air quality and proper mois- render the polymer insoluble [100,101]. Commonplace SAPs have
ture management has grown in proportion to implementing been synthesized using ionic acrylate/acrylamide homopolymers
energy-efficient envelope strategies. Tight envelopes reduce infil- and biopolymers, such as alginates, celluloses, and carrageenans
tration from outside air, frequently necessitating ventilation strate- [102]. Theoretically, these SAPs can boost moisture buffering with
gies to ensure sufficient quality of indoor air. Leveraging mois- much less surface area compared to other natural or synthetic
ture buffering—and perhaps designing and exploiting the multi- materials. However, it must be noted that the ultra-high mois-
functional potential (e.g., VOC removal) of innovative, high-MBV ture affinity of SAPs may impart slower desorption, which may re-
materials—would alleviate the frequency of air exchanges in build- duce the ability to buffer moisture when exposed to cycles of high
ings. In addition, failures of these high-performance envelopes and low RH. This phenomenon also makes some superabsorbent
from mold or water damage are costly. Therefore, managing mois- products, like common desiccants, non-ideal as moisture buffering
ture and preventing damage due to moisture accumulation has materials [68].
grown in significance, especially in cold climates that require As illustrated in Table 6, this review elucidated an increased in-
mandatory vapor retardation and where moisture swings are large terest in testing more innovative materials. Table 6 lists the MBV of
due to significant heating and time spent indoors during the win- innovative materials and categorizes their moisture buffering be-
ter months [98]. havior according to a negligible, limited, good, and excellent scale.
Given that emerging estimates of energy savings from passive New materials range from new composites made from traditional
moisture buffering effects have been positive (and non-trivial), it is building materials (e.g., clay and mortar plasters) with fiber rein-
anticipated that accounting for moisture buffering effects in build- forcement to highly absorbent SAP materials (i.e., sodium polyacry-
ing design and operation will become more established conven- late). Although MIL-100, a metal organic framework, and sodium
tion. As more buildings implement building automation systems polyacrylate, a commercial SAP, do exhibit high moisture buffer-
(BAS), it will be both less expensive and more feasible to include ing values of 15 g/RH/m2 and 9 g/RH/m2 respectively, they are
RH-sensing HVAC systems that account for the thermal conductiv- not yet commonly used as building materials and may not prove
ity and RH variations that occur within hygrothermal materials and cost-effective or viable for achieve adequate interior surface fin-
assemblies. By combining building energy models with new, in- ishes [59,68].
novative materials that exhibit exceptional moisture buffering ca- As expected, the MBVs of these new materials vary signifi-
pacity, additional, passive reductions in building energy consump- cantly. On one hand, studies of the carnauba wax particle, a nat-
tion could be realized. Studies with hemp-lime materials, for ex- ural, hydrophilic material that traps water on its rough surface,
ample, show a 5–30% cooling load reduction when using BIM to in- show it can boost MBV of traditional lacquered spruce board from
form HVAC systems [62]. Similarly, RH-sensing ventilation systems 0.3 g/RH/m2 to 1.1 g/RH/m2 [58]. On the other hand, adding
were shown to reduce ventilation 30–40% and energy consump- wetland phytomass in the form of typha or wool chips does not
tion 12–17% during cooling seasons [7]. In addition, full-scale wall necessarily improve the MBV above the 0.3 g/RH/m2 value base-
assemblies with various building materials (e.g., concrete, wood line of plain clay plaster [56], yet while adding olive fibers can in-
14 B.K. Kreiger and W.V. Srubar III / Energy & Buildings 202 (2019) 109394

Table 6
Summary of MBVs for non-conventional, innovative building materials.
CSP = Clay, Sand, Plaster.

Material MBV [g/RH/m2 ] Performance Category Reference

CSP + 2 wt% Typha Wool 0.32 Limited [56]

CSP + 2 wt% Typha Chips 0.32 Limited [56]
CSP + 1 wt% Typha Wool and 0.5 wt% Chips 0.38 Limited [56]
Carnauba Wax 1.085 Good [58]
Linseed Oil 0.7 Moderate [58]
Thick Wax Film 0.1 Negligible [58]
Sodium Polyacrylate 8.97 Excellent [59]
Cellulose 3.07 Excellent [59]
Bamboo Fiberboard 1.7 Good [53]
Bamboo Fiberboard (70%) and Bone Glue (30%) 3.2 Excellent [53]
Bamboo (70%) and Sodium Lignosulfonate (30%) 3.5 Excellent [53]
Rape-Straw-Lime Concrete 2.44 Excellent [30]
Flax-Lime Concrete 2.27 Excellent [30]
Hemp-Lime Concrete 2.02 Excellent [30]
Mortar + Date Palm Fiber 2.97 Excellent [66]
Clay Plaster + Olive Fibers 1.739 Good [67]
MIL-100 (Fe) 15 Excellent [68]

find application in buildings as moisture buffers. While plants pro-

vide carbon dioxide (CO2 ) and esthetic benefit to indoor environ-
mental quality, plants also evapotranspire and are well known to
add high humidity loads to the air [9,103–107]. Other benign living
organisms, such as lichen or moss, may have a potential to serve a
moisture buffering—and perhaps many other—beneficial functions
to indoor environments. Such beneficial functions beyond moisture
buffering could include carbon storage and sequestration, aesthet-
ics, occupant comfort, improved indoor air quality (e.g., VOC se-
questration) and an ability to indicate toxic levels of environmental
pollutants [108].
As one example, lichens possess exceptional moisture absorp-
tive properties and abilities to sequester CO2 and trap harmful
pollutants. Lichen are not plants but rather a symbiotic organism
comprised of fungi and algae and/or cyanobacteria. Since lichen
are not plants, they do not evapotranspire nor impart excess wa-
ter to the air that can increase the moisture load and contribute
to more health problems and mold potential. Regarding mois-
ture capacity, cyanobacteria-based gel lichens are able to absorb
20 0 0% water compared to their dry weight [109,110]. These in-
door water sorption properties are especially useful in cold cli-
mates where increased time spent indoors leads to higher mois-
ture loads. In addition to potential moisture sorption properties,
lichens are known for their interaction with airborne pollutants.
Fig. 7. Comparison between biotic, hybrid, and abiotic material moisture buffering
Studies show that lichen are the best accumulator of polyaromatic
effects.
hydrocarbons (PAHs) that are detrimental to indoor air quality and
found in high concentrations during wet winter conditions [111].
Because of this, the extraction of metals and organic pollutants in
lichen show that it can serve as a bio-indicator of pollutant con-
crease the MBV to a value of 1.7 g/RH/m2 [67]. An improvement
centrations that directly related to indoor environmental quality.
in MBV is expected, given the moisture affinity of natural fibers, as
Furthermore, a specific study of the species Ramalina maciformis
previously discussed, but this result is not always the case.
also shows photosynthesis will increase with increasing water con-
In fact, Fig. 7 shows that on average, natural materials have
tent, thereby increasing the uptake of CO2 [112]. It must be noted,
higher moisture buffering effects (1.75 g/RH/m2 ) than abiotic ma-
however, that, at a certain maximum water content, CO2 uptake is
terials (1.075 g/RH/m2 ). In distinguishing abiotic versus biotic ma-
inhibited, but these levels are not representative of indoor RH but
terials; wood, organic fibers, starches, earth and clay plasters,
of those found in heavy rain [111].
and plant derivatives were considered biotic materials. Concretes,
While true application of living materials may be more than
bricks, and gypsum, and other inorganics were considered abiotic
a decade away, the majority of new materials that are emerg-
materials, and any combination, including finish coatings (paints
ing in the field include composite formulations of natural, hy-
and lacquers) were viewed as a hybrid. With a 63% increase in
drophilic materials, such as olive fibers, rape straw, date palm
MBV on average when using natural, bio-based materials, explo-
fibers, flax, and hemp, with inorganic matrices, like earthen clay
ration of more organic materials will likely be the focus of future
plasters. However, more innovative materials, such as bamboo and
moisture buffering studies.
wax particles, have yielded MBVs in the excellent range and will
Given the high MBVs of biotic materials, living organisms are
likely garner more attention in future studies.
another emerging category of highly innovative materials that may
 B.K. Kreiger and W.V. Srubar III / Energy & Buildings 202 (2019) 109394 15

4.3. Advances in building energy modeling similar fashion in which thermal fluctuations are controlled with
the thermal buffering effects of insulating materials—has clearly
In 2008, the IEA ECBCS Annex 41 identified the major prob- demonstrated non-trivial building energy savings potential. These
lems yet to be solved in modeling heat, air, and moisture trans- findings highlight grand opportunities not only to develop and
fer [113]. The final remarks emphasized (1) establishing a balance characterize novel materials with hyperactive MBVs, but also to
between physical phenomena, (2) incorporating multi-dimensional fully understand, leverage, and quantify (e.g., measure, model) the
and/or transient affects, and (2) using distributed calculations for effects of moisture buffering effects at the assembly and building
computational efficiency. From 2008 to the present, advances have scale with improved accuracy.
been made to address some of these significant problems, while More specifically, the results of this review highlight a clear,
some remain to be solved in relation to moisture buffering. future trend in materials research, development, and characteriza-
Balancing physical phenomena (i.e., heat, air, and moisture) has tion on porous, natural, biotic (i.e., biological), and/or chemically
been addressed in multiple ways by coupling software [77] and hydrophilic (e.g., superabsorbent polymers, carnauba wax) mate-
simultaneously solving heat and moisture transfer equations [90]. rials that exhibit high moisture buffering values (MBVs). In ad-
Coupled computational fluid dynamics (CFD) software with numer- dition, improving the speed of HAMT modeling, while maintain-
ical models for solving heat and moisture transfer can now connect ing physical representativeness, will better inform building design,
building elements with indoor space [113], whereas previous mod- operation, and/or retrofit when the effects of moisture buffering
els made the common assumption that interior spaces act as one are considered. These numerical models have the potential to be
well-mixed zone. The well-mixed zone assumption, while physi- coupled with commercially available computational fluid dynamics
cally inaccurate, has be improved with CFD [77], thereby balancing software to improve assumptions and increase overall accuracy in
physical phenomena instead of focusing specifically on building el- their prediction of hygrothermal performance and energy savings
ement interactions with heat and moisture. potential.
Incorporating multi-dimensional and transient effects can now In summary, it is evident from the literature reviewed herein
be considered, but these effects are not always included due to that moisture buffering can yield tangible benefits by proactively
precedent and computational expense. Models can incorporate considering (and designing for) the moisture buffering effects of
multi-dimensional heat transfer and extend heat flux to a mois- hygroscopic materials. Future developments in materials, measure-
ture flux analogy to include multi-dimensional moisture transfer ment techniques, and modeling approaches will only maximize the
effects through building envelopes [114]. However, transient effects benefits of moisture buffering to reduce building energy consump-
(e.g., transient material properties, increased moisture loads in the tion and further improve indoor environmental quality.
first two years of construction) are not often considered. This cur-
rent omission should be evaluated in the future for its importance,
Declaration of Competing Interest
as moisture cycling is well-known to degrade materials. In addi-
tion, many other factors, like abrasion, hysteresis, and damage, can
The authors declare no conflicts of interest in the design, con-
alter the time-dependent hygroscopic properties of materials that
duct, and reporting of this work.
are originally included and assumed pristine in the modeling pro-
cess.
The IEA ECBCS Annex 41 also elucidated a need for niche pro- Acknowledgments
grams that address specialized concerns related to moisture buffer-
ing in buildings (e.g., furniture-scale mold problems). This dis- This research was made possible by the Department of Civil,
tributed calculation method would save computational time and Environmental, and Architectural Engineering and the Living Mate-
model complexity, allowing for more advanced modeling tech- rials Laboratory (LMLab) at the University of Colorado Boulder with
niques to be employed (e.g., 3D heat and moisture transfer, soft- financial support from the College of Engineering and Applied Sci-
ware coupling, degradation-induced transient material properties) ences Multifunctional Materials Interdisciplinary Research Theme
on a variety of scales. However, distributed modeling inputs would (IRT) Seed Grant Program. This work represents the views of the
need to be based off full-scale modeling results to capture realistic authors and not necessarily those of the sponsors.
physical phenomena—an effort that would be a significant contri-
bution to the field. Supplementary material

5. Conclusions Supplementary material associated with this article can be

found, in the online version, at doi:10.1016/j.enbuild.2019.109394.
This review highlights the materials, methods, and modeling
methodologies related to the moisture buffering effect in buildings References
and its potential impacts on human health, indoor environmental
quality, energy consumption, and occupant comfort. A multitude of [1] Energy Information Administration, Annual Energy Review, (2017). https:
materials in recent years have been characterized with respect to //www.eia.gov/totalenergy/data/annual/index.php (accessed November 17,
2018).
their moisture buffering capacities. While the most commonly em- [2] Office of Energy Efficiency & Renewable Energy, HVAC, Water Heat-
ployed NORDTEST method has emerged as the primary standard ing, and Appliances, (2018). https://www.energy.gov/eere/buildings/
for materials characterization of moisture buffering value (MBV), hvac- water- heating- and- appliances (accessed November 17, 2018).
[3] N. Mendes, F.C. Winkelmann, R. Lamberts, P.C. Philippi, Moisture effects on
differences in experimental facilities, material heterogeneity, and
conduction loads, 35 (2003) 631–644.
operator error remain sources of measurement error for reported [4] A.S. Nordby, A.D. Shea, Building materials in the operational phase, Impact Di-
MBVs. Multiple sources indicate that operator error is the principal rect Carbon Exch. Hygrotherm. Eff. 17 (2013) 763–776, doi:10.1111/jiec.12046.
[5] O.F. Osanyintola, C.J. Simonson, Moisture buffering capacity of hygroscopic
source of MBV variability. Materials-scale characterization studies
building materials: experimental facilities and energy impact, Energy Build.
of MBV are by far more common than assembly- or building-scale 38 (2006) 1270–1282, doi:10.1016/j.enbuild.2006.03.026.
moisture buffering experimentation. However, results from the [6] O. Adan, H. Brocken, J. Carmeliet, H. Hens, S. Roels, C.-.E. Hagentoft, De-
limited large-scale experiments and numerical studies conducted termination of liquid water transfer properties of porous building materials
and development of numerical assessment methods : introduction to the ec
to-date report that passively controlling RH fluctuations through hamstad project, J. Therm. Env. Bldg. Sci. 27 (2004) 253–260, doi:10.1177/
the moisture buffering capacities of hygroscopic materials—in a 1097196304042323.
16 B.K. Kreiger and W.V. Srubar III / Energy & Buildings 202 (2019) 109394

[7] M. Woloszyn, T. Kalamees, M. Olivier, M. Steeman, A. Sasic, The effect of com- [40] M. Rahim, O. Douzane, A.D.T. Le, G. Promis, T. Langlet, Charac-
bining a relative-humidity-sensitive ventilation system with the moisture- terization and comparison of hygric properties of rape straw con-
buffering capacity of materials on indoor climate and energy efficiency of crete and hemp concrete, Constr. Build. Mater. 102 (2016) 679–687.
buildings, 44 (2009) 515–524. doi:10.1016/j.buildenv.2008.04.017. doi:10.1016/j.conbuildmat.2015.11.021.
[8] T. Lu, X. Lu, M. Viljanen, Moisture and estimation of moisture generation rate, [41] R. Peuhkuri, C. Rode, Moisture buffer value: Analytical determination and use
Intech. 2 (2018) 25, doi:10.5772/32009. of dynamic measuremets, Technical University of Denmark, 2005.
[9] J. Christian, Moisture Control in Buildings: The Key Factor in Mold Prevention, [42] I. Gómez, S. Guths, R. Souza, J.A. Millan, K. Martín, J.M. Sala, Moisture buffer-
1st ed., American Society for Testing and Materials, Philadelphia, 1994. ing performance of a new pozolanic ceramic material: influence of the film
[10] Lieff, R. Trechsel, Moisture Migration in Buildings, 1st ed., American Society layer resistance, Energy Build. 43 (2011) 873–878, doi:10.1016/j.enbuild.2010.
for Testing and Materials, Philadelphia, 1982. 12.007.
[11] D. Erhardt, M. Mecklenburg, Relative humidity re-examined, Stud. Conserv. 39 [43] Y. Wu, G. Gong, C. Wah, Z. Huang, Proposing ultimate moisture buffering
(1994) 32–38, doi:10.1179/sic.1994.39.Supplement-2.32. value (UMBV) for characterization of composite porous mortars, Constr. Build.
[12] Q. Wu, C. Piao, Thickness swelling and its relationship to internal bond Mater. 82 (2015) 81–88. doi:10.1016/j.conbuildmat.2015.02.058.
strength loss of commercial oriented strandboard, For. Prod. J. 49 (1999) 50. [44] J. Kurnitski, T. Kalamees, J. Palonen, L. Eskola, O. Seppanen, Potential effects
[13] American Wood Council, National Design Specification for Wood Construction of permeable and hygroscopic lightweight structures on thermal comfort and
- 2018, American National Standards Institute, Leesburg, VA, 2018. perceived IAQ in a cold climate, Indoor Air 17 (2007) 37–49, doi:10.1111/j.
[14] S. Jacobsen, E.J. Sellevold, SELF healing of high strength concrete after deteri- 1600-0668.2006.00447.x.
oration by freeze / thaw, Cem. Concr. Res. 26 (1996) 55–62. [45] A. Holm, K. Lengsfeld, M. Krus, Moisture-Buffering Effect — Experimental In-
[15] P.L. Gaspar, J. De Brito, Quantifying environmental effects on cement-rendered vestigations and Validation, in: 6th Int. Conf. Indoor Air Qual. Vent. Energy
facades : a comparison between different degradation indicators, Build. Envi- Conserv. Build., Tohoku University Press, Holzkirchen, Germany, 2007: pp.
ron. 43 (2008) 1818–1828, doi:10.1016/j.buildenv.2007.10.022. 1011–1018.
[16] N. Lucas, C. Bienaime, C. Belloy, M. Queneudec, F. Silvestre, J. Nava-saucedo, [46] C.J. Simonson, M. Salonvaara, T. Ojanen, Indoor air and a permeable
Polymer biodegradation : mechanisms and estimation techniques, Chemo- and hygroscopic building envelope, J. Build. Phys. 28 (2004), doi:10.1177/
sphere. 73 (2008) 429–442, doi:10.1016/j.chemosphere.2008.06.064. 1097196304044395.
[17] F. Debieb, L. Courard, S. Kenai, R. Degeimbre, Mechanical and durability prop- [47] L. Shi, H. Zhang, Z. Li, X. Man, Y. Wu, C. Zheng, J. Liu, Analysis of moisture
erties of concrete using contaminated recycled aggregates, Cem. Concr. Com- buffering effect of straw-based board in civil defence shelters by field mea-
pos. 32 (2010) 421–426, doi:10.1016/j.cemconcomp.2010.03.004. surements and numerical simulations, Build. Environ. 143 (2018) 366–377,
[18] W. Srubar, An analytical model for predicting the freeze-thaw durability of doi:10.1016/j.buildenv.2018.07.018.
wood-fiber composites, Compos. Part B Eng. 69 (2015) 435–442, doi:10.1016/ [48] H. Zhang, H. Yoshino, K. Hasegawa, J. Liu, W. Zhang, H. Xuan, Practical mois-
j.compositesb.2014.10.1015. ture buffering effect of three hygroscopic materials in real-world conditions,
[19] H. Viitanen, J. Vinha, K. Salminen, T. OJanen, R. Peuhkuri, L. Paajanen, Energy Build. 139 (2017) 214–223, doi:10.1016/j.enbuild.2017.01.021.
K. Lahdesmaki, Moisture and bio-deterioration risk of building materials and [49] H.J. Moon, S.H. Ryu, J.T. Kim, The effect of moisture transportation on energy
structures, J. Build. Phys. 33 (2010) 33, doi:10.1177/1744259109343511. efficiency and IAQ in residential buildings, Energy Build. 75 (2014) 439–446,
[20] S. Klaus, M. Krus, K. Breuer, Mould Growth Prediction with a New Biohy- doi:10.1016/j.enbuild.2014.02.039.
grothermal Method and its Application in Practice, Fraunhofer-Institute for [50] W.V. S.rubar, C.W. Frank, S.L. Billington, Modeling the kinetics of water trans-
Building Physics Holzkirchen, 2003 Holzkirchen, Germany. port and hydroexpansion in a lignocellulose-reinforced bacterial copolyester,
[21] J.Y. Bergen, House air, Science 34 (1911) 407–408. Polymer (Guildf). 53 (2012) 2152–2161, doi:10.1016/j.polymer.2012.03.036.
[22] W. Kent, W. Crowell, L. Jones, The air we breathe, Science 33 (1911) 486–491. [51] M.O. Abadie, K.C. Mendonça, Moisture performance of building materi-
[23] L.R. Ingersoll, Indoor air, Science XXXVII (1913) 524–525. als: from material characterization to building simulation using the mois-
[24] L.M. Reinikainen, J.J.K. Jaakkola, Significance of humidity and temperature on ture buffer value concept, Build. Environ. 44 (2009) 388–401, doi:10.1016/j.
skin and upper airway symptoms, Indoor Air. 13 (2003) 344–352. buildenv.2008.03.015.
[25] C.G. Bornehag, J. Sundell, S. Bonini, A. Custovic, P. Malmberg, S. Skerfving, [52] T. Colinart, D. Lelievre, P. Glouannec, Experimental and numerical analysis of
T. Sigsgaard, A. Verhoeff, Dampness in buildings as a risk factor for health the transient hygrothermal behavior of multilayered hemp concrete wall, En-
effects, EUROEXPO : a multidisciplinary review of the literature (1998 – 20 0 0) ergy Build. 112 (2016) 1–11, doi:10.1016/j.enbuild.2015.11.027.
on dampness and mite exposure in buildings and health effects, Indoor Air. [53] D.M. Nguyen, A.C. Grillet, T.M.H. Diep, C.N. Ha Thuc, M. Woloszyn, Hy-
14 (, 4 ) (2004) 243–257, doi:10.1111/j.1600-0668.2004.00240.x. grothermal properties of bio-insulation building materials based on bamboo
[26] ASHRAE, ANSI/ASHRAE Standard 62.1-2013, American National Standards In- fibers and bio-glues, Constr. Build. Mater. 155 (2017) 852–866, doi:10.1016/j.
stitute, Atlanta, GA, 2013. conbuildmat.2017.08.075.
[27] A. Datta, R. Suresh, A. Gupta, D. Singh, P. Kulshrestha, Indoor air quality of [54] N.M.M. Ramos, J.M.P.Q. Delgado, V.P. De Freitas, Influence of finishing coatings
non-residential urban buildings in Delhi, India, Int. J. Sustain. Built Environ. 6 on hygroscopic moisture buffering in building elements, Constr. Build. Mater.
(2017) 412–420, doi:10.1016/j.ijsbe.2017.07.005. 24 (2010) 2590–2597, doi:10.1016/j.conbuildmat.2010.05.017.
[28] S. Jing, B. Li, M. Tan, H. Liu, Impact of relative humidity on thermal comfort [55] M. Palumbo, F. McGregor, A. Heath, P. Walker, The influence of two crop by-
in a warm environment, Indoor Built Environ 22 (2012) 598–607, doi:10.1177/ products on the hygrothermal properties of earth plasters, Build. Environ. 105
1420326X12447614. (2016) 245–252, doi:10.1016/j.buildenv.2016.06.004.
[29] C. Rode, R. Peuhkuri, L.H. Mortensen, K.K. Hansen, A. Gustavsen, T. Ojanen, [56] M. Maddison, T. Mauring, K. Kirsimäe, Ü. Mander, The humidity buffer capac-
J. Ahonen, K. Svennberg, L.-.E. Harderup, J. Arfvidsson, Moisture Buffering ity of clay-sand plaster filled with phytomass from treatment wetlands, Build.
of Building Materials, Nordisk Innovations Center, 2005 http://orbit.dtu.dk/ Environ. 44 (2009) 1864–1868, doi:10.1016/j.buildenv.20 08.12.0 08.
fedora/objects/orbit:75984/datastreams/file_2415500/content. [57] A. Arrigoni, A.C. Grillet, R. Pelosato, G. Dotelli, C.T.S. Beckett, M. Woloszyn,
[30] M. Rahim, O. Douzane, A.D.T. Le, G. Promis, B. Laidoudi, A. Crigny, B. Dupre, D. Ciancio, Reduction of rammed earth’s hygroscopic performance under sta-
T. Langlet, Characterization of flax lime and hemp lime concretes : hygric bilisation: an experimental investigation, Build. Environ. 115 (2017) 358–367,
properties and moisture buffer capacity, Energy Build 88 (2015) 91–99, doi:10.1016/j.buildenv.2017.01.034.
doi:10.1016/j.enbuild.2014.11.043. [58] A. Lozhechnikova, K. Vahtikari, M. Hughes, M. Österberg, Toward energy ef-
[31] K. Svennberg, K. Lengsfeld, Previous experimental studies and field measure- ficiency through an optimized use of wood: the development of natural hy-
ments on moisture buffering by indoor surface materials, J. Build. Phys. 30 drophobic coatings that retain moisture-buffering ability, Energy Build. 105
(2007) 30, doi:10.1177/1744259107073221. (2015) 37–42, doi:10.1016/j.enbuild.2015.07.052.
[32] T. Padfield, L.A. Jensen, Humidity buffering of building interiors by absorbent [59] S. Cerolini, M. D’Orazio, C. Di Perna, A. Stazi, Moisture buffering capacity of
materials, Conserv. Phys. 1(2010) 475-482. highly absorbing materials, Energy Build. 41 (2009) 164–168, doi:10.1016/j.
[33] JIS (Japanese Industrial Standard), A1470: Test method of Adsorp- enbuild.20 08.08.0 06.
tion/Desorption Efficiency For Building Materials to Regulate an Indoor [60] Z. Chen, M. Qin, Preparation and hygrothermal properties of composite phase
Humidity - part 1: Response Method of Humidity, JIS, Tokyo, Japan, 2002. change humidity control materials, Appl. Therm. Eng. 98 (2016) 1150–1157,
[34] ISO (International Standards Organization), ISO 24353: Hygrothermal per- doi:10.1016/j.applthermaleng.2015.12.096.
formance of Building Materials and Products - determination of Moisture [61] Z. Chen, M. Qin, J. Yang, Synthesis and characteristics of hygroscopic
Adsorption/Desorption Properties in Response to Humidity Variation., ISO, phase change material: composite microencapsulated phase change mate-
Geneva, Switzerland, 2008. rial (MPCM) and diatomite, Energy Build. 106 (2015) 175–182, doi:10.1016/
[35] DIN Deutsches Institut fur Normung, Earth plasters—Terms and definitions, j.enbuild.2015.05.033.
requirements, test methods. DIN 18947, Berlin, Germany, 2013. [62] E. Latif, M. Lawrence, A. Shea, P. Walker, Moisture buffer potential of exper-
[36] H. Kunzel, in: Die Feuchtigkeitsabsorption von Innenoberflachen und imental wall assemblies incorporating formulated hemp-lime, Build. Environ.
Inneneinrichtungen, Berichte Aus Der Bauforsch., 1965, pp. 102–116. 93 (2015) 199–209, doi:10.1016/j.buildenv.2015.07.011.
[37] J. Lima, P. Faria, Eco-efficient earthen plasters: the influence of the addi- [63] F. Mcgregor, A. Heath, A. Shea, M. Lawrence, The moisture buffering capacity
tion of natural fibers, RILEM Bookseries 12 (2016) 315–327, doi:10.1007/ of un fi red clay masonry, Build. Environ. 82 (2014) 599–607, doi:10.1016/j.
978- 94- 017- 7515- 1_24. buildenv.2014.09.027.
[38] F. McGregor, A. Heath, D. Maskell, A. Fabbri, J. Morel, A review on the buffer- [64] F. Collet, J. Chamoin, S. Pretot, C. Lanos, Comparison of the hygric behaviour of
ing capacity of earth building materials, Constr. Mater. 169 (2016) 169. three hemp concretes, Energy Build. 62 (2013) 294–303, doi:10.1016/j.enbuild.
[39] S. Roels, H. Janssen, A comparison of the nordtest and moisture buffering per- 2013.03.010.
formance, J. Build. Phys. (2006) 30, doi:10.1177/1744259106068101. [65] M. Palumbo, A.M. Lacasta, N. Holcroft, A. Shea, P. Walker, Determination
of hygrothermal parameters of experimental and commercial bio-based in-
 B.K. Kreiger and W.V. Srubar III / Energy & Buildings 202 (2019) 109394 17

sulation materials, Constr. Build. Mater. 124 (2016) 269–275, doi:10.1016/j. [88] P. Plathner, M. Woloszyn, Interzonal air and moisture transport in a test
conbuildmat.2016.07.106. house : experiment and modelling, Build. Environ. 37 (2002) 189–199.
[66] N. Chennouf, B. Agoudjil, A. Boudenne, K. Benzarti, F. Bouras, Hygrothermal doi:https://doi.org/10.1016/S0360-1323(0 0)0 0 096-2.
study of mortar with date palm fiber reinforcement, in: AIP Conf. Proc. 1988, [89] M. Woloszyn, C. Rode, Tools for performance simulation of heat, air
American Institute of Physics, 2018. doi:10.1063/1.5047607. and moisture conditions of whole buildings, Build Simul. 1 (2008) 5–24.
[67] S. Liuzzi, C. Rubino, P. Stefanizzi, A. Petrella, A. Boghetich, C. Casavola, G. Pap- doi:10.10 07/s12273-0 08-8106-z.
palettera, Hygrothermal properties of clayey plasters with olive fibers, Constr. [90] L. Soudani, A. Fabbri, J. Morel, M. Woloszyn, P. Chabriac, H. Wong, A. Gril-
Build. Mater. 158 (2018) 24–32, doi:10.1016/j.conbuildmat.2017.10.013. let, Assessment of the validity of some common assumptions in hygrother-
[68] X. Feng, M. Qin, S. Cui, C. Rode, Metal-organic framework MIL-100 (Fe) as a mal modeling of earth based materials, Energy Build 116 (2016) 498–511,
novel moisture bu ff er material for energy-e ffi cient indoor humidity con- doi:10.1016/j.enbuild.2016.01.025.
trol, Build. Environ. 145 (2018) 234–242, doi:10.1016/j.buildenv.2018.09.027. [91] G. Promis, O. Douzane, A.D. Tran Le, T. Langlet, Moisture hysteresis influence
[69] A. Arrigoni, C.T.S. Beckett, D. Ciancio, R. Pelosato, G. Dotelli, A. Grillet, Re- on mass transfer through bio-based building materials in dynamic state, En-
sources, conservation & recycling rammed earth incorporating recycled con- ergy Build. 166 (2018) 450–459, doi:10.1016/j.enbuild.2018.01.067.
crete aggregate : a sustainable, resistant and breathable construction solu- [92] M. Woloszyn, C. Rode, Common Exercises in Whole Building HAM Modeling,
tion, Resour. Conserv. Recycl. 137 (2018) 11–20, doi:10.1016/j.resconrec.2018. in: Proc. IEA ECBCS Annex 41 Closing Semin., Technical University of Den-
05.025. mark, Copenhagen Denmark, 2008: pp. 37–48.
[70] H. Zhang, H. Yoshino, A. Iwamae, K. Hasegawa, Investigating simultane- [93] C.-.E. Hagentoft, HAMSTAD - Final report: methodology of HAM-modeling, Re-
ous transport of heat and moisture in hygroscopic materials by a semi- port R-02:8, Gothenburg, Chalmers University of Technology, 2002.
conjugate CFD-coupled approach, Build. Environ. 90 (2015) 125–135, doi:10. [94] R. Judkoff, J. Neymark, Building energy simulation test (BESTEST) and diag-
1016/j.buildenv.2015.03.028. nostic method, NREL/TP-472-6231, NREL, Golden, CO, 1995.
[71] M. Zhang, M. Qin, C. Rode, Z. Chen, Moisture buffering phenomenon and its [95] R. Judkoff, J. Neymark, Twenty years on!: Updating the IEA BESTEST building
impact on building energy consumption, Appl. Therm. Eng. 124 (2017) 337– thermal fabric test cases for ASHRAE standard 140 NREL/CP - 5500-58487,
345, doi:10.1016/j.applthermaleng.2017.05.173. Chambery, France, 2013.
[72] C. Feng, H. Janssen, Y. Feng, Q. Meng, Hygric properties of porous building [96] K. Holzhueter, K. Itonaga, An evaluation of WUFI-Bio to predict mold growth
materials : analysis of measurement repeatability and reproducibility, Build. in straw bale walls in Japan, J. Asian Archit. Build. Eng. 16 (2017) 357–362.
Environ. 85 (2015) 160–172, doi:10.1016/j.buildenv.2014.11.036. doi:https//doi.org/10.3130/jaabe.16.357.
[73] S. Roels, P. Talukdar, C. James, C.J. Simonson, Reliability of material data mea- [97] S. Rouchier, M. Woloszyn, Y. Kedowide, T. Béjat, Identification of the hy-
surements for hygroscopic buffering, Int. J. Heat Mass Transf. 53 (2010) 5355– grothermal properties of a building envelope material by the covariance ma-
5363, doi:10.1016/j.ijheatmasstransfer.2010.07.020. trix adaptation evolution strategy, J. Build. Perform. Simul. 9 (2016) 1-14.
[74] D. Maskell, A. Thomson, P. Walker, M. Lemke, Determination of optimal plas- doi:10.1080/19401493.2014.996608.
ter thickness for moisture bu ff ering of indoor air, Build. Environ. 130 (2018) [98] ICC (International Code Council), Residential energy efficiency, 2018 IECC (In-
143–150, doi:10.1016/j.buildenv.2017.11.045. ternational Energy Conserv. Code), ICC, Washington, DC, 2018.
[75] R. Kramer, J. van Schijndel, H. Schellen, Simplified thermal and hygric building [99] M.J. Zohuriaan-Mehr, K. Kabiri, Superabsorbent polymer materials, Iran,
models: a literature review, Front. Archit. Res. 1 (2012) 318–325, doi:10.1016/ Polym. J. (2008) 451–477 http://journal.ippi.ac.ir.
j.foar.2012.09.001. [100] E.M. Ahmed, Hydrogel: preparation, characterization, and applications: a re-
[76] M. Barclay, N. Holcroft, A.D. Shea, Methods to determine whole building hy- view, J. Adv. Res. 6 (2015) 105–121, doi:10.1016/j.jare.2013.07.006.
grothermal performance of hemp-lime buildings, Build. Environ. 80 (2014) [101] F.L. Buchholz, A.T. Graham, Modern Superabsorbent Polymer Technology,
204–212, doi:10.1016/j.buildenv.2014.06.003. Wiley-VCH, 1998.
[77] H.J. Steeman, A. Janssens, J. Carmeliet, M. De Paepe, Modelling indoor [102] E. Feng, G. Ma, Y. Wu, H. Wang, Z. Lei, Preparation and properties of organic-
air and hygrothermal wall interaction in building simulation : cCompar- inorganic composite superabsorbent based on xanthan gum and loess, Carbo-
ison between cfd and a well-mixed zonal model, 44 (2009) 572–583. hydr. Polym. 111 (2014) 463–468, doi:10.1016/j.carbpol.2014.04.031.
doi:10.1016/j.buildenv.20 08.05.0 02. [103] T. Kalamees, J. Vinha, J. Kurnitski, Indoor humidity loads and moisture pro-
[78] J. Woods, J. Winkler, Field measurement of moisture-buffering model inputs duction in lightweight timber-frame detached houses, J. Build. Phys. 29
for residential buildings, Energy Build 117 (2016) 91–98, doi:10.1016/j.enbuild. (2006) 219–246, doi:10.1177/1744259106060439.
2016.02.008. [104] W.J. Angell, Home moisture sources, CD-FS-3396, University of Minnesota,
[79] J. Woods, J. Winkler, Effective moisture penetration depth model for resi- 1988.
dential buildings: sensitivity analysis and guidance on model inputs, Energy [105] CIBSE, Guide A: Environmental design, Chartered Institution of Building Ser-
Build. 165 (2018) 216–232, doi:10.1016/j.enbuild.2018.01.040. vices Engineers, London, UK,1999.
[80] J. Woods, J. Winkler, D. Christensen, Moisture modeling: effective moisture [106] A.P. Koch, B. Kvisgaard, J. Larsen, T. Nielsen, Fugt i boligen, Danish Technolog-
penetration depth versus effective capacitance, in: Proc. Int. Conf. Therm. Per- ical Institute,1986.
form. Exter. Envel. Whole Build. XII, Clearwater, FL, 2013, pp. 216–232. [107] BS 5250, British standard code of practice for control of condensation in
[81] J. Woods, J. Winkler, D. Christensen, Evaluation of the effective moisture pen- buildings, British Standards Institution, London, UK, 1989.
etration depth (EMPD) model for estimating moisture buffering in buildings, [108] R.M. Easton, Lichens and rocks: A review, Geosci. Canada. 21 (1995) 59-76.
NREL, Golden, CO, 2013. doi:10.2172/1219896. [109] Y. Gauslaa, Rain, dew, and humid air as drivers of morphology, func-
[82] H. Kunzel, Simultaneous heat and moisture transport in building compo- tion and spatial distribution in epiphytic lichens, Lichenol. 46 (2018) 1–16.
nents: One-and Two-dimensional calculations using simple paramters, IRB- doi:10.1017/S0 0242829130 0 0753.
Verlag, Holzkirchen, Germany, 1995. [110] P.-.A. Esseen, M. Ronnqvist, Y. Gauslaa, D.S. Coxson, Externally held water -
[83] C. Rode, K. Grau, Moisture buffering and its consequence in whole build- a key factor for hair lichens in boreal forest canopies, Fungal Ecol. 30 (2017)
ing hygrothermal modeling, J. Build. Phys. 31 (2008) 333–360, doi:10.1177/ 29–38, doi:10.1016/j.funeco.2017.08.003.
1744259108088960. [111] L. Van der Wat, P.B.C. Forbes, Lichens as biomonitors for organic air pollu-
[84] H. Kuenzel, A. Karagiozis, A. Holm, A hygrothermal design tool for architects tants, TrAC - Trends Anal. Chem. 64 (2015) 165–172, doi:10.1016/j.trac.2014.
and engineers, in: moisture analysis and condensation control in building en- 09.006.
velopes, ASTM Man. Ser., 50, 2001. [112] O.L. Lange, J.D. Tenhunen, Moisture content and CO2exchange of lichens. II.
[85] B. Seng, S. Lorente, C. Magniont, Scale analysis of heat and moisture transfer Depression of net photosynthesis in Ramalina maciformis at high water con-
through bio-based materials — application to hemp concrete, Energy Build. tent is caused by increased thallus carbon dioxide diffusion resistance, Oe-
155 (2017) 546–558, doi:10.1016/j.enbuild.2017.09.026. cologia. 51 (1981) 426–429, doi:10.10 07/BF0 0540917.
[86] S. Hameury, Moisture buffering capacity of heavy timber structures directly [113] A. Janssens, M. Woloszyn, C. Rode, A. Sasic-Kalagasidis, M. De Paepe, From
exposed to an indoor climate: a numerical study, Build. Environ. 40 (2005) EMPD to CFD – overview of different approaches for heat air and moisture
1400–1412, doi:10.1016/j.buildenv.2004.10.017. modeling in IEA annex 41, in: Proc. IEA ECBCS Annex 41 Closing Semin.,
[87] F. Tariku, K. Kumaran, P. Fazio, International journal of heat and mass trans- Copenhagen, Denmark, 2008.
fer transient model for coupled heat, air and moisture transfer through [114] Fraunhofer IBP, WUFI Plus, (2019). https://wufi.de/en/software/wufi-
multilayered porous media, Int. J. Heat Mass Transf. 53 (2010) 3035–3044, plus/(accessed May 30, 2019).
doi:10.1016/j.ijheatmasstransfer.2010.03.024.