Version of Record: https://www.sciencedirect.

com/science/article/pii/S0306261919307822
Manuscript_56cf37cf33c20458efd52c099cfaf228

Performance evaluation of a metamaterial-based new cool roof using

improved Roof Thermal Transfer Value model

Hong Fang a, Dongliang Zhao b, Jinchao Yuan a, Ablimit Aili b, Xiaobo Yin b, c, Ronggui Yang b,

and Gang Tan a,*

a Department of Civil and Architectural Engineering, University of Wyoming, Laramie, WY

82071, U.S.

b Department of Mechanical Engineering, University of Colorado, Boulder, CO 80309, U. S.

c Materials Science and Engineering Program, University of Colorado, Boulder, CO 80309, U. S.

*Corresponding author: Tel: +1 307 766 2390; E-mail address: [email protected]

© 2019 published by Elsevier. This manuscript is made available under the Elsevier user license
https://www.elsevier.com/open-access/userlicense/1.0/
Abstract:

A new cool roof with potential to generate significant energy savings in buildings has been

developed from a metamaterial film named as RadiCold. Considering that the RadiCold film has

unique optical and thermal characteristics and the current Roof Thermal Transfer Value model

neglecting the effect of roof thermal mass that may lead to overestimating the cooling load from

roofs, this work developed an improved Roof Thermal Transfer Value model and validated the

model for both RadiCold cool roof and traditional roofing structures. Data from the reduced-size

model building experiments showed that the improved Roof Thermal Transfer Value model can

accurately describe the heat gains or losses via the roofs. Under real-world weather conditions in

the United States (Tucson AZ, Los Angeles, CA, and Orlando FL), the improved Roof Thermal

Transfer Value model has been applied to three types of roof exterior finishing: shingle,

Thermoplastic Polyolefin (a cool roof material) and RadiCold. In a typical meteorology year, the

modeling results show that the shingle and Thermoplastic Polyolefin roof transfer 78.9 – 294.1

kWh/(m2·yr) and 8.5 – 128.2 kWh/(m2·yr) of heat into the building space, respectively, but the

RadiCold cool roof dissipates 137.6 – 268.7 kWh/(m2·yr) of heat from the building space to

outdoor environment (e.g., sky). The cooling load reduction from utilizing RadiCold cool roof

results in cooling electricity savings of 113.0 – 143.9 kWh/(m2·yr) compared to the shingle roof

and 88.0 – 92.4 kWh/(m2·yr) compared to the Thermoplastic Polyolefin roof for the three

analyzed locations with an assumed air conditioning system’s coefficient of performance of 3.0.

Key words: Cool roof, RTTV (Roof Thermal Transfer Value), radiative cooling, spectrum

selective, cooling load, energy saving.

NOMENCLATURE

2
A area (m2)

c speed of the light (m/s)

C thermal capacity (J/K)

cp specific heat of roof structure (J/kg K)

ɛ emissivity

Gsc solar constant (W/m2)

h natural convection heat transfer coefficient (W/m2 K)

hp Planck constant

IBB spectral radiance of the blackbody (W/m2)

k thermal conductivity (W/m K)

K thermal conductance of half portion of the roof structure (W/K)

kB Boltzmann constant

n nth day in a year

q heat flux (W/m2)

RH relative humidity (%)

RTTV roof thermal transfer value (W/m2)

s cloud effect parameter

t time (s)

T temperature (°C)

V speed (m/s)

α absorptivity

β angle between the normal direction of the surface and the sun

γ coefficient

3
δ declination angle (°)

θ angular over hemisphere (°)

λ wavelength (µm)

ρ roof structure board density (kg/m3)

φ latitude (°)

ω the sun’s angular deviation from south (°)

σ Stefan-Boltzmann constant

∆x thickness of the slab (m)

Superscript / subscript

amb ambient

atm atmospheric

c ceiling

cl, atm cloudy, atmospheric

con convection

con, in natural convection between the ceiling and indoor air

con, out natural convection between the roof and outdoor air

d dew point

e between the RadiCold surface and the ambient air

ex extraterrestrial

exp experimental data

film RadiCold film

i indoor

4
m average

o outdoor

oe the outer surface of the earth

rad radiation

solar solar

sun sun

wind wind

5
1. Introduction

Energy used by buildings accounts for a large portion of the worldwide energy consumption.

Building envelope components are considered as main sources of space cooling and heating

loads [1-2]. In a study of a low-rise residential building, 25.2% of the energy for the HVAC

system was used to overcome the heat gain from the envelope [3]. Many researches have

investigated new technologies to reduce building envelope related energy consumption, such as

dynamic building envelope with phase change material (PCM) [4], green roofs [5], and hydrogel

as regenerable “sweating skin” for building space cooling [6].

Since roofs (especially the flat and low-slope ones) usually receive more solar irradiation

than facades, reducing the heat gain via roofs can largely reduce the building energy

consumption. One of the major approaches to reduce heat gains through roof structures is

installing cool roof materials which can reflect more sunlight and absorb less heat than a

standard roof [7]. Previous studies have shown that a cool roof is capable of maintaining average

2.5 °C lower than the conventional roof under direct sun exposure [8]. In addition, cool roofs are

also engineered to have higher infrared emittance than the conventional roofs and more

effectively emit longwave radiation to the sky to avoid accumulated heat in the roof structure

entering the building space during night [9].

A large amount of research demonstrated that high surface emission within the atmospheric

window (8 µm < λ < 13 µm) is desirable since terrestrial radiation in this wavelength range is

relatively unaffected by absorption and scattering as it propagates upward through the

atmosphere to deep space [10]. It also has been recommended that widespread use of spectrally

selective surfaces that are characterized by low solar absorptivity and high atmospheric window

6
emissivity can provide a pathway to save energy efficiently [11, 12]. Although technology

breaking-through on metamaterial development has helped achieve sub-ambient cooling under

direct sunlight, the engineered nanophotonic materials haven’t opened an economically feasible

path for wide applications in buildings for energy saving purpose [13].

A recently developed scalable-manufacturing metamaterial (hereafter named RadiCold) film

can highly reflect solar irradiation and radiates infrared heat into space through the atmospheric

window [14]. This metamaterial consists of a 200 nm thick silver coating and a visibly

transparent polymer encapsulating randomly distributed silicon dioxide (SiO2) microspheres. The

encapsulated SiO2 microspheres have optical properties drastically different from that of the

surrounding matrix material at infrared wavelengths, due to the existence of strong phonon-

polariton resonances. The RadiCold film reflects about 96% of solar irradiation while possessing

a high emissivity (greater than 0.93) between 8 and 13 µm wavelengths which yields greater than

93 W/m2 radiative cooling power under direct sunlight [15]. Therefore, this RadiCold film

installed on a roof can achieve passive radiative cooling without electricity consumption during

the day.

When designing or implementing a cool roof, quickly and accurately predicting its thermal

performance is necessary. Roof Thermal Transfer Value (RTTV) model is prescribed for air-

conditioned, non-residential buildings as guidelines for thermal performance of roofs in

numerous Southeast Asia countries [2]. It is a quick tool to quantitatively estimate heat gain or

loss through roofs. Wong et al. used RTTV model to identify the benefits of light-colored roof in

reducing the heat gain into the building [16]. Chaiyapinunt et al. applied RTTV model to

estimate the amount of heat gain through building roofs and control the energy usage in

7
buildings [17]. However, the current RTTV model generally treats the cool roof material with

effective optical parameters (e.g., effective solar reflectance and thermal emittance) throughout

the corresponding spectrum ranges, and thus the current RTTV model may not be accurate to

estimate the heat gain or loss for the spectrum selective materials [18]. Another problem of the

current RTTV model is that it neglects the roof thermal mass effects and may lead to

overestimating of cooling load from the roof structures [19].

In this study, we extended the existing RTTV model into an improved RTTV (iRTTV) model

by incorporating the spectrum selective approach and the thermal mass effect of the roof

structure, which can be used as a tool to evaluate the thermal performance for various types of

roofs. The iRTTV model is validated with experimental measurements from the first completed

reduced-size model building tests to the metamaterial-based cool roof under real weather

conditions located in University of Wyoming, Laramie, Wyoming, USA. Besides the RadiCold

film, two market available materials, Thermoplastic Polyolefin (TPO) and shingle, have been

selected for performance comparison. Finally, by using the iRTTV model, RadiCold film was

applied to commercial building roofs in three locations (Tucson AZ, Los Angeles, CA, and

Orlando FL) to estimate the energy and cost savings.

2. Algorithm to improve the RTTV model

When developing the iRTTV model, the roof structure is simplified as a two-layered roof

that consists of a deck layer and the roof finishing layer (i.e., RadiCold film, TPO or shingle).

Several assumptions are made for the iRTTV model: (1) The contact thermal resistance between

the roof finishing layer and the deck layer is negligible and no air-gap exists between the two

roof layers. (2) The heat flow is perpendicular to the roof. (3) Each layer has a uniform thickness

8
and thermal property. (4) Since the RadiCold film is very thin (50 μm), its thermal mass is

negligible and the temperature of the film is uniform. (5) Since the thermal radiation between the

ceiling and the other indoor surfaces (e.g., walls and floor) is relatively small, this part of the

radiation is neglected.

2.1 The iRTTV model

As shown in Figure 1, the RadiCold cool roof is taken as an example to demonstrate how the

iRTTV model quantifies the heat flux through the roof. The main parameters used in iRTTV

model are illustrated in Figure 1(a). The roof structure can be simplified as a combination of the

thermal resistance and capacitance (RC) like an electric circuit (see Figure 1(b)) [20], where Ts is

the roof surface temperature that equals to the temperature of RadiCold film and Tc denotes the

ceiling surface temperature. For the indoor environment, heat is transferred between the ceiling

and indoor air through natural convection (i.e., qcon,in), which directly contributes to space

cooling load. The Roof Thermal Transfer Value (RTTV) is described as,

= , =ℎ − 1

where, hcon is the natural convection heat transfer coefficient between the ceiling and the indoor

air, and Ti is the indoor air temperature.

9
 Figure 1 (a) Sketch of the heat transfers through a flat deck roof with RadiCold finishing, and

(b) the equivalent RC diagram of the roof structure

2.2 Roof structure thermal model

For the RadiCold film, the energy balance is obtained as [21],

+ − = + , + 2

where, qrad is the radiative power of the RadiCold film, K is the thermal conductance of half

portion of the roof structure, Tm is the average temperature of the roof structure, qatm is the

atmospheric radiation, qcon,out is the convective heat transfer from outdoor air to the RadiCold

film, qsun is the solar irradiance absorbed by the RadiCold film, and A is the area of the RadiCold

cool roof.

Considering both thermal capacity and thermal resistance of the roof structure, a transient

heat transfer equation for the roof structure is given as [20],

= − + − 3

10
where,

2
=
∆"

= #$% ∆"

C refers to the thermal capacity of the roof structure, k is the thermal conductivity of roof

structure material, ∆x is the thickness of the deck, ρ is the density of the roof structure, cp is the

specific heat of roof structure, and t is the time.

For an imaginary ceiling surface without thermal mass, the ceiling surface has an energy

balance as following,

− =ℎ − 4

The natural convection heat transfer coefficient between the indoor air and the ceiling surface

is obtained by [22],

ℎ = 3.1 − (.))
5

Furthermore, the convection heat transfer between the outdoor air and the RadiCold film is

calculated by,

, = ℎ+ − 6

where, To is the outdoor air temperature, he denotes the convective heat transfer coefficient

between the roof exterior finishing (e.g., RadiCold film) and the ambient air, which is obtained

from the following equation [23],

ℎ+ = 8.3 + 2.5 . 7

11
where, Vwind is the wind speed at roof height.

2.3 Roof radiation model

In Equation (2), the radiative power of the RadiCold film is obtained by the angular

integration over the hemisphere [24],

:
9
)
= 20 1 1 233 , 4 56 7 4, 8 4 8 8
( (

where, εfilm is the RadiCold film emissivity.

In Equation (8), the total amount of radiance emitted by a blackbody at wavelength λ and

temperature Ts can be described by the Planck function as [25],

2ℎ% $ ) 1
233 ,4 = 9
4 exp ℎ% $ ⁄ 4
;
3 −1

where, IBB is the spectral radiance of a blackbody at temperature Ts, hp is the Planck constant, c is

the speed of the light, kB is the Boltzmann constant, λ is the wavelength, θ is the angle over the

hemisphere.

The absorbed thermal radiation from the atmosphere is obtained by [26, 27]

:
9
)
= 20 1 1 233 A, 4 56 7 4, 8 5 4, B, 8 4 8 10
( (

where, εatm is the atmospheric emissivity as a function of the radiation wavelength λ, the Relative

Humidity (RH) of the ambient air, and the hemisphere angle, which can be obtained by using the

calculation tool developed in [28]. Equation (10) shows that the atmospheric emissivity (or sky

12
radiance) is significantly dependent on the atmospheric water content, which indicates a high

ambient air relative humidity will reduce the radiative cooling power of the film.

The absorbed solar irradiance qsun is calculated by [29, 30]

= $DEF ∙ 7 ·I 11

where, β is the angle between the normal direction of the surface and the sun (i.e., solar

incidence angle), qsolar is the solar irradiance, and α is the solar absorptance of the surface.

2.4 Cloud model

When there is cloud in the sky, it is expected that the solar irradiation and roof surface

thermal radiation decrease and the atmospheric radiation increases. To account for the cloudiness

and modify the existing clear sky models, the cloud effect model has been incorporated. The

solar irradiation incident on a surface of θ=0 just outside the atmosphere is named as qoe (W/m2),

which is calculated from the solar constant qsc (1,367 W/m2) and the day number (the nth day in a

year) as follows [31]:

360 ∙ K
+ = J1 + 0.033 ∙ $DE L 12
365

Extraterrestrial radiation qex (W/m2) is the solar irradiation incident on the outer surface of

the earth. It is determined by the solar constant qsc and the zenith angle (θz). The latter one is a

function of latitude, time of day, and the day in a year. The extraterrestrial radiation is given by

[32],

+M = + ∙ $DE8N = + ∙ $DEO ∙ $DEP ∙ $DEQ + ERKO ∙ ERKP 13

13
where φ is the latitude, δ is the obliquity of the ecliptic which is equal to [33],

284 + K
P ≈ 23.45 ∙ ERK J360 ∙ L 14
365

ω is the sun’s angular deviation from south which is [33],

Q = 15° ∙ UDVWX RYZ − 12 15

ω is negative before Solar Noon and −180° ≤ Q ≤ 180°.

Once qex is obtained from Equation (13), the cloud effect parameter is estimated by [33],

(.;
+M%
E = J1.4286 − 0.3L 16
+M

where qexp is the obtained solar irradiation either from the pyranometer measurement in

experiments or from TMY3 weather data [34].

The atmospheric emissivity for a cloudy sky is adjusted according to [33]:

5 7, λ, B, θ = 5 λ, B, θ + γ · 1 − 5 λ, B, θ ∙E 17

where, εcl,atm (λ, RH, θ) is the atmospheric emissivity of a cloudy sky, and the coefficient γ=0.9 is

used in this work from experimental validation.

Then, the radiation from the atmosphere under a cloudy sky is obtained,

:
9
)
, = 20 1 1 233 A, 4 56 7 4, 8 5 7, 4, B, 8 4 8 18
( (

14
 Since the solar irradiation is zero at night but the effect of cloud on atmospheric emission still

exists, in this work the cloud effect parameter at night is assumed as the average value of the

daytime for that particular day.

3. Experimental setup for validation of iRTTV model

The experiments for validating the iRTTV model were performed at an open field on the

University of Wyoming campus in Laramie, Wyoming, USA. The site is located far away from

buildings and trees to eliminate potential shadings by surrounding objects. The experimental

platforms consist of two single-space model rooms with the doors facing south, as shown in

Figure 2. The Room A roof was covered by the RadiCold film and the Room B was used for

comparison as a baseline. Both model rooms were built identical except the roof exterior

finishing. The dimension of these two buildings is 2.44 m × 1.83 m × 2.44 m (8 ft × 6 ft × 8 ft)

(Length × Width × Height). In experiments, there were neither heating nor cooling sources inside

the model rooms.

For each model room, eighteen (18) thermocouples were installed to measure the surface and

the air temperatures, as shown in Figure 3. A weather station and a pyranometer were installed

next to the model buildings at roof height to record the real-time weather data such as the

ambient temperature, humidity, local wind speed, wind direction and solar irradiation.

15
Figure 2 The experimental platform consisting of two single - space model rooms

Figure 3 Locations of thermocouples installed to the experiment model rooms

16
 Both model rooms are wood-frame structure and R13 (2.29 m2 K/W) insulation is installed to

the walls and the floor [35]. The ceiling has no insulation. Detailed envelope parameters of the

model room are listed in Table 1.

Table 1 Detailed envelope parameters of the model room

Components Details (from interior to exterior)

2” × 4” wood framing
R13 insulation
Wall 6 mil plastic inside walls
Siding-3/8” Smart side OSB pre-finished
1 ¾” exterior door
2” × 6” wood framing
R13 Insulation
Floor
6 mil Plastic
5/8” OSB decking
2” × 4” wood framing
No Insulation
Roof
5/8” OSB deck
RadiCold film, 80 mil TPO or shingle
Three types of roof exterior finishing (RadiCold, TPO and shingle) were tested in

experiments. The shingle roof made by asphalt is one of the most widely used roofing covers in

the United States because it has a relatively low cost and is fairly simple to install. TPO is a

single – ply reflective roofing membrane made from polypropylene and ethylene – propylene

rubber polymerized together. TPO roof grows fast in commercial roofing membrane on the

market today. The TPO roof used in this study has an excellent Solar Reflectance Index (SRI),

which is an indicator on how well a roof rejects solar heat - higher value meaning a better solar

reflectance. In this work, the SRI value of the TPO was calculated at 99 [36], representing a top

market available cool roof material.

For the three roof finishing materials (RadiCold film, TPO and shingle), Figure 4(a) shows

the optical properties over the whole spectrum from 300 nm to 25 µm which clearly indicates

17
that the RadiCold metamaterial is a spectrum selective material giving extremely excellent

radiative cooling performance with a SRI of 123. Besides the uniquely high solar reflectivity and

high emissivity in atmospheric window (8-13 µm), the RadiCold film also demonstrates a special

better performance in the band of 13-17 µm with a low emissivity since it significantly reduces

the heat transfer from the sky (see Figure 4(a) and 4(b)). On average, the emissivity of the

RadiCold in the band of 13-17 µm is more than 60% lower than that of the shingle and TPO.

Researches show that the reflectance of many materials, especially metamaterials, presents a

strong angular dependence and thus has certain degradation when applied to tilt roofs [37].

However, the RadiCold film is produced by embedding resonant polar dielectric microspheres

randomly in a polymeric matrix [14] and it has preeminent emissivity/reflectivity. As shown in

Figure 4(c), the experiment measurement proves RadiCold film maintains quite high emissivity (>

0.85) in the atmospheric window throughout the tested -60° to 60° range (except two relatively

low values of 0.82 at -20° and 20°). Therefore, the RadiCold film demonstrates minor

degradation with tilt angle.

18
 Figure 4 (a) Optical emissivity of three roof finishing materials, (b) an example clear sky

atmospheric transmittance for 3~50 µm, and (c) Angular dependence of optical emissivity of

RadiCold within 8~13 µm.

4. Validation of the iRTTV model

To validate the iRTTV model, the roof finishing material’s optical properties, the measured

solar irradiation, the outdoor air temperature (To), the indoor air temperature (Ti), and the wind

speed were used as inputs (see Figure 5). The iRTTV model calculated the roof exterior surface

temperature (Ts) and the ceiling surface temperature (Tc), which were then compared against the

measured roof and ceiling temperatures from experiments.

Figure 5 Inputs (To, qsun, wind speed and Ti) and outputs (Ts and Tc) of the iRTTV model for

model validation

The comparison tests between RadiCold cool roof and TPO roof were performed on

September 4th to 5th, 2017 and August 10th to 11th, 2018. The comparison tests for RadiCold cool

19
roof and shingle roof were conducted on September 5th to 6th, 2017 and July 25th to 26th, 2018.

All the tests lasted for 24 hours as shown in Figure 6. The wind speed and solar irradiation data

were monitored and plotted in Figure 7. Also, during these testing days, Laramie Regional

weather station shows that a little cloudy sky generally appeared few (i.e., 1~3) hours for every

day (24-hour period) [38]. The most heavily cloudy sky lasted 18 hours on July 25th to 26th, 2018

[38]. In these four experiments, the indoor air temperatures gradually decreased from evening to

dawn. The indoor air temperature differences between the RadiCold room and the shingle/TPO

room were quite small at night because all the three roof finishing materials possess high infrared

emissivity, especially for the emissivity within the atmospheric window, as shown in Figure 4.

When the sun rose, the indoor air temperature differences dramatically increased due to these

three materials’ large difference in solar absorptivity and infrared emissivity outside of the

atmospheric window. As the absorptivity of shingle is the greatest, the indoor air temperature in

the shingle room increased fastest and it reached about 30 °C in the late afternoon. Even without

roof insulation, the indoor air temperature of the RadiCold room stayed below or close to the

ambient temperature for a daytime solar irradiation maximum at 840 W/m2.

20
Figure 6 Experimental measurements of (a)/(c) Outdoor and indoor temperature for RadiCold

and TPO; (b)/(d) Outdoor and indoor temperature for RadiCold and shingle

21
 Figure 7. Experimental measurements of (a)/(c) Solar irradiation and wind speed for RadiCold

and TPO; (b)/(d) Solar irradiation and wind speed for RadiCold and shingle.

Comparisons between the calculation results from the iRTTV model and the experimental

measurements for the roof exterior surface temperature and the ceiling temperature are given in

Figures 8 and 9. Overall, the iRTTV model can well predict the experimentally measured

temperatures. The minor difference between the experimental data and the iRTTV predictions is

possibly due to two reasons. First, clouds appearance in the sky is a very complicated and

stochastic natural phenomena and the simplified cloud impact model used in this work [33] may

not fully capture cloud’s effect on sky radiation towards the roof surface. Second, as shown in

22
Figure 7, the wind speed presents a high fluctuation and naturally the wind direction is expected

to be fluctuating too, which may introduce discrepancy into the simulation results compared with

the measurement data.

Figure 8. Exterior surface temperature of the roof and ceiling temperature comparison between

the results from the iRTTV model and the experimental data for tests in 2017.

23
 Figure 9. Exterior surface temperature of the roof and ceiling temperature comparison between

the results from the iRTTV model and the experimental data for tests in 2018.

Meanwhile, the iRTTV model can also predicts the heat transfer rate through the roof to the

indoor space, which equals to the roof thermal transfer value – RTTV. Figure 10 shows the

RTTVs of the three roofs for the experimental cases that have been used to validate iRTTV

model. Figure 10 (a) and (b) show that at night from 8:30 PM to 7:00 AM, the differences of the

RTTVs between RadiCold cool roof, TPO cool roof, and shingle roof are small. In contrast,

during the daytime from 7:00 AM to 8:30 PM, RTTVs significantly increase for both the TPO

and the shingle roofs, mainly caused by the absorption of solar irradiation. Due to the high solar

24
absorptivity, shingle roof produced the highest RTTVs with a rapid rising rate in the morning.

Combing the extremely high solar reflectivity and infrared emissivity, RadiCold cool roof

generated cooling even under direct sunlight and its RTTVs were always negative or close to

zero. The RTTVs of RadiCold cool roof exhibit a small convex at about 10:00 AM and a

concave at 3:00 PM in Figures 10(a) and (b), both of which were introduced by the fluctuating

natural indoor air temperature as there was no mechanical heating or cooling source inside. In the

early morning, the temperature of the RadiCold cool roof surface increases with sun rising while

the indoor air temperature doesn’t change that fast because of thermal mass effect, and thus

resulted in the convex phenomena. Then during the day, as more heat was transferred into the

indoor space from walls, indoor air temperature rose and became higher than the roof

temperature, resulting in a greater radiative cooling effect from the RadiCold cool roof, and a

concave was thus yielded around 3:00 PM.

In summary, Figure 10 shows the RadiCold cool roof always dissipates heat to the sky during

the test periods, especially on Sept. 4th to Sept 6th, 2017 when there is very little clouds in the sky

(see Figure 10(a) and (b)). In contrast, the cooling power of RadiCold cool roof was reduced to

about 10 W/m2 (see Figure 10(d)) under the heavily cloudy sky condition on July 25th - 26th,

2018, where solar irradiation demonstrated sharp fluctuations in the afternoon as Figure 7(d)

illustrates.

25
 Figure 10 RTTV comparisons between the RadiCold, TPO and Shingle roofs

5. Application of the iRTTV model

Three types of roof exterior finishing: RadiCold, TPO, and shingle above a 6 inch concrete

roof structure for three locations: Tucson (AZ), Los Angeles (CA) and Orlando (FL) have been

analyzed, with an assumed flat roof configuration. The TMY3 weather data is adopted as

environmental inputs in iRTTV model calculations [34]. The concrete has a thermal conductivity

of 1.65 W/(m·K), a specific heat of 1000 J/(kg·K) and a density of 2400 kg/m3 [39]. It is

assumed that space cooling set point temperature is 24 oC (75 oF).

26
 The RTTVs, estimates of heat gain/loss via roofs, are summarized for the cooling periods

throughout the year. Figure 11 shows the monthly cumulative RTTVs of three roof finishing for

Tucson, AZ, which is the sum of the total heat gain/loss through the roof during cooling periods

for a particular month. The shingle roof finishing has the largest monthly cumulative RTTV,

suggesting the highest heat gain from the outdoor environment to the indoor space through the

shingle roof finishing. The cumulative RTTV of TPO cool roof finishing is between those of the

shingle roof and the RadiCold cool roof. The monthly cumulative RTTV of the RadiCold cool

roof is always negative, indicating continuous heat dissipation from indoor to outdoor

environment. We also noticed that the TPO and shingle roof finishing have the same trend for

the monthly cumulative RTTVs across the modeled year and peaks in June while the RadiCold

cool roof provides greatest benefits (i.e., lowest RTTV or highest heat loss) during shoulder

seasons. The results verify that the RadiCold cool roof has the best performance to block heat

gain as compared to the other two roof materials.

Figure 12 shows the monthly cumulative RTTVs of the RadiCold cool roof for three

locations in the U.S. The results illustrate that applications of the RadiCold cool roof in these

locations will produce net cooling effect via the roof and reduce building’s cooling load since all

the monthly cumulative RTTVs are negative. It is found that the local weather condition affects

the performance of the RadiCold cool roof. Tucson has the maximum solar irradiation in July

which results in a peaking RTTV value. Compared to Tucson and Orlando, the lowest outdoor

temperature in Los Angles leads to lowest RTTVs from May to October.

27
Figure 11 Comparison of monthly cumulative RTTVs of the RadiCold, TPO and shingle roof

finishing in Tucson, AZ

Figure 12 Comparison of monthly cumulative RTTVs of RadiCold cool roof for three locations

28
 Since the RTTVs represent the heat gain/loss through roof structure, they are directly

associated with the cooling load introduced from the roof structure. During the cooling season of

a typical meteorology year in these three locations, the RTTV results show that the shingle roof

and TPO cool roof transfer 78.9 – 294.1 kWh/(m2·yr) and 8.5 – 128.2 kWh/(m2·yr) of heat into

the building space, respectively. The RadiCold cool roof withdraws 137.6 – 268.7 kWh/(m2·yr)

of heat from space and dissipates the heat to outdoor environment. Figure 13 shows the monthly

cumulative roof induced A/C cooling electricity for these three types of the roofs in the three

selected locations with an assumed coefficient of performance (COP) of 3.0. Compared to the

TPO and shingle roof finishing that consume cooling electricity due to heat gain through the

roofs, the RadiCold cool roof can save cooling electricity because of cooling load reduction. In

the hottest month in Tucson, Los Angeles and Orlando, compared to the shingle roof, the

RadiCold cool roof can save 18.9 kWh/m2, 23.1 kWh/m2 and 12.9 kWh/m2 of cooling electricity

per square meter of roof area, respectively. In comparison with the TPO cool roof, the RadiCold

cool roof still can save 11.7 kWh/m2, 18.4 kWh/m2 and 10.0 kWh/m2 of cooling electricity

during the hottest month.

Among the three locations analyzed, Orlando, although generally having a hotter climate

than Los Angeles based on cooling-degree-day statistics, produces lower annual cooling

electricity savings from application of RadiCold cool roof than those achieved at Los Angeles.

An important reason for this result is because of the high atmospheric water content in Orlando.

29
Figure 13 Comparison of monthly roof induced A/C cooling electricity from three roof

finishing in (a) Tucson (AZ), (b) Los Angeles (CA), (c) Orlando (FL)

30
 With adequate electricity prices assumed, the RadiCold cool roof’s energy cost savings

($/per annum) over the shingle roof and the TPO cool roof can be assessed, as shown in Table 2.

The electricity prices used in the evaluation are from DOE EIA [40] as 13.16 cents/kWh for

Tucson, AZ, 19.90 cents/kWh for Los Angeles, CA, and 11.36 cents/kWh for Orlando, FL.

Compared to the shingle roof, the RadiCold cool roof saves 113.0 – 143.9 kWh/(m2·yr) cooling

electricity, associated with cooling cost saving of $12.9 - $18.9/(m2·yr). Similarly, compared to

the TPO cool roof, the RadiCold cool roof saves 88.0 – 92.4 kWh/(m2·yr) cooling electricity and

cooling cost is reduced by $10.0 - $18.4/(m2·yr). Compared to the TPO cool roof, the acceptable

incremental cost of the RadiCold cool roof falls within $30.0 - $55.2/m2 and $50.1 - $91.9/m2 for

3-yr and 5-yr payback periods, respectively, depending on the regions and climates. This

indicates that the RadiCold cool roof will be an economically sound cool roof material.

Figure 14 illustrates the impact of A/C system COP on the roof induced annual cooling

electricity consumptions. For a lower COP (e.g., 2.5), the benefit of installing RadiCold cool roof

will increase further. The RadiCold cool roof saves 135.6 – 172.7 kWh/(m2·yr) as compared to

shingle roof finishing and saves 105.7 – 110.9 kWh/(m2·yr) as compared to TPO cool roof. The

reduction in cooling cost reaches $15.4 - $27.7 /(m2·yr) and $12.0 - $22.1 /(m2·yr) for shingle

roof and TPO cool roof, respectively.

31
 Figure 14 Comparison of annual roof induced A/C cooling electricity for the three roof

finishing for three locations

Table 2 Energy saving and economic analysis for RadiCold cool roof based on iRTTV

calculation for A/C COP=3.0

Location Tucson, AZ Los Angeles, CA Orlando, FL

Baseline roof finishing
Shingle TPO Shingle TPO Shingle TPO
(→)
Cooling electricity
143.9 88.6 115.9 92.4 113.0 88.0
saving (kWh/m2‧yr)
Cooling cost saving
18.9 11.7 23.1 18.4 12.9 10.0
($/m2‧yr)
Acceptable
incremental cost for 3-
56.8 35.0 69.2 55.2 38.6 30.0
yr simple payback
period ($/m2)
Acceptable
incremental cost for 5-
94.7 58.3 115.3 91.9 64.3 50.1
yr simple payback
period ($/m2)

32
6. Conclusion

By incorporating the spectrum-dependent approach and the roof thermal mass effect, the

iRTTV model can be employed to evaluate the thermal performance of different roofs, especially

for the RadiCold cool roof. The iRTTV model has been validated with measurements from the

reduced-size model building located at Laramie, Wyoming, USA under real-world conditions.

The validation of the iRTTV model for the shingle, TPO, and RadiCold cool roofs proved that

the model can accurately predict the roof exterior surface temperature and the ceiling

temperature. Compared to conventional roofs, the RadiCold cool roof has the smallest RTTVs

and it can efficiently dissipate heat throughout day- and nighttime. However, when designing the

metamaterial-based cool roof, more factors, such as the roof tilt angle and roof structures, need a

comprehensive evaluation in order to maximize the cooling benefits. The iRTTV tool will be

further developed with additional functionalities for this design purpose.

By applying the validated iRTTV model to a flat concrete roof structure without additional

insulation for three locations (Tucson AZ, Los Angeles, CA, and Orlando FL) in the United

States, the associated heat gains/losses through roof have been predicted for the cooling periods

in a typical year for three types of roof exterior finishing: shingle, TPO, and RadiCold. The

shingle finishing transfers 78.9 – 294.1 kWh/(m2‧yr) of heat into the building space, the TPO

finishing transfers 8.5 – 128.2 kWh/(m2‧yr) of heat into the building space, while the roof with

RadiCold finishing dissipates 137.6 – 268.7 kWh/(m2‧yr) of heat from the building space to

outdoor environment. Assuming the COP of a typical air-conditioning system is 3.0, the cooling

load reduction from utilizing RadiCold cool roof will result in cooling electricity savings of

113.0 – 143.9 kWh/(m2‧yr) compared to the shingle roof finishing and 88.0 – 92.4 kWh/(m2‧yr)

compared to the TPO roof finishing for the three analyzed locations.

33
Acknowledgement

This research was supported by the U.S. Department of Energy’s Advanced Research Projects

Agency-Energy (ARPA-E) grant DE-AR0000580.

References

[1] Chan KT. Chow WK. Energy impact of commercial-building envelopes in the sub-tropical

climate. Appl Energ. 1998; 60:21-39.

[2] Zingre KT, Wan MP, Yang X. A new RTTV (roof thermal transfer value) calculation method

for cool roofs. Energy. 2015;81:222-32.

[3] Fang Z, Li N, Li B, Luo G, Huang Y. The effect of building envelope insulation on cooling

energy consumption in summer. Energy and Buildings. 2014; 77:197-205

[4] Gracia A, Dynamic building envelope with PCM for cooling purposes – Proof of concept.

Appl Energ. 2019; 235: 1245-1253.

[5] Chemisana D, Lamnatou Chr, Photovoltaic-green roofs: An experimental evaluation of

system performance, Appl Energ. 2014, 119: 246-256.

[6] Cui S, Ahn C, Wingert MC, et al., Bio-inspired effective and regenerable building cooling

using tough hydrogels, Appl Energ. 2016; 168:332-339.

[7] Romeo C, Zinzi M. Impact of a cool roof application on the energy and comfort performance

in an existing non-residential building. A Sicilian case study. Energ Buildings. 2013;67:647-57.

34
[8] Kolokotroni M, Gowreesunker BL, Giridharan R. Cool roof technology in London: An

experimental and modelling study. Energ Buildings. 2013;67:658-67.

[9] Levinson R, Akbari H, Reilly JC. Cooler tile-roofed buildings with near-infrared-reflective

non-white coatings. Building and Environment. 2007;42:2591-605.

[10] Bergman TL. Active daytime radiative cooling using spectrally selective surfaces for air

conditioning and refrigeration systems. Sol Energy. 2018;174:16-23.

[11] Chu S, Cui Y, Liu N. The path towards sustainable energy. Nature materials. 2017;16:16.

[12] Wong KV. Anthropogenic heat generation and heat exhaust to the ultimate sink. Journal of

Energy Resources Technology. 2017;139:034701.

[13] Raman AP, Anoma MA, Zhu L, Rephaeli E, Fan S. Passive radiative cooling below ambient

air temperature under direct sunlight. Nature. 2014;515:540.

[14] Zhai Y, Ma Y, David SN, Zhao D, Lou R, Tan G, Yang R, Yin X. Scalable-manufactured

randomized glass-polymer hybrid metamaterial for daytime radiative cooling. Science.

2017;355:1062-6.

[15] Zhang K, Zhao D, Zhai Y, Yin X, Yang R, Tan G. Modelling study of the low-pump-power

demand constructal T-shaped pipe network for a large scale radiative cooled-cold storage system.

Appl Therm Eng. 2017;127:1564-73.

[16] Wong NH, Cheong DKW, Yan H, Soh J, Ong CL, Sia A. The effects of rooftop garden on

energy consumption of a commercial building in Singapore. Energ Buildings. 2003;35:353-64.

[17] Chaiyapinunt S, Phueakphongsuriya B, Mongkornsaksit K, Khomporn N. Performance

rating of glass windows and glass windows with films in aspect of thermal comfort and heat

transmission. Energ Buildings. 2005;37:725-38.

35
[18] Ventrillard I, Romanini D, Mondelain D, Campargue A. Accurate measurements and

temperature dependence of the water vapor self-continuum absorption in the 2.1 μ m

atmospheric window. The Journal of chemical physics. 2015;143:134304.

[19] Shaviv E, Yezioro A, Capeluto IG. Thermal mass and night ventilation as passive cooling

design strategy. Renew Energ. 2001;24:445-52.

[20] Praditsmanont A, Chungpaibulpatana S. Performance analysis of the building envelope: A

case study of the Main Hall, Shinawatra University. Energ Buildings. 2008;40:1737-46.

[21] Zhao D, Aili A, Zhai Y, Lu J, Kidd D, Tan G, Yin X, Yang R. Subambient Cooling of

Water: Toward Real-World Applications of Daytime Radiative Cooling. Joule. 2018.

[22] Khalifa A-JN, Marshall R. Validation of heat transfer coefficients on interior building

surfaces using a real-sized indoor test cell. Int J Heat Mass Tran. 1990;33:2219-36.

[23] Zhao D, Martini CE, Jiang S, Ma Y, Zhai Y, Tan G, Yin X, Yang R. Development of a

single-phase thermosiphon for cold collection and storage of radiative cooling. Appl Energ.

2017;205:1260-9.

[24] Kim JT, Todorovic MS. Tuning control of buildings glazing's transmittance dependence on

the solar radiation wavelength to optimize daylighting and building's energy efficiency. Energ

Buildings. 2013;63:108-18.

[25] Ding C, Liu X, Liu W, Liu M, Li Y. Mafic–ultramafic and quartz-rich rock indices deduced

from ASTER thermal infrared data using a linear approximation to the Planck function. Ore

Geology Reviews. 2014;60:161-73.

[26] Gao L, Qin Z. Research on the fitting relation of the planck equation expansion parameter

model in split window algorithm. Geography and Geo-Information Science. 2007;23:9-12.

36
[27] Lipton AE, Moncet J-L, Uymin G. Approximations of the Planck function for models and

measurements into the submillimeter range. IEEE Geoscience and Remote Sensing Letters.

2009;6:433-7.

[28] Steven D. Lord, A new software tool for computing Earth’s atmospheric transmission of

near- and far-infrared radiation, NASA Technical Memorandum 103957, December 1992.

[29] Pinker R, Laszlo I. Modeling surface solar irradiance for satellite applications on a global

scale. Journal of Applied Meteorology. 1992;31:194-211.

[30] Zhang K, Zhao D, Yin X, Yang R, Tan G. Energy saving and economic analysis of a new

hybrid radiative cooling system for single-family houses in the USA. Appl Energ. 2018;224:371-

81.

[31] Gueymard CA. A reevaluation of the solar constant based on a 42-year total solar irradiance

time series and a reconciliation of spaceborne observations. Sol Energy. 2018;168:2-9.

[32] Iqbal M. An introduction to solar radiation: Elsevier; 2012.

[33] Eicker U, Dalibard A. Photovoltaic–thermal collectors for night radiative cooling of

buildings. Sol Energy. 2011;85:1322-35.

[34] Wilcox S, Marion W. Users manual for TMY3 data sets: National Renewable Energy

Laboratory Golden, CO; 2008.

[35] ASHRAE AS. Standard 90.1-2016, Energy standard for buildings except low rise residential

buildings. American Society of Heating, Refrigerating and Air-Conditioning Engineers, Inc.

2016.

[36] Cavanaugh LM. Redefining the green roof. Journal of Architectural Engineering.

2008;14:4-6.
37
[37] Smith G, Gentle A, Arnold M, Cortie M. Nanophotonics-enabled smart windows, buildings

and wearables. Nanophotonics. 2016; 5:55-73

[38] Laramie Regional Station, WU Weather Underground. (February 21, 2019).

https://www.wunderground.com/history/daily/us/wy/laramie/KLAR/date

[39] ISO E. 10456: 2007 Building materials and products–Hygrothermal properties–Tabulated

design values and procedures for determining declared and design thermal values. Brussels: CEN.

2007.

[40] Electric Choice. Electricity Rates by State (Updated November 1, 2018).

https://www.electricchoice.com/electricity-prices-by-state/

38