A low temperature water-cooled radiation calorimeter for estimation of concentrated solar irradiance

Solar Energy 167 (2018) 194–209

Contents lists available at ScienceDirect

Solar Energy journal homepage: www.elsevier.com/locate/solener

A low temperature water-cooled radiation calorimeter for estimation of concentrated solar irradiance Ram Niwas Verma, Rajesh Kumar, Ambesh Dixit, Laltu Chandra

T

⁎

Indian Institute of Technology Jodhpur, Rajasthan 342037, India

A R T I C LE I N FO

A B S T R A C T

Keywords: Concentrated solar irradiance Radiation calorimeter Experiment Simulation Heat transfer Coating

Radiation calorimeter is a device for assessment of the incident local concentrated solar irradiance (CSI) onto a receiver surface. This is required to evaluate the performance of a concentrated solar thermal system using, for instance a volumetric receiver, for applications. In this paper a low-temperature water cooled radiation calorimeter (RADCAL) is presented. This is designed as a cavity based on the concept of blackbody to maximize the absorption of the incident CSI on its absorber surfaces. Solar selective coatings are deposited on its reflector and absorber surface to achieve the same and are characterized using standard methodologies. This depicts an absorptivity of 0.95 and a reflectivity of 0.87 in the desired range of the solar spectrum and is theoretically verified. The absorption of energy results in temperature rise of RADCAL, which is controlled by an external water jacket to limit its value at 100 °C for mitigating the use of pressurized water. At any instant and eventually at the steady-state the CSI is estimated following the conservation of energy principle. The developed RADCAL is experimentally evaluated up to a CSI of 800 Suns (1 Sun = 1 kW/m2) using Joule heating. The entire heat transfer process is analysed with the developed unsteady state one-dimensional mathematical model. A comparative assessment of the measured and calculated RADCAL body temperature provides the underlying uncertainty and confirms its design basis. Furthermore, the design and given consideration allows its use in arid desert condition with dust and wind. Thus, RADCAL is likely to serve in future for evaluating concentrated solar thermal system in arid deserts.

1. Introduction Estimation of concentrated solar irradiance (CSI) onto a receiver is required for both line and point focusing based concentrated solar thermal (CST) technologies Mouzouris et al., 2011; Ballestrín and Monterreal, 2004; Fu et al., 2016. The harnessed energy can be utilized for power generation, heating and cooling applications, which make these technologies versatile. For evaluation of such systems a reliable estimation of CSI is necessary (Ballestrín and Monterreal, 2004; Kaluza and Neumann, 2001). Various methods are used for this purpose, which are based on one or more sensors (Kaluza and Neumann, 2001; Skouri et al., 2013). The most common gages used for this purpose are the circular foil or Gardon-type (Gardon, 1953), also known as a thermogage or hycal and Schmidt-Bolter type flux sensors. The basic principle of the Gardon radiometer and Schmidt-Bolter sensor is similar. The former is based on the radial and the latter is based on the axial temperature difference by heat conduction through the sensing element and are calibrated using the black-body source; see Fig. 1a and b (Kaluza and Neumann, 2001; Ballestrín et al., 2006). Currently there

are few commercial sensors, from Vatell, Captec and Hukesflux (Chen et al., 2013). The range of hukseflux sensor is between 50 and 200 Suns and the time constant is about 450 ms (http://www.hukseflux.com/ product/sbg01-heat-flux-meter?referrer=/product_group/heat-fluxsensors). Vatell sensors are having a better time constant of 17μs . Its surface temperature is in between 350 and 600 °C (vatell.com/node/5). Gardon type radiometer has certain advantages in terms of the size and the response time (< 1 s) however, an over estimation of the CSI is reported (Ballestrin et al., 2003). Also, the need and the involved challenges of a standard calibration technique are reiterated by various authors (Murthy et al., 1998; Guillot et al., 2014). These are summarized in Fig. 1 including some of their advantages and disadvantages. Generally, all these sensors are calibrated at the steady-state with a liquid coolant, like water see Fig. 1c (Lawson and Mcgurien, 1953; Estrada et al., 2007). A similar method is used in a cavity-type calorimeter (CAVICAL) wherein the incident energy is absorbed on its inner wall. This sensor is reported to estimate up to 4000 Suns (1 Sun = 1 kW/m2); see Fig. 1d (Pérez-Rábago et al., 2006). Water-cooled advanced solar technology for estimation of radiation for improved

Abbreviations: CSI, Concentrated Solar Irradiance; MFR, Mass Flow Rate; RADCAL, Radiation Calorimeter ⁎ Corresponding author. E-mail addresses: [email protected], [email protected] (L. Chandra). https://doi.org/10.1016/j.solener.2018.04.006 Received 15 January 2018; Received in revised form 22 March 2018; Accepted 2 April 2018 0038-092X/ © 2018 Elsevier Ltd. All rights reserved.

Solar Energy 167 (2018) 194–209

R.N. Verma et al.

Nomenclature

Pr P1 Pe P qg″ q″ R Re ReDh Rq Ti Tf T0 Ts Tmf t Tms λ W ρs ρf εf θ α ε

aperture area of RADCAL Ao Asurf heat transfer surface-area of RADCAL to fluid cross section area of water-flow Acf Acs cross section area of solid B (λ, T) black body radiation spectrum cps specific heat capacity of RADCAL solid cpf specific heat capacity of water Dh hydraulic diameter F Faradays constant Gr Grashof number Gz lDh Graetz number h average convective heat transfer coefficient Ie electric current J current density thermal conductivity of solid material ks kf thermal conductivity of water l length of channel along the flow mf mass-flow-rate of fluid (water) ms mass of RADCAL solid m″ mass flux or deposition rate M molarity Nu Nusselt number n no. of electrons pv vacuum pressure

heat loss, as in a receiver (Prakash et al., 2009; Reddy and Kumar, 2009; Huang and Sun, 2016; Dave et al., 2013). A tilt angle from 0 to 45° may be preferred at a low wind speed (Wu et al., 2010). Such observations may be adopted for designing even a cavity-type radiation calorimeter. Generally, the absorbers in a water-cooled radiation calorimeter are exposed to atmosphere and their deterioration is expected with dust laden wind. Therefore, these sensors are preferred for indoor use with a higher accuracy in comparison to outdoor condition. Consequently, the use of such devices in arid desert regions with a high wind-speed needs special consideration. The deposition of dust on the exposed surface may lead to an underestimation of CSI. There are few CSI measuring devices, which are working at a high temperature range (∼850 °C) Ballestrin, 2002. Under such an extreme outdoor condition, absorber coating is limited and its durability is questionable (Selvakumar and Barshilia, 2012). Considering the detailed insight, challenges and limitations of such devices, an externally water-cooled low-temperature radiation calorimeter (RADCAL) is proposed, even for an outdoor application. Following literature review, this device is based on the concept of black body. Moreover, the calibration technique depending on water-cooling and the conservation of energy is adopted for its evaluation. In this design, the RADCAL temperature is limited up to 100 °C for a CSI of 800 Suns or 800 kW/m2 that relaxes the need of pressurized-water and mitigates heat loss to ambient. Joule heating is employed for evaluation of RADCAL. A coating is deposited on the reflecting surface of RADCAL to redirect the incident radiation onto the absorbing surface (Moreno et al., 2005; Lide, 2007). The absorber of RADCAL is coated with a solar selective material to maximize the useful heat gain from the reflected solar energy (Granqvist, 1991; Behar et al., 2013; Barshilia et al., 2006). These spectrally selective absorbers should exhibit a high absorptance of ≥0.95 in the spectral range 0.3–2.5 μm and a low thermal emittance of ≤0.05 in the infrared spectral range 2.5–25 μm (Kennedy, 2002). The ceramic metal (cermet) absorber structures are commonly used due to their high spectral selectivity and the ease of integrating metal in ceramic matrix. Various physical and chemical deposition processes are explored for synthesizing cermet structures. Zynolite is commonly used as an absorber layer on the exposed surface of sensors as it creates a

experiment (ASTERIX) was developed to address some of the reported limitations see Fig. 1e (Ferrier and Rivoire, 2000). In most of these devices the temperature of calorimeter is higher than 100 °C and therefore, pressurized-water is required for cooling. This needs special arrangement on-the-field and the safety standards must be strictly followed. In general, cavity-type design may be employed for mitigating

1

14 (d)

2

15 16 17

T T

Response Ɵme < 1 sec. Used for indoor measurement

T T

6

4 5 RadiaƟon calorimeter

(b) 7

18

8

9

(f) 11

10 (c)

12

19

(e) 22 20 21

New sensor approach ???

High accuracy of measurement Diĸculty in controlling water Ňow rate

3 2

(a)

Prandtl number perimeter of the insulation electric power perimeter of water channel heat flux (loss) through insulation heat-flux average radius of curved heat transfer area electric resistance Reynolds number root mean square roughness initial temperature (=inlet condition) temperature of fluid (water) temperature of water at outlet temperature of solid bulk temperature of fluid (water) time in second average temperature of RADCAL solid wavelength electrodeposited mass density of solid density of fluid (water) electrode efficiency azimuthal coordinate solar absorptance emittance

13

Fig. 1. Schematic of different types of measuring devices. (a) Gardon radiometer (1 constantan foil, 2 copper block (cooled), 3 shield, 4 +ve terminal, 5 −ve terminal); (b) Schmidt Boelter calorimeter (6 black coating, 7 Water cooling tube, 8 PTFE cable, 9 T type thermocouple); (c) Cold water calorimeter (10 diffuser, 11 inlet flow, l2 outlet flow, 13 circular stainless steel plate); (d) CAVICAL calorimeter (14 water inlet, 15 insulation, 16 stainless steel wall, 17 copper wall, 18 water outlet); (e) ASTRIX (19 incident beam, 20 water inlet, 21 water outlet, 22 littering). 195

Solar Energy 167 (2018) 194–209

R.N. Verma et al.

100 °C at the highest possible CSI (viz. 800 Suns or 800 kW/m2) and with the lowest convective heat transfer coefficient. The designed water-cooled jacket around the absorbing surface and a cross-section of RADCAL is shown in Fig. 3a and b. In this channel the water flow is laminar in nature. The minimum value of surface-area averaged Nusselt number (Nu ) is 4.86 in the thermally fully developed regime with a constant heat-flux condition. The hydraulic diameter of this cooling jacket is 3.74 mm. These results in the minimum heat transfer coefficient hmin of 793 W/m2 K from Eq. (1). Also, the relation between CSI and the cooling water temperature rise is provided in Eq. (2) at the steady-state by

rough surface, which improves the heat absorption beyond 3500 kW m−2 (Ballestrin et al., 2003). Fortunately, in the designed RADCAL there is a need of absorber coating, which is air and thermal stable at 200 °C. Therefore, the well-known black chrome solar selective coating, which is thermally stable up to 250 °C is deposited on the absorbing surface to ensure a high absorptivity in 0.3–3 µm (Cao et al., 2014; McDonald, 1974, 1975; Bogaerts and Lampert, 1983) and a low emissivity for thermal radiation beyond 3 µm. For this purpose the modified deposition technique on a curved surface is presented. Another consideration is that the heat loss from RADCAL can be mitigated using (a) a glass-cover on its aperture with an anti-reflective coating and (b) an external insulation. These are expected to reduce the measurement uncertainty on the field. At the steady-state, the CSI is estimated by the rate of heat removal with cooling water. Considering these aspects the design and evaluation of a low-temperature RADCAL is presented using both the detailed analysis and experiments.

h =

Nukf (1)

Dh

CSI = mḟ cpf

2. Design of radiation calorimeter

To−Ti hAsur = (Tms−Tmf ) Ao Ao

(2)

where Tms is the average surface temperature of RADCAL, Tmf is the bulk temperature of water at the inlet and outlet. In the RADCAL geometry Asur / A0 ∼ 27.3 allows limiting the temperature difference between solid and fluid. This is adopted in Eq. (2) as the temperature across the thickness is almost uniform, which is inferred from an extremely small value of Biot (Bi) number of 0.0036. It is noted that the flow inside cooling jacket will be developing and may attain the thermally fully developed condition. Thus, a better estimation of h is desired. For this purpose, the Graetz number Gz lDh Hausen, 1943; Mercer, 1960 is computed by

A schematic of the designed RADCAL along with its two-dimensional section is shown in Fig. 2. The concentrated solar radiation is incident on the reflecting surface depicted by 2 in this figure. Afterwards, the rays are redirected towards the absorbing surface at which the energy is absorbed. To ensure that the rays through the aperture AA' are incident on the reflecting surface, a tilt or an acceptance angle of about 20° is considered. This follows a recommendation from Wu et al. (2010) to design a cavity-type receiver. The ratio of absorber-surface to aperture area should be sufficient to limit the RADCAL temperature. The underlying assumption is that the incident rays are symmetrically distributed around the axis of RADCAL and are ideally normal to the aperture. The reflecting surface is represented by BOBˊ in Fig. 2b. The remaining inner surface depicts the absorber. The aperture width is assumed to be 20 mm along with an incidence angle of 20° to calculate the dimensions of RADCAL. This may be varied depending on applications and their requirements. Beyond this deviation around the normal or axis of RADCAL, the incident radiation may not be intercepted by the absorber. This is a constraint of the designed RADCAL, which is envisaged for a localized measurement. This may be relaxed with a wider aperture and its effect will be investigated in future. The centreline of RADCAL is indicated by z-axis at r = 0 as shown in Fig. 2b. The vertex of the cone (O) is aligned with this line to ensure symmetry of the reflected rays. The proposed design is analyzed with specular reflection from the reflecting surface of RADCAL, which is suitably coated for this purpose. The distance between cover plate with opening AAˊ and the origin (O) is selected to allow the confinement of incident rays inside RADCAL. For example, the straight line OK given by z = −0.36r is the reflection of the extreme ray H'O. A simple geometry based calculation reveals that the radial (r) coordinate of the point K should be greater than 10 mm, which is the same as the radius of aperture AA′. The cover plate is placed at a height of 6 mm from the vortex O with a margin of 0.6 mm. The ray at the other extreme namely, AB given by z = −2.74r + 33.47 is incident at B leading to its coordinate as (19.15, −19.15); see Fig. 2b. Hence, both the radius and height of the conical section BOB' are taken as 20 mm with a margin of 1 mm. Consequently, the distance between the cover and bottom plate of 26 mm along the z-direction is called the height of RADCAL. The solar selective absorber and reflector coatings are deposited to maximize the energy absorption from CSI to RADCAL. The shape of absorbing surface is adopted for limiting the temperature and to prevent the boiling of water at atmospheric pressure. The location of point D is carefully selected to maintain a gap with the point G along the line z = −7 . The point of intersection between this line and AB provides the coordinate (in mm) of G (14.7, −7) and that of D is (22, −7) on the r-z plane allowing a sufficient separation and flexibility in manufacturing. The selection of points E and F aims to limit the water temperature to

D Gz lDh = ⎛ ⎞ Re Dh Pr ⎝l⎠

(3)

where, Re Dh is the Reynolds number, Pr is the Prandtl number and Nu associated with the isothermal boundary condition in a circular tube is computed using Hausen correlation (Hausen, 1943) as

Nu = 3.66 +

0.0668Gz lDh 2/3 1 + 0.04Gz lD h

. (4)

Furthermore, Mercer correlation for the flow between flat plates with an insulated and an isothermal wall (Mercer, 1960) is given by

1 Rays 2 ReŇecƟng surface 3 Absorbing surface

H’ H r=0

r=-40 L

B’

r=40

A(10,6)

K A’ J

20

O (0,0)

F

GD

z=6 z=0 z=-7

z

I

z=-20 B(20,-20) E(40,-20)

r

Fig. 2. (a) Three and (b) two-dimensional views/section of RADCAL (dimensions in mm). 196

Solar Energy 167 (2018) 194–209

R.N. Verma et al.

3. Modelling: one-dimensional unsteady heat transfer The differential control volume of external water-cooling jacket around RADCAL corresponding to an angle dθ is shown in Fig. 5. The derived conservation of energy equation in the considered one-dimensional heat transfer approach is presented subsequently. Here, the properties are considered as independent of temperature, which is reasonable for the envisaged low temperature rise. The radius of heat transfer surface is calculated as an average of its inner and outer values at z = −7 & 6 mm, which is 31 mm. A part of the solid is not exposed to water-cooling due to manufacturing limitations; see Fig. 3a. Thus, a high solid temperature is expected at θ = 0° or 360°. The perimeter along θ from the inlet to outlet is 180 mm, excluding the unexposed section. 180 Thus, the average radius of water-cooling jacket is 2π = 28.6 mm . The total length of angular region at a cross section is 36 mm; see Fig. 3b. Therefore, the estimated heat transfer area is 6480 mm2, which is used to calculate heat-flux on the absorbing surface. The cross-section area of solid is non-uniform along the height as shown in Fig. 3b. Therefore, the average area is given by Acs = ms /(ρs × 2πr ) with ms and ρs as the mass and density of RADCAL solid, respectively.

(a) 40

2

31

1

27 22

3.1. Governing equation for fluid

(b)

Water flows along the azimuthal direction in the cooling jacket around the RADCAL solid. Therefore, heat is transferred from solid to the coolant by mixed or forced convection for the lowest or the highest 2 mass-flow-rate, which is inferred from 0.03 ⩽ Gr / ReDh ⩽ 0.35. Following simplifications are employed for one-dimensional analysis of the bulk fluid temperature along the azimuthal (θ) direction:

Fig. 3. (a) Isometric view of RADCAL with water jacket, (b) cross-section of RADCAL; dimensions in mm.

Nu = 4.86 +

1.2 0.0606Gz lD h 0.7 1 + 0.0909Pr 0.17Gz lD h

.

(5)

Finally, using the data as in Heaton et al. (1964) for the flow between flat plates with one wall at a constant heat-flux and the other as insulated the following correlation for Nu is obtained using 0.296 Nu = 1.864Gz lD . h

(a) heat conduction is along the azimuthal direction; (b) fluid properties are independent of temperature; (c) fluid temperature is uniform along the radial direction.

(6)

Applying the conservation of energy in the differential control volume over dθ results in

Let, the bulk temperature of water at the inlet is 25 °C and at the outlet is 62.5 °C. Using Eq. (2) for the minimum mass-flow-rate of water (0.0016 kg/s) and CSI of 800 Suns, results in Tmf of 43.8 °C. The allowable Tms of 90 °C including a safety margin of 10 °C is used for estimating the absorber surface-area of 6857 mm2. The designed area of 8600 mm2 includes a future possibility of using RADCAL for a higher CSI up to 1000 Suns. Therefore, the coordinates (in mm) of points E and F are taken as (40, −20) and (40, 6), respectively. Hence, it is confirmed that the water temperature will be lower than its boiling point and the requirement of pressurized-water is relaxed. Also, this will lead to reduced heat loss. The top and side views of RADCAL with overall dimensions are shown in Fig. 4a and b. As explained, both the absorbing and reflecting surface should satisfy the design criteria. These are high absorptivity and specular reflectivity of radiation having wavelength in the range of 0.3–3 µm. For reflection of the CSI in UV–Vis–NIR range towards the absorber surface, Silver is chosen as the coating material with a high reflectivity of ≥90%. However, a spectrally selective coating is required on the absorbing surface of RADCAL. Considering desirable optical properties and their thermal stability up to 200 °C, the widely used black-chrome is preferred (Bogaerts and Lampert, 1983; Driver, 1981; Smith et al., 1985; Sweet et al., 1984; Lampert, 1979; Holloway et al., 1980). However, as shown in Fig. 2, the absorbing surface is curved or highly non-linear and therefore, the electrochemical route is preferred for deposition over existing approaches. The adopted methodologies and characterization of the absorber and reflector coatings will be discussed in Section 6. In continuation with the design of RADCAL the heat transfer analysis and experimental evaluation of the device is presented in Section 3.

Fig. 4. (a) Top and (b) side views of RADCAL; dimensions are in mm. 197

Solar Energy 167 (2018) 194–209

R.N. Verma et al.

kf Δt kf Δt ṁ f Δt ⎞ j ⎛ ⎞ j ⎛ Tfji + 1 = ⎜ T + + T ρf cpf r 2 (Δθ)2 ⎟ fi + 1 ⎜ ρf cpf r 2 (Δθ)2 ρf Acf r Δθ ⎟ fi − 1 ⎝ ⎠ ⎝ ⎠ 2kf Δt ṁ f Δt hpΔt ⎞ j ⎛ hpΔt ⎞ j ⎛ +⎜ Ts + 1− − − T ρf Acf cpf ⎟ i ⎜ ρf cpf r 2 (Δθ)2 ρf Acf r Δθ ρf Acf cpf ⎟ fi ⎝ ⎠ ⎝ ⎠ (9)

ks Δt ⎞ (T j + T j ) + p1 k g ΔtTamb Tsji + 1 = ⎜⎛ si − 1 2 2 ⎟ si + 1 lg ρs Acs cps ⎝ ρs cps r (Δθ) ⎠ Δ p k t hpΔt ⎞ j ⎛ hpΔt ⎞ j 2ks Δt g + ⎜⎛1− 1 − − ⎟ Tf ⎟ Tsi + ⎜ 2 2 i ⎝ ρs Acs cps ⎠ ⎝ lg ρs Acs cps ρs cps r (Δθ) ρs Acs cps ⎠ q″pΔt + ρs Acs cps (10)

Fig. 5. A schematic depicting a differential control volume in the curved waterjacket for analysis.

∂2Tf + hp (Ts−Tf ) kf Acf 2 2   ∂θ  r  Convective heat Rate of conductive transfer rate from heat transfer solid to liquid ∂Tf ∂Tf mḟ cpf ρf Acf cpf − = . r θ t ∂ ∂    Rate of heat removedRate of energy stored by cooling water in fluid element

where, the indices i and j depicts the spatial and the temporal discretization, respectively. The Courant-Friedrich-Lewy (CFL) criterion is employed for ensuring the stability of the numerical approach. This is obtained based on the principle of positive coefficient as in Patankar (1980) gives 1

,⎞ ⎛ hp ṁ f 2kf ⎜ ρf Acf cpf + ρf Acf r Δθ + ρf cpf r 2 (Δθ)2 ⎟ Δt ⩽ min ⎜ ⎟ 1 p1 kg hp 2ks ⎜⎜ ⎟⎟ + + 2 2 ρs Acs cps lg ρs Acs cps ⎝ ρs cps r (Δθ) ⎠

(7)

3.2. Governing equation for solid

(11)

4. Experiment: heat transfer

To derive the governing equation of heat transfer from the solid surface to fluid, thermal uniformity along the thickness of solid is asserted by an extremely small value of Biot number ∼0.0036. This is expected due to (a) high thermal conductivity of solid material viz. copper and (b) mixed/forced convection in the water-cooling channel. Following simplifications are employed for analysing one-dimensional heat transfer:

A schematic of the experiment setup is depicted in Fig. 6a with the employed Joule heating and external water-cooling mechanisms. The Nichrome wire is internally wrapped for Joule heating on the absorber surface of RADCAL, which is covered with Mica and Kapton tape for electrical insulation. Special manual attention is given to avoid air gap in between the wire and the absorbing surface, as far as possible, in view of the surface curvature. The RADCAL is made of copper and is fabricated in three parts to allow the deposition of coating; see Fig. 6b and d. The use of copper with high thermal conductivity (385 W/m K http://www.engineeringtoolbox.com/thermal-conductivity-metals-d_ 858.html) promotes uniformity of the surface heat-flux and, therefore, mitigates a local hot-spot due to non-specular reflection. The adopted method allows (a) evaluation of the designed RADCAL and (b) its calibration with a known input power. The ratio of input power to the aperture area is equivalent to CSI onto the RADCAL. Sixteen equally spaced calibrated K-type thermocouples are located along θ on the middle part of RADCAL; see Fig. 6c. Moreover, four K-type thermocouples are used to measure inlet and outlet water temperature. Temperature difference at the inlet and outlet depends on the mass-flowrate (MFR) of water. For the device characterization and reducing heat loss to ambient, RADCAL is externally insulated with glass-wool. However, due to local air–gap and finite thermal conductivity of glasswool, heat loss to ambient is envisaged. Temperature is measured with an accuracy of ±0.1 °C. The cooling water flows around the RADCAL inside an external jacket; see Fig. 3a and 6a. The estimated uncertainty in the measured temperature difference of water at the inlet and outlet for 100 Suns with MFR of 0.0048 kg/s is ± 12.5% and for 400 Suns with MFR of 0.0016 kg/s is ± 1%. Thus, an overall uncertainty of ± 10% is associated with the measured temperature. Moreover, temperature variation of liquid and solid along the azimuthal direction is expected. Consequently, both the local and the surface-area averaged values are considered for the evaluation of RADCAL. The volume-flow-rate of water is measured with a known vessel and the time required for filling in the same. The recorded variation in the measurement of water MFR is ± 0.0002 kg/s, which is based on the two sets of experiment to be discussed at a later part. The resulting uncertainty in the measured MFR

(a) heat conduction in solid is along the azimuthal direction; (b) heat loss through the insulation is along the radial direction; Applying the conservation of energy results in,

∂ 2T q″P − + hp (Ts−Tf ) ks Acs 2 s2   ∂θ r    Rate ⏟ of Joule Convective heat Rate of conductive heating transfer rate from heat transfer solid to liquid ∂T ρs Acs cps s qg″ P1 − = t ∂    Rate of⏟ heat loss Rate of energy stored through insulation in solid element

(8)

where qg″ = k g (Ts−Tamb)/ lg with lg as 50 mm and initial condition is The boundary conditions are Tf (θ,0) = Ti = Ts (θ,0) . ∂Tf ,s

Tf (0,t ) = Ti , Ts (0,t ) = Ts (2π ,t ) and ∂θ (2π ,t ) = 0 . Here, Ti is the inlet temperature. This is considered to be the same for both solid and fluid at the initial condition. The measured value of Ti are used in simulations.

3.3. Discretization For numerical solution of the considered problem, the explicitmethod is employed. Here, the central-difference scheme is used for conduction term in solid and fluid domains. The backward-difference approach is used for convection term in fluid, which results in

198

Solar Energy 167 (2018) 194–209

R.N. Verma et al.

Fig. 7. A comparison between surface-area averaged Nusselt number (Nu ) with Reynolds number (Re Dh ) in two different experiments (a) E1 (b) E2; (c) increase in thermal resistance from E1 to E2.

5. Result and discussion: heat transfer

Fig. 6. (a) Schematic view of RADCAL with water jacket and cooling system, (b) manufactured parts of RADCAL (c) position of thermocouple on middle part and (d) coated parts of RADCAL.

The calculated values of surface-area averaged Nusselt number (Nu) basing on the measured solid and fluid temperature are presented in Fig. 7a and b. These include the two sets of performed experiments at the different points of time. The first-set of experiment (E1) was performed starting from January 2016 and the second-set (E2) was performed starting from December 2016. The detailed measurements in solid and fluid will be presented subsequently. Following are inferred from these figures:

of water is ± 12.5% for 0.0016 kg/s and is ± 4.2% for 0.0048 kg/s. The uncertainty in current and resistance measurement of the Nichrome wire is ± 0.05 A and ± 0.1 Ω, respectively. Consequently for the lowest power of 32 W an uncertainty of ± 4.3% is estimated using

ΔPe = Pe

2

2

⎛ 2ΔIe ⎞ + ⎛ ΔR e ⎞ ⎝ Ie ⎠ ⎝ Re ⎠

⎜

⎟

⎜

⎟

. (a) The estimated values of Nu depend on Re Dh and Pr. The temperature of water at the outlet increases with CSI and consequently Pr decreases. However, for a given Re Dh the values of Nu are practically independent with respect to CSI. This may be attributed to a small 199

Solar Energy 167 (2018) 194–209

R.N. Verma et al.

in the range 26–29 °C in E1 and 29–35 °C in E2 with the inlet temperature of about 25–26 °C. However, for the case with 400 Suns, the temperatures are, mostly, within 50–65 °C in E1 and 50–70 °C in E2 with some exceptions. These higher values may be considered as an indicator of local heat transfer deterioration. This may also include the contribution from unexposed calorimeter surface from 350 to 360° to the coolant. However, in the simulation this is treated as a continuous body by virtue of the employed boundary conditions. The computed values with Nu as in Eq. (12) are mostly comparable to measurement within ± 10%, which is inferred from the considered cases. The differences among measured values of RADCAL solid temperature, except a few θ positions in E1 & E2, are mostly within ± 10%, which reveal that the estimated uncertainty is indeed reasonable. Furthermore, consistent behaviour of elevated temperature in 180° ⩽ θ ⩽ 270° for E2 may indeed be attributed to the localized surface fouling. At the steady-state, the measured and calculated values of surfacearea averaged temperature of RADCAL-solid and bulk temperature of cooling water at the outlet for the different CSI up to 800 Suns with MFR of 0.0016 kg/s, 0.0048 kg/s and 0.015 kg/s are shown in Fig. 11. The calculated temperatures are mostly within ± 10% of the experimentally obtained values at the steady-state. The measured and calculated temperatures of RADCAL body and water decrease with increasing MFR. These figures show that the solid temperature decreases with increasing Re Dh or the corresponding Nu , which is indicated by the dotted line at 50 °C in Fig. 11a and b. In all the cases, the flow is in laminar regime and the RADCAL solid temperature is limited to 70 °C, substantially lower than that of the maximum value. Therefore, to measure a high value of CSI the device should be operated at a high MFR and vice versa. Attempts will be made to deduce a calibration chart to relate the CSI with a non-dimensional parameter. Fig. 12 shows the temperature rise (a) in solid from the initial to steady-state and (b) in water from the inlet to outlet for comparing the experiments E1 and E2. Here, two sets of experiments are performed starting from December 2016, denoted by 2/2.1 in the legend. The temperature rise in solid during the experiments E2 and E2.1 is higher than that of E1. There are no such practical differences between the measured values of E2 and E2.1. Temperature rise of water is similar for both the set of experiments viz. E1 and E2/2.1. Therefore, the heat removal rate by cooling water is practically the same during E1 and E2/ 2.1 depicting the same input power. These confirm a higher resistance to heat transfer between solid and water in E2/2.1 compare to E1, which is consistent with analysis of Fig. 7. Thus, it is recommended that such devices must be installed and regularly maintained. Moreover, the

variation in Pr in the temperature range with ± 10% error in estimation of Nu . Generally the values of Nu increases with Re Dh , as expected. The same is depicted by the trend lines at two different Re Dh in E1 and E2. (b) Higher values of Nu are obtained for E1 in comparison to E2. This may be attributed to the fouling effect on the copper body of RADCAL. To assess the same, the increase in thermal resistance is computed from these values of Nu as shown in Fig. 7c. The additional thermal resistance is obtained by subtracting convective resistance in E1 from the total resistance in E2 in which variations are observed due to the measurement uncertainty. The trend-line indicates that the effect of fouling will reduce with increasing Re Dh . Thus, for modelling and further analysis a correlation for Nu is derived from E2. The derived conservative correlation of Nu using the least-square method is shown in Eq. (12). It must be noted that the flow and thermal boundary layer will develop simultaneously along θ starting from the inlet. The measured and one-dimensional unsteady model based solid (copper) temperature of RADCAL at two different azimuthal positions viz. θ = 180° & 270° for E1 and E2 are shown in Fig. 9. These are depicted for a power of 128 W (400 Suns) and a MFR of water 0.0016 kg/ s. The solid temperature at the middle part of RADCAL is considered; see Fig. 6. The position dependent Nu correlation will be higher than that of Eq. (12). The derived correlation is compared with some of the wellknown expressions for laminar and developing internal flow; see Fig. 8. The experiment and correlation based Nu are in the range of 3–10 with Eq. (12) as the most conservative. Consequently, an over-estimation of copper temperature is envisaged with the same. This will allow framing a strategy for its mitigation. This experiment based Nu is employed for validation of the derived mathematical model and for the evaluation of RADCAL.

Nu = 0.6228 × Re D0.3855 . h

(12)

In general, the measured solid temperature at the upper-half of this section Ts _mu is higher than that of the lower-half Ts _ml . This may be attributed to the combined effect of non-uniform wire wrapping and even a slight misalignment of the injection port while manufacturing that lead to an unequal flow distribution. However, this can’t be visually confirmed with the opaque and insulated external surface. The measurement shows that Ts_mu in E1 is lower than that of E2. An unexpected high value of Ts_mu, perhaps, indicates the local deterioration of heat transfer due to fouling or even a bubble formation. Further investigations are needed to confirm the same. The model analyzed values are mostly within ± 15%. The differences between the measured and model analyzed values, are attributed to the use of conservative Nu , which should ideally be a function of θ. One limitation of the developed model is its failure to capture the temperature gradient beyond 100 s in solid by virtue of a lower time constant of 20 s. Therefore, further possibilities of improvement in the derived model exist and even a two-dimensional approach may be required for the same. A more detailed comparison between the measured temperatures at the steadystate for E1 and E2 will be presented at a later stage. The measured and one-dimensional unsteady model analysed steady-state solid temperature along θ is shown in Fig. 10a and b for a power of 32 W (100 Suns) with MFR of 0.0048 kg/s and for 128 W (400 Suns) with MFR of 0.0016 kg/s, respectively, for both the set of experiments E1 & E2. These correspond to Re Dh ∼ 290 and 97 , respectively. It is obvious that the water temperature increases from the inlet to the outlet and the same is expected for solid. However, as explained, the inlet and outlet are located at θ = 0° and 350°, respectively. Thus, a surface of about 10° remains unexposed to coolant. Moreover, the solid is a continuous body and therefore, the same temperature is expected at θ = 0° and 360°. The measured solid temperature at the middle-part with 100 Suns is

Fig. 8. A comparison between the existing and derived correlation for Nu with experiment E2. 200

Solar Energy 167 (2018) 194–209

R.N. Verma et al.

Fig. 9. Transient temperature rise of solid for power = 128 W (400 Suns or 400 kW/m2) and mass-flow-rate of water = 0.0016 kg/s (Re Dh ∼ 97) (a) at 180° and (b) at 270°.

use of demineralized water should be preferred to mitigate the scaling or fouling effect. 6. Deposition and characterization of coating materials for RADCAL The radiation calorimeter (RADCAL) consists of reflector and absorbing surfaces (see Fig. 2). Here, CSI should be reflected without any loss onto the absorber surface. The absorber surface is externally watercooled to estimate the concentration. The reflector surface should be nearly ideal and thus, a thermally stable coating should be optimized for entire solar spectral wavelength viz. 0.3–2.5 μm. This can be achieved by the selected metallic surfaces such as Aluminum, Silver and Gold. Aluminum reflecting surfaces may not withstand the environmental degradation such as, reaction with ambient oxygen and moisture. In contrast, Gold is very stable under such conditions, however, relatively very expensive and therefore not considered as a common reflecting surface. Thus, Silver is a common choice for reflecting surfaces not only because of its environmental inertness but also the relatively better reflecting properties ∼95–99% (Moreno et al., 2005; Lide, 2007; Rancourt, 1996) in solar spectral region viz. 0.3–2.5 µm. This is comparable to that of gold reflecting surfaces 98–99% at a wavelength of ≥0.5 µm.

Fig. 10. Azimuthal variation of the measured and the one-dimensional model analysed steady-state solid temperature (a) for 100 Suns at aRe Dh of 290 (b) for 400 Suns at aRe Dh of 97.

surface. Initially, the deposition of silver reflecting surfaces is optimized on 25 mm × 20 mm flat copper substrate to achieve a high reflectance in UV–Vis spectral range. The optimized conditions are used to deposit the silver layer on RADCAL reflector. The silver coins with 99% purity are placed on tungsten (W) boat for evaporation. The base pressure is maintained at ∼3.0 × 10−6 Torr prior to the deposition of silver (Ag) thin film. The distance between the source material viz. Ag and the substrate is maintained at 20 mm. The quartz crystal balance is used to

6.1. Methodology: reflector coating A semi-automated Semicore Triaxis (SC) make thermal evaporator system is used to synthesize the silver thin films on the reflecting 201

Solar Energy 167 (2018) 194–209

R.N. Verma et al.

Temperature (°C)

T_s_exp_Nu T_s_1D_mercer_Nu T_s_measured To_W_Hausen_Nu To_W_heaton_Nu Ti_W_Measured 60

T_s_1D_hausen_Nu T_s_heaton_Nu To_W_Exp_Nu To_W_Mercer_Nu To_W_Measured

50 °C

50 40 30

20 50

150

250 CSI (Suns)

T_s_exp_Nu T_s_1D_mercer_Nu T_s_measured To_W_Hausen_Nu To_W_Heaton_Nu Ti_W_Measured

350

450

T_s_1D_hausen_Nu T_s_Heaton_Nu To_W_Exp_Nu To_W_Mercer_Nu To_W_Measured

Temperature (°C)

80

Fig. 12. Temperature rise (a) from initial to the steady-state for RADCAL body (Ts) and (b) in water from inlet to outlet (To_W ) for_ aRe Dh = 97 and at different values of CSI.

60

50 °C 40

Table 1 The electrochemical bath details for the deposition of nickel and black chrome coatings on copper substrate.

20 50

150

250 CSI (Suns)

450

T_s_1D_hausen_Nu

70

Temperature (°C)

350

T_s_1D_mercer_Nu

60

T_s_Heaton_Nu T_s_exp_Nu

50

Chemical bath composition and condition for Ni deposition

Chemical bath composition and condition for BC deposition

NiSO4·6H2O - 250 g/l NiCl2·6H2O - 60 g/l H3BO3 - 40 g/l Temperature - 40 °C Time (t) - 30 s Current density (J) 0.00085 A/mm2

CrO3 - 275 g/l NaF - 0.2 g/l NaNO3 - 3 g/l Temperature - 15–17 °C Time (t) - 60 s Current density (J) 0.0035 A/mm2

T_s_measured 40

Table 2 The probable electrochemical reaction mechanism for nickel and black chrome coatings during deposition.

To_W_Hausen_Nu To_W_Mercer_Nu To_W_Heton_Nu

30 20 450

650 CSI (Suns)

850

To_W_Exp_Nu To_W_Measured Ti_W_Measured

Ni electro-deposition mechanism

Black chrome electro-deposition mechanism

Ni2 + Cl− → NiCl+

HCr3O10− + 6H+ + 6e− → Cr2O3 + HCrO4 + 3H2O HCr3O10− + 2H2O ↔ Cr (OH)2 + HCr2O7− + 2HF + H+ Cr(OH)2 ↔ CrO + H2O

NiCl+ + e− → NiCl

Fig. 11. Steady-state temperature of RADCAL body (Ts) and water at the outlet (T0_W) with (a) MFR of 0.0048 kg/s (Re Dh = 290) and (b) MFR of 0.0016 kg/s (Re Dh = 97) (c) MFR of 0.015 kg/s (Re Dh = 879).

NiCl + e− → NiClads NiClads + e− → Ni + Cl−

monitor the growth rate and thickness of the deposited thin films. The silver was heated at 3.46 kW DC power on the tungsten boat to realize the silver deposition on the desired substrates including conical reflecting surface; see Fig. 6d. The thickness of deposited structure is measured using the profilometer and is consistent with that of quartz crystal monitor. This is also verified by estimating the deposition using mass flux. Considering the negligible angular distribution of the evaporated particles/molecules on the substrate, Langmuir showed that the

HF

CrO + HF ←→ Cr + H2 O

rate at which vapor comes in contact with the metal surface is the same as evaporation rate of metal (Langmuir, 1913) and is given by

m′′ = p (M /2πRT )1/2

(13)

where m″ is the evaporation mass flux in kg m−2 s−1, p is the partial pressure of evaporating material and R is the gas constant (8.314 J K−1 mol−1). The calculated deposition mass flux is 202

Solar Energy 167 (2018) 194–209

R.N. Verma et al.

3.3 × 10−8 kg m−2 s−1, and is close to the experimentally calculated rate 3 × 10−8 kg m−2 s−1 with an uncertainty of 9%. 6.2. Methodology: absorber coating The absorbing surface of RADCAL should be coated with high spectrally selective absorber coatings to maximize the useful heat gain from the reflected CSI. The absorber layer needs to be thermally stable up to 100 °C without degrading the optical properties and to withstand the outdoor environmental conditions. SSCs are being developed for absorbers operating at a low (< 100 °C), medium (100–300 °C) and high temperature (> 300 °C) (Selvakumar and Barshilia, 2012; Kennedy, 2002; Cao et al., 2014; Zhang and Mills, 1992; Brunold et al., 2000; Muehlratzer et al., 1981). These are deposited using physical and electro-chemical methods (Bogaerts and Lampert, 1983; Driver, 1981; Smith et al., 1985; Sweet et al., 1984; Lampert, 1979; Teixeira et al., 2001; Fan and Spura, 1977) and the later is relatively favored (Andersson et al., 1980; Lai and Riley, 2008). An associated challenge is the current requirement and the resulting heat generation. For example, the electro-deposition of black chrome SSC is carried out at a current density of 0.002–0.005 A mm−2 (Daryabegya and Mahmoodpoora, 2006) that will increase with surface area. The RADCAL absorber surface area is ∼8600 mm2 and therefore, the required current is in the range of 17.2–43 A. This may locally alter the electrochemical bath conditions leading to the coating degradation. Thus, the process is optimized to achieve the desired optical properties. Coupling of surface with the infrared reflector enhances the spectrally selective response

Fig. 13. (a) The schematic of the electrodeposition setup for nickel and black chrome layers and (b) Zoomed electrode section explaining the different layers stacking representatively.

Fig. 14. The XRD pattern for Ag film deposited on copper substrate (Ag/Cu).

Fig. 15. (a, b) Optical image of bare Cu and Ag/Cu, (c, d) SEM image of bare Cu and Ag/Cu, (e, f) AFM 3-D image of bare Cu and Ag/Cu. 203

Solar Energy 167 (2018) 194–209

R.N. Verma et al.

electrodeposition setup is shown in Fig. 13a with the respective layer stacking in Fig. 13b including both nickel infrared reflector and black chrome spectrally selective coating as an absorber. The black chrome (Cr-Cr2O3) electrochemical bath consists of chromium trioxide (CrO3) as the main source of chromium and chromium oxide (Cr2O3) in conjunction with sodium fluoride (NaF) and sodium nitrate (NaNO3) as reaction catalysts. The probable chemical reactions during the electrochemical deposition of black chrome are given in Table 2 (Hoare and LaBoda, 1985). 7. Characterization measurements X-ray diffraction (XRD) measurements are performed to understand the development of crystallographic phases and structures of the synthesized reflector and absorber coatings. Bruker D8 Advance X-ray diffractometer is used in locked couple mode. XRD data is recorded in the range of 20–80° at angular speed 0.02°/s copper Kα (λ = 1.5406 Å) monochromatic radiation at 40.0 kV and 40.0 mA. The microstructural and surface properties are investigated using Carl Zeiss EVO 18 especial edition scanning electron microscope (SEM) and Park System XE-70 atomic force microscope (AFM) systems. AFM micrographs are collected on 5 × 5 µm2 scanning area at 256 × 256 pixel resolution. The elemental compositions is measured using the energy dispersive X-ray (EDX) instrument (OXFORD make) equipped with SEM system. The thicknesses of these films are characterized using Dektak XT stylus surface profiler (Bruker). The optical reflectance is measured using 110 mm integrated sphere based diffuse reflectance accessory with a Carry 4000 UV–Vis spectrophotometer within 0.3–0.9 µm (UV–visible) wavelength range to understand the spectral response. The measurements are normalized with respect to the polytetrafluoroethylene (PTFE) as a reference sample. A Bruker vertex 70 V FT-IR spectrophotometer is used to measure the reflectance in 2.5–25 µm wavelength range against gold as a reference sample. UV–Vis and FTIR reflectance measurements are used to calculate the room temperature solar absorptance α(λ) and thermal emittance ε(λ), respectively. Considering an opaque material, the spectral absorptance can be expressed in terms of spectral reflectance R (λ) as

Fig. 16. Reflectance as a function of wavelength of (a) Ag/Cu (flat surface) (b) Ag/Cu (inclined surface) [inset: optical image of conical reflecting surface].

and avoids the inter-diffusion of substrate elements. Here, a thin film of nickel is optimized as the infrared reflector having a lower emissivity compare to absorber. The black chrome absorber is electrochemically deposited on this reflector to achieve the desired optical performance. The nickel electrochemical bath consists of nickel sulfate, as the main source of nickel ions, and nickel chloride for distributing nickel ions homogeneously to achieve the uniform nickel film coverage on the entire surface. The pH of the electrochemical bath is maintained by adding the required boric acid. The bath conditions are summarized in Table 1. The electrodeposition of nickel involves two consecutive one electron charge transfer process and the formation of an adsorbed complex with the help of an anion (Saraby-Reintjes and Fleischmann, 1984). The probable chemical reaction mechanism of nickel ions in electrolyte solution is summarized in Table 2, where nickel sulfate is the main source of Ni+ ions. A schematic of the representative

α (λ ) = 1−R (λ )

(14)

where λ is the wavelength, and R(λ) is the spectral reflectance against wavelengths. The solar absorptivity can be calculated using the solar (λ) (http://rredc.nrel.gov/solar/spectra/am1.5/ spectrum ISun astmg173/astmg173.html) as λ

α=

∫λ1 2 [1−R (λ )] ISun dλ λ

∫λ1 2 ISun dλ

(15)

where λ1 = 0.3 μm, λ2 = 0.8 μm are the minimum and the maximum

Fig. 17. (a) Thickness versus time plot. (b) Emissivity versus thickness plot of Ni/Cu structure. 204

Solar Energy 167 (2018) 194–209

R.N. Verma et al.

Fig. 18. (a) Thickness versus current density. (b) Thickness versus time and (c) Emissivity versus thickness plot for BC/Ni/Cu structures.

copper substrates (Fig. 15c and d) that are visible on the deposited silver thin films. The respective EDX measurement on Ag/Cu structure is shown as an inset in Fig. 15c and d, depicting only silver and a fraction of copper, from substrate. There was no signature of oxygen in EDX measurement, substantiating the metallic nature of deposited silver thin film, without any silver oxide impurity. Further, atomic AFM measurements are carried out for probing local surface structure of both Cu and Ag/Cu and the collected 3D AFM micrographs are shown in Fig. 15e and f. The root mean square roughness (Rq) of scanned surface is estimated as Rq = |

wavelength of interest, respectively, for the present measurements. The thermal emittance is the ratio of emissive power of a hot body and a blackbody at same temperature. The hemispherical thermal emittance is assumed because of omnidirectional radiation nature (Tesfamicharel, 2000) and for an opaque object this is expressed as (16)

Considering the thermal emittance, the emissivity of a coating at temperature T is given by λ

ε (T ) =

∫λ1 2 [1−R (λ )] B (λ ) dλ λ

∫λ1 2 B (λ ) dλ

. (17)

−1 c2 c1 λ−5 e λ − 1

(

L

∫ |Z 2 (x )|dx |, where Z(x) describes the surface 0

profile of the sample over the investigated length L (Nilsson et al., 1970). The measurement reveals that surface roughness of Ag/Cu is of 16.74 nm which is nearly one half of the bare Cu substrate of 32 nm. This is attributed to deposition of silver on Cu substrate smoothing the substrate trenches. This silver surface roughness is much smaller than the wavelength of the incoming solar radiation. Thus, the reflecting properties of Ag/Cu surface are retained. The optical performance of Ag/Cu structure is evaluated by measuring UV–Vis reflectance in the wavelength range of 0.2–0.9 μm. The measured spectral reflectance is shown in Fig. 16 for all the deposited Ag/Cu structures. The reflectance of 100 nm thick film is 60% in the entire spectral range, shown in Fig. 16, marked as the pre-optimized sample. This is attributed to the poor surface quality of the deposited silver thin film. Considering the constraint, thickness of silver thin film is increased up to 500 ± 15 nm by increasing the deposition time up to 30 min under identical condition. This is repeated on several samples and the measured reflectance is shown in Fig. 16 for these (500 ± 15 nm thick) silver surfaces, which is 87.29 ± 4% for the spectral region > 0.3 µm. The surface roughness of these 500 nm thick silver films is 16.74 ± 1.00 nm, which is much smaller than that of 100 nm thick silver films ∼28.39 ± 1.83 nm, substantiating the observed poor reflectance properties (Fig. 16). Additionally, a sharp dip is also observed near 0.3 μm in the reflectance and is attributed to the plasmon resonance absorption because of the large (∼1022 cm−3) free electrons in silver metal (Nilsson et al., 1970). However, this should not affect the optical performance of the reflecting surface as the major fraction of solar spectrum lies beyond 0.3 μm wavelength. The measured reflectance values of silver against its thickness are summarized in the inset of Fig. 16a and are consistent with literature (Fekkai, 2014; http://www.filmetrics.com/reflectance). The observed small difference may be attributed to the surface properties such as roughness, which depends on the substrate surface quality and synthesis process used for depositing the silver reflecting surface. These depositions are done on flat copper substrates for 500 nm thick silver films. Considering the geometrical constraints of conical surface, the silver coating of 500 nm thickness is deposited under the identical conditions on four different samples. The measured optical reflectance (Fig. 16b) is comparable to that of the flat substrates (Fig. 16a). These observations depict that the silver reflecting properties are nearly identical and relatively insensitive to the geometry within the vapour coverage area. Finally RADCAL

Fig. 19. The XRD pattern of black chrome film deposited on bright nickel substrate (BC/Ni/Cu).

ε (λ ) = 1−R (λ ).

1 L

)

is the blackbody radiation spectrum at a where B (λ,T ) = and temperature T with c1 = 3.743 × 10−16 W µm4 m−2 c2 = 1.4387 × 10−16 µm K are the first and the second Planck’s radiation constants, λ1 and λ2 are 2.5 µm and 25 µm, respectively for measurements. 8. Results and discussion The representative X-ray diffractogram (XRD) is shown in Fig. 14 for the optimized silver thin film on a copper substrate. The diffraction peaks at 2θ = 37.7°, 43.8°, 63.7° and 76.4° can be are indexed with the reference ICDD PDF # 03-065-8428, which corresponds to (1 1 1), (2 0 0), (2 2 0) and (3 1 1) face centred cubic crystallographic planes, respectively, as marked in Fig. 14 for easy identification. These measurements confirm the presence of metallic silver thin film on copper substrates. The optical and scanning electron micrographs (SEM) are shown in Fig. 15 for both copper substrate and the silver deposited copper (Ag/ Cu) structures. The optical images do not reflect any surface roughness and processing imprints see Fig. 15a and b. However, the corresponding SEM micrographs clearly exhibit the cleaning process imprints on 205

Solar Energy 167 (2018) 194–209

R.N. Verma et al.

Fig. 20. (a, b) Optical image of Ni/Cu and BC/Ni/Cu & (c, d) SEM image of Ni/Cu and BC/Ni/Cu (e, f) 3-dimensional AFM image of Ni/Cu and BC/Cu/Ni.

which is defined as

reflecting surface is coated under an identical deposition conditions and the optical image of silver coated conical reflecting surface is shown as an inset in Fig. 16b. The spectrally selective coating on RADCAL absorber surfaces comprises of nickel infrared reflector layer, followed by the black chrome. These layers are deposited electrochemically and the detailed bath conditions and process are explained in Section 8. Initially, the nickel metal thin film is deposited for different times under the constant current density J = 850 A/m2 at ∼30 mm electrode separation. The films are deposited for 30, 60 and 120 s and the respective thicknesses are measured using thickness profilometer. The results are summarized in Fig. 17a showing the variation of thickness against the deposition time. The deposited nickel mass on the substrate is also estimated using Faraday's electrolysis laws as (Jensen, 2012; Paunovic et al., 2010).

W=

ItM nF

εf =

Wa × 100 Wth

(19)

where Wa is the actual and Wth is the theoretical weight of deposit. The estimated electrode efficiency of the nickel electrodes is ∼63% ± 9%. Considering the relative error bars, the minimum and the maximum electrode efficiency values will be about 54% and 72%, respectively. The measured thickness of the nickel thin films should lie between the minimum and the maximum limiting values. These limiting cases along with the present electrode efficiencies are summarized in Fig. 17 for electrodeposited nickel thin films. The measured thicknesses of nickel infrared reflector thin films are 1, 1.8 and 2.4 µm for 30, 60 and 120 s (data square, Fig. 17), respectively. The optical response of these thins films is analysed with FTIR reflectance in 2.5–25 µm and used to estimate the emissivity values. These values are summarized in Fig. 17b for different nickel thin films, suggesting high reflectivity in 2.5–25 µm with emissivity values are ≤0.03. Thus, one micrometer thick nickel film is used in conjunction with black chrome absorber layer for the desired spectrally selective coating response for the RADCAL reflecting surface. The black chrome layer is optimized after depositing one micron thick nickel infrared reflector layer on copper substrates. The electrochemical deposition process of black chrome absorber layer is described in Section 8 with

(18)

where W is the electrodeposited weight, I is deposition current, t is the deposition time, M is the molecular weight of deposited atom or molecule, F is Faraday constant (∼96,500 C mol−1) and n is the number of moles of electron. The calculated nickel weight is ∼0.080 g for 30 s electrodeposited thin film, whereas the measured weight for the same is ∼0.05 g. This observed difference is attributed to the lower electrode efficiency (εf), 206

Solar Energy 167 (2018) 194–209

R.N. Verma et al.

Fig. 21. (a) Reflectance versus wavelength for deposited BC/Ni/Cu structure. (b) HT experimental schematic. (c) BC/Ni/Cu-HT (d) BC/Ni/Cu and BC/Ni/Cu-HT.

and optical characterization to understand the structure – property – process correlations. The crystallographic structure of the developed BC/Ni/Cu structure is investigated using the XRD measurements and a representative X-ray diffraction graph is shown in Fig. 19 for the developed BC/Ni/Cu structure. The diffraction peaks at 2θ = 43.4, 50.4 and 74° correspond to (1 1 1), (2 0 0) and (2 2 0) diffraction planes for copper substrate. The additional diffraction peaks at 2θ = 44.7 and 51.8° correspond to (1 1 1) and (2 0 0) chromium metallic planes, respectively. The diffraction peaks are not observed for the chromium oxide phase, suggesting that the oxide phase is mostly present in the amorphous phase. This may be advantageous as there are no finite grains, thus minimizing the active interaction sites near grain boundaries for atmospheric contaminants and gases, protecting against environmental corrosive degradation. The optical, SEM and AFM micrographs are shown in Fig. 20 for Ni/Cu and BC/Ni/Cu top surfaces. The optical images suggest the smooth shining Ni/Cu surface and black BC/ Ni/Cu surface, Fig. 20b. However, SEM micrographs reveal otherwise and depict substrate imprints. The surface morphology of the nickel thin film is granular, whereas BC thin films comprise of non-uniform and randomly oriented grains. This is consistent with the observed amorphous nature chromium oxide from XRD studies. The respective EDX measurements are shown as the insets, suggesting the metallic nature of nickel infrared reflecting and chromium metal rich black chrome layers. The three dimensional AFM micrographs, shown in Fig. 20e and f, clearly exhibit the surface imprints for both the thin film structures. The surface roughness is reduced significantly in comparison to the compared copper substrate (∼32 nm). The observed root mean square (rms) roughness of Ni/Cu layer is 16.90 nm, much lower than that of the copper substrate. The same for BC/Ni/Cu surface has increased up to 21.05 nm, however lower than the copper substrate. This is attributed to the hillock-type surface morphological growth patterns on the black chrome thin films (see Fig. 20). The thicknesses of optimized Ni and BC layers are 1 ± 0.11 μm and 1.5 ± 0.031 μm, respectively, including both the

bath conditions and probable deposition reaction mechanisms in Tables 1 and 2, respectively. The black chrome thin films are deposited for 60 s at different current densities and at 0.0035 A cm−2 current density for different time intervals. The measured respective thicknesses are summarized in Fig. 18a and b for different current densities and different time intervals, respectively. The deposited BC weight is calculated using Eq. (18) and is 0.0.00766 kg for 60 s electrodeposited thin film. However, the measured weight of electrodeposited BC thin film is 0.00478 kg. The difference in calculated and measured weight is due to the lower electrode efficiencies, similar to that observed for case of nickel thin films. The respective thicknesses with limiting (the minimum and the maximum) and the present electrode efficiencies are plotted in respective Figures, together with 100% electrode efficiency. The observed behaviour of increase in thickness is consistent in all these estimates with the measured thickness and values with the actual electrode efficiencies are very close to the measured thicknesses. The optical response of black chrome on Ni/Cu structures is measured using FTIR reflectance and the estimated emissivity is summarized in Fig. 18c for BC/Ni/Cu structures with different thickness of BC absorber layer. The emissivity values are much larger for lower thickness BC thin films, which reduces very fast with increasing BC thickness either by increasing the current density or by increasing the deposition time. The large values of emissivity for thickness less than ≤1 μm is attributed to lower chromium metal fraction in chromium oxide ceramic matrix, which plays a critical role in achieving the desired optical and thermal response. Thus, a black chrome thin film with thickness 1.5 µm is opted for coating the RADCAL absorbing surface. The nickel and black chrome thin films with optimized thicknesses 1 and 1.5 µm are used for coating on the actual RADCAL absorbing surface with the electrochemical bath conditions as summarized in Table 1. The optical images of coated RADCAL absorber surface are shown in Fig. 6d. The black chrome spectrally selective coatings are further subjected to the intensive structural, microstructural, morphological 207

Solar Energy 167 (2018) 194–209

R.N. Verma et al.

thermal conductivity of solid allows volumetric redistribution of the absorbed thermal energy in RADCAL. The design basis depicts its capability for estimating say up to 800 Suns with a maximum temperature less than 100 °C. Experiments are performed with RADCAL using various mass-flow-rates of water and laminar flow condition. Based on the measured RADCAL solid temperature a conservative correlation for surface area averaged Nu is deduced for this special design. Subsequently, a one-dimensional unsteady heat transfer model is developed in which this correlation is employed. A comparison between the measured and calculated RADCAL temperature at the steady and unsteady state substantiates the design basis and depicts a probable uncertainty of 10%. Thus, the need of a more detailed analysis is proposed for future. Evaluation of the measured RADCAL solid and cooling water temperatures clearly reveals the repeatability and fouling effect. This is inferred with an elevated local RADCAL temperature at some azimuthal positions. The reflection of concentrated solar irradiance and the absorption of energy in RADCAL require suitable solar selective coatings. Silver as a reflector and black chrome as an absorber coatings are deposited on the RADCAL. Optimal deposition conditions are obtained based on experiments and analysis for these coatings. The absorptivity and emissivity of the absorber coatings are 0.95 and 0.05, respectively. This is adequately supported using both theory and experiments. The reflectivity of the corresponding surface is ∼0.87 in the desired special range. These coatings are found to be thermally stable and scalable. The developed RADCAL is envisaged for used in arid desert regions with dust and sand. Thus, special consideration is provided for the same. Finally, such a low-temperature RADCAL is expected to be useful in practice for evaluation of concentrated solar thermal systems in such regions, worldwide.

surface roughness and the measurement uncertainty. The optical performance of these fabricated structures is evaluated using UV–Vis and FTIR base reflectance measurements. The measured reflectance is shown in Fig. 21a, c and d for a set of BC/Ni/Cu structures under identical conditions. The reflectance of a pre-optimized sample is shown for a comparison. These are used to calculate the emissivity, using Eq. (17). At the room temperature ε ≤ 0.14 ± 0.02 in 2.5–25 μm with the optimized BC/Ni/Cu structures. The respective UV–Vis reflectance plots are shown as inset figures. This is used to estimate the absorptivity, which is ≥0.95 ± 0.01 in 0.3–0.9 μm for the optimized BC/Ni/Cu structures. Bruggeman model is generally used for large metal volume fraction in a ceramic matrix to compute the effective dielectric constant of the composite medium as (Nikalasson et al., 1981)

0 = fA

εA−ε ε −ε + fB B εA + 2ε εB + 2ε

(20)

where, ε is the effective dielectric constant of the composite medium, εA and εB are the dielectric constants of metal and dielectric, respectively, fA is the metal filling fraction in the composite medium with fB = 1-fA. The optical constant of the composite medium N can be computed as ε = N2, where N = n - ik with n as the real part of the refractive index and k the extinction coefficient. Considering the normal incidence, a system with m multi-layered structure, a matrix approach is adopted as (Han, 1989)

α ⎡ ⎤= ⎣β⎦

m

∏ j=1

⎡ cosδj ⎢ iN sinδj ⎢ ⎣ j

i sinδj ⎤ Nj ⎡

1 ⎤ ⎥ cosδj ⎥ ⎣ Nm + 1⎦ ⎦

(21)

where i = −1, Nj is the optical constant of the jth layer, and 2π δj = λ Nj dj (for normal incidence) is the phase thickness of the jth layer; β / α is the refractive index of the multilayer layer and substrate. Finally the reflectance R of multilayer system can be calculated as 2

Acknowledgements The authors’ are thankful Ms. Snehlata Joshi, Mr. Vishwadeepak Kumar, Mr. Gurveer Singh, M/s Borana Pumps for their valuable support and feedback. The authors’ are indebt to the Ministry of New and Renewable Energy, Government of India for the provided financial support vide Sanction No: 50/40/2010-11/ST and convey gratitude to IIT Jodhpur for the provided basic implimentation facilities.

∗

N −β / α ⎞ ⎛ N0−β / α ⎞ R = ⎜⎛ 0 ⎟⎜ ⎟ ⎝ N0 + β / α ⎠ ⎝ N0 + β / α ⎠

(22)

where N0 (=1) is the optical constant of air. The required optical constants are borrowed from the literature (Rakić et al., 1998; Werner et al., 2009; Ordal et al., 1987, 1985; Babar and Weaver, 2015) and used to compute the reflectance R of BC/Ni/Cu structure. The computed reflectance is in good agreement with the measured reflectance (Fig. 21a) in the higher wavelength range for 46% metal fraction in black chrome absorber layer. However, the computed reflectance differs in the lower wavelength region and is attributed to the lack of high quality materials data in this region in conjunction with the ignorance of surface and interface irregularities in the effective medium approach. The thermal response is important to understand the stability coatings and device in addition to their optical performance. The periodic heat treatment of BC/Ni/Cu structure is carried out in a controlled furnace from room temperature (RT) to 200 °C at constant heating rate of 0.16 °C/s and allowed to cooled down to RT in ambient. The periodic heating cycles are carried out up to for about 35 h, as depicted in Fig. 21b. The recorded infrared reflectance in 2.5–25 µm is shown in Fig. 21c with inset showing its UV–Vis reflectance. The reflectance plot are overlapping over the entire thermal wavelength region, suggesting the thermal robustness of the deposited black chrome structure. Reflectance for the prepared and the thermally treated BC/ Ni/Cu structures are summarized in Fig. 21d. The emissivity is ∼0.17 for the thermally treated sample which is practically unchanged. The absorptivity value is nearly same for thermally treated sample, suggesting that these coating structures are thermally stable.

References Andersson, Å., Hunderi, O., Granqvist, C.G., 1980. Nickel pigmented anodic aluminum oxide for selective absorption of solar energy. J. Appl. Phys. 51 (1), 754–764. Babar, S., Weaver, J.H., 2015. Optical constants of Cu, Ag, and Au revisited. Appl. Opt. 54, 477–481. Ballestrin, J., 2002. A non-water-cooled heat flux measurement system under concentrated solar radiation conditions. Sol. Energy 73 (3), 159–168. Ballestrín, J., Monterreal, R., 2004. Hybrid heat flux measurement system for solar central receiver evaluation. Energy, 29, 915–924. Ballestrin, J., Ulmer, S., Morales, A., Barnes, A., Langley, L.W., Rodriguez, M., 2003. Systematic error in the measurement of very high solar irradiance. Sol. Energy Mater. Sol. Cells 80, 375–381. Ballestrín, J., Estrada, C.A., Rodríguez-Alonso, M., Pérez-Rábago, C., Langley, L.W., Barnes, A., 2006. Heat flux sensors: calorimeters or radiometers? Sol. Energy 80 (10), 1314–1320. Barshilia, H.C., Selvakumar, A.N., Rajam, K.S., 2006. TiAlN/TiAlON/Si3N4 tandem absorber for high temperature solar selective applications. Appl. Phys. Lett. 89, 1919091–191909-3. Behar, O., Khellaf, A., Mohammedi, K., 2013. A review of studies on central receiver solar thermal power plants. Renew. Sustain. Energy Rev. 23, 12–39. Bogaerts, W.F., Lampert, C.M., 1983. Materials for photothermal solar energy conversion. J. Mater. Sci. 18, 2847. Bogaerts, W.F., Lampert, C.M., 1983. Materials for photothermal solar energy conversion. J. Mater. Sci. 18 (10), 2847–2875. Brunold, S., Frei, U., Carlsson, B., Möller, K., Köhl, M., 2000. Accelerated life testing of solar absorber coatings: testing procedure and results. Sol. Energy 68 (4), 313–323. Cao, F., McEnaney, K., Chen, G., Ren, Z., 2014. A review of cermet-based spectrally selective solar absorbers. Energy Environ. Sci. 7 (5), 1615–1627. Chen, K., Parker, N., Chun, W., Oh, S.J., Lim, S.H., 2013. Development and testing of a simple heat gauge for the measurement of high-intensity thermal radiation. Int. Commun. Heat Mass Transfer 46, 1–6. Daryabegya, M., Mahmoodpoora, A.R., 2006. Method of Manufacturing Absorbing Layers on Copper for Solar Applications (I). Isesaco science and Technology Vision, 2,

9. Conclusion A cavity type radiation calorimeter called RADCAL is designed basing on the concept of blackbody, which is made of copper. High 208

Solar Energy 167 (2018) 194–209

R.N. Verma et al.

Mercer, A.M., 1960. The growth of the thermal boundary layer at the inlet to a circular tube. Appl. Sci. Res. 9, 450–456. Moreno, I., Araiza, J.J., Avendano-Alejo, M., 2005. Thin-film spatial filters. Opt. Lett. 30 (8), 914–916. Mouzouris, M., Roberts, L.W., Brooks, M.J., 2011. Heat transfer analysis for high-flux solar measurements using a flat plate calorimeter. South African Inst. Mech. Eng. J. 27, 10–22. Muehlratzer, A., Goerler, G.P., Erben, E., Zeilinger, H., 1981. Selection of a black chrome bath for continuous tube-plating and the properties of the coatings deposited from it. Sol. Energy 27 (2), 115–120. Murthy, A.V., Tsai, B.K., Saunders, R.D., 1998. High-heat-flux sensor calibration using black-body radiation. Metrologia 35 (4), 501–504. Nikalasson, G.A., Granqvist, C.G., Hunderi, O., 1981. Effective medium models for the optical properties of inhomogeneous materials. Appl. Opt. 20, 26–30. Nilsson, P.O., Lindau, I., Hagstrqm, S.B.M., 1970. Optical plasma-resonance absorption in thin films of silver and some silver alloys. Phys. Rev. B, 1(2), 498–505. Ordal, M.A., Bell, R.J., Alexander, R.W., Long, L.L., Querry, M.R., 1985. Optical properties of fourteen metals in the infrared and far infrared: Al Co, Cu, Au, Fe, Pb, Mo, Ni, Pd, Pt, Ag, Ti, V, and W. Appl. Opt. 24, 4493–4499. Ordal, M.A., Bell, R.J., Alexander, R.W., Long, L.L., Querry, M.R., 1987. Optical properties of Au, Ni, and Pb at submillimeter wavelengths. Appl. Opt. 26, 744–752. Patankar, S., 1980. Numerical Heat Transfer and Fluid Flow. CRC Press. M. Paunovic, M. Schlesinger, D.D. Snyder: Modern Electroplating ch. Fundamental Considerations, 5th ed., John Wiley & Sons Inc, 2010, 1–32. Pérez-Rábago, C.A., Marcos, M.J., Romero, M., Estrada, C.A., 2006. Heat transfer in a conical cavity calorimeter for measuring thermal power of a point focus concentrator. Sol. Energy 80 (11), 1434–1442. Prakash, M., Kedare, S.B., Nayak, J.K., 2009. Investigations on heat losses from a solar cavity receiver. Sol. Energy 83 (2), 157–170. Rakić, A.D., Djurišic, A.B., Elazar, J.M., Majewski, M.L., 1998. Optical properties of metallic films for vertical-cavity optoelectronic devices. Appl. Opt 37, 5271–5283. Rancourt, J.D., 1996. Optical Thin Films: User Handbook. SPIE Press. Reddy, K.S., Kumar, N.S., 2009. An improved model for natural convection heat loss from modified cavity receiver of solar dish concentrator. Sol. Energy 83 (10), 1884–1892. Saraby-Reintjes, A., Fleischmann, M., 1984. Kinetics of electrodeposition of nickel from watts baths. Electrochim. Acta 29 (4), 557–566. Selvakumar, N., Barshilia, H.C., 2012. Review of physical vapor deposited (PVD) spectrally selective coatings for mid-and high-temperature solar thermal applications. Sol. Energy Mater. Sol. Cells 98, 1–23. Skouri, S., Bouadila, S., Salah, M.B., Nasrallah, S.B., 2013. Comparative study of different means of concentrated solar flux measurement of solar parabolic dish. Energy Convers. Manage. 76, 1043–1052. Smith, G.B., McPhedran, R.C., Derrick, G.H., 1985. Surface structure and the optical properties of black chrome. Appl. Phys. A Mater. Sci. Process. 36 (4), 193–204. Sweet, J.N., Pettit, R.B., Chamberlain, M.B., 1984. Optical modeling and aging characteristics of thermally stable black chrome solar selective coatings. Sol. Energy Mater. 10 (3–4), 251–286. Teixeira, V., Sousa, E., Costa, M.F., Nunes, C., Rosa, L., Carvalho, M.J., Gago, J., 2001. Spectrally selective composite coatings of Cr–Cr2O3 and Mo–Al2O3 for solar energy applications. Thin Solid Films 392 (2), 320–326. Tesfamicharel, T., 2000. Characterization of Selective Solar Absorbers. Experimental and Theoretical Modeling, ACTA Universitatis Upsaliensis Uppsala, 4–79. Werner, W.S.M., Glantschnig, K., Ambrosch-Draxl, C., 2009. Optical constants and inelastic electron-scattering data for 17 elemental metals. J. Phys Chem Ref. Data 38, 1013–1092. Wu, S.Y., Xiao, L., Cao, Y., Li, Y.R., 2010. Convection heat loss from cavity receiver in parabolic dish solar thermal power system: a review. Sol. Energy 84 (8), 1342–1355. Zhang, Q.C., Mills, D.R., 1992. Very low-emittance solar selective surfaces using new film structures. J. Appl. Phys. 72 (7), 3013–3021.

35–39. Dave, N., Azih, C., Yaras, M.I., 2013. A DNS study on the effects of convex streamwise curvature on coherent structures in a temporally-developing turbulent boundary layer with supercritical water. Int. J. Heat Fluid Flow 44, 635–643. Driver, P.M., 1981. An electrochemical approach to the characterisation of black chrome selective surfaces. Sol. Energy Mater. 4 (2), 179–202. Estrada, C.A., Jaramillo, O.A., Acosta, R., Arancibia-Bulnes, C.A., 2007. Heat transfer analysis in a calorimeter for concentrated solar radiation measurements. Sol. Energy 81 (10), 1306–1313. Fan, J.C., Spura, S.A., 1977. Selective black absorbers using rf-sputtered Cr2O3/Cr cermet films. Appl. Phys. Lett. 30 (10), 511–513. Fekkai, Zakia., 2014. Nazir Mustapha. Ali Hennache Optical, morphological and electrical properties of silver and aluminium metallization contacts for solar cells American Journal of Modern Physics 3 (2), 45–50. Ferrier, A., Rivoire, B., 2000. Measurement of concentrated solar radiation: the camorimeter ASTERIX. In: 10th International Symposium on Concentrated Solar Power and Chemical Energy Technologies, Australia. Fu, T., Zong, A., Tian, J., Xin, C., 2016. Gardon gauge measurements of fast heat flux transients. Appl. Therm. Eng. 100, 501–507. Gardon, R., 1953. An instrument for the direct measurement of intense thermal radiation. Rev. Sci. Instrum. 24 (5), 366–370. Granqvist, C.G., 1991. Solar-energy materials—overview and some examples. Appl. Phys. A: Mater. 52, 83–93. Guillot, E., Alxneit, I., Ballestrin, J., Sansand, L., Willsh, C., 2014. Comparison of 3 heat flux gauges and a water calorimeter for concentrated solar irradiance measurement. Energy Procedia 49, 2090–2099. Han, H., 1989. Spectrally Selective Tinoxide and Indiumoxide Coatings. Bibliotheek Technische University, pp. 1–57. Hausen, H., 1943. Darstellung des Warmeuberganges in Rohren durch verallgemeinerte Potenzbeziehungen. Z. VDI Beih. Verfahrenstech 4, 91–98. Heaton, H.S., Reynolds, W.C., Kays, W.M., 1964. Heat transfer in annular passages. Simultaneous development of velocity and temperature fields in laminar flow. Int. J. Heat Mass Transf. 7 (7), 763–781. Hoare, J.P., LaBoda, M.A., 1985. Fluoride-based catalysts for the electrodeposition of chromium. Electrochem. Soc. 132 (4), 798–801. Holloway, P.H., Shanker, K., Pettit, R.B., Sowell, R.R., 1980. Oxidation of electrodeposited black chrome selective solar absorber films. Thin Solid Films 72 (1), 121–128. Huang, W., Sun, L., 2016. Solar flux density calculation for a heliostat with an elliptical Gaussian distribution source. Appl. Energy 182, 434–441. Jensen, W.B., 2012. Faraday’s Laws or Faraday’s Law. J. Chem. Educ. 89 (9), 1208–1209. Kaluza, J., Neumann, A., 2001. Comparative measurements of different solar flux gauge types. J. Sol. Energy Eng. 123 (3), 251–255. Kennedy, C., 2002. Review of mid- to high- temperature solar selective absorber materials Technical report, Report No: NREL/TP-520-31267, National Renewable Energy Laboratory, pp. 1–58. Lai, M., Riley, D.J., 2008. Templated electrosynthesis of nanomaterials and porous structures. J. Colloid Interface Sci. 323 (2), 203–212. Lampert, C.M., 1979. Chemical, structural and optical characterization of a black chrome solar selective absorber (No. LBL-9123), California Univ., Berkeley (USA). Lawrence Berkeley Lab. Langmuir, I., 1913. The vapor pressure of metallic tungsten. Phys. Rev. 2, 329–342. Lawson, D.I., Mcgurien, J.H., 1953. A radiation calorimeter for the absolute measurement of radiation intensities between 0.4 and 12.5 W/cm2. Fire Safety Sci. 37 (1953), 1–8. Lide, D.R., 2007. CRC Handbook of Chemistry and Physics, 84th Edition 2003–2004. CRC Press. McDonald, G.E., 1974. Spectral Reflectance Properties of Black Chrome for use as a Solar Selective Coating, NASA TM X-71596. McDonald, G.E., 1975. Spectral reflectance properties of black chrome for solar collector use. Sol. Energy 17, 119–122.

209