Sunteți pe pagina 1din 9

Angular solar absorptance of

absorbers used in solar thermal collectors

Tuquabo Tesfamichael and Ewa Wäckelgård

The optical characterization of solar absorbers for thermal solar collectors is usually performed by
measurement of the spectral reflectance at near-normal angle of incidence and calculation of the solar
absorptance from the measured reflectance. The solar absorptance is, however, a function of the angle
of incidence of the light impinging on the absorber. The total reflectance of two types of commercial
solar-selective absorbers, nickel-pigmented anodized aluminum, and sputtered nickel兾nickel oxide coated
aluminum are measured at angles of incidence from 5° to 80° in the wavelength range 300 –2500 nm by
use of an integrating sphere. From these measurements the angular integrated solar absorptance is
determined. Experimental data are compared with theoretical calculations, and it is found that optical
thin-film interference effects can explain the significant difference in solar absorptance at higher angles
for the two types of absorbers. © 1999 Optical Society of America
OCIS codes: 160.1190, 260.3160, 310.1620, 310.6860, 310.6870, 350.6050.

1. Introduction dielectric matrix determines the durability of the ab-


Solar-selective absorbers for solar collectors absorb sorber by protecting the metal particles and the
incident solar radiation and convert it into thermal metal substrate from being oxidized.7–10 Because
energy with minimum thermal radiation losses. the metal particles are smaller than the wavelength
The possibility of producing practical selective ab- of the solar radiation, effective-medium theory can be
sorber surfaces was first shown in mid-1950’s. The applied to model the optical properties of solar-
first selective surfaces were based on coatings of some selective coatings.6
oxides and sulfides deposited on a metal substrate.1–3 The optical characterization of solar absorbers is
The most common type of selective absorber is a tan- usually performed by measurement of the spectral
dem absorber obtained when a highly solar absorbing reflectance at near-normal angle of incidence and de-
thin coating is deposited onto a highly infrared re- termination of the solar absorptance from the mea-
flecting metal substrate. The high absorptance can sured reflectance. Because the position of the Sun
be achieved with composite coatings containing small varies daily and seasonally, knowledge about the
metal particles in a dielectric matrix. The thin- angular-dependent solar absorptance of the absorb-
metal– dielectric coatings absorb the incoming solar ers is important but such data are not available for
radiation and transmit in the infrared wavelength most commercial absorbers.11,12 We have performed
range. Particle size, shape, and distribution in the reflectance measurements at variable angles of inci-
matrix control the high-absorptance to high- dence on two types of commercial selective coatings
transmittance crossover. The optical constants of on aluminum: nickel-pigmented aluminum oxide
the metal and the dielectric used in the coating are and sputtered nickel兾nickel oxide. The first-
also important for the optical performance.4 – 6 The mentioned coating is produced by an electrochemical
technique whereas the second is obtained by vacuum
deposition. The two coatings are deposited on the
same rolled aluminum substrate, but they differ in
T. Tesfamichael (tuquabo.tesfamichael@angstrom.uu.se) and E. layer structure and matrix oxide. Both coatings
Wäckelgård (eva.wackelgard@angstrom.uu.se) are with Solid
contain particles of nickel but of different shapes and
State Physics, Department of Materials Science, Uppsala Univer-
sity, Box 534, S-751 21 Uppsala, Sweden. distribution. We therefore make the hypothesis
Received 15 December 1998; revised manuscript received 13 that the difference in angular solar absorptance orig-
April 1999. inates from the difference in microstructures. We
0003-6935兾99兾194189-09$15.00兾0 have made a thorough theoretical investigation by
© 1999 Optical Society of America calculating the angular solar absorptance for differ-

1 July 1999 兾 Vol. 38, No. 19 兾 APPLIED OPTICS 4189


ent types of microstructures in order to sort out the
key features that govern the angular behavior of so-
lar absorptance of tandem solar absorbers.

2. Samples and Optical Measurements


We prepare nickel-pigmented aluminum oxide by
first anodizing an aluminum plate in diluted phos-
phoric acid solution to produce a porous alumina film,
followed by electroplating metallic nickel particles
within the pores from a nickel salt aqueous solution.
The nickel particles are formed as rods, ⬃30 nm in
average diameter, which start to grow from the pore
base13,14 and possess a volume fraction of ⬃30%.15
Commercial absorbers from TeknoTerm industry in
Sweden have a total coating thickness of ⬃700 nm
with a nickel-pigmented base layer 300 nm thick and
a nonpigmented alumina top layer of ⬃400 nm. The
substrate is a rolled aluminum sheet with pro-
nounced grooves from the rolling that are also
present at the front surface of the coating. The mi-
crostructure of the coating is shown in the inset of
Fig. 1共a兲.
The nickel兾nickel oxide coating is deposited by re-
active dc magnetron sputtering in a roll-coating de-
vice. The total thickness is ⬃200 nm, including a
top layer of an antireflection coating, ⬃50 nm thick.
It consists of a mixture of nickel and nickel oxide.
The oxygen inlet during the sputtering is designed to
give a gradually increasing amount of nickel oxide
from the substrate to the front surface. This pro-
duces a grading in the optical refractive index
through the coating.16 The structure of the sput-
tered nickel兾nickel oxide coating is shown in the inset
of Fig. 3共a兲 in Section 3.
Both the chemically produced and the vacuum-
produced commercial samples were cut in pieces of 5
cm ⫻ 5 cm for optical measurements. The rolled
aluminum substrate has a surface roughness that
scatters ⬃10% of the reflected light in nonspecular Fig. 1. Reflectances 共a兲 Rp and 共b兲 Rs for a nickel-pigmented alu-
directions, and it is therefore important to use an minum oxide coating as a function of wavelength in the solar
optical measuring system that is also able to detect spectrum recorded at different angles of incidence from near nor-
diffusely reflected light. The total spectral reflec- mal to 80°.
tance was therefore measured in a system equipped
with an integrating sphere. Both p- and s-polarized
light were used, and the corresponding polarized re- where I共␭兲 is the direct normal solar irradiance at air
flectances Rp and Rs were measured. The sample mass 1.5, i.e., for the Sun 48° from the zenith.18 Air
holder in the central of the sphere can be rotated so mass variation shows only small differences in the
that nonnormal reflectance can be measured between solar-weighted optical properties of selective materi-
5° and 80° in the UV–visible–near-IR region 共300 to als.19 The integration was done over the measured
2500 nm兲. The instrument and the measuring rou- wavelength range 共␭1 ⫽ 300 nm to ␭2 ⫽ 2500 nm兲.
tines are described in detail elsewhere.17
The angular solar absorptance was calculated from 3. Experimental Results
the measured reflectance with Figure 1 presents the p- and the s-polarized spectral
reflectances Rp and Rs of the commercial absorber of
nickel-pigmented anodized aluminum. For clarity


␭2 we have shown some selected angles of incidence.
兵1 ⫺ 0.5关Rp共␪, ␭兲 ⫹ Rs共␪, ␭兲兴其I共␭兲d␭ Because of the coating thickness, optical thin-film
␭1 interference within the measured wavelength range
␣␪ ⫽ , (1)


␭2 is inevitable.20 The interference maxima and min-
I共␭兲d␭ ima become very pronounced at high angles of inci-
␭1 dence. A small shift of the maxima and the minima

4190 APPLIED OPTICS 兾 Vol. 38, No. 19 兾 1 July 1999


Fig. 2. Reflectances Rp and Rs for the nickel-pigmented alumi- Fig. 3. Reflectances 共a兲 Rp and 共b兲 Rs for a sputtered nickel兾nickel
num oxide coating shown in Fig. 1 versus angles of incidence for oxide coating as functions of wavelength in the solar spectrum
two different wavelengths: 共a兲 500 nm, 共b兲 1200 nm. The non- measured at different angles of incidence from near normal to 80°.
polarized reflectance R ⫽ 共Rp ⫹ Rs兲兾2 is also shown.

larger than Rp for all higher angles of incidence,


to shorter wavelengths can be noted for the reflec- which is observed at the 500-nm wavelength. At
tance curves from 50° to 80°. This is because the 1200 nm the component of the p-polarized reflectance
layer at higher angles of incidence corresponds to a is higher, and this is caused by the thin-film inter-
coating with a thinner effective thickness, which is ference effect. As seen in Fig. 1共b兲, the s-polarized
characteristic behavior of optical interference coat- reflectance has a deep interference minimum at 1200
ings.21 Another feature of thin-film interference is nm, which makes it lower than the p-polarized reflec-
that reflectance minima and maxima do not occur at tance in Fig. 1共a兲, which has a shallow minimum at
the same wavelengths for p- as for s-polarized light.21 1000 nm.
This effect is also clearly seen in the curves of high- The p- and the s-polarized spectral reflectances of
angle spectral reflectance in Fig. 1. the sputtered nickel兾nickel oxide coating are shown
Figure 2 shows angular polarized reflectance for in Fig. 3 for different angles of incidence. Here the
two different wavelengths, 500 and 1200 nm. The interference pattern is suppressed by having a
data are extracted from the wavelength-dependent p- graded-index layer with an antireflection coating on
and s-polarized angular reflectances in Fig. 1. At top. The antireflection layer is necessary to reduce
near normal both the p- and the s-polarized reflec- front-surface reflection.16 Figure 4 presents the an-
tances coincide, but for higher angles the curves split. gular reflectances Rp and Rs for two different wave-
The normal case, as for bulk reflectance, is that Rs is lengths 共500 and 1200 nm兲. At near normal both Rp

1 July 1999 兾 Vol. 38, No. 19 兾 APPLIED OPTICS 4191


Fig. 5. Solar absorptance of nickel-pigmented aluminum oxide
and sputtered nickel兾nickel oxide commercial absorber coatings
determined from Figs. 1 and 3 versus angles of incidence from 5°
to 80°.

the anodized ones. Comparing Figs. 2 and 4, we can


observe that the polarization split between p and s
reflectance is larger for the sputtered nickel兾nickel
oxide absorber than for the nickel-pigmented anod-
ized aluminum coating.
From the results of Figs. 1 and 3, we have obtained
the solar absorptances of both coatings by using Eq.
共1兲. The solar absorptance of the nickel-pigmented
aluminum oxide absorber is constant over a wide
range of angles of incidence. It absorbs ⬃0.95 of the
solar radiation from near normal up to 50° but the
absorptance decreases steeply to ⬃0.70 when ap-
proaching 80° 共see Fig. 5兲. The solar absorptance of
Fig. 4. Reflectances Rp and Rs for the sputtered nickel兾nickel the nickel兾nickel oxide absorber decreases gradually
oxide coating shown in Fig. 3 versus angles of incidence for two from near normal up to 40°, and beyond this angle it
different wavelengths: 共a兲 500 nm, 共b兲 1200 nm. The nonpolar-
drops steeply, as shown in Fig. 5. The solar ab-
ized reflectance R ⫽ 共Rp ⫹ Rs兲兾2 is also shown.
sorptance of the sputtered nickel兾nickel oxide ab-
sorber begins to decrease at lower angles compared
with that of the nickel-pigmented aluminum oxide
and Rs have the same value and split as the angle of absorber. The angular solar absorptances are sig-
incidence increases, similar to bulk materials. With nificantly different for the two types of coatings.
increasing angle of incidence, we can see that Rs in- A significant gain in solar absorptance can be ob-
creases continuously but Rp first decreases until the served at higher angles of incidence because of surface
pseudo-Brewster angle and then increases at higher roughness of the coatings. As mentioned in Section 2,
angles. This trend is found to be the same for all rolled aluminum substrates were used for the two ab-
measured wavelengths for this type of coating. This sorbers. The rolling creates parallel grooves in the
is clear also from comparing Figs. 3共a兲 and 3共b兲. aluminum surface, which scatter the reflected light in
When the spectral angular reflectances of the two directions perpendicular to the direction of the rolling
types of absorbers are compared, it is found that thin- grooves.22 This one-dimensional texture introduces
film interference patterns dominate the spectral be- orientation-dependent reflectance, i.e., orientation-
havior of the two-layer-structure nickel-pigmented dependent solar absorptance. This is experimentally
alumina absorber while the graded-index nickel兾 observed when the absorbers are measured with
nickel oxide absorber has smoother reflectance groove orientation parallel or perpendicular to the
curves. It is also observed that the average reflec- plane of the incident light. As shown in Fig. 6, a
tance over the measured wavelength range is larger noticeable difference in solar absorptance at higher
at higher angles for the sputtered absorber than for angles was obtained when the groove orientation was

4192 APPLIED OPTICS 兾 Vol. 38, No. 19 兾 1 July 1999


smooth, and the effect of surface roughness on ab-
sorptance, polarization, or coating interference was
not considered. The optical constants of nickel,
nickel oxide, alumina, and the aluminum substrate
were taken from the literature.24 –27

A. Nickel-Pigmented Anodized Aluminum


We have used two different shapes, cylindrical and
spherical, for the nickel particles and two types of
layer structure 共two or seven layers兲 for the alumina
matrix. In the two-layer structure the thickness of
each layer was varied. The seven-layer calculation
was applied to model a coating with a graded-index
structure.

1. Two Layers, Model versus Experimental


This model structure is based on the experimental
characterization of nickel-pigmented anodized alumi-
num, as described in Section 2. We considered two
Fig. 6. Solar absorptance of nickel-pigmented aluminum oxide
layers of inhomogeneous media. A base layer, of
coating as a function of angle of incidence from 5° to 80°. Mea-
surements are shown for two different sample rolling orientations
thickness d2 ⫽ 300 nm, consists of nickel cylindrical
共grooves兲 of the aluminum substrate: perpendicular and parallel rods in alumina, and a top plain alumina layer, of
to the plane of the incident light. thickness d1 ⫽ 200 nm, contains cylinder-shaped air-
filled pores. The volume fraction of both the nickel
particles and the air-filled pores is 0.3, and the long
perpendicular to the plane of the incident light com- axis of the cylinders is perpendicular to the sample
pared with parallel-oriented grooves. The measure- surface 关see also inset of Fig. 1共a兲兴.
ments presented in Fig. 6 come from a nickel- The calculated nonpolarized reflectance R ⫽ 共Rs ⫹
pigmented anodized aluminum sample, but the effect Rp兲兾2 is compared with the experimental nonpolar-
is also seen for the nickel兾nickel oxide coated alumi- ized reflectance of the nickel-pigmented anodized alu-
num absorber with the same rolled aluminum sub- minum absorber, as shown in Fig. 7. The calculated
strate. It is therefore advantageous to install a solar reflectance 关Fig. 7共a兲兴 is higher than the experimental
collector with grooves of the absorber in the north– reflectance 关Fig. 7共b兲兴. A smoothing effect of exper-
south direction to gain more solar energy in the morn- imental spectra is caused by thickness variation
ings and afternoons. The small difference recorded at across the illuminated area on the sample, but this is
⬃5° angle of incidence in Fig. 6 is an artifact in which not observed from the calculations. The optical con-
light escapes from the integrating sphere through the stants used in the calculations as well as the assump-
entrance port when the grooves are not oriented par- tion of the smooth surface and interface of the coating
allel to the plane of incidence. This is because some of could also be other factors for obtaining higher cal-
the light that was reflected directly from the coating culated reflectance. However, for higher angles the
surface, as shown in the inset of Fig. 6, can easily position of the reflectance maximum at ⬃1600 nm is
escape through the port when the angle of incidence is reproduced in the calculations as well as the struc-
lower. All results presented in this paper come from ture between 600 and 1600 nm. Since the absolute
measurements with grooves parallel to the plane of the values of the calculated angular solar absorptance
incident light. obtained from the angular reflectance are lower than
those of the measured ones, we have chosen to
4. Theoretical Calculations present the angular solar absorptance normalized to
To investigate the impact of microstructure, the an- its near-normal value in Fig. 8. This makes it easier
gular solar absorptance was obtained from the calcu- to compare the angular dependence of different mod-
lated reflectance for different microstructures. The els with experimental values. The calculated solar
reflectance calculations were based on 2 ⫻ 2 matrix absorptance remains almost constant over a wide
technique by use of a surface impedance and admit- range of angles of incidence, with maximum solar
tance approach.23 The Fresnel amplitude coeffi- absorptance near 40°, and drops steeply beyond 60°,
cients were calculated for each interface in a similar to the experimental data. The existence of
multilayer stack of thin films. The modeling also this maximum is due to the more pronounced inter-
included the calculation of effective optical constants ference effects of the ideal model coating.
for two-component inhomogeneous materials by use
of a Bruggeman effective-medium model. The thick- 2. Two Layers, Different Particle Shapes
ness of the composite coating and the shape and the The next step was to investigate how particle shape
volume fraction of the particles inside the matrix influences the solar absorptance. Figure 9 shows
were varied. The surfaces were assumed to be calculated solar absorptance of a nickel-pigmented

1 July 1999 兾 Vol. 38, No. 19 兾 APPLIED OPTICS 4193


Fig. 8. Calculated solar absorptance versus angles of incidence of
nickel-pigmented aluminum oxide together with the experimental
results shown in Fig. 5. The solar absorptance is normalized to
the value at near-normal angle of incidence.

fect was observed for particles with spherical geom-


etry when the pigmented layer thickness d2 was
changed.
Figure 10共b兲 shows the normalized solar ab-
sorptance for cylindrical particle shape, constant pig-
mented layer 共d2 ⫽ 300 nm兲, and different values of
d1 共50, 200, and 400 nm兲 for the plain alumina top
layer. It is clear from the calculations that when d1
is thin or is thicker than 200 nm the solar ab-
sorptance begins to drop at lower angles of incidence
关Fig. 10共b兲兴. When d1 is changed, the position of the
interference pattern is shifted because of the change

Fig. 7. Unpolarized reflectance spectra R ⫽ 共Rp ⫹ Rs兲兾2 for 共a兲


calculated and 共b兲 experimental nickel-pigmented aluminum oxide
coating as functions of wavelength in the solar spectrum at differ-
ent angles of incidence from near normal to 80°.

aluminum oxide coating containing spherical and cy-


lindrical particles. The thicknesses of the plain and
the pigmented layers are the same in both cases. As
seen in Fig. 9, the particle shape has a noticeable
effect on the angular solar absorptance and the ex-
perimental values are between the solar absorptance
of the two curves representing different particle
shapes.

3. Two Layers, Film Thickness Variation


Variation of the pigmented film thickness d2 has a
small but noticeable effect on the solar absorptance at
higher angles of incidence for cylindrical particles.
Figure 10共a兲 shows the normalized solar absorptance Fig. 9. Calculated normalized solar absorptances 共␣兾␣5°兲 versus
for cylindrical particles versus angle of incidence with angles of incidence of nickel-pigmented aluminum oxide for two
d2 changed from 300 to 500 nm with the top plain different particle shapes 共cylindrical and spherical兲 and constant
alumina layer kept constant 共d1 ⫽ 200 nm兲. No ef- film thickness.

4194 APPLIED OPTICS 兾 Vol. 38, No. 19 兾 1 July 1999


Fig. 11. Calculated normalized solar absorptance 共␣兾␣5°兲 versus
angles of incidence of a graded-index layer of a nickel–alumina
mixture on an aluminum substrate. The curve is shown together
with the result of the nickel-pigmented aluminum oxide double-
layer structure modeled in Fig. 8.

tion of nickel at the front surface of the coating was


chosen to be zero, and it was increased toward the
substrate in steps of 0.1 between subsequent sublay-
ers. Figure 11 presents the angular solar ab-
sorptance of the graded-index structure together with
calculated solar absorptance of the double-layer
structure already shown in Figs. 8 –10. The graded
index with the seven-layer structure has a lower solar
absorptance at higher angles and shows behavior
similar to that of the double layer with the 50-nm
thin plain alumina top layer. The same result was
also obtained for spherical particles in a graded-index
coating.
Fig. 10. Calculated normalized solar absorptances 共␣兾␣5°兲 versus
angles of incidence of nickel-pigmented aluminum oxide for cylin- B. Nickel兾Nickel Oxide Coating
drical geometry 共a兲 at varying nickel–alumina thicknesses d2 and
This model structure is based on the experimental
constant air–alumina layer d1, 共b兲 constant d2 and different thick-
nesses of d1.
characterization of the sputtered nickel兾nickel oxide
coating, described in Section 2. We considered a
graded-index structure of spherical nickel particles in
a nickel oxide matrix with a total thickness of ⬃240
in coating thickness. Since the interference pattern nm. The thickness of the coating was divided into
at higher angles is more pronounced, more incident eight equally thick sublayers with increasing nickel
light can be absorbed when the position of the inter- content in the sublayers going from the front surface
ference minimum is adjusted to the wavelengths of to the substrate. For best fit to the experimental
maximum solar intensity. This means that there is reflectance, the filling factor must be greater than 0.6
an optimum thickness that gives the highest ab- at the substrate–film interface and zero at the film
sorptance at higher angles of incidence. A similar surface. The choice of the total thickness was not
result was also observed for spherical particles. critical enough to alter the results significantly.
There is no significant difference in the results when
4. Graded Index the number of sublayers exceeds eight. The calcu-
We also performed calculations for a graded-index lated nonpolarized reflectance is presented in Fig.
layer of cylindrical nickel particles in aluminum ox- 12共a兲. It shows the same increase with increasing
ide for a total film thickness of 500 nm. The film was angle of incidence as the experimental one in Fig.
divided into seven equally thick sublayers of 71 nm 12共b兲. The corresponding theoretical and experi-
and with increasing nickel content from the surface mental solar absorptances are shown in Fig. 13. It
toward the aluminum substrate. The volume frac- is seen from the figure that the angular solar ab-

1 July 1999 兾 Vol. 38, No. 19 兾 APPLIED OPTICS 4195


Fig. 13. Calculated and experimental normalized solar ab-
sorptances versus angles of incidence for sputtered nickel兾nickel
oxide coating determined from Fig. 12.

results in a lower average reflectance in the solar


range for the nickel-pigmented anodized aluminum
than for the nickel兾nickel oxide coating. It follows
that the solar absorptance of nickel兾nickel oxide
coated aluminum is lower than that for the nickel-
pigmented anodized aluminum at angles of incidence
higher than 40°.
The reason for this angular behavior has been ex-
plained by the results from theoretical modeling.
Both coatings consist of small nickel particles dis-
persed in an oxide matrix. It was found from the
calculations that the choice of oxide matrix, nickel
oxide or alumina, did not influence the angular be-
havior. The shape of the particles, either spherical
or cylindrical, did not alter the angular absorptance
Fig. 12. Unpolarized reflectance spectra R ⫽ 共Rp ⫹ Rs兲兾2 for 共a兲 significantly. It has been demonstrated that the so-
calculated and 共b兲 measured sputtered nickel兾nickel oxide coatings lar absorptance at higher angles of incidence mainly
as functions of wavelength in the solar spectrum at different angles depends on the layer structure. Thin-film interfer-
of incidence from near normal to 80°. ence effects in the solar wavelength range govern the
high-angle behavior of the solar absorptance. In a
graded-index coating with metal filling factors be-
sorptance of the model structure fits the experimen- tween 0.6 and 1.0 at the coating–substrate interface
tal results well, despite the difference in surface the interference is suppressed. The double-layer
smoothness. structure of the nickel-pigmented alumina coating
has a constant and lower filling factor in the nickel-
5. Discussion and Conclusions pigmented base layer, thereby causing interference
Optical characterization of two types of commercial with reflected light from both the front surface and
solar-selective absorbers as a function of angle of in- the plain oxide–nickel-pigmented oxide interface. It
cidence from 5° to 80° has been performed by mea- was found that the thickness of the plain alumina
surement of the angular spectral reflectance from 300 layer was crucial for the angular solar absorptance,
to 2500 nm by use of an optical measuring system which was not the case for the nickel-pigmented base-
equipped with an integrating sphere. A strong in- layer thickness. The position of the reflectance in-
terference pattern that was due to thin-film interfer- terference minimum in the solar range should be
ence was observed in the measured spectral made to coincide with the maximum intensity of the
reflectance of the nickel-pigmented anodized alumi- solar spectrum by a proper choice of the thickness of
num absorber but not for the nickel兾nickel oxide the plain alumina top layer and in order to optimize
coated aluminum one. The interference effect was the solar absorptance. Because of the contribution
seen to increase at higher angles of incidence, which of thin-film interference, the nickel-pigmented alumi-

4196 APPLIED OPTICS 兾 Vol. 38, No. 19 兾 1 July 1999


num oxide absorber is therefore found to have better 12. R. B. Pettit and R. R. Sowell, “Solar absorptance and emittance
solar optical performance than the sputtered nickel兾 properties of several solar coatings,” J. Vac. Sci. Technol. 13,
nickel oxide absorber at higher angles of incidence. 596 – 602 共1976兲.
13. E. Wäckelgård, “A study of the optical properties of nickel-
We acknowledge Arne Roos for reading the manu- pigmented anodic aluminum in the infrared region,” J. Phys.
script and interesting discussions and Per Nostell for Condens. Matter 8, 5125–5138 共1996兲.
the valuable assistance in the measurements. T. 14. G. H. Pontifex, P. Zhang, Z. Wang, T. L. Haslett, D. AIMaw-
Tesfamichael thanks the University of Asmara, Eri- lawi, and M. Moskovits, “STM imaging of the surface of small
trea, and the International Science Program of Upp- metal particles formed in anodic oxide pores,” J. Phys. Chem.
sala University for providing the opportunity to study 95, 9989 –9993 共1991兲.
at Uppsala University. This work was financially 15. S. Nakamura, M. Saito, L.-F. Hunga, M. Miyagi, and K. Wada,
supported by the Swedish Agency for Research Co- “Infrared optical constants of anodic alumina films with mi-
operation with Developing Countries, the Swedish cropore arrays,” Jpn. J. Appl. Phys. 31, 3589 –3593 共1992兲.
Building Research Council, and the Swedish Na- 16. E. Wäckelgård and G. Hultmark, “Industrially sputtered solar
absorber surface,” Solar Energy Mater. Solar Cells 54, 165–
tional Science Research Council.
170 共1998兲.
References 17. P. Nostell, A. Roos, and D. Rönnow, “Single beam integrat-
ing sphere spectrophotometer for R- and T-measurements
1. H. Tabor, “Selective radiation II. Wavelength discrimina-
tion,” Bull. Res. Counc. Isr. Sect. A: Chem. 5, 129 –134 共1956兲. versus angle of incidence in the solar wavelength range on
2. H. Tabor, “Selective radiation I. Wavelength discrimination,” diffuse and specular samples,” Rev. Sci. Instrum. 70, 2481–
Bull. Res. Counc. Isr. Sect. A: Chem. 5, 119 –128 共1956兲. 2494 共1999兲.
3. J. T. Gier and R. V. Dunkle, “Selective spectral characteristics 18. 9845-1 ISO, “Solar energy—reference solar spectral irradiance
as an important factor in the efficiency of solar collectors,” in at the ground at different receiving conditions-” 共International
Transactions of the Conference on the Use of Solar Energy Organisation for Standardization, Geneva, Switzerland,
共University of Arizona, Tucson, Ariz., 1955兲, Vol. 2, pp. 41–56. 1992兲.
4. C. M. Lampert, “Coatings for enhanced photothermal energy 19. M. A. Lind, R. B. Pettit, and K. D. Masterson, “The sensitivity
collection,” Solar Energy Mater. 2, 1–17 共1979兲. of solar transmittance reflectance and absorptance to selected
5. Å. Andersson, O. Hunderi, and C. G. Granqvist, “Nickel pig- averaging procedures and solar irradiance distributions,”
mented anodic aluminum oxide for selective absorption of solar Trans. Am. Soc. Mech. Eng. 102, 34 – 40 共1980兲.
energy,” J. Appl. Phys. 51, 754 –764 共1980兲. 20. E. Wäckelgård, T. Chibuye, and B. Karlsson, “Improved solar
6. G. A. Niklasson, C. G. Granqvist, and O. Hunderi, “Effective optical properties of a nickel pigmented anodized aluminum
medium models for the optical properties of inhomogeneous selective surface,” in Energy Conservation in Buildings, Pro-
materials,” Appl. Opt. 20, 26 –30 共1981兲. ceedings of NORTHSUN 90, A. A. M. Sayigh ed. 共Pergamon,
7. T. Tesfamichael, S. Andersson, T. Chibuye, and E. Wäckel- Oxford, 1990兲, pp. 177–182.
gård, “Study of oxidation kinetics for metal-dielectric films 21. A. Thelen, Design of Optical Interference Coatings 共McGraw-
using IR optical measurements,” in Optical Materials Technol- Hill, New York, 1989兲, Chap. 2, pp. 23–24 and Chap. 9, pp.
ogy for Energy Efficiency and Solar Energy Conversion XV, 177–178.
C. M. Lampert, C. G. Granqvist, M. Graetze, S. K. Deb, eds., 22. A. Roos, “Use of an integrating sphere in solar energy re-
Proc. SPIE 3138, 197–204 共1997兲. search,” Solar Energy Mater. Solar Cells 30, 77–94 共1993兲.
8. R. Gatt, G. A. Niklasson, and C. G. Granqvist, “Degradation
23. G. B. Smith, “Theory of angular selective transmittance in
modes of Cermet based selective solar absorber coatings,” in
oblique columnar thin films containing metal and voids,” Appl.
Optical Material Technology for Energy Efficiency and Solar
Opt. 29, 3685–3693 共1990兲.
Energy Conversion XI: Selective Solar Absorbers, A. H. Goff,
24. A. P. Lenham and D. M. Treherne, Optical Properties and
C. G. Granqvist, and C. M. Lampert eds., Proc. SPIE 1727,
Electronic Structure of Metals and Alloys 共North-Holland, Am-
87–101 共1992兲.
9. G. A. Niklasson, “Theoretical model for the durability of the sterdam, 1966兲.
solar selective absorber coatings at elevated temperatures,” in 25. P. J. Gielisse, J. N. Plendl, L. C. Mansur, G. R. Marshall, S. S.
Energy and the Environment into the 1990s, A. A. M. Sayigh, Mitra, R. Mykolajewycz, and S. Smakula, “Infrared properties
ed. 共Pergamon, Oxford, 1990兲, Vol. 3, pp. 1372–1376. of NiO and CoO and their mixed crystals,” J. Appl. Phys. 36,
10. E. Wäckelgård, “A comparative study of the optical properties 2446 –2450 共1965兲.
of nickel pigmented alumina films of different thicknesses ex- 26. T. S. Eriksson, A. Hjortsberg, G. A. Niklasson, and C. G. Gran-
posed to elevated temperature and humidity,” Solar Energy qvist, “Infrared optical properties of evaporated alumina
Mater. Solar Cells 54, 171–179 共1998兲. films,” Appl. Opt. 20, 2742–2746 共1981兲.
11. J. A. Duffie and W. A. Beckman, Solar Engineering of Thermal 27. J. H. Weaver, “Handbook of chemistry and physics,” in Optical
Processes, 2nd ed. 共Wiley-Interscience, New York, 1991兲, pp. Properties of Metals, D. R. Lide, ed. 共CRC Press, Boca Raton,
209 –210. Fla., 1993兲, Sect. 12, pp. 111–126.

1 July 1999 兾 Vol. 38, No. 19 兾 APPLIED OPTICS 4197

S-ar putea să vă placă și