You are on page 1of 3

Absorptivity of several metals at 10.

6 m: empirical
expressions for the temperature dependence
computed from Drude theory

Graham S. Arnold

Simple expressions for the temperature-dependent absorptivity at 10.6 m have been computed for silver,
aluminum, gold, copper, lead, and tungsten by means of a straightforward application of the Drude model
and experimental dc conductivity data over a wide temperature range. The results of these computations
are in reasonable agreement with experimental data where such are available.

1. Introduction where Eois the permittivity of free space, e is the electron


An important quantity governing the effects of laser charge, N is the free electron density, oOis the material's
irradiation of metals is the fraction of incident light electrical conductivity, and m* is the optical effective
deposited in the solid. Since data on the tempera- mass of the electron.2 It is from the temperature de-
ture-dependent absorptivity [A(T)] of solids are scarce, pendence of the conductivity that an implicit temper-
it is useful to have some simple approximate expressions ature dependence of the reflectivity is formally intro-
for A (T) which can be used in the modeling of laser in- duced.
teractions with metals. 11. Computations
For an opaque solid, the absorptivity at normal in-
cidence is given by' In practice, the quantities copand r are parametrized
or are themselves considered parameters whereby one
may cause the Drude model to describe the optical
(1 + n) 2 + k2 properties of a metal.2 Ordal et al. recently presented
where R is the solid's reflectivity, and n and k are the such a parametrization of the wavelength dependence
index of refraction and extinction coefficient, respec- of the optical constants of a number of metals in the
tively. For conducting materials in the longwave in- infrared and far infrared.4 A similar approach is used
frared, a simple free electron model, the Drude theory, here to calculate the temperature dependence of the
has been useful in calculating metal absorptivities.2 absorptivity at 10.6 Am of silver, aluminum, gold, cop-
Normally this theory is expressed in the equations3 per, lead, and tungsten.
n2- = 1- 2/((0 2
+ -2), (2) Equations (2) and (3) may be solved under the re-
quirement that 0p and r are real and positive to yield
2 2 2
2nk w /[Wr(0 + T- )], (3) (n 2 - 2n + k2 + 1)1/ 2 (n2 + 2n + k2 + 1)1/2W
=

(6)
where w is the angular frequency of the incident light, (k 2 - n 2 + 1)1/2
copis the plasma frequency of the material, and i- is the 1+ k2
-n2
electron relaxation time. In mks units, the plasma 2knw
(7)
frequency and relaxation time are given by
Values of copand r at room temperature were calculated
[Ne2 1/2
using Eqs. (6) and (7) and the room temperature values
cop= ~ , (4)
(m* exO
of n and k at or about 10gimas tabulated by Ordal et al.
N
1T
= 1 (5) The plasma frequency was considered to be indepen-
dent of temperature below the melting point and to
change as the square root of the fractional change in
molar volume upon melting [compare Eq. (8)]. The
variation in the relaxation time was obtained from the
temperature dependence of the bulk dc conductivity of
the metal.5 - 8
Equations (2) and (3) can be solved under the re-
The author is with Aerospace Corporation, Chemistry Physics quirement that n and k are real and positive to provide
Laboratory, P.O. Box 92957, Los Angeles, California 90009. the expressions
Received 17 October 1983.
0003-6935/84/091434-03$02.00/0. n I |( - Q)2 + ( Q) I] _ Q + 1l/2 (8)
© 1984 Optical Society of America.

1434 APPLIED OPTICS/ Vol. 23, No. 9 / 1 May 1984


100 A(T) = E a[T(kelvin)]i (11)
i=o
were fit to the calculated absorptivities. A quadratic
0.10 was required to describe the calculated absorptivity of
tungsten; straight-line segments (below and above the
melting points) were adequate for the other five metals.
Results of these fits are shown in Fig. 1 and summarized
in Table I.
G;
0.05 111. Discussion
Ag Because the plasma frequency and relaxation time
oAl at room temperature were constrained to produce the
o Au
Cu (average) room temperature values of n and k, agree-
* Pb ments between these computed absorptivities and ex-
W
perimentally measured values are quite good. Table
0 1000 2000 II compares the absorptivity at 295 K from the least-
TEMPERATURE
Kelvin) squares fits with the results of several measurements at
Fig. 1. Temperature dependence of the absorptivities of Ag, Al, Au, room temperature.
Cu, Pb, and W at 10.6 Aimcalculated from Drude parameters obtained Decker and Hodgkin have measured the temperature
from room temperature complex indices of refraction and tempera- dependence of the infrared reflectivities of solid alu-
ture-dependent bulk dc electrical conductivities. minum, copper, and silver.9 The measurement range
extended to 9 ptm. Table III compares the slopes of the
k= [1( - Q) 2 + (Q) /2 + Q - 1}1/2, (9) lines fit to the absorptivities computed here with their
experimentally determined values. Considering the
where simple nature of the model used here and the uncer-
Q = W2/(W2
+ T-2). (10)
tainty in the experimental observations, the agreement
for aluminum and silver is quite good. This comparison
Temperature-dependent absorptivities were calcu- for copper drives one to prefer the parametric expres-
lated using Eqs. (1), (8), and (9) and the prescription sion for the absorptivity of solid copper obtained using
presented above for the variation of copand r. Figure the parameters of Ordal et al.
1 shows the results of these computations. These cal- Hodgson has tabulated values of n*, the effective
culations were repeated using the room temperature number of electrons per atom, and the electron collision
values of copand - which Ordal et al. judged to produce frequency for several molten metals including alumi-
the best eyeball fit to log-log plots of the real and num, lead, and silver.14 From these quantities one may
imaginary parts of the metals' dielectric functions vs obtain values for wp and -. Table IV compares Hodg-
frequency. Only in the case of copper were there sig- son's values and the absorptivity one calculates from
nificant differences between the two sets of results. them with the results of the computations reported
Polynomials of the form here. For aluminum and silver the agreement is ex-

Table I. RoomTemperature Drude Parameters Used to Calculate 10.6-AimAbsorptivitiesof Metals (Second Entry in Each Case is the Best-Fit Value of
4
Ordal e a. ) and CoefficientsObtainedby Least-SquaresFits of the Form 100 X A( T) = asT'; Notation: 1.2(3) = 1.2 X 103

Up/1016 7TX 1014


Metal (rad/sec) (sec) Phase ao a, a2
Ag 1.43 3.49 solid -6.88(-3) 1.59(-3)
liquid 1.30(0) 1.64(-3)
1.37 3.66 solid -7.03(-3) 1.59(-3)
liquid 1.30(0) 1.66(-3)
Al 2.23 0.801 solid 1.47(-1) 3.20(-3)
liquid 2.91(0) 1.76(-3)
2.24 0.821 solid 1.34(-1) 3.13(-3)
liquid 2.88(0) 1.70(-3)
Au 1.37 2.47 solid 9.43(-2) 2.34(-3)
liquid 3.46(0) 1.48(-3)
1.32 2.46 solid 9.54(-2) 2.36(-3)
liquid 3.46(0) 1.50(-3)
Cu 1.17 1.25 solid 1.92(-1) 3.88(-3)
1.20 1.91 solid -2.83(-3) 2.59(-3)
Pb 1.12 0.322 solid 7.16(-1) 1.25(-2)
liquid 9.93(0) 2.47(-3)
1.17 0.366 solid 4.67(-1) 1.14(-2)
liquid 8.88(0) 2.25(-3)
W 0.869 1.29 solid -2.50(-1) 7.29(-3) -9.25(-7)
0.912 1.23 solid -2.10(-1) 7.16(-3) -9.09(-7)

1 May 1984 / Vol. 23, No. 9 / APPLIED OPTICS 1435


Table II. Comparisonof RoomTemperature Absorptivitlesfrom Least- factors of 1.70,1.45, and 1.63 for silver, aluminum, and
Squares Fits with ExperimentalValues (Literature Values at 10 Aim)
gold, respectively.
Metal This work Experimental Reference
IV. Summary
Ag 0.0046 0.0047 12
Al 0.011 0.0124 2 Simple expressions for the temperature-dependent
0.01 9 absorptivity at 10.6Am have been computed for silver,
0.013 10 aluminum, gold, copper, lead, and tungsten by means
0.0182 11 of a straightforward application of the Drude model and
Au 0.0078 0.006 10
0.0061 12 experimental dc conductivity data over the temperature
Cu 0.0076-0.013 0.0077 9 range of interest. The results of these computations are
0.011 10 in reasonable agreement with experimental data where
0.0114 11 such are available. These simple expressions and the
Pb 0.044
W 0.018
prescription presented for obtaining them should
comprise a useful tool for the modeling and analysis of
the interaction of infrared radiation with metal sur-
Table Ill. Comparisonof Temperature Dependenceof Reflectivitlesfrom
faces.
Least-Squares Fits with ExperimentalValues a

-dR/dT X 105 (K-1) This work was supported by the Aerospace Sponsored
Decker and Hodgkin Research Program.
Metal This work (at 9 Am)
References
Ag 1.59 2 0.7
Al 3.20 3 + 0.5 1. G. R. Fowles,Introduction to Modern Optics (Holt, Rinehart &
Cu 2.59-3.88 1.4 ± 0.5 Winston, New York, 1968), p. 168.
2. H. E. Bennett, N. Silver, and E. J. Ashley, J. Opt. Soc. Am. 53,
a Tabulated experimental value for silver is the average of 2.2 + 0.7 1089 (1963).
for a temperature change from 150 to 52.30 C and 1.8 ± 3.5 for a tem-
3. Ref. 1, p. 163.
perature change from 52.3 to 31.70 C. Tabulated experimental value
for copper is for a temperature change from 73 to 25.70 C after heating 4. M. A. Ordal et al., Appl. Opt. 22, 1099 (1983).
to 1600C.
13
Experimental uncertainty for copper is taken from Ref. 5. J. Babiskin and J. R. Anderson, in The American Institute of
9; uncertainty for aluminum is arbitrarily assigned to that of Physics Handbook, D. E. Gray, Ed. (McGraw-Hill, New York,
copper. 1972), pp. 9-39-9-40.

Table IV. Drude Parameters and Calculated Absorptivitlesfor Molten Metals a

Metal Reference T(K) wp(rad/sec) -(sec) A(T)


Ag This work 1273 1.39(16) 3.48(-15) 0.034
13 1.42(16) 3.47(-15) 0.033
Al This work 1173 2.12(16) 9.9(-16) 0.050
13 2.10(16) 9.02(-16) 0.054
Pb This work 843 1.10(16) 6.4(-16) 0.12
13 2.18(16) 2.26(-16) 0.11
a Relaxation time values for this work were obtained from straight-line interpolation or extrapolation
from temperatures nearest to those reported in Ref. 12; A(T) from least-squares fits.

cellent for all three quantities. Although the cp and


-r values for lead do not agree, the computed absorp- 6. A. Goldsmith et al., Handbook of Thermophysical Properties
of Solid Materials, Vol. 1 (MacMillan, New York, 1961).
tivities do.
7. R. Lyon, Ed., The Liquid Metals Handbook, The Atomic Energy
Sparks and Loh have used a similar approach to ob- Commission and Bureau of Ships, Department of the Navy
tain temperature-dependent absorptivities for silver, (1954), pp. 41-43.
gold, aluminum, and copper.15 They did not, however, 8. W. Hoffman,Lead and Lead Alloys, Properties and Technology
use experimental data for the temperature-dependent (Springer, Berlin, 1970), p. 20.
conductivity of aluminum and gold. Rather, they ob- 9. D. L. Decker and V. A. Hodgkin, in Proceedings, Symposium on
tained the temperature dependence of T for aluminum Laser Induced Damage in OpticalMaterials, Boulder, Colo.,30
and gold by analogy to the calculated variations for Sept.-1 Oct. 1980 (National Bureau of Standards, Oct. 1981), p.
silver and copper. Even with these approximations, the 190.
values of al obtained by Sparks and Loh for the solid 10. G. Hass, in Applied Optics and Optical Engineering, Vol. 3, R.
metals are in reasonable agreement with those reported Kingslake, Ed. (Academic, New York, 1965), p. 309.
11. J. J-G. Hsia, Ph.D. Thesis, Purdue U. (1968).
here. They made no attempt to model a temperature 12. J. M. Bennett and E. J. Ashley, Appl. Opt. 4, 221 (1965).
dependence of the molten metals assuming, rather, that 13. V. A. Hodgkin, Naval Weapons Center; private communica-
the increase in absorptivity upon melting for silver, tion.
aluminum, and gold is a factor of 1.9, which they de- 14. J. N. Hodgson, in Liquid Metals, Chemistry and Physics, S. Z.
scribe as the factor by which the resistivity of silver in- Beer, Ed. (Marcel Dekker, New York, 1972), p. 345.
creases upon melting. The least-squares fits reported 15. M. Sparks and E. Loh, Jr., J. Opt. Soc. Am. 69, 847 (1979).
here predict increases in absorptivity upon melting by 16. M. Sparks and E. Loh, Jr., J. Opt. Soc. Am. 69, 859 (1979).

1436 APPLIED OPTICS/ Vol. 23, No. 9 / 1 May 1984

You might also like