Volume 106, Number 5, September–October 2001 Journal of Research of the National Institute of Standards and Technology [J. Res. Natl. Inst. Stand. Technol. 106, 775–779 (2001)] Fraunhofer Diffraction Effects on Total Power for a Planckian Source Volume 106 Number 5 September–October 2001 Eric L. Shirley An algorithm for computing diffraction ef- Key words: diffraction; Fraunhofer; fects on total power in the case of Fraun- Planckian; power; radiometry. National Institute of Standards and hofer diffraction by a circular lens or aper- Technology, ture is derived. The result for Fraunhofer diffraction of monochromatic radiation is Gaithersburg, MD 20899-8441 Accepted: August 28, 2001 well known, and this work reports the re- sult for radiation from a Planckian source. [email protected] The result obtained is valid at all temper- atures. Available online: http://www.nist.gov/jres 1. Introduction Fraunhofer diffraction by a circular lens or aperture is tion, because the detector pupil is overfilled. However, a ubiquitous phenomenon in optics in general and ra- diffraction leads to losses in the total power reaching the diometry in particular. Figure 1 illustrates two practical detector. situations in which Fraunhofer diffraction occurs. In the All of the above diffraction losses have been a subject first example, diffraction limits the ability of a lens or of considerable interest, and they have been considered other focusing optic to focus light. According to geo- by Blevin [2], Boivin [3], Steel, De, and Bell [4], and metrical optics, it is possible to focus rays incident on a Shirley [5]. The formula for the relative diffraction loss lens to converge at a focal point. In practice, even with in spectral power is already well known, and it is noted the focal point at the center of a finite-sized circular below. However, a formula for the diffraction loss in the detector, some of the light incident on the lens fails to total power for the case of a Planckian source such as a be collected, giving rise to a “diffraction loss.” star or blackbody appears to have been given only in the The second example can occur when using a source, high-temperature limit. This article reports a formula such as a blackbody cavity source viewed through a for the diffraction loss in the power for such a source at small “pinhole” aperture, to calibrate optical systems all temperatures. The formula is useful for predicting intended for observing celestial objects or for use in the loss in cases such as the examples discussed, and it remote-sensing applications. In this case, the angle sub- can be used as an independent check for numerically tended by the cavity opening at the aperture is larger calculated diffraction losses. than the angle subtended by the detector pupil on the For the radiation present at a given wavelength ␭, the dark side of the aperture [1]. Because the former angle diffraction loss can be described in terms of a unit-less is larger than the latter one, geometrical optics suggests parameter, v =2␲⌿R/␭. R is the radius of the lens or that the total power detected depends only on the black- aperture. The angle ⌿ is defined in either of two ways: body temperature and a geometrical factor related to the either 2⌿ is the full angle subtended by the detection pinhole aperture, detector optics, and relative separa- pupil at the focusing optic, or 2⌿ is the full angle 775 Volume 106, Number 5, September–October 2001 Journal of Research of the National Institute of Standards and Technology Fig. 1. Two optical systems are shown, for which the (spectral) power detected is subject to Fraunhofer diffraction. In the first system (a), a lens concentrates radiation on the detector. In the second system (b), radiation from a blackbody cavity may pass through the blackbody cavity opening, pass through the pinhole aperture, and reach the detector pupil. subtended by the blackbody cavity opening at the 2. Derivation of Formula pinhole aperture. For a source in thermal equilibrium at temperature T, the effects of temperature on the In the Fraunhofer case, the ratio of the true spectral distribution of spectral power enter relevant equations power to the spectral power predicted by geometrical through the ratio, c2/(␭T) ≡ Av . Here, c2 = hc/k = optics is 1.438 7752(25) ϫ 10Ϫ2 m K is the second radiation con- 2 2 stant, where the number in parenthesis is the one-stan- F(v)=1Ϫ J0(v) Ϫ J1(v). (1) dard-deviation uncertainty in the last two digits. This implicitly defines another unit-less parameter, Here, Jm (v) is a cylindrical Bessel function. It is neces- A = c2/(2␲⌿RT). The relative effects of diffraction on sary to incorporate this wavelength-dependent factor the total power only depend on this one parameter, A. when evaluating the ratio of the total power detected to As a final note, if the detection pupil subtends a the total power predicted by geometrical optics. For a considerable angle, the diffraction losses can be approx- source whose spectral output obeys the Planck distribu- imately given as the weighted average of losses arising, tion, this unit-less ratio is given by in the case of an infinitesimal detector, for values of ⌿ ⌿ ϱ varying between minus the half-angle subtended by 3 Ϫ1 2 2 ͐dvv [exp(Av ) Ϫ 1] [1 Ϫ J0(v) Ϫ J1(v)] the detector and ⌿ plus the same half-angle. Analogous 0 F(A)= ϱ . (2) weighting can be used, in the context of the first exam- Ϫ ͐dvv 3[exp(Av ) Ϫ 1] 1 ple, to account for the finite angular diameter of a source 0 (smaller than 2⌿). Such averaging is discussed else- where [5] but is not discussed further in this work, The value of the denominator is familiar, being given which henceforth assumes a single effective value of ⌿. by 776 Volume 106, Number 5, September–October 2001 Journal of Research of the National Institute of Standards and Technology ϱ ϱ ϱ ͐dvv 3[exp(Av ) Ϫ 1]Ϫ1 = ͚ ͐dvv 3 exp(ϪnAv) High-Temperature Case 0 n=1 0 ϱ =6͚ (nA)Ϫ4 =6AϪ4␨(4). (3) To evaluate the numerator in the high-temperature n=1 case, the relation in Eq. (2) is rewritten to render Here, ␨(n) is the Riemann zeta function. When evaluat- ing the numerator, two techniques have been found to be Q(A)=6AϪ4␨(4)[1 Ϫ F(A)] helpful: one for the “low-temperature” case (defined as ϱ ϱ 3 2 2 A > 4) and one for the “high-temperature” case (defined = ͚ ͐dvv exp(ϪnAv)[J0(v)+J1(v)] as A < 4). By such definitions, “high” and “low” temper- n=1 0 ature cases arise depending on the average diffraction ϱ = ͚ Q1(nA), (8) loss for a source at temperature T. This depends on T n=1 and the geometry through A. In either case, one first with makes use of the relation, ϱ ϱ ͐ 3 Ϫ 2 2 2 2 Q1(A)= dvv exp( Av )[J0(v)+J1(v)]. (9) J0[2v sin(␪/2)] = J0(v)+2͚ Jm (v)cos(m␪), (4) 0 m=1 Use of Eq. (5) and the relation, from which follows, ϱ 1 1 2␲ ͐dvexp(Ϫ␣v)J (␤v)= , (10) 2 2 ͐ ␪ ␪ ␪ 0 2 2 J0(v)+J1(v)= d (1 + cos )J0[2vsin( /2)]. (5) 0 ͙␣ + ␤ 2␲ 0 one has Low-Temperature Case 3 2␲ ␪ ͩϪ d ͪ 1 ͐ ␪ 1 + cos To evaluate the numerator in the low-temperature Q1(A)= d . (11) dA 2␲ 0 ͙ 2 2 ␪ case, series expansion of the latter Bessel function in Eq. A + 4sin ( /2) (5) yields Making the abbreviation, w = 4/(A 2 + 4), one has 1 ϱ (Ϫ1)s v 2s 2␲ 2 v 2 v ͚ ͐ ␪ Ϫ 2 ␪ 2s ␪ 3 ␲/2 2␾ J0( )+J1( )=␲ 2 d [1 sin ( /2)]sin ( /2) ͩϪ d ͪͫ 4 ͐ ␾ cos ͬ s=0 (s!) 0 Q1(A)= d dA ␲͙A 2 +4 0 ͙1 Ϫ wcos2␾ ϱ (Ϫ1)s v 2s (2s Ϫ 1)!! (2s + 1)!! ͚ ͫ Ϫ ͬ 3 K w Ϫ E w =2 2 ͩϪ d ͪͭ 4 ͫ ( ) ( )ͬͮ s=0 (s!) (2s)!! (2s + 2)!! = dA ␲͙A 2 +4 w ϱ (Ϫ1)s v 2s (2s Ϫ 1)!! ͚ ͫ ͬ 3 ͙A 2 =2 2 (6) ͩϪ d ͪͭ +4 Ϫ ͮ s=0 (s!) (2s + 2)!! = [K(w) E(w)] , (12) dA ␲ This permits one to rewrite Eq. (2) as follows: where E(w)andK(w) are respectively complete elliptic Ϫ integrals of the first and second kind, defined according Q(A) ≡ 6A 4␨(4)[1 Ϫ F(A)] to the convention,7 ϱ ϱ ␲/2 ͸͵ 3 Ϫ 2 2 ͐ ␾͙ Ϫ 2␾ = dvv exp( nAv)[J0(v)+J1(v)] E(w)= d 1 wsin , n=1 0 0 ␲ ϱ /2 d␾ (Ϫ1)s(2s + 3)!␨(2s +4) (2s Ϫ 1)!! K(w)= ͐ . (13) ͸ ͫ ͬ 2 =2 2 2s+4 0 ͙1 Ϫ wsin ␾ s=0 (s!) A (2s + 2)!! ϱ (In a different convention, w is replaced by w 2 in the (2␲)2s+4(2s + 3)!(2s Ϫ 1)!!B = Ϫ ͸ 2s+4 , (7) integrand but nowhere else.) From the properties of (s!)2(2s + 4)!(2s + 2)!!A 2s+4 s=0 elliptic integrals [8], one may obtain In this expression, B is a Bernoulli number. The sum i 4 ͫ4(2 + A 2)E(w) Ϫ A 2K(w)ͬ converges for all A > 2, and one obtains an error in F(A) Q1(A)= 3 2 3/2 , (14) Ϫ ␲A (A +4) estimated to be as small as 10 14 for A > 4 if one sums up to s = 28.6 and 777 Volume 106, Number 5, September–October 2001 Journal of Research of the National Institute of Standards and Technology Ϫ Ϫ d ͫ 4E(w) ͬ One should include only the first m 1 terms in the Q1(A)= .