Abstract

It is reported an approximation to the scalar and vectorial Rayleigh-Sommerfeld diffraction integrals performed with a Fourier transform operation, as done in Fresnel and Fraunhoffer approximations, but suitable for paraxial, non-paraxial and off-axis regimes. High accuracy is obtained even for waves with low f numbers. The approximation becomes exact on-axis.

© 2017 Optical Society of America

1. Introduction

Diffraction calculus involves the use of Fresnel-Kirchoff or Rayleigh-Sommerfeld integrals in its scalar or vectorial forms [1]. Such calculus is generally performed numerically, which involves a double integral of a complex function with a fast oscillating argument [2]. In order to facilitate their calculus or to get some insight about diffraction properties, such integrals are approximated in the so called Fresnel or Fraunhoffer regimes, which are adequate for paraxial regions [3]. For non-paraxial optical systems other approximations have been developed like that due to Debye, well suited for optical systems with high Fresnel numbers [3,4].

So, apparently, there is not a reliable approximation to calculate the diffraction properties of high, intermediate and low numerical aperture optical systems. In this work we will propose an approximation that could fulfill such requirements. We propose to use tilted spheres to approximate the shifted spheres of the diffraction integral; in Fraunhoffer and Fresnel integrals planes and parabolas are used to perform such approximation. The beauty of our approximation is that we perform the calculations with a simple Fourier transform operation, as done with the previous two approximations, but without having the paraxial restrictions. Some immediate novel results, described in this work, obtained as a consequence of our proposal are: a non-paraxial formula for rotationally symmetric functions via a Hankel transform; off-axis diffraction analysis; non-paraxial vectorial formulas performed with Fourier transforms; and even the study of evanescent waves. Since our approximation becomes exact on-axis it is well behaved near the optical axis and could be a useful tool to study optical systems with high focus depth. An important feature is that our approximation is calculated with a Fast Fourier Transform algorithm and 2D images or even 3D volumes are easily obtained.

2. Alternative approximation to a diffraction integral

Let us consider the First Rayleigh-Sommerfeld formulation (R-S) in rectangular coordinates, as described by [3]. The complex amplitude field distribution G(xo,yo) of a function g(x,y) at a certain distance z is given by

GRS(x0,y0)=z2πΣg(x,y)(jk1R)exp(jkR)R2dxdy,
where
R=z2+(xx0)2+(yy0)2;
g(x,y)=f(x,y)gI(x,y), being gI(x,y)the illuminating wave and f(x,y) an aperture function; (x0,y0) are the image coordinates; Σ is the diffraction aperture and k=2π/λ, the wavenumber, where λ means the wavelength of the illuminating light (see Fig. 1). Note that exp(jkR)/Rare spheres with its center shifted to (x0,y0).

 figure: Fig. 1

Fig. 1 Geometry for the Rayleigh-Sommerfeld diffraction formula.

Download Full Size | PPT Slide | PDF

As it is known, instead of a direct evaluation of Eq. (1), which involves various numerical difficulties, usually the R in the argument of the exponential (phase function) is replaced by the first terms of the Taylor series

Rz+(xx0)2+(yy0)22z,
providing |xx0|and |yy0|be smaller than z (paraxial case) or z be large enough compared with such differences (very low numerical aperture systems). The amplitude term 1/R in Eq. (1) is generally approximated by 1/z. Assuming also that z>>λ we obtain the well-known Fresnel approximation (paraxial) of Eq. (1) given by
GF(y0,y0)=AΣg(x,y)exp(jkr2/2z)exp[j2π(xx0+yy0)/λz]dxdy,
where

A=jexp(jkz)λzexp[jπ(x02+y02)/λz]andr=x2+y2.

It can be noticed that the arguments in the exponential functions of Eq. (4) can be interpreted as tilting parabolas. When it is assumed that z>>r, such parabolas behaves like planes and we obtain the Fraunhoffer approximation.

In the following descriptions we will use the above explanations for comparison purposes.

Now, Taylor’s formula for a differentiable function of two-variables is used to expand R(Eq. (2)) up to second order to obtain

RRn(xx0+yy0)Rn+(y2+z2)x02+(x2+z2)y022xyx0y0Rn3,
with
Rn=z2+x2+y2.
Equation (6) has been used in [5] to obtain analytical expressions for a focused Gaussian beam, and later in [6], with the first two terms, to study radially polarized beams. Unfortunately, in those forms the resulting expressions are difficult to deal with, particularly due to the existence of Rnin the denominator.

Since it seems to me that no further simplifications to Eq. (6) have been done before, I will show that retaining only the first two terms of Eq. (6) and replacing the respectiveRnin the denominator by a constantAzwill be enough to simplify the calculus of Eq. (1) without sacrificing its accuracy significantly within a neighborhood of (x0,y0). Thus, our approximation would be simply given by

RRn(xx0+yy0)Az,
whereRnis given in Eq. (7).

Substituting Eq. (8) in Eq. (1) and setting RRn (the first term of Eq. (8)) in the denominators we obtain our new approximation to the diffraction integral of Eq. (1) given by

GN(x0,y0)=z2πΣg(x,y)(jk1Rn)exp(jkRn)Rn2exp[j2π(xx0+yy0)λAz]dxdy,
Equation (9) means that the diffracted field is just the Fourier transform of the original functiong(x,y), which is multiplied by the complex function (jk1/Rn)(1/Rn2)exp(jkRn). Note that the phase Rn(x,y)(xx0+yy0)/Az means tilted spheres. It can be appreciated that no quadratic phase term appears in the resulting diffracted field as it happens in Eq. (4). It can be noticed also that our approximation becomes exact on-axis since GN(x0=0,y0=0)=GRS(0,0).

Expanding Rn of Eq. (8) in Taylor series and using only its first terms in Eq. (9) we reach again, up to a constant factor, the Fresnel and Fraunhoffer approximations, Eq. (4). This way, our approximation can be regarded as a generalization of both.

Since we can define u=x0/λAzand v=y0/λAz, the ambiguity in choosing Az means that the diffracted fields, for a given z, are correctly calculated up to a scale factor. That scale could be unimportant when relative calculations are done or when the shape of the PSF is the goal. In general, we can estimate Az either numerically or analytically and their value will depend on the criteria used to calculate it. In order to achieve the calculations it must be noticed that a good choice of Az means that the linear function x/Az looks like the function x/(z2+x2+y2)1/2. With a least squares formulation such similarity can be optimized with the following expression

Az=[Σx2dx]/[Σ(x2/z2+x2)dx],
which for a circular aperture Σ of radius a becomes to
Az=3a44(a2+z2)1/212z2(a2+z2)1/2+8z3.
For more complicated diffraction apertures, Eq. (10) can be computed numerically. When circular apertures are chosen, settingAz=zin Eq. (9) provides errors lower than 2% for fnum2 , being fnum=z/D. In following sections the performance of Eq. (11) will be shown.

3. Error analysis in the approximation

To perform our error analysis, for simplicity we will contemplatex00,y0=0and Σ a circular aperture of diameter D. Now, considering again the first two terms of Eq. (6), Rn(xx0+yy0)/Rn, and the Taylor series expansion of x/Rnwe obtain

RRnx0[xzx32z3+o(x/z)5],
where neglecting higher orders o(x/z)3and defining Az=zwe obtain our Eq. (8). Taking into account that D/2 is the maximum value that can reach x we roughly estimate the maximum error between xx0/Rn and xx0/z like o(x0/fnum3), being fnum=z/D. Thus, the higher the fnumor the lower x0 (or y0), the higher the exactitude of Eq. (8). An important issue is that errors depends only onfnum, x0 and y0and not on the aperture D itself. This is an important result because it means that converging waves having equal fnum and different Fresnel numbers Fn=a/(2λfnum) must have the same accuracy.

Now we will compute integrals specified by Eqs. (1), (4) and (9) to compare the intensitiesIRS(x0,y0), IF(x0,y0)and IN(x0,y0) given by the squared modulus of the respective amplitudes. To perform such comparison we have chosen the following parameters: a circular apertureΣof diameter D=5mm, λ=0.6328x103mm, z=0.6D, f(x,y)=1, and gI(x,y)defined by a converging spherical wave of radius equal to z,

gI(x,y)=exp(jkRn)Rn.
This way our converging wave poses an f number equal to 0.6, which corresponds to a numerical aperture NA of 0.64. In order to perform a fair comparison, we evaluate the three integrals numerically using the Lobato’s algorithm within MatLab routines quad2d and quadl. Additionally, we chose our analysis range for the spatial coordinates x0 to be around three times the corresponding Airy radius (AiR), which for large f numbers was simply, max(|x0|)3(1.22)λz/D. The results are shown in Fig. 2, where it can be seen that among those PSFs, Fresnel is the most different, as expected.

 figure: Fig. 2

Fig. 2 Diffraction patterns IRS, IFand INof a circular aperture of semi-diameter a = 5mm. Various curves of IN are given to highlight thatAz is only a scale factor.

Download Full Size | PPT Slide | PDF

In Fig. 2, dotted-continuous line corresponds to IRS, continuous line to IF and dashed to IN. We show various graphs of IN corresponding to values of Az = 0.85z, 1.0z, 1.08z and 1.2z, being 1.2z the most accurate (this value was computed with Eq. (11)). With this graphics it is clarified that Az is just a scale factor of the diffraction pattern. Defining our error of calculus asE=max|(IRS/maxIRS)(IN/maxIN)|, we found for the best curve of IN shown in Fig. 2 thatE=1.85%, which may be considered small, taking into account that a high numerical wave has been analyzed. In addition to this, we show in Fig. 3 a curve specifying that errors decreases when fnumincreases, as predicted with the error analysis of Eq. (12). Errors of Fig. 3 can be reduced when the analysis range for the spatial coordinates x0 is reduced; for example the previous error E(fnum=0.6)=1.85% is reduced to E=1% when max(|x0|)1AiRwas considered. This result is expected since our approximation is truly exact on axis.

 figure: Fig. 3

Fig. 3 Percentage difference between normalized intensities ofIRS, and IN, corresponding to a circular aperture of semi-diameter a = 5mm and converging spherical waves of various f numbers.

Download Full Size | PPT Slide | PDF

With the purpose to show that effectively the exactitude of Eq. (9) depends only on fnum and not on D, we fixed fnum = 0.6, 1.0 and 2.0 and varied the corresponding aperture diameters D from 0.05mm to 1000mm, calculating the respective diffraction intensities and estimating the respective errors as done in curve of Fig. 3. Results are shown in Fig. 4 where it can be seen that, effectively, errors do not depend on the aperture diameter itself, being constant for the samefnum. Logarithmic scale (natural logarithm) has been used in order to show the wide range of values of D considered. Notice that log(D) = 3 in Fig. 4 means Fresnel number Fn(105). Additionally, we have obtained curves for fnum = 100 and fnum = 10000 (not shown) finding similar behavior: constant errors independently of D. The curve obtained for fnum = 10000, with errorsE0.001%, comprises analysis when Fresnel number goes from 0.002 to 39.

 figure: Fig. 4

Fig. 4 Percentage difference between normalized intensities ofIRS, and IN, corresponding to a circular aperture of variable diameter D as function of three f numbers. Continuous, dashed and dotted-dashed lines corresponds to f/0.6, f/1 and f/2, respectively.

Download Full Size | PPT Slide | PDF

From these results we conclude that our proposal can be used confidently in the analysis of low, medium or high numerical aperture optical systems with errors on the order of 2%.

4. Vectorial diffraction integrals

One application of our approximation is in the study of diffraction properties of polarized beams. Since the first kind vectorial Rayleigh-Sommerfeld formulas expressed in rectangular coordinates are given by [7]

Ux,y(x0,y0)=z2πΣEx,y(x,y)g(x,y)(jkn1R)exp(jknR)R2dxdy,
Uz(x0,y0)=12πΣEz(x,y,x0,y0)g(x,y)(jkn1R)exp(jknR)R2dxdy,
where R=[z2+(xx0)2+(yy0)2]1/2; n is the refractive index of the medium space; Ex(x,y)and Ey(x,y)denote the electric-field components along the x and y directions, respectively and Ez(x,y,x0,y0)=Ex(x,y)(xx0)+Ey(x,y)(yy0). Clearly,
Uz(x0,y0)=Uxy(x0,y0)+12πΣExy(x,y)g(x,y)(jkn1R)exp(jknR)R2dxdy,
with
Uxy(x0,y0)=x0zUx(x0,y0)+y0zUy(x0,y0)andExy(x,y)=[xEx(x,y)+yEy(x,y)].
Similarly to what has been done to Eq. (1), Eqs. (14) and (16) can be approximated with the following substitution
(jkn1R)exp(jknR)R2(jkn1Rn)exp(jknRn)Rn2exp[j2π(xnx0+yny0)λAz].
To show that such approximations can be used successfully to evaluate the vectorial diffraction fields of polarized light, we have used(Ex,Ey)=(y/[x2+y2]1/2,x/[x2+y2]1/2), which corresponds to an azimuthal polarized beam [8]. Then, we used Eq. (18) to calculate Eqs. (14), and (16) together with the parameters given previously like n = 1, a circular apertureΣof diameter D=5mm, λ=0.6328x103mm, z=0.6D, Az=1.2zand a spherical convergent functiongI(x,y) given by Eq. (13) . The results of using our approximation with a fast Fourier algorithm are presented in Figs. 5(a-c) and using the integration routines of MatLab in Fig. 5d, where the comparison with exacts integral is shown. Good agreement has been obtained again, with E = 2.2%.

 figure: Fig. 5

Fig. 5 Diffraction intensities at the focus plane obtained with vectorial diffraction equations and azimuthal polarization. (a)-(c) corresponds to|Ux|2+|Uy|2,|Uz|2and its sum, computed from our approximation Eq. (18). (d) are the plots of the azimuthal polarization calculated with Eq. (18) and with the exact Eqs. (14,16). The two intensities are indistinguishable on the scale in the figure.

Download Full Size | PPT Slide | PDF

5. Comparison with Debye approximation

Using the results of [9,10] we can write the vectorial diffraction integrals within the Debye approximation like

U(x0,y0)=jcλ0α02πPsinθcosθexp[jknH(z,x0,y0,θ,φ)dθdφ,

with

 U(x0,y0)=(Ux,Uy,Uz)(x0,y0),P=(P1,P2,P3),H(z,x0,y0,θ,φ)=zcosθ+x0sinθcosφ+y0sinθsinφ,P1=[1+(cosθ1)cosφ2]Ex+(cosθ1)cosφsinφEy,P2=(cosθ1)cosφsinφEx+[1+(cosθ1)sinφ2]Ey,P3=sinθcosφExsinθsinφEy,
whereαis the angular semi-aperture of the converging wavefield. Ex,Ey, and (x0,y0,z)poses the usual meaning within this writing. Using the same data to obtain Fig. 5, as described in section 4, we evaluated Eq. (19) and show in Fig. 6, with dashed lines, the corresponding calculated intensities |Ux|2+|Uy|2+|Uz|2. The other two curves correspond to those shown in Fig. 5d. It can be appreciated that both approximation, at focus, behaves similarly even that we are using a converging wave with a NA = 0.64. The error obtained with the Debye approximation was 2.05% which is a little lower than the obtained with our method (2.2%).

 figure: Fig. 6

Fig. 6 Vectorial diffraction intensities|Ux|2+|Uy|2+|Uz|2 at the focus plane, corresponding to an azimuthal polarization state. Continuous, dotted-continuous and dashed lines were obtained with Eqs. (14,16), (18) and (19), respectively.

Download Full Size | PPT Slide | PDF

It is interesting to study the behavior of our proposal when defocus is introduced. Without going deeper into this matter we show Fig. 7, obtained with the previous data of Fig. 6 (D=5mm, λ=0.6328x103mm,Az=1.2z), not at the focus plane but a distance of f/60=0.05mm from it towards the aperture, that is, atz=0.59D.

 figure: Fig. 7

Fig. 7 Vectorial diffraction intensities as obtained in Fig. 6 but with a defocus of f/60 mm towards the aperture. Continuous, dotted-continuous and dashed lines were obtained with Eqs. (14,16), (18) and (19), respectively.

Download Full Size | PPT Slide | PDF

It can be appreciated that at this defocused position our results (dotted-continuous line) were more accurate than those obtained with the Debye approximation (dashed line). When choosing another value for Az, i.e, Az=1.29z, the curves of Fig. 8 were obtained; in that figure it can be noticed that our approximation (I_N) was improved. At this point it must be emphasized that different values for Azmeans only different scale factors of the whole diffraction pattern.

 figure: Fig. 8

Fig. 8 Vectorial diffraction intensities as obtained in Fig. 7 but using Az=1.29z to obtain I_N. The other curves are equal to those shown in Fig. 7.

Download Full Size | PPT Slide | PDF

Although we have derived a formula to calculate Az for a focused plane, Eq. (11), a general expression must be found yet when the diffraction patterns are imaged onto others plane positions.

6. Diffraction integrals for rotationally symmetric functions

When a circular apertureΣof semi diameter a is used and g(x,y)=g(r=x2+y2) is defined as a rotationally symmetric function, it can be obtained straightforwardly a one dimensional integral form of Eq. (9), given by

GN(r0)=z0ag(r)(jk1Rr)exp(jkRr)Rr2J0[2πrr0/(λAz)]rdr,
with

Rr=z2+r2, r0=x02+y02and J0 a Bessel function of the first kind, zero order.

7. Off-axis diffraction analysis

We have demonstrated our approximation for small values of (x0,y0), which means calculations near the optical axis. However our proposal is not restricted to those values since we can define (xn0,yn0)=(xn+x0,yn+y0)with (x0,y0) small, as in previous analysis, and (xn,yn)any constant point on the image plane, not necessarily near (0,0). Then, it can be shown that our approximation can be described by

GN(xn0,yn0)=BΣg(x,y)(jk1Rn0)exp(jkRn0)Rn02exp[j2π(xx0+yy0)λBz]dxdy,
withB=(z/2π)exp[jk(x0xn+y0yn)/Bz], Rn0=z2+(xxn)2+(yyn)2 and Bz a constant, not necessarily equal to Az.

8. Diffraction integrals using planar illumination

It seems that our approximation could be used to analyze diffraction apertures illuminated with planar waves. To show this, let us define in Eq. (22), gI(x,y)=1.0, a unit planar wave parallel to the optical axis, andf(x,y)=1.0, a constant aperture function. Thus, what we have is the Fourier transform of (jk1/Rn0)(1/Rn02)exp(jkRn0). Using the following result [11,12]

{(jk1Rn)z2πRn2exp(jkRn)}=exp[jkz1(λu)2(λv)2],
where denotes the Fourier transform, (u,v) are the spatial frequency coordinates and Rn as defined in Eq. (7), it is readily shown from Eq. (22) that
GN(xn0,yn0)=exp{jk(z/Bz)Bz2(x02+y02)},
where u=x0/λBzand v=y0/λBz have been used.

From Eq. (24) it can be seen that for (x02+y02)1/2Bz, evanescent waves are obtained since k,z andBz are all real valued and positive. So, Eq. (22) may be helpful to describe the behavior of evanescent waves for general values of g(x,y).

9. Discussions

It can be noticed from Eq. (9) (and those of Eqs. (14-18)) that doing the usual assumptionz>>λ the second term of the evanescent function (jk1/Rn)can be neglected. If in addition to this we suppose that the illuminating light corresponds to a converging spherical wave behaves like Eq. (13) we obtain from Eq. (9)

GN(x0,y0)=jkz2πΣf(x,y)Rn3exp[j2π(xx0+yy0)λAz]dxdy.
Thus, remarkably, at the focus plane, the non-paraxial field propagation of a function f(x,y)diffracted by an aperture Σ is simply the Fourier transform of the aperture function f(x,y)divided by an apodizing function(z2+x2+y2)3/2. Therefore, results obtained from paraxial calculus would be useful for non-paraxial regimes also, providing the multiplication by the apodizing function. For relatively big f numbers, the term1/Rn3 does not modify the intensity significantly, for example, neglecting such term in Eq. (25) with a = 5mm, fnum=1, f(x,y)=1 and a circular aperture, an intensity error as little as 1.5% is introduced.

Neglecting also the evanescent term in Eq. (21), setting f(r)=1 and using the proper converging spherical wave, it turns out to be

GN(r0)=jkz0aJ0[2πrr0/(λAz)]Rr3rdr,
which is our non-paraxial formula for the diffraction of a circular aperture in the focal plane. It can be seen that only for the case when Rr behaves like a constant, such diffraction pattern will be given by the well-known formula G(r0)(1/r0)J1[2πar0/(λAz)], otherwise, the diffraction pattern will be enlarged due to the convolution with the Hankel transform of the apodizing function 1/Rn3. GN(r0)given by Eq. (26) is important since their zeros positions are strongly related with the definition of the resolution of optical systems.

It is convenient to mention that, according with the explanation given in section 7, we can calculate the diffraction patterns in a neighborhood of each out-off axis point, and by a stitching process we could obtain such diffraction, with the desired exactitude, in bigger areas beyond those neighborhoods at the cost of performing various Fourier transforms.

10. Conclusions

We have proved that an alternative approximation can be used to analyze paraxial and non-paraxial systems without restriction on the Fresnel number. The approximation simplifies scalar and vectorial diffraction integrals to be calculated with a Fourier transform operation. An expression for off-axis analysis is obtained as well. Consequences of our approximation like one-dimensional integral for rotationally symmetric functions or evanescent waves have been considered in the analysis. Although we have focused our approximation into RS a similar analysis can be carried out with Fresnel-Kirchhoff integrals since they share similar expressions.

References and links

1. M. Born and E. Wolf, Principles of Optics, 7th ed. (Cambridge University Press, 1999).

2. J. J. Stamnes, Waves in Focal Regions (IOP Publishing Limited, 1986).

3. J. W. Goodman, Introduction to Fourier Optics, 2nd ed. (McGraw-Hill, 1996).

4. C. J. R. Sheppard, “Limitations of the paraxial Debye approximation,” Opt. Lett. 38(7), 1074–1076 (2013). [CrossRef]   [PubMed]  

5. A. Ciattoni, B. Crosignani, and P. Di Porto, “Vectorial analytical description of propagation of a highly nonparaxial beam,” Opt. Commun. 202(1-3), 17–20 (2002). [CrossRef]  

6. D. Deng, “Nonparaxial propagation of radially polarized light beams,” J. Opt. Soc. Am. B 23(6), 1228–1234 (2006). [CrossRef]  

7. H. Ye, C.-W. Qiu, K. Huang, J. Teng, B. Luk’yanchuk, and S. P. Yeo, “Creation of a longitudinally polarized subwavelength hotspot with an utra-thin planar lens: vectorial Rayleigh-Sommerfeld method,” Laser Phys. Lett. 10(6), 1–8 (2013). [CrossRef]  

8. X. Hao, C. Kuang, T. Wang, and X. Liu, “Effects of polarization on the de-excitation dark focal spot in STED microscopy,” J. Opt. 12(11), 115707 (2010). [CrossRef]  

9. B. Richards and E. Wolf, “Electromagnetic diffraction in optical systems II. Structure of the image field in an aplanatic system,” Proc. R. Soc. Lond. 253(1274), 358–379 (1959). [CrossRef]  

10. L. E. Helseth, “Focusing of atoms with strongly confined light potentials,” Opt. Commun. 212(4-6), 343–352 (2002). [CrossRef]  

11. E. Lalor, “Conditions for the validity of the angular spectrum of plane waves,” J. Opt. Soc. Am. 58(9), 1235–1237 (1968). [CrossRef]  

12. W. H. Southwell, “Validity of the Fresnel approximation in the near field,” J. Opt. Soc. Am. 71(1), 7–14 (1981). [CrossRef]  

References

  • View by:
  • |
  • |
  • |

  1. M. Born and E. Wolf, Principles of Optics, 7th ed. (Cambridge University Press, 1999).
  2. J. J. Stamnes, Waves in Focal Regions (IOP Publishing Limited, 1986).
  3. J. W. Goodman, Introduction to Fourier Optics, 2nd ed. (McGraw-Hill, 1996).
  4. C. J. R. Sheppard, “Limitations of the paraxial Debye approximation,” Opt. Lett. 38(7), 1074–1076 (2013).
    [Crossref] [PubMed]
  5. A. Ciattoni, B. Crosignani, and P. Di Porto, “Vectorial analytical description of propagation of a highly nonparaxial beam,” Opt. Commun. 202(1-3), 17–20 (2002).
    [Crossref]
  6. D. Deng, “Nonparaxial propagation of radially polarized light beams,” J. Opt. Soc. Am. B 23(6), 1228–1234 (2006).
    [Crossref]
  7. H. Ye, C.-W. Qiu, K. Huang, J. Teng, B. Luk’yanchuk, and S. P. Yeo, “Creation of a longitudinally polarized subwavelength hotspot with an utra-thin planar lens: vectorial Rayleigh-Sommerfeld method,” Laser Phys. Lett. 10(6), 1–8 (2013).
    [Crossref]
  8. X. Hao, C. Kuang, T. Wang, and X. Liu, “Effects of polarization on the de-excitation dark focal spot in STED microscopy,” J. Opt. 12(11), 115707 (2010).
    [Crossref]
  9. B. Richards and E. Wolf, “Electromagnetic diffraction in optical systems II. Structure of the image field in an aplanatic system,” Proc. R. Soc. Lond. 253(1274), 358–379 (1959).
    [Crossref]
  10. L. E. Helseth, “Focusing of atoms with strongly confined light potentials,” Opt. Commun. 212(4-6), 343–352 (2002).
    [Crossref]
  11. E. Lalor, “Conditions for the validity of the angular spectrum of plane waves,” J. Opt. Soc. Am. 58(9), 1235–1237 (1968).
    [Crossref]
  12. W. H. Southwell, “Validity of the Fresnel approximation in the near field,” J. Opt. Soc. Am. 71(1), 7–14 (1981).
    [Crossref]

2013 (2)

C. J. R. Sheppard, “Limitations of the paraxial Debye approximation,” Opt. Lett. 38(7), 1074–1076 (2013).
[Crossref] [PubMed]

H. Ye, C.-W. Qiu, K. Huang, J. Teng, B. Luk’yanchuk, and S. P. Yeo, “Creation of a longitudinally polarized subwavelength hotspot with an utra-thin planar lens: vectorial Rayleigh-Sommerfeld method,” Laser Phys. Lett. 10(6), 1–8 (2013).
[Crossref]

2010 (1)

X. Hao, C. Kuang, T. Wang, and X. Liu, “Effects of polarization on the de-excitation dark focal spot in STED microscopy,” J. Opt. 12(11), 115707 (2010).
[Crossref]

2006 (1)

2002 (2)

L. E. Helseth, “Focusing of atoms with strongly confined light potentials,” Opt. Commun. 212(4-6), 343–352 (2002).
[Crossref]

A. Ciattoni, B. Crosignani, and P. Di Porto, “Vectorial analytical description of propagation of a highly nonparaxial beam,” Opt. Commun. 202(1-3), 17–20 (2002).
[Crossref]

1981 (1)

1968 (1)

1959 (1)

B. Richards and E. Wolf, “Electromagnetic diffraction in optical systems II. Structure of the image field in an aplanatic system,” Proc. R. Soc. Lond. 253(1274), 358–379 (1959).
[Crossref]

Ciattoni, A.

A. Ciattoni, B. Crosignani, and P. Di Porto, “Vectorial analytical description of propagation of a highly nonparaxial beam,” Opt. Commun. 202(1-3), 17–20 (2002).
[Crossref]

Crosignani, B.

A. Ciattoni, B. Crosignani, and P. Di Porto, “Vectorial analytical description of propagation of a highly nonparaxial beam,” Opt. Commun. 202(1-3), 17–20 (2002).
[Crossref]

Deng, D.

Di Porto, P.

A. Ciattoni, B. Crosignani, and P. Di Porto, “Vectorial analytical description of propagation of a highly nonparaxial beam,” Opt. Commun. 202(1-3), 17–20 (2002).
[Crossref]

Hao, X.

X. Hao, C. Kuang, T. Wang, and X. Liu, “Effects of polarization on the de-excitation dark focal spot in STED microscopy,” J. Opt. 12(11), 115707 (2010).
[Crossref]

Helseth, L. E.

L. E. Helseth, “Focusing of atoms with strongly confined light potentials,” Opt. Commun. 212(4-6), 343–352 (2002).
[Crossref]

Huang, K.

H. Ye, C.-W. Qiu, K. Huang, J. Teng, B. Luk’yanchuk, and S. P. Yeo, “Creation of a longitudinally polarized subwavelength hotspot with an utra-thin planar lens: vectorial Rayleigh-Sommerfeld method,” Laser Phys. Lett. 10(6), 1–8 (2013).
[Crossref]

Kuang, C.

X. Hao, C. Kuang, T. Wang, and X. Liu, “Effects of polarization on the de-excitation dark focal spot in STED microscopy,” J. Opt. 12(11), 115707 (2010).
[Crossref]

Lalor, E.

Liu, X.

X. Hao, C. Kuang, T. Wang, and X. Liu, “Effects of polarization on the de-excitation dark focal spot in STED microscopy,” J. Opt. 12(11), 115707 (2010).
[Crossref]

Luk’yanchuk, B.

H. Ye, C.-W. Qiu, K. Huang, J. Teng, B. Luk’yanchuk, and S. P. Yeo, “Creation of a longitudinally polarized subwavelength hotspot with an utra-thin planar lens: vectorial Rayleigh-Sommerfeld method,” Laser Phys. Lett. 10(6), 1–8 (2013).
[Crossref]

Qiu, C.-W.

H. Ye, C.-W. Qiu, K. Huang, J. Teng, B. Luk’yanchuk, and S. P. Yeo, “Creation of a longitudinally polarized subwavelength hotspot with an utra-thin planar lens: vectorial Rayleigh-Sommerfeld method,” Laser Phys. Lett. 10(6), 1–8 (2013).
[Crossref]

Richards, B.

B. Richards and E. Wolf, “Electromagnetic diffraction in optical systems II. Structure of the image field in an aplanatic system,” Proc. R. Soc. Lond. 253(1274), 358–379 (1959).
[Crossref]

Sheppard, C. J. R.

Southwell, W. H.

Teng, J.

H. Ye, C.-W. Qiu, K. Huang, J. Teng, B. Luk’yanchuk, and S. P. Yeo, “Creation of a longitudinally polarized subwavelength hotspot with an utra-thin planar lens: vectorial Rayleigh-Sommerfeld method,” Laser Phys. Lett. 10(6), 1–8 (2013).
[Crossref]

Wang, T.

X. Hao, C. Kuang, T. Wang, and X. Liu, “Effects of polarization on the de-excitation dark focal spot in STED microscopy,” J. Opt. 12(11), 115707 (2010).
[Crossref]

Wolf, E.

B. Richards and E. Wolf, “Electromagnetic diffraction in optical systems II. Structure of the image field in an aplanatic system,” Proc. R. Soc. Lond. 253(1274), 358–379 (1959).
[Crossref]

Ye, H.

H. Ye, C.-W. Qiu, K. Huang, J. Teng, B. Luk’yanchuk, and S. P. Yeo, “Creation of a longitudinally polarized subwavelength hotspot with an utra-thin planar lens: vectorial Rayleigh-Sommerfeld method,” Laser Phys. Lett. 10(6), 1–8 (2013).
[Crossref]

Yeo, S. P.

H. Ye, C.-W. Qiu, K. Huang, J. Teng, B. Luk’yanchuk, and S. P. Yeo, “Creation of a longitudinally polarized subwavelength hotspot with an utra-thin planar lens: vectorial Rayleigh-Sommerfeld method,” Laser Phys. Lett. 10(6), 1–8 (2013).
[Crossref]

J. Opt. (1)

X. Hao, C. Kuang, T. Wang, and X. Liu, “Effects of polarization on the de-excitation dark focal spot in STED microscopy,” J. Opt. 12(11), 115707 (2010).
[Crossref]

J. Opt. Soc. Am. (2)

J. Opt. Soc. Am. B (1)

Laser Phys. Lett. (1)

H. Ye, C.-W. Qiu, K. Huang, J. Teng, B. Luk’yanchuk, and S. P. Yeo, “Creation of a longitudinally polarized subwavelength hotspot with an utra-thin planar lens: vectorial Rayleigh-Sommerfeld method,” Laser Phys. Lett. 10(6), 1–8 (2013).
[Crossref]

Opt. Commun. (2)

A. Ciattoni, B. Crosignani, and P. Di Porto, “Vectorial analytical description of propagation of a highly nonparaxial beam,” Opt. Commun. 202(1-3), 17–20 (2002).
[Crossref]

L. E. Helseth, “Focusing of atoms with strongly confined light potentials,” Opt. Commun. 212(4-6), 343–352 (2002).
[Crossref]

Opt. Lett. (1)

Proc. R. Soc. Lond. (1)

B. Richards and E. Wolf, “Electromagnetic diffraction in optical systems II. Structure of the image field in an aplanatic system,” Proc. R. Soc. Lond. 253(1274), 358–379 (1959).
[Crossref]

Other (3)

M. Born and E. Wolf, Principles of Optics, 7th ed. (Cambridge University Press, 1999).

J. J. Stamnes, Waves in Focal Regions (IOP Publishing Limited, 1986).

J. W. Goodman, Introduction to Fourier Optics, 2nd ed. (McGraw-Hill, 1996).

Cited By

OSA participates in Crossref's Cited-By Linking service. Citing articles from OSA journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (8)

Fig. 1
Fig. 1 Geometry for the Rayleigh-Sommerfeld diffraction formula.
Fig. 2
Fig. 2 Diffraction patterns I RS , I F and I N of a circular aperture of semi-diameter a = 5mm. Various curves of I N are given to highlight that A z is only a scale factor.
Fig. 3
Fig. 3 Percentage difference between normalized intensities of I RS , and I N , corresponding to a circular aperture of semi-diameter a = 5mm and converging spherical waves of various f numbers.
Fig. 4
Fig. 4 Percentage difference between normalized intensities of I RS , and I N , corresponding to a circular aperture of variable diameter D as function of three f numbers. Continuous, dashed and dotted-dashed lines corresponds to f/0.6, f/1 and f/2, respectively.
Fig. 5
Fig. 5 Diffraction intensities at the focus plane obtained with vectorial diffraction equations and azimuthal polarization. (a)-(c) corresponds to | U x | 2 +| U y | 2 ,| U z | 2 and its sum, computed from our approximation Eq. (18). (d) are the plots of the azimuthal polarization calculated with Eq. (18) and with the exact Eqs. (14,16). The two intensities are indistinguishable on the scale in the figure.
Fig. 6
Fig. 6 Vectorial diffraction intensities | U x | 2 +| U y | 2 +| U z | 2 at the focus plane, corresponding to an azimuthal polarization state. Continuous, dotted-continuous and dashed lines were obtained with Eqs. (14,16), (18) and (19), respectively.
Fig. 7
Fig. 7 Vectorial diffraction intensities as obtained in Fig. 6 but with a defocus of f/60 mm towards the aperture. Continuous, dotted-continuous and dashed lines were obtained with Eqs. (14,16), (18) and (19), respectively.
Fig. 8
Fig. 8 Vectorial diffraction intensities as obtained in Fig. 7 but using A z =1.29z to obtain I_N. The other curves are equal to those shown in Fig. 7.

Equations (26)

Equations on this page are rendered with MathJax. Learn more.

G RS ( x 0 , y 0 )= z 2π Σ g(x,y)(jk 1 R ) exp(jkR) R 2 dxdy ,
R= z 2 + (x x 0 ) 2 + (y y 0 ) 2 ;
Rz+ (x x 0 ) 2 + (y y 0 ) 2 2z ,
G F ( y 0 , y 0 )=A Σ g(x,y)exp(jk r 2 /2z)exp[j2π(x x 0 +y y 0 )/λz]dxdy ,
A= jexp(jkz) λz exp[jπ( x 0 2 + y 0 2 )/λz]andr= x 2 + y 2 .
R R n (x x 0 +y y 0 ) R n + ( y 2 + z 2 ) x 0 2 +( x 2 + z 2 ) y 0 2 2xy x 0 y 0 R n 3 ,
R n = z 2 + x 2 + y 2 .
R R n (x x 0 +y y 0 ) A z ,
G N ( x 0 , y 0 )= z 2π Σ g(x,y)(jk 1 R n ) exp(jk R n ) R n 2 exp[ j2π(x x 0 +y y 0 ) λ A z ]dxdy ,
A z =[ Σ x 2 dx ]/[ Σ ( x 2 / z 2 + x 2 )dx ],
A z = 3 a 4 4 ( a 2 + z 2 ) 1/2 12 z 2 ( a 2 + z 2 ) 1/2 +8 z 3 .
R R n x 0 [ x z x 3 2 z 3 +o (x/z) 5 ],
g I (x,y)= exp(jk R n ) R n .
U x,y ( x 0 , y 0 )= z 2π Σ E x,y (x,y)g(x,y)(jkn 1 R ) exp(jknR) R 2 dxdy ,
U z ( x 0 , y 0 )= 1 2π Σ E z (x,y, x 0 , y 0 )g(x,y)(jkn 1 R ) exp(jknR) R 2 dxdy ,
U z ( x 0 , y 0 )= U xy ( x 0 , y 0 )+ 1 2π Σ E xy (x,y)g(x,y)(jkn 1 R ) exp(jknR) R 2 dxdy ,
U xy ( x 0 , y 0 )= x 0 z U x ( x 0 , y 0 )+ y 0 z U y ( x 0 , y 0 )and E xy (x,y)=[x E x (x,y)+y E y (x,y)].
(jkn 1 R ) exp(jknR) R 2 (jkn 1 R n ) exp(jkn R n ) R n 2 exp[ j2π(xn x 0 +yn y 0 ) λ A z ].
U ( x 0 , y 0 )= jc λ 0 α 0 2π P sinθ cosθ exp[jknH(z, x 0 , y 0 ,θ,φ)dθdφ ,
  U ( x 0 , y 0 )=( U x , U y , U z )( x 0 , y 0 ), P =( P 1 , P 2 , P 3 ), H(z, x 0 , y 0 ,θ,φ)=zcosθ+ x 0 sinθcosφ+ y 0 sinθsinφ, P 1 =[1+(cosθ1)cos φ 2 ] E x +(cosθ1)cosφsinφ E y , P 2 =(cosθ1)cosφsinφ E x +[1+(cosθ1)sin φ 2 ] E y , P 3 =sinθcosφ E x sinθsinφ E y ,
G N ( r 0 )=z 0 a g(r)(jk 1 R r ) exp(jk R r ) R r 2 J 0 [2πr r 0 /(λ A z )]rdr ,
G N ( x n0 , y n0 )=B Σ g(x,y)(jk 1 R n0 ) exp(jk R n0 ) R n0 2 exp[ j2π(x x 0 +y y 0 ) λ B z ]dxdy ,
{ ( jk 1 R n ) z 2π R n 2 exp(jk R n ) }=exp[ jkz 1 (λu) 2 (λv) 2 ],
G N ( x n0 , y n0 )=exp{ jk(z/ B z ) B z 2 ( x 0 2 + y 0 2 ) },
G N ( x 0 , y 0 )= jkz 2π Σ f(x,y) R n 3 exp[ j2π(x x 0 +y y 0 ) λ A z ]dxdy .
G N ( r 0 )=jkz 0 a J 0 [2πr r 0 /(λ A z )] R r 3 rdr ,

Metrics