Fundaments of optical far-field subwavelength

resolution based on illumination with surface

waves
Roberto Lopez-Boada,1 Charles J. Regan,2,3 Daniel Dominguez,3,4 Ayrton. A. Bernussi,2,3
and Luis Grave de Peralta3,4,*
1
ACE, Barry University, Miami Shores, FL 33161, USA
2
Department of Electrical and Computer Engineering, Texas Tech University, Lubbock, TX 79409, USA
3
Nano Tech Center, Texas Tech University, Lubbock, TX 79409, USA
4
Department of Physics, Texas Tech University, Lubbock, TX 79409, USA
*
[email protected]

Abstract: We present a general discussion about the fundamental physical

principles involved in a novel class of optical superlenses that permit to
realize in the far-field direct non-scanning images with subwavelength
resolution. Described superlenses are based in the illumination of the object
under observation with surface waves excited by fluorescence, the
enhanced transmission of fluorescence via coupling with surface waves,
and the occurrence of far-field coherence-related fluorescence diffraction
phenomena. A Fourier optics description of the image formation based on
illumination with surface waves is presented, and several recent
experimental realizations of this technique are discussed. Our theoretical
approach explains why images with subwavelength resolution can be
formed directly in the microscope camera, without involving scanning or
numerical post-processing. While resolution of the order of λ/7 has been
demonstrated using the described approach, we anticipate that deeper
optical subwavelength resolution should be expected.
©2013 Optical Society of America
OCIS codes: (110.0180) Microscopy; (240.6680) Surface plasmons; (240.6690) Surface waves;
(260.2510) Fluorescence; (260.6970) Total internal reflection.

References and links

1. S. Durant, Z. Liu, J. M. Steele, and X. Zhang, “Theory of the transmission properties of an optical far-field
superlens for imaging beyond the diffraction limit,” J. Opt. Soc. Am. B 23(11), 2383–2392 (2006).
2. E. Hetcht, Optics, 3rd edition (Addison Wesley, 1998).
3. M. Born and E. Wolf, Priciples of Optics, 5th edition (Pergamon Press, 1975).
4. J. B. Pendry, “Negative refraction makes a perfect lens,” Phys. Rev. Lett. 85(18), 3966–3969 (2000).
5. D. O. S. Melville, R. J. Blaikie, and C. R. Wolf, “Submicron imaging with a planar silver lens,” Appl. Phys. Lett.
84(22), 4403–4405 (2004).
6. W. Srituravanich, N. Fang, C. Sun, Q. Luo, and X. Zhang, “Plasmonic Nanolithography,” Nano Lett. 4(6), 1085–
1088 (2004).
7. O. E. Gawhary, N. J. Schilder, A. C. Assafrao, S. F. Pereira, and H. P. Urbach, “Restoration of s-polarized
evanescen waves and subwavelength imaging by a single dielectric slab,” New J. Phys. 14(5), 053025 (2012).
8. Y. Zhang, C. Arnold, P. Offermans, and J. Rivas, “Surface wave sensors based on nanometric layers of strongly
absorbing materials,” Opt. Express 20(9), 9431–9441 (2012).
9. Z. Liu, S. Durant, H. Lee, Y. Pikus, N. Fang, Y. Xiong, C. Sun, and X. Zhang, “Far-field optical superlens,”
Nano Lett. 7(2), 403–408 (2007).
10. Z. Liu, H. Lee, Y. Xiong, C. Sun, and X. Zhang, “Far-field optical hyperlens magnifying sub-diffraction-limited
objects,” Science 315(5819), 1686 (2007).
11. X. Zhang and Z. Liu, “Superlenses to overcome the diffraction limit,” Nat. Mater. 7(6), 435–441 (2008).
12. R. Rodriguez, C. J. Regan, A. Ruiz-Columbié, W. Agutu, A. A. Bernussi, and L. Grave de Peralta, “Study of
plasmonic crystals using Fourier-plane images obtained with plasmon tomography far-field superlenses,” J.
Appl. Phys. 110(8), 083109 (2011).

#186361 - $15.00 USD Received 5 Mar 2013; revised 28 Apr 2013; accepted 3 May 2013; published 8 May 2013
(C) 2013 OSA 20 May 2013 | Vol. 21, No. 10 | DOI:10.1364/OE.21.011928 | OPTICS EXPRESS 11928
13. C. J. Regan, R. Rodriguez, S. C. Gourshetty, L. Grave de Peralta, and A. A. Bernussi, “Imaging nanoscale
features with plasmon-coupled leakage radiation far-field superlenses,” Opt. Express 20(19), 20827–20834
(2012).
14. N. Garcia and M. Nieto-Vesperinas, “Left-handed materials do not make a perfect lens,” Phys. Rev. Lett. 88(20),
207403 (2002).
15. C. J. Regan, D. Dominguez, A. A. Bernussi, and L. Grave de Peralta, “Far-field optical superlens without metal,”
J. Appl. Phys. ((to be published).
16. M. H. Raether, Surface Plasmons on Smooth and Rough Surfaces and on Gratings (Springer-Verlag, 1988).
17. G. Christou and C. Mias, “Critique of optical negative refraction superlensing,” Plasmonics 6(2), 307–309
(2011).
18. S. P. Frisbie, C. Chesnutt, M. E. Holtz, A. Krishnan, L. de Peralta, and A. A. Bernussi, “Image formation in
wide-field microscopes based on leakage of surface plasmon-coupled fluorescence,” IEEE Photon. J. 1(2), 153–
162 (2009).
19. L. Grave de Peralta, C. J. Regan, and A. A. Bernussi, “SPP tomography: a simple wide-field nanoscope,”
Scanning n/a (2012), doi:10.1002/sca.21057.
20. I. Gryczinski, J. Malicka, K. Nowaczyk, Z. Gryczynski, and J. Lacowicz, “Effects of sample thickness on the
optical properties of surface plasmon-coupled emission,” J. Phys. Chem. B 108(32), 12073–12083 (2004).
21. I. Gryczynski, J. Malicka, Z. Gryczynski, and J. R. Lakowicz, “Surface plasmon-coupled emission with gold
films,” J. Phys. Chem. B 108(33), 12568–12574 (2004).
22. F. Goos and H. Hänchen, “Ein neuer und fundamentaler versuch zur totalreflexion,” Ann. Phys. 6(7-8), 333–346
(1947).
23. S. A. Taya, E. J. El-Farram, and T. M. El-Agez, “Goos-Hänchen shift as a probe in evanescent slab waveguides
sensors,” Int. J. Electron. Commun. 66(3), 204–210 (2012).
24. J. W. Goodman, Introduction to Fourier Optics (McGraw-Hill, 1968).
25. A. Drezet, A. Hohenau, D. Koller, A. Stepanov, H. Ditlbacher, B. Steinberger, F. R. Aussenegg, A. Leitner, and
J. R. Krenn, “Leakage radiation microscopy of surface plasmon polaritons,” Mater. Sci. Eng. B 149(3), 220–229
(2008).
26. L. Grave de Peralta, R. Lopez-Boada, A. Ruiz-Columbie, S. Park, and A. A. Bernussi, “Some consequences of
experiments with a plasmonic quantum eraser for plasmon tomography,” J. Appl. Phys. 109(2), 023101 (2011).
27. A. Houk, R. Lopez-Boada, A. Ruiz-Columbie, S. Park, A. A. Bernussi, and L. Grave de Peralta, “Erratum: some
consequences of experiments with a plasmonic quantum eraser for plasmon tomography,” J. Appl. Phys.
109(11), 119901 (2011).
28. J. J. Burke, G. I. Stegeman, and T. Tamir, “Surface-polariton-like waves guided by thin, lossy metal films,”
Phys. Rev. B Condens. Matter 33(8), 5186–5201 (1986).
29. Z. Liu, N. Fang, T. Yen, and X. Zhang, “Rapid growth of evanescent wave by silver superlens,” Appl. Phys.
Lett. 83(25), 5184–5186 (2003).
30. F. L. Pedrotti, L. S. Pedrotti, and L. M. Pedrotti, Introduction to Optics, 3rd Edition (Person Prentice Hall, 2007).
31. B. Schumacher, Quantum mechanics: the Physics of the Microscopic World (The Teaching Company, 2009).
32. E. J. Galvez, C. H. Holbrow, M. J. Pysher, J. W. Martin, N. Courtemanche, L. Heilig, and J. Spencer,
“Interference with correlated photons: five quantum mechanics experiments for undergraduates,” Am. J. Phys.
73(2), 127–140 (2005).
33. L. Grave de Peralta, “Phenomenological quantum description of the ultra fast response of arrayed waveguide
gratings,” J. Appl. Phys. 108(10), 103110 (2010).
34. Y. Chen, D. Zhang, L. Han, G. Rui, X. Wang, P. Wang, and H. Ming, “Surface-plasmon-coupled emission
microscopy with a polarization converter,” Opt. Lett. 38(5), 736–738 (2013).
35. W. T. Tang, E. Chung, Y. H. Kim, P. T. C. So, and C. J. Sheppard, “Surface-plasmon-coupled emission
microscopy with a spiral phase plate,” Opt. Lett. 35(4), 517–519 (2010).
36. C. J. Regan, O. Thiabgoh, L. Grave de Peralta, and A. A. Bernussi, “Probing photonic Bloch wavefunctions with
plasmon-coupled leakage radiation,” Opt. Express 20(8), 8658–8666 (2012).
37. C. J. Regan, L. Grave de Peralta, and A. A. Bernussi, “Equifrequency curve dispersion in dielectric-loaded
plasmonic crystals,” J. Appl. Phys. 111(7), 073105 (2012).
38. C. J. Regan, A. Krishnan, R. Lopez-Boada, L. Grave de Peralta, and A. A. Bernussi, “Direct observation of
photonic Fermi surfaces by plasmon tomography,” Appl. Phys. Lett. 98(15), 151113 (2011).
39. J. D. Joannopoulus, S. G. Johnson, J. N. Winn, and R. D. Meade, Photonic Crystals: Molding the Flow of Light
(Princeton University, 2008).

1. Introduction
The conventional wisdom, formed from the Abbe’s theory of image formation and Fourier
optics, is that traditional optical microscopy is diffraction-limited to spatial periods (p) larger
than ~λo/NA [1,2], or separation between two points (Δx) larger than λo/(2NA) [1–3], where
NA is the numerical aperture of the microscope objective lens, and λo is the free space
wavelength of the illuminating light. This explain why the theoretical prediction that a simple
thin layer of metal could be used to realize a practical superlens [4] was warmly received by

#186361 - $15.00 USD Received 5 Mar 2013; revised 28 Apr 2013; accepted 3 May 2013; published 8 May 2013
(C) 2013 OSA 20 May 2013 | Vol. 21, No. 10 | DOI:10.1364/OE.21.011928 | OPTICS EXPRESS 11929
the optical community searching for novel imaging techniques with subwavelength resolution
for nanotechnology and subcellular biological applications. Since then, optical subwavelength
resolution in the near-field using metallic superlenses has been demonstrated [5,6], exciting
new near-field superlenses without metal have been also proposed [7,8], and the original
concept was extended to the far-field [9–13]. Some demonstrated far-field superlenses require
intensive numerical image post processing [1,9], some require a structure much more
complicated than the original Pendry’s proposal [10], but until recently, all demonstrated far-
field superlenses require a structure with at least a thin metal layer. The unavoidable inherent
losses associated with the presence of the metal layer constitute a drawback characteristic of
these superlenses [14]; therefore, alternative methods for mitigation of the metal-related
losses are of great interest.
Recently, the first far-field optical superlens without metal has been demonstrated. The
details of the experiments will be published elsewhere [15]. While investigating the physical
principles responsible for the subwavelength resolution capabilities of the far-field
superlenses that resemble more closely the original Pendry’s proposal [12,13], we realized
that the specific nature of the surface plasmon polaritons (SPP) [16] excited in the superlens
was not a requirement for obtaining subwavelength resolution. This idea had been discussed
previously in the context of near-field superlenses [7,17] but its application to realize far-field
superlenses remains unexplored. In this work we show that the same physical principles are
responsible for the optical subwavelength resolution capabilities of demonstrated far-field
superlens without metal [15] and plasmonic superlenses [12,13]. In what follows, we group
both types of far-field superlenses in the concept of “surface wave illumination” superlenses
(SWIS). We show that the optical far-field subwavelength resolution obtained using SWIS is
based on three common principles: the illumination of the object under observation with
surface waves excited by fluorescence, the enhanced transmission of fluorescence via
coupling with surface waves, and the occurrence of far-field coherence-related fluorescence
diffraction phenomena. Therefore, SWIS are not ideal superlenses in the original sense of
being capable of infinite resolution [4,14,17], but they are far-field superlenses because SWIS
allow direct imaging in the far-field with subwavelength resolution, without involving
scanning or numerical post-processing. We present here, for the first time, a comprehensive
Fourier Optics description of the image formation in a SWIS-microscope arrangement. The
theory of image formation using SWIS proposed in this wok represents a generalization to
far-field superlenses of the well-known Abbe’s theory of image formation. This makes the
design of far-field superlenses easier and more intuitive by expanding the range of
applicability of the conventional wisdom about the image formation in optical instruments.
For instance, we show using these ideas how it might be possible to obtain deep optical
subwavelength resolution with SWIS.
This paper is organized as follows: In Section 2 we make a general description of a SWIS
and discuss the phenomena of enhanced transmission of fluorescence via excitement of
surface waves. Section 3 is dedicated to the Fourier optics description of the formation of the
Fourier plane (FP) images obtained in a SWIS-microscope arrangement. Several simple
examples are presented, and the analytical expressions are derived and compared with
experimental FP images. In Section 4, a Fourier optics description of the formation of the
surface emission (SE) images is presented, and the origin of the SWIS subwavelength
resolution is elucidated. Finally, the conclusions of this work are given in Section 5.
2. Enhanced transmission of fluorescence coupled to surface waves
Figure 1(a) shows a schematic illustration of the rather simple transversal structure of a
typical SWIS. The SWIS substrate is formed by a ~150 μm thick glass cover slip. At the top
of the SWIS is a ~150 nm thick layer of PolyMethylMethAcrylate (PMMA) doped with
Rhodamine-6G (R6G) with peak emission at λo~568 nm wavelength. As shown in Fig. 1(a),
an optional ~50 nm thick gold layer may be present between the substrate and the top layer.

#186361 - $15.00 USD Received 5 Mar 2013; revised 28 Apr 2013; accepted 3 May 2013; published 8 May 2013
(C) 2013 OSA 20 May 2013 | Vol. 21, No. 10 | DOI:10.1364/OE.21.011928 | OPTICS EXPRESS 11930
The object under observation (not shown in Fig. 1(a)) should be embedded in the PMMA +
R6G layer. The SWIS with the object under observation is then placed in a traditional inverse
microscope with an immersion oil high NA objective lens. A detailed description of a typical
SWIS-microscope arrangement have being previously published [12,13,15,18,19].

Fig. 1. (a) Cross-section schematic illustration of a sample used in a typical SWIS. (b)-(e)
illustrates different excitation schemes: conventional surface wave excitation [(b) and (d)] and
the excitation of the surface waves in SWIS [(c) and (e)] for superlenses with [(b) and (c)] and
without [(d) and (e)] a thin metal layer.

Typically, the R6G fluorophores in the PMMA layer are excited with light of a continuous
wave 532nm wavelength emitted by a diode pumped solid state laser. In SWIS, surface waves
are excited in the medium/PMMA interface by spontaneous emission events occurring
randomly in every direction near the interface, where the medium is gold in plasmonic
superlenses [12,13] or glass in superlenses without metal [15]. The surface waves are SPP in
plasmonic superlenses, or surface waves related with total internal reflection (TIR)
phenomena in superlenses without metal [15]. As shown in Figs. 1(b)-1(e), in SWIS the
excitation of surface waves by fluorescence results in a type of reverse phenomena when
compared to more standard methods to excite the surface waves. Figure 1(b) shows a
schematic illustration of SPP excitation using the well-known Kretschmann configuration
[16]. As it is illustrated in Fig. 1(c), plasmonic SWIS are based on the optical reversibility
principle [2], i.e., the excitation of SPP by fluorescence results in the occurrence of plasmon-
coupled leakage radiation [18,20,21]. Figure 1(d) shows a schematic illustration of TIR in a
conventional prism. The incident light produces a surface wave at the glass surface that
travels a few microns before leaking to the glass producing the reflected light [15, 22,23]. As
it is illustrated in Fig. 1(e), in SWIS without metal the output light is produced by the
fluorescence-excited surface waves without the presence of an incident beam. In SWIS, the
surface waves excited by fluorescence illuminate the object under observation; i.e. the surface
waves are the illumination source. This is of particular importance. A well known result of
the Abbe’s theory of image formation and Fourier optics is that the minimum resolvable
period using a traditional microscope is given by the following expression [1,2, 24]:
λ
pmin =
o
. (1)
NA
Equation (1) corresponds to a minimum resolvable separation between two points of
Δx~0.61pmin [1,19]. However, it is less known that Eq. (1) is derived assuming that the sample
is illuminated by a coherent monochromatic plane wave that propagates perpendicularly to
the surface of the sample under observation. The electric field associated with the surface
waves has a more prominent component, E† , which is perpendicular to the propagation
direction and can be perpendicular or parallel to the medium/PMMA interface; therefore, in

#186361 - $15.00 USD Received 5 Mar 2013; revised 28 Apr 2013; accepted 3 May 2013; published 8 May 2013
(C) 2013 OSA 20 May 2013 | Vol. 21, No. 10 | DOI:10.1364/OE.21.011928 | OPTICS EXPRESS 11931
SWIS the illumination source can be considered, in a first approximation, a superposition of
plane waves propagating in all directions along the medium/PMMA interface. As a
consequence, one should expect that Eq. (1) could not apply to SWIS. The component E† of
the electric field associated with a plane wave propagating along the medium/PMMA

interface in the direction defined by the wavevector, k w is described by the following
expression:
  
( ) (
E† r, t = E†o sin k w ⋅ r − ω t , ) (2)

where,
 λo
k w = k w = 2π
λ = ko neff , λ = neff , ω = 2πν . (3)

E†o , λ, and ν are the amplitude, wavelength, and frequency of the plane wave, respectively;

neff is the effective refractive index experienced by the surface wave, and r = ( x, y ) is
contained in the plane z = 0 (medium /PMMA interface). The surface waves are continuously
leaking to the glass substrate of the sample while propagating through the medium/PMMA
interface [25–27]. This is the light used for imaging in a SWIS-microscope arrangement
[12,13,15,18,19].


Fig. 2. Light leaks in the direction defined by kl , which forms an angle θl with respect to the
axis z. Leaked light is contained in the plane ρ, which also contains the z axis and the vectors
 
k w and kl .

As shown in Fig. 2, for a given value of k w , leakage radiation is very directional. It
occurs in a very narrow angle interval centered at a direction forming an angle θl respect to
the normal to the medium/PMMA interface [20,21, 25]. Therefore, the light leaking in the
direction θl can be described to a good approximation as a plane wave [25], which electric
field (El) is given by the following expression:
  
( ) (
El s, t = Elo sin kl ⋅ s − ω t , ) (4)

#186361 - $15.00 USD Received 5 Mar 2013; revised 28 Apr 2013; accepted 3 May 2013; published 8 May 2013
(C) 2013 OSA 20 May 2013 | Vol. 21, No. 10 | DOI:10.1364/OE.21.011928 | OPTICS EXPRESS 11932
where
 λo
kl = k ⊥2 + k2 = 2π
λs , λs = ns . (5)

Here ns is the refractive index of the substrate, s = ( x, y , z ) , and Elo is proportional to E†o

[25–28]. k ⊥ and k are the components of the wavevector of the leakage radiation ( kl ) in the

direction perpendicular and parallel to the medium/PMMA interface, respectively. k is also

parallel to k w [26], and its magnitude is given by the following expression [16,18,25]:

k = kl sin θ l = k w . (6)
SPP surface waves have a small component of the electric field in the direction of
propagation [25], which makes Eq. (2) only a first approximation of a SPP plane wave.
However, Eq. (4) is still valid for a SPP wave because Elo is proportional to the amplitudes
of both electric field components [25]. A signature of the occurrence of enhanced
transmission of fluorescence coupled to surface waves [15,20,21,29] is that the
Elo / E†o >> Et / E†o , where Et is the magnitude of the electric field corresponding to the
fluorescence radiation that pass through the medium/PMMA interface without exciting
surface waves. The surface and the leakage radiation plane waves described by the Eqs. (2)
and (4), respectively, are coherent waves. This description is then appropriate only if the
coherence length of the fluorescence (lc) is larger than the propagation length of the surface
waves (L). The value of lc can be estimated using the following expression [30]:
λo2
lc = , (7)
Δλ
where Δλ is the spectral width of the passband filter placed after the immersion oil objective
lens for filtering the laser excitation light [12,13,15,18,19]. For Δλ = 10 nm, we calculate a
value of lc ~30 μm, which is larger than the values of L~20μm [16,25,26] and few μm [22,23]
for SPP and surface waves related to TIR, respectively. Therefore, it is well justified to

consider that in SWIS the sample is illuminated in the direction determined by k w by a
coherent surface wave excited by fluorescence.
3. Formation of the Fourier plane images
3.1 Homogeneous sample
For a SWIS without object under observation, the optical disturbance (U) at the
medium/PMMA interface (z = 0) associated to El is given by the following expression [24]:
 
Uϕ ( x, y , z = 0) = Cϕ ei kw  r = Cϕ e (
i xk w cosϕ + yk w sin ϕ )
, (8)

where Cφ is a real constant, i is the imaginary unit, and φ is the angle between k w and the x-

axis (see Fig. 2). k w = 0 corresponds to a plane wave traversing perpendicularly and
 
undisturbed through the plane z = 0; therefore, the phase factor ei kw  r carries the information
about the directionality of the leakage radiation associated to a plane wave propagating in the

direction defined by k w along the medium/PMMA interface. The optical disturbance at the
Fourier plane (FP), Uφ(kx,ky), is given by the Fourier Transform (FT) of Uφ(x,y,z = 0) [24],
i.e.:

#186361 - $15.00 USD Received 5 Mar 2013; revised 28 Apr 2013; accepted 3 May 2013; published 8 May 2013
(C) 2013 OSA 20 May 2013 | Vol. 21, No. 10 | DOI:10.1364/OE.21.011928 | OPTICS EXPRESS 11933
 (
− i xk x + yk y )
Uϕ ( k x , k y ) ∝ FT Uϕ ( x, y , z = 0)  ∝  Uϕ ( x, y , z = 0)e dxdy. (9)
Here k = 2π/Λ and 1/Λ is the spatial frequency. Using Eqs. (8) and (9) results, for the
excitation in a homogeneous sample of the surface wave described by Eq. (2), the following
expression:
( )
− i  ( k x − k w cosϕ ) x + k y − k w sin ϕ y 
Uϕ ( k x , k y ) ∝  Cϕ e dxdy
(10)
∝ δ ( k x − k w cos ϕ , k y − k w sin ϕ ).
Where δ(k) is the delta of Dirac function [24]. The intensity distribution at the FP is
proportional to the square of the absolute value of Uϕ ( k x , k y ) , i.e.:
2
Iϕ (k x , k y ) ∝ Uϕ ( k x , k y ) . (11)
From Eqs. (9)-(11) then follows that:
Iϕ (k x , k y ) ∝ δ 2 ( k x − k w cos ϕ , k y − k w sin ϕ ). (12)
The intensity distribution in the FP images obtained in the SWIS-microscope arrangement
[12,13,15,18,19] is the part of Iϕ ( k x , k y ) inside of the reciprocal space region captured by the

microscope objective lens, k ≤ ko NA ; i.e.:

= 0, k w > ko NA
Iϕ , FPI ( k x , k y )  , (13)
∝ δ 2 ( k x − k w cos ϕ , k y − k w sin ϕ ) , k w ≤ ko NA
Therefore, the intensity distribution in a FP image, corresponding to the excitation in a
homogeneous sample of the surface wave described by Eq. (2), is a bright spot at the
 
reciprocal space point ( k w cos ϕ , k w sin ϕ ) at the extreme of the vector k = k w . Basically,
Iϕ , FPI ( k x , k y ) is the observed far-field diffraction pattern, which is produced by the
interference between photons that leaks from the medium/PMMA interface in the direction

defined by k w . The interference occurs because it is impossible to know the path of the
photons arriving to the FP image at the spot ( k w cos ϕ , k w sin ϕ ) . I.e., photons that leaks in the
direction, φ, through different points in the medium/PMMA interface, and arrive to the same
spot in the FP image, are undistinguishable; therefore, they interfere [31–33]. However,
photons that leaks in different directions can be distinguished by the direction of their
momenta; thus, they cannot interfere [31–33]. In consequence, the total intensity at the FP,
IT(kx,ky), must be calculated by adding the intensity contributions Iϕ ( k x , k y ) corresponding to
all the surface waves illumination directions φ; i.e, IT(kx,ky) is described by the following
expression:
I T ( k x , k y ) =  Iϕ ( k x , k y ). (14)
ϕ

#186361 - $15.00 USD Received 5 Mar 2013; revised 28 Apr 2013; accepted 3 May 2013; published 8 May 2013
(C) 2013 OSA 20 May 2013 | Vol. 21, No. 10 | DOI:10.1364/OE.21.011928 | OPTICS EXPRESS 11934
 Fig. 3. FP images obtained with a SWIS-microscope arrangement corresponding to a uniform
sample where (a) SPPs and (b) surface waves related to TIR, were excited. When the direction
of propagation of the surface waves along the medium/PMMA interface, φ, changes from 0 to

2π, k w rotates describing a ring.

Consequently, assuming that k w ≤ ko NA , for a homogeneous sample the intensity
distribution in the FP image is a bright ring of radius k = kw centered at the origin of the
reciprocal space. Figure 3 shows FP images obtained with a SWIS-microscope arrangement
corresponding to two homogeneous samples [15,18]. A thin (thick) bright ring in Fig. 3(a)
(3(b)) corresponds to the enhanced transmission of fluorescence coupled to SPP [20,21,29]
(surface waves related to TIR [15]). The weak background in the FP images is produced by
the fluorescence that passed through the medium/PMMA interface without exciting surface
waves. The difference in brightness between the rings and the background corresponds to the
fluorescence enhancement transmission condition Elo / E†o >> Et / E†o . As illustrated in Fig. 3,

when the direction φ of the surface wave illumination changes from 0 to 2π, the vector k w
rotates describing a ring. The width of the thin ring in Fig. 3(a) is not zero because in reality
the propagation length of a lossy surface wave is not infinite, as it is implicit in Eq. (2), but
finite. As a consequence, the leakage radiation occurs in a very narrow but finite angle
interval [20,21,25], the spot ( k w cos ϕ , k w sin ϕ ) gets elongated in the radial direction, and the
transversal section of the ring exhibits a Lorentzian profile with a width inverse proportional
to the propagation length of the surface wave [25]. The thick ring in Fig. 3(b) corresponds to
the occurrence of TIR for a broad range of incident angles [15]. This is because, as sketched
in Figs. 1(e) and 1(d), surface waves are excited by the fluorescence with a different value of
k w for every angle larger than the TIR angle. It should be pointed out that the sum of
intensities in Eq. (14) is different than alternative approaches where the electric field
amplitudes corresponding to light leaked in different directions were added [34,35]. A
comprehensive study of this apparent discrepancy will be published elsewhere.
3.2 Periodic sample
For a sample with a sinusoidal profile of period p along the x axes, the optical disturbance (U)
at the gold/PMMA interface (z = 0) associated to El is proportional to the Bloch wavefunction
corresponding to the periodical structure, i.e [36]:

#186361 - $15.00 USD Received 5 Mar 2013; revised 28 Apr 2013; accepted 3 May 2013; published 8 May 2013
(C) 2013 OSA 20 May 2013 | Vol. 21, No. 10 | DOI:10.1364/OE.21.011928 | OPTICS EXPRESS 11935
  
Uϕ , p ( x, y , z = 0) = Cϕ 1 + sin ( 2π
p )
x  ei kw  r
(15)
= Cϕ 1 + sin 2pπ x  e ( w (
i xk cosϕ + yk w sin ϕ )
. )
From Eqs. (9) and (15) follows that the optical disturbance at the FP is now given by the
following expression:
( )
Uϕ , p (k x , k y ) ∝  Cϕ 1 + sin ( 2π
p )
x  e
− i ( k x − k w cosϕ ) x + k y − k w sin ϕ y 
dxdy. (16)
Thus:

{ )}
        
 (
  )
Uϕ , p (k ) ∝ δ  k − k w  + δ  k − G + k w  + δ  k − −G + k w  ,
 ( (17)

where G is the reciprocal lattice vector corresponding to the grating, i.e.:

G = G = 2pπ . (18)

Uϕ , p ( k ) is a shifted version of the optical disturbance corresponding to the far-field two-
dimensional diffraction pattern that would be obtained under traditional out-of-plane
perpendicular illumination [24]. Therefore, Eq. (16) can be rewritten as:
  
Uϕ , p ( k ) = U ⊥ , p ( k − k w ), (19)

where:
     
{
U ⊥ , p ( k ) ∝ δ  k  + δ  k − G  + δ  k + G  . (20) }
Consequently, if k w + G ≤ ko NA , the BFP image is formed by three bright spots at the
reciprocal space points ( k w cos ϕ , k w sin ϕ ) ,
( k w cos ϕ + G , kw sin ϕ ) , and
    
( k w cos ϕ − G , kw sin ϕ ) at the extreme of the vectors k w , G + k w , and −G + k w , respectively.
Substituting Eq. (19) into Eq. (14) permits to calculate the total intensity distribution at the
FP, IT(kx,ky), corresponding to the incoherent superposition of numerous plane waves
propagating in all directions along the medium/PMMA interface.
Figure 4 shows a FP image corresponding to a plasmonic sample with a periodic profile
along the x-axis, with p = 4 μm, for a case where k w + G < ko NA [18]. As shown in Fig. 4,

when the direction φ of the surface waves illumination changes from 0 to 2π, the vectors k w
with origin at the points ( 0,0 ) , ( + G , 0 ) , and ( −G ,0 ) rotate simultaneously around their
origins. As a consequence, the three spots ( k w cos ϕ , k w sin ϕ ) , ( k w cos ϕ + G, k w sin ϕ ) , and
( k w cos ϕ − G , kw sin ϕ ) simultaneously describe circumferences of radius k w centered at the
points ( 0, 0 ) , ( + G ,0 ) , and ( −G ,0 ) , respectively. Due to the non-sinusoidal profile of the
sample, in addition to the zero and first order rings, two arc-segments of the second order
rings centered at the points ( + 2G ,0 ) , and ( −2G ,0 ) can be seen in the FP image shown in
Fig. 4. The second order rings do not appear complete because in the FP image only appears
the fraction of the total intensity distribution IT(kx,ky), which is inside of the reciprocal space

region k ≤ ko NA .

#186361 - $15.00 USD Received 5 Mar 2013; revised 28 Apr 2013; accepted 3 May 2013; published 8 May 2013
(C) 2013 OSA 20 May 2013 | Vol. 21, No. 10 | DOI:10.1364/OE.21.011928 | OPTICS EXPRESS 11936
 Fig. 4. FP image corresponding to a plasmonic sample with a periodic profile along the x-axis,
with p = 4 μm. When the direction φ of the surface waves illumination changes from 0 to 2π,

the extreme of each vector k w describes a circumference.

3.3 General case

In general, the optical disturbance at the medium/PMMA interface corresponding to a single
plane wave propagating along the medium/PMMA interface is given by the following
expression:
 
Uϕ ( x, y , z = 0) = f ( x, y )ei kw  r = f ( x, y )e (
i xk w cosϕ + yk w sin ϕ )
, (21)

where f(x,y) is the optical disturbance at the medium/PMMA interface that would exist under
traditional perpendicular out-of-plane illumination. The corresponding optical disturbance at
the FP is then given by the following expression:
( )
− i  ( k x − k w cosϕ ) x + k y − k w sin ϕ y 
Uϕ ( k x , k y ) ∝  f ( x, y )e dxdy. (22)

Therefore, Uϕ ( k ) is proportional to the two-dimensional FT of f(x,y) after a translation in

the reciprocal space by k w , i.e.:
  
Uϕ ( k ) = U ⊥ ( k − k w ), (23)

where, U ⊥ ( k ) would be the optical disturbance at the medium/PMMA interface if the sample
was traditionally studied with perpendicular out-of-plane illumination i.e.:

#186361 - $15.00 USD Received 5 Mar 2013; revised 28 Apr 2013; accepted 3 May 2013; published 8 May 2013
(C) 2013 OSA 20 May 2013 | Vol. 21, No. 10 | DOI:10.1364/OE.21.011928 | OPTICS EXPRESS 11937
   
U ⊥ ( k ) ∝ FT [ f ( x, y ) ] ∝  f ( x, y )e − i k  r dxdy. (24)
 
In general, Uϕ ( k ) is different than zero in a large area, therefore Uϕ ( k ) is not an slide in
the direction φ of the BFP image as it was proposed in Ref. 18. The intensity distribution
 
Iϕ ( k ) at the FP corresponding to Uϕ ( k ) is given by the Eq. (11). The total intensity

distribution at the FP, I T ( k ) , must be obtained by adding the intensity

contributions Iϕ ( k ) corresponding to all the surface waves illumination directions φ;
therefore, it is described by the Eq. (14). Consequently, from Eqs. (11) and (14), and Eq. (24)
we obtain:

I T ( k ) ≈  aϕ U ⊥ ( k x − k w cos ϕ , k y − k w sin ϕ ) .
2
(25)
ϕ
2
Each term aϕ U ⊥ ( k x − k w cos ϕ , k y − k w sin ϕ ) in Eq. (25) is a shifted version with weight
aϕ of the FP image that would be observed under traditional perpendicular out-of-plane

illumination. Therefore, one can find, from the distribution of points where U ⊥ ( k ) ≠ 0 , the

distribution of reciprocal space points ( k x , k y ) where I T ( k ) ≠ 0 . In order to achieve this, one

can imagine, as shown in Fig. 4, a vector k w attached to each point ( k x' , k y' ) such that
U ⊥ ( k x' , k x' ) ≠ 0 (represented by red dots in Fig. 4). When the direction φ of the surface wave

illumination changes from 0 to 2π, the vectors k w rotate simultaneously around their origins.
This resulting in each point ( k x' , k y' ) being substituted by a ring of radius k w centered in that
point. The collection of all these rings form the distribution of reciprocal space points
( k x , k y ) where IT (k x , k y ) ≠ 0 . The total intensity distribution in the FP image, IT,FPI(kx,ky), is
the part of I T ( k x , k y ) captured by the microscope high NA objective lens; i.e.:

= 0, k > ko NA
I T , FPI ( k x , k y )  . (26)
= I T ( k x , k y ) , k ≤ ko NA
It should be pointed out here that the observation of full rings in the FP images (see
instance in Fig. 4) is conditioned by the existence of surface waves traveling in all directions
in the medium/PMMA interface. However, this is not always the case. It is well known that
two dimensional plasmonic and photonic crystals may have directional gaps [37–39].
Directional gaps are formed when the propagation in the crystal of surface waves are
prohibited in a range of directions. When this happen the weight coefficient in Eq. (25) is
equal to zero for all prohibited directions.
4. Origin of subwavelength resolution
In the SWIS-microscope arrangement, each surface wave illumination direction φ produces
an independent image of the object under observation, Iϕ ,SEI ( x, y ) , at the image plane of the
microscope. The intensity distribution of this image is given by the following expression:
2
Iϕ , SEI ( x, y ) ∝ Uϕ , IP ( x, y ) , (27)

#186361 - $15.00 USD Received 5 Mar 2013; revised 28 Apr 2013; accepted 3 May 2013; published 8 May 2013
(C) 2013 OSA 20 May 2013 | Vol. 21, No. 10 | DOI:10.1364/OE.21.011928 | OPTICS EXPRESS 11938
where Uϕ , IP ( x, y ) is the optical disturbance at the image plane corresponding to Uϕ , FP ( k x , k y ) ,

which is the part of Uϕ ( k ) captured by the microscope lenses, i.e.:

= 0, k > ko NA
Uϕ , FP ( k x , k y )  . (28)
= Uϕ ( k x , k y ) , k ≤ ko NA
Therefore, Uϕ , IP ( x, y ) is given by the inverse FT of Uϕ , FP ( k x , k y ) , i.e [24]:

(
i xk x + yk y )
Uϕ , IP ( x, y ) ∝ FT −1 Uϕ , FP ( k x , k y )  ∝  Uϕ , FP ( k x , k y )e dk x dk y . (29)
The final intensity distribution in the SE image obtained with the SWIS-microscope
arrangement [12,13,15,18,19], I SEI ( x, y ) , is formed by addition of the images corresponding
to all directions, i.e.:
I SEI ( x, y ) =  Iϕ , SEI ( x, y ). (30)
ϕ

The sum of the intensities corresponding to different directions at the medium/PMMA

interface was previously introduced for plasmonic crystals by showing that plasmon-coupled
leakage radiation superlenses can be used to directly probe the Bloch wavefunctions of the
photons in a two-dimensional crystal [36]. The correspondence between the SE image and the
electric field distribution in the medium/PMMA interface is due to the proportionality
between Elo in Eq. (4) and E†o in Eq. (2). The origin of the subwavelength resolution
capabilities of SWIS [13,15,19] can be better illustrated for simple periodical structures. As
discussed above, when k w + G ≤ ko NA , the zero and first order rings are observed in the FP
image corresponding to a one dimensional periodic structure (see Fig. 4); therefore, using
Eqs. (17) and (19), and Eqs. (28)-(29), results:
 
Uϕ , IP ( x, y ) ∝ FT −1 U ⊥ , p ( k − k w ) 
(31)
{ )}
       
  (  )
∝ FT −1 δ  k − k w  + δ  k − k w + G  + δ  k − k w − G  .
 (
Therefore, it follows from the shift property of FT and Eq. (31) that for any illumination
direction φ, the absolute value of the optical disturbance at the image plane Uϕ , IP ( x, y ) is
proportional to the optical disturbance that would be at the object plane under traditional
perpendicular out-of-plane illumination, i.e.:

Uϕ , IP ( x, y ) ∝ 1 + sin
x  . ( 2π
p ) (32)
From Eqs. (30) and (32) follows that every SE image obtained using directional surface
waves illumination, Iϕ ,SEI ( x, y ) , independently of the illumination direction φ, matches the
image that would be obtained under traditional perpendicular out-of-plane illumination.
Therefore, in correspondence with experimental results [18], from Eq. (30) follows than in the
case k w + G ≤ ko NA , the total intensity distribution in the SE image closely matches the image
that would be obtained under perpendicular out-of-plane illumination.
The more interesting case of k w + G > ko NA but k w − G ≤ ko NA is depicted in Fig. 5. A
schematic illustration of the transversal structure of a plasmonic SWIS (SWIS without metal)
is shown in Fig. 5(a) (Fig. 5(b)). The fabrication details of these samples have been published
elsewhere [15,19]. The plasmonic SWIS is a plasmonic crystal with square symmetry and
period of p = 300 nm formed by patterning the top PMMA + R6G layer; therefore, the
patterned air holes with a diameter of 100 nm were the object under observation in this

#186361 - $15.00 USD Received 5 Mar 2013; revised 28 Apr 2013; accepted 3 May 2013; published 8 May 2013
(C) 2013 OSA 20 May 2013 | Vol. 21, No. 10 | DOI:10.1364/OE.21.011928 | OPTICS EXPRESS 11939
sample [19]. The SWIS without metal is a photonic crystal with square symmetry and period
of p = 220 nm formed by ~35 nm thick chromium (Cr) deposited over a ~150 μm thick
coverslip substrate. A 110 nm thick layer of PMMA doped with R6G was then spun on top of
the whole structure, and finally, in order to increase the subwavelength resolution, a drop of
water was placed on top of the fabricated SWIS; therefore, the Cr pillars were the object
under observation in this sample [15]. Cr was used here because it does not exhibit a well-
defined plasmonic signature, can be simply patterned after deposition on a glass substrate,
and it provides high SE contrast images. Therefore we denote the sample sketched in Fig.
5(b) as an SWIS without metal.

Fig. 5. Schematic illustration of the transversal structure of a (a) plasmonic SWIS, and (b)
SWIS without metal. (c) FP image corresponding to the plasmonic SWIS. (d) Intensity
distribution inside of the zero order ring resulting from subtracting the FP image corresponding
to a homogeneous SWIS without metal, from the FP image corresponding to the SWIS without
metal with a periodic patterned structure.

Figure 5(c) shows the FP image obtained with a SWIS-microscope arrangement

corresponding to the plasmonic SWIS with the structure sketched in Fig. 5(a). The zero order
ring and arc-segments of the first order rings distributed with square symmetry are clearly
seen in this image. A similar FP image was obtained using the SWIS without metal with the
structure sketched in Fig. 5(b) [15]; however, the first order rings were faint in the image.
This indicates that the in-plane scattering of the surface waves was more efficient in the
plasmonic SWIS than in the SWIS without metal [34]. Nevertheless, the presence of the faint
first order rings is revealed in Fig. 5(d), which shows the intensity distribution inside of the
zero order ring, resulting from subtracting the FP image corresponding to the homogeneous
SWIS without metal with the structure sketched in Fig. 1(a) (see Fig. 3(b)), from the FP
image corresponding to the SWIS without metal with the structure sketched in Fig. 5(b).
Similarly to the FP image shown in Fig. 5(c), arc-segments of the thick first order rings
distributed with square symmetry are clearly seen in Fig. 5(d).
The graphical compositions shown in Fig. 6 illustrate how the FP images shown in Fig. 5
are formed by the superposition of rings of radius k w centered at the spots corresponding to
the diffraction pattern that would be obtained using traditional perpendicular out-of-plane

#186361 - $15.00 USD Received 5 Mar 2013; revised 28 Apr 2013; accepted 3 May 2013; published 8 May 2013
(C) 2013 OSA 20 May 2013 | Vol. 21, No. 10 | DOI:10.1364/OE.21.011928 | OPTICS EXPRESS 11940
illumination. Only the arc-segments of these rings captured by the high NA objective lens are
observed in the FP image. Neighbors spots are separated in the reciprocal space a distance G
= 2π/p. A well-known result of the Abbe’s theory of image formation is that the periodic
structure of the sample would not appear in the image if the first order spots were not
captured by the microscope lenses [2]. This is what happens in the FP images shown in Figs.
5 and 6; therefore, the periodic structures existing in the samples sketched in Fig. 5(a) and
5(b) could not be observed in the SE images that were obtained using traditional
perpendicular out-of-plane illumination. This is in correspondence with the minimum
observable period of pmin~381 nm calculated from (1) with NA = 1.49. However, the periodic
structures with period of p = 300 nm and 220 nm sketched in Fig. 5(a) and 5(b), respectively,
have been successfully imaged (not shown here) using a SWIS-microscope arrangement
[15,19]. Consequently, the subwavelength resolution capabilities of SWIS have already been
unambiguously demonstrated.

Fig. 6. Graphical composition illustrating how the FP images shown in (a) Fig. 5(c), and 5(b)
Fig. 5(d) are formed. The red spots represent the diffraction pattern that would be observed
using traditional perpendicular out-of-plane illumination. Similar fractions of the first order
rings are observed in both images.

In what follows we will show that the subwavelength resolution capabilities of SWIS can
be easily understood using the Fourier optics description of the image formation in a SWIS-
microscope arrangement presented above. The diffraction pattern produced by illuminating
the samples with surface waves propagating in the direction φ is formed by the five spots

pointed by the vectors k w , which are represented by red arrows in Fig. 6. The rings are

formed when the vectors k w rotated around the red spots shown in Fig. 6. Only the
diffraction spots captured by the high NA objective lens are observed in the FP image. When
φ~3π/4 in the instance illustrated in Fig. 6(b), only the diffraction spot at the reciprocal space

( )
point S = − 2 2 k w , 22 k w is contained in the region k ≤ ko NA ; therefore, using Eq. (29),
results:

{ )}

4
U 5π , IP  (
( x, y ) ∝ FT −1 δ  k − − 2 2 k w , 22 k w  .

(33)
Therefore, from Eqs. (27) and (33) follow that I 5π 4 , SEI ( x, y ) is constant. This means that
the periodic structure in the object under observation cannot be observed in the image formed
by using only surface waves propagating in the direction φ = 3π/4. This is exactly what would
be seen using traditional perpendicular out-of-plane illumination. Nevertheless, the final

#186361 - $15.00 USD Received 5 Mar 2013; revised 28 Apr 2013; accepted 3 May 2013; published 8 May 2013
(C) 2013 OSA 20 May 2013 | Vol. 21, No. 10 | DOI:10.1364/OE.21.011928 | OPTICS EXPRESS 11941
intensity distribution in the image is formed by superposition of the images corresponding to
all directions. In the instances illustrated in Fig. 6, there are several directions for which first
order spots are captured by the microscope objective lens. As a result, the periodic structure
of the object under observation is stamped in the final intensity distribution in the
image, I SEI ( x, y ) . This is the origin of the subwavelength resolution capabilities of SWIS.
Consequently, for a SWIS-microscopy arrangement we can rephrase the traditional enunciate
of the Abbe’s theory of image formation saying that, if a fraction of the first order rings can
be captured by the high NA objective lens, then the periodic structure of the sample will be
visible in the image [13,15]. This results into the following expressions for the minimum
period observable using SWIS [13,15]:
λo
pSPP ,min = , (34)
NA + neff

λo
pTIR ,min = , (35)
nsup + nsub

where, neff is the effective refractive index experienced by the excited SPP in plasmonic
SWIS; and nsub and nsup are the refractive index of the media below and in top of the
medium/PMMA interface in the SWIS without metal. It is worth noting that the minimum
period observable is related with the Rayleigh resolution criteria [2] giving the minimum
resolvable separation between two points, Δx, by the relation Δx = pmin/2 [1,13,15]. Therefore,
from Eq. (1) with NA = 1.49, results a value of Δx ~λo/3~190 nm; meanwhile, Δx values as
small as λo/7~80 nm have been demonstrated using SWIS [13,15]. Although super-resolution
can be achieved for periodic and non-periodic samples [13,19] with any SWIS, the wave
excitation choice will depend on several factors such as complexity of sample fabrication,
subwavelength resolution limit, and contrast definition of the optical images. For instance,
surface waves related to total internal reflection require simpler samples dispensing the need
of metal layers when compared to samples investigated by SPP wave excitation. However, it
follows from Eq. (34) that deeper subwavelength resolution should be achievable using
plasmonic SWIS having a medium/PMMA interface where neff >>1. Values of neff much
larger than nsub and nsup are possible to obtain in plasmonic SWIS due to non-linear dispersion
relation of SPP [16].
5. Conclusions
We have presented, for the first time, a Fourier optics description of the image formation in a
SWIS-microscope arrangement. We found an excellent correspondence between the
analytical expressions obtained using the Fourier optics approach and the experimental
images. We have shown that the same physical principles are responsible for the
demonstrated subwavelength resolution capabilities of both plasmonic SWIS and SWIS
without metal. In general, SWIS are based in the illumination of the object under observation
with surface waves excited by fluorescence, the enhanced transmission of fluorescence via
coupling with surface waves, and the occurrence of far-field coherence-related fluorescence
diffraction phenomena. Our theoretical approach explains why images with subwavelength
resolution can be formed directly in the microscope camera, without involving scanning or
numerical post-processing. Finally, we have suggested alternative solutions to obtain deep
optical subwavelength resolution with a SWIS-microscope arrangement.
Acknowledgments
This work was partially supported by the NSF CAREER Award (ECCS-0954490).

#186361 - $15.00 USD Received 5 Mar 2013; revised 28 Apr 2013; accepted 3 May 2013; published 8 May 2013
(C) 2013 OSA 20 May 2013 | Vol. 21, No. 10 | DOI:10.1364/OE.21.011928 | OPTICS EXPRESS 11942