A high numerical aperture parabolic mirror as imaging device for confocal microscopy

M. A. Lieb; A. J. Meixner

doi:10.1364/OE.8.000458

1. Introduction

Parabolic mirrors with small numerical aperture (NA) are very common in astronomical telescope design and telecommunications. High NA parabolic mirrors have been used as efficient light collectors in single molecule spectroscopy [1–5], however, for imaging they have been avoided due to their poor off-axis imaging properties. The classical imaging devices for microscopy and imaging are objective lenses, because of their wide field of view and excellent resolution resulting from the high degree of correction and the high achievable NA. Parabolic mirrors focus a parallel beam, travelling along the optical axis, to one point without any aberrations in the geometrical optics approximation and in this sense they are ideal focusing devices. The drawback is that slight deviations of the incident beam from the optical axis or from parallelism give rise to huge aberrations for a high NA mirror resulting in a very small field of view.

Confocal microscopy is a point to point imaging technique and hence has to correct aberrations only for one single object point. This purpose is perfectly solved by parabolic mirrors in a stage scanning microscope where the focus rests on the optical axis. In addition it is fairly easy to achieve a high aperture without chromatic aberrations since the deflecting surface can be made highly reflective over a large spectral range.

In this work we investigate in the first section theoretically the focusing properties of a parabolic mirror with high NA and compare them with an aplanatic objective lens for linear and as well for radially polarized illumination. In a radially polarized beam the electric field vector is everywhere oriented radially with respect to the optical axis [6]. We show that high aperture parabolic mirrors illuminated with radially polarized light have the property that they produce a large longitudinal component of the electric field in the focus which is approximately fourteen times stronger then the transversal components also present near the focus. This property is especially interesting for all applications where a strong longitudinal field component is desirable, e. g. (1) for strong local field enhancement with a sharp metallic tip in near-field optics [7, 8], or (2) for polarization selective spectroscopy of nano-particles or single molecules in three spatial dimensions.

In the second section the confocal imaging of a dipole emitter, such as a single molecule, is discussed likewise for linear as well as for radially polarized excitation light. The use of radially polarized light and a reduced illumination aperture yields a distribution of longitudinal and in-plane components of the electric field which allows an easy determination of the emission dipole orientation.

2. Focusing

In this section the properties of a parabolic mirror as focusing element are discussed for linear and radial polarized incident light. The basic machinery by which polarized beams in focal regions can be analyzed was described by Richards and Wolf [9] using a decomposition of the incident field into an angular spectrum of plane waves. The adaptations for radially polarized beams are discussed in reference [6]. Basic work concerning the diffraction limited focusing with parabolic mirrors was published by Ignatovsky in 1920 [10] and numerical calculations for a linear polarized plane wave by Sheppard [11] and with more detail by Barakat [12] and recently for mirrors with arbitrary openings, particularly with openings larger then π, by Varga and Török [13, 14].

Fig. 1. Imaging geometry for a) a parabolic mirror and b) an aplanatic objective lens (fulfilling the Abbe sine condition r=f_o sinθ_o ). A ray incident at a distance r from the optical axis (z) reaches the sample under a certain angle θ_m in the mirror setup, while the same ray hits the sample under a much smaller angle θ_o in the lens setup. While the perpendicular component (s) of the electric field E changes sign when it is reflected on the mirror surface, it does not when it is refracted by the lens. p denotes the parallel component of E, f_m and f_o denote the focal spheres and s is a unit vector pointing from the focal sphere to the focus. c) Intensity distribution used for linear polarized light (solid line, Gaussian beam) and radially polarized light (dashed line, Bessel-Gauss beam), with the beam waist radius w₀ =3/2× pupil radius).

Download Full Size | PDF

We compare the parabolic mirror with an objective lens and emphasize the peculiarities of the mirror. Fig. 1 shows a parabolic mirror and an aplanatic objective lens both with a geometrical aperture of sinθ=1.0. Note that the focal length of the objective lens is twice that of the parabolic mirror. Hence the rays at the rim of the beam in both cases hit the focal point under an angle of incidence θ of π/2. The parabolic mirror reflects a ray incident at a given distance r from the optical axis to a large angle θ_m . The same ray is refracted by the objective lens to reach the focal point at a smaller angle θ_o . This means, due to energy conservation, that in a parabolic mirror much of the incident energy reaches the focus under high angles while the contrary is valid for an objective lens. This has noticeable consequences as will be seen later.

A second difference is, that the reflection at the mirror surface causes a phase jump of π to the component of the electric field which is polarized perpendicular to the plane of incidence (s) while the refraction by the objective lens does not. Note also the inverted sign of the p-polarized x,y-component: it points towards the optical axis for the objective and away from it for the mirror. Both together lead to inverted x,y-components for the mirror, which has only minor consequences. Furthermore, it can be seen that field components along the mirror axis arise in the focal region due to the high aperture.

2.1. Theory

We will give a short summary of the equations needed for the calculation of the optical field near the focus of a parabolic mirror, especially we will mention the differences to the case of an aplanatic lens as described in references [9] and [6] respectively. At optical frequencies, only the electric field is responsible for the interaction with matter (scattering, fluorescence, excitation, polarization, etc.). Hence only the electric fields will be discussed.

Fig. 1 shows the geometry of the problem with the used coordinate system. Starting point of the calculations is the angular spectrum representation of the incident field. The time independent part of the electric field at a point P (in polar coordinates r_P , θ_P and φ_P ) near the focus can then be written as (neglecting a constant phase term due to the total path length)

E (P) = - \frac{i k f}{2 π} \iint_{Ω} E_{f} (θ, φ) e^{i k s \cdot r_{P}} \sin θ d θ d φ

[15] with the wave number k (in the medium), the focal length f, the solid angle Ω, the electric field at the focal sphere E _f , which is calculated by geometrical optics rules, a unit vector in the direction of propagation s, the angles θ and φ as shown in Fig. 1. Complete derivations for an aplanatic lens are given in references [6, 9]. The analogue derivation for a parabolic mirror yields the electric field at a point P near the focus of an ideally conducting parabolic mirror (see section 5.).

a) For light which is linear polarized along the x-axis the field components are

E_{x} = \frac{i k f}{2} (I_{0, l} + I_{2, l} \cos 2 φ_{P}),

The integrals I _0,l, I _1,l and I _2,l are defined as

I_{0, l} = \int_{α_{0}}^{α_{1}} l_{0} (θ) \frac{2 \sin θ}{1 + \cos θ} (1 + \cos θ) J_{0} (k r_{P} \sin θ \sin θ_{P}) e^{- i k r_{P} \cos θ \cos θ_{P}} d θ,

where the J _n are Bessel functions of the first kind of order n. α₁ is the opening angle of the mirror, α₀ the smallest angle of acceptance not hidden by the sample, θ the angle θ_m as shown in Fig. 1 and l₀ (θ)=N exp(-r ²(θ)/ $w_{0}^{2}$ ) the electrical field amplitude of the illuminating laser beam. The normalization constant N is used to compare the objective lens and the mirror quantitatively. r is the distance from the optical axis as shown in Fig. 1 and w₀ is the waist radius of the illuminating Gaussian beam.

The first difference to the equations by Richards and Wolf [9] is the apodization factor 2/(1+cosθ) instead of the cos^½ θ. This factor is due to energy conservation. All energy flowing through a surface element in the parallel illuminating beam must flow through the corresponding surface element on the focal sphere f_m . A derivation can be found in [16]. The second difference is the positive sign for e_x and e_y that originates from the phase jump at the mirror surface. The z component arises from the reflection of components parallel to the plane of incidence (p-polarized) only and hence has no phase jump at all. The resulting effect is that the direction of rotation of the electric field vector in a plane perpendicular to the x, y-plane is inverted (see Fig. 4). A third difference is the negative exponent, which reflects just the inverted direction of propagation of light in the negative z direction.

b) We discuss the results for radially polarized illumination. Theoretical aspects concerning inhomogeneously polarized light, like radially polarized light, are discussed in references [6, 17–21], their generation in references [22–28], the simple generation from linear polarized light [28] should be mentioned specially. For this case we find the following equations for the electric field at a point P near the focus:

E_{r} = k f I_{1, r},

E_r is the radial component and E_a the azimuthal component showing the radial symmetry. The integrals are here defined as:

I_{0, r} = \int_{α_{0}}^{α_{1}} l_{0} (θ) \frac{2 \sin^{2} θ}{1 + \cos θ} J_{0} (k r_{P} \sin θ \sin θ_{P}) e^{- i k r_{P} \cos θ \cos θ_{P}} d θ,

with the field distribution l₀ (θ)=N exp(-r ²(θ)/ $w_{0}^{2}$ ) J₁(2 r(θ)/w₀ ) of a Bessel-Gauss beam.

The same differences between an objective lens [6] and a parabolic mirror are observed as mentioned above except that we have only parallel (p-polarized) components at the reflection in the case of radially polarized light and hence no phase jump.

2.2. Numerical field calculations

The results presented here are all calculated for the special case of a parabolic mirror with a numerical aperture of 1.515 (oil, index of refraction n=1.518), corresponding to a half opening angle of 86.4°, which is experimentally easily feasible. Such a mirror can be realized by a glas paraboloid with a high reflection coating on the curved surface. For the calculations a focal length f=4.5 mm was used and a central spot of 3.0 mm in diameter of the incident beam is blocked by a sample, corresponding to α₀ =18.9°. Although the results for an objective lens are well known, we present them here for direct comparison with the field distribution in the focal region of the parabolic mirror. In the case of the objective lens the same focal length f of 4.5 mm was used and a numerical aperture of 1.40 (oil, n=1.518) corresponding to a half opening angle of 67.3°, which is a typical value for an objective lens with high NA.

Fig. 2. Contour plots of constant |E|² in the x, y-plane of a focused Gaussian beam polarized along the x-direction in the upper row for an objective lens (NA=1.40, n=1.518, f_o =4.5 mm) and in the lower row for a parabolic mirror (NA=1.515, n=1.518, f_m =4.5 mm). a) and e) show |E|², b) and f) show |E_x|², c) and g) show |E_y|²×8, d) and h) show |E_z|². A logarithmic scale is used with a factor of two between adjacent contour levels. The length of the scale bar corresponds to one wavelength divided by the refractive index n.

Download Full Size | PDF

Fig. 3. Contourplots of constant |E|² in the focal region of a focused Gaussian beam polarized along the x-axis in the upper row for an objective lens and in the lower row for a parabolic mirror. a) and e) show sections of |E|² in the x,z-plane; b) and f) show sections of |E_x|² in the x,z-plane; c) and g) show sections of |E_y|2×8 along the z-axis and the diagonal between the x- and y-axis; d) and h) show sections of |E_z|2 in the x,z-plane. A logarithmic scale is used with a factor of two between adjacent contour levels. The length of the scale bar equals one wavelength divided by the refractive index n.

Download Full Size | PDF

2.2.1. Linear polarization

The illuminating beam has a Gaussian beam profile with a beam waist w₀ of 3/2 times the pupil radius. It is beyond the scope of this paper to discuss the optimum value of the beam waist.

Fig. 2a) and Fig. 2e) show contour plots of constant |E|² in the x, y-plane of a focused Gaussian beam polarized along the x-direction for an aplanatic objective lens and a parabolic mirror respectively. A logarithmic scale is used with a factor of two between adjacent contour levels. It can be seen that, (1) due to the higher aperture of the mirror system the size of the focal spot is slightly reduced but the distortion is more pronounced, (2) the energy is more spread out in the x, y-plane of the mirror. This fact is caused by the field incident under high angles. It forms a standing wave in the x, y-plane as Fig. 3e) and Fig. 4 show as well. And (3) the y and z components are enhanced by the mirror as Fig. 2c), g) and Fig. 2d), h) show. This fact is disturbing in conventional polarization microscopy. For single molecule detection however this might be an advantage as will be seen later.

Fig. 3 shows sections of |E|² parallel to the optical axis. While in the case of the objective lens the depth of the focus has a full width at half maximum (FWHM) of 1.51 λ/n, it is smaller in the mirror case namely FWHM=1.00 λ/n. The reason is again found in the energy incident from the high angles. Most of the energy is incident on a cone with a half opening angle of about 30° in the objective lens while this angle is about 50° for the mirror (see also Fig. 4). The interference region of this cone-shaped beam is either elongated in the optical axis, as for the objective lens, or more spread out in the x, y-plane, as for the mirror.

Fig. 4. (400kB) Movie showing a section through |Re(E(t))|² in the x,z- plane close to the focus for an objective lens (a) and a parabolic mirror (b) with linear polarized illumination along the x-axis. A logarithmic scale is used with a factor of v2 between adjacent contour levels. The vectors show the orientation of the electric field. The time between successive images is one 12^th of the optical period. The length of the scale bar is equal to one wavelength λ/n.

Download Full Size | PDF

Fig. 5. Contourplots of constant |E|² in the x, y-plane of a focused radial polarized Bessel-Gauss beam in the upper row for an objective lens and in the lower row for a parabolic mirror. a) and d) show |E|², b) and e) show the radial component |E_r|², c) and f) show the longitudinal component |E_z|². A logarithmic scale is used with a factor of two between adjacent contour levels. The length of the scale bar equals one wavelength λ/n.

Download Full Size | PDF

Fig. 6. Contourplots of constant |E|² in the focal region of a focused radially polarized Bessel-Gauss beam in the upper row for an objective lens and in the lower row for a parabolic mirror. a) and d) show |E|², b) and e) show the radial component |E_r|², c) and f) show the longitudinal component |E_z|². A logarithmic scale is used with a factor of two between adjacent contour levels. The length of the scale bar corresponds to one wavelength λ/n.

Download Full Size | PDF

Fig. 4 shows a movie of the time evolution of the electric field near the focus of a Gaussian beam, which is linearly polarized along the x-axis, formed by an objective lens (a) and a parabolic mirror (b). Contour plots of |Re(E(t))|² in the x,z-plane, where Re denotes the real part, are overlaid by the orientations of the field vectors at several points.

The formation of the interference of the intensity from the angles which contribute most of the energy appears very clearly in this movie. In the case of the objective lens the wavefronts shown in Fig. 4 has been measured by Schrader and Hell [29].

2.2.2. Radial polarization

For radially polarized illumination a Bessel-Gauss beam with a beam waist w₀ =3/2×pupil radius was used.

Fig. 5 and Fig. 6 compare contour plots of |E|² for radial polarization for an objective lens and a parabolic mirror in the x, y-plane and perpendicular to it. In the case of radial polarization major differences can be seen: The central spot is very sharp in the parabolic mirror and rather blurred for the objective lens. By comparing the radial and longitudinal components the effect is obvious: The radial components are formed by the intensity incident under small angles θ while the longitudinal components emerge from the large angles (see Fig. 6b), c), e) and f)). Due to the higher NA and the high concentration of energy in the large angles the parabolic mirror produces predominantly longitudinal components, while the objective lens produces longitudinal and radial components in comparable amounts. The blurred spot in the objective lens is formed by the superposition of the central longitudinal and the annular radial components. We observe here as well a reduced longitudinal size of the focal region in the mirror (mirror: FWHM=1.18 λ/n, objective lens: FWHM=1.65 λ/n). The ratio between the longitudinally and laterally polarized intensities is 14.3 in the parabolic mirror while it is only 3.50 in the objective lens.

To draw a quantitative comparison between objective lens and parabolic mirror the total intensities of the parts of the illuminating beams which are focused, have to be normalized. In Fig. 7 the intensities are normalized to 1 mW. The parabolic mirror shows a slightly larger intensity of the electric field for linear polarized light and a strongly enhanced intensity at the center for radially polarized light as compared to an aplanatic objective lens.

Fig. 7. Variation of the electric field intensity, |E|², in the x, y-plane of an objective lens (a) and a parabolic mirror (b) for linear polarized illumination along the x-axis (along the x-axis --, along the y-axis …) and radial polarization (full line). Both figures are normalized to the same optical power (1 mW) focussed by the objective lens and by the parabolic mirror. Hence the field strength can directly be obtained from √|E|².

Download Full Size | PDF

2.3. Alignment and coma

The drawback of a parabolic mirror is that it shows strong coma if it is not illuminated exactly along its optical axis. A slight misalignment of the propagation direction of the illuminating beam causes already a strong distortion of the field distribution near the focus. Hence a proper alignment of the illuminating beam is critical for the use of a parabolic mirror. Fig. 8 shows movies of |E|² in the focal region of a parabolic mirror illuminated with radially polarized light with a wavelength of 500 nm for different angles of incidence. From image to image the angle of incidence is increased by 0.00318° starting with 0°. Calculations for linear polarized light are comparable. To calculate the field in the focal region an additional phase term is included into the integrals of equation 5. This phase term takes the tilted wave front into account. The integration is performed over the focal sphere f _m, as proposed by Sheppard [30]. A deviation of the angle of incidence by more than about 0.006° causes a severe asymmetry of the field and reduces the maximum intensity by more then a factor of two. Hence the propagation direction of the illuminating beam has to be aligned to the optical axis of the mirror better than 0.006°. This accuracy can be reached by a steering mirror equipped with micrometer adjustment screws.

Fig. 8. Movies showing the formation of coma in the focal region of a parabolic mirror illuminated with radially polarized light. Successive images show the electric field intensity |E|² for an increasing angle (0°, 0.00318°, 0.00637°, 0.00955°, 0.01273° and 0.01592°) between the mirror axis and the k-vector of the incident wave. a) (91kB) section in the x, y-plane and b) (97kB) section in the x, z-plane. A logarithmic scale is used with a factor of two between adjacent contour levels. The length of the scale bar equals one wavelength λ, which is set to 500 nm, divided by the refractive index n.

Download Full Size | PDF

3. Single molecule imaging

The observation of single molecules and the determination of their orientation earns growing interest [31–34]. We analyze the detection of a dipole emitter in a homogeneous environment through a telecentric optical system consisting of a parabolic mirror and an aplanatic objective lens with small NA. A detector with limited detection area, as the commonly used avalanche photo diodes, measures the intensity in the image focus. Telecentric systems have been studied by Sheppard [35]. A comparable study for an objective lens instead of a parabolic mirror has also been published by Sheppard [36] and recently by Enderlein [37].

3.1. Fundamentals

The major difference to the objective lens is here again the apodization factor of the mirror and the phase jump of π for the perpendicular components at the reflection as well as the aberrations for an emitter located outside the geometrical focus. The electric far-field caused by a dipole emitter can be written as

E_{de} = \frac{1}{\sqrt{4 π ε_{o}}} k^{2} (s_{m} \times p) \times s_{m} \cdot \frac{1}{r} \cdot \exp (i k r - ω t)

in SI-units, where ε₀ is the electric permittivity of vacuum, k the wave number, p is the dipole moment and s_m a unit vector pointing in the direction of observation, r the distance from the dipole to the point of observation, ω the angular frequency and t the time [38]. Transversal components (s-polarized, indexed ⊥) suffer the phase jump of π while parallel components (indexed ‖) are deflected by the mirror and the objective lens. Using equations 1 and 6 and neglecting a constant phase term due to the total path length, the time independent part of the electric field in the secondary focus can be written as

Fig. 9. Scheme illustrating the confocal imaging geometry of a point-dipole emitter located in the focus of a parabolic mirror with a focal length f_m . s_m is a unit vector pointing from the focus to the mirror surface. The light is collected by a lens with focal length f_o and s_o is a unit vector pointing from the surface of the focal sphere of the lens to the image focus on the optical axis.

Download Full Size | PDF

E (r_{o}) = - \frac{i k^{3} f_{o}}{\sqrt{16 π^{3} ε_{o}}} \int_{θ_{o, min}}^{θ_{o, max}} \int_{0}^{2 π} \frac{(1 + \cos θ_{m})}{2 f_{m}} \sqrt{\cos θ_{o}} [(p \cdot e_{m, ∥}) e_{o, ∥} -

where f _o and f _m are the focal length of mirror and objective lens respectively. A schematic representation of the geometry with the angles θ_m and θ_o is shown in Fig. 9. k is the vacuum wave number, n_m the refractive index of the mirror material, r _m is the vector from the geometrical focus of the mirror to the location of the molecule and r _o the vector from the geometrical focus of the objective lens to the point of observation. The unit vectors e _⊥, e _m,_‖, e _o,‖, s_m and s_o are defined as follows:

e_{⊥} = (- \sin φ, \cos φ, 0),

2/(1+cosθ_m ) is the apodization factor of the mirror, cos^½ θ_o that of the lens. The term in square brackets describes the orientation of the electric field on the focal sphere of the lens, the exponential factor accounts for the phase corrections for points outside the mirror- and lens-focus and sin θ_o dθ_o d φ is the solid angle element. Using the magnification M=f_o /f_m , θ_m and the integration limits θ _{o, min} and θ _{o, max} can be given by

θ_{m} = 2 \arctan (\frac{M}{2} \sin θ_{o}),

where NA_m is the numerical aperture of the mirror, n_m the refractive index and r_samp the radius of the central part of the mirror which is covered by the sample holder.

In the case of an objective lens as collection optics the linear relation sinθ_m =M sinθ_o is valid. So a shift of the molecule out of the optical axis can be simulated by just moving the image of the in-focus dipole by -M times this shift. This can be seen by analyzing the exponential term of equation 7. The relation connecting sinθ_m and sinθ_o in the case of a parabolic mirror as collector is strongly nonlinear in M (see equation 9). The images of off-axis lying molecules will hence be distorted. The longitudinal magnification as well does not equal M ², as it would in the paraxial approximation. A molecule located on the optical axis, but slightly moved out of the geometrical focus, does not produce the same wave front in the parallel beam as it would by passing through a thin lens.

For calculating the intensity measured by a detector positioned in the geometrical focus of the lens while scanning the dipole, equation 7 must be integrated over the detection area. The ratio of the detected to the emitted intensity as a function of the emitters position yields the collection efficiency function (CEF). When the numerical aperture of the image forming objective lens is small, the electric and magnetic fields in the focal region are almost perpendicular to each other and their absolute values have a nearly constant ratio. Hence the energy flow 1/(2µ₀ ) Re(E×B*) is proportional to E·E* (the asterisk denotes complex conjugation and µ₀ is the vacuum permeability) and it is sufficient to know the electric field in the focus. The detected intensity is thus proportional to the spatial integral ∫ _A E·E*da over the detector area A. Since the field distribution in the detector plane depends strongly on the position of the emitter in the focal region, the field distribution must be calculated for each position of the dipole to obtain the CEF. To reduce computational effort symmetry considerations can be taken into account.

The fluorescence rate R(r _m) of a dipole emitter located at r _m can be written as

R (r_{m}) = c {∣ p \cdot E (r_{m}) ∣}^{2}

at sufficiently low intensities, where saturation effects can be neglected [39]. E(r _m) is the electric field vector of the excitation field at the position of the emitter, p its absorption dipole moment and c a constant. For the following calculations it is assumed, that the absorption and emission dipole moments are equal and the fluorescence quantum efficiency is unity. The confocal image of the dipole emitter is obtained by multiplication of the detection efficiency function with the fluorescence rate caused by an illuminating electrical field.

3.2. Image distortion and collection efficiency

All numerical results are calculated for a parabolic mirror with focal length 4.5 mm, numerical aperture 1.515 (oil, refractive index=1.518) and a shadowed center of 3 mm in diameter. The objective lens has a focal length of 225 mm yielding a magnification M of 50. The illuminated aperture of the objective lens is small, NA=0.0375, and hence the approximation of the energy flow by c·E·E* is still good (where c is a constant).

Fig. 10. Images of |E|², for a point dipole emitter in the focal region of a parabolic mirror. In the images a) - c) the dipole lies in the focal point at r _m=(0, 0, 0) and is oriented along the x, y and z-axis, respectively. For the images d) to f) the dipole lies one wavelength away from the optical axis at r _m =λ/n (cos 30°, sin 30°, 0) indicated by the arrows. The black circles in the images a) and c) have a diameter of 1.22 λ/NA_o , corresponding to the diameter of the Airy disk. A logarithmic scale is used with a factor of 2 between adjacent contour levels. The scale bar length is the magnification M times the vacuum-wavelength λ.

Download Full Size | PDF

Fig. 10 shows images of |E|² of a dipole emitter in the focal region of a parabolic mirror as seen in the plane of the detector. The dipole lies in the focal point in Fig. 10a) - c) and is moved by one wavelength away from the optical axis in Fig. 10d) - f). Depending on the dipole orientation different images are obtained (see also [35]). The symmetry of the field distribution is reduced when the dipole is moved away from the focus. Hence, in contrast to an aplanatic objective lens, a high NA parabolic mirror has one focal point instead of a focal plane. Note as well the similarity of the intensity (and also polarization) distribution of a dipole oriented along the optical axis (Fig. 10c)) to that of a radially polarized beam.

The black circles in Fig. 10a) and c) have a diameter of 1.22 λ/NA_o corresponding to the diameter of the Airy disk produced by a plane wave in paraxial approximation. While for a dipole oriented in the x, y-plane almost all energy of the main maximum is collected within this circle it is not for a longitudinally oriented dipole. To collect light from longitudinally oriented dipoles with high efficiency the diameter of the detector has to be increased slightly at the expense of a slightly reduced resolution. For the following results a factor of 1.2 was used. The resulting detection efficiency is 0.39 for a longitudinally oriented dipole and 0.37 for a dipole oriented perpendicular to the optical axis. In the case of the objective lens (as described in paragraph 2.2.) as collector the detection efficiency reduces to about 0.30 for the dipole in the x, y-plane and to 0.16 for the longitudinally oriented one, due to the smaller aperture.

Fig. 11 shows contour plots of the collection efficiency function (CEF) for different orientations of a dipole emitter scanned either in the x, y-plane, a), or in the x, z-plane, b), of a parabolic mirror collector. Only orientations of the dipole axis with an azimuth angle φ between 0 and 90 degree are shown. Orientations between 90 and 360 degree can be obtained by symmetry considerations. Note that a linear intensity scale is used in these figures.

Fig. 11. Contour plots of the collection efficiency function (CEF) for different orientations of a dipole emitter scanned a) in the x, y-plane and b) in the x, z-plane of a parabolic mirror collector. The angles between the dipole moment p and the coordinate system, θ and φ, are shown in the lower left images. A linear intensity scale is used with a difference of 0.05 between adjacent contour levels. The maximum of the scale corresponds to a collection efficiency of 0.5. The scale bar has a length of one wavelength λ/n.

Download Full Size | PDF

The reason for the asymmetry of the collection efficiency for the in-plane scan (Fig. 11a) is given by the asymmetry of the corresponding distribution of energy (see Fig. 10) due to the sin² θ direction dependence of the dipole radiation. An elongation of the main maximum in x-direction causes an elongation of the CEF in y-direction for instance. The point symmetry in the x, z-scan can be understood by the geometrical optics image inversion: light on the left side in front of the focus is found on the right side behind focus (Fig. 11b), θ=45°, φ=0°, 45°).

Due to this asymmetries the resolution limit is not a constant but depends on the relative distance and orientation of two dipole emitters. The shape of the CEF shows a low distortion at the half maximum value. Hence it is reasonable to describe the resolution in terms of the full width at half maximum (FWHM). The CEF of an in-plane emitter has a lateral FWHM of 1.11 λ/n in the direction of the dipole axis, 1.04 λ/n perpendicular to it and a longitudinal FWHM of 1.46 λ/n. For a longitudinally oriented dipole the in-plane FWHM is increased to 1.20λ/n while the longitudinal FWHM is decreased to 1.20 λ/n.

3.3. Confocal imaging

Combining the collection efficiency with a given excitation, according to equation 10, scanning confocal optical images of single fluorescent molecules can be calculated. Fig. 12 shows results for linear, a) and b), and radial, c) and d), polarization. The molecule is either scanned in the x,y-plane, as in a) and c) or in the x, z-plane, as in b) and d). In the case of linear polarized excitation along the x axis every orientation of the dipole produces a different confocal image, by which its orientation can be identified. However the intensities vary by up to about two orders of magnitude which makes it very difficult to detect all orientations of emitters simultaneously at a reasonable signal to noise level. The depth resolution of the mirror setup is good. The full width at half maximum is 0.80 λ/n while the lateral FWHM of a dipole lying along the polarization direction is 0.43 λ/n and 0.35 λ/n, respectively.

The intensity patterns in the images of an emitter which is radially excited allow the determination of its orientation. But the intensities of the lateral and longitudinal fields differ also by one order of magnitude. By reducing the illumination aperture of the mirror this ratio can be reduced to obtain comparable intensities. Fig. 12c) and d) show results calculated with the illumination aperture reduced to NA_m =1.21 and the ratio between entrance pupil and beam waist of the Bessel-Gauss beam kept constant. The ratio of the in-plane polarized intensity to the longitudinal one is 0.44 in this case. By reducing the numerical aperture a slight reduction of the resolution must be accepted but in return the orientation of the dipole emitter can be easily determined from the intensity pattern from an image. In order to determine the polar angle θ of the dipole the patterns must be analyzed quantitatively.

Fig. 12. Confocal images, |E|², of a dipole emitter for different orientations. a) and b) linear excitation polarization along the x-axis, c) and d) radial polarization. The illuminated aperture is reduced to 1.21 for radial polarization. In a) and c) the dipole is scanned in the x, y-plane, whereas in b) and d) it is scanned in the x, z-plane. The dipole orientation angles as well as the axis are shown in the lower left images. A logarithmic scale is used with a factor of 2 between adjacent contour levels. All images in a) and b) as well as in c) and d) are scaled identically with the maximum given by the highest intensity which is reached for a) and b) by the dipole oriented parallel to the polarization direction and for c) and d) by the dipole oriented along the optical axis z. The scale bar length is one wavelength λ/n.

Download Full Size | PDF

4. Conclusion

It has been shown, that a parabolic mirror has a good depth and lateral resolution due to the high numerical aperture which can be easily reached. However, for generating a clean focus, the alignment of the illumination beam is critical due to strong coma.

The use of a concave parabolic mirror with high NA could also have advantages in fempto second spectroscopy due to the absence of chromatic aberration.

By using radially polarized light strong and well localized longitudinal components of the electric field of the light can be produced with only weak disturbing lateral fields. This feature might be very useful for some Near-field Scanning Optical Microscopy (NSOM) applications, where a sharp tip is placed in a laser focus for further confining the optical field. If the polarization of the laser is along the tip axis a very strong field enhancement occurs at the tip end which allows imaging of the sample with super resolution [7]. For this application a concave mirror must be used to place a tip in the focus.

The longitudinal component in combination with linear polarized light along two perpendicular directions yields a platform for investigation of nanometer sized particles or single molecules with optical fields in all three spatial dimensions.

The use of a parabolic mirror for single molecule detection is mainly suitable because of its high numerical aperture which allows the detection of arbitrary oriented dipole emitters, such as single molecules, with a constant high detection efficiency. For determining the orientation of a dipole emitter the use of radially polarized excitation with an optimized illumination NA is best suited.

5. Appendix

Calculation of the diffraction integrals

For calculating the electrical field on the focal sphere f_m the rules of ray optics are used. According to Fig. 1 the unit vector pointing in the direction of propagation s, the p-polarized and s-polarized electrical field components after reflection, E _p and E _s , and the vector r _P pointing from the geometrical focus of the mirror to the point of observation can be written as:

s = - (\sin θ_{m} \cos φ, \sin θ_{m} \sin φ, \cos θ_{m}),

in Cartesian coordinates, with the field components before reflection being denoted as E _0,p and E _0,s. In addition the power density of the incident field increases towards the focal region. This factor can be calculated using the fact, that the energy flowing through a differential surface perpendicular to a ray must be constant on its path to the focus. A derivation of this so called apodization factor is given in [16]. In the case of a parabolic mirror it is 2/(1+cosθ).

For linear polarized light we have E _0,p=cosφ l₀ (θ) and E _0,s=-sinφ l₀ (θ) with the field strength of the incident beam l₀ (θ) we find equations 2 and 3. For radially polarized light the incident beam has no s-polarized components. The resulting field on the focal sphere f_m is (neglecting a constant phase factor due to the path length)

E_{f} = - \frac{2 l_{0} (θ)}{1 + \cos θ_{m}} {\begin{matrix} \cos θ_{m} \cos φ \\ \cos θ_{m} \sin φ \\ - \sin θ_{m} \end{matrix}

The evaluation of the scalar product s·r _p in the exponential term of equation 1 yields

\exp [i k s \cdot r_{p}] = \exp [- i k \cos θ \cos θ_{P}] \cdot \exp [- i k \sin θ \sin θ_{P} \cos (φ - φ_{P})]

With the use of the following identities [9] the electric field in the focal region yields the expressions given in equations 4 and 5.

\int_{0}^{2 π} \cos n φ \exp (i ρ \cos (φ - ϕ)) d φ = 2 π i^{n} J_{n} (ρ) \cos n ϕ,

Acknowledgments

We are much obliged to Alexander F. Nowikow from the St.Petersburg State Technical University, St. Petersburg, Russia for sending us copies of the Ignatovsky paper.

References and links

1. W. P. Ambrose, T. Basché, and W. E. Moerner, “Detection and spectroscopy of single pentacene molecules in a p-terphenyl crystal by means of fluorescence excitation,” J. Chem. Phys. 95, 7150–7163 (1991). [CrossRef]

2. L. Fleury, P. Tamarat, B. Lounis, J. Bernard, and M. Orrit. “Fluorescence spectra of single pentacene molecules in p-terphenyl at 1.7 K,” Chem. Phys. Lett. 236, 87–95 (1995). [CrossRef]

3. H. van der Meer, J. A. J. M. Disselhorst, J. Koehler, A. C. J. Brouwer, E. J. J. Groenen, and J. Schmidt, “An insert for single-molecule magnetic-resonance spectroscopy in an external magnetic field,” Rev. Sci. Instrum. 66, 4853–4856 (1995). [CrossRef]

4. Y. Durand, J. C. Woehl, B. Viellerobe, W. Göhde, and M. Orrit, “New design of a cryostat-mounted scanning near-field optical microscope for single molecule spectroscopy,” Rev. Sci. Instrum. 70, 1318–1325 (1999). [CrossRef]

5. J. Enderlein, T. Ruckstuhl, and S. Seeger, “Highly efficient optical detection of surface-generated fluorescence,” Appl. Optics 38, 724–732 (1999). [CrossRef]

6. K. S. Youngworth and T. G. Brown, “Focusing of high numerical aperture cylindrical-vector beams,” Opt. Express 7, 77–87 (2000), http://www.opticsexpress.org/tocv7n2.htm. [CrossRef] [PubMed]

7. L. Novotny, E. J. Sánchez, and X. S. Xie, “Near-field optical imaging using metal tips illuminated by higher-order Hermite-Gaussian beams,” Ultramicroscopy 71, 21–29 (1998). [CrossRef]

8. E. J. Sánchez, L. Novotny, and X. S. Xie, “Near-field fluorescence microscopy based on two-photon excitation with metal tips,” Phys. Rev. Lett. 82, 4014–4017 (1999). [CrossRef]

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

10. V. S. Ignatovsky, “Diffraction by a parabolic mirror having arbitrary opening,” Trans. Opt. Inst. Petrograd 1, paper 5 (1920), in Russian.

11. C. J. R. Sheppard, A. Choudhury, and J. Gannaway, “Electromagnetic field near the focus of wide-angular lens and mirror systems,” IEE J. Microw. Opt. Acoust. 1, 129–132 (1977). [CrossRef]

12. R. Barakat, “Diffracted electromagnetic fields in the neighborhood of the focus of a paraboloidal mirror having a central obscuration,” Appl. Opt. 26, 3790–3795 (1987). [CrossRef] [PubMed]

13. P. Varga and P. Török, “Focusing of electromagnetic waves by paraboloid mirrors. I. Theory,” J. Opt. Soc. Am. A 17, 2081–2089 (2000). [CrossRef]

14. P. Varga and P. Török, “Focusing of electromagnetic waves by paraboloid mirrors. II. Numerical results,” J. Opt. Soc. Am. A 17, 2090–2095 (2000). [CrossRef]

15. E. Wolf, “Electromagnetic diffraction in optical systems. I. An integral representation of the image field,” Proc. Roy. Soc. A 253, 349–357 (1959). [CrossRef]

16. J. J. Stamnes, Waves in focal regions, (Hilger, Bristol, UK, 1986), Sec. 16.1.2.

17. R. H. Jordan and D. G. Hall, “Free-space azimuthal paraxial wave-equation: the azimuthal Bessel-Gauss beam solution,” Opt. Lett. 19, 427–429 (1994). [CrossRef] [PubMed]

18. D. G. Hall, “Vector-beam solutions of Maxwell’s wave equation,” Opt. Lett. 21, 9–11 (1996). [CrossRef] [PubMed]

19. P. L. Greene and D. G. Hall, “Diffraction characteristics of the azimuthal Bessel-Gauss beam,” J. Opt. Soc. Am. A 13, 962–966 (1996). [CrossRef]

20. P. L. Greene and D. G. Hall, “Properties and diffraction of vector Bessel-Gauss beams,” J. Opt. Soc. Am. A 15, 3020–3027 (1998). [CrossRef]

21. S. Quabis, R. Dorn, M. Eberler, O. Glöckl, and G. Leuchs, “Focusing light to a tighter spot,” Opt. Commun. 179, 1–7 (2000). [CrossRef]

22. J. J. Wynne, “Generation of the rotationally symmetric TE₀₁ and T_M01 modes from a wavelength-tunable laser,” IEEE J. Quant. Elec. QE-10, 125–127 (1974). [CrossRef]

23. R. Yamaguchi, T. Nose, and S. Sato, “Liquid crystal polarizers with axially symmetrical properties,” Jpn. J. Appl. Phys. Pt. 1 28, 1730–1731 (1989). [CrossRef]

24. S. C. Tidwell, D. H. Ford, and W. D. Kimura, “Generating radially polarized beams interferometrically,” Appl. Opt. 29, 2234–2239 (1990). [CrossRef] [PubMed]

25. E. G. Churin, J. Hoßfeld, and T. Tschudi, “Polarization configurations with singular point formed by computer generated holograms,” Opt. Commun. 99, 13–17 (1993) [CrossRef]

26. S. C. Tidwell, G. H. Kim, and W. D. Kimura, “Efficient radially polarized laser-beam generation with a double interferometer,” Appl. Opt. 32, 5222–5229 (1993). [CrossRef] [PubMed]

27. M. Stalder and M. Schadt, “Linearly polarized light with axial symmetry generated by liquid-crystal polarization converters,” Opt. Lett. 21, 1948–1950 (1996). [CrossRef] [PubMed]

28. K. S. Youngworth and T. G. Brown, “Inhomogenous polarization in scanning optical microscopy,” in Three-Dimensional and Multidimensional Microscopy: Image Acquisition Processing VII, J.-A. Conchello, C. J. Cogswell, and T. Wilson, editors, Proc. SPIE3919, 75–85 (2000).

29. M. Schrader and S. W. Hell, “Wavefronts in the focus of a light microscope,” J. Microsc. 184, 143–148 (1996).

30. C. J. R. Sheppard, “Aberrations in high aperture optical systems,” Optik 105, 29–33 (1997).

31. J. Sepiol, J. Jasny, J. Keller, and U. P. Wild, “Single molecules observed by immersion mirror objective. The orientation of terrylene molecules via the direction of its transition dipole moment,” Chem. Phys. Lett. 273, 444–448 (1997). [CrossRef]

32. R. M. Dickson, D. J. Norris, and W. E. Moener, “Simultaneous imaging of individual molecules aligned both parallel and perpendicular to the optic axis,” Phys. Rev. Lett. 81, 5322–5325 (1998). [CrossRef]

33. T. Ha, T. A. Laurence, D. S. Chemla, and S. Weiss, “Polarization spectroscopy of single fluorescent molecules,” J. Phys. Chem. B 103, 6839–6850 (1999). [CrossRef]

34. B. Sick, B. Hecht, and L. Novotny, “Orientational imaging of single molecules by annular illumination,” Phys. Rev. Lett. 85, 4482–4485 (2000). [CrossRef] [PubMed]

35. C. J. R. Sheppard and M. Gu, “Imaging by a high aperture optical system,” J. Mod. Opt. 40, 1631–1651 (1993). [CrossRef]

36. C. J. R. Sheppard and P. Török, “An electromagnetic theory of imaging in fluorescence microscopy, and imaging in polarization fluorescence microscopy,” Bioimaging 5, 205–218 (1997). [CrossRef]

37. J. Enderlein, “Theoretical study of detection of a dipole emitter through an objective with high numerical aperture,” Opt. Lett. 25, 634–636 (2000). [CrossRef]

38. J. D. Jackson, Classical Electrodynamics, second edition (Wiley & Sons, New York, USA, 1975), Chap. 9.2.

39. T. Plakhotnik, E. A. Donley, and U. P. Wild, “Single-molecule spectroscopy,” Annu. Rev. Phys. Chem. 48, 181–212 (1997). [CrossRef] [PubMed]

A high numerical aperture parabolic mirror as imaging device for confocal microscopy

Abstract

1. Introduction

2. Focusing

2.1. Theory

2.2. Numerical field calculations

2.2.1. Linear polarization

2.2.2. Radial polarization

2.3. Alignment and coma

3. Single molecule imaging

3.1. Fundamentals

3.2. Image distortion and collection efficiency

3.3. Confocal imaging

4. Conclusion

5. Appendix

Acknowledgments

References and links

Supplementary Material (3)

Cited By

Figures (12)

Equations (32)

Optics Express