Spectral Fringes in Nanostructures

Assessing Pulse Propagation in Nanowires Using TIR Micro-Spectroscopy

  • Fig. 1: Scheme of the modified TIR micro- spectrometer. Fig. 1: Scheme of the modified TIR micro- spectrometer.
  • Fig. 1: Scheme of the modified TIR micro- spectrometer.
  • Fig. 2 (a-i): Image of a 5 μm NW using conventional illumination; (a-ii) The same NW using evanescent wave illumination; (b-i) Spectral image of the NW; (b-ii) Spectra after shift correction.
  • Fig. 3 (a): Spectrum of the overlapping spots on the slit plane; (b) Reconstruction of the pulse from the spectra; (c): Cosine variation associated with the peak of the reconstructed pulse.

Spectral fringes naturally arise when micro-spectrometers observe nanowires propagating broad-band light by plasmon coupling and uncupling processes. In this work we propose an extension of the technique to assess pulse propagation in more general structures. As an example, we show the registration of spectral fringes in long multimetal nanowires exited and observed using a total internal reflection microscope coupled with a spectrometer. Recent years have seen an increased interest in nanowires (NWs) based on the ability of these nanoparticles to confine and conduct light with sub-diffraction resolution [1].

The spectroscopic analysis of the light emitted by the NWs can provide information of the surrounding medium and how a light pulse changes when propagates in the NW as a plasmonic wave; the concepts of low-coherence interferometry –the interference of wide spectrum light with itself– can be applied here [2].
The evanescent wave generated in the total internal reflection (TIR) process in a prism is a typical source of choice for plasmon excitation in metallic particles since the generated photons have enough momentum to couple with plasmons. However, the use of TIR in a high NA immersion objective instead of a prism is a convenient way to simultaneously observe the nanostructures since the excitation light does not enter the collection part of the microscope not hampering the observation.

In this work we present a modified microscope that uses TIR microscopy and a spectrometer to carry out low-coherence interferometry measurements. The system is demonstrated using the light entering and exiting a 5 μm metallic NW to measure the delay of a pulse.

Material & Methods

Figure 1 shows the experimental system. It uses an inverted microscope (Nikon Eclipse) as platform equipped with a 60X, 1.4 NA oil immersion objective. A super-continuum white laser (NKT, SuperK) beam is conditioned to obtain a thightly focused beam on the back focal plane of the objective after reflection on a beam splitter cube. To surpass the critical angle -needed to generate an evanescent wave on top of the cover slip- the laser beam axis is displaced respecting the microscope optical axis.

A second removable microscope objective is placed on an axial micrometric translation stage after the tube lens. The slit of a gratin spectrometer (PI, Acton 2500i) is placed at the image plane of the microscope camera port or moved after the second objective as in the configuration depicted in figure 1. This image plane is transported without augments through the spectrometer to the CCD (Andor, Ixon). The spectrometer allows the rotation of the diffraction grating (150 g/mm). When the zero-order is selected the grating acts as a mirror and the spectrometer behaves as a 4-f optical system. The sample consists on multimetal NWs (DropSens, Pt-Au-Ni) dried on top of the cover slip.


In the first part of the experiment the second objective is not present in the system and the grating tilted to align the zero order and the slit full open. This configuration allows the acquisition of images using two illuminations: Figure 2 (a-i) shows the image of a NW using the conventional built-in halogen lamp of the microscope and figure 2(a-ii) the image using the evanescent wave illumination when the laser is on and the lamp is switched off. The fabrication process of the NWs produces particles with sub-micrometric length precision allowing the calibration of the pixel size. The second image shows two spots: one corresponding to the scattering at the entrance end of the NW and the other to the far end of the NW where the uncoupling of the plasmons with propagating waves occurs. Given the limited extension of the spots, the spectrum of each source can be acquired simultaneously without closing the slit.

The spectral span of the emission is much higher that the spectral range that can be acquired in a single CCD exposure. To cover the full spectrum, several acquisitions are needed for consecutive angles of the grating. The images are post-processed to stich them together to obtain the full spectrum shown in figure 2 (b-i).

Figure 2 (b-ii), shows the spectra of the two sources in the same plot (the solid line corresponding to the upper spectrum and the dotted line to the lower) obtained averaging two rows around the vertical position of the wavelength spread spots. Given the different horizontal locations of the spots with respect to the slit center, to place them correctly in the wavelength scale a red/blue shift of the spectra is needed.

Assuming that one spectrum corresponds to the entrance and the other to the pulse exiting the NW, one can try to measure the time that spends traveling. If a light pulse interferes with a delayed copy of itself, the spectrum of the combined light shows an oscillatory modulation which frequency proportional to the delay time [3]. The processing of theses fringes –called spectral or channel fringes– is the basis of several modalities of low-coherence (or broadband) interferometry.

In the particular case of NWs, after some controversy, the spectral fringes phenomena explain the fringes observed in the spectra of NWs that were applied to determine the group velocities in 1 μm gold and silver NWs [2]. In this cited work, the excitation of plasmons is carried out using evanescent waves generated by a prism and the spectral analysis using a micro-spectrometer. Given that their microscope augments are low and the NWs short, the light from both ends overlaps on the slit causing the spectral fringes.

In our system, this superposition is not possible since the NWs are much longer and the TIR illumination forces us to use a high NA objective producing higher augments which separates the spectra (fig. 2 (b-i)) making impossible to directly register spectral fringes. To solve this problem, the second microscope objective (Nikon, PF 20X) -mounted in a linear translation stage to control the defocus- is used to compress the NW image on the slit plane. Previously, using the zero diffraction order and the microscope x-y sample translation stage, the NW was carefully positioned to ensure that the image of the NW´s ends where placed equidistant from the slit center when closed. After that, to register the interference spectra the grating is tilted lo align the first diffraction order with the camera axis and the slit closed to the minimum aperture. Figure 3 (a) shows the spectrum in the interval between 550 and 630 nm. The reason to choose this interval is because the spectra of the light emitted from the two ends of the NW is similar and, compared with other intervals, higher contrast is found. To obtain the time delay of a pulse with this spectral composition, the spectrum is processed by firstly averaging the central rows of the spectrogram where the overlapping of light for the two defocused spots occurs. To pass from the wavelength domain to the time domain, the Fourier transform is computed. The result is shown in figure 3(b). The peak at 150 fsec can be identified as the center of the delayed pulse. Figure 3(a) shows for comparison the cosine variation with wavelength corresponding to this peak. From this value and the NW length, the median group velocity in the spectral range can be estimated as a ten percent of the light speed.


We have described a system to measure pulse delay in long metallic nanowires using a TIR micro-spectrometer. The system can be generalized to measure the interference between pairs of emission sources in a nanostructure. This could be accomplished by using an additional demagnifying 4-f system to relay the image plane at the camera port to the image plane at the interferometer slit. In this way, by selectively masking the image at the camera port, the light that can pass to the spectrometer can be controlled to form spectral fringes probing different pairs of sources.


Ignacio Iglesias

This research was supported by the Spanish MINECO grant ENE2016-7982-C5-4-R. We thank Lisa Almonte for preparing the samples.

Dr. Ignacio Iglesias

Universidad de Murcia,
Dpto. Fisica
Campus de Espinardo
Murcia, Spain


More information on nanowires

More about Total Internal Reflection Fluorescence Microscopy

[1]  H. Xu, ed., Nanophotonics: Manipulating Light with Plasmons (Pan Stanford, 2018).
[2]  I. Iglesias, H. S. Chen, K. D. Mills, D. S. Dilworth, and E. N. Leith, "Electronic channel fringe holography for depth and delay measurements," Appl. Opt., AO 38, 2196–2203 (1999) DOI 10.1364/AO.38.002196
[3]  B. Wild, L. Cao, Y. Sun, B. P. Khanal, E. R. Zubarev, S. K. Gray, N. F. Scherer,  and M. Pelton, "Propagation Lengths and Group Velocities of Plasmons in Chemically Synthesized Gold and Silver Nanowires," ACS Nano 6, 472–482 (2012) DOI 10.1021/nn203802e


Universidad de Murcia
Departomento de Fisica (CIOyN)
30100 Murcia

Register now!

The latest information directly via newsletter.

To prevent automated spam submissions leave this field empty.