**2. Experimental setup for Raman spectroscopy in nanowires**

to free‐standing III–V NWs are explained, since this is the most widely used technique to grow high‐quality nanowires. Each gold droplet represents the nucleation site of a NW, so there are approximately as many NWs as the droplets are. The gold droplets can be directly deposited on the substrate or result from the annealing of an Au thin film. To achieve a perfect control on the NW position, an array of gold particles can be even prepared by using lithography

Thanks to the high degree of control reached on the NW growth process, nowadays most of NW properties can be finely tuned, to such an extent that the creation of NWs tailor fit to specific applications is close to be achieved. Due to the several technological applications enabled by NWs, the interest of the scientific community on them is rapidly growing, as testified by the exponentially growing number of papers published on NWs in the last two decades. The field where the peculiar shape and dimensions of NWs have revealed to have great potential in enabling new functions and/or simply enhancing performances of existing devices is very broad, as it includes electronics, photonics, biosensing, energy conversion, and storage [7]. Therefore, we will pick few examples taken from such a huge variety. For instance, the capability to controllably dope NWs is routinely exploited in field effect transistors [8] while the low mass peculiar to NWs renders them ideal to be used in cantilever force sensors [9] and the NW flexibility makes them easily integrable in devices like flexible displays and artificial skin [10]. Moreover, the NW diameter, smaller than the light wavelength in the visible range, allows NWs to confine electromagnetic waves in the radial direction while guiding light in the axial direction, a property that has been exploited to build NW‐based antennas, lasers, and light‐emitting diodes (LEDs). An example of a multicolor NW‐based LED is shown in **Figure 3(a)** [11]. In the field of energy conversion, we mention that the typically low thermal

**Figure 3.** Two very recent examples of the technological power of NWs. (a) Schematic of monolithically integrated multicolor single InGaN/GaN dot‐in‐nanowire LED pixels on a single chip. Light emission wavelength is tuned across the visible spectrum by varying the nanowire diameter. Reproduced with permission from [11]. Copyright 2016, Amer‐ ican Chemical Society. (b) SEM image showing the deformation of an array of InP NWs in direct contact with the body of a phytopathogen Xylella fastidiosa cell (colorized in green). The ordered NW array allows to evaluate single cell adhesion forces and to explore their dependence on organochemical surface compositions. Reproduced with permis‐

sion from Ref. [14]. © 2016, American Chemical Society.

techniques.

84 Raman Spectroscopy and Applications

Whenever an electromagnetic radiation heats a sample, the light interacts with the sample. It may be reflected, absorbed or scattered. Raman spectroscopy relies on inelastic scattering of the radiation by the sample. It is a versatile and nondestructive technique based on the interaction between the radiation and the vibrational and/or rotational motions of the ions and it provides information such as crystal symmetry, composition, strain, lattice dynamics, and electronic band structure. In a Raman experiment, a monochromatic light is usually sent on the sample and the scattered light is collected and analyzed. When the frequency of the scattered radiation is analyzed, there will be not only the incident radiation wavelength (elastic or Rayleigh scattering component) but also radiation scattered at frequencies lower (Stokes) or higher (anti‐Stokes) than the elastic one (inelastic scattering or Raman components). The intensity of the elastic component is much higher than the inelastic ones, thus special tricks are used to detect the weak Raman signal. Furthermore, the Stokes peak has intensity higher than the anti‐Stokes peak, and their intensities depend on the temperature. The inelastic peaks appear at frequencies that differ from the incident one by a quantity called Raman shift, independent of the excitation frequency. The Raman shift is the most significant information in a Raman experiment.

In a typical Raman setup, the excitation energy is provided by a monochromatic laser source emitting in the visible range (488, 514, and 633 nm are the most common wavelengths). The laser light is filtered by a laser band pass filter and its polarization is cleaned from possible depolarized contributions with the use of a polarizer. Then, the beam is guided toward the sample with a set of mirrors and a polarization preserving beam splitter (typically 50:50). If nanostructures are investigated, like in our case, the laser light is focused via a microscope objective. A good objective has a high numerical aperture (∼0.9 for 100× magnification), which results in a diffraction‐limited laser spot able to provide submicrometric resolution. The nanostructure is positioned on an *x*–*y* piezoelectric stage, in such a way that its surface can be automatically scanned with high precision. The scattered Raman signal is collected by the same objective in the so‐called backscattering geometry and focused to the entrance slit of a spec‐ trometer equipped with diffraction gratings (typically with 1800 lines/mm). The signal is dispersed in the spectrometer and then sent to a Si multichannel charge coupled device (CCD) detector. When standard, namely single stage spectrometers are used, a notch filter before the spectrometer is necessary to reject the intense elastic component, and only Raman shift larger than ∼50 cm−1 can be detected. The best spectrometers existing for Raman spectroscopy are triple‐stage spectrometers, where no notch is required, because the first two gratings are used for the rejection of the elastic contribution and of the stray light (in the so‐called "subtractive mode"), while the third grating disperses the radiation components and determines the resolution. To provide an idea, a 1800 lines/mm grating mounted on a spectrometer with 0.5 m focal length gives a resolution of ∼0.7 cm−1. After the measurements, an accurate frequency calibration of the Raman spectra is usually performed exploiting the well‐known energies of the emission lines of a neon lamp.

**Figure 4.** Sketch of a typical backscattering geometry used in polarization‐resolved Raman experiments on a single ZB nanowire whose growth axis is in the crystallographic direction [111], aligned with the *z*‐axis of the reference system. (a) The incident light wavevector *ki* is parallel to the *x*‐axis and its polarization vector *εi* varies, forming an angle *θ* with *z*. (b) The scattered light wavevector *ks* is along ‐*x* and the components of its polarization vector *εs* either parallel or perpendicular to *z* are selected. (c) Four common scattering configurations are indicated in the Porto notation: *ki* (*εi* , *εs*)*ks*.

In **Figure 4**, a typical geometry for performing polarization‐resolved Raman measurements on a single NW is sketched. Let us consider a NW with zinc blende (ZB) phase, grown along the [111] direction and having a hexagonal cross section, with facets of the {110} family. After transferring the wire on a substrate, the flat facet of the family {110} is perpendicular to the incident light wavevector (*ki* ) and the NW long symmetry axis is aligned with the *z*‐axis of our reference system, as schematized in (a). In the used backscattering geometry, *ki* is parallel to the *x*‐axis and the scattered light wavevector (*ks*) is opposite to it, see panel (b). As a conse‐ quence, all light polarization vectors lie in the *yz* plane. During the measurements, the polarization of the excitation (*εi* ) is rotated by an angle *θ* with respect to the nanowire growth axis. The scattered radiation is analyzed by selecting the component of the polarization either parallel to the NW growth direction (*εs,//*) or perpendicular to it (*εs,*┴). In the following, we will indicate the scattering configuration in the so‐called Porto notation, *ki* (*εi* ,*εs*)*ks*, where the outer terms, from left to right, refer to the directions of exciting and scattered light, respectively, and the inner bracket refers to the excitation and detection polarizations. Some examples of the most common scattering configurations as indicated in Porto notation are given in panel (c). From the experimental point of view, *εi* of the incoming light is rotated by an angle *θ* with respect to the *z*‐axis by using a lambda half plate whose fast axis forms an angle *θ*/2 with the *z*‐axis, while a polarizer before the spectrometer is used to select the scattered radiation with components of the polarization either parallel or perpendicular to the *z*‐axis. Since the efficiency of a spectrometer depends on the polarization of the entering light, a lambda half plate at the entrance of the spectrometer is used to flip the polarization of the light into the most efficient direction.

dispersed in the spectrometer and then sent to a Si multichannel charge coupled device (CCD) detector. When standard, namely single stage spectrometers are used, a notch filter before the spectrometer is necessary to reject the intense elastic component, and only Raman shift larger than ∼50 cm−1 can be detected. The best spectrometers existing for Raman spectroscopy are triple‐stage spectrometers, where no notch is required, because the first two gratings are used for the rejection of the elastic contribution and of the stray light (in the so‐called "subtractive mode"), while the third grating disperses the radiation components and determines the resolution. To provide an idea, a 1800 lines/mm grating mounted on a spectrometer with 0.5 m focal length gives a resolution of ∼0.7 cm−1. After the measurements, an accurate frequency calibration of the Raman spectra is usually performed exploiting the well‐known energies of

**Figure 4.** Sketch of a typical backscattering geometry used in polarization‐resolved Raman experiments on a single ZB nanowire whose growth axis is in the crystallographic direction [111], aligned with the *z*‐axis of the reference system.

*z*. (b) The scattered light wavevector *ks* is along ‐*x* and the components of its polarization vector *εs* either parallel or perpendicular to *z* are selected. (c) Four common scattering configurations are indicated in the Porto notation: *ki*

In **Figure 4**, a typical geometry for performing polarization‐resolved Raman measurements on a single NW is sketched. Let us consider a NW with zinc blende (ZB) phase, grown along the [111] direction and having a hexagonal cross section, with facets of the {110} family. After transferring the wire on a substrate, the flat facet of the family {110} is perpendicular to the

the *x*‐axis and the scattered light wavevector (*ks*) is opposite to it, see panel (b). As a conse‐ quence, all light polarization vectors lie in the *yz* plane. During the measurements, the

axis. The scattered radiation is analyzed by selecting the component of the polarization either parallel to the NW growth direction (*εs,//*) or perpendicular to it (*εs,*┴). In the following, we will

terms, from left to right, refer to the directions of exciting and scattered light, respectively, and the inner bracket refers to the excitation and detection polarizations. Some examples of the most common scattering configurations as indicated in Porto notation are given in panel (c).

reference system, as schematized in (a). In the used backscattering geometry, *ki*

indicate the scattering configuration in the so‐called Porto notation, *ki*

) and the NW long symmetry axis is aligned with the *z*‐axis of our

) is rotated by an angle *θ* with respect to the nanowire growth

(*εi*

is parallel to the *x*‐axis and its polarization vector *εi*

varies, forming an angle *θ* with

(*εi* ,

is parallel to

,*εs*)*ks*, where the outer

the emission lines of a neon lamp.

86 Raman Spectroscopy and Applications

(a) The incident light wavevector *ki*

incident light wavevector (*ki*

polarization of the excitation (*εi*

*εs*)*ks*.

By measuring the dependence of the scattered intensity on the incident and scattered polari‐ zation directions one can deduce the symmetry of the Raman tensors, thus the symmetry of the corresponding phonon, based on the comparison with the theoretical calculation of the intensity of the Raman signal for that specific experimental geometry. The intensity of the scattered light s into a solid angle dΩ is given by [15],

$$\mathbf{I}\_{\mathbf{s}} = \mathbf{I}\_{\mathbf{i}} \cdot \mathbf{k} \cdot |\boldsymbol{\varepsilon}\_{\mathbf{i}} \cdot \mathbf{R} \cdot \boldsymbol{\varepsilon}\_{\mathbf{s}}|^{2} \text{ d}\Omega \tag{1}$$

with i the irradiance of the radiation incident on the sample, = 2*πa*<sup>2</sup> *ω*s (*a* = 1/137 and *ω*<sup>s</sup> frequency of the scattered light), and *R* the Raman scattering tensor, which is defined as = d d0 ∙ , where is the crystal polarizability, and the vector displacement of an atom

induced by a phonon. It is usually convenient to transform the Raman tensors in a basis made by the three main crystallographic axes of the sample and to express the polarization vectors in that basis. From Eq. (1) one can calculate the selection rules for all modes in specific scattering geometries.

Once the basic principles of Raman spectroscopy in NWs and measurement setup have been summarized, the most significant results of Raman spectroscopy applied to NWs can be described.
