Photobleaching profile of Raman peaks and fluorescence background
Laser-induced fluorescence is the most common source of interfering baseline signal encountered in Raman measurements. It shows itself as a slowly changing background in a spectrum. One of the challenges in the successful use of Raman spectroscopy is to extract Raman signatures from this, orders of magnitude stronger, broadband fluorescence emission. Irradiating the sample with intense laser light, ie, photobleaching, is one effective technique to reduce the level of fluorescence emission, thus increasing the signal to noise (S/N) ratio of a spectrum.
The fluorescence interference in Raman spectroscopy may result from the compound analysed or from fluorescent impurities in the sample. It is an absorption process that causes molecules to be excited to a higher electronic state, which requires high-energy photons. Fluorescence light is then emitted while molecules relax back to the lower energy level. The phenomenon depends strongly on the excitation wavelength and appears only at a fixed frequency, while the Raman shifts are independent of the laser’s wavelength.
Raman scattering and fluorescence emission may compete with each other when the excitation laser energy is close to the electronic transition energy of the material. Higher energy green or red excitation sources, such as 514nm or 633nm visible laser, produce stronger fluorescence background. As the laser line moves to the near-infrared (NIR) region, as with 785nm or 1,064nm laser, the fluorescence effect subsides or completely disappears since the energy of these wavelengths may not be sufficient to excite the molecule to higher electronic state or may not be enough to destroy fluorescing molecules in the material. This effect is illustrated in Figure 1.
Figure 1 shows spectra of an inactive pharmaceutical ingredient, hydroxypropyl methylcellulose (HPMC), a common pharmaceutical excipient, which is measured at three different laser lines; 514nm, 633nm and 785nm. The figure illustrates the loss of Raman peaks in the presence of increasing levels of fluorescence with high energy excitation sources. In contrast to the 514nm laser excitation line that produced significant fluorescence baseline, a better S/N ratio HPMC spectrum is acquired with 785nm excitation wavelength. Advances in compressive detection strategy have recently been made to facilitate Raman classification and quantitation in the presence of fluorescence background.1 This is reported to be a better alternative to conventional subtraction strategies2-4 by virtue of its compatibility with automated high-speed chemical analysis in the presence of fluorescence background.1
It is well known that Raman scattering is an inefficient phenomenon. Typically, one Raman photon is generated for every 106 to 109 laser photons incident upon the sample. As a result, very low amounts of fluorescent species in the sample may be enough to mask the low Raman scattered photons or make it difficult to interpret the spectrum. If the fluorescence baseline is high, the shot noise generated by this signal may be of a similar order, or even greater than the Raman signal alone and will completely mask the signal from the Raman photons. Consequently, proposed mathematical techniques2-4 to eliminate the fluorescence background will only make larger peaks more visible against the background while smaller peaks may still remain undiscernible against the noise.
In many cases, Raman signal can still be acquired from samples producing autofluorescence. In the fluorescence process, the sample is excited to the higher electronic state by the absorption of a photon and subsequently relaxes back to the ground electronic state by emitting a fluorescence photon. When compared to the rate of fluorescence emission, which typically occurs in…