Background-Free Stimulated Raman Spectroscopy and Microscopy

We propose a three-color, double-modulation scheme for the background-free detection of stimulated Raman scattering (SRS). We call the scheme stimulated Raman gain and opposite loss detection (SRGOLD). It exploits the symmetric nature of potential parasitic signals (cross phase modulation, two-photon absorption, and thermal effects) to the end of suppressing them. Conversely, the antisymmetric nature of SRS provides for a twofold increase in the magnitude of the SRS signal. We experimentally demonstrate SRGOLD spectroscopy and microscopy on test samples as well as on mice skin samples.

Figure 1

SRGOLD principle and practical implementation. [(a)–(d)] Sketch of incident pumps (${\omega }_0+\mathrm{\Omega }$, red, or ${\omega }_0-\mathrm{\Omega }$, blue) and probe (${\omega }_0$, orange) and modulation transfer to the probe in (a) SRG scheme, (b) SRL scheme, (c) stimulated Raman opposite loss ($\mathrm{SRL}\text{−}180°$) scheme, and (d) SRGOLD scheme. Comparison of (a) to (c): The modulations induced on the probe due to SRS are in phase while the artifacts induced by XPM, TPA, and the thermal effect are in opposite phase. Conclusion on (d): In the SRGOLD scheme, the SRG and SRL signals add up while the artifacts cancel out. (e) Practical SRGOLD implementation.

Figure 2

SRGOLD spectroscopy and microscopy. (a) Stimulated Raman spectra of chlorobenzene as recorded in stimulated Raman gain (SRG, blue circles), stimulated Raman loss ($\mathrm{SRL}\text{−}180°$, red triangles), and SRGOLD (black squares) schemes. The spectral offsets seen in SRG and $\mathrm{SRL}\text{−}180°$ are canceled in SRGOLD over the $160{\mathrm{cm}}^{-1}$ recorded band. (b) $20\mu \mathrm{m}$ polystyrene bead imaging with the SRG, $\mathrm{SRL}\text{−}180°$, and SRGOLD schemes. At C–C molecular resonance ($\mathrm{\Omega }=1003{\mathrm{cm}}^{-1}$) and off resonance ($\mathrm{\Omega }=930{\mathrm{cm}}^{-1}$); the background offsets seen in SRG and $\mathrm{SRL}\text{−}180°$ are absent from the SRGOLD images while the signal is the sum of the SRG and $\mathrm{SRL}\text{−}180°$ ones. The ${\omega }_0$ image is the unmodulated ${\omega }_0$ signal, with the sample present and with ${\omega }_0+\mathrm{\Omega }$ and ${\omega }_0-\mathrm{\Omega }$ blocked. Image $101\times 17\text{pixels}$, $I_{{\omega }_0+\mathrm{\Omega }}=70\mathrm{mW}$, $I_{{\omega }_0}=160\mathrm{mW}$, $I_{{\omega }_0-\mathrm{\Omega }}=30\mathrm{mW}$, pixel dwell time 5 ms.

Figure 3

SRGOLD in scattering test samples. (a) Model sample with a scattering top layer. (b) Image of the back aperture of the collection objective showing how ${\omega }_0$ is clipped in the presence of scattering. (c) SRG, $\mathrm{SRL}\text{−}180°$, and SRGOLD images of $20\mu \mathrm{m}$ polystyrene beads in the presence of scattering layer; the color scale is the same as that in Fig. 2. Image $101\times 17\text{pixels}$, $I_{{\omega }_0+\mathrm{\Omega }}=60\mathrm{mW}$, $I_{{\omega }_0}=50\mathrm{mW}$, $I_{{\omega }_0-\mathrm{\Omega }}=50\mathrm{mW}$, pixel dwell time 10 ms. (d) SRG, $\mathrm{SRL}\text{−}180°$, and SRGOLD images of polystyrene beads for weak Raman bands at $\mathrm{\Omega }=1034$ and $1041{\mathrm{cm}}^{-1}$.

Figure 4

SRGOLD in biological samples. (a) white light, SRG, $\mathrm{SRL}\text{−}180°$, and SRGOLD images of a $20\mu \mathrm{m}$ thick mouse skin sample for the amide II ($\mathrm{\Omega }=1450{\mathrm{cm}}^{-1}$) resonance. SRGOLD artifact balance is performed at point 0. The ${\omega }_0$ image is the ${\omega }_0$ signal with the sample present and with ${\omega }_0+\mathrm{\Omega }$ and ${\omega }_0-\mathrm{\Omega }$ beams blocked. Image $30\times 30\text{pixels}$, $I_{{\omega }_0+\mathrm{\Omega }}=86\mathrm{mW}$, $I_{{\omega }_0}=33\mathrm{mW}$, $I_{{\omega }_0-\mathrm{\Omega }}=40\mathrm{mW}$, pixel dwell time 50 ms. (b) SRG (blue circles), $\mathrm{SRL}\text{−}180°$ (red triangles), and SRGOLD (black squares) spectra as recorded in points 1 and 2. Superimposed is the ${\omega }_0$ signal only (gray circles). SRGOLD proves to cancel artifact over the $1350{\mathrm{cm}}^{-1}<\mathrm{\Omega }<700{\mathrm{cm}}^{-1}$ band.