Apparatus and method for oblique photothermal microscopy

Abstract

An oblique photothermal microscopy system and method includes generating a pulsed infrared excitation light beam for exciting a sample and a probe light beam for illuminating the sample. The infrared excitation light beam and the probe light beam are combined into a combined beam and directing the combined beam along an optical axis of the microscopy system onto an objective element, the objective element generating a sample beam from the combined beam and directing the sample beam along the optical axis toward the sample. A split detector disposed in proximity to the sample includes a plurality of detection elements. The sample beam is directed from the objective element through the space between two of the detection elements and striking the sample, the two detection elements receiving probe light scattered by the sample and generating signals indicative of probe light scattered by the sample.

Description

PATENT Attorney Docket No.264003.0073PCT BU-2024-026 APPARATUS AND METHOD FOR OBLIQUE PHOTOTHERMAL MICROSCOPY CROSS REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U. S. Provisional Application No. 63 / 571,777, filed on March 29, 2024, the entire contents of which are incorporated herein by reference. FIELD OF THE TECHNOLOGY

[0002] The subject disclosure relates generally to microscopy and, particularly, to oblique photothermal microcopy, in particular, oblique infrared photothermal microscopy. BACKGROUND OF THE TECHNOLOGY

[0003] In-vivo infrared (IR) spectroscopy faces challenges due to poor sensitivity in reflection mode and low resolution at micrometer scale. Classic photothermal measurement captures only a small fraction of probe photons through a pinhole to extract the photothermal signal. SUMMARY OF THE TECHNOLOGY

[0004] According to one aspect, an oblique photothermal microscopy system is provided. A source of pulsed infrared light generates a pulsed infrared excitation light beam for exciting a sample, such that the pulsed infrared light beam selectively heats the sample by absorption of the pulsed infrared light. A source of probe light generates a probe light beam for illuminating the sample. A combining element combining the infrared excitation light beam and the probe light beam into a combined beam and directing the combined beam along an optical axis of the microscopy system onto an objective element, the objective element generating a sample beam from the combined beam and directing the sample beam along the optical axis toward the sample. A split detector disposed in proximity to the sample, the split detector comprising a plurality of detection elements with at least one space being disposed between at least two adjacent detection elements of the plurality of detection elements, thesample beam being directed along the optical axis from the objective element through the space between two of the detection elements and striking the sample, the two detection elements receiving probe light scattered by the sample and generating signals indicative of probe light scattered by the sample. A processor processes the signals indicative of probe light scattered by the sample to generate an image of the sample.

[0005] In some exemplary embodiments, the processing of the signals includes generating a sum of the signals.

[0006] In some exemplary embodiments, the processing of the signals includes generating a difference between the signals.

[0007] In some exemplary embodiments, the detection elements are disposed along the optical path of the microscopy system between the objective element and the sample.

[0008] In some exemplary embodiments, the sample is disposed along the optical path of the microscopy system between the objective element and the detection elements.

[0009] In some exemplary embodiments, the source of probe light comprises a laser.

[0010] In some exemplary embodiments, the probe light has a wavelength of 532 nm.

[0011] In some exemplary embodiments, the source of excitation light comprises a laser.

[0012] In some exemplary embodiments, the excitation light has a wavelength in a range of 4.28 µm to 4.29 µm.

[0013] In some exemplary embodiments, the excitation light has a wavelength in a range of 5.4 µm to 12.99 µm.

[0014] In some exemplary embodiments, the excitation light has a wavelength in a range of 266 nm to 2400 nm.

[0015] In some exemplary embodiments, the method further comprises a scan unit disposed along the optical axis of the microscopy system between the combining element and the objective element for scanning the combined beam received from the combining element.

[0016] In some exemplary embodiments, the combining element comprises a dichroic mirror.

[0017] According to another aspect, an oblique photothermal microscopy method is provided. The method includes generating a pulsed infrared excitation light beam for exciting a sample, such that the pulsed infrared light beam selectively heats the sample by absorption of the pulsed infrared light; generating a probe light beam for illuminating the sample; combining the infrared excitation light beam and the probe light beam into a combined beam and directing the combined beam along an optical axis of the microscopy system onto an objective element, the objective element generating a sample beam from the combined beam and directing the sample beam along the optical axis toward the sample; providing a split detector for being disposed in proximity to the sample, the split detector comprising a plurality of detection elements with at least one space being disposed between at least two adjacent detection elements of the plurality of detection elements, the sample beam being directed along the optical axis from the objective element through the space between two of the detection elements and striking the sample, the two detection elements receiving probe light scattered by the sample and generating signals indicative of probe light scattered by the sample; and processing the signals indicative of probe light scattered by the sample to generate an image of the sample.

[0018] In some exemplary embodiments, the processing of the signals includes generating a sum of the signals.

[0019] In some exemplary embodiments, the processing of the signals includes generating a difference between the signals.

[0020] In some exemplary embodiments, the detection elements are disposed along the optical path of the microscopy system between the objective element and the sample.

[0021] In some exemplary embodiments, the sample is disposed along the optical path of the microscopy system between the objective element and the detection elements.

[0022] In some exemplary embodiments, the source of probe light comprises a laser.

[0023] In some exemplary embodiments, the probe light has a wavelength of 532 nm.

[0024] In some exemplary embodiments, the source of excitation light comprises a laser.

[0025] In some exemplary embodiments, the excitation light has a wavelength in a range of 4.28 µm to 4.29 µm.

[0026] In some exemplary embodiments, the excitation light has a wavelength in a range of 5.4 µm to 12.99 µm.

[0027] In some exemplary embodiments, the excitation light has a wavelength in a range of 266 nm to 2400 nm.

[0028] In some exemplary embodiments, the method further comprises a scan unit disposed along the optical axis of the microscopy system between the combining element and the objective element for scanning the combined beam received from the combining element.

[0029] In some exemplary embodiments, the combining element comprises a dichroic mirror. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figs.1(a) through 1(f) illustrate an oblique photothermal microcopy OPTM) technique and simulation results, according to exemplary embodiments.

[0031] Figs.2(a) through 2(e) schematically illustrate OPTM microscope instrumentation and signal processing, according to exemplary embodiments.

[0032] Figs.3(a) through 3(g) illustrate OPTM imaging of microparticles in a scattering medium, according to exemplary embodiments.

[0033] Figs.4(a) through 4(f) illustrate sensitivity improvements over MIP spectroscopy by OPTM of the current disclosure.

[0034] Figs.5(a) through 5(c) illustrate depth-resolved in vivo OPTM imaging of mouse skin, according to exemplary embodiments.

[0035] Figs.6(a) through 6(g) illustrate in vivo investigation of topical drug pathway inside mouse skin.

[0036] Figs.7(a) through 7(i) illustrate in vivo OPTM imaging of human skin.

[0037] Fig.8 is a schematic functional block diagram of a mid-infrared photothermal (MIP) microscope.

[0038] Fig.9 includes images acquired by OPTM at off-resonance of infrared absorption (1770 cm-1).

[0039] Figs.10(a), 10(b), 10(c) illustrate DC images of phase gradient, absorption, and MIP modalities in 3(a) through 3(g). Fig.10(d) illustrates statistical analysis of pixel- wise amplitudes in DC images of MIP and absorption, representing photo currents detected by remote photodiode and split detectors based on photon counts. Fig.10(e) illustrates improvement factor, calculated by dividing the amplitude of absorption DC images by that of MIP DC images, demonstrating scaling improvements of up to 700-fold.

[0040] Fig.11 illustrates in-vivo OPTM imaging of mouse ear skin at a penetration depth of 100 µm.

[0041] Figs.12(a) through 12(e) illustrate mid-infrared photothermal spectrum of benzoyl peroxide. Fig.12(a) is an image of a skin-care product used. Fig.12(b) is a close-up image indicating that the active ingredient is 10% benzoyl peroxide. Fig.12(c) illustrates the molecular structure of drug molecule, benzoyl peroxide. Fig.12(d) illustrates an experimental set-up for measuring infrared photothermal spectrum of samples. Fig.12(e) illustrates normalized infrared photothermal spectrum of benzoyl peroxide gel.

[0042] Fig.13 illustrates control group images without benzoyl peroxide administrated, including phase gradient DC images and PTPG images at different penetration depth in mouse skin.

[0043] Figs.14(a) through 14(d) illustrate depth-resolved phase gradient DC images and PTPG images at four field of views for unveiling drug pathway and quantitative evaluations of drug delivery efficiency.

[0044] Fig.15 illustrates depth-resolved images illustrating endogenous chemicals, including sebaceous glands, and the distribution of topical BPO beneath the skin.

[0045] Figs.16(a) through 16(e) illustrate in vivo OPTM imaging of human skin without benzoyl peroxide administration.

[0046] Fig.17 illustrates sub-video-rate in vivo OPTM imaging of protein in mouse abdominal skin, capturing skin movement during breathing.

[0047] Figs.18(a) through 18(h) illustrate OPTM imaging of a mouse brain slice.

[0048] Figs.19(a) through 19(c) illustrate OPTM imaging of 0.5-µm PMMA particles in a scattering medium at C=O absorption peak 1729 cm-1.

[0049] Figs.20(a) through 20(c) illustrate visualizing bulk features in 10-µm PMMA particles by Hilbert transformed analysis of the phase gradient images.

[0050] Figs.21(a) through 21(c) illustrate exemplary dimensions of a split detector PD1, PD2, according to some exemplary embodiments.

[0051] Figs.22(a) through 22(d) illustrate in vivo imaging of mouse skin.

[0052] Figs.23(a) and 23(b) are schematic functional block diagrams illustrating an oblique photothermal microcopy OPTM) microscope and technique and simulation results, according to some alternative exemplary embodiments. DETAILED DESCRIPTION

[0053] In-vivo infrared (IR) spectroscopy faces challenges due to poor sensitivity in reflection mode and low resolution at micrometer scale. To address these drawbacks, the technology of the current disclosure provides oblique photothermal microscopy (OPTM) including an oblique photothermal microscope to enable ultrasensitive IR spectroscopic imaging of live subjects at sub-micron resolution.

[0054] Classic photothermal measurement captures only a small fraction of probe photons through a pinhole to extract the photothermal signal. In contrast, OPTM of the current disclosure uses a differential split detector placed on, adjacent to, or otherwise inclose proximity to the sample surface to collect about 500-fold more photons and suppress the laser noise by about 12-fold via balanced detection. Leveraging its improved sensitivity, OPTM of the current disclosure enables low-dose IR imaging of skin without photodamage. Depth-resolved in vivo OPTM imaging of metabolic markers beneath mouse and human skin is described herein in detail. Furthermore, in vivo OPTM tracking of topical drug contents within mouse and human skin is described herein in detail. Collectively, OPTM of the current disclosure presents a highly sensitive imaging platform for in vivo and in situ molecular analysis.

[0055] Label-free imaging of chemicals in live animals and human subjects is important for both basic research and clinical translation. Vibrational spectroscopic imaging has emerged as a highly sensitive, label-free platform to visualize molecular contents in living systems, advancing the study of biology and medicine. In particular, spatially-offset Raman spectroscopy and coherent Raman scattering microscopy have been developed to image chemicals in the skin layers of live subjects. Compared to spontaneous Raman scattering, infrared (IR) absorption provides an approximately eight orders of magnitude larger cross-section, making it particularly sensitive to fingerprint vibrations. Since William Coblentz introduced the IR spectrometer in the early 1900s to explore molecular structures, IR spectroscopy has been significantly advanced. Among the various techniques, Fourier- transform infrared (FTIR) spectroscopy is extensively used in the fields of biological research, function materials, and pharmaceuticals. The invention of quantum cascade laser as a room-temperature semiconductor laser has facilitated highly sensitive IR spectroscopic imaging. Traditional IR spectroscopic imaging measures the loss of IR photons, which is not suitable for imaging live animals because IR photons are attenuated and cannot penetrate through or be reflected by an intact animal body. Photoacoustic IR microscopy circumvents the issue of low photon collection efficiency by using the weak-scattering photoacoustic wave as a readout of IR absorption. However, because IR photons attenuate in the acoustic coupling medium, photoacoustic IR microscopy cannot operate in an epi-mode, making it not applicable for universal in vivo imaging scenarios. FTIR and photoacoustic IR also face limitations in spatial resolution due to the diffraction limit of long IR wavelengths, making them inefficient for nanoscale chemical visualization in vivo. Using a near-field probe strategy to overcome the IR diffraction limit, atomic force microscope-infrared (AFM-IR) spectroscopy achieves 10-nm resolution. However, AFM-IR's limited penetration depth makes it unsuitable for volumetric imaging of biological samples.

[0056] Recently-developed mid-infrared photothermal (MIP) microscopy addresses these limitations by using a visible light to probe the mid-infrared-induced thermal effects. Over the past few years, advances have been made towards widefield measurement, video-rate imaging speed, micromolar-level sensitivity in fingerprint and silent windows, and 3D tomographic imaging capability. Despite these advances, in vivo MIP imaging has not yet been feasible due to limited sensitivity in visualizing chemical content in live animals.

[0057] Figs.1(a) through 1(f) illustrate an oblique photothermal microcopy OPTM) technique and simulation results, according to the current disclosure. Fig.1(a) is a schematic functional diagram of a mid-infrared photothermal (MIP) microscope. Fig.1(b) is a schematic functional diagram of a portion of an OPTM microscope 10, according to some exemplary embodiments. Photons propagating to PD1 are shown in the box labeled “Sample”, i.e., the sample being examined. Fig.1(c) schematically illustrates the principle of OPTM according to the disclosure, illustrating the probe propagations at the status of IR- pump-on and IR-pump-off. Fig.1(d) illustrates Monte Carlo simulation results of photon distribution on the surface of a uniform scattering layer. Fig.1(e) is a curve illustrating photon collection efficiency of the split detector and objective of the microscope. Fig.1(f) is a curve illustrating the improvement factor in photon collection efficiency with oblique detection of the current disclosure compared to the classic method. For Figs.1(a) through 1(f), BS: beam splitter. DM: dichroic mirror. PD: photodiode. RL: reflective objective.

[0058] The epi-detection geometry is important for in vivo imaging because photons cannot penetrate through the entire animal body due to strong scattering and absorption. In a classic epi-detected MIP microscope (Fig.1(a)), the pump (IR) and probe beam are coaligned and focused by a reflective objective (RL) onto a sample. The scattered photons from the sample are collected by the same objective RL. An iris is employed before a remote photodetector (PD) to maximize the photothermal signals. However, the probe photons experience enormous scattering events in complex tissues, resulting in scattered photons losing their original propagation directions. Thus, using an objective and an iris to collect scattered photons is ineffective. While classic MIP can visualize chemical content in opaque samples, such as pharmaceutical tablets and thin tissue slices, it is not sensitive for in vivo measurement due to significant loss of highly scattered probe photons during back- propagation to the remote detector (PD).Described herein in detail is an oblique photothermal microscope (OPTM) 10 to enable high-sensitivity in vivo infrared spectroscopic imaging at sub-micron resolution, illustrated schematically in Fig.1(b)). According to the technology of the current disclosure, as illustrated in Fig.1(b), instead of an iris before a remote detector, oblique photothermal detection is achieved by placing a differential split detector (PD1, PD2) above the sample surface to collect epi-propagated probe photons with high efficacy. OPTM microscope 10 includes a source 12 of visible probe light which emits a probe beam and a source 14 of infrared (IR) light which emits a pulsed IR pump beam. In some exemplary embodiments, the pulsed IR pump laser source 14 operates within the wavelength ranges of 4.28–4.29 µm and 5.4–12.99 µm, which can be extended to cover the ultraviolet to near-infrared regions (266 nm to 2400 nm). In some exemplary embodiments, the probe laser source 12 has a fixed wavelength of 532 nm, which can be varied / adjusted to other wavelengths as needed. The probe beam remains spectrally distinct from the IR pump beam. Signals indicative of detections by detection elements PD1, PD1 of the split detector are received by processing system or processor 22. Processing system or processor 22 carries out the processing and control used to effect the OPTM microscopy approach of the current disclosure. Processing system or processor 22 can include processing circuitry and associated elements such as memory, input / output facilities, interface facilities, etc. used to implement the current technology. Processing system or processor 22 receives the signals indicative of detections made by detection elements PD1, PD2 and processes those signals as described herein to generate an image of the sample. Processing system or processor 22 can include, for example, a general purpose computer, a microprocessor system, a microcontroller system, or any processing system used to process data and provide control and input / output and user interface capability associated with the technology described herein.

[0059] The probe beam and IR pump beam are combined by a dichroic mirror (DM), and the resulting combined beam is directed onto reflective objective (RL). After multiple scattering within tissue, the forward scattered photons from the focus are redirected to the backward direction before reaching the split detector. The propagation path of collected photons aligns obliquely from the focus to the split detector. The oblique optical detection or illumination using two fibers provides phase gradient information of samples for improvedcontrast, in which phase gradient signal is proportional to refractive index variations (^^⁄ ^^ ,^ is refractive index of sample, ^ is the lateral distance). Yet, oblique detection itself only gives the phase gradient information, without molecular sensitivity.

[0060] In OPTM of the current disclosure, photothermal expansion of an object decreases its refractive index, thus reducing the deflection angle of the original probe path. Consequently, with photothermal effect, one photodiode receives more photons while the other receives fewer. Thus, photothermal modulations from each half of the split detector PD1, PD2 have opposite signs. Subtracting the two signals of the split detector not only enhances the photothermal signal but also suppresses the laser noise via a balanced detection operation. In comparison, classic photothermal microscopy only collects a small fraction of back scattered photons from the focus, which could be overwhelmed by the laser noise and detector noise. The approach of the current disclosure increases the photon collection efficiency by 500 times and suppresses the laser noise by a factor of 12 via balanced detection. Leveraging its enhanced sensitivity, OPTM of the current disclosure allows low- dose IR spectroscopic imaging of animal skin, thereby avoiding the risk of photodamage. Consequently, OPTM of the current disclosure enables in vivo IR spectroscopic imaging of metabolic markers within animal and human skin. Moreover, OPTM allows for depth- resolved monitoring of topical drugs under mouse and human skin, unveiling the topical drug pathways and allowing quantitative evaluation of drug delivery efficiency. These advances highlight OPTM’s potential in biomedical research and in situ molecular analysis.

[0061] In an epi-detected MIP microscope, IR and visible probe beam are combined via a dichroic mirror (DM), focused by a reflective objective (RL), then delivered onto a sample, as illustrated in Fig.1(a). The scattered probe photons are collected by the same objective. An iris is employed before the remote photodetector PD to maximize the photothermal signals. In OPTM of the current disclosure, both pump and probe beams are coaxially aligned. A split detector PD1, PD2 is positioned upon the sample surface, allowing it to collect more epi-propagated probe photons, as illustrated in Fig.1(b). The photon propagation path aligns obliquely from the focus to the detector. The difference in intensity between the two halves of the split detector PD1, PD2 reveals the phase gradient information of an object within its surrounding medium, while the sum provides optical absorption information of the same object. Upon photothermal modulation, the intensity difference between both detectors PD1, PD2 yields photothermal phase gradient (PTPG) contrast. For comparison with PTPG images, sum images are acquired simultaneously by taking the sum of the split detector’s signals.

[0062] Referring to Fig.1(c), with the IR pump off, the object deflects the beam path due to different refractive indices between the object and its environment (Fig.1(c), top). The intensity difference between the split detector is proportional to the variation ofrefractive index in samples (^^⁄ ^^ , ^ is refractive index of sample, ^ is the lateral distance).With the IR pump on, the object absorbs IR photon energy, undergoes thermal expansion, and creates a thermal lens in its surrounding environment (Fig.1(c), bottom). Because the thermal lens decreases the refractive index and increases the dimension of the object, it reduces the deflection angle of the original probe path. Compared to the IR-off status, one photodiode receives more photons while the other receives fewer photons. Consequently, the photothermal signals from each half of the split detector PD1, PD2 have opposite signs. Therefore, the subtraction operation enhances the photothermal signals. Meanwhile, equivalent to balanced detection, the subtraction operation suppresses the common-mode fluctuation which is contributed by the probe laser noise.

[0063] To quantitatively demonstrate an improved photon collection efficiency, a Monte Carlo simulation was performed to analyze the propagation of four million photons within a uniform scattering layer, with reference to Fig.1(d). On the layer surface, the objective only collects the central photons within a finite angle of 30 degrees corresponding to the numerical aperture (NA) of the objective (RL). In contrast, the split detector PD1, PD2 collects the scattered photons in a larger area as indicated with black dash box in Fig.1(d). The photons on the detection areas of the split detector and the objective were summed, and the sum was divided by the total photons on sample surface, and the photon collection efficiencies of the split detector and objective were generated, as indicated by red (upper) and black (lower) curves in Fig.1(e). Although the slit within split detectors PD1, PD2 causes the leakage of photons, the collection efficiency of split detector PD1, PD2 can still reach to 39% while that of the objective is only ~0.1%. The improvement factor was computed by dividing the collection efficiency of the split detector PD1, PD2 to that of objective RL. The results indicate oblique detection offers around 500-times larger collection efficiency of scattered photons than the objective detection method, as illustrated in Fig.1(f). In summary, instead of pinhole filtering, OPTM of the current disclosure uses a split detector PD1, PD2 to collect 500-fold more epi-propagated probe photons. Harnessing the inversed photothermal modulations on the split detector, OPTM employs the phase-gradient signal as an efficient readout of photothermal contrast by subtracting the signals from the split detector PD1, PD2.

[0064] Figs.2(a) through 2(e) schematically illustrate OPTM microscope 10 instrumentation and signal processing, according to exemplary embodiments. Fig.2(a) is a schematic functional diagram of a portion of an OPTM microscope 10, according to some exemplary embodiments. Fig.2(b) is a schematic functional block diagram of a scan unit of the microscopy 10. Fig.2(c) is a schematic circuit diagram of split detector for collecting photocurrent from each photodiode independently of microscope 10. Fig.2(d) is an image of split detector PD1, PD2 of microscope 10. Fig.2(e) includes a signal processing flowchart of generating absorption and photothermal phase gradient images from split detector PD1, PD2 signals of microscope 10. For Figs.2(a) through 2(e), DM: dichroic mirror. RL: reflective objective. PD: photodiode. CW: continuous wave. GM: galvo mirror. CM: concave mirror.

[0065] As illustrated in Fig.2(a), the OPTM of the current disclosure uses a pulsed IR laser source 14 with a visible continuous wave laser source 12 as the pump and probe sources, respectively. The resulting pulsed IR and visible CW beams are combined by a dichroic mirror (DM), and the combined beams are delivered to a scan unit 16 for beam scanning and focused by a reflective objective RL onto a sample. Referring to Fig.2(b), which is a detailed schematic illustration of the scan unit 16, the scan unit 16 includes a 2D galvo scanner for scanning laser beams, including two galvo mirrors GM1, GM2, followed by reflective relay optics that includes two concave mirrors CM1, CM2. The use of reflective relay optics and a reflective objective mitigates the achromatic aberrations of the visible and IR beams. To enable oblique photothermal detection in the setup, a split detector including of two identical photodiodes PD1, PD2, is placed above sample surface (Fig.2(a)). A slit between the detectors allows the laser beam to pass through. To facilitate oblique photothermal detection, a detection circuit is provided on a customized printed circuit board (PCB), as illustrated in Figs.2(c)-2(d). Because oblique detection collects more probe photons, to avoid saturation, the split detector is biased by a 100-V DC source (VB) to increase the saturation threshold of photodiodes. A RC lowpass filter is used to remove the high-frequency noise in the DC source. Signals indicative of detections by detection elements PD1, PD1 of the split detector are received by processing system or processor 22. Processing system or processor 22 carries out the processing and control used to effect the OPTM microscopy approach of the current disclosure. Processing system or processor 22 can include processing circuitry and associated elements such as memory, input / output facilities, interface facilities, etc. used to implement the current technology. Processing system orprocessor 22 receives the signals indicative of detections made by detection elements PD1, PD2 and processes those signals as described herein to generate an image of the sample. Processing system or processor 22 can include, for example, a general purpose computer, a microprocessor system, a microcontroller system, or any processing system used to process data and provide control and input / output and user interface capability associated with the technology described herein.

[0066] The signals from two halves of split detector are recorded independently, following the signal processing flowchart shown below. For contrast, an epi-detected MIP microscope is shown in Fig.8, where a beamsplitter is used to reflect the backward- propagated photons, then filtered by an iris before reaching a remote photodiode. Fig.8 is a schematic functional block diagram of a mid-infrared photothermal (MIP) microscope. In the detection path, a non-polarizing beamsplitter BS (R50:T50) is used to reflect the backward propagated probe photons, which is filtered by an iris before reaching a remote photodiode PD. The signal from photodiode PD is separated into DC and AC components by a biased tee. The DC component is recorded by a DAQ card. The AC component is amplified by a RF amplifier and delivered into the lock-in amplifier to generate photothermal signals. For Fig.8, BS: beam splitter. PD: photodiode. DM: dichroic mirror. RL: reflective objective.

[0067] Continuing to refer to Figs.2(a) through 2(e), a signal processing flowchart illustrating an approach to generating photothermal phase gradient and sum images, according to the current technology, is illustrated in Fig.2(e). Signals from each half of the split detector PD1, PD2 are split into AC and DC components. The DC components are recorded by a data acquisition card (DAQ). The sum of DC channels yields optical absorption information of probe wavelength, while their difference provides phase gradient of the objects. Each AC component is amplified and delivered to a lock-in amplifier for frequency- dependent demodulation of photothermal signals. The in-phase (X) and quadrature (Y) components are then output and recorded by a DAQ system. Through vector subtraction of AC components, OPTM images are generated to provide photothermal phase gradient (PTPG) information. The sum images are generated by a vector summation operation for comparison with PTPG images.

[0068] To validate the principle of oblique photothermal detection, the OPTM of the current disclosure was employed to image well-defined microparticles in a scattering medium. Specifically, 10-µm polymethyl methacrylate (PMMA) beads were embedded in ascattering medium composed of 1% intralipid and polydimethylsiloxane (PDMS) mixture, then imaged using the OPTM of the current disclosure, with reference to Figs.3(a) through 3(g).

[0069] Figs.3(a) through 3(g) illustrate OPTM imaging of microparticles in a scattering medium, according to exemplary embodiments. For the imaging of Figs.3(a) through 3(g), 10-µm PMMA beads were embedded in a scattering medium and imaged by the OPTM microscope 10. Fig.3(a) illustrates DC images of PD1 and PD2. Fig.3(b) illustrates phase gradient and absorption DC images. Fig.3(c) illustrates line plotted indicated by the white arrows in Fig.3(b). Fig.3(d) illustrates in-phase (X) and quadrature (Y) images of each photodiode from the lock-in amplifier. Fig.3(e) illustrates scatter plotted in region of interest indicated by black arrow in Fig.3(d). Fig.3(f) illustrates photothermal phase gradient (PTPG) and absorption images. Fig.3(g) illustrates spectra at position indicated by the white arrow in Fig.3(f). For Figs.3(a) through 3(g) scale bar 10 µm. Fieldof view 75× 75 µm2. Probe power on sample was 10 mW. IR power on sample was ~2 mWwith repetition rate 390 kHz, pulse width 80 ns. The imaging speed was 2.5 frame per second with a pixel dwell time of 10 µs. SNR: signal-to-noise ratio. Std: standard deviation.

[0070] In the DC images, both halves of the split detector PD1, PD2 yield the phase gradient information of the microparticles (Fig.3(a)). The DC signals show an inversed polarity at the two ends of particles. The DC1 and DC2 images show the inversed phase gradients, akin to taking the first derivative of the refractive index from opposite horizontal directions. By subtracting and summing the DC1 and DC2 images, phase gradient DC and absorption DC images are generated, with reference to Fig.3(b). Fig.3(c) shows the plots of white lines in Fig.3(b) to illustrate that the phase gradient measurement improves the contrast for visualizing particles in a scattering medium compared to absorption measurements. Because PMMA microparticles have negligible absorption at the probe wavelength of 532 nm, no visible features appear in the absorption DC images.

[0071] The photothermal images are acquired with the IR wavenumber of 1729 cm-1, contributed by the stretching absorption of C=O bond in PMMA particles. In the photothermal AC images, a lock-in amplifier demodulates frequency-dependent photothermal signals and generates both in-phase (X) and quadrature (Y) components of each split detector to form a complex photothermal response (X+iY), with reference to Fig.3(d). The X and Y channels form a vector signal, where X is the real part and Y is the imaginary part. Fig.3(e)shows the vector signals of the split detector in a 2D coordinates from the region of interest indicated by the black arrow in Fig.3(d). The vector signal of PD1 (X1+iY1) has an inversed direction compared to the vector signal of PD2 (X2+iY2). This indicates that the photothermal modulations of both photodiodes have opposite signs. Compared with the steady status of the IR pump off, with IR pump heating, one photodiode receives more photons while the other receives fewer. These results confirm the principle of oblique photothermal detection. Subsequently, through a post-processing step involving vector subtraction and summation, photothermal phase gradient (PTPG) images and sum images are derived, with reference to Fig.3(f). In the results, the signal-to-noise ratios (SNR) of PTPG and sum images reach 417 and 31, which indicates that OPTM is capable of imaging PMMA particles in a scattering medium with a 13-fold improvement of SNR over sum image and 7- fold reduced laser noise. Additionally, the off-resonance images were acquired at 1770 cm-1to demonstrate the bond-selectivity of OPTM imaging, with reference to Fig.9, which includes images acquired by OPTM at off-resonance of infrared absorption (1770 cm-1). ForFig. 9, scale bar 10 µm. Field-of-view 75× 75 µm2. Probe power on sample was 10 mW. IRpower on sample was ~2 mW with repetition rate 390 kHz, pulse width 80 ns. The imaging speed was 2.5 frame per second with a pixel dwell time of 10 µs. Fig.3(g) shows the spectra of PTPG and sum at the region of interest indicated with the white arrow in Fig.3(f). It indicates that OPTM enhances the spectroscopic intensity of particles in a scattering medium by over 10-fold in comparison with the sum measurement. Together, the results validate the principle of oblique photothermal detection, leveraging the differential response of split detectors to enhance photothermal signals while effectively suppressing laser intensity noise.

[0072] Different lipid compositions are important indicators for skin cancer, such as sebaceous carcinoma. Coherent Raman scattering microscopy has allowed imaging of total amount of lipids using C-H vibrational signals. OPTM of the current disclosure can differentiate free fatty acid and lipid ester in individual adipocytes under live mouse skin. To compare the sensitivity of OPTM of the current disclosure versus the MIP microscope, adipocyte cells in the back skin of a live mouse were imaged using OPTM and MIP. Adipocytes on the epidermal layer of animal skin are vital to health as they form a barrier against external pathogens and serve as indicators of skin biology and disease. High- sensitivity adipocyte imaging in live animals can facilitate the study and diagnosis of health status.

[0073] The OPTM of the current disclosure and the MIP microscope were employed to image the same region in the epidermis layer of a live mouse’s back skin at a penetration depth of 50 µm. To differentiate different lipid contents, multispectral OPTM and MIP images were acquired from 1700 to 1770 cm-1with a step size of 2 cm-1, which is contributed by C=O vibration absorption of lipids. The DC images of OPTM and MIP were acquired simultaneously, measuring phase gradient, absorption, and back scattering information, with reference to Figs.10(a) through 10(e). Figs.10(a), 10(b), 10(c) illustrate DC images of phase gradient, absorption, and MIP modalities in 3(a) through 3(g). For Figs. 10(a), 10(b), 10(c), scale bar 30 µm. Fig.10(d) illustrates statistical analysis of pixel-wise amplitudes in DC images of MIP and absorption, representing photo currents detected by remote photodiode and split detectors based on photon counts. Fig.10(e) illustrates improvement factor, calculated by dividing the amplitude of absorption DC images by that of MIP DC images, demonstrating scaling improvements of up to 700-fold.

[0074] Following the flowchart in Fig.2(e), photothermal phase gradient images were acquired by OPTM. Meanwhile, sum images were generated for a comparison purpose. MIP images were acquired independently using pinhole filtering approach as illustrated in Fig.8. The photothermal images at two IR wavenumbers, 1704 and 1742 cm-1, are selected to present the fatty acid and lipid ester-rich distributions, with reference to Figs.4(a)-4(c). The absorption peaks of the fatty acid and lipid ester are contributed by the vibrational absorption of acidic and esterified C=O bond in lipids, which are consistent with FTIR spectroscopy measurement in animal skin samples. Off-resonance images at wavenumber 1770 cm-1were acquired to confirm bond-selective imaging of OPTM and MIP. Spectra in two regions of interest in Fig.4(a) are plotted to illustrate the absorption peaks of fatty acids and lipid esters at IR wavenumbers 1704 and 1742 cm-1, with reference to Fig.4(d).

[0075] Figs.4(a) through 4(f) illustrate sensitivity improvements over MIP spectroscopy by OPTM of the current disclosure. OPTM unveils two types of lipids in adipocytes under live mouse skin. Figs.4(a) through 4(f) illustrate OPTM and MIP imaging of adipocyte cells in the back skin of a live mouse at a penetration depth of 50 µm. Fig.4(a) illustrates photothermal phase gradient (PTPG). Fig.4(b) illustrates absorption, and Fig.4(c) illustrates MIP images acquired in a same field of view. Fig.4(d) illustrates PTPG spectra of two regions of interest in Fig.4(a) illustrating the absorption peaks of the fatty acid and the lipid ester. Fig.4(e) illustrates line 1 and Fig.4(f) illustrates line 2 plotted in the fatty acidand the lipid ester images. For Figs.4(a) through 4(f), scale bar 30 µm. Field of view 150×150 µm2. Probe power on sample was 10 mW. IR power on sample was ~2 mW with repetition rate 390 kHz, pulse width 80 ns. The imaging speed was 1.6 frame per second with a pixel dwell time of 10 µs.

[0076] With reference to Figs.4(e) and 4(f), to quantify the sensitivity improvement of OPTM over MIP, intensities were plotted along two dash lines in the fatty acid and lipid ester images. Regarding the noise levels, PTPG signals show a 12-fold and 8-fold suppression of noise compared to the sum signals and the MIP signals, respectively. These results indicate that, OPTM can effectively suppress the common-mode noise, primarily contributed by laser intensity noise, via a balanced detection strategy. In parallel to noise suppression, OPTM enhances the photothermal signal by up to 8-fold over MIP owing to the higher photon collection efficiency. As indicated in dashed black rows in Figs.4(e) and 4(f), OPTM enables visualization of lipid structures not detectable in MIP. The peak at the PTPG image of lipid ester has a full width at half maximum of 0.86 µm, indicating sub-micron lateral resolution of OPTM.

[0077] Depth-resolved infrared spectroscopic imaging is important for accurately visualizing and analyzing chemical contents in the intricate layers and structures within complex tissues. Despite its importance, in vivo infrared depth-resolved imaging remains an unmet need in the field. The OPTM technique of the current disclosure addresses this gap by providing optical sectioning capability at visible light resolution through a pump-probe approach.

[0078] Figs.5(a) through 5(c) illustrated in vivo depth-resolved OPTM imaging. In connection with Figs.5(a) through 5(c), OPTM was utilized to image the ear of a live mouse at different penetration depths. At each depth, protein images were acquired at an IR wavenumber of 1553 cm-1, contributed by the absorption of the Amid II band. Two lipid contents, fatty acid and lipid ester, were imaged at IR wavenumbers of 1704 and 1742 cm-1due to the vibrational absorption of acidic and esterified C=O bond in lipids. Additionally, off-resonance images at 1770 cm-1were acquired to confirm bond-selectivity of OPTM imaging.

[0079] Figs.5(a) through 5(c) illustrate depth-resolved in vivo OPTM imaging of mouse skin, according to exemplary embodiments. Fig.5(a) illustrates phase-gradient DCimages and photothermal phase gradient images at imaging depth of 10 µm (stratum corneum layer). Fig.5(b) illustrates phase-gradient DC images and photothermal phase gradient images at imaging depth 50 µm (viable epidermis layer). Fig.5(c) illustrates phase-gradient DC images and photothermal phase gradient images at imaging depth 75 µm (epidermis layer). For Figs.5(a) through 5(c), scale bar 50 µm. Field of view 160×160 µm2. Step size 0.4 µm. Probe power on sample was 10 mW. IR power on sample was ~2 mW with repetition rate 390 kHz, pulse width 100 ns. The imaging speed was 0.5 frame per second with a pixel dwell time of 10 µs.

[0080] On the stratum corneum (SC) layer at a depth of 10 µm beneath skin surface, which is the outmost layer of skin, the mouse hair is visualized (Fig.5(a)). SC shows sparse lipid and protein distributions because it includes mostly of dead cells. The viable epidermis layer, right beneath the SC layer, mainly consists of live cells providing nutrition to the skin. Thus, the protein and lipid content are uniformly distributed at a depth of 50 µm (Fig.5(b)). As the penetration depth increases to 75 µm into the epidermis layer, sebaceous glands become visible, predominantly composed of lipid esters and some fatty acids (Fig.5(c)). At a penetration depth of 100 µm, which is equal to one mean free length of 532-nm photons, the phase gradient DC image shows a good contrast for resolving deep tissue features, with reference to Fig.11, which illustrates in-vivo OPTM imaging of mouse ear skin at a penetration depth of 100 µm. In Fig.11, Scale bar 50 µm. Field-of-view 160×160 µm2.

[0081] Photothermal phase gradient images still show differentiable chemical distributions. Thus, the penetration depth limit of OPTM is determined to be 100 µm for imaging animal’s skin. These results collectively validate OPTM's depth-resolved imaging capability in live animals.

[0082] Topical transdermal drug delivery presents a promising alternative to both oral administration and hypodermic injections. In vivo investigation of topical drug delivery inside animal skin facilitates analysis of the drug pathway, guiding the development of new biomedicine. Quantitative evaluation of effective transdermal drug delivery is important to optimize therapeutic efficacy while minimizing systemic side effects. Figs.6(a) through 6(g) include images which demonstrate that OPTM facilitates high-resolution and depth-resolved in vivo tracking of topical drug pathways in mouse skin with quantitative evaluation capabilities. Specifically, the skin penetration behavior of benzoyl peroxide (BPO), acommonly used active pharmaceutical ingredient, is explored. BPO is particularly effective for acne treatment, as it penetrates hair follicles to exert antimicrobial and keratolytic effects.

[0083] Figs.6(a) through 6(g) illustrate in vivo investigation of topical drug pathway inside mouse skin. Fig.6(a) illustrates pseudo-color hyperspectral OPTM images of skin of a live mouse at a depth of 32 µm with benzoyl peroxide administrated. Fig.6(b) illustrates reference spectra of fatty acid, lipid ester, and benzoyl peroxide. Fig.6(c) includes images of fatty acid, lipid ester, and benzoyl peroxide after applying LASSO spectral unmixing method to the hyperspectral image of Fig.6(a) to differentiate different contents. Fig.6(d) illustrates a captured 3D view to illustrate mouse skin chemical structure and topical BPO distribution. Fig.6(e) illustrates phase gradient DC images and PTPG images at different penetration depths in mouse skin. Fig.6(f) illustrates efficiency of topical BPO delivery at different depths, based on analysis of 5 FOV images shown in Fig.6(e) and Figs.14(a) through 14(d). Figs.14(a) through 14(d) illustrate depth-resolved phase gradient DC images and PTPG images at four field of views for unveiling drug pathway and quantitative evaluations of drug delivery efficiency. Scale bar 50 µm. Field-of-view 180×180 µm2. Probe power on sample was 10 mW. IR power on sample was ~2 mW with repetition rate 390 kHz, pulse width 100 ns. The imaging speed was 1.1 frame per second with a pixel dwell time of 10 µs. FA, fatty acid. LE, lipid ester. BPO, benzoyl peroxide.

[0084] . Fig.6(g) illustrates endogenous chemical structure and topical BPO distribution within the skin layers at a depth of 32-48 µm, where the sebaceous glands are located. Additional images of the same region at different depths are presented in Fig.15, which illustrates depth-resolved images illustrating endogenous chemicals, including sebaceous glands, and the distribution of topical BPO beneath the skin. The protein images were acquired at 1553 cm⁻¹, corresponding to the Amide II band absorption. FA, fatty acid. LE, lipid ester. BPO, benzoyl peroxide. For Fig.15, scale bar 50 µm. For Figs.6(a) through 6(g), scale bar 50 µm. Field of view 180×180 µm2. Probe power on sample was 10 mW. IR power on sample was ~2 mW with repetition rate 390 kHz, pulse width 100 ns. The imaging speed was 1.1 frame per second with a pixel dwell time of 10 µs. FOV, field of view. FA, fatty acid. LE, lipid ester. BPO, benzoyl peroxide.

[0085] To visualize both BPO and endogenous chemical contrasts, multispectral in vivo OPTM imaging was performed at various depths in mouse skin, following theadministration of 5-µL BPO samples. Fig.6(a) shows a pseudo-colored multispectral image acquired at a depth of 32 µm beneath the mouse skin surface. To quantitatively disentangle the different chemical components from the composite spectrum, a least absolute shrinkage and selection operator (LASSO) was employed for spectral unmixing. In LASSO, it is important to have reference spectra of different chemical contents. As illustrated in Fig.6(b), the reference spectra of fatty acid and lipid ester are acquired from the regions indicated by the red (upper) and green (lower) arrows in Fig.6(a), which are rich in fatty acid and lipid ester, respectively. The reference spectrum of BPO is acquired from the infrared photothermal spectrum of drug samples, with reference to Figs.12(a) through 12(e), which illustrate mid-infrared photothermal spectrum of benzoyl peroxide. Fig.12(a) is a photo of a skin-care product used. Fig.12(b) is a close-up photo to indicate that the active ingredient is 10% benzoyl peroxide. Fig.12(c) illustrates the molecular structure of drug molecule, benzoyl peroxide. Fig.12(d) illustrates experimental set-up for measuring infrared photothermal spectrum of samples. For Figs 12(a) through 12(d), IR, infrared. RL, reflective objective. Fig.12(e) illustrates normalized infrared photothermal spectrum of benzoyl peroxide gel. The peak at 1760 cm-1is from the esterified C=O bond.

[0086] Fig.6(c) shows the spectral unmixing results of LASSO that separate the fatty acid, lipid ester, and administrated BPO contents from the raw multispectral image. An overlap of three chemical components shows the relative distributions of endogenous lipid and the drug contents. LASSO spectral unmixing was applied to other depth-resolved imaging data to trace topical BPO pathway. A control experiment, conducted without drug administration, shows no visible contrast in the BPO channels, with reference to Fig.13, which illustrates control group images without benzoyl peroxide administrated, including phase gradient DC images and PTPG images at different penetration depth in mouse skin. For Fig.13, scale bar 50 µm. Field-of-view 180×180 µm2.Step size 0.6 µm. Probe power on sample was 10 mW. IR power on sample was ~2 mW with repetition rate 390 kHz, pulse width 100 ns. The imaging speed was 1.1 frame per second with a pixel dwell time of 10 µs. FA, fatty acid. LE, lipid ester. BPO, benzoyl peroxide.

[0087] Through depth-resolved infrared spectroscopic imaging and LASSO spectral unmixing, it is demonstrated that OPTM enables three-dimensional visualization of chemical content in mouse skin, illustrating endogenous skin chemical structure and administrated BPO distribution.

[0088] Fig.6(d) shows a captured three-dimensional (3D) view, in which the mouse hairs and the hair follicles opening can be seen. Protein images were acquired at 1553 cm-1contributed by the absorption of Amid-II band. To unveil the BPO pathway under mouse skin, the depth-resolved phase gradient DC and different chemical images illustrating the skin morphology, chemical structure, and drug / skin distributions at different skin layers are shown (Fig.6(e)). At the stratum corneum layer with a depth 0-8 µm, which is the skin surface, the hair follicle locates at one end of the mouse hair as indicated by white dash circle in the phase gradient DC image. There is topical BPO content overlapping with the hair follicle opening. At the epidermis layer with depths of 16-24, 32-40, and 48-56 µm, BPO content overlaps with the hair follicles indicated by white dash circles. This confirms that BPO penetrate into the hair follicles in the skin through its openings, as the acnes happen in the hair follicles. At a deeper penetration depth of 64 µm, the endogenous chemical and BPO images still have visible contrast, demonstrating the capability of OPTM to trace drug molecules in deep skin layer. Through all skin layers, the distribution of administrated drug content is not uniform but aggregates. It was also determined that BPO content also penetrates sweat pores as indicated by the solid white arrows in Fig.6(e). To quantitatively evaluate the delivery efficiency of topical BPO, another four group images at different fields of view were taken, with reference to Figs.14(a) through 14(d). Since BPO is effective to treat acne by penetrating hair follicles to perform antimicrobial and keratolytic activities, the delivery efficiency is quantified by computing the ratio between the amount of drug content within hair follicles and within the whole field of view at a same depth (Fig.6(f)). From the skin surface, with a deeper penetration depth, the BPO content within the hair follicle increases at first, then followed by a decreasing. It indicates that topical BPO enters the hair follicles through its opening and remains inside the follicles. The delivery efficiency of BPO is less than 30% in skin layers. Such a low efficiency indicates the treatment of skin acne needs a long-term administration of topical BPO, which is aligned with previous studies of acne treatment by BPO (58).

[0089] The interaction between topical BPO and sebaceous glands has not been clarified, and the side effects of BPO remain controversial. Here, OPTM of the current disclosure is used for monitoring the spatial distributions of BPO and sebaceous glands. Fig. 6(g) shows the skin layers, including the sebaceous glands, at a depth of 32-48 µm. The sebaceous glands are visualized at endogenous lipid images indicated by white dashenclosures, which does not overlap with BPO. Fig.15 shows the depth-resolved images in the same region for illustrating the drug pathway within the skin layers.

[0090] In summary, OPTM shows that BPO penetrates the mouse skin through hair follicles opening, sweat pores, and other skin areas. The low efficiency of BPO delivery via hair follicles, which is less than 30%, indicates a need for long-term administration of BPO for the treatment of skin acne. In addition, the observations show that topically applied BPO does not reach sebaceous glands. Collectively, OPTM enables in vivo depth-resolved tracking of topical drug pathway with quantitative evaluation of delivery efficiency, providing guides to the development of new transdermal biomedicine.

[0091] Non-invasive chemical analysis of live human subjects plays an important role in biomedical research and clinical diagnostics. Providing fingerprint information with high spatial resolution, OPTM is a good candidate for non-invasive chemical analysis of human skin. In OPTM, the use of visible probe offers sub-micron resolution, while the improved sensitivity allows low-dose IR spectroscopic imaging without the concern of photodamage. Figs.7(a) through 7(i) demonstrate OPTM’s ability to non-invasively image the metabolic markers under live human skin and the topical transdermal drug content on the skin surface.

[0092] Figs.7(a) through 7(i) illustrate in vivo OPTM imaging of human skin. Fig. 7(a) includes a schematic diagram and a photographic image of in vivo OPTM imaging of a human’s forearm skin, according to exemplary embodiments. Fig.7(b) includes photothermal phase gradient images of endogenous proteins and lipids in the viable epidermis layer with a depth of 40 µm without topical drug administration. Fig.7(c) includes a phase gradient DC image of the stratum corneum showing a hair on the skin surface and the hair follicle opening. Figs.7(d) through 7(g) include photothermal phase gradient images of protein (Fig.7(d), 1553 cm-1), fatty acid (Fig.7(e), 1704 cm-1), lipid ester (Fig.7(f), 1742 cm-1), and benzoyl peroxide (Fig.7(g), 1760 cm-1) in the stratum corneum layer with a depth of 8 µm with topical drug administration. Fig.7(h) includes an overlap image of phase gradient DC and drug distribution. Fig.7(i) includes an overlap image of endogenous lipid and administrated drug contents. For Figs.7(a) through 7(i), FA, fatty acid. LE, lipid ester. BPO, benzoyl peroxide. Scale bar 50 µm. Field of view 120×120 µm2. Probe power on sample was 10 mW. IR power on sample was ~2 mW with repetition rate 390 kHz, pulse width 100 ns. The imaging speed was 2 frame per second with a pixel dwell time of 8 µs. PG, phase gradient. FA, fatty acid. LE, lipid ester. BPO, benzoyl peroxide.

[0093] A volunteer’s forearm was imaged by OPTM to visualize the skin chemical structure and the topical drug content upon the skin surface, with reference to Fig.7(a). With reference to Fig.7(b), OPTM visualizes endogenous proteins and lipids in the viable epidermis layer with a depth of 40 µm without topical drug administration. Off resonance imaging at 1780 cm-1confirms the bond-selective imaging of human skin. The relative distribution of protein and different lipid contents under skin surface are an important health indicator, such as skin aging and inflammatory skin disease.

[0094] Additionally, it was demonstrated that OPTM enables to trace the topical drug pathway on human skin surface. Figs.7(c) through 7(g) show phase gradient DC and PTPG images of endogenous chemicals and administrated benzoyl peroxide contents on human arm skin with a depth of 8 µm. From phase gradient DC image in Fig.7(c), the hair follicle opening located at the end of one human hair can be seen. Fig.7(g) shows that topically administrated BPO contents on human skin surface is also not uniform but aggregates, which aligns with the observations of mouse imaging in Figs.6(a) through 6(g). The BPO content overlap with hair follicles opening is shown in Fig.7(h), which demonstrates that BPO penetrates skin through hair follicle opening and other skin areas. The percentage of BPO through follicle is estimated to be 46.5% based on two different field of views.

[0095] Figs.7(d) through 7(f) show a mixture of chemical content within the hair follicles, such as sebum. Fig.7(i) shows the overlap of endogenous chemicals and topical BPO, demonstrating that OPTM allows to map both topical biomedicine and skin chemical structure. To verify the imaging fidelity, a control experiment was performed by using OPTM to image human arm skin without BPO administration, with reference to Figs.16(a) through 16(e), which illustrate in vivo OPTM imaging of human skin without benzoyl peroxide administration. Fig.16(a) illustrates phase gradient DC image of the stratum corneum showing a hair on the skin surface and the hair follicle opening. Figs.16(b) through 16(e) illustrate photothermal phase gradient images of protein (Fig.16(b), wavenumber 1553 cm-1), fatty acid (Fig.16(c), wavenumber 1704 cm-1), lipid ester (Fig.16(d), wavenumber 1742 cm-1), and benzoyl peroxide (Fig.16(e), wavenumber 1760 cm-1). For Figs.16(b) through 16(e), scale bar 50 µm.

[0096] . The BPO image in the control group shows no visible contrast within the hair follicles. It confirms that the contrast of Fig.7(g) comes from the administrated BPO. Collectively, we demonstrate that OPTM allows non-invasive infrared spectroscopic imagingof human skin in vivo. OPTM visualizes the metabolic biomarker beneath human skin surface, facilitating heath diagnosis in clinical application. OPTM also allows monitoring of topical drug content on human skin, facilitating the development of transdermal biomedicine.

[0097] Oblique photothermal detection challenges the conventional wisdom of photothermal sensing by fundamentally altering the photon collection strategy. Traditional photothermal methods typically measure a small fraction of backward-scattered photons filtered through an iris, which inherently limits the number of collected photons. This limitation arises because only photons scattered in specific directions are detected, reducing the signal intensity and detection sensitivity. In contrast, oblique photothermal detection employs a split detector positioned on the sample surface, enabling the collection of both backward and forward-scattered photons. The forward-scattered photons are redirected into the backward direction through multiple scattering events within the sample. Essentially, this method significantly enhances the photon collection efficiency.

[0098] Furthermore, OPTM suppresses the laser noise via balanced detection. By leveraging the inversed photothermal modulations on the split detector, OPTM amplifies the photothermal signal amplitude by analyzing the differential response between the two halves of the split detector. Since the photothermal signal predominantly occupies the low-frequency region due to its low-duty cycle nature, traditional photothermal detection methods are compromised by 1 / f laser noise. OPTM collects 500 times more photons from tissue, and mitigates the laser noise through a subtraction operation akin to the balanced detection strategy.

[0099] OPTM could detect weak scatterers in scattering samples by measuring the photothermal phase gradient of the object rather than back scattering signals. The amount of back scattering photons from the object is limited by the minimal refractive index difference within tissue. Traditional photothermal detections, which also gather other diffuse and specular reflected photons, could have the scattering signals from the focus overwhelmed by the laser noise. Oblique photothermal detection addresses this challenge by utilizing a split detector near sample surface to acquire phase gradient as an ultrasensitive readout of photothermal signals. By measuring the photothermal phase gradient rather than back scattered photons, OPTM boosts the imaging sensitivity, enabling visualization of weak scatterers in complex tissue. Due to the enhanced sensitivity, in vivo sub-video-rate imaging of protein in mouse skin is shown in Fig.17, which illustrates sub-video-rate in vivo OPTMimaging of protein in mouse abdominal skin, capturing skin movement during breathing. This result demonstrates the capability of OPTM for longitudinal imaging of a dynamic physiological process. The images were acquired at the Amide II band (1553 cm⁻¹). For Fig. 17, scale bar: 10 µm. The image consists of 80 × 80 pixels with a step size of 0.6 µm and a pixel dwell time of 9 µs. To ensure linear scanning, the fast-axis scan includes an additional 80 pixels. The imaging frame rate is 8 fps.

[0100] Though skin applications have been discussed, OPTM of the current disclosure is applicable to other tissues. Figs.18(a) through 18(h) illustrate OPTM imaging of a mouse brain slice. A 500-µm thick coronal brain slice was placed between a calcium fluoride (CaF₂) glass substrate and a coverslip, then positioned on a Teflon block to serve as a scattering substrate. Images were acquired by OPTM in an epi configuration. Fig.18(a) illustrates phase-gradient direct current (DC) images of the brain slice. The blood vessel is outlined by dashed white lines, and two representative soma cells are marked with dashed white circles. Figs.18(b) through 18(d) illustrate photothermal phase gradient (PTPG) images of endogenous chemical components. Fig.18(b) illustrates a protein image acquired at IR wavenumber 1657 cm-1, contributed by the absorption of Amid I band. The blood vessel wall is visualized to have rich protein content, which is contributed by elastin and collagen. Figs.18(c) through 18(d) illustrate fatty acid and lipid ester image acquired at 1704 and 1742 cm-1, contributed by the absorption of acidic and esterified C=O bond in lipids. The endogenous protein and lipid contents within soma cells were detected. Fig.18(e) illustrates off-resonance image acquired at 1770 cm-1to validate the bond-selective imaging of OPTM. For Figs.18(a) through 18(e), scale bar 30 µm. Fig.18(f) illustrates a three-dimensional protein image acquired at the same region of Figs.18(a) through 18((e) acquired by depth- resolved OPTM measurement. Fig.18(g) illustrates the line plotted of the red dash arrow in Fig.18(b) to illustrate the sub-micron lateral resolution of OPTM imaging. Fig.18(h) illustrates the axial intensity at the location indicated by white arrow in Fig.18(b) to demonstrate the axial resolution 5.4 µm according to the depth intensity profile of the smallest protein structure, equivalent to the axial resolution of MIP microscopic imaging.

[0101] Figs.18(a) through 18(h) demonstrate that OPTM facilitates highly sensitive IR spectroscopic imaging of a mouse brain slice with fine axial sectioning capability. OPTM allows for in situ molecular analysis within the soma and blood vessels, revealing the potential for detailed molecular profiling in complex tissues.

[0102] OPTM exhibits an improved sensitivity to image heterogenous structures, demonstrating a fine impulse response to a point source, with reference to Figs. 19(a) through 19(c), which illustrate OPTM imaging of 0.5-µm PMMA particles in scattering medium at C=O absorption peak 1729 cm-1. The scattering medium includes 1% intralipid in PDMS mixture. Fig.19(a) illustrates the photothermal phase gradient (PTPG). Fig.19(b) illustrates the sum image. Fig.19(c) illustrates the intensity profiles of line 1 and 2 in the Fig.19(a) PTPG and Fig.19(b) sum images.

[0103] Moreover, bulk chemical features can be effectively resolved through Hilbert transform reconstruction, as described in detail below in the Additional Descriptive Material and illustrated in Figs.20(a) through 20(c), which illustrate visualizing bulk features in 10-µm PMMA particles by Hilbert transformed analysis of the phase gradient images. Fig.20(a) (left) illustrates a phase gradient (PG) DC image, with reference to Fig. 3(a) (middle) phase image reconstructed by Hilbert transform. Fig.20(a) (right) illustrates line plotted in PG DC and reconstructed phase image. Fig.20(b) (left) illustrates a PTPG image, with reference to Fig.3(f) (middle) the phase map of PTPG signals. Fig.20(b) (right) illustrates line plotted in PTPG and phase map image. Fig.20(c) (left) illustrates a vector PTPG image. Fig.20(c) (middle) illustrates the reconstructed photothermal image. Fig.20(c) (right) illustrates line plotted in vector PTPG and reconstructed photothermal image. For Figs.20(a) through 20(c), scale bar 10 µm.

[0104] Other advancements, such as the measurement of orthogonal phase gradients, may facilitate the reconstruction of images with minimized computational artifacts.

[0105] There is potential for further enhancing OPTM's penetration depth. The absorption of IR photons and the scattering of probe photons are two main factors of OPTM to penetrate deeper. However, the attenuation length of IR photons (< 200 µm) is still larger than the attenuation length of 532-nm probe photons (125 µm). Thus, the main limiting factor is the strong scattering of probe photons within tissue due to the short visible wavelength used. Employing a longer probe wavelength could facilitate deeper imaging by reducing optical scattering.

[0106] In summary, by providing reduced laser noise, enhanced imaging contrast, and improved sensitivity, OPTM enables high-sensitivity infrared spectroscopic imaging of chemicals in live animals and human subjects with sub-micron resolution. OPTMopens translational opportunities for molecule-based clinical diagnosis and study of topical drug delivery through human skin. As a platform technology, OPTM is not limited to the mid-infrared window and can be broadly adapted to other photothermal microscopy techniques such as overtone photothermal microscopy and visible photothermal microscopy.

[0107] A Monte Carlo simulation was conducted to explore the propagation path of 4-million photons in a uniform scattering layer. The sizes of the layer were (width) 25 mm, (length) 25 mm, and (depth) 10 mm. A focused beam with a numerical aperture (NA) of 0.5 was launched on the surface of simulation layer. The NA of simulated focused beam is equal to the NA of reflective objective used in the OPTM and MIP microscope. To mimic theoptical properties of skin, the scattering coefficient µ s was set from 20 to 125 cm-1, referringto the attenuation length (^ = 1⁄ µ^ ) of 500 to 80 µm. The absorption coefficient was set to0.1 cm-1.

[0108] In some exemplary embodiments, the OPTM of the current disclosure (Figs.2(a) through 2(e) couples a 532-nm continuous wave probe laser (Samba, HUBNER photonics, for example) with a pulsed quantum cascade laser (QCL) (MIRcat 2400, Daylight Solutions, for example) tunable from 900 cm−1to 2,300 cm−1via a Germanium window (GEBBAR-3-5-25, Andover Corporation, for example). The visible / IR combined beam is directed to a 2D galvo scanner (Saturn 5B, ScannerMax, for example) and then through a reflective relay optics, primarily comprising a concave mirror with a 150-mm focusing length (CM1, CM254-150-P01, Thorlabs, for example) and a concave mirror with a 250-mm focusing length (CM2, CM254-250-P01, Thorlabs, for example). Subsequently, both beams are focused onto the sample by a reflective objective (LMM40X-P01, Thorlabs, for example). The objective is mounted on a piezo stage (MIPOS 100 SG RMS, Piezosystem Jena, for example), allowing precise adjustment of the IR and visible foci along the depth axis. In the detection part, a split detector including two identical photodiodes (S3994-01, Hamamatsu or FDS1010, Thorlabs, for example) is biased with a 100-volt voltage to elevate its saturation threshold. A 402 kΩ resistor (R) and a 0.1 µF capacitor (C) form an integrated RC low-pass filter to reject high-frequency noise from the biased voltage supply. A lab-designed printed circuit board integrates the photodiodes and other electrical components, with reference to Fig.21(a). The split detector is mounted on a homebuilt printed circuit board. A copper mesh (PSY406, Thorlabs, for example) is used as a cover for electromagnetic shielding, with reference to Fig.21(b). Here, a 2-mm slit width between two photodiodes is created to allowsthe combined pump and probe beam to pass through. A 3D printed holder is glued at the other side of PCB board so that it can be mounted on the imaging objective, with reference to Fig.21(b). The output of each photodiode is then split into DC and AC signals using a biased tee (ZFBT-4R2GW+, Mini-circuits, for example). The DC signals are collected by a low- noise data acquisition card (DAQ) (Oscar, Vitrek, for example), while the AC signals are amplified by a RF low-noise amplifier (CMP61665-1, Gain 40 dB, NF Corporation, for example). Then, the amplified AC signals are fed into a lock-in amplifier (HF2LI, Zurich, for example) for frequency-dependent demodulation of photothermal signals. To ensure phase- sensitive measurement, the lock-in amplifier generates a TTL trigger for the QCL laser, synchronizing the IR pulse with its internal reference phase. Consequently, the phase of the AC signal is determined by the IR-induced photothermal response. A multichannel DAQ board (National Instruments, PCIe-6363, for example) is used for real-time data acquisition. A complete instrument interface was developed in LabVIEW (LabVIEW 2020, National Instruments Corporation, for example) to enable hardware synchronization, data acquisition, and fast pre-processing.

[0109] Polydimethylsiloxane (PDMS) (Sylgard 184, Dow Corning Corporation, for example) was prepared with a base-to-curing agent ratio of 10:1. Then, 10- µm Polymethylmethacrylate (PMMA) beads (17136-5, Polysciences, for example) were added to the PDMS mixture along with 1% intralipid (I141, Sigma-Aldrich, for example). The attenuation length at 532 nm is 125 µm, which is comparable to that of skin tissue. The mixture was degassed in a vacuum chamber for 30 minutes to remove residual air bubbles. Following this, the composite was then solidified by heating it in an oven at 70°C for 30 minutes.

[0110] Male adult C57BL / 6J and BALB / c mice were used in this study. The details of mouse preparations for imaging are illustrated in Figs.22(a) through 22(d), which illustrate in vivo imaging of mouse skin. Fig.22(a) illustrates step 1: anesthetize a mouse via isoflurane, hair removal, then inject Ketamine and xylizine solution. Fig.22(b) illustrate step 2: use biocompatible glue to fix back skin of a mouse on Teflon block. Fig.22(c) illustrates step 3: imaging. Fig.22(d) illustrates step 4: after experiment, mouse skin was detached from Teflon block without visible hurt. In step 1 (Fig.22(a)), the mouse was anesthetized with isoflurane, and the hair was removed or shaved to clean the skin surface. Subsequently, the mouse was given a subcutaneous injection of 50-100 mg / kg Ketamine and 5-10 mg / kgXylazine, enabling the transition to the imaging process. The level of anesthesia was assessed by the loss of response to pinching the base of the tail. In step 2 (Fig.22(b)), a biocompatible glue was applied to fix the mouse's skin to a Teflon block, which served as a scattering substrate and mitigated breathing artifacts during imaging. A warming bag was placed under the mouse to maintain its body temperature. In step 3 (Fig.22(c)), the mouse was transferred to the OPTM for imaging. In step 4 (Fig.22(d)), after imaging, the mouse skin was detached from the Teflon block without visible hurt. Then, the animals were euthanized via isoflurane anesthesia followed by cervical dislocation. The preparation process for imaging ear skin of mouse is similar to the process described above, excluding the removal of hair. By driving the piezo stage mounted on the objective, the IR and visible foci can be precisely adjusted along the depth axis. This enables depth-resolved OPTM imaging of specific skin layers and facilitates volumetric chemical imaging. Topical benzoyl peroxide drug samples were applied for 1 minute before OPTM

[0111] To mitigate the motion artefact during imaging, the volunteer’s arm was fixed on a holder by transparent tapes. The protein, fatty acid, lipid ester, and benzoyl peroxide contents were imaged at the IR wavenumber of 1553 cm-1, 1704 cm-1, 1742 cm-1, and1760 cm-1, respectively, for shortening imaging time without losing information. No visible damage can be seen on human skin after the imaging experiment. To perform OPTM imaging of human skin at a specific depth, we first locate the skin surface using DC and photothermal signals, then precisely adjust the IR and visible foci to the desired depth by controlling the piezo stage. Topical benzoyl peroxide drug samples were applied for 1 minute before OPTM imaging, following the drug product instructions.

[0112] A skin-care drug was acquired from a drug store which is used to treat skin acne (Fig.12(a)). It includes 10% benzoyl peroxide (BPO) as the active pharmaceutical ingredient (Fig.12(b)). One BPO molecule includes two phenol rings and two C=O bonds (Fig.12(c)). To explore the spectral features of the drug molecule, we used a co-propagation MIP spectroscope to measure the infrared photothermal spectrum of drug (Fig.12(d)). The drug sample was sandwiched by a coverslip and a CaF2 glass substate which is IR transparent. A photodiode was placed in the forward direction with an iris to measure photothermal signals because most probe photons went through such transparent samples. Fig.12(e) shows the normalized infrared photothermal spectrum of benzoyl peroxideindicating the peak locates at 1760 cm-1, which is contributed by vibrational absorption of C=O bond in the BPO molecule.

[0113] Least absolute shrinkage and selection operator (LASSO) used in this study unmixes hyperspectral OPTM imaging data into different chemical components.

[0114] In LASSO, given the hyperspectral OPTM dataset registered as spatial domain, ^^, ^^, spectral domain, ^^, to unmix into multiple chemical components,the 3D hyperspectral stack is first reshaped to a 2D matrix (^ ∈ ^^^^^×^^) by arranging thepixels in the raster order. Assuming the number of interested chemical channels as ^, amodel is used to decompose the data matrix into the multiplication of concentration maps ^ ∈^^^^^×^ and spectral profiles of pure chemicals ^ ∈ ^^×^^:^ = ^^^ + ^ 1!where E is the error. Here, a L1-norm regularization is added to each row of the concentration matrix, and the original inverse problem is solved in a row-by-row manner through LASSO regression: % 15^",: = &'()*^+,,:-. / − + 612^3,:21 2!where ^ is a k-matrix, ^% is the3,: 9:hyperparameter that tunes the level of sparsity. Additional Descriptive Material

[0115] The Hilbert transform is widely used as an approximate integral method for retrieving phase images from phase gradient signals in differentiated intensity contrast microscopy. This approach facilitates the visualization of bulk features in both phantom and biological samples.

[0116] Since the oblique photothermal microscope (OPTM) of the current disclosure captures photothermal-modulated phase gradient information, the Hilbert transform can be applied to reconstruct bulk chemical features. Figs.20(a) through 20(c) illustrate the Hilbert transform pipeline. Fig.20(a) shows the phase gradient DC image of a 10-µm PMMA particle embedded in a scattering phantom, obtained from Fig.3(a). Thephase map Φ ^! is reconstructed from the phase gradient image < ^! using the Hilbert transform (1): Φ^! = =>< ^!? = 7EB ^C!@ A DE ^CD^!^^F(3) where =>< ^!? represents the Hilbert transform operating in a one-dimensional (1D) real space. As shown in Fig.20(a) (middle, right), the reconstructed phase map reveals central features within the PMMA particle, which remain undetected in the original phase gradient DC image.

[0117] In Fig.20(b) (left), the absolute amplitude of the photothermal phase gradient (PTPG) signal is extracted, following the processing pipeline in Fig.2(e), to indicate the infrared (IR) absorption of the 10- µm PMMA particle (derived from Fig.3(f)). The phase map G of PTPG signal is computed as: G= tanD7 KLMN (4)where X and Y denote the in-phase and quadrature components from the lock-in amplifier. As shown in Fig.20(b) (middle, right), the phase map G reveals an inverse phase relationship at the particle’s edges.

[0118] To further enhance the representation of bulk chemical features, the phase information G is incorporated into the PTPG images to construct the vector PTPG mapP OP⃗ :P OP⃗ = ^ ∙ sign G! (3)

[0119] Following Equation (3), the photothermal image is reconstructed. As shown in Fig.20(c) (middle, right), the processed image unveils the bulk chemical features in the PMMA particle’s center, which appear as null signals in the raw PTPG image.

[0120] Currently, OPTM measures only the 1D phase gradient signal, which introduces reconstruction artifacts. However, in the field of quantitative phase microscopy, such artifacts have been effectively removed by acquiring 2D phase gradients. A quantitative phase image can be accurately retrieved from 2D phase gradient measurements without computational artifacts. By integrating an additional split detector to capture the phase gradient along an orthogonal direction, it becomes feasible to reconstruct photothermal images that resolve bulk chemical features without computational artifacts (1).

[0121] Figs.21(a) through 21(c) illustrate exemplary dimensions of a split detector PD1, PD2, according to some exemplary embodiments. Fig.21(a) is an image of a printed circuit board (PCB). Fig.21(b) illustrates dimensions of the PCB. Fig.21(c) illustrates a 2× cross-sectional view of the cut line in Fig.21(b), illustrating the thickness profile of the PCB with electronics. Fig.21(d) includes an image of the split detector PD1, PD2 mounted on the PCB. The detector unit is covered with a copper mesh to isolate external electrical noise. Fig.21(e) includes an image of the detector unit mounted on the imaging objective using a 3D-printed holder. RL: reflective objective.

[0122] Figs.23(a) and 23(b) are schematic functional block diagrams of an oblique photothermal microcopy OPTM) microscope 100 and microscopy technique, according to alternative exemplary embodiments. Fig.23(a) is analogous to Fig.1(b), and Fig.23(b) is analogous to Fig.2(a). The alternative embodiment of Figs.23(a) and 23(b) differs from the embodiment of Figs.1(b) and 2(a) in that the split detector PD1, PD2 of the alternative embodiment of Figs.23(a) and 23(b) is below the sample instead of above the sample. That is, in the alternative embodiment of Figs.23(a) and 23(b), the sample is disposed between along the optical path of the microscopy system between the objective element and the split detector PD1, PD2, in contrast with the embodiment of Figs.1(b) and 2(a), in which the split detector PD1, PD2 is disposed along the optical path of the microscopy system between the objective element and the sample. The entirety of the current disclosure applies to the alternative embodiment illustrated in Figs.23(a) and 23(b), to the extent the disclosure is consistent with the alternative embodiment illustrated in Figs.23(a) and 23(b).

Claims

CLAIMS 1. An oblique photothermal microscopy system, comprising: a source of pulsed infrared light for generating a pulsed infrared excitation light beam for exciting a sample, such that the pulsed infrared light beam selectively heats the sample by absorption of the pulsed infrared light; a source of probe light for generating a probe light beam for illuminating the sample; a combining element for combining the infrared excitation light beam and the probe light beam into a combined beam and directing the combined beam along an optical axis of the microscopy system onto an objective element, the objective element generating a sample beam from the combined beam and directing the sample beam along the optical axis toward the sample; a split detector for being disposed in proximity to the sample, the split detector comprising a plurality of detection elements with at least one space being disposed between at least two adjacent detection elements of the plurality of detection elements, the sample beam being directed along the optical axis from the objective element through the space between two of the detection elements and striking the sample, the two detection elements receiving probe light scattered by the sample and generating signals indicative of probe light scattered by the sample; and a processor for processing the signals indicative of probe light scattered by the sample to generate an image of the sample.

2. The oblique photothermal microscopy system of claim 1, wherein the processing of the signals includes generating a sum of the signals.

3. The oblique photothermal microscopy system of claim 1, wherein the processing of the signals includes generating a difference between the signals.

4. The oblique photothermal microscopy system of claim 1, wherein the detection elements are disposed along the optical path of the microscopy system between the objective element and the sample.

5. The oblique photothermal microscopy system of claim 1, wherein the sample is disposed along the optical path of the microscopy system between the objective element and the detection elements.

6. The oblique photothermal microscopy system of claim 1, wherein the source of probe light comprises a laser.

7. The oblique photothermal microscopy system of claim 1, wherein the probe light has a wavelength of 532 nm.

8. The oblique photothermal microscopy system of claim 1, wherein the source of excitation light comprises a laser.

9. The oblique photothermal microscopy system of claim 1, wherein the excitation light has a wavelength in a range of 4.28 µm to 4.29 µm.

10. The oblique photothermal microscopy system of claim 1, wherein the excitation light has a wavelength in a range of 5.4 µm to 12.99 µm.

11. The oblique photothermal microscopy system of claim 1, wherein the excitation light has a wavelength in a range of 266 nm to 2400 nm.

12. The oblique photothermal microscopy system of claim 1, further comprising a scan unit disposed along the optical axis of the microscopy system between the combining element and the objective element for scanning the combined beam received from the combining element.

13. The oblique photothermal microscopy system of claim 1, wherein the combining element comprises a dichroic mirror.

14. An oblique photothermal microscopy method, comprising: generating a pulsed infrared excitation light beam for exciting a sample, such that the pulsed infrared light beam selectively heats the sample by absorption of the pulsed infrared light; generating a probe light beam for illuminating the sample; combining the infrared excitation light beam and the probe light beam into a combined beam and directing the combined beam along an optical axis of the microscopysystem onto an objective element, the objective element generating a sample beam from the combined beam and directing the sample beam along the optical axis toward the sample; providing a split detector for being disposed in proximity to the sample, the split detector comprising a plurality of detection elements with at least one space being disposed between at least two adjacent detection elements of the plurality of detection elements, the sample beam being directed along the optical axis from the objective element through the space between two of the detection elements and striking the sample, the two detection elements receiving probe light scattered by the sample and generating signals indicative of probe light scattered by the sample; and processing the signals indicative of probe light scattered by the sample to generate an image of the sample.

15. The oblique photothermal microscopy method of claim 1, wherein processing the signals includes generating a sum of the signals.

16. The oblique photothermal microscopy method of claim 1, wherein processing the signals includes generating a difference between the signals.

17. The oblique photothermal microscopy method of claim 1, wherein the detection elements are disposed along the optical path of the microscopy system between the objective element and the sample.

18. The oblique photothermal microscopy method of claim 1, wherein the sample is disposed along the optical path of the microscopy system between the objective element and the detection elements.

19. The oblique photothermal microscopy method of claim 1, wherein the source of probe light comprises a laser.

20. The oblique photothermal microscopy method of claim 1, wherein the probe light has a wavelength of 532 nm.

21. The oblique photothermal microscopy method of claim 1, wherein the source of excitation light comprises a laser.

22. The oblique photothermal microscopy method of claim 1, wherein the excitation light has a wavelength in a range of 4.28 µm to 4.29 µm 23. The oblique photothermal microscopy method of claim 1, wherein the excitation light has a wavelength in a range of 5.4 µm to 12.99 µm.

24. The oblique photothermal microscopy method of claim 1, wherein the excitation light has a wavelength in a range of 266 nm to 2400 nm.

25. The oblique photothermal microscopy method of claim 1, further comprising providing a scan unit disposed along the optical axis of the microscopy system between the combining element and the objective element for scanning the combined beam received from the combining element.

26. The oblique photothermal microscopy method of claim 1, wherein the combining element comprises a dichroic mirror.

Images