1. Introduction

As a nondestructive and accurate analyzing method, Raman spectroscopy has become more and more popular for measuring both organic and inorganic substances in many fields such as biology [1], chemistry [2,3], medicine [4,5], food science [6], geology [7,8], and industry [9]. The spatial heterodyne spectrometer (SHS) is a modern type of Fourier interferometer that offers multiple advantages including high throughput, high spectral resolution or a broad measurement range, no moving parts, and compactness [10–12]. The combination of the SHS and Raman spectroscopy is very promising for material identification and analysis and is widely used in research on chemical compounds [13–15], minerals [16–18], and biomarkers [19].

In the conventional SHS, there is an inverse relationship between the need for high spectral resolution and a broad spectral range due to the Nyquist sampling theorem [12]. For example, an SHS with a high spectral resolution of 2.47 cm⁻¹ and 1024 pixels can only achieve a spectral range of 1262.8 cm⁻¹ [18], which cannot measure a lot of characteristic Raman peaks of organic solutions beyond 2500 cm⁻¹ [20–22]. An SHS with a broad spectral range from 157 cm⁻¹ to 3249 cm⁻¹ and 1024 pixels can only achieve a spectral range of 3.86 cm⁻¹, which is not suitable for the measurement of mixtures of inorganic powders or organic solutions [23]. Replacing the conventional grating with a multi-grating in the SHS can break this inverse relationship, and corresponding experiments have been conducted [24–26]. However, the production of multi-gratings is somewhat difficult. First, when sub-gratings with different groove densities are ruled on a grating master plate, an error during the ruling process of one sub-grating requires scrapping the entire multi-grating. Second, the monitoring process of spliced grating production is also quite complicated [27]. Thus, it is not only time-consuming to rule a multi-grating on one grating master plate, but it is also very expensive.

In response to the issues mentioned above, we proposed the concept of a multi-Littrow-angle multi-grating (MLAMG) and conducted the measurement by using a spatial heterodyne Raman spectrometer (SHRS) based on an MLAMG. The MLAMG is made by splicing multiple sub-gratings with different Littrow angles, which is why we call the multi-grating produced in this way a multi-Littrow-angle multi-grating. Each sub-grating can be produced individually and is easy to mass-produce by grating replication. In addition, we can choose the mass-produced sub-gratings with the best performance to splice, so there is no need to worry about the problem of error in the production of the sub-gratings. In this study, each MLAMG in the SHS is made using four sub-gratings with same groove density and different Littrow angles. The basic principle and calibration theory of the MLAMG-SHRS are described in detail. In the experiment, the Raman spectra of the chemicals for different laser powers, integration times, contents, and containers are discussed in detail. The resulting Raman spectra of a mixture of inorganic powders and organic solutions are also presented. Raman spectra of minerals are also measured.

2. Principle

2.1 Basic theory

Figure 1 is the schematic diagram of the optical measurement layout of the MLAMG-SHRS, which can be mainly divided into two parts. In the excitation and collection optical path, the optical layout between the laser and the dichroic beam splitter is convenient for adjusting the laser power and focusing it on the sample. The excited Raman signal is collected by collimation lens L₁, filtered by the filters, and then converged by collimation lens L₂ into a fiber. The signal light is then transmitted to the MLAMG-SHRS via the fiber. The principal prototype of the MLAMG-SHRS integrates a collimation lens L₃, a beam splitter, two MLAMGs, imaging optics, and a charge-coupled detector (CCD). A collimation lens L₃ is placed at the focal distance from the optical fiber splice so that it can collimate signal light emitted from the fiber. The collimated incident light then enters the beam splitter and is divided into two light beams going to the gratings in the two light arms. The two gratings in the two light arms of the conventional SHS are replaced with MLAMGs, which consist of N_SG (N_SG ≥ 2) sub-gratings. In the MLAMG-SHRS we designed, N_SG = 4. The sub-gratings in the multi-grating can have different groove densities to achieve the broadening of the measurement band.

 

Fig. 1. Optical layout of the spatial heterodyne Raman spectrometer measurement based on a multi-Littrow-angle multi-grating.

Download Full Size | PDF

As shown in Fig. 2, since we designed each sub-grating with different Littrow angles α_Lj (j = 1, 2, 3, 4), the groove densities of all sub-gratings can be the same, which greatly simplifies the manufacturing of the multi-grating. At the same time, it is worth mentioning that the directions of the groove lines on all sub-gratings should be parallel, and the rotation position should be at the center of the grating so as to guarantee the positions of the zero-path difference of each sub-grating are the same to the maximum extent. Each of the sub-gratings SG_ij (i = 1, j = 1, 2, 3, 4) is arranged along the direction of the groove line with a rotation angle Δα, and the rotation angle Δα_j is same for both corresponding sub-gratings SG_1j and SG_2j in the two multi-gratings.

 

Fig. 2. Schematic diagram of the multi-Littrow-angle multi-grating: (a) main view; (b) top view.

Download Full Size | PDF

The relationship between rotation angle and Littrow angle for every two adjacent sub-gratings can be expressed as

Analyzing the definition of the resolution limit yields the following expression for the theoretical spectral resolution δ_σj of each sub-grating SG_ij:

2.2 Calibration theory

The performance parameter of the MLAMG-SHRS in the experimental breadboard system can be estimated using a calibration procedure based on a standard calibration light source with known wavelengths (e.g., a mercury lamp and sodium lamp). The calibration process is achieved by matching the spatial frequency in the Fourier transform of the two-dimensional interference fringe with the corresponding wavelength. For the two pairs of sub-gratings SG_1j and SG_2j used for calibration, the Littrow wavelength and Littrow wavenumber can be calculated from

3. Experiment

3.1 Breadboard

The optical layout of the experimental breadboard system is illustrated in Fig. 3, while Table 1 provides a list of the essential parameters of all the commercially-available optical components utilized in constructing the experimental breadboard system. The experimental breadboard can be divided into two parts. The first part is the Raman signal excitation and optical collection system. The power of the solid-state 532 nm green laser (Changchun New Industries Optoelectronics Tech. Co., Ltd) can be adjusted linearly from 0 to 400 mW with the linear variable filter. During the experiments, the sample was positioned at the focal plane of a 25-mm-diameter collimation lens L₁ with a 15 mm focal length. One 700 nm short-pass filter (84-714, Edmund) was used for filtering the ambient light and the fluorescent light at wavelengths higher than 700 nm, and two 532 nm long-pass edge filters (532-LAB-80AC-2.5, CNI) were used for laser line rejection. The optical fiber splice of the 1.5-mm-diameter silica fiber was positioned at the focal plane of a 25-mm-diameter collimation lens L₂ with a 75 mm focal length to transfer signal light.

 

Fig. 3. Breadboard system of the spatial heterodyne Raman spectrometer based on a multi-Littrow-angle multi-grating.

Download Full Size | PDF

Table 1. Essential parameters of the optical elements used in the experimental breadboard system

View Table

The second part is the experimental prototype of the MLAMG-SHRS. The core of the MLAMG-SHRS was constructed using a 50.8 × 50.8 × 50.8 mm³ cubic beam splitter (model no. 20BC17MB.1, Newport), eight reflective diffraction sub-gratings with 300 grooves/mm and a size of 40 × 9 × 12 mm³ (Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences), and two fixing frames for the sub-gratings with four different inclination grooves each to produce different rotation angles (Changchun UP Optotech (Holding) Co., Ltd). The grating facet of one MLAMG in the MLAMG-SHRS was tilted by a small angle to generate two-dimensional fringes. The collimation lens L₃ had a 200 mm focal length and was positioned so that the optical fiber splice was at its focal plane. A CCD with a 13.3 × 13.3 mm² sensor size (iKon-M 934, Andor) and a camera lens (SP AF 60 mm F/2 Di II Macro 1:1, Tamron) were used to record the two-dimensional fringe images. The CCD was cooled to −80 °C to reduce the thermal noise within the sensor chip.

The flat fielding correction was carried out in conjunction with the calibration process [28]. The Fourier transforms of the two-dimensional fringe images were executed through a 2-D fast Fourier transform (FFT) function. For the Fourier-transform spectroscopy, the phase correction method [29] and wavelet threshold function [30] were utilized to accurately recover the spectral information.

3.2 Calibration

The standard light source used in the calibration was a mercury lamp. Figure 4(a) shows the interference fringes of the mercury lamp produced by the MLAMG-SHRS. We used the interferogram on each sub-grating SG_i3 for the calibration. From Fig. 4(b) and 4(c), the two spatial frequencies of the interference fringes are 496 and 417, and the corresponding two known spectral lines of the mercury lamp are 576.961 nm and 579.067 nm. Equations (13)–(15) give a Littrow wavelength λ_L3= 590.4432 nm, spectral resolution δ_σ3 = 0.7979 cm^-1, and Littrow angle α_L3= 5.0811° for each sub-grating SG_i3. The imaging width of the CCD, W_I, is then calculated to be 35.237 mm.

 

Fig. 4. (a) Interference fringes of the mercury lamp produced by the MLAMG-SHRS. (b) Spatial frequency obtained from the FFT of the interference fringes on SGi3. (c) Measured spectrum on SGi3 of the mercury lamp after calibration.

Download Full Size | PDF

Using α_L3 and the rotation angles shown in Table 1, the Littrow angles of the other sub-gratings can be calculated as α_L1= 4.6346°, α_L2= 4.8527°, and α_L4= 5.3234°. Using the imaging width with Eq. (7) and Eq. (8), we calculate the spectral resolution of the other sub-gratings as δ_σ1 = 0.8752 cm⁻¹, δ_σ2 = 0.8357, and δ_σ4 = 0.7614 cm⁻¹. The overall spectral range was from −215.08 cm⁻¹ to 3019.20 cm⁻¹ (i.e., the wavenumber of the 532 nm laser was set as the zero wavenumber). The calibration results of the experimental breadboard system were close to the designed MLAMG-SHRS performance.

4. Results and discussion

4.1 Raman spectrum analysis of sulfur with different integration times

Figure 5(a) shows the recovered Raman spectra of solid sulfur with different integration times at the same laser power of 20 mW. As shown in Fig. 5(a), the Raman signals at 153 cm⁻¹, 218 cm⁻¹, and 472 cm⁻¹ were well measured, and were assigned to the dominant vibration of the S₈ molecule as the antisymmetric bond-bending mode, the symmetric bond-bending mode, and the symmetric bond-stretching mode, respectively [31]. A relatively low laser power of 20 mW was used in the experiment to study the difference in the vibrational Raman spectrum at various integration times. From Fig. 5(a), we can see that the three peaks were still visible when the integration time was 1 s, and became clearer as the integration time increased.

 

Fig. 5. (a) Recovered Raman spectra of solid sulfur with different integration times at the same laser power of 20 mW. (b) Measured SNR of sulfur versus integration time.

Download Full Size | PDF

The signal performance of the spectrum measured by the MLAMG-SHRS here can be quantified using the signal-to-noise ratio (SNR) given by

4.2 Raman spectrum analysis of calcium carbonate powder and methanol with different laser powers

Figure 6(a) shows the recovered Raman spectra of calcium carbonate (CaCO₃) powder with different laser powers at the same integration time of 60 s. The strongest Raman peak signal at 1085 cm⁻¹ corresponds to the symmetric stretching vibration v₁ (CO₃^2-), and the Raman peak signals at 711 cm⁻¹ and 281 cm⁻¹ correspond to the vibration mode v₄ (CO₃^2-) and the bending mode v₅ (CO₃^2-) [32], respectively. The Raman peak signal at 155 cm⁻¹ is the corresponding lattice mode [33]. The spectrum around 711 cm⁻¹ is enlarged in Fig. 6(a), and the lowest Raman peak of calcium carbonate is clearly visible when the laser power is higher than 60 mW. Other Raman peaks that are higher are still clearly visible when the laser power is 20 mW, even though the peak value decreases.

 

Fig. 6. (a) Recovered Raman spectra of methanol with different laser powers at the same integration time of 40 s. (b) Measured SNR of methanol versus laser power.

Download Full Size | PDF

Figure 6(b) shows the measured SNR of the characteristic Raman peak of calcium carbonate at 1085 cm⁻¹ for different laser powers at the same integration time of 40 s. Figure 6(b) shows that the SNR grows almost linearly with increasing laser power at the integration time of 40 s.

Figure 7(a) shows the recovered Raman spectra of methanol with different laser powers at the same integration time of 40 s. The Raman peak of methanol at 1038 cm⁻¹ belongs to the =CO stretching vibration mode, the relatively weak Raman peak at 1454 cm⁻¹ belongs to the =CH₃ deformation vibration mode, and the two high Raman peaks at 2835 cm⁻¹ and 2945 cm⁻¹ correspond to the symmetric and antisymmetric stretching vibration modes of the =CH₃ group, respectively [22]. All the vibrational Raman peaks are clearly visible from 20 mW to 200 mW at the same integration time of 30 s.

 

Fig. 7. (a) Recovered Raman spectra of methanol with different laser powers at the same integration time of 30 s. (b) Measured SNR of methanol versus laser power.

Download Full Size | PDF

Figure 7(b) shows the measured SNR of the characteristic Raman peak of methanol at 2835 cm⁻¹ for different laser powers at the same integration time of 30 s. Figure 7(b) shows that the SNR grows almost linearly with increasing laser power at the integration time of 30 s, while the SNR of methanol increases more slowly than that of calcium carbonate. The above Raman spectrum and SNR analyses of calcium carbonate and methanol demonstrate that the MLAMG-SHRS has the capability to detect inorganic compounds and organic solutions.

4.3 Raman spectrum analysis of suspensions with different titanium dioxide contents

Figure 8(a) shows the recovered Raman spectra in the range from 0 cm⁻¹ to 800 cm⁻¹ for suspensions of titanium dioxide with different contents for a laser power of 180 mW and the same integration time of 10 s. The 144 cm⁻¹, 197 cm⁻¹, and 639 cm⁻¹ Raman peaks are assigned to the E_g modes. The Raman peak at 144 cm⁻¹ is assigned to v₆(E_g), which is very intense and sharp, and the Raman peaks at 197 cm⁻¹ and 639 cm⁻¹ correspond to v₅(E_g) and v₁(E_g), respectively. The Raman peaks at 397 cm⁻¹ and 515 cm⁻¹ correspond to v₄(B_1g) and v₂(B_1g), respectively. v₁(E_g) and v₂(B_1g) correspond to the Ti–O bond stretching vibration, and v₅(E_g), v₆(E_g), and v₄(B_1g) correspond to the O–Ti–O bending vibrations [34].

 

Fig. 8. (a) Recovered Raman spectra of titanium dioxide for different contents with a laser power of 180 mW and the same integration time of 10 s. (b) Measured SNR of titanium dioxide versus content.

Download Full Size | PDF

In Fig. 8(a) the characteristic Raman peak of the suspension of titanium dioxide at 144 cm⁻¹ is clearly visible and grows higher from 3.2 g/L to 160 g/L. The highest value at 144 cm⁻¹ is in the spectrum of the powder, and its SNR is 440. Figure 8(b) shows the measured SNR of the Raman peak at 144 cm⁻¹. The measured SNR grows rapidly when the content of the titanium dioxide grows from 3.2 g/L to 96 g/L. When the content is larger than 96 g/L, the measured SNR starts fluctuating and is much smaller than the SNR of the powder of the titanium dioxide. The Raman spectrum and the corresponding SNR of the suspension of titanium dioxide analyzed above demonstrate the capability of the MLAMG-SHRS in environmental monitoring applications.

4.4 Raman spectrum of potassium sulfate powder and acetone in different containers

Figure 9(a) and (b) shows the Raman spectra of potassium sulfate (K₂SO₄) and acetone in different containers under the same laser power and integration time, respectively. All spectra of potassium sulfate were measured at a laser power of 100 mW and an integration time of 30 s, and all spectra of acetone were measured at a laser power of 200 mW and an integration time of 20 s. The thickness of the plastic bag was 0.2 mm, the thickness of the glass bottle was 1.5 mm, and the thickness of the plastic bottle was 2 mm. The plastic bottle and plastic bags used in the experiment were both made of polyethylene (PE).

 

Fig. 9. (a) Recovered Raman spectra of potassium sulfate stored in different containers with a laser power of 100 mW and integration time of 30 s. (b) Recovered Raman spectra of acetone stored in different containers with a laser power of 200 mW and integration time of 20 s.

Download Full Size | PDF

In Fig. 9(a), the Raman vibration bands of potassium sulfate are obvious. The main Raman peak corresponds to the symmetric stretching mode v₁ (SO₄^2-) centered at 992 cm⁻¹. The Raman peak at 449 cm⁻¹ is assigned to the bending vibration v₂ (SO₄^2-), and the Raman peaks at 619 cm⁻¹ and 627 cm⁻¹ are assigned to the bending vibration v₄ (SO₄^2-). Both the v₂ and v₄ modes are detected in the low-wavenumber region [35]. The Raman peaks at 1106 cm⁻¹ and 1145 cm⁻¹ are assigned to the anti-symmetric stretching mode v₃ (SO₄^2-), and the Raman peaks at 1240 cm⁻¹ and 1254 cm⁻¹ are assigned to 2v₄ (SO₄^2-) [36]. Both the v₃ and 2v₄ modes are detected in the high-wavenumber region. The Raman peak at 111 cm⁻¹ is assigned to the lattice mode. In the recovered Raman spectrum of potassium sulfate in a plastic bag, the Raman peaks at 2848 cm⁻¹ and 2885 cm⁻¹ correspond to methylene (CH₂) groups in the PE [37]. The measured SNRs of the main Raman peaks in the glass bottle, plastic bag, and plastic bottle are 467, 397, and 321, respectively.

In Fig. 9(b), the Raman vibration bands of acetone are obvious. The Raman peak at 530 cm⁻¹ is assigned to C–(=O)–C in plane deformation. The Raman peak at 786 cm⁻¹ is assigned to C–C(=O)–C symmetric stretching. The Raman peak at 1428 cm⁻¹ is assigned to asymmetric CH₃ deformation vibration (CH₃–C(=O–)). The Raman peak at 1706cm⁻¹ is assigned to C = O stretching. The Raman peaks at 2848 cm⁻¹ and 2924 cm⁻¹ are assigned to CH₃ symmetric stretching. The Raman peaks at 2966 cm⁻¹ and 3006 cm⁻¹ are assigned to CH₃ asymmetric stretching [38]. The measured SNRs of the highest Raman peaks at 2924 cm⁻¹ in the glass bottle, plastic bag, and plastic bottle are 175, 188, and 112, respectively.

While the potassium sulfate and acetone were in different containers, the recovered Raman peaks remained clearly visible. There are no significant differences among the spectrums and spectral resolutions in different containers under the same laser power and integration times. However, the intensity and SNR of the Raman spectrums in the plastic bottle are lower than those in other containers owing to the higher reflection of the laser. From the above analysis of the spectra of potassium sulfate and acetone in different containers, the MLAMG-SHRS has the capability to detect the Raman spectra of the targets in the plastic bag, glass bottle and plastic bottle, which will be very helpful in the detection of hazardous and toxic chemicals that have been sealed in transparent containers.

4.5 Raman detection of a mixture of inorganic powders and organic solutions

Figure 10(a) shows the recovered Raman spectra of sodium sulfate (Na₂SO₄), sulfur, titanium dioxide (TiO₂), and a mixture of these inorganic solid powders with a laser power 120 mW and integration time of 30 s. The main Raman peaks were well detected. In the recovered Raman spectrum of the mixture, the Raman characteristic peaks are mainly distributed in the spectral range of the sub-grating SG_i1.

 

Fig. 10. (a) Recovered Raman spectra of a mixture of inorganic powders and of single inorganic powders with a laser power of 120 mW and integration time of 30 s. (b) Recovered Raman spectra of a mixture of organic solutions and of single organic solutions with a laser power of 200 mW and integration time of 15 s.

Download Full Size | PDF

The Raman peaks at 449 cm⁻¹, 467 cm⁻¹, 620 cm⁻¹, 632 cm⁻¹, 647 cm⁻¹, 993 cm⁻¹, 1101 cm⁻¹, 1131 cm⁻¹, and 1152 cm⁻¹ in the spectrum of the mixture correspond to the spectrum of sodium sulfate [39]. Some weak peaks of sodium sulfate in the mixture assigned to the v₂ mode centered at 449 cm⁻¹ and the v₄ mode between 610 cm⁻¹ and 650 cm⁻¹ are not easy to identify owing to their proximity to the Raman peaks of other inorganic powders. The Raman peaks assigned to the v₁ mode at 993 cm⁻¹ and the v₃ mode between 1100 cm⁻¹ and 1160 cm⁻¹ are distributed in the spectral range of the sub-grating SG_i2 and are all clearly visible. The Raman peaks at 153 cm⁻¹, 218 cm⁻¹, and 472 cm⁻¹ in the spectrum of the mixture correspond to the spectrum of sulfur. The Raman peaks at 144 cm⁻¹, 197 cm⁻¹, 397 cm⁻¹, 515 cm⁻¹, and 639 cm⁻¹ in the spectrum of the mixture correspond to the spectrum of titanium dioxide. The Raman peak at 144 cm⁻¹ for titanium dioxide and the Raman peak at 153 cm⁻¹ for sulfur are well distinguished in the recovered Raman spectrum of the mixture.

Figure 10(b) shows the recovered Raman spectra of carbon tetrachloride (CCl₄), acetone, cyclohexane, and a mixture of these organic solutions with a laser power of 200 mW and integration time of 15 s. The main Raman peaks were well detected. In the recovered Raman spectrum of the mixture, the spectral ranges of all sub-gratings are used to measure the Raman characteristic peaks.

The Raman peaks at 218 cm⁻¹, 314 cm⁻¹, 452 cm⁻¹, 762 cm⁻¹, and 789 cm⁻¹ in the spectrum of the mixture correspond to the spectrum of carbon tetrachloride [40], and the high-intensity Raman peaks of the carbon tetrachloride distributed in the spectral range of sub-grating SG_i1 are well measured in the spectrum of the mixture. The Raman peaks at 786 cm⁻¹, 1428 cm⁻¹, 1706cm⁻¹, 2848 cm⁻¹, 2924 cm⁻¹, and 3006 cm⁻¹ in the spectrum of the mixture correspond to the spectrum of acetone. The Raman peaks at 803 cm⁻¹, 1030 cm⁻¹, 1268 cm⁻¹, 2853 cm⁻¹, 2923 cm⁻¹, and 2938 cm⁻¹ in the spectrum of the mixture correspond to the spectrum of cyclohexane. Many Raman peaks of the acetone and the cyclohexane are very close to each other. By distinguishing the intensity differences between the strongest characteristic Raman peak at 2924 cm⁻¹ of acetone and the strong characteristic Raman peaks at 2853 cm⁻¹, 2923 cm⁻¹, and 2938 cm⁻¹ of cyclohexane distributed in the spectral range of sub-grating SG_i4, we still can easily identify the acetone and the cyclohexane in the mixture. The Raman peaks at 762 cm⁻¹ and 789 cm⁻¹ of carbon tetrachloride, the Raman peak at 786 cm⁻¹ of acetone, and the Raman peak at 803 cm⁻¹ of cyclohexane are overlapped with each other, but this does not prevent us from analyzing the composition of the organic solution according to the strong Raman characteristic peaks in the spectrum of the mixture. The Raman spectrum analyses of the mixture, single organic solution, and single inorganic powder demonstrate that the MLAMG-SHRS has the capability to detect the mixture of organic solutions and inorganic powders.

4.6 Raman spectra of minerals

Figure 11 shows the recovered Raman spectra of rose quartz, celestine, and calcite. The celestine is mainly composed of cesium sulfate (SrSO₄). The Raman peak at 122 cm⁻¹ is assigned to the M–O₁₂ vibration, and the Raman peak at 451 cm⁻¹ is assigned to the v₂ mode. The Raman peaks at 619 cm⁻¹ and 654 cm⁻¹ are assigned to the v₄ mode. The Raman peak at 1000 cm⁻¹ is assigned to the v₁ mode. The Raman peaks at 1110 cm⁻¹, 1159 cm⁻¹, and 1185 cm⁻¹ are assigned to the v₃ mode [41]. The main Raman characteristic peaks of rose quartz at 126 cm⁻¹, 204 cm⁻¹, and 452 cm⁻¹ and calcite at 155 cm⁻¹, 281 cm⁻¹, 711 cm⁻¹, and 1085 cm⁻¹ are also well detected [42]. The measured SNRs of the main Raman peaks of rose quartz, celestine, and calcite are 237, 384, and 449, respectively. The minerals we measured are essential for many modern industries. The detection performance for these minerals demonstrates the MLAMG-SHRS has wide application potential in geochemical and environmental measurements and industrial fields.

 

Fig. 11. Recovered Raman spectra of minerals: rose quartz with a laser power of 80 mW and integration time of 50 s; celestine with a laser power of 100 mW and integration time of 15 s; calcite with a laser power of 100 mW and integration time of 30 s.

Download Full Size | PDF

5. Summary and conclusions

In this paper, we described the optical layout and corresponding experiment for a broadband, high-resolution spatial heterodyne Raman spectrometer based on an MLAMG. The designed MLAMG-SHRS provides a new, convenient method for producing multi-gratings to realize broadband, high-resolution Raman spectral measurement by splicing sub-gratings with different Littrow angles but the same groove density, which greatly decreases the difficulty in ruling the multi-grating. In the designed MLAMG-SHRS, the MLAMG consists of four sub-gratings with a groove density of 300 gr/mm and Littrow angles of 4.6355°, 4.8536°, 5.0820°, and 5.3253°. Each of the four sub-gratings is fixed by a fixing frame to make it more convenient to adjust the optical layout of the proposed spectrometer.

From the experimental results for sulfur, calcium carbonate, methanol, and the suspension of titanium dioxide, we can see that the MLAMG-SHRS is capable of detecting chemicals with different integration times, laser powers, containers, and contents, making it suitable for environmental monitoring and the detection of hazardous chemicals. The MLAMG-SHRS is also capable of analyzing the Raman spectrum of a mixed target with high spectral resolution and meeting the requirement of a sufficiently broad spectral range, as evidenced by the detection of Raman spectra in a mixture of inorganic powders and organic solutions. Minerals such as celestine, rose quartz, and calcite are also well detected. All the experiments and advantages mentioned above show that the MLAMG-SHRS is suitable for broadband spectral measurements at high resolution in a wide range of potential applications.