Analysis on the terahertz spectral characteristics of two types of sialic acids in solid state and aqueous solution Authors: Zhuang Peng, Yang Gao, Xinrui Zhang, Huiyu Yang, Bo Su, Kai Li, Cunlin Zhang Published: 2025 ABSTRACT AbstractIn this study, N-acetylneuraminic acid (Neu5Ac) and N-glycolylneuraminic acid (Neu5Gc) were chosen as representative sialic acids to study their terahertz (THz) spectral characteristics in solid and aqueous states. Firstly, the structures of the two sialic acids were determined and analyzed in detail using powder X-ray diffraction (PXRD) laying the foundation for the subsequent simulation calculations. Secondly, the THz time-domain spectroscopy (THz-TDS) system was used to measure their spectral properties in solid state. Thirdly, the molecular structures of the two sialic acids were simulated using Material Studio (MS) software, and their vibrational modes were analyzed using solid-state density functional theory (DFT). Then, to compare the THz spectral characteristics in solid and aqueous states, microfluidic technology was introduced to measure the THz spectra of the sialic acids in aqueous solutions. The results indicate that there are differences in the THz spectra of the two sialic acids in their solid and aqueous solution states. Finally, THz spectroscopy was conducted on a common edible sialic acid powder, revealing that the absorption coefficient spectrum of the powder exhibited characteristic absorption peaks of Neu5Ac. This study provides a reference for the analysis of biomolecules.Keywords: THz time-domain spectroscopy, Density functional theory, Sialic acid, Microfluidic chipSubject terms: Biophysics, Optics and photonics, Physics Introduction Sialic acids are found in the saliva of humans and animals and represent a crucial class of biomolecules with various physiological functions. They are located on the glycoproteins on cell surfaces and play a key role in intercellular communication, participating in biological processes such as cell adhesion and immune responses1. Furthermore, sialic acids have significant functions in the nervous system, where they are involved in communication between nerve cells and in the development of neurons2. Sialic acids are also widely utilized in the biomedical field, serving as cell surface markers and drug delivery vehicles. Studying the structure and function of sialic acids is essential for understanding intercellular interactions, the mechanisms of disease development, and developing new drugs3. Neu5Ac and Neu5Gc are two common sialic acids. Neu5Ac is involved in the normal physiological processes of the human body. In contrast, Neu5Gc is present in the normal physiological processes of certain mammals and is not present in large amounts in the normal human body. Research has demonstrated that Neu5Gc is taken up by human tumor cells through macropinocytosis and transported to the cytoplasmic compartment via lysosomal transporters, thereby promoting the growth of tumor cells and ultimately exerting an adverse impact on human health4. THz waves refer to electromagnetic radiation with frequencies between 0.1 and 10 THz and wavelengths ranging from 30 to 3000 μm5. This frequency range lies between microwaves and infrared waves, belonging to the far-infrared band, and has broad application prospects in the study of material properties, high-speed communication, non-destructive imaging, and security inspections6. With the continuous development of THz detection technology, using THz spectroscopy to study the properties of biochemical samples has become an important method in this field7. The photon energy of THz waves is relatively low, at only 4 meV, which does not ionize samples, allowing for non-destructive testing of materials8. The photon energy of THz waves matches the rotational and vibrational energy levels of many organic substances, making it possible to obtain characteristic absorption spectra of these samples using THz spectroscopy for non-destructive quantitative and qualitative analysis9. Additionally, many biological macromolecules exhibit specific absorption and scattering characteristics in the THz frequency range, creating new avenues for the detection of biochemical substances10. PXRD refers to the diffraction patterns obtained by subjecting a sample to X-ray diffraction. These diffraction patterns result from the interaction between the incident X-rays and the atoms within the sample, reflecting the structure of the sample. PXRD spectra contain a series of diffraction peaks, each corresponding to a specific diffraction angle of a particular structure within the sample. By analyzing the position, intensity, and shape of these diffraction peaks, information about the structure, orientation and phase of the sample can be obtained. For instance, Li et al. used PXRD to detect metal gluconates and confirmed the polycrystalline nature of gluconate samples11. Most biomolecules need to be in an aqueous environment to exert their biological activity. However, using THz technology to study biological samples in aqueous solutions often faces challenges. This is primarily because water strongly absorbs THz waves, making it difficult for THz technology to detect biological samples in aqueous solutions12. To overcome this limitation, microfluidic chip technology was introduced in this study. Microfluidic technology involves controlling the flow of fluids with volumes ranging from picoliters to nanoliters within low-dimensional channel structures that have at least one dimension in the micrometer or even nanometer scale13. Microfluidic chip technology offers many advantages, such as high throughput, low consumption, portability, integration, ease of operation, and sample savings. It is expected to become a promising detection method in the future14. In this study, we used copolymer of cycloolefin (COC), which has high THz wave transmittance, to fabricate the detection area of the chip. The outer frame of the chip was made using PMMA. We successfully prepared a THz microfluidic chip with visible light transparency, enhancing the utilization of COC material. Additionally, the chip has the advantages of easy manufacturing, strong sealing, convenient cleaning, and reusability, greatly facilitating experimental operations. This study selected two commonly sialic acids, Neu5Ac and Neu5Gc, to investigate their THz spectral characteristics. Firstly, the THz spectra in their solid state were measured, and the structures of the two sialic acids were analyzed using PXRD technique. Secondly, molecular dynamics simulations were performed, and through data processing, the absorption peak information of the samples was calculated. The absorption characteristics of their THz spectra and the mechanisms of absorption peak formation were investigated from the perspective of molecular vibrations. Then, the THz spectral characteristics of these two sialic acids in aqueous solutions were measured using microfluidic chips and compared with their solid-state THz absorption spectra. It was found that there are differences between them. Finally, a common edible sialic acid powder was analyzed using THz spectroscopy. After data processing, the absorption coefficient spectrum was obtained and compared with that of Neu5Ac. The results showed that the absorption peaks of both spectra were very similar, demonstrating that THz spectroscopy can effectively identify composition of substance. This study integrates PXRD, microfluidic technology, and THz-TDS to investigate the spectral characteristics of two sialic acids, ensuring the accuracy of the research findings. The results demonstrate new prospects and significant value for the application of THz technology in the detection of biomolecules such as sialic acids. Materials and experiments Experimental samples The chemical structures of Neu5Ac and Neu5Gc are shown in Fig. 1. Neu5Ac was purchased from Jiangsu Aikang Biopharmaceutical, with a molecular formula of C11H19NO9, a molecular weight of 309.27, and a density of 1.6 g/cm3. Neu5Gc was purchased from Shanghai Aladdin Biochemical Technology LTD, with a molecular formula of C11H19NO10, a molecular weight of 325.269, and a density of 1.9 g/cm3. The samples and polyethylene (PE) were first weighed using a precision balance at a mass ratio of 1:5. Then, the weighed samples and PE were thoroughly ground into a fine powder using an agate mortar. Finally, the ground powder was poured into a pellet die and placed in a tablet press, where the sample was pressed with a 6-ton force for 3 min, resulting in a disc with a thickness of approximately 1 mm and a diameter of about 12 mm. Molecular structure diagrams of two samples (a) Neu5Ac (3D), (b) Neu5Gc (3D), (c) Neu5Ac (2D), (d) Neu5Gc(2D). THz-TDS system This study utilized a self-constructed THz-TDS system with an effective bandwidth range of 0.2–2.2 THz. In the unloaded state, the terahertz time-domain spectroscopy system was used to measure the time-domain spectra when acquiring unloaded terahertz signals and blocked terahertz signals (background noise). Their frequency-domain spectra were obtained through fast Fourier transform (FFT). Further, the power spectral ratio (SN) between the system signal and background noise—i.e., the signal-to-noise ratio (SNR) spectrum—was calculated. By determining the two intersection points where the SNR spectrum equals 20 dB, the frequencies at these intersection positions, denoted as "f₁" and "f₂", can be used to define the spectral range of the system. The system is composed of a femtosecond laser, a THz generation device, a delay device, and a THz detection device. The laser source used in this experiment is the "self-mode-locked fiber femtosecond laser" independently developed by Peking University, with a central wavelength of 1550 nm, output power of 130 mW, pulse width of 75 fs, and pulse repetition frequency of 100 MHz. We employed SketchUp 2025 (https://www.sketchupchina.com) to create the experimental optical path, as shown in Fig. 2. The femtosecond pulse laser is split into two beams after passing through a polarization beam splitter (PBS). One beam serves as the pump pulse and is coupled into a fiber-coupled photoconductive antenna (BATOP bPCA-100-05-10-1550-c-f) via a delay control system to generate THz waves. The other beam serves as the detection pulse, which is converged by a lens and then coupled into a fiber-coupled photoconductive antenna (BATOP bPCA-180-05-10-1550-c-f) to detect THz waves. The solid samples or microfluidic chips are fixed between off-axis parabolic mirrors. As the THz waves are focused and collimated through off-axis parabolic mirrors and then pass through the sample, they carry information about the measured object. This information is subsequently received by the detection antenna and transmitted to a lock-in amplifier for amplification. Finally, the data is collected and processed by a computer. To ensure the accuracy of the experiment, the entire process is conducted in a nitrogen-filled environment at room temperature, with a relative humidity of 10%. THz-TDS system. Fabrication of microfluidic chip For THz microfluidic chips, material transparency in the THz band and processability are critical. Polydimethylsiloxane (PDMS), although widely used in conventional microfluidics, exhibits low THz transmittance, which may distort spectral data during THz time-domain spectroscopy measurements. Polyethylene (PE), while transparent to THz waves, is opaque to visible light, making real-time optical observation of liquid sample behavior within the chip channels impossible—a fundamental requirement for monitoring dynamic processes such as chemical reactions or fluid flow patterns. In contrast, cycloolefin copolymer (COC) offers both high THz transparency to minimize signal attenuation and visible-light transparency for direct optical inspection. Additionally, its chemical inertness ensures minimal interaction with various sample substances, preserving the integrity of experimental results. However, due to the high cost of COC sheets, we adopted a hybrid approach: using COC exclusively in the detection area where THz-sample interaction occurs, while employing polymethyl methacrylate (PMMA) for the remaining structural components. PMMA is cost-effective, easily machinable, and achieves a balance between performance and cost. We first used a laser engraver to cut PMMA sheets into 40 × 40 × 2 mm thin slices. A 20 × 20 × 2 mm square hole was precisely cut in the designated sample detection area to form a cavity for accommodating the COC detection module. Leveraging the dot-matrix function of the laser engraver, we drilled two 10-mm-long, 0.7-mm-diameter cylindrical channels on the left and right sides of the PMMA slice to serve as the main inlet and outlet for liquid samples. To facilitate multi-directional fluid control and potential sample mixing, 0.7-mm-diameter vertical channels were added at the top left and right, vertically connecting to the main cylindrical channels. The COC material was then cut to fit exactly into the square hole of the PMMA base. Using a high-precision milling machine, a 20 × 8 × 0.025-mm detection area was carved at the center of two COC pieces, which were then bonded with acrylic glue to form a 0.05-mm-deep liquid channel. This dimension was carefully selected to optimize THz wave penetration and interaction with the sample while minimizing signal loss. During final assembly, the COC detection module was embedded into the square hole of the PMMA base. To ensure airtightness and prevent sample leakage, the interfaces between COC and PMMA, as well as the side holes of the channels, were sealed with hot melt glue. The left L-shaped channel of the PMMA was designated as the input channel to guide samples into the chip, while the right L-shaped channel served as the output channel for controlled sample removal. The fabrication process of the THz microfluidic chip is illustrated in Fig. 3. Injecting the sample solution into the chip and allowing it to stand confirmed no liquid leakage, validating the effectiveness of our sealing method and the structural integrity of the chip. This ensures reliable and consistent THz measurements. Preparation process of THz microfluidic chip. Calculation method for absorption coefficient The unique information of biomolecules is often reflected in optical parameters such as the THz absorption coefficient and refractive index. Therefore, it is necessary to establish a parameter extraction model for the measured THz-TDS signals to calculate the corresponding THz absorption coefficient, refractive index, and other optical parameters. This also has important practical significance for the application of THz-TDS detection systems. The method used in this study to calculate the absorption coefficient is similar to that of Liu et al.15. The complex refractive index of a material is generally represented by: where is the refractive index of the sample, representing the dispersive properties of sample and is the extinction coefficient of the sample, representing the absorptive properties of sample. is the angular frequency. Before testing the sample, the THz time-domain signal of pure PE is measured. This signal, denoted as , serves as the reference. Next, the sample to be tested is placed in the optical path, and the THz time-domain signal passing perpendicularly through the sample is measured, denoted as , which is the sample signal. According to the Lambert–Beer law, the reference signal in the frequency domain can be expressed as: where c is the speed of light in a vacuum. is the THz frequency-domain signal generated by the system when the THz wave propagates through a uniformly distributed medium (air in this study) and l is the distance the THz wave travels in the uniformly distributed medium. In practical measurements, terahertz waves undergo multiple internal reflections within the sample due to the Fabry–Perot effect, leading to oscillatory tails in the time-domain signal. To eliminate this artifact, a time-domain truncation method was applied during data processing. This approach involves selecting an appropriate time window containing the main transmitted pulse while excluding subsequent reflected signals. The truncated time-domain signal—comprising only the initial transmission pulse—was then converted into the frequency domain via Fourier transform. Based on this procedure, the optical parameters of the sample in the frequency domain could be accurately derived. Thus, under the assumption of neglecting Fabry–Perot effects, the sample signal in the frequency domain can be expressed as: where d is the actual thickness of the sample. T12 is the transmission coefficient of the THz wave as it enters the sample from the external medium (air in this study) and T21 is the transmission coefficient of the THz wave as it propagates from the sample into the external medium. The complex transmission coefficient in the frequency domain can be expressed as the ratio of the sample signal to the reference signal in the frequency domain. Considering that the external medium is air, the complex refractive index of the reference signal can be set to 1, and the complex transmission coefficient is given by: The complex transmission coefficient can also be expressed in another form: where is the amplitude ratio between the sample signal and the reference signal and is the phase difference between the sample signal and the reference signal. If the extinction coefficient of the sample is much smaller than its refractive index , under this condition, by combining formulas (1) and (4) and converting the result into the form shown in Eq. (5), we can obtain: by combining formulas (7) and (8), we can obtain: The extinction coefficient and absorption coefficient satisfy the following relationship: Finally, according to formulas (10) and (11), the absorption coefficient of the detection sample can be obtained: Powder X-ray diffraction (PXRD) This experiment utilized an XRD-6000 X-ray diffractometer (Rigaku Co., Japan), which employs a Cu Kα (λ = 1.5418 Å) radiation source. The operating conditions of the X-ray tube were set to a tube voltage of 40 kV and a tube current of 30 mA. The diffraction angle range was set from 2° to 50°, with a scanning speed of 10°/min. During measurement, Neu5Ac and Neu5Gc were evenly spread on a glass slide to ensure a flat surface. Results and discussion PXRD analysis of two solid sialic acids We utilized PXRD technology to analyze the samples, with the experimental measurements PXRD spectra presented in Fig. 4. The spectra reveal distinct diffraction peaks, indicating that the crystal structure of the samples is highly ordered, with no significant impurities or doping. According to the measured PXRD pattern, the relevant parameters of the substance can be obtained, which will provide help for the subsequent study of the THz spectral characteristics of Neu5Ac and Neu5Gc and the simulation calculation. PXRD spectra (a) Neu5Ac, (b) Neu5Gc. THz spectra of two solid sialic acids Figure 5a displays the normalized THz absorption spectrum of solid Neu5Ac at room temperature, revealing two absorption peaks at 1.58 and 2.07 THz. Figure 5c shows that Neu5Gc also has two absorption peaks at 1.40 and 1.94 THz. These data demonstrate the unique spectral characteristics of Neu5Ac and Neu5Gc in the THz range, revealing their molecular properties. Figure 5b and d present the refractive index spectra of Neu5Ac and Neu5Gc, respectively. The refractive index of Neu5Ac ranges from 1.48 to 1.50, while the refractive index of Neu5Gc ranges from 1.47 to 1.49. The refractive index at the locations of the absorption peaks decreases with increasing frequency, exhibiting anom-alous dispersion, which is consistent with the Kramers–Kronig relations16. The THz absorption spectrum of solid Neu5Ac is similar to the results measured by Guan et al.17, both showing the same absorption peak at 1.58 THz. In addition, the measurement range of this paper is wider, and a new absorption peak at 2.07 THz is obtained. (a) The absorption spectrum of Neu5Ac. (b) The refractive index spectrum of Neu5Ac. (c) The absorption spectrum of Neu5Gc. (d) The refractive index spectrum of Neu5Gc. Simulation analysis of THz spectra of two solid sialic acids To further investigate the THz spectral characteristics of Neu5Ac and Neu5Gc and to elucidate the complex interactions between THz spectra and molecular vibrations, we performed structural and energy optimizations of Neu5Ac and Neu5Gc molecules using the CASTEP module in MS software (Materials Studio 2020). The unit cell parameters of Neu5Ac were obtained from the CCDC, with the CSD ID code 1102953 (https://www.ccdc.cam.ac.uk/structures/Search?Compound=Nacetylneuraminic%20acid&DatabaseToSearch=Published), while the unit cell of Neu5Gc was self-constructed. Based on density functional theory, the molecular structures, energies, and frequencies of Neu5Ac and Neu5Gc were optimized using the Generalized Gradient Approximation (GGA) and Perdew-Burke-Ernzerhof (PBE) parameters. In the calculations, the plane-wave cutoff energy was set as 571.4 eV and Brillouin zone sampling of electronic states was performed on a 6 × 6 × 6 Monkhorst–Pack grid18. Figure 6 shows the comparison between the experimental and simulated normalized absorption peaks of Neu5Ac. From the simulation results, the simulated absorption peaks of solid Neu5Ac were found to be at 1.58 and 2.09 THz. The experimental absorption peaks of solid Neu5Ac were found to be at 1.58 and 2.07 THz. The experimental and simulated absorption peaks are very similar in position. Comparison of absorption peaks of Neu5Ac (a) Experiment, (b) Simulation. The vibrational modes of Neu5Ac at different THz bands are shown in Fig. 7a and b, with Fig. 7c illustrating the crystal structure of Neu5Ac. The absorption peak at 1.58 THz is attributed to the rotation of –CH3, –OH and –COOH groups, as well as the wagging of =O and –OH. The absorption peak at 2.09 THz is due to the rotation of –CH3, –OH and –COOH groups, as well as the wagging of =O and –OH. Vibrational modes of Neu5Ac at different THz bands (a) 1.58 THz, (b) 2.09 THz. The crystal structure of Neu5Ac (c). Figure 8 compares the experimental and simulated normalized absorption peaks of Neu5Gc. It was found that the positions of the absorption peaks are quite similar. From the simulation results, the simulated absorption peaks of solid Neu5Gc were found to be at 1.40 and 1.94 THz. The experimental absorption peaks of solid Neu5Gc were found to be at 1.40 and 1.97 THz. The positions of the simulated and experimental absorption peaks correspond closely. Comparison of absorption peaks of Neu5Gc (a) Experiment, (b) Simulation. The vibrational modes of Neu5Gc at different THz bands are shown in Fig. 9a and b, with Fig. 9c illustrating the crystal structure of Neu5Gc. The absorption peak at 1.40 THz is attributed to the rotation of –CH2OH, –OH and –COOH groups, as well as the wagging of –OH. The absorption peak at 1.97 THz is due to the rotation of –CH2OH and –OH groups, as well as the wagging of –CH2OH, –COOH and –OH. Vibrational modes of Neu5Gc at different THz bands (a) 1.40 THz, (b) 1.97 THz. The crystal structure of Neu5Gc (c). The absorption peak positions of Neu5Ac and Neu5Gc in both simulated and experimental results correspond closely, but the minor differences between the simulated and experimental results can be explained by several factors. Firstly, the calculations were based on ideal crystal structures, whereas actual experimental measurements are subject to various interfering factors. Secondly, the absorption peaks of THz radiation can be influenced by intermolecular interactions, particularly hydrogen bonding19. These interactions might not be fully captured in the simulations, leading to slight variations in the absorption peaks. We have consolidated the vibrational assignments of the two sialic acids within the THz frequency range into Table 1, providing a clearer understanding of the unique vibrational modes of these substances in the THz spectrum. Table 1THz vibration modes of Neu5Ac and Neu5Gc.SampleExperimental result (f/THz)Theoretical simulation (f/THz)Vibration mode attributionNeu5Ac1.581.58r(–CH3, –OH, –COOH), w(–OH, =O)2.072.09r(–CH3, –OH, –COOH), w(–OH, =O)Neu5Gc1.401.40r(–CH2OH, –OH, –COOH), w(–OH)1.941.97r(–CH2OH, –OH), w(–OH, –CH2OH, –COOH)The abbreviations adopted for the vibrations are r(rotation) and w(wagging). THz vibration modes of Neu5Ac and Neu5Gc. The abbreviations adopted for the vibrations are r(rotation) and w(wagging). THz spectra of two sialic acid solutions Water exhibits strong absorption in the THz range20, which poses significant challenges for spectral measurements. Therefore, in this study, microfluidic technology was employed to precisely control the liquid thickness at 50 μm, thereby reducing water absorption. The solid samples were first dissolved in deionized water to create a saturated solution, and then the supernatant was injected into a microfluidic chip. The absorption coefficients of the two sialic acid solutions were measured at room temperature. Figure 10 presents the normalized absorption spectra of the two sialic acid solutions. It can be observed that the absorption of THz waves by sialic acid solutions differs from that by solid sialic acid, with the overall absorption being higher for sialic acid solutions. This indicates that sialic acid molecules are influenced by water molecules in the solution, resulting in differences between the absorption spectra of the solution and the solid state. These differences are attributed to the weakened intra- and intermolecular interactions in the solution state. Additionally, the hydroxyl groups and other functional groups of sialic acid form hydrogen bonds with water molecules, increasing the absorption in the THz range, thus leading to greater absorption for the solution. Additionally, the absorption peaks of the two sialic acid solutions differ, with the Neu5Ac solution exhibiting an absorption peak at 2.04 THz, while the Neu5Gc solution did not show obvious absorption peaks, indicating that the absorption of terahertz by the aqueous solution affected the experimental results. It is worth mentioning that in Guan’s previous study, the absorption spectrum of the Neu5Ac solution in the 0.2–1.2 THz band was measured, and no obvious absorption peak was found, which is consistent with our experimental results. We measured the absorption peak of the Neu5Ac solution at 2.04 THz, providing a new idea for the qualitative analysis of sialic acid content21,22. Absorption spectra of (a) Neu5Ac solution, (b) Neu5Gc solution. THz spectra of salivary acid powder We randomly procured a common edible sialic acid powder. Figure 11 presents the normalized THz absorption spectrum of this powder. The absorption peak in the spectrum is consistent with that of Neu5Ac, and no absorption peak similar to that of Neu5Gc is observed. This indicates that the main component of the powder is Neu5Ac and does not contain Neu5Gc. Absorption spectra of salivary acid powder. Conclusion This study utilized THz-TDS systems and DFT methods to systematically detect and deeply analyze the vibrational absorption of the two sialic acids, Neu5Ac and Neu5Gc. To ensure the repeatability and accuracy of the experiment, we took the average of the three measurement results, calculated the standard deviation, and presented them in the form of mean plus error bars. Experimental results show that these sialic acid molecules exhibit obvious characteristic absorption peaks in the terahertz range, corresponding to the abnormal dispersion phenomenon of their refractive index distribution. The characteristic peaks of Neu5Ac appear at 1.58 THz and 2.07 THz, while those of Neu5Gc appear at 1.40 THz and 1.94 THz. Furthermore, to investigate the spectral characteristics of Neu5Ac and Neu5Gc in greater depth and elucidate the complex interactions between THz spectra and molecular vibrations, we obtained optimized molecular structures and vibrational properties through simulation calculations. The high consistency between experimental results and simulation calculations demonstrates the accuracy of the experiments and the relevance of the calculations. The vibrational modes of the two sialic acids were analyzed in detail. In addition, we used microfluidic chip technology to determine the spectral characteristics of the two sialic acids in solution and conducted a comparative analysis between their solid and solution states. The study found that the absorption of THz waves by sialic acid solutions differs from that of solid sialic acid, with the overall absorption being higher in the solution. The intra- and intermolecular interactions of Neu5Ac and Neu5Gc are weakened under the influence of water molecules, and the hydroxyl groups and other functional groups of sialic acid form hydrogen bonds with water, leading to changes in the absorption peaks. Additionally, the Neu5Gc solution did not show obvious absorption peaks, indicating that the absorption of terahertz by the aqueous solution affected the experimental results. Finally, THz spectroscopy was used to analyze the main components of the randomly purchased sialic acid powder. The positions of the absorption peaks indicate that its primary component is Neu5Ac. Through the study of Neu5Ac and Neu5Gc samples, the THz spectral characteristics of sialic acid were elucidated. This work provides a novel perspective and value for THz technology in the field of biological molecules detection. Acknowledgements The author would like to thank the national Natural Science foundation of China (NSFC) (61575131), National Key R&D Program of China (2021YFB3200100) and Natural Science Foundation of Beijing Municipality (4232066) for their funding. Author contributions Conceptualization, Z.P.; Methodology, Z.P. and Y.G.; Validation, H.Y.; Formal analysis, X.Z.; Resources, X.Z.; Data curation, Z.P. and Y.G.; Writing—original draft, Z.P.; Writing—review & editing, B.S., K.L. and C.Z. All authors have read and agreed to the published version of the manuscript. Funding This research was funded by National Natural Science Foundation of China (NSFC) (61575131), the National Key Research and Development Program of China (2021YFB3200100) and the Natural Science Foundation of Beijing Municipality (4232066) for their funding. Data availability Data underlying the results presented in this paper are not publicly available at the time of publication, but may be obtained from the corresponding author (Email Address: subo75@cnu.edu.cn) upon reasonable request. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement Data underlying the results presented in this paper are not publicly available at the time of publication, but may be obtained from the corresponding author (Email Address: subo75@cnu.edu.cn) upon reasonable request.