By means of the quantum chemical modeling within the ORCA 5.03 software package with the HF3c/MINIS/MINIS1+1(d)(Cl)/def2-SV(P)ECP(Pt) level of theory, the electronic structure and binding energy of adducts on the basis of carboplatin, fullerenol, quinine and its esters, as well as their three-component systems were studied. The performed analysis of the total energies of systems and calculated diagrams of the energies of the highest occupied and lowest vacant molecular orbitals of both the initial components and the molecular ensembles they form allowed identifying the most probable combinations of them in terms of stability. Features of synergistic effects have been indicated and prospects for using the three-component system carboplatin–fullerenol C₆₀(OH)₂₄–quinine (or quinine esters) during chemotherapy in oncological practice have been outlined as well.

Spectral methods for the analysis of equilibrium and kinetic characteristics of complexation processes of the known photosensitizer meta-tetrakis(hydroxyphenyl)chlorine (mTHPC) with monomeric and polymeric derivatives of β-cyclodextrin (β-CD) have been developed. The study of binding isotherms showed that mTHPC has a higher affinity for the polymeric derivatives of β-CD - carboxymethyl-β-cyclodextrin-polymer (CM-β-CDPD) and β-cyclodextrin-polymer (β-CDPD) – than for monomeric methyl-β-cyclodextrin (M-βCDMD). Profiles of changes in the relative content of mTHPC inclusion complexes with β-CDs in solution and in the presence of lipid vesicles were obtained, and dissociation constants of PS molecules were calculated from the composition of inclusion complexes with cyclodextrins. It was shown that the dissociation process of mTHPC from inclusion complexes with M-β-CDMPD takes 1–2 minutes, and with polymeric CDs (CM-β-CDMPD and β-CDMPD) – more than 1 hour. The results suggest that the peculiarities of complexation with polymeric and monomeric cyclodextrins may play an essential role in the delivery of the photosensitizer to cellular/tissue structures. 

The optical absorption spectra of (E)-2-(2-aryl-1-cyanovinyl)-4-cyclopropylthiazole derivatives, with different substituents in the para-position of the aryl ring, have been studied. The measurements were carried out in dimethylformamide solutions. It is shown that the acceptor substituents introduction leads to a bathochromic shift of absorption. This is explained by the fact that they promote the delocalization of electrons of unoccupied pairs of nitrogen and sulfur of the thiazole ring. Absorption bands in the range of 250—450 nm arise from HOMO→LUMO and HOMO-1→LUMO transitions. 

Hydrophilic polyurethane-urea films were obtained using cycloaliphatic diisocyanates and nonionic surfactants such as ethoxylated alcohols and alkyl phenols. The effect of the selected modifiers on the phase structure of the obtained materials was estimated using the IR-ATR spectroscopy method. 

For the first time, the phenomena of light emission during the sonolysis of liquid ammonia and solutions of cerium and silver salts in it at – 70 °C were considered. The presence of a Ce³⁺ luminescence band with a maximum of 400 nm in the spectrum of moving single-bubble sonoluminescence generated by ultrasound in the solution of 5 * 10^-4 M Ce(SO₄)₂ in ammonia was discovered. This band was found in addition to the bubble luminescence continuum usual for liquids (240—800 nm in ammonia). This band is quenched in the presence of 10^-2 M AgNO₃ in the solution. This indicates the suppression of Ag⁺, the electron acceptor, of the probable reaction of sonochemiluminescence Ce⁴⁺ + e_s^– →η(Ce³⁺)* + (1– η)Ce³⁺, (Ce³⁺)*→Ce³⁺ + hν, responsible for the detected Ce³⁺ glow. The concentration of Ce⁴⁺ in the solution decreases when it is reduced by this reaction, and the intensity of sonochemiluminescence in the Ce³⁺ band decreases during sonolysis, but reaches a constant level a few minutes after its start. This is explained by the achievement of dynamic equilibrium between the rates of reduction and oxidation reactions of cerium ions: Ce⁴⁺→Ce³⁺ and Ce³⁺→Ce⁴⁺, respectively. The one-electron oxidizing agent in the last reaction is presumably the NH₂ radical, another primary product of ammonia sonolysis, in addition to e_s^– . The formation of Ce³⁺ during sonolysis of a Ce(SO₄)₂ solution was confirmed by registration of Ce³⁺ photoluminescence in it. The excitation yield in the sonochemiluminescence reaction was estimated to be η = 2 * 10^-6. 

Phononic band gaps of one-dimensional phononic crystals based on porous anodic aluminum oxide were numerically simulated. The spectral position of phononic band gaps of anodic aluminum oxide impregnated with various liquids is determined. Air and liquid filled nanopores display markedly different phonon dispersion relations. Calculations of the dispersion characteristics of samples under study were carried out, the group velocity of phonons, as well as their effective mass, were determined. 

Quartz from pegmatite mining wastes in northern Karelia (Russia) were studied using inductivelycoupled plasma atomic emission spectroscopy, IR-spectroscopy, optical and scanning electron microscopy. The study is aimed to reveal the main technological characteristics of quartz that determine the possibility of its utilization as a promising raw material for the high-purity quartz concentrate production. It is shown that mineral inclusions in quartz are mainly represented by feldspars, muscovite, calcite, iron oxides, borosilicates, which is confirmed by the trace element composition of quartz dominated with Al, Ca, K, Na. Along with molecular water, which is part of fluid inclusions, quartz contains OH-complexes associated with structural impurities of Al, Li and B. As a result of unsorted quartz processing, including electromagnetic separation, microwave exposure and chemical leaching, a quartz concentrate with a total impurity content of 160 ppm was obtained. High-purity quartz concentrate production with total impurity content of <50 ppm on quartz grit with the fineness of 0.5–0.1 mm turned out to be difficult, due to the presence of hard-to-remove micron-sized mineral inclusions and structural impurities. 

By the method of high-frequency repetitively pulsed f ~ 10–12 kHz laser radiation with wavelength λ = 1.064 μm and power density q = 33 MW/cm²  on the piezoceramics PZT-19 target at pressure in the vacuum chamber p = 2.2 * 10^–2 mm Hg nanostructured thin films on a silicon and glass substrates have been obtained. The morphology of thin piezoceramics films was studied using atomic force microscopy. Transmission spectra of the obtained films were investigated in the visible, near and mid-IR regions. The electrophysical characteristics of piezoceramics on Si structures are analyzed. 

The results of investigating the spectral and luminescent characteristics of optical ceramics from thulium and holmium coactivated yttrium sesquioxide (Tm,Ho:Y₂O₃) are presented. The ceramic samples were fabricated using solid-state vacuum sintering of nano-sized particles with a complex chemical composition synthesized by laser evaporation. The transmittance of Tm,Ho:Y₂O₃ ceramics in the region of 2 μm exceeds 82 % indicating their high optical quality. The redistribution of line intensities in the luminescence spectra of ³F₄→³H₆ transitions of Tm³⁺ ions and ⁵I₇→⁵I₈ transitions of Ho³⁺ ions observed under excitation of Tm³⁺ ions to the ³H₄ level in ceramics with different Tm³⁺/Ho³⁺ balances indicates the occurrence of a nonradiative energy transfer process from ³F₄ level of Tm³⁺ ions to ⁵I₇ level of Ho³⁺ ions. 

A comparative study of the structural and phase state of Ti–Al–N systems obtained in the form of a coating and a bulk material is carried out. The methods of X-ray diffraction, energy-dispersive spectroscopy (EDS) and Raman spectroscopy (RS) are used to characterize the structure. The use of EDS and RS makes it possible to verify the X-ray diffraction data on the phase composition of the samples, which is especially important in the study of multiphase Ti–Al–N coatings, and to identify the features of the surface microstructure of the coating. The Ti–Al–N coating is obtained by chemical deposition of Ti and Al in an N₂ atmosphere on a Ti substrate with subsequent annealing in vacuum at 700, 800, 900 and 1000 ºС. The bulk Ti–Al– N sample is obtained by reaction sintering of Ti, Al, TiN powders in vacuum at 1200 and 1300ºС. Ti₂AlN MAX phase appears in the coating at a lower temperature than in the bulk sample and is characterized by lower thermal stability. Ti–Al–N coating is characterized by a greater multiphase nature; after annealing in vacuum at 900 ºС, the following phases are registered in it: Ti₂AlN, Ti₄AlN₃, TiN, Ti₂N, AlN. The destruction of the MAX phase structure occurs at 1000 ºС. In the bulk sample after annealing in vacuum at 1300ºС, the main phase is Ti₂AlN with a small admixture of TiN and TiAl, the destruction of the Ti₂AlN MAX phase occurs at 1400 ºС. 

Carbon-palladium nanocomposite films with a thickness of ~140 nm with a palladium nanoparticle size of up to 2 nm were synthesized by the method of ion sputtering of composite targets. The optical properties of carbon-palladium composite films in the visible spectrum region of were studied. It has been established that the optical absorption of these nanocomposite films is described by the characteristic Tauc dependence. In this case, the presence of palladium nanoparticles leads to a sharp decrease in the optical gap in composite films, which is explained by an increase in the average size of graphene nanoclusters in the carbon matrix. 

The switching of lasing from the ground state to the excited state with current increasing in InGaAs/GaAs lasers with quantum well-dots (QWDs) and quantum wells with different waveguide widths has been studied. It is shown that, due to the peculiarities of the density of states in QWDs, lasers with waveguide of 0.45 μm wide do not switch to the lasing mode through an excited state at current densities up to 48 kA/cm² , in contrast to lasers with quantum wells. Generation through an excited state in QWD lasers occurs when the waveguide width increases up to 0.78 μm and is accompanied by switching from the fundamental to the 2nd order optical mode. 

The approach is developed to describe the optical properties of layered systems (multilayers) consisting of monolayers of spherical particles in a light-absorbing medium. It takes into account multiple scattering of waves in individual monolayers and multiple reflections between monolayers. The amplitude transmission and reflection coefficients of the selected layers of light-absorbing medium containing monolayers of particles are determined in the framework of the recently developed method based on utilizing the quasicrystalline approximation. They are used to calculate coefficients of transmission and reflection of multilayer in the framework of the transfer matrix method. The calculation results are presented for the absorption coefficient of weakly absorbing medium layer containing 1, 4, 8, and 32 monolayers of titanium oxide (TiO₂) particles under the incidence of a plane electromagnetic wave along the normal, as applied to the enhancement of the photocatalytic reactions in photochemical microreactors. 

Based on the analysis of high-resolution FTIR-spectra recorded at the St. Petersburg State University atmospheric monitoring station in 2009–2022, the possibility of long-term monitoring of NH3 total column in the atmosphere from the results of ground-based measurements of direct solar radiation in the mid-IR range is demonstrated. Estimates of random and systematic uncertainties of the NH₃ total column are evaluated. The results of 14-year FTIR monitoring of NH₃ in the atmosphere are presented, and a comparison with data from independent satellite and ground-based measurements is made. 

The mixing process in Rudner–Levitov chains, which are a one-dimensional lattice of bosonic modes with designed losses in every second mode, has been studied. It is shown that by choosing the initial excitation in the system in a special way, it is possible to achieve a significant reduction in the mixing time. For an initial state localized in one mode, it is necessary to choose the eigenvector of the effective Hamiltonian, that corresponds to a non-zero eigenvalue of the effective Hamiltonian with the smallest absolute value of the imaginary part. The initial state, localized in two or more modes, must be chosen so that it is orthogonal to this eigenvector. 

A generalized theoretical model of the coherent interaction of linearly polarized (1+1)D light beams in a cubic optically active photorefractive Bi₁₂SiO₂₀ cut crystal (1‾ 1‾ 0) placed in an external constant electric field is presented, which includes, in addition to the traditional electro-optical effect, the inverse piezoelectric and photoelastic effects. The interaction of identically linearly polarized one-dimensional light beams with a transverse Gaussian distribution of the light field at the entrance to the crystal has been studied. It is shown that taking into account additional parameters and effects in the mathematical model makes significant changes to the obtained theoretical results, which will allow for a more correct interpretation of experimental data. Numerical modeling of the interaction of driven and leading Gaussian light beams with mutually orthogonal linear polarizations at the entrance to the crystal has been performed. 

An improved method for determining the temperature of a laser diode and the thermal resistance of the main elements of an equivalent thermal circuit based on measuring transient temperature-sensitive characteristics of direct voltage at the p-n junction in response to a step-like effect of a heating current pulse is proposed. The individual components and the total thermal resistance of the laser diode were experimentally studied and analyzed, the latter was ~7.4 K/W. It was found that the main contribution to the total thermal resistance is made by the laser crystal layer itself from the p-n junction to the lower plane (~2.8 K/W) and the AlN-switching thermal conductive substrate (~2.6 K/W), for which no further reduction paths are visible. At the same time, it was shown that without significant overheating ΔT < 40 K, a continuous mode of operation is realized with 6 times exceeding the threshold I ≈ 2 A and generation power P ≈ 2.5 W, efficiency ~30 % and differential quantum output η ≈ 60 %. 

A focus depth optimization method was proposed for signal enhancement and reproducibility improvement in pump-probe collinear double-pulse femtosecond laser-induced breakdown spectroscopy. This method was based on the optimization of the focus depth of the second pulse. The spectral signal intensity showed a stable enhancement with an enhancement factor of ~3, and the signal reproducibility exhibited minimal fluctuations when the focus depth of the second pulse was within the range of 0–2 mm below the sample surface. Additionally, the possible mechanisms behind the observed signal enhancement and improvement in reproducibility were discussed based on the measured plasma temperature and electron density. 

Accurate resonance energies and quantum defects for various Rydberg series originating from the 3s² 3p⁵  (² P°_3/2) ground and 3s² 3p⁵  (² P°_3/2) metastable states of K III ions are reported up to n = 30. Theoretical photoionization is performed in the framework of the screening constant per unit nuclear charge (SCUNC) method. The SCUNC predictions, believed to be the first calculations, agree very well with the only available advanced light source (ALS) measurements. Some of the overlap peaks in the ALS measurements are clearly identified via the present SCUNC calculations. The present work demonstrates the strength of the SCUNC formalism to assist experimenters in the analysis of measurements from synchrotron radiation. New data are tabulated for n = 15–30. 

Hydroxyapatite (HA) has been successfully coated on alkaline-treated titanium (NaOH-treated Ti) through an electrochemical method using a solution containing Ca(NO₃)₂ ⋅ 4H₂O and NH₄H₂PO₄ at 80°C. Pure Ti substrates were immersed in 5M NaOH solution at 80°C for 24 h. XRD results confirmed the formation of an HA layer on the surface of the Ti substrates. The SEM result has shown that the structure of the HA layer contains a ribbon-like morphology with a diameter of ~120 nm. HA coating enhanced the hydrophilic properties compared with the surface of bare Ti substrates. Moreover, HA coating showed excellent biocompatibility as demonstrated by cell attachment in vitro at 72 h of culturing, which is a promising coating method of HA on NaOH-treated Ti for titanium implants. 

Recently Zinc sulphides (ZnS) have been widely used in research due to their distinct catalytic properties. These investigations have demonstrated that the quick electron-hole pair production caused by photoexcitation makes ZnS nanocrystals effective photocatalysts. When ZnS is prepared using a surfactant, a change of phase is observed from ZnS-cubic to wurtzite. Ethylene diamine is used as a surfactant for the preparation of ZnS in a hydrothermal method. The prepared sample was evaluated for photophysical characteristics by X-ray diffraction method, XPS, and UV/Vis spectra. Under UV light, the photocatalytic activity of ZnS with and without surfactant is accessed for photocatalytic decomposition of different dyes. When comparing the photocatalytic activity of ZnS with and without surfactant, ZnS with surfactant exhibits greater catalytic activity. 

Nitrogen-doped carbon quantum dots (N-CQDs) were prepared by a one-step hydrothermal method using urea, aniline, and ethylenediamine as raw materials and nitrogen sources which exhibited good fluorescence performance. Spectral characteristics of the samples were analyzed using UV spectrophotometry, fluorescence spectrophotometry, and Fourier transform infrared spectroscopy. Optimal conditions, such as pH, reaction time, and ion concentration were investigated for N-CQDs. Under the optimal conditions (40 μL N-CQDs, buffer solution at pH 6 with a reaction time of 30 min), it was observed that cobalt ions caused a significant fluorescence quenching effect on the prepared N-CQDs, as determined by fluorescence spectrophotometry. The spiked recovery rates ranged from 95.8 to 104.6%, indicating that the N-CQDs could serve as a fluorescent probe for detecting the content of Co(II) in the actual environment. This method could be applied to the monitoring of heavy metal pollution in mine groundwater, the detection results of Co(II) using this method were very similar to ICP-MS, indicating that this method had good application prospects.

Nickel oxide nanoparticles (NiO) were produced using the sol-gel dip-coating method to create NiO/glass and silicon thin film. Fourier transform infrared spectroscopy, ultraviolet-visible spectroscopy, and scanning electron microscopy were used to examine the film’s structural and optical properties. The nonlinear optical characteristics of the nickel oxide thin film were examined using the open aperture and close aperture Z-scan methods at wavelengths of 532 and 635 nm. Manipulating the laser optical intensities and wavelengths altered the refractive index of the samples in that increasing the laser intensities and wavelengths led to a decrease in the nonlinear refractive index. The nonlinear absorption coefficient exhibited a negative correlation with laser intensity and a small decreasing relationship with the wavelength. 

The diagnosis of lung cancer has always been a challenging clinical issue. In this work, we use laserinduced breakdown spectroscopy (LIBS) combined with machine learning to differentiate samples of lung cancer tumors from those of normal tissues. Sample plasma was collected by laser ablation at 1064 nm to obtain the characteristic spectra of lung tumor and normal tissue samples. Twelve lines of C, Mg, Ca, C-N, Na, and K were selected for the diagnosis of malignancy. Principal component analysis (PCA), support vector machine (SVM), k-nearest neighbors (KNN), Decision Tree, and Bagged Tree were used to establish the discrimination model for tumors and normal tissue. A 10-fold cross-validation method was used to evaluate the discrimination model. The results showed that the integrated learning Bagged Tree model performed best, with an overall accuracy of 98.9%, sensitivity and specificity of 98.6 and 99.3%, respectively, and an area under the curve (AUC) of 0.982. This study suggests that LIBS can be used as a fast and accurate means of identifying human lung tumors. 

Zr and Cl double-doped carbon dots (Zr/Cl-CDs) were prepared by a one-step hydrothermal method using citric acid, urea, and ZrCl₄ as precursors. Zr/Cl-CDs were characterized by ultra-violet-visible spectroscopy (UV-Vis), transmission electron microscopy, Fourier transform infrared spectroscopy, X-ray diffraction, and X-ray photoelectron spectroscopy (XPS). Experimental analysis revealed that the average size of Zr/Cl-CDs was analyzed experimentally to be 2–3 nm, while XPS analysis revealed the presence of C-Cl and Zr-O bonds. The fluorescence quantum yield of the final CDs was 17.5% at the excitation of 360 nm. In addition, Zr/Cl-CDs exhibited excellent performance in the detection of rutin with a concentration of 30–110 and 100–160 μmol/L, respectively. Linear correlation coefficients (R² ) were 0.9931 and 0.9962. The underlying mechanism was that rutin had an excellent quenching effect on the emission of the Zr/Cl-CDs as the fluorescence intensity exhibited a good linear relationship between rutin concentration and fluorescence intensity. These findings provide a new strategy for synthesizing double-doped CDs for rutin detection. 

The aim of this study was to develop an ultraviolet spectrophotometry-based analysis method for determining naringin in bulk solutions, as well as to validate it for linearity, precision, accuracy, LOD, and LOQ. The study devised and validated a procedure to accurately detect naringin in a bulk solution using a comparatively economical, more reliable, and efficient technique. We are expecting similar results with comparable accuracy to the existing expensive techniques but with minimized risks. We utilized pure naringin (API) and methanol (solvent) to develop the method and further used water, isopropyl alcohol, chloroform, ethyl acetate, butan-1-ol, mustard oil, neem oil, citronella oil, tea tree oil, and lemongrass oil to check its solubility. The calibration curve obtained for naringin using the developed method was found to be robust (standards relative standard deviation (%RSD) = 1.54–1.95% at 1 nm difference in wavelength); exhibited linearity (R²  = 0.9974); had a precision of %RSD 1.02–2.89% (intra-day) and 1.09–3.4% (inter-day); and an accuracy revealed by the % recovery range 98.43–107.62% at three levels, i.e., 80, 100, and 120%. These values are comparable to those obtained for HPLC, LC-MS, and LC. The solubility test was carried out using this developed method, and naringin proved to be the most soluble in methanol (10.470±0.03 mg/mL) and in tea tree oil (4.332±0.015 mg/mL) among the polar and nonpolar solvents. The method is applied for naringin solubility determination. 

The development of simple to use, cost-effective, accurate, and precise UV spectrophotometry is the objective of this study. Validation characteristics such as specificity, linearity, accuracy, precision, limit of detection, limit of quantification, and robustness were evaluated in accordance with ICH Q2(R1) guidelines. The drug was measured to have the maximum absorbance at a wavelength of 310 nm. The solution remains stable for up to 12 h. The equation y = 0.1724x – 0.0443, with an R²  of 0.9995, was used to demonstrate the linear response for tafamidis concentration ranges of 2–12 μg/mL. It was found that the accuracy ranged from 100–103%. The intraday and interday precision for repeatability fell within the acceptable parameters (<2%). The technique’s limit of detection was 0.112 μg/mL and the limit of quantification was 0.34 μg/mL. The proposed validated method can be applied to standard quality control purposes as well as tafamidis analysis in pharmaceutical doses form. 

Cefpodoxime proxetil (CFP) is a third-generation cephalosporins group antibiotic which is an antibacterial beta-lactam agent that is used in the treatment of acute otitis, pharyngitis, tonsillitis and lower respiratory tract infections (e.g. pinomonia and bronchitis). A simple, fast and sensitive spectrophotometric method was developed for the determination of cefpodoxime proxetile in bulk and pharmaceutical preparations. The proposed method was based on the charge transfer interactions of cefpodoxime proxetil as n-electron donor with 7,7,8,8-tetracyanoquinodimethane (TCNQ) as the π-acceptor. In addition, the described method was validated in accordance with the requirements of ICH guidelines. The absorbance was measured at 596 nm. The parameter of linearity, LOD, LOQ, accuracy, precision and recovery were also evaluated. The assay was linear over the concentration range of 3.0–15.0 µg/mL. The correlation coefficient was obtained 0.9997 for CFP. LOD and LOQ values were found to be 0.036 and 0.108 µg/mL, respectively. Recovery values were obtained ranging from 99.97 to 101.3% according to the standard addition method. The performance of the proposed method was evaluated in terms of the student’s t-test and variance ratio Ftest to find out the significance of the proposed methods over the reference spectrophotometric method. The developed method was successfully applied for estimation of CFP in a pure form and as a pharmaceutical preparate.

Based on paraxial and Wentzel-Kramers-Brillouin (WKB) approximations, we investigated density transition-based self-focusing of asymmetric finite Airy–Gaussian (AiG) beams in plasma. The relativistic nonlinearity in the dielectric constant of plasma is considered. The electric field distribution of asymmetric finite AiG beams is expressed in terms of beam-width parameters f₁ and f₂ with respective modulation parameters a and b in transverse dimensions of the beam. The coupled differential equations governing the behavior of f₁ and f₂ with dimensionless propagation distance ξ are established from the parabolic equation approach. The effect of the exponential density ramp profile on the self-focusing of asymmetric finite AiG beams is presented. The sensitiveness of asymmetry in a and b parameters of finite AiG beam and exponential density ramp of plasma is specifically highlighted. 

The present communication explores the impact of the self-focused q-Gaussian laser beam on the second harmonic generation (SHG) in collisional plasma. In the case of collisional plasma, the generation of density gradients takes place in a perpendicular direction due to carrier redistribution on account of nonuniform heating. Due to density gradients, electron plasma wave (EPW) is produced at pump beam frequency. EPW interacts with the pump beam, producing SHG. The well-known Wentzel-Kramers-Brillouin approximation and paraxial theory are used for deriving 2^nd-order ODE for the beam waist of the main beam and the efficiency of 2^nd harmonics. A numerical solution of equations is carried out by using the supreme variables of laser and plasma. It is observed from analysis that the beam’s focusing ability and the efficiency of the 2^nd harmonics are enhanced with the increase in laser intensity, density of plasma electrons, radius of beam and q-values. 

The extraction of feature variables can significantly reduce the dimension, simplify the features, and improve the accuracy of quantitative models, which is of great significance in spectral data preprocessing using laser-induced breakdown spectroscopy (LIBS). A feature variable extraction method based on convolutional neural network was proposed. The LIBS spectral data was subjected to multilayer two-dimensional convolution through the series connection of the convolution structure and an InceptionV2 module, and the multi-level features were gradually extracted to find the optimal feature variable combination. Finally, the prediction results were obtained by the fully connected layer. In order to verify the applicability of the extracted features, the extraction results of the convolutional neural network were directly input into random forest to construct a quantitative model. In this paper, the concentration of K element in the mixed solution prepared by the laboratory was tested. The determination coefficients R²  of the CNN-RF training set and the test set reached 0.993 and 0.990, respectively. The root mean square error (RMSEC) of the training set and the root mean square error (RMSEP) of the test set were 0.0067 and 0.0084 wt.%, respectively, and the average relative error reached 8.533%. The evaluation parameters were significantly better than the extraction results of least absolute shrinkage and selection operator, principal component analysis and SelectKBest. The results show that the convolutional neural network can effectively extract LIBS characteristic spectral lines and improve the accuracy of quantitative analysis.