Spectroscopic Study of Electrolytic-Plasma Discharge During Hardening of 20GL Steel and Its Effect on Microstructure and Mechanical Properties

Abstract

This study investigated the electrolytic-plasma hardening (EPH) of cast 20GL steel, used for railway spring beams. The main objective was to analyze the spectral characteristics of the cathodic discharge and establish correlations between the plasma parameters, processing regimes, and resulting surface properties. Optical emission spectroscopy revealed that the plasma at 260 V exhibited a high-energy state with an electron density of ~5.3 × 10¹⁶ cm⁻³ and an electron temperature of 10,031 K. Using these parameters, the heat flux from the plasma to the steel surface was estimated at ~1.5 × 10⁷ W/m², confirming that the discharge provides sufficient energy for surface austenitization. Microstructural analysis demonstrated that the electrolyte flow rate, which determines the cooling rate, is the key parameter controlling phase transformations. At low flow rates, ferrite–pearlite and bainitic structures formed, while a fully martensitic structure and maximum hardness (1046 HV) were achieved at 10 L/min. Tribological tests confirmed the superior wear resistance of the martensitic layers, showing a friction coefficient of 0.454 and a wear volume 3.4 times lower than in the as-cast state. These findings verify that EPH offers an energy-efficient, low-cost method for improving the surface performance and service life of 20GL steel components in heavy-duty railway applications.

1. Introduction

Frequent failures of spring beams in railcars lead to substantial operating costs and compromise traffic safety. The primary cause is the beam’s operation under dynamic loads and intense friction, resulting in accelerated wear and a shortened service life [1]. Cast steel 20GL is widely used for manufacturing these components. While it possesses adequate strength, its fatigue resistance is relatively low due to structural heterogeneity [2]. To enhance reliability, quenching is employed, creating a strengthened martensitic structure that increases surface hardness and wear resistance. Subsequent tempering reduces internal stresses and ensures the necessary ductility [3]. Therefore, the appropriate selection of quenching and tempering conditions significantly extends the service life of spring beams and improves their resistance to fatigue failure.

In [4], it is noted that quenching remains one of the primary methods of steel heat treatment, ensuring martensite formation and increased hardness. However, the process is accompanied by the development of internal stresses, necessitating tempering. Low tempering maintains high hardness, medium tempering provides a balance of strength and ductility, and high tempering promotes the formation of a ductile structure resistant to impact loads. Insufficient cooling, however, can lead to the formation of troostite or sorbite, which reduces hardness. Conversely, optimal heating parameters (800–900 °C) and cooling in water, oil, or sprayed water ensure the achievement of maximum mechanical properties.

Research [5] confirms that the greatest strengthening is achieved by quenching without the formation of large grains, while subsequent heat treatment merely stabilizes the properties. However, the process of quenching alloy steels is technologically complex and can lead to structural heterogeneity and residual stresses in the absence of control over phase transformations and the distribution of alloying elements. Conversely, the study by [6] demonstrated that the choice of cooling mode has a decisive influence on the formation of the microstructure—from a ferrite–pearlite mixture to martensite, which directly determines the range of mechanical properties.

Traditional methods of strengthening spring beams, such as surfacing, metallization, and welding restoration, are expensive and do not always ensure structural homogeneity [7,8]. A promising alternative is electrolytic plasma hardening (EPH) [9,10,11], based on localized heating and cooling of the steel surface in an electrolyte under the influence of an electric discharge. This method allows for the formation of a finely dispersed martensitic structure, enhances the hardness, wear, and corrosion resistance of steel, and combines thermal strengthening with chemical-thermal processes. EPH is characterized by its relative simplicity and low cost, positioning it as a competitive solution for heavily loaded mechanical engineering components. However, the underlying processes during EPH are not fully understood, necessitating further research to optimize parameters and evaluate the method’s effectiveness for railcar spring beams.

There is growing interest in characterizing the plasma and discharge phenomena that occur during electrolytic-plasma treatment. For example, the authors of [12] examined the electrophysical and thermal processes during a liquid-cathode discharge. The authors demonstrated the distribution of temperatures and electric fields, but their research focused primarily on discharge characteristics, without analyzing the influence of the regimes on the structure of steels.

In [13], the authors performed a critical review of published data on gas temperature and electron density in electrolyte cathode atmospheric glow discharges (ELCADs). They highlighted the wide range of results associated with the use of different spectroscopic methods, line selection, and assumptions regarding local thermodynamic equilibrium. This analysis demonstrates the need for a more rigorous approach to selecting a methodology for estimating electron temperature and correctly accounting for background radiation.

The research in [14] investigated the electrolytic-plasma treatment (EPT) of low-carbon steels, with a specific focus on the electrolytic-plasma nitrocarburizing of Steel 20 under varying thermal conditions. The findings indicated that the microstructure and hardness of the resulting surface layer depended on the treatment parameters, with optimal characteristics observed at approximately 750 °C. While spectral analysis provided insights into the plasma’s electron density, the precise determination of electron temperature and a comprehensive understanding of electrolyte flow rate were not fully clarified.

Existing studies confirm the importance of spectroscopic analysis in electrolytic-plasma treatment. However, these studies also reveal gaps, specifically insufficient information on electron temperature and limited knowledge of how cooling conditions influence phase transformations in low-carbon steels. This study aimed to address these deficiencies by investigating the spectral characteristics of the discharge during the electrolytic-plasma hardening of 20GL steel specimens. Furthermore, it examined the effect of electrolyte flow rate on the structure and mechanical properties of 20GL steel after hardening.

2. Materials and Methods

The experiments were carried out at the Engineering Center “Hardening Technologies and Coatings”, affiliated with Shakarim University in Semey. Electrolytic-plasma treatment was performed using an experimental setup (Plasma Science LLP, Ust-Kamenogorsk, Kazakhstan) (Figure 1a) consisting of a 50 kW DC power source, an electrolyte bath with a circulation system, an electric pump, and a cone-shaped stainless steel anode. A 20% aqueous sodium carbonate (Na₂CO₃) solution was used as the electrolyte, with a flow rate ranging from 1 to 10 L/min. The specimens were parallelepipeds of as-cast 20GL steel (20 × 20 × 30 mm), with the bottom face submerged in the electrolyte. The specimen was placed at the cone’s geometric center, coaxial with the cone axis. The anode was configured as a truncated cone with a lower base diameter of d1 = 1 cm, an upper base radius of d2 = 18.8 cm, and a height of h = 17 cm. The specific chemical composition of the 20 GL steel is detailed in

Table 1. Chemical composition of 20GL steel.

Element	C	Si	Mn	S	P	Fe
Content, %	0.15–0.25	0.2–0.4	1.2–1.6	up to 0.04	up to 0.04	~98

. Prior to treatment, the specimen surfaces were rendered smooth using a TROJAN GP-1A grinding machine (Trojan Material Technology Co., Suzhou, China) and subsequently polished manually on a flat glass surface coated with abrasive paper. This grinding process involved a progression from coarser to finer abrasive grits, specifically from P100 to P2500. Any residual surface imperfections from the grinding were eliminated through polishing with a free-textured cloth and a specialized paste containing a suspension of dispersed chromium oxide particles.

The power source provided voltages up to 380 V and currents up to 150 A, controlled by a digital module with a PC interface. Electrolyte circulation supplied the electrolyte to the anode and returned it to the settling chamber, where the workpiece, acting as the cathode, was located (Figure 1b). When voltage was applied, Na⁺ and OH⁻ ions were accelerated in the electric field, causing intense heating of the electrolyte’s cathode layer and the formation of a vapor-gas shell. Electrical discharges generated within this environment resulted in rapid, localized heating of the sample surface, enabling the electrolyte-plasma hardening process.

Considering the EPT process from the standpoint of the current–voltage relationship, three distinct regions can be identified on the characteristic graph (Figure 2). The first region (Phase I) is characterized by the Faradaic electrolysis process, where the current peaks before entering a current interruption phase (Phase II). Upon reaching a specific critical voltage, denoted as V_B, active gas bubble formation commences near the cathode surface. This phenomenon leads to a reduction in the average current as nucleate boiling of the electrolyte occurs in the near-cathode zone. With a subsequent increase in voltage (Phase III), the coalescing vapor bubbles form a stable vapor-gas shell around the cathode. Within this shell, film boiling of the electrolyte takes place. This stage, marked by the onset of the cathodic process when a second critical voltage, V_D, is attained, involves the heating of the component by the resultant plasma.

Bubble formation arises from two primary mechanisms: the electrolysis of the electrolyte solution, which liberates gases, and the evaporation of the electrolyte’s aqueous base. The mixing of these bubbles establishes the foundation for the vapor-gas shell within the cathode space [15,16].

The electrolytic-plasma hardening process was visually documented utilizing an EVERCAM 2000-16-C high-speed camera, manufactured by General Optics LLC (Moscow, Russian Federation). Concurrently, the voltage and current parameters were logged via the electrolytic-plasma processing control panel (Figure 1a).

Optical emission spectra of plasma radiation generated during electrolytic-plasma hardening were recorded using an ATP3030 fiber-optic spectrometer (Optosky, Jimei, Xiamen, China) in the spectral range of 250–900 nm with a resolution of 0.2–0.4 nm. Spectral lines were identified by comparing the obtained spectra with reference data from the National Institute of Standards and Technology (NIST) database [17]. The spectrometer detector was positioned approximately 20 cm from the plasma generation zone. Radiation was collected from the entire plasma discharge volume, allowing for assessment of the plasma composition and components without reference to a specific point within the discharge. Measurements were performed once the plasma discharge reached a stable cathodic mode (Phase III in Figure 2). The integration time was 10 ms, and the interval between scans was 500 ms. The spectrometer, factory-calibrated by the manufacturer, was applied without additional calibration. Prior to integrating the line profiles, the spectra were pre-processed using the instrument’s built-in algorithms: “Dark Spectra Deducted”—subtraction of the dark signal and readout noise; and “Nonlinear Correction”—compensation for detector response nonlinearity based on the counters–exposure calibration function. Both procedures were performed automatically in the Optosky Spectra software (version 3.6.22).

The microhardness of the initial and hardened steel samples was measured using a Vickers HV-1 DT microhardness tester (Shanghai Hualong Test Instruments Corporation, Shanghai, China) at a load of 1 N and a holding time of 10 s (GOST 9450-76 [18]). The mean value was determined as the average of ten indentations per sample.

The microstructure and chemical composition of the hardened samples were analyzed using an SEM3200 scanning electron microscope (CIQTEK Co., Ltd., Hefei, China) equipped with a tungsten filament. Observations were carried out at an accelerating voltage of 15 kV under low-vacuum conditions. Both backscattered electron (BSE) and secondary electron (SE) detectors of the scanning electron microscope were employed to capture micrographs at magnifications of 35×, 250×, and 500×. Prior to imaging, the samples were etched in a 4% nitric acid solution in ethanol for 30 s to reveal the microstructural features.

Tribological tests were carried out on an Anton Paar TRB3 tribometer (Anton Paar, Graz, Austria) in a ball-on-disk configuration with the following parameters: wear track radius—3 mm; friction path—60 m; sample rotation speed—0.02 m/s; and normal load—6 N. A 100Cr6 steel ball with a diameter of 6 mm was used as the counterbody. The wear volume was measured using a HY2300 precision profilometer (Anytester, Hefei, China) [19,20,21].

The experiments were carried out using a constant electrolyte composition, namely an aqueous Na₂CO₃ solution with a fixed concentration of 20% while varying the applied voltage (U, V), treatment duration (t, s), and electrolyte flow rate (G, L/min). The electrolyte flow rate was measured with a Piusi K24 Meter electronic flowmeter (Piusi, Suzzara, Italy). The parameters of the electrolyte-plasma hardening process are summarized in

Table 2. Electrolytic-plasma hardening modes.

Sample No.	Voltage, V	Average Current, A	Time, s	Electrolyte Flow Rate, L/min
1	240	20	6	1
2	240	20	8	2
3	240	20	10	5
4	260	30	8	10

.

3. Results and Discussion

The plasma emission spectrum of 20GL steel subjected to electrolytic-plasma treatment at 260 V (Figure 3) showed a pronounced presence of iron (Fe I, Fe II) across almost the entire spectral range, which is consistent with the base composition of the steel. Manganese, present in the alloy as an alloying element, was also detected, with a distinct line at ~403 nm.

Of particular note were the sodium (Na I) emission lines, most prominently the D-lines at 589.0 and 589.6 nm, as well as the line at 616 nm. Their intensity reached the spectrometer’s upper detection limit (65,535 relative units), which can be attributed to the exceptionally strong sodium emission characteristic of the cathodic discharge.

The overall background spectrum at 260 V was markedly higher compared to lower voltages. In the 300–400 nm range, a broad spectral feature (“hump”) was observed, which can be attributed to electron recombination and bremsstrahlung processes that become more pronounced at elevated plasma temperatures and densities [22]. These observations indicate that at 260 V, the plasma was in a high-energy state: the emission intensity was elevated, the electron temperature was higher, and energy transfer to the sample surface was intensified, confirming the cathodic discharge regime and strong surface heating.

To estimate the plasma electron density, the broadening of the Hα line (656.28 nm) was analyzed. Considering the spectrometer characteristics (instrumental broadening with a full width at half maximum of about 1 nm), the experimental line profiles were fitted using the Voigt function. This approach allowed us to separate the Gaussian and Lorentzian contributions to the line shape and attribute the Lorentzian component to Stark broadening. The Olivero & Longbotham approximation [23] was applied to accurately distinguish the two contributions, enabling the extraction of the true Lorentzian width associated with the linear Stark effect. The Lorentzian full width at half maximum (FWHM) was then used to determine the electron density.

The Olivero–Longbotham approximation is expressed as follows:

$${∆\lambda }_V=0.5346·{∆\lambda }_L+\sqrt{0.2166·{{∆\lambda }_L}^2+{{∆\lambda }_G}^2},$$

(1)

where ∆λ_G is the FWHM of the Gaussian profile, ∆λ_V is the FWHM of the Voigt profile, and ∆λ_L is the FWHM of the Lorentzian profile.

From the measured spectrogram data, the FWHM of the Voigt profile was obtained (∆λ_V = 1.438 nm). Using Equation (1), the FWHM of the Lorentzian profile was calculated as ∆λ_L = 0.7177 nm. Since for the Lorentzian profile the values of FWHM and full width at half area (FWHA) coincided, the resulting ∆λ_L was directly used in the formula proposed by [24] to calculate the electron density from the Hα line:

$$n_e={10}^{17}·{\left({\displaystyle \frac{{∆\lambda }_L}{1.098}}\right)}^{1.47135},$$

(2)

where n_e—electron density in cm⁻³.

As a result, the calculated value of electron density was n_e ≈ 5.3 × 10¹⁶ cm⁻³, which is in good agreement with previously reported data on the spectroscopic analysis of electrolytic plasma [12,25].

The electron temperature of the plasma formed during EPH was determined using the spectral line intensity ratio method [26]. The analysis was based on the experimental spectrum recorded at 260 V. Neutral sodium (Na I) lines at 568.82 nm and 616.07 nm were chosen for evaluation, as they originated from the same ionic state, were well separated from neighboring strong lines, and were within the 560–620 nm region where the spectrometer response is essentially wavelength-independent. Therefore, no correction for spectral sensitivity was required. It should be emphasized that the Na I resonance line at 589.26 nm reached the detector’s saturation limit in several spectra, producing an artificial plateau at the line maximum. This effect was not related to self-absorption but resulted from spectrometer saturation. Consequently, the 589.26 nm line was excluded from the temperature calculation, and the analysis was performed using the weaker Na I lines at 568.82 nm and 616.07 nm, which exhibited sharp, undistorted peaks without signs of saturation.

For each line, a ± 1.0 nm window was selected. After background correction, the line intensity was calculated by trapezoidal integration, which yielded the integrated values I₁ and I₂ (in arbitrary units, counts·nm). The transition parameters were taken from the NIST database: for Na I (568.82046 nm), the upper level energy was E₁ = 34,548.729 cm⁻¹, the statistical weight q₁ = 6, and the Einstein coefficient A₁ = 1.21 × 10⁷ s⁻¹; for Na I (616.0747 nm), the upper level energy was E₂ = 33,200.673 cm⁻¹, the statistical weight q₂ = 2, and the Einstein coefficient A₂ = 4.98 × 10⁶ s⁻¹. The electron temperature, T_e, was determined using the Boltzmann relation [27]:

$$T_e=-{\displaystyle \frac{E_1-E_2}{k_B·\mathit{ln}\left(\frac{I_1·{\lambda }_1·g_2·A_2}{I_2·{\lambda }_2·g_1·A_1}\right)}},$$

(3)

where k_B is the Boltzmann constant (8.625 eV/K).

After background correction and integration, the following relative intensities were obtained: I₁(562.82 nm) = 9601.6 counts·nm; I₂(616.0747 nm) = 1475.6 counts·nm. The energy difference was ΔE = 0.167 eV. Substituting these values into Equation (3) gave an estimate of the electron temperature during EPH: T_e = 10,031 K.

Taking into account the dark spectrum and applying nonlinear correction, the uncertainty in the integrated relative intensities in the investigated range (560–620 nm) was estimated to be within 5%. This corresponded to a relative uncertainty in the calculated electron temperature of about ±10–12%. It should be noted that the main source of error is related to registration quality, as the small energy difference between the selected Na I transitions (ΔE = 0.167 eV) increased the sensitivity of the method to intensity errors. Future work will improve the electron temperature estimation accuracy by using spectral lines with larger excitation energy differences, multi-line Boltzmann analysis, and higher-resolution spectrometry. Nevertheless, the obtained electron temperature of ~10,000 K is consistent with the literature data for EPP and atmospheric plasma discharges [13,28], confirming both the reliability of the chosen approach and the representativeness of the results for describing the electron component of the plasma under the studied conditions.

Given the estimated electron temperature T_e, the approximate heat flux from the plasma, q_p, to the cathode surface can be evaluated using the relation proposed by [29]

$$q_p=j·U-{\displaystyle \frac{j}{e}}·\left(A_{eff}+3.2k_B{T}_e\right),$$

(4)

where j—current density, U—potential drop across the cathodic sheath, A_eff—effective work function of the cathode material, and e—elementary charge. Substituting the measured parameters, the heat flux to the cathode surface was estimated as q_p ≈ 1.5 × 10⁷ W/m². This estimated heat flux represents the energy transfer from the plasma to the steel surface. Its value may vary within one order of magnitude due to the oscillatory character of the discharge, breakdown–recovery cycles of the vapor-gas envelope, and heat removal into the electrolyte. Nevertheless, the obtained value indicates that the plasma is in a high-energy state capable of heating steel to temperatures above 1000 °C. Rapid quenching in the circulating electrolyte from these elevated temperatures then provides very high cooling rates, which ensures martensitic transformation of the surface layer.

Figure 4a presents a section of the experimental setup (the electrolytic cell) in its initial state: the sample is fixed in the holder, immersed in the electrolyte, and connected to the electrodes, while processing has not yet begun. In contrast, Figure 4b, taken 0.8 s after applying a voltage of 260 V, shows the appearance of an intense glow accompanied by the release of gas bubbles at the sample–electrolyte interface. This phenomenon is attributed to the formation of a vapor-gas envelope around the sample and the onset of the electrolyte-plasma discharge, which locally raises the steel surface temperature to very high values. As a consequence, phase transformations occur, including austenitization of the surface layer, followed by rapid quenching in the electrolyte, ultimately leading to the development of a hardened microstructure and an increase in material hardness.

The initial hardness of the 20GL cast steel was estimated at 150–160 HV. The Vickers hardness (HV) measurements of the treated samples showed significant variation, reflecting different extents of structural transformation (Figure 5). The first sample reached 266 HV, corresponding to a predominantly ferrite–pearlite structure with only slight hardening. The second sample exhibited slightly higher values (275 HV), but their scatter and instability suggest only partial formation of bainitic areas without a pronounced strengthening effect. More substantial changes were observed in the third sample, where the hardness increased to 492 HV. This rise is attributed to the development of a mixed structure consisting of bainite and partially transformed martensite; however, incomplete austenite transformation limited the achievable strength. The most pronounced strengthening effect was obtained in the fourth sample, with a hardness of 1046 HV, confirming the intensive formation of martensite. This sample demonstrates the most favorable balance of mechanical properties among those studied.

Figure 6 illustrates the microstructural evolution and phase transformations induced by electrolytic plasma hardening. The initial state of the cast steel (Figure 6a) is represented by a ferrite–pearlite mixture. Sample 1 (Figure 6b) displayed a structure essentially identical to the as-cast condition, consisting of ferrite and pearlite. Ferrite, with its low dislocation density and soft body-centered lattice, contributes ductility but little hardness, whereas pearlite, with its alternating ferrite–cementite lamellae, moderately enhances strength. The measured hardness of 266 HV reflects these features and indicates a low hardening efficiency. Sample 2 (Figure 6c) revealed the development of acicular bainite, characterized by fine carbide precipitates within a ferritic matrix and a higher density of lattice defects. The literature [30] values place bainite hardness between 250–400 HV, consistent with the observed 275 HV. This confirms the intermediate character of bainite between pearlite and martensite, combining enhanced hardness with retained toughness. Sample 3 (Figure 6d) exhibited a mixed structure of partially formed martensite and acicular bainite. Martensite, a supersaturated solid solution of carbon in α-iron formed under rapid quenching conditions, is associated with a very high defect density. Its presence produced a pronounced rise in hardness to 492 HV, clearly exceeding the pearlitic and bainitic levels. However, incomplete austenite transformation restricted the hardness, reflecting an intermediate phase state. Sample 4 (Figure 6e) demonstrated the most significant transformation. Under optimal EPH parameters (260 V, 8 s, electrolyte flow rate of 10 L/min), rapid cooling promoted the complete transformation of austenite into martensite. This phase is recognized as the hardest of quenched structures, and the measured hardness of 1046 HV is in excellent agreement with the literature data (900–1200 HV for martensitic structures in low-alloy steels). Thus, EPH facilitates the formation of a hardened surface layer with extreme properties through the predominance of martensitic transformation.

Qualitative phase identification in carbon–manganese steels typically relies on conventional metallographic criteria, such as morphology, etching response, and hardness values. For the present work, which aimed to determine optimal processing parameters for the electrolytic-plasma treatment of 20GL steel, this approach—combining etched SEM cross-sections with hardness profiling and tribological data—was deemed sufficient. This methodology is well-established for distinguishing martensitic, bainitic, and pearlitic/ferritic regions in steels with similar compositions and treatment histories. Furthermore, the approach was validated by the clear correlation between microstructural morphology and the microhardness levels characteristic of these phases, as reported in [31,32]. Quantitative methods such as XRD, EBSD, and related advanced characterization techniques for precise phase identification will be addressed in our subsequent studies. Similar approaches to microstructural characterization using EBSD and in situ microscopy have been discussed in the literature [33].

The structural transformations observed in the samples after EPT are closely linked to the processing conditions, particularly the electrolyte flow rate, which governs the cooling rate. At low flow rates (1–2 L/min), cooling was insufficient to retain austenite, resulting in a microstructure composed mainly of ferrite–pearlite and bainitic constituents. At intermediate flow rates (around 5 L/min) combined with prolonged treatment time, partial transformation of austenite into bainite and martensite occurred, leading to intermediate hardness values. The most pronounced strengthening was achieved at the highest flow rate (10 L/min), where the cooling rate exceeded the critical threshold for 20GL steel, enabling complete martensitic transformation and hardness values above 1000 HV. These results demonstrate that the electrolyte flow rate is the key process parameter controlling phase transformations during electrolytic plasma hardening.

In addition to governing the cooling rate, the electrolyte flow rate strongly influences the dynamics of vapor-gas shell formation. At low flow rates, the intervals between shell breakdown (discharge) and recovery are prolonged. Under these conditions, shell reformation around the sample surface occurs more slowly, ionization becomes less stable, and plasma–surface contact is intermittently disrupted. Consequently, heating is non-uniform and accompanied by localized temperature fluctuations. In contrast, higher flow rates promote faster shell recovery, leading to more frequent discharges, stable plasma–surface interaction, and uniform heating of the near-surface layer. This facilitates more efficient phase transformations and the development of a homogeneous hardened structure.

To evaluate the effect of EPH on the chemical composition of the near-surface layer, energy-dispersive microanalysis was carried out on cross-sections of the samples. Due to technical limitations of the sample preparation procedure (embedding in epoxy resin and subsequent analysis), the quantitative data for carbon (C) were overestimated because of background contributions from the organic matrix and therefore cannot be considered reliable. For this reason, the analysis primarily focused on changes in the alloying element manganese (Mn). The results revealed a decrease in manganese concentration in the near-surface region after EPT, which can be attributed to its enhanced chemical activity under plasma heating. During treatment, manganese undergoes partial oxidation and volatilization within the vapor-gas shell, leading to surface depletion. Moreover, spectroscopic analysis of the EPT plasma detected manganese lines in both atomic (Mn I) and ionized (Mn II) states. Their presence provides additional evidence that under the energy-intensive cathodic mode, manganese actively participates in discharge-zone processes.

It should also be noted that the reduction in manganese content in the solid solution may be attributed not only to surface oxidation and diffusion-driven redistribution, but also to the formation of secondary carbides. Due to its pronounced affinity for carbon, manganese can form carbide phases such as Mn₃C or (Fe,Mn)₃C, particularly under conditions of local carbon enrichment followed by ultra-rapid cooling during EPH. The precipitation of such carbides decreases the manganese concentration in the austenitic/martensitic matrix and further contributes to the depletion of manganese in the solid solution. To clarify these mechanisms, future work will involve the depth profiling and mapping of manganese and carbon distributions, transmission electron microscopy (TEM) for the identification of dispersed carbide and oxide phases, and chemical characterization of oxide inclusions formed in the discharge zone.

The structural analysis of the 20GL steel cross-section after EPH at 260 V for 8 s revealed the formation of a distinct zonal structure (Figure 7). Three regions can be identified: I—the hardened surface layer, II—the transition zone, and III—the original matrix. Detailed observations using scanning electron microscopy at ×250 and ×2000 magnifications confirmed the microstructural differences. The surface layer (zone I) exhibited a fine acicular martensitic structure, formed as a result of short-term plasma heating followed by rapid quenching. This indicates that sufficiently high temperatures were reached during EPH to induce complete austenitization and subsequent martensitic transformation. The transition zone (II) showed a heterogeneous morphology with features of partial recrystallization, grain refinement, and distortion, reflecting temperature and stress gradients. In contrast, the deeper zone (III) retained the ferrite–pearlite structure typical of 20GL steel after conventional heat treatment. The preservation of this structure confirms that the thermal influence of EPH is localized within the upper 450–500 μm layer, where the complete austenitization and subsequent martensitic transformation occur. This observation is consistent with the microhardness distribution shown in Figure 5: the maximum hardness values (up to ~800 HV) extended to a depth of 0.5–0.7 mm, which correlates well with the hardened layer thickness revealed in the SEM images.

Figure 8a illustrates the dependence of the friction coefficient on sliding distance for 20GL steel in the initial state and after EPH. In the untreated state, the average coefficient was 0.568, indicating intense surface interaction and pronounced adhesive wear. EPH reduced friction: for Sample 1, the coefficient decreased to 0.558 (–1.8%), for Sample 2 decreased to 0.536 (–5.6%), for Sample 3 reduced to 0.547 (–3.7%), while the greatest reduction was achieved for Sample 4 to 0.454 (–20.1%). The friction behavior of the initial sample and Samples 1–3 was similar, as their surface layers predominantly contained ferritic–pearlitic or bainitic structures with comparable hardness and wear resistance. In contrast, Sample 4 exhibited a markedly lower and more stable coefficient of friction. This improvement stems from the formation of a fully martensitic layer, which, due to its high hardness and homogeneous surface morphology, effectively minimizes adhesive wear. The increased microhardness of the hardened layer suppresses plastic deformation, limits adhesive wear, and ensures stable antifriction behavior under dry sliding conditions, which is particularly important for improving the reliability of spring beams in rolling stock.

Figure 8b presents the wear test results. The untreated sample showed the highest wear volume—about 1.16 × 10⁻⁵ mm³—confirming the inherently low wear resistance of 20GL steel without surface treatment. After EPH, a marked improvement in tribological performance was observed: wear volume decreased to approximately 1.05 × 10⁻⁵ mm³ for Sample 1, 1.12 × 10⁻⁵ mm³ for Sample 2, and 1.19 × 10⁻⁵ mm³ for Sample 3. The best performance was demonstrated by Sample 4, where the wear volume was reduced to 5.12 × 10⁻⁶ mm³, representing a 3.4-fold decrease relative to the initial state. This improvement can be explained by the formation of a dense, hardened layer with high microhardness and a uniform distribution of carbide phases, which significantly enhances wear resistance. Thus, the processing regime applied to Sample 4 can be considered the most effective for extending the service life of 20GL steel components.

4. Conclusions

The conducted study confirmed the high efficiency of electrolytic-plasma hardening as a promising method for the surface treatment of spring beams made of 20GL steel. It was established that the key process parameter influencing the phase transformations, and consequently the performance properties of the material, is the electrolyte flow rate. This factor governs the cooling rate and the stability of the vapor-gas envelope formed in the processing zone. Optical emission spectroscopic analysis of the plasma at 260 V revealed its high-energy state, characterized by an electron density of about nₑ ≈ 5.3 × 10¹⁶ cm⁻³ and an electron temperature of approximately 10,000 K. These parameters provide an intense heat flux to the steel surface. Microstructural studies demonstrated that by varying the EPT conditions, it is possible to tailor the surface layer structure—from ferrite–pearlite and bainitic to fully martensitic. The most pronounced hardening (up to 1046 HV) was achieved under the regime ensuring the highest cooling rate (electrolyte flow rate of 10 L/min), which led to complete martensitic transformation. Tribological testing showed that samples with a martensitic structure exhibited the lowest friction coefficient (0.454) and a wear volume 3.4 times lower than that of the original material. This improvement is attributed to the high hardness and homogeneity of the hardened layer, which limit adhesive wear and ensure stable antifriction performance. These results highlight the strong potential of electrolytic-plasma hardening for extending the service life and reliability of 20GL steel components used in rolling stock. Additionally, surface depletion of manganese was observed, which is associated with its high chemical activity under plasma heating conditions. This phenomenon may result both from oxidation processes and from the formation of secondary carbides. This effect requires further in-depth study.

Thus, electrolytic-plasma hardening can be regarded as a technologically efficient and economically viable method for enhancing the wear resistance and service life of highly loaded components, particularly spring beams for rolling stock. Future research should focus on optimizing EPT parameters for large-scale components and on elucidating the role of alloying elements in the kinetics of phase transformations under ultra-fast heating and cooling, which is of both scientific and practical significance.