Elucidating O and Cr Elemental Transfer Behavior in Submerged Arc Welding with Cr₂O₃-Bearing Fluxes

Abstract

This study investigates the influence of Cr₂O₃-bearing fluxes on the transfer behavior of O and Cr during the submerged arc welding process. A series of fluxes with varying Cr₂O₃ content are prepared and applied in submerged arc welding. A cross-zone model is developed to separately evaluate the transfer of O and Cr in both droplet and weld pool zones. The results reveal significant O enrichment in the droplet zone due to the decomposition of Cr₂O₃ under arc heating, followed by deoxidation in the weld pool. Cr transfer is found to be inhibited by the high oxygen potential in the droplets and further affected by evaporation loss. A comparison of predicted ΔCr values shows that the gas–slag–metal equilibrium model overestimates Cr transfer level, while the cross-zone model provides predictions more consistent with experimental results. This study highlights the critical role of Cr₂O₃ in regulating transfer behaviors O and Cr and provides valuable insights for flux design aimed at achieving precise compositional control and improved weld quality in welding applications.

1. Introduction

Submerged arc welding (SAW) is recognized as a highly flexible and efficient automated arc welding process. SAW is extensively utilized in heavy industrial fields due to its remarkable deposition rate and operational stability [1,2,3]. During the SAW process, intense chemical interactions occur among different consumables, including the slag, base metal (BM), molten droplets, etc. [4,5,6,7]. Given the significance of weld metal (WM) composition control in achieving high-quality welds, it is essential to understand the underlying mechanisms that govern compositional variations in the submerged arc welded metal [8,9,10,11].

The flux performs several essential functions for SAW [12]. Apart from stabilizing the plasma and facilitating slag formation and detachment, the WM composition is also governed by the flux through complex chemical interactions with the molten pool and arc plasma [13,14,15,16]. Therefore, a comprehensive understanding of the elemental transfer mechanisms during the SAW process is vital for accurate compositional control, which in turn governs submerged arc welded metal’s mechanical properties [17,18].

Cr₂O₃ is an essential component of flux, particularly in applications where enhanced corrosion resistance and hardness of the WM are required [19,20,21,22,23,24]. To investigate the effects of Cr₂O₃ on element transfer behavior during welding, several studies have been conducted. Mitra et al. [19] performed SAW using Cr₂O₃-containing flux and conducted thermodynamic analysis of the associated chemical reactions. Quintana et al. [25] examined the transfer efficiency of alloying elements during SAW with Cr₂O₃-bearing flux. Zhang et al. [26], on the other hand, designed a series of Cr₂O₃-containing fluxes and developed a method capable of constraining the transfer behavior of essential elements.

However, previous studies on Cr₂O₃-bearing fluxes have mainly concentrated on the weld pool zone. Although these studies conducted a systematic quantitative analysis of element transfer behavior, the role of molten droplets in influencing element transfer during the SAW process has been overlooked [19,26,27].

The contributions of SiO₂ and FeO to the O potential have been well documented, whereas the role of Cr₂O₃ has received far less attention. Cr₂O₃ is of particular interest because its high-temperature decomposition can generate lower-valence chromium oxides (e.g., CrO) and release O₂, thereby producing an O potential comparable to—or even higher than—that of SiO₂ or FeO. As a result, Cr₂O₃ can exert a substantial influence on the O potential under SAW conditions [12,19,28,29].

The previous model only considered the weld pool zone, which led to considerable prediction errors [7]. In contrast, the cross-zone model proposed in this study accounts for the significant O enrichment in the droplet zone and its influence on element transfer, while also considering possible metal evaporation during the SAW process.

In the present study, a series of fluxes are prepared with progressively increasing Cr₂O₃ content. By quantifying the transfer behavior of elements in different zones and incorporating thermodynamic simulations, this work aims to comprehensively elucidate the mechanisms by which Cr₂O₃ in the flux affects elemental transfer, with particular attention to O and Cr elements, thereby enhancing the understanding of its role in controlling WM composition and guiding the optimization of flux design for improved weld quality.

2. Materials and Methods

2.1. Flux Preparation

For each designed flux composition, 1 kg of analytical-grade powders was weighed according to the formulations shown in

Table 1. Measured compositions of fluxes (wt pct) [26].

Fluxes	Cr2O3	SiO2	Na2O	CaF2
F-1	12.58	7.99	0.52	78.91
F-2	22.89	7.56	0.48	69.07
F-3	32.88	7.63	0.55	58.94
F-4	41.99	7.88	0.61	49.52
F-5	53.19	7.66	0.49	38.66

; apart from Cr₂O₃, CaF₂ was introduced as an oxygen-free non-oxide diluent; the powders were then thoroughly blended in a V-type mixer operating at 0.5 Hz for one hour; after mixing, a 150 g sodium silicate solution was added as a binder, introducing SiO₂ and Na₂O, which improve slag removal and arc stability [26]. The bonded mixture was subsequently pelletized and dried in a muffle furnace at 973 K for three hours; finally, the sintered material was crushed and sieved to obtain particles between 14 and 100 mesh [26,30]. It should be noted that, in this study, the term flux refers to the raw material before SAW, whereas slag denotes the molten or solidified flux formed during or after the SAW process [12]. The flux samples were characterized for their elemental composition through X-ray fluorescence (XRF) analysis with an S4 Explorer spectrometer; the analytical compositions of the fluxes are presented in Table 1, while the procedures for determining the chemical compositions of the metal have been described in detail in previous work [26,31].

The ΔO and ΔCr values were obtained under stable welding conditions after the submerged arc welded metal composition reached a steady state [30]. Because these values were measured after the metal composition reached a stable state, they reflect the steady elemental transfer of O and Cr under controlled welding conditions. As a result, repeated experiments and statistical indicators (e.g., standard deviation or standard error) are not applicable, which is consistent with previous studies in this field [1,12,26,28,32]. A single arc heat input and one BM grade (Q345A) were used to avoid adding unnecessary variables. This allows the study to focus purely on the influence of Cr₂O₃ on the transfer behavior of O and Cr, which is the major objective. This approach aligns with established practices in related studies [1,12,26,28,32].

2.2. Welding Experiment

The base material used in this study was Q345A, a widely applied low-alloy structural steel. Bead-on-plate welding was conducted via a single-pass, double-electrode SAW process (Lincoln Electric Power Wave AC/DC 1000 SD, Lincoln Electric, Cleveland, OH, USA) under a total heat input of 60 kJ/cm. The current and voltage settings were DC 850 A/32 V for the leading electrode and AC 625 A/36 V for the trailing one, with a travel velocity of 500 mm/min.

Specifically, 60 kJ/cm was chosen because it represents a typical high heat-input level in SAW, and Q345A is widely used and compositionally stable, making it a representative BM [12,26,32]. Using one heat input and one BM grade reduces unnecessary variability and allows the essential thermodynamic trends to be identified more clearly, which is appropriate for generalizing the underlying mechanisms [19,32].

2.3. Chemical Composition Analysis

The concentrations of metallic elements were measured using inductively coupled plasma optical emission spectrometry (ICP-OES), while the O content was analyzed with a LECO analyzer (LECO Corporate, St. Joseph, MI, USA). The chemical compositions of both the electrode and BM are listed in

Table 2. Measured chemical compositions of BM and electrode (wt pct) [26].

	C	Si	Mn	Ti	Cr	O	Fe
BM	0.112	0.142	1.54	0.015	0.018	0.003	Balanced
Electrode	0.127	0.049	1.65	0.015	0.015	0.003	Balanced

. Data on metal content measurements can be found in our previous publication [26]. The measured Δ value (the contribution of the flux to elemental transfer) for O and Cr is summarized in

Table 3. Quantified Δ values of O (ppm) and Cr (wt pct) [26].

Weld Metals	WM-1	WM-2	WM-3	WM-4	WM-5
Fluxes	F-1	F-2	F-3	F-4	F-5
ΔO	460	710	1150	1440	1590
ΔCr	1.034	1.603	1.783	2.063	2.604

.

3. Thermodynamic Calculation

In our previous work, a thermodynamic analysis of SAW using Cr₂O₃-bearing fluxes was conducted [26]. However, previous study focused solely on the local thermodynamic behavior within the weld pool zone, neglecting the reactions occurring in the molten droplet zone. Furthermore, when quantifying element transfer level, the prior analysis only considered compositional changes in the weld pool relative to the original materials (i.e., BM and electrode), which may lead to misinterpretation of the underlying transfer mechanisms.

To address the limitations of previous studies, the SAW process will be simulated using our previously proposed “cross-zone model.” This model enables separate quantification of element transfer behavior in both the droplet and weld pool zones, with particular attention to O and Cr elements [32].

Figure 1 illustrates the essential reaction interfaces involved in the metallurgical process of SAW process. A comparison reveals that conventional thermodynamic equilibrium models consider only the localized equilibrium within the weld pool, whereas the multi-zone model accounts for cross-zone element transfer, particularly the significant O enrichment in the droplets under the arc [32].

In the present work, the thermodynamic modeling was carried out using FactSage 7.3, a software platform grounded in the CALPHAD methodology [33,34]. This program integrates several functional modules that enable the processing of thermodynamic information from a variety of databases. For phase-specific modeling, the Fstel, FactPS, and FToxide modules were applied to represent steel phases, gaseous phases, and slag or flux phases, respectively [33,34]. Chemical equilibrium states were calculated using the Equilib module, which estimates the distribution of chemical species when the system reaches thermodynamic equilibrium [32].

In thermodynamic studies of the SAW process, researchers commonly use the so-called “effective equilibrium temperature” to perform calculations [1,12]. This temperature does not correspond to the actual measured equilibrium temperature in SAW but rather represents the temperature at which the experimental mass action index equals the equilibrium constant. In the present study, the SAW process temperature for the droplet zone and weld pool zone was selected based on previous literature [1,12,32].

3.1. Droplet Zone

Equilib module of FactSage was employed to conduct the thermodynamic modeling subject to the Droplet Zone. The databases FToxid, Fstel, and FactPS were selected to represent the relevant phases [34]. In the thermodynamic calculations, the molten slag and steel systems were represented using the ASlag-liq (all oxides), S (FToxid-SLAGH), and LIQUID (Fstel-Liqu) solution phases [32].

• The equilibrium temperature in the submerged arc welding (SAW) process was taken as approximately 2500 °C, corresponding to the typical temperature of the arc plasma.
• The chemical composition of the BM served as the input for the metallic phase.
• An equilibrium simulation was carried out using Fe and O as the representative input elements to estimate the O concentration in the droplet zone. During this calculation, the O partial pressure P_O2 (The O₂ partial pressure obtained from the previous step) is fixed according to the values listed in

  Table 4. Simulated results for the oxygen partial pressure (P_O2, atm) and oxygen content (O, ppm) in the droplet.

  Flux	PO2	O Concentration in Droplet
  F-1	7.64 × 10−6	3642
  F-2	9.68 × 10−6	4100
  F-3	1.04 × 10−5	4250
  F-4	1.08 × 10−5	4331
  F-5	1.11 × 10−5	4391

  . The resulting O concentrations and simulated P_O2 values are also presented in Table 4.

4.

It is well known that metal evaporation tends to occur during the SAW process due to the presence of arc plasma. As noted by Kou [13], thermodynamic analysis alone cannot fully predict such losses. Given that detailed metal evaporation models are still limited in the literature, a recent empirical approach was applied to estimate metal evaporation. As such, the relationships used in Equations (1) and (2) were adopted here to approximate the evaporation behaviors of Mn and Si under high-temperature SAW conditions, which is consistent with previous papers [32,35]. The extent of metal evaporation was evaluated using Equations (1) and (2), as proposed by Zhu et al. [36], where η represents the burn-off ratio of the metal subject to the droplet zone.

$${\eta }_{\mathrm{Si}}=(38.55-0.33{\mathrm{M}}_{\mathrm{Mn}}-0.58{\mathrm{M}}_{\mathrm{Si}})/100$$

(1)

$${\eta }_{\mathrm{Mn}}=(49.04-0.32{\mathrm{M}}_{\mathrm{Mn}}-0.49{\mathrm{M}}_{\mathrm{Si}})/100$$

(2)

5.

Based on the above calculations, the metal input composition for thermodynamic simulation of the weld pool zone is summarized in

Table 5. Data of nominal compositions used for modeling.

	C	Si	Mn	Ti	Cr	O	Fe
F-1	0.12	0.06	0.82	0.015	0.017	0.184	Balanced
F-2	0.12	0.06	0.82	0.015	0.017	0.207	Balanced
F-3	0.12	0.06	0.82	0.015	0.017	0.214	Balanced
F-4	0.12	0.06	0.82	0.015	0.017	0.218	Balanced
F-5	0.12	0.06	0.82	0.015	0.017	0.221	Balanced

.

3.2. Weld Pool Zone

After completing the simulation of the droplet zone, the weld pool is subsequently modeled. At this stage, M_N is introduced to represent the dilution effect of the droplet on the weld pool, as described in our previous work. M_N refers to the nominal composition, which reflects dilution effects of the droplet and weld pool [1,30,37]. The gas–slag–metal equilibrium was simulated using the Equilib module of FactSage considering mass ratio of slag to WM [26]:

• The thermodynamic databases FToxid, Fstel, and FactPS were applied in this work.
• The molten slag and steel systems were described through the solution models ASlag-liq (all oxides), S (FToxid-SLAGA), and LIQUID (Fstel-Liqu).
• The equilibrium temperature in SAW of 2000 °C was set. All other settings were kept consistent with our previous work.

4. Results and Discussion

4.1. Transfer of O

In submerged arc welding, O plays a key role in determining the WM microstructure and its resulting mechanical properties [1,12,28,30]. It is well known that excessive O reduces toughness and hardenability, whereas insufficient O suppresses acicular ferrite formation, resulting in degraded toughness [12].

Traditional basicity index model and gas-slag-metal model consider only chemical interactions in the weld pool zone [12]. However, numerous experimental studies have shown that the O content is significantly higher in the molten droplets. For instance, Lau et al. [14,15] conducted extensive work demonstrating that O reaches its peak concentration in the droplets, which are subsequently diluted by the BM during welding.

In the following section, the transfer behavior of O across different zones is quantitatively analyzed. The calculation is based on standard industry formulas. Equations (3) and (4) are applied to evaluate the O transfer level, where M_DO and M_WO correspond to the O levels within the droplet and weld pool zones. Δ_DO refers to the amount of O transferred in the droplet zone and Δ_WO denotes the transfer occurring in the weld pool zone. M_NO means the nominal composition of O element [35].

$${\mathrm{\Delta }}_{\mathrm{D}}{\mathrm{O}=\mathrm{M}}_{\mathrm{D}}\mathrm{O} -{ \mathrm{M}}_{\mathrm{N}}\mathrm{O}$$

(3)

$${\mathrm{\Delta }}_{\mathrm{W}}{\mathrm{O}=\mathrm{M}}_{\mathrm{W}}\mathrm{O} -{ \mathrm{M}}_{\mathrm{D}}\mathrm{O}$$

(4)

Table 6. Simulated transfer behavior of O (ppm) subject to droplet zone and weld pool zone.

Flux	ΔDO	ΔWO
F-1	3612	−1732
F-2	4070	−1470
F-3	4220	−1440
F-4	4301	−1461
F-5	4361	−1611

shows the data of Δ_DO and Δ_WO.

Figure 2 presents the quantified levels of O transferred across different zones. In Figure 2a, the predicted values are obtained using the multi-zone model. A considerable transfer of O into the metal occurs in the droplet zone, after which deoxidation takes place in the weld pool. In contrast, Figure 2b shows the results predicted by the equilibrium-based model, which does not account for the strong oxidation effect occurring in the droplet zone. This model assumes that oxidation takes place solely within the weld pool zone. Tip means the tip of electrode.

In fact, Lau et al. [14,15] experimentally confirmed the high O content in droplets by extracting and analyzing them. Similarly, Sengupta et al. [1] emphasized the elevated O levels in droplets in their review. Figure 3 describes the typical cross-zone transfer behavior of O. A comparison between Figure 2 and Figure 3 clearly indicates that Figure 2a provides a more realistic representation of the O transfer pathway during the SAW process.

Among various arc welding techniques, SAW exhibits a relatively high degree of O enrichment due to the involvement of flux. According to Chai et al. [5,28,39], The high temperature of the welding arc causes the oxides in the flux to decompose, releasing O₂, significantly increasing the O content in the metal. Based on the thermodynamic modeling, Cr₂O₃ and CrO are identified as the primary chromium-containing oxides in the slag. Therefore, it can be inferred that Cr₂O₃ decomposes into CrO and O₂ under arc heating, leading to a substantial rise in O content within the droplets [28].

$${2(\mathrm{Cr}}_2{\mathrm{O}}_3{)=4(\mathrm{CrO})+\mathrm{O}}_2(\mathrm{g})$$

(5)

Lau et al. [14,15] further suggested that the partial pressure of O₂ within the arc reflects the O potential of the flux-i.e., the flux’s capacity to supply O to the metal. To validate this hypothesis, a comparison was made between the measured O content and the simulated O partial pressure. As shown by Figure 4, the O₂ partial pressure increases with rising Cr₂O₃ content in flux.

4.2. Transfer of Cr

The subsequent discussion focuses on the transfer characteristics of chromium. According to the preceding analysis, chromium transfer in Cr₂O₃-containing fluxes is mainly controlled by Reactions (6) and (7) [19,26].

$${(\mathrm{Cr}}_2{\mathrm{O}}_3)=2[\mathrm{Cr}]+3[\mathrm{O}]$$

(6)

$$\left(\mathrm{CrO}\right)=\left[\mathrm{Cr}\right]+\left[\mathrm{O}\right]$$

(7)

Figure 5 shows the relationship between Cr₂O₃ content in the flux and the activities of Cr₂O₃ and CrO. Figure 5 illustrates that the activities of Cr₂O₃ and CrO both rise as the Cr₂O₃ content in the flux increases.

Nevertheless, focusing solely on oxide activity cannot adequately account for the chromium transfer behavior [21,22,23]. Given the relatively high O potential of Cr2O3 and the increased O concentration in the droplet region, chromium transfer may be partially hindered. Figure 6a displays the predicted ΔCr values derived from both the traditional gas–slag–metal equilibrium model and the proposed cross-zone approach. The black symbols denote the ΔCr results from the conventional model, whereas the red symbols correspond to those obtained by the cross-zone model. When the O enrichment in the droplet is considered, the predicted ΔCr values exhibit markedly better agreement with the experimental data.

In addition to thermodynamic factors, Cr evaporation also occurs during the SAW process [19,27]. This phenomenon is closely related to the interaction between the high-temperature arc and the molten metal. Currently, no model is available to accurately predict the amount of Cr evaporation, and the underlying mechanism remains not fully understood. In this study, the evaporation loss of Cr was further estimated based on mass balance calculations to refine the prediction of ΔCr [26]. As shown in Figure 6b, the accuracy of the predicted ΔCr values is further improve, except the level of Cr₂O₃ is higher than 50 wt pct. Therefore, future work should focus on developing a more advanced metal evaporation model to further improve the overall accuracy of the simulation [9,40]. In addition, the model should be further optimized to clarify the relationships between the thermodynamic model and element transfer, element evaporation, and the thermodynamic behavior during solidification, thereby improving the model’s accuracy and general applicability [41,42,43,44,45,46].

5. Conclusions

This work investigates the impact of Cr₂O₃-containing fluxes on the elemental transfer of O and Cr during the SAW process. The following conclusions can be drawn:

• A multi-zone thermodynamic model is established to simulate element transfer in the droplet and weld pool zones. Compared to the conventional gas–slag–metal equilibrium model, this approach more accurately reflected the actual distribution and evolution of O and Cr during welding.
• The results revealed that O is significantly enriched in the molten droplet zone due to the decomposition of Cr₂O₃ under high-temperature arc conditions. This O enrichment leads to subsequent deoxidation reactions in the weld pool, which were well captured by the proposed cross-zone model.
• The transfer behavior of Cr was found to be strongly influenced by both thermodynamic factors and evaporation Cr loss. While the activities of Cr₂O₃ and CrO increase with rising Cr₂O₃ content in the flux, the accompanying increase in O potential within the droplet zone tends to inhibit Cr transfer into the weld metal. Incorporating evaporation loss into the cross-zone model further improved the accuracy of ΔCr prediction.

It is noted that the contribution of Cr evaporation is estimated from mass balance and empirical formulas rather than direct measurements; therefore, it should be regarded as a probable (or significant) contributor rather than a definitively established mechanism.