Study on the Thermal Fatigue Effect of Carboxymethylcellulose Solution Media Dissolved in Water as a Quenching Cooling Medium

Abstract

The degradation of a quenching cooling medium is a particularly important technical aspect in the technology of primary and final thermal treatments. This paper studied the effect of the thermal cycles of heating and cooling on a tempering medium type of 2.5% carboxymethyl cellulose solution in water. The main characteristic of a cooling medium is the absorption of heat from the part, which is influenced by the physico-chemical characteristics of the cooling liquid according to the degree of thermal fatigue. For this, the main properties of the environment are analyzed, such the chemical composition, corrosion capacity, wetting capacity and cooling characteristics. To analyze the change in properties during the operation, we studied the effect of thermal cycles on the physico-chemical characteristics of the quenching medium to determine the optimal period when the quenching medium could function in good parameters without the necessary corrections. For this purpose, cyclic and linear corrosion tests, FTIR compositional analyses and contact angle measurements were conducted.

1. Introduction

The cooling media used up to now do not fully satisfy the requirements imposed by modern treatment techniques that must ensure a broad spectrum of the physical–chemical properties of heat-treated parts as a requirement of the continuous development of the machine building industry; i.e., to have a low cost price, not to be flammable, not to be toxic and to keep the cooling properties constant over time. Therefore, as an alternative to classic tempering media, synthetic tempering media have been tested and used, usually chosen from the residual substances obtained from the chemical industries of paper processing, petroleum products, etc. By mixing certain substances with water, gels are formed. Among these substances, we mention carboxymethyl cellulose (CMC), which is the object of study of the present article. The degradation effects of synthetic carboxymethyl cellulose 2.5% hardening medium in water after multiple uses have been studied [1,2]. This is useful because in large quench pools, the medium is only changed after long intervals, which means that after a certain number of cooling cycles, quenches are performed using thermally worn quench media.

The quenching media for quenching must meet certain specific characteristics such as being cheap and having high cooling rates in the range of the minimum stability of supercooled austenite at 800–500 °C and having low cooling rates in the range of the beginning of the transformation of austenite into martensite in order not to introduce dangerous structured stresses for cracking and deformation [3,4]. Water corresponds with both from an economic perspective and the cooling speed in high-temperature ranges, but it has too high speeds in the temperature ranges of structural transformations. Thermal treatment oil perfectly corresponds technically, but it is expensive and flammable. In this context, the palette of cooling media has been diversified, creating polymeric cooling media and organic cooling media resulting from secondary products from the production process. Carboxymethyl cellulose belongs to the last category, being a secondary product obtained from paper processing [5].

Carboxymethyl cellulose is an organic polymeric substance obtained as a by-product from the manufacture of paper, having the chemical formula C8H16O8 [6]. Dissolved in water with or without additions in a percentage of less than 5%, it is used industrially as a quenching medium with a medium cooling rate, which is favorable for quenching parts made of steels sensitive to cracking or parts with shapes that generate tension concentrators [7,8]. Due to the use in high-capacity industrial tempering basins, it is necessary to study the maintenance of the quality and physical–chemical characteristics of the substance dissolved in water after a certain number of reuses [9]. The exact number of cycles that can be used without damaging the tempering process is not well determined.

In this paper, the effect of thermal wear on the quenching characteristics of carboxymethyl cellulose media was studied; namely, the cooling curves, the corrosion capacity, the contact angle between the non-thermally used and used quenching fluids [10,11] and the surfaces of improved steel parts [12]. This result established the suitable number of cycles of the CMC quenching medium used in industrial processes [13].

If the tempering medium does not correspond in terms of properties after a certain period of use, it is recommended to regenerate the tempering bath by adding additives for correction [14,15]. These additives can be preservative substances because carboxymethyl cellulose is organic and, therefore, decomposes over time, or they can be soluble or insoluble substances that can increase the number of vapor formation centers slowing down cooling or they can reduce the surface tension of the liquid [16].

2. Materials and Methods

A solution of 2.5% residual CMC from a paper manufacturing process in distilled water was made. For the analyses, a solution of 2.5% fresh CMC and 5 heat-treated solutions with 5 different cycles were obtained: CMC 10 (10 cycles), CMC 20 (20 cycles), CMC 30 (30 cycles), CMC 40 (40 cycles) and CMC 50 (50 cycles). The cycles were carried out by heating some blank samples at 800 °C and cooling in a specific environment. The cooling curves T = f (t) from Figure 1 were drawn with a standard sample of Ag with a diameter of 12.5 mm and a height of 25 mm, inside which was mounted a chromel–alumel thermocouple connected to a millivoltmeter indicator.

The contact angle measurements were obtained using a Kruss Easy Drop goniometer. The sessile drop method was chosen to measure the static contact angles of distilled water and the six solutions of CMC on a metal surface. Steel A 1040 1035 (ASTM) was chosen (

Table 1. Chemical composition of steel A 1040 1035 (ASTM) (%).

Elements	C	Si	Mn	P	S	Cr	Ni	Cu	W	Fe
Percent	0.279	0.25	0.662	0.0082	0.0135	0.065	0.011	0.053	0.035	Rest

). Using drop shape analysis software (DSA version 1.90.0.14), the average between the right and left contact angles was calculated. Before each measurement, the sample was cleaned with isopropanol and dried in the atmosphere. The volume of the liquid dropped on the surface was 2 μL. For each sample, at least 7 measurements were made every 2 s for 40 s. It was observed that 10 s was enough for the drop to stabilize on the surface. After that, the results could be considered valid to create a mean value. The standard deviation was calculated.

Fourier transform infrared (FTIR) was applied to identify the presence of metallic and organic oxides in the analyzed calibration media. FTIR spectra were registered on a Bruker Vertex 70 FT-IT spectrophotometer in total attenuated reflectance mode in the wavenumber range of 1900–650 cm⁻¹.

Electro-corrosion resistance was obtained using a VoltaLab-21 potentiostat (Radiometer, Denmark) with a three-electrode cell. A Pt auxiliary electrode and a calomel-saturated one were used. A working electrode area of 0.8 cm² was exposed to the electrolyte solution. This experimental setup was used to analyze the linear and cyclic curve determinations. Linear plots were registered at a scan rate of 1 mV/s and the cyclic plots at a scanning rate of 10 mV/s.

The corrosion potential at zero corrosion currents of Eo ≡ E (I = 0), Tafel slopes (ba and bc), polarization resistance (Rp), corrosion current density (Jcor) and corresponding corrosion rate (Vcor) were evaluated using the facilities offered by VoltaMaster4 software (version 6.0.225310).

The material used for linear and cyclic voltammetry, the cylindrical samples, the quality of carbon steel and improvement steel and the chemical composition are listed in Table 1.

After corrosion in the fresh and altered environments, scanning electron microscopy (SEM) (VegaTescan LMH II, SE detector, 30 kV) was performed. The chemical composition of the surface was established with an EDS detector (Bruker X-flash).

3. Results

From a technical perspective, the study of the cooling curves for tempering is of high importance and can provide indications about the characteristics of the use of the tempering medium. In the cooling chart in Figure 1, water was used as the comparison medium. Water has a very high cooling speed in all working ranges between 800 °C and 20 °C. Analyzing the cooling curves of carboxymethyl cellulose solutions dissolved in water of 2.5% fresh medium and a thermally used cooling medium with multiple heating–cooling cycles, it was observed that with thermal fatigue, the synthetic quenching medium became slower and slower in changing, both in the cooling speed in the areas of high temperatures in the area of the elbow of minimum stability of the austenite as well as in the area of temperatures below 200 °C that represented the beginning of the transformation of austenite into martensite [17]. The slower speed was due both to the increase in the viscosity of the medium due to intense evaporation and the presence of oxide-type residues in the quenching bath [18].

The fresh and used solutions with 10 and 20 cycles had relatively close curves, which did not create major problems involving correcting the tempering bath. This appeared to be necessary only from 30 thermal cycles upwards after studying the tempering curves. Starting with the thermal properties of the CMC–water-cooling environment and the microstructural and chemical characterization of the liquid compounds, this article presents the behavior of the liquid after a different number of cycles up to 50.

On the cooling curves, a modification of the cooling rate could be observed in the case of all CMC cooling media at around 300 °C. This may have been due to the thermal degradation of the CMC, special to the main decomposition stage at 295 °C [19]. The weight losses were assigned to fragmentation associated with pyrolytic decomposition, leading to a decomposition to lower molecular-weight molecules. The pyrolytic decomposition led to fragmentation into lower-weight molecules; thus, weight loss was recorded. Due to the thermal instability, the degraded molecules were further decomposed. The reaction continued to the ends of the polymer chains, forming aromatized units and, subsequently, cross-linked carbon skeletons.

3.1. CMC Microstructure

Figure 2 shows the basic structure of the carboxymethyl used for the experiment, easily observing the characteristics of plant fibers because the substance was collected as waste from a paper factory that used cane as the raw material.

The preliminary results on the CMC microstructure and on the cooling environments based on water combined with CMC are presented in Figure 3 and Figure 4, respectively. The experiments were performed under a vacuum and the film that remained after evaporation was analyzed. The aspect of the medium before cooling was homogeneous (Figure 3), presenting particles with the same size. The structure had a compact appearance with well-defined bonds, particles and even bubbles forming continuous and coherent film-like films. After 50 cycles, the material presented the effects of the thermal influence (Figure 4) such as cracks or fissures and the modification of particular elements of the material. Discontinuities, a lack of coherence in the structure and foreign particles, such as metal oxides, were easily observed.

In similar experimental conditions, energy-dispersive X-ray (EDX) experiments were performed on the cooling dry medium before and after 50 cooling cycles.

Table 2. Chemical composition of the carboxymethyl cellulose solution in water after CMC drying and CMC 50 medium.

Chemical Elements	Cooling Medium (CMC)		Used Cooling Medium (CMC 50)
Oxygen	80.40	89.54	49.32	55.14
Sodium	15.53	11.63	35.39	27.54
Chlorine	3.45	1.67	5.17	2.61
Nitrogen	0.85	0.24	0.41	0.16
Carbon	1.23	1.84	9.82	14.62
Iron	0.62	0.19	0.31	0.1

presents the results of the energy-dispersive X-ray analysis for the environment with 2.5% CMC.

The difference between the used and unused media from a chemical point of view was firstly given by the loss of water from the medium through vaporization, which decreased the oxygen content and increased the content of Na and Cl (probably through the formation of a NaCl compound). A decrease to half of the nitrogen element was also observed after 50 cycles and was ascribed to the partial degradation of the Na–carboxymethyl cellulose.

In the used environment, we observed the appearance of the carbon chemical element in a bigger proportion, with 9.817 wt% comparative to the initial liquid that appeared from the carbonization of the polymer under the thermal effect.

3.2. Wettability

Table 3. Mean values and standard deviation of contact angle (°) of the samples.

Sample	Water	CMC	CMC 10	CMC 20	CMC 30	CMC 40	CMC 50
Contact angle, θ	63.7 ± 2.9°	74.3 ± 2.9°	61.2 ± 2.3°	48.4 ± 2.8°	57.1 ± 0.3°	56.8 ± 0.9°	43.8 ± 3.6°

shows that the metal surface had a different hydrophilic behavior (θ < 90°), depending on the type of liquid used. The contact angle for the CMC solution had the highest value in the series. A linear decrease in the contact angle values was observed up to the CMC 30 solution, where it slightly increased. From this moment, the contact angle values had a downward trend, reaching a minimum of 43.8° for CMC 50.

From the data presented in Figure 5, it could be concluded that the wetting of the metal surface was very favorable for CMC 20 and CMC 40 liquids, and especially for CMC 50. These liquids formed a continuous thin liquid film on the hard surface. The CMC solution minimized the contact with the metal surface the most and had the lowest coefficient of friction.

The contact angle is considered to be a method for characterizing the wettability of surfaces and, consequently, of the way that cooling liquids spread on a metal part [20].

The temperature of the CMC solution increased during the metal quenching process. For this reason, a process of degradation of the polymer solution occurred, which started with the breaking of the hydrogen bonds between the macromolecular chains, followed by the breaking of the covalent bonds in the basic chain (backbone) and, consequently, a decrease in the molecular mass [19,21]. The acetal bond in polysaccharides is easily hydrolysable; the reaction is strongly favored by the presence of acidic or basic substances, but also of some natural enzymes (microbiological destruction). Depending on the stage of degradation and the fractions of reduced molecular mass, the solutions would have had different viscosities. Due to the depolymerization and decomposition processes of the cellulose solutions and also of the particles, resulting from the hardening of metals, the intermolecular interactions were different from case to case. It is, therefore, understandable why the CMC solutions had a different behavior in terms of surface wetting. Considering the variation in the contact angle—determined by the analysis of the samples from CMC 10–CMC 50, where, after CMC 30, a slight decrease in the hydrophobic properties was observed—and considering the even greater oxidation of the tempering medium if we exceeded 50 cycles, it could be concluded that the hydrophobic properties linearly decreased as the number of cycles increased.

This conclusion was confirmed by the variation in other properties, such as pH. There was a correlation between the value of the contact angle and the pH of the solution; these two parameters had a similar variation, certainly determined by the different composition of the carboxymethyl cellulose solutions in separate stages of degradation [21]. The variation in the pH of NaCMC with heat treatment procedures was established and determined. The pH value decreased to a given point, after which the pH again took a high value and then decreased, demonstrating a different chemical composition of the CMC solution after the heat treatment.

A strong change in the properties of the liquids was observed for the cooling liquid used with 20 thermal cycles. This manifested itself as a destabilization of the properties due to the formation of oxides, which worsened the cooling properties during quenching (the contact angle decreased from 61.25° to 48.4°). After another 10 cycles in the new conditions of increased viscosity through water vaporization and the appearance of iron oxides, silicon oxides and other compounds, a new organization of the coolant structure appeared. This was also observed in the linear and cyclic corrosion curves as well as in the cooling curves.

The cooling liquid with an organic solution was generally slightly hydrophilic (the contact angle varied between 70° and 45°), indicating the formation and preservation of a vapor jacket; i.e., the contact film during heating with more weight and the cooling period through the film of the gas that surrounded the part (i.e., by radiation) was greatly reduced. The cooling was achieved mainly by boiling and conduction and, therefore, was much more intense.

Along with the change in the composition and viscosity of the medium following the thermal cycles of use during cooling for repeated tempering, the hydrophilic effect of the liquid on the surface of the hardening steel was strongly accentuated and the contact angle dropped to values below 45° at 50 thermal cycles. This caused the vapor film that was created in the first stage of cooling to be much more adherent to the part, increasing the heating period and negatively modifying the cooling parameters.

This corroborated with the increase in the oxidation effect, which showed that at a number greater than 50 cooling cycles, a synthetic medium of an organic nature must be added to restore its properties as a cooling medium during tempering.

3.3. FTIR Analysis

Two samples of 2.5% carboxymethyl cellulose solution in water were analyzed; the first was unused liquid (red curve) and the second was the medium used for 50 cooling cycles (black curve) (Figure 6). The liquid was placed between two potassium bromide tablets and irradiated in the infrared spectrum (4000–400 cm⁻¹). The two curves of the absorption bands overlapped when looking for the areas with peaks (local maxima and minima) and were highlighted with the help of the deconvolution performed by the device. There was a substantial difference between the two curves; the areas with spectral changes were 990 cm⁻¹, 929 cm⁻¹, 876 cm⁻¹ and 777 cm⁻¹.

The FTIR analysis showed the main signals on the spectrum between the wavelengths of 700 cm⁻¹ and 1900 cm⁻¹. The analyzed bands showed the characteristic spectrum of the bonds of the asymmetric and symmetrical groups of carboxymethyl cellulose at 1582 cm⁻¹ and at 1390 as well as the C-H bond at 1409 cm⁻¹ of organic compounds with the hydroxyl group O-H. Signals also appeared at 1320 cm⁻¹ and 1010 cm⁻¹, indicating the presence of C-H and C-O bonds, which are characteristics of the polymer structure of carboxymethyl cellulose. The C-O bond appeared at 1730 cm⁻¹, characteristic of the group of pectic compounds specific to cellulose plant cells.

The peaks at 1638 cm⁻¹, 1640 cm⁻¹ and 1648 cm⁻¹ presented the bonds H-O-H, C-O-O and C-OH (the presence of the hydroxyl group). Also, the group of signals at 1415 cm⁻¹ indicated the presence of bonds of C-O-C. Unlike the absorption band of the freshly prepared carboxymethyl cellulose solution, the one used in multiple heating and cooling thermal cycles showed some variations that indicated the presence of iron and silicon oxides and other compounds present as residues after cooling during the heat treatment of tempering. On the curve of the multiple medium-used CMC 50, there was the bond of Si-O-Si at 980 cm⁻¹, the presence of FeOH at 920 cm⁻¹, FeS at 800 cm⁻¹ Si-O-H and, at 605 cm⁻¹, Fe-O.

The new bonds of compounds that appeared after the thermal decomposition of the CMC, established by FTIR analysis, influenced the surface tension of the liquid, decreased the viscosity and provided a contact angle characteristic of strongly hydrophilic substances.

3.4. Analysis of Chemical Corrosion by Linear and Cyclic Voltammetry

Tafel diagrams and cyclic voltammograms were presented to exemplify how to evaluate the electro-chemical parameters of the corrosion process generated by the unaltered 2.5% CMC quenching medium and the thermally fatigued one for the selected steel (Figure 7). In the case of linear voltammetry, the parameters according to which we could appreciate the degree of corrosion of an environment with respect to a certain metal were the corrosion potential (Ecor (mV)); the oxidation-reduction potential on the Tafel diagram; the anodic branch ba (mV), which provided indications about the oxidizing characteristics of the environment; the cathodic branch bc (mV), which provided indications about the reducing characteristics; the corrosion rate (Vcor (mm/year)); and the current density (Jcor (mA/cm²)).

From the Tafel slopes, the transfer coefficient could be determined, which was different from 0.5 in the case of slow reactions. From the Evans diagram, it could be seen that the anodic branch of the polarization curve had a steeper slope than the cathodic branch in both cases.

The high value of the ba constant, especially at CMC 2.5% (ba = 1023 mV, thermally fatigued) indicated that the anodic oxidation process was high, so the spent quenching medium had a strong oxidizing effect. It was observed that the altered quenching medium had more significant oxidation characteristics (ba = 1023 mV) than the fresh quenching medium (ba = 773 mV). The reduction potential was higher (cathodic branch; bc = −399 mV) for the 2.5% CMC quenching medium unaltered by thermal wear compared with the reduction potential of the worn 2.5% CMC (bc = −380 mV). The corrosion potential provided indications about the thermodynamic stability of the metal-corrosive medium interface and specified the physical–kinetic conditions for breaking through the superficial layer for the appearance of the corrosion effect.

The instantaneous corrosion current indicated that the current flow that was present at the equilibrium potential was E = Eo (when the superficial film on the metal sample was pierced). In the used medium, it was Ecor = 1102 mV, whilst in the fresh medium, it was Ecor = 1028 mV. The higher the Jcor, the higher the effort required to damage (corrosion) the surface layer.

From

Table 4. Corrosion characteristics in linear voltammetry.

Sample	E0 (mV)	ba (mV)	bc (mV)	Rp (ohm·cm2)	Jcor (µAA/cm2)	Vcor (mm/an)
CMC	−1028.5	773.2	−399.8	454.50	0.11	1.324
CMC 10	−1075.5	548.8	−162.5	666.76	87.79	1.011
CMC 20	−1061.5	185.4	−114.4	452.46	50.35	5.799
CMC 30	−1102.0	1027.3	−380.1	594.77	0.28	3.329
CMC 40	−1111.8	345.1	−174.8	604.40	87.50	1.007
CMC 50	−1135.2	318.8	−192.4	584.97	92.58	1.066

, it can be seen that the Jcor for unaltered 2.5% CMC was 0.1 mA/cm², whilst for the treated medium with 50 thermal cycles, it was 0.28 mA/cm², which was significantly high.

The higher the current density (Jcor), the lower the polarization resistance, so the more aggressive the liquid medium was on the metal surface.

The polarization resistance is used to determine the instantaneous corrosion current that occurs at the metal-corrosive medium interface and cannot directly be measured by electro-chemical methods. The polarization resistance provides indications about the thermodynamic stability characteristics of the metal. It can be seen from Table 4 that the metal–CMC clean interaction was Rp = 594 ohm/cm², whilst the CMC-used metal action was Rp = 454 ohm/cm², which highlighted the corrosive nature of the thermally fatigued biological tempering environment. The corrosion rate was determined as the thickness of the metal layer removed by corrosion per unit time. From the table, it can be seen that the corrosion rate was 2.5 times higher in the case of the altered quenching medium compared with the clean quenching medium. This may have been due to the decrease in molecular mass and the formation of acids in the CMC 50 solution following the stronger thermal degradation of the CMC compound [22].

Cyclic voltammetry showed the nature of the corrosion. It was observed that the unaltered 2.5% CMC achieved generalized corrosion on the steel sample, which was desirable because it occurred over a long time and the entire surface was affected and, therefore, visible. After the corrosion analysis, the steel samples were analyzed from a microstructural and chemical perspective.

From the structural analysis, shown in Figure 8, for the sample of improvement of steel corroded in fresh CMC 2.5%, pitting was observed, much less than for the sample corroded in CMC 50. The greater amount of pitting on the sample corroded in CMC 50 provided premises for a faster appearance of cracks on the surface. Also, it can be seen from Figure 8c,d that deposits of the conglomerate type were distributed both on the surface and inside the pitting. These were due to the presence of particles in the suspension attributed to residues from previous tempering cycles.

Table 5. Point analysis for areas with pitting corrosion.

Element	Fe wt%	C wt%	O wt%	Si wt%	F wt%	Mn wt%	EDX Error
Point 1 (Figure 9a)	93.61	3.48	-	0.43	1.33	1.14	1.3
Point 2 (Figure 9a)	74.07	2.78	21.70	-	-	1.45	0.5
Point 3 (Figure 9a)	67.07	2.52	29.3	0.53	-	-	0.1

shows the chemical analysis of surface affected by corrosion in three different areas.

Table 6. Spot analysis for areas with pitting corrosion.

Chemical Elements	Fe wt%	C wt%	O wt%	Cl wt%	Mn wt%	Si wt%
Point 1 (Figure 10a)	88.11	1.99	9.89	-	-	-
Point 2 (Figure 10a)	77.09	1.27	18.82	2.05	-	0.75
Point 3 (Figure 10a)	72.57	1.65	19.37	5.77	0.61	-

shows the chemical analysis of the first point on a surface not affected by corrosion, the second point on the areas with deposit conglomerates and the third point inside the pitting. There was a decrease in the percentage of Fe and an increase in oxygen, mainly due to the formation of oxides, but the presence of Cl and Si was a result of the formation of chlorides following the corrosion process.

4. Conclusions

Fresh and used CMC types of CMC 10 and CMC 20 cooling media had good average quenching rate cooling curves for tool steels that were sensitive to cracking.

CMC 30, CMC 40 and CMC 50 had too low cooling curves and required the correction of the cooling bath with salt-type elements and industrial detergents, with the role of increasing the cooling capacity and intensifying convection in all thermal ranges.

From the SEM comparative analysis of fresh CMC and CMC 50, a deterioration of the structure could be observed after 50 thermal cycles. Also, the non-formity and incoherence of the thermally fatigued CMC structure due to the presence of residual oxides and substitutes from the thermal processing process were highlighted.

The FTIR analysis showed variations in the spectrum that indicated the presence of iron and silicon oxides and other compounds present as residues after cooling during the heat treatment, which changed some physico-chemical properties such as the viscosity of the medium, the thermal conductivity and corrosion.

The contact angle was used as a method to characterize the wettability performance of different solutions of CMC. The results showed that the solutions that were subjected to 30, 40 and 50 heating cycles had the best spread on metal surfaces. Instead, water and unused solutions or solutions that had been subjected to 10 or 20 cycles of hardening metal parts had the highest contact angles in the series.

Both from the analysis of the electro-corrosion parameters, especially of the polarization resistance and the corrosion speed, and from the SEM and EDX analyses of the analyzed samples, it was observed that the quenching medium maintained its corresponding technological properties up to a level of CMC 30 end wear. After this level of wear, the quenching medium had to be corrected.