Wetting and Cooling Performance of Mineral Oils for Quench Heat Treatment of Steels

Transcription

1 , pp Wetting and Cooling Performance of Mineral Oils for Quench Heat Treatment of Steels Gopalan RAMESH and Kotekar Narayan PRABHU* Department of Metallurgical and Materials Engineering, National Institute of Technology Karnataka, Srinivasnagar, Mangalore, India. (Received on December 20, 2013; accepted on February 19, 2014) In the present work, wetting kinetics, kinematics and heat transfer characteristics of mineral s having varying thermo-physical properties sourced from different suppliers were investigated using contact angle, online video imaging and cooling curve analysis techniques. The relaxation behavior of mineral s of low viscosity and surface tension on Inconel substrate indicated improved wettability and fast spreading kinetics while mineral s of high viscosity and surface tension showed reduced wettability and slower spreading kinetics. Further, the spreading behavior of mineral s of lower viscosity and density showed the absence of viscous regime. During rewetting, formation of double wetting fronts and more uniform nature of wetting front were observed with mineral s of high viscosity and flash point whereas no additional wetting front was observed for mineral s of low viscosity and flash point. Among the convectional/fast/ hot mineral s, higher wetting front velocity and cooling rate were obtained for low viscosity mineral. The heat extracting capability of high viscosity mineral s was higher during vapour and nucleate bing and lower during liquid cooling stage. Further, highly viscous mineral s showed uniform heat transfer compared to mineral s having low viscosity. KEY WORDS: quenching; mineral s; wetting; contact angle; cooling curve analysis; heat flux transients. 1. Introduction Quenching is a critical step of heat treating process which defined as rapid cooling of components in-order to form martensitic/bainitic microstructure and to avoid the transformation of pearlite/ferrite in the case of steel. The cooling conditions during quenching play an important role on phase transformation and development of mechanical properties of the components. 1) Factors that influence the cooling conditions of component during quenching are categorized into three groups (i) workpiece characteristics (composition, mass, geometry, surface roughness and condition) (ii) quenchant characteristics (density, viscosity, specific heat, thermal conductivity, bing temperature) and (iii) quenching facility (bath temperature, agitation rate, flow direction). Of these, the quench medium used to extract the heat from the hot component at particular rate is a significant factor for heat treating engineer to alter/achieve the desired cooling condition of components in both technical and economical consideration. 2) Different liquid quench media used in heat treating industries are water, brine, aqueous polymer and mineral quenchants. Even though mineral possesses several environmental problems and fire risks, still it is most widely (almost 85%) used quench medium in heat treating industries. This is due to its better aging stability, thermal stability and oxidation resistance. Further, mineral quenching resulted in more uniform cooling, reduced distortion and cracking of steel components. 1) The cooling behavior of mineral during quenching is same like water which involves three stages of cooling namely, vapour blanket, * Corresponding author: prabhukn_2002@yahoo.co.in DOI: nucleate bing and convective cooling stages. However, the cooling performance of mineral is much lower than water. Further, formation of double wetting fronts on quench probe was reported for mineral quenching. 3,4) Mineral s are petroleum byproducts, generally mixtures of chemical structures with a range of molecular weights and do not contain any fatty components. Generally the mineral s are distilled from the C26 to C38 fraction of petroleum and composed of branched paraffins (C nh 2n+2) and cycloparaffins (C n H 2n ) together with a small amount of aromatics (benzene ring and its derivatives). Within an individual molecule, there are some cycloparaffin rings, aromatic rings and the necessary paraffin and olefin side or connecting groups. 5) The wetting agents, accelerators and anti-oxidant may be added to achieve specific quenching characteristics of mineral s. Mineral s can be grouped into distinctive groups based on the composition, the presence of additives and application temperature. They are classified as conventional s, fast or accelerated s, martempering or hot quenching s. Conventional quenching s are usually composed of paraffinic and naphthenic fraction with a viscosity ranging from 100 to 110 SUS (Saybolt Universal Seconds) at 40 C (some s may have viscosities of upto 200 SUS at 40 C). These s may contain antioxidants to reduce the rates of oxidative and thermal degradation but do not contain additives to increase the cooling rate. The conventional quenching tends to show a prolonged vapor blanket stage, a short nucleate bing stage and finally a very slow cooling of convective stage. Fast quenching s have low viscosities in the range of 50 and 110 SUS at 40 C. These s contain one or more additives to enhance the wetting and quenching speed and often contain antioxidants. The fast quenching shows a high quenching speed during the vapor blanket stage and in some situations approaching 2014 ISIJ 1426

2 the initial speed of water followed by a moderately fast cooling rate in the nucleate bing range. The cooling rate in convection stage is usually about the same as provided by conventional quenching s. However, some fast quenching s containing special additives provide faster cooling rates in convection stage. Martempering or hot quenching s are solvent-refined paraffin-base mineral s with good thermal stability and oxidation resistance. They are used at temperatures between about 95 C and 230 C. They may also contain antioxidants to improve their aging stability. 4,6,7) Totten et al. 8) discussed the importance of chemistry of quench and its significance on heat transfer performance during quenching. Ma et al. 9) investigated the performance of a series of mineral based quenchants having viscosities ranging from 10.4 to 120 mm 2 /s using medium carbon alloy steel (AISI 4140) probe of 9.5 mm diameter by 38.1 mm length. They observed that peak cooling rate and heat transfer coefficient obtained for the probe were increased with decrease in viscosity of quenchant. Similarly, the hardening power of the quenchant was increased with a decrease in viscosity of quenchant. Asada and Fukuhara 10) observed that the length of vapour stage during quenching was influenced by the viscosity and molecular weight of the mineral. The higher viscosity and molecular weight of the mineral resulted in early collapse of vapour film at higher temperature. Yokota et al. 11) investigated mineral based quenchants having identical viscosities and additives formulation but different types of mineral base stocks. They observed that cooling of 0.45% C steel (especially in the temperature range C) was significantly influenced by difference in the base stocks of quenchants. Fernandes and Prabhu 12) showed that the blending of palm with mineral increased the spreading rate as well as the quench severity. The literature showed wide range of quenching performance of mineral can be obtained through careful formulation and blending. The mineral s from the different producers are formulated to obtain different chemical structures with different additives. The varying chemical compositions of mineral quenchants have significant influence on its cooling performance and wetting behavior during quenching. A detailed understanding of wetting and cooling behavior of mineral base quenchants is therefore necessary for judicious selection of quenchant to obtain superior properties of components with reduced distortion and cracking. The present work is aimed at the study of wetting kinetics, kinematics and cooling performance of mineral s sourced from different suppliers and assessment of the suitability of these s for industrial heat treatment. 2. Experimental In the present work, different kind of mineral s were used as quench media and denoted as MQ-1, MQ-2, MQ-3, MQ-4, MQ-5, MQ-6, MQ-7 and MQ-8. The mineral quenchants were obtained from four different suppliers. The mineral s, MQ-1, MQ-5 and MQ-7 were procured from one supplier. Similarly, MQ-3, MQ-4 and MQ-8 were procured from one supplier. The remaining mineral s, MQ-2 and MQ-6 were procured from different sources. Among the mineral s, MQ-7 and MQ-8 are classified as hot s whereas MQ-4, MQ-5 and MQ-6 are classified as accelerated/fast s. Remaining mineral s, MQ-1, MQ-2 and MQ-3 are classified as normal/bright s. The viscosity and thermal conductivity of quenchants were measured using Brookfield LVDV-IIIU Rheometer and KD2 Pro thermal property analyser respectively. Weight displacement method was used determine the density of the fluids. The fire and flash points of quenchants were determined using Cleveland open cup apparatus. Pendant drop method was to determine the surface tension of quenchants. A 2.5 ml syringe with 0.9 mm diameter needle having a precision flow control valve was used for this purpose. For spreading studies, a droplet of quenchant was dispensed on to the Inconel 600 substrate. The spreading phenomenon was recorded using dynamic contact angle analyzer FTA 200 (First Ten Angstroms, USA) equipment. Captured images were analyzed using the FTA image analysis software to determine the interfacial tension, contact angle, droplet base diameter and spread area. The surface texture of the Inconel 600 substrate was similar to the Inconel 600 probe used for quenching experiments. The experiments were carried out at an ambient temperature of 30 C. For cooling curve analysis, two quench probes of 12.5 mm diameter and 60 mm length were prepared from Inconel 600 material. To assess the axial variations in heat flux transients, holes of 1 mm diameter were drilled at different heights located at 2 mm from the surface (probe I) as shown in Fig. 1(a). Holes designated as A1, A2, A3, A4, A5, A6, A7 and A8 were located at 7.5, 15, 22.5, 30, 37.5, 45, 52.5 and 40 mm ± 1 mm from the top surface of the quench probe respectively. For determination of heat flux variations in the radial direction, holes of 1 mm were drilled at different azimuth angles to a depth of 30 mm ± 1 mm and were located at 2 mm from the surface (probe II) as shown in Fig. 1(b). These holes, designated as R1, R2, R3, R4, R5, R6, R7 and R8, were located at angles of 0, 45, 90, 135, 180, 225, 270 and 315 respectively. Holes of diameter 1 mm (A9 for probe I and R9 for probe II) were drilled at geometric centers of both probes. Quench probes were conditioned by heating and quenching in quench for several times inorder to obtain reproducible results. Calibrated K-type Inconel thermocouples were inserted into the quench probe. The other ends of thermocouples were connected to a PC based temperature data acquisition system (NI 9213). Vertical tubular electric resistance furnace open at both ends was used to preheat the quench probe to 850 C. The heating zone of the furnace was 80 mm in diameter and having a length of 190 mm. During heating, the top and bottom parts of the furnace were covered with insulating blanket. A quench tank of internal diameter of 115 mm and length of 210 mm with ml of quenchant was kept below the furnace during quenching. The quench probe support operated through guide pins was designed such a way that, the probe was positioned at centre of heating zone during heating and at 50 mm from the bottom surface of quench tank during quenching. Once the probe attained the preheating temperature, it was directly quenched into fluid without any significant time delay (<0.35 s). The probe temperatures were recorded at a time interval of 0.1 s during quenching. The schematic of the experimental setup is shown in Fig. 2. A high performance smart camera (NI 1774C) was used for online video monitoring of the quenching process. The scanning rate was 3 images per second. The metal/quenchant interfacial heat flux transients were estimated from the measured temperature histories and thermo-physical properties of probe material by solving the inverse heat conduction problem (IHCP). The equation that governs the two-dimensional transient heat conduction is given below. Fig. 1. Schematic of (a) quench probe I and (b) quench probe II ISIJ

3 For axial location: 1 r T T T C... (1) r r r + z z = λ λ ρ p t For radial location: 1 1 r T T T C... (2) 2 r r r + r = λ ϕ λ ρ ϕ p t The above equations were solved inversely with the following initial and boundary conditions using the finite element based TmmFE inverse solver software (TherMet Solutions Pvt. Ltd., Bangalore, India), for estimating metal/quenchant heat flux transients. Initial condition Trz (, )= at t = 0... (3) T i and boundary conditions T λ λ = =... (4) r n T z n q r z t k p l r z k(,, ) on ΓK; 12,,...,..., T ι λ λ =... (5) r n T z n r z 0onΓ T ιι λ λ =... (6) r n T z n h T T r z ( )on Γ The mathematical description of the serial solution to IHCP is given in Ref. 13). 13) Figure 3(a) shows the solution domain of half symmetrical shape of the quench probe I used for estimation of heat flux components in the axial direction. The geometry was discretized using four node quadrilateral and four side linear, uniform mesh. The total number of elements used was (25 120). Figure 3(b) shows the solution domain of the quench probe II used for estimation of heat flux components in the radial direction. The geometry was discretized using three node triangle and three side curved, uniform mesh. The total number of elements in this case was The thermo-physical properties of the probe material used in the inverse model are given in Table 1. 14) For both probes the surface in contact with the liquid was divided into eight segments which were assigned an unknown heat flux boundary. The convergence limit for Gauss-Siedel iterations was set at Fig. 2. Schematic of experimental setup. 3. Results and Discussion The measured thermo-physical properties of quenchants are presented in Table 2. The density, thermal conductivity, surface tension, viscosity, flash point and fire point of the mineral quenchants obtained from the different sources were found to be in the range of kg/m 3, W/mK, mn/m, cp, C and C respectively. Among the mineral s, MQ-8 showed higher values of thermal conductivity, surface tension, viscosity, density, fire point and flash point while lower values were obtained with MQ-4. Fig. 3. Solution domain of (a) quench probe I and (b) quench probe II used in IHCP Contact Angle and Spreading Behavior Figure 4 shows images of mineral s droplet on an Inconel 600 substrate during spreading. A droplet form a triple phase contact point, also known as the contact line front/ advancing front, upon dispensing which starts to move from Table 1. Thermo-physical properties of Inconel 600 used in IHCP. 14) Temperature ( C) Thermal conductivity (W/mK) Specific heat (J/kgK) Density (Kg/m 3 ) Table 2. Thermo-physical properties of various mineral s used in the present study. Quenchant Density (kg/m 3 ) Thermal conductivity (W/mK) Flash point ( C) Fire point ( C) Surface tension (mn/m) Viscosity at 30 C (cp) MQ MQ MQ MQ MQ MQ Remarks Normal mineral Fast mineral MQ Hot MQ mineral 2014 ISIJ 1428

4 Fig. 4. Images showing contact angle relaxation in (a) MQ-1 (b) MQ-2 (c) MQ-3 (d) MQ-4 (e) MQ-5 (f) MQ-6 (g) MQ-7 and (h) MQ-8 spreading on Inconel 600 substrate. its initial position as spreading proceeds. The movement of advancing front was fast in the initial stage and slows down in the later stage before attaining equilibrium. The relaxation of droplet spreading was presented by using the time dependence of contact angle and spread area and is shown in Fig. 5. All quenchant droplets showed similar relaxation of contact angle and spreading behavior over the Inconel substrate. The contact angle decreases from the initial high value while spread area increases with time. The decrease in contact angle value and increase in spread area were rapid during initial stage and become gradual as the system approached equilibrium. The equilibrium contact angle θ e dθ (defined as the value of θ beyond which /ms) dt for all quenchant spreading was determined from the plot of contact angle relaxation. The θ e values of 13.79, 18.05, 20.84, 11.18, 15.85, 16.27, and were obtained for MQ-1, MQ-2, MQ-3, MQ-4, MQ-5, MQ-6, MQ-7 and MQ-8 respectively. Higher contact angle and lower spreading area were obtained for MQ-8 while lower contact angle and higher spreading area were obtained for MQ-4. This is due to the difference in the physical properties of. The high viscosity and surface tension of offers higher resistance to flow resulted in high contact angle and low spreading area. On the other hand, the low viscosity and surface tension of offers lower resistance to flow resulted in low contact angle and high spreading area. The high equilibrium contact angle and low spread area of is indication of lower wettability and slower spreading of on Inconel substrate. Further, having high density showed higher transition contact angle (i.e., change of contact angle from the rapid initial stage). For example, MQ-8 having high density showed transition contact angle of while MQ-4 having low density showed transition contact angle of Figure 6 shows the plot of natural logarithm of drop base diameter vs natural logarithm of relaxation time. It was observed that mineral s having high viscosity and density showed all three regimes of spreading namely capillary, gravity and viscous regimes. On the other hand, spreading behavior of mineral s having low viscosity and density on Inconel substrate consisted of capillary, gravity regimes and absence of viscous regime. The presence of viscous regime in spreading of MQ-3, MQ- 7 and MQ-8 indicates that relaxation of contact angle was almost completed. The absence of viscous regime in spreading of MQ-1, MQ-2, MQ-4, MQ-5 and MQ-6 indicates that spreading of was still active. The results clearly indicate that improved wettability and fast spreading kinetics for mineral s of low viscosity and surface tension while reduced wettability and slow spreading kinetics for mineral s of high viscosity and surface tension. Fig. 5. Fig. 6. (a) Contact angle relaxation and (b) spread area of quenchants droplet during spreading on Inconel 600 substrate. Plot of ln (droplet base radius) verses ln (time) during spreading of s Wetting Behavior The video images taken during the quenching of hot Inconel probe in mineral s are shown in Fig. 7. Due to the ISIJ

5 Fig. 7. Photographs of Inconel probe heated to 850 C quenched in different mineral s. dark nature of color, video imaging of the mineral s, MQ- 2, MQ-5 and MQ-6 were not possible. The thermal histories of Inconel probe measured at various axial and radial locations during quenching in different mineral s were used to determine nature of wetting front, wetting front velocity and rewetting temperature for all quenchants. Figure 8 shows typical time temperature data of Inconel probe measured during quenching in mineral. All mineral s show the formation of stable vapour film around the quench probe surface. Rewetting of the fluid begins after a time period at the bottom surface of quench probe and resulted in formation of wetting front. The wetting front then ascends to top. The time taken to start the rewetting was about 6.0, 4.4, 3.8, 2.7, 3.2, 3.8, 1.1, and 2.6 s for MQ-1, MQ-2, MQ-3, MQ-4, MQ-5, MQ-6, MQ-7 and MQ-8 respectively. Fast quenching s (MQ-4, MQ-5 and MQ-6) show less time to start the rewetting than conventional quenching s. However, it was observed that hot s show short time to start rewetting than fast and conventional s. The rewetting times at different radial locations were determined to assess the nature of wetting front. Figure 9 shows the variation of rewetting time at 30 mm in different radial locations of probe. It indicates that nature of wetting front was not uniform over the probe surface. To assess the nature of the wetting front, a wetting front uniformity parameter was defined as the difference between the maximum rewetting time and minimum rewetting time measured at 30 mm in different radial locations of probe. Wetting front uniformity parameters of about 2.6, 1.2, 1.6, 1.6, 1.7, 0.6, 0.9 and 0.4 s were obtained for MQ- 1, MQ-2, MQ-3, MQ-4, MQ-5, MQ-6, MQ-7 and MQ-8 respectively. A lower value of the wetting front uniformity indicates more uniform of nature of the wetting front. The nature of the wetting front was more uniform with mineral s of high viscosity and surface tension while less uniform with mineral s of low viscosity and surface tension. For example, MQ-8 having higher viscosity ( cp) and surface tension (42.28 mn/m) showed wetting front uniformity value of of about 0.4 s while MQ-1 having lower viscosity (23.56 cp) and surface tension (33.1 mn/m) showed rewetting time variation of about 2.6 s. The wetting front uniformity decreased monotonically with increase in the Fig. 8. Typical thermal histories of quench probe measured at different (a) axial locations and (b) radial locations during quenching in MQ-1. surface tension and viscosity of. The low viscosity and surface tension of mineral s offer improved wettability and faster spreading which expected to collapse of vapour film easily resulted in less uniformity in nature of wetting 2014 ISIJ 1430

6 Fig. 9. Variation of rewetting time at different radial locations of probe during quenching in various mineral s. front than higher viscosity and surface tension of mineral s which offer greater resistance to flow and collapse of vapour film. The video imaging of quenching process showed formation of the additional wetting front at the top surface of quench probe and started moving downwards for MQ-3, MQ-7 and MQ-8. On the other hand no additional wetting front was observed for MQ-1, MQ-4. The rewetting times (transition from vapour to nucleate bing stages) at different axial locations and the corresponding temperatures (rewetting temperature) of the probe were measured. The wetting front velocity was calculated by dividing the axial location by rewetting time at that location. Figure 10 shows variations of wetting front velocity and rewetting temperature on axial locations of quench probe for all mineral s. The velocity of the wetting front increases with distance from the bottom surface of probe. In the case of MQ-3, MQ- 6, MQ-7 and MQ-8, the maximum wetting front velocity was obtained at intermediate locations of probe surface. This is due to the formation of additional wetting front at the top of quench probe. The heat extraction at the metal/quenchant interface is by thermosyphon effect in the case of low viscosity quenching whereas by heating of thin layer of at the part surface in the case of high viscosity of quench. 7) The mineral s, MQ-1, MQ-2, MQ-4 and MQ-5 were low viscosity quenching s and flash/fire points of these s were also low. Heat extraction at the probe surface after collapse of vapour film in these mineral s are expected by thermosyphon effect which resulted in strong convection at the metal quench interface. This causes continuous movement of wetting front. This is also possible reason for less uniformity of nature of wetting front in low viscosity. Further, these mineral s showed improved wettability and fast spreading. Thus no additional wetting front formation in these mineral s. On the other hand, MQ-3, MQ-6, MQ-7 and MQ-8 were high viscosity quenching s and flash/fire points of these s were also high. Heat extraction at the probe surface after collapse of vapour film in these mineral s is expected by heating of thin layer of at the part surface. This causes slow movement of wetting front from the bottom of quench probe and more uniform nature of wetting front. At that same time, the increase in quenchant temperature resulting in lower viscosity of. This causes the formation of additional wetting front at top of quench probe which started moving downward. The schematic of the two wetting phenomena are shown in Fig. 11. The values of wetting front velocity and rewetting temperature measured at different axial locations (Fig. 10) were used to calculate the average wetting front velocity and rewetting temperature. Averages wetting front velocities of 7.85, 2.34, 3.88, 9.98, 15.68, 11.98, 9.61 and 6.97 mm/s were obtained for MQ-1, MQ-2, MQ-3, MQ-4, MQ-5, MQ-6, MQ-7 and MQ-8 respectively. The corresponding average rewetting temperatures were found to be 543, 625, 611, 676, 695, 708, 726 and 705 C respectively. Fast quenching s showed higher wetting front velocities than convention and hot quenching s. Fig. 10. Fig. 11. Variation of (a) wetting front velocity and (b) rewetting temperature on axial locations of quench probe surface during quenching in various mineral s. Schematic of rewetting phenomena in (a) low viscosity and (b) high viscosity mineral s quenching. This is due to the presence of accelerating additives (such as calcium naphthenate, alkenyl succinimide and sodium sulfonate etc.) in the fast mineral s. However, among the convectional/fast/hot mineral s, higher wetting front velocity were obtained for low viscosity while lower wetting front velocity was obtained for high viscosity. Further, the rewetting of the fluid occurs at higher temperature with increase in viscosity of the mineral Cooling Behavior Figure 12 shows cooling curves obtained at geometric centers during quenching of quench probe I and probe II respectively in the reference fluid before and after quenching experiments with mineral s. Maximum cooling rate differences of 3 and 5 C/s were observed for probe-i and probe-ii respectively. The results confirm the repeatability of experiments with Inconel probe. In order to compare the cooling performance of mineral s, the thermal histories measured at geometric center of Inconel probe were plotted. Figure 13 shows cooling and cooling rate curves measured at the geometric center of quench probe for all mineral s used in the present study. Cooling curves showed three stages of cooling namely vapour blanket, nucleate bing and convective cooling for all the quench media. Cooling rate was significantly higher in nucleate bing stage. The presence of accelerating additives in the fast quenching s ISIJ

7 Table 3. Cooling curve parameters determined for various quenchants. Critical cooling parameters Normal mineral Fast mineral Hot mineral MQ-1 MQ-2 MQ-3 MQ-4 MQ-5 MQ-6 MQ-7 MQ-8 CR peak ( C/s) T CRpeak ( C) CR 705 ( C/s) CR 550 ( C/s) CR 300 ( C/s) CR 200 ( C/s) t (s) Fig. 12. Fig. 13. Comparison of cooling curves of (a) probe-i and (b) probe-ii against reference before and after quenching experiments with mineral s. Cooling curves (thin line) and cooling rate curves (thick) line of quench probe at the geometric centre during quenching. resulted in short duration of vapour film compared to normal quenching s. However, hot s show even less duration of vapour film than fast s due to its higher flash points. On the other hand normal quenching s show longer vapour film stage due to its lower flash points. The duration of vapour film was found to be 8.00, 8.40, 6.90, 5.80, 5.20, 6.30, 2.70 and 4.80 s for MQ-1, MQ-2, MQ-3, MQ-4, MQ-5, MQ-6, MQ-7 and MQ-8 respectively. From the plots, the critical cooling curve parameters such as peak cooling rate (CR peak), temperature of the peak cooling rate (T CRpeak ), time to cool from 730 to 260 C (t ), cooling rates at 705 C (temperature at which austenite transformation starts to occur for the most of the carbon steels), 550 C (temperature which is at or near the nose of TTT curves for many steels), 300 C and 200 C (temperatures which are in the region of the martensitic transformation for many steels) (which were denoted as CR 705, CR 550,CR 300 and CR 200 respectively) were determined and are presented in Table 3. Higher cooling rates at critical temperatures and lower t were obtained for fast quenching s than hot s and conventional quenching s which indicating the fast cooling performance. Hot quenching s show higher cooling rates at peak cooling and 705 C than conventional quenching s. The cooling rates at 550 C, 300 C and 200 C of hot s and conventional quenching s were comparable. Further, it was observed that temperature at which peak cooling occurs and time to cool from 730 to 260 C of hot s were higher than fast and conventional quenching s. However, among the convectional/fast/hot mineral s, s having high viscosity showed lower cooling rates at critical temperatures and higher time to cool from 730 to 260 C while higher cooling rates at critical temperatures and lower time to cool from 730 to 260 C were observed for low viscosity s. Hardening power (HP) of s were determined using the following equations 6) HP = TVP CR 3. 85TCP... (7) where T VP temperature of film bing to nucleate bing transition ( C) CR cooling rate over the temperature range of 600 to 500 C ( C/s) T CP temperature of nucleate bing to convective cooling ( C) The HP values of 522, 589, 345, 1 067, 800, 619, 156 and 44 were obtained for MQ-1, MQ-2, MQ-3, MQ-4, MQ-5, MQ-6, MQ-7 and MQ-8 respectively. Even though hot quenching s showed higher cooling rates at peak cooling and 705 C than conventional quenching s, lower values of HP were obtained due to its high viscosity. Higher value of HP was obtained for low viscosity while lower value of HP was obtained for high viscosity Metal/quenchant Heat Flux The thermal histories measured at various axial and radial locations and thermo-physical properties of quench probe were input to the inverse heat conduction problem (IHCP) to estimate the spatial dependence of metal/quenchant heat flux transients for all mineral s quenching. The timetemperature data measured at different axial locations except at A8 (40 mm from the top surface) were input to the inverse program and the temperatures measured at A8 location was used to compare the estimated temperatures at the same location. Figure 14 shows a good agreement between the measured and estimated time temperature data at the A8 location. In the case of radial locations, all thermal histories measured were input to the inverse program. The overall error in estimated temperatures for the whole domain was calculated using the equation: 13) %Error in Estimated Temperatures = n 1 Tmeasured T estimated 100 n i= 1 Tmeasured i... (8) 2014 ISIJ 1432

8 Fig. 14. Measured and estimated temperature profile of quench probe-i at A8 location. Fig. 16. Table 4. Estimated metal/quenchant heat flux transients at (a) axial and (b) radial locations of probe surface during quenching in MQ-5. Estimated multiple peak heat flux components during vapour blanket stage. Probe Quench medium q vapour (kw/m 2 ) q 1 q 2 q 3 q 4 q 5 q 6 q 7 q 8 Remarks MQ Normal Fig. 15. Overall % error in estimated temperatures of (a) probe-i and (b) probe-ii. where n is the number of unknown heat fluxes assigned at the quench probe surface. Figure 15 shows overall % error in the estimated temperatures obtained from the solution of IHCP. The maximum overall % errors in the estimated temperatures of axial and radial locations were found to be less than 5%. Figure 16 shows typical spatially dependent metal/quenchant interfacial heat flux transients estimated at axial and radial locations of the quench probe surface. Heat flux curve showed initial peak followed by decrease in heat flux to a value. This is due to initial wetting of liquid and subsequent vaporization resulted in formation of vapour around the probe surface. Heat flux value then increased sharply to higher value due to collapse of vapour film and formation of bubble bing on the quench probe. Thereafter, it starts to decrease with surface temperature of the probe. Tables 4 and 5 show the estimated peak heat fluxes during vapour and nucleate bing stage respectively in the axial as well as radial locations. The average peak heat flux values of Probe-I (axial variation) Probe-II (radial variation) MQ mineral MQ MQ Fast MQ mineral MQ MQ-7 MQ Hot mineral MQ Normal MQ mineral MQ MQ Fast MQ mineral MQ MQ-7 MQ Hot mineral 623, 482, 597, 614, 549, 519, 713 and 545 kw/m 2 were obtained for MQ-1, MQ-2, MQ-3, MQ-4, MQ-5, MQ-6, MQ-7 and MQ-8 respectively during film bing. The corresponding peak heat flux values during nucleate bing were found to be 1 292, 1 309, 1 233, 2 138, 2 031, 1 899, and kw/m 2 respectively. Higher values of peak heat flux were obtained for fast quenching s and lower values of peak heat flux were obtained for conventional quenching. Hot s show intermediate peak heat flux ISIJ

9 values. The variations of heat flux values at critical temperatures 705, 550, 300 and 200 C at different locations of quench probe surfaces for all quenchants were determined. Figure 17 show typical spatially dependent heat flux at critical temperatures in axial and radial locations on the probe surface during quenching. The plots show heat transfer during quenching was not uniform over the surface of probe. For example, standard deviation of heat flux values at 705 C in axial locations of the probe were found to be 162, 138, 147, 236, 274, 197, 93, 81 kw/m 2 for MQ-1, MQ-2, MQ-3, MQ- 4, MQ-5, MQ-6, MQ-7 and MQ-8 respectively. Heat flux variations at axial locations were found to be higher than at radial locations for all quenchants. The heat flux variations at 705 and 550 C were found to be higher than at 300 and 200 C. It indicates that heat transfers during liquid cooling stage was more uniform than film and nucleate bing stage. Further, no definite relation between heat flux variation and properties of was found. However, lower standard deviations of heat flux values were observed for hot s having high viscosity indicating more uniform heat transfer than conventional and fast mineral s. The amount of heat removed during quenching was determined by plotting the integral heat flux curve for all quench media. The amounts of heat extracted by the quenchants to cool the probe from 850 to 200 C were found to be 9.45, 9.49, 9.75, 9.43, 9.41, 9.46, 9.44 and 9.57 MJ/m 2 for MQ-1, MQ-2, MQ-3, MQ-4, MQ-5, MQ-6, MQ-7 and MQ-8 respectively. Figure 18 shows the average time for the known fraction of heat removal during quenching in mineral s. The time for known fraction of heat removal was lower for fast quenching mineral s indicating superior heat extracting capability compared to conventional and hot quenching s. The hot s quenching showed lower time up to about 65% heat removal and higher time for remaining percent heat removal compared to conventional. Cooling curve analyses parameters also showed higher cooling rates at peak cooling and 705 C and comparable cooling rates at 550 C, 300 C and 200 C for hot s compared to conventional quenching s. Further, the times to cool from 730 to 260 C were higher for hot s. This is the reason for low HP values of hot s quenching though it showed higher peak heat flux values and peak cooling rates than conventional. It indicates that heat extracting capability of the hot was faster during vapour and nucleate bing and slower during liquid Fig. 18. Time for specified heat removal from the quench probe. Fig. 19. Experimentally measured cooling curves of mineral s superimposed on CCT curve of AISI 1040 steel. Table 5. Estimated multiple peak heat flux components during nucleate bing stage. Probe Quench medium q nucleate (kw/m 2 ) q 1 q 2 q 3 q 4 q 5 q 6 q 7 q 8 Remarks Fig. 17. Typical spatially dependent heat flux at critical temperatures in (a) axial and (b) radial locations on the probe surface during quenching in MQ-5. Probe-I (axial variation) Probe-II (radial variation) Normal MQ mineral MQ MQ MQ Fast mineral MQ MQ-7 MQ Hot mineral MQ Normal MQ mineral MQ MQ Fast MQ mineral MQ MQ Hot mineral MQ ISIJ 1434

10 Table 6. Resultant micro constituents (in volume%) of AISI 1040 steel for mineral s quenching. Micro constituent Normal mineral Fast mineral Hot mineral MQ-1 MQ-2 MQ-3 MQ-4 MQ-5 MQ-6 MQ-7 MQ-8 Ferrite Pearlite Bainite Martensite cooling stage. To quantify the effect of cooling performance of quench media on the formation of micro-constituents and hardness in steel, cooling curves measured at geometric centre of the probe were superimposed on the CCT diagram. It should be noted that cooling curve analysis was carried out using Inconel 600 probe. However, superimposition of cooling curve on the steel CCT diagram can be used to quantify the effect of cooling performance of quench media. Figure 19 shows the CCT diagram of AISI 1040 obtained using JMatPro software (Sente Software Ltd., UK). The resultant microconstituents of steel are given in Table 6. Fast mineral quenching resulted in higher amount of transformed products (bainite and martensite) than normal and hot s. Further, normal quenching resulted in higher amount of transformed product than hot s. It indicates that higher quenching severities for fast mineral s, lower quench severities for hot s and intermediate quench severities for normal s. With varying cooling rates obtained with mineral quenchants used in the present work, the resultant hardness of the AISI 1040 steel estimated from the CCT diagram was found to be 277, 271, 266, 288, 289, 273, 268 and 261 HV for MQ-1, MQ-2, MQ-3, MQ-4, MQ-5, MQ-6, MQ-7 and MQ-8 respectively. 4. Conclusions The following conclusions were drawn based on the results and discussion. (1) Higher contact angles on Inconel substrate were obtained for mineral s of high viscosity and surface tension whereas mineral s of low viscosity and surface tension showed lower contact angle. It indicates that wettability of mineral improved with a decrease in viscosity and surface tension of mineral. Similarly, relaxation of contact angle and spreading area of droplet on Inconel substrate indicated faster spreading kinetics for mineral s of low viscosity and surface tension and slower spreading kinetics for mineral s of high viscosity and surface tension. (2) Spreading behavior of mineral s having high viscosity and density showed all three regimes of spreading namely capillary, gravity and viscous regimes. On the other hand, spreading behavior of mineral s having low viscosity and density on Inconel substrate consisted of capillary, gravity regimes and absence of viscous regime. (3) Hot quenching showed less time delay to start rewetting phenomenon while longer time was taken to start rewetting with conventional quenching. Fast s quenching showed intermediate time to start rewetting of the fluid. (4) The nature of the wetting front was more uniform with mineral s of high viscosity and surface tension while less uniform with mineral s of low viscosity and surface tension. (5) Mineral s of high viscosity and flash point showed the formation of additional wetting front at the top of quench probe which started moving downward during rewetting of fluid. On the other hand, no additional wetting front was observed for mineral s of low viscosity and flash point (6) Among the convectional/fast/hot mineral, higher wetting front velocity was obtained for low viscosity while lower wetting front velocity was obtained for high viscosity. Further, the rewetting of the fluid occurs at higher temperature with an increase in viscosity of the mineral. (7) Higher cooling rates at critical temperatures and lower t were obtained for fast quenching s than hot s and conventional quenching s. Hot quenching s showed higher cooling rates at peak cooling and 705 C than conventional quenching s. The cooling rates at 550 C, 300 C and 200 C of hot s and conventional quenching s were comparable. (8) Higher values of peak heat flux were obtained for fast quenching s and lower values of peak heat flux were obtained for conventional quenching. Hot s show intermediate peak heat flux values. (9) Among the conventional/fast/hot s, having higher viscosity showed more uniform heat transfer than low viscosity. (10) Fast quenching mineral s showed faster heat extracting capability compared to conventional and hot quenching s. The cooling performance of the hot was better during vapour and nucleate bing stages and slower during liquid cooling stage compared to conventional quenching s. Acknowledgement One of the authors (KNP) gratefully acknowledges the financial support provided by the Science and Engineering Research Board (SERB), Department of Science and Technology (DST), Government of India, New Delhi, India under a R&D project. REFERENCES 1) B. Liscic, H. M. Tensi, L. C. F. Canale and G. E. Totten: Quenching Theory and Technology, CRC Press, Boca Raton, FL, (2010). 2) B. Liscic: Quenching and Carburising, ed. by P. D. Hodgson, The Institute of Materials, London, (1993), 1. 3) H. M. Tensi, A. Stich and G. E. Totten: Steel Heat Treatment, ed. by G. E. Totten, CRC Press, Boca Raton, FL, (2006), ) H. E. Boyer and P. R. Cary: Quenching and Control of Distortion, ASM International, Materials Park, OH, (1988). 5) S. Ma: PhD Thesis, Worcester Polytechnic Institute, Worcester, USA, (2002). 6) G. E. Totten, C. E. Bates and N. A. Clinton: Handbook of Quenchants and Quenching Technology, ASM International, Materials Park, OH, (1993). 7) C. E. Bates, G. E. Totten and R. L. Brennan: ASM Handbook - Heat Treatment, Vol. 5, ASM International, Materials Park, OH, (1991), ) G. E. Totten, H. M. Tensi and L. C. F. Canale: Proc. of 22nd Heat Treating Society Conf. and 2nd Int. Surface Engineering Cong., ASM International, Materials Park, OH, (2003), ) S. Ma, M. Maniruzzaman and R. D. Sisson, Jr.: Proc. of 1st ASM Int. Surface Engineering and 13th IFHTSE, ASM International, Material Park, OH, (2000), ) S. Asada and K. Fukuhara: Proc. of 20th Heat Treating Society Conf., ASM International, Materials Park, OH, (2000), ) H. Yokota, H. Hoshino, S. Satoh and R. Kanai: Proc. of 20th Heat Treating Society Conf., ASM International, Materials Park, OH, (2000), ) P. Fernandes and K. N. Prabhu: J. Mater. Sci. Eng. B, 3 (2013), No. 2, ) T. S. P. Kumar: Numer. Heat Tr. B-Fund., 45 (2004), ) R. N. Penha, L. C. F. Canale, G. E. Totten, G. S. Sarmiento and J. M. Ventura: J. ASTM Int., 3 (2006), No. 5, JAI ISIJ