ISIJ International
Online ISSN : 1347-5460
Print ISSN : 0915-1559
ISSN-L : 0915-1559
Fundamentals of High Temperature Processes
Formation Mechanism of Sintering Crack in Mo–W–ZrO2 Sensor for Molten Steel Temperature Caused by Water Vapor
Jiu Zhang Qixian XieGuohui MeiZhizhong Mao
著者情報
ジャーナル オープンアクセス HTML

2022 年 62 巻 4 号 p. 626-631

詳細
Abstract

Mo–W–ZrO2 cermet, which is usually sintered in H2, has been used for molten steel temperature sensor. However, the crack caused by water vapor during its sintering limits the industrial application. Thus, the mechanism of crack formation has been investigated. Cracks are mainly caused by a significant volume expansion which is attributed to that the metal powders of the green body of cermet are oxidized by water vapor. The obvious expansion of approximately 17% in diameter occurs below 1200°C, reducing the bonding strength. Meanwhile, the support frame for the sintering restricts the expansion. Both of them are the main reasons for the crack formation. At higher temperature, the volume of above oxidized metal powders decreases caused by the reduction of H2. The volume of the metal powders increases first and then decreases, which leads to a more difficult sintering densification. Moreover, the ceramic powders are sintered earlier and form a skeleton. As a result, the sintering densification is restrained and the pores form. As the water vapor is eliminated through the H2 filling and pumping vacuum, the problems of cracks and pores are addressed, which is helpful to solve the main contradiction in industrial application of the Mo–W–ZrO2 sensor.

1. Introduction

Mo–W–ZrO2 cermet has potential applications in the metallurgical industry. It solves two key problems of thermal shock resistance and corrosion resistance for metallurgical applications (such as nozzle, stopper and molten steel temperature sensor).1,2,3) Furthermore, it has a much higher strength and lower porosity than those of the traditional refractories (such as Al2O3–C, MgO–C and Al2O3–ZrO2–C).4,5,6,7) Addition of the ductile metal phase exhibits a toughening effect and improves the thermal shock resistance significantly. As a thermal shock fracture occurs, the metal phase bridges the crack interface of the brittle ceramic phase to prevent the crack from propagating.8,9,10) Simultaneously, the plastic deformation of the metal phase occurs, which consumes the fracture energy and further prevents the crack from propagating. In addition, the ceramic ZrO2 phase solves the corrosion issue, because the ZrO2 cannot be wetted and dissolved by molten steel.11,12,13) As a result, the wall thickness of a Mo–W–ZrO2 sensor for molten steel temperature can be prepared by only 3 mm. By contrast, the wall thickness of the traditional Al2O3–C sensor is 20–30 mm.2,3,14) As the thin-walled sensor is heated by molten steel, it can form a blackbody cavity faster and emits a stable infrared thermal radiation according to molten steel temperature. An infrared probe receives the thermal radiation and calculates the temperature.

Many investigations into oxide-metal cermets have focused on compositions to address the problems of thermal shock resistance and corrosion resistance. Various metal phases (including Mo, W, Ni, Ti and Nb) have been investigated for toughening the ceramic phase.8,9,10,15) Furthermore, the corrosion behavior of the cermet (such as Mo–MgO and Mo–ZrO2) has been investigated.16,17)

However, the manufacturing process that also has a significant impact on performance has been relatively less investigated. Sintering is a particularly critical process in the manufacture processes of Mo–W–ZrO2 sensor, which transforms the green body into compact cermet with metallurgical bonding. The green body that is only bonded by surface friction force between the powders after the cold isostatic pressing (CIP) has a poor bonding strength. Generally, the cermet is sintered in H2 to prevent the metal powder from being oxidized.

Before sintering, air is removed by H2 replacement. Although air can be easily replaced before sintering, the water absorbed by the furnace is difficult to remove and leads to cracks of the cermet. The water mainly originates from the heat insulation cotton and corundum bricks in the furnace that have previously absorbed moisture from air. It can be turned into water vapor as the furnace is heated, part of which can be removed with H2. However, the other part of the water vapor will be condensed as the furnace wall is cooled by cooling water at low temperature. Furthermore, the furnace absorbs new water from air. This cracking issue becomes more serious in summer owing to a higher air humidity, which has become the main contradiction, restricting the industrial application of the Mo–W–ZrO2 sensor.

The aim of this work is to elucidate the mechanism of crack formation in Mo–W–ZrO2 cermet sensor caused by water vapor during sintering. Initially, the cracking of the cermet green body caused by water vapor has been investigated. Concretely, the chemical thermodynamic of the interaction between the cermet and water vapor and the composition evolution have been investigated. Secondly, a method for removing water vapor has been proposed and the cermet sensor with a good microstructure has been sintered. Furthermore, the Mo–W–ZrO2 sensor has been used to measure molten steel temperature in a steelmaking mill for an application test.

2. Experimental Procedure

2.1. Raw Materials

Mo–W–ZrO2 sensor was prepared by the powder metallurgy method. Its composition was 30 wt.% Mo, 35 wt.% W and 35 wt.% ZrO2. The raw materials comprised the following powders: (1) calcium-stabilized zirconia (synthesized by the electrofusion method, purity: 98%, Astron Ltd, Shenyang, China) with an average particle size d50 of 45 μm. (2) 3Y-ZrO2 (synthesized by the chemical doping method, 10 wt.%, purity: 99.7%, specific surface area: 10 m2/g, L. Feng Ltd, Jiaozuo, China) with an average particle size d50 = 50 nm. This is partially stabilized zirconia and is used to improve the thermal shock resistance through the phase transformation toughening.18) (3) Molybdenum (purity: 99.5%, Cemented Carbide Corp., Ltd, Chengdu, China) with an average particle size d50 = 3–4 μm. (4) Tungsten (purity: >99.9%, Cemented Carbide Corp., Ltd, Chengdu, China) with an average particle size d50 = 2.5–3 μm.

2.2. Sample Preparation

The cermet sensor samples were tubes with hemispheric bottoms (diameter × thickness × length: Φ25 mm × 3 mm × 100 mm). Their manufacturing processes were conducted as follows: (1) The initial raw material powders were placed in a polyurethane pot and ball milled for 48 h with WC grinders; (2) the powders were molded into a tube by CIP at 180 MPa; (3) the sensor was sintered in an H2 atmosphere (purity: 99.99%, flux: 50 mL/min, pressure: 0.2 MPa) at 1800°C for 1 h with a heating rate of approximately 100°C/h. The temperature was measured by a W-Re thermocouple that was close to the middle layer of the furnace. The cermet sensor samples were located in the top, middle and bottom layers of the furnace, with a distance of 250 mm. The temperatures at the upper layer and the bottom layer were about 1820°C and 1785°C respectively, which were measured by temperature measuring rings (RTC 1600–1900°C, Hanqun Electronic Materials Co., Ltd, Suzhou, China).

During the sintering of the sensor, the H2 entered the furnace from the top and left from the bottom, as illustrated in Fig. 1. The water absorbed by the furnace was heated and converted into water vapor, leading to crack. The water vapor was observed indirectly as follows: Some part of water vapor was removed with H2 during sintering and would be condensed and observed in a water vapor condenser bottle.

Fig. 1.

Sintering furnace for the industrial production of cermet sensor. (Online version in color.)

2.3. Property Characterization

To investigate the crack formation of cermet sensor caused by water vapor, microstructure evolution, chemical thermodynamics analysis and X-ray diffraction (XRD) analysis were used. The microstructure was checked by using scanning electron microscopic (SEM, SSX-550). And the chemical thermodynamics was analyzed using the HSC 6.0 thermodynamics analysis software package.19) Besides, the crystalline-related property was characterized by an X-ray diffractometer (PW3040/60). It provided CuKα radiation and its step interval was set at 0.02°.

3. Results and Discussion

3.1. Crack Formation Caused by Water Vapor

As the water vapor is generated inside the furnace during the sintering, the cermet sensors sintered at 1800°C show an obvious crack problem, as illustrated in Fig. 2(a). Meanwhile, about 180 mL liquid water is condensed and observed in the water vapor condenser bottle, reflecting the water vapor concentration in the furnace indirectly.

Fig. 2.

(a) Cermet sensors sintered in the presence of water vapor, (b) microstructure (SEM) of the 1# sensor, P: pore, C: ceramic phase, M: metal phase. (Online version in color.)

It can be also found that the cermet sensors sintered at different layers in the furnace exhibit different degrees of crack. At the bottom layer, the 1# and 2# sensors have the most obvious cracks. Their microstructure shows many large size pores, which is illustrated in Fig. 2(b) (microstructure of the 1# sensor). At the middle layer, the 3#–5# sensors exhibit less obvious cracks. At the top layer, only small cracks or no cracks of the 6#–8# sensors are generated. Although the 8# sensor has no cracks, it has many obvious pores, as shown in Fig. 3. Compared with the 1# sensor, the pores of the 8# sensor are much smaller and fewer.

Fig. 3.

Pores in the microstructure of the #8 sensor (SEM) caused by the water vapor.

The above different crack degrees of the cermet sensors are associated with the different water vapor concentrations at different layers in the furnace during the sintering. This may be due to the fact that water vapor has a much higher density compared with H2. As a result, the cermet sensor that is sintered at the bottom layer in the furnace exhibits a more severe crack problem.

3.2. Mechanism of Crack Formation

To clarify the mechanism of cracks formation, the chemical thermodynamic of the interaction between cermet and water vapor has been discussed. There are several oxidation reactions between the metal powers and water vapor, which are the basic reasons for the crack problem. The Gibbs free energy changes (ΔG°) from the oxidation reactions (Reactions (1)–(4)) of the metals (Mo and W) are calculated. The results are illustrated in Fig. 4.   

1/3Mo+ H 2 O(g)1/3Mo O 3 + H 2 (g) (1)
  
1/3W+ H 2 O(g)1/3W O 3 + H 2 (g) (2)
  
1/2W+ H 2 O(g)1/2W O 2 + H 2 (g) (3)
  
1/2Mo+ H 2 O(g)1/2Mo O 2 + H 2 (g) (4)
Fig. 4.

Gibbs free energy changes (ΔG°) of the oxidation reactions of the metals (W and Mo). (Online version in color.)

Due to the fact that ΔG° for the reaction (1) which produces MoO3 is > 0 in the temperature range of 0–1800°C, MoO3 is not produced. ΔG° for the oxidation reaction (2) which produces WO3 is < 0 when the temperature is less than 824°C, meaning that this reaction can proceed in this temperature range. Below approximately 1104°C, ΔG° for the oxidation reactions (3)–(4) which produce WO2 and MoO2 are also < 0, indicating that WO2 and MoO2 can be produced.

The above oxidation reactions will lead to an obvious volume increase of the metal powders. According to the change about density and weight of the metals (W and Mo) caused by oxidation, the changes of their volumes are illustrated in Table 1 when the metal powders are oxidized completely. The density and molar mass are derived from the HSC 6.0 thermodynamics analysis software package. It can be found that their volumes will increase to approximately 2.1 times (W→WO2 and Mo→MoO2) and 3.4 times (W→WO3) below 1104°C, respectively. In other words, these oxidation reactions will lead to a significant volume expansion of the green body.

Table 1. Volume variation of the metals before and after oxidation.
WMoWO2MoO2WO3W→WO2Mo→MoO2W→WO3
Density g/cm319.3510.210.86.477.167.16
Molar mass g/mol183.8495.94215.85127.94231.85231.85
1 mol volume cm39.509.4119.9919.7732.38
Volume variation9.50→19.99
(2.1 times)
9.41→19.77
(2.1 times)
9.50→32.38
(3.4 times)

To investigate the expansion caused by above oxidation reactions, the green body variation of the cermet has been investigated in the sintering process. It can be found from Fig. 5 that the green body sintered at 1200°C has an obvious expansion of approximately 17% in diameter (25 mm→29.3 mm). Meanwhile, the obvious crack is also observed. In addition, the weight change is checked to verify the occurrence of oxidation reaction. The weight of the green body of the cermet increases from 135.0 g to 150.2 g. These are largely consistent with the chemical thermodynamic analysis, implying that the cracks of the cermet sensors sintered at 1800°C already exist at 1200°C.

Fig. 5.

Expansion and cracks of cermet sensor sintered in the presence of water vapor at 1200°C: (a) bottom, (b) top. (Online version in color.)

In order to further clarify the oxidation mechanism, the XRD analysis has been used. After the cermet sensor is sintered in the presence of water vapor at 1200°C, its compositions have been checked by the XRD analysis (see Fig. 6). It can be found that the reflection peaks of W and Mo, MO2 and WO2 are near respectively, which may be due to the fact that they have same crystal structure and similar lattice parameters.20) Meanwhile, WO3 is not found while the W and Mo still exist, which may be attributed to that the metal is not oxidized completely and the reaction to produce WO3 is not sufficient to occur. In summary, the oxidized products are WO2 and MoO2.

Fig. 6.

XRD patterns of the sample oxidized by water vapor (B, 1200°C) and original sample (A). (Online version in color.)

It can be reasoned that the crack generation proceeds via the following stages. (1) As the water vapor is generated by the furnace during the sintering, it oxidizes the metal powders of the green body of cermet sensor, causing a significant volume increase of the metal powders. (2) The expansion reduces the bonding strength of the green body. And the support frame for the cermet sintering restricts the expansion. Both of them are the main reasons for the crack formation. (3) As the expansion may not be enough to cause cracks, it can also cause many pores. Specifically, at higher temperature, the volume of above oxidized metal powders will decrease caused by the reduction of H2, but the volume variation increases the distance between the metal powders of the cermet sensor. This leads to that the sintering densification of the metal powders is more difficult and occurs later than that of the ceramic powders. Meanwhile, the ceramic powders that are sintered earlier and more easily form a skeleton before the sintering densification of the metal powders. These restrain the sintering densification and further lead to pores formation in Fig. 3.

3.3. Elimination of Cracks

In brief, the expansion caused by water vapor is the key to form the crack. To remove the water vapor means to eliminate the crack. According to the analysis of chemical thermodynamic, the volume expansion caused by water vapor occurs below 1104°C approximately, so the water vapor should be removed in this stage. However, the water absorbed by the heat insulation cotton and corundum bricks of the furnace is difficult to remove through a passive method by heating the furnace.

Therefore, a new active method of filling H2 and pumping vacuum is proposed to remove water vapor, which is illustrated in Fig. 7. A new sintering furnace for removing water vapor is designed and made, as shown in Fig. 8. With the protection of H2, the furnace is heated and vacuumized synchronously as the temperature is increased to 1200°C. This temperature is higher than the temperature analyzed by chemical thermodynamic (1104°C) to ensure that the green body of cermet will not be oxidized by water vapor as much as possible. Above 1200°C, the furnace is still filled with H2 to reduce the oxide film on the surface of metal powders of raw material and protect metals from oxidation. Raw materials of metal powders may be slightly oxidized by air in the process of storage and transportation, forming oxide film on the surface.

Fig. 7.

Method and processes for removing the water vapor.

Fig. 8.

A new sintering furnace for removing the water vapor. (Online version in color.)

As the water vapor is removed, the metal powders will not be oxidized and their volume does not initially increase and then decrease. Under these conditions, the cracks of the cermet sensors caused by water vapor are no longer observed. It can be found from Fig. 9(a) that all of the cermet sensors sintered in the whole furnace exhibits no cracks. Figure 9(b) shows that these cermet sensors exhibit good microstructure. The XRD patterns of these cermet sensors are same to the XRD patterns of the original sample that has been illustrated in Fig. 6 (curve A). The reflection peaks of MO2 and WO2 disappear. Compared with Figs. 2(b) and 3, this microstructure only exhibits a small number of small pores. This is due to the fact that the sintering densification of the cermet is not restrained by expansion, and the pores caused by water vapor are no longer observed. These remaining small pores may be attributed to that the sealed gas inside the cermet cannot be eliminated completely during sintering.

Fig. 9.

(a) Cermet sensors sintered in the absence of water vapor, (b) their microstructure (SEM), P: pore, C: ceramic phase, M: metal phase. (Online version in color.)

3.4. Application Test

The cermet sensors have been prepared in the absence of water vapor. Their performance has been tested in a steelmaking mill to measure the molten steel temperature. As a reference, the 8# cermet sensor is also used, which exhibits no cracks but is sintered in the presence of water vapor. The test is designed as inserting these cermet sensors into molten steel suddenly, withstanding an intense thermal shock from room temperature to approximately 1500°C.

Figure 10(a) shows that the structure of the cermet sensor sintered in the absence of water vapor exhibits good condition without damage after the test. On the contrary, the cermet sensor sintered in the presence of water vapor exhibits a thermal shock fracture, as illustrated in Fig. 10(b). This may be due to the existence of these large size pores in Fig. 3, leading to a decrease in high temperature strength and thermal shock resistance.

Fig. 10.

Performance test of the cermet sensors: (a) sintered in the absence of water vapor, (b) sintered in the presence of water vapor. (Online version in color.)

4. Conclusions

The crack formation mechanism of Mo–W–ZrO2 sensor for molten steel temperature during the sintering has been investigated. The crack formation processes are as follows:

(1) The cracks are mainly caused by a significant volume expansion that is attributed to that the metal powders of the green body of cermet sensor are oxidized by water vapor. This expansion of approximately 17% in diameter occurs below 1200°C during the sintering, which reduces the bonding strength of the green body. Meanwhile, the support frame for the sensor sintering restricts the expansion. Both of them are main reasons for the crack formation.

(2) As the expansion is not enough to cause cracks, it also leads to many pores. At higher temperature, the volume of above oxidized metal powders decreases caused by the reduction of H2. It increases the distance between the metal powders and leads to a more difficult sintering densification of the metal powders. Moreover, the ceramic powders are sintered earlier and more easily, forming a skeleton. These restrain the sintering densification of the cermet sensor and further lead to pores formation.

In brief, the expansion caused by water vapor is the key reason to form crack. A new method by filling H2 and pumping vacuum has been proposed to remove the water vapor. As the water vapor is eliminated through the H2 filling and pumping vacuum (10 Pa) below 1200°C, the problems of cracks and pores are addressed, which has solved the main contradiction in industrial application of the Mo–W–ZrO2 sensor for molten steel temperature.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (Nos. 61903074 and 61873051), Fundamental Research Funds for the Central Universities (No. N2104012), and Taihe Metallurgical Measurement and Control Technologies Co., Ltd, China (No. 2018021000036).

References
 
© 2022 The Iron and Steel Institute of Japan.

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs license.
https://creativecommons.org/licenses/by-nc-nd/4.0/
feedback
Top