14 July/August 2004 Refractories Applications and News, Volume 9, Number 4 Feature Article. . . KEY PROPERTIES FOR THE OPTIMIZATION OF REFRACTORY CASTABLE DRYING M. M. Akiyoshi, F. A. Cardoso, M. D. M. Innocentini and V. C. Pandolfelli, Department of Materials Engineering, Federal University of São Carlos 13565-905, S. Carlos, SP, Brazil, vicpando@power.ufscar.br Drying is one of the most complex steps in refractory castable processing due to the considerable risk of damage or explosive spalling during the first heat-up. To provide a basis to optimize the drying step of refractory castables, this work correlates the mechanical strength and permeability with the mass loss rate and surface temperature profiles of high-alumina, ultra-low cement castables cured at different temperatures. 1. INTRODUCTION The increasing demand for better and cheaper pre-cast refractory products has motivated the search for safer and shorter drying schedules in the refractory industry. However, drying is still a cru- cial step in castable processing, because improper heating sched- ules may lead to mechanical damage or even explosions when the tensile stress generated by pressurized vapor inside the refractory exceeds the material’s mechanical strength. Low curing temperatures of less than 20ºC [1-8], followed by fast heating rates, have been identified as one of the most important factors to promote spalling. More comprehensive analyses have revealed that the curing temperature and its time affect permeabil- ity and mechanical strength, both associated with the likelihood of spalling [2, 3]. This work therefore aims to correlate the perme- ability and mechanical strength with the mass loss rate and surface temperature profiles of high-alumina, ultra-low cement castables cured at 10ºC and 50ºC, in order to provide the basis for a better understanding of this important step in the production of refracto- ry castables. 2. EXPERIMENTAL PROCEDURE The castables tested here consisted of high-alumina, ultra-low cement (2 wt%), containing 4.50 wt% of water (dry basis). The particle size distribution was adjusted to a theoretical curve, based on Andreasen’s packing model with a coefficient of distribution of q=0.21 (maximum aggregate size: 4.5 mm). Alcoa S.A. supplied all the raw materials, and further details of the composition are given elsewhere [3]. Castable samples having a 7.5 cm diameter and 2.5 cm thickness were prepared for the permeability tests, while those for the mechanical strength and drying tests were cast in the shape of 4 cm x 4 cm cylinders. The samples subjected to permeability and mechanical strength evaluations were pre-dried in silica gel at the curing temperature [7]. To evaluate the surface temperature, a thin K-type thermocouple (0.2 mm diameter) was inserted near the surface (1 mm depth) to record the sample’s actu- al heating profile. The castables were cured at 10ºC or 50ºC (rela- tive humidity of ~ 100%) for 48 h in a climatic chamber (Vötsch). Considering the same length of time [7], low curing temperatures provide a smaller degree of hydration than higher temperatures. Therefore, the castable samples were cured up to 16 days, allowing the ones cured at 10ºC and those cured at 50ºC to be exposed to a similar degree of hydration. Mechanical strength was evaluated through a splitting test [9], at a loading rate of 42 N/s to keep the stress rate within a range of 690-1380 kPa/min. The splitting tensile strength was calculated by: where P (N) is the maximum load, and d (mm) and h (mm) are the samples’ diameter and height, respectively. Air permeability at room temperature was evaluated based on Forchheimer’s equation [10] for compressible fluids and adjusting the k 1 (Darcian) and k 2 (non-Darcian) constants: where P i (Pa) and P o (Pa) are, respectively, the absolute air pres- sures at the entrance and exit of the sample, v s (m/s) is the fluid velocity, L (m) is the sample’s thickness, μ (Pa•s) is the air viscos- ity and ρ (kg/m 3 ) is the air density, evaluated for P o = 690 mmHg (92•10 3 Pa) and T = 25ºC. Dewatering tests were performed in a thermogravimetric appara- tus [11] consisting of a digital scale (400±0.001 g) coupled to a fur- nace (maximum working temperature: 1000ºC). The tendency for explosive spalling was evaluated at a heating rate of 20ºC/min, while the drying profiles were performed at 10ºC/min. Mass loss during drying was assessed through the parameters W and W d , defined as: where M is the instantaneous mass recorded at time t i during the heating stage, M o is the initial mass and M f is the final (dry) mass of the tested sample. The variable W evaluates the cumulative frac- tion of water expelled during the heat-up per total amount of water initially present in the body (W varied from zero to 100% during the test), while W d relates the water loss to the dry weight of the body (in the range of 4.5% in the composition studied). The mass loss and surface heating rates were then evaluated using equations (5) and (6), respectively.