TUNING CASTABLE MIXES BY CHOOSING PROPER CHEMICAL ADDITIVES Joachim v. Seyerl*, BASF Construction Polymers GmbH, 83308 Trostberg, Germany
INTRODUCTION Refractory castables are hydratable and / or ceramically hardened materials withstanding temperatures in the upper 1000 °C range in contact with molten metals or slag. Improving the properties of castable mixes for refractory applications is an iterative process. It starts with choosing raw material combinations according to the application requirement, adding a binder system, water and chemical additives for processing. Workability and placement properties are very often related to the mix design. Water demand of raw materials and especially their selected particle size distribution determine the basic behavior. Chemical additives can help to improve the overall handling during placement. In the next 20 minutes I will concentrate on the two marked bullets (Workability / Placement and addition of Water / Chemicals) which deal with the chemistry at ambient temperature before ceramisation. The use of phosphates, citric acid or polyacrylates for dispersing and water reduction in castable or ceramic mixes is known since long. These products work because of their electrostatic dispersing mechanism. This presentation deals with polycarboxylate ethers sold under the BASF trade name Castament. They can be used as chemical additives in high performance mix designs where small amounts offer big effects. Dispersing Mechanism The new molecules consist of a backbone having anionic carboxylic groups responsible for electrostatic interaction. Additionally the molecules have included nonionic polyether side chains which are responsible for sterical interaction. The molecules may be described as a round brush where anionic and steric segments have been combined to an all-in-one dispersant with electro-sterical dispersing mechanism. The molecules can be designed in different sizes in terms of overall length of the backbone as well as the side-chains. It is also state of the art to modify the molecules by adding more or less side-chains thus influencing the charge density of the anionic backbone. A model for the interaction of polycarboxylate ethers with particle surfaces is shown in Fig. 1: Whenever such a molecule gets close to a cationic particle surface it will dock on to it by means of the backbone related anionic carboxy groups and will leave the outer particle sphere covered with nonionic, surfactant-like parts. This – let us call it “hedgehog mechanism” - will give a
creamy rheology to a high solids suspension of well dispersed or deflocculated inorganic particles. Whenever such a new dispersant molecule gets close to an anionic or nonionic surface it will dock onto the surface by means of a hydrogen bridging mechanism related to the polyether side chains leaving the outer sphere of a particle covered with anionic charges. Let us call this “octopus mechanism”.
Fig. 1. Model for Dispersing Particles with PCE. For many castables like a typical high alumina microsilica-free LCC it is essential to add a water amount as low as possible to receive a decrease in porosity as well as an increase in density which will add enhanced durability to the refractory shape. Adding a small amount of water to a mix design will give a crumbly mass not very good to handle and to place. Adding a very small amount of a polymer like Castament to the same amount of water containing mix will enable to launch a vibratable, free flowing, or pumpable mix with all advantages for downstream usage. Examples will be presented how “Hedgehog” or “octopus mechanism” may decide over compatibility between dispersant and matrix fines interaction. Tuning PCE Additives basically have three effects when incorporated in an aqueous castable mix design containing for example CAC, Reactive Alumina and sometimes Microsilica. 1. Water reduction will help in getting better physical properties through a denser packing. 2. Wetting of the inorganic matrix will improve mixing as well as placement. 3. Last not least setting can be adjusted to certain needs to improve economy of production.
Improved deflocculation of a mix can lead to either improved flow when the total amount of water is kept constant or the water demand is reduced when spread remains constant. In practice mostly a medium way is chosen because the exact dosage determined in a lab sometimes cannot be transferred so easily into the plant. Using a proper PCE can help in dispersing raw materials difficult to disperse in water because of a hydrophobic nature of their surface. Taking a mix which has incorporated some percentage of “C” in form of micronized or flaky graphite and adding an inpropriate dispersant will give the typical spots of nonwetted / non-dispersed material which are observed when a hydrophobic material is swimming on top of an aqueous phase. Using a modified PCE dispersant can significantly reduce this effect in “C”- containing mixes. By introducing long polyether side chains or hydrophobic groups into the backbone Castament polycarboxylate ether polymers provide a more hydrophobic character. Incorporating such a product in a castable mix will minimize mixing times and improve the wettability of hydrophobic components like e.g. chromiumoxide mineral raw materials and even graphite. It has to be realized that the dispersing mechanism for hydrophobic surfaces will probably go with the before mentioned “Octopus Mechanism”. This means that the surface is covered by the side-chains leaving the outer sphere anionic by the carboxylate groups. This can result in a mix which does not show free-flowing properties but is a perfect vibra-flow mix design. Vibration can destroy the uncongested gel formation which occurs when “hedgehog” and “octopus” mechanism come together in one mix. The adjustment of flow, working time and set time in green body refractories is depending on different factors. No. 1 is the binder matrix which is usually formed by CAC, Reactive Alumina and Microsilica. A typical lapse of strength development over time is related to this composition. The binder matrix also gives an important indication for choosing a certain type of dispersant. Where Castament FS20 is currently the only product for mixes having high amounts of microsilica incorporated all other PCEs are more or less suitable in microsilica-free surrounding. In order to determine the three basic effects, water reduction, wetting and setting we follow following procedure. Raw materials according to Tab. 1 are weighed in according to the desired particle size distribution and the dry dispersant is added. A Hobart mixer and a 3-5kg quantity may be used. After dry and wet mixing the compound is placed in a flow cone with 10 cm diameter, the cone is lifted after distinct times and the spread is measured. Besides that a part of the
wet castable is placed in a dewar and connected with a thermo couple to measure the heat development of the hydrating process. Last not least another part is placed to a mold to receive bars which are tested after 24 hours for strength development. Tab. 1. Typical High Alumina LCC Recipe Tabular Alumina 78 % Reactive Alumina 17 % CAC Cement 5% Dispersant 0.2 % Water 4.0-4.4 % To give an indication how the main factor “Binder Matrix” affects the adjustment of flow, working time and set time in green body refractories we compared 5 different reactive aluminas in a recipe according Tab. 1. We have chosen a rather high amount of different reactive aluminas to better show the differences. In each mix a PCE with long side chains has been used. It is evident that a multimodal alumina needs much less water due to a better fit in particle size distribution. But some differences related to the specific water demand of the fines can be observed. Green body strength development after 24 hours is pretty similar but a significant shift in set time development / heat of hydration is realized (Fig. 2). Taking the products with most different behavior we can see a band width of almost 12 hours in which we can tune our castable mix by just choosing an adequate reactive alumina type. We think this may also indicate a surface treatment of the fines during processing or some traces of soluble alkali remaining on the surface. 29 27 ] 25 C ° [ . p 23 m e T
21 19 A17NE
CTC50
PBR
CL370C
PFR 20
17 0:00
6:00
12:00
18:00
24:00 30:00 Time [h:m]
36:00
42:00
48:00
Fig. 2: Heat of Hydration for different Reactive Aluminas When this test series is performed by using other PCE types a similar picture will be received. Comparing the green strength development after 24 hours based on different PCEs a differentiated picture may be observed. As the binder matrix in some mixes has not yet hydrated within the defined 24 hours also CCS may
not be detected. The long chain PCE however shows the most balanced behavior in all test conditions. As an intermediate result it can be summarized: FS10 is extremely retarding but gives some initial 24 hour strength. FS30 is more like a water reducer showing retardation without major flow decrease. FS60 behaves more like a plasticizer with minor retardation. Number two in our consideration about flow, working time and set time in green body refractories is temperature. If we take again our previous mix design and check the influence of outside temperature related to Iceland, Germany or India we can see the various PCEs contributing to retardation. It is more or less negligible what matrix minerals are taken, the shorter the side chain length of a PCE the more delay in receiving the maximum of heat of hydration is determined. Even for Castament FS10, the PCE with the shortest side chain we can see a distinct peak after more than 32 hours when we measure the mix in a climate chamber of 35 °C. In practice it is therefore a simple possibility to apply some heat during drying to enhance set time development. This takes us to No.3 in our consideration about tuning flow, working time and set time in green body refractories. Set time is not only determined by the reactivity of a given binder matrix and the temperature surrounding it but can also be adjusted by choosing a well directed polycarboxylate ether type according to its retarding or accelerating effect. Besides workability controlled by water addition as well as wetting of raw materials the setting of the mix is an important factor for refractory castables because of time to demolding and early green strength development. In the tests described so far it was also pointed out that different PCE types give different placing – setting windows. To illustrate this, a typical high alumina mix design is compared using the three basic PCE types. In line with the heat of hydration over time one can see the 24 hour CCS (Cold Crushing Strength) being OK for mixes with heat increase occurring before 24hours. FS10 shows some strength although the heat increase has not yet taken place. The placing – setting window can also be shifted by the amount of dispersant applied. (Fig. 3) Doubling the dispersant dosage may result in an up to 6 hour shift of
additional retardation, thus affecting strength development at constant temperature after 24 hours. Of course it is also a possibility to combine products based on PCE technology to play on the different retardation phenomena of different PCEs. 32
FS60 (0,1%) FS60 (0,15%)
30
FS60 (0,1%)+FW10 (0,01%) FS60 (0,15%)+FW10 (0,01%)
] 28 C ° [ p 26 m e T
24 22 20 0: 00
3: 00
6:0 0
9: 00
12: 00
15: 00
1 8: 00
2 1: 00
24: 00
Time [h:m]
Fig. 3: Effects on Hydration / Set Time It is state of the art to combine a non-accelerated PCE with an accelerated PCE version (FW10). Sometimes this becomes a rather sensitive undertaking because small quantities may be required to fulfill the tuning job. Using an accelerated polymer version does have the advantage that Lithium, which is part of the neutralized carboxylic group in the polymer, can be added very homogeneously. Small amounts however already influence flow and placing properties. Summary In manufacturing of refractory castables adverse surface properties of raw materials, differing lots of one raw material or inadequate mixing sometimes can cause problems. Small quantities of chemical additives are able to tune castable formulations in a significant manner. The addition of PCE-based dispersants affects the performance of castable mixes in conjunction with the fine materials incorporated as follows: - Increasing the amount of PCE > Increase flow / reduce water, increase retardation - Changing the PCE type > Shift wetting properties, modify rheology - Combining PCE types > shift retardation - Selecting cement > shift retardation - Selecting Reactive Alumina > shift flow / water reduction and retardation Consistent mixing at low water dosage, good wet-out time, enough time for placing and appropriate set time are features for reliable industrial scale preparation of high performance refractory castables. This relates not only to monolithic parts with special shape like nozzles or plates but also to bricks, or repair mixes which may be gunned to cold or hot furnace walls.
