Reaction between Steel-Making Slag and Carbonaceous Materials While Mixing with High Density Polyethylene

Experimental apparatus

A horizontal resistance furnace was employed in the current study to provide the high temperature environment for the reaction, of which schematic diagram is shown in Figure 1. Both ends of the furnace tube were sealed to ensure the furnace was air-tight, allowing the easy manipulation of the internal environment. An alumina sample holder, connected to a stainless steel rod, was used to transport samples to and from the isothermal zone of the furnace. The outlet gas was led to a CO/CO₂ IR gas analyzer to monitor the off-gas composition for further analyzing the reaction mechanism.

Figure 1

Schematic diagram of experimental setup

Experimental materials

Synthesized graphite powder, industry coke fines and HDPE particles were chosen as the starting raw materials for the study. The purity of graphite (mean size 44 µm) and HDPE (mean size 3 mm) was more than 99.9%. The composition of the coke is listed in Table 1. The coke was grinded to get an average size of about 100 µm before use. Three types of carbonaceous blocks around 10 g each were prepared for the experiments using high pressure hydraulic pressing (2.2 ╳ 10⁸ Pa). These include the blocks of graphite, coke and the mixture of HDPE and coke with a mixing ratio of 1:1 based on the results of a previous work that it is the maximum mixing ratio for HDPE without causing operation disorders. The mixture of HDPE and coke was made by mixing the pre-weighed grinded coke powder and HDPE homogeneously before being hot-pressed at 423 K. Steel-making slag composed of 29.5% CaO, 34.7% FeO, 8.2% Al₂O₃, 9.3% MgO, 12.9% SiO₂ and 5.4% MnO was used to react with carbonaceous materials. One thing should be mentioned that the iron oxide is present in the steel-making slag as both FeO and Fe₂O₃. Due to the difficulty distinguishing these two phases, iron ions are counted only as FeO. The initial weight of slag was around 0.1 g. The experiments were carried out in argon atmosphere and the gas flow rate was set at 1.67 ╳ 10⁻⁵ m³/s, equivalent to a linear velocity of 8.5 ╳ 10⁻³ m/s in a Φ50 ╳ 1000 mm corundum tube.

Experimental procedure

After around 0.1 g of slag was put on a pre-weighed carbonaceous block located on the alumina sample holder in the furnace, the furnace tube was sealed. Then the sample was pulled back to the cold zone close to the water-cooled front flange to keep it at low temperature before the commencement of the experiment. Then Ar gas was purged to remove the air inside the furnace tube for 1800 s before heating was started. When the temperature of the isothermal zone reached the pre-set temperature of 1823 K, the sample was pushed to the middle of the furnace tube to initiate the experiment, and at this moment the CO/CO₂ IR gas analyzer was started as well to monitor the CO and CO₂ contents in the off-gas originated from the reaction between carbon and iron oxide in the slag. After a certain period of time, the reaction was stopped by pulling the sample holder, on which the sample was located, back to the cold zone.

Experimental results and discussion

The measured CO and CO₂ in the off-gas are plotted against reaction time for the carbonaceous substrates of HDPE and coke mixture together with coke and graphite in Figure 2. It is assumed that the measured gas composition was of the local source since the atmosphere gas flow rate was set large to eliminate the influence of gas phase diffusion on the overall reaction rate. As similar initial weight for all the carbonaceous substrates was used, it is reasonable to compare the reaction rate qualitatively by the CO and CO₂ contents in the off-gas shown in the graphs for different carbonaceous materials.

Figure 2

(a) Change of CO and CO₂ contents in the off-gas with time for graphite. (b) Change of CO and CO₂ contents in the off-gas with time for coke. (c) Change of CO and CO₂ contents in the off-gas with time for the mixture of HDPE and coke

There are two distinguishable reaction patterns classified according to the dominant reaction products, which are called the coke type and the graphite type in the present study. The output of the oxidation reaction of graphite was mainly CO₂ in a relatively long period of the reaction time, whereas the dominant reaction product for the oxidation reaction of the mixture of HDPE and coke or coke only was CO almost throughout the whole duration of the reaction except a few seconds in the very beginning. The amounts of CO and CO₂ originating from the reaction between the slag and coke containing substrates, HDPE and coke mixture or coke only, were much greater than those from the reaction between the slag and graphite. Both CO and CO₂ contents rose to their respective peaks quickly after the start of the reaction and fell down steeply to a gradually changing stage in the case of graphite substrate. In contrast to this, the CO content in the off-gas in the cases of HDPE and coke mixture or coke only rose to a near plateau maximum value zone lasting for around 600 s, then from the maximum value zone dropped to a gradually decreasing zone in a very short time; on the other hand, CO₂ content increased to a peak value after starting the reaction, and then decreased gradually along the reaction time. The CO and CO₂ contents in the off-gas for the substrate of HDPE and coke mixture, and coke only substrate follow a similar trend.

The reactions between carbonaceous materials and iron oxide (Fe₂O₃ and FeO) in molten slag are generally accepted following the four reactions forming CO and CO₂ as indicated below step-wisely, which is Fe₂O₃ is reduced firstly to FeO and then to Fe.

Since the CO and CO₂ formation reactions (5) and (6) have been proved to be primary ones by isotopic studies [9], the above four reactions (1) to (4) should occur independently.

However, the following three reactions may occur but not necessarily depending on the conditions.

The mean size of graphite powder used in this work is 44 µm and the average grain size of coke is 100 µm, the specific surface area of coke is less than that of graphite. However, coke containing substrates, coke and the mixture of HDPE and coke, released CO and CO₂ more quickly than graphite substrate. This is because graphite is chemically more stable than coke. The differences in the crystalline structure of graphite and coke are considered to be one of the causes. There are two types of planes in crystallized or partially crystallized carbonaceous materials: basal planes and prismatic or edged planes. The surface energies of these two types of crystallographic direction differ largely: about 0.11 J/m² in the basal planes and 5 J/m² in the prismatic planes [10]. Therefore basal planes are very stable and prismatic or edged planes tend to have lower reaction activation energies. The crystalline size of graphite is larger than that of coke, which also has a highly scattered crystalline direction, as shown in Figure 3. For the typical carbonaceous materials peak at the scanning angle of around 26°, it is very sharp for graphite suggesting the crystalline size of graphite is very large and the crystals are well-oriented; while the peak of coke is very broad, corresponding to the fact that in coke only a small portion is graphitized whilst its dominant part remains amorphous with small crystals. As a result, coke is more reactive than graphite due to its larger proportion of edged planes and hence more products CO and CO₂ are generated in the case of coke compared to graphite. The reaction of prismatic or edged planes may be predominant during the quick reaction period, while the slow reaction stage may be attributed to the fact of the reaction area being occupied mostly by basal planes after the prismatic or edged planes initially exposed to the atmosphere have been consumed. Another reason for the quicker reaction rate of coke than graphite may exist in the fact that coke contains 15% ash composed of oxides such as lime and iron oxide which may act as catalysts during the reaction [11].

Figure 3

XRD results of various carbonaceous materials

The standard free energy changes of CO and CO₂ generation reactions (5) and (6) are plotted in Figure 4. It is clear from this figure that the formation reaction of CO₂ proceeds preferably compared to the formation reaction of CO as expressed by eq. (5). Thus, more CO₂ was generated than CO in the initial stage of the oxidation for all the carbonaceous substrates studied in the current research. This should correspond to the reduction period of Fe₂O₃ to FeO according to eq. (7). However, the duration of favored production of CO₂ over CO was longer for graphite whereas in the cases of coke and the mixture of HDPE and coke this phenomenon only appeared for seconds in the very beginning of the reaction. This may be due to the fact that graphite is less active than coke, therefore longer time is needed to reduce all Fe₂O₃ to FeO for graphite than coke. Afterwards CO production dominated over CO₂ production, especially for coke and the mixture of HDPE and coke, where CO level was markedly higher than CO₂.

Figure 4

Standard free energy change as a function of temperature

The index of Pco/Pco₂, an indicator for the ratio of CO to CO₂ partial pressure, is plotted as a function of time in Figure 5. The reaction time at which P_co/P_co2 = 1 for all of these three carbonaceous substrates shows no correlation with their respective reaction time in Figure 2(a)–(c), at which the oxidation reaction transits from the fast reaction rate zone to the slow reaction rate zone of the dominant oxidation product (CO₂ for HDPE + coke mixture and coke, CO for graphite).

Figure 5

Change of P_co/P_co2 value with time

Since carbon was in excess near the reaction sites, the reactions (7) and (8) may not necessarily proceed while the Boudouard Reaction or the solution-loss reaction (9) occurs inevitably. Therefore the solution-loss reaction of carbon is considered to be the reason of CO overtaking CO₂ later on.

The solution-loss reaction (9) plays a key role in many carbothermic reduction processes of oxides and is judged as at least one of the main rate controlling steps by many researchers [12]. The value of pco2/pco2 for the reaction system of HDPE + coke mixture, coke and graphite with iron oxide in molten slag is calculated and plotted in Figure 6. In this graph, the dotted line is the value of pco2/pco2 when the solution-loss reaction is at equilibrium. The mixture of HDPE and coke approached the equilibrium of the solution-loss reaction most quickly from the beginning followed by coke, while the oxidation of graphite hardly reached the equilibrium of the solution-loss reaction. As a result, CO content soon surpassed CO₂ content in the gas phase in the cases of HDPE + coke mixture and coke, whereas for graphite and iron oxide reaction system more CO₂ was detected than CO in the off-gas during most time of the reaction. The fact that HDPE + coke mixture reached more quickly than coke is correlated to the pyrolysis characteristics of HDPE at high temperature. Many researchers have investigated the pyrolysis behavior of PE. I. Kayacan et al. [13] showed that the pyrolysis of HDPE proceeded vitally from below 700 K. At low temperatures, the pyrolysis products included solid carbon residue, oil or wax and light organic and inorganic gases [14] and their proportions varied with experimental conditions such as temperature, heating rate, presence of catalysts, etc. With the increase in the temperature, the wax or oil started to break down to small molecules. When the temperature was very high as iron- and steel-making ones, the reaction between PE and hematite yielded H₂ and CO with small amounts of H₂O and methane in the gas phase as shown by T. Matsuda et al. [15]. H₂O and CO gas must be the reaction product of hydrogen-iron oxide and carbon-iron oxide, respectively. For the current case, since the temperature is very high as 1823 K, it should not bring much dispute if reaction (10) is used to represent the overa(CH2)n=nCs+nH2g ll pyrolysis process of HDPE to simplify the role played by HDPE in this study as to provide reducing agents H₂ and solid residue carbon.

Figure 6

Comparison of pco2/pco2 of carbonaceous substrates with the equilibrium value of solution-loss reaction

Although the current work has not monitored the H₂ and H₂O contents in the off-gas, the release of H₂ gas from PE decomposition has been confirmed by many other researchers. T. Murakami et al. [7] indicated that PE played a significant role for the reduction of iron oxide by supplying reducing gases, H₂ and CO, through pyrolysis when utilizing PE in combination with graphite as reducing agents. The temperature of the gas generation peak originated by the reduction of graphite, 1370 K, was much higher than those of the other two gas generation peaks, 770 K and 1070 K, caused by the reaction of PE. They also showed the residual carbon after PE pyrolysis was ready to be oxidized by CO₂ from a much lower temperature than coke or graphite, which implies that the derived carbon from PE pyrolysis is chemically more active than coke and graphite. In addition, A. Bazargan et al. [6] reviewed that the residual carbon after PE pyrolysis was very porous and possessed a large specific surface area compared to coke or graphite. The current result graphed in Figure 6 that the mixture of HDPE and coke approached the equilibrium of solution-loss reaction the most quickly among the carbonaceous materials investigated in this study can be explained correspondingly.

Another important result to be pointed out in Figure 6 is that once in a short time period, the mixture of HDPE and coke showed a larger pco2/pco2 value than the equilibrium value of the solution-loss reaction. This may be due to the fact that hydrogen, one of the pyrolysis products of HDPE according to reaction (10), can react further with CO₂ according to eq. (11) to give a higher CO content and a lower CO₂ content compared to their respective equilibrium values according to the solution-loss reaction.

The release of hydrogen gas from HDPE is a very fast process starting from low temperatures as low as 650 K [16]. Therefore, it is highly possible that after 1000 s, there were no more hydrogen in the reaction system. However, the pco2/pco2 value of the mixture of HDPE and coke was still above that of coke, much closer to the equilibrium value of the solution-loss reaction than that of coke oxidation at the latter stage of the reaction. This is probably due to that in the mixture of HDPE and coke, after the pyrolysis of HDPE, there was still porous solid carbon residue chemically more active than coke or graphite getting involved in the reaction in the system.