Occurrence, migration pathways, and multifactorial regulation of chlorine in coking coal

2.1. Determination of total chlorine in coal and coke

Total chlorine in raw coal and coke was determined according to GB/T 3558-2014, utilizing the Eschka method. A sample of 1.0±0.1 g (or >1.0 g for low-chlorine coke) was mixed with 3 g of Eschka mixture in a crucible and covered with an additional 2 g of the mixture [23]. The crucible was heated in a muffle furnace at 680°C for three hours. The calcined residue was dissolved in boiling water. In an acidic medium, a known excess of AgNO₃ solution was added, and the unreacted Ag⁺ was back titrated with a standard KHSO₄ solution using NH₄Fe(SO₄)₂·12H₂O as an indicator. The precision of the method, expressed as the repeatability standard deviation (RSD), is within 0.010% for both coal and coke samples.

2.2. Comprehensive equipment assembly for a simulated coking system

Pilot-ccale coking experiments were performed in a 5 kg pilot-scale coke oven (SYD-5CL, Anshan SYD Science and Technology Co. Ltd.), as depicted in Figure 1. For each run, a 5 kg coal sample was charged into the retort when the oven wall temperature reached 800°C. This temperature was recovered within 30 min, then raised to 1050°C at a rate of 1.39°C min^-1 and held for one hour. Gaseous by-products were passed through a three-stage purification and condensation system to collect tar and condensate water.

Close

Figure 1.

Device for simulating pilot-scale coking experiments. (a) Support structure for coal loading box; (b) Mobile sample carriage; (c) Temperature controller; (d) Main structure of the coke oven; (e) Three-stage purification system; (f) Low-temperature thermastal.

Export to PPT

2.3. Coking gas collection for chlorine analysis

To analyze the gas-phase chlorine, 50 g of coal was coked in a small retort (Figure 2). The evolved gas was passed through two absorption bottles immersed in a -5°C cooling bath to condense liquids. The non-condensable gas was collected in sampling bags, and its volume was measured to calculate the chlorine content.

Close

Figure 2.

Simulated coking gas production system: (a) Main structure of the coke oven; (b) Coal retort; (c) Low-temperature thermastal; (d) empty absorption bottle; (e) Gas sampling bag.

Export to PPT

2.4. Analysis of organic and inorganic chlorine

The chemical states of chlorine in raw coal and coke were analyzed using an X-ray photoelectron spectrometer (XPS). Samples were ground to <0.2 mm. Spectra were calibrated using the C 1s peak (284.6 eV) as an internal standard, and the Cl 2p orbital region (195–210 eV) was analyzed.

2.5. Determination of total chlorine in tar

Total chlorine in tar was measured using a microcoulometric chlorine analyzer (JF-WK-2000A) via high-temperature combustion hydrolysis and potentiometric titration. The instrument was calibrated with 10 mg L^-1 and 20 mg L^-1 chlorine standards. Operating conditions were: N₂ carrier gas at 110 mL min^-1, inlet O₂ at 30 mL/min, pyrolysis O₂ at 350 mL min^-1, and an injection volume of 20 μL. Table 1 [24]shows the repeatability limits of the microcoulometric method for different chlorine concentration ranges.

2.6. Determination of total chlorine in gas

Total chlorine in the gas samples was determined using the same microcoulometric analyzer coupled with a semi-automatic gas sampler (JYQ-2). A 5 mL gas sample was introduced into the cracking furnace with a carrier gas flow of 20–25 mL min^-1. This method can convert all organic chlorine (such as chlorobenzene) and inorganic chlorine (such as HCl, Cl₂) in the gas into measurable ions through high-temperature cracking, achieving direct measurement of the total chlorine in the gas and overcoming the traditional alkaline absorption method that overlooks the content of organic chlorine. To improve the scientific validity of the research, gas samples from each condition were analyzed in three to four replicates to ensure methodological precision.

3.1. Coal properties and chlorine speciation

The proximate and ultimate analyses of the nine coking coals are presented in Tables 2 and 3. The data clearly show that with an increasing degree of coal metamorphism, which is indicative of higher rank, the volatile matter (V_daf) decreases, the carbon content (C) increases, and the hydrogen content (H) decreases. Table 4 details the chlorine, metal, and organic chlorine contents. As shown in Figure 3, generally, a positive correlation was observed between the chlorine and the contents of sulfur and nitrogen, suggesting a common association with the organic matrix of the coal, a hypothesis proposed by Fan et al [25]. Furthermore, the molar quantity of Cl in all samples was substantially lower than that of alkali and alkaline earth metals, indicating that most Cl is not present as simple inorganic salts but rather as organic chlorine or Cl^- in pore water. Cl can also form complex salts with alkali metals and alkaline earth metals, with lower molar amounts, but it does not exist as organochlorine [26]. This was confirmed by XPS analysis (Figure 4), which revealed that organic chlorine constitutes the majority of chlorine in coking coal, with an average proportion of 66.3%. This indicates that chlorine is predominantly bound to the coal’s organic structures rather than its mineral phase.

Close

Figure 3 (a,b).

Correlation between chlorine content in coking coal and N and S [25].

Export to PPT

Close

Figure 4.

Chemical forms of chlorine in coking coal and coke from sample No. 5, as detected by XPS.

Export to PPT

3.2. TG-MS analysis of chlorine release

The pyrolysis behavior of the representative prime coking coal (No. 5) was investigated using TG-MS under an N₂ atmosphere (Figure 5). In the temperature range below 200°C, a slight mass loss is observed within the low-temperature zone, primarily attributed to water evaporation. Between 434°C and 594°C, the largest changes are related to the softening and subsequent melting of the main coking coal, which flows as a very liquid gel. At the same time, the macromolecular structures present in coal, such as the fatty side chains and oxygen-containing functional groups, undergo depolymerization and decomposition reactions that concentrate and release large amounts of tar and volatile matter. This is one of the main factors that affect the yield of tar and gas.

Close

Figure 5.

Results of (a) TG-DTG and (b) TG-MS of prime coking coal from No. 5.

Export to PPT

Between 594°C and 1000°C, semi-coke becomes coke; in this process, the rate of mass reduction decreases, but mainly releases gases. If the horizontal TG line of this process is reached, the residual carbon content is still relatively high, around 79%. This result is consistent with the results of dry total coke rate tests conducted on samples weighing 5 kg of coking coal. It also shows characteristics of high-quality coking coal. Cl₂ is continuously released with a weak intensity in the range of 400-800°C. The release of hydrogen chloride has three peaks: 300-400°C, 400-600°C, and 700-800°C. In the first stage, hydrogen chloride mainly comes from unstable organic chlorides existing in coal. These organic chlorides decompose or hydrolyze at low temperatures.

The second stage was characterized by a high pre-exponential factor (LnA), indicating a high frequency of C-Cl bond cleavage reactions (Table 5). There is obvious thermal cracking and polycondensation of macromolecules in coal (such as aromatic fused rings and bridge bonds) in the temperature range of 400-600°C [27-29]. In this process, pyrolysed covalently bonded organic chlorine (such as Cl on the aromatic ring) precipitates HCl together with the liberation of macromolecular fragments (including tar and gas) [30]. This peak is relatively wide and overlaps with the temperature range of tar formation. This may lead to an increase in the chlorine content of both tar and gas. In the third stage, the semi-coke or coke formed from the matrix of coal has been basically settled. The lower activation energy (E) value suggests reactions with lower energy barriers, such as the combination of Cl⁻ with Na⁺/Ca²⁺ to form NaCl/CaCl₂. It is inferred that inorganic chlorine in the coal partially polymerizes into NaCl/CaCl₂ crystals during high-temperature pyrolysis. At this time, the precipitation of HCl mainly depends on the decomposition and gasification reactions occurring inside inorganic minerals. The reaction rate in this process is relatively slow, and the intensity of hydrogen chloride is lower than that of the main pyrolysis stage.

3.3. Chlorine migration into coke

The chlorine content and conversion rate in the coke were calculated using Eq. (1) and are presented in Table 6. As shown in Figure 6, chlorine content in the raw coal generally increased with the degree of coal metamorphism. However, the chlorine content in the resulting coke did not show a clear correlation with coal type (Figure 7). Notably, despite higher initial chlorine concentrations in some prime coking coals, the final chlorine content in the coke was low.

Close

Figure 6.

Correlation between the chlorine content of coking coal and coke, and the metamorphism of coal.

Export to PPT

Close

Figure 7.

Correlation between chlorine conversion rate to coke and coal rank.

Export to PPT

where, $\Delta Cl$ is the conversion rate of chlorine, %; $K$ is the coking yield of 5kg of dry coking coal, %; ${Cl}_{t,d,co\text{a}l}$ is the total chlorine content in coal (dry basis), %; ${Cl}_{t,d,coke}$ is the total chlorine content in coke (dry basis), %.

The stable temperature at the center of the coke cake, shown in Table 6, indicates that the thermal field inside the oven shows good uniformity during the stage when high-temperature coking occurs. The trend in temperature at the center of the coke cake that this study finds can be compared with data from the industrial coke oven with 7.63 m dimensions, which has a capacity of 50-60 t for a single charge, as report by Gao Mingjie et al [31]. This comparison shows that the cycle for coking in the oven used in this experiment is significantly shorter than the industrial oven. However, the heating trends of the coke cake center temperature in both are consistent. The range of temperature where the main reactions occur in this experiment, which is 950–1050°C, aligns with the range of temperature where chlorine movement occurs in the process used in industrial production.

The chlorine conversion rate into coke varied significantly with coal type (Figure 8). The fat coal (No. 2) exhibited the highest conversion rate (65.78%), while the gas fat coal (No. 1) showed the lowest (7.71%). In general, fat coals had higher conversion rates, while prime coking and lean coals had lower rates. This suggests that the chemical structure and pyrolysis behavior inherent to different coal ranks play a crucial role in chlorine retention.

Close

Figure 8.

Correlation between chlorine rate of content in coke and volatile matter in coal.

Export to PPT

3.4. Chlorine migration into gas and condensate water

The conversion rate of chlorine to tar (C₁), gas (C₂) and condensate water (C₃) were calculated and summarized in Table 7. The total chlorine content in gas varied between 2.067 mg m^-3 (No. 6) and 12.825 mg m^-3 (No. 7). There was no clear correlation between the chlorine content in the gas phase and coal rank (Figures 9 and 10).

Close

Figure 9.

Correlation between the chlorine conversion rate (C₂) of gas and the metamorphism of coal.

Export to PPT

Close

Figure 10.

Correlation between chlorine conversion rate in gas and volatile matter content.

Export to PPT

Figure 10 shows a general trend of an increasing chlorine migration ratio to gas (C₂) with an increasing V_daf during the coal-to-coke conversion process. This trend can mainly be attributed to the interaction between the pyrolysis characteristics of coal and the forms of chlorine occurrence in coal. Coals with high volatile content usually have a large amount of aliphatic side chains, oxygen-containing functional groups and unstable organic structures. These compounds are more likely to undergo bond cleavage and depolymerization reactions during pyrolysis, which cause a large amount of volatile matter (including tar and gas) to be released.

Organic chlorine, which constitutes the dominant form of chlorine in coal, may partially evolve directly into the gas phase in the form of chlorinated hydrocarbons or Cl₂, leading to an elevated chlorine conversion rate in the gas phase. Furthermore, the pyrolysis of high-volatile coal is more intense, accompanied by a higher gas release rate. This rapid devolatilization likely reduces the opportunity for chlorine to react with metallic cations (such as Na⁺, K⁺) to form stable inorganic chlorides, thereby promoting the transfer of chlorine into the gas phase.

The overall distribution of chlorine among the pyrolysis products is depicted in Figure 11. Chlorine primarily partitioned into the coke and condensate water, which together accounted for over 60% of the initial chlorine in most cases. The high proportion of chlorine captured in the condensate water (C₃ values up to 87.86%) indicates that chlorine is released from the coal primarily as highly water-soluble species like HCl. As the hot gas cools in the collection system, these species dissolve in the condensed water, leaving only a minor fraction in the non-condensable gas phase. Moreover, the TG-MS results presented in Figure 5 provide strong evidence that the predominant volatile chlorine species in the gas phase is HCl.

Close

Figure 11.

Conversion rate distribution of chlorine in pyrolysis products.

Export to PPT

3.5. Influence of heating system on gas-phase chlorine

The heating protocol significantly impacted the final chlorine concentration in the gas phase (Figure 12). Three heating systems were compared: System 1 (High-temp coking): Coal was charged into the oven at 800°C and heated to 1050°C. This rapid heating resulted in the highest gas-phase chlorine concentrations (e.g., 12.825 mg m^-3 for coal No. 7). System 2 (Room-Temp Coking): Coal was charged at room temperature and heated to 1050°C. This slower heating rate yielded significantly lower gas-phase chlorine (e.g., 2.13 mg/m³ for coal No. 7). System 3 (Low-temp coking): Coal was charged at room temperature and heated to a lower final temperature of 730°C. The chlorine concentration was intermediate.

Close

Figure 12.

The influence of heating systems on the chlorine content in gas. High-temp coking is based on the temperature rise of a small coke oven. A 50g small coal cup is introduced into the oven at a high temperature of 800°C, achieving a final temperature of 1050°C.

Export to PPT

The high chlorine content under rapid heating (System 1) suggests that the instantaneous high energy input accelerates gas production, allowing some organic chlorine to enter the gas phase before it can fully decompose into ions that would otherwise be retained in the solid or liquid phases. The slower rate of heating in System 2 provides more time for reactions that break down and retain components. Some organic chlorine that is present in the material shows sufficient time to break bonds that hold it and to form chloride ions before the material enters the gas phase. This allows more of the chlorine present within the coal to decompose into chloride ions. These ions either release with the water vapor or deposit within the coke during the process. System 3, fails to reach the complete pyrolysis threshold (>800°C), thereby suppressing chlorine release especially during the coke condensation stage above 700°Cand resulting in lower total chlorine evolution compared to high-temperature conditions.

3.6. Chlorine migration into tar

In general, these results are basically consistent with the reported results for the generation of tar from different coal types. Generally, low-rank coal produces more tar. There are still big differences in the degree of variation for different coal types and even the same type.

The content of chlorine in tar is generally relatively high. Take the chlorine concentration of coal type No. 8 and coal type No. 4 as an example. The chlorine concentration of coal type No. 8 can reach 3673.71 mg L^-1, while the chlorine concentration of coal type No. 4 is much lower, only 803.65 mg L^-1. Figures 13 and 14 show the correlation between chlorine concentration and conversion rate of tar relative to different kinds of coal. Generally, this correlation is weak; however, it is obvious that low rank coals with low volatile matter content have higher conversion rate. So, in the coking process of low-rank coal, there is more tar produced which can help to remove more chlorine component from the system. This conclusion is also supported by Figure 13.

Close

Figure 13.

Correlation between tar chlorine concentration and coal type.

Export to PPT

Close

Figure 14.

Correlation between the chlorine in tar and coal type.

Export to PPT

It is worth noting that the chlorine content concentration of low rank coal is not very high in total; however, there are large fluctuations in primary coking coal types, indicating that both pyrolysis process and chlorinated compounds formation are complicated.

3.7. Optimization of coal blending based on chlorine migration control

Building upon the afore mentioned understanding of chlorine occurrence and migration patterns in different coals. Two optimized coal blending schemes are proposed based on a steel plant’s current Scheme 1^# (comprising 44% low-rank coal and 56% high-rank coal), with the ratios of 1/3 coking coal (coded “a”) and prime coking coal (coded “b”) maintained to provide properties for caking and coking that are important. Table 8 shows the specific blending ratios.

The optimization is grounded in the distinct chlorine migration characteristics of different coals, with the specific changes as follows:

Prime coking coal ratio adjustment (Scheme 1): Coal No. 5 shows a rate for conversion to coke chlorine that is significantly lower than Coal No. 4, while the two coals show rates for conversion to tar chlorine (C₁) that are similar. Coal No. 4 also shows a rate for conversion to coke chlorine that is higher than Coal No. 3. Scheme 1 increases the proportion of Coal No. 3 and reduces the proportion of Coal No. 4.

Lean coal ratio optimization (Scheme 2): Coal No. 9 shows a market price that is relatively lower, and this coal also shows low levels for both rates of conversion to coke chlorine and tar chlorine. Scheme 2 increases the lean coal amount from 10% to 12%. This scheme reduces the proportion of Coal No. 4 and increases the proportion of Coal No. 5. This adjustment further mitigates the risk of chlorine migration within the system while concurrently controlling blending costs.