Chemosphere 134 (2015) 279–285
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Technical Note
Disposal of historically contaminated soil in the cement industry and the evaluation of environmental performance Yeqing Li a,⇑, Jiang Zhang b, Wenjuan Miao b, Huanzhong Wang b, Mao Wei a a b
Huaxin Cement Co. Ltd., Tower 5, Int’L Enterprise Center, No. T1 Guanggu Avenue, Wuhan 430073, PR China Huaxin Environment Engineering Co. Ltd., Tower 5, Int’L Enterprise Center, No. T1 Guanggu Avenue, Wuhan 430073, PR China
h i g h l i g h t s Contaminated soil was co-processed without sacrificing cement clinker quality. PCDD/PCDF emissions ranged from 0.0023 to 0.0085 ng I-TEQ Nm
3
.
DRE and DE of DDTs/HCHs were better than 99.9999% and 99.99%, respectively.
a r t i c l e
i n f o
Article history: Received 14 July 2014 Received in revised form 9 April 2015 Accepted 17 April 2015 Available online 15 May 2015 Keywords: Co-processing Cement kiln DDTs/HCHs-contaminated soil DRE
a b s t r a c t Approximately 400 000 t of DDTs/HCHs-contaminated soil (CS) needed to be co-processed in a cement kiln with a time limitation of 2 y. A new pre-processing facility with a ‘‘drying, grinding and DDTs/ HCHs vaporizing’’ ability was equipped to meet the technical requirements for processing cement raw meal and the environmental standards for stack emissions. And the bottom of the precalciner with high temperatures >1000 °C was chosen as the CS feeding point for co-processing, which has rarely been reported. To assess the environmental performance of CS pre- and co-processing technologies, according to the local regulation, a test burn was performed by independent and accredited institutes systematically for determination of the clinker quality, kiln stack gas emissions and destruction efficiency of the pollutant. The results demonstrated that the clinker was of high quality and not adversely affected by CS co-processing. Stack emissions were all below the limits set by Chinese standards. Particularly, PCDD/PCDF emissions ranged from 0.0023 to 0.0085 ng I-TEQ Nm 3. The less toxic OCDD was the peak congener for CS co-processing procedure, while the most toxic congeners (i.e. 2,3,7,8-TeCDD, 1,2,3,7,8PeCDD and 2,3,4,7,8-PeCDD) remained in a minor proportion. Destruction and removal efficiency (DRE) and destruction efficiency (DE) of the kiln system were better than 99.9999% and 99.99%, respectively, at the highest CS feeding rate during normal production. To guarantee the environmental performance of the system the quarterly stack gas emission was also monitored during the whole period. And all of the results can meet the national standards requirements. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction In recent years, China has devoted great effort to remediate soil contaminated by industrial enterprises after they have relocated during rapid urbanization. The soil we excavated for treatment was from a closed pesticide manufacturing plant where the main contaminants were DDTs and HCHs. An integrated and systematic sampling and pollutants analysis of the brown land was taken place for environmental impact assessment by the land owner. The land was divided into 45 small squares by 50 m 50 m, soil ⇑ Corresponding author. E-mail address: [email protected] (Y. Li). http://dx.doi.org/10.1016/j.chemosphere.2015.04.048 0045-6535/Ó 2015 Elsevier Ltd. All rights reserved.
samples from each square have been taken for different height varying from 0 to 1.8 m as the first layer, 1.8–5 m as the second layer and 5–9 m as the third layer. The highest concentration of DDT was 33548.14 mg kg 1 in the first layer and 4661.46 mg kg 1 for HCHs in the second layer. The average concentration for DDT and HCHs was 554.852 mg kg 1 and 23.819 mg kg 1 in the first layer, 139.169 mg kg 1 and 56.07 mg kg 1 in the second layer, and 1.367 mg kg 1 and 0.533 mg kg 1 in the third layer. Conventional methods for destroying organic pollutants comprise biological, chemical and thermal treatments (Cravotto et al., 2007). Biological and chemical technologies are often time consuming and not cost-effective for treating large volumes of polluted materials (Cravotto et al., 2007; Lai et al., 2009; Zhang
280
Y. Li et al. / Chemosphere 134 (2015) 279–285
CSP
Crusher
Silo
Storage hall Clinker
Fig. 1. Flow chart of the CS pre and co-processing unit operations.
et al., 2010; Venny et al., 2012). Thermal decontamination of soil is widely used due to many advantages, including removal efficiencies above 99%, short remediation times and applicability to a wide range of organic contaminants (Lee et al., 1998; Gan et al., 2009; Chien, 2012). In this field-scale remediation case, the survival of the degrading organisms can be affected by the toxicity associated to highly contaminated sites (Rein et al., 2007; Perelo, 2010; Megharaj et al., 2011). The cement kiln is one thermal treatment that has attracted much attention for the environmentally sound destruction of persistent organic pollutants (POPs), owing to its inherent features such as high temperatures, long residence time and surplus oxygen as well as the elimination of slag, ash or liquid residue byproducts (Chadbourne, 1997; Reijnders, 2007; Karstensen, 2008; Kookos et al., 2011). Several documents have demonstrated that this process can simultaneously offer more disposal capacity, maintain good destruction efficiencies and avoid the formation of PCDD/PCDFs (Karstensen et al., 2006, 2010; Weber, 2007; Karstensen, 2008; Khumsaeng et al., 2013). Because the site was located in the city center, the site cleanup was required to be completed within 2 y. To meet this time limitation, co-processing by cement kiln was chosen as the disposal method for the soil with DDT and HCH concentrations higher than 50 mg kg 1. The total amount was estimated to be as high as 400 000 t. For the soil with DDT and HCH concentration lower than 50 mg kg 1 was remedied in-situ by biological method. The chemical components of DDTs/HCHs-contaminated soil are similar to clay, thus after pre-processing, this soil can be co-processed as an alternative raw material in the cement kiln to destroy DDTs/HCHs and recover valuable SiO2 components during clinker sintering. Generally, alternative raw materials are fed at the raw mill (GTZ and Holcim, 2006; Karstensen, 2008). In this manner, their fineness, moisture and homogenization with normal raw meal can be controlled to guarantee the clinker quality. However, alternative raw materials containing components that are volatile at low temperatures have to be fed into the high temperature zones of the kiln system. If the soil would have been fed to traditional points at low temperature, such as the raw mill and the raw meal feeding point at the outlet of preheator, DDTs/HCHs would not have decomposed immediately. Instead there would have been a high risk of incomplete destruction or escape of DDTs/HCHs to the stack. Therefore, a higher temperature location, the bottom of the precalciner (>1000 °C) (Li et al., 2009), was modified and chosen as the feeding point. This feeding point enables instant homogenization with raw meal and can assure DDTs/ HCHs destruction efficiency and clinker quality. So far, feeding CS into the precalciner of a cement kiln has rarely been reported in the literature (Yan et al., 2014).
CS cannot be used directly as an alternative material, as it must first undergo a preparation process. This step produces a waste product with defined characteristics that complies with the technical specifications of cement production and guarantees that environmental standards are met (GTZ and Holcim, 2006). Due to the large particle size and high moisture content of virgin CS, grinding and drying procedures are required in a specialized treatment facility with high pre-processing capacity. To prevent the output of DDTs/HCHs during the drying procedure, the thermal desorption system, ‘‘drying, grinding and DDTs/HCHs vaporizing’’ facility (Li et al., 2011), was designed and employed to fulfill the task. Based on the above analysis and previous experience that precalciners are known to be efficient and environmentally sound feeding points for complex organic pollutants (GTZ and Holcim, 2006; Karstensen et al., 2006), the precalciner of the cement plant was chosen as the feed point and particularly modified to serve this task. A CS co-processing system with 35 t h 1 pretreatment capacity was constructed, including a CS pre-processor (CSP) with drying/grinding ability and an efficient, environmentally sound precalciner was equipped for co-processing (Li et al., 2012a). After running for more than 1 y, 218 078 t of CS have been co-processed. In this period, a 2 d test burn and regularly kiln stack gas emission monitoring were implemented. Third parties monitored the clinker quality, CSP effectiveness, destruction and removal efficiency (DRE), destruction efficiency (DE) for DDTs and HCHs and stack emissions (including organic compounds, acid gases and particulates) to assess the performance of the system during the test burn. Quarterly kiln stack gas has been sampled and analyzed for DDTs, HCHs, TVOC and dust emission. A baseline of DDTs, HCHs, PCDD/PCDFs emission was measured on October 30, 2011 before CS disposal.
2. Materials and methodology 2.1. CS pre and co-processing procedure As shown in Fig. 1, first, CS is crushed to a size less than 100 mm and stored in a pre-homogenization hall. Second, the crushed soil is dosed and fed into the CSP, where the soil is ground to be sifted through a 200 lm sieve and dried using 300–350 °C air from the tertiary air duct and circulating fan. Because the boiling points of DDTs and HCHs range from 260 to 330 °C, the hot air in the CSP could partially vaporize the contaminants. These vaporized components are immediately taken to the precalciner and destroyed. Thus, grinding, drying and removing a part of the organic pollutants are realized in parallel. The pretreatment process could be
281
Y. Li et al. / Chemosphere 134 (2015) 279–285
expanded to removing a wide range of organic pollutants by adjusting the percentage of circulating and tertiary air. At the CSP outlet, the fine soil with size below 200 lm, with a specific surface area as high as 500 m2 kg 1, is lifted into a cyclone by air, where the gaseous and solid substances are separated. The soil falling into the silo is sealed, mixed within the silo and then fed into the precalciner by pneumatic system. In the precalciner, the soil is homogenized and mixed promptly with the raw meal in a turbulence model by the kiln system gas. By such technical procedure, the fine soil can be well dispersed by the hot gas with temperature above 1000 °C and the gas/solid ratio is about 2.838. The gas containing the vaporized DDTs and HCHs is transported back to the tertiary air duct again, and these compounds are finally completely destroyed at >1000 °C in the precalciner. The treatment system is under negative pressure and well-sealed, which completely prevents the POPs from escaping to the external environment. The detailed technical parameters related to the pre and co-processing system can refer to some patents (Li et al., 2009, 2011, 2012a). 2.2. Operation conditions The CS should not negatively affect the smooth and continuous kiln operation, the product quality or the site’s environmental performance. Therefore, the CS must be fed at a constant rate. According to a statistical analysis, feed rates of 35 t h 1 to the CSP and 12–14 t h 1 to the kiln were the highest rates to match the raw material composition, and they provided stable operation capacity. Due to the effect of raw meal on the absorption of dioxins (Li et al., 2012b), the raw mill was kept running during the test burn for CS co-processing with an hourly average meal feed rate of 365 t h 1. The clinker output stayed at 5570 t d 1. The utilization of CS could reduce sandstone/shale consumption approximately by 34%. 2.3. Sampling and monitoring The test burn measurement of the CS co-processing technology was implemented in kiln #1 of the Huaxin (Wuxue) cement plant. The exact time was from 7:00 to 21:00 on November 13–14th, 2012 with an average ambient temperature of 15 °C. Measurements focused on four general categories: clinker quality, stack emissions, CSP efficiency and DE/DRE. The solid samples, i.e., CS, clinker and bag filter (BF) dust, were collected by trained plant staff, and the gaseous samples were trapped by accredited third parties. The procedure strictly followed national standards and China Metrology Accreditation (CMA) certified the results. Samples to determine clinker quality were obtained by blending subsamples every 24 h from the clinker conveyor as the conveyor left the cooler. A vibration mill ground 500 g of clinker for chemical analysis, and a 10-kg combined sample was prepared by adding 5% gypsum (w/w) with a laboratory ball mill for the physical properties test. The test items were inorganic constituents, soundness, fineness, strength, water demand, etc. A variety of chemicals were measured in the main stack emissions: organic compounds, i.e., DDTs, HCHs, PCDD/PCDFs, VOCs, benzene; inorganic gases, i.e., O2, CO, CO2, SO2, NOX, HCl, fluorides, NH3; and particulates. The effect of the CSP was determined by monitoring the variations in moisture, fineness and DDTs/HCHs concentrations before and after pre-processing. The interval between sampling at the CSP inlet and outlet was 30 min, ensuring that the samples corresponded with each other.
Hot air partially vaporized the DDTs/HCHs and directly transferred these compounds into the precalciner for complete destruction. The samples related to DE/DRE measurements were DDTs/ HCHs vapor entering the kiln, CS fed into kiln, clinker, BF dust and stack emission. 2.4. QA and QC All sampling and analysis devices were calibrated in advance, and the records were saved for traceability. Filters, adsorbent resin, sampling probe and other components of the sampling train were rinsed with water, methanol, methylene chloride and toluene. The rinsing was performed until no analyze was observed above the detection limit. Three different blank measurements were collected, i.e., trip blanks, reagent blanks and method blanks, to identify potential interference introduced by the progression of sampling and analysis. When measuring the DDTs/HCHs, spiked samples were detected simultaneously, and within 82–116% of the standards were recovered. Isotopically marked PCDD/PCDF standards were added to the sorbent prior to sampling and to the filter immediately before sample extraction. All PCDD/PCDF surrogate recoveries were within 75–111%. Each sampling site was measured three times every day to guarantee the reliability of the results. 3. Results and discussion 3.1. Clinker quality The results for the daily clinker samples are in Table 1. The sample quality is based on standardized limits in GB 175-2007 and GB/ T 21372-2008. The data show that all the properties of the clinker can satisfy the national standards, and the clinker was of high quality during CS co-processing. The four chief minerals were calculated by SiO2/Al2O3/Fe2O3/CaO/MgO/SO3/f-CaO content. The first and second days of combined samples contained 62.71% and 61.74% of C3S, 12.43% and 14.31% of C2S and 6.98% and 7.14% of C3A, respectively, and both days contained 10.70% of C4AF. The Table 1 Results of the clinker quality test. Item
Unit
Density Specific surface Sieve residue Water requirement Setting time 1 Setting time 2 Soundness SO3 in cement
Limit
Result 1#
2#
g/cm3 m2/kg % % h:min h:min / %
/ 350 ± 10 64 / P0:45 66:30 / 2.0–2.5
3.16 340 1.9 24.2 1:45 2:39 Good 2.4
3.16 329 2.2 24.0 1:48 2:52 Good 2.1
Breaking strength
3d 28 d
MPa MPa
/ /
6.0 8.6
5.9 8.8
Compression strength
3d 28 d
MPa MPa
P26.0 P52.5
35.3 57.4
35.0 55.1
Components
Loss SiO2 Al2O3 Fe2O3 CaO MgO SO3 Ff-CaO K2O Na2O Cl P2O5
% % % % % % % % % % % % %
61.5 / / / / 65.0 61.5 / 61.5 / / 60.06 /
0.12 20.72 4.88 3.52 65.70 2.56 0.74 0.07 0.80 0.81 0.12 0.008 0.12
0.10 21.12 4.94 3.52 65.69 2.66 0.70 0.05 1.20 0.78 0.12 0.004 0.10
282
Y. Li et al. / Chemosphere 134 (2015) 279–285
above data demonstrate that feeding CS to the bottom of the modified precalciner guaranteed the uniformity of the raw meal; therefore, clinker quality was not adversely affected.
3.2. Emissions 3.2.1. Organic compounds Sampling for PCDD/PCDFs was performed according to Chinese standard HJ 77.2-2008. Due to different detector responses to each isomer, PCDD/PCDFs sample detection limits (SDLs) ranged from 0.0009 ng Nm 3 to 0.003 ng Nm 3. US EPA method 23 (Federal Register, 2000) was followed for the DDTs/HCHs sampling. This method has been proven to be effective for sampling a wide range of semi-volatile organic compounds from combustion systems, including PCBs, PAHs, HCB and pesticides (Karstensen et al., 2006). The SDLs were 0.39–0.78 ng Nm 3 for DDTs and 0.17–0.33 ng Nm 3 for HCHs. The PCDD/PCDFs concentration in the stack emission during CS co-processing is presented in Table 2. In general terms, the test burn results expressed in TEQ were far below the limit of 0.1 ng I-TEQ Nm 3 established by the Chinese standard (GB49152004), with values ranging from 0.0023 to 0.0085 ng I-TEQ Nm 3, with an average value and median of 0.0036 and 0.0026 ng ITEQ Nm 3 (n = 6), respectively. The TEQ concentrations in stack emissions collected within a few hours of each other on the same day could, in some cases, be very different. The first sample measured to be 0.0085 ng I-TEQ Nm 3 was much higher than the other five samples. Even though, comparing with the baseline 0.00187 ng I-TEQ Nm 3, all of the data still keep in the same level, which are also in the same range as the results shown by Yan et al. (2014). During the time period stack gas was sampled the plant was running normally and spike recoveries of each sample were within the normal range. One possible explanation of this variability in emissions would be unevenly distribution of normal raw material organics. Thermal processes with organochlorines normally arouse the greatest emission concerns about the formation of dioxins and furans. During the CS co-processing, DDTs/HCHs fed with the soil may be regarded as the potential precursors to PCDD/PCDFs. However, the modern suspension preheater and precalciner kiln possesses several advantages that minimize the emissions of PCDD/PCDFs despite high concentrations of organochlorines in the CS feed, such as long residence time, high temperatures, rapid cooling procedure and intensive gas cleaning by the raw meal. Moreover, feeding CS to the bottom of the precalciner (>1000 °C) avoided the addition of DDT/HCH as part of raw-material-mix through the traditional points at low temperatures. The availability of organics in the CS
was controlled, and the risk stimulating formation of PCDD/ PCDFs was thus eliminated. The results are consistent with those reported in the literatures. The emission of PCDD/PCDFs for responsible organics burning in cement kilns is normally below the value of 0.1 ng I-TEQ Nm 3 (Abad et al., 2004; Van Loo, 2008; Karstensen, 2008). Cement kilns operating in Germany revealed an average concentration of 0.02 ng I-TEQ Nm 3 (Kuhlmann et al., 1996). Data on the analysis of PCDD/PCDFs and dioxin-like PCBs revealed emission values from cement kilns around 0.0016 ng WHO-TEQ Nm 3 (Luthardt et al., 2002). Data from the U.K. revealed average values of 0.058 ng ITEQ Nm 3 (n = 14) and median values of 0.032 ng I-TEQ Nm 3 (n = 14) (Alcock et al., 1999). Conventional analysis of congeners and their distribution, commonly called chemical fingerprint analysis, is widely used as an important tool to link the presence of these contaminants to a specific source. A representative congener-specific concentration profile of the 2,3,7,8-chloro-substituted PCDDs/PCDFs is shown in Fig. 2. For purposes of profile construction, PCDD/PCDF congeners that were not detected in any given test-run were presumed to be present at one-half of the lower limit of sample detection. Within each test-run, each congener’s profile level was calculated
100% OCDF
90%
1,2,3,4,7,8,9- HpCDF 1,2,3,4,6,7,8-HpCDF
80%
2,3,4,6,7,8- HxCDF
70%
1,2,3,7,8,9-HxCDF 1,2,3,6,7,8- HxCDF
60%
1,2,3,4,7,8-HxCDF 2,3,4,7,8- PeCDF
50%
1,2,3,7,8-PeCDF 2,3,7,8-TCDF
40%
OCDD 1,2,3,4,6,7,8-HpCDD
30%
1,2,3,7,8,9-HxCDD
20%
1,2,3,6,7,8-HxCDD 1,2,3,4,7,8-HxCDD
10% 0%
1,2,3,7,8-PeCDD 2,3,7,8-TCDD
1
2
3
4
5
6
Test-runs Fig. 2. Congener distribution of 2,3,7,8-PCDD/PCDFs (ng/Nm3) in the cement kiln during CS co-processing.
Table 2 Emissions from the stack gas in the test burn. Emissions (mg Nm
3
)
PCDD/Fs (ng TEQ Nm DDTs (ng Nm 3) HCHs (ng Nm 3) TVOCs Benzene CO CO2 (V V 1%) Dust SO2 NOX HCl HF NH3
3
)
1#
2#
3#
4#
5#
6#
ELV
0.0085 ND (
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Technical Note
Disposal of historically contaminated soil in the cement industry and the evaluation of environmental performance Yeqing Li a,⇑, Jiang Zhang b, Wenjuan Miao b, Huanzhong Wang b, Mao Wei a a b
Huaxin Cement Co. Ltd., Tower 5, Int’L Enterprise Center, No. T1 Guanggu Avenue, Wuhan 430073, PR China Huaxin Environment Engineering Co. Ltd., Tower 5, Int’L Enterprise Center, No. T1 Guanggu Avenue, Wuhan 430073, PR China
h i g h l i g h t s Contaminated soil was co-processed without sacrificing cement clinker quality. PCDD/PCDF emissions ranged from 0.0023 to 0.0085 ng I-TEQ Nm
3
.
DRE and DE of DDTs/HCHs were better than 99.9999% and 99.99%, respectively.
a r t i c l e
i n f o
Article history: Received 14 July 2014 Received in revised form 9 April 2015 Accepted 17 April 2015 Available online 15 May 2015 Keywords: Co-processing Cement kiln DDTs/HCHs-contaminated soil DRE
a b s t r a c t Approximately 400 000 t of DDTs/HCHs-contaminated soil (CS) needed to be co-processed in a cement kiln with a time limitation of 2 y. A new pre-processing facility with a ‘‘drying, grinding and DDTs/ HCHs vaporizing’’ ability was equipped to meet the technical requirements for processing cement raw meal and the environmental standards for stack emissions. And the bottom of the precalciner with high temperatures >1000 °C was chosen as the CS feeding point for co-processing, which has rarely been reported. To assess the environmental performance of CS pre- and co-processing technologies, according to the local regulation, a test burn was performed by independent and accredited institutes systematically for determination of the clinker quality, kiln stack gas emissions and destruction efficiency of the pollutant. The results demonstrated that the clinker was of high quality and not adversely affected by CS co-processing. Stack emissions were all below the limits set by Chinese standards. Particularly, PCDD/PCDF emissions ranged from 0.0023 to 0.0085 ng I-TEQ Nm 3. The less toxic OCDD was the peak congener for CS co-processing procedure, while the most toxic congeners (i.e. 2,3,7,8-TeCDD, 1,2,3,7,8PeCDD and 2,3,4,7,8-PeCDD) remained in a minor proportion. Destruction and removal efficiency (DRE) and destruction efficiency (DE) of the kiln system were better than 99.9999% and 99.99%, respectively, at the highest CS feeding rate during normal production. To guarantee the environmental performance of the system the quarterly stack gas emission was also monitored during the whole period. And all of the results can meet the national standards requirements. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction In recent years, China has devoted great effort to remediate soil contaminated by industrial enterprises after they have relocated during rapid urbanization. The soil we excavated for treatment was from a closed pesticide manufacturing plant where the main contaminants were DDTs and HCHs. An integrated and systematic sampling and pollutants analysis of the brown land was taken place for environmental impact assessment by the land owner. The land was divided into 45 small squares by 50 m 50 m, soil ⇑ Corresponding author. E-mail address: [email protected] (Y. Li). http://dx.doi.org/10.1016/j.chemosphere.2015.04.048 0045-6535/Ó 2015 Elsevier Ltd. All rights reserved.
samples from each square have been taken for different height varying from 0 to 1.8 m as the first layer, 1.8–5 m as the second layer and 5–9 m as the third layer. The highest concentration of DDT was 33548.14 mg kg 1 in the first layer and 4661.46 mg kg 1 for HCHs in the second layer. The average concentration for DDT and HCHs was 554.852 mg kg 1 and 23.819 mg kg 1 in the first layer, 139.169 mg kg 1 and 56.07 mg kg 1 in the second layer, and 1.367 mg kg 1 and 0.533 mg kg 1 in the third layer. Conventional methods for destroying organic pollutants comprise biological, chemical and thermal treatments (Cravotto et al., 2007). Biological and chemical technologies are often time consuming and not cost-effective for treating large volumes of polluted materials (Cravotto et al., 2007; Lai et al., 2009; Zhang
280
Y. Li et al. / Chemosphere 134 (2015) 279–285
CSP
Crusher
Silo
Storage hall Clinker
Fig. 1. Flow chart of the CS pre and co-processing unit operations.
et al., 2010; Venny et al., 2012). Thermal decontamination of soil is widely used due to many advantages, including removal efficiencies above 99%, short remediation times and applicability to a wide range of organic contaminants (Lee et al., 1998; Gan et al., 2009; Chien, 2012). In this field-scale remediation case, the survival of the degrading organisms can be affected by the toxicity associated to highly contaminated sites (Rein et al., 2007; Perelo, 2010; Megharaj et al., 2011). The cement kiln is one thermal treatment that has attracted much attention for the environmentally sound destruction of persistent organic pollutants (POPs), owing to its inherent features such as high temperatures, long residence time and surplus oxygen as well as the elimination of slag, ash or liquid residue byproducts (Chadbourne, 1997; Reijnders, 2007; Karstensen, 2008; Kookos et al., 2011). Several documents have demonstrated that this process can simultaneously offer more disposal capacity, maintain good destruction efficiencies and avoid the formation of PCDD/PCDFs (Karstensen et al., 2006, 2010; Weber, 2007; Karstensen, 2008; Khumsaeng et al., 2013). Because the site was located in the city center, the site cleanup was required to be completed within 2 y. To meet this time limitation, co-processing by cement kiln was chosen as the disposal method for the soil with DDT and HCH concentrations higher than 50 mg kg 1. The total amount was estimated to be as high as 400 000 t. For the soil with DDT and HCH concentration lower than 50 mg kg 1 was remedied in-situ by biological method. The chemical components of DDTs/HCHs-contaminated soil are similar to clay, thus after pre-processing, this soil can be co-processed as an alternative raw material in the cement kiln to destroy DDTs/HCHs and recover valuable SiO2 components during clinker sintering. Generally, alternative raw materials are fed at the raw mill (GTZ and Holcim, 2006; Karstensen, 2008). In this manner, their fineness, moisture and homogenization with normal raw meal can be controlled to guarantee the clinker quality. However, alternative raw materials containing components that are volatile at low temperatures have to be fed into the high temperature zones of the kiln system. If the soil would have been fed to traditional points at low temperature, such as the raw mill and the raw meal feeding point at the outlet of preheator, DDTs/HCHs would not have decomposed immediately. Instead there would have been a high risk of incomplete destruction or escape of DDTs/HCHs to the stack. Therefore, a higher temperature location, the bottom of the precalciner (>1000 °C) (Li et al., 2009), was modified and chosen as the feeding point. This feeding point enables instant homogenization with raw meal and can assure DDTs/ HCHs destruction efficiency and clinker quality. So far, feeding CS into the precalciner of a cement kiln has rarely been reported in the literature (Yan et al., 2014).
CS cannot be used directly as an alternative material, as it must first undergo a preparation process. This step produces a waste product with defined characteristics that complies with the technical specifications of cement production and guarantees that environmental standards are met (GTZ and Holcim, 2006). Due to the large particle size and high moisture content of virgin CS, grinding and drying procedures are required in a specialized treatment facility with high pre-processing capacity. To prevent the output of DDTs/HCHs during the drying procedure, the thermal desorption system, ‘‘drying, grinding and DDTs/HCHs vaporizing’’ facility (Li et al., 2011), was designed and employed to fulfill the task. Based on the above analysis and previous experience that precalciners are known to be efficient and environmentally sound feeding points for complex organic pollutants (GTZ and Holcim, 2006; Karstensen et al., 2006), the precalciner of the cement plant was chosen as the feed point and particularly modified to serve this task. A CS co-processing system with 35 t h 1 pretreatment capacity was constructed, including a CS pre-processor (CSP) with drying/grinding ability and an efficient, environmentally sound precalciner was equipped for co-processing (Li et al., 2012a). After running for more than 1 y, 218 078 t of CS have been co-processed. In this period, a 2 d test burn and regularly kiln stack gas emission monitoring were implemented. Third parties monitored the clinker quality, CSP effectiveness, destruction and removal efficiency (DRE), destruction efficiency (DE) for DDTs and HCHs and stack emissions (including organic compounds, acid gases and particulates) to assess the performance of the system during the test burn. Quarterly kiln stack gas has been sampled and analyzed for DDTs, HCHs, TVOC and dust emission. A baseline of DDTs, HCHs, PCDD/PCDFs emission was measured on October 30, 2011 before CS disposal.
2. Materials and methodology 2.1. CS pre and co-processing procedure As shown in Fig. 1, first, CS is crushed to a size less than 100 mm and stored in a pre-homogenization hall. Second, the crushed soil is dosed and fed into the CSP, where the soil is ground to be sifted through a 200 lm sieve and dried using 300–350 °C air from the tertiary air duct and circulating fan. Because the boiling points of DDTs and HCHs range from 260 to 330 °C, the hot air in the CSP could partially vaporize the contaminants. These vaporized components are immediately taken to the precalciner and destroyed. Thus, grinding, drying and removing a part of the organic pollutants are realized in parallel. The pretreatment process could be
281
Y. Li et al. / Chemosphere 134 (2015) 279–285
expanded to removing a wide range of organic pollutants by adjusting the percentage of circulating and tertiary air. At the CSP outlet, the fine soil with size below 200 lm, with a specific surface area as high as 500 m2 kg 1, is lifted into a cyclone by air, where the gaseous and solid substances are separated. The soil falling into the silo is sealed, mixed within the silo and then fed into the precalciner by pneumatic system. In the precalciner, the soil is homogenized and mixed promptly with the raw meal in a turbulence model by the kiln system gas. By such technical procedure, the fine soil can be well dispersed by the hot gas with temperature above 1000 °C and the gas/solid ratio is about 2.838. The gas containing the vaporized DDTs and HCHs is transported back to the tertiary air duct again, and these compounds are finally completely destroyed at >1000 °C in the precalciner. The treatment system is under negative pressure and well-sealed, which completely prevents the POPs from escaping to the external environment. The detailed technical parameters related to the pre and co-processing system can refer to some patents (Li et al., 2009, 2011, 2012a). 2.2. Operation conditions The CS should not negatively affect the smooth and continuous kiln operation, the product quality or the site’s environmental performance. Therefore, the CS must be fed at a constant rate. According to a statistical analysis, feed rates of 35 t h 1 to the CSP and 12–14 t h 1 to the kiln were the highest rates to match the raw material composition, and they provided stable operation capacity. Due to the effect of raw meal on the absorption of dioxins (Li et al., 2012b), the raw mill was kept running during the test burn for CS co-processing with an hourly average meal feed rate of 365 t h 1. The clinker output stayed at 5570 t d 1. The utilization of CS could reduce sandstone/shale consumption approximately by 34%. 2.3. Sampling and monitoring The test burn measurement of the CS co-processing technology was implemented in kiln #1 of the Huaxin (Wuxue) cement plant. The exact time was from 7:00 to 21:00 on November 13–14th, 2012 with an average ambient temperature of 15 °C. Measurements focused on four general categories: clinker quality, stack emissions, CSP efficiency and DE/DRE. The solid samples, i.e., CS, clinker and bag filter (BF) dust, were collected by trained plant staff, and the gaseous samples were trapped by accredited third parties. The procedure strictly followed national standards and China Metrology Accreditation (CMA) certified the results. Samples to determine clinker quality were obtained by blending subsamples every 24 h from the clinker conveyor as the conveyor left the cooler. A vibration mill ground 500 g of clinker for chemical analysis, and a 10-kg combined sample was prepared by adding 5% gypsum (w/w) with a laboratory ball mill for the physical properties test. The test items were inorganic constituents, soundness, fineness, strength, water demand, etc. A variety of chemicals were measured in the main stack emissions: organic compounds, i.e., DDTs, HCHs, PCDD/PCDFs, VOCs, benzene; inorganic gases, i.e., O2, CO, CO2, SO2, NOX, HCl, fluorides, NH3; and particulates. The effect of the CSP was determined by monitoring the variations in moisture, fineness and DDTs/HCHs concentrations before and after pre-processing. The interval between sampling at the CSP inlet and outlet was 30 min, ensuring that the samples corresponded with each other.
Hot air partially vaporized the DDTs/HCHs and directly transferred these compounds into the precalciner for complete destruction. The samples related to DE/DRE measurements were DDTs/ HCHs vapor entering the kiln, CS fed into kiln, clinker, BF dust and stack emission. 2.4. QA and QC All sampling and analysis devices were calibrated in advance, and the records were saved for traceability. Filters, adsorbent resin, sampling probe and other components of the sampling train were rinsed with water, methanol, methylene chloride and toluene. The rinsing was performed until no analyze was observed above the detection limit. Three different blank measurements were collected, i.e., trip blanks, reagent blanks and method blanks, to identify potential interference introduced by the progression of sampling and analysis. When measuring the DDTs/HCHs, spiked samples were detected simultaneously, and within 82–116% of the standards were recovered. Isotopically marked PCDD/PCDF standards were added to the sorbent prior to sampling and to the filter immediately before sample extraction. All PCDD/PCDF surrogate recoveries were within 75–111%. Each sampling site was measured three times every day to guarantee the reliability of the results. 3. Results and discussion 3.1. Clinker quality The results for the daily clinker samples are in Table 1. The sample quality is based on standardized limits in GB 175-2007 and GB/ T 21372-2008. The data show that all the properties of the clinker can satisfy the national standards, and the clinker was of high quality during CS co-processing. The four chief minerals were calculated by SiO2/Al2O3/Fe2O3/CaO/MgO/SO3/f-CaO content. The first and second days of combined samples contained 62.71% and 61.74% of C3S, 12.43% and 14.31% of C2S and 6.98% and 7.14% of C3A, respectively, and both days contained 10.70% of C4AF. The Table 1 Results of the clinker quality test. Item
Unit
Density Specific surface Sieve residue Water requirement Setting time 1 Setting time 2 Soundness SO3 in cement
Limit
Result 1#
2#
g/cm3 m2/kg % % h:min h:min / %
/ 350 ± 10 64 / P0:45 66:30 / 2.0–2.5
3.16 340 1.9 24.2 1:45 2:39 Good 2.4
3.16 329 2.2 24.0 1:48 2:52 Good 2.1
Breaking strength
3d 28 d
MPa MPa
/ /
6.0 8.6
5.9 8.8
Compression strength
3d 28 d
MPa MPa
P26.0 P52.5
35.3 57.4
35.0 55.1
Components
Loss SiO2 Al2O3 Fe2O3 CaO MgO SO3 Ff-CaO K2O Na2O Cl P2O5
% % % % % % % % % % % % %
61.5 / / / / 65.0 61.5 / 61.5 / / 60.06 /
0.12 20.72 4.88 3.52 65.70 2.56 0.74 0.07 0.80 0.81 0.12 0.008 0.12
0.10 21.12 4.94 3.52 65.69 2.66 0.70 0.05 1.20 0.78 0.12 0.004 0.10
282
Y. Li et al. / Chemosphere 134 (2015) 279–285
above data demonstrate that feeding CS to the bottom of the modified precalciner guaranteed the uniformity of the raw meal; therefore, clinker quality was not adversely affected.
3.2. Emissions 3.2.1. Organic compounds Sampling for PCDD/PCDFs was performed according to Chinese standard HJ 77.2-2008. Due to different detector responses to each isomer, PCDD/PCDFs sample detection limits (SDLs) ranged from 0.0009 ng Nm 3 to 0.003 ng Nm 3. US EPA method 23 (Federal Register, 2000) was followed for the DDTs/HCHs sampling. This method has been proven to be effective for sampling a wide range of semi-volatile organic compounds from combustion systems, including PCBs, PAHs, HCB and pesticides (Karstensen et al., 2006). The SDLs were 0.39–0.78 ng Nm 3 for DDTs and 0.17–0.33 ng Nm 3 for HCHs. The PCDD/PCDFs concentration in the stack emission during CS co-processing is presented in Table 2. In general terms, the test burn results expressed in TEQ were far below the limit of 0.1 ng I-TEQ Nm 3 established by the Chinese standard (GB49152004), with values ranging from 0.0023 to 0.0085 ng I-TEQ Nm 3, with an average value and median of 0.0036 and 0.0026 ng ITEQ Nm 3 (n = 6), respectively. The TEQ concentrations in stack emissions collected within a few hours of each other on the same day could, in some cases, be very different. The first sample measured to be 0.0085 ng I-TEQ Nm 3 was much higher than the other five samples. Even though, comparing with the baseline 0.00187 ng I-TEQ Nm 3, all of the data still keep in the same level, which are also in the same range as the results shown by Yan et al. (2014). During the time period stack gas was sampled the plant was running normally and spike recoveries of each sample were within the normal range. One possible explanation of this variability in emissions would be unevenly distribution of normal raw material organics. Thermal processes with organochlorines normally arouse the greatest emission concerns about the formation of dioxins and furans. During the CS co-processing, DDTs/HCHs fed with the soil may be regarded as the potential precursors to PCDD/PCDFs. However, the modern suspension preheater and precalciner kiln possesses several advantages that minimize the emissions of PCDD/PCDFs despite high concentrations of organochlorines in the CS feed, such as long residence time, high temperatures, rapid cooling procedure and intensive gas cleaning by the raw meal. Moreover, feeding CS to the bottom of the precalciner (>1000 °C) avoided the addition of DDT/HCH as part of raw-material-mix through the traditional points at low temperatures. The availability of organics in the CS
was controlled, and the risk stimulating formation of PCDD/ PCDFs was thus eliminated. The results are consistent with those reported in the literatures. The emission of PCDD/PCDFs for responsible organics burning in cement kilns is normally below the value of 0.1 ng I-TEQ Nm 3 (Abad et al., 2004; Van Loo, 2008; Karstensen, 2008). Cement kilns operating in Germany revealed an average concentration of 0.02 ng I-TEQ Nm 3 (Kuhlmann et al., 1996). Data on the analysis of PCDD/PCDFs and dioxin-like PCBs revealed emission values from cement kilns around 0.0016 ng WHO-TEQ Nm 3 (Luthardt et al., 2002). Data from the U.K. revealed average values of 0.058 ng ITEQ Nm 3 (n = 14) and median values of 0.032 ng I-TEQ Nm 3 (n = 14) (Alcock et al., 1999). Conventional analysis of congeners and their distribution, commonly called chemical fingerprint analysis, is widely used as an important tool to link the presence of these contaminants to a specific source. A representative congener-specific concentration profile of the 2,3,7,8-chloro-substituted PCDDs/PCDFs is shown in Fig. 2. For purposes of profile construction, PCDD/PCDF congeners that were not detected in any given test-run were presumed to be present at one-half of the lower limit of sample detection. Within each test-run, each congener’s profile level was calculated
100% OCDF
90%
1,2,3,4,7,8,9- HpCDF 1,2,3,4,6,7,8-HpCDF
80%
2,3,4,6,7,8- HxCDF
70%
1,2,3,7,8,9-HxCDF 1,2,3,6,7,8- HxCDF
60%
1,2,3,4,7,8-HxCDF 2,3,4,7,8- PeCDF
50%
1,2,3,7,8-PeCDF 2,3,7,8-TCDF
40%
OCDD 1,2,3,4,6,7,8-HpCDD
30%
1,2,3,7,8,9-HxCDD
20%
1,2,3,6,7,8-HxCDD 1,2,3,4,7,8-HxCDD
10% 0%
1,2,3,7,8-PeCDD 2,3,7,8-TCDD
1
2
3
4
5
6
Test-runs Fig. 2. Congener distribution of 2,3,7,8-PCDD/PCDFs (ng/Nm3) in the cement kiln during CS co-processing.
Table 2 Emissions from the stack gas in the test burn. Emissions (mg Nm
3
)
PCDD/Fs (ng TEQ Nm DDTs (ng Nm 3) HCHs (ng Nm 3) TVOCs Benzene CO CO2 (V V 1%) Dust SO2 NOX HCl HF NH3
3
)
1#
2#
3#
4#
5#
6#
ELV
0.0085 ND (
