Short Stories

A pilot-plant study for destruction of PCBs in contaminated soils using fluidized bed combustion technology

A pilot-plant study for destruction of PCBs in contaminated soils using fluidized bed combustion technology
of 6
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  A pilot-plant study for destruction of PCBs in contaminated soils usingfluidized bed combustion technology Dilip L. Desai, Edward J. Anthony, Jinsheng Wang  CANMET Energy Technology Centre, Natural Resources Canada, 1 Haanel Dr., Ottawa, Ontario, Canada K1A 1M1 Received 28 June 2005; received in revised form 20 February 2006; accepted 5 June 2006Available online 9 August 2006 Abstract Destruction of polychlorinated biphenyls (PCBs) in contaminated soils and wastes using circulating fluidized bed combustion (CFBC)technology was studied using a pilot plant and simulated waste material. The results show that the technology is effective and particularlypromising for treatment of PCB-containing materials like the toxic sludge from a large contaminated site. Destruction of the toxics in thegas phase appears to be very fast, and over 99.9999% destruction and removal efficiency can be achieved in the temperature range875–880 1 C. Heat transfer in the fluidized bed also appears adequate. Toxic residues in treated soil can be reduced to very low levels.Rate-controlling factors of the decontamination process are analyzed, and key issues for determination of the process conditions arediscussed.Crown Copyright r 2006 Published by Elsevier Ltd. All rights reserved. Keywords:  PCBs; Soils; CFBC; Decontamination; Dichlorobenzene 1. Introduction Polychlorinated biphenyls (PCBs) are persistent organicpollutants (POPs). There are many sites in North Americacontaining PCB-contaminated soils or wastes. A well-known and long-standing case is the Sydney Tar Ponds inNova Scotia (east coast of Canada), which is regarded asNorth America’s largest hazardous waste site. The TarPonds contain 700,000 tones of toxic sludge with PCBs asone of the major contaminants (Greenhalgh, 2002; Vinson, 2004; Rainham, 2002). The disposal of diverse sources of  PCBs and remediation of such sites remain challengingproblems especially with the development of new environ-mental standards in Canada.Incineration is a reliable method for destruction of PCBs.It can destroy over 99.9999% of PCBs and is widely used(UNEP, 1998; Environment Canada, 2003a). However, there are concerns over the potential formation of poly-chlorinated dibenzo dioxins/furans (PCDD/Fs) from incin-erators, and regarding the cost of PCB destruction byincineration. Some studies concluded that incineration canonly be used on PCB-containing equipment and contami-nated liquid, and is not suitable for the decontamination of affected soils (Wikipedia, 2005). Nevertheless, under certaincircumstances, incineration can be the most feasible method.One such case is the Sydney Tar Ponds, where the toxicsludge will be incinerated in a 10-year remediation program(Government of Nova Scotia, 2004; PWGSC, 2005). Circulating fluidized bed combustion (CFBC) is aneffective technology for the incineration of wastes anddestruction of toxic hydrocarbons. CFBC uses high-velocity air to entrain circulating solids and create a highlyturbulent combustion zone, thus providing very goodsolids mixing and gas–solid contact. It also provides thelong solid residence time necessary for thorough destruc-tion of the toxics in solid wastes. CFBC can burn high-moisture wastes (Yaverbaum, 1977; Patumsawad and Cliffe, 2002) and use calcium-based sorbents (limestoneor dolomite) for in situ sulfur removal. These are inherentadvantages for CFBC incineration of sludge from theSydney Tar Ponds, which is contained in a tidal estuaryand has high salt-water content. The calcium-basedsorbents may help somewhat reduce emissions of chlorine ARTICLE IN PRESS 0301-4797/$-see front matter Crown Copyright r 2006 Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.jenvman.2006.06.006  Corresponding author. Tel.:+16139437772; fax: +16139929335. E-mail address: (J. Wang).  species formed from the salt water, and can have a positiveeffect on reduction of CO emissions (Leckner et al., 1992).The uniform, low operating temperature enables lower fuelconsumption and lower emissions of NO x . It also helps toreduce vaporization of heavy metals from the contami-nated sludge. Finally, at lower temperature the concentra-tion of Cl 2  produced during the combustion process maybe lower, and hence the formation of PCDD/Fs may bereduced (Procaccini et al., 2003; Stanmore, 2004). In this work we present the results of pilot-plantincineration tests, which were carried out to obtain keydata for a large-scale PCB destruction project. The projectwas contracted by the government of Quebec, Canada, forincineration by CFBC of highly chlorinated wastescontaining solids (mostly soils) and liquids with highconcentrations of PCBs, chlorobenzenes (CBs) and per-chloroethylene. The results of these tests have also been of use for experimental investigations of CFBC incinerationof the saline sludge from the Sydney Tar Ponds. Based onthe results, we analyze the important factors for usingCFBC in large-scale remediation of soil/sludge contami-nated by toxics such as PCBs with environmentallyacceptable performance. We also develop a model todescribe the destruction rate of the toxics in CFBC andevaluate the decontamination potential of the process. 2. Experimental A 0.8MWt CFBC was used for the incineration tests. Asthis unit was not permitted to incinerate high-PCBcontaining materials, a non-carcinogenic compound, 1,3-dichlorobenzene (DCB), was employed as a surrogate forPCBs. The compound has similarities to PCBs in structureand composition, as shown in Fig. 1. Because of thesimilarities the compound (DCB) was expected to showsimilar incineration behavior to that of PCBs. Theanalytical data of the DCB used in the tests are given inTable 1. The feed of PCB-contaminated soils was simulatedby sand and liquid DCB. A schematic of the CFBC unit isshown in Fig. 2. The combustor is 6.6m high, with an innerdiameter of 0.405m. Other main components are cyclone,L-valve, flue gas cooler, baghouse, solids and liquidshandling and feed system, and air supply system. The rigis equipped with a comprehensive instrumentation, controland data acquisition system. To start the test runs, thecombustor and return leg were charged with sand and airwas supplied to fluidize the solids in the combustor. Thenatural gas ignition burner was fired and the refractory andsolids in the combustor were heated gradually. When thetemperature reached about 600 1 C, fuel oil with a calorificvalue of 45MJ/kg was fed and the gas burner was shut off.The continuous sand feed was started and graduallyincreased to the specified amount of 300kg/h. Sandwithdrawal from the cyclone return leg then commencedat the same rate. The combustion temperature wasmaintained at 875–880 1 C. DCB feed was started afterstable conditions were established. The gas residence timewas about 1.5s. During steady state operation periods, allimportant test parameters were measured and flue gas andsolid samples were taken. Analyses for DCB, CBs (whichcould be formed from DCB decomposition) and PCDD/Fswere performed using Environment Canada method EPS1/RM/2 for semi-volatile organic compounds (Environ-ment Canada, 2005) and US Environmental ProtectionAgency (EPA) method 0030 for volatile organic com-pounds (EPA, 2005).Four tests were performed. Two of them wereconducted with the addition of limestone to study itseffect on the emissions. The destruction and removalefficiency (DRE), defined as (DCB input  –CB output )/DCB input  100%, and the combustion efficiency, definedas [CO 2 ]/([CO 2 ]+[CO])  100%, were evaluated from theresults. 3. Results and discussion The test conditions and results are summarized in Table2. The DCB content in the feed, DCB/(DCB+fuel+sand),was at the level of 6wt%. The initial DCB concentration inthe gas was about 3600vppm. Even with such highconcentrations, the DRE reached 99.99995–99.99999%,with CB concentration in the flue gas below 0.1 m g/m 3 . The ARTICLE IN PRESS Fig. 1. Structures of PCB and DCB.  combustion efficiency was above 99.9% in all cases. Thisperformance is quite encouraging (EPA requires that thedestruction efficiency is above 99.9999%, and combustionefficiency above 99.9% for rotary kiln incinerators;see (Environment Canada, 2003b). See also (UNEP, 1998) for required destruction efficiency). The PCDD/Femission level (2,3,7,8 TCDD equivalent) in Tests 1, 2 and4 was from 7.5 to 42ng/Nm 3 , which exceeded thepermitted regulated level of 5ng/Nm 3 . However, it isgenerally accepted that PCDD/Fs mainly form in the ARTICLE IN PRESS Table 1Analysis of DCB used in the testsConstituent ConcentrationBenzene  o 0.5%Chlorobenzene  o 0.5%1,2-Dichlorobenzene  o 0.5%1,3-Dichlorobenzene 100%1,4-Dichlorobenzene  o 0.5%Calorific value 9.9 MJ/kgFig. 2. Schematic of the CFBC pilot plant.Table 2Test conditions and resultsConditions Test 1 Test 2 Test 3 Test 4DCB feed rate kg/h 21.6 20.7 18.8 21.1Chlorine feed rate kg/h 10.4 10 9.1 10.2Fuel oil feed rate kg/h 30 29 32 NASand feed rate kg/h 300 300 255 284Limestone feed rate kg/h 0 0 45 16Ca:Cl 2  Mole ratio 3 1Combustor temp.  1 C 875 877 885 880O 2  % 6.7 6.7 6.9 6.6 Measured results CO ppmv 84 97 21 36CO 2  % 10.9 10.8 12.2 11.3NO x  ppmv 30 30 38 29N 2 O ppmv 0.53 NA 0.27 0.48SO 2  ppmv 150 154 43 86Combustion efficiency % 99.92 99.91 99.98 99.97DRE % 99.99995 99.99994 99.99997 99.99999PCDD/Fs a ng/Nm 3 42 7.5 3.3 8Total CB in solids,  m g/g  m o 0.04  o 0.003 0.002  o 0.002PCDD/Fs in solids, a ng/g 0.37 0.024 0.035  o 0.01 a Reported as 2,3,7,8-TCDD equivalent.  post-combustion section (Stanmore, 2004; Preto et al., 2005). Under properly controlled post-combustion condi-tions, or by using a dry scrubber at the back end (Domingoet al., 2000), PCDD/F emissions can be effectively reduced.It is noteworthy that the levels of CBs and PCDD/Fs in thetreated solids (Table 2) suggest that the regulation for PCBincineration ( o 0.5 m g/g for PCB and o 1ng/g for PCDD/Fs) can be met.The high destruction efficiency shown in Table 2 suggeststhat, under the test conditions, combustion of the organictoxics is very effective. To evaluate the rate of combustion,the combustor can be taken as a plug flow reactor (Niessen,1995). Based on plug flow and instant vaporization of theliquid DCB we obtain the following relation:d C  d t  ¼ k  g C  , (1)where  C   is the concentration of DCB,  t  time and  k  g  the rateconstant which includes oxygen concentration. In Eq. (1), theoxidation of DCB is taken as first order in DCB concentra-tion, as is close to the observed order of hydrocarbonsburning in oxygen (Kanury, 1975). With Eq. (1) and our testresults we obtain  k  g E 10s  1 for the rate constant for DCBoxidation in the combustor, which is in reasonable agreementwith the value calculated from literature data for DCBoxidation in pure air, 13s  1 (Niessen, 1995).In the case of the targeted contaminated soil, the toxicsneed to be eliminated from the soil and hence desorption of the toxics is the prerequisite. In addition to adequate masstransfer for the desorption process, heat transfer is alsoimportant, particularly for sludge where large amounts of moisture will be vaporized.For heat transfer, the CFBC has the advantage that thesolids being treated are surrounded by hot gas and hotparticles. This enables not only more effective heat transferfrom gas to solid, but also intensive heat transfer betweensolids by radiation, and conduction—which is not attainablein the solid bed of a rotary kiln. As a result, adequate heattransfer can be achieved in a CFBC at lower temperatures.As has been mentioned earlier, lower incineration tempera-tures can reduce NO x  emissions and vaporization of hazardous metals. Moreover, the lower incineration tem-peratures allow large potential savings of coal for incinera-tion of materials like Sydney Tar Pond sludge—estimatesbased on sample data (Jia et al., 2005) and a temperature of 1100 1 C typically required for incineration of halogenatedsubstances in rotary kilns (UNEP, 1998) show thatthousands of tones of coals could be saved for incinerationof the 700,000 tones of sludge using CFBC.The major issue is then the removal of the toxics fromthe soil, and we discuss the conditions here.As a simplified treatment, we take the rate of desorptionof the toxics as proportional to the amount of toxicscontained in the soild M  d t  ¼ k  s M  , (2)where  k  s  is the coefficient dependent on temperature andthe properties of the soil and the toxics, and  M   the amountof the toxics. A similar treatment has been found valid forthe release of a number of volatiles from soil (Bucala ´ et al.,1996). In Eq. (2), the effect of the gas phase concentrationof the toxics on the desorption rate is not considered, as thegas phase concentration can be expected to be very low,owing to the fast oxidation rate discussed above. Based onEq. (2) the toxics remaining, expressed as a fraction of theinitial amount of toxics, is obtained as a function of theresidence time  y ¼  M M  0 ¼ exp ð k  s t Þ , (3)where  M  0  denotes the initial amount. For a CFBC unit inwhich the solids are fully recirculated, the discharged solidswill have experienced a distribution of residence times. Thiseffect on the toxics content of the treated solids can bediscussed in the following way.In the CFBC, the fraction of solids, which experiencedonly one cycle of circulation, may be given as  f  1  ¼  F  0 F  R þ F  0 ¼  g 1 þ g , (4)where  F  0  and  F  R  denote the mass flow rate of the feed andthe recirculated solids, respectively, and  g ¼ F  0 = F  R  is theratio of the two rates. The fraction experiencing two cyclesof circulation would be  f  2  ¼  f  1 F  R F  R þ F  0 ¼  f  1 1 þ g ¼  g ð 1 þ g Þ 2  . (5)Similarly, the fraction experiencing  n  cycles of circula-tion is  f  n  ¼  g ð 1 þ g Þ n  . (6)Based on Eq. (3), the level of the toxic remaining for theparticles that experienced  n  cycles may be given as  y n  ¼ exp ð k  s t rs n Þ , (7)where  t rs  is the effective residence time of the solid particlesin one cycle. The average level of toxic remaining in thedischarged solids can then be given as ¯  y ¼ X 1 n ¼ 1  y n  f  n  ¼ X 1 n ¼ 1 g  exp ð k  s t rs n Þð 1 þ g Þ n  ¼  g  exp ð k  s t rs Þ 1 þ g  exp ð k  s t rs Þ .(8)Accordingly, when  t rs , which is dependent on the velocityof gas, is fixed, the level of toxics remaining in the treatedsolids can be controlled by adjusting  g , the ratio of thesolids feed to the recirculated solids in the CFBC. Clearly, ¯  y  decreases with decreasing  g . By lowering the value of   g ,sufficiently low levels of the toxic residue in the solids canbe achieved.Now we consider the overall effect of desorptionfrom the soil and gas phase oxidation on concentrationof the toxics in the gas phase. The concentration can be ARTICLE IN PRESS  expressed byd C  d t  ¼ R d  R o  ¼ k  s ¯  yM  0 r  k  g C  , (9)where  R d  and  R o  denote the rate of desorption andoxidation of the toxics, respectively, and  r  is the mass of the soil in a unit volume of the combustor. For aconservative estimation we take  r  as not varying alongthe height of the combustor. In this way we obtain fromEq. (9), C  g  ¼ k  s ¯  yM  0 r k  g 1  exp ð k  g t rg Þ   , (10)where  C  g  is the gas phase concentration of the toxics at theexit of the combustor, and  t rg  is the gas residence time.With a value of   k  g  on the order of that for DCB (10s  1 ) asevaluated from the present work, the gas phase concentra-tion of the toxics  C  g  will essentially be  k  s ¯  yM  0 r = k  g , as canbe seen from the plot of   C  g = ð k  s ¯  yM  0 r = k  g Þ  as a function of residence time (Fig. 3). Therefore, the concentration of thetoxics can be controlled by controlling the value k  s ¯  yM  0 r = k  g . So long as  k  s ¯  yM  0 r  k  g , the concentrationof the toxics in the flue gas will be very low. Since the valueof   k  s  can be evaluated experimentally (by thermogravi-metric analysis, for example), and the other parametervalues are either known or can be calculated (in the case of  ¯  y ), operating conditions for the required DRE can bedetermined to achieve the desired performance. Moreover,Fig. 3 suggests that increasing the residence time above 1swould not increase the DRE when treating the contami-nated soil.The effect of limestone addition in this work is notconclusive. As can be seen from Table 2, the test with alarge amount of limestone (Test 3) showed low PCDD/Fs,but the test with a smaller amount of limestone (Test 4) didnot show a significant difference. By contrast, limestoneaddition might have a greater effect on CO emissions—theCO level appears to decrease with increasing limestoneaddition as has been observed elsewhere in the literature(Leckner et al., 1992).With the key data obtained in the above tests, 17,000tones of PCB-contaminated material in Quebec wassuccessfully treated in a full-scale incinerator (NaturalResources Canada, 2000). Moreover, by using a dryscrubber with hydrated lime injection at the back end,the PCDD/F emissions from the stack, which were largelyrelated to post-combustion conditions of the incinerator(Stanmore, 2004; Preto et al., 2005), were well below the regulated level.Test results for treatment of the sludge from Sydney TarPonds using CFBC have been reported elsewhere (Jia et al.,2005). Promising results have also been achieved. More-over, the concentrations of PCDD/Fs, particulate matter,metals and HCl were all below both the current limits andthe gas release limits to be implemented in Canada in 2008,indicating the high potential of this technology for thelarge-scale remediation program. 4. Conclusions CFBC incineration appears to be effective for remedia-tion of contaminated soil/sludge. Results of the testsreported here suggest that the destruction rate of organictoxics in the CFBC is very fast. PCDD/F emissions can besuppressed. The toxics in treated soil can be reduced tovery low levels. The heat transfer appears to be adequateunder the operating temperature of the typical CFBC. Thiscould result in savings of thousands of tones of coal inincineration of materials such as those of the contaminatedsludge from the Sydney Tar Ponds. Acknowledgment Funding of the test work by Cintec Environment Inc. isgratefully acknowledged. References Bucala ´, V., Saito, H., Howard, J.B., Peters, W.A., 1996. Productscompositions and release rates from intense thermal treatment of soil.Industrial and Engineering Chemistry Research 35, 2725–2734.Domingo, J.L., Schuhmacher, M., Mu ¨ller, L., Rivera, J., Granero, S.,Llobet, J.M., 2000. Evaluating the environmental impact of an oldmunicipal waste incinerator: PCDD/F levels in soil and vegetationsamples. Journal of Hazardous Materials 76, 1–12.Environment Canada, 2003a. Destruction technologies for polychlori-nated biphenyls. Available at  / S Environment Canada, 2003b. Guidelines for the Management of WastesContaining Polychlorinated Biphenyls (PCBs) Available at  / S Environment Canada, Regulations, 2005. Available at  / S EPA, 2005. Test Methods. Available at  / S Government of Nova Scotia, 2004. $400 Million committed for Tar Pondscleanup. Available at  / S ARTICLE IN PRESS 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 t rg,  [s]    C   g    /   (      k     s    y     M     0            k     g    ) ,   [  -   ] Fig. 3. Dependence of gas phase concentration of toxics on gas residencetime.
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks