Comparison of Pb, Cd, Zn, and Cu chlorination during pyrolysis and incineration

Fuel 194 (2017) 257–265

Contents lists available at ScienceDirect

Fuel journal homepage: www.elsevier.com/locate/fuel

Full Length Article

Comparison of Pb, Cd, Zn, and Cu chlorination during pyrolysis and incineration Si-Jia Wang a,b, Pin-Jing He b,c, Wen-Tao Lu a,b, Li-Ming Shao b,c, Hua Zhang a,⇑ a

State Key Laboratory of Pollution Control and Resources Reuse, Tongji University, 1239 Siping Road, Shanghai 200092, PR China Institute of Waste Treatment and Reclamation, Tongji University, 1239 Siping Road, Shanghai 200092, PR China c Centre for the Technology Research and Training on Household Waste in Small Towns & Rural Area, Ministry of Housing and Urban-Rural Development of PR China, 1239 Siping Road, Shanghai 200092, PR China b

h i g h l i g h t s
PVC can directly react with ZnO or CuO, while PbO or CdO reacts with HCl from PVC.
PbO was more easily chloridized than other metal oxides during thermal treatment.
NaCl alone can convert PbO to PbCl2 via a liquid-solid reaction at >801 °C.

a r t i c l e

i n f o

Article history: Received 26 August 2016 Received in revised form 25 November 2016 Accepted 7 January 2017 Available online 12 January 2017 Keywords: Heavy metals Chloride Thermal treatment Volatilization Reaction

a b s t r a c t The mechanisms of chlorination of PbO, CdO, ZnO, and CuO by poly(vinyl chloride) (PVC) and sodium chloride (NaCl), including their reaction temperatures, pathways, and products, were studied and compared. It was found that PVC can chloridize the four oxides via different mechanisms, producing corresponding chlorides. The heavy metal oxides in PVC–PbO and PVC–CdO were chloridized by gas–solid reaction with HCl, while direct chlorination by PVC occurred at 190 °C in PVC–ZnO and PVC–CuO, as their initial temperatures for weight loss were 35–44 °C lower than that of PVC decomposition. The relatively facile chlorination of PbO as compared with the other oxides might be a reason why Pb was more volatile than the other metals. NaCl had no chlorination effect on CdO, ZnO, or CuO in the absence of other media. It was found for the first time that NaCl alone could convert PbO to PbCl2 via a liquid–solid reaction when the temperature was higher than the melting point of NaCl (801 °C), and oxygen was not involved. The chlorination effect of NaCl was markedly weaker than that of PVC. Since both PVC and NaCl are the most important chlorine sources in solid waste, their chlorination effects on heavy metals cannot be ignored. Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction Waste-to-energy conversion has been an important technology for the treatment of mixed wastes such as municipal solid waste (MSW), medical waste, and hazardous waste. The migration of lead, cadmium, zinc, and copper in wastes during thermal treatments, including pyrolysis and incineration, can lead to secondary environmental pollution [1,2]. Chlorination reactions are crucial to the distribution of heavy metals into gas or solid phases. Chlorine in wastes is mainly from poly(vinyl chloride) (PVC) and sodium chloride (NaCl) respectively [3–6]. High temperature (650–850 °C) and the presence of PVC can enhance volatilization of Pb, Cd, Zn, and Cu [4]. In some cases, ⇑ Corresponding author. E-mail address: [email protected] (H. Zhang). http://dx.doi.org/10.1016/j.fuel.2017.01.035 0016-2361/Ó 2017 Elsevier Ltd. All rights reserved.

PVC can be used to remove heavy metals from fly ash (FA) produced by MSW incineration [7], sewage sludge ash (SSA) [8,9], electric arc furnace dust (EAFD) [10,11] and other wastes [12]. Thermal treatment (850 °C) using PVC (5% by weight) has been found to increase volatilization of the heavy metals Pb, Cd, and Zn by 10–15% [7] in FA; recovery ratios for the metals reached >97.0% at 1000 °C upon addition of PVC to EAFD [11]. Other researchers [8,9] have found that increasing the temperature and PVC content of the SSA could increase the ratios of Pb, Cd, Zn, and Cu removed. An increased partitioning tendency of some heavy metals into FA or flue gas has also been found when NaCl was added to simulate MSW [4], co-combusted bituminous coal, and recovered solid fuel [12], FA [13], EAFD [14] or other system [15]. In Chiang’s study, the effect of inorganic chloride on metal volatilization was less significant than that of organic chloride because Na has a stronger

258

S.-J. Wang et al. / Fuel 194 (2017) 257–265

were used in the study, and both were calibrated with standard compounds. The FTIR spectral region was scanned at 4000– 800 cm1 at a rate of 0.6329 cm s1 and a resolution of 8 cm1. The temperature of the gas line was set at 220 °C.

affinity for Cl [4] than heavy metals have. Two kinds of chemical reactions have been proposed, i.e., the formation of heavy metal chlorides by the reaction of heavy metal oxides with (1) HCl produced from the reaction of NaCl and H2O or SO2 [5,16] and mineral matrixes (e.g., SiO2 and Al2O3) [14,16]; and with (2) gaseous Cl2 from the reaction of NaCl with mineral matrixes (e.g., SiO2 and Al2O3) [14]. The heavy metal chlorides formed can significantly influence the volatilization of heavy metals during thermal treatment and depend on the temperature and residence time. Studies on the treatment of mixed wastes have generally focused on the volatilization of heavy metals; the present study, however, paid attention to chlorination reactions, which are fundamental to volatilization. Thermal analysis and simulated thermal treatment were utilized to investigate the mechanism of chlorination of heavy metal oxides (PbO, CdO, ZnO, and CuO) by organic chloride (PVC) and inorganic chloride (NaCl). This work examined (1) the temperatures of chlorination and (2) stoichiometry and products of incineration and pyrolysis, and (3) a comparison of chlorination of heavy metal oxides during these processes.

A horizontal tube furnace system was used for the simulated thermal treatment experiments. Each sample (1.0–2.0 g) was transferred to a corundum or platinum crucible to prevent interference by Al2O3 in the chlorination reactions for NaCl–PbO, NaCl– CdO, NaCl–ZnO, and NaCl–CuO mixtures. It was then heated from 50 to 190, 350, 450, 600, 780, 850, or 900 °C at a rate of 10 °C min1; residence times were 30–180 min. The residues in the crucibles and the condensates on the surface of the glass rings placed at the end of the quartz tube were collected and analyzed by X-ray diffraction (XRD) measurements (D8 Advance diffractometer, 40 mA and 40 kV; Bruker Corporation, Germany) using Cu Ka radiation. The step size was 0.02° and the scanning time was 0.1 s.

2. Materials and methods

3. Results

2.1. Materials

3.1. Thermal analysis of PVC

Oxides, carbonates, and acetates of the heavy metals Pb, Cd, Zn, and Cu, the most common heavy-metal compounds of concern in solid waste, usually decompose into oxides during heating. Therefore, PbO, CdO, ZnO, and CuO were used as representative heavy metal compounds in this study. Analytical reagent grade PVC (Sigma-Aldrich Corporation, Shanghai, China), NaCl, and metal oxides (Sinopharm Chemical Reagent Co. Ltd., Shanghai, China) with particle sizes below 150 lm were dried and blended at ratios indicated in Table 1.

All characteristic temperatures in the TGA, DSC, and DTG curves are summarized in Table 2. According to Fig. 3a and b, PVC decomposition during incineration and pyrolysis can be divided into two stages. During the first stage (225–350 °C), PVC loses 62–64% of its weight, resulting in endothermic peaks at 277 °C (incineration) or 289 °C (pyrolysis) on the DSC curves. The main products of PVC decomposition are HCl and benzene from polycondensation of carbon chains during both incineration and pyrolysis, similar to products observed in other investigations [17,18]. As shown in Fig. 1, HCl started to form at 229 or 250 °C, reaching maximum levels at 285 °C during incineration and 290 °C during pyrolysis. During the second stage of incineration (350–554 °C), PVC loses 36% of its weight (Fig. 3a) because of carbon oxidation and CO2 volatilization (Fig. 1a). During pyrolysis, the weight loss reached 26% (Fig. 3b), and the main gaseous products were alkanes and alkenes (Fig. 1c), which were also observed by Ma et al. and Yannick et al. [19,20].

2.2. Thermal analysis experiments Experiments using simultaneous thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) (Q600 SDT; TA Instrument Corporation, USA) were conducted on PVC and NaCl, as well as PVC–PbO, PVC–CdO, PVC–ZnO, PVC–CuO, NaCl–PbO, NaCl–CdO, NaCl–ZnO, and NaCl–CuO mixtures (<40 mg). The heating temperatures for TGA were increased from 50 to 190, 350, 450, 600, 850, or 900 °C at a rate of 10 °C min1. Gases used for simulating incineration and for pyrolysis, air and N2 respectively, were introduced at a flow rate of 100 mL min1. TGA (STA6000; Perkin Elmer Corporation, USA) integrated with Fourier transform infrared (FTIR) spectroscopy (Frontier; Perkin Elmer Corporation, USA) was also used to analyze the gaseous products. In this case, the TGA temperature program for PVC–ZnO and PVC–CuO mixtures consisted of heating from 50 to 190 °C, holding at 190 °C for 30 min, and final heating from 190 to 900 °C. Two TGA instruments

2.3. Simulated thermal treatment experiments

3.2. Thermal analysis of PVC and the heavy metal oxide mixtures The thermal analysis results for the PVC–PbO, PVC–CdO, PVC– ZnO, and PVC–CuO mixtures are summarized in Fig. 3 and Table 2. The relative weight losses of all mixtures in the first stage were less than that of PVC of the same mass. The aforementioned endothermic peak for PVC decomposition during the first stage was replaced by an exothermic peak or was diminished in the presence of heavy metal oxides. In contrast to the TGA curves of PVC, those of the four

Table 1 Mass and molar fractions of PVC, NaCl, and metal oxides in the samples. Sample

PVC-PbO PVC-CdO PVC-ZnO PVC-CuO NaCl-PbO NaCl-CdO NaCl-ZnO NaCl-CuO

PVC

NaCl

PbO

CdO

ZnO

CuO

wt.%

mol%

wt.%

mol%

wt.%

mol%

wt.%

mol%

wt.%

mol%

wt.%

mol%

45.5 59.4 69.7 70.1 – – – –

75.0 75.0 75.0 75.0 – – – –

– – – – 34.4 47.7 59.0 60.0

– – – – 66.7 66.7 66.7 66.7

54.5 – – – 65.6 – – –

25.0 – – – 33.3 – – –

– 40.6 – – – 52.3 – –

– 25.0 – – – 33.3 – –

– – 30.3 – – – 41.0 –

– – 25.0 – – – 33.3 –

– – – 29.9 – – – 40.0

– – – 25.0 – – – 33.3

259

S.-J. Wang et al. / Fuel 194 (2017) 257–265 Table 2 Characteristic decomposition temperatures and gaseous products of PVC and of PVC–PbO, PVC–CdO, PVC–ZnO, and PVC–CuO mixtures during incineration and pyrolysis. Sample

PVC

*

b c

PVC-CdO

PVC-ZnO

PVC-CuO

Parameter

Air

N2

Air

N2

Air

N2

Air

N2

Air

N2

Initial temperature for weight loss (°C)

230

225

260

265

250

255

186

187

194

190

Temperature for DTG peaks (°C)

276 435 523 542 552

285 328 454

298 431 481 583 677

300 329 460 667

286 518 616

290 447 555 614

226 387 574

225 277 513

268 352 435 506

272 447

–c

301b

314b 478b 577b

317b 501a

222b 392b 580b

221b 309b 519b

358b 499b

272b 447b

277a 444b 525b 541b 551b

289a

Initial temperature for HCl emission (°C)

229

250

277

286

260

281

190*

190*

246

268

Temperature for maximum HCl emission (°C)

285

290

305

310

300

306

291

286

297

300

Total weight loss (%)

100

90.4

91.8

75.5

82.0

91.8

85.4

68.4

85.2

68.5

Remained solid (%)

0

9.6

8.2

24.5

18.0

8.2

14.6

31.6

14.8

31.5

Temperature for DSC peaks (°C)

a

PVC-PbO

The temperature was increased from 50 to 190 °C and held for 30 min, then increased from 190 to 900 °C. Endothermic peak on DSC curve. Exothermic peak on DSC curve. Not found.

0.08

CO 2 Benzene

(a) HCl 0.06 285 °C 3431-2526cm-1 CO 2

0.04

Absorbance (a.u.)

0.02 0.00 0.04

750-600cm-1

2400-2233cm-1

H2O

1800-1200cm-1

0.00 8

390 °C

4

2240-2021cm-1

H2O

CO 2

H2O

0.00 0.8 585 °C 0.4

20

0.2

CO 2

H2O 0.0 4000 3500 3000 2500 2000 1500 1000 500

Wavenumber (cm-1) 0.08

(c)

0.06 290 °C

Absorbance (a.u.)

2 0 40 CO2 (2400-2233cm-1) 30

CO2

0.6

HCl (3131-2526cm-1 )

6

CO 2 CO

0.02

(b)

0.04 Gram-Schmidt 0.02

H2O

3780-3530cm-1

0.06

HCl 3431-2526cm-1

0.04

H2O

0.02

3780-3530cm

CO 2 -1

H2O

10 0

0

20

40

60

80

100

120

100

120

Time (min) CO2 Benzene

0.020

750-600cm-1

0.015 Gram-Schmidt

(d)

0.010

1800-1200cm-1

2400-2233cm

-1

0.005

0.00 0.04 410 °C

0.000 8

HCl (3131-2526cm-1 )

6

0.02 0.00 0.04

CO 2

H2O 472 °C

CO 2

2 0 -1 4 Alkane (3137-2872cm )

Alkanes Alkenes

0.02

H2O

4

H2O

CO 2

H2O

2

CO 2

0.00 4000 3500 3000 2500 2000 1500 1000 500 -1

Wavenumber (cm )

0

0

20

40

60

80

Time (min)

Fig. 1. FTIR spectra of PVC decomposed products during incineration and pyrolysis. (a) Gaseous products at different temperatures during incineration; (b) Evolution of characteristic products during the whole incineration process; (c) Gaseous products at different temperatures during pyrolysis; (d) Evolution of characteristic products during the whole pyrolysis process.

260

S.-J. Wang et al. / Fuel 194 (2017) 257–265

190 °C

190 °C

(a)

100

100

PVC PVC-CuO

96

96

94

94

92

PVC PVC-CuO

92

PVC-ZnO

PVC-ZnO

90

90 88

190 °C

98

98

Weight (%)

190 °C

(b)

0

10

20

30

40

50

Time (min)

60

70

80

88

0

10

20

30

40

50

60

70

80

Time (min)

Fig. 2. Thermal analysis curves of PVC and of PVC–ZnO and PVC–CuO mixtures during incineration (a) and pyrolysis (b).

Fig. 3. Thermal analysis curves of PVC and of PVC–PbO, PVC–CdO, PVC–ZnO, and PVC–CuO mixtures during incineration (a, c, e, g, i) and pyrolysis (b, d, f, h, j).

S.-J. Wang et al. / Fuel 194 (2017) 257–265

TGA (%)

DTG (%/min)

261

DSC (W/g)

Fig. 3 (continued)

mixtures are of two types. The initial temperatures for weight loss (Ti) of PVC–PbO (260–265 °C) and PVC–CdO (250–255 °C) mixtures were 25–40 °C higher than that of PVC, while those for weight loss of PVC–ZnO (186–187 °C) and PVC–CuO mixture (190–194 °C) were markedly lower than that of PVC. These results indicate that the reactions between PVC and the former two heavy metal oxides occurred after PVC decomposition, whereas those between PVC and the latter two oxides occurred before decomposition. To confirm these conclusions, thermal analysis experiments at 190 °C for 30 min were conducted. As shown in Fig. 2, pure PVC lost about 1% of its weight in the entire process, while the PVC–ZnO and PVC– CuO mixtures respectively lost about 5% and 11% during both incineration and pyrolysis. These results prove that ZnO and CuO can promote PVC decomposition. The FTIR spectra show that the gaseous products from each mixture were almost the same. The initial temperatures for HCl emission from the PVC–CuO mixture due to incineration or pyrolysis were higher than that from PVC. HCl from decomposition of the PVC–ZnO mixture was detected while heating at 190 °C for 10 min (Table 2), during which the weight-loss rate of the PVC–ZnO mixture changed. The relationship between the integral absorbance from FTIR spectroscopy and the yield of gaseous component is linear [21], as shown in Eq. (1); therefore, Eq. (2) could be used for quantitative comparison of HCl yields.

Z

IHCl ¼ x

t2

t1

Z v2 v1

½Aðm; tÞdmdt

. ¼ IHCl IHCl RHCl x x PVC  100%

 Mx

ð1Þ ð2Þ

where x is the sample mixture (PVC–PbO, PVC–CdO, PVC–ZnO, or and IHCl PVC-CuO mixture); IHCl x PVC are the integrated absorbances of HCl formed by decompositions of sample x and PVC, respectively (min cm1 mg1), which reflect the levels of HCl; t is the heating time (min), with t1 and t2 comprising the time ranges of HCl evolution; m is the wavenumber (cm1), with m1 and m2 comprising the wavenumber range of HCl; A(m, t), which is dimensionless, is the absorbance at wavenumber m at time t; Mx is the amount of PVC in the sample x (mg); RHCl is the ratio of IHCl to IHCl x x PVC , which reflects the tendency of sample x to react with HCl. Results are summarized in Table 3. According to Table 3, IHCl values for PVC–PbO, PVC–CdO, PVC– ZnO, and PVC–CuO mixtures are less than that of PVC during incineration and pyrolysis, similar to the trend for relative weight loss. RHCl PVCPbO for incineration is 49.1% and that for pyrolysis is 43.3%, both of which are lower than corresponding values for PVC–CdO and PVC–CuO mixtures. RHCl PVCZnO is not discussed because of the long residence time of the PVC–ZnO mixture at 190 °C. In summary, these results confirm the reactions between heavy metal oxides and PVC or HCl from PVC decomposition. The tendency of PbO to react with HCl was greater than that of CdO and CuO; their values during incineration are higher than those during RHCl x pyrolysis. 3.3. Simulated thermal treatment of PVC and heavy metal oxides Products of chlorination and physicochemical reactions under different final temperatures and atmospheres, as revealed by XRD analyses, may be compared using Table 4. The chlorination

262

S.-J. Wang et al. / Fuel 194 (2017) 257–265

Table 3 Weight losses of PVC in each system during the first stage.

a b

Sample

Atmosphere

Weight loss in the first stage (%)

Weight loss of mixture/weight loss of PVC (%)

IHCl (min cm1 mg1)

RHCl (%)

PVC PVC PVC-PbO PVC-PbO PVC-CdO PVC-CdO PVC-ZnO PVC-ZnO PVC-CuO PVC-CuO

Air N2 Air N2 Air N2 Air N2 Air N2

64.2 61.4 37.1 41.5 54.0 47.5 42.2 43.3 59.6 59.5

100 100 57.8 67.6 84.1 77.4 65.7 70.5 92.8 96.9

2.77 3.00 1.36 1.30 1.85 1.48 1.56 1.08 1.78 1.36

100a 100a 49.1a 43.3a 66.8a 49.3a 56.3b 36.2b 64.3b 45.3b

The temperature was increased from 50 to 900 °C. The temperature was increased from 50 to 190 °C and held for 30 min, then increased from 190 to 900 °C.

Table 4 Products of the mixture samples in the simulated thermal treatment experiments. Sample

Atmosphere

Temperature (°C)

Residue

Condensate

PVC-PbO

Air N2 Air N2

350 350 900 900

PbCl2 PbCl2, PbO PbCl2 PbO, Pb2O

– – PbCl2 PbCl2

PVC-CdO

Air N2 Air N2

350 350 900 900

CdCl22H2O, CdO, Cd3Cl2O2 CdCl22H2O, CdO, Cd3Cl2O2 CdO Amorphous

– – CdCl22H2O CdCl22H2O, Cd

PVC-ZnO

Air N2 Air N2 Air N2

190 190 350 350 900 900

ZnO, Zn2OCl22H2O ZnO, Zn2OCl22H2O ZnO, Zn2OCl22H2O Zn2OCl22H2O, Zn5(OH)8Cl2H2O ZnO ZnO

– – – – ZnCl2, Zn ZnCl2

PVC-CuO

Air N2 Air N2 Air N2

190 190 350 350 900 900

CuO, CuCl CuO, CuCl CuO, CuCl CuO, CuCl, Cu CuO Cu2O, Cu

– – – – CuClOH –

–: Not found.

products found in the residues at 350 °C were heavy metal chlorides or oxychlorides such as PbCl2, CdCl22H2O, Cd3Cl2O2, Zn2OCl22H2O, Zn5(OH)8Cl2H2O, and CuCl. Chlorination of ZnO and CuO at 190 °C may explain the reaction between PVC and ZnO or CuO. When the temperature was raised to 900 °C, heavy metal chlorides were detected in the condensates. The chlorination products of incineration and pyrolysis were similar, but heavy metal oxides were reduced only during pyrolysis.

3.4. Thermal analysis of NaCl The melting point of PbO is 886 °C, whereas those of CdO, ZnO, and CuO are higher than 900 °C. In order to determine the possibility of the reaction between NaCl and heavy metal oxides, the final temperatures for the thermal analysis experiment were set at 850 °C for the NaCl–PbO mixture and at 900 °C for the other three mixtures. If there was no chlorination reaction between NaCl and heavy metal oxides, then the total weight loss of the mixtures would match the mass fraction of NaCl because of nonvolatilization of heavy metal oxides below 850 or 900 °C. The pure NaCl sample gradually lost weight from its melting point of 801 °C and endothermic peaks on its DSC curves at the same temperature (Fig. 4a and b). NaCl completely volatilized when the temperature was increased to 850 °C for 60 min (incineration) or 80 min (pyrolysis).

3.5. Thermal analysis of the mixtures of NaCl and heavy metal oxides The thermal analyses results of the mixtures of NaCl and heavy metal oxides may be classified into two types, as shown in Fig. 4 and Table 5. Weight loss of the NaCl–PbO mixture started at 801 °C, corresponding to endothermic peaks on the DSC curves. When the temperature was increased from 50 to 850 °C and then maintained for 180 min, the weight losses of the mixtures were approximately 86% (incineration) and 65% (pyrolysis), which exceed the mass ratio of NaCl (34.4%) in the NaCl–PbO mixture. These results suggest physical or chemical changes in the NaCl– PbO mixture, the mechanism of which is discussed in Section 4.2. When the weight losses of the mixtures reached equilibrium, they almost matched (<5% absolute difference) the NaCl mass in each mixture. These results suggest a minor effect on or no chlorination of CdO, ZnO, and CuO. Simulated thermal treatment experiments were thus necessary to confirm the absence of chlorination products. 3.6. Simulated thermal treatment experiments on NaCl and heavy metals oxides The compositions of the residues from the simulated thermal treatment experiments were analyzed by XRD measurements, as shown in Table 6. The above thermal analysis results imply physical or chemical changes in the NaCl–PbO mixture, which can be

S.-J. Wang et al. / Fuel 194 (2017) 257–265

TGA (%)

DSC (W/g)

263

Temperature (°C)

Fig. 4. Thermal analysis curves of NaCl and of NaCl–PbO, NaCl–CdO, NaCl–ZnO, and NaCl–CuO mixtures during incineration (a, c, e, g, i) and pyrolysis (b, d, f, h, j).

264

S.-J. Wang et al. / Fuel 194 (2017) 257–265

Table 5 Thermal analysis results for NaCl and for NaCl–PbO, NaCl–CdO, NaCl–ZnO, and NaCl–CuO mixtures (The holding times were set according to the time required for the weight loss to reach a constant value). NaCl

Initial temperature for weight loss (°C) Temperature for DSC peak (°C) Holding time (min) Weight loss (%) a b

NaCl-PbO

NaCl-CdO

NaCl-ZnO

NaCl-CuO

Air

N2

Air

N2

Air

N2

Air

N2

Air

N2

801 801 60a 100

801 801 80a 100

801 801 180a 86

801 801 180a 65

801 801 120b 46

801 801 120b 52

801 801 120b 60

801 801 120b 60

801 801 120b 65

801 801 120b 64

Final temperature was 850 °C. Final temperature was 900 °C.

Table 6 XRD results for the products in the simulated thermal treatment experiments.

*

Sample

Atmosphere

Temperature (°C)

Residue

Condensate

NaCl-PbO

Air, N2 Air, N2

780 850

NaCl, PbO NaCl, PbO, Pb4O3Cl2*

– NaCl, PbCl2

NaCl-CdO

Air, N2 Air, N2

650 900

CdO, NaCl CdO

– NaCl

NaCl-ZnO

Air, N2 Air, N2

190 900

ZnO, NaCl ZnO

– NaCl

NaCl-CuO

Air, N2 Air, N2

190 900

CuO, NaCl CuO

– NaCl

Only found during incineration (Air); –: Not found.

confirmed by the XRD results. When the final temperature was set at 780 °C, the residue was found to contain only NaCl and PbO; when the final temperature was increased to 850 °C, NaCl, PbO, and Pb4O3Cl2 (only found during incineration) were detected in the residues, while the corresponding condensates from incineration and pyrolysis were found to be a mixture of NaCl and PbCl2. The observed Pb4O3Cl2 and PbCl2 evidence the chlorination reaction between NaCl and PbO. The same method was used to study the reactions of NaCl–CdO, NaCl–ZnO, and NaCl–CuO mixtures. When the temperature was lower than 801 °C, the residues included heavy metal oxides and NaCl; when the temperature was 900 °C, only NaCl volatilized into the condensates formed at 801 °C, leaving heavy-metal oxides in the residues.

4. Discussion 4.1. Chlorination reactions between PVC and heavy metal oxides On the basis of the results of thermal analysis and simulated thermal experiments, we can easily conclude that heavy metal oxides reacted with PVC. The reactions proceeded via two possible mechanisms: (1) PVC directly reacted with heavy metal oxides and (2) HCl from PVC decomposition reacted with heavy-metal oxides. In the present study, Ti values for PVC–PbO and PVC–CdO mixtures during incineration and pyrolysis were 25–40 °C higher than that of PVC, but those of PVC–ZnO and PVC–CuO mixtures were 35–44 °C lower than that of PVC. Similarly, Zhang and Takashi [22,23] found a lower Ti value for the PVC–ZnO mixture, although they studied PVC dechlorination instead of the reactions between PVC and heavy metals. Evidently, ZnO and CuO can directly react with PVC; on the contrary, PbO or CdO must react with HCl from PVC decomposition. The FTIR spectra support these conclusions. According to the dechlorination mechanism of PVC [24], HCl is generated along with polyene via 1,4-elimination from a carbon– carbon double bond. The next step is HCl-catalyzed isomerization of the polyene to regenerate the initial structure. For the PVC– PbO and PVC–CdO mixtures, minor amounts of HCl from PVC

decomposition initially reacted with PbO or CdO, thus increasing the Ti and initial temperature for detection of HCl in the PVC– PbO system. Some of the excess HCl volatilized, as manifested by the weight loss in the TGA curve. ZnO and CuO in the PVC–ZnO and PVC–CuO mixtures can directly capture chlorine atoms from the PVC structure. The RHCl value of PbO suggests that it is more x easily chloridized; this behavior may be a reason why tended to Pb volatilize more than the other heavy metals did. A possible explanation of the observed behavior of Pb is the characteristic of non-transitional heavy metals. The chlorination of heavy metal oxides with PVC demonstrated that the transformation of oxides to chlorides can be realized at a relatively low temperature (<300 °C) and that chlorides can be recovered by increasing the temperature to above their melting points. Therefore, PVC can be used as a chlorinating agent for the separation or removal of heavy metals in hazardous waste, e.g., FA [7] and EAFD [12]. For MSW and other waste incineration, in which lead volatilization is undesirable, the existence of PVC in waste should be taken carefully controlled preferably by limiting its source.

4.2. Chlorination reaction between NaCl and heavy metal oxides Ti values and the corresponding endothermic peaks from incineration and pyrolysis in the thermal analysis and simulated thermal treatment experiments on the NaCl–PbO mixture are the same. Some reports suggest that NaCl can promote volatilization of heavy metals in the presence of SO2, SiO2, Al2O3, or H2O [5,15,25]; however, the detection of Pb4O3Cl2 and PbCl2 in these studies suggests that the reaction could happen without other substances. If NaCl and PbO form a eutectic mixture, then the eutectic point is lower than the melting point of NaCl (801 °C) and PbO (886 °C), contradicting the experimental results. The TGA and DSC curves indicate no weight change, lacking an endothermic or exothermic peak; the only lead-containing compound in the residues at 780 °C was PbO. These results suggest that chlorination was not thermodynamically spontaneous below 801 °C. When NaCl melts or vaporizes and reacts with solid-phase PbO, NaCl

S.-J. Wang et al. / Fuel 194 (2017) 257–265

forms free ions, thus increasing the conductivity; NaCl often forms ion groups, which are clusters with a net charge [26]. Therefore, free-state or ionic-state chloride reacts with PbO. The ionic distance in gaseous or liquid NaCl is smaller than that in the solid phase, possibly resulting in lower reactivity of the solid phase as compared with that of gaseous or liquid NaCl. In summary, NaCl alone enhanced PbO volatilization during incineration or pyrolysis. A pilot-scale two-stage fluidized-bed incineration at lowtemperature (up to 800 °C) [27] has been shown to have lead emissions that are lower than those from a traditional hightemperature (up to 1000 °C) one-stage MSW incineration system. This reduction can be explained by the absence of chlorination by NaCl below 850 °C. Results of thermal analysis and simulated thermal treatment experiments for the NaCl–CdO, NaCl–ZnO, and NaCl–CuO mixtures suggest that NaCl cannot chloridize CdO, ZnO, and CuO. The difference between PbO and the other three oxides also indicates high affinity of PbO to chloride, similar to the affinity leading to the reaction between PVC and oxides. In general, the temperature of waste incineration is higher than 850 °C, while the most common temperature for pyrolysis is in the range 400– 700 °C [28]. Therefore, the NaCl content must be lowered to minimize Pb emission during waste incineration; in waste pyrolysis, however, the direct chlorination of heavy metals by NaCl can be ignored. 5. Conclusions (1) PVC can chloridize heavy metal oxides via different mechanisms. The similarity of the chlorination temperatures for PVC–PbO and PVC–CdO mixtures to the PVC decomposition temperature indicates a gas–solid reaction that forms the products PbCl2 and CdCl2. The Ti values of PVC–ZnO and PVC–CuO mixtures were 35–44 °C lower than that of PVC, indicating that the chlorination of ZnO and CuO is a solid– solid reaction at 190 °C. The chlorination of heavy metal oxides by PVC at low temperature was instantaneous, but the reaction between PVC and PbO or CdO was more facile than that between PVC and ZnO or CuO. Pb was more volatile than the other oxides probably because PbO was more easily chloridized than other oxides, which may explain why Pb was more volatile than the others. (2) NaCl could be a Cl donor for PbO only, but not for CdO, ZnO, and CuO. The mechanism of the reaction of NaCl with PbO was a liquid–solid reaction. The final product was liquid PbCl2, which volatilized and transferred to flue gas or condensed onto particles once it formed. A strategy that includes the possible important role of NaCl in PbO chlorination can be proposed. This reaction may be used in the treatment and management of waste and the control of heavy metal migration. Since the most common temperatures in waste pyrolysis is below the reaction temperature, direct chlorination of Pb, Cd, Zn, and Cu on NaCl can be ignored. To minimize Pb emissions in waste incineration, lowering the NaCl content in addition to lowering the lead content of byproducts is desirable. This study establishes the effect and mechanism of chlorination by organic and inorganic chlorides. The reaction is proposed mainly for heavy metal removal during incineration and pyrolysis.

Acknowledgements This study was financially supported by the National Natural Science Foundation of China (No. 21577102), the National Basic Research Program of China (973 Program, No. 2011CB201500),

265

the Fundamental Research Funds for the Central Universities, the Collaborative Innovation Center for Regional Environmental Quality. References [1] Zhang H, He PJ, Shao LM. Fate of heavy metals during municipal solid waste incineration in Shanghai. J Hazard Mater 2010;156:365–73. [2] Verhulst D, Buekens A. Thermodynamic behavior of metal chlorides and sulfates under the conditions of incineration furnaces. Environ Sci Technol 1995;30:50–6. [3] Wang KS, Chiang KY, Chu WT. Fate and partitioning of heavy metals affected by organic chloride content during a simulated municipal solid waste incineration process. J Environ Sci Heal A 1997;32:1877–93. [4] Chiang KY, Wang KS, Lin FL, Chu WT. Chloride effects on the speciation and partitioning of heavy metal during the municipal solid waste incineration process. Sci Total Environ 2007;203:129–40. [5] Jakob A, Stucki S, Struis RPWJ. Complete heavy metal removal from fly ash by heat treatment: influence of chlorides on evaporation rates. Environ Sci Technol 1996;30:3275–83. [6] Belevi H, Moench H. Factors determining the element behavior in municipal solid waste incinerators. 1. Field studies. Environ Sci Technol 2000;34:2501–6. [7] Rio S, Verwilghen C, Ramaroson J, Nzihou A, Sharrock P. Heavy metal vaporization and abatement during thermal treatment of modified wastes. J Hazard Mater 2007;148:521–8. [8] Vogel C, Adam C. Heavy metal removal from sewage sludge ash by thermochemical treatment with gaseous hydrochloric acid. Environ Sci Technol 2011;45:7445–50. [9] Vogel C, Exner RM, Adam C. Heavy metal removal from sewage sludge ash by thermochemical treatment with polyvinylchloride. Environ Sci Technol 2012;47:563–7. [10] Al-Harahsheh M, Al-Otoom A, Al-Makhadmah L, Hamilton LE, Kingman S, AlAsheh S, et al. Pyrolysis of poly(vinyl chloride) and—electric arc furnace dust mixtures. J Hazard Mater 2015;299:425–36. [11] Lee GS, Song YJ. Recycling EAF dust by heat treatment with PVC. Miner Eng 2007;20:739–46. [12] Wu H, Glarborg P, Frandsena FJ, Dam-Johansena K, Jensena PA, Sander B. Trace elements in co-combustion of solid recovered fuel and coal. Fuel Processing Technol 2013;105:212–21. [13] Nowak B, Rocha SF, Aschenbrenner P, Rechberger H, Winter F. Heavy metal removal from MSW fly ash by means of chlorination and thermal treatment: influence of the chloride type. Chem Eng J 2012;179:178–85. [14] Yoo JM, Kim BS, Lee JC, Kim MS, Nam CM. Kinetics of the volatilization removal of lead in electric arc furnace dust. Mater Trans 2005;46:323–8. [15] Wang SJ, He PJ, Xia Y, Lu WT, Shao LM, Zhang H. Role of sodium chloride and mineral matrixes in the chlorination and volatilization of lead during waste thermal treatment. Fuel Process Technol 2016;143:130–9. [16] Matsuda H, Ozawa S, Naruse K, Ito K, Kojima Y, Yanase T. Kinetics of HCl emission from inorganic chlorides in simulated municipal wastes incineration conditions. Chem Eng J 2005;60:545–52. [17] Levchik SV, Weil ED. Overview of the recent literature on flame retardancy and smoke suppression in PVC. Polym Adv Technol 2005;16:707–16. [18] Zhu HM, Jiang XG, Yan JH, Chi Y, Cen KF. TG-FTIR analysis of PVC thermal degradation and HCl removal. J Anal Appl Pyrol 2008;82:1–9. [19] Ma S, Lu J, Gao J. Study of the low temperature pyrolysis of PVC. Energy Fuel 2002;16:338–42. [20] Soudais Y, Moga L, Blazek J, Lemort F. Coupled DTA-TGA-FTIR investigation of pyrolytic decomposition of EVA, PVC and cellulose. J Anal Appl Pyrol 2007;78:46–57. [21] Marsanich K, Baronitini F, Cozzani V, Petarca LG. Advanced pulse calibration techniques for the quantitative analysis of TG-FTIR data. Thermochim Acta 2002;390:153–68. [22] Zhang B, Yan XY, Shibata K, Uda T, Tada M, Hirasawa M. Thermogravimetricmass spectrometric analysis of the reactions between oxide (ZnO, Fe2O3 or ZnFe2O4) and polyvinyl chloride under inert atmosphere. Mater Trans JIM 2002;41:1342–50. [23] Kosuda T, Okada T, Nozaka S, Matsuzawa Y, Shimizu T, Hamanaka S, et al. Characteristics and mechanism of low temperature dehydrochlorination of poly(vinyl chloride) in the presence of zinc(II) oxide. Polym Degrad Stabil 2012;97:584–91. [24] Amer AR, Shapiro JS. Hydrogen halide-catalyzed thermal decomposition of poly(vinyl chloride). J Macromol Sci A 1980;14:185–200. [25] Wey MY, Liu KY, Tsai TH, Chou JT. Thermal treatment of the fly ash from municipal solid waste incineration with rotary kiln. J Hazard Mater 2006;137:981–9. [26] Petravic J, Delhommelle J. Conductivity of molten sodium chloride and its supercritical vapor in strong dc electric fields. J Chem Phys 2003;118:7477–85. [27] Peng TH, Lin CL. Influence of various chlorine additives on the partitioning of heavy metals during low-temperature two-stage fluidized bed incineration. J Environ Manage 2014;146:362–8. [28] Chen DZ, Yin LJ, Wang H, He PJ. Pyrolysis technologies for municipal solid waste: a review. Waste Manage 2014;34:2466–86.