Chemosphere 119 (2015) 682–689

Contents lists available at ScienceDirect

Chemosphere journal homepage: www.elsevier.com/locate/chemosphere

Hydrothermal carbonization of poly(vinyl chloride) J. Poerschmann ⇑, B. Weiner, S. Woszidlo, R. Koehler, F.-D. Kopinke UFZ-Helmholtz Center for Environmental Research, Department of Environmental Engineering, Permoserstr. 15, D-04318 Leipzig, Germany

h i g h l i g h t s  PVC was subjected to hydrothermal carbonization.  Quantitative hydrodechlorination was observed beyond temperatures of 235 °C.  An array of PAHs and O-functionalized breakdown products was detected.  The sorption potential of the hydrochar for organic sorbates proved very low.

a r t i c l e

i n f o

Article history: Received 25 February 2014 Received in revised form 30 June 2014 Accepted 5 July 2014

Handling Editor: Caroline Gaus Keywords: Hydrothermal carbonization Poly(vinyl chloride) Dehydrochlorination Polycyclic aromatic hydrocarbons

a b s t r a c t Poly(vinyl chloride) (PVC) was subjected to hydrothermal carbonization in subcritical water at 180–260 °C. Dehydrochlorination increased with increasing reaction temperature. The release of chlorine was almost quantitative above 235 °C. The fraction of organic carbon (OC) recovered in the hydrochar decreased with increasing operating temperature from 93% at 180 °C to 75% at 250 °C. A wide array of polycyclic aromatic hydrocarbons (PAHs) could be detected in the aqueous phase, but their combined concentration amounted to only 140 lg g1 PVC-substrate at 240 °C. A pathway for the formation of cyclic hydrocarbons and O-functionalized organics was proposed. Chlorinated hydrocarbons including chlorophenols could only be identified at trace levels (low ppb). Polychlorinated dibenzodioxins (PCDDs) and dibenzofurans (PCDFs) could not be detected. The sorption potential of the hydrochar turned out to be very low, in particular for polar organic pollutants. Our results provide strong evidence that hydrothermal carbonization of household organic wastes which can be tied to co-discarded PVC-plastic residues is environmentally sound regarding the formation of toxic organic products. Following these findings, hydrothermal treatment of PVC-waste beyond operating temperatures of 235 °C to allow complete release of organic chlorine should be further pursued. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Hydrothermal carbonization (HTC) is a relatively new technique to treat wet biomasses, in particular waste biomasses (Funke and Ziegler, 2009). The wet matrices are allowed to react in subcritical water at 180–260 °C to produce a highly carbonaceous product called hydrochar. Subcritical water functions as a solvent and reagent for reactions of organic compounds (Kuhlmann et al., 1994). Significant decomposition processes during the hydrothermal treatment of organic matrices include hydrolysis as the initial step, followed by defunctionalization such as dehydration and decarboxylation, and finally recondensation and aromatization (Libra et al., 2011). Appropriate HTC substrates may also include

⇑ Corresponding author. Tel.: +49 341 235 1761; fax: +49 341 235 2492. E-mail address: [email protected] (J. Poerschmann). http://dx.doi.org/10.1016/j.chemosphere.2014.07.058 0045-6535/Ó 2014 Elsevier Ltd. All rights reserved.

waste biomasses such as municipal organic wastes (Oliveira et al., 2013); (Berge et al., 2011). The recycling service in Germany comprises some dedicated bins, among them a recycling bin for organic waste. Plastics, which are frequently associated with organic household waste, may be co-discarded into these recycling bins. Thus, a question about the fate of co-discarded plastics during the HTC treatment of the organic waste arises. This objective is of special significance in case of halogenated materials including PVC, which is widely used in packaging, wrappings, bottles and containers. It is well-known that thermal combustion of PVC results in the formation of charred residues and airborne particulate smoke along with toxic emissions of PAHs, PCDDs and PCDFs (McKay, 2002), as well as chlorophenols and chlorobenzenes (Font et al., 2010). These organic compounds are formed either from the plastics itself or from additives. The latter group is tied to a multitude of functionalities comprising ‘‘softeners’’ such as phthalates, antioxidants such as phenols and

J. Poerschmann et al. / Chemosphere 119 (2015) 682–689

organophosphatides, UV-stabilizers such as amines and piperidyl esters, antistatic agents such as ethoxylated amines, heat stabilizers such as organotins and pigments (Endo, 2002). It was also shown that oxidative coupling products, which can act as precursors for PCDDs and PCDFs and their derivatives (chiefly hydroxylated surrogates, Poerschmann et al., 2009) are formed under certain circumstances in aqueous solutions containing chlorinated organic compounds (Poerschmann et al., 2009). Hence, this contribution is firstly aimed at clarifying if priority pollutants such as PAHs and PCDDs/PCDFs are formed during hydrothermal treatment of PVC at 180–260 °C. At present, PVC is commonly disposed of in landfills and by incineration. Recycling of polymeric solid waste can be performed by re-extrusion, mechanical, chemical and energetic recovery (Al-Salem et al., 2009). All these methods have specific advantages and disadvantages (Sadat-Shojai and Bakhshandeh, 2011). As an example, combustion/pyrolysis methods suffer from toxic emissions as well as the release of inorganic chlorine causing corrosion in the furnace. Thus, a reliable alternative method to treat PVC waste safely and efficiently in an environmentally sound way is worth pursuing. HTC is expected to meet these requirements. This remediation-related focus gains momentum once combined with the formation of HTC char as a valuable end-of-pipe product. However, its application as a sorbent and for soil enhancement requires the hydrochar to be devoid of organic chlorine. Following this line, special attention was secondly given to the degree of dechlorination of PVC during the HTC process. A third objective of this contribution was studying the sorption potential of hydrochars which is essential considering their potential application as sorbents in the remediation field, e.g. in wastewater treatment. Sorption coefficients of analytes covering a wide range of hydrophobicities and functionalities on HTC chars were determined by the solventless Solid Phase Microextraction (SPME) technique (Kopinke et al., 1999). The basic premise of using the non-depleting SPME technique to study sorption phenomena is that only the freely dissolved analyte fraction is sampled by the fibre. 2. Material and methods 2.1. Chemicals PVC was purchased from Aldrich (Munich/Germany; Cat.No. 18,958-8), average MW 62 000 Da. The plastic was ground in a planetary ball mill to pass a 0.25-mm screen. Analytical grade organic solvents, the silylation reagent bis(trimethylsilyl) trifluoroacetamide (BSTFA; derivatization grade) as well as authentic standards (purity beyond 97%) were purchased from Sigma Aldrich, and Merck (Darmstadt/Germany). 2.2. Elemental analysis Elemental analyses (C, H, and N) of the hydrochar and the native PVC were performed with a Perkin–Elmer PE 2400 Elemental Analyzer (Perkin–Elmer Corp., Norwalk, CT). Chlorine in solids (nontreated PVC, hydrochar) was determined by a home-made combustion approach, which was calibrated prior to the application to PVC. In a nutshell, after a two-step combustion of the sample under a mild air stream at 550 °C and 750 °C in a quartz tube, the combustion gases were flushed from the furnace into an alkaline thiosulfate solution, which is capable of absorbing HCl as well as Cl2 and Cl-oxides. Afterwards, the chloride concentration was determined by ion chromatography (IC; see Section 2.3.). Prior to organic chlorine determination, the hydrochars were washed with

683

deionized water until the chloride concentration in the washing solution proved chloride-free, then washed with acetone. The concentration of dissolved organic carbon (DOC) in the HTC solutions (pH  2.5 to remove inorganic carbon: pH measured with a pH-electrode WTW Typ 3310 SenTix41, Germany) was determined by a TOC (total organic carbon) analyzer (Shimadzu, Duisburg/Germany). The carbon content of the solids was determined by combustion using a carbon analyser (C-mat 5500, Stroehlein Instruments, Duesseldorf/Germany). The ash content of the hydrochar was determined gravimetrically from the residue remaining after combustion according to standard guidelines (DIN, 1978).

2.3. Chloride determination in HTC process water Chloride in HTC water was determined by IC with KOH eluent generator (DX 600, anion exchange column: AS 19 with precolumn AG 19; ThermoFisher, Dreieich/Germany). Likewise, low-molecular weight carboxylic acids were analyzed by IC using a gradient technique with an AS 11 stationary phase.

2.4. HTC-experiments HTC experiments were conducted in closed glass vials with a volume of 22 mL (closed test tube length 200 mm, diameter 12 mm). The operating temperatures were between 180 °C and 260 °C, which represents a range typical of HTC experiments. Near the top of the vial, a hole of 3 mm diameter was drilled into the glass to fill in the PVC sample (25 mg) along with 17 mL deionized water (corresponding to 1470 mg L1 PVC). This concentration is regarded appropriate in case of a potential hydrothermal treatment of organic waste from a recycling bin. The filled glass vial was inserted into a high-pressure steel autoclave as described in Poerschmann et al. (2013), tightly closed and subjected to the operating temperature for 15 h inside a GC oven. Water solubility of macromolecular PVC is considered negligible even at a maximum temperature of 260 °C. Experiments were performed with citric acid (100 lg mL1) as well as without citric acid. HTC experiments were conducted with both non-treated PVC to simulate practically relevant circumstances and with PVC extracted with acetone/benzene (1:1, v/v) to remove low-molecular weight organic compounds. The latter protocol is aimed at analyzing organic breakdown products which are formed exclusively during hydrothermal treatment of the macromolecular PVC. The application of closed vials fitted up with a small hole near the top proved superior compared to open glass vials because it minimized the potential contaminations from the autoclave. Solvent extraction of the aqueous HTC slurry (including hydrochar) was performed twice using benzene. Prior to solvent extraction, isotopically labelled internal standards [2H10]anthracene (anthracene-d10), phenol-d6, hydroquinone-d4, succinic acid-d4, and PCB 13C-labelled EC 1418 (octachlorobiphenyl; Promochem) were spiked into the aqueous phase (1 lg g1, 1 lg g1, 1 lg g1, 10 lg g1 and 0.01 lg g1 referred to PVC-substrate, respectively). The combined extracts were dried over sodium sulphate, rotary evaporated to about 100 lL, and derivatized with BSTFA (silylation at 80 °C for 2 h). The extracts were then subjected to GC/MS analysis. Native PVC was extracted twice by short-term ultrasonic extraction (60 s) using acetone/benzene at room temperature (ultrasonic bath Emmi-12 Hc, ATP, Ettenheim, Germany). Alkaline saponification of the hydrochar was performed by refluxing with 3 M KOH in an inert atmosphere (He) for 90 min. The hydrochar resulting from saponification was neutralized with diluted hydrochloric acid (0.1 M), then dried at 105 °C for 4 h.

684

J. Poerschmann et al. / Chemosphere 119 (2015) 682–689

2.5. Determination of sorption coefficients on HTC chars The basics and fundamentals of the use of the SPME technique to characterize sorption phenomena are highlighted in a recent publication (Poerschmann and Schultze-Nobre, 2014). In the batch approach, conventional liquid immersion SPME using both polar polyacrylate (film thickness 85 lm; Sigmaaldrich) and non-polar poly(dimethylsiloxane) (film thickness 100 lm) coatings was conducted in 40 mL amber screw cap vials completely filled with deionized water adjusted to pH = 6 with potassium dihydrogen orthophosphate buffer, which came close to real-world wastewater treatment conditions. Vials were sealed with Teflon-lined silicon septa. Hydrochar content (as dry matter) was set to be 1000 lg mL1, and 3000 lg mL1; the hydrochar for sorption studies was obtained from HTC at 220 °C. One 40 mL aqueous sample was devoid of the hydrochar sorbent, serving for external calibration. Prior to insertion of the SPME fibres into the aqueous phase, the analytes (see below) were spiked and allowed to equilibrate onto the sorbent overnight under stirring. Sorbate concentrations were 1 lg mL1 for phenols and 10 ng mL1 for PAHs. The extraction time of the analytes by the fibre was 2 h each under intense stirring (500 rpm), which comes close to equilibrium conditions (Kopinke et al., 1999). During this procedure, the freely dissolved analytes were extracted. The extracted analyte fraction was desorbed from the fibre in the hot GC injector for 1 min. Sorption coefficients were calculated on the basis of analyte uptake by the fiber in the HTC char slurry as compared to the uptake in the sorbentfree sample according to Kopinke et al. (1999). The latter uptake reflects the total analyte concentration, whereas uptake from the hydrochar-amended phase corresponds to the free analyte concentration. The reproducibility of absolute peak areas of selected ions (e.g. m/z = 107 amu for xylenols, m/z = 128 amu for naphthalene) from SPME analyses was better than ±13% (relative standard deviation of single values, RSD). Each data point for a given fibre was run in duplicate. Sorption coefficients obtained for a given hydrochar concentration with both fibres were averaged.

2.6. GC/MS analysis A GC-column 30 m  0.25 mm  0.25 lm coated with a nonpolar DB-1 MS stationary phase was used (Agilent, Waldbronn/ Germany). The column was run in the temperature programmed mode: starting at 40 °C, final temperature 300 °C with a linear ramp of 10 °C min1. Pulsed splitless injection mode was used (injector temperature 290 °C). The MSD HP 5973B was operated in full scan mode for data acquisition. Authentic standards for

identification purposes included PAHs (16 PAHs from EPA list), phenols (EPA 8270 phenols mix), and aliphatic aldehydes (nC8– nC12). Further structural identification was performed based on molecular ions and characteristic fragments along with a comparison of Kovacs retention indices listed in NIST- and NBS-libraries. Response factors of the analytes to be quantified were calculated by referring the sum of relative abundances of the three most abundant fragment ions to the sum of relative abundances of the three most abundant fragment ions of the internal isotopically labelled standard. Individual PAHs were calibrated against anthracene-d10. As an example, the authentic standard pyrene (sum of abundances of m/z = 202 amu, m/z = 201 amu, m/z = 101 amu, which are diagnostic for pyrene) was calibrated against the isotopically labelled standard anthracene-d10 (sum of abundances of m/z = 188 amu, m/z = 158 amu, m/z = 80 amu, which are diagnostic for anthracene-d10). Similarly, phenols were calibrated against phenol-d6. As an example, phenol (sum of abundances of m/z = 151 amu, m/z = 166 amu, m/z = 77 amu, which are diagnostic for the trimethylsilylether of phenols) was calibrated against the isotopically labelled standard phenol-d6 (sum of abundances of m/z = 156 amu, m/z = 171 amu, m/z = 82 amu, which are diagnostic for the trimethylsilylether of phenol-d6). Finally, low-molecular mass organic acids were calibrated against the isotopically labelled standard succinic acid-d4 (as trimethylsilyl ether), and diols were calibrated against hydroquinone-d4 (as trimethylsilyl ether). A two-point calibration was applied in each case: aqueous solutions contained 10 lg mL1 or 1 lg mL1 each of authentic standards, the concentration of the isotopically labelled standards was kept constant at 1 lg mL1 each. The standard solutions were extracted with benzene as detailed in Section 2.4., then silylated with BSTFA.

3. Results and discussion 3.1. Mass balance of carbon and chlorine Mass balance data are given in Table 1 for HTC experiments devoid of citric acid. Mass balances were based on the elemental composition of non-treated PVC: 38.6% w/w C, 56.8% w/w Cl, 4.7% w/w H, 0.1% w/w N. Fig. 1 depicts the release of organic chlorine into the aqueous HTC solution as measured by IC. Chlorine can be released from PVC by dehydrochlorination via elimination of HCl or by dechlorination via nucleophilic substitution with water molecules as nucleophiles. As shown in Fig. 1, the hydrothermal treatment of PVC proved to be an efficient approach for the release of organic chlorine. Complete dechlorination of PVC was observed at HTC temperatures above 235 °C. Citric acid addition did not

Table 1 Mass balance for HTC of PVC.

a b c

T (°C)

mChar/mPVC(0) (%, m m1)

mOC, Char/mChar (%, m m1)

180 190 200 210 220 220-sapon. 230 240 240-sapon. 245 250 260

92 85 64 52 45

43.3 46.3 59.1 65.5 72.4

39 35

75.8 (77) 80.7 (73)

33 33 32

84.0 (72) 85.2 (73) 84.5 (70)

(101) (102) (98) (88) (84)

a

mHydrog, Char/mChar (%, m m1)

H/C (mol mol1)

mCl, Char/mChar (%, m m1)

mOxyg, Char/mCharc (%, m m1)

2.1

15 30 54 75 88

4.71 4.83 4.93 4.43 4.51

(92) (87) (68) (59) (44)

1.30 1.25 1.01 0.82 0.75

nn 3.9 5.0 13.0 12.1

2.9

96 105

5.09 (42) 5.25 (39)

0.80 0.75

99 98 98

5.14 (36) 5.30 (37) 5.02 (34)

0.73 0.74 0.71

54 (87) b 45 (68) 31 (35) 19 (17) 6.5 (5.2) 4.1 (3.2) 4.1 (2.8) 1.6 (1.0) 0.75 (0.46) 0.78 (0.46) 0.50 (0.30) 0.43 (0.24)

mDOC, Water/mOC, (%, m m1) 1.4 1.7

4.0 4.2

PVC(0)

mCl, Water/mCl, (%, m m1)

PVC(0)

In parenthesis: recovery of OC, i.e. mOC, Char/mOC, PVC(0) (%, m/m). In parenthesis: recovery of chlorine, i.e. mCl, Char/mCl, PVC(0) (%, m/m). Ash content below 0.6% throughout all chars, thus only C, H, Cl used to calculate oxygen content.

15.0 12.5 10.1 9.1 10.0

J. Poerschmann et al. / Chemosphere 119 (2015) 682–689

CCl, Water/CCl, PVC(0) (%) 100

80

60

40

PVC without citric acid PVC with 100 ppm citric acid

20

0 170

190

210

230

250

T (ºC) Fig. 1. Recovery of chloride in HTC process water versus the temperature of hydrothermal treatment.

affect the release of chloride into the aqueous phase over the entire HTC temperature range. This finding indicates that hydroxyl ions do not play a significant role as reactants under the applied conditions. Complete removal of organic chlorine demonstrates the advantages of using HTC over (dry) pyrolysis, which results in incomplete removal, e.g. 83% chlorine removal at 340 °C as described by Lu et al. (2002). Upon hydrothermal treatment, the pH dropped from circumneutral values to pH = 2.4 and pH = 3.2 at 260 °C and 180 °C, respectively, chiefly due to the release of HCl. However, low-molecular weight organic acids including acetic, formic, and (traces of) propionic and succinic acids also contributed to acidification. The extent of formation of lowmolecular weight organic acids was proportional to the operating temperature: the most abundant acetic acid amounted to 0.10% referred to PVC substrate mass at 180 °C. The concentration of acetic acid increased to 0.43% at 240 °C (corresponding to 4440 lg OCHac g1 OC,PVC). Correspondingly, the concentration of propionic acid increased from 45 lg OCProp g1 OC,PVC at 180 °C to 135 lg OCProp g1 OC,PVC at 240 °C. Visual observations showed that the hydrochar progressively changed from light-brown to brown and finally black with increasing operating temperature, as the colour of the hydrochar likely correlated to the number of double bonds and the incorporation of oxygen into the macromolecular network. As listed in Table 1, the H/C elemental ratios decreased from 180 °C to 220 °C: upon further heating only a minor reduction in the H/C ratios occurred (for a comparison, H/C elemental ratio for non-treated PVC amounts to 1.53). Moreover, the hydrogen content in the hydrochar did not decrease concomitantly with the chlorine content. Removal of organic chlorine due to dehydrochlorination (HCl elimination) decreases mHydrogen,char/mChar ratio, whereas a nucleophilic substitution would result in an increased mHydrogen,char/mChar ratio (associated with an increased mOxygen,char/mChar ratio). The mass balance of the PVC-substrate showed a decline in the hydrochar recovery with increasing operating temperature (from 92% at 180 °C to 32% at 260 °C, see Table 1). Addition of citric acid also proved insignificant with respect to hydrochar recovery (explicit data not shown here). A similar dependence was observed for the OC-balance referred to the mass of OC in the non-treated PVC (OCPVC). At operating temperatures of 180–200 °C, the OCPVC was almost quantitatively converted into the hydrochar. However, the decline of OC recovered was less pronounced (from 100% at 180 °C to 70% at 260 °C, see Table 1). This finding was (i) due to the almost linear increase of the OC-fraction in the char with an increase in the operating temperature (from 43.3% at 180 °C to 84.5% at 260 °C), irrespective of the addition of citric acid, and

685

(ii) due to the loss of the substrate OC via the aqueous HTC phase (4.2% at 250 °C). The loss of OC due to the formation of water–soluble products is of minor significance at low operating temperatures (e.g. 1.4% at 180 °C, see Table 1). The OC-balance was almost complete until operating temperatures of about 200 °C. Headspace analysis, which will be detailed in a forthcoming contribution, provided strong evidence that the formation of carbon dioxide (and carbon monoxide to a lesser extent) was largely responsible for the gap in the OC balance at higher operating temperatures. The mass balance of chlorine (Fig. 1, Table 1) provided strong evidence that the chlorine which might be covalently bound inside the macromolecular HTC char network was of minor (if any) significance at temperatures above 235 °C. The question whether organic chlorine was eliminated quantitatively in the hydrochars produced above 235 °C or present in small concentrations could not unambiguously answered on the basis of the chloride analysis alone (in our experience, IC is accurate for Cl determination within 0.1–0.2%.) This objective is significant, though, because even very low concentrations of (bioavailable) chlorinated organic compounds can result in severe environmental hazards (Guillén et al., 2012). Thus, a complementary combustion method to directly determine organic chlorine in the hydrochars was applied. Data in Table 1 confirm that the removal of chlorine above 235 °C was almost quantitative: residual chlorine in the char obtained upon HTC at 240 °C amounted to 1.6%, and that in char obtained at 260 °C was 0.24% referred to the chlorine in PVC (see Table 1: RSD > 25% at chlorine contents of less than 3%). The results obtained by chlorine determination in the hydrochars corresponded well to the IC-based chloride determination in the process water. The residual organic chlorine in hydrochars may be resistant to hydrothermal treatment (covalently bound to sp3-carbon). Other mechanisms operative for the incorporation of organic chlorine into the hydrochars such as entrapment in narrow pores or adsorption can be considered insignificant due to the solvent extraction of the hydrochars prior to combustion. Alkaline saponification of two hydrochars (resulting from HTC at 220 °C and 240 °C) released a significant fraction of the residual chlorine (see Table 1: from 6.5% to 4.1%, and from 1.6% to 0.75%, respectively). However, a small, recalcitrant interior chlorine fraction remained inside the char network which was not accessible even to the applied harsh saponification reaction. This residual organic chlorine is assumed to be bound to sp2-hybridized carbon atoms in olefinic and aromatic structures, which are known to be recalcitrant against nucleophilic substitution by hydroxyl ions. In contrast to that, alkaline saponification is known to release chlorine bound to sp3-carbon (Bylaska et al., 2010). 3.2. Organic compounds arising from HTC of PVC Table 2 provides qualitative and quantitative information on organic compounds in aqueous HTC solutions. The identification of HTC products revealed a wide array of dechlorinated, chiefly cyclic compounds. Basically, no chlorinated organics exempt traces of dichlorophenols – known as precursors of PCDDs and PCDFs – could be observed. PCDDs and PCDFs were absent (detection limit 0.1 ng mL1). These findings could be confirmed by coupling GCAED (atomic emission detection) using the carbon emission line at 193 nm and the chlorine line at 479 nm (detailed in a forthcoming contribution). Analytes written in italics in Table 2 (first column) constitute monomolecular impurities in the native macromolecular PVC-substrate. Prominent analytes in this respect include trichlorobenzenes, and fluorene-9-one, which could not be formed during the HTC treatment. A consideration of naphthalene and fluorine is difficult in this respect because they can both be formed during the HTC treatment and be present as native impurities.

686

J. Poerschmann et al. / Chemosphere 119 (2015) 682–689

Table 2 Total concentrations of organic compounds in HTC-slurries of PVC (concentration data given in lg g1 PVC, data averaged of three, RSD between 10% and 20%).

a

Trichlorobenzenes Dichlorophenols Lindane Ethyleneglycol Phenol Cyclohexene-1-ol Benzylalcohol R Benzenediols R Naphthols R TeH-Naphthols R Phenanthrenols R TeH-Phenanthrenols Glycolic acidb Lactic acidb Succinic acid Adipic acid Benzoic acid R M-2-Cyclopentenonesc R DM-2-Cyclopentenones R C3-2-Cyclopentenones Benzaldehyde Indene Acetophenone Naphthalenea R DH-Naphthalenes TeH-Naphthalene Indanone Isobenzofuranon R M-Naphthalenes 1-Tetralon Biphenyl R Naphthalene carboxyaldehydes Fluorenea Diphenylmethane Dibenzofuran C12H8O (Naphthofuran?) Phenanthrene Fluoren-9-onea R TeH-phenanthrenes Benzophenone R DH-oxo-phenalenes Pyrene Naphthalene-2-phenyl R TeH-pyrenes Anthracenedione OH-Anthracenedione R C17H12 (Benzofluorenes?) R C18H12 a b c d

MW (Da)

180 °C

200 °C

220 °C

240 °C

250 °C

260 °Cd

180 162 290 62 94 98 108 110 146 150 194 198 76 90 118 146 122 96 110 124 106 116 120 128 130 132 132 134 142 146 154 156 166 168 168 168 178 180 182 182 182 202 204 206 208 216 216 228

1.3 + 4.8 n.d. n.d. 5 2.0 n.d. Fig. 2. Aromatic products of hydrothermal carbonization. Sample: solvent extract of HTC solution (220 °C). Data presentation: sum of selected molecular ions (in parenthesis).

687

significant differences. Pyrolysis patterns of PVC are characterized by high abundances of gaseous HCl, low molecular weight hydrocarbons and cyclic hydrocarbons (styrenes, indenes, PAHs) with the virtual absence of hydrolysis products (Simoneit et al., 2005; Font et al., 2011 and references cited therein). As an estimate, vacuum pyrolysis of PVC (550 °C) yields 15% light liquid oil, 19% heavy liquid oil and a solid residue of 9% (Miranda et al., 2001). 3.3. Proposed pathways for the formation of stable intermediates As described in the literature, thermal dechlorination of PVC mostly follows a free radical mechanism (see Starnes (2012) and references cited therein). The reaction may start with the production of free radicals and chloroallylic structures characterized by low thermal stabilities. A subsequent stepwise HCl elimination gives rise to polyene formation, which is a non-free radical reaction mechanism (Starnes and Ge, 2005). Obviously, conjugated double bonds are created by a ‘‘zipper’’ mechanism: once a double bond has formed, the allylic chlorine atom on the C-atom adjacent to the double bond splits off HCl forming two double bonds in the process, which in turn activates adjacent chlorine to propagate the dehydrochlorination process. In aqueous suspensions, PVC degradation via ionic chain reactions and through cracking is assumed (Nagai et al., 2007). Beyond that, nucleophilic substitution with water as the nucleophile proceeds to generate alcohols, diols, and polyols. Reactive polyenes, as well as diol/polyol structures, are viewed as precursors of aromatic compounds and O-functionalized low molecular weight compounds. As emphasized by Nagai et al. (2007), temperatures below 450 °C and high water densities favour OH-nucleophilic substitution with water acting as a nucleophilic agent. Fig. 3 summarizes the proposed pathways to produce both stable organic intermediates and a condensed hydrochar. 3.4. Sorption of hydrophobic pollutants onto HTC char formed from PVC Fig. 4 shows sorption coefficients referred to the OC of the hydrochar versus the hydrophobicities of the analytes expressed by their octanol–water partition coefficients (KOW). For a comparison, sorption coefficients on a hydrochar originating from HTC of brewer’s spent grain (BT) at 220 °C are given alongside. It should be noted that the sorption behaviour of hydrochars obtained from BT was similar to those obtained by HTC of other feedstocks including digestates, maize silage, beet slices, and model substrates such as sucrose (explicit data not shown here). The labels of data in Fig. 4 were chosen to be different for PAHs (empty boxes) on the one hand and phenols/chlorinated analytes (shorthand designation in Fig. 4: ‘‘phenols’’, filled boxes) on the other hand: Both sorbate groups were chosen because PAHs have been known to exercise predominantly non-specific interactions to amorphous organic sorbents, while phenols are capable of entering into specific interactions (H-donor, dipole–dipole). Data in Fig. 4 and Table 3 reveal a strong correlation for both hydrochars between sorption coefficients and hydrophobicities for the hydrophobic PAHs. In case of the PVC-derived hydrochar, the strong dependence of sorption coefficients on hydrophobicities is represented by the dashed line (see Fig. 4). Sorption data of phenols fit approximately into this arbitrary ‘‘non-polar PAH-line’’. This provides a strong indication that the contribution of polar interactions to the overall sorption is of minor significance. Thus, it is safe to assume that the hydrochar formed by HTC of PVC at 220 °C is not capable of exercising polar interactions towards functionalized sorbates. On the other hand, the low sorption potential of the hydrochar towards hydrophilic organic compounds arising from the hydrothermal process (e.g. hazardous benzenediols)

688

J. Poerschmann et al. / Chemosphere 119 (2015) 682–689

2.5

1-Naphthol

Naphthalene 4-n-Propylphenol

3.0

3,4-Xylenol

3.5

p-Cresol p-Chlorophenol

4.0

2,3,5-Tri-M-phenol

4.5

Phenanthrene

Log KOM

Fluorene

5.0

Tri-Cl-benzene ß-HCH Acenaphth.

Fig. 3. Assumed reaction pathway of PVC decomposition in subcritical water, based on Nagai et al. (2007).

PVC-phenols BT-phenols PVC-PAHs BT-PAHs

2.0 1.5 2.0

Log KOW 2.5

3.0

3.5

4.0

4.5

5.0

Fig. 4. Sorption coefficients of selected organic pollutants on chars obtained by HTC of PVC and brewers’ spent grain versus sorbate hydrophobicity.

seems to be beneficial. Ideally, reversibly sorbed organic compounds should be removed from hydrochars prior to their application for remediation purposes. Given the low sorption coefficients for all sorbates under study (Table 3), it must be concluded that a prospective utilization of PVC-related chars as sorbents in wastewater remediation is of little significance. Overall, sorption coefficients on the PVC-derived hydrochar were lower by an order of magnitude as compared to the brewers’ spent grain hydrochar (see Fig. 4), and several orders of magnitude lower compared to those on activated carbon (Zhang et al., 2010). As for phenanthrene, only 30% of the total concentration was sorbed onto the PVC hydrochar (see Table 3; given a char concentration of 662 lg g1, which results from CPVC,0 = 1470 lg g1 and a hydrochar mass recovery of 45% at 220 °C). On this basis, the freely dissolved phenanthrene concentration in the aqueous phase can be calculated to 3.85 lg g1 PVC-substrate (total concentration at 220 °C: 5.5 lg g1 PVC, see Table 2). In addition to higher sorption coefficients, the Log KOM–Log KOW slopes were different for both PAH-lines (1.15 for brewer’s spent grain versus 0.70 for PVC). This finding provides evidence that the density and strength of centres of forces were significantly higher with the

hydrochar from brewer’s spent grain. The low slope for phenols with hydrochar from brewer’s spent grain (0.75) points to a minor contribution of polar forces to the sorption of OH-functionalized organics. Several years ago, a one-parameter concept based on Hildebrand solubility parameters was developed to calculate hydrophobic partitioning by considering both the affinity of the analyte with its amorphous host and its ‘‘incompatibility’’ with water (Kopinke et al., 1999). This successfully validated one-parameter concept (Poerschmann and Kopinke, 2001) offers the possibility to calculate solubility parameters of amorphous sorbents as a characteristic feature based on sorption coefficients (see Eq. (1)): dChar ¼ di 

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2:3  R  T ðLogK OW  LogK Char Þ  LogqChar : ðdi  dOctanol Þ2 þ Vm ð1Þ

Vm is the molar volume of the analyte studied (mL mol1), di,char the solubility parameter for solute, and char, respectively ((J cm3)0.5), and qChar is the char density to accommodate dimensions KOW (mL g1) and KChar (g g1), estimated to be 0.8 g mL1. R = 8.31 J K1 mol1, T = 293 K, dOctanol = 21.0 (J cm3)0.5. Solubility parameters for the PVC-derived char, calculated from the data given in Table 3 were very similar for all PAHs (averaged dChar = 12.2 (J cm3)0.5). This value was surprisingly low (for a comparison, solubility parameters for the very non-polar polyethylene average to dPE = 16.4 (J cm3)0.5). Hence, it can be concluded on the basis of these calculations that a surface-related adsorption mechanism prevails over hydrophobic partitioning. Typically, char formation under HTC conditions can proceed by (i) a sequence of dissolution and condensation/precipitation steps or by (ii) a solid state carbonization reaction. Based on the sorption data (see Table 3), it can be concluded that the latter pathway dominates. This hypothesis is in line with the very low specific surface area of the hydrochars from PVC (below 1 m2 g1). 4. Conclusions From the environmental perspective, hydrothermal treatment at operating temperatures above 235 °C, which ensures almost complete dechlorination, offers a prospective alternative for treat-

Table 3 Sorption characteristics for the PVC-derived char (220°C).

a b

Sorbate

Vm (mL mol1) Spurlock and Biggar (1994)

Log KOW Poerschmann and Kopinke (2001)

Log KChar (mL g1)

Naphthalene Acenaphthene Fluorene Phenanthrene 1-Naphthol

124.6 142.0 154.8 160.6 123.5

3.30 3.92 4.18 4.44 3.11

1.93 2.30 2.63 2.81 2.10

(5)a (12)a (22)a (30)a (8)a

di ((J cm3)0.5) Poerschmann and Kopinke (2001)

dCharb ((J cm3)0.5)

20.2 20.0 19.9 20.0 24.1

12.25 11.90 12.27 12.33 16.60

In parenthesis: analyte fraction (in %) sorbed onto the hydrochar (e.g. CPhenanthrene,sorbed/CPhenanthrene,total = 30%). Calculated from Eq. (1).

J. Poerschmann et al. / Chemosphere 119 (2015) 682–689

ing solid PVC waste. The formation of chlorinated organic compounds proved negligible. PCDDs and PCDFs could not be detected at all. Based on this view, PVC plastics which may be co-discharged with organic household waste do not pose a significant environmental threat. However, the formation of PAHs should be taken into consideration. Future work should be focused on hydrothermal treatment of real-world PVC plastics mixed with real-world organic waste, and wet oxidation of the HTC process water rich in DOC. The cascade use of the aqueous HTC product stream, e.g. by molecular oxygen or hydrogen peroxide, is expected to destroy PAHs as well as improve the energetic self-sustainability of the HTC process. Further on, HTC of PVC under basic conditions (e.g. brought about by NaOH or aqueous ammonia) with abundant hydroxyl ion concentrations acting as strong nucleophile is expected to allow complete dechlorination at much lower temperatures. References Al-Salem, S.M., Lettieri, P., Baeyens, J., 2009. Recycling and recovery routes of plastic solid waste (PSW): a review. Waste Manag. 29, 2625–2643. Berge, N.D., Ro, K.S., Mao, J., Flora, J.R.V., Chappell, M.A., Bae, S., 2011. Hydrothermal carbonization of municipal waste streams. Environ. Sci. Technol. 45, 5696– 5703. Bylaska, E.J., Glaesemann, K.R., Felmy, E.R., Vasiliu, M., Dixon, D.A., Tratnyek, P.G., 2010. Free energies for degradation reactions of 1,2,3-trichloropropane from ab initio electronic structure theory. J. Phys. Chem. A 114, 12269–12282. Comisar, C.M., Savage, P.E., 2004. Kinetics of crossed aldol condensations in hightemperature water. Green Chem. 6, 227–231. DIN 51719, German Institue of Standardization, 1978. Testing of solid fuelsdetermination of ash content. Endo, K., 2002. Synthesis and structure of poly(vinyl chloride). Progr. Polym. Sci. 27, 2021–2054. Font, R., Gálvez, A., Moltó, J., Fullana, A., Aracil, I., 2010. Formation of polychlorinated compounds in the combustion of PVC with iron nanoparticles. Chemosphere 78, 152–159. Font, R., Moltó, J., Egea, S., Conesa, J.A., 2011. Thermogravimetric kinetic analysis and pollutant evolution during the pyrolysis and combustion of mobile phone case. Chemosphere 85, 516–524. Funke, A., Ziegler, F., 2009. Hydrothermal carbonization of biomass: a summary and discussion of chemical mechansims for process engineering. Biofuels Bioprod. Bioref. 4, 160–177. Guillén, D., Ginebreda, A., Farré, M., Darbra, R.M., Petrovic, M., Gros, M., Barceló, D., 2012. Prioritization of chemicals in the aquatic environment based on risk assessment: analytical, modeling and regulatory perspective. Sci. Total Environ. 440, 236–252. Hall, W.J., Williams, P.T., 2006. Fast pyrolysis of halogenated plastics recovered from waste computers. Energy Fuels 20, 1536–1549. Keiluweit, M., Kleber, M., Sparrow, M.A., Simoneit, B.R.T., Prahl, F.G., 2012. Solventextractable polycyclic aromatic hydrocarbons in biochar: influence of pyrolysis temperature and feedstock. Environ. Sci. Technol. 46, 9333–9341. Kislov, V.V., Sadovnikov, A.I., Mebela, A.M., 2013. Formation mechanism of polycyclic aromatic hydrocarbons beyond the second aromatic ring. J. Phys. Chem. A 117, 4794–4816. Kopinke, F.-D., Poerschmann, J., Georgi, A., 1999. Application of SPME to study sorption phenomena on dissolved humic organic matter. In: Pawliszyn, J. (Ed.),

689

Applications of solid phase microextraction, RSC, Hertfordshire, UK, pp. 111– 128. Kubátová, A., Lagadec, A.J.M., Hawthorne, S.B., 2002. Dechlorination of lindane, dieldrin, tetrachloroethane, trichloroethene, and PVC in subcritical water. Environ. Sci. Technol. 36, 1337–1343. Kuhlmann, B., Arnett, E.M., Siskin, M., 1994. Classical organic reactions in pure superheated water. J. Org. Chem. 59, 3098–3101. Kus, N.S., 2012. Organic reactions in subcritical and supercritical water. Tetrahedron 68, 949–958. Le Moan, G., Chaigneau, M., 1971. Food Cosmet. Toxicol. 9, 450–451. Libra, J.A., Ro, K.S., Kammann, C., Funke, A., Berge, N.C., Neubauer, Y., Titirici, M.-M., Fühner, C., Bens, O., Kern, J., Emmerich, K.H., 2011. Hydrothermal carbonization of biomass residuals: a comparative review of the chemistry, processes and applications of wet and dry pyrolysis. Biofuels 2, 89–124. Lu, J., Ma, S., Gao, J., 2002. Study on the pressurized hydrolysis dechlorination of PVC. Energy Fuels 16, 1251–1255. McKay, G., 2002. Dioxin characterisation, formation and minimisation during municipal solid waste (MSW) incineration. Chem. Engn. J. 86, 343–368. Miranda, R., Pakdel, H., Roy, C., Vasile, C., 2001. Vacuum pyrolysis of commingled plastics containing PVC. II. Product analysis. Polym. Degr. Stabil. 73. 47–47. Nagai, N., Smith, R.L., Inomata, H., Arai, K., 2007. Direct observation of polyvinylchloride degradation in water at temperatures up to 500 °C and at pressures up to 700 MPa. J. Appl. Polym. Sci. 106, 1075–1086. Oliveira, I., Blöhse, D., Ramke, H.G., 2013. Hydrothermal carbonization of agricultural residues. Biores. Technol. 142, 138–146. Poerschmann, J., Kopinke, F.-D., 2001. Sorption of very hydrophobic organic compounds (VHOCs) on dissolved humic organic matter (DOM). 2. Measurement of sorption and application of a Flory–Huggins concept to interpret the data. Environ. Sci. Technol. 35, 1142–1148. Poerschmann, J., Schultze-Nobre, L., 2014. Sorption determination of phenols and polycyclic aromatic hydrocarbons in a multiphase constructed wetland system by solid phase microextraction. Sci. Total Environ. 482–483, 234–240. Poerschmann, J., Trommler, U., Gorecki, T., Kopinke, F.-D., 2009. Formation of chlorinated biphenyls, diphenyl ethers and benzofuranes as a result of Fentondriven oxidation of 2-chlorophenol. Chemosphere 75, 772–780. Poerschmann, J., Weiner, B., Baskyr, I., 2013. Organic compounds in olive mill wastewater and in solutions resulting from hydrothermal carbonization of the wastewater. Chemosphere 92, 1472–1482. Sadat-Shojai, M., Bakhshandeh, G.-R., 2011. Recycling of PVC wastes. Polym. Degr. Stabil. 96, 404–415. Simoneit, B.R.T., Medeiros, P.M., Didyk, P.M., 2005. Combustion products of plastics as indicators for refuse burning in the atmosphere. Environ. Sci. Technol. 39, 6961–6970. Spurlock, F.C., Biggar, J.W., 1994. Thermodynamics of organic chemical partition in soils. Environ. Sci. Technol. 28, 1003–1009. Starnes, W.H., 2012. How and to what extent are free radicals involved in the nonoxidative thermal dehydrochlorination of poly(vinyl chloride)? J. Vinyl Addit. Technol. 18, 71–75. Starnes, W.H., Ge, X., 2005. Mechanism of autocatalysis in the thermal dehydrochlorination of poly(vinyl chloride). Macromolecules 37, 352–359. Sullivan, R.F., Boduszynski, M.M., Fetzer, J.C., 1989. Molecular transformations in hydrotreating and hydrocracking. Energy Fuels 3, 603–612. Takeshita, Y., Katob, K., Takahashib, K., Satoc, Y., Nishi, S., 2004. Basic study on treatment of waste polyvinyl chloride plastics by hydrothermal decomposition in subcritical and supercritical regions. J. Supercrit. Fluids 31, 185–193. Zhang, S., Shao, T., Kose, H.S., Karanfil, T., 2010. Adsorption of aromatic compounds by carbonaceous adsorbents: a comparative study on granular activated carbon, activated carbon fiber, and carbon nanotubes. Environ. Sci. Technol. 44, 6377– 6383.