PAH in Forest Humus Types

Research Articles

Polycyclic Aromatic Hydrocarbons (PAH) in Different Forest Humus Types 1Martin Pichler, 2Georg Guggenberger, 3Robert Hartmann, 2Wolfgang Zech 1 Lehrstuhl f/.ir Bodenkunde, Technische Universit/~t Mfinchen, D-85350 Freising-Weihenstephan, Germany 2 Institute of Soil Science and Soil Geography, University of Bayreuth, D-95440 Bayreuth, Germany 3 Institut Dr. Neumayr, Hiigelstraf~e 3-5, D-74564 Crailsheim, Germany

Corresponding author: Dipl.-Geo6kol. Martin Pichler, e-mail: [email protected]

Abstract The behavior of 20 PAH in the organic layers of a L mull, an Of mull and a mor was assessed by a combined approach of a soil profile study, and the analysis of particle-size separates. Increasing PAH concentrations with depth in the mor profile (L, 866 lag kg ~; Of, 2902 lag kg-l; Oh, 10489 lag kg-l) were assigned to selective enrichment during organic matter decomposition. PAH were further highly enriched within the finer separates. For the L horizons, significant positive correlations were established between the enrichment of individual PAH (as observed from the decomposition gradient between the >2-mm fraction and the Table 1). Organic surface layers were defined as F-Mull (F mull; Gels), L-Mull (L mull; All) and Rohhumus (mor; HoM) according to the German classification (AG Boden, 1994). For comparison, data from four other sites (Tegernsee, Teg; Forchheim, Forch; Schweinsbacher Sattel, Schwe; Farrenleite, Farr) obtained from HAP,TMANN (1993) was included. Teg is located in the Mangfallgebirge (northern Alps), Forch represents a site near Nuremberg, and Schwe and Farr are situated in the Fichtelgebirge. PAH deposition at these sites was low (Teg), moderate (Forch), high (Schwe), and very high (Farr). Humus type was moder at Forch, and mor at Teg, Schwe and Farr. Table 2 gives a compilation of important parameters for the investigated horizons. Table 2:

Important features of the investigated forest floor horizons Gels L"

thickness

AllL

HoML~HoMOf

HoMOh

[cm]

3

3

1.5

5

4

bulk density [g cm -3]

0.02

0.02

0.06

0.09

0.17

4.5

4.8

9.2

43.8

66.2

pH (H20)

6.2

6.3

4.1

4.0

3.6

pH (CaCI2)

5.4

5.7

3.3

3.1

2.7

~'orgr" [%, weight]

49.3

47.7

51.7

47.8

44.3

S/N

33.7

23.9

28.1

27.0

27.6

stock

[t ha-1]

a includes Geisberg Of, which was too thin to be sampled separately (see 2.2). Abbreviations are: Geis: Geisberg; All: Allersdorf; HeM: Hohe Matzen

ESPR - Environ. Sci. & Pollut. Res. 3 (1) 1996

2.2

Typic Haplorthod Cambi - Carbic Podzol Spruce forest Picea abies (L.) KARST.

mor

Soil Sampling, Simple Analyses and Calculation of Stocks

Forest floor horizons were taken as mixed samples in December 1993 (All, HoM) and in September 1994 (Gels), with snail-shells, twigs and roots already being removed in the field. Analyses of C t and Nt of 105 ~ and ground samples were carried out on an Elementar Vario EL - CHNS Analyzer. Since a check for carbonate was negative, C t represented C org. Relative standard errors were 0.6 % for C org and0.8 % f o r N v For determination of pH values, material from the forest floor horizons was suspended with deionized water or a 0.01 M CaC12 solution (weight ratio 1:10). Bulk density and stocks of organic matter in the organic layers of the H o M forest floor were determined from volume samples (4 replicates) taken with a cut-off frame (16x8x6 cm). Relative standard errors for the individual horizons ranged from 5 % to 9 %. In order to obtain the bulk density and stocks of the L horizons at Gels and All, whole layers, excluding snail-shells, twigs and roots, were sampled from a defined area (40x40 or 60x60 cm, respectively) in 5 replicates. After drying of the material at 105 ~ stocks could be determined directly. Relative standard errors were 1 1 % for Gels and 5 % for All. The Of horizon of the Gels forest floor was very thin and indistinct, thus no proper sampling could be carried out. This material was sampled together with the L horizon, and all results concerning the Gels L horizon refer to Gels L and Of.

25

PAH in Forest Humus Types 2.3

Particle-Size Separation

Organic matter was dispersed following the method of DINEL et al. (1976) by reciprocal shaking of air-dried material with deionized water (10 g air-dried sample and 500 ml H2Odeion) in a 1-1 glass bottle with a teflon seal for 16 h. Dispersion by sonication should be avoided due to destruction of primary organic particles (GuGGENBERGER, 1993, unpublished). Only the HoM Oh horizon was used field moist. Preliminary studies showed that air-drying resulted in formation of aggregates about 1 to 2 cm in diameter, which were not dispersed even after 16 h of shaking. All data was determined on a 105 ~ base. Particle-size separation was conducted on a Retsch Retac3D wet-sieving machine. The apparatus was equipped with a set of sieves (Retsch, sieve diameter 200 ram) exhibiting mesh widths of 2, 0.5, 0.2 and 0.05 mm. Below the 0.05mm sieve there was a 10-1 glass bottle. For separation, dispersed organic matter samples were put on the coarsest sieve. Sieving was performed by the combined action of riddling and sprinkling H2Odeio n (1.8 1 min -1 at an angle of 120 ~ onto the sample. After 1 min, the 2-mm sieve was removed and all other sieves were sprinkled one after another. Subsequently, the organic matter remaining on the sieves could be freeze-dried directly, whilst the smallest fraction (< 0.05 mm) was found in the outflow of the sieves, and was collected in the glass bottle. The suspension was coagulated by addition of CaCI2 and subsequently centrifuged at 3600 g. The supernatant was discarded whilst the residue was freeze-dried. Recovery of the sum of separates was: Geis L horizon, 99.5 %; All L horizon, 94.9 %; HoM L horizon, 98.1%; HoM Of horizon, 95.1%; and HoM Oh horizon, 88.5 %. Losses were partly assigned to dissolved organic carbon (DOC), which was discarded after centrifugation of the smallest separates. The highest loss (All L) amounted to 1.7 % of total Cotr Further, there was some loss of particulate organic matter. Most of the loss occurring during separation of the H o M Oh horizon was probably an artifact, and could be attributed to the use of field moist samples. The high variability of the gravimetric water content (322-376%) of this sample could have resulted in a wrong estimation of the test portion (10 g on a 105 ~ base), and hence in a miscalculation of the recovery. 2.4

PAH Analysis

Within this study, the 16 EPA PAH (KEITHet al., 1979) plus benzo(e)pyrene, perylene, benzo(j)fluoranthene and triphenylene were investigated. Since benzo(b)fluoranthene, benzo(j)fluoranthene, and benzo(k)fluoranthene could not be separated completely, they were bulked together and abbreviated as Bbjk. Triphenylene (T) and chrysene (C) were coeluting at gas chromatography and also evaluated together, abbreviated as C+T. Other abbreviations used are naphthalene (Np), acenaphthylene (Ay), acenaphthene (Ace), fluorene (FI), phenanthrene (Phe), anthracene (Ant), fluoranthene (Fla), pyrene (Pyr), benzo(a)anthracene (BaA), benzo(e)pyrene (BeP), benzo(a)pyrene (BaP), pery-

26

Research Articles lene (Per), indeno(1,2,3-cd)pyrene (IdP), dibenzo(a,h)anthracene (DbA), benzo(ghi)perylene (Bghi). Analysis of PAH was conducted according to HARTMANN (1995), with minor modifications. Only organic solventcleaned and heated (300 ~ glass and high-grade steel vessels were used for sample preparation. The method is based on extraction of PAH by saponification of soil organic matter with methanolic-aqueous KOH in an ultrasonic bath. Purification of the extracts was carried out by solid-phase extraction with SiO 2 and AI203. A Hewlett Packard 5890 II GC equipped with a 5971A mass selective detector was used for PAH determination. Eight deuterated PAH (NpD 8, Ace-D10, FI-D10, Ant-D10, Pyr-D10, C-D12, Per-D12, Bghi-Dxz) were used as internal standards for quantification. Most analyses were carried out in duplicate by investigating two subsamples, and all results were presented on a 105 ~ base. Detection limits were determined as the substance amount that yields a peak intensity three times greater than the background noise. For the external standards, this amount was ascertained to 2-3 pg. Recovery of low-molecular PAH was increased by collecting the total effluent of the silica-aluminum oxide columns. By using a deuterated injection standard (Fla-Dl0), recoveries of internal standards were quantified as follows (means • standard errors, n=10): Np-Ds, 76• %; Ace-D10, 82• %; FI-D10, 82• %; Ant-D10, 94• %; Pyr-Di0 , 88• % ; C-D12 , 8 8 + 3 % Per-D12, 90• %; Bghi-D12 , 89• %. Reproduceability of the analyses was calculated as the geometrical mean (SACHS, 1984) of the relative standard errors of each duplicate analysis, performed for 23 samples and 17 PAH or PAH groups (n=391). Mean error was 1.8 % (n=368; 23 analyses below quantification limit), ranging from 0.6 % for C+T to 4 % for Ay. Since weighted sums of PAH concentrations in the size separates generally agreed well with the values for the unfractionated samples (L horizons) or, as discussed later, were even greater (HoM Of and Oh), PAH losses due to solubilization in water and due to freeze-drying were considered to be insignificant. 3 3.1

Results and Discussion PAH Concentrations and Stocks in Forest Floor Horizons

Table 3 shows that the PAH concentrations (E 20 PAH) in the L horizons increased in the order Geis