Direct Analysis of Lignin Phenols in Freshwater Dissolved Organic

Nov 15, 2017 - A novel approach for the analysis of dissolved lignin in freshwaters is presented. Lignin concentrations in natural water samples are l...
0 downloads 10 Views 2MB Size
Download PDF
Article Cite This: Anal. Chem. 2017, 89, 13449−13457

pubs.acs.org/ac

Direct Analysis of Lignin Phenols in Freshwater Dissolved Organic Matter Hendrik Reuter,*,†,@ Julia Gensel,†,‡,@ Marcus Elvert,§ and Dominik Zak†,∥,⊥ †

Department of Chemical Analytics and Biogeochemistry, Leibniz-Institute of Freshwater Ecology and Inland Fisheries, D-12587 Berlin, Germany ‡ Department of Chemistry, Humboldt-Universität zu Berlin, D-12489 Berlin, Germany § MARUM Center for Marine Environmental Sciences & Department of Geosciences, University of Bremen, D-28359 Bremen, Germany ∥ Department of Bioscience, University of Aarhus, DK-8600 Silkeborg, Denmark ⊥ Institute of Landscape Ecology and Site Evaluation, University of Rostock, D-18059 Rostock, Germany S Supporting Information *

ABSTRACT: A novel approach for the analysis of dissolved lignin in freshwaters is presented. Lignin concentrations in natural water samples are low, and a lignin extraction is usually required to obtain sufficient material for analysis. In this study, the alkaline CuO oxidation, which liberates a set of ligninderived phenols, is performed directly on 15 mL of water sample in a microwave digestion system, hence reducing the required sample amount and preparation time considerably. These features make the method particularly suitable to study diagenetic changes of dissolved lignin in small-scale laboratory or field experiments. Phenol separation and quantification by gas chromatography tandem mass spectrometry lead to method detection limits between 22.7 and 1260 ng/L for single phenols, which corresponds to minimum lignin concentrations in the range of 8.5 μg/L (Σ8), offering applications for wetland, river, and lake waters with high terrestrial dissolved organic matter inputs. As a general method improvement, we present the addition of EDTA during phenol workup. EDTA binds remaining copper, thereby speeding up sample flow through the solid phase sorbent during phenol extraction and, furthermore, prevents substantial phenol losses, which occur if a water sample contains nitrate. Three natural water samples, a fresh leaf leachate and two humic-rich lake waters, were analyzed by the direct method presented here and in comparison with the established C18 extraction approach. Results show a similar reproducibility of both methods but reveal lower absolute lignin phenol yields in the humic-rich lake water samples upon C18 extraction.

L

total DOM with yields of 0.001−0.01 mg/100 mg DOC in the North Pacific and North Atlantic ocean surface waters.12,13 Compared to the analysis of lignin in soils or sediments, the analysis of dissolved lignin in water samples requires higher effort during sample preparation. DOM concentrations in aquatic systems are low (about 1−100 mg/L DOC), and a concentration routine is usually needed to extract enough organic substrate for CuO oxidation.14 Today, the most widely used DOM concentration method is solid-phase extraction (SPE) on C18 functionalized silica sorbents by which lignin phenols from 0.5 to 50 L water sample are quantitatively recovered.15,16 Other concentration methods include SPE on XAD resins, reverse osmosis, or the direct dry down by freezedrying or rotary evaporation.17,18 In this study, we present a novel approach for the analysis of dissolved lignin that directly uses the aqueous water sample for

ignin is a structural component of vascular plants and, after cellulose, the second most abundant biopolymer on earth. In senescent plant litter it accounts for approximately 20% dry mass. Owing to its decay-resistant structure, intact, or moderately altered lignin macromolecules constitute a part of the organic matter in soils, sediments, and dissolved organic matter (DOM).1,2 One technique to analyze lignin in such samples is the alkaline CuO oxidation which releases a set of lignin-derived phenolic monomers that are subsequently separated and quantified mostly using gas chromatography/ mass spectrometry.3,4 For DOM, the quantification of lignin phenols is a powerful approach to estimate the degree of terrestrial organic matter inputs to an aquatic system.5−9 Typical lignin yields in freshwater DOM range from 0.24 to 3.12 mg/100 mg dissolved organic carbon (DOC).10,11 The upper range of dissolved lignin is usually encountered in wetland waters where organic matter inputs from the vegetation are high.5,6,9 In contrast, dissolved lignin in the oceans constitutes a smaller part of the © 2017 American Chemical Society

Received: September 12, 2017 Accepted: November 15, 2017 Published: November 15, 2017 13449

DOI: 10.1021/acs.analchem.7b03729 Anal. Chem. 2017, 89, 13449−13457

Article

Analytical Chemistry CuO oxidation, omitting the initial DOM concentration step. We combined several method improvements reported over the last years which lead to low blank values and significantly improved detection limits. While still limited to water samples with comparatively high dissolved lignin content, this approach considerably reduces the sample preparation time and the amount of water sample required. This makes the method particularly suitable for laboratory decomposition experiments. Applications as well as limitations of this approach are presented. Furthermore, we investigated the impact of DOC concentration on the lignin yield and report negative effects on lignin yield, if the water sample contains dissolved nitrate. The addition of EDTA to the reaction mixture is presented as an approach to counteract nitrate effects and as a general improvement of the phenol extraction procedure.

Prepacked C18 cartridges (Mega-Bond Elut-C18, 10 g, 60 mL, Agilent Technologies) were pretreated with 100 mL methanol followed by 50 mL acidified water (pH 2). 500 mL Water sample, acidified to pH 2, was passed through the C18 cartridges under slightly elevated head-pressure. Loaded C18 cartridges were rinsed with 500 mL acidified water to remove salts and dried by passing a mild stream of argon through the sorbent for 10 min. Finally, the DOM was eluted from the cartridge into a 100 mL volumetric flask in one fraction of 90 mL methanol and made up to volume. An aliquot (10 mL) of the methanolic DOM extract was used for determination of the DOC recovery. A second aliquot (3 mL), corresponding to DOM in 15 mL of the initial water sample, was directly transferred to the PTFE reaction vessel, and the methanol was removed at 45 °C under a mild stream of nitrogen. Fifteen milliliters of water (sparged with argon for 60 min) was added to the reaction vessel in addition to the later described reagents for the CuO oxidation. Lignin Oxidation. The lignin oxidation procedure was adapted after Goñi and Montgomery23 using a microwave digestion system (Microprep A, MLS GmbH, Germany) equipped with 10 PTFE reaction vessels (100 mL volume). 500 mg CuO, 150 mg (NH4)2Fe(SO4)2·6H2O, and 10 mg glucose were added to each reaction vessel. The aqueous DOM sample was sparged with argon for 60 min to remove dissolved oxygen, and 15 mL sample was transferred to each vessel. Finally, 50 μL internal standard solution (ethylvanilline and cinnamic acid, c = 80 μg/mL in pyridine), and 1.76 mL NaOH (50%) were added and all reaction vessels were transferred into a 520 L glovebag (Atmosbag, Sigma-Aldrich) which was thoroughly flushed with argon for 5 min. After sealing and intense shaking, the vessels were placed into the microwave system and heated to 150 °C within 10 min with a temperature hold time of 90 min. Lignin Phenol Extraction. After lignin oxidation, the content of each PTFE vessel was transferred to a 50 mL glass centrifuge tube (custom-made), centrifuged at 750g for 5 min, and the supernatant was decanted. The step was repeated after rinsing the PTFE vessel with 5 mL 2 M NaOH. In order to remove remaining traces of CuO, the combined supernatants were centrifuged once more and transferred to a final 50 mL glass tube. Twenty milligrams of ethylenediaminetetraacetic acid (EDTA) were added to the cooled reaction mixture (ice bath) before slowly adding 5.50 mL HCl (25%), thereby avoiding any heating up of the solution. The pH of each solution was tested before adding further HCl, if required (pH > 2). Lignin phenol extraction was adapted from Kaiser and Benner4 using Oasis HLB extraction cartridges (60 mg, 3 mL, Waters) placed on a 12-port extraction manifold (J.T. Baker). The HLB cartridges were conditioned twice with 2 mL methanol and twice with 2 mL acidified water (pH 2). Acidified samples (pH 2) were passed through the HLB cartridges under gravity flow and the glass centrifuge tubes were rinsed with 0.5 mL acidified water. Salts were removed from the HLB cartridges with two rinses of 2 mL acidified water. Residual water was removed from the HLB cartridges by centrifugation at 7000g for 5 min. At this stage, the HLB cartridges were frozen at −20 °C until further use. For elution, Teflon needle liners (disposable flow control liners for Visiprep DL, Supelco) were installed into the extraction manifold to avoid possible contamination. Prepacked anhydrous sodium sulfate drying cartridges (1 g, Agilent) were

■

EXPERIMENTAL SECTION Chemicals. All used reagents and solvents were of analytical grade, HPLC grade or LC−MS grade and were, unless otherwise stated, obtained from Merck KGaA (Darmstadt, Germany), Sigma-Aldrich Co. (St. Louis, MO) or VWR International GmbH (Darmstadt, Germany). Pyridine was stored over KOH and freshly distilled every day. Every acidification procedure mentioned was performed using 25% HCl. To avoid contamination, all glassware was heated to 450 °C for 4 h before use. Water was obtained from an arium pro Ultrapure Water System from Sartorius (Göttingen, Germany). Leaf Extracts and Natural Water Samples. Leaf leachates were prepared from senescent Phragmites australis leaves collected in autumn 2014 from the rewetted fen Stangenhagen, southwest of Berlin, Germany.19 We only collected brown P. australis leaves that were still connected to the plant. Leaves were stored under dry and dark conditions until further usage. 5.25 g Dry leaves were leached for 24 h in 3.5 L 3.5 mM NaCl solution at room temperature. A second leachate was prepared using 15 g P. australis leaves from the kettle-hole mire Kablow-Ziegelei, situated south of Berlin, under the same leaching procedure. That leachate was sequentially diluted before CuO oxidation to study the effect of DOC concentration on lignin yield. Natural water samples were taken from the experimental brown-water Lake Grosse Fuchskuhle (northeastern Germany) on March 21, 2017. Since 1989, the lake has been artificially divided into four basins with distinct catchment areas and thus contrasting DOM characteristics.20 In the present study, we sampled two basins with different hydrochemistry: the southwest basin that receives high influxes of humic-rich organic matter from an adjacent Sphagnum mire, and the northeast basin that receives no water from the mire.21,22 Lake water samples and leachates were passed through 0.2 μm PTFE membrane filters (Omnipore, Millipore) and stored in the dark at 4 °C until analysis within 48 h. Subsamples were analyzed for DOC concentration on a Shimadzu TOC analyzer. Comparative DOM Extraction. In order to evaluate the performance of the direct analysis of dissolved lignin to an established analytical routine, we analyzed three natural water samples directly and after isolation of the dissolved lignin using C18 cartridges after Louchouarn et al.16 Briefly, we extracted lignin from 500 mL water samples and used an aliquot of the methanolic lignin extract which corresponded to the amount of dissolved lignin used in the direct approach. This volume precludes sample size effects which have been reported to affect intrinsic lignin parameters in former studies.4 13450

DOI: 10.1021/acs.analchem.7b03729 Anal. Chem. 2017, 89, 13449−13457

Article

Analytical Chemistry

Table 1. Retention Times, Collision Energies, and m/z-values with Proposed Fragmentation Patterns of the Silylated Lignin Phenols RTa

precursor ion

product ion

CE

(min)

(m/z)

(m/z)

(eV)

CiAD

19.45

EVAL

20.07

PAL

17.30

PON

18.55

PAD

20.37

VAL

19.35

VON

20.35

VAD

21.89

SAL

21.27

SON

22.01

SAD

23.27

DiOHBA

22.39

CAD

23.78

FAD

25.22

205 [M−CH3]+ 205b [M−CH3]+ 161 [M−CH3−CO2]+ 195 [M−CH3−C2H4]+ 195 [M−CH3−C2H4]+ 167b [M−CH3−C2H4−CO]+ 179b [M−CH3]+ 151 [M−CH3−CO]+ 151 [M−CH3−CO]+ 208 [M]+· 208 [M]+· 193b [M−CH3]+ 282 [M]+· 267 [M−CH3]+ 267b [M−CH3]+ 209 [M−CH3]+ 209 [M−CH3]+ 194b [M−C2H6]+· 223 [M−CH3]+ 223b [M−CH3]+ 193 [M−CH3−C2H6]+ 312 [M]+· 297 [M−CH3]+ 297b [M−CH3]+ 254 [M]+· 254b [M]+· 224 [M−C2H6]+· 268 [M]+· 253b [M−CH3]+ 238 [M−C2H6]+· 342 [M]+· 327b [M−CH3]+ 327 [M−CH3]+ 370b [M]+· 355 [M−CH3]+ 355 [M−CH3]+ 293 [M−CH3]+ 293b [M−CH3]+ 219 [M−C3H9SiO]+ 323 [M−CH3]+ 323b [M−CH3]+ 308 [M−C2H6]+·

161 [M−CH3−CO2]+ 131b [M−CH3−C2H6SiO]+ 145 [M−CH3−CO2−CH4]+ 179 [M−CH3−C2H4−CH4]+ 167 [M−CH3−C2H4−CO]+ 151b [M−CH3−C2H4−CO−CH4]+ 151b [M−CH3−CO]+ 95 [M−CH3−CO−C2H4Si]+ 75 [M−CH3−CO−C6H4]+ 193 [M−CH3]+ 73 [M−C8H7O2]+ 73b [M−C8H7O2]+ 267 [M−CH3]+ 223 [M−CH3−CO2]+ 193b [M−CH3−C2H6SiO]+ 193 [M−CH3−CH4]+ 165 [M−CH3−C3H8]+ 137b [M−C2H6−C2H5Si]+· 208 [M−CH3−CH3]+· 193b [M−CH3−C2H6]+ 137 [M−CH3−C2H6−CH3−CHO]+ 297 [M−CH3]+ 282 [M−CH3−CH3]+· 267b [M−CH3−C2H6]+ 239 [M−CH3]+ 224b [M−C2H6]+· 195 [M−C2H6−CHO]+ 238 [M−C2H6]+· 238b [M−CH3−CH3]+· 195 [M−C2H6−C2H3O]+ 327 [M−CH3]+ 253b [M−CH3−C2H6SiO]+ 223 [M−CH3−C3H9SiO−CH3]+ 281b [M−C3H9SiO]+ 311 [M−CH3−CO2]+ 281 [M−CH3−C2H6SiO]+ 249 [M-CH3−CO2]+ 219b [M−CH3−C2H6SiO]+ 191 [M−C3H9SiO−CO]+ 293 [M−CH3−C2H6]+ 249b [M−CH3−C2H6SiO]+ 293 [M−C2H6−CH3]+

5 17 5 8 8 11 8 8 17 8 29 20 8 11 20 20 29 29 5 17 23 8 5 14 5 17 20 14 5 40 8 23 32 23 11 17 8 17 11 14 17 5

compound

a

Retention time on DB-5ms Ultra Inert column (60 m, 0.25 mm ID, 0.25 μm film thickness; Agilent Technologies). bQuantification transitions.

Lignin Phenol Quantification. Lignin phenol quantification was carried out using a 7000C Triple Quadrupole GC/MS system from Agilent (Palo Alto, CA) equipped with a 7890B GC oven, a multimode inlet, and an automated liquid sampler (ALS). Prior to analysis, lignin phenols were derivatized by transferring a 5 μL sample into a 2 mL crimp top vial equipped with a 250 μL vial insert. 50 μL N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA) with 1% trimethylchlorosilane (TMCS) was added and the reagents were mixed by pipetting up and down about 5 times. The vials were closed with crimp caps, stored at 75 °C for 20 min to complete the derivatization reaction, and transferred to the ALS. One microliter sample was injected under splitless mode with an inlet temperature of 300 °C. The septum purge flow was 3 mL/min. The purge flow to

installed on the disposable liner, followed by the HLB cartridge. Lignin phenols were eluted from the HLB cartridge with three rinses of 1 mL dichloromethane/methyl acetate/pyridine (70/ 25/5, v/v/v%). Residual solvent was removed from the drying cartridge by applying vacuum to the extraction manifold. The sodium sulfate drying cartridge was removed, and the HLB cartridge was installed directly on the disposable flow liner. Phenol elution from the HLB cartridge was completed with two rinses of 0.5 mL dry methanol. The solvents in the combined eluates were evaporated under a mild stream of nitrogen at room temperature. Samples were redissolved in 150 μL dry pyridine, transferred into a 1.5 mL sample vial, and stored frozen at −20 °C until further use. 13451

DOI: 10.1021/acs.analchem.7b03729 Anal. Chem. 2017, 89, 13449−13457

Article

Analytical Chemistry

Figure 1. Tandem mass chromatogram of the Lake Grosse Fuchskuhle (southeast) water sample in MRM mode (black) and in full scan mode (gray, multiplied with 0.2). The inserted picture highlights the coelution of VON and PAD and the three MRM transitions recorded for each analyte.

a calibration mixture and 50 μL of the internal standard mixture before CuO oxidation. Concentrations were determined as relative response factors to the internal standards EVAL (for VAL, VON, and SAL) or CiAD (remaining phenols). Using a 1:2 dilution pattern, 7 calibration points ranging in concentration from 70 to 25.000 pg/μL per phenol were determined to generate quadratic calibration curves based on 5 selected data points and dependend on the target analyte concentration in the specific data set.

split vent was set to 100 mL/min after 3 min. 870 μL Ultra inert inlet liners with glass wool were used. Separation was achieved on a DB-5ms Ultra Inert capillary column (60 m length, 0.25 mm inner diameter, 0.25 μm film thickness) at constant flow mode (1.5 mL/min, helium). The start temperature of the GC oven was 50 °C, held for 3 min, followed by a temperature ramp of 10 °C/min, a final temperature of 300 °C, and a hold time of 5 min. The transfer line temperature was held at 250 °C. On the MS side, we used the 7000C electron ionization (EI) ion source heated at 230 °C and operated at 70 eV, quadrupoles held at 150 °C, a nitrogen collision flow of 1.5 mL/min, and a helium quench flow of 2.25 mL/min. Measurements were carried out in multiple reaction monitoring (MRM) mode following Louchouarn et al.15 with modifications. The chromatographic run duration was subdivided into 12 time segments corresponding to the retention times of the different lignin phenols: 16.5−18.0 min (phydroxybenzaldehyde, PAL), 18.0−19.1 min (p-hydroxyacetophenone, PON), 19.1−19.9 min (vanillin, VAL and cinammic acid, CiAD), 19.9−20.25 min (ethyl vanillin, EVAL), 20.25− 20.95 min (p-hydroxybenzaldehyde, PAD and acetovanillon, VON), 20.95−21.75 min (syringaldehyde, SAL), 21.75−22.2 min (vanillic acid, VAD and acetosyringone, SON), 22.2−22.9 min (3,5-dihydroxybenzoic acid, DiOHBA), 22.9−23.55 min (syringic acid, SAD), 23.55−24.5 min (p-coumaric acid, CAD), 24.5 min-end (ferulic acid, FAD). One quantification and two qualification transitions were monitored for each target compound (Table 1). The dwell times were set to 50 μs per transition for time segments with three monitored transitions and 40 μs for segments with six transitions. Lignin phenols were quantified using the Agilent MassHunter Workstation Software (Quantitative Analysis, Version B.07.01 for QQQ). Calibration points were determined for each lignin phenol by spiking blanks with increasing amounts of

■

RESULTS AND DISCUSSION Method Development Strategy. In previously reported methods, the analysis of dissolved lignin in natural water samples requires a concentration of the DOM.16 The obtained dry DOM extract is redissolved in 2 M NaOH which serves as a solvent in the subsequent lignin oxidation reaction. The critical method modification presented here omits the DOM concentration. 1.76 mL 50% NaOH is directly added to 15 mL of natural water sample to obtain the lignin-containing 2 M NaOH for the oxidation reaction. Advantages of this approach include the use of less water sample and the avoidance of an additional sample preparation step like C18 SPE or direct dry down of the water sample. Our approach leads to considerably smaller analyte concentrations compared to previously described methods. This was addressed with the use of a microwave digestion system which allows higher sample volumes than the often used reaction minibombs. Lignin phenols were extracted using polymer-based extraction cartridges4 to obtain lower procedural blank values compared to the liquid−liquid extraction technique utilizing ethyl acetate. Phenol quantification was performed by tandem mass spectrometry in MRM mode which leads to higher selectivity and sensitivity compared to single quadrupole instruments. 13452

DOI: 10.1021/acs.analchem.7b03729 Anal. Chem. 2017, 89, 13449−13457

Article

Analytical Chemistry Chromatography, MRM Transitions, and Detection Limits. A tandem GC/MS chromatogram of the Lake Grosse Fuchskuhle water sample is presented in Figure 1. No chromatographic separation was achieved for VON and PAD, which was not required for quantification due to the high selectivity of the MRM transitions (Figure 1, inserted picture). An overview of the MRM method settings of the silylated lignin phenols is presented in Table 1. Tandem mass spectrometry is based upon the selective filtering of a specific precursor ion in the first quadrupole followed by the collision-activated dissociation of the precursor in the second quadrupole and the quantification of a specific product ion in the third quadrupole. Such effective ion filtering results in very low signal-to-noise values and a sensitive and selective quantification of the target analytes. Three transitions were measured for each analyte and consistent peak ratios of the quantifier and qualifier signals were constantly monitored by the quantification software. Only PAL and PON required manual inspection of a correct peak integration in a few natural samples with low lignin concentrations. These analytes provided a comparably poor fragmentation pattern due to the low number of functional groups present, thus more unspecific product ions had to be chosen [m/z = 75, (CH3)2SiOH+ and m/z = 73, (CH3)3Si+]. Instrument detection limits (IDL) and method detection limits (MDL) were determined following guidelines of the U.S. Environmental Protection Agency (EPA).24 The IDL was calculated from 7 replicate injections of the lowest calibration standard (4.2−7.4 pg/μL) as IDL = tα × σ where tα refers to the critical value of the t distribution (α = 0.01) and σ is the standard deviation. The MDL was calculated analogously using 7 spiked samples (10−50 ng phenol) that ran through the complete sample preparation procedure. IDLs are presented in pg analyte per μL pyridine/BSTFA solution (Table 2), ranging

DOC content in order to analyze a water sample directly for dissolved lignin. The total dissolved lignin content of a water sample is defined as the sum of vanillyl, syringyl, and cinnamyl phenols (Σ8 = V+S+C). This leads to a Σ8-MDL of 2.8 μg/L and a method quantification limit Σ8-MQL of 8.5 μg/L. In a natural aquatic system with terrestrial DOM sources, total dissolved lignin accounts for about 0.5 mg/100 mg DOC what would lead to a minimal DOC concentration of 1.7 mg/L in order to quantify lignin. This estimation, however, underestimates the DOC concentration as it is solely based on the MDL and neglects the natural distribution of lignin phenols released during oxidation of a natural water sample. Consequently, the direct approach of dissolved lignin analysis can be applied to lignin-rich waters from wetlands,18 to riverine water with high terrestrial DOM concentrations, such as the Congo river,9 as well as to soil porewater or to laboratory experiments studying dissolved lignin decomposition on appropriate sample types. However, waters from many rivers, such as the arctic rivers,7 as well as estuarine and marine waters12,13 generally do not contain sufficient dissolved lignin for direct analysis. Lignin Phenol Extraction. The liberated lignin phenols are extracted from the aqueous phase after CuO oxidation. Alternatively to the liquid−liquid extraction with ethyl acetate, we followed the SPE approach with Oasis HLB cartridge after Kaiser and Benner.4 Deviating from the original method, the microwave digestion system used here processes higher sample volumes than the classical reaction minibombs. We therefore investigated the phenol retention capacity of the HLB resin by spiking equal amounts of a standard containing between 0.5 and 1.5 μg of each phenol into different water volumes (i.e., 2, 10, 25, and 50 mL) treated with base, followed by acidification and finally extracted according to the lignin protocol. Syringic phenols and those bearing an acidic functional group revealed a lower retention on the resin which led to considerable analyte breakthrough at higher sample volumes (Figure 2 and Table S1). In particular SAD, FAD, and DiOHBA showed recoveries below 50% when extracted from 50 mL water volume, whereas VAD and PAD were quantitatively recovered. In accordance with the original lignin oxidation protocol in a microwave digestion system,23 we routinely used a volume of

Table 2. Instrumental Detection Limit (IDL) and Method Detection Limit (MDL) of Single Lignin Phenols Using MRM by Tandem Mass Spectrometry compound

IDLa (pg/μL)

MDLb (ng/L)

PAL PON PAD VAL VON VAD SAL SON SAD DiOHBA CAD FAD

0.37 0.24 0.28 0.40 0.15 0.39 0.29 0.28 0.55 0.24 0.43 0.60

230 64.2 225 57.7 22.7 91.3 166 93.5 294 1100 1260 836

a

IDL refers to unsilylated analyte concentration in the injected pyridin/BSTFA mixture. bMDL refers to concentration of lignin oxidation products in a water sample.

from 0.15 to 0.60 pg/μL. MDL values are presented as lignin phenol concentrations in a water sample after CuO oxidation and ranged from 22.7 to 1260 ng/L. The lowest values between 22.7 and 93.5 ng/L were determined for phenols bearing a ketone group (PON, VON, and SON). These analytes showed a high and reproducible extraction recovery and low blank values. MDL values allow a rough estimation of the minimal

Figure 2. SPE recoveries of a lignin phenol standard mixture (0.5−1.5 μg of each phenol) extracted from 2, 10, 25, and 50 mL water. Error bars represent standard deviations of duplicate analyses. Σ Aldehydes = PAL + VAL + SAL, Σ Ketones = PON + VON + SON, Σ Acids = PAD + VAD + SAD + CAD + FAD. 13453

DOI: 10.1021/acs.analchem.7b03729 Anal. Chem. 2017, 89, 13449−13457

Article

Analytical Chemistry 15 mL for CuO oxidation of water samples. Recovery rates for calibration standards that ran through the complete analysis ranged from 76.7 to 111.6% except DiOHBA with a low recovery of 48.5% (Table S2). Loading time of the lignin phenols on the HLB sorbent was between 2 and 4 h. Even though no residual CuO was visually detected in the samples after centrifugation, we noticed the accumulation of CuO residuals on the SPE sorbent during loading thus reducing sample flow through the sorbent. Addition of EDTA to the sample before the acidification step counteracted this effect as no accumulation of CuO on the sorbent occurred. Intentionally added to prevent phenol losses in water samples that contain nitrate (see below), we therefore suggest the general use of EDTA before acidification. No negative effects of EDTA on phenol extraction were observed, but the HLB cartridge loading time is reduced and catalytically active copper(II) is removed at an early stage. Nitrate Interference. The C18 extraction of DOM from water samples not only concentrates lignin phenols but also removes dissolved inorganic salts from the sample matrix. As we omitted this step, inorganic salts remain present in the reaction mixture during CuO oxidation. Thus, potential matrix effects of the latter need to be addressed which we confidently believe have no effect on the lignin phenol analysis of aqueous DOM samples for two reasons. First, the method is commonly applied to dried sediments, soils or freeze-dried DOM, thus to samples that similarly bear a strong inorganic matrix;4,17 and second, (NH4)2Fe(SO4)2·6H2O is routinely added to the reaction mixture prior to the CuO oxidation step in order to remove residual oxygen, a step that generally increases the inorganic matrix considerably. As one exception to the otherwise low matrix effects, however, we noticed discrepancies in the lignin yield of water samples that contained high amounts of dissolved nitrate. To specifically investigate this effect, we spiked a P. australis leaf leachate with increasing amounts of sodium nitrate before CuO oxidation. Results of this experiment indicate a significant drop in lignin yield at nitrate concentrations of 15 mg/L nitrate-N or higher (t test, p < 0.001) (Figure 3a). Phenol losses at high nitrate concentrations were particularly pronounced for FAD and DiOHBA whose yields decrease by about 90 and 73%, respectively (Table S3). Different extraction experiments with aqueous phenol standard solutions indicated that phenol losses occure during the acidification step of the lignin workup. Furthermore, no significant phenol loss was detected if only nitrate was added to the aqueous solution before acidification and extraction, but, in contrast, phenol losses increased significantly if nitrate was existing in the presence of traces of CuO. Ice cooling of the sample during acidification slightly increased phenol recovery, but losses were still substantial compared to samples free of nitrate. A likely cause for the phenol losses is therefore the carry-over of trace amounts of CuO and the catalytic activity of the latter in mild phenol nitration reactions in the presence of nitrate and acid.25,26 To prevent phenol losses in the presence of nitrate, we undertook two approaches. First, we took advantage of the high intermolecular selectivity of the nitration reaction at room temperature25 and spiked the samples with a more electron-rich “sacrificial” phenol (for example 3′,4′,5′-trimethoxyacetophenone) before CuO oxidation. Results of this approach indicated only marginally increased lignin phenol recoveries, but negative impacts on lignin phenol blank values or the chromatographic

Figure 3. Effect of dissolved nitrate. (a) Lignin phenol yields of a P. australis leaf leachate with increasing nitrate concentrations. Sodium nitrate was artificially added to the leachate before CuO oxidation to final NO−3 -N-concentrations of 1, 5, 15, 25, and 50 mg/L. Error bars represent the standard deviation of duplicate analyses on different days. (b) Lignin phenol recoveries of the leaf leachate (1) in the absence of nitrate and the addition of 20 mg EDTA before acidification, (2) with 50 mg/L nitrate added before CuO oxidation of the sample, and (3) in the combination of 50 mg/L nitrate initially added and 20 mg EDTA added before acidification. Recovery rates are in relation to the phenol yield of the unmodified leachate; error bars represent the standard deviation of duplicate analyses on the same day. Σ Aldehydes = PAL + VAL + SAL, Σ Ketones = PON + VON + SON, Σ Acids = PAD + VAD + SAD + CAD + FAD.

phenol separation why this approach was not further pursued, respectively. The second approach we explored was the addition of 20 mg EDTA to the sample mixture before acidification (i.e., after lignin oxidation and removal of CuO by centrifugation). EDTA is a chelating agent that efficiently binds dissolved copper and thereby suppresses its catalytic activity. The results of this approach clearly indicate a high recovery of the lignin phenols in the presence of nitrate without negative effects on the overall analytical procedure (Figure 3b). We therefore suggest the addition of EDTA to suppress nitrate interferences. Method Performance. One major concern in previous studies was the sample size used for CuO oxidation. It has been shown that a decrease in sample size below 2 mg organic carbon leads to a disproportionally high decrease in phenol yield and a shift of certain phenol ratios.4,16,27 Especially syringyl and cinnamyl phenols are prone to this effect, and 13454

DOI: 10.1021/acs.analchem.7b03729 Anal. Chem. 2017, 89, 13449−13457

Article

Analytical Chemistry besides a decrease in Λ8 (the carbon normalized lignin yield), decreases in C/V- and S/V-ratios and changing acid-toaldehyde ratios are commonly observed. The reduction of CuO from 1 g to 250 mg and the addition of 10 mg glucose as an additional carbon source reduce these losses to some extent. In the here-described method, the amount of organic carbon is predetermined by the DOC content of the sample. Therefore, we studied the effect of DOC concentration on lignin yield by analyzing a P. australis leaf leachate with DOC concentrations of 5, 30, and 100 mg/L, prepared by sequential dilution. No significant differences were detected between the intrinsic lignin parameters for samples with 30 and 100 mg/L DOC (t test, p > 0.05) (Figure 4a). The sample with 5 mg/L

These changes in phenol yield for low organic carbon samples follow the reported trends in previous studies,4,16 which recommended a minimum sample amount of 1.5 mg organic carbon to avoid such critical phenol losses. However, the 30 mg/L DOC corresponds to only 0.45 mg organic carbon in the here-described method, and no significant deviations in phenol yield were detected. Up to our knowledge, no comprehensive studies on the effect of low organic carbon on the phenol yield have been performed using lignin oxidation in a microwave digestion systems.23 Even though data presented here are limited, it suggests that low DOC concentrations will only have minor negative impacts on the direct lignin analysis of natural DOM samples. Carbon normalized lignin yields in most waters are commonly lower than in fresh leaf leachates (Λ8 = 6.79 mg/100 mg DOC), thus DOC concentrations in the range of 5 mg/L will more likely contain certain lignin phenols in concentrations below the method quantification limit before such nonlinear phenol losses take place. To illustrate the overall method performance we analyzed a fresh P. australis leaf leachate and two water samples from the Lake Grosse Fuchskuhle, an experimental brown-water lake in northeast Germany that is artificially divided into different basins.20 Due to different catchment areas of the lake basins, the hydrochemistry of the different water bodies has diverged over time which has made Lake Grosse Fuchskuhle subject to various studies in the field of aquatic microbial ecology.22,28,29 We sampled the southwest lake basin that receives high influxes of humic-rich, allochthonous organic matter from an adjacent Sphagnum mire, and the northeast basin that receives no water from the mire and has DOM with a higher autochthonic character of algal origin.21 Not suprisingly, the highest dissolved lignin yield of Λ8 = 4.34 mg/100 mg DOC was measured in the leaf leachate (Table 3). Total dissolved lignin in the lake Grosse Fuchskuhle accounted for 1.12 mg/100 mg DOC in the southwest basin and 0.500 mg/100 mg DOC in the northeast basin opposite to the mire. Notably, DiOHBA was the most abundant lignin oxidation product in both basins. The most likely source of DiOHBA in this system are tannic acids, a substance group commonly present in Sphagnum peatlands that releases about 1.5 mg DiOHBA per 100 mg organic carbon upon CuO oxidation.18,30 Data of comparative analysis of the three water samples by the traditional C18 extraction approach are also shown in Table 3. Standard deviations of replicate analyses (n = 4) were not significantly different for both methods (t test, p > 0.05). The lignin phenol yields, however, were generally lower when DOM was processed via the C18 extraction approach prior to CuO oxidation. Using the direct lignin analysis of the water samples as reference, the dissolved lignin C18 recovery accounted for 89% for the leaf leachate and 48 and 79% for the samples Lake Grosse Fuchskuhle southwest and northeast, respectively. These recoveries stand in contrast to the total DOC recoveries of 39% for the leachate and 75 and 73% for the lake samples. Recoveries from C18 sorbents are sensitive to DOM type and usually decrease with increasing polarity of the DOM.14,17 The low C18 DOC recovery of the fresh leachate might therefore reflect high amounts of organic acids being not retained on the sorbent, while leached leaf lignin is. Dissolved lignin from the brown-water lake is less well-recovered with single phenol recoveries ranging between 40 and 54% in the humic-rich southwest basin and between 72 and 85% in the northeast basin. Particularly the low lignin recovery of the southwest

Figure 4. Lignin phenol yields and phenol ratios of a P. australis leaf leachate, sequentially diluted to 5, 30, 100 mg/L DOC before CuO oxidation. (a) Carbon normalized yields of dissolved phenol groups: P = PAL + PON + PAD, V = VAL + VON + VAD, S = SAL + SON + SAD, C = CAD + FAD. (b) Acid-to-aldehyde ratios of the vanillyl and syringyl phenols and the coumaric-to-ferulic acid ratio vs DOC concentration. 10 mg glucose was routinely added to all samples before CuO oxidation. Error bars represent standard deviations of triplicate analyses on the same day.

DOC, however, indicated a significant decrease in syringyl and cinnamyl phenol yield (t test, p < 0.05). Accordingly, this loss led to a highly significant decrease of the syringyl acid-toaldehyde ratio from 0.67 to 0.17 in the 5 mg/L DOC leachate (t test, p < 0.01) whereas the vanillyl acid-to-aldehyde ratio remained stable (Figure 4b). 13455

DOI: 10.1021/acs.analchem.7b03729 Anal. Chem. 2017, 89, 13449−13457

Article

Analytical Chemistry Table 3. Concentrations of Dissolved Lignin Phenols of Three Natural Water Samplesa P. australis leaf leachate

Lake Fuchskuhle (southwest)

Lake Fuchskuhle (northeast)

29.6 mg/L

41.9 mg/L

14.7 mg/L

DOC: direct compound PAL (μg/L) PON PAD VAL VON VAD SAL SON SAD DiOHBA CAD FAD Σ8 (μg/L)b Λ8 (mg/100 mg DOC)c P/Vd S/Ve C/Vf (Ad/Al)vg (Ad/Al)sh CAD/FAD

C18

direct

C18

direct

C18

58.84 26.76 50.77 137.9 43.28 114.9 107.6 71.47 115.6 42.91 233.3 461.5

± ± ± ± ± ± ± ± ± ± ± ±

8.49 1.92 2.07 9.3 3.91 3.2 8.1 5.71 4.8 10.64 2.7 6.5

36.35 22.85 40.17 103.9 39.11 108.9 91.33 57.58 104.0 33.21 211.0 428.9

± ± ± ± ± ± ± ± ± ± ± ±

2.85 1.35 2.77 9.4 5.85 3.2 11.63 5.64 9.6 10.42 4.1 12.8

49.92 65.34 50.18 112.0 53.27 96.03 52.45 23.82 38.40 230.8 43.95 49.44

± ± ± ± ± ± ± ± ± ± ± ±

3.39 3.50 2.28 5.6 2.58 2.98 1.91 0.90 1.09 14.3 1.00 0.57

43.97 55.73 45.63 50.55 27.33 49.35 20.93 12.80 18.22 209.6 21.10 24.81

± ± ± ± ± ± ± ± ± ± ± ±

1.42 1.83 2.11 1.98 1.16 1.84 1.07 0.58 1.36 8.6 0.82 1.24

7.079 9.906 10.73 12.92 5.504 13.09 6.491 6.070 6.138 52.22 8.077 15.27

± ± ± ± ± ± ± ± ± ± ± ±

1.036 1.497 0.89 1.34 0.482 0.95 0.917 0.449 1.083 5.22 0.573 1.06

10.08 9.358 11.08 9.938 4.359 10.89 4.713 5.152 4.542 42.75 6.287 11.95

± ± ± ± ± ± ± ± ± ± ± ±

1.27 1.541 0.97 1.039 0.482 0.86 0.526 0.192 0.294 3.66 0.253 0.49

1285.5 4.34 0.46 0.99 2.35 0.84 1.08 0.51

± ± ± ± ± ± ± ±

39.2 0.13 0.02 0.01 0.12 0.04 0.05 0.01

1144.7 3.87 0.39 1.00 2.55 1.05 1.15 0.49

± ± ± ± ± ± ± ±

60.6 0.20 0.01 0.03 0.13 0.07 0.09 0.01

469.4 1.12 0.63 0.44 0.36 0.86 0.73 0.89

± ± ± ± ± ± ± ±

16.0 0.04 0.01 0.00 0.01 0.02 0.01 0.01

225.1 0.537 1.14 0.41 0.36 0.98 0.87 0.85

± ± ± ± ± ± ± ±

9.9 0.024 0.02 0.01 0.00 0.01 0.03 0.01

73.57 0.500 0.88 0.59 0.74 1.02 0.94 0.53

± ± ± ± ± ± ± ±

6.63 0.045 0.03 0.03 0.03 0.03 0.06 0.01

57.83 0.393 1.21 0.57 0.73 1.10 0.97 0.53

± ± ± ± ± ± ± ±

3.16 0.021 0.05 0.03 0.08 0.03 0.10 0.03

a

Each DOM sample was analyzed using the direct approach presented here, and the traditional approach that includes DOM extraction using C18 SPE columns. All phenol concentrations are presented in μg/L. Errors are standard deviations (n = 4). bDissolved lignin, here as the sum of 8 lignin phenols (V, S, and C). cDissolved lignin normalized to DOC. dP = PAL + PON + PAD, V = VAL + VON + VAD. eS = SAL + SON + SAD. fC = CAD + FAD. gRatio of VAD to VAL. hRatio of SAD to SAL.

basin water sample results in a drop of lignin yield (Λ8) from 1.12 to 0.537 mg/100 mg DOC and thus creates an artificial lignin similarity between the different environmental basins if a C18 lignin extraction step is included in the analytical routine. A second notable trend is the high C18 recovery of the phydroxy phenols (PAL, PON, and PAD) and DiOHBA for the Lake Fuchskuhle samples with up to 104%, phenols that mainly originate from dissolved humic substances (tannic acids) and not from dissolved lignin.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Hendrik Reuter: 0000-0002-3145-6196

■

Author Contributions

CONCLUSIONS The newly developed method pursues the direct oxidation of dissolved lignin in water samples, omitting the classic C18 concentration step. Being applicable to lignin-rich freshwaters, this approach reduces the required amount of water sample and the lignin phenol preparation time considerably. The method provides a tool for lignin studies at the terrestrial−aquatic interface which will improve the knowledge of DOM leaching processes and diagenesis but is also applicable to wetland waters and soil porewaters.

■

australis leaf leachate, spiked with increasing amounts of sodium nitrate before CuO oxidation (Table S3). (PDF)

@

H.R. and J.G. contributed equally.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Jörg Gelbrecht and Hans-Peter Grossart for helpful discussions about sampling sites and Michael Linscheid for directing our thoughts toward the addition of EDTA to counteract nitrate effects. We furthermore appreciate comments by Tobias Goldhammer on a previous version of the manuscript. This research was supported by the Deutsche Forschungsgemeinschaft (ZA 742/2-1).

ASSOCIATED CONTENT

■

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.7b03729.

REFERENCES

(1) Thevenot, M.; Dignac, M.-F.; Rumpel, C. Soil Biol. Biochem. 2010, 42, 1200−1211. (2) Chemical Biomarkers in Aquatic Ecosystems; Princeton University Press, 2011; pp 248−266. (3) Hedges, J. I.; Ertel, J. R. Anal. Chem. 1982, 54, 174−178. (4) Kaiser, K.; Benner, R. Anal. Chem. 2012, 84, 459−464.

Recoveries of lignin phenols extracted from different water volumes (Table S1), and recovery rates of lignin phenols (Table S2), and percentual phenol yields of a P. 13456

DOI: 10.1021/acs.analchem.7b03729 Anal. Chem. 2017, 89, 13449−13457

Article

Analytical Chemistry (5) Woods, G. C.; Simpson, M. J.; Pautler, B. G.; Lamoureux, S. F.; Lafrenière, M. J.; Simpson, A. J. Geochim. Cosmochim. Acta 2011, 75, 7226−7241. (6) Eckard, R. S.; Hernes, P. J.; Bergamaschi, B. A.; Stepanauskas, R.; Kendall, C. Geochim. Cosmochim. Acta 2007, 71, 5968−5984. (7) Amon, R.; Rinehart, A.; Duan, S.; Louchouarn, P.; Prokushkin, A.; Guggenberger, G.; Bauch, D.; Stedmon, C.; Raymond, P.; Holmes, R.; McClelland, J.; Peterson, B.; Walker, S.; Zhulidov, A. Geochim. Cosmochim. Acta 2012, 94, 217−237. (8) Ward, N. D.; Richey, J. E.; Keil, R. G. Biogeochemistry 2012, 111, 629−645. (9) Spencer, R. G. M.; Hernes, P. J.; Dinga, B.; Wabakanghanzi, J. N.; Drake, T. W.; Six, J. Global Biogeochemical Cycles 2016, 30, 1105− 1121. (10) Perdue, E. M.; Ritchie, J. D. Treatise on Geochemistry, 2nd ed.; Elsevier, 2003; Vol. 5, p 605 (11) Dalzell, B. J.; Filley, T. R.; Harbor, J. M. J. Geophys. Res. 2005, 110, 2156−2202. (12) Hernes, P. J.; Benner, R. Deep Sea Res., Part I 2002, 49, 2119− 2132. (13) Hernes, P. J.; Benner, R. Mar. Chem. 2006, 100, 66−79. (14) Minor, E. C.; Swenson, M. M.; Mattson, B. M.; Oyler, A. R. Environ. Sci.: Processes Impacts 2014, 16, 2064−2079. (15) Louchouarn, P.; Amon, R. M.; Duan, S.; Pondell, C.; Seward, S. M.; White, N. Mar. Chem. 2010, 118, 85−97. (16) Louchouarn, P.; Opsahl, S.; Benner, R. Anal. Chem. 2000, 72, 2780−2787. (17) Spencer, R. G.; Aiken, G. R.; Dyda, R. Y.; Butler, K. D.; Bergamaschi, B. A.; Hernes, P. J. Org. Geochem. 2010, 41, 445−453. (18) Gandois, L.; Teisserenc, R.; Cobb, A.; Chieng, H.; Lim, L.; Kamariah, A.; Hoyt, A.; Harvey, C. Geochim. Cosmochim. Acta 2014, 137, 35−47. (19) Zak, D.; Gelbrecht, J.; Steinberg, C. E. W. Biogeochemistry 2004, 70, 357−368. (20) Koschel, R. Int. Rev. Gesamten Hydrobiol. 1995, 80, 507−518. (21) Sachse, A.; Babenzien, D.; Ginzel, G.; Gelbrecht, J.; Steinberg, C. Biogeochemistry 2001, 54, 279−296. (22) Buck, U.; Grossart, H.-P.; Amann, R.; Pernthaler, J. Environ. Microbiol. 2009, 11, 1854−1865. (23) Goñi, M. A.; Montgomery, S. Anal. Chem. 2000, 72, 3116− 3121. (24) Smith, R. L. Chemical Concentration Data Near The Detection Limit; EPA903/8-91/001; Environmental Protection Agency, 1991. (25) Gaude, D.; Goaller, R. L.; Pierre, J. L. Synth. Commun. 1986, 16, 63−68. (26) Hernando, E.; Castillo, R. R.; Rodríguez, N.; Gómez Arrayás, R.; Carretero, J. C. Chem. - Eur. J. 2014, 20, 13854−13859. (27) Amelung, W.; Flach, K. W.; Zech, W. Soil Sci. Soc. Am. J. 1999, 63, 1222−1228. (28) Grossart, H.-P.; Jezbera, J.; Horňaḱ , K.; Hutalle, K. M. L.; Buck, U.; Šimek, K. Environ. Microbiol. 2008, 10, 635−652. (29) Casper, P.; Chim Chan, O.; Furtado, A. L. S.; Adams, D. D. Aquat. Sci. 2003, 65, 36−46. (30) Dickens, A. F.; Gudeman, J. A.; Gélinas, Y.; Baldock, J. A.; Tinner, W.; Hu, F. S.; Hedges, J. I. Org. Geochem. 2007, 38, 1256− 1276.

13457

DOI: 10.1021/acs.analchem.7b03729 Anal. Chem. 2017, 89, 13449−13457