the potential distortion of sedimentary d15n and c /n · pdf filethe potential distortion of...

$: The potential distortion of sedimentary d15N and C /N · PDF fileThe potential distortion of sedimentary d15N and C org/N ratios by NH 4 ... ammonium isotope fractionation nitrogen$
The potential distortion of sedimentary d15N and Corg/N ratiosby NH4

þ and the effects of pre-analysis sample treatment

Sandra Meisel and Ulrich Struck*

Museum f�r Naturkunde, Leibniz-Institut f�r Evolutions- und Biodiversit�tsforschung an der Humboldt-Universit�t zu Berlin, 10115 Berlin,Germany, e-mail: [email protected]

Introduction

Nitrogen isotope abundance ratios (d15N) and C/N ra-tios are both important instruments in the study of mar-ine food webs. Their value – as with every proxy in-dicator – depends on the reliable recording of theoriginal phenomenon in the absence of the direct meas-ure. In this regard, the deceptive influence of samplepreparation on the original C/N ratios and d15N-signalsis of great interest and has already been addressed byseveral authors (Nieuwenhuize et al. 1994; Bunn et al.1995; Holmes et al. 1999; Jacob et al. 2005; Kennedyet al. 2005; Carabel et al. 2006; Ng et al. 2007).

Various studies have shown, for instance, that the re-moval of carbonate via acidification not only alters

d13C but may likewise bias d15N (Bunn et al. 1995; Ja-cob et al. 2005; Carabel et al. 2006; Ng et al. 2007). Asa consequence, many authors now refrain from pre-ana-lysis decalcification when measuring d15N and preferto conduct independent measurements using untreatedsubsamples (Goering et al. 1990; Thornton & McManus1994; Bunn et al. 1995; Bouillon et al. 2002; Carabelet al. 2006).

Our findings clearly indicate that this approach isnot applicable to all kinds of material. The goal ofthis study is to investigate the effects of common pro-cessing methods on the organic-rich sediments fromthe central Namibian shelf and to establish which canbe most trusted with respect to the d15N- and C/N re-cord.

# 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Received 15 November 2010Accepted 22 February 2011Published 3 August 2011

Key Words

ammoniumisotope fractionationnitrogen isotopessample treatment

Abstract

We examined the susceptibility of d15N-signals and Corg/N ratios in organic-rich sedi-ments to pre-analysis sample treatment. Each sample was subjected to three differentkinds of processing. For comparative purposes, the first measurement series (MS-1)was carried out on untreated sediment. In MS-2, the sediment was rinsed with distilledwater. In MS-3, analyses were carried out on decalcified and rinsed material, in MS-4the samples were decalcified without being subsequently washed. The sediment yieldedconspicuously different results depending on the type of processing it was subjected to.Rinsing, irrespective of whether acidification was included or not, induced substantialmodifications in d15N accompanied by a pronounced loss of NH4

þ (up to 14 wt% ofthe initial N-content). Molar Corg/N ratios, on the other hand, were only affected by acombination of acidification and rinsing. The discrepancies are ascribed to the influ-ence of decomposition-derived ammonium (NH4

þ). In untreated sediment (MS-1),NH4

þ seems to produce misleading shifts in both d15N-signals and Corg/N ratios. Pre-sumed mechanisms involved are as follows: Firstly, nitrogen isotopes fractionate duringNH4

þ-volatilisation in the heating oven, where the sediment is put to desiccate. Sec-ondly, NH4

þ-ions are able to escape that fractionation when adsorbed to negatively-charged SiO2-surfaces. The adsorption capacity of SiO2 increases with increasing pHof the pore water and hence with increasing carbonate content. Our findings raise ser-ious doubts about whether untreated sediment (MS-1) can provide reliable Corg/N andd15N-records. Pre-analysis acidification plus rinsing (MS-3) seems to eliminate the de-ceptive influence of NH4

þ-adsorption and -outgassing.

Fossil Record 14 (2) 2011, 141–152 / DOI 10.1002/mmng.201100004

* Corresponding author

Material and methods

Core 180

Core 180 was retrieved from the northern Benguela Upwelling areaduring METEOR cruise M57-3 (15/03/2003 to 08/04/2003) (Fig. 1).The 5 m long core consists of organic-rich diatomaceous ooze (aver-age opal 54.3 %, average organic matter 9.8 %, average carbonate5.9 %; Bremner & Willis 1993), which coats the inner shelf as a beltrunning parallel to the coast at a depth of up to 150 m (Shannon &O’Toole 1998; Struck et al. 2002; Br�chert et al. 2004). Its presencetestifies to the high productivity of the region. The deposit has anaverage thickness of 5.1 m, an estimated age of 5000 years with amean sedimentation rate amounting to 1 mm/a (Bremner & Willis1993).

The total organic carbon content of core 180 averages 5.5 wt%.The sediment is primarily made up of phytoplankton remains, mainlydiatom frustules (Vavilova 1990; Graml 2001). This explains the highpercentage of biogenic silica (opal) and why unaltered d15N shouldmore or less correspond to the isotopic signature of the first trophiclevel. A dilution by higher trophic levels as well as terrestrial detritus(i.e. clay minerals) is considered negligible (Holmes et al. 1998; Pi-chevin et al. 2005).

Sample processing and isotopic measurements

A total of 100 syringe samples were taken at 5 cm-intervals. The sedi-ments were desiccated in a heating oven (40 �C) and homogenisedusing an agate mortar. Each sample was subsequently split into fourparts with each subsample being subjected to an individual pre-analy-sis sample treatment (Fig. 2).

(1) The first measurement series (MS-1) was carried out on� 8 mg of untreated sediment wrapped in tin-foil cups. [N] and [TC](total carbon) measured relate to the bulk sediment volume.

Meisel, S. & Struck, U.: Ammonia loss and distortion of d15N of marine sediments142

Figure 1. The provenance of core 180 (22.88�S/14.41�E; waterdepth 48 m). The organic-rich diatomaceous ooze coating theinner shelf is shaded grey.

Figure 2. The single measurement series(MS) in brief. MS-1 and MS-2 yield thetotal carbon content (TC), whereas, dueto the acidification process, MS-3 andMS-4 yield the amount of organic car-bon (Corg).

museum-fossilrecord.wiley-vch.de # 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

(2) In MS-2, the material was rinsed, shaken, and centrifuged a to-tal of 3 times at 3000 rpm before being desiccated a second time(40 �C). The measurement was carried out on � 8 mg of sedimentwrapped in tin-foil cups. Note that only about a quarter of the sam-ples was subjected to this type of treatment (n ¼ 22). As some com-pounds may be lost during rinsing, [N] and [TC] measured in MS-2relate to a smaller sediment volume than [N] and [TC] in MS-1. Thismakes [N] and [TC] appear proportionally larger compared to MS-1.

(3) In order to determine the amount of organic carbon (Corg), car-bonate needs to be eliminated via acid treatment prior to measure-ment. Acidification was conducted in two different ways: In the firstcase (MS-3), the sediment was put into a centrifuge tube and treatedwith 2N hydrochloric acid. It was then thoroughly rinsed with dis-tilled water and centrifuged a total of 3 times at 3000 rpm. Measure-ments were made on � 8 mg of dried (40 �C) sample material wrappedin tin-foil cups. As some compounds (including carbonate) are lostduring rinsing, [N] and [Corg] in MS-3 again relate to a smaller sedi-ment volume than [N] and [TC] in MS-1.

(4) In MS-4 the samples were directly acidified in tarred silvercups by adding 2N hydrochloric acid until no more CO2 was released.Prior to analysis, the acid was allowed to evaporate in the heatingoven (40 �C). Different from MS-3, [N] and [Corg] relate to the bulksediment volume since no subsequent washing was applied (NB: Notethat the sample’s weight was taken before adding the acid.).

The isotopic composition of nitrogen (d15N) and total weight per-cent (wt%) of carbon and nitrogen were determined with a ThermoFlash EA 1112 Elemental Analyser connected to an isotope-ratiomass-spectrometer (Finnigan, Delta V). The reference gas was pureN2 nitrogen from a cylinder calibrated against IAEA-standards N-1and N-2. The external reproducibility of the isotope measurementswas tested with an internal standard (peptone) after every 5th meas-urement. As regards the concentrations of C and N, reproducibility is� 3% of the original amount present. For d15N, the analytical preci-sion of the lab standard was � 0.2 ‰ (1s standard deviation). Theisotope abundance ratios are reported in the conventional d-notationin per mil (‰) with respect to atmospheric N2 (AIR).

Ammonia concentration in the rinsing water

Ammonium (NH4þ) is a by-product of organic matter decomposition.

In organic-rich and anoxic sediments NH4þ is abundantly produced

and easily preserved.As outlined below, our findings gave strong reason to believe that

NH4þ exerts a deceptive influence on the original d15N- and Corg/N

record. In order to substantiate that assumption, we decided to testthe rinsing water in MS-2 and MS-3 for NH4

þ. The measurementswere made on the same selection of samples subjected to MS-2(n ¼ 22). The concentration of NH4

þ was measured by means of thephotometrical test ‘LCK 303 Ammonium Cuvette Test’ of Hach-Lange (http://www.hach-lange.co.uk). For MS-3, the pH was adjustedto 7 by the addition of 1 molar NaOH-solution. The amount of NH4

þ

always lay within the accepted methodological range (2–47 mg/l).

Results

Delta15N vs. depth

The down-core distributions of d15NHCl and d15Nuntreated

are similar in both quality (pattern) and quantity (iso-tope composition) (Fig. 3A). Both records show largebut comparable variations spanning roughly 8 to 15 ‰(d15Nuntreated) and 9 to 15 ‰ (d15NHCl). Median d15NHCl

(11.4 ‰) is slightly higher than median d15Nuntreated

(10.2 ‰) (Fig. 3B; Table 1).Washing with distilled water, irrespective of whether

prior acid treatment is included (MS-3) or not (MS-2),entails strong modifications in the down-core develop-ment of d15N (Fig. 3A; Table 1). In both measurementseries, the range of d15N-variation is reduced to withinthe standard error inherent in the method. Both datasets appear much smoother than d15Nuntreated.

The lowering from mean 10.2 ‰ (or 10.0 ‰; seeTable 1) in MS-1, to 8.5 ‰ in MS-3 and 8.4 ‰ inMS-2 suggests a mobile phase with comparably highd15N-values (Fig. 3B). The offsets distinctly exceed themeasurement precision and are almost equivalent to theisotopic difference between two trophic levels (Fry1988). Given the magnitude of this discrepancy, the re-cords would lead to completely different conclusionsregarding nutrient supply and consumption.

Corg/N ratios vs. depth

With MS-3 and MS-4 it is possible to directly calculateCorg/N. In contrast, MS-1 and MS-2, not having beenacidified, provide no information about [Corg]. Thismeans that we must derive this information from else-where. As regards MS-1, we can easily use [Corg] ofMS-4, as both [N] from MS-1 and [Corg] from MS-4refer to the bulk sediment volume (the d13C-values inMS-4 oscillate minimally around �19.5 ‰ (not shown),which indicates that the removal of carbonate was com-plete.).

This is not possible with MS-2 because [N] refers toan amount smaller than the bulk sediment volume. Forthe sake of legibility, the calculation of [Corg] in MS-2is shown elsewhere (see caption next to Fig. 4).

Fossil Record 14 (2) 2011, 141–152 143

Table 1. Standard deviation, median, minimum and maximum values of Corg/N and d15N. MS-2 is limited to a selection ofsamples (n ¼ 22). Values in brackets are based on this same selection of samples. This is in order to ensure a better compara-bility of MS-2 with the other measurement series.

MS-1

untreated

MS-2

H2O

MS-3

HCl þ H2O

MS-4

HCl

MS-1

untreated

MS-2

H2O

MS-3

HCl þ H2O

MS-4

HCl

Corg/N d15N (%�)

standard deviation 0.6 (0.5) 0.7 0.2 (0.2) 0.4 (0.4) standard deviation 1.7 (1.8) 0.2 0.4 (0.5) 1.4 (1.5)

median 7.2 (7.3) 7.4 7.5 (7.5) 7.3 (7.3) median 10.2 (10.0) 8.4 8.5 (8.5) 11.4 (11.2)

minimum 5.4 (5.9) 5.5 7.1 (7.2) 5.7 (6.9) minimum 7.9 (8.3) 8.0 7.8 (7.9) 8.8 (9.3)

maximum 8.9 (8.2) 8.6 8.5 (8.4) 8.7 (7.7) maximum 14.8 (14.1) 8.7 9.9 (9.9) 14.8 (14.0)

# 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim museum-fossilrecord.wiley-vch.de

Corg/N ratios are neither affected by washing norby in-situ acidification. [Corg/N]H2O and [Corg/N]untreated

show a markedly similar development (Fig. 4, rightchart) and [Corg/N]HCl deviates only sporadically from[Corg/N]untreated. Corg/N strongly reacts, however, to acid-ification plus rinsing (MS-3). The combination of acid-ification and rinsing leads to pronounced downcoredata-smoothing, accompanied by a slight increase ofmedian Corg/N (from 7.2 in untreated material to 7.5 inacidified plus rinsed material; Table 1).

Ammonium in rinsing water

The rinsing water from MS-2 and MS-3 contains – asexpected – significant amounts of ammonium nitro-gen. In MS-2, the loss averages 3.5 wt%, ranging from0.4 to 10.8 wt% (Fig. 5). In MS-3, the loss averages6.3 wt%, spanning 1.1 to 14 wt%. The offset betweenthe losses in MS-2 and MS-3 averages 2.1 wt%. Ob-viously, prior acidification (MS-3) somehow mobilisesan even greater quantity of NH4

þ.

Discussion

A study of the available literature shows that the effectsof sample preparation on d15N and C/N are far fromunanimous. Bunn et al. (1995), for instance, found thatacidification and subsequent rinsing increased the meand15N-signatures of shrimp tissue and decreased those ofseagrass to a degree that may confuse the interpretationof food webs. Rinsing is a common practice applied inorder to remove acid remains (Stoner & Zimmerman1988; Kang et al. 2003) but the fact of the matter isthat several authors strongly advise against it in orderto minimise the loss of dissolved organic matter (Nieu-wenhuize et al. 1994; Bunn et al. 1995; Jacob et al.2005; Carabel et al. 2006). Still, acidification alonedoes not guarantee the stability of organic compounds,either: Ng et al. (2007) and Jacob et al. (2005) reportedelevated C/N ratios in decalcified sample volumes, sug-gesting a proportionally greater loss of N relative to C.Moreover, the d15N-signature of sedimentary matter(Kennedy et al. 2005; Carabel et al. 2006), marine algae


Figure 3. A. Downcore development of d15N. For reasons of clarity, d15NH2O (MS-2) is shown separately; B. Delta15N-data pooledin boxplots. For better comparison of MS-2 with the other measurement series, the boxplots to the left are based on the sameselection of samples processed in MS-2.


and cyanobacteria (Ng et al. 2007), invertebrates andfish (Jacob et al. 2005) decreased significantly in re-sponse to in-situ acidification. In contrast, mollusc tis-sue (Ng et al. 2007) and algal material examined byKennedy et al. (2005) proved rather resistant.

In tests conducted by Holmes et al. (1999) neitherthe application of weak hydrochloric acid nor subse-quent rinsing significantly affected sedimentary d15N.

It clearly shows that the influence of sample prepara-tion is as variable and manifold as the types of materialinvestigated. According to our own results, washingmay leave pronounced marks on the d15N-record whileCorg/N is only affected when samples are subjected toboth acidification and rinsing. The questions are: first,which record can be trusted to provide the prime isoto-pic signal required for palaeoceanographic reconstruc-tions? Second, can the loss of NH4

þ be held responsi-ble for the changes of Corg/N and d15N occurring in thecourse of sample processing? We pre-empt the answer –yes, there is a variety of evidence that substantiates theinfluence of ammonium on both Corg/N and d15N.

The potential impact of NH4þ on Corg/N and the role

of carbonate

Corg/N ratios are a widely used instrument in distin-guishing between marine and terrestrial organic matter(Schubert & Calvert 2001). The observation that NH4

þ

is capable of biasing the original C/N ratio and causingmisinterpretation is not new. Several studies have al-ready reported fixation and adsorption of NH4

þ in andonto clay minerals and accordingly low C/N ratios (e.g.M�ller 1977 and references therein; Rosenfeld 1979;

Fossil Record 14 (2) 2011, 141–152 145

Figure 4. Downcore development of Corg/N. The right chart shows [Corg/N]H2O (MS-2) along with the corresponding [Corg/N]untreated data (MS-1), revealing their striking similarity. Since rinsing eliminates no carbonate, the amount of organic carbon inMS-2, i.e. [Corg]MS-2, had to be calculated in an indirect manner. [Cinorg] ¼ [TC]MS-1 – [Corg]MS-4; [TC]MS-1 and [Corg]MS-4 both referto the bulk sediment volume and are thus comparable; [Corg]MS-2 ¼ [TC]MS-2 – [Cinorg]. Note that [TC]MS-2 refers to a smallersediment volume (due to the loss of various compounds in the course of rinsing) than [Cinorg] (which refers to the bulk sedimentvolume). For this reason, [Cinorg] turns out to be slightly underestimated when applied in the equation. The underestimation of[Cinorg] results in an overestimation of [Corg]MS-2 and consequently [Corg/N]H2O. Overall, however, the error is expected to be small.Besides that, our conclusions are based on down-core patterns rather than absolute values.

Figure 5. The amount of NH4þ-N washed away by the rinsing

water in MS-2 (H2O) and MS-3 (HCl þ H2O). The losses relateto the amount of nitrogen present in untreated sediment (MS-1). The measurements were made on the same selection ofsamples subjected to MS-2 (n ¼ 22). On average, acidificationplus rinsing mobilises 2.1 % more NH4

þ-N than rinsing alone.


von Breymann & Suess 1988; Schubert & Calvert2001; Morse & Morin 2005). In core 180, clay mineralsare negligible and so is their significance in this con-text. There must be another reason for the observeddeviations of Corg/N. On the evidence of the excellentcorrelation between [CaCO3] and Corg/N in MS-1(Fig. 6A), carbonate appears to be the key to the an-swer.

Due to the alkaline nature of the carbonate ion(CO3

2�), the pH of the interstitial water is slightlyraised while carbonate dissolves (Faure 1992). An in-creasing pH promotes the adsorption of dissolved am-monium ions to silicate surfaces (Brunelle 1978). Bio-genic silica is fairly abundant in the form of diatomfrustules. The isoelectronic point of silica is aroundpH 1. For further explanation: The isoelectric point(IEP) is the pH at which a particular molecule carriesno net electrical charge. The surfaces of mineral oxidesthat are surrounded by a solution with a pH higher thanits IEP result negatively charged which is compensatedfor by adsorbed cations. The reverse is true if the parti-cle is dipped in a medium with a pH lower than its IEP(Brunelle 1978). However, the negative polarisation ofsilicate surfaces only becomes significant at a pH above7. This is the threshold above which the amount ofNH4

þ adsorbed to silica surfaces increases rapidly(Brunelle 1978). The pH in marine sediments oscillatesaround 8, held more or less stable due to the bufferingqualities of ammonium and carbonate (Ben-Yaakov1973). At these pH-ranges (pH > 7) even slight changesin the Hþ/OH––-ratio have pronounced effects onNH4

þ-adsorption (Brunelle 1978).The negative correlation between Corg/N and

[CaCO3] in MS-1 (Fig. 6A) is ascribed to the combina-tion of NH4

þ-adsorption and -outgassing: The amountof adsorbed ammonium rises along with the carbonatecontent as the adsorption capacity of SiO2 is enhancedat higher pH-values. At the same time, it is predomi-nantly adsorbed (as opposed to dissolved) ammonium

that is kept from outgassing in the form of NH3 in theheating oven (more on the ammonium loss due to out-gassing in the following chapter).

Note that our reasoning requires the pore water to beundersaturated with respect to carbonate (if the solutionwas already supersaturated, rising [CaCO3] would notlead to enhanced dissociation of CaCO3), a prerequisite


"

Figure 6. Corg/N vs. [CaCO3]: [CaCO3] was calculated by sub-tracting [Corg] (MS-4) from the total amount of carbon (MS-1).The resulting amount of inorganic carbon was multiplied by8.33, according to the molar relationships in CaCO3. CaCO3

¼ ([TC]MS-1 – [Corg]MS-4)�8.33. The negative correlation in(A) testifies to the increasing number of NH4

þ-ions ad-sorbed to negatively charged SiO2-surfaces, the higher theamount of carbonate. The dissolution of carbonate lifts the pHin the interstitial water, which increases the surface polarisationand adsorption capacity of SiO2. The higher the percentage ofadsorbed NH4

þ-ions the lower the N-loss due to outgassing.Washing with distilled water (B) does not change anythingabout the negative correlation. Instead, acidification lowers theadsorption capacity of SiO2, causing the release of adsorbedNH4

þ-ions (C and D). During subsequent rinsing they arewashed away with the rinsing water. This is why the negativecorrelation is completely erased in MS-3 (C). Acidificationwithout rinsing seems to release previously adsorbed NH4

þ

during later drying as NH3 (D).


that is met by the high partial pressures of CO2 (pro-duced by OM matter decay) and the comparably lowtemperatures (Faure 1992).

Also note that NH4þ has NH3 as a conjugate base.

This implies that the amount of NH4þ decreases in fa-

vour of NH3 as the pH rises. However, this effect isseemingly outweighed by the concomitantly increasingadsorption capacity of SiO2 as the pH increases. Thefact that the solubility of SiO2 increases as the porewaters get more basic (Brunelle 1978; Faure 1992) doesnot disturb the pattern either.

Acidification plus rinsing has a considerable impacton the Corg/N ratios. In MS-3, the originally inversecorrelation between [CaCO3] and Corg/N is completelyerased (Fig. 6C). Lower Corg/N ratios of carbonate-richsample splits are adapted to the Corg/N ratios in carbo-nate-poor sample splits and mean Corg/N is slightlyraised (see also Table 1).

What happens in the course of acidification plus rin-sing? As the pH is lowered, the SiO2-adsorption capacitydiminishes and the SiO2-surfaces result oversaturatedwith respect to ammonium. Ammonium is released intothe pore water and carried away in the course of rin-sing. Corg/N consequently rises.

In-situ acidification (MS-4) brings about the same re-lease of formerly adsorbed NH4

þ. However, in contrastto MS-3, the liberated NH4

þ-molecules are not washedaway. In the heating oven some of them outgas to formgaseous NH3 (see below). This outgassing is consideredthe reason for the disturbance of the originally inversecorrelation between [CaCO3] and Corg/N (Fig. 6D).

Note that the amount of ammonium lost in MS-2 stillaverages more than half of the amount lost in MS-3(Fig. 5). Assuming that rinsing alone releases hardlyany adsorbed NH4

þ, the bulk of the ammonium lost inMS-2 represents dissolved NH4

þ. If this hypothesis isright, how is it that the leaching of dissolved NH4

þ

does not result in rising Corg/N ratios? [Corg/N]untreated

and [Corg/N]H2O are virtually identical (Figs 4, 6B). Themost likely explanation is the simultaneous loss of dis-solvable organic matter with high Corg/N ratios: Organicmatter that has already been subjected to some micro-bial activity usually exhibits higher Corg/N ratios thanfresh organic matter. This is because nitrogen is moresusceptible to decay than carbon and, therefore, prefer-entially lost (Rosenfeld 1981; Lehmann et al. 2002).Apparently these compounds are sufficiently unstableto be leached away together with dissolved NH4

þ, thusinhibiting the increase in Corg/N. In contrast, acidifica-tion plus rinsing is capable of additionally mobilisingoriginally adsorbed NH4

þ; this additional loss leads tothe observed increase of Corg/N ratios in MS-3.

The weakness of the correlations notwithstanding,Figure 7 substantiates our hitherto held ideas and con-firms the control of carbonate, or rather pH, on themobility of NH4

þ in MS-2 and MS-3. The less CaCO3,the less adsorbed NH4

þ and the greater the loss duringrinsing (Fig. 7A). The positive correlation in Figure 7bfurther corroborates the assumption that acidification

plus rinsing (MS-3) mobilises both adsorbed and dis-solved NH4

þ, while rinsing alone (MS-2) mobilisesonly the latter.

The potential impact of NH4þ on d15N

Rinsing caused average d15N to drop by 1.6 ‰(Fig. 3B; Table 1). According to our above line of rea-soning, rinsing does predominantly (if not exclusively)mobilise dissolved NH4

þ. This implies a d15N-enrich-ment in dissolved NH4

þ compared to the residual se-diment. The 15N-enrichment in dissolved NH4

þ may bedue to nitrogen isotope fractionation during the dryingprocess. When NH4

þ transforms to gaseous NH3, thelight isotope (14N) may accumulate in the gaseous phase,thus leaving 15N-enriched NH4

þ behind (Fig. 8B). Theprocess is comparable to kinetic isotope fractionation

Fossil Record 14 (2) 2011, 141–152 147

Figure 7. The control of CaCO3 on the fate of NH4þ during

sample treatment. Both charts substantiate the ability of CaCO3

to promote NH4þ-adsorption; A. Samples containing lower car-

bonate tend to loose more NH4þ during rinsing (MS-2). This is

attributed to the slightly lower pH of the pore waters, the conse-quently less negative surface charge of the SiO2-particles andthe higher percentage of NH4

þ dissolved in interstitial waters;B. Chart (B) testifies to the rising proportion of adsorbed NH4

þ

as carbonate and the pH increases. Acidification plus rinsingmobilises both adsorbed and dissolved NH4

þ while rinsingalone mobilises only dissolved NH4

þ. At low CaCO3, whereonly few NH4

þ is adsorbed, the offset between N lostHCl þ H2O

and N lostH2O tends to be low. The increasing offset with higher[CaCO3] is due to the rising proportion of adsorbed NH4

þ andits loss in the course of acidification plus rinsing.


of oxygen isotopes during the vaporisation of water andcan reach pronounced proportions. In a study conductedon simulated urine patches on grassland soils, the iso-topic composition of the NH3 volatilised was depletedby up to almost 30 ‰ compared to the original urea-Nadded (Frank et al. 2004).

The inverse relation between d15N and [N] in un-treated material (Fig. 9A) substantiates that ammoniavolatilisation – accompanied by the discriminationagainst 15N – is taking place. The outgassing of 15N-depleted NH3 results both in the reduction of [N] andan increase of the d15N-signatures. After rinsing, thenegative correlation is completely erased (Fig. 9B); thisagrees with the leaching of 15N-enriched ammoniumions as illustrated in Figure 8C.

Acidification plus rinsing (MS-3) and rinsing alone(MS-2) yield extremely similar d15N-patterns (Figs 3,9B–C). Apparently, the mobilisation of originally ad-sorbed NH4

þ in MS-3 has no additional effect on theisotopic signature of the sediment. This suggests thatadsorbed NH4

þ is resistant to both the outgassing andfractionation during the drying process (Figs 8B, D).

On the basis of the available evidence, d15N is stronglycontrolled by dissolved NH4

þ. As outlined in the abovesection, the amount of dissolved NH4

þ depends on theadsorption capacity of SiO2, which co-varies with theamount of carbonate, or rather, the pH of the interstitialwaters. The lower the carbonate content, the lower thepH and the higher the amount of dissolved NH4

þ, whichis the fraction liable to fractionation in the heatingoven. Figure 10 reflects the anticipated relationship be-tween carbonate and d15N reasonably well. The lowerthe carbonate content, the more pronounced the shifttowards higher d15Nuntreated-values (Fig. 10A). Due tothe leaching of isotopically “heavy” NH4

þ-moleculesin both MS-2 and MS-3, the negative correlation is lostand average d15N is again lowered (Figs 10B–C).

In-situ acidification (MS-4) increases the amount ofdissolved NH4

þ-molecules (Fig. 8E) compared to theamount already present in the original sediment(Fig. 8B). This enhances the loss of 15N-depleted NH3

during desiccation and might explain why average


3

Figure 8. Schematic summary of the processes occurring dur-ing various methods of sample preparation. Arrows mark thefractions lost during processing; A. Chart (A) identifies thefractions that are relevant as regards the modification of Corg/Nand d15N during sample treatment. NH4

þ is excreted in thecourse of organic matter (OM) decayal, leaving partially de-composed OM with elevated Corg/N behind. If not stated other-wise, Corg/N and d15N-values reflect unaltered OM; B. In theheating oven, some of the dissolved NH4

þ is lost in the formof 15N-depleted NH3, thus leaving a 15N-enriched residual be-hind; C. Rinsing mobilises intermediate and by-products result-ing from deanimation reactions, exept for adsorbed NH4

þ.Adsorbed NH4

þ-ions are released during acidification; D. Thesubsequent rinsing in MS-3 induces their loss; E. In MS-4, un-fixed NH4

þ are not washed away but fractionate during desic-cation, thereby increasing the 15N-enrichment in the residual.


d15NHCl is slightly higher than average d15Nuntreated

(Fig. 3B). Figure 10D furthermore reveals that the in-crease of d15N is more pronounced where CaCO3 isabundant, i.e. where we expect a higher proportion ofNH4

þ to be released.The presence of NH4

þ seems equally problematicwith d15N as with Corg/N. NH4

þ is generated duringOM decayal. As our data indicates, the decayal itselfdoes not seem to produce the observed 15N-enrichmentin the residual sediment. This is worth noting because

it disagrees with the widely held belief in nitrogen iso-tope fractionation during protein hydrolysis (e.g. Mel-ander 1960; Gaebler et al. 1966; Saino & Hattori 1980;Altabet & McCarthy 1985; Sch�fer & Ittekkot 1993;Altabet & Francois 1994; Montoya 1994; Ostrom et al.1997; Sachs & Repeta 1999).

What makes us so certain that deanimation occurredwithout fractionation is that MS-2 and MS-3 yield somarkedly similar d15N-signatures (Fig. 3B). The acid-induced mobilisation of adsorbed NH4

þ in MS-3 doesnot additionally influence the isotopic signature of thesediment. Apparently, adsorbed NH4

þ is unfractionatedper se. Its d15N-signature seemingly corresponds to thatof fresh organic matter, otherwise its loss during acidi-fication plus rinsing should have induced additionalisotopic shifts. Our data clearly demonstrates that theobserved shift towards higher d15N-signals in MS-1truly results from kinetic isotope fractionation duringthe drying process, i.e. during the transformation ofNH4

þ to gaseous NH3.

Which record to trust

It appears that the use of untreated material for d15N-analyses is not, contrary to popular belief (see above),advisable a priori. Our findings have shown that thedown-core scattering of [Corg/N]untreated and d15Nuntreated

is primarily to be ascribed to the misleading influenceof NH4

þ, rather than being indicative of environmentalfluctuations.

If interested in the original signal, it is recommendedthat NH4

þ be removed prior to measuring d15N andCorg/N. This appears to be done best by a combinationof acidification and subsequent rinsing (MS-3). Acidifi-cation plus rinsing eliminates the effects of NH4

þ-ad-sorption and -outgassing and restores the original signalof both Corg/N and d15N. If d15N is the sole parameterof interest, it is enough to remove dissolved NH4

þ as itis only dissolved NH4

þ that undergoes fractionationduring desiccation; then rinsing alone suffices (MS-2)in order to eliminate the decomposition-, or more pre-cisely, outgassing-derived shift towards higher d15N-sig-nals.

Fossil Record 14 (2) 2011, 141–152 149

3

Figure 9. d15N vs. [N]. Dissolved NH4þ is lost in the form of

gaseous, 15N-depleted NH3; A. Its outgassing results in the si-multaneous reduction of [N] and accumulation of 15N-enrichedNH4

þ in the residual, hence the inverse relation; B–C. The ne-gative correlation is erased in both MS-2 and MS-3, implyingthat rinsing alone provokes the leaching of 15N-enriched NH4

þ.The acid-induced loss of adsorbed NH4

þ in MS-3 has no addi-tional influence on d15N, because adsorbed NH4

þ has not un-dergone any fractionation; D. As 15N-enriched NH4

þ is notwashed away in MS-4, the negative correlation remains. NB:In MS-1 and MS-4, [N] refers to the bulk sediment volume. InMS-2 and MS-3, the sample splits have already undergone pro-cessing and consequently lost a certain amount of componentswhen being weighed. This is why the percentage of [N] appearssomewhat higher in MS-2 and MS-3 (B, C).


Conclusions and recommendations

[1] Our data provide strong evidence of nitrogen iso-tope fractionation during desiccation when NH4

þ trans-forms into gaseous NH3. Thereby, 14N accumulates inthe gaseous phase, thus leaving 15N-enriched NH4

þ be-hind.

[2] The outgassing and fractionation only concernsNH4

þ-ions dissolved in the pore waters. In contrast,NH4

þ-molecules adsorbed to negatively-charged SiO2-surfaces remain unaffected. The adsorption capacity ofSiO2 increases with increasing pH, i.e. rising carbonatecontent.

[3] We conclude that excessive confidence in un-treated material is inappropriate when NH4

þ-moleculesare involved: Ammonia volatilisation is capable of pro-ducing misleading fluctuations in d15Nuntreated withoutany environmentally-triggered background. Our resultsclearly indicate that the higher the amount of dissolvedNH4

þ, the more we must suspect unwanted shifts to-wards higher d15N in untreated sediment. At the sametime, it remains unclear how many of the dissolvedNH4

þ-ions turn into gaseous NH3. Due to this un-controllable and variable outgassing, [Corg/N]untreated islikewise considered unreliable.

[4] Acidification plus rinsing seems capable of wip-ing out the effects of NH4

þ-adsorption and -outgassingand associated distortions of d15N and Corg/N: Acidifi-cation reduces the negative surface polarisation of SiO2

and liberates adsorbed ammonium ions. Subsequent rin-sing ideally mobilises all decomposition-derived inter-mediate products, including both dissolved and for-merly bound ammonium. Certainly we cannot excludethe possibility of parts of “fresh” organic matter beingconcomitantly washed away.

[5] In organic-rich and anoxic sediments, where theconditions are favourable for the presence of NH4

þ, weassume acidified and rinsed material to deliver thetrustworthiest record with decayal-derived influencesbeing eliminated.

Several authors have called for a standardisation ofpre-analysis sample treatment, fearing that the differentmethodology may hamper the comparability of data(e.g. Jacob et al. 2005; Ng et al. 2007). We do onlysupport this view as long as it is the same type of ma-terial analysed.

Due to compositional differences, however, sedimentsmay require completely different approaches in order toovercome possible misinterpretations of Corg/N ratiosand d15N-values (e.g. Schubert & Calvert 2001). Thechemistry of the interstitial waters, the porosity and thesolid/water ratio are further aspects to consider.

Our results certainly do not warrant any generalisa-tion. Instead, they highlight the importance of carefullyinvestigating the varying effects of different processing


3

Figure 10. d15N vs. [CaCO3]. The negative correlation in MS-1(A) reflects the enhanced outgassing of 15N-depleted NH3 thelower the amount of carbonate, i.e. the lower the negative sur-face polarisation of SiO2. Only dissolved NH4

þ is able to out-gas and fractionate. For further details see text. Washing alone(B) or washing and acidification (C) reduces the d15N-variabil-ity completely. Here the remaining fraction of porewater-NH4

þ

after its incomplete degassing is washed away. The treatmentwith HCl alone (D) does not remove this fraction completely.


methods, rather than following a simple recipe. It is re-commended that random measurements be conductedon acidified plus rinsed sample splits in addition to un-treated material, in order to better evaluate the potentialinfluence of ammonium on d15N and Corg/N.

Acknowledgements

The authors thank Ewgenija Kuhl, Nina Holzner and Michelle Mohrfor sample preparation and taking measurements. We are furthermoreindebted to Uwe Reimold as well as Cameron Paul for proof-reading.Special thanks to the reviewers and Volker Br�chert for constructivecriticism, helpful discussions and advice. The research was funded bythe German Research Foundation (DFG) in the frame of the projectSTR356/3.

References

Altabet, M. A. & Francois, R. 1994. Sedimentary nitrogen isotopicratio as a recorder for surface ocean nitrate utilization. – GlobalBiogeochemical Cycles 8: 103–116.

Bakun, A. & Weeks, S. J. 2004. Greenhouse gas buildup, sardines,submarine eruptions and the possibility of abrupt degradation ofintense marine upwelling ecosystems. – Ecology Letters 7:1015–1023.

Ben-Yaakov, S. 1973. PH Buffering of Pore Water of Recent AnoxicMarine Sediments. – Limnology and Oceanography 18 (1): 86–94.

Bouillon, S., Raman, A. V., Dauby, P. & Dechairs, F. 2002. Carbonand nitrogen isotope ratios of subtidal benthic invertebrates in anestuarine mangrove ecosystem (Andhra Pradesh, India). – Estuar-ine, Coastal and Shelf Science 54: 901–913.

Bremner, J. M. & Willis, J. P. 1993. Mineralogy and geochemistry ofthe clay fraction of sediments from the Namibian continental mar-gin and the adjacent hinterland. – Marine Geology 115: 85–116.

Brunelle, J. P. 1978. Preparation of catalysts by metallic complex ad-sorption on mineral oxides. – Pure and Applied Chemistry 50:1211–1229.

Br�chert, V., Altenbach, A. V., Bening, G., Bockelmann, F., Currie, B.,Donath, J., D�becke, J., Endler, R., Erdmann, S., Ertan, T., Fuchs, B.,Klockgether, G., Kr�ger, S., Kuypers, M. M. M., Lass, H. U.,Lavik, G., Lilienthal, S., Leipe, T., Nickel, G., Noli-Peard, K.,Ohde, T., Schulz, B., Schulz, H., Siegel, H., Struck, U., Wulf, J.,Zitzmann, S. & Zonneveld, K. 2004. The Benguela UpwellingSystem 2003, Part 3, Cruise No. 57, Leg 3, March 15 – April 13,2003, Walvis Bay, Namibia – Dakar. – Meteor Berichte 4: 1–53.

Bunn, S. E., Loneragan, N. R. & Kempster, M. A. 1995. Effects ofacid washing on stable isotopes ratios of C and N in penaeidshrimp and seagrass: implications for food-web studies using mul-tiple stable isotopes. – Limnology and Oceanography 40 (3):622–625.

Carabel, S., God�nez-Dom�nguez, E., Ver�simo, P., Fern�ndez, L. &Freire, J. 2006. An assessment of sample processing methods forstable isotope analyses of marine food webs. – Journal of Experi-mental Marine Biology and Ecology 336: 254–261.

Emeis, K.-C., Br�chert, V., Currie, B., Endler, R., Ferdelman, T.,Kiessling, A., Leipe, T., Noli-Peard, K., Struck, U. & Vogt, T.2004. Shallow gas in shelf sediments of the Namibian coastal up-welling ecosystem. – Continental Shelf Research 24 (6): 627–642.

Faure, G. 1992. Principles and applications of inorganic geochemistry– a comprehensive textbook for geology students. Maxwell Mac-millan International Editions, New York, Toronto, Oxford.

Frank, D. A., Evans, R. D. & Tracy, B. F. 2004. The role of ammoniavolatilization in controlling the natural 15N abundance of a grazedgrassland. – Biogeochemistry 68: 169–178.

Fry, B. 1988. Food web structure on Georges Bank from stable C, N,and S isotopic compositions. – Limnology and Oceanography 33:1182–1190.

Gaebler, O. H., Vitti, T. G. & Vukmirovich, R. 1966. Isotope effectsin metabolism of 14N and 15N from unlabeled dietary proteins. –Canadian Journal of Biochemistry 44: 1249–1257.

Goering, J., Alexander, V. & Haubenstock, N. 1990. Seasonal variabil-ity of stable carbon and nitrogen isotope ratios of organism in aNorth Pacific Bay. – Estuarine, Coastal and Shelf Science 30:239–260.

Graml, M. 2001. Sedimentfazielle- und isotopenanalytische Untersu-chungen an rezenten Schelfsedimenten im Auftriebsgebiet desS�d-Ost-Atlantiks vor Namibia. Diploma Thesis. Institut f�r Pa-l�ontologie und Historische Geologie, Ludwig-Maximilians-Uni-versit�t, Munich: 63 pp.

Holmes, M. E., M�ller, P. J., Schneider, R. R., Segl, M. & Wefer, G.1998. Spatial variations in euphotic zone nitrate utilization basedon d15N in surface sediments. – Geo-Marine Letters 18 (1): 58–65.

Holmes, B., Eichner, C., Struck, U. & Wefer, G. 1999. Reconstruc-tions of surface ocean nitrate utilization using stable nitrogen iso-topes in sinking particles and sediments. In Fischer, G. & Wefer,G. (eds). Use of Proxies in Paleoceanography: Examples from theSouth Atlantic. Springer-Verlag, Berlin: pp. 447–468.

Jacob, U., Mintenbeck, K., Brey, T., Knust, R. & Beyer, K. 2005.Stable isotope food web studies: a case for standardized sampletreatment. – Marine Ecology Progress Series 287: 251–253.

Kang, C.-K., Kim, B. J., Lee, K.-S., Kim, J. B., Lee, P.-Y. & Hong, J.-S.2003. Trophic importance of benthic microalgae to macrozoo-benthos in coastal bay systems in Korea: dual stable C and N iso-tope analyses. – Marine Ecology Progress Series 259: 79–92.

Kennedy, P., Kennedy, H. & Papadimitriou, S. 2005. The effect of acid-ification on the determination of organic carbon, total nitrogen andtheir stable isotopic composition in algae and marine sediment. –Rapid Communications in Mass Spectrometry 19: 1063–1068.

Lehmann, M. F., Bernasconi, S. M., Barbieri, A. & McKenzie, J. A.2002. Preservation of organic matter and alteration of its carbonand nitrogen isotope composition during simulated and in situearly sedimentary diagenesis. – Geochimica et CosmochimicaActa 66: 3573–3584.

Morse, J. W. & Morin, J. 2005. Ammonium interaction with coastalmarine sediments: influence of redox conditions on K*. – MarineChemistry 95: 107–112.

Melander, L. 1960. Isotope effects on reaction rates. Ronald Press,New York.

Montoya, J. P. 1994. Nitrogen isotope fractionation in the modernocean: Implications for the sedimentary record. In Zahn, R., Pe-dersen, T. F., Kaminski, M. A. & Labeyrie, L. (eds). Carbon Cy-cling in the Glacial Ocean: Constraints on the Oceans Role inGlobal Change 17, NATO ASI Series. Springer-Verlag, Berlin:pp. 259–279.

M�ller, P. J. 1977. C/N ratios in Pacific deep-sea sediments: Effect ofinorganic ammonium and organic nitrogen compounds sorbed byclays. – Geochimica et Cosmochimica Acta 41 (6): 765–776.

Ng, J. S. S., Wai, T.-C. & Williams, G. A. 2007. The effects of acidi-fication on the stable isotope signatures of marine algae and mol-luscs. – Marine Chemistry 103: 97–102.

Nieuwenhuize, J., Maas, Y. E. M. & Middelburg, J. J. 1994. Rapidanalysis of organic carbon and nitrogen in particulate materials. –Marine Chemistry 45: 217–224.

Ostrom, N. E., Macko, S. A., Deibel, D. & Thompson, R. J. 1997.Seasonal variation in the stable carbon and nitrogen isotope bio-geochemistry of a coastal cold ocean environment. – Geochimicaet Cosmochimica Acta 61 (14): 2929–2942.

Fossil Record 14 (2) 2011, 141–152 151


Pichevin, L., Martinez, P., Bertrand, P., Schneider, R., Giraudeau, J. &Emeis, K. 2005. Nitrogen cycling on the Namibian shelf andslope over the last two climatic cycles: Local and global forcings.– Paleoceanography 20: 2.

Rosenfeld, J. K. 1979. Ammonium adsorption in nearshore anoxic se-diments. – Limnology and Oceanography 24 (2): 356–364.

Rosenfeld, J. K. 1981. Nitrogen diagenesis in Long Island Sound sedi-ments. – American Journal of Sciences 281: 436–462.

Sachs, J. P. & Repeta, D. J. 1999. Oligotrophy and nitrogen fixationduring eastern Mediterranean sapropel events. – Science 286:2485–2488.

Saino, T. & Hattori, A. 1980. 15N natural abundance in oceanic sus-pended particulate matter. – Nature 283: 752–754.

Sch�fer, P. & Ittekkot, V. 1993. Seasonal variability of d15N in settlingparticles in the Arabian Sea and its paleogeochemical signifi-cance. – Naturwissenschaften 80: 511–513.

Schubert, C. J. & Calvert, S. E. 2001. Nitrogen and carbon isotopiccomposition of marine and terrestrial organic matter in ArcticOcean sediments: implications for nutrient utilization and organicmatter composition. – Deep-Sea Research I 48: 789–810.

Shannon, L. V. & O’Toole, M. J. 1998. Integrated Overview of theOceanography and Environmental Variability of the BenguelaCurrent Region: Thematic Report No. 2, Synthesis and Assessmentof Information on BCLME. October 1998. UNDP, Windhoek.

Stoner, A. W. & Zimmerman, R. J. 1988. Food pathways associatedwith penaeid shrimps in a mangrove-fringed estuary. – FisheriesBulletin 86: 543–552.

Struck, U., Emeis, K.-C., Alheit, J., Schneider, R., Eichner, C. & Alten-bach, A.-V. 2002. Changes of the upwelling rates of nitrate pre-served in the d15N-signature of sediments and fish scales fromthe diatomaceous mud belt off Namibia. – GeoBios 35: 3–11.

Thornton, S. F. & McManus, J. 1994. Application of organic carbonand nitrogen stable isotope and C/N ratios as source indicators oforganic matter provenance in estuarine systems: evidence fromthe Tay Estuary, Scotland. – Estuarine, Coastal and Shelf Science38: 219–233.

Von Breymann, M. T. & Suess, E. 1988. Magnesium in the marinesedimentary environment: Mg–NH4 ion exchange. – ChemicalGeology 70: 359–371.

Vavilova, V. V. 1990. Marine biology: phytoplankton of the Benguelaupwelling system off Namibia in the austral summer of 1985. –Oceanology 30: 472–477.

Weeks, S. J., Currie, B., Bakun, A. & Peard, K. R. 2004. Hydrogensulphide eruptions in the Atlantic Ocean off southern Africa: im-plications of a new view based on SeaWiFS satellite imagery. –Deep-Sea Research I 51: 153–172.



the potential distortion of sedimentary d15n and c /n · pdf filethe potential distortion of...

Documents