Advances in agronomy, volume 133

In these conditions,the resonant absorption of the incident beam happens so close to the surface fluores-of the sample that all fluores-of thefluorescence photons are able to escape from th

A DVANCES IN

AGRONOMY

Texas A&M University

Emeritus Advisory Board Members

Newark, Delaware, USA

AMSTERDAM • BOSTON • HEIDELBERG • LONDON NEW YORK • OXFORD • PARIS • SAN DIEGO SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO Academic Press is an imprint of Elsevier

525 B Street, Suite 1800, San Diego, CA 92101-4495, USA

125 London Wall, London EC2Y 5AS, UK

The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK

First edition 2015

Copyright Ó 2015 Elsevier Inc All Rights Reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing As new research and experience broaden our understanding, changes in research methods, professional practices,

or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

ISBN: 978-0-12-803052-3

ISSN: 0065-2113

For information on all Academic Press publications

visit our website at http://store.elsevier.com/

Fernando Auat Cheein

Autonomous and Industrial Robotics Research Group (GRAI), Advanced Center of Electrical and Electronic Engineering (AC3E), Department of Electronic Engineering, Universidad Técnica Federico Santa María, Valparaíso, Chile

Canadian Light Source Inc., Saskatoon, SK, Canada; Department of Soil Science,

University of Saskatchewan, Saskatoon, SK, Canada

Kodigal A Gopinath

ICAR-Central Research Institute for Dryland Agriculture, Hyderabad, India

viij

Indian Council of Agricultural Research, New Delhi, India

Cherukumalli Srinivasa Rao

ICAR-Central Research Institute for Dryland Agriculture, Hyderabad, India

Matthias Stettler

Bern University of Applied Sciences, School of Agricultural, Forest & Food Sciences HAFL, Zollikofen, Switzerland

Jan J.H van den Akker

Alterra, Wageningen University and Research, Wageningen, The Netherlands

Volume 133 contains five first-rate reviews dealing with contemporarytopics important in the crop and soil sciences Chapter 1 is a comprehensivereview on the advances that have occurred in the use of synchrotron-basedsoft X-ray spectroscopy to study biogeochemical processes of importantlight elements such as carbon, nitrogen, and phosphorus in soils Chapter 2introduces a new hybrid land evaluation system to assess climate changeeffects on the suitability of an agricultural area for maize production.Chapter 3 is a timely review on progress in using structured light sensors

in precision agriculture and livestock farming Chapter 4 covers the potentialand challenges of rainfed farming in India including features of rainfedecosystems and rainfed crops and cropping systems Chapter 5 presents aDriver-Pressure-State-Impact-Response (DPSIR) analysis and risk assess-ment for soil compaction from a European perspective

I am grateful for the authors’ fine contributions

Donald L SparksNewark, Delaware, USA

xij

Advances in Using Soft X-Ray

Spectroscopy for Measurement

of Soil Biogeochemical Processes

Adam W Gillespie*,x,1, Courtney L Phillipsx, James J Dynes*,

David Chevrier*, Thomas Z Regier*and Derek Peakx

*Canadian Light Source Inc., Saskatoon, SK, Canada

xDepartment of Soil Science, University of Saskatchewan, Saskatoon, SK, Canada

1 Corresponding author: E-mail: adam.gillespie@lightsource.ca

Contents

3 Slew Scanning of Radiation-Sensitive Solids 12

4.1 Current Applications and Future Prospects 29

of atoms important in biogeochemical processes X-ray absorption spectroscopy (XAS) probes the local bonding and coordination environment of these elements in whole samples Bulk XAS techniques permit for high throughput, the study of whole soils, and high sampling density These analyses are complementary to X-ray transmission microscopy techniques which are limited by low throughput, thin particles ( <100 nm), and low sampling density In many projects, these bulk XAS measurements may be essential to understanding large-scale processes in soils such as the global

C cycle.

Despite these important applications, bulk soft XAS has not been extensively applied to environmental samples until recently The primary reasons for this gap is the lack of beamline endstations that are suitable for “dirty” samples and the technical challenges related to acquiring and normalizing spectra from dilute samples Many of these technical challenges have now been overcome through the development of Advances in Agronomy, Volume 133

ISSN 0065-2113

All rights reserved 1j

energy-resolving detectors, proper detector positioning, and development of liquid cell applications Technical developments and recent applications will be presented, showing how bulk soft X-ray XAS is now positioned to contribute signi ficantly to advancing the characterization of soils and environmental samples.

1 INTRODUCTION

The field of biogeochemistry describes how the biotic and abioticworlds interact More specifically, it studies the chemical and biochemicalprocesses in which elements are transformed and moved within and bet-ween living and nonliving parts of the ecosystem On a global scale, thecycles of major elements including carbon, nitrogen, oxygen, and iron arelinked with global climate change, food security, water security, and eco-nomic development (Godfray et al., 2010; Lal, 2009; McBratney et al.,

2014) Understanding the forms, pool sizes, and transformation rates of theseelements is critical to assessing their behavior on a global scale The availabil-ity and cycling of nutrients is determined by the interaction of physical,chemical, and biological processes in an ecosystem This interaction of pro-cesses is important as it determines the forms, transformations, and ultimatefate of nutrients in a given system Therefore, an assessment of their chem-istry is a key in determining how these elements will react, cycle, and persist

in the environment

Research that targets biogeochemical cycling in terrestrial ecosystems is akey component to understanding global systems Examples of biogeochem-ical processes include global biological processes of photosynthesis, nitrogenfixation, and soil organic matter decomposition, as well as critical redox pro-cesses of nitrification, denitrification, iron reduction, sulfate reduction, andthe myriad of transformations associated with phosphorus cycling Indeed,soil carbon pools are greater than both the atmospheric and biomass carbonpools combined (Lal, 2007; Stevenson and Cole, 1999), and soils contain themajority of reactive N as organic forms in soil organic matter (Olk, 2008;Stevenson, 1996) Soils also harbor an enormously diverse and dense micro-bial population which is ultimately responsible for organic matter transfor-mations The composition of soil organic matter and the composition ofthe soil microbial community are tightly coupled They participate in adynamic system where organic matter composition is constantly changing

in response to biological processes and, in turn, microbial communitycomposition changes in response to available substrates and nutrients In

addition, there is growing evidence describing the role of soil minerals asreactive surfaces that act as reaction catalysts, microbial niches, and adsor-bents/stabilizers of organic molecules.

Light or low atomic number (Z) elements are particularly important inbiogeochemical processes Light elements (C, N, O, P, and S) are the pri-mary atomic building blocks of all biological systems, and are implicated

in global climate change, acidification, and eutrophication Light basic ions (Na, Mg, K, Ca) are essential nutrients for plant growth, and mineralsare primarily composed of Al, Si, O, and Fe Finally, third row transitionmetals comprise essential micronutrients and contaminants of interest Anideal analytical approach for studying biogeochemical processes in environ-mental samples would therefore target the chemistry of these low Z ele-ments and 3rd row transition metals

cat-Among the tools available to probe biogeochemical processes, tron-based X-ray absorption spectroscopy (XAS) is an especially promisingchoice XAS is a technique that uses X-rays to probe the local electron struc-ture, thus elucidating the speciation and bonding environment of anelement Since the core electron energies of different elements are unique,specific elements can be probed in a complex sample by scanning the X-raybeam energy across the binding energies of the core electrons in theelements of interest Promotion of these core level electrons into unoccu-pied (i.e., valence) electron orbitals occurs when the incident X-ray energycoincides with an accessible transition energy, resulting in the resonantabsorption of the X-ray photons This resonant absorption results in a reduc-tion in the penetration depth of the incident beam Observations of subtlechanges in the resonant absorption energies can be used to differentiateoxidation states, bonding environments, and neighboring atoms

synchro-X-rays in the“soft” X-ray region (250e2500 eV, or wavelength 5 to0.6 nm) access the chemistry of lighter elements of biogeochemical interest.These include the K-edges (i.e., excitation of the 1s electrons) of elementsranging from C to Si and the L2,3edges (i.e., excitation of the 2p electrons)from K to Se (Figure 1)

Soft X-rays provide a unique way to probe the electronic structure of

C, N, and O at their K-edges in environmental samples, and is a powerfulcomplement to established and widely used NMR and IR spectroscopy.XAS offers advantages because it is sensitive to all isotopes of a particularelement, and is not limited by13C or15N abundances Indeed, for analysis

of N chemistry in environmental samples, XAS currently offers the only toolswhich are sensitive to most N-species and largely free of technique-induced

Figure 1 Biogeochemically signi ficant elements and core electron shells accessible by soft X-ray spectroscopy Black lines are K-edges (1s), red (light grey in print versions) lines are L-edges (2p), and blue (grey in print versions) line is N-edge (4f).

sample modifications It is also not necessary, to enrich organic matter centrations by separating it from the mineral fraction or to digest samples in

con-HF to remove minerals, paramagnetic materials, or other components forthe analysis to proceed

The soft X-ray region also accesses the L-edges of the transition metals.Forfirst row transition metals, soft X-ray absorption can result in the exci-tation of a 2p electron into unoccupied 3d orbitals In transition metals, theelectrons in the 3d orbital participate directly in ligandemetal complexation(de Groot and Kotani, 2008) Probing the L-edge of organometal com-plexes therefore provides direct information on the bonding and coordina-tion environment of those complexes The energy of the absorptionresonance can be used to determine the energy of the lowest unoccupiedmolecular orbitals (LUMOs) of the complex From ligandfield theory, allligands behave as either sigma donors, pi acceptors, or pi donors, and theircomplexation with a metal center will change the relative (splitting) andabsolute (ELUMO) positions of the LUMOs These effects can be measureddirectly by the intensity, position, and number of peaks in the L-edge XASspectrum of the metal-ligand complex

Soil biogeochemical processes occur at all levels of spatial variability,from the global scale to the nanoscale Indeed, the quality of biogeochemicalresearch depends not only on the techniques available, but also upon thesampling strategy, study location, and spatial resolution in many dimensions,including time XAS offers theflexibility to provide analytical information atmany spatial scales, and this depends on a particular beamline design Forexample, most ecological soil sampling is conducted to understand therole of spatial information such as landscape position/topography (soilcatenas) or vertical soil horizons (cores/sampling from pits) Bulk soft X-ray XAS measurements can probe aw1  0.1 mm area in 10e100 subsam-ples from such a sample set and provide quantitative speciation of an elementthat fits nicely into field research However, there may not always be aunique solution to bulk speciation in heterogeneous samples such as soils;having mm- or nm-scale spectromicroscopy capabilities with a soft X-raymicroprobe (10mm pixel) or scanning transmission X-ray microscope(STXM, 20e50 nm pixel) can provide additional information into the vari-ability of the sample and improve confidence in quantitative fittingapproaches In the example ofFigure 2, the bulk Fe L-edge X-ray measure-ments on the right were performed with a beam roughly 1 mm 0.1 mm,the microprobe image in the lower portion shows the distribution of Fe overthe same sample at 0.5 mm 0.5 mm scale with 5 mm resolution, and the

STXM map covers a 5mm  5 mm spot (roughly 1 pixel of the microprobe)

in detail sufficient to observe microbial colonization of defect sites on Fe(III)-rich particles

Synchrotron-based studies are advantageous because they require littlesample preparation before measurement There are, however, several chal-lenges that have limited the adoption of this technique for soil and environ-mental samples First of all, chemical speciation on environmental samplesshould ideally be conducted at ambient moisture and gas pressure/compo-sition, and at dilute concentrations The current widespread practice ofbulk soft X-ray spectroscopy requires samples to be under vacuum, whichrequire them to be dry and pulverized Next, elements in environmentalsamples are typically present at concentrations nearing detection limits andmodel systems with higher concentrations suffer from saturation effects

Finally, radiation damage is a significant and well-documented problem atsoft X-ray energies (Beetz and Jacobsen, 2003; Wang et al., 2009;Zubavichus et al., 2004) In this document, we describe solutions to thesechallenges which have been implemented at the bulk soft X-ray beamline,11ID-1 SGM (Spherical Grating Monochromator) at the Canadian LightSource (CLS) (Regier et al., 2007a,b) What follows describes the use ofenergy-resolving detectors at soft X-ray energies, the development of aflow-through liquid cell, the application of fast (slew) scanning, and thespectral acquisition and normalization procedures specific to carbonmeasurements.

2 DETECTOR ADVANCEMENTS

XAS is most often measured using transmission, total electron yield(TEY) by measuring the drain current through the sample (Gudat andKunz, 1972), or using total fluorescence yield (TFY) using either micro-channel plates or solid state diode detectors (Jaklevic et al., 1977) Transmis-sion methods are not generally suitable for bulk measurements in the softX-ray region because the attenuation of the photons at these energiesrequires samples to be uniform and less than 100 nm thick for measurements

at the C K-edge, and ca 2mm at the Si K-edge TEY and TFY are moreoften used because the analyst is not restricted by sample thickness TEYand TFY are measurements of secondary decay processes which operate

on the principle that, after an electron core hole is produced, the corehole will decay with the emission of either an Auger electron or afluores-cence photon As the incident X-ray energy is scanned through an absorp-tion resonance, the penetration depth is reduced and more of the secondaryelectrons orfluorescence photons are produced nearer to the surface of thesample and will exit the sample and make it to the detector Thus, underideal conditions, the measured intensity of the secondary emission will beproportional to the linear attenuation coefficient

Electron yield is a surface sensitive detection method, whereas TFY isconsidered a bulk-sensitive detection method Electrons are much largerthan photons, and have an escape depth of 10 nm before they are reabsorbed

by surrounding atoms Fluorescent photons however, have an escape depth

on the order of 100 nm (Frazer et al., 2003; Katsikini et al., 1997) Electronyield is also sensitive to the conductivity of the sample material Many envi-ronmental materials, including soil, contain some insulating components,

which will restrict the movement of charge through a sample, causingcharge buildup and release (sample charging) The measured current will

be dependent on exactly how the sample was mounted and what nents of the sample happen to rest on a conducting substrate material Fluo-rescence yield does not suffer from charging effects, making it more suitablefor the measurement of insulating sample materials

compo-TFY measurements, however, have several limitations These ments are commonly obtained using either microchannel plates or solid statediode detectors, both of which output a signal proportional to the powerdeposited in a detection medium As the TFY is comprised of both the reso-nant and nonresonant emission from the sample, it has two major draw-backs The first is that, for a dilute system (where saturation effects aresmall) the resonant emission from the target element is very weak compared

measure-to the nonresonant emission (background) from the matrix This candramatically reduce the concentration limits of the detection method Thesecond drawback is that the background emission from the matrix canexhibit an energy dependence at resonance due to a change in the penetra-tion depth of the incident photons (Achkar et al., 2011) This effect canresult in a severe reduction in fluorescence intensity that can create a sub-background in the pre-edge of some samples that depends on the concen-tration of the sample Finally, in grating-based beamlines, higher order light

is transmitted, and the TFY will contain spectral contributions fromelements excited by this light As a specific example, this is especially prob-lematic in C measurements of soil and environmental samples in whichoxygen-containing clay minerals are present Measurement at the C K-edge (at 285 eV) in natural systems will always have artifacts in TFY arisingfrom the C K-edge coincidence with the second order O K-edge(at 260 eV)

Another limitation offluorescence yield-based absorption spectroscopy

is saturation of the fluorescence signal due to overabsorption, commonlyreferred to as self-absorption (note that this term is somewhat misleading

as it suggests that the saturation arises due to the reabsorption of thecence signal from the sample when, in fact, this is not true) This type ofsaturation introduces large distortions on the measuredfluorescence spectrawhen the attenuation coefficient of the element of interest becomes compa-rable to the total attenuation coefficient of the sample In these conditions,the resonant absorption of the incident beam happens so close to the surface

fluores-of the sample that all fluores-of thefluorescence photons are able to escape from thesurface of the sample, resulting in a loss of contrast in the measured spectrum

These main complications with TFY measurements can be overcomewith the use of an energy dispersive detector, capable of resolving the reso-nant, nonresonant, and higher order emission from the sample This allowsfor the measurement of the partial fluorescence yields (PFY), which arefree from background distortions and much more sensitive than total yieldmethods (Figure 3) Silicon drift detectors (SDD) provide such energy-resolved detection, and have been used routinely in the hard X-ray regionfor many years More recently, improvements in thin-window manufacturinghas resulted in the extension of their use to soft X-ray detection.

On the SGM beamline at the CLS, the SDD’s are configured for twooperating modes: pulse counting mode and full spectrum acquisition Inpulse counting mode, a region of interest is specified in the detector hard-ware to electronically restrict the energy range of reported photons at thetime of data acquisition In this scenario, the detector reports only photonsarising fromfluorescence of a particular atom This setup can be optimizedfor faster counting because it does not require the complete readout of thedetector electronics at each step of the scan.Figure 3shows the X-rayfluo-rescence spectrum of a soil when excited by 2000 eV photons Thefluores-cence produced by single elements in the mixture can be isolated duringanalysis, offering the ability for rapid data acquisition and lowered detectionlimits

In full spectrum acquisition, the entire X-ray fluorescence spectrum isread out and recorded for each excitation energy across the absorptionedge being probed In these experiments, thefluorescence associated with

Figure 3 Fluorescence spectrum of a soil excited by 2000 eV photons as measured with

a silicon drift detector Detector output can be set to an emission energy range to isolate the fluorescence of a single element, allowing for rapid data acquisition with minimal background interference.

resonant atoms can be isolated to lower detection limits of dilute materials.

In the example illustrated inFigure 4, copper nitrate is present at mM centrations in a liquid sample, and TFY is unable to distinguish resonancesfrom the strong oxygen background (note that liquid cell constructionand spectroscopy will be detailed later in this work) By using an energy-resolved detector, the fluorescence from Cu can easily be extracted from

con-an excitation/emission map by applying a softwarefilter to the SDD output

In addition, using an energy discriminating detector, the inverse partialfluorescence yield (IPFY) method can be applied (Achkar et al., 2011).IPFY is a method of detection which is bulk sensitive and does not sufferfrom saturation effects This method consists of scanning across the absorp-tion edge of the element of interest, but monitors the change influorescenceintensity of another (spectator) element which is present in the mixture Oneimportant requirement for this method to work is that the spectator elementmust have a core hole binding energy below that of the element of interest.IPFY operates on the principle that the nonresonant emission will beinversely proportional to the total attenuation coefficient of the sample.For a full description and mathematical proof, see Achkar et al (2011).This method resolves the problem of saturation because it is based on the

Figure 4 Silicon drift detector output from a 100 mM copper (II) nitrate solution sample across the Cu L-edge Both the O and Cu emissions are labeled in the excitation- emission matrix (above) Analysis across the Cu edge includes strong emission line from O, which dominates the total fluorescence, shown in blue in the lower panel Copper spectrum (in red (black in print versions)) is accessible by isolating only the

Cu emission line.

nonresonant scattering intensity from the sample, it is directly proportional

to the linear attenuation coefficient as it does not contain distortions fromthe modulation of the resonant scattering cross section

IPFY is applicable for the measurement of iron and other transition metaloxides because oxygen is always present in environmental samples and canact as a suitable spectator This technique is especially suited to exploringthe chemistry of iron oxides and oxyhydroxides Peak and Regier (2012)

used IPFY to describe the structure of ferrihydrite (FeOOH), ultimatelyreporting a model in which ferrihydrite contained some tetrahedral ironwithin a structure of predominantly octahedral iron This experiment wasconducted at the Fe L2,3 edge, on solid state powdered samples, and theuse of energy-resolving detectors was critical to the success of the study

Figure 5shows results from the analysis of ferrihydrite at the Fe L-edge

Figure 5 Structure of ferrihydrite shown to contain some tetrahedral Fe as determined using inverse partial fluorescence yield Reprinted with permission from Peak and Regier (2012) Copyright 2012 American Chemical Society.

The Fe L-edge is excited above ca 705 eV These energies are above thebinding energy for O, and so strong fluorescence from O in the sample isdetected These two emission phenomena can be resolved with anenergy-resolving SDD, as seen in Figure 5 These correspond to the Fe

Laand Lbemission (625725 eV) and O Kaemission (450550 eV) lines.The importance of this emission resolution is evident when looking at theTFY inFigure 5 In this case, this plot is the sum of all Fe and Ofluorescence

at a particular excitation energy This is analogous to TFY as reported bymicrochannel plate or diode-type detectors The ability to separate thefluo-rescence from Fe and O shows that the totalfluorescence signal is a convo-lution of increasing Fefluorescence and decreasing O fluorescence Key tothe interpretation of this data is the principle that fluorescence from Fe isdistorted by saturation, whereas fluorescence from O is not The inverse

of the PFY of O in the sample is proportional to the total attenuationcoefficient, and is thus a high quality Fe spectrum that is free from saturationeffects

3 SLEW SCANNING OF RADIATION-SENSITIVE SOLIDS

Once purported as a nondestructive method (Vairavamurthy andWang, 2002), XANES studies have shown degradation and transformation

of some reference compounds through beam exposure over time,particularly when using high photon flux undulator beamlines (Cody

et al., 2009; Leinweber et al., 2007; Zubavichus et al., 2004) Theseproblems associated with radiation-induced decomposition also arecompounded at soft X-ray energies because X-ray absorption cross sectionsare higher compared to hard X-ray cross sections This increased crosssection manifests in a smaller radiation absorption volume concentrated

at the sample surface and results in a higher absorbed radiation dose.Second, for lower Z elements, the probability that a core hole will befilledthrough a nonradiative process (i.e., Auger electron emission) is muchhigher than the probability of radiative emission (Hubbell et al., 1994).Since the escape depth of electrons is very shallow, the energy of Augerelectrons is mostly deposited within the sample, whereasfluorescence pho-tons, have a longer escape depth, and have a lower probability of interact-ing with the sample before leaving Together, the increased X-rayabsorption cross section, strong probability of Auger electron processesover fluorescence, and the decreased likelihood that electrons will leave

a sample before depositing kinetic energy, all contribute to the problem ofradiation damage in samples at soft X-ray energies.

Radiation damage manifests itself in mass loss, chemical changes in thesample through the formation of free radicals, and reordering of bonding ar-rangements and changes to overall structural orders in crystalline and semi-conductor systems (Beetz and Jacobsen, 2003; Leontowich et al., 2012;Wang et al., 2009; Wilks et al., 2009; Zubavichus et al., 2004) Cryogenicshave been successfully employed to reduce radiation damage due to massloss in hard X-ray systems and are routinely employed for protein crystallog-raphy Mass loss, however, is not necessarily the main source of radiationdamage at soft X-ray energies, and studies evaluating the use of cryogeniccooling have not shown appreciable improvements in radiation-induceddamage artifacts over ambient temperature (Beetz and Jacobsen, 2003;Terzano et al., 2013)

A more viable approach in addressing problems of radiation-induceddamage is to reduce the dose incident on the sample The total absorbeddose rate (DR) for the top 10 nm of sample is estimated as mega Gray persec (MGy$s1, or 106J$kg1s1) using:

DR¼ ðF$EÞ=ðr$d$AÞwhere F is the incident photon flux, E is the photon energy, r is thematerial density, d is the depth, and A is the beam spot area One solution

to reducing the dose is to enlarge the beam spot size via defocusing, thusdistributing the photonflux throughout a larger area A second would be

to reduce the overall photonflux by attenuating the incident beam usingfilters or slits; this, however, negates the advantages afforded from thehigher flux of an undulator-based beamline A third way therefore toreduce the overall photon flux is to reduce the time required to record asingle scan Reducing the scan time on a sample can be accomplished byincreasing the photon use efficiency of a beamline This requires severalmodifications to the methods by which spectral data are normallycollected

Spectra can be acquired by operating the beamline in a slew scanningmode, which continuously scans the energy of the monochromator and ad-justs the gap of the beamline undulator while acquiring data These scans canvary in length, and typically are 5e20 s in duration This is in contrast to stepscan mode where an absorption spectrum is obtained across an absorptionedge by a stepwise movement of the monochromator and undulator to po-sitions which optimize the target photon energies before data collection

begins These scans take on the order of 7e10 min due to the time required

to move optical components and operate with a duty cycle (time spentacquiring data over the time that the sample is exposed to the beam) ofapproximately 50% In slew scan mode, the optics and detectors are runcontinuously, and data is streamed to a database buffer during the scan’scollection time Real-time monochromator encoder feedback is included

in the exported dataset so that after the scan is complete, the stream ofdata can be parsed into energy-resolved bins and output as a data file Forthis technique to be feasible, detector and encoder output must be acquiredrapidly and amplified cleanly In practice, the signal to noise ratio is low, andmany slew scans must be recorded and averaged for a single sample Tominimize radiation dose each scan is recorded on a fresh spot using a roboticsample manipulator Multiple slew scans are then combined, averaged, andnormalized Figure 6 shows a comparison between multiple summed andaveraged slew scans (n¼ 60) and a single step scan of 10% citric acid

Figure 6 Comparison of a single 10 min step scan to the average of 60 slew scans of

20 s on fresh sample spots of ca 10% citric acid in Al2O3 Radiation-induced artifacts are evident in the step scan.

(w/w) in Al2O3 Radiation-induced artifacts are clearly present in the stepscan spectrum, as identified with arrows.

An important question in evaluating the effectiveness of slew scanning is theneed to determine if indeed radiation damage is occurring before the comple-tion of thefirst scan This can be evaluated empirically and theoretically.Empirically, radiation damage can be assessed by measuring a candidatecompound repeatedly at a single spot without moving, and evaluating anytransformation or change in the spectra with respect to time.Figure 7shows

a time series scan of glycine (an amino acid) and bovine serum albumin (aprotein) at the N K-edge Proteins and peptides are known to be susceptible

to radiation damage, however, a single 20 s slew scan does not produce anydiscernible artifact peaks

Theoretical calculations using density functional theory of model pounds can be compared to experimental measurements to ascertainwhether features are present experimentally which are not predicted theo-retically Figure 8shows the calculated X-ray transition intensities of citricacid obtained using the StoBe density functional theory package alongwith the slew scan X-ray absorption measurement of citric acid Individualtransition intensities for the nonequivalent carbon sites on the citric acidmolecule are plotted to illustrate that there is a set of low energy transitions(red and blue) arising from core level excitation of the atoms labeled 2 and 3

Figure 7 Radiation damage to glycine and albumin monitored at the N K-edge, as a function of increasing dose Each spectrum represents a single slew scan.

These transitions give rise to a pre-edge shoulder in the citric acid spectrumthat was previously believed to be due beam damage that occurred in theduration of the first scan The calculation was able to confirm that thispre-edge feature is not due to damage but is inherent to the molecule’sX-ray absorption spectrum.

3.1 CarbonK-Edge Analysis

Measurements at the C K-edge present a special set of challenges to theanalyst First, samples must be prepared on C-free substrates Second, carboncontamination and the transmission of higher order light in beamline mono-chromators complicate the measurement of the incidentflux (Io) during thescan and can result in normalized spectra that exhibit distortions The spec-trum becomes more sensitive to the normalization procedure at low carbonconcentrations Therefore, it is extremely important to use proper samplepreparation protocols and incident flux characterization for soils measure-ment where the C concentrations can be low

For X-ray analysis of most elements, samples can be simply mounted ondouble-sided conductive carbon tape Indeed, a C-derived substrate isunsuitable for conducting C analysis on any sample Instead, samples aremounted on gold-coated silicon wafers In this procedure, gold pellets areheated above 650C in a vacuum chamber containing the silicon wafers.

Figure 8 Calculated electronic transition intensities for citric acid shown with the measured XAS The inset diagram and legend indicate the speci fic atoms giving rise

to each transition.

Sample material is then drop-coated onto the wafers by dissolving or rying in water, placing 8e10 mL of the mixture on the wafer, and allowing

slur-to dry (Figure 9)

The incident flux of X-ray photons on a sample is modulated by Ccontamination of the beamline optics This C contamination, which arisesdue to the beam-induced adsorption of residual hydrocarbon gases in thebeamline vacuum, will therefore introduce its own structure into the mea-surement (Figure 10) Ideally, the adventitious carbon would be removed

Figure 9 Sample preparation procedure for measurement at the C K-edge Soil is weighed into microcentrifuge tube, slurried in water, and evaporated onto a carbon- free substrate On site preparation of gold-coated silica is a suitable substrate for this measurement.

Figure 10 Lineshape of the incident beamline flux across the C K-edge Measurement

is of light scattered by an Au-coated silicon wafer and a silicon drift detector.

through cleaning of the beamline optics In practice, there are often cations or uncertainties related to both in situ and ex situ cleaning proce-dures that may prevent their application Also, even once optics arecleaned, they can quickly start to become contaminated again Therefore,most beamlines will exhibit some level of carbon contamination on theoptics.

compli-The line shape of the structure introduced by C contamination is theconvoluted transmission spectrum of the contaminants on each of the beam-line optics This spectral intensity is manifest on each measured X-rayabsorption spectrum and must be accurately measured and divided outfrom the measured signal to produce a normalized spectrum For soft X-ray experiments, this is typically accomplished by concurrent measurement

of the total drain current from a gold mesh-placed upstream of the sample.Alternatively, the incidentflux across an edge can be measured either before

or after the sample measurement using the photon-induced current from aphotodiode While these methods are satisfactory for XAS measurements ofmost elements, they are not adequate for measurements at the C K-edge Aumesh measurements of the Io are affected by contamination of the meshitselfdcarbon buildup on the mesh affects the drain current measured onthe mesh

The diffraction gratings used in soft X-ray monochromators are prone

to the transmission of higher order light This means that along with thedesired first order radiation the incident beam will also contain someamount of light with two and three (and higher) times the energy of thefirst order Different strategies are used to limit the amount of higher orderlight transmitted by a monochromator such asfilters or harmonic rejectionmirrors But significant reduction of the higher orders will also attenuatethe intensity of thefirst order light and complete rejection of higher orders

is not possible

The presence of the second (and higher) order light in the incidentbeam presents a significant problem with respect to the accurate measure-ment of the incidentflux The response of the drain current from a cleangold mesh as well as that of a photodiode are both dependent on photonenergy, so the high order light will contribute significantly to the signalmeasured Therefore, the resulting line shape measured by these devicesacross the C K-edge is a convolution of the signal from the significantlychangingfirst order light (due to reduction in flux by the carbon contam-ination on the optics) and the gradually changing higher order light Accu-rate normalization requires that the intensity of only thefirst order light be

known, so mesh drain current and photodiode current are not suitable formeasuring Io at the C K-edge.

These complications can be avoided by measuring the incidentflux with

an energy-resolving detector such as a SDD, which is able to discriminatebetween photons offirst order and higher orders SDDs become saturated

by high incident photonfluxes so they are not able to accept the full photonflux from this beamline To reduce the incident count rate on the detector

to an acceptable level, the elastic scattering or reflected beam intensity from acarbon-free substrate can be measured For a reflectivity measurement, a C-free mirror can be prepared by in situ annealing of a Si wafer The intensity

of the reflection can then be measured across the C K-edge and adjusted forthe reflectivity of Si to obtain an accurate Io

The intensity of the scattering of the incident beam from a freshly tered (carbon free) Au surface can also be used to determine Io The scat-tering cross section for Au has no features in the energy region aroundthe C K-edge and shows only a slowly increasing intensity with photonenergy Therefore, the measured scattering intensity from a clean Au surfacecan be used to determine the incidentflux profile very accurately Becausethe absolute cross section for the scattering is quite low, several hundredscans must be averaged together to obtain proper statistics It is also impor-tant to note that the scattering cross section drops off with the distance be-tween the incident beam and the detector so the latter must be placed close

sput-to the incident beam sput-to get sufficient counts

Higher order light can also be problematic when trying to measureaccurate carbon PFY from soils and sediments Aluminosilicate clayscontain appreciable amounts of crystalline oxygen; indeed, O is the mostabundant element in the earth’s crust (Sparks, 2003) The energy of thesecond order light for carbon measurements (ca 280 eV * 2¼ 560 eV) iscoincident with the O K-edge extended X-ray absorptionfine structure(EXAFS) Because of the high O content of clays, there is the potentialfor a strong background from the second order structure from O in the

C K-edge spectrum Fortunately, this problem is easily overcome using aSDD for measurement of the PFY, and by using Ti or Vfilters to attenuatethe signal from second order O with respect to C The use of a Tifilter inline with the incident beam will significantly improve the performance ofthe SDD because the Ti L-edge occurs at ca 450 eV, effectively attenu-ating second order light which would otherwise excite O atoms(Figure 11) Indeed, the 200 nm thick Ti filter offers ca 80% attenuation

of light at the O K-edge compared to 50% at the C K-edge

The use offilters for the reduction of higher order content in the dent beam is also important due to the possibility of C excitation by thislight While the probability of a higher energy photon exciting C is smallcompared to the probability of a first order photon exciting C; there will

inci-be some C fluorescence that arises due to these higher orders This willintroduce a background to the PFY measurement that will have to be sub-tracted before normalization Fortunately, the use offilters is able to reducethis higher order excitation to negligible levels

A further complication encountered influorescence-yield measurements

at the C K-edge is the background resulting from the elastic scattering ofphotons Scattered photons track the energy of the incident beam andartificially increase the counts in a PFY measurement as they cannot bedistinguished from fluorescence using the coarse energy resolution of anSDD This affects the normalization procedure because it adds an energy-dependent offset to the measured spectrum that cannot be unambiguouslyisolated and subtracted To reduce this scattering background, the C K-edge measurements of the sample can be made with the detector at the

90 horizontal position with respect to the incident beam (for horizontally

polarized sources) This concept takes advantage of the polarized nature ofthe synchrotron beam because scattering is forbidden along the orientation

of the incident beam polarization This careful detector placement canalmost completely eliminate the contribution of elastic scattering to thePFY spectrum Fluorescence arising from K-edges (p to s transitions), how-ever, propagates outward spherically from the excited atom Detector andFigure 11 Attenuation curve of a 0.2 mm Ti filter Filter used to attenuate second order light, reducing O fluorescence which is coincident with C K-edge fluorescence.

sample geometry can therefore be used to minimize the contribution of tering in C K-edge PFY measurements (Figure 12), which is critical foraccurate measurements, especially for samples with low C concentrations.

scat-In general, the proper normalization for C K-edge measurementsrequires a sample measurement that has minimal scatter, little or no back-ground fluorescence from other elements in the sample, and an accuratemeasurement of beamline flux for the Io In practice, it is not possible tocompletely eliminate scattering from a measurement due to the finite size

of the detector elements Also, as the concentration of the element in thesample decreases the portion of the detected signal that is due to scatteringwill increase Moreover, it is only possible to place a single detector element

in the optimal position, limiting the utility of multielement detector systemsand restricting count rates In order to handle the situation of nonzero scat-tering, the scattering has to be subtracted from the measurement

The XAS spectrum of the sample is calculated as:

XAS¼ IPFY=Io

where IPFYis the intensity of the PFY of carbon, and Iois the beamlineflux.Consider also that the intensity measured by the SDD, ISDD, is the sum of

Figure 12 Detector is placed at 90 with respect to the incident beam Fluorescence

(yellow (white in print versions)) propagates in all directions from the sample, whereas scattering (blue (grey in print versions)) approaches zero at 90 due to the horizontally

polarized beam.

the IPFY,the elastic scattering (ISc), Cfluorescence from higher order light(IPFY(2)), and detector noise:

ISDD¼ IPFYþ IScþ IPFY ð2Þþ noise:

If detector geometry is optimized (seeFigure 12), then IScz 0, and ifhigher order light is attenuated using a Ti filter, then IPFY(2) also z0 Ifthe Io is measured accurately with no background, the absorption (ABS) isgiven by

ABS¼ IPFY=Io¼ ðISDD AÞ=Io;where A is a scaling factor that is from higher order light exciting C andsome detector noise The scaling factor can be determined empirically byextracting the average counts measured from the C K-edge pre-edge region(i.e., below 284 eV) Since by definition, it is known that below theabsorption resonance IPFY should be exactly zerodthere should be nofluorescence occurring for excitation below threshold

Finally, converting the detector output to absorption

IPFY¼ ðISDD AÞ=Io

This procedure is illustrated inFigure 13

This normalization procedure, when used in conjunction with properdetector geometry, beamline flux measurement, and beamline filters toattenuate higher order light, should provide the correct absorption spectrum

of dilute C at the C K-edge in biogeochemical samples which containappreciable amounts of oxygen

4 SOFT X-RAY LIQUID CELLS

While soft X-ray measurements of powder samples greatly expand theutility of synchrotron-based XANES for light elements, there are alwaysconcerns associated with studying environmental samples without waterpresent In natural systems, the presence of water dramatically affects themetal coordination environment for metaleligand complexes and thesolidewater interface in soils represents the primary site of chemicalreactions (Sparks, 2003) In addition, it is known that solution conditions(pH, ionic strength, analyte concentration) can change the structure andaggregation of humic and fulvic acids (Myneni et al., 1999) As an example

of the effect of liquid versus powder measurements, compare the Cu L-edge

XANES spectra of two Cu-ligand compounds inFigure 14 In the case ofCu-acetate, it is clear that traditional powder XAS measurements result inphoto-reduction and artifacts However, even in the case of Cu(NO3)2

where additional Cu(I) and Cu(0) peaks are not present, the edge position

is nonetheless shifted and distorted in the powder sample compared to theaqueous Cu solution Thus, the measurement of many environmentalsamples in vacuum can lead to results that may not be representative ofthe true chemical environment that is present in nature Accordingly, softX-ray absorption studies on liquids have been performed at many synchro-tron sources around the world using a range of sampling approaches

60 single scans averaged together PFY by de finition should show zero counts in the pre-edge region; however, some counts are registered because of detector noise and from higher order light exciting C This value (A) is subtracted from the averaged spec- trum before normalization (panel B) Normalization to the beamline flux is then carried out using the spectrum of light scattered from a C-free Au-coated Si substrate (panel C), producing the resultant XAS spectrum (panel D).

Figure 14 Photo-reduction of Cu-ligand complexes by soft X-ray irradiation observed in the solid state (dotted lines) compared to liquid state measurements in a continuous flow cell.

One method for measurements of liquids is the microjet in which water

is sprayed through the vacuum chamber at a high pressure The main itation with this apparatus is the dependence on the vapor pressure of theliquid Because the liquid sample is entering a low atmospheric chamber,vaporization happens rapidly Inevitably, vapor particles become part ofthe spectrum causing inaccuracies in the data (Ukai et al., 2008, 2009;Winter and Faubel, 2006) Many reviews for this type of liquid cell identifythe limitations for use in environmental studies; they are not adapted toperform XAS, they are limited to the range of solutions that can be studied,and they are not able to study large organic molecules (Winter, 2009;Winter et al., 2008; Winter and Faubel, 2006) More recently, static cellshave emerged that allow for the sample to be isolated from the vacuumenvironment (Hay and Myneni, 2010) The main advantage of this type

lim-of apparatus is the ability to use a variety lim-of samples and solvents, since vaporpressure is no longer of concern However, these static cells remain suscep-tible to beam-induced damage because the sample is not refreshed duringanalysis Also, controlling sample conditions (i.e., pH, temperature, and con-centration) is difficult within a static cell as it completely contained in thevacuum system For these reasons,flow cells have become popular, and staticcells are less common

Another approach to measuring liquids used bySiegbahn and Sodergren(1995)included a rotating disk in a box of liquid that created a thinfilm ofliquid on the disk The design’s purpose was for better control of sampleconditions and the elimination of gas phase contributions to the spectrum.This same concept was applied to XAS measurements of liquid water in

2002 where water was passed over a metal plate and measured withcence yield (Myneni, 2002) Since then, there have been several differentflow cells designed and built at several synchrotrons and used with varyingdegrees of success

fluores-Flow cells have a distinct advantage over the static cell, which is theability to control the sample environment With aflowing cell, the samplesolution can be present outside of the vacuum chamber allowing for theexternal monitoring and adjustment of the sample if needed It also provides

a natural temperature control since the exposed liquid, which has enced an increase in temperature from interaction with the beam, is replaced

experi-by new sample liquid at the proper temperature (Fuchs et al., 2008)

At the CLS SGM beamline, we have designed and implemented a liquidflow cell with a 100 nm thick Si3N4window with a 1 1 mm clear aperture(SPI supplies, Pennsylvania) mounted on a cell body printed from a 3D

photopolymer (http://www.stratasys.com/materials/polyjet) Two miniaturesolenoid valves (Bio-Chem Fluidics, New Jersey) were installed inline beforeand after the cell and interlocked to the chamber pressure sensors to prevent alarge volume offluid from entering the vacuum system upon window failure.Tubing and fittings were constructed from polytetrafluoroethylene, anonreactive material The cell was positioned in the SGM absorptionendstation and samples were pumped through using a peristaltic pump Solu-tions were contained in a small beaker outside of theflow cell and vacuumchamber, where both the inlet and outlet lines allowed for recirculation ofthe sample Sample volume was just large enough (w30 mL) to allow forthe insertion of a pH meter for monitoring Continuous mixing via a stir-bar, sparging to eliminate dissolved CO2via Ar or N2gas, measurement ofoxidation-reduction potential via sensor, and other control/instrumentationare all performed in the reaction vessel A figure illustrating the liquid celldesign is shown inFigure 15.

A recent publication on Cueligand complexation by Phillips et al.(2013) using the SGM liquid cell is a good case study for how liquid L-edge measurements have great potential for environmental biogeochemicalresearch Copper (as Cu2þ) is a trace micronutrient that is essential for life atlow levels but can become toxic at higher levels In natural ecosystems, boththe toxicity and mobility of Cu2þare strongly affected by its chemical speci-ation; Cu is often found in nature complexed with organic materials (e.g.,dissolved natural organic matter, plant root exudates, and soil humic

Figure 15 Liquid flow cell construction is compatible with vacuum system endstation Silicon nitride window embedded in printed polymer cell with integral flow paths, connected to a peristaltic pump and beamside wet chemistry station.

substances) (Thurman, 1985) These organic ligands are all macromoleculeswith varied molecular weights, acid/base solubility, and poorly characterizedstructures, but it is known that they all contain carboxylate (ReCOO),alcohol/phenol (ReOH), and amide (ReNH2) functional groups Despiteseveral spectroscopic studies of Cu-organic bonding with hard X-rays,disagreement exists in the literature (Karlsson et al., 2006; Manceau andMatynia, 2010) about the importance of N-bearing amide groups for copperbonding in natural systems.

Cu L-edge XANES arise from a 2p to 3d electron transition Since the3d orbitals in Cu participate in the Cu-organic bond, this transition acts as aprobe into the Cu-organic molecular bond Also, it is well known thatorganic ligands affect the degeneracy and energy position of metal d-orbitals(Atkins et al., 2010), thus a change in the Cu L-edge should occur withchanges in the organic ligand involved in the complex The researchapproach in this study was to probe a range of increasingly complex Cu-organic aqueous complexes at the Cu L-edge A range of carboxylate,alcohol, and amide ligands were systematically chosen to explore the effects

of complex structure (linear versus ring complexes), number of functionalgroups on the ligand, and functional group identity (i.e., carboxyl, alcohol,and amine) on the ligand field Cu L-edge XANES were collected at theSGM beamline at the CLS using a SDD array and the custom liquidflowcell in Figure 15

The energy position of the Cu L3-edge peak has a positive shift when

Cu is bound to an organic ligand (Figure 16) The relative position of theCu-organic peak to the Cu2þ(aq)peak is dependent on the type of functionalgroups involved in the complex, the pH, and the orientation of those groups

to Cu The effect of the functional group mirrors that of a spectrochemicalseries, that is, the positive shift increases in the order ReCOO<ReOH < ReNH2 thereby providing a convenient method to discrimi-nate N bonding in the system There is also a higher positive shift withincreasing pH in each system, and the shift has a linear correlation to the

pKa of the ligand involved (R2¼ 93%) This indicates that Cu-organicbonding is related to the Lewis acid/base properties of the ligand; specif-ically, it reflects the Lewis base properties of the O and N atoms that arecomplexed with Cu Finally, the orientation of these groups to Cu can beeither on one of two locations on the Cu octahedron; the axial or the equa-torial planes Due to JahneTeller distortion of the Cu octahedron, thesepositions are not equal, and instead the location of ligand bonding to Cuhas implications on sterics and orbital overlap, which affect the observed

L-edge energy position This is best demonstrated with the anomalous case

of Cu-EDTA Normally thought of as a hexadentate coordination wheretwo N ligands coordinate on the axial plane, this complex should give rise

to a large positive L-edge shift, yet it does not Complementary FTIR measurements indicated a pentadentate ligand coordination withone axial water, and N groups on different planes, which interferes withthe electronic orbital overlap making the shift smaller than predicted As

ATR-pH is lowered and two carboxyl groups are protonated, the complexchanges conformation and the axial N shifts to the equatorial plane and con-tributes electron density more strongly This is most likely why the L-edgeremains relatively stable as the pH is lowered (Figure 16), unlike other Cu-organic ligands that have a smaller shift

Three major conclusions could be drawn from this research First of all,

Cu complexation and electronic properties correlate quite strongly to theLewis acid/base properties of the coordinated ligand functional groups

Figure 16 (A) Cu L3-edge XANES of Cu-ligand complexes and the resulting tion environments (center) (B) Effect of changing pH on Cu L3-XANES of Cu-ligand com- plexes Adapted with permission from Phillips et al (2013) Copyright 2013 American Chemical Society.

coordina-This is an important finding, because previously the stability constants forcomplex formation were used to evaluate the strength of coppereligandcomplexes, but these do not correlate to the electronic structure (L-edgeshift) of the complex Next, our complementary ATR-FTIR studiesconfirmed the importance of water molecules in the structures of aqueouscomplexes, afinding that stresses the importance of performing in situ mea-surements Finally, as expected, Cu L-edge XANES is quite sensitive toCueN coordination even with ligands that have mixtures of amide andcarboxyl groups, and we observed direct CueN bonding in all of ouramide-bearing ligands Furthermore, there is a large difference in XANESfor Cu bonded with primary versus tertiary amines This suggests that naturalorganic matter samples should be characterized not just for elemental C:N:Ocontent but also for the type of N groups that are present.

4.1 Current Applications and Future Prospects

Advances in soft X-ray spectroscopy have provided new opportunities forunderstanding the dynamics of many elements of interest to biogeochemists,including C, the major nutrient anions and cations, and the transition metals.These advances have been made possible through the use of new SDD,improved endstation design, slew scanning, and in situ liquid cell develop-ment Recently, several studies have shown the ecological utility of softX-ray spectroscopy Gillespie et al (2014b) showed how C chemistry isaffected in subarctic soils by burial through frost action, and found thatburied high C soils are resistant to decomposition because of modifications

to C chemistry In fertility studies, the composition of soil C and N has beenfound to change depending on the type of N source used as a fertilizer(Gillespie et al., 2014a) It was found that long-term manure applicationproduced C and N which were chemically stabilized compared to fertiliza-tion with synthetics or in a legume rotation Finally, Phillips et al (2013)

used a liquid cell analysis for a fundamental description of coppereligandchemistry In this study, they showed that LUMO energy is governed bythe ligandfield strength and is related to Lewis acid/base properties of theligand functional groups This research approach is now being applied tospectroscopically tracking Cu dynamics in environmental systems

The complexity of biogeochemical systems research requires a mensional and multidisciplinary approach Indeed, it is common and advis-able for environmental scientists to employ a range of analytical techniques

multidi-to answer questions of biogeochemical importance Soft X-ray spectroscopycontinues to develop as an environmental technique and advancements in

the types of experiments, sample types, and spatial resolutions will continue

to offer biogeochemists, the tools needed to better understand the istry of natural systems over a wide range of time and spatial scales

biochem-REFERENCES

Achkar, A.J., Regier, T.Z., Wadati, H., Kim, Y.-J., Zhang, H., Hawthorn, D.G., 2011 Bulk sensitive x-ray absorption spectroscopy free of self-absorption effects Phys Rev B e Condens Matter Mater Phys 83 (8) Available from: http://www.scopus.com/ inward/record.url?eid¼2-s2.0-79960981128&partnerID¼40&md5¼a27596f587b72d4 fed03c3421d299d43

Atkins, P., Overton, T., Rourke, J., Weller, M., Armstrong, F., 2010 Inorganic Chemistry Oxford University Press, Oxford, MA.

Beetz, T., Jacobsen, C., 2003 Soft X-ray radiation-damage studies in PMMA using a STXM J Synchrotron Radiat 10 (3), 280e283.

cryo-Cody, G.D., Brandes, J., Jacobsen, C., Wirick, S., 2009 Soft X-ray induced chemical fication of polysaccharides in vascular plant cell walls J Electron Spectrosc Relat Phenom 170 (1e3), 57e64.

modi-Frazer, B.H., Gilbert, B., Sonderegger, B.R., De Stasio, G., 2003 The probing depth of total electron yield in the sub-keV range: TEY-XAS and X-PEEM Surf Sci 537 (1e3), 161e167.

Fuchs, O., Maier, F., Weinhardt, L., Weigand, M., Blum, M., Zharnikov, M., Denlinger, J., Grunze, M., Heske, C., Umbach, E., 2008 A liquid flow cell to study the electronic structure of liquids with soft X-rays Nucl Instrum Methods Phys Res Sect A: Accel Spectrom Detect Assoc Equip 585 (3), 172e177.

Gillespie, A.W., Diochon, A., Ma, B.L., Morrison, M.J., Kellman, L., Walley, F.L., Regier, T.Z., Chevrier, D., Dynes, J.J., Gregorich, E.G., 2014a Nitrogen input quality changes the biochemical composition of soil organic matter stabilized in the fine fraction:

a long-term study Biogeochemistry 117, 337e350.

Gillespie, A.W., Sanei, H., Diochon, A., Ellert, B.H., Regier, T.Z., Chevrier, D., Dynes, J.J., Tarnocai, C., Gregorich, E.G., 2014b Perennially and annually frozen soil carbon differ

in their susceptibility to decomposition: analysis of Subarctic earth hummocks by bioassay, XANES and pyrolysis Soil Biol Biochem 68, 106e116.

Godfray, H.C.J., Beddington, J.R., Crute, I.R., Haddad, L., Lawrence, D., Muir, J.F., Pretty, J., Robinson, S., Thomas, S.M., Toulmin, C., 2010 Food security: the challenge

of feeding 9 billion people Science 327 (5967), 812e818.

de Groot, F., Kotani, A., 2008 Core Level Spectroscopy of Solids In: Advances in Condensed Matter Science CRC Press, Boca Raton, FL.

Gudat, W., Kunz, C., 1972 Close similarity between photoelectric yield and tion spectra in the soft-x-ray range Phys Rev Lett 29 (3), 169e172.

photoabsorp-Hay, M.B., Myneni, S.C.B., 2010 X-ray absorption spectroscopy of aqueous organic complexes J Phys Chem A 114 (20), 6138e6148.

aluminum-Hubbell, J.H., Trehan, P.N., Singh, N., Chand, B., Mehta, D., Garg, M.L., Garg, R.R., Singh, S., Puri, S., 1994 A review, Bibliography, and Tabulation of K, L, and higher atomic shell X-Ray fluorescence yields J Phys Chem Reference Data 23 (2), 339e364.

Jaklevic, J., Kirby, J.A., Klein, M.P., Robertson, A.S., Brown, G.S., Eisenberger, P., 1977 Fluorescence detection of exafs: Sensitivity enhancement for dilute species and thin films Solid State Commun 23 (9), 679e682.

Karlsson, T., Persson, P., Skyllberg, U., 2006 Complexation of copper(II) in organic soils and in dissolved organic matter e EXAFS evidence for chelate ring structures Environ Sci Technol 40 (80), 2623e2628.

Katsikini, M., Paloura, E.C., Fieber-Erdmann, M., Kalomiros, J., Moustakas, T.D., Amano, H., Akasaki, I., 1997 N K-edge X-ray-absorption study of heteroepitaxial GaN films Phys Rev B 56 (20), 13380e13386.

Lal, R., 2009 Soils and food sufficiency A review Agron Sustain Dev 29 (1), 113e133 Lal, R (Ed.), 2007 Encyclopedia of Soil Science CRC Press, Boca Raton, FL.

Leinweber, P., Kruse, J., Walley, F.L., Gillespie, A., Eckhardt, K.U., Blyth, R.I.R., Regier, T., 2007 Nitrogen K-edge XANES e an overview of reference compounds used to identify “unknown” organic nitrogen in environmental samples J Synchrotron Radiat 14, 500e511.

Leontowich, A.F.G., Hitchcock, A.P., Tyliszczak, T., Weigand, M., Wang, J., Karunakaran, C., 2012 Accurate dosimetry in scanning transmission X-ray microscopes via the cross-linking threshold dose of poly(methyl methacrylate) J Synchrotron Radiat.

mol-Myneni, S.C.B., Brown, J.T., Martinez, G.A., Meyer-Ilse, W., 1999 Imaging of humic stance macromolecular structures in water and soils Science 286 (5443), 1335e1337 Olk, D.C., 2008 Organic forms of soil nitrogen In: Schepers, J.S., Raun, W.R (Eds.), Nitrogen in Agricultural Systems American Society for Agronomy, Crop Science Society of America, Soil Science Society of America, Madison, WI, pp 57e100 Peak, D., Regier, T., 2012 Direct observation of tetrahedrally coordinated Fe(III) in ferrihydrite Environ Sci Technol 46 (6), 3163e3168.

sub-Phillips, C.L., Regier, T.Z., Peak, D., 2013 Aqueous Cu(II)-organic complexation studied

in situ using soft X-ray and vibrational spectroscopies Environ Sci Technol 47 (24), 14290e14297.

Regier, T., Krochak, J., Sham, T.K., Hu, Y.F., Thompson, J., Blyth, R.I.R., 2007a mance and capabilities of the Canadian Dragon: the SGM beamline at the Canadian Light Source Nucl Instrum Methods Phys Res Sect A: Accel Spectrom Detect Assoc Equip 582 (1), 93e95.

Perfor-Regier, T., Paulsen, J., Wright, G., Coulthard, I., Tan, K., Sham, T.K., Blyth, R.I.R., 2007b Commissioning of the Spherical Grating Monochromator soft X-ray spectroscopy beamline at the Canadian Light Source AIP Conf Proc 879 (1), 473e476.

Siegbahn, H.O.G., S €odergren, S., 1995 Structures at liquid surfaces studied by core electron spectroscopy J Electron Spectrosc Relat Phenom 76 (C), 417e420.

Sparks, D.L., 2003 Environmental Soil Chemistry, second ed Academic Press, San Diega, CA.

Stevenson, F.J., 1996 Nitrogen e organic forms In: Sparks, D.L (Ed.), Methods of Soil Analysis Part 3 Chemical Methods Soil Science Society of America, Madison, WI,

pp 1185e1200.

Stevenson, F.J., Cole, M.A., 1999 Cycles of Soil: Carbon, Nitrogen, Phosphorus, Sulfur, Micronutrients John Wiley & Sons, Inc., Toronto, ON.

Terzano, R., Mimmo, T., Vekemans, B., Vincze, L., Falkenberg, G., Tomasi, N., Schnell Ramos, M., Pinton, R., Cesco, S., 2013 Iron (Fe) speciation in xylem sap by XANES

at a high brilliant synchrotron X-ray source: opportunities and limitations Anal Bioanal Chem 405 (16), 5411e5419.

Thurman, E.M., 1985 Organic Geochemistry of Natural Waters Kluwer Academic Publishers, Hingham, MA.

Ukai, M., Yokoya, A., Fujii, K., Saitoh, Y., 2008 X-ray absorption spectrum for

guanosine-5 0-monophosphate in water solution in the vicinity of the nitrogen K-edge observed in

free liquid jet in vacuum Radiat Phys Chem 77 (10e12), 1265e1269.

Ukai, M., Yokoya, A., Nonaka, Y., Fujii, K., Saitoh, Y., 2009 Synchrotron radiation electron studies for primary radiation effects using a liquid water jet in vacuum: total and partial photoelectron yields for liquid water near the oxygen K-edge Radiat Phys Chem 78 (12), 1202e1206.

photo-Vairavamurthy, A., Wang, S., 2002 Organic nitrogen in geomacromolecules: Insights on speciation and transformation with K-edge XANES spectroscopy Environ Sci Technol 36 (14), 3050e3056.

Wang, J., Morin, C., Li, L., Hitchcock, A.P., Scholl, A., Doran, A., 2009 Radiation damage

in soft X-ray microscopy J Electron Spectrosc Relat Phenom 170 (1e3), 25e36 Wilks, R.G., MacNaughton, J.B., Kraatz, H.B., Regier, T., Blyth, R.I.R., Moewes, A.,

2009 Comparative theoretical and experimental study of the radiation-induced position of glycine J Phys Chem A 113 (18), 5360e5366.

decom-Winter, B., 2009 Liquid microjet for photoelectron spectroscopy Nucl Instrum Methods Phys Res Sect A: Accel Spectrom Detect Assoc Equip 601 (1e2), 139e150 Winter, B., Aziz, E.F., Ottosson, N., Faubel, M., Kosugi, N., Hertel, I.V., 2008 Electron dynamics in charge-transfer-to-solvent states of aqueous chloride revealed by Cl- 2p resonant auger-electron spectroscopy J Am Chem Soc 130 (22), 7130e7138 Winter, B., Faubel, M., 2006 Photoemission from liquid aqueous solutions Chem Rev.

106 (4), 1176e1211.

Zubavichus, Y., Fuchs, O., Weinhardt, L., Heske, C., Umbach, E., Denlinger, J.D., Grunze, M., 2004 Soft X-ray-induced decomposition of amino acids: an XPS, mass spectrometry, and NEXAFS study Radiat Res 161 (3), 346e358.