The One-Sample PARAFAC Approach Reveals Molecular Size Distributions of Fluorescent Components in Dissolved Organic Matter

Abstract

Molecular size plays an important role in dissolved organic matter (DOM) biogeochemistry, but its relationship with the fluorescent fraction of DOM (FDOM) remains poorly resolved. Here high-performance size exclusion chromatography (HPSEC) was coupled to fluorescence emission-excitation (EEM) spectroscopy in full spectral (60 emission and 34 excitation wavelengths) and chromatographic resolution (<1 Hz), to enable the mathematical decomposition of fluorescence on an individual sample basis by parallel factor analysis (PARAFAC). The approach allowed cross-system comparisons of molecular size distributions for individual fluorescence components obtained from independent data sets. Spectra extracted from allochthonous DOM were highly similar. Allochthonous and autochthonous DOM shared some spectra, but included unique components. In agreement with the supramolecular assembly hypothesis, molecular size distributions of the fluorescence fractions were broad and chromatographically unresolved, possibly representing reoccurring fluorophores forming noncovalently bound assemblies of varying molecular size. Samples shared underlying fluorescence components that differed in their size distributions but not their spectral properties. Thus, in contrast to absorption measurements, bulk fluorescence is unlikely to reliably indicate the average molecular size of DOM. The one-sample approach enables robust and independent cross-site comparisons without large-scale sampling efforts and introduces new analytical opportunities for elucidating the origins and biogeochemical properties of FDOM.

Glacier Outflow Dissolved Organic Matter as a Window Into Seasonally Changing Carbon Sources: Leverett Glacier, Greenland

“…While C 483 and C 425 are typical of DOM from plant and soil sources, and C 317 and C 368 are similar to protein‐like fluorescence, traditionally believed to be indicative of algal and microbial sources, C 416 is a somewhat ambiguous component, with short excitation wavelengths but long emission wavelengths, and only matched two components in OpenFluor (Murphy et al, ). One study that described this component deconvoluted PARAFAC components from Pony Lake fulvic acid, a lake considered an end‐member of microbial and algal production (Kellerman et al, ; McKnight et al, ), run on a high‐pressure size‐exclusion chromatograph with a fluorescence detector (Wünsch et al, ). The second study that described this component is from a watershed influenced by both urban and agricultural inputs (Osburn, Handsel , et al, ).…”

“…The PARAFAC model was validated using split‐half validation (supporting information Figure S2) and inspection of residuals (Murphy et al, ), and it explained 99.4% of the variance. The five components are as follows and abbreviated using their fluorescence emission maximum wavelengths as in Wünsch et al (). The first component had excitation/emission maxima at 265 nm/317 nm (C 317 ), the second component at <260 nm and 365 nm/483 nm (C 483 ), the third component at <260 nm and 315 nm/425 nm (C 425 ), the fourth component at <260 nm and 280 nm/416 nm (C 416 ), and the fifth component at 265 nm and 300 nm/368 nm (C 368 ).…”

Emerging patterns in the global distribution of dissolved organic matter fluorescence

“…photochemistry or chromatography) in a manner that allows robust statistical descriptions ("one-sample PARAFAC"). 15,16 Future studies can benet from such methods and we hypothesize that this will result in a convergence towards more similar uorescence components across independent studies. 15…”

Electrochemical characterization of natural organic matter by direct voltammetry in an aprotic solvent

“…The various methods for fractionation of NOM have been optimized, standardized, compared, and criticized over many years of study, 7,[20][21][22][23][24] and during this process they have become deeply embedded in the literature on all aspects of NOM. The methods used to ngerprint NOM have expanded with advancements in the availability of high-resolution instrumentation, beginning with Fourier transform infrared spectroscopy (FTIR) and nuclear magnetic resonance (NMR), 15,[25][26][27][28][29][30][31][32][33][34] then uorescence excitation emission matrices (EEMs), [35][36][37][38][39] and most recently Fourier transform ion cyclotron resonance mass spectrometry (FTICR-MS). 5,6,9,10,19,20,22,24,[40][41][42][43] The increasingly wholistic characterizations of NOM obtained from ngerprinting methods have renewed concerns over the representativeness of NOM samples obtained by extraction methods, 7,21,24 and growing recognition that sample preparation for ngerprinting can introduce biases due to fractionation.…”