Fluorescence

19 downloads 0 Views 9MB Size Report
May 14, 2018 - Of C=O/C=C groups, electron-deficient C=O. (1740 cm. −1. ) ..... electron-withdrawing groups toward higher frequency,22 as the pyrolysis ...
Research Article pubs.acs.org/journal/ascecg

Cite This: ACS Sustainable Chem. Eng. 2018, 6, 8055−8062

In Situ and Ex Situ 2D Infrared/Fluorescence Correlation Monitoring of Surface Functionality and Electron Density of Biochars Minori Uchimiya,*,†,∥ Isao Noda,‡,§ Alexander Orlov,∥ and Girish Ramakrishnan∥ †

USDA-ARS Southern Regional Research Center, 1100 Robert E. Lee Boulevard, New Orleans, Louisiana 70124, United States Department of Material Science and Engineering, 211 DuPont Hall, University of Delaware, Newark, Delaware 19716, United States § Danimer Scientific, 140 Industrial Boulevard, Bainbridge, Georgia 39817, United States ∥ Department of Material Science and Engineering, State University of New York, 100 Nicolls Road, Stony Brook, New York 11794, United States ‡

S Supporting Information *

ABSTRACT: Carboxyl, hydroxyl, and other oxygen-containing functional groups play key roles in the interfacial reactions of soil surfaces including biochar (solid-phase slow pyrolysis product) soil amendment. Intensity and directionality in both real (synchronous) and imaginary (asynchronous) coordinates of 2D infrared correlation spectra were confirmed by the time courses of pyrolysis reaction (temperature × wavenumber × absorbance; 10 °C min−1, 1 h residence time at 500 °C) utilizing high-density (74 total spectra) in situ diffuse reflectance Fourier transform (DRIFTs) monitoring. Similar primary trends were observed for four different lignocellulosic biomass feedstocks: cottonseed hull, cotton ginning waste, flax shive, and pecan shell. In the OH stretch region (3100−3750 cm−1), free OH was most sensitive to pyrolysis temperature and reacted before H-bonded OH indicating the evaporation of water, followed by the cleavage of interchain H-bonds. Aromatic CH (R=CHn) was the primary CH functionality (within 2700−3100 cm−1) impacted by the pyrolysis temperature perturbation and formed as the aliphatic CHx was removed. Of C=O/C=C groups, electron-deficient C=O (1740 cm−1) was most sensitive to pyrolysis, reacted synchronously (in the same direction) with the aromatic C=C (1510 cm−1), and was formed after the most electron-rich C=O (1620 cm−1). This electron-density trend in the C=O/C=C (1400−1800 cm−1) region of infrared coincided with the formation of aromatic extractable carbon before aliphatic structures in 2D fluorescence emission−emission correlation spectra using 340 nm excitation wavelength. Results could be used to drive biomass pyrolysis toward desirable solid- (carboxyl-enriched biochar) and liquid-phase (less hydrophilic bio-oil) products. KEYWORDS: Thermochemical conversion, Agricultural commodity, Waste management, Torrefaction, Bio-oil



INTRODUCTION Carboxyl, hydroxyl, and other oxygen-containing functionalities of soils and sediments are known to control the environmental processes including the metal speciation, aggregation of colloids, and redox reaction1,2 and are equally important to biochar produced as the soil amendment.3,4 Fourier transform infrared (FTIR) spectroscopy is the primary technique used to identify oxygen-containing surface functional groups on the solid-phase biomass pyrolysis product, biochar.5,6 Peak integration at a characteristic wavenumber (in cm−1) is often used to characterize the pyrolysis temperature dependence in the formation/removal of carboxyl and other essential ionizable functional groups.5,6 However, the majority of oxygencontaining polar functionalities partition into the liquid (tar/ boil−oil) phase by 500 °C.7 Empirical stoichiometry indicated that only 4% of the oxygen in the feedstock remained on biochar after the slow pyrolysis of pecan shell (10 °C min−1 up to 500 °C).7 Therefore, thermodynamics drives biomass © 2018 American Chemical Society

pyrolysis toward major drawbacks in utilizing solid (low surface functionality on biochar) as well as liquid (hydrophilicity of biooil) product.7,8 Featureless FTIR spectra of higher pyrolysis temperature (700−800 °C) biochars9 resemble that of graphite lacking characteristic bands.10 Baseline shifts caused by the condensed aromatic structure (diffuse absorption)10 and overlapping peaks11 further complicate the FTIR-based monitoring of surface functionality on biochar. Perturbation-induced generalized two-dimensional correlation spectroscopy (2DCOS)12 provides an avenue to overcome those challenges. Application of 2DCOS is not limited to FTIR and has included the resolution of chromatographic peaks, construction of partial least-squares (PLS) calibration models,13−15 and elucidation of conformational changes in Received: April 17, 2018 Revised: May 8, 2018 Published: May 14, 2018 8055

DOI: 10.1021/acssuschemeng.8b01720 ACS Sustainable Chem. Eng. 2018, 6, 8055−8062

Research Article

ACS Sustainable Chemistry & Engineering

Figure 1. Synchronous and asynchronous in situ DRIFTs (10 °C min−1, 1 h residence time at 500 °C) 2D correlation maps of OH stretching region (3100−3750 cm−1) for (a, b) cottonseed hulls, (c, d) cotton ginning waste, (e, f) flax shive, and (g, h) pecan shell. Only wavenumber ranges with visible peaks are presented, and shaded area indicates negative sign. Wavenumbers (i, j) of characteristic peaks are color-coded for positive (red) and negative (blue) signs, where the corrected sign (after taking synchronous sign into account) is provided in the asynchronous map. then held at 500 °C for 1 h. The total number of collected spectra was 74 for each biomass. 2D Infrared Correlation Spectroscopy. In situ DRIFTs spectra were separately preprocessed for each biomass (cotton ginning waste, pecan shell, cottonseed hull, or flax shive) by Savitzky−Golay smoothing and then automatic Whittaker filter-based baseline correction using MATLAB version 8.6.0.267246 (R2015b; Mathworks, Natick, MA) with PLS toolbox version 8.61 (Eigenvector Research, Manson, WA). Preprocessed spectra were used to plot (1) 3D (temperature, wavenumber, absorbance) pyrolysis kinetics using SigmaPlot 13.0 and (2) generalized 2D correlation spectra using 2D Shige software.20 Generalized 2D correlation analyses were performed on dynamic spectra with the average spectrum as a reference,12 separately for OH (3100−3750 cm−1), CH (2700−3100 cm−1), and C=O/C=C (1400−1800 cm−1) stretch regions,5,21,22 and were interpreted using Noda’s rules.23 As described in detail elsewhere,23 the synchronous 2D map represents simultaneous/coincidental changes of intensity at wavelength i (x-axis) versus j (y-axis).23 The asynchronous map represents sequential/successive changes of intensity, and an asynchronous cross peak exists only if intensities (at i, j) change out of phase with one another, i.e., are delayed or accelerated.12,23 2D Fluorescence Correlation Analysis. Preparation and characterization of biochars, extraction methods, and fluorescence excitation emission spectrophotometry with parallel factor analysis (EEM/PARAFAC) procedures were described in detail previously24 and are summarized in section I of the Supporting Information. On the basis of the previous EEM-PARAFAC analysis of sequential hot water (80 °C) and base (50 mM NaOH at room temperature) extracts,24 2D emission/emission (Em/Em) correlation spectra were constructed for 340 nm excitation (Ex) wavelength using emission spectra obtained from biomass feedstocks and biochars produced at four different temperatures between 350 and 800 °C. That is, five emission spectra were used to build Em/Em 2D maps for each biomass: pecan shell, cottonseed hull, and broiler litter. Prior to the 2D correlation analysis, all EEM spectra were preprocessed by the blank subtraction and the removal of additional regions dominated by Rayleigh/Raman peaks25−27 using MATLAB with PLS toolbox.

enzyme.16 Furthermore, advanced 2DCOS techniques allow identification of pertinent perturbation variable regions for more focused correlation analysis (moving window) and interpretation based on complementary techniques (heterocorrelation) such as FTIR and fluorescence.12 Prior detailed 2D infrared correlation analysis of biomass pyrolysis relied on a small number (10) of spectra obtained from ex situ (postproduction) FTIR characterization of 200− 650 °C solid pyrolysis products (biochars).5 When a larger number of spectra was obtained by in-situ FTIR (primarily FTIR coupled with thermogravimetric analyzer, TGA-FTIR) for coal17 or cellulose,11,18 interpretation of 2D correlation maps was confirmed solely by the integration of characteristic peaks. However, the characteristic infrared band is known to shift as the pyrolysis progresses,11 primarily because of the relative intensity changes of overlapping peaks.19 The objective of the present study was to elucidate simultaneous/sequential direction and intensity trends in OH, CH, and C=O/C=C surface functionalities. Our previous report on in situ kinetic monitoring of biomass pyrolysis provided a visual evidence for the critical temperature range (200−500 °C) of surface functional group variation.9 However, overlapping and congested DRIFTs peaks make the interpretation of spectroscopic response to pyrolysis temperature challenging. The present study re-examined the DRIFTs across two-dimension (2DCOS) to extract and interpret the kinetic trends of specific functional groups; 74 spectra were obtained for each of four different lignocellulosic biomass (cotton ginning waste, pecan shell, cottonseed hulls, and flax shive).



MATERIALS AND METHODS

In Situ DRIFTs Monitoring of Slow Pyrolysis. As described in detail previously for cottonseed hull9 and pecan shell7 feedstocks, finely ground biomass sample (2 mg cotton ginning waste or flax shive) was pyrolyzed in a time-resolved (0.5 s) DRIFTs environmental chamber (Nicolet 6700; Thermo Scientific, Waltham, MA) with a SMART Collector accessory under 30 cm3 min−1 N2 flow rate. The temperature of the environmental chamber was manually increased at 10 °C min−1 (within the slow pyrolysis regime) until 500 °C and was



RESULTS AND DISCUSSION 2D Correlation Spectra of in-Situ DRIFTs for Different Biomass. Figure 1 presents synchronous and asynchronous 2D spectra for the OH stretch region (3100−3750 cm−1) of four 8056

DOI: 10.1021/acssuschemeng.8b01720 ACS Sustainable Chem. Eng. 2018, 6, 8055−8062

Research Article

ACS Sustainable Chemistry & Engineering

Figure 2. Smoothed and baseline-corrected raw (before 2DCOS) DRIFTs of four different biomass in the OH region.

indicate the formation of i before j, while negative cross-peaks indicate the formation of j before i, and those orders are reversed if the corresponding sign of the synchronous peak is negative. The synchronous map of cottonseed hulls contains a single autopeak of the free OH (3590 cm−1 in Figure 1a), while the asynchronous map shows a +3590/3530 cross-peak (Figure 1b), indicating that free OH is formed before H-bonded OH. To confirm those observations in 2D correlation maps (Figure 1a, b), Figure 2a shows a 3D plot of raw DRIFTs (74 total smoothed and baseline-corrected spectra plotted as wavenumber × temperature × absorbance) in the OH stretch region collected over the course of the pyrolysis reaction. Figure 2a indicates free OH (3600 cm−1) as the dominant peak occurring before (at lower pyrolysis temperature than) H-bonded OH at 3540 cm−1, in agreement with 2DCOS in Figure 1a, b. This observation is in agreement with TGA-FTIR monitoring (25 °C min−1 up to 1100 °C) of coal pyrolysis;17 2DCOS showed free OH (3621 cm−1) before H-bonded OH at 3525 and 3443 cm−1. Furthermore, heat treatment (up to 370 °C) of cotton

different biomass: cottonseed hulls (a, b), cotton ginning waste (c, d), flax shive (e, f), and pecan shell (g, h). Free OH of alcohols and phenols occurs at 3600−3650 cm−1, while Hbonded OH appears as a broad peak at 3200−3500 cm−1.22 Within the range of H-bonded OH at 3200−3500 cm−1,22 shifts toward higher wavenumbers indicate the lower degree of Hbonding.28,29 Figure 1 only shows wavenumber regions having visible peaks, where shaded peaks indicate the negative sign. Wavenumbers (i, j) of characteristic peaks are color-coded for positive (red) and negative (blue) signs, where the signs in the asynchronous map are corrected for the sign in the synchronous map of the same region. The peak intensity and direction are interpreted as follows:23 (1) intensity of autopeak (along the diagonal) in the synchronous map is proportional to the overall degree of sensitivity to the pyrolysis temperature perturbation, (2) positive cross-peaks (at off-diagonal wavenumbers i, j) on the synchronous map indicate the changes in the same direction at i and j, while negative cross-peaks indicate the changes in opposite direction as a function of pyrolysis temperature, (3) positive cross-peaks on the asynchronous map 8057

DOI: 10.1021/acssuschemeng.8b01720 ACS Sustainable Chem. Eng. 2018, 6, 8055−8062

Research Article

ACS Sustainable Chemistry & Engineering

Figure 3. (a) Synchronous and (b) asynchronous maps and (c) 3D DRIFTs time courses of slow pyrolysis reaction in the CH region of cottonseed hull feedstock.

Figure 4. (a, b) 2DCOS maps for the fingerprint (C=O/C=C) region of cottonseed hull slow pyrolysis monitored in situ by (c) DRIFTs.

fiber30 showed a free OH peak attributed to the evaporation of water (3604 cm−1), followed by the cleavage of interchain Hbond (3524 cm−1) in the synchronous spectrum. Both interand intramolecular H-bond of OH in cellulose could break at temperatures as low as >120 °C, although intermolecular Hbond is expected to be slightly stronger11 than the intramolecular H-bond. A water molecule in cellulose could be free or H-bonded, and as a result, its vibration appears at 3200− 3700 cm−1.31

Cotton ginning waste (Figure 1c, d) showed similar but weaker synchronous and asynchronous OH peaks as compared to cottonseed hulls (Figures 1a, b). This trend was reflected in a less visible peak for H-bonded OH in the 3D DRIFTs of cotton ginning waste (Figure 2c), relative to cottonseed hulls (Figure 2a). The synchronous map of flax shive (Figure 1e) showed an additional autopeak for H-bonded OH (3540 cm−1) having a lower intensity than free OH (3600 cm−1); cross-peaks for 8058

DOI: 10.1021/acssuschemeng.8b01720 ACS Sustainable Chem. Eng. 2018, 6, 8055−8062

Research Article

ACS Sustainable Chemistry & Engineering

Figure 5. Synchronous and asynchronous fluorescence emission−emission (340 nm excitation) maps of (a−d) cottonseed hulls and (e−h) broiler litter feedstocks and biochar (350, 500, 650, and 800 °C) samples sequentially extracted by 80 °C water (a, b, e, f) and 50 mM NaOH at room temperature (c, d, g, h).

cm−1), and additional oxygen-containing functionalities, for example, O−CH3 at 1450 cm−1. This wavenumber range is often used to highlight the “fingerprint region” of biochars attributable to the ionizable functional groups.5,6,21 Characteristic peaks of C=O stretching at 1680−1740 cm−1 are expected to shift by (1) conjugation (e.g., phenyl, alkene, and α, βunsaturated carbonyls) toward lower frequency and (2) electron-withdrawing groups toward higher frequency,22 as the pyrolysis reaction progresses. The synchronous map for cottonseed hulls in Figure 4a shows 1740 cm−1 (aldehyde, ester, or other C=O with electronwithdrawing group) as the dominant autopeak, which was most sensitive to pyrolysis. The following cross-peaks are observable in the synchronous map (in cm−1): +1510 (C=C)/1740, −1690 (conjugated acid/aldehyde/ketone)/1740, and +1620 (C=C, C=O)/1740. Therefore, electron-deficient C=O (1740 cm−1) is most sensitive to pyrolysis, changes in the same direction as the aromatic C formation (1510), and in the opposite direction compared to the electron-rich C=O (1690). These observations in the synchronous map of 2DCOS are confirmed in raw DRIFTs time courses (Figure 4c). Figure 4c shows a gradual reduction in carboxyl (1740 cm−1), C=O/C=C (1620 cm−1), and C=C (1520 cm−1) peaks as a function of pyrolysis temperature. On the other hand, conjugated acid/ aldehyde/ketone (1680 cm−1) formation peaked near 300 °C (Figure 4c). The asynchronous map in Figure 4b indicates that the most electron-rich C=O (1620) formed before the electron-deficient C=O (1690 and 1740), in agreement with Figure 4c. In addition, C=O functionalities reacted before C=C (+1750/ 1520 and +1620/1520 in Figure 4b). In conclusion, Figures 1−4 utilized in-situ pyrolysis monitoring in 3D (wavenumber × temperature × absorbance) to confirm both simultaneous (synchronous map) and sequential (asynchronous map) intensity and directional trends in 2DCOS. Electron-rich C=O (1620) formed before electron-deficient C=O (1690 and 1740 cm−1), and carboxyl C=O was the most sensitive to

those wavenumbers are negative, indicating the opposite effects of pyrolysis temperature. This trend is reflected in the 3D DRIFTs (Figure 2b) showing a consistent decrease in dominant free OH peak, as opposed to the H-bonded OH reaching the maximum near 300 °C. At 3600−3640 cm−1 free OH region, a lower wavenumber peak (3600 cm−1) is formed before the higher wavenumber peak (3640 cm−1) in Figure 2b, as indicated in the asynchronous map for flax shive (Figure 1f). The pecan shell showed two auto peaks within the free OH region occurring in the opposite direction (3650, 3590 cm−1 in the synchronous map, Figure 1g). The asynchronous map indicates the formation of a higher wavenumber peak before a lower wavenumber peak (3660 before 3540, and 3640 before 3610), in agreement with raw DRFITs kinetics presented in Figure 2d. Because similar primary trends (free OH was most sensitive to pyrolysis temperature and began to disappear as H-bonded OH formed) were observed for different biomass samples in Figures 1 and 2, Figures 3 and 4 show synchronous (a) and asynchronous (b) maps and 3D DRIFTs (c) of CH and C=O/ C=C regions for cottonseed hulls as a representative biomass. The synchronous map of the CH region shows three autopeaks with the following decreasing intensity trend: aromatic C−H (3030 cm−1) > aliphatic CHx (2950 cm−1) > aliphatic CHx (2880 cm−1, Figure 3a). Aromatic C−H had negative crosspeaks with both of aliphatic CHx, while two aliphatic CHx peaks had a positive cross-peak (Figure 3a). The asynchronous map indicates that the lowest wavenumber aliphatic CHx (2890 cm−1) reached the maximum at a lower temperature than the aromatic C−H (3030 cm−1) and aliphatic CHx at 2950 cm−1. Those trends in 2DCOS (Figure 3a, b) are observable in the time course of pyrolysis kinetics (Figure 3c): consistent decrease in the aliphatic CHx peak as a function of temperature, as opposed to the aromatic C−H reaching the maximum at a higher temperature range. Figure 4 explores the wavenumber range (1400−1800 cm−1) attributable to carboxyl C=O (1700−1740 cm−1), C=C (1600 8059

DOI: 10.1021/acssuschemeng.8b01720 ACS Sustainable Chem. Eng. 2018, 6, 8055−8062

Research Article

ACS Sustainable Chemistry & Engineering

Base extracts (Figure 5c, d) were obtained sequentially, after the hot water extraction (Figure 5a, b). The autopeak of the base extract is shifted to lower wavelength, indicating a slightly less aromatic structure; otherwise, both synchronous and asynchronous spectra are similar to the hot water extract. Low extractable carbon in pecan shell24 resulted in low 2DCOS peak intensity (section IV of the Supporting Information). To provide comparison with another biomass type, Figure 5e and h presents fluorescence 2DCOS of broiler letter biochars previously shown to have a high amount of polar functionality.24 As shown in Figure 5, synchronous and asynchronous maps of water/base sequential extracts for broiler litter biochars are similar to cottonseed hull biochars. Fluorescence 2DCOS has been explored (to a much lesser extent than infrared) using excitation wavelength, analyte concentration, and florescence quenching as the perturbation parameters.16,36,37 In fluorescence 2DCOS, inner-filtering could contribute to the asynchronous cross-peaks.36 Inner-filtering is an artifact caused by the absorption of excitation and emission light and is greatest at the shortest EEM wavelengths.38 Innerfiltering effects could either be (1) intentionally removed by sample dilution or algorithmic correction methods39 or (2) utilized as the intrinsic characteristic of the sample.40 To understand the feedstock dependence in the abovedescribed 2DCOS (Figures 1−5), Figure S8 of the Supporting Information presents a Pearson’s correlation map (p < 0.05 marked by yellow circle) for the following properties of four different biomass (pecan shell, cottonseed hulls, cotton ginning waste, and flax shive): cellulose, hemicellulose, and lignin contents; C/H/N/S/O; and proximate analysis results (ash, fixed carbon, moisture, and volatile matter (VM); detailed characterization methods and results are described in sections I and V of the Supporting Information). In agreement with minimal feedstock dependence on 2DCOS (Figures 1−5 and section III of the Supporting Information), oxygen content of different lignocellulosic feedstocks did not correlate with lignin/hemicellulose/cellulose contents in Figure S8. Cellulose content is negatively correlated with lignin (Figure S8), in agreement with the negative cross-peaks observed in the synchronous map of the CH region (Figure 3a). Hemicellulose positively correlated with VM, suggesting the contribution of hemicellulose in labile carbon (as opposed to recalcitrant fixed carbon) composition of biomass feedstock. In conclusion, the present study utilized perturbationinduced generalized two-dimensional correlation infrared spectroscopy to understand the direction and intensity of OH, CH, and C=O/C=C functionality formation and disappearance during biomass pyrolysis. Interpretation of 2D correlation maps was confirmed by 3D kinetic plots (temperature, wavenumber, absorbance) of raw spectra (before 2D correlation analysis). In addition, 2DCOS was conducted on the fluorescence excitation−emission spectra of biochar’s DOC to compare the degree and directionality of functional group formation in the solid (char) and residual liquid-phase products on biochar. Electron-dense structures formed/reacted before (at lower temperature than) electron-deficient carbon in bulk biochar as well as its DOC. Regardless of lignocellulose source (cottonseed hull, cotton ginning waste, flax shive, or pecan shell), similar simultaneous/coincidental (synchronous) and sequential/successive (asynchronous) trends were observed. Within the fingerprint region (1400−1800 cm−1), electrondonating and electron-withdrawing C=O functionalities showed distinctive formation and removal phases. Observed temper-

the temperature perturbation among C=O/C=C functional groups. On the other hand, aliphatic CHx (2890 cm−1) reached the maximum at a lower temperature than the aromatic C−H (3030 cm−1). Free OH was most sensitive to pyrolysis temperature and began to disappear before H-bonded OH formed. Previous FTIR-based 2DCOS utilized the integrated areas of characteristic bands to support observations in synchronous and asynchronous maps and to deduce pyrolysis reaction mechanisms.5,11,17 However, as described in detail in section II of the Supporting Information (for cottonseed hulls), the characteristic wavenumber is known to shift as the pyrolysis reaction progresses.28−30 The 3D time courses of pyrolysis kinetics provide a visual comparison with 2DCOS, eliminating the need to integrate individual peaks as a function of perturbation parameter, that is, temperature. Different infrared methods are known to shift characteristic wavenumbers.32−34 In particular, attenuated total reflectance (ATR-FTIR) method enhances lower wavenumber bands, because of the deeper sample penetration depths.32−34 In addition, ATR-FTIR spectra are sensitive to the refractive index of the sample in a wavenumber-dependent fashion, which results in further distortion of the peak profiles compared to DRIFTs. Section III of the Supporting Information examined 2DCOS on ATR-FTIR spectra of cottonseed hull biochars obtained ex situ (post-pyrolysis). Interpretation of 2DCOS based on ATR-FTIR reflects the enhancement of lower wavenumber peaks and is in agreement with a prior report utilizing 2DCOS on ex situ FTIR analysis of biochar samples:5 (1) cleavage of H-bonded OH to yield free OH, which is putatively oxidized to carboxyl, and (2) decomposition of cellulose R-CH2 (2940 cm−1) before lignin R-OCH3 (2890 cm−1). 2D Correlation Fluorescence Emission−Emission Spectra. To further explore the magnitude and direction of pyrolysis temperature perturbation effects, 2D correlation maps were built for the fluorescence EEM of biochar extracts. The water-extractable portion of biochar carbon represents dissolved organic carbon (DOC) that could be released to amended soil to impact aquatic chemistry ranging from metal speciation to pH buffering.24,35 The excitation wavelength (340 nm) was selected to match PARAFAC fingerprint previously attributed to the transient pyrolysis products including carboxyl.24 Figure 5 shows synchronous (a) and asynchronous (b) 2D correlation fluorescence emission−emission spectra of cottonseed hulls at 340 nm excitation; pyrolysis temperature is the perturbation parameter. Figure 5 was constructed to only show the wavelength ranges having visible peaks; the region affected by the first-order Rayleigh scatter (Ex = Em = 340 nm) was excluded.27 The synchronous map of the hot (80 °C) water extract (Figure 5a) shows a single autopeak at 430 nm as the emission wavelength most sensitive to pyrolysis. The corresponding asynchronous map (Figure 5b) shows a positive 450/390 cross-peak within the positive region of the synchronous map (Figure 5a), indicating that the aromatic structure (higher emission wavelength) formed before the aliphatic structure. This directional observation is in agreement with 2D infrared correlation analysis (Figure 4b). Electronenriched C=O appeared (formed/reacted) before electrondeficient C=O (+1680/1740, + 1620/1740, and +1620/1690; Figure 4b). 8060

DOI: 10.1021/acssuschemeng.8b01720 ACS Sustainable Chem. Eng. 2018, 6, 8055−8062

Research Article

ACS Sustainable Chemistry & Engineering

(11) Leng, E.; Zhang, Y.; Peng, Y.; Gong, X.; Mao, M.; Li, X.; Yu, Y. In situ structural changes of crystalline and amorphous cellulose during slow pyrolysis at low temperatures. Fuel 2018, 216, 313−321. (12) Noda, I. Advances in Two-Dimensional Correlation Spectroscopy (2DCOS). In Frontiers and Advances in Molecular Spectroscopy; Laane, J., Ed.; Elsevier: Amsterdam, 2018; Chapter 2, pp 47−75. (13) Van Orden, A.; Keller, R. A. Fluorescence correlation spectroscopy for rapid multicomponent analysis in a capillary electrophoresis system. Anal. Chem. 1998, 70, 4463−4471. (14) Wang, G.; Geng, L. Two-dimensional fluorescence correlation in capillary electrophoresis for peak resolution and species identification. Anal. Chem. 2000, 72, 4531−4542. (15) Phong, D. D.; Hur, J. Using two-dimensional correlation size exclusion chromatography (2D-CoSEC) and EEM-PARAFAC to explore the heterogeneous adsorption behavior of humic substances on nanoparticles with respect to molecular sizes. Environ. Sci. Technol. 2018, 52, 427−435. (16) Fukuma, H.; Nakashima, K.; Ozaki, Y.; Noda, I. Twodimensional fluorescence correlation spectroscopy IV: Resolution of fluorescence of tryptophan residues in alcohol dehydrogenase and lysozyme. Spectrochim. Acta, Part A 2006, 65, 517−522. (17) Hu, J.; Chen, Y.; Qian, K.; Yang, Z.; Yang, H.; Li, Y.; Chen, H. Evolution of char structure during mengdong coal pyrolysis: Influence of temperature and K2CO3. Fuel Process. Technol. 2017, 159, 178−186. (18) Watanabe, A.; Morita, S.; Ozaki, Y. Study on temperaturedependent changes in hydrogen bonds in cellulose Iβ by infrared spectroscopy with perturbation-correlation moving-window twodimensional correlation spectroscopy. Biomacromolecules 2006, 7, 3164−3170. (19) Ryu, S. R.; Noda, I.; Jung, Y. M. What is the origin of positional fluctuation of spectral features: True frequency shift or relative intensity changes of two overlapped bands? Appl. Spectrosc. 2010, 64, 1017−1021. (20) Morita, S. 2Dshige, 1.3. 2004−2005, Kwansei-Gakuin University (https://sites.google.com/site/shigemorita/home/2dshige) (accessed 18 May 2018). (21) Keiluweit, M.; Nico, P. S.; Johnson, M. G.; Kleber, M. Dynamic molecular structure of plant biomass-derived black carbon (biochar). Environ. Sci. Technol. 2010, 44, 1247−1253. (22) Silverstein, R. M.; Bassler, G. C.; Morrill, T. C. Spectrometric Identification of Organic Compound, 5th ed.; John Wiley & Sons, Inc.: New York, 1991. (23) Noda, I.; Dowrey, A. E.; Marcott, C.; Story, G. M.; Ozaki, Y. Generalized two-dimensional correlation spectroscopy. Appl. Spectrosc. 2000, 54, 236A−248A. (24) Uchimiya, M.; Ohno, T.; He, Z. Pyrolysis temperaturedependent release of dissolved organic carbon from plant, manure, and biorefinery wastes. J. Anal. Appl. Pyrolysis 2013, 104, 84−94. (25) Christensen, J. H.; Hansen, A. B.; Mortensen, J.; Andersen, O. Characterization and matching of oil samples using fluorescence spectroscopy and parallel factor analysis. Anal. Chem. 2005, 77, 2210− 2217. (26) Bahram, M.; Bro, R.; Stedmon, C.; Afkhami, A. Handling of Rayleigh and Raman scatter for PARAFAC modeling of fluorescence data using interpolation. J. Chemom. 2006, 20, 99−105. (27) Rinnan, Å.; Andersen, C. M. Handling of first-order Rayleigh scatter in PARAFAC modelling of fluorescence excitation-emission data. Chemom. Intell. Lab. Syst. 2005, 76, 91−99. (28) Imamura, K.; Sakaura, K.; Ohyama, K. I.; Fukushima, A.; Imanaka, H.; Sakiyama, T.; Nakanishi, K. Temperature scanning FTIR analysis of hydrogen bonding states of various saccharides in amorphous matrixes below and above their glass transition temperatures. J. Phys. Chem. B 2006, 110, 15094−15099. (29) Wolkers, W. F.; Oldenhof, H.; Alberda, M.; Hoekstra, F. A. A Fourier transform infrared microspectroscopy study of sugar glasses: Application to anhydrobiotic higher plant cells. Biochim. Biophys. Acta, Gen. Subj. 1998, 1379, 83−96. (30) Kokot, S.; Czarnik-Matusewicz, B.; Ozaki, Y. Two-dimensional correlation spectroscopy and principal component analysis studies of

ature dependence could be used to manipulate the surface functionality and electron density of biochar soil amendment.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acssuschemeng.8b01720. Production and ex-situ characterization of biochar, Gaussian integration of characteristic DRIFTs bands, 2D infrared correlation analysis of ex situ ATR-FTIR spectra for cottonseed hull biochars, 2D fluorescence correlation spectra of pecan shell biochar, correlation map of feedstock properties (PDF)



AUTHOR INFORMATION

Corresponding Author

*Fax: (504) 286-4367; phone: (504) 286-4356; e-mail: sophie. [email protected]. ORCID

Minori Uchimiya: 0000-0001-9117-7431 Notes

Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. USDA is an equal opportunity provider and employer. The authors declare no competing financial interest.



REFERENCES

(1) Stumm, W.; Morgan, J. J. Aquatic Chemistry; Wiley-Interscience: New York, 1996. (2) Sposito, G. The Chemistry of Soils; Oxford University Press: New York, 1989. (3) Pignatello, J. J.; Uchimiya, M.; Abiven, S.; Schmidt, M. W. I. Evolution of Black Carbon Properties in Soil. In Biochar for Environmental Management: Science, Technology, and Implementation; Lehmann, J., Joseph, S., Eds.; Taylor & Frances: London, 2015; pp 195−234. (4) Xiao, X.; Chen, B.; Chen, Z.; Zhu, L.; Schnoor, J. L. Insight into multiple and multilevel structures of biochars and their potential environmental applications: a critical review. Environ. Sci. Technol. 2018, 52, 5027−5047. (5) Harvey, O. R.; Herbert, B. E.; Kuo, L. J.; Louchouarn, P. Generalized two-dimensional perturbation correlation infrared spectroscopy reveals mechanisms for the development of surface charge and recalcitrance in plant-derived biochars. Environ. Sci. Technol. 2012, 46, 10641−10650. (6) Chen, Z.; Xiao, X.; Chen, B.; Zhu, L. Quantification of chemical states, dissociation constants and contents of oxygen-containing groups on the surface of biochars produced at different temperatures. Environ. Sci. Technol. 2015, 49, 309−317. (7) Leblanc, J.; Uchimiya, M.; Ramakrishnan, G.; Castaldi, M. J.; Orlov, A. Across-phase biomass pyrolysis stoichiometry, energy balance, and product formation kinetics. Energy Fuels 2016, 30, 6537−6546. (8) Crossley, S.; Faria, J.; Shen, M.; Resasco, D. E. Solid nanoparticles that catalyze biofuel upgrade reactions at the water/oil interface. Science 2010, 327, 68−72. (9) Uchimiya, M.; Orlov, A.; Ramakrishnan, G.; Sistani, K. In situ and ex situ spectroscopic monitoring of biochar’s surface functional groups. J. Anal. Appl. Pyrolysis 2013, 102, 53−59. (10) Mochidzuki, K.; Soutric, F.; Tadokoro, K.; Antal, M. J., Jr; Tóth, M.; Zelei, B.; Várhegyi, G. Electrical and physical properties of carbonized charcoals. Ind. Eng. Chem. Res. 2003, 42, 5140−5151. 8061

DOI: 10.1021/acssuschemeng.8b01720 ACS Sustainable Chem. Eng. 2018, 6, 8055−8062

Research Article

ACS Sustainable Chemistry & Engineering temperature-dependent IR spectra of cotton-cellulose. Biopolymers 2002, 67, 456−469. (31) Fengel, D. Influence of water on the OH valency range in deconvoluted FTIR spectra of cellulose. Holzforschung 1993, 47, 103− 108. (32) Application Note: 50581 Advanced ATR Correction Algorithm, Thermo Electron Scientific Instruments LLC: Madison, WI, 2008. (33) Rosenthal, R. J.; Lowry, S. R. Effects of sampling methodologies on FT-IR data base searching. Microchim. Acta 1986, 89, 291−302. (34) Shimadzu Services & Support, Kyoto, Japan. ATR Precautions [Online Database] Available: https://www.shimadzu.com/an/ftir/ support/ftirtalk/letter1/atr1.html (accessed 18 May 2018). (35) Yuan, J. H.; Xu, R. K.; Zhang, H. The forms of alkalis in the biochar produced from crop residues at different temperatures. Bioresour. Technol. 2011, 102, 3488−3497. (36) Nakashima, K.; Yashuda, S.; Ozaki, Y.; Noda, I. Twodimensional fluorescence correlation spectroscopy. I. Analysis of polynuclear aromatic hydrocarbons in cyclohexane solutions. J. Phys. Chem. A 2000, 104, 9113−9120. (37) Roselli, C.; Burie, J. R.; Mattioli, T.; Boussac, A. Twodimensional correlation method applied to Yb3+ vibronic sideband spectroscopy: Discrimination of fluorescence spectral features arising from different Yb3+ binding sites. Biospectroscopy 1995, 1, 329−339. (38) Kothawala, D. N.; Murphy, K. R.; Stedmon, C. A.; Weyhenmeyer, G. A.; Tranvik, L. J. Inner filter correction of dissolved organic matter fluorescence. Limnol. Oceanogr.: Methods 2013, 11, 616−630. (39) Ryder, A. G.; Stedmon, C. A.; Harrit, N.; Bro, R. Calibration, standardization, and quantitative analysis of multidimensional fluorescence (MDF) measurements on complex mixtures (IUPAC Technical Report). Pure Appl. Chem. 2017, 89, 1849−1870. (40) Gu, W.; Pei, X.; Cheng, Y.; Zhang, C.; Zhang, J.; Yan, Y.; Ding, C.; Xian, Y. Black phosphorus quantum dots as the ratiometric fluorescence probe for trace mercury ion detection based on inner filter effect. ACS Sensors 2017, 2, 576−582.

8062

DOI: 10.1021/acssuschemeng.8b01720 ACS Sustainable Chem. Eng. 2018, 6, 8055−8062