Environ. Sci. Technol. 2004, 38, 2680-2684
Accurate Quantification of Aromaticity and Nonprotonated Aromatic Carbon Fraction in Natural Organic Matter by 13C Solid-State Nuclear Magnetic Resonance J.-D. MAO AND K. SCHMIDT-ROHR* Department of Chemistry, Gilman Hall, Iowa State University, Iowa 50011
An improved approach for accurately determining the aromatic carbon fraction (fa) and nonprotonated aromatic carbon fraction (faN) in natural organic matter by solidstate 13C NMR is described. Quantitative peak areas are obtained from direct polarization 13C nuclear magnetic resonance (NMR) under high-speed magic angle spinning (MAS). The problem of overlap between aromatic and alkyl carbon resonances around 90-120 ppm in 13C NMR spectra is solved by a 13C chemical shift anisotropy (CSA) filter technique. After correction for residual spinning sidebands, an accurate value of the aromaticity fa is obtained. To obtain a quantitative faN fraction, dipolar dephasing was adapted for high-speed MAS 13C NMR; the separation of the signals of nonprotonated alkyl and aromatic carbons was achieved by CSA filtering plus dipolar dephasing. The method is demonstrated on a peat humic acid, yielding fa ) 45 ( 2% and faN ) (0.64 ( 0.07) × 45%.
Introduction A large fraction of organic matter in the natural environment contains considerable amounts of aromatics, which contribute significantly to the earth’s carbon pool. In an effort to better understand this material, which is thought to derive significantly from lignin and charcoal (1, 2), reliable quantification of its amount is a crucial first step. The percentage of aromatic carbon is also useful as a characteristic for classifying samples of natural organic matter. The importance of the aromatic carbon fraction (fa), which is commonly referred to as “aromaticity” in soil organic matter (1), is indicated by the various, more or less direct methods that have been proposed to estimate fa in natural organic matter. They include thermogravimetric analysis (3, 4), X-ray diffraction (5), and 13C nuclear magnetic resonance (NMR) (6-10), to name a few. Of these approaches, 13C solid-state NMR is the most direct, the most frequently employed, and the most widely accepted. Nevertheless, most reported NMR estimates of fa have not been very accurate. Using 13C solid-state NMR with magicangle spinning (MAS) to quantify aromatics in natural organic matter is more challenging than commonly acknowledged, because the signals of alkyl O-C-O and of aromatic carbons overlap between 90 and 120 ppm. Therefore, aromaticity estimated based on routine 13C NMR has inevitably included some of alkyl carbon signals, namely those of O-C-O groups, * Corresponding author phone: (515)294-6105; fax: (515)294-0105; e-mail:
[email protected]. 2680
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 9, 2004
or excluded some aromatics. Furthermore, in most investigations, cross polarization (CP) from 1H to 13C was employed, but it has been shown that CP techniques are not reliably quantitative for complex aromatic structures, due to differential CP efficiencies (11, 12). Even tedious studies with 8 or more CP contact times (13) fail to consistently eliminate problems arising from weak C-H dipolar couplings, differential T1F relaxation (14, 15), and signal loss due to paramagnetic species (16-18). Furthermore, the spinning sidebands of aromatic carbons have often not been taken into account correctly. A more accurate quantification of aromatics in natural organic matter requires a new combination of NMR methods that is not only quantitative (12) but also separates alkyl from aromatic carbon signals. A useful parameter related to aromaticity is the nonprotonated aromatic carbon fraction, faN. For instance, the ratio of faN/fa increases toward unity with an increasing size of fused aromatic rings and thus provides an estimate of the typical size of polycyclic aromatic components in highly aromatic natural organic matter, including charcoal (19). Previous 13C solid-state NMR studies have been aimed at estimating faN (7, 8, 20), but the systematic underrepresentation of nonprotonated aromatic carbons in CP/MAS 13C NMR spectra makes most of the reported numbers unreliable. In this study, we employ improved solid-state 13C NMR methods to overcome the shortcomings of previous attempts of determining aromaticity and faN values. To obtain quantitative fa values, we adopt 13C DP/MAS (Direct Polarization Magic Angle Spinning) at a high spinning frequency of 14 kHz, instead of the notoriously nonquantitative 13C CP/MAS NMR. We provide a table with the corrections for residual spinning sidebands at various common experimental conditions. To separate the overlapping sp3-hybridized O-C-O carbon signals from the aromatic bands, we acquire a CP/ MAS spectrum with a 13C CSA (chemical shift anisotropy) filter (21) that selects the signals of sp3-hybridized carbons. To quantify faN, we introduce a simple recoupling method that makes dipolar dephasing (also known as gated decoupling) effective at spinning frequencies of 10-20 kHz and apply it in combination with the quantitative 13C DP/MAS NMR. We separate the nonprotonated-aromatic from the nonprotonated O-C-O carbon signals by CP/MAS NMR with a 13C CSA filter and dipolar dephasing (21). The procedures for accurate quantification are described step by step and applied to a typical peat humic acid.
Experimental Section Samples. The substituted aromatic model compound, 3-methoxybenzamide (m-anisamide), was selected for its reported relatively low toxicity and purchased from SigmaAldrich-Fluka. Amherst humic acid (HA) is a fairly typical peat humic acid (22). It was extracted from the surface layer (0-20 cm) of a peat soil in Amherst, MA; its characterization has been described in detail elsewhere (22). Its elemental composition is approximately C100H102O57N4. NMR Spectroscopy. All the experiments were performed using a Bruker DSX400 spectrometer at 100 MHz for 13C. The high-speed DP and CP 13C NMR experiments were run using a Bruker 4-mm double-resonance probehead, while all other experiments were performed with 7-mm sample rotors in a Bruker double-resonance probehead. Quantitative 13C DP/MAS spectra, with the pulse sequence of Figure 1(a), were run at a spinning speed of 14 kHz. The 90° 13C pulse length was 3.4 µs. Recoupled dipolar dephasing for DP/MAS 13C NMR at a spinning speed of 14 kHz, with the pulse sequence of Figure 1(b), was developed, see below, to 10.1021/es034770x CCC: $27.50
2004 American Chemical Society Published on Web 03/20/2004
FIGURE 1. (a) Pulse sequence for quantitative 13C DP/MAS NMR, consisting of a 90° pulse for excitation and a 180° pulse after a rotation period tr, to form a Hahn echo at 2 tr, which avoids baseline problems due to probehead deadtime. (b) Pulse sequence for dipolar dephasing with recoupling, to be used at spinning frequencies of 10-20 kHz. TPPM decoupling is used during detection. Again, the 13C chemical-shift evolution is refocused into a Hahn echo at 2 t . r (c) CP/T1/TOSS pulse sequence for efficiently determining the recycle delay to be used in DP/MAS experiments. 13C magnetization generated by CP from protons is stored along the (z direction during the period tz. The signal decreased by exp(-tz/T1) is detected after total suppression of sidebands (TOSS). During the 180° pulses of TOSS, the decoupling power is increased for improved proton decoupling. (d) Standard pulse sequence for dipolar dephasing (gated decoupling) at slow MAS. While 13C one-pulse direct polarization (DP) is shown in (b) and cross polarization (CP) in (d), either can be used in either dipolar-dephasing experiment to generate the 13C magnetization. (e) 13C chemical-shift-anisotropy filter pulse sequence that selects the signals of sp3-hybridized carbons. It consists of a 180°-pulse and an incremented z-period (“γ-integral”) followed by four-pulse TOSS. The optional period with decoupling gated off (“tgade”) is shown dashed. obtain quantitative information on the nonprotonated aromatic carbon fraction faN. The 90° 13C pulse-length was 3.4 µs. To save time, in the demonstration of this new dipolar dephasing technique on 3-methoxybenzamide, CP instead of DP was employed. The dipolar dephasing time was 67 µs, the 1H 90° pulse-length, 3.5 µs, and the contact time, 1 ms. The recycle delay for quantitative DP/MAS 13C NMR of Amherst HA was chosen as 30 s, based on CP/T1/TOSS experiments (22).
NMR Background Quantitative DP/MAS 13C NMR. Quantitative 13C NMR spectra can be obtained by 90°-pulse excitation of equilibrium 13C magnetization after a sufficiently long recycle delay between scans. High spinning frequencies can be used without a problem in combination with the DP/MAS method, ensuring that aromatic-carbon spinning sidebands will be small (see below) and will not overlap with the centerbands of other carbon resonances. Figure 1(a) shows the pulse sequence used. To avoid distortions from deadtime effects, a 180°-pulse is applied at time tr and produces a Hahn echo at 2 tr, where acquisition is started. To avoid spectral distortions from B1 inhomogeneity, the “EXORCYCLE” phase cycling (11, 23, 24) is used: For each phase value of the 90° pulse, e.g. +x, the 180° pulse is cycled through all four phases, i.e., +y, -x, -y +x, and the receiver phase is alternated, e.g. +y, -y, +y, -y in our example. Due to the high spinning frequency νr, the period 2tr ) 2/νr
before detection is too short for differential T2 relaxation to become significant. We have checked that this holds even in samples with high ash content, using a special pulse sequence with variable, short pre-echo delay. DP/MAS NMR is robust to small spectrometer mis-settings, much more so than is cross-polarization 13C NMR, since no Hartmann-Hahn match of the radio frequency field strengths is required. Recycle Delay from CP/T1/TOSS NMR. The recycle delay is an important parameter in quantitative DP/MAS 13C NMR experiments. If it is too short, nonquantitative relative peak intensities will result, but if it is chosen too long, the spectrum will be noisy or the measurement time excessive. Before running the DP/MAS experiment, we use the CP/T1/total suppression of sidebands (TOSS) (25) technique (22) with the pulse sequence of Figure 1(c) in order to make sure that all carbon sites will be relaxed to
