Multiharmonic electron paramagnetic resonance for extended samples with both narrow and broad lines

Abstract

Multiharmonic electron paramagnetic resonance spectroscopy was demonstrated for two samples with both narrow and broad lines: (i) α,γ-Bisdiphenylene-β-phenylallyl (BDPA) with ΔBpp of 0.85 G plus ultramarine blue with ΔBpp of 17 G, and (ii) a nitroxide radical immobilized in sucrose octaacetate. Modulation amplitudes up to 17 G at 41 kHz were generated with a rapid scan coil driver and Litz wire coils that provide uniform magnetic field sweeps over samples with heights of 5mm. Data were acquired with a 2-D experiment in the Xepr software through the transient signal path of a Bruker E500T and digitized in quadrature with a Bruker SpecJet II. Signals at the modulation frequency and its harmonics were calculated by digital phase-sensitive detection. The number of harmonics with signal intensity greater than noise increases as the ratio of the modulation amplitude to the narrowest peak increases. Spectra reconstructed by the multiharmonic method from data obtained with modulation amplitudes up to five times the peak-to-peak linewidths of the narrowest features have linewidths that are broadened by up to only about 10% relative to linewidths in spectra obtained at low modulation amplitudes. The signal-to-noise improves with increasing modulation amplitude up to the point where the modulation amplitude is slightly larger than the linewidth of the narrowest features. If this high a modulation amplitude had been used in conventional methodology the linewidth of the narrowest features would have been severely broadened. The multiharmonic reconstruction methodology means that the selection of the modulation amplitude that can be used without spectral distortion is no longer tightly tied to the linewidth of the narrowest line.

Keywords: BDPA; Digital phase-sensitive detection; EPR; Multiple harmonics; Nitroxide; Overmodulation.

Figure 1

Data processing procedure. For each magnetic field step in the spectrum, digital phase-sensitive detection is applied to both quadrature channels from the digitizer, which produces complex arrays p_n(B) and p_n’(B). The modulation phase is adjusted to null the imaginary component of each array. Fourier transformation of both arrays produces S_n(u) and S_n’(u). Reconstruction of the Fourier transform of the spectrum (Eq. (2)) and low pass filtering produces F(u) and F’(u). Inverse Fourier transformation gives the complex spectrum f(B). Microwave phase correction produces absorption and dispersion spectra.

Figure 2

130 gauss scans of the X-band spectra of the sample containing UMB and BDPA acquired with modulation amplitude of 0.17 G, B₁ of 2.5 mG, and the same 130 s data acquisition time. A) Signal acquired through the SPU, followed by 3-point smoothing, S/N = 27, B) Signal in A after Fourier filtering with the same cutoff as used for multiharmonic reconstruction, S/N = 30, C) s₁(B) from the multiharmonic data set, S/N = 80, and D) f(B) obtained by multiharmonic reconstruction with N_H = 100, S/N = 80.

Figure 3

Comparison of 130 gauss scans of the X-band spectra of the sample containing BDPA and UMB obtained with B₁ = 2.5 mG and 130 s data acquisition times, reconstructed with n = 1 and multiharmonic analysis for modulation ratios of 0.2, 1.0, and 5.0, respectively. Spectral amplitudes were normalized for modulation amplitude and y-axis scales are the same for all portions of the figure.

Figure 4

Comparison of 150 gauss scans of X-band spectra of ¹⁴N-PDT in sucrose octaacetate obtained with B₁ = 2.5 mG and 90 s data acquisition time, reconstructed with n = 1 and multiharmonic analysis for modulation ratios of 0.2, 1.0, and 5.0, respectively. Spectral amplitudes were normalized for modulation amplitude and y-axis scales are the same for all portions of the figure.