Biradical rotamer states tune electron J coupling and MAS dynamic nuclear polarization enhancement.

Cross Effect (CE) Dynamic Nuclear Polarization (DNP) relies on the dipolar (D) and exchange (J) coupling interaction between two electron spins. Until recently only the electron spin D coupling was explicitly included in quantifying the DNP mechanism. Recent literature discusses the potential role of J coupling in DNP, but does not provide an account of the distribution and source of electron spin J coupling of commonly used biradicals in DNP. In this study, we quantified the distribution of electron spin J coupling in AMUPol and TOTAPol biradicals using a combination of continuous wave (CW) X-band electron paramagnetic resonance (EPR) lineshape analysis in a series of solvents and at variable temperatures in solution - a state to be vitrified for DNP. We found that both radicals show a temperature dependent distribution of J couplings, and the source of this distribution to be conformational dynamics. To qualify this conformational dependence of J coupling in both molecules we carry out Broken Symmetry DFT calculations which show that the biradical rotamer distribution can account for a large distribution of J couplings, with the magnitude of J coupling directly depending on the relative orientation of the electron spin pair. We demonstrate that the electron spin J couplings in both AMUPol and TOTAPol span a much wider distribution than suggested in the literature. We affirm the importance of electron spin J coupling for DNP with density matrix simulations of DNP in Liouville space and under magic angle spinning, showcasing that a rotamer with high J coupling and optimum relative g-tensor orientation can significantly boost the DNP performance compared to random orientations of the electron spin pair. We conclude that moderate electron spin J coupling above a threshold value can facilitate DNP enhancements.

Endogenous dynamic nuclear polarization NMR of hydride-terminated silicon nanoparticles.

Silicon nanoparticles (SiNPs) are intriguing materials and their properties fascinate the broader scientific community; they are also attractive to the biological and materials science sub-disciplines because of their established biological and environmental compatibility, as well as their far-reaching practical applications. While characterization of the particle nanostructure can be performed using 29Si solid-state nuclear magnetic resonance (NMR) spectroscopy, poor sensitivity due to low Boltzmann population and long acquisition times hinder in-depth studies of these potentially game-changing materials. In this study, we compare two dynamic nuclear polarization (DNP) NMR protocols to boost 29Si sensitivity in hydride-terminated SiNPs. First, we assess a traditional indirect DNP approach, where a nitroxide biradical (AMUPol or bCTbk) is incorporated into a glassing agent and transferred through protons (e- → 1H → 29Si) to enhance the silicon. In this mode, electron paramagnetic resonance (EPR) spectroscopy demonstrated that the hydride-terminated surface was highly reactive with the exogenous biradicals, thus decomposing the radicals within hours and resulting in an enhancement factor, ε, of 3 (TB = 15 s) for the 64 nm SiNP, revealing the surface components. Secondly, direct DNP NMR methods were used to enhance the silicon without the addition of an exogenous radical (i.e., use of dangling bonds as an endogenous radical source). With radical concentrations <1 mM, 29Si enhancements were obtained for the series of SiNPs ranging from 3 to 64 nm. The ability to use direct 29Si DNP transfer (e- → 29Si) shows promise for DNP studies of these inorganic nanomaterials (ε = 6 (TB = 79 min) for 64 nm SiNPs) with highly reactive surfaces, showing the sub-surface and core features. These preliminary findings lay a foundation for future endogenous radical development through tailoring the surface chemistry, targeting further sensitivity gains.

Direct dynamic nuclear polarization of 15N and 13C spins at 14.1 T using a trityl radical and magic angle spinning.

We investigate solid-state dynamic nuclear polarization of 13C and 15N nuclei using monoradical trityl OX063 as a polarizing agent in a magnetic field of 14.1 T with magic angle spinning at ∼100 K. We monitored the field dependence of direct 13C and 15N polarization for frozen [13C, 15N] urea and achieved maximum absolute enhancement factors of 240 and 470, respectively. The field profiles are consistent with polarization of 15N spins via either the solid effect or the cross effect, and polarization of 13C spins via a combination of cross effect and solid effect. For microcrystalline, 15N-enriched tryptophan synthase sample containing trityl radical, a 1500-fold increase in 15N signal was observed under microwave irradiation. These results show the promise of trityl radicals and their derivatives for direct polarization of low gamma, spin-½ nuclei at high magnetic fields and suggest a novel approach for selectively polarizing specific moieties or for polarizing systems which have low levels of protonation.

Bis-Gadolinium Complexes for Solid Effect and Cross Effect Dynamic Nuclear Polarization.

High-spin complexes act as polarizing agents (PAs) for dynamic nuclear polarization (DNP) in solid-state NMR spectroscopy and feature promising aspects towards biomolecular DNP. We present a study on bis(Gd-chelate)s which enable cross effect (CE) DNP owing to spatial confinement of two dipolar-coupled electron spins. Their well-defined Gd⋅⋅⋅Gd distances in the range of 1.2-3.4 nm allowed us to elucidate the Gd⋅⋅⋅Gd distance dependence of the DNP mechanism and NMR signal enhancement. We found that Gd⋅⋅⋅Gd distances above 2.1 nm result in solid effect DNP while distances between 1.2 and 2.1 nm enable CE for 1 H, 13 C, and 15 N nuclear spins. We compare 263 GHz electron paramagnetic resonance (EPR) spectra with the obtained DNP field profiles and discuss possible CE matching conditions within the high-spin system and the influence of dipolar broadening of the EPR signal. Our findings foster the understanding of the CE mechanism and the design of high-spin PAs for specific applications of DNP.

Frequency swept microwaves for hyperfine decoupling and time domain dynamic nuclear polarization.

Hyperfine decoupling and pulsed dynamic nuclear polarization (DNP) are promising techniques to improve high field DNP NMR. We explore experimental and theoretical considerations to implement them with magic angle spinning (MAS). Microwave field simulations using the high frequency structural simulator (HFSS) software suite are performed to characterize the inhomogeneous phase independent microwave field throughout a 198GHz MAS DNP probe. Our calculations show that a microwave power input of 17W is required to generate an average EPR nutation frequency of 0.84MHz. We also present a detailed calculation of microwave heating from the HFSS parameters and find that 7.1% of the incident microwave power contributes to dielectric sample heating. Voltage tunable gyrotron oscillators are proposed as a class of frequency agile microwave sources to generate microwave frequency sweeps required for the frequency modulated cross effect, electron spin inversions, and hyperfine decoupling. Electron spin inversions of stable organic radicals are simulated with SPINEVOLUTION using the inhomogeneous microwave fields calculated by HFSS. We calculate an electron spin inversion efficiency of 56% at a spinning frequency of 5kHz. Finally, we demonstrate gyrotron acceleration potentials required to generate swept microwave frequency profiles for the frequency modulated cross effect and electron spin inversions.

Efficient Dynamic Nuclear Polarization at 800 MHz/527 GHz with Trityl-Nitroxide Biradicals.

Cross-effect (CE) dynamic nuclear polarization (DNP) is a rapidly developing technique that enhances the signal intensities in magic-angle spinning (MAS) NMR spectra. We report CE DNP experiments at 211, 600, and 800 MHz using a new series of biradical polarizing agents referred to as TEMTriPols, in which a nitroxide (TEMPO) and a trityl radical are chemically tethered. The TEMTriPol molecule with the optimal performance yields a record (1) H NMR signal enhancement of 65 at 800 MHz at a concentration of 10 mM in a glycerol/water solvent matrix. The CE DNP enhancement for the TEMTriPol biradicals does not decrease as the magnetic field is increased in the manner usually observed for bis-nitroxides. Instead, the relatively strong exchange interaction between the trityl and nitroxide moieties determines the magnetic field at which the optimum enhancement is observed.

Topical Developments in High-Field Dynamic Nuclear Polarization.

We report our recent efforts directed at improving high-field DNP experiments. We investigated a series of thiourea nitroxide radicals and the associated DNP enhancements ranging from ε = 25 to 82 that demonstrate the impact of molecular structure on performance. We directly polarized low-gamma nuclei including 13C, 2H, and 17O using trityl via the cross effect. We discuss a variety of sample preparation techniques for DNP with emphasis on the benefit of methods that do not use a glass-forming cryoprotecting matrix. Lastly, we describe a corrugated waveguide for use in a 700 MHz / 460 GHz DNP system that improves microwave delivery and increases enhancements up to 50%.

Balance improvements in healthy subjects are independent to postural strategies involved in the training.

BACKGROUND: Adequate postural strategies have a pivotal role in ensuring balance during the performance of daily or sport activities. These strategies are responsible for the management of center of mass kinematics and depend on the magnitude of perturbations and posture assumed by a subject. RESEARCH QUESTION: Are there differences in postural performance after a standardized balance training performed in sitting versus standing posture in healthy subjects? Does a standardized unilateral balance training with the dominant or non-dominant limb improve balance on trained and untrained limbs in healthy subjects? METHODS: Seventy-five healthy subjects reporting a right-leg dominance were randomized into a Sitting, Standing, Dominant, Non-dominant or Control groups. In the Experiment 1, Sitting group performed a 3-week balance training in seated posture, whereas Standing group performed the same training in bipedal stance. In the Experiment 2, Dominant and Non-dominant groups underwent a 3-week standardized unilateral balance training on the dominant and non-dominant limbs, respectively. Control group underwent no intervention and was included in both experiments. Dynamic (Lower Quarter Y-Balance Test with the dominant and non-dominant limbs and trunk and lower limb 3D kinematics) and static (center of pressure kinematics in bipedal and bilateral single-limb stance) balance were assessed before and after the training, and at 4 weeks follow-up. RESULTS: A standardized balance training in sitting or standing posture improved balance without between-group differences, while a unilateral balance training with the dominant or non-dominant limb improved postural stability on the trained and untrained limbs. Trunk and lower limb joints range of motion increased independently to their involvement in the training. SIGNIFICANCE: These results may allow clinicians to plan effective balance interventions even when a training in standing posture is not possible or in subjects with restricted limb weight-bearing.

DNPSOUP: A simulation software package for dynamic nuclear polarization.

Dynamic Nuclear Polarization Simulation Optimized with a Unified Propagator (DNPSOUP) is an open-source numerical software program that models spin dynamics for dynamic nuclear polarization (DNP). The software package utilizes a direct numerical approach using the inhomogeneous master equation to treat the time evolution of the spin density operator under coherent Hamiltonians and stochastic relaxation effects. Here we present the details of the theory behind the software, starting from the master equation, and arriving at characteristic operators for any section of density operator time-evolution. We then provide an overview of the DNPSOUP software architecture. The efficacy of the program is demonstrated by simulating DNP field profiles on small spin systems exploiting both continuous wave and time-domain DNP mechanisms. Examples include Zeeman field profiles for the solid effect, Overhauser effect, and cross effect, and microwave field profiles for NOVEL, off-resonance NOVEL, the integrated solid effect, the stretched solid effect, and TOP-DNP. The software should facilitate a better understanding of the DNP process, aid in the design of optimized DNP polarizing agents, and allow us to examine new pulsed DNP methods at conditions that are not currently experimentally accessible, especially at high magnetic fields with high-power microwave pulses.

Direct measurement of the triple spin flip rate in dynamic nuclear polarization.

A previous study of the effect of Gadolinium doping on the dynamic polarization (DNP) of 13C using trityls showed that the rate at which the polarization builds up is almost independent of the Gadolinium concentration, while the electron spin-lattice relaxation rate varies over an order of magnitude. In this paper we analyze the polarization build-up in detail and show that in this case DNP is due to the cross-effect (CE) and that the build-up rate can be quantitatively interpreted as the rate of the triple spin flips responsible for the CE. Thus this build-up rate presents a direct measurement of this triple spin flip rate.

Role of electron spin dynamics and coupling network in designing dynamic nuclear polarization.

Dynamic nuclear polarization (DNP) has emerged as a powerful sensitivity booster of nuclear magnetic resonance (NMR) spectroscopy for the characterization of biological solids, catalysts and other functional materials, but is yet to reach its full potential. DNP transfers the high polarization of electron spins to nuclear spins using microwave irradiation as a perturbation. A major focus in DNP research is to improve its efficiency at conditions germane to solid-state NMR, at high magnetic fields and fast magic-angle spinning. In this review, we highlight three key strategies towards designing DNP experiments: time-domain "smart" microwave manipulation to optimize and/or modulate electron spin polarization, EPR detection under operational DNP conditions to decipher the underlying electron spin dynamics, and quantum mechanical simulations of coupled electron spins to gain microscopic insights into the DNP mechanism. These strategies are aimed at understanding and modeling the properties of the electron spin dynamics and coupling network. The outcome of these strategies is expected to be key to developing next-generation polarizing agents and DNP methods.

Numerical recipes for faster MAS-DNP simulations.

Numerical simulations of Magic Angle Spinning Dynamic Nuclear Polarization (MAS-DNP) have transformed the way the DNP process is understood in rotating samples. In 2012, two methods were concomitantly developed to simulate small spin systems (< 4 spin-1/2). The development of new polarizing agents, including those containing metal centers with S > 1/2, makes it necessary to further expand the numerical tools with minimal approximations that will help rationalize the experimental observations and build approximate models. In this paper, three strategies developed in the past five years are presented: an adaptive integration scheme, a hybrid Hilbert/Liouville formalism, and a method to truncate the Liouville space basis for periodic Hamiltonian. Each of these methods enable time savings ranging from a factor of 3 to > 100. We illustrate the code performance by reporting for the first time the MAS-DNP field profiles for "AMUPol", in which the couplings to the nitrogen nuclei are explicitly considered, as well as Cross-Effect MAS-DNP field profiles with two electrons spin 5/2 interacting with a nuclear spin 1/2.

Monte Carlo study of the spin-spin interactions between radicals used for dynamic nuclear polarization.

The spectrum of the electron spin-spin interactions largely determines which mechanism is responsible for the growth of the nuclear spin polarization in dynamic nuclear polarization (DNP). When electron spin-spin interactions are weak and their spectrum is narrow, the solid effect (SE) dominates the process. When they are stronger, the cross effect (CE) and thermal mixing (TM) come into play. Then a narrow spectrum favours the CE-that is an exchange of electron Zeeman energy with the nuclear spins-and a broad spectrum also TM-that is an exchange of electron spin-spin interaction energy with the nuclear spins. Moreover, the spectrum of the electron spin-spin interactions critically determines the rate of spectral diffusion of electron spin polarization across the electron spin resonance (ESR) line, and the associated conversion of electron Zeeman energy into electron spin-spin interaction energy. This way electron spin-spin interactions indirectly influence the DNP process. The present work describes Monte Carlo simulations of the spectrum of these interactions for approximately spherical radicals in glasses and analytical approximations of the simulation results. As an example application expressions for the relative strengths of the energy flows due to the CE and TM are derived.

The distance between g-tensors of nitroxide biradicals governs MAS-DNP performance: The case of the bTurea family.

Bis-nitroxide radicals are common polarizing agents (PA), used to enhance the sensitivity of solid-state NMR experiments via Magic Angle Spinning Dynamic Nuclear Polarization (MAS-DNP). These biradicals can increase the proton spin polarization through the Cross-Effect (CE) mechanism, which requires PAs with at least two unpaired electrons. The relative orientation of the bis-nitroxide moieties is critical to ensure efficient polarization transfer. Recently, we have defined a new quantity, the distance between g-tensors, that correlates the relative orientation of the nitroxides with the ability to polarize the surrounding nuclei. Here we analyse experimentally and theoretically a series of biradicals belonging to the bTurea family, namely bcTol, AMUPol and bcTol-M. They differ by the degree of substitution on the urea bridge that connects the two nitroxides. Using quantitative simulations developed for moderate MAS frequencies, we show that these modifications mostly affect the relative orientations of the nitroxide, i.e. the length and distribution of the distance between the g-tensors, that in turn impacts both the steady state nuclear polarization/depolarization as well as the build-up times. The doubly substituted urea bridge favours a large distance between the g-tensors, which enables bcTol-M to provide ∊on/off>200 at 14.1 T/600 MHz/395 GHz with build-up times of 3.8 s using a standard homogenous solution. The methodology described herein was used to show how the conformation of the spirocyclic rings flanking the nitroxide function in the recently described c- and o-HydrOPol affects the distance between the g-tensors and thereby polarization performance.

Dynamic nuclear polarization via the cross effect and thermal mixing: A. The role of triple spin flips.

In dynamic nuclear polarization (DNP) via the cross effect (CE) and thermal mixing (TM) a microwave field first reduces the polarization of some electron spins, so the electron spin system deviates from thermal equilibrium with the lattice. Next, the mutual interactions combine with their interaction with the nuclear spins to transform this deviation into nuclear spin polarization. Two approaches were introduced to describe the latter process. The fluctuating field approach considers the electron spins to fluctuate between their up and down states due to their mutual interactions. This results in a classical fluctuating field at the position of the nuclear spins, and the component of this field at the NMR frequency induces flips of the nuclear spins. The scrambled states approach considers the electron and nuclear spin states to be mixed by the hyperfine and super-hyperfine interaction. Next the mutual interaction between the electron spins induces transitions between these mixed states and thus flips nuclear spins. Some authors considered the fluctuating field approach and the scrambled states approach to be just two equivalent methods to describe exactly the same process. Other authors considered the two approaches to describe two separate processes, the former exchanging electron interaction energy, the latter transferring differences of electron spin polarization to the nuclear spins. The present work introduces a generalized approach that first calculates the mixing of electron spin states exactly. Next it considers the hyperfine or super-hyperfine interaction to induce transitions involving these mixed states and the nuclear spin states. It is found that the scrambled states approach and the fluctuating field approach are neither fully equivalent, nor completely independent processes, but rather represent two distinct limits of a single process. The former corresponding to very weak mutual interactions between electron spins and the latter to very strong mutual interactions. It extends the treatment to the whole range of mutual interactions and shows that this single process simultaneously exchanges electron Zeeman energy and electron interaction energy with the nuclear spins. Expressions for these two flows as a function of the strength of the mutual interaction are derived.

Dynamic nuclear polarization via the cross effect and thermal mixing: B. Energy transport.

The fundamental process of dynamic nuclear polarization (DNP) via the cross effect (CE) and thermal mixing (TM) is a triple spin flip, in which two interacting electron spins and a nuclear spin interacting with one of these electron spins flip together. In the previous article (Wenckebach, 2018) these triple spin flips were treated by first determining the eigenstates of the two interacting electron spins exactly and next investigating transitions involving these exact eigenstates and the nuclear spin states. It was found that two previously developed approaches-the scrambled states approach and the fluctuating field approach-are just two distinct limiting cases of this more general approach. It was also shown that triple spin flips constitute a single process causing two flows of energy: a flow originating in the electron Zeeman energy and a flow originating in the mutual interactions between the electron spins. In order to render their definitions more precise, the former flow was denoted as the CE and the latter as TM. In this article the treatment is extended to a glass containing NI equivalent nuclear spins I=12 and NS randomly distributed and oriented electron spins S=12. Rate equations are derived for the two flows of energy to the nuclear spins. The flow originating in the electron Zeeman energy-i.e. the CE-is found to lead to the same stationary state as was previously predicted by the scrambled states approach, though the rate may be smaller due to limitations imposed by conservation of energy. The flow originating in the mutual interactions between the electron spins-i.e. TM-is found to involve the full spectrum of the mutual interactions between the electron spins, while the fluctuating field approach only accounts for the component of this spectrum at the nuclear magnetic resonance (NMR) frequency. Still, TM is found to induce equal spin temperature for different nuclear spin species during nuclear spin-lattice relaxation and, at least in some cases also during polarization. It is also confirmed that TM couples local nuclear spins near the electron spins so strongly to the mutual interactions between electron spins, that they may constitute a single energy reservoir (Cox et al., 1973). Hence such local nuclear spins may have to be included in treatments of the dynamics of the electron spins.

Different radiation dose rate as radioprotection and the cross effect with chromium using in vivo somatic cells of Drosophila.

Ionizing radiation plays a key role in the adaptation of an individual organism to environmental pollution, at the same time, it has biological effects that depend on radiation intensity or dose rate (DR). Although the effect of DR has been studied in vitro, the phenomenon known as the inverse effect of DR, which indicates as it decreases that the induction of damage is greater, has not been widely studied in vivo. The present study is aimed to test 0.5 and 1 Gy in somatic cells of the wing of D. melanogaster, administered at 5.4 or 34.3 Gy/h and from 0.037 to 0.3 mM of CrO3 as conditioning treatment. No changes were found in larva-to-adult viability. A protective as well as a cross effect of pre-exposure to different DR and CrO3 concentrations against genetic damage induced by 20 Gy or 1 mM CrO3 was evident.

Verdazyl-ribose: A new radical for solid-state dynamic nuclear polarization at high magnetic field.

Solid-state dynamic nuclear polarization (DNP) using the cross-effect relies on radical pairs whose electron spin resonance (ESR) frequencies differ by the nuclear magnetic resonance (NMR) frequency. We measure the DNP provided by a new water-soluble verdazyl radical, verdazyl-ribose, under both magic-angle spinning (MAS) and static sample conditions at 9.4 T, and compare it to a nitroxide radical, 4-hydroxy-TEMPO. We find that verdazyl-ribose is an effective radical for cross-effect DNP, with the best relative results for a non-spinning sample. Under non-spinning conditions, verdazyl-ribose provides roughly 2× larger 13C cross-polarized (CP) NMR signal than the nitroxide, with similar polarization buildup times, at both 29 K and 76 K. With MAS at 7 kHz and 1.5 W microwave power, the verdazyl-ribose does not provide as much DNP as the nitroxide, with the verdazyl providing less NMR signal and a longer polarization buildup time. When the microwave power is decreased to 30 mW with 5 kHz MAS, the two types of radical are comparable, with the verdazyl-doped sample having a larger NMR signal which compensates for its longer polarization buildup time. We also present electron spin relaxation measurements at Q-band (1.2 T) and ESR lineshapes at 1.2 and 9.4 T. Most notably, the verdazyl radical has a longer T1e than the nitroxide (9.9 ms and 1.3 ms, respectively, at 50 K and 1.2 T). The verdazyl electron spin lineshape is significantly affected by the hyperfine coupling to four 14N nuclei, even at 9.4 T. We also describe 3000-spin calculations to illustrate the DNP potential of possible radical pairs: verdazyl-verdazyl, verdazyl-nitroxide, or nitroxide-nitroxide pairs. These calculations suggest that the verdazyl radical at 9.4 T has a narrower linewidth than optimal for cross-effect DNP using verdazyl-verdazyl pairs. Because of the hyperfine coupling contribution to the electron spin linewidth, this implies that DNP using the verdazyl radical would improve at lower magnetic field. Another conclusion from the calculations is that a verdazyl-nitroxide bi-radical would be expected to be slightly better for cross-effect DNP than the nitroxide-nitroxide bi-radicals commonly used now, assuming the same spin-spin coupling constants.

Spectral diffusion and dynamic nuclear polarization: Beyond the high temperature approximation.

Dynamic Nuclear Polarization (DNP) has proven itself most powerful for the orientation of nuclear spins in polarized targets and for hyperpolarization in magnetic resonance imaging (MRI). Unfortunately, the theoretical description of some of the processes involved in DNP invokes the high temperature approximation, in which Boltzmann factors are expanded up to first order, while the high electron and nuclear spin polarization required for many applications do not justify such an approximation. A previous article extended the description of one of the mechanisms of DNP-thermal mixing-beyond the high temperature approximation (Wenckebach, 2017). But that extension is still limited: it assumes that fast spectral diffusion creates a local equilibrium in the electron spin system. Provotorov's theory of cross-relaxation enables a consistent further extension to slower spectral diffusion, but also invokes the high temperature approximation. The present article extends the theory of cross-relaxation to low temperature and applies it to spectral diffusion in glasses doped with paramagnetic centres with anisotropic g-tensors. The formalism is used to describe DNP via the mechanism of the cross effect. In the limit of fast spectral diffusion the results converge to those obtained in Wenckebach (2017) for thermal mixing. In the limit of slow spectral diffusion a hole is burnt in the electron spin resonance (ESR) signal, just as predicted by more simple models. The theory is applied to DNP of proton and 13C spins in samples doped with the radical TEMPO.

Mechanisms of dynamic nuclear polarization in insulating solids.

Dynamic nuclear polarization (DNP) is a technique used to enhance signal intensities in NMR experiments by transferring the high polarization of electrons to their surrounding nuclei. The past decade has witnessed a renaissance in the development of DNP, especially at high magnetic fields, and its application in several areas including biophysics, chemistry, structural biology and materials science. Recent technical and theoretical advances have expanded our understanding of established experiments: for example, the cross effect DNP in samples spinning at the magic angle. Furthermore, new experiments suggest that our understanding of the Overhauser effect and its applicability to insulating solids needs to be re-examined. In this article, we summarize important results of the past few years and provide quantum mechanical explanations underlying these results. We also discuss future directions of DNP and current limitations, including the problem of resolution in protein spectra recorded at 80-100 K.