the future of epr - weizmann institute of science

149

The Future of EPR

Sandra S. Eaton and Gareth R. Eaton

Department of ChemistryUniversity of Denver

Denver, Colorado 80208

ContentsI. Introduction 150

A. Results of the 1987 Workshop 151B. Research Advances 152C. Goals for the 1992 Workshop 153D. A Perspective on EPR 154E. Current Themes in EPR 155F. Software 155G. Applications of EPR 155H. The Literature of Magnetic Resonance 157

II. State of the Art Lecture - New EPR Methodologies: James S. Hyde 159A. Q-Band EPR 160B. Pseudomodulation 161C. Multiquantum EPR 161D. Respondent - Melvin P. Klein 162E. Discussion 163

III. State of the Art Lecture - In Vivo EPR: Harold M. Swartz 163A. The Scope of In Vivo EPR 163B. Respondent - Lawrence Berliner 165C. Discussion 165

IV. State of The Art Lecture - FT EPR and High-Field EPR: Jack H. Freed 166A. Comparison with NMR 166B. FT EPR 166C. High Frequency EPR 167D. Respondent - Linn Belford 169E. Discussion 169

V. State of The Art Lecture - Pulsed EPR: Arthur Schweiger 169A. Comparison with NMR 169B. New EPR Detection Schemes 170C. Recent Instrumental Innovations in Pulsed EPR 173D. Respondent - David Singel 173E. Discussion 174

VI. Panel Discussion - High resolution EPR 174A. Kinetics 174B. Longitudinal Detection 175C. Signal to noise 175

150 Bulletin of Magnetic Resonance

D. Ex Vivo EPR; Aqueous Samples in Flat Cells 175E. Dielectric Resonators 175F. Small and/or Dedicated EPR Spectrometers 176

VII. Panel Discussion — In Vivo EPR and Imaging 176A. The Question of Sample Size 176B. Frequency Scaling 176C. Interpretation of In Vivo Spectra 177D. Magnetic Field and Magnetic Field Gradient Control 177E. Low Frequency and Imaging Spectrometers 177F. Nitric Oxide In Vivo 177G. Noise in FT EPR, EPR Imaging and In Vivo EPR 178

VIII. Panel Discussion — New Perspectives on Spins 179A. SQUIDs in EPR 179B. Multiquantum EPR 179C. Microwave Source Phase Noise 179D. Pulsed ENDOR 180E. Dissemination of Modern Techniques 180F. Software for Visualization of EPR Data 180

IX. Summary on Instrumentation and Methodology 181

X. The Funding Agency Perspective 181A. Questions Regarding Funding of EPR in the USA 181B. Information from the Presentation by John Beisler, DRG, NIH 182

XI. The Vendor Perspective 183A. Bruker (Dieter Schmalbein) 183B. JEOL (Jack Francis) 184C. Micro-Now (Clarence Arnow) 184D. Oxford Instruments (Mark Woolfrey) 184

XII. Summary Perspective 184A. The Horizons of EPR 185B. Where EPR is Today 185C. The Future 186

XIII. Acknowledgment 186

XIV. References 187

Vol. 16, No. 3/4 151

I. IntroductionAn NIH-sponsored Workshop on the Future of

EPR was held in Denver, Colorado, August 7, 1992,following the 15th International EPR Symposium.Participants in the Workshop included about 65 re-searchers from several countries, representatives ofsix corporations, and a representative of NIH. Thisreview of the state of EPR and expression of con-cerns, hopes, and predictions for the future is basedon the contributions of the participants in the Work-shop. Those who presented state-of-the-art lectures,served as respondents, or participated in panel dis-cussions, are identified in appropriate places in thetext. When a comment/question from another par-ticipant in the Workshop presented information thatshould be identified with that person, the person isidentified in the text. Otherwise, the questions andanswers are summarized rather than quoted. Thepanel members, and Colin Mailer, Keith Madden,Carmen Arroyo, Ralph Weber, Francisco Jent, RonMason, and Peter Hofer were particularly active inthe discussions.

Some references to the literature have been pro-vided to lead the reader to more extensive discus-sions. In addition to the well-known review series inmagnetic resonance, five recent books provide sum-maries of individual topics in EPR (1-5).

A Workshop on the Future of EPR has a lot tocover because EPR is such an extensive field. Ourfocus for this Workshop is a vision of the future.Scientists should not shy away from predicting thefuture of their field. In fact, scientists have to bebetter at this than the self-styled futurists. It istheir profession - scientists spend much of their timepreserving the best of the past and creating a betterfuture.

A. Results of the 1987 WorkshopFive years ago researchers gathered to say what

the cutting edge results from research laboratoriesimplied for the future needs of EPR spectroscopy(6,7). At the first Workshop in 1987 there was alot of controversy and a lot of discussion, but thereemerged from the Workshop a fairly clear statementthat the EPR field as it was then certainly neededthe very best sensitivity and signal-to-noise (S/N)that one could get in the standard X-band regionof the spectrum. This was a very high priority.

Also important, was to exploit the information thatcould come from broadband EPR. Researchers re-ally wanted a frequency range of 60 MHz to greaterthan 250 GHz, but suggested that 1-18 GHz wouldbe a nice goal for the commercial instruments thatwould end up in all laboratories. As a short termmatter the goal was narrowed to 3-15 GHz. In sum-mary, the designs that emerged from the 1987 Work-shop included two types of spectrometer, with thefeatures listed: First, a 9-9.6 GHz EPR optimizedfor sensitivity and S/N. Second, a broadband EPRspectrometer, with the features:

• 3-15 GHz (1-18 GHz preferred; ultimately, 60MHz to >250 GHz)

• not an automatic frequency control (AFC)-locked cavity system (except for in vivo stud-ies)

• encode the data from the whole modulationcycle

• solid state microwave source

• resonators designed specifically for applica-tions

• open architecture

• computer control (with human interface);

• standard computer interface

• computer(s) integrated with the hardware torun the spectrometer

• attached workstation

• two bridges: continuous wave (CW) and satu-ration recovery (SR) electron-spin-echo (ESE)and Fourier transform (FT)

• 1-2 GHz and 35 GHz accessories

• electromagnet-based

There were a number of issues and concerns withthese choices:

• narrow-band components versus broadbandcomponents

• loss of sensitivity with broadband components


• loss of magnetic field homogeneity away fromthe center field

• maybe there would need to be a g=2 spec-trometer and a metals spectrometer

There was vigorous debate in 1987 aboutwhether the limited R&D effort available for EPRshould be applied to ultimate sensitivity X-bandCW; broadband EPR; or pulsed EPR.

In 1987 many other needs were expressed, fromneed for a better high-power travelling-wave-tubeamplifier (TWT), to need for a better measure ofwhat the temperature really is at the sample, toneeds relating to educating the next generation ofscientists who will apply EPR to solving importantproblems in materials science and biomedical prob-lems. Some of the additional needs expressed in1987 were discussed in the report on that Workshop(6).

The design criteria for an instructional EPRspectrometer synopsizing the desires expressed in1987 would be:

• a 3 or 4-inch electromagnet, with power supplythat could be run from a standard 110 V walloutlet;

• no cooling water required for operation for upto 3 hours at 3400 G;

• >600 G scan range;

• 100-150 mG homogeneity over the sample;

• microwave performance of the Bruker EMS104;

• data output in a format that students couldtake home to work with on a PC or Mac.

While variable-temperature, etc., are nice, as apractical matter one would not do much of this inan undergraduate lab. The principles of rigid vs.fluid solution can be demonstrated with two sam-ples, rather than freezing one sample. Membranemelting can even be done with two different samplesrather than one sample as a function of temperature.

With regard to proposals concerning instruc-tional and routine spectrometers, there was contro-versy: some people think there should not be a large"low tech" market for EPR (as there is in NMR) be-cause of the relative spectral anisotropies, and theattendant spectral interpretation difficulty.

1. Progress Since 1987

In spite of the gap between aspirations and real-ity, there has been an incredible amount of progressin the past five years (Table 1). The prototypeof the Bruker ESP380 pulsed EPR spectrometerwas on display at the Symposium the week of the1987 Workshop. Now this has matured into an in-strument that can revolutionize EPR spectroscopyfor those who rely upon commercial instruments.Bruker also responded to the needs expressed at thefirst Workshop for a low-cost instrument with theECS 106. Bruker has produced the EMS 104, the firstEPR designed for quantitative analysis. JEOL hasproduced a pulsed EPR spectrometer also, and hasdeveloped low-noise solid state microwave sources sothat they do not use klystrons in their spectrome-ters. A small EPR spectrometer developed in Russiais being marketed by Norell in the USA. SumitomoSpecial Metals has two versions of very small EPRspectrometers for instructional use. A small EPRwhich nearly matches the specifications set forth in1987 for an instructional EPR spectrometer was ondisplay by Micro-Now at the Symposium immedi-ately preceding the Workshop. Note that these cri-teria were strictly for a CW spectrometer. Thereis need to introduce time-domain techniques to stu-dents if EPR is to prosper. Software specifically forEPR has been enhanced greatly by the efforts ofBruker and of Scientific Software Services. OxfordInstruments introduced new products for tempera-ture control in response to needs expressed at thefirst Workshop.

Progress since 1987 Workshop has been striking.EPR spectroscopists recognized the benefits of bet-ter communication, and an enhanced community ofusers, with the result that the International EPR(ESR) Society formed. It now has ca. 1000 mem-bers in >35 countries.

B. Research Advances

The advances in EPR during the past five yearshave been phenomenal. A totally new branch ofEPR, multiquantum EPR (MQEPR), has been de-veloped by James S. Hyde and his colleagues, sonow we should categorize EPR in three modes: CW,pulse and multiquantum.

There has been a rebirth of spin labeling (a tech-nique which had become thought of as "the old

Vol. 16, No. 3/4 153

Table 1: Summary of Commercial EPR Products

BrukerECS-106 low-cost EPRESP 300E fully computer-controlled research spectrometerESP 380 pulsed ESE and FTEPR, first demonstrated in 1987EMS 104 first EPR designed for quantitative analysisPulsed ENDOR and stochastic ENDOR

JEOLpulsed EPRL-band EPRPC-based EPR data stationlow-noise Gunn diode sourcecavity for aqueous samples

Norellsmall EPR built by St. Petersburg Instruments, Ltd.

Micro-NowModel 8400 on display at the 1992 EPR Symposium

Sumitomo Special Metals - Spin-X and Spin-XXScientific Software Services

PC-based acquisition/manipulation software for Bruker, Varian, and Micro-Now spectrometersMedical Advances - resonators, and development effort on an S-band bridgeOxford Instruments - variable temperature accessories

ESR900 can now use nitrogen as well as heliumautomatic transfer linesCF935 dewars for wide range of S- to Q-band cavitiessensor to measure temperature at the sample position

Wilmad Glass - quartzware accessorieshigh precision EPR tubesa standard sample for EPR

stuff') due to combinations of loop gap resonators(LGRs) and site-directed mutagenesis.

Both academic and industrial laboratories havedeveloped lower noise oscillators.

Indeed, there have been so many major advancesin EPR since 1987 Workshop that there is room onlyto list a few keywords (Table 2). The vendors ofEPR equipment and software have an almost im-possible task of predicting the needs of those whodevelop and those who use EPR. Part of the pur-pose of this Workshop was to have researchers sharetheir aspirations with manufacturers.

C. Goals for the 1992 Workshop

With this background, the goals for the 1992Workshop were:

• Participants would learn about the power ofnew spectroscopic techniques that they couldapply in their research.

• Critique the predictions and desires expressedduring the first Workshop in 1987.

• Provide a new, updated, perspective on theEPR instrumentation needs of research.

• Present a new set of predictions and criteria.


Table 2: Listing of Recent Research Advances

multiquantum EPRmultiple resonance

(especially pulsed ENDOR and multiquantum ENDOR)multifrequency

saturation recoveryENDOR (especially high frequency CW ENDOR)high-field EPR and high-frequency EPRlow frequency EPRspin echo at frequencies other than X-band

new types of resonatorsLGR and bridged LGR, dielectric resonator

multidimensional imagingin vivo EPRdetection of radical adducts in biological fluidsrebirth of spin labeling via

site specific mutagenesis and oximetrymany new pulsed EPR techniques

FT-EPR (Bruker pulsed FTEPR)pulsed field gradientspulsed electron nuclear double resonance (ENDOR) using Davies sequenceFT-electron-electron double resonance (FT-ELDOR)electron spin transient nutationESEEM sequences for improved modulation depths, etc.

ENDOR signals in small, lossy protein crystalslow phase noise oscillatorslow noise microwave preamplifiersdigital oscilloscopes for signal processinguseful level of computer power at each spectrometermuch more sophisticated data acquisition and analysis software availablepseudomodulation - modeling of the transfer characteristics of an instrumentrebirth of solutions to biological problems by applying the above advances

To focus discussion, the Workshop was or-ganized around seeking answers to two ques-tions:

• What are the EPR instrumental or softwarelimits to important experiments in science?

• What are the technology limits on instrumen-tation and software for EPR?

These goals were not as crisply addressed as washoped in advance, in part because some participants

were too focused on what they had accomplishedwith limited resources. In addition, EPR spectro-scopists have become acculturated to cleverly work-ing within boundary conditions imposed by com-mercial instruments and funding, and had difficultyexpressing what these limits are.

D. A Perspective on EPR

As an overall perspective on EPR, consider that sofar most EPR has been CW, and most studies have

Vol. 16, No. 3/4 155

been done in the linear response region, using homo-geneous magnetic fields, using magnetic field scans,and almost all of this has been done in TE102 cavi-ties (Table 3). Most of what we celebrate as bench-mark results are exceptions to this generalization.

EPR is becoming (as was revealed at the 15th In-ternational EPR Symposium in the days precedingthe Workshop) multi-frequency, multi-dimensional,multi-everything; it is non-linear, time-domain;most of the experiments now are being done withhome-built resonators designed specifically for thepurpose; often experiments are being done in gradi-ent fields for imaging or in vivo (Table 4). Thesechanges are becoming necessary because of themany applications for EPR.

E. Current Themes in EPR

Pulsed, fourier transform, non-linear CW, imag-ing, and in vivo techniques are increasingly impor-tant. Resonators are being designed to fit the needsof the experiment, rather than fitting the exper-imental design to the resonator (or deciding notto do the experiment). Very low (e.g., 250 MHz)and very high (e.g., 250 GHz) frequency EPR haveexpanded our view of spins. Multiple pulse tech-niques are bringing to EPR powerful insights anal-ogous to those that are becoming commonplace inNMR. Imaging and in vivo techniques are letting ussee EPR spectra at each location in space, permit-ting us to perform, for example, oximetry in livinganimals. EPR without magnetic field modulationopens new vistas, in saturation recovery EPR, fast-response EPR, and multiple-quantum EPR. These,and other current themes in EPR are listed in Table5.

F. Software

In modern EPR, software is so important that itdeserves special emphasis in this report. One mustpay as much attention to the quality of the soft-ware as to the hardware. Increasingly scientists seethat software is a central and crucial part of EPRspectroscopy. Color graphics displays can help vi-sualization of the information content of the EPRdata, but can also deflect attention from the com-putational artifacts. The field needs a series of well-posed problems against which new software can betested. For example, in the field of image analy-

sis there are standard problems such as the Shepp-Logan head phantom, against which each new algo-rithm is tested.

The crucial issues with regard to quality of EPRsoftware are highlighted in Table 6.

G. Applications of EPR

There are many fascinating aspects of spin physicsto be explored, and impressive new tools with whichto explore them. However, the funding needed forthese exploratory voyages will come largely becausethe insights to be gained have such important ap-plications.

To emphasize the immediate relevance of EPR tobiomedical research, some of the applications in Ta-ble 7 are categorized by NIH Institute. Far beyondthe simple characterization of organic free radicalsand transition metal complexes, there are applica-tions in dental research, research on aging, researchon the eye, etc. The list is of things that people havealready done. This Workshop was more concernedto look toward the future - applications that are notwell-known yet.

Consider for a moment the EPR spectrum in Fig-ure 1. The spectrum in Figure 1 has terrible S/N,but it is an important sample - a sample of braintissue. The purpose in showing it is not to discussthe questions of artifacts, etc., but to point out thatthe interpretation would be enormously improved ifone could achieve at least an order of magnitude im-provement in S/N. Then, consider how much moremeaningful it would be if it were in vivo insteadof dissected tissue. Then consider potential appli-cations to heart, lung, etc. As we look toward thefuture, we should envision making this measurementof the EPR spectrum of brain tissue not on excisedtissue but in vivo, localized, with at least an order ofmagnitude improvement in S/N, using the panoplyof EPR spectroscopy techniques that have been re-ported at the EPR Symposium.

H. The Literature of Magnetic Reso-nance

According to information provided by ChemicalAbstracts, about 4 times as many NMR as EPRpapers were cited in CA in 1991. In 1986 the ratiowas 3.2. It is possible that the leveling off in recentyears reflects a maturing of CW EPR techniques to


Table 3

EPR has beenCWLINEAR RESPONSE REGIONMAGNETIC FIELD SCANHOMOGENEOUS MAGNETIC FIELDTE102 CAVITY

Table 4

EPR is becomingMULTI-FREQUENCYMULTI-DIMENSIONALNON-LINEARTIME DOMAINPURPOSE-BUILT RESONATORSHETEROGENEOUS SAMPLESGRADIENT FIELDS

3225.0 3235.0 3245.0 3255.0 3265.0 3275.0 3285.0 3295.0 3305.0 3315.0 3325.0

Figure 1: X-band EPR spectrum of a piece of excised brain tissue, frozen and kept at ca. -70°C until theEPR spectrum was recorded at — 160°C.

Vol. 16, No. 3/4 157

Table 5: Current Themes in EPR

Multifrequency and multi-dimensional EPR.Very low frequency EPR.Low-frequency in vivo spectroscopy.Very high frequency EPR.High-magnetic-field EPR.Non-linear CW EPR.EPR imaging.Oximetry combined with imaging.EPR without magnetic field modulation.Multiple-quantum EPR.Saturation recovery EPR.Fast-response EPR.Emphasis on the time domain as well as on the frequency domain.Pulsed EPR.Multiple pulse techniques.Fourier transform EPR.New EPR pulse sequences.Magnetic field dependence in ESEEM studies.Multiple frequency electron spin echo.Applications of ESEEM to metalloenzymes.Practical aspects of spectrometer construction.Resonators designed to fit the needs of the experiment.Slow-wave and non-resonant microwave structures.Design and construction of loop-gap resonators.In vivo EPR.Spin-trapping studies in vivo.ENDOR of metal ions.Software standards and portability in the EPR community.Calculational and experimental aspects of molecular motion.Mathematical methods for the interpretation of time-domain EPR.Interpreting electron spin echo data.EPR simulation problems.

the point that EPR does not get mentioned in thetitle or abstract, even though it was central to thepaper.

There is a troubling concern among EPR spec-troscopists about financial support. Many peoplefeel that the applications of EPR are more impor-tant than has become common knowledge. One per-spective uses number of published papers as a mea-sure of the overall importance to science. Table 8compares numbers of papers published in that partof science that explores electron spins and that part

of science that uses other analytical methodologies.These numbers are just for the topics covered inChemical Abstracts, which covers only about 12,000journals (a small subset of science). Over a five-year period EPR has grown but some other topicsare growing very rapidly. Vendor decisions abouttheir allocation of effort and resources, and fundingagencies deciding upon allocation of resources, areresponses to perceptions of whether this quantifica-tion of journal articles also reflects importance.

C. P. Poole, in Vol 4, no. 2 of the EPR Newslet-


Table 6: Software Issuesa

Function (applicability, boundary conditions)Performance (accuracy, speed, throughput)Operational Characteristics (user friendly?!)InstallabilityData Security and ProtectionCompatibility (and migratability)Serviceability (updating, etc.)DocumentationSupport (especially when "locally written")

'(Some would like to add "infallibility" to this list!)

ter, (August 1992) reviewed the EPR (ESR) liter-ature covered in Physical Abstracts, Georef, Med-line, Chemical Abstracts, etc. His review revealsthat there are some 55,000 papers that have EPRor ESR in the title or abstract. Only a hundred ofthem are about ELDOR. But ELDOR is very im-portant. One can't apply these numbers directly toinform funding decisions. However, the numbers arereadily available, and will be used (and misused),so people concerned about planning for the futureshould be aware of them and learn to use them inan intelligent way.

II. State of the Art Lecture- New EPR Methodologies:James S. Hyde

Since the 1987 Workshop there has been a majoradvance in Q-band (35 GHz) EPR technology (8,9). In the past Hyde has focused the community ongoing to lower frequency (S-band or L-band) (10-14), but at this Workshop Hyde presented an em-phasis on going to higher frequency. The messageis the same - there are advantages in doing EPRaway from X-band. The vendors should lead withappropriate products for research. Multifrequencycapability should be widely available.

A. Q-Band EPR

A recent paper in RSI (9) brought several recentadvances together to greatly improve Q-band per-formance. The contributing advances were each first

developed for or demonstrated at a lower EPR fre-quency, but now they jointly revolutionize Q-bandEPR. The key contributors are low-noise microwavesources, loop-gap resonators, low-noise GaAsFETmicrowave preamplifiers, and pseudomodulation forresolution enhancement. The basic ideas were ex-pressed at the Workshop 5 years ago. When tak-ing advantage of modern low-noise GaAsFET mi-crowave amplifiers, overall system improvement re-quires also decreasing the phase noise of the os-cillator (15, 16, 17). The improved Q-band spec-trometer incorporates two essential components - aGaAsFET preamplifier and a Gunn diode source.There is also a physically small 125 cm long refer-ence arm electrical length equalizer that was createdin a block by making two halves with a numericallycontrolled mill and screwing them together. Thefollowing equation, from the book by Robbins (18),summarizes the phase noise problem, as the phasenoise density to carrier ratio:

N,op 1 FkT /fo\2

P 2P4QUWwhere Nop is the phase noise at frequency fm rela-tive to a reference frequency fo; F is the noise figurecharacteristic of the device, Q is the quality factor ofthe tank to which the device is coupled. The powerP is on both sides of the equation. The key messagefrom this equation is that the phase noise increasesas the square of the microwave frequency, and de-creases as the square of the Q of the cavity to whichthe device is coupled.

Hyde used a high-Q TEQH cavity with the Gunn

Vol. 16, No. 3/4 159

Table 7: Applications of EPR

Materials SciencesMagnetic interactions

polymer super-paramagnetsferrimagnetsferromagnetsanti-ferromagnets

SuperconductivityConducting polymers

Chemistry and PhysicsDetermination and characterizationSpin distributionsOrbital interpretationsKineticsSpin trapping

Topics arranged in accordance with NIH Institutes:

General BiomedicalStudy normal and abnormal physiological function and disease states

as directly, and as non-invasively as possible.Characterize metalloproteins, motion of biomolecules, free radical production, etc.Measurement of O2 in each organ system.Spin labeling to study organ-specific macromolecules and reactions.

Heart, Lung, and BloodOxidative reactions - ischemia and reperfusion injuryStudy directly the free radicals in heart tissueRadical generation during cardiac surgeryOxidative damage of lipoproteinsRadicals in phototherapyDetection of NO2 exposure

Diabetes and Digestive and Kidney DiseasesDiabetes mellitus, Type IIschemia-reperfusion gastric lesionsThe liver and kidney, along with the heart, contain the highest

concentrations of free radicals (other than pigmented tissue).Halocarbon metabolism

Arthritis and Musculoskeletal and Skin DiseasesInflammationMagnetic resonance imaging of extremitiesCharacterization of contrast agents for MRISpin labeling study of muscle function


Table 7: Applications of EPR (continued)

CancerFree radical generation - cancer initiation and promotionVascularization and tumor necrosisIn vivo measurement of oxygen concentration, as a function

of growth of tumors, and in relation to therapyToxicity of anticancer drugs (e.g., AZQ)

Neurological Disorders and StrokeBrain ischemiaMembrane studiesThe Parkinson-like impact of MPTP has been postulated to involve free radicals.Role of neuromelanin in Parkinson's disease

AgingRadical reactions in the aging processFree radical reactions implicated in Alzheimer's disease.

Dental ResearchRadiation-induced defects in teethFree radicals in diseased teethDosimetry based on radiation-induced radicals in teeth

Eye InstituteStructure and dynamics of rod outer segments, rhodopsin, etc.Free radicals in Green's melanoma

Table 8a

Subject

atomic spectroscopygas chromatographyhigh performance liquid chromatographyinfrared spectroscopy (organic aspects)infrared spectroscopy (physicochemical aspects)mass spectrometryRaman spectroscopyultraviolet and visible spectroscopyX-ray analysis and spectroscopycarbon & heteroatom NMRproton magnetic resonancesolid state NMRelectron spin resonance (chemical aspects)

Number of abstracts198647422819373822715454284027934140405543296250

3329

199148852762426424066915474635904326448356158707895

3834

aThe data in this table were provided by Chemical Abstracts Service and are based on the number of abstractsin their CA Selects categories.

Vol. 16, No. 3/4 161

diode oscillator. The phase noise turned out to beabout 23 dB lower than that of the klystron used inthe Varian Q-band EPR spectrometers. To make anoscillator a functional unit of a spectrometer one hasto have it respond to the 70 kHz AFC (automaticfrequency control) system. This was accomplishedwith piezoelectric devices, since the displacementsneeded at Q-band are very small.

Another aspect of phase noise is that its impacton the overall system can be reduced by decreasingthe demodulation of phase noise by the resonator.This can be done by using a low-Q resonator, suchas a LGR.

The LGR implementation used at Q-band is cou-pled to the waveguide via an iris. The microwavesare coupled into the large hole of the LGR first, andthen into the small hole, so the coupling is effec-tively a 2-step transformer. The LGR holds ca. 30nL of liquid sample. The phase noise contributionto the noise in the detected EPR signal is 13 dBbetter with the LGR than with the standard TEoncavity at Q-band.

The Varian detection system had a higher noisefigure than had been realized, and when the low-noise preamplifier was used a factor of 10 to 20 im-provement was realized over a wide range of con-ditions. At high power the noise is from the os-cillator. At low power the signal is 25 dB higher(due to the GaAsFET), but the noise is only 14 dBhigher. Note that with the low noise amplifier thesystem becomes more sensitive to phase noise, be-cause other noise sources are less important.

Table 9 (9) contains the main lessons from thiswork: the power was adjusted to get the largest sig-nal from the sample. The best geometry known atX-band yielded S/N = 1815. The best geometryknown at Q-band yielded S/N = 86, about 20 timesworse. The minimum number of spins detected wasreduced by a factor of 200 from X- to Q-band. Ona molarity basis nothing beats a flat cell in a TMcavity at X-band - it is 200 times better than the op-timum at Q-band on a molarity basis. These swingsof 200 either way create opportunities to optimizean EPR measurement for a particular problem. Theimprovements in the Q-band system have made itabout 10 times better than the old system for de-tecting nitroxyl radicals in aqueous solution.

In a Q-band saturation transfer EPR (STEPR)study Johnson and Hyde noted that in the disper-

sion mode the signal intensity increased by a factorof 10, and the noise increased by a factor of 10 (20).The resultant S/N was about 10 times worse thanhad been demonstrated at X-band. With the recentimprovements in the Q-band S/N, it can now be pre-dicted that one should be able to achieve equivalentS/N in STEPR experiments at X-band and Q-band.

B. Pseudomodulation

The use of pseudomodulation (21, 22) to providemore features in the spectrum that can be parame-terized, combined with dispersion mode STEPR inthe new Q-band system, make possible major ad-vances in the use of STEPR. Pseudomodulation isthe convolution of a sinusoidally modulated deltafunction with the digitized data.

fn(x, t) = f(x) * 6(x — (ax/2)cosu;t) = Efn(x)ncosnu;xtfn(x)

These terms are derivative-like terms convoluted byfilter functions:

This filter function is rather like a Gaussian filter- it is sharp in one domain and doesn't ring in theother domain.

This is a formal expression of what happens whenone has field modulation in an EPR spectrome-ter. When one uses pseudomodulation one getsthe derivative effect of the modulation simultaneouswith filtering, with a distortion that is about thesame as would be caused by the field modulationitself.

C. Multiquantum EPR

Multiquantum EPR (MQEPR) (23-28) is an ex-citing new opportunity. It looks especially promis-ing for Q-band because operation of Q-band EPRsystems at liquid He temperature is very difficultwith 100 KHz magnetic field modulation. Magneticfield modulation is a severe technical problem for thedesign of EPR resonators. For the future one should


Table

P(mW)a

S/Nactive volumeof sample (//L)no. of spins inthe active volumeminimum detectableconcentration (M)minimum detectableno. of spinsb

9: CW EPR Sensitivity Comparisons

X-BandTMno cavitywith flat cell

651815

162

1.6xlO14

8.8xlO~10

8.8xlO10

LGR

1186

1.42

1.4xlO12

8.6xlO~9

7.5xlO9

Q-BandTEon cavity

4.1103

0.31

3xlOu

1.5xlO~8

2.9xlO9

LGR

0.1686

0.031

3xlO10

1.9xlO"8

3.5xlO8

aIncident power yielding most intense signalbExtrapolated to S/N = 1. For a single line (note that this data is for the 15N doublet) the minimumdetectable number of spins would be 50% of the value in the table.

consider multiquantum EPR as a practical alterna-tive to magnetic field modulation. One could pseu-domodulate to get the normal derivative display. In-deed, multiquantum EPR (MQEPR) is proposed formany types of experiments, such as high pressure,low temperature, etc., where it is technically diffi-cult to get modulation to the sample.

The bridge for MQEPR uses two sources lockeda specific frequency apart. Irradiation with two mi-crowave frequencies is equivalent to irradiating witha single frequency that has been sinusoidally mod-ulated. Non-linear response of the spin system canresult in intermodulation sidebands, which can bedetected. The outputs are the multiquantum transi-tions, which can be combined in various ways to getuseful displays. MQEPR may be a useful method-ology in the future of Q-band EPR.

Multifrequency saturation-recovery (SR) EPRmeasurements of Ti of nitroxyl radicals in fluid so-lution have been measured from ca. 2.5 GHz to 18GHz. Ti has been found to be linearly dependent onmicrowave frequency. If the lengthening continuesto 35 GHz many EPR experiments (SR, STEPR,MQEPR) will work better at Q-band than at lowerfrequencies. Everything is handier for liquid phaseEPR if the Tis get longer. The construction of aSR EPR spectrometer at Q-band is now practical

because pin diode switches and other componentshave improved enough.

All of these advances taken together (low phasenoise sources, low-noise preamplifiers, MQEPR, andpseudomodulation) lead to the prediction that in thenext five years Q-band EPR will increase in signifi-cance.

Hyde designed the Varian Q-band system in 1962in the V-line series of spectrometers. In 1970 thiswas converted to the E-line series of spectrometers.About 10 of these were sold each year, making acommercial success within the small EPR market.The new Q-band design produced in Hyde's lab isclose to a commercial design. Drawings have beendistributed to labs that have requested them. Thishas been produced with federal grant funding asso-ciated with the mission of the National BiomedicalESR Center.

D. Respondent - Melvin P. Klein

In photosynthesis research EPR signals extendover thousands of gauss and have to be observedat temperatures below 10 K. G-anisotropy and ex-change coupling cause signals to be spread over widemagnetic field ranges. Much of biological spectros-copy has to be done at very low temperature. With

Vol. 16, No. 3/4 163

present Q-band systems it is difficult to get below20 K. The variable temperature technology is an im-portant current effort in which there is need for a lotof development. There is a severe problem gettingmagnetic field modulation to the sample - with onedewar assembly Klein could get only a few mG ofmodulation at the sample. Since the spectra extendover a couple of kG, the very small magnetic fieldmodulation does not provide much S/N. The idea ofmultiquantum spectroscopy to enable one to scan atrue absorption curve is very attractive for studyingthese broad signals.

The critical factors in the development of theNMR field were the use of multiple resonance (e.g.,13C while irradiating 1H), then the FT techniques,and now the various multiple pulse technologies. Wenow see a parallel evolution going on in EPR - forexample, in the recent work of the Freed laboratoryand the Schweiger laboratory. An important ques-tion is the extent to which these techniques can becombined with the MQEPR Hyde is developing toget even better insights. Most of the work of theHyde laboratory has addressed CW and SR EPR.Pulsed methods have a very important future, in-cluding at high frequency.

It is always helpful to be able to use smaller sam-ples, so new resonators that can be more efficientand more effective are very important.

Higher-Q resonators have been made using su-perconducting materials. Possibly the use of high-Tc materials will help stabilize solid-state sources.

E. DiscussionFor earlier types of EPR spectrometers, Har-

vey Buckmaster and coworkers (29, 30) analyzedthe relation between noise and balance of microwavebridges that incorporate a magic T. They also mea-sured the characteristics of crystal detectors andthe improvements obtainable with phase-lock mi-crowave frequency stabilizers. The sensitivity in1967 of a spectrometer in Buckmaster's lab at 35GHz was the same as at 9 GHz.

Buckmaster has always used oscillator synchro-nizers to decrease the source phase noise. The spec-tral purity of the sources is better than 10 Hz at 35GHz. In his spectrometers, the use of a circulatorin a bridge does not give enough bridge balance toachieve the needed phase noise. To achieve the 100dB balance needed it was necessary to use a magic

T, adjust the impedance of the arms, and use crit-ical coupling to the resonator. The bridge balancedepends on the Q of the resonator. Most commer-cial oscillator synchronizers cannot be used with 100kHz magnetic field modulation; one has to use muchlower frequencies, of the order of 1 to 10 kHz. Proofthat this system works well is the fact that up to theavailable 1 W source power the S/N is proportionalto power. The system does not have a microwavepreamplifier. Description of the 35 GHz system wasnot published because it was done exactly the sameway the 9 GHz work was done, with comparableresults.

Twenty years ago Roger Isaacson had resultswith oscillator synchronizers similar to those re-ported here by Buckmaster. Isaacson emphasizedthat the key goal is to decrease klystron noise. It iseasy to stabilize klystrons with crystal-locked oscil-lators. Beginning many years ago they have per-formed EPR with 4 Hz modulation frequency oflight in photosynthesis, where long signal decaytimes don't allow higher frequencies. They were ableto get the noise figure quite low by using a crystal os-cillator lock on the klystron. Jack Freed uses phaselocked oscillators to produce low noise at 250 GHz.

Hyde disagrees with the statement that a cir-culator cannot be used to achieve low phase noise.Roger Isaacson also agrees that a magic T is notneeded. D. A. Knoll in Hyde's lab worked on im-proving isolation in circulators (38). Instinctively,one wants to improve the isolation of the circulatorby the amount of the gain of the GaAsFET am-plifier. This is not attained by most commercialcirculators. Colin Mailer reported that he recentlybought an X-band circulator with high isolation at9.0 GHz (Pacific Microwave Technology, Camarillo,CA XYG1044-50).

III. State of the Art Lecture - InVivo EPR: Harold M. Swartz

A. The Scope of In Vivo EPR

Exciting in vivo EPR is being done in manylaboratories around the world (4, 39-49). Some3'ears ago it appeared that in vivo EPR imaging wasnot going to be worthwhile, but it is now provid-ing important new information. Unlike NMR imag-ing, where the high proton density in the body can


be used for the image, in most cases EPR imag-ing requires adding spins to the biological system.This apparent disadvantage is an advantage in someclasses of experiments, since there is no backgroundinterfering signal. Thus, one can know what isadded, and sometimes direct it to the location inthe body where one wants it.

The scope of in vivo EPR encompasses (4):

• low frequency, low resolution, EPR imaging invivo

• high resolution microscopic imaging in vitro

• in vivo spectroscopy, with and without spatiallocalization

Important information can be obtained from invivo imaging even if the resolution is low. The key isto keep in mind the biological goals of the measure-ment. High resolution microscopic imaging of bio-logical systems is difficult to do at frequencies below9 GHz. A useful perspective on in vivo EPR is thatimaging and high resolution spectroscopy are differ-ent ends of a continuum of multidimensional spec-troscopy. For a particular problem one optimizes atradeoff between spatial resolution and spectral res-olution. This is an important problem that needs tobe addressed over the next few years.

There are a variety of detector configurations forin vivo EPR. The best detector is the one that givesthe best result for a particular experiment. Theoptimum might be a surface coil, a cavity, a LGR,a coupled loop, an implanted loop or antenna, etc.

Increasingly, the information one can expect toget from in vivo EPR is the full spectrum that onecan get from non-in vivo EPR of model systems. Inaddition, one gets information that is pertinent tocomplex tissues. This includes:

• oximetry

• distribution of MRI contrast agents

• distribution of spin-labeled drugs

• redox metabolism

• detecting reactive intermediates via spin trap-ping

• biophysical measurements such as fluidity

In vivo EPR can be accomplished with a straight-forward L-band bridge and resonator interfaced to acommercial spectrometer. The main technical prob-lem is water. There is no optimum microwave fre-quency - high frequency is desired for S/N and lowfrequency is desired for penetration of the body. Awide range of frequencies needs to be available soone can select for a particular application. The 250MHz spectrometer in Halpern's laboratory is prob-ably as low as will give useful S/N for in vivo EPRfor small animals.

Naturally occurring radicals are not at highenough concentration to be studied with currentEPR technology. Added radicals are of two types:(1) soluble radicals such as nitroxides, which dis-tribute more or less uniformly, albeit with sometargeting possible though not yet well exploited;and (2) particulate species, which are well localized.Each has advantages and disadvantages. The opti-mum type of paramagnetic material will depend onthe experiment. Nitroxides continue to be impor-tant because there is a lot of flexibility in design anda lot of information has been accumulated about ni-troxides. As the focus on specific targeting of spinlabels increases, there will be increasing dependenceon organic chemists to construct the specific labelsneeded.

By using nitroxide radicals and surface coils, onecan monitor accumulation in organs such as liverand bladder, and one can monitor redox metabolismas well. In addition to pharmacokinetics, one canstudy, via the effect on the EPR signal, temperatureand oxygen concentration. The future of measure-ment of temperature and oxygen concentration invivo by EPR is bright. EPR is probably the besttechnique available for oximetry in vivo. The mea-surement of oxygen concentration is very importantmedically, and the existing methods do not give theneeded information, especially at medically signifi-cant low levels.

New particulate probes for oxygen concentrationbased on lithium phthalocyanine, fusinite, or car-bohydrate chars report oxygen concentrations, viaEPR line broadening, at very low oxygen levels (42,43, 46-48). They appear to be largely inert (non-toxic) in vivo. The EPR linewidth response to oxy-gen of a fusinite sample has been shown to be re-versible in vivo over a period of six months. Mea-surements of oxygen concentration in heart mus-

Vol. 16, No. 3/4 165

cle have been performed. The potential is clearlypresent for oximetry of tumors to provide clinicallyrelevant parameters to tailor radiation therapy andchemotherapy.

Aspirations for the future include performing si-multaneous assays of multiple sites using magneticfield gradients. The possibility of this has been es-tablished. The use of EPR to measure oxygen con-centration may be the first to reach routine clinicalapplication.

The main focus for in vivo EPR in the future islikely to be in the areas listed below. These are areasin which EPR is likely to provide useful information,and information that is not likely to be as readilyaccessible by other techniques.

• biophysical parameters, similar to those usedfor in vitro systems

• pharmacokinetics, using the paramagneticspecies as the tracer

• redox metabolism, using metabolism of nitrox-ides as the parameter

• oximetry, emphasizing repeated non-invasivemeasurements in tissues

• viability of cells

• temperature and distribution of temperature

B. Respondent - Lawrence Berliner

In vivo spectroscopy involves engineers, chemists,and medical professionals. The future dependsrather strongly on the chemists, because in solv-ing specific biomedical problems a major difficultyis producing a specific paramagnetic probe.

L-band is the more appropriate frequency if youwant to put a small animal into a spectrometer. Invivo EPR is a wonderful technique for studying thehealth of mice or rats, and it has the potential ofbeing applied to larger animals and, perhaps, pa-tients. Ex vivo EPR, e.g., on biopsy samples, onblood, or on fluid emissions already can yield im-portant results for larger animals (humans). De-pending upon sample size, one might use X-band, S-band, or L-band spectroscopy. Unfortunately, theseapplications are limited so far to labs that haveenough engineering support to build their own res-onators, since commercial instrument vendors are

not supporting the instrumentation needs of thisarea. JEOL is working with a few labs in Japan,but no such industrial collaborations are known inthe US.

An important problem in in vivo EPR is cou-pling the microwaves with the sample. It is help-ful to communicate with medical personnel to usethe word "detector" to describe the EPR resonator,since this is the nomenclature familiar from radiol-ogy-

Low resolution EPR imaging is also a chemicalproblem. The more specifically targeted the spin la-bel, the higher resolution EPR imaging that is pos-sible.

There are lots of biological problems with theuse of nitroxides, especially with regard to theirmetabolism. However, the pharmacokinetics willteach us about redox metabolism. There will beadvantages to starting with a non-radical precursorwhich could be biologically reduced to a free radical.

The use of solid particle probes, such as fusiniteor lithium phthalocyanine, is limited at present be-cause they have to be placed physically in the tissue.However, once they are in place they can provide in-formation without further invasive procedures. It isattractive to contemplate the analogy with MRI ofthe use of magnetite coupled to antibodies as thefuture of this type of probe. Oximetry by any ofthese means holds great promise - the vision for thefuture is clinical application.

C. Discussion

One problem with in vivo spectroscopy so far isthat researchers have not been able to achieve S/Nany where near as good as one can with a flat cellin a TM cavity. Many have tried unsuccessfully todetect radical adducts in vivo. The Swartz lab hasdetected EPR of melanin in frog skin. Because freeradical based lung damage is an important humanhealth problem, the use of EPR oximetry to monitoroxygen in the lungs is a goal worth pursuing. Di-gestive track and fecal material should be a sourceof EPR signals for the study of biological systems.Unfortunately, these desired target tissues and othermaterials do not have radicals in high enough con-centration for current spectrometers.

In vivo could mean plants as well as animals.Lawrence Berliner has published examples of EPRimaging in plants (49). There are interesting prob-


lems, but research in the US is driven by fundingsources, and most of the funds available are not ori-ented toward study of plants.

Using the organic chemistry developed by LeonidVolodarsky, it is possible to use diamagnetic mole-cules which will become paramagnetic in vivo.

The reason Howard Halpern is using 250 MHzEPR is to be able to apply it to humans. Penetra-tion of the microwaves into the body is necessary.Even at 250.MHz the skin depth is only 7 cm inmuscle and a bit deeper in fat. Limited penetrationdepth is not a barrier to all in vivo applications ofEPR. For example, human skin is an important or-gan. The EPR of skin of living humans, even EPRimaging of skin, is accessible to X-band EPR.

Because of the limitations of current spectrom-eters much of the discussion emphasized the greatefforts to obtain even a simple CW EPR spectrum ofthese in vivo samples. Consider the insights possi-ble if one could use, for example, FT EPR on thesesamples. As noted later, there is no advantage ofFT if one is observing a single-line resonance unlessthe data acquisition rate can be increased.

IV. State of the Art Lecture- FT EPR and High-FieldEPR: Jack H. Freed

A. Comparison with NMR

The developments in the last 15 years in NMRthat have led to the revolutionary importance ofNMR in many branches of science include:

• NMR: high resolution via high magnetic fieldsand associated frequencies - e.g. from 100MHz up to 750 MHz; EPR has available aneven larger jump in frequency, from 9 GHz to250 GHz.

• FT NMR and 2D FT NMR; EPR - analogousdevelopments have been realized.

• MRI and its applications to both materialsscience and medicine; EPR imaging has notbeen applied to humans yet, but already hasmany applications in materials science andbiomedicine.

B. FT EPRWork in the Freed lab has been driven by a

need for better techniques for improved resolutionin dynamics and structure in the chemical physicsof biophysical problems. Five years ago at the Work-shop some very new results in these areas were in-troduced. There has been a great deal of progresssince then, including 2D FT EPR (50, 51). The S/Nachievable in FT EPR is illustrated with a 0.75 mMsample of perdeuterated tempone in 16 microlitersof a smectic liquid crystal. The effective decay rateof the FID following a microwave pulse for this sam-ple is 200 ns. This is T?J - it includes both homoge-neous and inhomogeneous broadening. With pulsewidths of 5.5 ns and time resolution of 5 ns, some40,000 FIDs can be averaged in 6 s. The FID canbe observed for more than 10 times the T£. Figure2 shows the sensitivity possible with FT EPR.

In addition to the possibilities for enhanced S/N,FT EPR can also be applied to the study of tran-sient species (52, 53). One can measure radicalswith submicrosecond lifetimes, generated for exam-ple by a laser pulse, by recording the single pulseFID and performing Fourier transforms.

The modern era in FT EPR, including initial re-alization of 2D FT EPR starts in about 1986 (54).Applications of 2D FT EPR in the short intervalsince then include:

• nationally narrowed - FT EPR, 2D ELDOR;diffusion in liquid crystals and model mem-branes (55);

• viscous fluids and powders - experiments aremore challenging but they yield more micro-scopic details about motion - applicable tech-niques include 2D ESE, 2D FT, SECSY, 2DELDOR; these techniques are also useful forstructural studies via nuclear modulation.

• 2D FT EPR imaging with pulsed field gradi-ents - spatially resolved 2D FT EPR (56, 57);

Now one can with 2D FT EPR obtain nuclearspin flip rates, Heisenberg exchange rates, and fromthem molecular rotational and translational diffu-sion coefficients. All of these are measured simulta-neously on the same sample, so there is no problemwith comparisons due to sample preparation or sam-ple conditions - and, they are obtained quickly.

Vol. .16, No. 3/4 167q

cs

LdQ

<N

in1

ood

0.0 300.0 600.0 900.0 1200.0 1500.0 1800.0 2100.0 2400.0TIME (ns)

Figure 2: FID from 0.75 mM perdeuterated tempone in 16 piL of a smectic A phase at 35°C. The microwavepulse width was 5.5 ns. 40,000 FIDs were averaged with 5 ns resolution in 6 seconds total experiment time.The experimental dead-time (lack of data at the beginning of the FID) was 80 ns. T2 for this sample wasca. 200 ns. Note that the vertical display was magnified by 10 at ca. 700 ns and again at ca. 1350 ns.Unpublished data provided by Jack Freed.

Lipid vesicles and biological membranes are chal-lenging because they are macroscopically disor-dered, and hence there is a great deal of inhomo-geneous broadening. One takes advantage of theextra resolution of the second spectral dimension.Because the T2 is only 30 ns, one uses 1 ns time res-olution in the data acquisition. For dilute samplesthe cross peaks show that electron nuclear dipolarterms dominate for short mixing times and Heisen-berg exchange becomes important for long mixingtimes. In spite of the short decay times it is stillpossible to get detailed information on combined ro-tational and translational motions, as is shown inFigure 3.

A major goal in the study of molecular dynam-ics is to obtain the homogeneous T2. This can bedone in a spin echo experiment in which one simul-taneously irradiates all three lines in the nitroxide

EPR spectrum. The SECSY experiment shows thatthe T2 changes abruptly across an inhomogeneousline. In the slow motional regime systems with T£ asshort as 10 ns has been studied by these techniques(58).

The slow-motional 2D ELDOR experiment isbased on spin echoes, not the FIDs. One can watchthe complete rotational dynamics - how a moleculeoriented such that it gives a certain spectral fre-quency moves so that it contributes to another re-gion of the spectrum within a given mixing time.For example, 2D ELDOR of a spin-probe in a vis-cous solvent reveals the residual rotational motion,whereas for an irradiated crystal one observes slowdynamic modes in the crystal.

The large proton hyperfine splitting pattern ofirradiated malonic acid can be excited in a singlepulse. The predicted exchange peaks in the 2D dis-


1GPC/P0PC ELDOR 66C T-200nj silt 16PC/POPC ELDOfl 66C T-60Onf ellt

taPC/POPC ELOOR 66C T-1200ni «lfw 16PC/POPC ELOOR 66C T-2000ns elfv

Figure 3: 2D-FT-ESR spectra of nitroxide radicals in lipid dispersions. For this sample, Tj is ca. 20-30 ns.Cross peaks in the upper left spectrum (for short mixing times) result from electron-nuclear dipolar inter-actions. At longer mixing times (lower right-hand spectrum) Heisenberg exchange dominates. Unpublishedresults provided by Jack Freed.

play due to cross relaxation can be observed growingin as a function of mixing time.

The technical challenge with performing pulsedFT EPR imaging was creating pulsed magnetic fieldgradients of 100 G/cm that persist for less than 100ns. All of the advantages of FT NMR imaging be-come available to research on electron spins via thisFT EPR and pulsed field gradient technology. Spa-tially resolved 2D FT EPR (thus three dimensions)has been demonstrated for samples containing 15Nand 14N nitroxides.

C. High Frequency EPR

High frequency EPR yields:

• higher g-factor resolution - one can read thethree nitroxyl g-values directly from the spec-trum at 250 GHz; even for the nearly free elec-trons trapped in solids the g-tensors can bemeasured.

• greater sensitivity to dynamics - one can mea-sure picosecond motions, since the sensitivityof line widths to motion is about a thousand

Vol. 16, No. 3/4 169

times greater at 250 GHz than at 9 GHz;

• transition metal spectra with large zero-fieldsplittings (ZFS) (63) - e.g., a Mn(II) complexwith a ZFS of 5800 G can be analyzed in termsof second order perturbation theory.

• better absolute sensitivity will eventuallybe realized, once spectrometer upgrades de-scribed elsewhere are made.

The transition to high frequency EPR bringsa new vocabulary to EPR. The spectrometers arebuilt using quasioptics, and techniques are those offar infrared not microwave technology. Both hard-ware and software have to be developed to performand interpret these experiments (64-67).

Although the above techniques are available toothers (visitors are encouraged to come to Cornell tolearn about unpublished details), they are not fullydeveloped, since there has been little funding forthis work. There are definite needs to improve thetechnology. The most important technical problemin FT EPR is spectrometer deadtime. With 1 KWpulses the current deadtime is 60 ns. The ringingtime of the low-Q resonator used implies one shouldbe able to reach a deadtime of 25 ns. It is also impor-tant to extend these techniques to multi-frequencies,because there are advantages and disadvantages forvarious experiments at different frequencies.

D. Respondent - Linn BelfordThese 2D FT EPR techniques are beautiful.There are benefits to high-field high-frequency

that come principally from having the high fre-quency, and other benefits that come from havingthe very high magnetic fields. One advantage of highfrequency is that one can cover large ZFSs. Oneexpects extensive applications to important prob-lems in metalloproteins. The high sensitivity ex-pected at high frequency holds out the possibilityof studying very small samples. The benefits fromhigh field EPR come from the fact that the impor-tance of the Zeeman term relative to the ZFS termsin the Hamiltonian increases at high field. The morenearly first-order spectra at high field increase thechance of interpretation of the spectra.

Very few high frequency EPR spectrometers areavailable in the world. There are spectrometers inRussia, France, Germany, Netherlands, Japan, and

the US. The highest frequency at which conventionalresistive magnets are useful is 60-70 GHz. At higherfrequencies than this there are four spectrometers inthe US, the far-infrared spectrometer at the NavalResearch Laboratory, the 95 GHz spectrometer atIllinois, the 140 GHz spectrometer at the MIT BitterMagnet Laboratory and the 250 GHz spectrometerat Cornell.

Some questions posed regarding high-frequencyEPR instrumentation:

• why are there not more spectrometersin the 30-70 GHz region, where non-superconducting magnets can be used?

• should frequencies >250 GHz be pursued vig-orously?

• is it reasonable to expect that high frequency(millimeter range) EPR spectrometers couldbecome viable commercial products?

• can the difficulty of sweeping the superconmagnets be overcome?

• is there possibility of using modern FT IR in-struments with magnets installed for Zeemansplitting (66-68)?

• can the sensitivity (especially for aqueoussamples) be enhanced by new resonator de-signs?

E. DiscussionJack Freed at Cornell had access to a far infrared

laser with the same lines as are being used at Greno-ble, but the work was quickly abandoned becausethe lasers did not provide the degree of spectral pu-rity and stability that EPR spectroscopy uses in themicrowave region. The instability of Bitter magnetsis also a problem. Consequently, the high frequencyspectrometer at Grenoble performs low resolutionEPR relative to what is needed for molecular dy-namics studies.

In the Cornell system, a second supercon magnetmay readily be swept ±500 G about the center field.This is adequate for studying organic species, but itis clearly inadequate for studying inorganic species,for which the main magnet is swept.


V. State of the Art Lecture —Pulsed EPR: Arthur Schweiger

A. Comparison with NMR

Until recently, conventional CW methods ofmeasurement prevailed in EPR spectroscopy. Thiscontrasts with the situation in NMR spectroscopywhere the CW techniques have been superseded al-most entirely by an impressive variety of elegantpulse techniques. Although pulse methods were in-troduced in EPR at about the same time as in NMR,only a small number of research groups applied pulsetechniques to EPR in the first three decades (69, 70).The slow growth of pulsed EPR is probably due tothe expensive instrumentation that was needed, andto the lack of digital electronics sufficiently fast forany but a restricted range of experiments. How-ever, the situation has changed radically within thepast few years, and pulsed EPR is undergoing ex-traordinary rapid development. New instrumentalcapabilities and new pulse techniques make it pos-sible to reduce the measurement times, to increasesensitivity, to improve resolution, and to simplifycomplicated spectra.

Today almost all topic areas of EPR spectros-copy are, or will soon be, affected by various pulsemethods. Techniques of particular importance in-clude time-resolved EPR spectroscopy, methods formeasuring relaxation times, techniques for study-ing molecular motions, methods for the indirect de-tection of nuclear transition frequencies, electron-nuclear double resonance, and EPR imaging.

B. New EPR Detection Schemes

The following topics and references will focus onEPR of materials in the solid state. The State of theArt Lecture by Freed provided references to ID and2D EPR techniques applied to species in solution.The annotated list of references provides very briefcomments on the new EPR techniques introducedin the last couple of years.

1. Electron Spin Echo

Following an initial emphasis on saturation recov-ery measurements (71), the majority of recent pulseEPR experiments in the solid state measure the res-onance phenomena via the electron spin echo (1-5,

69, 70, 72-76).The popularity of the electron spin echo ap-

proach is due to the fact that, with a very few excep-tions, the EPR lines of solids are strongly inhomo-geneously broadened. As a consequence, the trans-verse magnetization caused by a microwave pulse,called the free induction decay (FID), rapidly de-cays. The instrumental deadtime usually preventsobservation of the FID in solids, and the dephasingof the transverse magnetization has to be refocusedby performing an electron spin echo (ESE) experi-ment.

2. FID detected hole burning

In FID-detected hole-burning (77-81), a transientspectral hole burnt into an inhomogeneously broad-ened EPR line by means of a selective microwavepulse is shifted or broadened by various types ofperturbations (radio-frequency field, Bo-field jump,electric field, sample rotation, etc.), and is subse-quently recorded in a single experiment via an FIDfollowing a nonselective microwave pulse. The FID-detected hole-burning experiment can be applied toany EPR spectrum with inhomogeneously broad-ened lines, provided the relaxation times are suf-ficiently long. Many of the well-known ESE pulsesequences have an analogous FID-detected hole-burning sequence that is often superior to the ESEexperiment.

3. CW Detection

The detection schemes described above involvemonitoring the transient signals (echoes, FIDs)emitted by the sample after pulsed excitation. Analternative approach is to get information about theperturbed spin system by measuring on-resonancemagnetization of the spin ensemble by using weakCW microwave irradiation (78, 82-84).

4. Longitudinal Detection

Longitudinal detection (85, 86) is based on theobservation of rapid changes in the z-magnetizationeffected by microwave pulses. Pickup coils withtheir normal oriented parallel to the static mag-netic field are used to record the time-dependentz-magnetization during the pulse sequence. Longi-tudinal detection is free of artifacts caused by theinstrumental dead-time.

Vol. 16, No. 3/4 171

5. New Methods for the Measurement ofthe Nuclear Modulation Effect

The standard electron spin echo envelope modu-lation (ESEEM) experiments suffer from several dis-advantages. A number of pulse schemes have beendeveloped recently to improve resolution and sensi-tivity, to separate overlapping ESEEM spectra, andto overcome various types of instrumental distor-tions (76, 78, 87).

6. ESEEM at Frequencies Other Than X-Band

Going to lower or higher microwave frequenciesthan X-band (88-94) may increase the depth of themodulation and reduce or eliminate the dispersionof nuclear frequencies. This ESEEM cancellationeffect has been analyzed (91-94).

7. Phase-Shifted Excitation

The modulation depth may be increased byeliminating the decay caused by dipolar interactionamong unpaired electrons (95).

8. 5-Pulse ESEEM

With the 5-pulse ESEEM sequence (96), themodulation amplitude can be up to a factor of eightlarger than in the corresponding 3-pulse experiment.The echo signal contains no unmodulated part.

9. Extended Time Excitation

The entire two-pulse echo modulation can beobtained by a single experiment using a coherent, astochastic, or a pulse-burst stimulation followed bya strong refocusing pulse (97,98).

10. Coherent Raman Beats

This experiment allows one to record the en-tire three-pulse modulation in a single experimentby detecting nuclear coherences with a weak probepulse (84).

11. Soft ESEEM

By using two microwave frequencies, ESEEM canbe accomplished with low microwave power (milli-watts instead of watts). Because of the use of two

microwave frequencies one does not have to exciteallowed and forbidden transitions simultaneously toget echo modulation. "Soft" ESEEM (99, 100) doesnot suffer from blind spot artifacts and the modu-lation frequency is not limited by the pulse band-width.

12. Remote Echo Detection

In this pulse scheme transverse magnetizationrepresenting the echo is converted into longitudinalmagnetization (101). A two-pulse echo sequence isthen used to read this magnetization. The proce-dure is insensitive to the deadtime of the spectrom-eter.

13. Echo Modulation Echoes

With this special three-pulse sequence the shapeof broad hyperfine lines can be restored (102).

14. 4-Pulse ESEEM

The ESEEM peaks that correspond to sumsof frequencies contain important information aboutthe magnetic parameters of the nuclei. The 4-pulseESEEM approach allows one to measure highly re-solved sum peak spectra of disordered systems (76,98, 103).

15. HYSCOREDOR

Hyperflne Selected EN-

HYSCORE (hyperfine sublevel correlation spec-troscopy) is a very powerful technique to study weakhyperfine interactions, in particular in disorderedsystems (94, 104-107). The technique is distin-guished by a high spectral resolution in both dimen-sions and allows one to disentangle the correlationfeatures over two quadrants of the 2D frequency do-main.

16. 2D FT-EPR in Solids

For EPR spectra covering a small field range, as isoften the case for radicals, 2D FT-EPR techniqueshave been applied successfully for the measurementof the nuclear modulation effect (60, 108).


17. Fourier Transform EPR-Detected NMR

FT-EPR detected NMR is based on the burn-ing of transient holes into the EPR line by excitingforbidden EPR transitions and detecting the entirehole pattern via an FID (81). The procedure allowsthe observation of all nuclear transition frequenciesin a single experiment. The sensitivity may exceedthat of an ESEEM experiment by up to an order ofmagnitude.

18. Phase Cycling

Phase cycling is of great importance in pulsedEPR to record undistorted echo or FID signals (105,109).

19. Double Resonance Experiments

Along with the rapid developments in pulsed EPRspectroscopy, there has also been a fast-growing in-terest in pulsed ENDOR and related double reso-nance techniques (87, 110-115). There have beenseveral recent reviews of the field.

terns that do not contain nonsecular hyperfine in-teractions (e.g., liquid solutions) (119).

24. EPR-Detected Nuclear Transient Nuta-tions and Multiple Quantum ENDOR

EPR-detected nuclear transient nutations andmultiple quantum ENDOR are closely related tech-niques (82, 113, 120). They can be applied to deter-mine the multiplicity in ENDOR spectra as well asthe hyperfine spectral density in different sectionsof an ENDOR spectrum.

25. Time-Domain ENDOR

Strong rf pulses used in the technique of time-domain ENDOR excite a spectral width of about 1MHz (121, 122). The FID of the nuclear spins isrecorded via an electron spin echo. The sensitiv-ity and resolution achieved with this pulse sequencemay exceed that obtained with standard pulse tech-niques.

20. Optimized ENDOR

By using a new mixing scheme the polariza-tion transfer between nuclear and electron spins isimproved, and an optimum ENDOR efficiency isachieved (116).

21. Triple Resonance

In a triple resonance experiment, nuclear transi-tions are excited with two rf pulses of different fre-quencies (110). The technique is used to determinerelative signs of hyperfine coupling constants and toseparate overlapping ENDOR spectra.

22. Hyperfine-Selective ENDOR

The procedure allows the measurement of EN-DOR subspectra originating exclusively from nucleiwith a predetermined hyperfine coupling constant(117, 118).

23. Radio-Frequency Driven ESEEM

The radio-frequency driven ESEEM pulse schemecan create echo modulations in paramagnetic sys-

26. Coherence Transfer ENDOR

Coherence transfer ENDOR is an interestingexperiment from the point of view of spin dynam-ics (123, 124). However, the technique suffers frompoor spectral resolution and is therefore not of verygeneral practical use.

27. SEDOR-ENDOR Spectroscopy

SEDOR-ENDOR is basically a SEDOR experi-ment for the nuclear spins (125). The electron spinsare used only for the polarization of the nuclei andfor detection. The technique allows the measure-ment of nuclear-nuclear dipole couplings.

28. Fourier-Transform Hyperfine Spectros-copy

Fourier-transform hyperfine spectroscopy is basedon the FID-detected hole-burning approach (80). Inthe spectrum obtained each group of equivalent nu-clei is represented by one peak at the hyperfine fre-quency, independent of the nuclear spin quantumnumber.

Vol. 16, No. 3/4 173

29. ENDOR-Edited-ESEEM Spectroscopy

A combined ENDOR-ESEEM experiment allowsthe correlation of different nuclei (126).

30. 2+1 Pulse Train ESEThe ESE pulse sequence termed "2+1" can be

used to determine electron dipole-electron dipole in-teractions between paramagnetic centers (127-130).

31. ID and 2D Pulsed ELDOR

Pulsed ELDOR uses either two microwave fre-quencies, or a jump in the magnetic field strength,for the measurement of relaxation times, spatial dis-tributions of paramagnetic centers, and magnetiza-tion transfer (131-137).

32. EPR Imaging

Although early EPR imaging experiments wereperformed with CW techniques, pulsed EPR tech-niques recently have become important in EPRimaging (56, 57, 138-142).

33. Resolution Enhancement of Field-Swept EPR

A number of methods are under developmentto disentangle field swept EPR spectra using pulsedEPR techniques (143).

34. Electron-Zeeman-Resolved EPR

An EPR spectrum can be resolved in a seconddimension based on differences in the electron Zee-man interaction of different paramagnetic centers ordifferent orientations in a disordered system (79).

35. Anisotropy-Resolved EPR

Methods have been developed to make use ofthe anisotropy of the magnetic parameters to disen-tangle powder EPR spectra by rapidly changing theorientation between the static field and the sample(144, 145).

36. Electron Spin Transient Nutations

Transient nutation techniques are applied to sep-arate overlapped EPR spectra, to determine spin

quantum numbers and to study photoinduced elec-tronic states (78, 83, 146-148).

C. Recent Instrumental Innovations inPulsed EPR

Over the past few years instrumentation in pulsedEPR has made enormous progress. The follow-ing discussion is restricted to resonator design andto spectrometers working at microwave frequenciesother than X-band.

1. Resonator Design

The most significant innovation in resonator de-sign (124, 149-154) in recent years is the introduc-tion of the EPR loop-gap resonator (LGR) by Hydeand coworkers (149, 150), and the development ofrelated structures for different types of pulsed EPRexperiments, including pulsed ENDOR, and mag-netic field jumps (144). In addition to lumped-circuit resonators of the LGR type, increasingly di-electric resonators are finding application in EPR(155-157).

2. Spectrometer FrequencyMost pulsed EPR spectrometers operate with a

microwave frequency of ca. 9 GHz. The develop-ments up to 1987 were reviewed in (160). Recentlyseveral pulse EPR spectrometers operating at higheror lower frequencies (88, 90, 158, 159) have been de-scribed, including ENDOR at 97 GHz (161-163).Other innovations in instrumentation over the lastfew years include:

• miniaturization of spectrometers, e.g., forstudying irradiated foods, dosimetry, etc.(164, 165)

• Fabry-Perot resonator design (64)

D. Respondent - David SingelThere have been many illustrations of the util-

ity of multifrequency ESEEM during the Workshopand the preceding Symposium. Ultimately, varyingthe magnetic field and using some of the new pulsetechniques may accomplish much the same thing insorting out nuclear hyperfine and quadrupole fre-quencies. The balance between nuclear Zeeman andhyperfine interactions determines the amplitude of


the modulation effect. Some of the new pulse tech-niques may change this balance, but the effect de-pends on a resonant phenomenon, so the experimen-tal magnetic field strength is very important. TheHyscore and echo-modulation-echo pulse sequencesare ways to deal with broad lines. Ways to get ridof broad lines include cancellation of hyperfine andZeeinan interactions and cancellation of first orderlinewidths that show up in the 14N double quantumfrequencies.

Assignment of frequencies to a particular nucleuscan be made by observing the field dependence ofthe frequencies. See for example the recent study ofpyruvate kinase by Peisach in which distinction be-tween coordination by N or P was made (166). Thesum combination peak shift is inversely proportionalto frequency; this suggests important applications ofS-band ESEEM.

E. Discussion

The 2D FT EPR spectrometer developed byFreed and coworkers at Cornell has been describedin a review article (51). Recently developed highpower microwave switches have not been published- the inventor at Cornell is applying for a patent,and they lack funds and time to do some of thecharacterizing experiments needed to write a paperabout the switches. Freed invites people to come tothe lab to learn about these things.

The real question about these new 2D FT ex-periments is whether new information can be ob-tained. For example, is there any evidence for an-gular dependences of nuclear relaxation? These arejust the type of questions to which these techniqueswere applied by Freed and coworkers in 1989, wherethey demonstrated anisotropy of the nuclear spinrelaxation and interpreted it in terms of moleculardynamics. An experimental and theoretical studyof Heisenberg exchange in oriented liquid crystalsshowed that there is no reason to expect muchanisotropy (55, 167). Currently, ESEEM as a func-tion of frequency is often necessary to make thespectral patterns comprehensible. Are there newpulse sequences that could make the frequency de-pendence measurements unnecessary? The 5-pulseexperiment can be continued with more and morepulses to increase the modulation depth. However,with more pulses there are limits on relaxation timesthat can be studied and one loses sensitivity. FID

detected hole burning also gives deeper modulation,sometimes even in cases where one would not seemodulation in normal 2- or 3-pulse ESEEM. If theTi trend observed by Hyde continues and Ti islonger at Q-band and higher frequency, then someof the pulse sequences demonstrated at X-band areeven more useful at higher frequency. Work is inprogress in the Schweiger lab on pulsed Q-band.

Applications of high-field EPR would appear tobe extensive for species whose spectral linewidthsdo not scale with field, where one is removing, forexample, second-order fine structure broadening. Itis not obvious that the spectrum of a Cu(II) complexwill be improved at high frequency.

The major application of high field EPR tometalloproteins will likely be for those that havelarge ZFS, including non-Kramers even-spin sys-tems. Even though the lines may be broad, high-field EPR will be important if a transition is ob-served at all, since they cannot be seen at X-band.To see a signal that one could not otherwise seeis an enormously good reason for doing high fre-quency EPR. G-strain is considerably larger at highfrequency for something like Cu(II) in frozen solu-tion. For example, in the Cu(II) species that havebeen studied at 250 GHz at Cornell, g-strain scaledwith field, so there was no improvement in resolu-tion. Hyperfine structure resolution can even getworse at high frequency. Despite the g-strain line-broadening with increasing frequency, there are ex-amples (see, e.g., Nilges, et al., in previous EPRSymposia) of considerably enhanced spectral infor-mation content for powdered Cu(II) specimens at95 GHz. The prospects for such enhancement arevery case-dependent, in the experience of the Illi-nois group.

Often the incentive for going to higher field is toget better g-tensor information. One expects sen-sitivity to scale roughly as frequency squared, withmaybe + or — 1/2 in the exponent. One problem isthat as the frequency increases and the spectra getbroader, the magnetic field modulation amplitude asa fraction of linewidth decreases, so sensitivity doesnot improve as much for the normal phase-sensitivedetected CW spectrum when the lines are broad.

Vol. 16, No. 3/4 175

VI. Panel Discussion - High res-olution EPR

Panel Members: Jack Freed, Ronald Ma-son, Roger Isaacson, Lowell Kispert, Arthur Heiss(Bruker Instruments), Clarence Arnow (Micro-Now), Philip Morse (Scientific Software), MarkWoolfrey (Oxford Instruments).

The topic "High Resolution EPR" for the pur-poses of this review encompasses most of the ap-plications of "normal" CW EPR, whether to or-ganic radicals or metals, in solid phase or in so-lution. Issues include: research and instructional,portable and application-dedicated, multifrequency,S/N, data manipulation, simulation, visualization.

The following paragraphs summarize commentsand questions from the audience and the panel.

A. Kinetics

Real-time kinetics measurements of radicals is animportant and expanding area of EPR, and one towhich FT EPR is making important contributions.See for example the work of van Willigen, Turro,Dinse, etc. Microsecond kinetics can be studied,because in this time one can obtain an FID.

Fast-response conventional (CW) EPR is also be-ing developed. Bruker has a microwave transientbridge which, combined with a split-ring resonatorin a matched condition (critically coupled, not over-coupled) with a low Q, results in a system with200 MHz bandwidth for these types of experiments.This bridge has many of the components of the pulsebridge, without the switches.

B. Longitudinal Detection

The sensitivity for longitudinal detection is abouta factor of 10 worse than normal detection. Thedetection coil is resonant at the frequency of therepetition rate of the pulse experiment.

C. Signal to noise

Although EPR is more sensitive than NMR on aper spin basis, the species of interest in biomedicalfields are not very abundant. Therefore, there is avery serious S/N problem, especially for samples of,e.g., a microliter of protein solution. In the biomed-

ical area one of the key priorities is improved S/N.Spectrometer improvements are needed.

Concerns were expressed that the treatment ofnoise in FT EPR may not yet fully reflect the na-ture of the experiment. For example, is it possibleto define the noise in a time-domain experiment andapply it in an unbiased fashion to the FT spectrum?Based on the discussion of noise in FT NMR byErnst in 1966 (168), Freed discussed some aspectsof noise in FT EPR (51). One problem with S/Nenhancement via FT in EPR relative to NMR is theneed in EPR to decrease the resonator Q (to ca. 40)in order to get adequate bandwidth (e.g., 100 MHzat 9 GHz). One expects that the signal loss is pro-portional to Q. On the other hand, NMR has to haveslow pulse repetition rates because of the long nu-clear Ti values. In EPR, Ti is short enough in mostcases that some S/N improvement relative to NMRcan be regained by faster data collection. However,no commercial digitizer can accept repetition ratesas fast as EPR Tis would permit.

D. Ex Vivo EPR; Aqueous Samples inFlat Cells

Ex vivo EPR got a bad reputation a long time agobecause of artifacts created by grinding or lyophyl-izing the sample. However, these problems are nowrecognized and ex vivo EPR studies can be done re-liably. For example, bile or urine can be studiedin flat cells in TM cavities, via cannulae if desired.S/N is a problem for in vivo EPR, even with spintraps.

It has been reported that the surface of normalflat cells is rough enough that it introduces vortex-ing and resultant noise in some spectra when usedas a flow cell. Specially made cells with smootherinterior construction work better, but are more ex-pensive. A newly redesigned flat cell with muchtighter tolerances on flatness gives much better per-formance than the older flat cells. Wilmad is work-ing on a redesigned flat cell, which should be avail-able in a few months, to solve this problem at areasonable cost.

Loop gap resonators are worth considering forpulsed EPR studies of aqueous samples becausethere is fairly good separation of B and E fields ina LGR, especially relative to a cavity resonator. Inaddition, the Q used for pulsed EPR is low enoughthat a large amount of water can be put in the res-


onator without having much further effect on theQ.

E. Dielectric Resonators

During the Symposium preceding the Work-shop, results presented by Roger Isaacson empha-sized the desirability of using dielectric resonatorsto get even better separation of B and E field, whileretaining the benefit of a higher Q where it can beused. Bruker markets a dielectric resonator at X-band in the Flex-line resonator series. This Brukerresonator uses sapphire in the dielectric resonatorsbecause other materials have too many impuritiesto be useful for CW EPR. Sapphire cut in the rightdirection, and turned in the right direction in theEPR probe, provides a magnetic field region of ca.200 G in which there are no impurity signals. Ifyou cool the sapphire resonator, lines from impuritylevels of Fe, Cr, etc., will increase in intensity, butnot so much that they will distort the spectrum. Inpulsed EPR these impurities do not interfere withthe signal at all because they are in such low abun-dance and their relaxation times are so short. Foraqueous solutions in a small cylindrical capillary adielectric resonator yields a factor of 6.7 improve-ment in S/N relative to a standard resonator. Thedielectric resonator is slightly better in this regardthan the LGRs with which it has been compared. Ifenough sample is available, better S/N will be ob-tained for aqueous samples in a large flat cell in aTM cavity.

Peter Hofer reported that tests at Stuttgartshowed that UV light did not have any effect onthe sapphire resonator, but gamma radiation wasnot tested.

F. Small and/or Dedicated EPR Spec-trometers

The EPR field has been looking for a longtime for a market for dedicated EPR spectrometers.The largest market that ever occurred was the saleof about 50 FRAT (by Syva; Syntex-Varian) spec-trometers for drug testing, but other techniques re-placed the use of EPR for that application. Dia-mond companies have purchased a portable EPRto screen for synthetic diamonds. Varian producedtwo 1 GHz spectrometers, and Micro-Now producedthree 1 GHz spectrometers, for screening crude oil

for vanadium many years ago.Dosimetry is a possible market. An ASTM com-

mittee is working on a standard that will permitEPR use in dosimetry. The Bruker EMS104 was de-veloped for radiation dosimetry and is being testedfor monitoring irradiated food in Europe.

Clinical oximetry is a likely application for anEPR spectrometer. This will probably have to bea portable, low frequency spectrometer, not just aversion of a standard spectrometer.

VII. Panel Discussion — In VivoEPR and Imaging

Panel Members: Lawrence Berliner, HaroldSwartz, Howard Halpern, Sandra Eaton, DieterSchmalbein (Bruker), Mark Woolfrey (Oxford In-struments).

A. The Question of Sample Size

The hardware and software issues for in vivo EPRand EPR imaging are very different from those forthe standard high-resolution experiment. Thus, wediscuss together "high resolution" spectra in vivoand multidimensional imaging. The colloquial ques-tion is "When can we get the elephant into the EPRspectrometer?" That is, how do we get to real appli-cations with samples bigger than mm size, or mousesize?

In counterpoint, Hal Swartz asserts that thequestion is wrong - people are too pessimistic. Withthe existing technology and relatively simple devel-opment one can do a large fraction of what needsto be done. One can look at the elephant if onlythe first cm or so of the elephant is to be exam-ined. Many interesting things are within that sur-face layer. At 250 MHz 80-90% of the things one isinterested in from a clinical point of view are alreadyaccessible. The problems remaining are not funda-mental, but merely the nitty gritty things that needto be sorted out. No one significant break-throughis needed.

For information on the use of surface resonators(e.g., the volume sensitivity), for cases in which thesample is too large to put into a resonator, see (44,172).

Larry Berliner suggests another point of view- Why don't we try to put the EPR spectrome-

Vol. 16, No. 3/4 177

ter inside the elephant? The hardware developmentneeded is miniaturization such that the probe couldbe inserted by catheterization.

B. Frequency Scaling

Clearly, while the in vivo and imaging experimentsare stimulating creative approaches to solving excit-ing problems, there remain some very fundamentalquestions. For example, it is not obvious what fre-quency scaling is appropriate to these experimentson complex living tissue. In NMR the penetrationseems to scale nearly linearly with frequency, andnot according to the square law that early literaturewould lead one to expect. One should not read con-flict into the decision to perform imaging at differ-ent frequencies in different labs. Since most exper-iments were started with little or no funding, eachlab worked with what was available. The Halpernspectrometer at 250 MHz is widely viewed as a closeto optimum choice. Work in other labs at higher fre-quency than 250 MHz is not a statement that higherfrequency is better - it is what is available and is giv-ing good results on an important set of problems. Itis a mistake to assume that one cannot obtain EPRspectra on almost all except the trunk of a humanbeing, if one works at 250 MHz.

C. Interpretation of In Vivo Spectra

Extracting information from in vivo spectra prob-ably requires a spectral fitting approach (169, 170).A reasonable fit hypothesis can be used to focus theentire spectral information on the few parametersassociated with the hypothesis. This approach al-lows one to determine very small variations betweenvery noisy spectra.

One of the main problems with animal imag-ing experiments is suppressing the noise caused bymovement of the animal. Attempts to capture themotional information electronically to be able to useit for corrections gives the side benefit that there isnow a record of, e.g., the depth of respiration of theanimal.

D. Magnetic Field and Magnetic FieldGradient Control

One of the main challenges for imaging experi-ments is the magnetic field control. In systems that

use Hall probes, the current practice is laboriouspositioning to put the Hall probe in a nodal planeof the imaging gradient field. This becomes verydifficult for more than one imaging dimension.

A related problem is the quality of the gradi-ent field. At the very high gradients used in EPRimaging, great care must be taken to ensure linear-ity of the gradient over the sample volume of inter-est. At very low RF frequencies the gradient coilswould produce a larger field than the main Zeemanfield. In the 250 MHz imaging spectrometer, theHelmholtz coils used to create the Zeeman field aresplayed to create the gradient field (40, 171).

It is attractive to use current control of copperHelmholtz coils to avoid the problems of Hall probepositioning on iron-core electromagnets (hysteresisproblems prevent current control of iron-core mag-nets). However, the perturbations of the field byferromagnetic materials in the vicinity is a problem.One has to keep ferromagnetic materials far away;even an infusion pump used to inject the spin probeinto the animal can cause interference with the spec-trometer.

E. Low Frequency and Imaging Spec-trometers

Dieter Schmalbein reported that Bruker is watch-ing the EPR imaging field, but until a clear appli-cation market develops they cannot afford the de-velopment costs. It would require several milliondollars to develop a professional EPR imaging sys-tem. They have made several experiments, and ten-tatively would expect to use a frequency below 1GHz, and would expect to design a resonator thatwould accommodate a whole rat. However, it isjudged premature to build a commercial product.The current market for the Bruker L-band EPRbridge is near zero. A new 2 to 8 GHz multifre-quency bridge has been built using the most modernmicrowave equipment available. It has much bettersensitivity than the L-band system, which was de-signed about 10 years ago. Until a commercial in-strument becomes available, researchers who wantto enter this field need to obtain information fromone of the labs that developed instrumentation andsoftware for imaging. The NIH-funded Illinois ESRCenter (which now has a branch at Dartmouth) ishappy to assist people, or to put them in touch witha lab that can assist with a specific problem outside


the experience of the Illinois Center.

F. Nitric Oxide In Vivo

There is much current interest in measuring NOin vivo, but estimated concentrations of NO in thebody are less than micromolar. As the simple di-atomic molecule it cannot be studied by EPR invivo. Possibly it could be studied via its paramag-netic effect, analogous to oxygen, or by trapping it.But are either of these approaches likely to producean image? Harold Swartz has unpublished demon-strations that one can trap NO with lithium ph-thalocyanine. Hemoglobin is a naturally occurringtrap for NO. It seems very unlikely that it will bepossible to monitor NO in vivo by EPR, let aloneimage it. If it could be done, the importance of NOin the body makes monitoring NO by EPR a likelyclinical application of EPR (173).

At the Lovelace Institute human volunteersbreathed NO2, then their lungs were washed withsaline and the cells studied by EPR. Heme-NO wasobserved with good enough S/N to serve as a mon-itor of NO2 exposure (174).

G. Noise in FT EPR, EPR Imaging andIn Vivo EPR

Multiple fast scan vs. slow scan data collection isone of the key decisions for in vivo EPR imaging ex-periments. This is one of the data collection param-eters that is optimized against the rates of motionof the animal, and other inherent time constants ofthe system. In the current 250 MHz imaging sys-tem, typically 15 sec scans are used. The currentlimit is the monitoring of the frequency of the field-frequency lock system, and the fact that IEEE488communication is used.

Colin Mailer emphasized that talk about improv-ing S/N to do in vivo imaging should face the real-ity that the signal relates to two parameters - thenumber of spins and Bi. In current technology thereis a tradeoff between sample volume and Bi - thelarger the resonator the smaller the Bi at the sam-ple. With LGRs the technology appears to be closeto the fundamental limit.

The key to solving the S/N problem is to under-stand the noise source. For in vivo EPR, the noisesource is likely to be the animal. Beyond the limitsjust discussed, there are problems such as how to

make the resonator less sensitive to animal motions.Use of a dielectric resonator to further decouple theanimal from the resonator might help.

If the problem of animal motion is solved, onestill has to work to decrease other noise sources, suchas the source and the detection system. The recentintroduction of a balun between the transmissionline and the resonator in the 250 MHz imaging sys-tem decreased the noise by a factor of 4. The systemis not fully optimized yet.

The fundamental limits discussed by ColinMailer have yet to be approached by the 250 MHzEPR system in Halpern's lab. The primary reasonis animal motion. If one operates under conditionsoptimized for the nitroxide radical, magnetic fieldmodulation amplitudes of 0.5 to 1 G can be used.With an input power of 100 mW, it is estimatedthat the Bi in the animal is ca. 0.3 G, a valuethat approaches saturation of the spin system. (Ifthe deuterated form of the nitroxide is used, dif-ferent conditions - e.g., Bi = ca. 0.01 to 0.03 G- are required for optimization, because of nar-rower linewidths.) Under these conditions, surfacecurrents (eddy currents) induced by the magneticfield modulation and by the RF are further modu-lated by the animal-motion-induced microphonics.These contributions increase the breathing-relatedartifact. A balanced power delivery system is onepossible approach to electronic suppression of thisartifact. About two orders of magnitude noise sup-pression will be required before encountering thelimits referred to by Colin Mailer.

The value of Bi that is useful in CW imaging islimited by the relaxation time of the electron spins.Possibly there is an application for contrast reagentsto shorten the relaxation times in EPR so that largerBi can be applied without saturating the spin sys-tem.

Ernst's analysis for NMR was that there wouldbe little advantage to performing FT spectroscopyfor single-line spectra. The imaging experiment in-herently is not a single line. FT EPR or rapid scanspectroscopy followed by mathematical deconvolu-tion (which was useful in NMR just before FT NMRwas developed) might be used to advantage in EPR.

100 MHz spectral width at 9 GHz requires a Q ofca. 40. If the center frequency drops by a factor often, but the bandwidth stays the same, then the Qrequired is 8. If the focus were on a narrower line,

Vol. 16, No. 3/4 179

so that the Q could remain at 80, then the issuebecomes one of sensitivity. It is not clear that thechoice of FT vs. CW in this case is a black andwhite issue. Among the tradeoffs are the relationof Q to the amount of aqueous sample that can beput in the resonator, the repetition rate that canbe used, etc. Imaging is an additional perturbationon the judgment. At low frequencies the bandwidthneeded for even narrow lines necessitates substantialpower.

In any non-pulsed experiment we throw awaya lot of information because we only look at oneFourier coefficient of the EPR signal. More sen-sitivity could be obtained by stacking synchronousdemodulators. Some years ago Hyde taught us (175)that we should digitize the entire 100 kHz modula-tion signal and try to get all of the information outof each modulation cycle. Bruker markets pream-plifiers and digitizers that have adequate speed toperform this type of analysis.

VIII. Panel Discussion - NewPerspectives on Spins

Panel Members: James Hyde, Melvin Klein,Arthur Schweiger, Bruce Robinson, Harvey Buck-master, Edward Reijerse, Hans Thomann, DieterSchmalbein (Bruker).

Arbitrarily gathered under this umbrella are awide variety of pulse, time-domain, multiple res-onance, and multiple modulation techniques thatshare the feature of exploiting non-linear behaviorand relaxation phenomena.

A. SQUIDs in EPR

SQUID devices are almost noiseless detectors, sothey are attractive wherever they can be used. InNMR SQUIDs are superior detectors up to about30 MHz, but at higher frequencies the standardmethods are better. This severely limits the typeof EPR experiment for which they could be useful.It is attractive to consider using a SQUID for zero-field EPR. Zero-field measurements would eliminatesome of the anisotropy problems often encounteredin EPR.

B. Multiquantum EPRModern microwave technology permits the gen-

eration of multiple CW frequencies with a commontimebase. In principle, the same irradiation frequen-cies could be generated by suitable time-modulationof a single frequency, but summing of distinct fre-quencies seems technologically preferable. Multi-quantum EPR is readily generalized from two orthree frequencies (as in Hyde's papers so far), toN frequencies. The potential is very great. Thereare two general thrusts: as a practical alternativeto magnetic field modulation for improved systemstability, and as a way to obtain information on re-laxation rates.

Among the many applications envisaged, fewhave been explored yet, since the technique is sonew. EPR imaging is one potential application.Image reconstruction algorithms require absorptionspectra (not derivative spectra). The fact that MQEPR yields absorption spectra directly makes it at-tractive to consider MQ EPR imaging. The need forabsorption spectra is another reason for the use ofFT EPR imaging instead of CW EPR imaging. Al-ternatives to CW for EPR imaging are imperative.

C. Microwave Source Phase Noise

What EPR applications would there be for amicrowave source with 20-40 dB lower phase noise?

At low frequency the wideband tunable mi-crowave sources have poor phase noise, and with alow noise GaAsFET amplifier in the detection sys-tem one finds that the source noise dominates. Thisis a case in which reducing the phase noise of thesource would be important. Lower phase noise atlower modulation frequency could be important -e.g., Roger Isaacson used 4 Hz modulation for ex-periments where the EPR signals under study willnot respond rapidly. Also, with the increased useof microwave preamplifiers the 1/f noise of the crys-tal detector is overcome, and in many experimentsone can more advantageously use field modulationin the region of 100-25,000 Hz.

Does phase noise scale with frequency? Thereseems to be little comparison data, but the usualassumption is that phase noise at all frequenciesrelative to the center frequency scales with themicrowave frequency. This may not be true forklystrons, where there could be mechanical vibra-


tions at particular frequencies. However, at 100KHz away from the carrier, phase noise appears toscale for klystrons. During the Symposium MarkNilges showed curves for a 100 GHz oscillator, andsome of them look like they are not scaling.

In many cases 100 kHz modulation is no longernecessary, and for species with long relaxation times100 KHz modulation is not desirable. Lower mod-ulation frequencies are likely to become more com-monly used. Thus, phase noise at 100 KHz may notbe the best comparison to make.

One should be aware that using phase lockingtechniques may introduce new sources of noise. Thereference oscillator has to be a very clean source,or it could become the limiting noise source in thesystem. This occurred in some cases in Buckmas-ter's lab when a synthesizer was used to phase-locka source.

James Hyde encouraged reference to Robins'book (18), which considers ways of handling phasenoise. With incomplete data available regardingphase noise characteristics of various microwavesources, an overall impression is that currentlyphase-locking to a quartz oscillator is preferable be-low about 4 or 5 GHz, and a fundamental oscilla-tor locked to a high-Q tank circuit is preferable athigher frequencies.

A comprehensive search of the literature by Hydedid not uncover a device at Q-band that had betterphase noise than the one he described (9).

One has to build the right system even to testthe phase noise - no commercial spectrum analyzeris satisfactory for the measurement.

D. Pulsed ENDOR

In CW ENDOR the rule of thumb is that the CWEPR spectrum S/N should be greater than 100:1 toget reasonable ENDOR results. Also one usuallyassumes that ca. 1 mM solutions are needed. Incontrast, if one can see a pulsed EPR signal (echo)one can obtain pulsed ENDOR for the sample. Inideal cases one can invert the spins and get a 100%ENDOR effect using pulsed EPR techniques. How-ever, the pulsed EPR signal usually has poorer S/Nthan the CW EPR signal, so a 100% ENDOR effectmay not result in better S/N than the smaller effectobserved in CW ENDOR.

Typically for metalloprotein solutions one ob-serves about a 5-10% ENDOR effect. How much

signal is lost during the polarization transfer pe-riod depends on cross relaxation and other relax-ation times. Cu(II) proteins at liquid He tempera-ture typically have cross relaxation times of 10 msor less. Tis are several hundred microsec. One doesnot want a high spin concentration, since then thephase relaxation time becomes short. For Cu pro-teins shortening of the phase relaxation time canbe observed starting at about 1 mM, depending onwhere the metal is in the protein - when they areabout 20 A apart one starts to see effects. Overall,the sensitivity is roughly a factor of 5 lower than forCW EPR.

E. Dissemination of Modern Techniques

A colleague once commented to Hyde with re-gard to a lecture presentation of exciting new tech-niques, "another experiment I cannot do." Engi-neers are not available in all labs to implementnew techniques. Of the techniques that Hyde hasdeveloped, the most generally applied, because itcan be implemented on largely standard spectrom-eters, is STEPR. Possibly a double-quantum EPRexperiment using double sideband/suppressed car-rier techniques could be implemented with a simpleaccessory. Commercial suppliers cannot do every-thing, but some things can be done, even thoughnot financially justified by themselves, because theyhelp carry the main product line. One possibilityfor introducing new techniques would be for groupsof investigators to submit a joint proposal to a fund-ing agency to purchase x number of accessories, andhave the vendor produce a batch of x of them atone time. Another approach'would be to have, asan outgrowth of a Workshop such as this, an interna-tional commission make a recommendation betweencompeting alternative demands on the limited de-velopment resources available.

Many of the techniques can be done with existingcommercial boxes if one knows how to put the boxestogether in the right way. Maybe someone shouldpublish the details of how to do these experimentswith existing boxes.

F. Software for Visualization of EPRData

Often software is the key to success. Without acombined hardware/software system one won't get

Vol. 16, No. 3/4 181

many results. It used to be that when a lab neededsoftware, someone just went home and wrote it atnight, but now software needs are too sophisticatedfor this approach.

There are some thorny issues about software.For example, even if you can write it in a night, itwill take a week to document it in such a way thatsomeone can use it. There is a lesson in commer-cial spread-sheet software. Sometimes it is betterto force an application into some documented andsupported commercial software rather than writingyour own special-purpose software. It is difficultto make excellent general-purpose software. Maybethe emphasis should be on subroutine libraries, andeasily modified software.

The new EPR spectrometers and experimen-tal methodologies described at the Symposium andWorkshop will provide enormous amounts of infor-mation (or at least raw data that somehow mustbecome information). Relative to slow-scan CWEPR, the new EPR technologies produce data atsuch a prodigious rate that data storage and subse-quent manipulation becomes a larger problem thanEPR labs have had to deal within the past. Al-though trivial by comparison with data generationrates in other fields of science (e.g., MRI, parti-cle physics, or the space program), the amounts ofdata require qualitatively different computationalapproaches than are available in most EPR labs.Some labs already approach this problem by us-ing data compression techniques, which can makethe data storage requirements modest. For exam-ple, Jack Freed's FT EPR can produce a few 1 MBspectra per hour of spectrometer operation. Hugeamounts of data are transferred to a supercomputerfor the most substantive analyses. Linear predictivemethods are used to reduce the volume of data forstorage. Specialized software is needed for visualiza-tion of the multidimensional information that nowcan be generated so quickly, in order that it be com-municated to human beings. Another approach is torecognize that the result of an experiment may be aseries of Fourier coefficients, and these are what onewould store, not all of the raw data. Others mightbe uncomfortable with the irreversible interim inter-pretation imposed on the data by these approaches.

Now that EPR has a standard (Bruker BES3T)for storage and transfer of EPR data we need toconsider how to present the data for visualization.

This is a major problem. The solutions in otherareas of science, where the visualization problem isanalogous, are very large software packages whichare very expensive because of the development effortto create them. A key issue is whether the EPRcommunity will be able to support the effort neededto develop this software.

Reef Morse pointed out that Scientific SoftwareServices from the beginning has always providedsource code for the marketed software. Customersmake significant modifications. Dieter Schmalbeinpointed out that the software provided with theBruker ESP380 pulse spectrometer represents 32man-years of effort. More sophisticated softwarecould be developed, but Bruker is limited in theeffort that can be invested in EPR software by theprofits that can be made in the EPR business. Morethan 90% of the EPR spectrometers are delivered touniversities or government institutes. In contrast,80-90% of the customers for NMR spectrometersare in industry. In the NMR field a professionalsoftware package can be sold to industry for a rea-sonable amount of money, because industry can seethe cost savings in terms of time saved by the soft-ware.

IX. Summary on Instrumenta-tion and Methodology

The horizons of new EPR techniques are phenom-enal. References were given above to many ways toapply pulses to spins. Many of the techniques arevery expensive to implement. The excitement is inapplying these techniques to problems that are nowonly being approached by the use of CW EPR. Alot of the discussion at this or other meetings aboutapplying EPR is about just getting a CW spectrum.The S/N problem is bad enough for some samplesthat we sometimes struggle just getting a spectrum,sometimes for days at a time. But problem solving,in systems to which ways of studying electron spinscan be applied, requires some of these new tech-niques. How soon can we get there? What is in ourway? The general response at the Workshop was -Money. So now let us consider the money aspects.


X. The Funding Agency Per-spective

Once upon a time there was to be a fourth panel,but the government ran out of money and fundingagency representatives could not attend the Work-shop. Dr. John Beisler, Executive Secretary, Bio-physical Chemistry Study Section, Division of Re-search Grants, NIH, was the only person invited whocould attend.

A. Questions Regarding Funding of EPRin the USA

The questions posed regarding funding are:

1. What funding is available for the new researchopportunities presented at this workshop, andfor solving the instrumentation and softwareproblems highlighted?

2. How many EPR spectrometers were fundedin recent years? What is the average dollaramount of such grants? Is there an historicaltrend?

3. What characterizes a successful EPR instru-mentation proposal?

4. What types of referee comments characterizeEPR instrumentation proposals that are notfunded?

5. How many grants (and what dollar volume)have been awarded in which a major focus ofthe research proposed is the development ofEPR instrumentation and/or methodology?

6. How many grants (and what dollar volume)have been awarded in which EPR is an im-portant technique even if not a major focus ofthe grant?

7. There is a tendency to compare funding ofEPR with funding of NMR, since they areboth magnetic resonance techniques. Does thestructure of instrumentation grant programsmake funding of an NMR proposal more prob-able than funding of an EPR proposal?

8. What do the long-range planning processeson-going at the federal funding agencies por-tend for research in or using EPR?

B. Information from the Presentation byJohn Beisler, DRG, NIH

The CRISP data base at NIH is the source of thefactual information he presented. He also providedhis observations and perspective as Executive Sec-retary of the Biophysical Chemistry Study Sectionat NIH.

The most recent fiscal year for which data wasavailable is FY92 (ended June 30, 1992). To querythe data base one has to use terms that are in itsthesaurus. The terms used were electron spin res-onance spectroscopy, electron nuclear double res-onance, and nuclear magnetic resonance spectros-copy. The number of grants includes RO1, PO1,P41, etc., types. Grants are listed as having e.g.,EPR as the primary, secondary, or tertiary thrust.

In FY92 General Medical Sciences funded 179projects (43% of the total awarded) in EPR. TheHeart Institute, with 50 awards, is far in secondplace. The Cancer Institute made 24 awards. TheAging Institute made only three awards in the twofiscal years examined. There is a lot of opportu-nity for applications of EPR in some of the otherInstitutes.

In the shared instrument program in FY87 about2/3 of the proposals were funded, but this was un-usual.

Because of the small number of applications forEPR spectrometers, they get reviewed by a panelfor "other spectroscopy." This results in relativelyfew of the reviewers being expert in EPR, which isviewed by some researchers as a liability, but in ahomogeneous review panel as for NMR, there is atendency to rank all of the applications.

Most EPR-related proposals tend to be reviewedby three study sections, Physical Biochemistry, Bio-physical Chemistry, and Metallobiochemistry. Ofthe roughly 80 applications per review cycle in bio-physical chemistry, about 24 are in NMR, 24 in crys-tallography, and a few in EPR. There is usually oneperson with specialization in EPR and a few othersknowledgeable about EPR on the study section.

Dr. Beisler asked various other people at NIHand members of study sections (past and present)about some of the questions asked for this Work-shop. Some of the impressions and opinions offeredwere:

Vol. 16, No. 3/4 183

FY87

FY92

EPR

EPR

NMR

327$45

Table

awards8M total

415 grants$57.3M total1763 grants

1U

16%

15%

20%

primary,

primary,

primary,

80%

81%

77%

tertiary

tertiary

tertiary

Table 11: NIH Shared Instrument Program

FY87NMR 51 applications reviewed, 31 funded, $7.9M totalEPR 3 applications reviewed, 2 funded, $356K total

FY92In FY92 the shared instrumentation program was cut from $32M/year to ca. $8M/year.3 awards for EPR, $600K total

1. NMR and EPR proposals fare about equallywell.

2. The perception is that the real richness of EPRapplications to lipid or membrane research hasbeen mined. There is a low opinion of EPR inlipid research.

3. EPR proposals need to emphasize what infor-mation on a particular problem EPR can givethat other techniques such as NMR, fluores-cence, X-ray, etc., cannot.

4. For greater success, put EPR in the broadercontext of other spectroscopies. How doesit complement the information available fromother spectroscopies? For structural biochem-istry, for example, what does EPR revealabout distances, angles, etc.

5. Remember to speak to the reviewers ratherthan making assumptions that they have abackground in EPR.

6. The advantages EPR has relative to NMR aresmall sample size and high sensitivity relativeto NMR.

7. In the context of discussion about new hard-ware development, it is well to keep in mindthat one can often get very reasonable datafrom a 15-year-old Varian spectrometer. Ele-gant solutions to problems can often be donewith very simple instruments.

Many scientists, hearing this opinion attributedto peer reviewers, wish to communicate the largermessage of this workshop, that elegant new EPRtools are now available for more powerful problem-solving than was possible with the older EPR tech-niques.

XI. The Vendor Perspective

At the close of the Workshop the communitysought the response of instrument and software ven-dors to the challenges and opportunities presented.

A. Bruker (Dieter Schmalbein)As a manufacturer, Bruker finds Q-band unprof-

itable but has decided not to discontinue it! TheBruker Q-band system has switched from klystrons,which are no longer available, to Gunn oscillators,


and the sensitivity is about the same. With thenew helium FlexLine cryostat, which is also usedby the L-, S- and pulsed X-band systems, the Q-band system can operate to 1.8 K, with magneticfield modulation from 1.5 KHz to 100 KHz withoutproblem.

Bruker continues to offer a diverse range of mi-crowave bridges for many specific applications. Thephase noise of X-band sources (klystron and Gunnoscillator) is suppressed 130-140 dB at 10 KHzfrom the carrier. The high output 2-8 GHz Multi-Frequency Bridge should meet the needs of manyresearchers.

Over the seven-year period, 1985-1992, 37% ofthe EPR spectrometers produced by Bruker weredelivered in the US, 19.4% in Germany, 6.5% Japan,4.4% England, in terms of dollars, not number ofspectrometers. In the past 12 months the situationhas changed, and 59% of EPR sales (in dollars) havebeen to Europe, 20% to Japan, and only 10% to theUS. This may change in the near future.

Since EPR is a very low volume market, Brukerhas to be very careful in selecting the areas in whichto invest development effort and capital. In therecent past they have put this effort into develop-ing the most advanced spectrometers that are pos-sible in a commercial market, culminating in theESP380E. About 40 of these have been sold so far(only five in the US). Bruker has the impression thatthe ESP380E is ahead of the users - people cannotexploit the capabilities of the ESP380E. There is aneed for more institutes around the world to teachpeople how to use non-stationary EPR techniquesand to provide service to people to help them startusing these techniques. No EPR service center inthe US (e.g., NIH Research Resource Centers) has amodern commercial pulsed EPR spectrometer. Be-fore Bruker can invest in making these spectrome-ters even more complicated, with capabilities suchas pulsed ENDOR, multifrequency pulsed EPR, andpulsed EPR imaging, there has to be more use ofthe existing capabilities of the spectrometer. Then,Bruker can consider the commercial implementationof these new techniques.

Up to now Bruker has not charged for EPR soft-ware - it was delivered as part of the spectrometer.Now there is evident need for much more profes-sional, sophisticated, and diversified software, andBruker anticipates having to hire more programmers

and hence to have to charge for the software.Bruker has tried very hard to listen to what the

researchers and other customers say they want in anEPR instrument. In the past everyone wanted thebest spectrometer possible, and the specificationsof the spectrometer were very important. Recently,this has changed in a few markets, especially in theUS. In the US people seem to want the lowest pricespectrometer, and the specifications usually are aminor consideration. In Europe price/performanceis the most important consideration. In Japan, how-ever, performance is most important. Bruker has todecide whether to develop two types of spectrome-ters, one with the highest possible performance andsophistication for part of the market, and anotherspectrometer at the lowest possible price. Duringthe Workshop scientists have expressed desire tohave spectrometers with higher performance andto have hardware and software from the manufac-turer. But the part of the market not represented atthe Workshop may require Bruker to put its devel-opment effort not into spectrometers for advancedtechniques but into spectrometers at lower prices.

Bruker will continue to improve the FT spec-trometers. Pulsed ENDOR will come on the marketnext year. They will experiment with imaging tech-niques. At the moment there are no plans to gointo high-field EPR, because they cannot foresee amarket in this area.

B. JEOL (Jack Francis)

JEOL will be involved in the development andmarketing of EPR spectrometers for a long time,and hopes to be more involved in conferences likethis by next year, and possibly add a bit to what isbeing discussed.

C. Micro-Now (Clarence Arnow)

Micro-Now has been involved in EPR instrumen-tation for over 25 years, mostly with accessories.They built an L-band spectrometer and a Q-bandspectrometer about 20-25 years ago. In the last 5years they have put more effort into building EPRspectrometers. They have built four types of spec-trometers - for teaching, for dosimetry, a more com-plete system in modular form, and a new spectrome-ter, demonstrated at the Symposium this year. Thisnew spectrometer incorporates a magnet built in

Vol. 16, No. 3/4 185

Russia, and is very portable. The spectrometer usesa Gunn source.

Their effort will generally be in the direction ofspectrometers such as this new one, which essen-tially address the part of the market that once wasserved by the Varian E-4.

D. Oxford Instruments (Mark Woolfrey)

There has been no comment at the Work-shop of limits to research due to the performanceof cryostats, in contrast to the discussion in 1987.Special versions of cryostats can be made when-ever modification of the standard cryostats wouldbe helpful.

XII. Summary PerspectiveA. The Horizons of EPR

Much of the discussion of commercial instru-mentation, and of NIH funding, even at this Work-shop has been about relatively standard CW, linear,single-frequency (X-band), field swept EPR. Thehorizons of EPR are much different. The fundingsituation is as if you were looking East from Den-ver, and the reality of research needs is as if youwere looking west from Denver. There is a lot ofdifficult terrain to get through to do such thingsas pulsed magnetic field jump, pulsed ENDOR, ormultiquantum EPR.

It is surprising, maybe even distressing, thatchanges in EPR as practiced in most laboratoriesand as described in most spectroscopy texts are somuch slower than some of the other changes goingon in society, especially internationally. If one looksback at the design criteria set forth by the 1987Workshop, one would make relatively few changestoday. The priorities remain about the same.

One can hope that some time not too far inthe future at another Workshop the focus will havechanged to the now largely unexplored regions ofEPR spectroscopy: 4D, multiquantum, multifre-quency, etc. What will be possible when we cansee EPR spectra of brain tissue, in vivo, localizedin a living animal, using all of the advanced EPRtechniques we learned about at the Symposium andWorkshop? This is where EPR is really going to beable to solve problems. The future has some excit-ing possibilities. Some day we will look back on the

current S/N and wonder why people say, as theyhave for at least 20 years, that EPR is near the the-oretical limits. Almost nothing that was reportedtoday could have been done even a few years ago.

Harold Swartz from time to time reminds us (andhimself) that some years ago he declared that "theproblem with EPR imaging is that there is nothingto image and no way to image it." At the Sympo-sium and the Workshop he was a strong advocate ofthe current research and imminent clinical applica-tion of EPR imaging. His earlier comments, alongwith such famous quotations as "I think there is aworld market for about five computers" (ThomasWatson, 1943) should be engraved on the portals ofNIH to serve as a reminder to those who serve onstudy sections.

B. Where EPR is Today

The EPR perspective on a problem is very broadindeed (Table 12). Even though, as stated at theoutset, much of what has been done has been CW,linear response, field swept, in homogeneous mag-netic fields, and in one dimension, a few labs haveshown the way with pulsed time domain EPR, ex-tending into 2 and 3 dimensions. Hyde has recentlyopened our eyes to the possibilities of multiquantumEPR.

The goals set in 1987 were ambitiously forward-looking. With all of the exciting new developmentsin EPR, instrumentation and software are still waybehind the needs of researchers. In fact we haven'tcome very far in five years toward the goals set in1987. This statement, which is true with respectto the full scope of the demonstrated possibilities ofEPR, is not meant to in any way detract from thealmost revolutionary advances made by instrumentvendors in the past five years. The Bruker ESP380Ehas capabilities for pulsed X-band EPR that usershave not yet learned to exploit. The Micro-Now8400 bench-top EPR makes it possible to expandthe applications (and importantly the instruction)of CW X-band EPR into labs that previously couldnot afford a spectrometer. The Bruker EMS 104 isthe first spectrometer built for quantitative EPR,a severely under-exploited area. The software be-coming available is of a sophistication well beyondanything even dreamed of a few years ago. Thehopes for extracting information from spectra in thenear future are very bright. At the time of the 1987


Table 12: The EPR Perspective

CW ID 2D 3D 4Dmultifrequency (MHz to THz)multiresonance (ELDOR, ENDOR, TRIPLE)field-swept, frequency-sweptlinearnon-linear (ST-EPR, saturated)ODMR, etc.

pulsedmultifrequencyESE (Ti, T2)saturation recoveryFT-EPRpulsed ENDORpulsed magnetic field

multiquantummult ifrequency

Workshop it was a valid point of view to declare thatthe "new" spectrometers of the day were not enoughof an improvement over existing spectrometers tojustify replacement of a functioning old spectrome-ter. Now, these new instrumentation and softwarecapabilities change the situation entirely.

The extreme importance of multifrequency, mul-tidimensional (and, we project, multiquantum)EPR leaves many regions of the matrix of EPR ob-servables not served by commercial instrumentation.

The crucial issues expressed in 1987 remain - forexample, should the limited R&D effort that is avail-able in EPR be applied to creating the ultimate X-band CW EPR, should it be applied to broadbandEPR, should it be applied to pulsed EPR? In this re-gard it should be recognized that the total R&D ef-fort by commercial EPR instrument manufacturersis about the same as (maybe less than) the instru-mentation research effort in the handful of leading-research labs around the world. In the USA in par-ticular, Bruker and MicroNow are investing in EPRas strongly as they can prudently do so consideringthe magnitude of the EPR instrumentation marketas it exists today. Given the severe limitations onresearch funding, it is not surprising that there is

a "lowest bidder" attitude among purchasers, butthis very attitude causes the capability limitationsabout which people complain.

These companies cannot invest enough to lead inall areas of EPR. The task for the EPR communityis to set some priorities on how the EPR perspec-tive should advance to fill out the matrix of possibleexperiments to enhance problem solving in science.Guiding the selection of research progress for com-mercialization requires a collective wisdom for thegood of science - and for the good of the fields inwhich the results of EPR will be applied. TheseWorkshops are a small step toward channeling thebest ideas of workers in EPR to guide priorities forour future.

C. The Future

It is clear that some applications need techniquesfor identifying many electron spin sites in complexphysical or biological structures. A specific exam-ple is photosynthetic reaction centers. It is also clearthat simulation and visualization of experimental re-sults lag behind the ability to acquire data.

A fundamental quandary for researchers is that

Vol. 16, No. 3/4 187

marketing of spectrometers is demand-driven, butresearch is resource-driven. We hope that the ag-gregate market will permit manufacturers of EPRspectrometers to provide some leadership via mar-keting of spectrometers with capabilities that manyresearchers don't yet know that they need. TheBruker ESP380E is such an example - it has morefunctionality for pulsed X-band EPR than most pur-chasers have been able to exploit.

XIII. Acknowledgment

In this paper GRE and SSE serve as re-porters/reviewers of the information (some of it un-published) and opinions presented at the Workshop.Where researchers' names are associated with par-ticular sections, they reviewed the section beforethe draft was submitted for publication. Numer-ous comments, corrections, and additions by MelvinKlein, Roger Isaacson, Arthur Heiss, Ralph Weber,Philip Morse, Linn Belford, Ron Mason, HarveyBuckmaster, and Howard Halpern helped transformtape recordings and notes from a meeting into thiswritten report. Especially helpful comments andadditional references were provided by Jack Freed,Arthur Schweiger, James Hyde, and Harold Swartz.In some places the wording is nearly verbatim asstated by a presenter or a discussant in the au-dience, or as provided by a participant after theWorkshop. In other parts, the report is a synop-sis and even a rearrangement of order from the oralpresentation. The comments and references in thesection on pulsed EPR are largely as provided byArthur Schweiger. Without the extensive contribu-tions of many people this report would be less com-plete. However, SSE and GRE are responsible forthe final version and the overall focus and emphasisof this prospective on the future of EPR.

Partial support of the Workshop was providedby NIH grant GM46669. Support of the preceding15th International EPR Symposium by Bruker In-struments Inc., Medical Advances Inc., Norell Inc.,Wilmad Glass Inc., Scientific Software Services, andMicro-Now Instruments Inc. also contributed to thesuccess of the Workshop.

XIV. References1 "Pulsed EPR: A new field of applications," C. P.

Keijzers, E. J. Reijerse, and J. Schmidt, eds., NorthHolland, Amsterdam, 1989.

2 "Advanced EPR: Applications in Biology andBiochemistry," A. J. Hoff, ed., Elsevier, Amsterdam,1989.

3 "Modern Pulsed and Continuous-Wave ElectronSpin Resonance," L. Kevan and M. K. Bowman,eds., Wiley, New York, 1990.

4 "EPR Imaging and In Vivo EPR," G. R. Eaton,S. S. Eaton, and K. Ohno, eds., CRC Press, BocaRaton, Florida, 1991.

5S. A. Dikanov and Yu. D. Tsvetkov, "ElectronSpin Echo Envelope Modulation (ESEEM) Spec-troscopy," CRC Press, Boca Raton, Florida, 1992.

6G. R. Eaton and S. S. Eaton, Bull. Magn. Re-son. 10, 3 (1988).

7G. R. Eaton and S. S. Eaton, Spectroscopy 3,34 (1988). Also republished in Spectroscopy Inter-national.

8W. Froncisz, T. Oles, and J. S. Hyde, Rev. Sci.lustrum. 57, 1095 (1986).

9J. S. Hyde, M. E. Newton, R. A. Strangeway, T.G. Camenisch, and W. Froncisz, Rev. Sci. Instrum.62, 2969 (1991).

10J. S. Hyde and W. Froncisz, National Electron-ics Conf. Proc. 35, 602 (1981).

UJ . S. Hyde and W. Froncisz, Ann. Rev. Bio-phys. Bioeng. 11, 391 (1982).

12J. S. Hyde, W. E. Antholine, and R. Basosi,"Sensitivity Analysis in Multifrequency EPR Spec-troscopy, Biological and Inorganic Copper Chem-istry," K. D. Karlin and J. Zubieta eds., AdeninePress, 1985, page 239.

13W. K. Subczynski, S. Lukiewicz, and J. S.Hyde, Magn. Reson. Med. 3, 747 (1986).

14J. S. Hyde, M. Pasenkiewicz-Gierula, R. Ba-sosi, W. Froncisz, and W. E. Antholine, J. Magn.Reson. 82, 63 (1989).

15R. A. Strangeway, T. K. Ishii, and J. S. Hyde,IEEE Trans. Microwave Theory & Technique 36,792 (1988).

16P. Lesniewski and J. S. Hyde, Rev. Sci. In-strum. 61, 2248 (1990).

17T. Oles, R. A. Strangeway, J. Luglio, W. Fron-cisz, and J. S. Hyde, Rev. Sci. Instrum. 63, 4010(1992).


18W. P. Robins, "Phase Noise in Signal Sources(Theory and Applications)," Peter Peregrinus Ltd.on behalf of the Institute of Electrical Engineers,London, 1982, equation 5.27, p. 53.

19W. Froncisz, T. Oles, and J. S. Hyde, Rev. Sci.lustrum. 57, 1095 (1986).

20M. E. Johnson and J. S. Hyde, Biochemistry20, 2875 (1981).

21J. S. Hyde, M. Pasenkiewicz-Gierula, A. Jes-manowicz, and W. E. Antholine, Appl. Mqgn. Re-son. 1, 483 (1990).

22J. S. Hyde, A. Jesmanowicz, J. J. Ratke, andW. E. Antholine, J. Magn. Reson. 96, 1 (1992).

23J. S. Hyde, P. B. Sczaniecki, and W. Froncisz,J. Chem. Soc. Faraday Trans. 7 85, 3901 (1989).

24W. Froncisz, P. B. Sczaniecki, and J. S. Hyde,Physica Medica 5, 163 (1989).

25P. B. Sczaniecki, J. S. Hyde and W. Froncisz,J. Chem. Phys. 93, 3891 (1990).

26P. B. Sczaniecki, J. S. Hyde and W. Froncisz,J. Chem. Phys. 94, 5907 (1991).

27H. S. Mchaourab and J. S. Hyde, J. Chem.Phys. 98, 1786 (1993).

28H. S. Mchaourab, S. Pfenninger, W. E. Antho-line, J. S. Hyde, and P. M. H. Kroneck, Biophys. J.64, 1576 (1993).

29H. A. Buckmaster and J. C. Dering, IEEETrans. lustrum. Meas. IM-16, 13 (1967).

30H. A. Buckmaster and M. J. Kloza, IEEETrans. lustrum. Meas. IM-18, 237 (1969).

31H. A. Buckmaster and J. C. Dering, Can. J.Phys. 43, 1088 (1965).

32H. A. Buckmaster and J. C. Dering, Can. J.Phys. 45, 107 (1967).

33H. A. Buckmaster and J. C. Dering, J. Sci.lustrum. 43, 554 (1966).

34H. A. Buckmaster and J. C. Dering, Proceed-ings of the XlVth Colloque Ampere, Ljubljana,1966, p. 1017, North Holland Publ., 1967.

35H. A. Buckmaster and R. S. Rathie, Can. J.Phys. 49, 853 (1971).

36H. A. Buckmaster and J. D. Skirrow, Can. J.Phys. 50, 1092 (1972).

37G. Feher, R. A. Isaacson, and J. D. McElroy,Rev. Sci. lustrum. 40, 1640 (1969). This paperwas reproduced starting on page 341 of "AdvancedEPR: Applications in Biology and Chemistry," A.J. Hoff, ed., Elsevier, Amsterdam, 1989.

38Douglas A. Knoll, M.S. Thesis, Marquette Uni-versity, March 1989.

39L. Berliner, "The Scope of In Vivo EPR inEPR Imaging and In Vivo EPR," G. R. Eaton, S. S.Eaton, and K. Ohno, eds., CRC Press, Boca Raton,Florida, 1991.

40H. J. Halpern and M. K. Bowman, Low-Frequency EPR Spectrometers: MHz Region, Chap-ter 6 in "EPR Imaging and In Vivo EPR," G. R.Eaton, S. S. Eaton, and K. Ohno, eds., CRC Press,Boca Raton, Florida, 1991.

41L. Berliner and J. Koscielniak, Low FrequencyEPR Spectrometers: L-Band, Chapter 7 in "EPRImaging and In Vivo EPR," G. R. Eaton, S. S.Eaton, and K. Ohno, eds., CRC Press, Boca Ra-ton, Florida, 1991.

42H. M. Swartz and J. F. Glockner, Measure-ment of Oxygen by EPRI and EPRS, chapter 24in "EPR Imaging and In Vivo EPR," G. R. Eaton,S. S. Eaton, and K. Ohno, eds., CRC Press, BocaRaton, Florida, 1991.

43H. M. Swartz and J. F. Glockner, Measure-ments of the Concentration of Oxygen in Biologi-cal Systems Using EPR Techniques, Chapter 21 in"Advanced EPR: Applications to Biology and Bio-chemistry," A. J. Hoff, ed., Elsevier, Amsterdam,1989.

44L. Berliner, Applications of In Vivo EPR,Chapter 25 in "EPR Imaging and In Vivo EPR,"G. R. Eaton, S. S. Eaton, and K. Ohno, eds., CRCPress, Boca Raton, Florida, 1991.

45G. Bacic, M. J. Nilges, R. L. Magin, T. Wal-czak, and H. M. Swartz, Magn. Reson. Med. 10,266 (1989).

46H. M. Swartz, S. Boyer, P. Gast, J. F. Glock-ner, H. Hu, K. J. Liu, M. Moussavi, S. W. Norby,N. Vahidi, T. Walczak, M. Wu, and R. B. Clarkson,Magn. Reson. Med. 20, 333 (1991).

47H. M. Swartz, S. Boyer, D. Brown, K. Chang,P. Gast, J. F. Glockner, H. Hu, K. J. Liu, M. Mous-savi, M. Nilges, S. W. Norby, A. Smirnov, N. Vahidi,T. Walczak, M. Wu, and R. B. Clarkson, The Useof EPR for the Measurement of the Concentrationof Oxygen In Vivo In Tissues Under PhysiologicallyPertinent Conditions and Concentrations, in "Oxy-gen Transport to Tissue XIV," W. Erdmann andD. F. Bruley (eds.), Plenum Publishing Corp., NewYork, 1992, p. 221.

48K. J. Liu, P. Gast, M. Moussavi, S. W. Norby,

Vol. 16, No. 3/4 189

N. Vahidi, M. Wu, and H. M. Swartz, Proc. Natl.Acad. Sci. U.S.A. 90, 5438 (1993).

49L. J. Berliner and H. Fujii, Science 227, 517(1985).

50J. Gorcester and J. H. Freed, J. Chem. Phys.88, 4678 (1988).

5 1J. Gorcester, G. L. Milhauser, and J. H.Freed, Two-Dimensional Electron Spin Resonance,in "Modern Pulsed and Continuous-Wave ElectronSpin Resonance," L. Kevan and M. K. Bowman,eds., Wiley-Interscience, 1990.

52K. P. Dinse, M. Pluschau, G. Kroll, T. Prisner,and H. van Willigen, Bull. Magn. Reson. 11, 174(1989).

53A. Angerhofer, M. Toporowicz, M. K. Bow-man, J. R. Norris, and H. Levanon, J. Phys. Chem.92, 7164 (1988).

54J. Gorcester and J. H. Freed, J. Chem. Phys.85, 5375 (1986).

5 5J. Gorcester, S. B. Rananavare, and J. H.Freed, J. Chem. Phys. 90, 5764 (1989).

56U. Ewert, R. H. Crepeau, C. R. Dunnam, D.Xu, S.-Y. Lee, and J. H. Freed, Chem. Phys. Lett.184, 25 (1991).

57U. Ewert, R. H. Crepeau, S.-Y. Lee, C. R. Dun-nam, D. Xu, and J. H. Freed, Chem. Phys. Lett.184, 34 (1991).

5 8B. R. Patyal, R. H. Crepeau, D. Gambiel, andJ. H. Freed, Chem. Phys. Lett. 173, 443 (1990).

5 9B. R. Patyal, R. H. Crepeau, D. Gambiel, andJ. H. Freed, Chem. Phys. Lett. 175, 453 (1990).

60S.-Y. Lee, B. R. Patyal, and J. H. Freed, A J.Chem. Phys. 221, 397 (1993).

61D. H. Shin, J. L. Dye, D. E. Budil, K. A. Earle,and J. H. Freed, J. Chem. Phys. 97, 1213 (1993).

6 2D. E. Budil, K. E. Earle, and J. H. Freed, J.Chem. Phys. 97, 1294 (1993).

63D. E. Budil, K. A. Earle, W. B. Lynch, andJ. H. Freed, Electron Paramagnetic Resonance at 1Millimeter Wavelengths, in "Advanced EPR: Appli-cations in Biology and Chemistry," A. J. Hoff, ed.,Elsevier, Amsterdam, 1989.

64W. B. Lynch, K. A. Earle, and J. H. Freed,Rev. Sci. lustrum. 59, 1345 (1988).

6 5J. H. Freed, J. Chem. Soc. Faraday Trans. 86,3173 (1990).

66 "Proceedings of the National High MagneticField Laboratory Workshop on Nuclear MagneticResonance," edited by E. R. Andrew, T. Mareci,

and N. S. Sullivan, Published by the University ofFlorida, Gainesville, Florida, 1992. This book con-tains contributions on high-field EPR by J. H. Freed,G. R. Eaton and R. Weber.

6 7S. S. Eaton and G. R. Eaton, Magn. Reson.Rev. 16, 157 (1993).

68P. L. Richards, W. S. Caughey, H.Eberspaecher, G. Feher, and M. Malley, J. Chem.Phys. 47, 1187 (1967).

69W. B. Minis, Electron Spin Echoes, in "Elec-tron Paramagnetic Resonance," S. Geschwind, ed.,Plenum, New York, 1972, p. 263.

70K. M. Salikhov, A. G. Seminov, and Yu. D.Tsvetkov, "Electron spin echoes and their applica-tions," Nauka, Novosibirsk, 1976 (in Russian).

71K. J. Standley and R. A. Vaughan, "Elec-tron Spin Relaxation Phenomena in Solids," PlenumPress, New York, 1969.

72H. Thomann, L. R. Dalton, and L. A. Dalton,Biol. Magn. Reson. 6, 143 (1984).

73H. Thomann and W. B. Mims, Pulsed electron-nuclear spectroscopy and the study of metallopro-tein active sites, in "Pulsed Magnetic Resonance:NMR, ESR, and Optics," D. M. S. Bagguley, ed.,Clarendon Press, Oxford, 1992, p. 362.

74Yu. D. Tsvetkov and S. A. Dikanov, Met. Ions.Biol. Syst. 22, 207 (1987).

75S. A. Dikanov and Yu. D. Tsvetkov, J. Struct.Chem. 26, 766 (1985).

76A. Schweiger, Angew. Chem. Int. Ed. Engl.30, 265 (1991).

77Th. Wacker, G. A. Sierra, and A. Schweiger,Isr. J. Chem. 32, 305 (1993).

78A. Schweiger, Pure Appl. Chem. 64, 809(1992).

79G. A. Sierra, A. Schweiger, and R. R. Ernst,Chem. Phys. Lett. 184, 363 (1991).

80Th. Wacker and A. Schweiger, Chem. Phys.Lett. 191, 136 (1992).

81Th. Wacker and A. Schweiger, Chem. Phys.Lett. 186, 27 (1991).

82C. Gemperle, A. Schweiger, and R. R. Ernst,Chem. Phys. Lett. 145, 1 (1988).

83A. V. Astashkin and A. Schweiger, Chem.Phys. Lett. 174, 595 (1990).

84M. K. Bowman, R. J. Massoth, and C. S. Yan-noni, Coherent Raman Beats in Electron Paramag-netic Resonance Spectroscopy, in "Pulsed MagneticResonance: NMR, ESR, and Optics," D. M. S. Bag-


guley, ed., Clarendon Press, 1992, p. 423.85A. Schweiger and R. R. Ernst, J. Magn. Reson.

77, 512 (1988).86A. Calligiani, M. Giordano, D. Leporini, M.

Lucchesi, M. Martinelli, L. Pardi, and S. Santucci,Appl. Magn. Reson. 3, 107 (1992).

87K. P. Dinse, Pulsed ENDOR, Chapter 17 in"Advanced EPR. Applications in Biology and Bio-chemistry," A. J. Hoff, ed., Elsevier, Amsterdam,1989.

88R. B. Clarkson, M. D. Timken, D. R. Brown,H. C. Crookham, and R. L. Belford, Chem. Phys.Lett. 163, 277 (1989).

89R. B. Clarkson, D. R. Brown, J. B. Cornelius,H. C. Crookham, W.-J. Shi, and R. L. Belford, PureAppl. Chem. 64, 893 (1992).

9 0B. E. Sturgeon and D. M. Britt, Rev. Set. In-strum. 63, 2187 (1992).

91G. J. Gerfen and D. J. Singel, J. Phys. Chem.93, 4571 (1990).

9 2D. J. Singel, Multifrequency ESEEM: Perspec-tives and Applications, in "Advanced EPR: Appli-cations in Biology and Biochemistry," A. J. Hoff,ed., Elsevier, Amsterdam, 1989, p. 119.

93R. G. Larsen, G. J. Gerfen, and D. J. Singel,Appl. Magn. Reson. 3, 369 (1992).

94E. J. Reijerse and S. A. Dikanov, Pure Appl.Chem. 64, 789 (1992).

9 5S. L. Tan, J. S. Waugh, and W. H. Orme-Johnson, J. Chem. Phys. 81, 576 (1984).

96C. Gemperle, A. Schweiger, and R. R. Ernst,Chem. Phys. Lett. 178, 565 (1991).

97A. Schweiger, L. Braunschweiler, J.-M. Fauth,and R. R. Ernst, Phys. Rev. Lett. 54, 1241 (1985).

98A. Schweiger, New Trends in Pulsed ElectronSpin-Resonance Methodology, in "Modern Pulsedand Continuous-Wave Electron Spin Resonance," L.Kevan and M. K. Bowman, eds., Wiley-Interscience.New York, 1990, p. 43.

99A. Schweiger, C. Gemperle, and R. R. Ernst,J. Magn. Reson. 86, 70 (1990).

100E. J. Hustedt, A. Schweiger, and R. R. Ernst,J. Chem. Phys. 96, 4954 (1992).

101H. Cho, S. Pfenninger, C. Gemperle, A.Schweiger, and R. R. Ernst, Chem. Phys. Lett.160, 31 (1989).

102J.-M. Fauth, A. Schweiger, and R. R. Ernst,J. Magn. Reson. 81, 262 (1989).

103C. Gemperle, Thesis, ETH 1992.

104P. Hofer, A. Grupp, H. Nebenfuhr, and M.Mehring, Chem. Phys. Lett. 132, 279 (1986).

105C. Gemperle, G. Aebli, A. Schweiger, and R.R. Ernst, J. Magn. Reson. 88, 241 (1990).

106P. Hofer, Experimental Aspects of 2D ESEEMTechniques, Proceedings EMARDIS 91, World Sci-entific Publ., Singapore, 1991, p. 1.

107J. J. Shane, P. Hofer, E. J. Reijerse, and E.de Boer, J. Magn. Reson. 99, 596 (1992).

108D. Gamiel and J. H. Freed, J. Magn. Reson.89, 60 (1990).

109J. M. Fauth, A. Schweiger, L. Braunschweiler,J. Forrer, and R. R. Ernst, J. Magn. Reson. 66, 74(1986).

110M. Mehring, P. Hofer, and A. Grupp, Phys.Chem. 91, 1132 (1987).

111 A. Grupp and M. Mehring, Pulsed ENDORSpectroscopy in Solids, in "Modern Pulsed andContinuous-Wave Electron Spin Resonance," L. Ke-van and M. K. Bowman, eds., Wiley-Interscience,New York, 1990, p. 195.

112C. Gemperle and A. Schweiger, Chem. Rev.91, 1481 (1991).

113H. Thomann and M. Bernardo, Isr. J. Chem.32, 323 (1993).

114H. Thomann and M. Bernardo, Spectrosc. Int.8, 119 (1990).

115H. Thomann and M. Bernardo, Advances inElectron Nuclear Double Resonance Spectroscopy,Chapter 4 in "Magnetic Resonance of Carbona-ceous Solids," R.E. Botto and Y. Sanada, eds., ACSSymposium Series 229, American Chemical Society,Washington, D.C., 1993, pp. 65-90.

116C. Gemperle, O. W. Sorensen, A. Schweiger,and R. R. Ernst, J. Magn. Reson. 87, 502 (1990).

U 7 C . Buhlman, A. Schweiger, and R. R. Ernst,Chem. Phys. Lett. 154, 285 (1989).

U 8H. Thomann and M. Bernardo, Chem. Phys.Lett. 169, 5 (1990).

119H. Cho, S. Pfenninger, J. Forrer, and A.Schweiger, Chem. Phys. Lett. 180, 198 (1991).

120M. Mehring, P. Hofer, H. Kaas, and A. Grupp,Europhys. Lett. 6, 463 (1988).

121P. Hofer, A. Grupp, and M. Mehring, Phys.Rev. A33, 3519 (1986).

122H. Cho, J. Chem. Phys. 94, 2482 (1991).123M. Mehring, P. Hofer, and A. Grupp, Phys.

Rev. A33, 3523 (1986).

Vol. 16, No. 3/4 191

124J. Forrer, S. Pfenninger, J. Eisenegger, and A.Schweiger, Rev. Sci. lustrum. 61, 3360 (1990).

125H. Thomann and M. Bernardo, unpublishedresults.

126H. Thomann and M. Bernardo, Methods En-zymol. 227, 118 (1993).

127V. V. Kurshev, A. Raitsimring, and Ytf. D.Tsvetkov, J. Magn. Reson. 81,441 (1989).

128Z. Levi, A. Raitsimring, and D. Goldfarb, J.Phys. Chem. 95, 7830 (1991).

129V. V. Kurshev, A. M. Raitsimring, and T.Ichikawa, J. Phys. Chem. 95, 3564 (1991). Notethat this paper corrected a few errors in the equa-tions presented in previous papers.

130A. Raitsimring, J. Peisach, H. C. Lee, and X.Chen, J. Phys. Chem. 96, 3526 (1992).

131J. S. Hyde, W. Froncisz, and C. Mottley,Chem. Phys. Lett. 110, 621 (1984).

132J. P. Hornak and J. H. Freed, Chem. Phys.Lett. 101, 115 (1983).

133S. K. Rengen, V. R. Bhagat, V. S. S. Sastry,and B. Venkataraman, J. Magn. Reson. 33, 227(1979).

134A. D. Milov, K. M. Salikhov, and M. D. Shchi-rov, Sov. Phys. Solid State 23, 565 (1981).

135G. Volkel, S. A. Dzuba, A. Bartl, W. Brunner,and J. Frohner, Phys. Stat. Sol. 85, 257 (1984).

136P. Dirksen, A. Henstra, and W. Th. Wenke-bach, J. Phys. Condens. Matter 1, 8535 (1989).

137C. G. Maresch, M. Weber, A. A. Dubinskii,and H. W. Spoess, Chem. Phys. Lett. 193, 134(1992).

138S. S. Eaton and G. R. Eaton, ElectronSpin Resonance Imaging, in "Modern Pulsed andContinuous-Wave Electron Spin Resonance," L. Ke-van and M. K. Bowman, eds., Wiley-Interscience,New York, 1990, p. 405.

139S. S. Eaton and G. R. Eaton, Pulsed EPRImaging, in "EPR Imaging and In Vivo EPR," G.R. Eaton, S. S. Eaton, and K. Ohno, eds., CRCPress, Boca Raton, 1991, p. 73.

140D. J. Sloop, H.-L. Yu, and T.-S. Lin, Chem.Phys. Lett. 124, 456 (1986).

141A. D. Milov, A. Yu. Pusep, S. A. Dzuba, andYu. D. Tsvetkov, Chem. Phys. Lett. 119, 421(1985).

142G. R. Eaton and S. S. Eaton, J. Magn. Reson.67, 73 (1986).

143J.-L. Du, K. M. More, S. S. Eaton, and G. R.Eaton, Isr. J. Chem. 32, 351 (1993).

144S. Pfenninger, A. Schweiger, J. Forrer, and R.R. Ernst, Chem. Phys. Lett. 151, 199 (1988).

145G. A. Sierra, A. Schweiger, and R. R.Ernst, Anisotropy-Resolved EPR, Proceedings 26thCongress Ampere, Athens, 1992, p. 270.

146D. Stehlik, C. H. Bock, and M. C. Thur-nauer, Transient EPR - Spectroscopy of Photoin-duced Electronic Spin States in Rigid Matrices, in"Advanced EPR: Applications in Biology and Bio-chemistry," A. J. Hoff, ed., Elsevier, Amsterdam,1989, p. 371.

147J. Fessmann, N. Rosch, E. Ohmes, and G.Kothe, Chem. Phys. Lett. 152, 491 (1988).

148J. Isoya, H. Kanda, J. R. Norris, J. Tang, andM. K. Bowman, Phys. Rev. B41, 3905 (1990).

149W. Froncisz and J. S. Hyde, J. Magn. Reson.47, 515 (1982).

150J. S. Hyde and W. Froncisz, Loop Gap Res-onators, in "Advanced EPR: Applications in Biol-ogy and Biochemistry," A. J. Hoff, ed., Elsevier,Amsterdam, 1989, p. 277.

151S. Pfenninger, J. Forrer, and A. Schweiger,Rev. Sci. lustrum. 59, 752 (1988).

152J. Forrer, S. Pfenninger, B. Wagner, and Th.Weiland, Pure Appl. Chem. 64, 865 (1992).

153M. Mehring and E. Freysoldt, J. Phys. E: Sci.lustrum. 13, 894 (1980).

154J. P. Hornak and and J. H. Freed, J. Magn.Reson. 62, 311 (1985).

155R. W. Dykstra and G. D. Markham, J. Magn.Reson. 69, 350 (1986).

156W. M. Walsh and L. W. Rupp, Jr., Rev. Sci.lustrum. 57, 2278 (1986).

157P. Hofer, Bruker's Dielectric Resonator forFT-EPR, Bruker Report 2/1989, p. 4.

158R. T. Weber, J. A. M. Disselhorst, L. J. Prevo,J. Schmidt, and Th. Wenkenbach, J. Magn. Reson.81, 129 (1989).

159A. Yu. Bresgunov, A. A. Dubinskii, V. N.Krimov, Yu. G. Petrov, O. G. Poluektov, and Ya.S. Lebedev, Appl. Magn. Reson. 2, 715 (1987).

160R. W. Quine, G. R. Eaton, and S. S. Eaton,Rev. Sci. lustrum. 58, 1709 (1987).

161O. Burghaus, A. Toth-Kischkat, R. Klette,and K. Mobius, J. Magn. Reson. 80, 383 (1988).

162O. Burghaus, M. Rohrer, T. Gotzinger, M.Plato, and K. Mobius, Meas. Sci. Technol. 3, 765


(1992).163O. Burghaus, M. Plato, D. Bumann, B. Neu-

mann, W. Lubitz, and K. Mobius, Chem. Phys.Lett. 185, 381 (1991).

164T. Kojima, Y. Haruyama, H. Tachibana, R.Tanaka, J. Okamoto, H. Hara, and Y. Yamamoto,Appl. Rad. Isotopes 44, 361 (1993).

165D. Maier and D. Schmalbein, Appl. Rad. Iso-topes 44, 345 (1993). A description of the BrukerEMS 104 spectrometer. See also, Bruker Report91/92.

166P. A. Tipton, J. McCracken, J. B. Cornelius,and J. Peisach, Biochemistry 28, 5720 (1989).

167A. Nayeem, S. B. Rananavare, V. S. S. Sastry,and J. H. Freed, J. Chem. Phys. 91, 6889 (1989).

168R. R. Ernst, Adv. Magn. Reson. 2, 1, 1966.169B. L. Bales, J. Magn. Reson. 38, 193 (1980).170H. J. Halpern, M. Peric, T-D. Nguyen, D.

Spencer, B. A. Teicher, Y. J. Lin, and M. K. Bow-man, J. Magn. Reson. 90, 40 (1990).

m H . J. Halpern, D. P. Spencer, J. vanPolen, M.K. Bowman, A. C. Nelson, E. M. Dowey, and B. A.Teicher, Rev. Sci. Instrum. 60, 1040 (1989).

172M. J. Nilges, T. Walczak, H. M. Swartz, Phys-ica Medica 5, 195 (1989).

173J. S. Stamler, D. J. Singel, and J. Loscalzo,Science 258, 1898 (1992).

174K. R. Maples, T. Sandstroem, Y. F. Su, andR. Henderson, Am. J. Respir. Cell Mol. Biol. 4,538 (1991).

175J. S. Hyde and J. Gajdzinski, Rev. Sci. In-strum. 59, 1352 (1988).

the future of epr - weizmann institute of science

Documents