We proposed a public sensitivity test of Bruker and JEOL Electron Spin Resonance instruments at the Annual Meeting of the Oxygen Club of California, 1998 World Congress Free Radical School (February 5-8, St. Barbara, CA) as part of the Electron Spin Resonance workshop conducted there.

The central problem with biological Electron Spin Resonance is, in our opinion, the low sensitivity of Electron Spin Resonance instruments relative to the concentration of free radicals in vivo or even in vitro. It is our impression that the sensitivity of Electron Spin Resonance instruments has not increased since the TM110 cavity was introduced by Dr. Hyde of Varian in the early Seventies.

A Varian instrument has a sensitivity of 1.7 x 1010 Δ H spins. Δ H is defined as the original width in Gauss without saturation at half maximum of absorption signal with a 1-second integration time. Assuming the sample having negligible dielectric loss with Δ H=1.0 Gauss the sensitivity would be

1.7 x 10 10 spins = 2.8 x 10-14 moles

Avogadro´s number

Unfortunately the dielectric loss of water, and the resultant low Q and small sample size results in much poorer sensitivity than might be expected from that of the weak pitch sample. Assuming a maximum aqueous volume of 200 ml (TM110 cavity with a 17-mm flat cell), 2.8x10-14 moles would correspond to 1.4 x 10-10 M. In practice the molar sensitivity of Electron Spin Resonance for detecting free radicals in biological systems is nowhere near 10-10 M.

The proposed performance tests consist of three parts. The first test compares calculated sensitivity of Electron Spin Resonance instruments based on the traditional Varian weak pitch signal-to-noise (S/N) test as developed by Dr. Hyde. The second test uses the spectrometer settings for ice-like sample under low power conditions. The third test compares signal-to-noise values obtained under optimized conditions using a low concentration (10-7 M) nitroxide sample in aqueous buffer. We ran these tests under with our Varian and Bruker spectrometers located in the Free Radical Metabolite Section at NIEHS/NIH. At the conference, Bruker model EMX and JEOL model FR30 were subjected to identical tests. The detailed test plans and results are presented in following sections. The biggest difficulty in these comparisons is the subjective determination of the rms noise, which in practice was determined by Dr. Yang Fann with the help of a ruler and not as the p-p observed noise divided by 2.5 as indicated in the Test instructions.

Test 1

Standard High Power S/N Ratio of Varian Weak Pitch

 Note that according to the Varian traditional signal-to-noise test, the signal amplitude is measured at 12.5mW, the highest microwave power level at which the pitch signal amplitude can be reliably measured (i.e. without saturation). To obtain the unsaturated signal amplitude for the 200 mW level, multiply the 12.5 mW amplitude by the factor (200/12.5)½ = (16)½ = 4. THE NOISE LEVEL IS OBSERVED AND MEASURED DIRECTLY AT 200 mW. The rms noise amplitude is assumed to be the p-p observed noise divided by 2.5. The computed signal to noise ratio is (4 x 2.5) or 10 times the observed p-p signal height (at 12.5 mW) divided by the observed p-p noise amplitude (200 mW) when both signal and noise are measured at the same spectrometer settings. The receiver gain should be increased, if necessary, during noise level measurements in order to obtain approximately one major chart division of noise to insure that a sluggish recorder cannot generate an unduly optimistic test result. (The noise level is normalized to a gain setting consistent with that used to observe the signal level when the S/N is computed).

Spectrometer settings for the high power S/N test:

  1. Record Electron Spin Resonance weak pitch signal by setting controls as follows:
    • ATTENTUATION/POWER 12 dB (12.5 mW)
    • FIELD SET 3395 G (approx.)
    • GAIN Adjust for full-scale recorder signal (5000 approx.)
    • MODULATION 8.0 Gauss
    • TIME CONSTANT 1.0 sec
    • SCAN RANGE 40 G
    • SCAN TIME 4 min
  2. Record noise trace:
    • Reset following controls:
    • FIELD SCAN 1500 G
    • SCAN RANGE OFF
    • ATTENUATION/POWER 0 dB (200 mW)
  3. The S/N ratio is calculated as the p-p weak pitch amplitude divided by the rms noise amplitude and corrected by the Varian "gold standard" sample correction factor.

The results of this performance test are summarized in Table I. Using this traditional test the new Bruker high Q cavity with an EMX instrument was a revolutionary advance over all previous instruments. The JEOL FR30 instrument was severely handicapped because it operates at a maximum microwave power of 16 mW and the S/N was not scaled to 200mW thus resulting in a significant loss in calculated signal-to-noise. The JEOL TE series Electron Spin Resonance spectrometer is capable of producing 200mW microwave power and this instrument should be used next time this comparison is done.

Table I.

Performance Test 1

(weak pitcha)

Spectrometer Varian E-109 Bruker ESP300 Bruker ECS106 Bruker EMX JEOL FR30
S/N (cavity)b 207

(TE)
301d

(TE)
283

(TE)
3389

(SHQ)
303

(TE011)
Conditions c RG(S): 4 x 105

RG(N): 4 x 106
RG(S): 5 x 104

RG(N): 1 x 105
RG(S): 1 x 105

RG(N):2.5x105
S: 1 mW RG(S): 1 x 105

N: 200 mW

RG(N): 1 x 106
S: 4 mW

RG(S): 200

N: 16 mW

RG(N): 2000
  1. The same Varian weak pitch with a correction factor of 0.895 was used through out the tests.
  2. Cavity used: TE ­ TE102 ; SHQ ­ super high Q; TE011
  3. Standard conditions except otherwise noted: microwave power 12.5 mW (or under saturation) for signal (S), 200 mW for noise (N), modulation amplitude (MA), 8 G; time constant (TC), 1.0 sec; sweep width (SW), 40 G; scan time, 4 min or less; receiver gain (RG)
  4. This spectrometer did not meet the Bruker S/N specification which is 330.

The traditional weak pitch test assumes the sample does not saturate so the larger the B1 field at the sample the better. We asked Bruker and JEOL to perform the weak pitch test at lower power. The chosen power was set where the weak pitch signal is just below saturation. Where saturation is defined as no longer linear in the square root of microwave power, which according to Varian is 12.5 mW in a TE102 cavity. In theory this S/N should be (200 mW/12.5 mW)1/2 =161/2 = 4 times less than the traditional one, if the noise is microwave power independent. This test removes the bias of the larger B1 is better in the traditional test. 12.5 mW is a typical power for frozen biological line samples at 77 K and we thought this low power weak pitch test would be more relevant for Electron Spin Resonance measurements of HbNO and related nitric oxide complexes.

Test 2

Sensitivity Test for Low Dielectric (Ice-like) Sample at Biologically Relevant Microwave Power.

  1. All spectrometer settings are the same for the signal trace as used in the high power S/N test described in the previous section. The controls were carefully adjusted to maximize the signal amplitude at the calibrated 12.5 mW microwave power level and 8 G p-p modulation level. According to Varian 12.5 mW is just under saturation (i.e. the highest power that the signal is proportional to the square root of the power level). This condition was experimentally determined for the power settings when cavities other than the TE102 were used. 
  2. Reset controls to the same settings as for the noise trace in the high power S/N test except the same microwave power was used for both signal and noise traces.

The test results are shown in Table II. In this test the JEOL FR30 did much better than the Varian or Bruker instruments in our laboratory but again the new high Q cavity of Bruker gave the best performance. 

Table II.

Performance Test 2

(Low Powera)

Spectrometer Varian
E-109
Bruker
ESP300
Bruker
ECS106
Bruker
EMX
JEOL
FR30
S/N (cavity)b 54

(TE)
87

(TE)
82

(TE)
580

(SHQ)
281

(TE011)
Conditions c RG(S): 4 x 105

RG(N): 4 x 106
RG(S): 5 x 104

RG(N): 1 x 105
RG(S): 1 x 105

RG(N):2.5x105
S: 1 mW

RG(S): 1 x 105

N: 1 mW

RG(N): 4 x 106
S: 4 mW

RG(S): 200

N: 4 mW

RG(N): 2000
  1. Varian weak pitch with a correction factor of 0.895 was used through out the tests.
  2. Cavity used: TE ­ TE102 ; SHQ ­ super high Q; TE011
  3. Standard conditions except otherwise noted: microwave power 12.5 mW (or under saturation) for signal (S) and noise (N), modulation amplitude (MA), 8 G; time constant (TC), 1.0 sec; sweep width (SW), 40 G; scan time, 4 min or less; receiver gain (RG)

For the last 25 years the TM110 cavity with a 17mm flat cell as developed by Dr. Hyde has provided the highest S/N for free radicals in aqueous samples. We proposed sensitivity tests on an aqueous sample at low concentration to challenge manufacturers of Electron Spin Resonance instruments to improve the molar sensitivity of dilute aqueous nitroxide samples, which would greatly facilitate the spin trapping studies in our laboratory.

Test 3

Aqueous Sensitivity Test Protocols

A dilute aqueous of 2,2,6,6-tetramethyl-4-hydroxypiperidine-1-oxyl (TEMPOL, a stable nitroxide free radical) in pH=7.4 phosphate buffer treated with Chelex and 50 mM DTPA was provided. The best S/N without obvious distortion of the line heights (the positive deflection of the line will be no more than 105% of the negative deflection) is the objective. The microwave power and modulation amplitude should be adjusted to maximize the S/N. The total scan time will be 20 minutes or less. The cavity type and sample holder are open. For the scan used for the noise trace an aqueous sample without nitroxide will be used with the gain increased to give 3 cm of noise on the recorder chart and the S/N scaled as necessary. In this test, the nitroxide sample was prepared on-site right before the test and the same sample and buffer solutions were used for both Bruker EMX and JEOL FR30 instruments. The test results are presented in Table III.

Table III.

Performance Test 3

(10-7 M TEMPOL a)

Spectrometer Varian
E-109
Bruker
ESP300
Bruker
ECS106
Bruker
EMX
JEOL
FR30
S/N (cavity)b 34

(TM)
39

(TM)
31

(TM)
121/61d

(SHQ)
12

(TE011)
Conditionsc S(N): 150 mW

MA: 3.2 G

TC: 0.5 sec
 
RG(S): 5 x 104

RG(N):2.5x106
S(N): 100 mW

MA: 3.6 G

TC: 1.3 sec

RG(S): 5 x 104

RG(N): 5 x 105
S(N): 100 mW

MA: 3.2 G

MF: 50 KHz

TC: 1.3 sec

RG(S): 2.5x104

RG(N): 5 x 104
S(N):50.16mW

MA: 3.2 G

TC: 0.082 sec

RG(S): 5 x 104

RG(N): 5 x 104114 scans
S (N): 4 mW

MA: 1.6 G

TC: 0.3 sec

RG(S): 5000

RG(N): 500010 scans
  1. A 10-7 M solution of TEMPOL in pH=7.4 phosphate buffer treated with Chelex and 50 mM DTPA was used for the test (actual concentration was 7.6 x 108 M determined by UV-VIS at 241 nm with an extinction coefficient of 2915 M-1cm-1).
  2. Cavity used: TM­TM110 with a 200 ml flat cell; SHQ ­ super high Q with a 100 ml flat cell; TE011 with a 100 ml flat cell.
  3. Spectrometer conditions in each case were optimized for maximum signal height without distortion, 100 kHz modulation frequency (MF) was used unless noted; sweep width (SW), 50 G; number of scan, 1 unless noted; modulation amplitude (MA); time constant (TC); receiver gain (RG); total scan time, 20 min or less.
  4. A background signal at g=2.00 was detected in the noise trace. Two S/N numbers were reported with and without counting this background signal as noise.

Again the new Bruker high Q cavity gives highly significant improvement with 4 times the signal-to-noise of instruments in our laboratory. Unfortunately the high Q cavity has a background signal at twice the signal-to-noise, which if counted as noise decreases this advantage to a factor of 2. This improvement was gained with a flat cell with an active volume of 100 ml instead of the 200 ml of the 17-mm flat cell used in the TM110 cavity. The S/N of the JEOL FR30 was disappointing. Perhaps with its higher power the JEOL TE 200 would be more sensitive.

We have tried to make these tests as fair and open as possible, we believe they represent the capabilities of the instruments tested. We wish to thank both JEOL and Bruker for their full cooperation and their significant expenditures of both time and money.

After we wrote up the results of the public Electron Spin Resonance test we received information from JEOL, which we have paraphrased below. Their measurements in Japan show the sensitivity results for the TE-200 evaluated for weak pitch (sensitivity (I) and (II) and for TEMPOL in aqueous solution of 10-7M (sensitivity (III) are much superior to the FR30, which is limited by a maximum incident power of 16 mW.

On the basis of this information, we believe that the JEOL TE-200 and the EMX with its high Q cavity should be publicly compared side by side at the 21st International EPR Symposium in Denver in July in order to determine the relative sensitivity of these instruments.

Sincerely

Drs. Ronald P. Mason and Yang C. Fann
Free Radical Metabolite Section
Laboratory of Pharmacology and Chemistry
National Institute of Environmental Health Sciences
National Institutes of Health

Contact

Michael B. Fessler, M.D.
Michael B. Fessler, M.D.
Chief, Immunity, Inflammation, and Disease Laboratory and Principal Investigator
Tel 984-287-4081
[email protected]
P.O. Box 12233
Mail Drop D2-01
Durham, NC 27709