F19 Detection

F19 is a sensitive nucleus to detect. On most probes it is detected through the same RF and preamp as for H1. Unlike H1, F19 has an extremely wide chemical shift range, spanning +/- 500ppm depending on the molecule and structure. In a majority of flourinated organic compounds, the chemical shift ranges from 100 to -250ppm. F19 also has strong homonuclear (among adjacent F19 nuclei) and heteronuclear (to H1, C13, etc.) J-couplings. Note the following when setting up a F19 experiment:

• Get a good estimate of chemical shift range (relative to CFCl3 at 0.0ppm) for your compound and center your spectrum accordingly (see a F19 chemical shift table here).
• For an unknown chemical shift range, first try a large spectral width (~ 500000 Hz or largest allowed) with the transmitter offset (tof, center of spectrum) set to the center of the guess range, or to 0.0ppm if unknown. The large spectral width prevents any aliasing (or wrapping around of peaks) from outside the window.
• Note pulse bandwidth limitation and pulse offset effects on phasing, excitation and integration accuracy: The F19 pulse (~ 10 usec) is not able to excite all peaks across a very large window (see example below). To locate all F19 peaks, try to scan through the entire range by collecting a spectrum array with the center offset changing from ~ +500ppm to -500ppm with a ~25000 Hz stepsize (see procedure below). After the shift range is known, use movesw to narrow down the window to signal region. For extremely large chemical shift span, you may want to set window to one narrow area at a tme and collect several spectra. Baseline rolling and phase distortion are common with large spectral window.
• In a coupled spectrum (no H1 decoupling applied) and for a spectral window of ~ 10000Hz or less, F19 integration can be as good as for H1 if proper recycle time and pulse width are set according to T1max. Much larger window will limits integration accuracy (see integration results in example below)

Referencing

Choose one of the following (among others) to reference your F19 spectrum:

1. For most accurate and convenient chemical shift referencing, use TMS at < 1% as an internal 0.0ppm reference for H1. F19 is indirectly referenced relative to the TMS H1 peak accordingly to IUPAC standard.
2. Alternatively, for rough referencing, type setref to use the lock signal to reference any nucleus in the direct dimension. Note the setref command needs solvent to be set correctly and locked. See this page for more details.
3. If locking is not possible, use neat CFCl3 as an external reference at 0.0ppm and run experiment unlocked.

Procedure for H1 coupled F19 experiment on nmr500

Step 1: Optional: Collect and reference H1 spectrum

• Lock and shim as usual.
• Under exp1 (type jexp1 to go to exp1) or another experiment, collect a 1H spectrum. Reference TMS signal to 0.0pm. For rough referencing with the lock signal, simply type setref.
• Join a different experiment (i.e. type jexp2 to join exp2. To create it, type cexp(2) jexp(2).
• Type mf(1,2) wft to move data over from exp1 to exp2. Double check referencing.

Step 2: Optional: Estimate F19 transmitter offset to use (Skip this step if you do not know what ppm range your F19 signals cover)

• Type dn='F19' to set channel 2 to F19.
• Type setrefppm
• Enter 1 for 1st decoupler nucleus
• Enter the ppm value you want to use as center of F19 spectrum: (i.e. -100)
• Write down the offset value printed on screen.

Step 3: Change F19 filter and tune probe

• Type setexp('F19') to load standard F19 detection parameters. Type su.
• Replace the H1 bandpass filter on the right side of the H1/F19 preamp with a F19 bandpass filter (labeled).
• Tune channel 1 to F19 following the same procedure as you would for H1. Make sure cables are reconnected back to normal configuration after tuning.

Step 4: Set up F19 experiment

• If you have done Step 2, type tof=### where ### is the value from Step 2. Otherwise, use default value (set to ~ -92ppm).
• For first time NMR work on a compound, try a few test runs to see where the signals are and set an appropriate spectral width (sw) and center (tof).
• Type sw=500000 (or maximum allowed by software). For konwn chemical range, set sw and tof accordlingly.
• Set gain=36 for quick run (gain='n' for autogain is fine but slower). Set nt=4 (or larger if signal is weak). Type go to collect data, and wft f full when done. If receiver or ADC overflows, drop gain by 2 at a a time.
• Type aph to phase spectrum. You may see some baseline rolling due to the large spectral width and phasing.
• Put box cursor to enclose the peak regions but leave ~10% baseline area on both sides. Type movesw to set new spectral window to this area.
• Type go to re-collect data and wft f full to process it.
• To reference spectrum:
• Type mref(1,2) to use referencing in exp1 (1H spectrum) to reference F19 spectrum in exp2 (current)
• OR: type setref to use lock signal for rough referencing

Default parameters

• d1=4.8 at=2.5 tpwr=57 pw=10 sw=100000 nt=4
• tof centered at ~ -92.5 ppm

Step 5: Collect final spectrum

• Once an appropriate spectral window is set, set nt to multiple of 4 and collect data.
• For uniform excitation, and best baseline and integration accuracy involving multiple F19 peaks, it is better to collect several spectra with a smaller (~10000-20000 Hz) sw centered around different signal regions.

Step 6: Processing

• Type wft f full aph. Adjust lb (lb=0.2 or lb=0.5 ..), followed by wft, to smooth spectrum without adding too much line-broadening.
• Double check referencing with mf(1,2) or setref
• You may see baseline rolling due to the large spectrum width of F19 after phasing. To minimize its impact, try the following:
• Make sure you have not accidentally applied a very large first order phase correction (lp parameter). Type lp?. If the value is over a few hundreds, it means it may have caused the baseline rolling, sometime caused by incorrect manual phasing. Type lp=0 to turn the 1st-order phasing to zero. Then, type aph0 to only apply zero-order phasing. If that does not phase all peaks across spectrum,, try aph to include 1st order phasing.
• To fix small baseline rolling (as shown in example below), try baseline correction with bc command.
• Optional: To reduce the requirement of 1st order phase correction, adjust certain dealys in the pusle sequence and re-collect data (note this may not work well for very large bandwidth):
• First, use aph or manually phase all peaks up across spectra.
• Type calfa. Recollect and process data. lp required should be reduced now.

Step 7: Recable and tune probe back for H1

• Replace F19 filter with H1 filter
• Type tn='H1' su and tune channel 1 back for H1 detection

How to scan through the entire F19 shift range in an arrayed experiment

• Find out tof value of 0ppm for F19 according to procedure in Step 2, or set tof to a guessed center value
• Type sw=500000 (or maximum allowed by software)
• Under a F19 detection experiment, type array
• Enter parameter to array: tof
• Enter number of steps: 25
• Enter starting value: -250000 (subtract 0.5*tof)
• Enter stepsize: 20000
• Type nt=1 gain=36 go to collect data.
• Click Process->Text Ouput to view array index and values
• Type wft dssh dssl to display all spectra while data being collected. Type ds(#) f full to display spectrum number #. Type vs=vs*0.5 to shrink peak height by half, dssh dssl to redisplay spectra.
• Once the signal region is located, display the best spectrum enclosing all peaks. Use box cursor to enclose region with ~10% baseline region on each side. Type movesw to set spectral window to box cursor region.
• Re-collect data with nt=4 (or multiple of 4). Readjust window size if needed.

Example

1: spectrum with sw=500000 centered at -201 ppm. Note the baseline rolling and phase abnormality due to the large bandwidth and pulse offset effect.
2: spectrum with sw=500000 centered at 0 ppm. Compare with 1 and note the F19 peak at -211ppm did not get excited and is missing from the spectrum due to the large offset from transmitter center. Also note the changes in relative intensity of the peaks from 1.
3: Narrow down spectral window with movesw to enclose all peaks found in 1 and recollect data (nt=4). Note baseline rolling due to large 1st order phase correction.
4: Apply baseline correction with bc command
5:Following 4, reset integrals and scale integral of peak at -110ppm to 2. The integral for each multiplet from left to right, is: 2.7, 0.92, 0.90, 2.0, and 0.79.
6: Window size is further cut down to only include the four peaks on the left with transmitter centered at -92.5ppm. The integrals for the multiplets from left to right, are: 2.97, 1.02, 1.02, 2.0. These values are much more accurate than in 5.

H. Zhou updated Dec 2010