Hi everybody,
I have received a lot of very useful information of the peak width for 90% H2O samples from both sides of the Atlantic!!! Thanks a lot to Chris, Rajan, Robert, Bob, David, Charlie, Clemens, Kristian, Fred, Simon, Robert, Kirk, Toshi, Greg, Debbie, George, Ben and Rainer. I will summerize the responses here.
Regards,
Heike
Original Query:
I was wondering, if anybody could tell me what a "normal" (well shimmed) peak width for the water peak in a 90% H2O+ 10% D2O sample should be? We are using a Bruker DRX 500. Any input would be very much appreciated.
Responses:
On my Varian Inova (ca 1998 console, shims and probe) I use a 2mM sample of sucrose in 90% H2O and 10% D2O. Following a one-shot transient (no dummy or steady state pulses) with a 90 degree flip angle and 20 seconds of presaturation with a 50 Hz CW saturation field on water resonance I get a 27 Hz residual linewidth at the height of the anomeric proton signal and a signal to noise ratio of 130:1 or better for this proton (0.25 second acquisition time with a COS-squared smoothing). This is non-spinning of course.
If the same experiment is repeated but you allow for a 2 second acquisition and a COS-squared smoothing function then I find that the anomeric proton is J-split into two signals with about 80% baseline resolution (non-spinning).
There are many other ways to do this - you could acquire one round of phase cycled transients (four ?) with steady state pulses and shorter pre-sat delays (2 seconds ?) and a short acquisition time (0.25s). This is more real world, of course, but then you need some standard to compare your instrument to. My experience has been that this is poorly documented but perhaps you will find otherwise.
The gain is relatively high, DSP is used.
-------------------------
On our 500, the water linewidth is around 30 Hz. That's true even if the
natural linewidth is around 1Hz. The line is artificially broadened by
radiation damping. If you're using a cryoprobe it could be as big as
100Hz.
-------------------------
A 1H spectrum of the H2O signal will show broadening resulting from radiation damping when the probe is tuned tuned. As a matter of fact, you will see these effects even when using the decoupler coil of a 5mm broadband probe for 1H detection on a 300 MHz NMR. I typically see a full width at half-maximum linewidth of the water peak of about 35 Hz under these conditions with an inverse probe at 500 Mhz.. Detuning the probe results in narrower lines.
I have attached two papers on the subject of radiation damping which hopefully will shed some light on what is happening.
-------------------------
Radiation damping from the strong water signal will make it appear many times wider than it really is at half height, in the absence of water suppression. The important point is when you suppress the water, how much of the spectrum is not useful because the residual water is covering it. That is mainly a matter of shimming. I would think you are doing OK if the base of the water peak is 100-150 hertz wide with optimized presat. In practice, I use excitation sculpting or W5 for everything on my DRX-500; the region of no signal is probably a little wider than that. The correlations still show up in 2D spectra.
-------------------------
It will be 5-7Hz half bandwidth
-------------------------
I am assuming that you are not talking about the line width after any kind of water suppression. On a DRX500 system like yours, the plain water peak width at 50% is about 20 Hz. Since this is a radiation damping related figure, a smaller volume of water will give you smaller line widths, for instance, for a sample of .25 ml volume in a shigemi tube.
-------------------------
Someone out there might know the actual linewidth, although I would guess that it will be dependent on the concentration (i.e., is it really 90% ?) as well as the actual tube type (the i.d. of the tube). Any sample this concentrated (actually, on a 500, anything > ~ 1%) is relaxed by a phenomena called radiation dampening, particularly important for protons (although other nuclei can undergo the same type of relaxation, esp. 31P and 19F). I would guess the linewidth would be 7 Hz or more, full-width at half-height. Typically, you'd want to presat or watergate or other type of solvent suppression. Other signals in the sample should shim to much narrower linewidth, depending on the solute. Peptides and proteins to 1-2 Hz, small organics to < 1Hz, low concentration (1% or <) to 0.3Hz or better with a very good shim.
-------------------------
the linewidth of the water signal of a 90% H2O sample is determined by the radiation damping. I have no exact numbers at hand but on our systems (700MHz and 800 MHz) equipped with a cryoprobe they are larger than 150 Hz. For room temperature probes the line width are also larger than 30-50 Hz. Therefore the H2O linewidth in a pulse-aquire 1D NMR experiment is not a suitable parameter for a well shimmed system. The Bruker standard test is the sample 2mM sucrose in 90% H2O and running a presaturation 1D experiment (zgpr). From the lineshape of the signal of the anomeric proton together with the hump of the residual water signal the shim quality can be deduced. For samples without a sharp reference signal (e.g. protein samples) I always take the data from the gradient shimming procedure as a measure for good shimming.
-------------------------
I encourage our users to check the quality of their shimming by collecting a 1d presaturation experiment (zgpr) with 1.5s d1 and 100Hz B1 field for the presat power (pl9). We do biologicals here and with a 1mM protein sample I would expect the residual water line to be less than 80Hz at the height of the resonances (alpha protons etc) near the water. If it's over 100-120Hz the shimming is bad and water suppression will be poor no matter what method you use to 'remove' the water signal. The best I've shimmed water to (assessing using the above method) on our DRX-500 is 60Hz. I use presaturation as radiation damping then doesn't dominate the line shape.
-------------------------
It really depends on your probe and other factors--what's in your sample. Because of the radiation damping, the H2O peak is going to be broad anyway. I have seen H2O peak widths ranging form 15-30Hz in 500MHz.
-------------------------
Our Varian standards are the opposite concentration 20% H2O 80% D20 and they used to be marked 2Hz samples. I believe even though we really don't have an official shimming spec' for this sample it was always used first for lineshape and customers would get hung up on the 2Hz nomenclature demanding more time from the installers before they moved on to CDCl3/Acetone. Now the same concentration sample is marked 4Hz.
Anyway I am still able to easily meet the 2Hz ½ height width on most instruments spinning. I believe good numbers are 2, 50, 200 Hz _at_ 50, .55, .11 % height. The last number is most important and is improved iteratively mainly through a combination of high order spins and non-spins.
-------------------------
the width of the water line in a 90% H2O sample has almost nothing to do with shimming but is completely dominated by radiation damping.
The T1 of pure water is in the order of several seconds, this should result in a very sharp line.
Normal line width for water willbe anywhere between 20 - over 100 Hz depending on the sensitivity of the probe used to observe the water line.
A multinuclear or BBO probe would give the sharpest line, an inverse probe a somewhat broader one and a cryoprobe the broadest.
The "true' line can be obtained by massively detuning the probe and using a small flip angle.
-------------------------
If you are observing 90% H2O with a 90 degree pulse (using a well tuned probe), you may worry about the radiation damping. The line could be broad but shim is good.
-------------------------
This depends completely on the type of probe and how many shims your shim supply has. In general, for an indirect triple resonance type probe ( this would also hold true for any indirect proton observe probe), I would use the 2mM sucrose standard sample ONLY to assess the water suppression. I would think you should get something like </= 50Hz at ½ height and </= 150Hz at the height of the anomeric proton peak (doublet around 5.3ppm). This is a standard test, and you should have specs in your probe documentation, but these should be quite close. Often times, you can do better than that. The critical parameters are:
Use only a 90 degree pulse
It is best to use the standard parameter set which you should be able to copy from the user "accept" or found using the atp interface, and then adjust final presat power for a 50Hz presat field, rg as necessary, and pw to be 90 degrees.
Make sure the o1 is properly set to minimize the area of the fid in gs mode.
I would avoid the temptation to use any other sample to check your water suppression as this one is carefully designed to assess water suppression. It will give you information on resolution (expressed as % splitting from baseline for the anomeric proton - something like 15-20% is good), and will also check for temperature gradients because if you have a temperature gradient you will be able to meet water suppression specs but the anomeric proton will be broad, or vice versa. If everything is working, and the shims are good, you should get good resolution on the anomeric peak AND meet the water linewidth specs I gave above.
You should also find a good tutorial on how to run this test in NMRGuide, which is provided with your instrument, and can also be run from the Bruker website. This is indispensable if you are not familiar with this test.
-------------------------
Unless you severely de-tune the probe, the linewidth of 90% H2O is governed by radiation damping, not shim (unless your shim is REALLY bad). Typical linewidth at 500 MHz would be 30 Hz.
-------------------------
Your question is difficult to answer.
The width of the water peak is very dependant on radiation dampening effects. The degree of radiation damping will depend on the Q of the
probe: eg. an inverse geometry probe with a very high-Q will have a broader line while a BBO-probe or a probe with a poorly tuned proton
channel (therefore having a lower Q) will show a narrower line.
I remember seeing an excellent slide at a Bruker Pre-ENC workshop where they showed how the water peaks got broader as they went higher in
field. The final slide showed the water peak in a cryoprobe at 800MHz. It was HUGE!
So, I guess to answer your question regarding what is "normal", you will need to ask people who have similar systems about what they observe.
Question: how sharp are the solute peaks in your 90%H2O/D2O sample?
-------------------------
The water linewidth in this kind of sample is much larger than the magnet-determined homogeneity, because of "radiation damping". Very
strong samples at high field, and particularly in cold probes can have linewidths of 10-80 Hz. The stronger the radiation damping, the broader the line.
You can do a presat experiment and look at a solute peak. This would have, typically, the homogeneity-determined linewidth.
-------------------------
For a TXI probe I would expect the linewidth to be ~20-30 Hz.
-------------------------
Funny enough that varies between as little as 10-20Hz and as much as 100-200Hz. Apart of the (general) line shape, I don't think the
watersignal is a good measure for the evaluation of shim quality. Factors of influence seem to be sample-height, exchange with other
protons in the sample, kind of buffer, temperature, etc...
Received on Wed Oct 03 2007 - 12:23:41 MST