Not a very good idea to my opinion. You are going to lose homogenity causing some extent of line-broadening. Why don't use molecular sieves beforehand and pipet the clear solution into the NMR tube. 

I would also suggest to filter either your deuterated solvent or compound in solution trough a pipette with a cotton plug and some anhydrous sodium sulfate or anhydrous magnesium sulfate. Also storing your deuterated solvents in inert atmosphere and molecular sieves may help.

molecular sieves might help if you put it before measuring process  in H-NMR tube.then you can draw your sample for measuring .also the position of H2O peak in NMR are known so you can easily identified .

Korupalli,

No you cannot add molecular sieves in your NMR tube but as colleagues mentioned dry your solvent over molecular sieves; dissolve your sample under inert atmosphere and transfer your sample using pipette as mention by Ariza to obtain clear solution ready for NMR run.

Good luck

You cannot add directly molecular sives to dry the solevnt because it needs clear solution & no suspended particles.Tube is spinning if  it does not spin properly the broadening  of the peaks occurs. Another problem is that the breaking of the tube may take place.

No, You can not add molecular sieves directly into the NMR tube this is a wrong practice. It is always advisable to dry your sample 1st and this can be done either by applying efficient vacuum at 40-60 degree in case your sample is not heat sensitive or you can pass it through a small column (column of silica/ anhyd. sodium sulfate/ magnesium sufate etc.) . In case water peak is due to residual water present in the deutrated solvent only then you can use molecular sieves to remove the water of solvent prior to dissolving your sample in it.  

Molecular Sieves take a while to sequester the water. Depending on your relative humidity, the solvent is hygroscopic and will pull water from the atmosphere. Obviously this is defeating the purpose of the sieves. As pointed out much earlier, you can use dry box techniques and predry solvents, pump everything for sample prep into a glove bag or box and prep you sample. The sample will need to be sealed with at least a cap and parafilm, if not a valved joint. By introducing sieves into the NMR tube, you can mess up the homogeneity and get broader lines. There are times when water is a big problem, but you can shift a water peak by warming or cooling the sample or adding D2O. Otherwise, water is a good check for calibration. Embrace it!

I have had good results in the past in 60 MHz instruments by adding a single pellet of Molecular Sieve 4A to the tube remove water and acid from deteriorated CDCl3, letting it work and then shaking it up to the top of the tube. Capillary action would then hold it away from the sensitive zone and allow a normal spectrum. (This also works very well with damp DMSO-d6.)  Desperate situations often call for desperate methods. I have not tried this on a higher-field instrument - but if you do, consult the manager first!

"It's all very well in practice, but in theory it just doesn't work."

You can not  add directly  molecular sieves to the NMR tube to remove water in the sample. This is because the NMR tube will not spin properly & bulging may occur, broadening of the spectrum takes place (Unresolved spectrum).

In my experience one must use molecular sieves in very small rod form and ensure that the pellet is shaken up into the higher part of the tube and held onto the wall by capillary action, preferably  in the spinner where the unbalance effect will be minimal. This is somewhat desperate technique and does not always work.

I must admit, I add a single 4A bead often and it settles to the bottom, out of the collection window and remains symmetrical during spinning- I have not witnessed any detriment to the spectra quality. This will only help with a small amount of residual water however

Perhaps a better method is to set d-DMSO in HIVAC desiccator for 5-10min prior to use then prepare sample. you can even pump on it once prepared as well. 

My old technique of shaking the MS 4A pellet to the top of the tube worked well because of the relatively high surface tension of DMSO which would hold it in place. I have not tested the method on a high-field instrument and am pleasantly surprised that the pellet sinks happily and does not degrade the spectrum. Perhaps if it has an effectively lo.wer density than the sample it may "float" to the xis of spin.