400 MHz spectrometer user manual

Transcription

1 400 MHz spectrometer user manual january 2017 Sandrine Denis-Quanquin

2 1. THE NMR SPECTROMETER MANUAL MODE / AUTOMATION SAMPLE CHANGER MANUAL MODE AUTOMATION PRELIMINARY SETTINGS LOCK PROBE TUNING SHIM GAIN ACQUISITION SOME TIPS SOME ROUTINE EXPERIMENTS WHICH 13 C 1D TO CHOOSE? HOW TO SUPPRESS A STRONG SOLVENT SIGNAL? HOW TO CALIBRATE A 13 C, 31 P OR OTHER X SPECTRUM? «STANDARD» 2D EXPERIMENTS SOME OTHER 2D PROCESSING OF 1D DATA SOME INTERESTING PROCESSING PARAMETERS PHASE CORRECTION BASELINE CORRECTION SPECTRUM CALIBRATION PROCESSING OF 2D DATA CONTOUR LEVELS PHASE CORRECTION D PROJECTIONS TROUBLESHOOTING sur 21

3 1. THE NMR SPECTROMETER The spectrometer is equipped with a nitrogen-cooled cryoprobe that is more sensitive that room-temperature 500MHz probes. There is a factor 2 gain in sensitivity for both proton and X nuclei compared to our old equipement. Rq: 19 F experiments can't be performed on the 400MHz. The 400MHz is equipped with a temperature regulation unit (from -40 to +80 C). For temperatures lower than 0 C the air inlet must be replaced by the liquid nitrogen evaporator. A ceramic spinner and even a pyrex tube must be used for experiments with large variations of temperature. Standard parameters sets are copied in the «MANIPS MODELES» directory and may be read with the command > rpar. You are strongly advised to start from these parameter sets and then you may optimize them for your sample. Each Topspin command may either be clicked or entered in the pink command line or found in the upper menus/submenus. In this manual I will show you the clickable buttons and commands to enter (they will always be preceded by > ) 3 3 sur 21

4 400 MHz spectrometer user manual MANUAL MODE / AUTOMATION 2.1 Sample changer ATTENTION check that the sample changer light is GREEN before adding/taking a tube from the carousel. WHITE light: automatic operation running, DO NOT TOUCH BLUE light: carousel not in place RED: technical problem NEVER LEAVE CAROUSEL!!! AN EMPTY SPINNER IN THE 2.2 Manual mode Check that no automation run is going on. If the automation window is opened, check that no experiment is running and close the automation window (see next ). If an experiment is running and IF IT IS YOUR SCHEDULED TIME, HALT the experiment and close the automation window. The command > sx 3 ejects the tube in case there is one in the magnet AND inserts the tube at position 3... The carousel has a capacity of 16 samples, you can put a tube at any available position. ATTENTION if a tube is already in the magnet it will be ejected back to the position IN FRONT of the nmr access!!! The command > sx ej position. ejects the tube from the magnet and put it back in its 2.3 Automation Click on the «AUTOMATION» button to start the software IconNMR then click on «Automation». Select your name in the pop up window then OK. In the IconNMR window you may schedule and set up the experiments. 4 4 sur 21

5 First column is for the position number on the carousel. Double click for modification. Then fill in the sample name (Name), the solvent and select the experiments to run. The experiments named «totale» consist in a 1D proton, a COSY, a decoupled carbon, an HSQC and an HMBC. there is a variant with a udeft instead of the decoupled carbon. When you select «totale» the different experiments are automatically created with following experiment numbers. Some parameters may me optimized by clicking on the blue button (Par column).you may edit the title (Title/Orig). Data will be stored in the directory of the USER selected when starting IconNMR. If several people want to scedule experiments they must change the USER (Change User). Once the experiments are set up for the samples on the carousel click on Submit then on Start (green «cogwheel» on the upper left corner of the window). NB: you may add samples once the run is started. If you want to modify a submitted sample, first click cancel then edit. The last sample is ejected at the end of the run. To stop IconNMR at the end of the run you need to be logged as the USER that started the run. Do not save the set up. Close all IconNMR windows. 5 5 sur 21

6 3. PRELIMINARY SETTINGS Whatever experiment you plan, it is strongly advised to always start with a proton spectrum. A quick proton spectrum is useful to check that lock is ok, shims are good, sample concentration is ok for the planned experiments... First you need to create a new dataset: or >new Change NAME and make sure that DIR is your directory. EXPNO is the number of the experiment. For example with the sample «ananas» you may create 1 - a proton dataset, 2 - a COSY, 3 - an HSQC... NB: the option «use current parameters» creates an experiment identical as the one you start from. To create a different experiment read the parameters afterwards with >rpar SDQ* Once your sample is in the magnet and your data set is ready you need to optimize a few parameters before running your experiment 3.1 Lock Click on Lock or >lock and select your solvent once for each sample The aim of the lock is to guarantee a stable magnetic field along the experiment. The lock system is a spectrometer in the spectrometer, dedicated to the observation of deuterium. The system compares the solvent deuterium signal frequency with a theoretical value and corrects the magnetic field strength accordingly. This correction is repeated at a high frequency as long as the system is locked, thus compensating for the magnetic field fluctuations. 6 6 sur 21

7 Furthermore the lock system calibrates the spectrum with the solvent proton residual signal chemical shift as a reference (NMR solvents are > 99,5% deuterated). 3.2 Probe tuning Click on Tune or >atma for each nucleus! Each coil in the probe is a circuit «tuned» to the resonance frequency of the observed nucleus ( 1 H or X). Tuning the probe may be compared to tuning a radio receiver to a FM frequency. A poor tuning leads to a sensitivity loss. Matching consists in adjusting the probe impedance (resistance) until it matches that of the receiver circuit in the spectrometer. It should guarantee a maximum transmission of the signal. As electrical properties differ between solvents the probe must be tuned for each sample. The X coil needs to be tuned just as much as the proton coil, especially with the BBO. The previous user might have run a 31 P spectrum and tuned the X coil for a maximal sensitivity at 202,4 MHz. If you forget to tune the probe and intent to run a 13 C experiment - frequency 125,7MHz - you will barely see any signals... The tuning is optimal when there are only green diodes on the grey box near the magnet or when the minimum of the curve on the computer screen is on the red line i.e. tuned at the resonance frequency of the observed nucleus. 3.3 Shim Click on Shim or > topshim to perform a routine optimization of the magnetic field homogeneity (topshim as described below). High resolution NMR experiments require a uniform magnetic field over the whole of the sample volume that sits within the detecting coil. B 0 being sensitive to many factors it must be adjusted to obtain signals as narrow and symmetric as possible. To this end «shims» coils are used. They carry electrical currents thus generating small local magnetic fields that are tuned until they compensate B 0 inhomogeneities. 7 7 sur 21

8 An automatic program is used to shim: Topshim tunes shim coils Z to Z5 routinely. By default topshim is configured as follows: dimension 1D, solvent s default, no Z6, TUNE before and after off. NB: to compensate for transverse field inhomogeneities spinning of the tube may be used ONLY FOR 1D SPECTRA. ERROR MESSAGES Topshim uses a complex algorithm needing a very good starting point to converge. If the field homogeneity is really poor the algorithm won t find a valid result and an error message such as Echo time must reduced or FieldMap - signal-to-noise is too low may be displayed on the screen. -> You can load another set of shims (>rsh and choose a recent file, the solvent is not important) and start Topshim again. You can also tune before on Z-X-Y or even Z-X-Y-XZ-YZ-Z to optimize the transverse shims before starting the optimization procedure NB: scratched or dirty tube, poorly soluble sample, too concentrated sample, really small or really large volume, paramagnetic impurities... may deteriorate the field homogeneity. Shim tuning can t compensate for everything, so be careful with your sample! 3.4 Gain The NMR signal is weak and must be amplified by the receiver. The receiver gain - RG - characterizes the signal amplification: a weak RG shows the sample is concentrated - the signal doesn t need to be strongly amplified. Click on Gain or > rga for an automatic setting of the gain (the gain is the amplification factor of the recorded signal, it depends of the sample concentration). Remember to run rga for 31 P spectra. For 13 C experiments rg is always maximal so no need for rga. For protons 2D (COSY, TOCSY, NOESY, ROESY) you may copy the rg value determined for the preceding 1 H spectrum. 8 8 sur 21

9 3.5 Acquisition To start the acquisition after all the adjustements are done: > zg Be careful zg = zero + go, this command erases any possible previous data before starting the acquisition. If you wish to keep the previous data - e.g. 16 scans are already acquired and you want to add some more because of a poor S/N - you may enter >go. If you have prepared a set of experiments for the same sample (be careful as they must have following EXPNOs), from the 1st one enter > multizg number_of_experiments. Example: from experiment 1 > multizg 4 Example 2: from experiment 16 > multizg Some tips To optimize your experiment you may change some parameters: the spectral window SW and its middle O1P the number of scans NS the relaxation delay D1 the number of points TD Most of these parameters may be found in the AcquPars window. Click on A to display all the acquisition parameters, and click to return to the first default display. If you don t remember where to find a parameter, enter its name in the command line: for O1P enter > o1p, a dialog box with the parameter to change will open. Reminder: TD = 2 x AQ x SWH (where AQ: acquisition time and SWH: SW in Hz) and TR (repetition time) = D1 + AQ 9 9 sur 21

10 AQ must be long enough to record the full FID and TR must be long enough for relaxation of the observe nucleus - for a quantitative spectrum. This means you should increase D1 if you observe inconsistencies in the integration of your proton spectrum. The number of scans - NS - and the relaxation delay - D1 - are decisive for the experiment time click on to check the total time. Raw data transform For 1D Fourier Transform > ft. Routinely > efp combines ft + exponential multiplication + phase according to the parameters in ProcPArs (see about process). During acquisition you must transfer the firsts scans if you want to treat them: > tr then >efp. With 2D > xfb sur 21

11 4. SOME ROUTINE EXPERIMENTS 4.1 Which 13 C 1D to choose? «standard» spectrum: 1 H decoupled carbon (zgpg) spectrum with no decoupling (zg) to observe J CH couplings. Sensitivity may be an issue and the increased number of peaks makes the spectrum more complicated to assign. udeft for molecules with quaternary carbons giving weak signals or even no signals with the standard experiment jmod: edited spectrum where Cq and CH 2 signals are positive whereas CH and CH 3 signals are positive. dept135: edited spectrum like jmod but more sensitive thanks to a polarization transfer from 1 H to 13 C. As a consequence no signals are observed for C q. dept135 jmod decoupled 13 C udeft, 4h total decoupled 13 C, d1=20s, 10h total First acquire a decoupled spectrum before thinking about a non decoupled or a udeft. For synthesis follow up that doesn t need quaternary carbons observation run a dept or an HSQC (2D even more sensitive than dept) sur 21

12 4.2 How to suppress a strong solvent signal? In some cases - often with biological samples in solution in D 2 O or even H 2 O/D 2 O - the solvent residual proton signal is really intense compared to the signals from the sample. The signal may be suppressed by applying a low power pulse at the solvent resonance frequency (this is called presaturationduring the relaxation delay d1. First acquire a «standard» proton sectrum and note down the frequency of the signal to suppress (in Hz, not ppm). create a new proton dataset and change the PULPROG from zg30 to zgpr. Don t forget to run > getprosol to adjust the presaturation power level. > o1 : in the pop up window enter the resonance frequency of the signal to suppress (in Hz). ATTENTION O1 is also the middle of the spectrum! Check that the spectral window is large enough, if it is not change SW. If you re not satisfied with the signal suppression change O1 of a few Hz. 4.3 How to calibrate a 13 C, 31 P or other X spectrum? You need to acquire a proton spectrum right before your X spectrum and calibrate it carefully. Write down the precise value of SF (spectrum frequency) that you will find in the ProcPars of the proton spectrum. Then multiply this value by the ratio - taken from the table below - for the wanted nucleus (Ξ= for referencing a 31 P sur 21

13 spectrum to H 3 PO 4 for example). The result is to be used as SF value for your X spectrum. Nucleus Frequency ratio Ξ Ref molecule 13 C TMS 19 F CCl3F 29 Si TMS P H 3 PO 4 For other nuclei see ref Pure Appl. Chem., Vol. 73, No. 11, pp , 4.4 «Standard» 2D experiments Parameters sets fitted for observation of organic molecules are saved in the «MANIPS MODELES» directory and may be read with > rpar SDQ*. The parameters to optimize and the adjustements not to forget are summarized in the Title window. Remember to acquire a proton spectrum before any 2D experiment to check shims, gain, spectral window... COSY : > rpar SDQ_COSY COSY is a 2D experiment that shows correlations between protons that share a scalar coupling ( 2 J HH or 3 J HH ). F2 column is for the direct dimension (X axis), F1 is for the indirect dimension (Y axis). To prevent 2D experiments from being time consuming the number of points in the indirect dimension is significantly lower than in the direct dimension. Try to change TD (F2 and F1), AQ and SW and see what happens. For each change check the duration of the experiment. HSQC : > rpar SDQ_HSQC HSQC is a heteronuclear experiment that shows 1 J H-C scalar interactions. The indirect dimension is now 13 C so be careful of spectral window. The default value is 160 ppm centered at 80 ppm which is enough for most organic compounds. According to the observed molecule it might be relevant to change these values (e.g. no signals are expected in the aromatic region). Reminder: no signals are observed for quaternary carbons in HSQC! sur 21

14 4.5 Some other 2D HMBC > rpar SDQ_HMBC HMBC shows 2 J H-C and 3 J H-C (even 4 J H-C ) scalar heteronuclear correlations. The value for the observed long distance coupling may be changed - it is the CNST13 parameter. Furthermore CNST2 is the value of the residual 1 J C-H couplings that are filtered from the spectrum. HMBC is a low sensitivity experiment so remember to use twice the number of scans needed for HSQC. It is really helpful with HMBC interpretation to superimpose the two spectra (HSQC + HMBC). Be careful as 3 J CH USUALLY lead to MORE intense signals than 2 J CH. However the intensities happen to be similar in some cases. NOESY / ROESY NOESY and ROESY experiments rely on the existence of a dipolar intercation between 2 nuclei at close distance (<5Å) that leads to an effect called NOE. NOE depends on the distance between the nuclei as well as on correlation time of the molecule τ c. The sign of the NOE depends on τ c. For middle size molecules (600 < MW<1500 g.mol -1 depending on solvent viscosity) NOESY might be inconclusive and it is advised to acquire also a ROESY. NOESY and ROESY spectra show signals due to a NOE and exchange signals. The signs of both kinds of cross peaks are detailed in the following table for a spectrum with a negative diagonal: ω 0 τ c < 1 ω 0 τ c 1 ω 0 τ c >1 NOESY cross peak ROESY cross peak exchange signals NB: other experiments are available on this spectrometer. 2D HSQC_10ppm and HMBC_10ppm for example give spectra with a 3Hz resolution in the carbon dimension (instead of 40Hz at best with a standard HSCQ) sur 21

15 5. PROCESSING OF 1D DATA Basic treatment after fourier transform consists in phase correction and baseline correction. > apk and > abs n perform automatic phase and baseline corrections that are often satisfying. à buttons for vertical scaling (intensity) à buttons for horizontal scaling (zooming) 5.1 Some interesting processing parameters Look in the ProcPars window. SI = TD or 2*TD to have enough points for well defined signals SR is the offset in Hz due to manual calibration of the spectrum (see 6.4). The default «window function» is an exponential (EM) with a Line Broadening factor (LB) of 0.3 Hz. The multiplication of FID with an exponential function increases S/B ratio but it also broadens the signals - which might result in a loss of coupling information. The optimal LB value is measured at half height of the peaks in a spectrum obtained with «ft». With ef LB is always POSITIVE. These parameters are taken into account when applying the fourier transform > ef (= ft + em) or > efp (to use also the phase correction parameters PH0 and PH1) sur 21

16 Multiplication of FID with a gaussian function may also be helpful. WDW, LB and GB have to be modified in ProcPars. WDW = GM, then LB and GB are adjusted according to the expected result. The example below illustrates the impact of LB and GB on the separation of overlapping peaks. With gf, LB is always NEGATIVE and 0 < GB < 1. Be careful as high values of GB and LB may lead to the loss of weak signals. n ft n ef, LB=0,3 n ef, LB=1 n ft n gf, GB=0,1, LB=-0,3 n gf, GB=0,1, LB=-1 n ft n gf, GB=0,5, LB=-0,3 n gf, GB=0,5, LB=-1 n ft n gf, GB=1, LB=-0,3 ef, n LB=2 gf, GB=1, LB=-1 ef, LB=1 ef, LB=0,3 ft sur 21

17 5.2 Phase correction Pivot point If you re not satisfied with the automatic phase correction (apk) you may correct the phase manually left-click-hold on 0 and move the mouse up or down for zero order correction (phase correction of the biggest peak where the pivot point is set) left-click-hold on 1 and move the mouse up or down for first order correction (phase correction of the distant peaks) 5.3 Baseline correction If you re not satisfied with the baseline correction obtained with > abs n you may correct it manually. Click on to display the baseline correction window. Zoom out to see all the spectrum in the window. LEFT-CLICK-HOLD on A and roll the mouse up or down to distort the baseline. Do the same with B, C, D and E sur 21

18 5.4 Spectrum calibration Topspin calibrates the chemical shifts using the solvent signal as a reference. This calibration may be unsatisfying - e.g. D2O chemical shift depends on temperature and ph. You may calibrate the spectrum manually. Expand the region of the reference signal (chloroform...) Click on Calib axis to display the calibration window. Then click on the signal and enter the right frequency in ppm in the dialog box. The shift applied to the initial spectrum is stored in the process parameter SR (in Hertz) sur 21

19 6. PROCESSING OF 2D DATA Once your data have been processed with > xfb remember to correct the baseline in the indirect dimension > abs1 only if your experiment needs a manual phase correction (HSQC, NOESY, ROESY). NB: COSY and HMBC routine experiments on this spectrometer are not phase sensitive and thus are processed in magnitude mode: all signals are positive. 6.1 Contour levels For a better resolution of 2D spectra the number of contour levels may be increased. o right-click on the spectrum o select «Edit contour levels» o replace 1.8 by 1.3 and 8 by 20 o click, in that order, on Fill, Apply and OK 6.2 Phase correction For phase correction in the direct dimension click on Adjust phase to display the phase correction window. Right-click and «Add» on 2 distant signals. Then click on (row) and phase the extracted rows as they should be (here CH2 signals are supposed to be negative) sur 21

20 6.3 1D projections You may replace the 1D projections with previously acquiried 1D spectra. Right click on the projection and select External projection. Write the experiment number of the spectrum to be displayed. Copy the SR value from the 1D spectrum used as external projection and paste it in the corresponding SR (F2 or F1) in the ProcPars. If the 2D spectrum still looks a bit off you may calibrate it manually: click and do as for a 1D spectrum sur 21

21 7. TROUBLESHOOTING Bruker hotline number: Tell you are a working in the ENS Chemistry lab in Lyon and explain your problem. û HOW TO REBOOT THE SAMPLE CHANGER? First switch it off (press a few seconds the blue button in front). Then gently remove the carousel. Switch on the blue button WITHOUT the carousel. Wait for the upper light to go from orange to blue. Put the carousel back into place, gently! The carousel should turn around and the light should go from blue to white then green. û HOW TO CHANGE A DATASET LOCATION If you start from an experiment in MANIPS MODELES remember to change the USER!!!!! If you forget and your dataset is saved in the wrong directory use the linux files explorer to move it in the right folder. û TOPSHIM COMMON ERROR MESSAGES If you encounter an error message such as Echo time must reduced or FieldMap - signal-to-noise is too low this means that the shims you are trying to optimize are not good enough for the optimization algorythm. --> You can load another set of shims (>rsh and choose a recent file for the right probe, the solvent is not important) and start Topshim again. You can also tune before on Z-X-Y or even Z-X-Y-XZ-YZ-Z to optimize the transverse shims before starting the optimization procedure. û OTHER SHIM ISSUES >rsh if for shims reading and >wsh is for shims writing. You may write shims for a specific solvent or sample if you need it. Don t overwrite the existing shims files. û HOW TO STOP A PROCESS (MULTIZG FOR EXAMPLE) >kill is used to stop a multizg. It opens a window showing the running processes. Select the multizg line and click Kill. You may need to repeat this once or twice. û ii restart >ii restart allows you to reset the electronic connections in the spectrometer sur 21