NMR Hardware Outline Magnet Lock Shims Gradient Probe Signal generation and transmitters Receiver and digitizer Variable temperature system Solids hardware Instrumentation: Magnet Often the most impressive (and expensive) part of the NMR system Must be large, stable and homogeneous Superconducting wire typical Made more homogenous with Cryoshims Room temperature shims Increasing magnetic field increases Sensitivity, by power of 3/2 Dispersion, linearly 1
T, G, MHz, Hz and ppm Magnetic field is measured in Tesla (T) or Gauss (G) (1 T = 10 4 G), but in NMR usually just refer to 1 H frequency at the given field magnetic field chemical shift T G MHz Hz ppm 1.0 10,000 42.6 42.6 1 7.05 70,500 300 300 1 23.5 235,000 1000 1000 1 Independent of the field d ppm = (ν ν 0 ) 10 6 ν 0 The advantages of greater magnetic field strength Signal dispersion Sensitivity ol-acetate Oestradiol-acetate 900 MHz 900 MHz 900 MHz 900 1350 1300 1250 1200 1150 1100 Hz Current Data P 1400 1350 1300 1250 1200 1150 1100 Hz NAME txioe EXPNO 1450 1400 1350 1300 1250 1200 1150 1100 Hz Estradiol Current Data Parameters 600 NAME MHz acetate PROCNO txioestradiol Current Data Parameters EXPNO 3 F2 - Acquisition NAME txioestradiol PROCNO 1 Date_ 200 EXPNO 3 600 600 MHz Time 1 PROCNO 1 600 MHz F2 - Acquisition Parameters INSTRUM Date_ 20010126 PROBHD 5 m F2 - Acquisition Parameters Time 16.50 PULPROG Date_ 20010126 INSTRUM spect TD 65 Time 16.50 PROBHD 5 mm TXI 13C Z SOLVENT INSTRUM spect PULPROG zg NS PROBHD 5 mm TXI 13C ZTD 65536 DS PULPROG zg SOLVENT CDCl3 SWH 74 950 900 850 800 750TD 65536 NS 4 FIDRES 0 SOLVENT CDCl3 HzDS 0 AQ 4.40 NS 4 SWH 7440.476 Hz RG 1 DS 0 950 900 850 800 750 FIDRES 0.113533 Hz DW 67 Hz SWH 7440.476 Hz AQ 4.4040694 sec DE 6 950 900 850 800 750 FIDRES 0.113533 Hz Hz RG 180 TE 30 400 AQ MHz 4.4040694 sec DW 67.200 usec D1 3.000 RG 180 400 DE 6.00 usec DW 67.200 usec TE 300.0 K ======== CHA 400 MHz DE 6.00 usec D1 3.00000000 sec NUC1 TE 300.0 K P1 10 400 MHz D1 3.00000000 sec ======== CHANNEL f1 ======== PL1-4 NUC1 1H SFO1 900. ======== CHANNEL f1 ======== P1 10.00 usec NUC1 1H PL1-4.00 db F2 - Processing P1 10.00 usec SFO1 900.0236581 MHz SI 655 PL1-4.00 db SF 900.02 SFO1 900.0236581 MHzF2 - Processing parameters WDW SI 65536 SSB 650 600 550 Hz F2 - Processing parameters SF 900.0200000 MHz LB -0. SI 65536 1.65 1.60 1.55 1.50 1.45 1.40 1.35 1.30 1.25 ppm WDW GM GB 0 SF 900.0200000 MHz SSB 0 PC 4 650 600 550 Hz WDW GM LB -0.30 Hz SSB 0 650 600 550 GB 0.3 Hz LB -0.30 Hz PC 4.00 GB 0.3 PC 4.00 Lock The lock is a system to keep the magnetic field stable It works by monitoring the deuterium NMR signal arising from the solvent on a nearly continuous basis If the deuterium signal shifts, implying that the magnetic field had shifted, then a compensatory field would be adjusted The compensatory field is called z0 on Varian instruments The lock field on instruments 2
What does the lock do? Stabilises the field: Corrects for drift (decay of superconducting magnet) Corrects for transient disturbances Makes the field correct Means unreferenced spectra have meaningful chemical shifts TMS should appear at ~ 0 ppm How does the lock work? The lock monitors the deuterium spectrum of the solvent signal, and keeps the frequency constant by adjusting the magnetic field. If the signal is on resonance, the dispersion mode spectrum has zero intensity at the O1 frequency: If the field changes, the intensity of the spectrum becomes non-zero, and the field is adjusted accordingly. Shimming Purpose: create a homogeneous magnetic field across sample Means: adjust current in shim coils putting current in a shim coil establishes a magnetic field each shim coil has its own geometry (Z, X, Z2, etc.) sum of main magnetic field and effect of shim coils should be a homogeneous magnetic field shim sample c b a bad field good field lineshape 3
Good shimming 1.3 Hz (0.3 Hz line broadening) 29 Si satellites 1 H 1.0 Hz (0.3 Hz line broadening) 1 H 1 H Sample preparation Sample should be Clear (filter if necessary, for example through a ball of cotton at the neck of a glass pipette) Homogeneous 5 cm long Labeled on top with marker only NMR tube should be Clean Dry (use oven up to 50 C, keep tubes standing as upright as possible) Free of chips Wiped with a Kimwipe immediately before insertion in magnet Why? Shimming Better observation of small J-couplings Better S/N Avoiding finger grease in probe The grease contaminates your sample and other people s, too Positioning sample in depth gauge It is better to dilute sample to correct depth than to run with a short sample (because concentrated, short sample will have a hump at the bottom of the peak; the added concentration won t increase peak intensity) But, if you have a short sample anyway 4
Pulsed magnetic field gradients A pulsed magnetic field gradient is a temporary distortion of the magnetic field. The gradient can be turned on and off quickly is linear and reproducible can be scaled in amplitude (+100% to 100%) normal field + - Gradient field -+ sample Normal evolution In a normal NMR magnet all spins experience the same field and evolve at the same rate. normal field signal The sum of all spins is coherent, giving a strong, sharp signal Gradient pulse Applying a gradient pulse means each section of the sample sees a different field; nuclei in each section therefore evolve at a different frequency gradient field no net signal At the end of a typical gradient pulse the spins are completely distributed in phase, giving zero net magnetisation for the whole sample 5
Second gradient pulse Local coherence has not been lost! Apply a gradient of the same strength but opposite polarity - the spins rewind and recreate the observable signal negative gradient signal refocussed Dephasing unwanted signals and rephasing wanted signals is the key to most gradient experiments Gradient shimming Automated shimming routine using gradients to identify regions of inhomogeneity within the magnetic field On instruments, also touch up off-axis shims using the lock level (as you would do manually) Gradients in 2D experiments Apply gradients to both wanted and unwanted magnetisation Allow wanted and unwanted magnetisation to evolve differently Apply gradients to refocus desired magnetisation only 1 H pulses, delays pulses, delays 13 C pulses, delays pulses, delays Grad ratio of gradients = ratio of H / X frequency 6
Inverse 2D with Gradients no gradients 22 mins, t1-noise Inverse (proton-detected) experiments are intrinsically more sensitive than carbon-detected versions But they require good spectrometer stability to suppress the 12 C- 1 H signal Using gradients, unwanted signals can be eliminated at source, removing the suppression problem. This gives much quicker and cleaner experiments gradients 10 mins, clean Instrumentation: Probe The heart of the NMR system Transmits and receives rf between spectrometer console and sample Optimised for a particular nucleus or set of nuclei Determines the maximum diameter NMR tube that can be used Can be used for special applications Flow Solids Extra sensitivity with a cryoprobe Probes and channel configurations The transmit / receive coils of an NMR probe are each part of an rf circuit that can be tuned to transmit or receive to a particular range of frequencies Tuning changes as a function of solvent, especially for the most sensitive nuclei Most recent probes have gradient capabilities Most recent probes have VT capabilities Nucleus Frequency (MHz) in 7.0 T magnet Frequency (MHz) in 9.4 T magnet Frequency (MHz) in 11.7 T magnet 1 H 300.0 400.0 500.0 19 F 282.2 376.3 470.4 31 P 121.4 161.9 202.4 13 C 75.4 100.5 125.7 2 H 46.1 61.5 76.8 15 N 30.4 40.5 50.7 7
Ten minute 13 C comparison: 3 mg strychnine Varian Mercury 300 Avance 500 Ten-minute 13 C solution: HSQC??? Varian Mercury 300 Avance 500 Indirect probes Indirect probe: 1 H observe, X decouple Most sensitive for 1 H (and sometimes for 19 F) Good for water suppression (residual water signal is very narrow) Can detect X ( 13 C, 15 N, 31 P, etc.) directly, but insensitively Can detect X indirectly, or decouple X so run 2D experiments (HSQC, HMQC, HMBC) In biological work, especially, three channel inverse probes are common Tuned to 1 H, 13 C, and 15 N (proteins) Or 1 H, 13 C, and 31 P (RNA) Four channel and even five channel equivalents are also possible: 1 H, 13 C, 15 N, and 19 F, or 1 H, 13 C, 15 N, 31 P, etc. 2 H decoupling sometimes included 8
Observe probes Standard observe probes Optimized for X ( 13 C, 15 N, 31 P) observe Can also observe 1 H (and, often, 19 F), but water suppression may not be ideal Varian 500 MHz (McGill) Has tuning rods to change the X frequency range (good for 15 N, usable but not ideal for 31 P) High frequency ( 1 H) channel can also be used for 19 F Varian 300 MHz (McGill) X channel tuned to 13 C as standard High-frequency channel always tuned to both 1 H and 19 F SmartProbes Broadband probes (X optimized, but can also observe 1 H) X range stretches up to 19 F Reasonable water suppression Automated tuning on both channels Cryoprobes Transmit/receive coil is cooled to reduce thermal noise Helium-cooled Transmit/receive coil cooled to ca. 20 K Obtain an increase in S/N by a factor of 4 Savings of factor of 16 in terms of time Nitrogen-cooled Transmit/receive coil cooled with liquid nitrogen Obtain an increase in S/N by a factor of 2-3 Savings of factor of 4-9 in terms of time Some sensitivity figures McGill Chemistry QANUC Test Varian 300 Varian 400 400 Varian 500 500 Cryoprobe 800 1 H 154 358 476 420 645 8600 13 C without decoupling 103 266 249 325 1550 31 P 30 253 241 246 Concordia Sherbrooke UQAM U de Montréal Test Fourier 300 300 600 (BBFO) 600 (TXI) 400 (AV I) 400 (Prodigy) 700 DCH 1 H 150 201 706 1266 368 1100 3900 13 C without decoupling 111 128 357 230 500 2900 31 P 144 352 214 9
What spectrometer should I use? SNR for peak at 8.1 ppm of same sealed strychnine sample (approximately 13 mm or 3 mg), using default acquisition parameters Varian Mercury 300: 135 16 scans, ca. 5 min. Varian Mercury 400: 308 (435) 8 (16) scans, ca. 3 min. Avance 400: 368 16 scans, ca. 7 min. Varian VNMRS 500: 260 (368) 8 (16) scans, ca. 5 min. Avance 500: 508 16 scans, ca. 7 min. Why tune probes? Varian 500, detuned: 104 (147) 8 (16) scans, ca. 5 min. Varian 500, tuned: 260 (368) 8 (16) scans, ca. 5 min. High-resolution magic-angle spinning appropriate for samples in which magnetic susceptibility anisotropy limits resolution interaction removed by spinning at the magic angle liquid B 0 Motional averaging of anisotropic interactions semi-solid Solid Liquid Solid Solid solid 10
Solid phase synthesis: Tetrapeptide Ala-Ile-Gly- Met bound to a Merrifield resin swollen in DMF Spin-half nuclei in presence of MAS static + H 3 N O - O 350 300 250 200 150 100 50 0-50 -100-150 -200 C O MAS, 5 khz θ m 200 150 140 100 5040 0 180 160 120 80 60 20 ppm 13 C / ppm over all crystallite orientations and spinning at the magic angle yields Generating a pulse Signal Generating Unit (SGU) creates shape of pulse Amplifier amplifies it, maintaining rf profile and phase 11
Hard pulses Shaped pulses Selective Broadband (usually adiabatic) Preamplifiers Amplify the NMR signal shortly after it is detected by the probe, so that the signal will be proportionately larger than thermal noise picked up in the rest of the rf circuitry Contain the tuning circuitry 12
Receiver Detects signal excited from the z direction into the xy plane Needs to work in concert with a signal shifted by 90 in order to know which side of the centreband the signal is from But, if the two channels are uneven, will get mirror image peaks about the centre So, use phase cycling Phase cycling The appearance of the line depends on the phase of the initial pulse (which determines where the signal ends up) and the phase of the receiver By making the receiver follow the phase and use all four directions in four sequential co-added experiments, any imbalance between the two channels will be eliminated, removing so-called quadrature artifacts Modern receivers Another solution to the problem of knowing where the signal is supposed to be is accomplished by: Using a very wide receiver bandwidth Centering that bandwidth well to the side of the expected signals Using a digital filter to reduce the bandwidth to the range needed This requires a receiver with a very high sampling rate (6 MHz) 13
Receiving signal: Digitisation Data is stored digitally but generated as analog data Analog-to-Digital Converter (ADC) converts data Small and large measurements must fit within ADC bandwidth receiver gain setting determines whether it fits too large a signal: ADC overflow (artefacts in spectrum) too small a signal: poor digitisation of noise (poor signal-to-noise ratio) Too high receiver gain / too much signal Digitisation in action SW Hz FT DW =1 / (2 SW) AQ = TD DW 14
Modern receiver and digital filter 15