Solid-state NMR: why a powder gives broad lines, what magic-angle spinning and cross-polarisation do, and what a CP-MAS spectrum can and cannot count
Share
Solid-state NMR: why a powder gives broad lines, what magic-angle spinning and cross-polarisation do, and what a CP-MAS spectrum can and cannot count
The same nuclei and the same chemical-shift scale as in solution, measured on a powder spun at many thousands of revolutions per second in a rotor inclined at 54.74° to the magnetic field
Dissolving a drug substance destroys its crystal form. A solution spectrum therefore cannot show which form is present or how much of the substance is amorphous. A cured resin does not dissolve at all. Solid-state NMR measures the powder as it is. The spectrum has the same chemical-shift scale as in solution, but the lines are about a thousand times broader. They are narrowed by spinning the sample and decoupling the protons. The peak areas depend on how the spectrum was recorded.
Five terms used in this article. An anisotropic interaction changes with the orientation of the molecule in the magnetic field. Magic-angle spinning (MAS) rotates the sample about an axis inclined at 54.74° to the field. Cross-polarisation (CP) passes the polarisation of the protons to a rare nucleus such as 13C through their dipolar coupling. A spinning sideband is a copy of a signal, displaced from it by a whole multiple of the spinning rate. Polymorphs are different crystal forms of one compound.
At a glance: the observation and its cause
| Observation | Cause | Remedy |
|---|---|---|
| A powder gives a line kilohertz wide where a solution gives a few hertz | Without molecular tumbling, the dipolar couplings and the shift anisotropy are not averaged | Magic-angle spinning and proton decoupling |
| Small peaks appear at equal distances on both sides of a signal | Spinning sidebands, at multiples of the spinning rate | A second spinning rate, faster spinning, or sideband suppression (TOSS) |
| CP-MAS peak areas do not match the composition | Each carbon takes up polarisation from the protons at its own rate | Direct polarisation, multiCP, or calibration with known mixtures |
| One carbon gives two or more sharp peaks | Several independent molecules in the crystal, or a mixture of forms | Spectra of the pure forms for comparison |
| A broad signal lies under the sharp ones | Amorphous material, with lines about ten times broader | Quantification against the crystalline signal |
| An 11B or 27Al line remains broad at any spinning rate | Second-order quadrupolar broadening | A higher field or the two-dimensional MQMAS experiment |
| The 1H spectrum of a rigid solid is a single unresolved band | Every proton is coupled to many others | About 100 kHz spinning, in a rotor that holds about 1 mg |
Why a powder gives broad lines
A nuclear spin is subject to the same interactions in a solid as in a liquid.[1] The difference lies in molecular motion. The chemical shift, the through-space coupling between two nuclei (the dipolar coupling) and, for nuclei with spin greater than ½, the quadrupolar coupling all depend on the orientation of the molecule in the field. In solution the molecule tumbles so rapidly that only the averages are observed: the isotropic shift that a solution spectrum reports, and zero for the dipolar coupling. In a powder each crystallite keeps its orientation and with it its own frequency, so the signal of a single site is spread over a range of frequencies, the powder pattern.[2] The intensity is spread over the same range, and resolution and sensitivity are lost together.
The two scales differ by three orders of magnitude: shifts and scalar couplings need a resolution of a few hertz, whereas the dipolar broadening in a solid is normally several kilohertz.[3]
The interactions of a nuclear spin in a solid
| Interaction | Typical size | In a liquid | In a spinning powder |
|---|---|---|---|
| Chemical shift | isotropic range about 10 ppm for 1H, 200 ppm for 13C | averaged to the isotropic shift | isotropic line plus spinning sidebands |
| Scalar coupling | 0–1 kHz | resolved as multiplets | usually hidden within the line |
| Dipolar coupling | 0–20 kHz | averaged to zero | reduced by spinning; 1H–13C removed by decoupling |
| Quadrupolar coupling | 100 kHz to tens of MHz | relaxation and broad lines | removed to first order only |
Sizes from [1]. For the quadrupolar coupling in solution see Part 7.
Figure 1. The signal of one site in a powder at rest, under slow spinning and under fast spinning. Schematic.
The magic angle
The dipolar coupling between two nuclei contains the factor (3cos2θ − 1), where θ is the angle between the line joining them and the field.[4] At cos θ = 1/√3, that is at 54°44′ or 54.74°, the factor is zero. The internuclear vectors of a powder cannot all be oriented at this angle, but the whole sample can be rotated about an axis at this angle. Over one rotation the factor averages to zero for every vector, whatever its own direction. Andrew, Bradbury and Eades reported in 1958 that fast rotation narrows the spectrum of a solid,[5] and showed a year later that rotation at an angle α to the field reduces the width of the central line by the factor ½(3cos2α − 1), which is zero at 54°44′.[4] The anisotropic part of the chemical shift is averaged in the same way, so each site in a spinning powder gives a line at its isotropic shift.[1] For the routine experiment the European Pharmacopoeia names a rotor of about 4 mm and rotation at 4–30 kHz;[6] the upper part of that range requires a smaller rotor, as the table below shows.
Figure 2. Left: the rotor is inclined at 54.74° to the field B0 and spins about its own axis. Schematic. Right: the factor ½(3cos2θ − 1) against the angle; it is 1 along the field, −½ at 90° and zero at 54.74°. Our calculation.
The anisotropy is not removed completely: besides the central line, the rotating sample gives spinning sidebands at whole multiples of the spinning rate, and these become distant and weak only when the rotation is fast compared with the static line width.[4] In a 13C spectrum, sites with a large shift anisotropy still show such sidebands below about 20 kHz;[1] the carbonyl carbon of glycine at 10 kHz is an example.[7]
A sideband can be mistaken for a signal, but its position can be calculated. At 9.4 T, 13C resonates at 100.63 MHz,[8] so 1 ppm is 100.63 Hz, and at 12 kHz the first sidebands lie 119 ppm on either side of the centre band. By our arithmetic a carbonyl carbon at 170 ppm has sidebands at 51 and 289 ppm, the first of them among the aliphatic carbons. A second spectrum at another spinning rate distinguishes sidebands from signals, because the sidebands shift and the signals do not. The pulse sequence TOSS (total suppression of sidebands) removes them from the spectrum, as in the 13C spectra of a patent on crystal forms of lorlatinib maleate.[9] Sidebands are also informative, since the anisotropy of the shift can be calculated from their intensities.[1, 10]
Rotors, spinning rates and the amount of sample
The powder is packed into a small zirconia cylinder, the rotor,[8] which is spun by a stream of gas. Smaller rotors reach higher spinning rates but hold less sample. The absolute signal falls with rotor size, although the smaller coil detects each microlitre of sample more efficiently.[11] A recent methods primer gives sample quantities of about 500 mg for a 7 mm rotor and about 1 mg for a 0.7 mm rotor.[1]
Rotor diameter and maximum spinning rate
| Diameter (mm) | Maximum rate (kHz) | Sample volume |
|---|---|---|
| 4.0 | 15 | more than 30 µL |
| 3.2 | 24 | more than 30 µL |
| 2.5 | 35 | 10–30 µL |
| 1.9 | 42 | about 10 µL |
| 1.3 | 67 | 1–10 µL |
| 0.7 | 111 | less than 1 µL |
Rotors of one manufacturer, with the volume classes of the same review, which places 4.0 and 3.2 mm together in its largest class; 1.9 mm lies on the boundary between two classes.[11] The 4 mm rotors of one of the studies cited here held 50 µL.[8]
Figure 3. Maximum spinning rate for six rotor diameters of one manufacturer; the circles show the diameters in proportion to each other.[11] Our drawing of published values.
Friction with the gas heats the rotor in proportion to the square of the spinning rate. Lead nitrate, used as a thermometer, indicated a rise of about 40 °C when the rate was increased to 40 kHz,[12] and the heating is not uniform: in a zirconia rotor at 12 kHz, regions of the sample differed by 5 K.[13] Strong proton decoupling causes additional heating in moist samples.[14] For a hydrate, a solvate or a metastable form, the spinning rate and the sample temperature are therefore part of the result.
The angle must be set to within a fraction of a degree. This is done with potassium bromide, because 79Br resonates so close to the 13C frequency that the probe needs only a small retuning. The salt is spun at about 5 kHz and the angle is adjusted until the rotational echoes in the bromine signal, recurrences of the signal at multiples of the rotation period, are as many and as strong as possible; judged by the sidebands of the bromine spectrum, the angle can usually be set to within 0.1°.[7, 15] The same signal serves as a thermometer, because its shift moves by −0.025 ppm per kelvin.[14]
Proton decoupling and cross-polarisation
Spinning narrows a line only when it is fast compared with the interaction that broadens it.[4] Dipolar couplings of up to about 20 kHz[1] are of the same order as the spinning rates of the common rotors, so spinning alone cannot remove the coupling of a carbon to its protons. This coupling is removed instead by strong irradiation of the protons while the carbon signal is acquired.[3] In the examples cited here the decoupling fields are 57 to 86 kHz, applied as phase-modulated sequences such as TPPM and SPINAL-64,[7, 8, 9, 16, 17] and a rigid CH2 group can need as much as 100 kHz.[18]
The remaining problem is sensitivity. 13C is rare and, in a solid, slow to relax. After a pulse the nuclei return to equilibrium with the time constant T1, which in polystyrene was 1.55 s for the protons and more than 60 s for the carbons (estimated).[16] Pines, Gibby and Waugh solved both problems by transferring polarisation from the protons.[3] For a few milliseconds, the contact time, both nuclei are irradiated with radio-frequency fields B1 matched so that γHB1H = γCB1C, where γ is the magnetogyric ratio, the strength of a nucleus’s magnetic moment relative to its spin. Under this condition, named after Hartmann and Hahn,[19] protons and carbons precess about their fields at the same frequency and exchange polarisation through the dipolar coupling. The carbon signal can grow by a factor of up to γH/γC ≈ 4, and that of 15N by up to about 10. The experiment can also be repeated as soon as the protons have recovered, because their T1, and not that of the carbons, determines the repetition rate.[3]
The first experiments were done at 2.28 T, without spinning the sample.[3] Schaefer and Stejskal added magic-angle spinning in 1976,[20] and the combination, CP-MAS, is still the standard 13C experiment for an organic solid. Spinning complicates the matching: the condition splits into bands separated by the spinning rate, and its centre weakens as the rate rises.[21] Sequences that vary the amplitude of one field during the contact widen the range of matching again;[21] three of the studies cited here used a ramp of this kind,[8, 16, 22] a method introduced by Metz, Wu and Smith.[23] The condition is set on adamantane, which also serves as the shift reference.[7]
Figure 4. The CP-MAS experiment. A 90° pulse on the protons is followed by the contact period, in which both nuclei are irradiated, and the carbon signal is then acquired under proton decoupling. Schematic; the durations are not to scale.
Conditions and measuring times of published CP-MAS spectra
| Sample | Field (1H, MHz) | Spinning (kHz) | Contact (ms) | Delay (s) | Scans | Time (h) |
|---|---|---|---|---|---|---|
| Cellulose (Avicel and others)[22] | 200 | 6.5 | not stated | 1.0 | 10 000–20 000 | 2.8–5.6 |
| Indomethacin, crystalline and amorphous[8] | 400 | 12 | 3 | 40.3 | 300 or 600 | 3.4 or 6.7 |
| Mebendazole, forms A and C[8] | 400 | 12 | 3 | 72.8 | 300 | 6.1 |
| Lorlatinib maleate, anhydrous (Form 1)[9] | 400 | 8.0 | 5 | 5 | 5760 | 8.0 |
| Lorlatinib maleate hydrate (Form 2)[9] | 400 | 10.0 | 7 | 10.5 | 5269 | 15.4 |
The last column is our arithmetic, scans × delay, without the acquisition time. The indomethacin and mebendazole spectra were recorded for quantification, hence delays of eight times the proton T1 of the slowest-relaxing form.[8]
Why CP-MAS peak areas are not quantitative
In a solution spectrum recorded with a long enough delay, every nucleus counts equally (Part 9). In a CP-MAS spectrum the carbons do not start with equal polarisation. Each acquires it from the protons during the contact, at a rate set by its dipolar coupling to them, and so by distance and motion. The proton polarisation itself decays during the contact, under the radio-frequency field, with the time constant T1ρ.[8, 24] A carbon bearing hydrogen reaches its maximum early and then declines as the protons decay. A quaternary carbon, or a carbon in a mobile group, builds up slowly and reaches its maximum later. At any single contact time the carbons are therefore at different points of their build-up curves.
Weathered polystyrene, which has been partly oxidised by light, illustrates the time scales. Its CH and CH2 carbons build up with a time constant TIS of about 90 µs, carboxyl carbons need about 1 ms and a quaternary carbon without attached hydrogen about 1.7 ms, while T1ρ of the protons is 1.8–3.9 ms.[16] At a contact time of 2 ms its oxidised carbons were over-represented, and their areas had to be multiplied by 0.6–0.7 to agree with the multiCP experiment described below.[16] For a material of unknown composition, a ratio of two CP-MAS areas is not a ratio of amounts.
Figure 5. Carbon signal against contact time for a carbon that takes up polarisation quickly and one that takes it up slowly. Schematic: the curves are drawn with time constants inside the ranges reported for polystyrene[16] and are not fitted to data. At a short contact time the first carbon is over-represented, at a long one the second.
A second source of bias is the delay between scans. Solid forms of one compound usually differ in proton T1,[25] and the less crystalline a form, the faster it relaxes;[26] a short delay therefore favours the faster-relaxing form. Quantitative work needs a delay of at least five times the longest T1 in the sample.[26] The same difference is used deliberately as a relaxation filter, which suppresses the signal of one form and retains that of the other.[25, 27]
Three ways to record the carbon spectrum of a solid
| Experiment | Source of the signal | Delay set by | Areas |
|---|---|---|---|
| CP-MAS | proton polarisation, one contact | T1 of the protons | need calibration |
| Direct polarisation (single pulse) | the carbons themselves | T1 of the carbons, which can be a minute or more | quantitative, but slow |
| MultiCP | proton polarisation, many short contacts | T1 of the protons | close to quantitative |
After [1, 16, 28].
Direct polarisation was for a long time the only reliable quantitative method for an unlabelled solid. MultiCP replaces the single contact by a series of short ones, between which the carbon polarisation is stored along the field while the protons recover. More than 10 ms of contact is reached in this way without the losses of one long contact, and the measuring time is reduced by a factor of more than 50 compared with direct polarisation.[28] For polystyrene, blocks of 150–250 µs separated by 2–3 s gave areas accurate to about 10 % and a detection limit of roughly 1 % for carbon at natural abundance.[16]
The third route retains ordinary CP-MAS and calibrates it. With spectra of the pure forms, or from a set of mixtures alone, the proportions of mebendazole forms A and C and of crystalline and amorphous indomethacin were recovered with mean errors of about 2–5 %.[8] Silicon behaves in the same way: a single-pulse 29Si spectrum counts every site, whereas a cross-polarisation spectrum favours silicon near protons (Part 10), which is why CP was applied early to the surface of silica gel.[29]
A CP-MAS spectrum identifies a solid form reliably. A statement of amounts needs direct polarisation, multiCP or a calibration, and the report should say which was used.
What the solid-state spectrum shows
Crystal forms and amorphous material
An estimated 80–90 % of drugs are delivered as solids,[8] and the crystal forms of one substance can differ in melting point, chemical reactivity, apparent solubility and dissolution rate.[30] Regulators expect the forms to be characterised. ICH Q6A names X-ray powder diffraction, thermal analysis, microscopy and spectroscopy as examples.[31] The FDA guidance on polymorphism and the European Pharmacopoeia both list solid-state NMR, and the pharmacopoeia states that several techniques must be used together.[30, 32] Two reviews by Harris cover polymorphs, hydrates and solvates in depth.[33, 34]
NMR separates the forms because the shift of a carbon responds to the conformation of the molecule and to its neighbours in the crystal. A different packing moves a signal by 1–2 ppm and a different conformation by about 5 ppm, and a shift in the solid can lie as much as 10 ppm from its value in solution.[26] A drug substance of average size gives 10 to 30 13C peaks, enough to identify a form.[25] The methyl carbon of mebendazole, for example, lies at 51.2 ppm in form A and at 54.2 ppm in form C.[8] If the crystal holds two molecules that are not related by symmetry, each carbon gives two peaks, and three or four such molecules give three or four peaks.[25] In many crystalline drug substances the lines are 0.5 ppm wide or less, and an amorphous substance gives lines about ten times broader:[25] the methylene carbon of indomethacin lies at 28.1 ppm in the crystalline γ form and at 29.4 ppm in the amorphous substance.[8]
Figure 6. Left: the methyl carbon of mebendazole in form A, in form C and in a mixture. Right: the methylene carbon of indomethacin in the crystalline γ form, in the amorphous substance and in a mixture. Peak positions from the source;[8] the line shapes are schematic, with the amorphous line ten times broader than the crystalline one.[25] Our drawing of published values.
A tablet can be measured intact. Signals of drug substances fall mostly at 0–50 and 110–200 ppm and those of common excipients at 60–110 ppm, so the two often do not overlap.[25] Patents make use of the same property: the lorlatinib maleate patent characterises its crystal forms by lists of 13C and 19F shifts and gives the usual variability of a peak position as ± 0.2 ppm.[9]
Published limits of detection or quantification for a minor solid form
| Case | Experiment | Limit |
|---|---|---|
| Typical crystalline drug substance, general estimate[25] | 13C solid-state NMR | 0.1–1 %; about ten times higher for amorphous content, for which an earlier article gives 10–20 %[26] |
| Ranitidine hydrochloride in lactose monohydrate[26] | 13C CP-MAS, 8 hours | about 0.5 %, against about 5 % by X-ray powder diffraction |
| Drug substance in a low-dose formulation[35] | 13C CP-MAS | detected at 0.7 % by weight; the polymorph could not be identified |
| Cimetidine form A in its hydrochloride salt[36] | 1H double-quantum MAS, 850 MHz, 30 kHz, about 10 mg | 1 % w/w |
| Crystalline drug in an amorphous formulation[27] | 19F direct polarisation and CP | 0.1 % w/w |
| The same, with a relaxation filter[27] | 19F, amorphous signal suppressed | 0.04 % w/w |
The general estimate assumes a molar mass of about 400 g/mol, lines of 0.5 ppm or less and a proton T1 of 10 s or less. w/w: by weight.
Fluorine is the most favourable nucleus for this purpose. About a fifth of commercial drugs contain it;[37] the nucleus is 100 % abundant, resonates close to the proton frequency,[1] and is absent from excipients, so the spectrum of a tablet shows the drug substance alone.[25]
Salt or co-crystal
Whether an acid has transferred its proton to a basic nitrogen atom, giving a salt, or is only hydrogen-bonded to it, giving a co-crystal, has regulatory and legal consequences for a drug substance.[38] The 15N shift distinguishes the two cases, as it does in solution, where a pyridine-type nitrogen moves by about 100 ppm on protonation (Part 5). In a study of 17 acid–base complexes, protonation of a heterocyclic nitrogen moved its 15N signal by 47–118 ppm to lower frequency. Hydrogen bonding alone moved it by a few ppm, although one strong hydrogen bond reached 33 ppm. For N9 of theophylline the figures were 48–50 ppm in two salts and 1–5 ppm in five co-crystals.[17]
Where the shift is not conclusive, the N–H distance can be measured through the dipolar coupling. It was 1.17 Å in a salt, 1.62 Å in a co-crystal and 1.25 Å in an intermediate case, on 0.8 mg of unlabelled sample, although each measurement took 3.5 to 5.8 days at 700 MHz and 60–70 kHz spinning.[38]
Insoluble materials
For a material that cannot be dissolved, the solid-state spectrum is the only NMR spectrum available. In cellulose the C4 carbon gives a peak at 89 ppm from ordered chains and one at 84 ppm from disordered chains, and the fraction of the combined area between 80 and 93 ppm that belongs to the 89 ppm peak serves as a crystallinity index. For the microcrystalline cellulose Avicel PH-101 this gave 56.7 %, whereas three ways of evaluating the X-ray diffraction pattern of the same material gave 60.6, 77.7 and 91.7 %.[22] A crystallinity value is therefore meaningful only together with the method by which it was obtained.
What other nuclei distinguish in a solid
| Nucleus | Distinction | Values |
|---|---|---|
| 29Si | number of siloxane bonds at silicon | Q3 near −100 ppm, Q4 near −110 ppm (Part 10) |
| 27Al | coordination number of aluminium | four-coordinate about 40–70 ppm, five-coordinate about 20–40 ppm, six-coordinate near 0 ppm[39, 40] |
| 11B | trigonal or tetrahedral boron | quadrupole coupling constant 2.66–3.29 MHz in boronic acids and esters,[41] about 0.4 MHz at tetrahedral BO4 (Part 7) |
| 15N | protonated or hydrogen-bonded nitrogen | 47–118 ppm to lower frequency on protonation of a heterocyclic nitrogen[17] |
| 19F | crystal form of a fluorinated compound | two peaks for each of two forms of triamcinolone, in a formulation with 5 % drug substance[25] |
27Al shifts are relative to [Al(H2O)6]3+, itself six-coordinate; the ranges are the assignments of a catalyst patent, whose range for six-coordinate aluminium is misprinted, so that we give only the position of the reference. In the zeolite Na-Y the four-coordinate aluminium lies at 56.5 ppm.[40]
Boron and aluminium are quadrupolar, and for them spinning is only partly effective. It removes the quadrupolar broadening to first order and leaves a second-order part, which decreases with increasing field.[1] The position of a 27Al peak therefore changes with the field and is not simply the chemical shift,[40] and trigonal boron, whose coupling constant is several times larger, remains far broader than tetrahedral boron. The usual remedy is the multiple-quantum MAS (MQMAS) experiment of Frydman and Harwood, which removes the second-order broadening in a two-dimensional spectrum.[1, 42]
Gels and swollen resins: HR-MAS
A resin from solid-phase synthesis, swollen in a solvent, is neither a solution nor a rigid solid. The molecules bound to it are mobile, yet in an ordinary NMR tube they give broad lines, chiefly because the magnetic susceptibility changes abruptly at every boundary between bead and solvent.[43] This broadening has the same angular dependence as the dipolar coupling, so magic-angle spinning at a few kilohertz removes it. The probe is built for narrow lines and usually has a deuterium lock, and the technique is called high-resolution MAS, HR-MAS.[44] It developed from work on swollen polystyrene gels,[45] and in 1996 it was used to identify a compound bound to a single synthesis bead.[46]
An instrument manual gives a representative example: 2 mg of a compound on TentaGel resin, swollen in CDCl3 and spun at 3.2 kHz in a 4 mm rotor, gave proton lines as narrow as 2.7 Hz at 400 MHz. A CPMG sequence suppressed the broad signals of the resin itself.[47] An intermediate can thus be examined on the support, without cleavage. HR-MAS is not applicable to a rigid solid, because the dipolar couplings of such a sample are too large to be averaged at these spinning rates.[44]
Proton spectra and the limits of sensitivity
The proton is the most sensitive nucleus and, in a rigid solid, the hardest to resolve. At spinning rates up to about 50 kHz the 1H lines remain too broad to detect directly. At about 100 kHz they become narrow enough, although the resolution remains 10 to 100 times worse than in solution.[1] Such rates require 0.7 mm rotors and about 1 mg of sample.[1, 11] The cimetidine result in the table above was obtained at 30 kHz with a two-dimensional experiment that selects pairs of neighbouring protons.[36]
When the signal is too weak, dynamic nuclear polarisation (DNP) transfers the polarisation of unpaired electrons, whose magnetogyric ratio is 658 times that of the proton, to the nuclei. Gains of two orders of magnitude are routine in materials science, but the sample must then be measured at 20–110 K with a radical added to it.[1] In meta-aminobenzoic acid the method revealed a minor polymorph within the major one: the difference between two 13C spectra, recorded at 110 K after polarisation times of 12 s and 350 s, was the spectrum of that minor form. According to the commentary on that work, the method detects such impurity phases at a few per cent by weight.[48, 49]
Practical points for a measurement
- The question. Identifying a form, measuring a ratio of forms, detecting a few per cent of amorphous material and counting the structural units of a resin require different experiments, and measuring times range from minutes to days.[26]
- Amount. A standard rotor of 3.2 or 4 mm holds some tens of microlitres of powder;[8, 11] a smaller one holds less and gives a weaker signal.
- Sample integrity. Pressure and heat can change a solid form. Compaction into tablets measurably reduced the crystallinity of one fluorinated drug candidate,[50] and the sample in a spinning rotor is above room temperature.[12] A hydrate or a low-melting substance should be declared on the request form.
- Referencing. Shifts are referenced externally, usually to solid adamantane: either its high-frequency 13C line is set to 38.5 ppm or its low-frequency line to 29.46 ppm on the tetramethylsilane scale.[7, 9] Done carefully, this is accurate to better than ± 0.03 ppm.[51]
- Conditions in the report. Field, rotor size, spinning rate, contact time, delay, number of scans, decoupling and reference belong in the report. Without the spinning rate a sideband cannot be distinguished from a signal, and without contact time and delay a ratio cannot be judged.
- The solution spectrum. Identity and chemical purity of the same batch are established faster and with narrower lines in solution. The solid-state spectrum answers the remaining question: in which form the substance is present.
Summary
- In a powder the dipolar couplings and the orientation dependence of the shift are not averaged, so one site gives a line several kilohertz wide.
- Rotation about an axis at 54.74° to the field averages both and leaves lines at the isotropic shifts, with sidebands at multiples of the spinning rate.
- Faster spinning needs a smaller rotor and leaves less sample: 15 kHz at 4 mm, 111 kHz at 0.7 mm.
- Cross-polarisation transfers polarisation from the protons to the carbons. 13C gains a factor of up to about 4, and the repetition rate is set by the proton T1.
- CP-MAS areas depend on contact time and delay. Amounts need direct polarisation, multiCP or a calibration.
- Crystal forms differ by up to a few ppm, independent molecules in the crystal multiply the peaks, and amorphous material gives lines about ten times broader.
- Published limits of detection for a minor crystalline form lie between 0.04 and 1 %, the lowest for 19F; for amorphous content they are higher.
- Protonation moves the 15N signal of a heterocyclic nitrogen by 47–118 ppm, which separates a salt from a co-crystal.
- 11B and 27Al keep a second-order broadening under spinning, which a higher field or MQMAS reduces.
- Swollen resins and gels give nearly solution-like spectra by HR-MAS at a few kilohertz.
Solid-state NMR from NorrChemica
Our NMR service measures 1H, 13C, 19F, 31P and 11B in solution at 400 MHz, priced per sample; long acquisitions and further nuclei are quoted separately. Solid-state NMR is not on the published list. If you need a spectrum of a solid, please describe the material and the question on the request form (the crystal form, a ratio of forms, amorphous content or the structural units of an insoluble resin, for example), and we will reply with what can be measured and under what conditions. The service is available worldwide for chemical and pharmaceutical R&D. Every report is signed and comes with the native data. Research use only.
- 4-Fluorophenylboronic acidtrigonal boron, one fluorine
- Potassium (4-fluorophenyl)trifluoroboratetetrahedral boron
- 2-Fluoropyridine-5-boronic acidboron, fluorine and a basic nitrogen
- 5-Fluorouracilfluorinated drug substance
- 5 mm NMR tubes for quantitative NMRthe solution spectrum of the same batch
NMR tubes made in Germany by Deutero and shipped from Finland. See the NMR tube guide →
References
- Reif B, Ashbrook SE, Emsley L, Hong M. Solid-state NMR spectroscopy. Nat. Rev. Methods Primers 2021, 1, 2. DOI
- Polenova T, Gupta R, Goldbourt A. Magic angle spinning NMR spectroscopy: a versatile technique for structural and dynamic analysis of solid-phase systems. Anal. Chem. 2015, 87, 5458–5469. DOI
- Pines A, Gibby MG, Waugh JS. Proton-enhanced NMR of dilute spins in solids. J. Chem. Phys. 1973, 59, 569–590. DOI
- Andrew ER, Bradbury A, Eades RG. Removal of dipolar broadening of nuclear magnetic resonance spectra of solids by specimen rotation. Nature 1959, 183, 1802–1803. DOI
- Andrew ER, Bradbury A, Eades RG. Nuclear magnetic resonance spectra from a crystal rotated at high speed. Nature 1958, 182, 1659 (abstract). DOI
- European Pharmacopoeia 11.0, chapter 2.2.33, Nuclear magnetic resonance spectrometry, section “NMR in the solid state”.
- Taylor RE. Setting up 13C CP/MAS experiments. Concepts Magn. Reson. A 2004, 22A (author’s copy). eScholarship
- Sabena C, Rosso C, Castellino L, Gallo A, Alladio E, Gobetto R, Chierotti MR. Making cross polarization magic angle spinning quantitative: a solid-state NMR/multivariate curve resolution-alternating least squares approach for quantifying crystalline and amorphous phases. Chem. Methods 2025, e202500042. DOI
- Crystalline forms of lorlatinib maleate. US Patent 11,078,215 (passage on the solid-state NMR measurements). Text
- Herzfeld J, Berger AE. Sideband intensities in NMR spectra of samples spinning at the magic angle. J. Chem. Phys. 1980, 73, 6021.
- Nishiyama Y, Hou G, Agarwal V, Su Y, Ramamoorthy A. Ultrafast magic angle spinning solid-state NMR spectroscopy: advances in methodology and applications. Chem. Rev. 2023, 123, 918–988. DOI
- Samoson A, Tuherm T, Past J, Reinhold A, Anupõld T, Heinmaa I. New horizons for magic-angle spinning NMR. In: Klinowski J (ed.), New Techniques in Solid-State NMR, Topics in Current Chemistry 246. Springer, 2005 (opening pages). DOI
- van Gorkom LCM, Hook JM, Logan MB, Hanna JV, Wasylishen RE. Solid-state lead-207 NMR of lead(II) nitrate: localized heating effects at high magic angle spinning speeds. Magn. Reson. Chem. 1995, 33 (abstract). DOI
- Thurber KR, Tycko R. Measurement of sample temperatures under magic-angle spinning from the chemical shift and spin-lattice relaxation rate of 79Br in KBr powder. J. Magn. Reson. 2009, 196, 84–87. DOI
- Frye JS, Maciel GE. Setting the magic angle using a quadrupolar nuclide. J. Magn. Reson. 1982, 48, 125–131. DOI
- Mauel A, Pötzschner B, Meides N, Siegel R, Strohriegl P, Senker J. Quantification of photooxidative defects in weathered microplastics using 13C multiCP NMR spectroscopy. RSC Adv. 2022, 12, 10875–10885. DOI
- Stevens JS, Byard SJ, Seaton CC, Sadiq G, Davey RJ, Schroeder SLM. Proton transfer and hydrogen bonding in the organic solid state: a combined XRD/XPS/ssNMR study of 17 organic acid–base complexes. Phys. Chem. Chem. Phys. 2014, 16, 1150–1160. DOI
- Solid-state NMR. Lecture notes “NMR Spectroscopy”, no author named, hosted by University College London. PDF
- Hartmann SR, Hahn EL. Nuclear double resonance in the rotating frame. Phys. Rev. 1962, 128, 2042.
- Schaefer J, Stejskal EO. Carbon-13 nuclear magnetic resonance of polymers spinning at the magic angle. J. Am. Chem. Soc. 1976, 98, 1031–1032. DOI
- Solid state NMR: cross-polarization. Lecture notes, chapter 17, no author named, hosted by the Department of Chemistry, University of British Columbia. PDF
- Park S, Baker JO, Himmel ME, Parilla PA, Johnson DK. Cellulose crystallinity index: measurement techniques and their impact on interpreting cellulase performance. Biotechnol. Biofuels 2010, 3, 10. DOI
- Metz G, Wu XL, Smith SO. Ramped-amplitude cross polarization in magic-angle-spinning NMR. J. Magn. Reson. A 1994, 110, 219–227.
- Kolodziejski W, Klinowski J. Kinetics of cross-polarization in solid-state NMR: a guide for chemists. Chem. Rev. 2002, 102, 613–628. DOI
- Munson E. Solid-state NMR spectroscopy of drug substances and drug products. European Pharmaceutical Review, 29 October 2020. Article
- Offerdahl TJ. Solid-state nuclear magnetic resonance spectroscopy for analyzing polymorphic drug forms and formulations. Pharmaceutical Technology, 1 February 2006 (supplement). Article
- Du Y, Phyo P, Li M, Sorman B, McNevin M, Xu W, Liu Y, Su Y. Quantifying micromolar crystallinity in pharmaceutical materials utilizing 19F solid-state NMR. Anal. Chem. 2022, 94, 15341–15349 (abstract). DOI
- Johnson RL, Schmidt-Rohr K. Quantitative solid-state 13C NMR with signal enhancement by multiple cross polarization. J. Magn. Reson. 2014, 239, 44–49 (abstract). DOI
- Maciel GE, Sindorf DW. Silicon-29 NMR study of the surface of silica gel by cross polarization and magic-angle spinning. J. Am. Chem. Soc. 1980, 102, 7606–7607.
- US Food and Drug Administration, CDER. Guidance for industry. ANDAs: pharmaceutical solid polymorphism. Chemistry, manufacturing, and controls information. July 2007. PDF
- ICH guideline Q6A. Specifications: test procedures and acceptance criteria for new drug substances and new drug products: chemical substances. Decision tree 4. PDF
- European Pharmacopoeia, chapter 5.9, Polymorphism, as reproduced in the British Pharmacopoeia 2012, Appendix I F.
- Harris RK. NMR studies of organic polymorphs & solvates. Analyst 2006, 131, 351–373 (abstract). DOI
- Harris RK. Applications of solid-state NMR to pharmaceutical polymorphism and related matters. J. Pharm. Pharmacol. 2007, 59, 225–239. DOI
- Virtanen T, Maunu SL. Quantitation of a polymorphic mixture of an active pharmaceutical ingredient with solid state 13C CPMAS NMR spectroscopy. Int. J. Pharm. 2010, 394, 18–25 (abstract). DOI
- Maruyoshi K, Iuga D, Watts AE, Hughes CE, Harris KDM, Brown SP. Assessing the detection limit of a minority solid-state form of a pharmaceutical by 1H double-quantum MAS NMR spectroscopy. J. Pharm. Sci. 2017, 106, 3372–3377 (abstract). DOI
- Inoue M, Sumii Y, Shibata N. Contribution of organofluorine compounds to pharmaceuticals. ACS Omega 2020, 5, 10633–10640 (extract). DOI
- Rajput L, Banik M, Yarava JR, Joseph S, Pandey MK, Nishiyama Y, Desiraju GR. Exploring the salt–cocrystal continuum with solid-state NMR using natural-abundance samples: implications for crystal engineering. IUCrJ 2017, 4, 466–475. DOI
- Method and system for producing low aromatic hydrocarbon products. US Patent 10,227,535 (Example 16). Text
- Smith ME. Application of 27Al NMR techniques to structure determination in solids. Appl. Magn. Reson. 1993, 4 (opening pages). DOI
- Weiss JWE, Bryce DL. A solid-state 11B NMR and computational study of boron electric field gradient and chemical shift tensors in boronic acids and boronic esters. J. Phys. Chem. A 2010, 114, 5119–5131 (extract). DOI
- Frydman L, Harwood JS. Isotropic spectra of half-integer quadrupolar spins from bidimensional magic-angle spinning NMR. J. Am. Chem. Soc. 1995, 117, 5367–5368.
- Lippens G, et al. Study of compounds attached to solid supports using high resolution magic angle spinning NMR. Curr. Org. Chem. 1999, 3, 147–169 (abstract).
- Alam TM, Jenkins JE. HR-MAS NMR spectroscopy in material science. In: Advanced Aspects of Spectroscopy. InTech, 2012. DOI
- Stöver HDH, Fréchet JMJ. NMR characterization of cross-linked polystyrene gels. Macromolecules 1991, 24, 883–888.
- Sarkar SK, Garigipati RS, Adams JL, Keifer PA. An NMR method to identify nondestructively chemical compounds bound to a single solid-phase-synthesis bead for combinatorial chemistry applications. J. Am. Chem. Soc. 1996, 118, 2305–2306.
- HR-MAS manual, section 7.1, HR-MAS of small organic molecules. Copy at IMSERC, Northwestern University. Page
- Chen Y, Mi J, Rossini AJ. A focus on detection of polymorphs by dynamic nuclear polarization solid-state nuclear magnetic resonance spectroscopy. Chem. Sci. 2023, 14, 11296–11299. DOI
- Cousin SF, et al. Exploiting solid-state dynamic nuclear polarization NMR spectroscopy to establish the spatial distribution of polymorphic phases in a solid material. Chem. Sci. 2023. DOI
- Liu J, Nagapudi K, Kiang YH, Martinez E, Jona J. Quantification of compaction-induced crystallinity reduction of a pharmaceutical solid using 19F solid-state NMR and powder X-ray diffraction. Drug Dev. Ind. Pharm. 2009, 35, 969–975 (abstract). DOI
- Morcombe CR, Zilm KW. Chemical shift referencing in MAS solid state NMR. J. Magn. Reson. 2003, 162, 479–486 (abstract). DOI
About this article. We read sources 1, 3, 4, 7, 8, 14, 16, 17, 22, 25, 26, 38, 44 and 48 in full, and the lecture notes 18 and 21. In 16 the equation of the build-up curve and the table of time constants are not in the text we could read, so we give only the values stated in its text; in 26 the tables were likewise not readable. Sources 2 and 11 were read in part: the opening sections and, in 11, the tables of rotors. Of 6, 12, 30, 31, 32, 40 and 47 we read the passages we cite, 32 as reproduced in the British Pharmacopoeia, and of the patents 9 and 39 the passages on the NMR measurements; we cite patents because they state measurement conditions that journal articles seldom give. Sources 5, 13, 27, 28, 33, 35, 36, 37, 41, 43, 50 and 51 were read as abstracts or extracts only. Sources 10, 15, 19, 20, 23, 24, 29, 34, 42, 45, 46 and 49 are bibliographic records, cited for the origin of a method or for further reading; what we say about 49 comes from 48. We found no readable source for the width of a 13C powder pattern in ppm or for the equation of the cross-polarisation build-up, and give neither. Our service is described from the service page of our shop. The positions of the sidebands, the measuring times in the table of published conditions and the curve in Figure 2 are our own calculations from the values cited. Figures 3 and 6 are our drawings of published values, with schematic line shapes in Figure 6; Figures 1, 4 and 5 are schematic. All sources were read on 14 November 2026.
NMR Series
- Part 1 — Reading 1H and 13C NMR
- Part 2 — Reading 19F, 31P and 11B NMR
- Part 3 — Reading 2D NMR: COSY, HSQC and HMBC
- Part 4 — Quantitative NMR: how a purity figure is made
- Part 5 — 15N NMR: which nitrogen is alkylated, protonated or coordinated?
- Part 6 — NMR tubes by experiment: which tube for which measurement
- Part 7 — Quadrupolar nuclei in NMR: broad lines and missing carbons
- Part 8 — Two sets of NMR signals: rotamers, diastereomers or an impurity
- Part 9 — NMR yield and isomer ratios: relaxation and the delay between pulses
- Part 10 — Silicon NMR: the glass signal, the negative NOE and M, D, T, Q units
- Part 11 — Solid-state NMR: magic-angle spinning, cross-polarisation and what a CP-MAS spectrum counts (this article)
- Part 12 — Enantiomeric excess by NMR: when NMR is the best method, the chiral auxiliaries and the detection limit for the minor enantiomer
- Part 13 — Predicted NMR shifts: how accurate ChemDraw, MestReNova and other predictors are
- Part 14 — Phosphorus NMR in biology: phospholipids, phosphorylated proteins, oligonucleotide drugs and phosphates in food
- Part 15 — Multiplets in ¹H NMR: coupling trees, second-order patterns and two signals from one CH₂
- Part 16 — NMR for cosmetics: adulterated essential oils, free formaldehyde, aloe vera, rancid lipsticks and mineral oil in lip products
- Part 17 — Polymers by NMR: composition, sequence, chain ends and branches, and what a spectrum of many chains can and cannot measure
- Part 18 — Benchtop NMR: what a permanent-magnet spectrometer does well, and when a sample needs high field
- Part 19 — NOESY, ROESY and TOCSY: regioisomers, E and Z isomers, anomers, and the signals of one compound in a mixture
- Part 20 — Rapid-injection NMR: starting a reaction inside the spectrometer, how fast the spectra can follow it, and the intermediates it has revealed
- Part 21 — Fluorine NMR for PFAS: total organofluorine, what targeted LC-MS/MS does not detect, and the detection limit
- Part 22 — Polysiloxanes by proton NMR: silicone oil in medicines, food and cosmetics, the reactive groups before curing, and the rubber network after it
© 2026 NorrChemica. All rights reserved.
Content may not be reproduced without written permission.