Cubic three-dimensional hybrid silica solids for nuclear hyperpolarization

Porous network architecture of hybrid silicas containing TEMPO radicals along their pores is key for increased hyperpolarization performances.

For each sample, the material was impregnated with 1,1,2,2-Tetrachloroethane and filled in a 3.0 mm quartz tube. All spectra were recorded at 110K using a nitrogen flow cryostat. Attenuation was varied from 32 to 20 dB. The EPR spectrum of a nitroxide radical consists of three lines due to strong hyperfine interaction with the 14N nucleus. For the line-width measurements, we have used the central line. Since it is least broadened by the g-tensor and hyperfine anisotropies, it is therefore the most sensitive to the dipolar broadening. For the obtained signal to noise ratios, the estimated linewidth errors were within the order of 5 %.

Spin count.
The samples were filled in a 3.0 mm quartz tube with a maximum sample height of 3mm. The sample position in the cavity was carefully optimized. The spectra were recorded at room temperature and with a sweep width of 600 Gauss and attenuation between 26 and 14 dB. The amount of radical was determined by double integration of the CW spectra and referencing to calibration curve of TEMPO in toluene solutions measured for the concentration range between 0.4 and 80 mM. An additional correction for the difference in the incident microwave power has been taken into account. Data was processed with OriginLab.

EPR analysis of filtered solutions after dissolution.
After dissolution, the solution was filtered and subjected to centrifugation (5000 rpm for 5 min). The supernatant was filled in a Hirschmann glass capillary with a sample height of 18 mm, which was then closed with putty. The sample position in the cavity was carefully optimized. The spectra were recorded at room temperature with a sweep width of 600 Gauss and attenuation at 20 dB. The number of scans per spectrum was adjusted to have a reasonable signal-to-noise ratio. The amount of radical was determined by double integration of the CW spectrum and referencing to the calibration curve of 4hydroxy TEMPO aqueous solutions measured of the concentration range between 0.002 and 0.2 mM. An additional correction for the difference in the incident microwave power has been taken into account. Data was processed with MATLAB ® (R2011a, MathWorks Inc.).

Transmission Electron Microscopy (TEM).
Conventional TEM micrographs were performed at the "Centre Technologique des Microstructures", UCBL, Villeurbanne, France, using a JEOL 2100F electron microscope. The acceleration voltage was 200 kV. The samples were prepared by dispersing a drop of the ethanol suspension of a ground sample on a Cu grid covered by a carbon film.

Nitrogen Adsorption-desorption (BET).
The Nitrogen adsorption and desorption measurements were achieved at 77 K using a BELSORB-Max from BEL-JAPAN. Before N 2 adsorption, the samples were outgassed at 10 -5 mbar at 408 K for 12 h. The pore diameter distribution and the mean pore diameter (d p ) were calculated using Barrett-Joyner-Halenda (BJH) method. The specific surface area (S BET ) was according to Brunauer-Emmett-Teller (BET) equation.

Sample Preparation
Samples were prepared by incipient wetness impregnation (see the movie http://pubs.acs.org/JACSbeta/scivee/index.html#video2) using 8.5-16.2 mg of dry powder with 20-30 µL of an analyte containing solution or with pure solvent (H 2 O or 1,1,2,2-tetrachloroethane, EtCl 4 ). The total mass of impregnated material was determined, and the sample was mixed using a glass-stirring rod to obtain a homogeneous distribution of the solution in the powder. The impregnated powder was then packed into a 3.2 mm sapphire NMR rotor to maximize MW penetration into the sample. The mass of impregnated material inside the rotor was determined, and a tight polyfluoroethylene plug was inserted to prevent any leakage of the solvent during spinning. The rotor was capped with a zirconia drive tip and quickly inserted into the DNP spectrometer.
All spectra were acquired on a Bruker Avance III 400 MHz DNP NMR spectrometer equipped with a 263 GHz gyrotron microwave system (B 0 = 9.4 T, ω H /2π = 400 MHz, ω C /2π = 100 MHz, ω Si /2π = 79.5 MHz). The field sweep coil of the NMR magnet was set so that MW irradiation occurred at the DNP enhancement maximum of TOTAPOL (263.334 GHz), with an estimated 4 W power of the MW beam at the output of the probe waveguide. 1 H, 13 C, and 29 Si spectra were recorded using a triple resonance low-temperature CPMAS probe with a sample temperature of 99 K and sample spinning frequencies of 8 kHz for the 13 C and 29 Si spectra. SPINAL-64 heteronuclear decoupling was applied during acquisition (ω 1 H /2π = 100 kHz). 1 H, 13 C and 29 Si chemical shifts are referenced to TMS at 0 ppm. Standard cross-polarization (CP) was used for 1D carbon-13 and silicon-29 spectra. For 13 C CPMAS, the 1 H π/2 pulse length was 2.5 μs (100 kHz). A linear amplitude ramp (from 90% to 100% of the nominal RF field strength) was used for the 1 H channel, with a 2 ms contact time and a nominal RF field amplitude of ν 1 = 76 kHz for 1 H and 40 kHz for 13 C. SPINAL-64 proton decoupling was applied during the acquisition of the 13 C signal with an RF field amplitude of ν 1 = 100 kHz. The 13 C acquisition time was 20 ms with 1608 complex points.
For 29 Si CP-MAS, the 1 H π/2 pulse length was 2.5 μs (100 kHz). A linear amplitude ramp (from 90% to 100% of the nominal RF field strength) was used for the 1 H channel, with a 2 ms contact time and a nominal RF field amplitude of ν 1 = 76 kHz for 1 H and 35 kHz for 29 Si. SPINAL-64 proton decoupling was applied during the acquisition of the 29 Si signal with an RF field amplitude of ν 1 = 100 kHz. The 29 Si acquisition time was 20 ms with 1608 complex points.
Spectra were acquired with 32 -256 scans for 13 C and 32-1024 scans for 29 Si in order to obtain a good signal on noise ratio.
Processing of the spectra was done using the Topspin software package.

Calculation of DNP Enhancement Factors (ε).
DNP enhancement factors on the nucleus X (ε X ) were determined by scaling the intensities of the spectra of the nuclei X obtained under the same experimental conditions with or without MW irradiation.

Sample preparation
A solution of 20 vol% H 2 O and 80 vol% D 2 O is prepared for 1 H polarization study. 20 to 25 mg of HYPSO is impregnated with this solution so that the volume of solution corresponds to 90-100 % of the HYPSO total pore volume (P/P 0 <0.99), determined by N 2 adsorption at 77K. The impregnated material is then transferred to a home-built Teflon sample holder, then placed in the polarizer and cooled down to 4.2 K or 1.2 K.

Experimental method
4. DNP is performed at 4.2 K or 1.2 K by microwave irradiation (188.3 GHz, set for negative DNP, and 87.5 mW) in a field of 6.7 T (285.23 MHz for protons). Frequency modulation can be applied or not. When it was applied, the amplitude of the frequency modulation was set to Δν µw = 100 MHz with a modulation frequency f mod = 10 kHz. The used setup is described in the experimental description.
5. The DNP build-up of 1 H spins is measured with 1° nutation angle pulses followed by an acquisition period of 1ms. The resulting free induction decay is Fourier transformed, phased and integrated to obtain a value of 1 H spin for the corresponding acquisition period. This sequence is applied every 5 seconds during at least 250 seconds (up to 1500 seconds for long build up time), thus, obtaining at least 50 values of 1 H spin.

The resulting DNP build-up curve is fitted with function defined in function of time (t) as
leading to the values P max , and to the polarization build-up time !"# . The fit is performed using the asymptotic-symmetry based method, chi-square minimization being the procedure used to minimize the residual sum of square (refer to "Inference in Nonlinear Fitting" guide line of origin lab for more details) between the experimental points and the calculated values. The 1 H polarization P max hence obtained is used for comparing the materials and hereunder referred as P( 1 H). The build-up time τ DNP (seconds) corresponds to time required to reach 63% of polarization, 5×τ DNP corresponding to 99% of P max = P( 1 H).

nm 200 nm
The peak at ca. 2100 cm -1 corresponds to -N 3 stretching vibrations.
Diffusive Reflectance Infra-Red Fourier Transformed (DRIFT) analysis of the powder allowed to evaluate the efficiency the cycloaddition on HYPSO-3 (referred to as CuAAC yield) by monitoring the decrease of the intensity of the -N 3 absorption peak, at 2210 cm -1 (see Figure S3).