Parahydrogen-induced polarization with a metal-free P–P biradicaloid

The activation of parahydrogen by a metal-free P–P biradicaloid leads to 1H and 31P nuclear spin hyperpolarization.

. 1 H NMR spectra obtained after a 5 s parahydrogen bubbling through a 0.04 M toluene-d8 solution of [P(µ-NTer)]2 (a), and after the relaxation to thermal equilibrium (b) at 293 K. Both spectra were acquired using π/4-pulses. ...3 Figure S2. Experimental 31 P NMR spectra of adduct [HP(µ-NTer)]2 acquired after parahydrogen bubbling at 353 K through the equilibrium solution after the full conversion of [P(µ-NTer)]2 (a), and after the relaxation of the nuclear spins (b). The spectrum (b) is multiplied by a factor of 32 relative to spectrum (a). The comparison of the spectra reveals the signal enhancement on the order of magnitude from 60 to 300, depending on a transition line. Both spectra were acquired using π/4-pulses. The signal-to-noise ratio was lower at 353 K as compared to 293 K ( Figure Figure S3. A simulated thermal 1 H NMR multiplet corresponding to H2 originating atoms in [HP(µ-NTer)]2 adduct (blue trace) and all possible non-zero contributions of the AA'XX' spin states to the multiplet's most intense transition lines (red, yellow, violet, green traces Figure S4. (a) Experimental 1 H decoupled 31 P NMR spectrum obtained by applying a weak-power (63 Hz RF field amplitude) 42 ms pulse after parahydrogen bubbling. (b) 1 H decoupled 31 P NMR spectrum obtained by using a hard π/4pulse after the relaxation of spins to thermal equilibrium. The spectrum (b) is multiplied by a factor of 32 relative to spectrum (a). The small "satellite" signals in (

General information
NMR experiments with parahydrigen were performed on a 300 MHz Bruker AV 300 NMR spectrometer equipped with a broad-band 10 mm RF probe. The standard temperature control unit of the NMR spectrometer was used for cooling and heating the samples. The kinetic measurements were performed on a 400 MHz Spectrometer Bruker AV 400 NMR equipped with a broad-band 5 mm RF probe.
[P(µ-NTer)]2 was synthesized using procedures described in Refs. S1,2 High-purity commercially available H2 gas was used for producing parahydrogen-enriched H2 referred to in the main text as simply parahydrogen. The enrichment was performed with a Bruker parahydrogen generator, which produced H2 gas with 91 % of parahydrogen.

NMR experiments
In a typical workflow, parahydrogen was bubbled trough a 0.04 M solution of the biradicaloid [P(µ-NTer)]2 in degassed toluene-d8 in a 5 mm sample inside the NMR magnet for ca. 5 s, and then the parahydrogen flow was abruptly switched off and an NMR experiment was started. The bubbling procedure was performed under atmospheric pressure (1 bar) in the same manner as explained in detail in Ref. S3 Parahydrogen was supplied to the bottom of the sample tube through a 1/32" PTFE tubing. The sample temperature was varied in the experiments when it was required.
Since [P(µ-NTer)]2 is highly sensitive to both air and moisture, the sample preparation procedures were done under inert Ar atmosphere.

Full range 1 H spectra acquired at 293 K
A full-range 1 H NMR spectrum acquired after parahydrogen bubbling through the 0.04 M solution of [P(µ-NTer)]2 in toluene-d8 is shown in Figure S1a, with part of this spectrum also presented in Figure 1a. For comparison, a thermal equilibrium spectrum after the relaxation is shown in Figure S1b. The detection of unusual antiphase signals after the parahydrogen bubbling serves as a solid justification for the formation of the hyperpolarized state.
The P-P and H-H pairs form the symmetric AA'XX' spin system in the resulting [HP(µ-NTer)]2 adduct. Nevertheless, like in common PASADENA experiments, S4 π/4 RF pulses provided higher signal amplitudes compared to those obtained with π/2-pulses. The exact 1 H signal enhancement measurement was difficult to perform, because thermally polarized signals of the adduct [HP(µ-NTer)]2 were hampered by the signals of aromatic groups in the measured 1 H NMR S3 spectra. The more precise estimation of enhancement was done using 31 P NMR spectra as described in the main text. Figure S1. 1 H NMR spectra obtained after a 5 s parahydrogen bubbling through a 0.04 M toluene-d8 solution of [P(µ-NTer)]2 (a), and after the relaxation to thermal equilibrium (b) at 293 K. Both spectra were acquired using π/4-pulses.

31 P NMR spectra acquired at 353 K after parahydrogen bubbling
As described in the main text, the polarized signals begin to appear after every bubbling of parahydrogen upon heating the sample to 353 K, while at 293 K polarization effects are no longer visible after a few bubbling repetitions because of the full conversion of [P(µ-NTer)]2 to [HP(µ-NTer)]2. This implies that the reaction at this temperature becomes reversible "enough" so that parahydrogen can replace the "relaxed" hydrogens in [HP(µ-NTer)]2. Figure S2a shows 31 P NMR spectra observed reproducibly after the parahydrogen bubbling at 353 K through the reaction mixture. The 31 P signal multiplet measured at 353 K ( Figure S2a) looks very similar to that measured at 293 K ( Figure 2a of the main text). The corresponding thermal spectrum is depicted in Figure S2b. It should be noted that mainly due to the sensitivity of the 31 P RF probe to the temperature in our setup, the thermal polarization 31 P signal at 353 K ( Figure S2b . Experimental 31 P NMR spectra of adduct [HP(µ-NTer)]2 acquired after parahydrogen bubbling at 353 K through the equilibrium solution after the full conversion of [P(µ-NTer)]2 (a), and after the relaxation of the nuclear spins (b). The spectrum (b) is multiplied by a factor of 32 relative to spectrum (a). The comparison of the spectra reveals the signal enhancement on the order of magnitude from 60 to 300, depending on a transition line. Both spectra were acquired using π/4-pulses. The signal-to-noise ratio was lower at 353 K as compared to 293 K ( Figure 2b of the main text) due to the sensitivity of 31 P RF probe to the temperature.

Simulations of NMR spectra of [HP(µ-NTer)] 2
Technically, the NMR simulations of the expected line-shape after parahydrogen addition to [P(µ- should be noted that JHH has a relatively large value for a long-range HH coupling constant through four chemical bonds. Likely, it is due to the specific structure of the biradicaloid adduct with H2, which is relatively rigid. According to our preliminary computational studies, the main contribution to the J-coupling between the two hydrogens arises from the Fermi contact interaction. In addition, all molecular orbitals that are likely to partake in the coupling path show contributions to the P-H and P-N bonds. Therefore, it is hard to say whether the J-coupling is purely through-space or bondmediated. It seems that a cooperative effect results in the large JHH value. As this communication is focused on nuclear hyperpolarization effects, the structural features of [HP(µ-NTer)]2 will be addressed elsewhere.

Spin states of AA'XX' spin system of H-H and P-P pairs in [HP(µ-NTer)] 2
The spin states of the resulting AA'XX' spin system is convenient to build in the representation of triplet and singlet nuclear spin pairs made out of 1 H (A nuclei) and 31 P (X nuclei) similarly like it is done in Ref. S6 Since parahydrogen accommodates only the singlet nuclear spin state, it is obvious that all states with singlet H pairs in the AA'XX' (ST+, ST-, SS, ST0) will be overpopulated after the parahydrogen addition. The last two states are not the eigenstates since they are premixed to T0S and T0T0 states by the coupling network. Therefore, these latter states should be overpopulated as well,

Kinetic measurements
The measurements of kinetic constants were performed using spin saturation transfer method S7 on a Bruker Avance III 400 MHz spectrometer. A heavy-wall 5 mm NMR tube equipped with a tight plug was used in the experiments.

SLIC spectra
The spin-lock induced crossing (SLIC) S8 tests were done by applying a single RF pulse of weak power on the 31 P resonance frequency (121.5183837 MHz, ca. 194 ppm) after the parahydrogen bubbling. The employed NMR pulse sequence is shown in Scheme S1. The optimized values for the length and the power level were estimated by using numerical simulations. For example, Figure 4a shows the 1 H decoupled (WALTZ16) 31 P NMR spectrum acquired after applying 42 ms 31 P pulse of 63 Hz RF field amplitude. The small "satellite" signals in this spectrum is a manifestation of the non-ideality of the 1 H decoupling. The corresponding 1 H decoupled 31 P NMR spectrum obtained after the relaxation is shown in Figure 4b for comparison.  Figure S4. (a) Experimental 1 H decoupled 31 P NMR spectrum obtained by applying a weak-power (63 Hz RF field amplitude) 42 ms pulse after parahydrogen bubbling. (b) 1 H decoupled 31 P NMR spectrum obtained by using a hard π/4pulse after the relaxation of spins to thermal equilibrium. The spectrum (b) is multiplied by a factor of 32 relative to spectrum (a). The small "satellite" signals in (a) are artifacts of non-ideal 1 H decoupling.

Computational details
Electronic structure computations were carried out using Gaussian09 [S9] and ORCA 4.0.1. [S10,11] To estimate the activation barrier of H2 release, the structures of the model compounds  [S12-16] All structures were confirmed as minima or transition states by frequency analyses.
T h e o p t i m i z e d s t r u c t u r e s w e r e u s e d f o r s i n g l e -p o i n t D L P N O -C C S D ( T ) / a u g -c c -p V T Z calculations [S17-23] to obtain more reliable estimates of the electronic energies. The single point S9 energies were then added to the thermal corrections obtained from the frequency analyses to estimate the enthalpies H and Gibbs free energies G (Table S1).  Hence, the activation barrier for H2 release amounts to ΔH ‡ = 22.6 and ΔG ‡ = 24.4 kcal/mol, in good agreement with the experimental values. The activation barrier for the reverse reaction, i.e. addition of H2 to the biradical, is much smaller (ΔH ‡ = 9.7, ΔG ‡ =18.1 kcal/mol), explaining the quick reaction at ambient temperature. The addition of H2 is an exothermic and exergonic process (ΔH° = −12.9, ΔG° = −6.3 kcal/mol). Notably, the trans isomer of [HP(μ-NPh)]2 is thermodynamically only slightly less favoured than the cis isomer (ΔG° = 2.7 kcal/mol). Thus, the exclusive formation of the cis isomer is a kinetic effect.