This manuscript presents a new pulse sequence which offers a way to probe the nuclear polarization obtained via DNP by transferring it back to electron polarization after a step of saturation of the electron spins. The authors call this experiment reversed DNP. This is an elegant and original idea and the technique, based on the NOVEL DNP sequence,  will be  useful  in the future to study nuclear spin diffusion and may find applications in ENDOR. The presentation is generally clear (though many experimental details are missing, see below) .The authors present a convincing series of experiments, which clearly demonstrate that their idea works and do the right control. I liked the work and I think it fits very well to MR. I recommend publication after addressing the following:

1. The new method is demonstrated at Q-band frequencies (for the electrons) , ~ 50 MHz for protons, which is way below the standard NMR frequencies at which DNP is applied. This is a general problem of NOVEL. It will be nice to have some outlook regarding possibilities of using this at more relevant frequencies.
2. The technique is demonstrated on trityl, which at Q-band has a rather narrow EPR signal and with the help of chirp pulses the authors managed to excite the whole spectrum. At a high field the spectrum of trityl broadens and full excitation may not be possible. What will be the consequences of partial excitation (saturation) ? please add  to the discussion
3. Related to point 2 – will this work for nitroxides, which are often used as polarizers for DNP?
4. Many experiential details are not given, like length of pulses, delays, properties of the chirp pulses when used, details on the train of saturating pulses, their number and interval , how long it takes to acquire the data. Please give these details in the fig. captions or in the Experimental part for each figure shown in the manuscript and the SI.
5. Figure 5b –it is not clear to me what was the sequence , where are inversion pulses used ? , the sequences show only saturation train.
6. In the introduction – around line 20,  maybe mention that schemes different than just CW irradiation exist.
7. S9 – please remove the application ruler.

This is an original article by Nino et al. The authors used a coherent DNP pulse sequence to obtain information about the protons surrounding the electron spins. They qualitatively described the effect of deuteration and radical concentration on the decay of the nuclear polarization surrounding the unpaired electrons, as well as used "electron decoupling". The authors prove that the proton polarization decay around the radical is mostly driven by spin diffusion away from the said electron spins. To the best of my knowledge, this is the first time such experiments are carried out. The article is well written and its stimulating content warrants publication in MR. I have minor comments that should be addressed in order to improve the manuscript.

1. The introduction states that the experiments could be used for ENDOR and to probe the DNP mechanism which I entirely agree with. However, the discussion is rather succinct. How would the reverse DNP be used for ENDOR? My understanding is that the NOVEL condition requires large mw power, as a consequence would it be possible to have “resolution”? Would an ENDOR require a frequency sweep of a rf pulse?
2. In addition, could we imagine using different pulse sequences? e.g. can we use an eNCP (Rizzato, 2013) to generate the nuclear polarization and then carry out the same reverse CP? If yes, this may be more practical for high field? (Bearing in mind that using offset to have HH condition may be similar as doing off-resonance NOVEL).
3. Could the first step of the nuclear hyperpolarization be replaced by a long ELDOR pulse? Then read the polarization with reverse NOVEL (or other pulse seuqence)? Same question, with DAVIES ENDOR, Iz generated after the rf pi pulse, so DAVIES could also be used?
4. In general, having more experimental details, such as the delays used, the timing in the pulse sequence figures, would be appreciated. Since MR is a specialized journal, I do not think this would make the manuscript less valuable.
5. “These findings strongly indicate that nuclear polarisation is generated during the first NOVEL block, even without direct proton NMR detection.” True but that’s like ELDOR, isn’t it? So "strongly" is probably not needed.
6. Why is the “minimum of the electron spin echo intensity at 184 ns”?
7. “Note that even in deuterated solvent, there are still 48 non-exchangeable protons in OX063. We speculate that the transfer efficiency could be improved for less abundant nuclei, but experiments are needed to test this hypothesis.” I am not sure I understand what is meant here. Do you think that a fully deuterated OX063 would enable higher transfer efficiency? If so, could you explain a bit more?
8. Figure 2, would it be possible to add on the pulse sequence figure the timings in between the pulses? Not as numerical values, but as variables.
9. Figure 3, a, is this deuterated or protonated? In the text it is mentioned but not in the caption. This might improve the readability.
10. Figure S9 has an issue
11. Figure S13, the caption could be improved. I think it would be better if it was described by a separate paragraph.

The manuscript by Wili et al describes a new polarization transfer experiment between electron and nuclear spins that consists of two steps. The first step has been reported in the literature and is based on electron-nuclear cross-polarization at the NOVEL condition. The second step is a reverse transfer of polarization from nuclei to electron spins at the same NOVEL condition. The experiment is very interesting as it provides a tool to monitor the dynamics of nuclear polarization after a pulse DNP step, which is otherwise difficult to access. The manuscript is overall clearly written, the experimental effects are robust and seem rationalized.  Sometimes details are missing. I would recommend publication in MR after adding the following details.

• The first step of the sequence contains a subtle difference to the DNP experiment proposed by the Griffin group, (Can 2015, Mathies 2016). In Figure 2, the NOVEL step is repeated only once, whereas in DNP it is repeated n times (n >~ 1000). In my understanding, this multiple contact is required as one electron spin polarizes a large number of nuclear spins. The question is how much the nuclear spins are polarized in average after only one contact and what is expected then in the reverse polarization step. The authors observe an increase in the intensity of ‘repolarized’ electron spin in deuterated solvent, indicating that the number of nuclear spins plays indeed a role, see also comment on page 5 (‘we speculate that the transfer efficiency could be improved for less abundant nuclei….’). Could the contact time be repeated, which might enhance the efficiency, or are there any hardware constraints to perform this check?
• In context of point (1), in SI 9 the authors report an experiment in which the NOVEL sequence is repeated several times. Perhaps I have missed this, but I could not find a reference to this figure in the main text. This experiment is called ‘saturation behavior during depolarization’ but it is not entirely clear. Please give the sequence and explain better, also in the main text, what is normalized to ‘one’ and why the echo is larger than 1 only in the deuterated sample.
• Electron polarization decays as a function of waiting times. The authors write (page 6) that the decay is ‘supposably’ assigned to proton polarization decay. It would be nice to know what other competitive pathways they suspect.
• Figure 4 shows a clear effect by the solvent deuteration on the ‘electron’ spin re-polarization dynamics. The effect is tentatively assigned to the faster proton polarization decay in protonated solvent. It is not clear whether we can learn about the nuclear T1n, which would be very important also for ENDOR.
• Effect of ‘decoupling sequence’ on the polarization decay. The effect of ‘decoupling’ is interesting but it would be good to see the sequence in Figure 5 and have a rationale of the chosen parameters (length, intensity, repetition rate, setting of adiabatic pulses) also based on the hyperfine couplings (how large are they ?)

Finally, the authors should add some comments about how this experiment could be beneficial for ENDOR, particularly whether it could be extended to radicals with (much) larger hyperfine interaction and the feasibility at higher fields.