N-type organic thermoelectrics: demonstration of ZT > 0.3

The ‘phonon-glass electron-crystal’ concept has triggered most of the progress that has been achieved in inorganic thermoelectrics in the past two decades. Organic thermoelectric materials, unlike their inorganic counterparts, exhibit molecular diversity, flexible mechanical properties and easy fabrication, and are mostly ‘phonon glasses’. However, the thermoelectric performances of these organic materials are largely limited by low molecular order and they are therefore far from being ‘electron crystals’. Here, we report a molecularly n-doped fullerene derivative with meticulous design of the side chain that approaches an organic ‘PGEC’ thermoelectric material. This thermoelectric material exhibits an excellent electrical conductivity of >10 S cm−1 and an ultralow thermal conductivity of <0.1 Wm−1K−1, leading to the best figure of merit ZT = 0.34 (at 120 °C) among all reported single-host n-type organic thermoelectric materials. The key factor to achieving the record performance is to use ‘arm-shaped’ double-triethylene-glycol-type side chains, which not only offer excellent doping efficiency (~60%) but also induce a disorder-to-order transition upon thermal annealing. This study illustrates the vast potential of organic semiconductors as thermoelectric materials.

How is this consistent with the picture provided earlier where these side chains order at low temperatures, and upon heating? 12) line 272. A doping efficiency of 60% is calculated. Are these polarons formed per dopant molecule or free charges? And how does this value compare to the proposed disappearance of aggregates (line 200), which leads to a higher conductivity. Does the doping efficiency increase upon annealing? The authors quote carrier densities at 120 and 150 deg but only one doping efficiency. Is this because the doping efficiency does not actually increase, which would be inconsistent with the earlier argument that more dopant is taken up by the fullerene material? 3) The doped fullerene material is not stable under ambient conditions. All electrical characterization was done under nitrogen. But how about the thermal conductivity measurement.
How did the authors ensure that the sample had not degraded? The Linseis system requires that the sample is handled in air. Was the instrument placed in a glove box?
Reviewer #3: Remarks to the Author: Liu et al. claimed that they developed OTE materials with ZT value over 0.3 following PGEC concept. In fact, this concept is widely used in inorganic materials, but its applications in OTE materials remain unclear. Although the basic concept seems interesting, I have the follow concerns about the results.
(1) For typical PGEC concept in inorganic materials, the designed atoms are introduced into crystal cell to ensure high electrical conductivity and suppress the lattice thermal conductivity. Although the authors claimed their designed materials is similar the concept, I failed to find a clear definition of the concept in organic materials. Notably, several doped organic semiconductors have been confirmed to possess ordered molecular packing and the dopant is located at side-chain regime. Once again it raises a question, how could define PGEC materials in conjugated systems? Whether these previous reported materials also follow the concept?
(2) The ultralow thermal conductivity and high ZT are the key evidences for the so-called PGEC concept. I have several concerns about measurement of the κ and ZT values: i) As a well-known fact, it is challenge to measure sample with low in-plane thermal conductivity (κ￡0.1 W/m K). Even with the technique provided in the supporting information, I'm not sure about the accuracy of results. A prefered thickness of the sample should be > 400 nm for a sample with κ of 0.4 W/m K (Journal of Electronic Materials 2018, 47 (6) , 3203-3209. DOI: 10.1007/s11664-017-5989-4). I believe that the situation is very challenge for PTEG-2 film since much thicker film is required. The author should provide the exact thickness of the sample used in TFA measurement and evaluate the deviation.
ii) I note the measurement of σ and S were performed in a N2-controlled environment while κ∥was obtained in vacuum by TFA. What about the air stability of the sample during the sample transfer? Such kind of n-type materials usually possess poor air stability, which will lead to rapid dedoping and lower κ∥because of reduced contribution from electron transport. In fact, TFA can characterize all the parameters and even temperature dependent performance in one chip. Why the author prefers to utilize different techniques to characterize the samples?
iii) It is difficult to understand that the σ shows a large variation versus temperature (>30%) while the κ∥almost keeps nearly unchanged in Figure 5. The electronic contribution of the κ is related to the σ and follows Wiedemann-Franz law. In other words, one should observe the temperature dependent κ even for PGEC materials. The authors should identify the electronic contribution and lattice contribution to the κ to confirm and understand the phenomenon.
3) The authors found that the electrical conductivity first rises then falls along with the increase temperature. They attribute the decreased σ to the enhanced disorder of the side chain. Providing the energetic disorder before and after phase transition will make it clearer. In fact, dedoping can also contribute to the decrease of the conductivity. What about the thermal stability of doped films?
4) The doped PTEG-2 film shows high crystallinity. Since the author use 'electron crystal' to describe the materials, I suggest the authors to evaluate of the carrier concentration and mobility by TFA to support the definition.
Based on these concerns, I am afraid that I cannot recommend its publication in the present form.
We thank the reviewers for their constructive comments and positive appraisal of our manuscript. We have performed a number of additional experiments and have added text and plots to address the issues raised by the reviewers.

Reviewer #1:
The authors carry out an excellent study of the thermoelectric properties of a series of fullerene derivatives. The striking result is a record high ZT for n-type organic materials (and perhaps higher than the p-type record). The work is very thorough and I do not have any major comments on the detailed technical work. 1). A minor point is that I do not really think this is an electron-crystal/phonon glass. There is no evidence of extended state conduction, i.e. an electron crystal. Likely the conduction is by hopping which, to me, is not really an electron crystal. The paper cited on disorder is for a polymer that arguably has higher disorder, but a much higher electrical conductivity which somewhat contradicts the statement. This is not meant to detract from the major results in the paper.

Response:
We thank the review for the nice comments.  the charge carrier mobility should reach the crystalline limit of this particular material.
In this manuscript we show that the thermal conductivity does indeed reach (it's actually even lower than) the amorphous limit. For the related fullerene derivative PCBM there are several reports in the literature that show that the pristine material (i.e. undoped) does indeed satisfy this part of the definition. Here we show that 1) PTEG-2 has a similarly low thermal conductivity, and 2) it is not affected by doping. In sum, doped PTEG-2 is a phonon glass.
Regarding the charge carrier mobility: Frankevich et al. (Frankevich, E.;Maruyama, Y.;Ogata, H. Chem. Phys. Lett. 1993, 214, 39.) have found a (time-of-flight) mobility of 0.5±0.2 cm 2 /Vs for singlecrystal C60 and Bao et al. reported a high field-effect transistor mobility of 5.2±2.1 cm 2 /Vs for a needlelike single crystal (J. Am. Chem. Soc. 2012, 134, 2760. It stands to reason that the mobility of C60 is an upper limit to that of fullerene derivatives as the side-chains will dilute the conjugated part of the system. In our manuscript, we find a bulk mobility of 1.2 cm 2 /Vs which is on the same order of magnitude as that of single crystal C60. In other words, PTEG-2 comes close to (if not actually fulfills) the mobility requirement.
The referee is right, however, in saying that doped PTEG-2 probably does not show band transport. We do note, however, that the conductivity (see Fig 5a) shows little temperature dependence. Above 120 o C the conductivity even shows a negative temperature dependence. Figure 5b (Seebeck coefficient) shows that the Seebeck coefficient does not change with temperature, suggesting that the number of charge carriers remains constant. Taken together, this suggests that the conductivity might even be metallic in the sense that it shows negative temperature dependence. However, we do not wish to claim that this is the case as that would require a more detailed study.
All in all, we have revised our claim by stating that our material system approaches the PGEC concept.
2) Can the authors compare these results to other doped fullerenes, e.g. potassium-doped materials, in literature to determine the role of the functional group? It would be nice to see the Lorenz factor here; the electrical conductivity is likely too low for the thermal conductivity of electrons to have much of an impact.

Response:
We compared the best doped PTEG-2 film with reported alkali metal doped C60 films in terms of electrical conductivity. It is interesting to find that the electrical conductivity (>8 S cm -1 ) for the 5 wt%doped pteg-2 film is within the conductivity range of various alkali metal doped C60 films (4 S cm -1 for Cs-doped C60, 10 S cm -1 for Li-doped C60, 20 S cm -1 for Na-doped C60, and 500 S cm -1 for K-doped C60 film). (Nature, 1991, 350, 320) The K-doped C60 film displays metallic nature and becomes superconducting at Tc of 18 K. Note that the excellent conduction of alkali metal doped C60 films is most likely achieved at a much higher doping level than that of the doped PTEG-2 film: for example, the K-doped C60 film exhibits a low Seebeck coefficient of -11 μVK -1 and C60 molecule is triple charged (a doping level of 3) upon K-doping (Phys. Rev. Lett. 1992, 69, 3797) while the doped PTEG-2 film has an Seebeck coefficient of -223 μV K -1 and a doping level of ~0.12. Additionally, the doped PTEG-2 film with n-DMBI is more conductive than previously reported co-evaporated C60/n-DMBI blend film (with an electrical conductivity of 5.5 S cm -1 ) (JACS, 2012, 134, 3999). These results suggest that meticulously designed side chains could not only provide the fullerene derivative with much improved solution processability but also enable a comparable charge transport with that of C60 fullerene upon doping.
By following an empirical L-S relation for non-degenerate semiconductors reported by Snyder et al (APL Mater., 2015, 3, 041506), we estimated the Lorenz number of L=1.65×10-8 WΩK -2 for the doped PTEG-2 film. Then the electronic contribution to the thermal conductivity (KE) is calculated to be 0.004 Wm -1 K -1 , which is very small compared to the total thermal conductivity K. Therefore, we agree with the reviewer that the electrical conductivity is too low to impact the thermal conductivity of electrons KE and the total K (KE+KL) originates from the lattice thermal conductivity (KL).
We have changed the manuscript accordingly.

Reviewer #2:
The manuscript entails an experimental study of the thermoelectric performance of a molecularly doped fullerene derivative. A branched ethylene glycol side chain is found to result in a, for fullerenes, record electrical conductivity while preserving a low thermal conductivity. Overall, a for organic materials very high thermoelectric figure of merit is obtained; the value is among the highest values reported to date. The thermoelectric data are certainly very good, and should be published in a journal such as Nature Communications. However, the analysis of the data both in terms of (heat and electrical) transport physics and physical chemistry needs improvement. Therefore, I cannot recommend publication of the manuscript in its current form. I am sorry for not being more positive, but hope that my comments below will help the authors to further sharpen their manuscript. 1) My major concern is the proposition that the presented material is a phonon-glass electron-crystal. While this argument is feasible, the manuscript lacks proof that could substantiate such a claim. As the authors state, this insight would be a considerable advance for the field of organic semiconductors. At a minimum, the authors should present temperature dependent thermal and electrical conductivity data. That should allow to identify phonon-glass electron-crystal type behavior.

Response:
We thank the reviewer for stating that the thermoelectric data should be published in a journal such as Nature Communications.
The temperature dependent thermal conductivity and electrical conductivity can be found in Fig. 5A and 5C. Those data show that the thermal conductivity hardly depends on temperature, while the electrical conductivity shows a maximum around 120 o C.
After carefully considering the reviewer's point, we drop the claim that we fully realized an organic 'PGEC' in the present study, and instead claim that we are close to it.
Unfortunately, there is no clear, strict definition of an organic PGEC in the current literature. Even for an inorganic PGEC it is difficult to find a proper definition. Therefore, in the revised version of our manuscript, we propose the following definition of an organic PGEC:  the thermal conductivity reaches the amorphous limit (https://journals.aps.org/prb/abstract/10.1103/PhysRevB.46.6131) of this material;  the charge carrier mobility should reach the crystalline limit of this particular material.
In this manuscript we show that the thermal conductivity does indeed reach (it's actually even lower than) the amorphous limit. For the related fullerene derivative PCBM there are several reports in the literature that show that the pristine material (i.e. undoped) does indeed satisfy this part of the definition. Here we show that 1) PTEG-2 has a similarly low thermal conductivity, and 2) it is not affected by doping. In sum, doped PTEG-2 is a phonon glass.
Regarding the charge carrier mobility: Frankevich et al. (Frankevich, E.;Maruyama, Y.;Ogata, H. Chem. Phys. Lett. 1993, 214, 39.) have found a (time-of-flight) mobility of 0.5±0.2 cm 2 /Vs for singlecrystal C60 and Bao et al. (J. Am. Chem. Soc. 2012, 134, 2760−2765 reported a high field-effect transistor mobility of 5.2 ± 2.1 cm 2 /Vs for a needle-like single crystal. It stands to reason that the mobility of C60 is an upper limit to that of fullerene derivatives as the side-chains will dilute the conjugated part of the system. In our manuscript, we find a mobility of 1.2 cm 2 /Vs which is of the same order of magnitude as that of single crystal C60. In other words, PTEG-2 comes close to (if not actually fulfills) the mobility requirement.
We note that the conductivity (see Fig 5a) shows little temperature dependence. Above 120 o C the conductivity even shows a negative temperature dependence. Figure 5b (Seebeck coefficient) shows that the Seebeck coefficient does not change with temperature, suggesting that the number of charge carriers remains constant. Taken together, this suggests that the conductivity might even be metallic in the sense that it shows negative temperature dependence. However, we do not wish to claim that this is the case as that would require a more detailed study.
These properties indicate that the doped PTEG-2 film is very close to an organic 'PGEC' material. The corresponding changes have been made and marked in red in the main text.
2) the authors mention the doping efficiency repeatedly. I recommend that the authors define this term more carefully. Do the authors refer to a charge transfer event? Not each generated polaron will lead to a free charge. Ionization and dissociation of charges should be discussed.

Response:
We agree with the reviewer that not every generated polaron will lead to a free charge. The doping efficiency is defined as the free charge density to the number of introduced dopant molecules (marked in red on page 14 in the main text). We directly determined the free charge density of doped PTEG-2 films (annealed at different temperatures) by using Mott-Schottky analysis on ion-gel based metalinsulator-semiconductor devices, which is used to estimate the doping efficiency (47% for annealed at 120 o C and 60% for annealed at 150 o C). Based on the reviewer's comment, we added the electron paramagnetic resonance spectra of the pristine and doped PTEG-2 films annealed at different temperatures ( Figure S7c). We find that 40% more polarons are generated upon annealing at 150 o C instead of 120 o C. The annealing temperature does not seem to strongly influence the dissociation probability of polarons into free charges. Therefore, more polarons lead to more free charges and, hence, increased doping efficiency. These results are consistent with the improved mixing between the host and the dopant molecules as shown by the morphology study.
3) in the introduction the authors write that organic materials are not toxic. That is possibly true for the semiconductors. However, molecules like DMBI are highly reactive. The SDS provided by Sigma states: acute toxicity. I suggest that the authors reconsider their claim.

Response:
We thank the reviewer for pointing out the toxicity of the dopant. We changed the 'non-toxic' into 'abundant' in the introduction part.
4) line 121. The authors speculate that the fullerene material changes its "arrangement" upon annealing. The authors should provide evidence that no solvent is trapped in the films. Even chloroform may be trapped, despite a low boiling point, which could explain the observed decrease in film thickness.

Response:
Yes, it is very important to exclude any solvent trapping for dynamic ellipsometry measurement. We took care to do so when preparing the samples. After spin-coating various fullerene derivative based films, we kept the samples in high vacuum (<10 -6 mbar) for two days in order to remove any trapped solvent. This strategy has proven very effective to remove high-boiling-point additives in polymer solar cells, for example (J. Phys. Chem. C 2013, 117, 14920−14928). Additionally, we performed the same measurement for vacuum-treated pristine PTEG-2 film pre-annealed at 90 ℃ (higher than the boiling point of chloroform) for 1 h and the result is displayed in Figure S1f. the d-T plot for this pristine PTEG-2 film shows the same feature: inflection point at T=131 ℃, and the film shrinks in the 131 ℃ <T<155 ℃ range. Regarding this, we believe that the variation in thickness is attributed to the arrangement of fullerene derivatives rather than the removal of any trapped solvent molecules.
5) The film thickness during annealing is extracted from ellipsometry measurements. Detailed information about the model should be provided that was used to convert the ellipsometric angles into film thickness. How were the raw data fitted and how was the accuracy of the fit determined? At minimum, a least square analysis of the fit should be provided. And what range of wavelengths were fitted?

Response:
More detailed information about the model and the fitting were added in the Supplementary Information. The fitting parameters and the mean squared error (MSE) of fits are summarized in Table S1. We used the model, that is silicon substrate/silicon oxide/Cauchy layer, for fitting the spectroscopic data of all the fullerene derivative based samples. The thickness of the native oxide layer was determined by fitting the ellipsometry data of a blank silicon substrate with a native oxide layer. The Cauchy dispersion function (n(λ)=An+Bn/λ+Cn/ λ, k=0, where n is the refractive index, k extinction coefficient, An, Bn and Cn are Cauchy parameters) was used to fit the optical property of the fullerene derivative based layers at various wavelengths. The ellipsometry data were fitted in the wavelength range from 700 nm to 1700 nm, in which the fullerene derivative based films show no light absorption. Figures S1a-e shows the experimental data and the corresponding model fits, which show very good overlap with each other. From the Figure S1a-e and the small MSE values, we think the experimental data were well fitted with a high accuracy.
6) The authors conclude from their measurements that the fullerene materials undergo a phase transition, and propose that side chain melting occurs. Line 171: "T_inf can be physically interpreted as the melting point of the side chains". I am not quite sure what a "side chain melting" is supposed to mean. In any case, a melting event would lead to a sudden decrease in density, and therefore an increase in thickness. The authors observe a decrease in thickness though. In my opinion, the suggested thermal behavior is pure speculation and must be substantiated with other types of measurements.

Response:
We thank the reviewer for pointing out this. We tried to express that at Tinf the side-chains become so flexible that the molecular arrangement of entire PTEG-2 molecules can change, which leads to a more densely packed film. We were, however, not suggesting that the entire compound would melt, which would indeed lead to a sudden decrease in density. However, a proper understanding of the exact roles of the side-chains is beyond the scope of this manuscript and we have removed the statements relating to the melting of the side-chains from the main text. 7) related to 6), the comment on line 174: "melting…provide[s] energy for rearrangement" makes no sense. How does melting provide energy? Do the authors mean the heat that is released by a melting event? Please measure that heat.

Response:
We measured the DSC for various fullerene derivative powder samples. Unfortunately, we did not observe apparent melting behavior of the side chains as reported by a previous study on a glycolated polythiophene (Li et al. Org. Electron. 2016, 33, 23). One possible reason might be the low weight percentage of the side chain in fullerene derivatives, as compared to that in the glycolated polythiophene. So, we deleted the part "Linear ethylene-glycol-type side chains of three units have been reported to melt at 98 o C. 35 The side chains of PPEG-1 and PTEG-2 are longer or more bulky and should therefore melt above 98 o C. The Tinf values of the fullerene derivatives vary with the side chains in the order PPEG-1< PTEG-2 < F2A, which correlates well with the melting points of these side chain types. Thus, we surmise that Tinf can be physically interpreted as the melting point of the side chain of a fullerene derivative. This speculation is reasonable because the melting of a side chain should increase the spatial freedom of fullerene-based molecules and provide energy for rearrangement into more energetically favorable positions." 8) line 171. The authors insinuate that the polarity of side chains improves the thermal stability. Only one of the polar side chains (those of PTEG-2) show good thermal stability. Surely, it is not the polarity but some other parameter that determines the stability. That aside, I agree with the authors that their observation of a thermally stable doped system is significant.

Response:
We thank the reviewer for pointing out this. We measured the thermal stability of the doped F2A with the alkyl side chain under the same condition (at 150 o C) and the result is displayed in Figure 2b together with those for other fullerene derivatives (PTEG-1, PPEG-1, and PTEG-2). Of these materials, the ones with polar side chains are more thermally stable than the one with an alkyl side chain (F2A), which agrees with previous reports. Therefore, the polarity of the side chain appears to play some role in the thermal stability of the doped films. We do not exclude the effect of other parameters like the geometry of the side chain. We have changed the text accordingly. 9) line 200. "aggregates gradually disappear. How is the presence and disappearance of aggregates included in the ellipsometric model that was used to determine the film thickness. Did the authors assume a multilayer stack? If the model does not include at least a bilayer (the top layer represents the aggregates) that disappears upon annealing, then the observed changes in thickness could easily be an artefact. I must thus question the ellipsometry analysis.

Response:
From the dynamic ellipsometry measurement, we obtained a plot of the ellipsometry parameter (ψ) as a function of the temperature. The ellipsometry parameter ψ (T) is related to contribution of the entire organic film, which is one layer for the un-doped film or bilayer for the doped one. As the organic systems do not undergo decomposition during the annealing process, the weight of the sample is constant, and thus the thickness variation corresponds to the volume change with the temperature, which inversely scales with the density variation of the film. Therefore, the thickness should be a parameter translated by the volume of the organic material excluding the void part. For this reason, we used one Cauchy layer to describe the organic film and extract the d(T) plots for various pristine and doped fullerene derivatives in Figure 2a.

We also used a bilayer model for the doped PTEG-2 film and the bilayer includes a Cauchy layer and an atop composite layer consisting of the Cauchy material and void with a certain percentage (f). The thickness (d2) of the composite layer was considered to be the height (30 nm-70 nm) of the aggregates measured by the AFM and is set to 50 nm in the model. By fitting the experimental data, the thickness (d1) of the Cauchy layer and the percentage of the void were
determined to be 89.7 nm and 96.4%, respectively. Therefore, the volume-translated thickness (d) of organic material could be expressed by: d=d2+d1 • (1-f) and we obtain d=91.5 nm, which is very close to d=91.6 fitted by the one layer model. Therefore, we think the one layer model provides a good approximation for the volume-translated thickness for the doped PTEG-2. In addition, the valley in the d(T) curve for the doped PTEG-2 film was also observed in the d(T) plot for the pristine PTEG-2 film that has no aggregates on the surface. As such, we believe the inflection point in the d(T) plot corresponds to the phase transition of the organic material. Figure 3 and paragraph starting at line 235. The MD simulations are very illustrative. I would like to see a simulation of the diffraction pattern in Fig 3b, which substantiates the proposed unit cell.

We computed (procedure in the methods and SI of the revised manuscript) the diffraction patterns for the MD unit cells along the qz and qy directions which can be compare to the experimental ones-see updated Figures 3c and 3d:
The agreement with the experimental scattering line cuts (Fig 3c and 3d, respectively) is very good. We note here that the calculated scattering curves come from averaging over 240 perfectly oriented 100nm sized crystals, while this is not the case for the experimental scattering signals, where some "misaligned" crystals can: 1) broaden the peaks; 2) introduce some of the peaks which would belong to the qz direction into the qy scattering signal and vice versa. We think that this is the reason why, for example, the peak observed in the low-q region along the qy (Fig 3d.) is not present in the simulated scattering pattern, that peak being likely coming from a misaligned (i.e., rotated by 90 degree) PTEG-2 crystallite. 11) line 249. Now the ethylene glycol side chains are described as having different configurations. How is this consistent with the picture provided earlier where these side chains order at low temperatures, and upon heating?

Response:
We acknowledge that the description in the text was perhaps misleading and we rephrased now the text to (previous line 245-250):

common feature of the multiple configurations obtained from the MD simulations is a staggered arrangement for the C60 bilayers interposed by the ethylene glycol phase. However, the MD simulations do not converge all into one specific configuration of the ethylene glycol chains, but rather give an ensemble of similar ones. The convergence into a single ethylene glycol configuration is likely prevented by the fact that the ethylene glycol chains are quite flexible due to the low energy barriers between their different configurations.'
12) line 272. A doping efficiency of 60% is calculated. Are these polarons formed per dopant molecule or free charges? And how does this value compare to the proposed disappearance of aggregates (line 200), which leads to a higher conductivity. Does the doping efficiency increase upon annealing? The authors quote carrier densities at 120 and 150 deg but only one doping efficiency. Is this because the doping efficiency does not actually increase, which would be inconsistent with the earlier argument that more dopant is taken up by the fullerene material?

Response:
The doping efficiency is defined by the ratio of the free charge carrier density to the number of the dopant molecules. We used the ion-gel based metal-insulator-semiconductor (MIS) devices along with Mott-Schottky analysis to extract the free charge carrier density. The effectiveness of this approach has been proven in our previous work (Adv. Mater. 2018, 30, 1704630). The doping efficiency of doped PTEG-2 film is calculated to be around 47% after annealing at 120 ℃ and 60% after annealing at 150 ℃.
The increased doping efficiency at the higher annealing temperature is likely to result from the improved mixing between the host and the dopant molecules. This was further substantiated by the EPR measurement that indicates more polarons are generated upon annealing at 150 o C ( Figure S7c). We changed the text accordingly.
13) line 298. Now the side chain melting occurs at 120 deg. On line 169 it is 98 deg.

Response:
In the original version of the manuscript, On line 169, 98 ℃ is the melting point of the linear triethylene glycol type side chain, reported in the Ref. 35 (Li et al. Org. Electron. 2016, 33, 23). On line 298, 120 ℃ is assumed to be the melting point of double triethylene glycol type side chain, which is more bulky than the linear triethylene glycol type side chain. However, as the reviewer pointed out, we lack sufficient evidence to support this point. Therefore, we deleted the corresponding part discussing whether or not the melting of the side chain initiated the phase transition.
2) line 259 and 267. Provide error bars for the Seebeck coefficient and carrier densities.

Response:
The error margins for the Seebeck coefficient and the carrier densities are provided in the main text.
3) The doped fullerene material is not stable under ambient conditions. All electrical characterization was done under nitrogen. But how about the thermal conductivity measurement. How did the authors ensure that the sample had not degraded? The Linseis system requires that the sample is handled in air. Was the instrument placed in a glove box?

Response:
We agree with the reviewer that most n-doped organic semiconductors including doped fullerene materials are not stable under ambient conditions. Importantly, the main doping process of fullerene derivative/n-DMBI system does not occur directly by blending with a solution process and requires postannealing above 90 ℃ to be activated. The as-cast PTEG-2/n-DMBI blend (5 wt% n-DMBI) film exhibits an electrical conductivity of 2.6×10 -4 S cm -1 . We found that air exposure of as-cast fullerene derivative films has little influence on the eventual electrical conductivity if the as-cast fullerene derivatives are post-annealed in an inert environment or vacuum. This applies to most fullerene derivatives such as PCBM, PTEG-1, PPEG-1, and PTEG-2, which are doped by n-DMBI. Air exposure of 5 wt%-doped PTEG-2 film for 1 h causes a deviation of the eventual conductivity less than 10%. Therefore, before the thermal conductivity measurement, we transferred the as-cast PTEG-2/n-DMBI blend film into the Linseis setup and carried out the in situ thermal annealing at 150 ℃ in vacuum to complete the doping process. The thermal conductivity was also measured by the IIT group. They mounted the samples into an air-tight chamber with optical windows in a N2-filled glovebox. Before thermal conductivity measurement, the air-tight sample chamber was pumped down with a vacuum of <10 -5 torr. There was no air exposure during the measurement.

Reviewer #3:
Liu et al. claimed that they developed OTE materials with ZT value over 0.3 following PGEC concept. In fact, this concept is widely used in inorganic materials, but its applications in OTE materials remain unclear. Although the basic concept seems interesting, I have the follow concerns about the results.
(1) For typical PGEC concept in inorganic materials, the designed atoms are introduced into crystal cell to ensure high electrical conductivity and suppress the lattice thermal conductivity. Although the authors claimed their designed materials is similar the concept, I failed to find a clear definition of the concept in organic materials. Notably, several doped organic semiconductors have been confirmed to possess ordered molecular packing and the dopant is located at side-chain regime. Once again it raises a question, how could define PGEC materials in conjugated systems? Whether these previous reported materials also follow the concept?

Response:
We agree with the reviewer that there is no clear definition of the PGEC concept as applied to organics. In the new version of our manuscript, we have attempted to introduce a suitable definition:  the thermal conductivity reaches the amorphous limit (https: //journals.aps.org/prb/abstract/10.1103/PhysRevB.46.6131)

of this material;
 the charge carrier mobility should reach the crystalline limit of this particular material.
In this manuscript we show that the thermal conductivity does indeed reach (it's actually even lower than) the amorphous limit. For the related fullerene derivative PCBM there are several reports in the literature that show that the pristine material (i.e. undoped) does indeed satisfy this part of the definition. Here we show that 1) PTEG-2 has a similarly low thermal conductivity, and 2) it is not affected by doping. In sum, doped PTEG-2 is a phonon glass.
Regarding the charge carrier mobility: Frankevich et al. (Frankevich, E.;Maruyama, Y.;Ogata, H. Chem. Phys. Lett. 1993, 214, 39.) have found a (time-of-flight) mobility of 0.5±0.2 cm 2 /Vs for singlecrystal C60 and Bao et al. (J. Am. Chem. Soc. 2012, 134, 2760−2765 reported a high field-effect transistor mobility of 5.2 ± 2.1 cm 2 /Vs for a needle-like single crystal. It stands to reason that the mobility of C60 is an upper limit to that of fullerene derivatives as the side-chains will dilute the conjugated part of the system. In our manuscript, we find a mobility of 1.2 cm 2 /Vs which is on the same order of magnitude as that of single crystal C60. In other words, PTEG-2 comes close to (if not actually fulfills) the mobility requirement.
All in all, we have revised our claim by stating that our material system approaches the PGEC concept.
A few doped organic semiconductors have been confirmed to possess ordered molecular packing and the dopant is located at the side-chain regime. We are not sure whether or not they can be qualified as 'organic electron crystals' only on the basis of the microstructure, as a proper PGEC also has very good electrical transport properties (see definition above).
(2) The ultralow thermal conductivity and high ZT are the key evidences for the so-called PGEC concept. I have several concerns about measurement of the κ and ZT values: i) As a well-known fact, it is challenge to measure sample with low in-plane thermal conductivity (κ ￡0.1 W/m K). Even with the technique provided in the supporting information, I'm not sure about the accuracy of results. A prefered thickness of the sample should be > 400 nm for a sample with κ of 0.4 W/m K (Journal of Electronic Materials 2018, 47 (6) , 3203 3209. DOI: 10.1007/s11664-017-5989-4). I believe that the situation is very challenge for PTEG-2 film since much thicker film is required. The author should provide the exact thickness of the sample used in TFA measurement and evaluate the deviation.

Response:
Yes, it is really a challenge to measure the thermal conductivity of the fullerene derivative materials. We measured the thermal conductivity of the pristine and doped PTEG-2 films which are several micrometers thick, for several times and similar results are obtained. For the result shown in Figure 5c, the thickness is 6.71 ± 0.23 μm for the doped PTEG-2 film and 3.71± 0.35 μm for the pristine PTGE-2 film. The obtained thermal conductivity values are in agreement with the reported (by two different groups) literature values for PCBM (0.03 Wm -1 K -1 Phys. Rev. Lett. 2013, 110, 015902 and0.06 Wm -1 K -1 Phys. Rev. B. 2013, 88, 075310), which supports our findings.
ii) I note the measurement of σ and S were performed in a N2-controlled environment while κ∥ was obtained in vacuum by TFA. What about the air stability of the sample during the sample transfer? Such kind of n-type materials usually possess poor air stability, which will lead to rapid de-doping and lower κ∥ because of reduced contribution from electron transport. In fact, TFA can characterize all the parameters and even temperature dependent performance in one chip. Why the author prefers to utilize different techniques to characterize the samples?

Response:
We agree with the reviewer that moderately/heavily doped organic semiconductors are not stable under ambient conditions. The main doping process within fullerene derivative/n-DMBI films does not occur directly after the spin-coating process and requires post-annealing above 90 ℃ to be activated. The ascast PTEG-2/n-DMBI blend (5 wt% n-DMBI) film exhibits an electrical conductivity of 2.6×10 -4 S cm -1 . We found that air exposure of as-cast fullerene derivative films has little influence on the eventual electrical conductivity if the as-cast fullerene derivative films are post-annealed in an inert environment or vacuum. This applies to most fullerene derivatives such as PCBM, PTEG-1, PPEG-1, and PTEG-2, which are doped by n-DMBI. Air exposure of as-cast 5 wt%-doped PTEG-2 film for 1 h causes a less than 10% deviation of the eventual conductivity upon annealing at 150 ℃ for 1 h. Therefore, before thermal conductivity measurement, we transferred the as-cast PTEG-2/n-DMBI blend film into the Linseis TFA setup and carried out the thermal annealing at 150 ℃ i n vacuum to complete the doping process. In this way, we avoid to underestimate thermal conductivity because of the electronic contribution loss. The thermal conductivity was also measured by the IIT group. They mounted the samples into an air-tight chamber with optical windows in the N2-filled glovebox. Before thermal conductivity measurement, the air-tight sample chamber was pumped down with a vacuum of <10 -5 torr. There was no air exposure during the measurement.
As for the thermoelectric characterization, we performed all the electrical conductivity and Seebeck coefficient measurement in the N2-filled glove-box in our lab in the University of Groningen. The data shown in this work have been reproduced for several times. Furthermore, we would like to keep this work consistent with our previous reports, where the same method was used to measure the electrical conductivity and Seebeck coefficient. However, we have no facility to measure the thermal conductivity of the organic film. Therefore, we collaborated with Prof. Derya Baran's group at KAUST for the thermal conductivity measurement of the pristine and doped PTEG-2. They are experts in studying thermal transport of thin film samples (Advanced Science, 2020, 1903389). In order to make sure the sample quality, they checked the electrical conductivity of the doped film, which is >8 S cm -1 very close to the value obtained at the University of Groningen. The measurement of the thermal conductivity has been repeated for several times and consistent results were obtained, which is displayed in the main text.
iii) It is difficult to understand that the σ shows a large variation versus temperature (>30%) while the κ∥ almost keeps nearly unchanged in Figure 5. The electronic contribution of the κ is related to the σ and follows Wiedemann-Franz law. In other words, one should observe the temperature dependent κ even for PGEC materials. The authors should identify the electronic contribution and lattice contribution to the κ to confirm and understand the phenomenon.

Response:
We would like to first estimate the electronic and lattice contribution to the thermal conductivity of the doped PTEG-2 film. By applying an empirical L-S relation for non-degenerate semiconductors reported by Snyder et al (APL Mater., 2015, 3, 041506) to this study, we estimated the Lorenz number of L=1.65×10-8 WΩK -2 for the doped PTEG-2 film. Based on the Wiedemann-Franz law, the electronic contribution to the thermal conductivity (KE) is calculated to be 0.004 Wm -1 K -1 , which is very small compared to the total thermal conductivity K. The thermal conductivity of the doped PTEG-2 film is dominated by the lattice contribution.
3) The authors found that the electrical conductivity first rises then falls along with the increase temperature. They attribute the decreased σ to the enhanced disorder of the side chain. Providing the energetic disorder before and after phase transition will make it clearer. In fact, dedoping can also contribute to the decrease of the conductivity. What about the thermal stability of doped films?

Response:
The doped PTEG-2 film is very stable under the thermal stress as shown in Figure 2b. Actually, after cooling the sample down to room temperature, we measured the electrical conductivity of σ= 8.0 S cm -1 at 25 ℃ (the red star in Figure 5a), which is very close to its original value. It indicates that the decrease of electrical conductivity with temperature above 120 ℃ is a reversible process and was not caused by the de-doping of the sample. We speculate that, at higher temperatures, molecular vibrations hamper charge transport, somewhat similar to phonon scattering in an inorganic semiconductor. We made corresponding changes in the main text, which was marked in red.
4) The doped PTEG-2 film shows high crystallinity. Since the author use 'electron crystal' to describe the materials, I suggest the authors to evaluate of the carrier concentration and mobility by TFA to support the definition.