Review Reports - A Novel Thiosemicarbazide-Based Fluorescent Chemosensor for Hypochlorite in Near-Perfect Aqueous Solution and Zebrafish

Round 1

Reviewer 1 Report

I would reccomend the paper after major revisions - please see below.

Formal issues>

replace hyphen with minus sign, e.g. in ClO-, or concentrations
improve English
use multiplication sign ‘×’ instead of the letter 'x'

Sections 2.2, 2.3>

Are the syntheses of both compounds new? If they are new, add at least one additional analysis, e.g. combustion analysis or high-resolution MS. If not, add references.
HR-MS for AFC is wrong. It should be probably [C20H17N5OS + H⁺ + DMSO]⁺ instead of [C20H17N5OS - H⁺ + DMSO]⁺.
How is it possible to have DMSO adduct in MS if you did not use DMSO during the synthesis? If you would use DMSO-d₆ from the NMR analysis, then it should be adduct with DMSO- d₆ instead of DMSO.
add ¹³C NMR for FHC
add MS for FHC
round ¹³C NMR shifts to one decimal place

Section 2.4 >

Which source of NaClO did you use for the preparation of your solutions? A less common option is NaClO 5H₂O (if stored at low temperatures) or a solution.
How do you know that AFC contained 0.33% DMSO when you used a DMSO solution (according to 2.4. General procedures)?

Sections 2.6, 3.3. >

Pople basis sets, e.g. 6-31g(d,p) are not generally recommended for any type of calculations due to the presence of 'sp' orbitals instead of 's' and 'p' orbitals, please repeat the calculation with more recent and better-ballanced basis set, e.g. cc-pvdz (or higher) or def2-TZVP. I would also recommend other DFT functional, e.g. BMK or CAM-B3LYP.
I do not think the TD-DFT results suggest by any means that ClO(-) first hydrolyse C=N bond and then oxidise N from the acridine. If so, please explain. I did not see any energy barriers, which could suggest it (thermodynamic explanation). If it would be a matter of any kinetic effect that the oxidation is faster, then, the explanation based on theory would be even more difficult.
Although dihedral angle could be defined as it is, it is not commonly used in this way and usually, it is defined for atoms connected by bonds and next to each other, for example here, it would be 'unlabelled carbon next to N1' -- 2C – ‘unlabelled carbon between 2C and 4N' -- 4N. Moreover, I think it would be better to have both angles positive which can be easily done by changing the definition of the angle as the molecule is symmetric (swap 2C and 3C). Then, the angle would be 178.250 deg.
To get better results, one should use non-equilibria solvation for TD-DFT calculations. Maybe it was used, maybe was not – please describe how did you get your results.

Sections 3.1. >

what is the time stability of AFC under the same conditions without ClO(-) ion? I mean the same pH, concentration, and other ion presence (probably chloride and ions from the buffer), etc. It could bring some light into the mechanism. Personally, I would expect that there could be oxidation of acridine nitrogen caused by ClO(-) and the rest is simple hydrolysis due to pH of the environment. Also, ClO(-) could oxidise also the other moieties in the molecule.
In the UV-Vis figures (Fig 2., Fig 3), highlight and label the curve with the highest concentration of ClO(-) and zero concentration of ClO(-).
Don't use abbreviation 'Flu. Intensity at 455 nm (a.u.)', use only 'Intensity (a.u.)' and explain the rest in the Figure caption (Fig 5., Fig 6).
Correct the phase of the NMR spectra in Figure 4. what is the solvent used for NMR titration in Figure 4? There could be also the effect of changing H to D caused by the NMR solvent.
Explain why proton labelled as 5 disappeared after addition of 0.5 eq of ClO(-).
what is ‘MJ’ abbreviation in Figure 5?

Author Response

Question 1: replace hyphen with minus sign, e.g. in ClO-, or concentrations

Answer: As the reviewer 1 suggested, we replaced hyphen with minus sign.

Question 2: Improve English

Answer: We looked carefully into the whole text with a grammar-check program.

Question3: use multiplication sign ‘×’ instead of the letter 'x'

Answer: We changed the letter ‘x’ to multiplication sign ‘×’.

Question 4: Are the syntheses of both compounds new? If they are new, add at least one additional analysis, e.g. combustion analysis or high-resolution MS. If not, add references.

Answer: We added the reference for FHC in the section 2.2 in page 3, because it was already known. The sensor AFC is new. Therefore, we carried out its elemental analysis and added the results in the section 2.3 in page 3.

Question 5: HR-MS for AFC is wrong. It should be probably [C20H17N5OS + H+ + DMSO]+ instead of [C20H17N5OS - H+ + DMSO]+.

Answer: We really thank the reviewer 1. We corrected it.

Question 6: How is it possible to have DMSO adduct in MS if you did not use DMSO during the synthesis? If you would use DMSO-d6 from the NMR analysis, then it should be adduct with DMSO- d6 instead of DMSO.

Answer: DMSO was used to dissolve AFC for the preparation of the MS sample.

Question 7: add 13C NMR for FHC

Answer: We added the reference in the section 2.2 in page 3, because FHC is a known compound.

Question 8: add MS for FHC

Answer: We added the reference in the section 2.2 in page 3, because FHC is a known compound.

Question 9: round 13C NMR shifts to one decimal place

Answer: As the reviewer 1 suggested, we rounded ¹³C NMR shifts to one decimal point in the section 2.3 in page 3.

Question 10: Which source of NaClO did you use for the preparation of your solutions? A less common option is NaClO 5H2O (if stored at low temperatures) or a solution.

Answer: We used a 11 % NaClO solution. We described the source of NaClO in the section 2.4 in page 3.

Question 11: How do you know that AFC contained 0.33% DMSO when you used a DMSO solution (according to 2.4. General procedures)?

Answer: We rewrote the part of AFC solution for the easy understanding in the section 2.5 in page 4.

Question 12: Pople basis sets, e.g. 6-31g(d,p) are not generally recommended for any type of calculations due to the presence of 'sp' orbitals instead of 's' and 'p' orbitals, please repeat the calculation with more recent and better-ballanced basis set, e.g. cc-pvdz (or higher) or def2-TZVP. I would also recommend other DFT functional, e.g. BMK or CAM-B3LYP.

Answer: As the reviewer 1 suggested, we performed additional DFT/TD-DFT calculations for AFC with BMK and CAM-B3LYP level. As shown in Table 1 (see below), the theoretical excitation energy of AFC based on B3LYP level showed the most consistent wavelength with experimental UV-vis (9 nm deviation) compared to the result of BMK and CAM-B3LYP (43 and 46 nm deviation). Thus, we chose B3LYP functionals as the suitable methods for the following DFT and TDDFT calculations.

We carried out DFT/TD-DFT calculations for AFC with cc-pVDZ and def2-TZVP basis set in B3LYP level, and found that there were only 3 nm (cc-pVDZ) and 1 nm (def2-TZVP) differences from the experimental UV-vis in wavelength (excited state 1) as shown in Table 2 (see below). Additionally, we also re-calculated AAD-O with cc-pVDZ and def2-TZVP in B3LYP level, and found that there were 27 nm and 40 nm differences, respectively, from the experimental UV-vis as shown in Table 3 (see below).

Although TD-DFT results of AFC with def2-TZVP showed the most consistent wavelength with the experimental UV-vis, those of AFC and AAD-O with 6-31g(d,p) was much more suitable wavelength compared to the results with cc-pVDZ and def2-TZVP. Therefore, we performed all DFT/TD-DFT calculations based on B3LYP/6-31g(d,p) due to its efficiency and accuracy.

Table 1. The first excited state of AFC with 6-31g(d,p) basis set in various DFT functionals.

AFC	Excited state 1	Wavelength (nm)	Percent (%)	Oscillator strength
B3LYP	HOMO → LUMO	390.55 (-9)†	98	0.1768
BMK	HOMO → LUMO	356.10 (-43)	98	0.2288
CAM-B3LYP	HOMO → LUMO	353.67 (-46)	98	0.2329

† The difference of wavelength between the theoretical excitation energy and the experimental UV-vis of AFC (400 nm).

Table 2. The first excited state of AFC with various basis sets in B3LYP level.

AFC	Excited state 1	Wavelength (nm)	Percent (%)	Oscillator strength
6-31g(d,p)	HOMO → LUMO	390.55 (-9)†	98	0.1768
cc-pVDZ	HOMO → LUMO	397.50 (-3)	98	0.2288
def2-TZVP	HOMO → LUMO	398.52 (-1)	98	0.1588

† The difference of wavelength between the theoretical excitation energy and the experimental UV-vis of AFC (400 nm).

Table 3. The first excited state of AAD-O with various basis sets in B3LYP level.

AAD-O	Excited state 1	Wavelength (nm)	Percent (%)	Oscillator strength
6-31g(d,p)	HOMO → LUMO	499.39 (-1)†	99	0.1054
cc- pVDZ	HOMO → LUMO	527.36 (+27)	99	0.0699
def2-TZVP	HOMO → LUMO	540.14 (+40)	99	0.072

† The difference of wavelength between the theoretical excitation energy and the experimental UV-vis of AAD-O (500 nm).

Question 13: I do not think the TD-DFT results suggest by any means that ClO(-) first hydrolyse C=N bond and then oxidise N from the acridine. If so, please explain. I did not see any energy barriers, which could suggest it (thermodynamic explanation). If it would be a matter of any kinetic effect that the oxidation is faster, then, the explanation based on theory would be even more difficult.

Answer: Yes, we agreed to the reviewer 1’ suggestion. Therefore, we just proposed the cleavage of AFC and further oxidation of the resultant AAD by ClO^-, based on ESI-mass, ¹H NMR titration and TD-DFT calculations. We re-wrote this proposal in the second paragraph of page 14.

Question 14: Although dihedral angle could be defined as it is, it is not commonly used in this way and usually, it is defined for atoms connected by bonds and next to each other, for example here, it would be 'unlabeled carbon next to N1' -- 2C – ‘unlabeled carbon between 2C and 4N' -- 4N. Moreover, I think it would be better to have both angles positive which can be easily done by changing the definition of the angle as the molecule is symmetric (swap 2C and 3C). Then, the angle would be 178.250 deg.

Answer: As the reviewer 1 suggested, we revised dihedral angles by using connected atoms, and angles to be positive in Fig. 8 in page 14.

Question 15: To get better results, one should use non-equilibria solvation for TD-DFT calculations. Maybe it was used, maybe was not – please describe how did you get your results.

Answer: As the reviewer 1 suggested, we used non-equilibria solvation for TD-DFT calculations of AFC and AAD-O (see below).

Table 4. The first excitation energies of AFC and AAD-O applied to the equilibrium or the non-equilibrium solvation based on the TD-DFT calculations (B3LYP/6-31g(d,p)).

	Absorption (nm)
	Vertical excitation (equilibria solvation)	External iteration (non-equilibria solvation)
AFC	390.55	386.21
AAD-O	499.39	479.85

Question 16: what is the time stability of AFC under the same conditions without ClO(-) ion? I mean the same pH, concentration, and other ion presence (probably chloride and ions from the buffer), etc. It could bring some light into the mechanism. Personally, I would expect that there could be oxidation of acridine nitrogen caused by ClO(-) and the rest is simple hydrolysis due to pH of the environment. Also, ClO(-) could oxidise also the other moieties in the molecule.

Answer: We checked the time-dependent UV-vis change (400 nm) of AFC (10 μM) with/without ClO^- as shown in Fig. S2. Under the same conditions without ClO(-), AFC was stable enough for 1 h. We described this result in the last paragraph of page 6.

Question 17: In the UV-Vis figures (Fig 2., Fig 3), highlight and label the curve with the highest concentration of ClO(-) and zero concentration of ClO(-).

Answer: As the reviewer 1 suggested, we highlighted and labeled the curves with the highest concentration of ClO^-and zero concentration of ClO^-.

Question 18: Don't use abbreviation 'Flu. Intensity at 455 nm (a.u.)', use only 'Intensity (a.u.)' and explain the rest in the Figure caption (Fig 5., Fig 6).

Answer: As the reviewer 1 suggested, we changed 'Flu. Intensity at 455 nm (a.u.)' to ‘Intensity (a.u.)’ and explained the rest in the Figure captions (Figs. 5 and 6).

Question 19: Correct the phase of the NMR spectra in Figure 4. what is the solvent used for NMR titration in Figure 4? There could be also the effect of changing H to D caused by the NMR solvent.

Answer: As the reviewer 1 suggested, we corrected the phase of the NMR spectra in Figure 4. We used DMSO-d₆ as a solvent and mentioned it in the Figure caption.

Question 20: Explain why proton labelled as 5 disappeared after addition of 0.5 eq of ClO(-).

Answer: After addition of 0.5 eq of ClO^-, the proton 5 did not disappear. It just became smaller because of the cleavage of the imine bond.

Question 21: what is ‘MJ’ abbreviation in Figure 5?

Answer: We are really sorry to make a mistake. We corrected MJ to AFC.

Reviewer 2 Report

This manuscript details the preparation of a new fluorophore based on acridine. The authors were able to modify the fluorophore to make it water-soluble, and this useful for cellular and other in vivo type of imaging. The fluorophore is a "turn-off" sensor that is sensitive to the reactive oxygen species ClO^- and relatively insensitive to many other analytes. Overall, this is a well-written article detailing clear and concise preparation steps, evaluation of the fluorophore as a sensor, and real-world demonstration of the sensor. It is recognized that English is not the authors' first language. Having said that, the article is well written in that regards, though one more editing for English might be desirable (though not required). This reviewer suggests publication of this article after some minor revisions and clarifications. These are the points that should be addressed:

Page 2, line 47: ESI-MS should be defined. This would be those readers who may not be familiar with that type of mass spectrometry.
Page 2, Line 52-53: Could the authors' please provide the Varion model of the NMR as well as the magnet power and frequency?
Page 2, Section 2.2: There is no issues with the description of the synthetic procedures. However, could the authors provide the expected yield (in %) for the synthesis of both compounds (FHC and AFC, respectively).
Page 4, Line 115: The authors state that the detection limit is 58.7 µM. How does this compare with other sensors of this type? If possible, it would help to show this type of comparison.
Page 8, Figure 5: Could the authors provide a bit of clarity when describing this figure? It seems that this is a comparison of samples with just the fluorophore (labelled MJ), and samples with only ClO^-, and then samples with both ClO^- and other analytes in equal concentrations. The purpose is to show that the fluorophore is selective to ClO^-. Is this correct? A better description of the figure in the text would be warranted. Also, could the authors define MJ?

Author Response

Question 1: Page 2, line 47: ESI-MS should be defined. This would be those readers who may not be familiar with that type of mass spectrometry.

Answer: As the reviewer 2 suggested, we defined ESI-MS in the last paragraph of page 2.

Question 2: Page 2, Line 52-53: Could the authors' please provide the Varion model of the NMR as well as the magnet power and frequency?

Answer: We provided the Varion model of the NMR as well as the magnet power and frequency in the section 2.1 in page 3.

Question 3: Page 2, Section 2.2: There is no issues with the description of the synthetic procedures. However, could the authors provide the expected yield (in %) for the synthesis of both compounds (FHC and AFC, respectively).

Answer: We provided the yields of FHC and AFC into the sections 2.1 and 2.2 in page 3.

Question 4: Page 4, Line 115: The authors state that the detection limit is 58.7 µM. How does this compare with other sensors of this type? If possible, it would help to show this type of comparison.

Answer: As the reviewer 2 suggested, we made Table S1 for a comparison of the detection limit of AFC with other sensors.

Question 5: Page 8, Figure 5: Could the authors provide a bit of clarity when describing this figure? It seems that this is a comparison of samples with just the fluorophore (labelled MJ), and samples with only ClO-, and then samples with both ClO- and other analytes in equal concentrations. The purpose is to show that the fluorophore is selective to ClO-. Is this correct? A better description of the figure in the text would be warranted. Also, could the authors define MJ?

Answer: We are really sorry to make a mistake. We corrected MJ to AFC.

Round 2

Reviewer 1 Report

Recommendation: Major revision still required

Message to authors:

The authors replied to all questions and improved the manuscript, however, I still found several issued, which are not explained. Especially, the TD-DFT part was performed in not recommended fashion – ‘we need the best match but it does not matter if it is physically correct’. Moreover, the predicted spectra are not matching well the theoretical as several peaks are missing, which makes the DFT part quite unreliable.

The second issue is NMR titration – I personally feel that more experiments are required. It would be nice to compare the results to pure compound, e.g. aminoacridine, aminoacridine-N-oxide, etc.

Q10 In section 2.4> I would say 'aqueous solution', not 'dissolved in H2O' in the case of NaClO.

Q12 I disagree with the authors. If you compare different DFT functionals with not recommended basis set (6-31g(d,p), you may get wrong results and due to happy coincidence that some errors were cancelled by another errors, B3LYP and 6-31g(d,p) gave you the best match. Later, you use B3LYP DFT functional with better-ballanced basis sets than 6-31g(d,p) and due to the same reason, you might get wrong results.

Moreover, I would not say you have the best match - if I compare the theoretical and experimental spectrum of AFC, there are some missing peaks about 400 nm - you have only one calculated, but in experimental are three maxima. The same is true for ADD-O (Figure S4, S6) where are missing peaks.

Q13 Could you please suggest the mechanism how is cleaved the imine bond by ClO-? ClO- is usually used for oxidations, not for hydrolysis.

Q14 I do not think the H-NMR spectrum in Figure 4 is explained. Clearly visible proton 5 of AFC (brown line), after addition of 0.5 eq of ClO-, the proton 5 disappeared (authors said the intensity is lower, but I cannot find that proton in the second spectrum even if zoomed, dark yellow colour). does it mean that 0.5 eq of ClO- hydrolyse 1 eq of AFC? or oxidise N of acridine? later, green spectrum, 3 eq of ClO- , Hb proton appeared. later, Hb proton is changing its chemical shift. Why? What happens after the addition of 1 eq of ClO-? ofter 2 eq? after 10 eq?

Author Response

Question 10: In section 2.4> I would say 'aqueous solution', not 'dissolved in H₂O' in the case of NaClO.

Answer: As the reviewer 1 suggested, we changed ‘dissolved in H₂O’ to ‘aqueous solution’ in section 2.4_.

Question 12: I disagree with the authors. If you compare different DFT functionals with not recommended basis set (6-31g(d,p), you may get wrong results and due to happy coincidence that some errors were cancelled by another errors, B3LYP and 6-31g(d,p) gave you the best match. Later, you use B3LYP DFT functional with better-ballanced basis sets than 6-31g(d,p) and due to the same reason, you might get wrong results.

Answer: We absolutely agree to the comment of the reviewer 1 on our calculation. Quantitative coincidence of calculations with experiments is not sufficient to judge whether models and theories used for the calculations are physically rigorous. Moreover, as the referee pointed out, since calculations with crude models and theories often result in nice quantitative consistency with experiments thanks to fortunate error cancellations, we should have been cautious to rule out the suggestion of the reviewer 1 in previous comment only based on the numerical accuracy. Therefore, we carried out additional calculations using all basis sets (6-31G(d,p), cc-pVDZ, and def2-TZVP) and DFT functionals (B3LYP, BMK, and CAM-B3LYP) that the referee mentioned. We found that regardless of what basis sets and DFT functionals are used, we can qualitatively rationalize our experiments. However, we found that physical accuracy in the basis sets and the DFT functionals did not guarantee a good quantitative consistency with experiments. In addition, our current methods cannot capture all aspects of experiments such as the splitting of the peaks around 400 nm of AFC. Therefore, with toning down our argument, we significantly modified the subsection 3.3 Calculations to clarify this important point raised by the reviewer 1.

Question 13: Could you please suggest the mechanism how is cleaved the imine bond by ClO-? ClO- is usually used for oxidations, not for hydrolysis.

Answer: This is a really difficult question for us. We also know that ClO^- is a strong oxidant. Actually, we tried to figure out the mechanism of how the imine bond is cleaved by ClO^-. However, it was not successful. In addition, we examined other references related to oxidative cleavage of imine bond by ClO^-, but none was reported. At this stage, it is very difficult to propose the oxidative cleavage mechanism of imine bond by ClO^-.

Question 14: I do not think the H-NMR spectrum in Figure 4 is explained. Clearly visible proton 5 of AFC (brown line), after addition of 0.5 eq of ClO-, the proton 5 disappeared (authors said the intensity is lower, but I cannot find that proton in the second spectrum even if zoomed, dark yellow colour). does it mean that 0.5 eq of ClO- hydrolyse 1 eq of AFC? or oxidise N of acridine? later, green spectrum, 3 eq of ClO- , Hb proton appeared. later, Hb proton is changing its chemical shift. Why? What happens after the addition of 1 eq of ClO-? ofter 2 eq? after 10 eq?

Answer: We really thank the reviewer 1 for pointing out ¹H NMR titration. We re-investigated the ¹H NMR titration and found that our previous assignment was not right. We replaced Fig. 4 with the new one including ADD, ADD-O and 1, 2 and 6 equiv of ClO^-. 10 equiv of ClO- was not applied due to solubility problem.

Round 3

Reviewer 1 Report

Please phase NMR spectra in Figure 4.

Spectrum in Fig S6, combination B3LYP/cc-pvdz , has different experimental spectrum compared to the others.

The DFT part

at this moment, I am not sure if this part is really important to this paper. In the beginning of the 3.3 section, it is said ‘To further understand the sensing mechanisms .. ‘ but I am missing how it help to understand the mechanism. According to me, the mechanism is still unclear as the authors said in their response.

The peaks about 400 nm are related to acridine moiety (e.g. https://webbook.nist.gov/cgi/cbook.cgi?ID=C90459&Mask=400#UV-Vis-Spec) and the splitting is related to vibronic features. It is not so difficult to calculate it, on the other hand, simple explanation would be sufficient.
If you compare the results immediately after the calculations, one might say B3LYP/6-31g(d,p) has the best match. On the other hand, you might want to scale your results (I mean energies of transitions, not wavelengths) by some scaling factor in similar manner as it is used for calculated infrared frequencies. Then, the match could be better for the other combinations.
Personally, I also do not like that authors use ‘the best match was achieved with B3LYP/def2tzvp’ and later ‘B3LYP/6-31g(d,p)’ – it is like mixing ‘pears and apples’.

Frankly, I would suggest to remove DFT part and focus more on the application part of the paper. If you insist on DFT, then rationalise the results according to following scheme:

In general discussion, use only the one model - means both DFT functional and basis set.
The comparison with the other models put into SI, comment how and why you chose the best model and not the others.
Model also the other possible products (right now, you have only two species included – AFC and ADDO) but there are also the other species – at least ADD, FHC.

I do understand how it is difficult to confirm the mechanism. However, in this case, it would be helpful if you could provide at least separate UV-Vis spectra of following pure compounds (not in reaction mixtures):

AFC (already have)
ADD (known compound)
ADD-O
- Would be great to prove that ClO- can oxidise ADD into ADD-O without any byproducts.
  - Otherwise, prepare it by other method in sufficient purity.
- FHC
  - Would be great to prove that ClO- cannot oxidise FHC into anything fluorescent and with significant absorption.

Author Response

Question 1. Please phase NMR spectra in Figure 4.

Answer: As the reviewer suggested, we revised the phase of the NMR spectra in Figure 4.

Question 2. Spectrum in Fig S6, combination B3LY P/cc-pvdz , has different experimental spectrum compared to the others.

The DFT part

The peaks about 400 nm are related to acridine moiety (e.g. https://webbook.nist.gov/cgi/cbook.cgi?ID=C90459&Mask=400#UV-Vis-Spec) and the splitting is related to vibronic features. It is not so difficult to calculate it, on the other hand, simple explanation would be sufficient.
If you compare the results immediately after the calculations, one might say B3LYP/6-31g(d,p) has the best match. On the other hand, you might want to scale your results (I mean energies of transitions, not wavelengths) by some scaling factor in similar manner as it is used for calculated infrared frequencies. Then, the match could be better for the other combinations.
Personally, I also do not like that authors use ‘the best match was achieved with B3LYP/def2tzvp’ and later ‘B3LYP/6-31g(d,p)’ – it is like mixing ‘pears and apples’.

Frankly, I would suggest to remove DFT part and focus more on the application part of the paper. If you insist on DFT, then rationalise the results according to following scheme:

In general discussion, use only the one model - means both DFT functional and basis set.
The comparison with the other models put into SI, comment how and why you chose the best model and not the others.
Model also the other possible products (right now, you have only two species included – AFC and ADDO) but there are also the other species – at least ADD, FHC.

Answer: As the reviewer suggested, we removed the DFT part and focused more on the application part.

Question 3. I do understand how it is difficult to confirm the mechanism. However, in this case, it would be helpful if you could provide at least separate UV-Vis spectra of following pure compounds (not in reaction mixtures):

AFC (already have)
ADD (known compound)
ADD-O
Would be great to prove that ClO- can oxidise ADD into ADD-O without any byproducts.
Otherwise, prepare it by other method in sufficient purity.
FHC
Would be great to prove that ClO- cannot oxidise FHC into anything fluorescent and with significant absorption.

Answer: As the reviewer suggested, we investigated fluorescent and UV-vis changes of AAD with/without ClO^- and FHC with/without ClO^-, and showed them in Figs. S4-S7. As shown as Fig. S4, fluorescent intensity of AAD was distinctly decreased, suggesting the oxidation of AAD into AAD-O. UV-vis spectra of AAD showed increase of absorbance at 490 nm (Fig. S5). Moreover, FHC with/without ClO^- showed no fluorescence intensity (Fig. S6). In UV-vis spectra, absorbance of FHC with ClO^-at 280 nm increased (Fig. S7). We described these results in the first paragraph of page 9.