Enhancing multiphoton upconversion through interfacial energy transfer in multilayered nanoparticles

Photon upconversion in lanthanide-doped upconversion nanoparticles offers a wide variety of applications including deep-tissue biophotonics. However, the upconversion luminescence and efficiency, especially involving multiple photons, is still limited by the concentration quenching effect. Here, we demonstrate a multilayered core-shell-shell structure for lanthanide doped NaYF4, where Er3+ activators and Yb3+ sensitizers are spatially separated, which can enhance the multiphoton emission from Er3+ by 100-fold compared with the multiphoton emission from canonical core-shell nanocrystals. This difference is due to the excitation energy transfer at the interface between activator core and sensitizer shell being unexpectedly efficient, as revealed by the structural and temperature dependence of the multiphoton upconversion luminescence. Therefore, the concentration quenching is suppressed via alleviation of cross-relaxation between the activator and the sensitizer, resulting in a high quantum yield of up to 6.34% for this layered structure. These findings will enable versatile design of multiphoton upconverting nanoparticles overcoming the conventional limitation.

modelling such phenomena with rate equations may bot be trivial (to account for core-shell), but maybe could help to understand the physics.
Reviewer #2 (Remarks to the Author): Evaluation: • Data is technically sound: Data is mostly sound, but there are some holes and missing data that would complete the manuscript. • Paper provides strong evidence: The manuscript mostly contains strong evidence, but there are some questions and concerns.
• Results are novel: Results are sufficiently novel • Manuscript is important to scientists in the specific field: Manuscript is sufficiently important to advancing the field. • Recommendation: The novelty of the results and findings are sufficient and relevant to advancing the field of upconversion nanoparticles for consideration. Considerable for Nature Communications with major revisions addressing the concerns laid out below.
This manuscript reports a two-pronged approach to enhancing upconversion nanoparticle quantum yield, claiming the highest reported quantum yield for these materials at low power density excitations. This is truly a remarkable find, especially if the findings provide such a generalizable strategy as is claimed in the manuscript; however, there are a few concerns regarding some of the background theory, the suggestions implied by the mechanism, some technical details, and some potentially missing data points that would fully round out the manuscript and meet the standards of a Nature Communications publication: 1) The author stated that the rationale, at least partially, for this core-shell-shell structure is that it would alleviate the backward energy transfer pathway in these upconverting nanoparticles systems. Is this pathway significant enough to address directly? It would be appreciated if the authors can supply a stronger rationale and supporting literature to back up this point. Checking citations 11 and 20 as cited by the author, it is not clear that this pathway is so detrimental to the author's goals, and as a matter of fact, it is implied that this back transfer actually assists stronger red and blue emissions, as is shown later in the manuscript.
2) The author mentions a short Yb3+-Er3+ distance at the interface, yet later in the manuscript they state that "the rise in the decay times for the Er3+ emission are obviously prolonged" due to the same structural feature. This seems inconsistent, and careful wording or explanation would be appreciated.
3) The author states that the canonical structure for UCNPs is a codoped NaYF4 system, which is true; however, my impression is that the canonical structures are usually β-hexagonal as opposed to α phase used in this paper. It would be therefore important to include the β phase canonical structure in your quantum yield measurements. Furthermore, checking your quantum yield comparisons table  (Supplementary Table 1), it is interesting to note that there is a large size discrepancy between the β phased particles and the α phased particles. This is of particular concern considering that within the α phase particles, there is a ~1% difference in quantum yield for a difference of ~6.6 nm, while there is a ~16.1 nm difference between the optimal core-shell-shell structure and composition showcased in the paper, and the β phased version. Please provide further data to supplement and clear up this inconsistency or provide strong arguments as to why this comparison was made. The author should also provide a more consistent size comparison for the canonical structures as well.
4) The severe quenching due to the lack of protection from an inert shell is to be expected, but it would be interesting to see a comparison between the CY:10% Er@SYb structure and an unprotected codoped core (canonical structure/composition) and see the effect without removing the surface quenching aspect.
5) The finding that increasing Er3+ concentrations continues to increase the upconversion emissions is fascinating. Why did the author choose 10% as the optimal composition if the luminescent output is higher in other showcased compositions? Furthermore, a comparison of a completely doped (100% Er3+) system would also be interesting. What are the quantum yield results for these compositions? If indeed a 10% Er3+ doping ratio provides the highest quantum yield, then this needs to be demonstrated with quantum yield data. Please provide the complete data set showing quantum yield measurements for these showcased compositions and consider a wider range.
6) The temperature dependent photoluminescent observations are very interesting. It would be appreciated if the author can include measurements showing the photoluminescence at higher temperatures and comment on the effect of temperature on the peak ratios.
7) The finding that this core-shell-shell structure can be a general strategy for enhancing upconversion luminescence is fascinating! There should be quantum yield comparisons for the LiYF4 structures for consistency, to showcase this interesting discovery. 8) To further emphasize previously mentioned points, there needs to be a more consistent approach to comparisons. Why is there no quantum yield data comparing the Tm3+ and Tb3+ doped systems to their canonical counterparts? This dataset would be crucial for showcasing these amazing discoveries.
9) The author is unclear in their description of the canonical Tm doped upconversion nanoparticle system. Are these structures α or β phased particles? The author should be specific for consistency. 10) Is there a functional utility for these core-shell-shell structured upconversion nanoparticles? It would be appreciated if the author can use a simple demonstration to showcase these great enhancements.
In this work, the authors attempted separating sensitizers (Yb 3+ ions) and activators (Er 3+ ions, Tm 3+ ions) in different layers (ie shells) to enhance multiphoton upconversion emission. Even though the authors have carried out good amount of work and decent control experiments, it is to my opinion that the novelty of this work is not strong enough for publication in Nature Communications. Some experiments are inconsistent and not comparable, please see my further remarks below.
1. Separation of sensitizers and activators in different layers has been widely reported and accepted in Nd 3+ sensitized systems, which showed superior upconvsion performance and less heating up effect. 2. The authors should demonstrate and prove cross-relaxation (CR) between Yb 3+ and Er 3+ ions since the authors claimed that the intrinsic cross-relaxation energy loss in UCNPs has not yet been properly addressed (line 57-58, page 3). The reason of concentration quenching is commonly attributed to that the increase of concentration of activators will induce more CR between activators, instead of CR between activator and sensitizer the authors claimed in abstract.
3. In comparison of quantum yields of different UCNPs, β-NaYF4:10%Er@NaYbF4@NaYF4 showed lower QY than that of α-NaYF4:10%Er@NaYbF4@NaYF4, which is different from the former reported that β-UCNPs showed stronger UC emission than α-UCNPs. The authors should explain the reason for this. The sizes of UCNPs of α and β phases are different, which is not comparable and hence an appropriate control of equal size should be selected. . Figure 1.e, the emission of CY:10%Er@SYb@SY and CY:20%Yb, 10%Er@SYb@SY did not differ too much, which is not consistent with the authors' claim.

As shown in
5. The authors tried to prove the enhancement of multiphoton upconversion emission, especially 407 nm. However, the significance of such enhancement is not fully emphasized, and one potential application should be demonstrated in this communication to show that after enhancement the nanostructured nanoparticles could still be appropriately applied to a particular field.

Manuscript no. NCOMMS-18-38464
Response to reviewers' comments Firstly, we would like to thank all reviewers for their precious comments, which helped us to improve the work dramatically.
The major changes to manuscript in this revision are briefly listed as follows: 1. The comparison of peak ratios was replaced with relative intensity of emission bands in Figure 1f. The pictures were taken with and without filters again for consistence and Figure 1g was updated.
In the following pages, the reviewers' comments were replied point to point with the response highlighted in blue.

Reviewer #1
The

Response: Sample information including thickness of NaYbF 4 shell and NaYF 4 shell has been supplemented in all figures and tables including those in
Supplementary Information. 6. The spectra are not corrected for spectral sensitivity (which includes grating, PMT, optics etc.) -this is not a problem for relative changes, but is crucial if the authors claim the UC emission is X times larger than the green of red one.
Response: Taking Tm 3+ doped nanoparticles as an example, versatile UCL spectra have been reported previously (Figure 1-2 Fig.4c presents uncorrected data and makes the comparison more difficult. 12. Why the lifetimes rise from 2-30% of Er (Fig.S14) Fig. 2b), Supplementary Fig. 3), Yb lifetimes vs. rising Er concentration (Fig. S15a), Er QY vs displacement between Yb and Er shell with yet another empty shell (Fig. S17 and   2) The author mentions a short Yb3+-Er3+ distance at the interface, yet later in the manuscript they state that "the rise in the decay times for the Er3+ emission are obviously prolonged" due to the same structural feature. This seems inconsistent, and careful wording or explanation would be appreciated.

TEM-EDX for the core (inset in
Response: In our activator doped core@sensitizer shell@inert shell structure, the distance of Yb 3+ -Er 3+ at the interface between NaYF 4 :10%Er core and NaYbF 4 shell is shorter than that in the canonical structure because higher concentration  Supplementary Fig. 23b).
In fact, the length of these nanorods is larger than α phased particles, but their diameter is smaller than α phased particles. Thus, we modified the synthesis method slightly and acquired nanoparticles with isotropic structure. Fig. 31d). There are smaller nanoparticles with dark core, indicating partly unsuccessful epitaxial growth of NaYbF 4 . Therefore, the detected quantum yield is ~4.99%, smaller than the corresponding α phased particles. We believe the lower quantum yield is an artifact induced by our synthesis. We have specified the reason for lower quantum yield of β phased particles in the revised manuscript (Table 1).

4) The severe quenching due to the lack of protection from an inert shell is to be expected, but it would be interesting to see a comparison between the CY:10%
Er@SYb structure and an unprotected codoped core (canonical structure/composition) and see the effect without removing the surface quenching aspect.

Response: We supplemented comparison between the C Y :10%Er@S Yb structure
and an unprotected codoped core in Supplementary Fig. 4. Obviously, much stronger emission can also be observed from the former than the latter. In fact, NaYbF 4 sell acts as not only a sensitizing layer but also an inert layer for Er 3+ emission in this case since emission from Yb 3+ is much less sensitive to surface quenching. In other word, there still exists a protection layer in C Y :10%Er@S Yb structure, different from unprotected codoped core. Response: The quantum yields for core-shell-shell nanoparticles with Er 3+ concentration increasing from 2 to 100% are supplemented in Table S4. The α-NaYF 4 :10%Er@NaYbF 4 @NaYF 4 UCNPs exhibit the highest quantum yield. The corresponding UCL spectrum, downconversion luminescence spectrum and decay curves of α-NaErF 4 @NaYbF 4 @NaYF 4 have also been supplemented in Supplementary Information (Supplementary Fig. 14-16 (Table S4, 0.48±0.04%). The quantum yields of the other three are smaller than 0.1%, which is lower than the detection limit of our spectroscopy. The experimental results show that 1 O 2 production activity of C Y :2% Er@S Yb @S Y UCNPs reaches 3.8 times that of C Y :20% Yb, 2% Er@S Y UCNPs.

Considering the neglected enhancement in green and red emission for C Y :2%
Er@S Yb @S Y UCNPs as compared with C Y :20% Yb, 2% Er@S Y UCNPs, the activity promotion should be attributed to enhanced violet emission. Related discussion is also supplemented in the text (last paragraph, page 13-14).

UCNPs has not yet been properly addressed (line 57-58, page 3). The reason of concentration quenching is commonly attributed to that the increase of concentration of activators will induce more CR between activators, instead of CR between activator and sensitizer the authors claimed in abstract.
Response: Generally, the concentration quenching induced by elevating the concentration of the activators is widely accepted, the contribution of CR between activator and sensitizer (mainly for Yb 3+ ) to concentration quenching has long been neglected. First, we demonstrate here that all of the multilayered structures with Yb 3+ and Er 3+ codoped into the core present a much lower UC luminescent intensity than our C Y :10% Er@S Yb @S Y structure despite their particle sizes, sensitizer layers and inert layers all being nearly the same (Fig. 1).
In particular, C Y :10%Er@S Yb @S Y sample presents much more stronger  Supplementary Fig. 23, the diameter of our β-NaYF 4 :10%Er@NaYbF 4 @NaYF 4 nanorods (26.1±4.3 × 37.2±3.9) is really smaller than α-NaYF 4 :10%Er@NaYbF 4 @NaYF 4 particles (32.2±2.2), but their length is larger than the latter. Thus, we think the anisotropic growth of β-NaYbF 4 shell may contribute more to their lower quantum yield. We tried to synthesize larger β-NaYF 4 :10%Er@NaYbF 4 @NaYF 4 particles, but we found the growth of β-NaYbF 4 shell is still uncontrollable. There is a large deviation in the sizes of β-particles and independent growth of NaYbF4 also appears ( Supplementary Fig. 31d). Consequently the detected quantum yield of newly synthesized nanoparticles is also lower than α phased particles. We believe the lower quantum yield is an artifact induced by our synthesis. We have specified the reason for lower quantum yield of β phased particles in the revised manuscript (Table 1). Figure 1.e, the emission of C Y :10%Er@S Yb @S Y and C Y :20%Yb, 10%Er@S Yb @S Y did not differ too much, which is not consistent with the authors' claim.

Reviewers' Comments:
Reviewer #1: Remarks to the Author: I appreciate the response from the authors, who improved the manuscript significantly. There are however still some points to clarify and add, before the MS is ready for acceptance.
1. Why Fig.2b stops at d_SYb=10nm as the curve continuously rises up ? The same question is valid for the inset in Fig.2b, Fig.3d, 3b? Such unprecedented high values of QY deserves versatile characterisation even at higher thickness of Yb intermediate shell as well as excitation power in this case. If this is going to "change the game", the authors cannot stop "in the middle of the trip". 2. I have requested to study TEM-EDX maps -i.e. the composition maps of the nano-crystals (similat to, but more detailed to the one in Fig.1c) -this is critically important for understanding the observed behaviour and relate it to the structure of the core-shell-shell composition. Why? For example, we have synthesized beta-NaYF4 Yb@Er core shell materials for a project we have, and they were much weaker than YbEr or Yb@YbEr, which is in opposite to the claims made in this manuscript. Therefore, having 20%Yb and 10%Er of such high concentration makes them eligible for NC composition mapping with TEM-EDX (i.e. the concentrations are high enough) to check if the Yb and Er ions colocalize in some intermediate space between core and the first shell. Without this result, the data presented by the authors are just general observations and hypotheses only. Pure single Er doped upconversion is of course possible, but is significantly enhanced with Yb sensitizer, which is intuitively correct when you know how APTES work. However, the claims made by the authors, that separating Er from Yb enhance the up-conversion are somehow against intuition and definitely require deeper understanding and more facts. 3. What is the mechanism behind the PDT in this MS ? Usually, this is FRET between Er donors and the PS acceptor, but here Er is inside the core, too far away (a few times Forster distance) from the PS acceptor at the surface. Therefore, this is most probably reabsorption of light, but this hypothesis would require validation (e.g. lifetimes vs concentration of PS for example). I understand the reason to put PDT experiment in this manuscript, but this, rather trivial demonstration, does not explain anything else than the fact, the new NPs are brighter than conventional core only NPs. 4. Actually I am astonished that it is more easy to control the shape and size in alpha phase NaYF4 then in beta nano-crystals.

Evaluation:
 Data is technically sound: Satisfactory data accomplished with the addition of significant key data points in the SI as well as the amendment of the text to provide more relevant figure sets.  Paper provides strong evidence: The manuscript contains strong evidence and rationale.  Results are novel: Results are sufficiently novel  The manuscript is important to scientists in the specific field: Manuscript is sufficiently important to advancing the field.  Recommendation: The novelty of the results and findings are sufficient and relevant to advancing the field of upconversion nanoparticles for consideration. Considerable for Nature Communications with minor revisions addressing the concerns laid out below.
The authors have given a strong effort at addressing the concerns of the reviewers. More specifically, they have added a small utility demonstration as well as significantly expanding certain parts of the SI to address the reviewer's concerns; however, while many changes are made, not all of them are positive, and there remains room for improvement.
1) It is not necessarily a bad/incorrect for the B phase to have a lower quantum yield in the case of your structure, but it is interesting to note and it would be interesting if there might be some other way to think about this enhancement. The clarification and addition of data are appreciated. 2) It might be an interesting addition to add a photograph of the main canonical compositions compared with your uniquely structured nanoparticle to showcase improvements.   . Supposing cations intermixing does not occur at the interface between core and the first shell given here, the composition at core/NaYbF 4 shell interface is approximately NaYF 4 : 50%Yb, 5%Er if one NaYF 4 :10%Er layer and one NaYbF 4 layer are considered. Such a doping concentration is higher than that for canonical codoped structure (20%Yb, 2%Er), making it possible for the former (core/NaYbF 4 shell interface) to possess shorter Yb-Er distance than that for the latter. As a result, more efficient APTES happens at the interface between core and NaYbF 4 shell. According to our experimental results, C Y :10% Er@S Yb core-shell UCNPs emit much weaker UCL than C Y :20%Yb, 2%Er@S Y core-shell and C Y :10% Er@S Yb @SY core-shell-shell UCNPs (Figure 2e and Figure 3c). This finding indicates surface quenching significantly affects UCL of these UCNPs since their particle sizes are similar (27.7±2.4 and 32.0±2.5 nm for the former and the latter, respectively). In other words, highly efficient energy transfer upconversion can only be realized when the surface quenching is suppressed and negligible. Similarly, more severe surface quenching for beta-NaYF 4 :Yb@Er core-shell material than that for YbEr or Yb@YbEr structure may lead to its much weaker UCL than the latter two structures. On the other hand, larger Yb-Er distance (the doping concentration in separated core-shell structure is the same as that in codoped one) for the former than the latters may also lead to its weak UCL.

What is the mechanism behind the PDT in this MS? Usually, this is FRET between
Er donors and the PS acceptor, but here Er is inside the core, too far away (a few times Forster distance) from the PS acceptor at the surface. Therefore, this is most probably reabsorption of light, but this hypothesis would require validation (e.g. lifetimes vs concentration of PS for example). I understand the reason to put PDT experiment in this manuscript, but this, rather trivial demonstration, does not explain anything else than the fact, the new NPs are brighter than conventional core only NPs.
Response: We have supplemented luminescence decay curves for Tween 20 modified NaYF 4 :Er@NaYbF 4 @NaYF 4 nanoparticles (d SYb = 8.3±0.7 nm, d SY = 2.3±0.1 nm) before and after loading the photosensitizer, HMME (Supplementary Fig.35). The invariable lifetimes exclude Förster Resonance Energy Transfer (FRET) mechanism. PDT experiment herein demonstrates the trilayered nanoparticles do emit much stronger light at shorter wavelength and efficiently initiate photoreactions, which are necessary for their potential application in deep-tissue biophotonics such as optogenetics. Related discussion has been supplemented on page 15 (line 5-10).

Actually I am astonished that it is more easy to control the shape and size in alpha phase NaYF4 then in beta nano-crystals.
Response: In our previous work, we found relatively smaller beta-NaYF 4 nanoparticles could be synthesized using lanthanide trifluoroacetates as the precursors, but the sizes of these nanoparticles vary from one batch to another. We had to repeat the experiments several or tens of times to acquire the same sized nanoparticles. Whereas, the sizes of small alpha-NaYF 4 nanoparticles synthesized using the same precursors are nearly invariable for different batches. We believe thermal decomposition of sodium trifluoroacetate (NaTFA) before beta-core formation varies from one batch to another because excess NaTFA is necessary for minimizing nanoparticle size of beta phase (NaTFA decomposes at 250 °C and beta-core forms at a temperature higher than 300 °C). In contrast, there is no excess NaTFA and the temperature for formation of alpha-core is much lower than that for beta-core, so thermal decomposition of NaTFA can be neglected and the synthetic condition is invariable for different batches. According to these experiences, we chose alpha-phase to investigate the structure-dependent energy transfer upconversion of Yb/Er doped NaYF 4 UCNPs in order to get reliable results. To make sure that the conclusion is also suitable for beta-phase NaYF 4 , the widely used host lattice, we also synthesized beta-phase using the same method. Unfortunately, the growth of beta-NaYbF 4 layer seems to be more difficult to control than NaYF 4 counterpart. To clarify which one is more favorable for photon upconversion for this trilayered structure: alpha-phase or beta-phase, we tried to re-synthesize beta-phase trilayered nanostructures following a procedure reported by Steven Chu et al.
[Nature Photonics 12, 5488-553 (2018)] with a slight modification and measured their quantum yield. The detected quantum yield for these better controlled beta-phase UCNPs is higher than their alpha-phase counterparts (6.82±0.50 versus 5.42±0.43, Table 1), indicating beta-phase is still the better host. Accordingly, Supplementary Fig. 31 has been updated.

Reviewer #2
The authors have given a strong effort at addressing the concerns of the reviewers. More specifically, they have added a small utility demonstration as well as significantly expanding certain parts of the SI to address the reviewer's concerns; however, while many changes are made, not all of them are positive, and there remains room for improvement.
Response: We are grateful for the reviewer's helpful comments about improving the manuscript and addressed the comments accordingly below.
1) It is not necessarily a bad/incorrect for the B phase to have a lower quantum yield in the case of your structure, but it is interesting to note and it would be interesting if there might be some other way to think about this enhancement. The clarification and addition of data are appreciated.
Response: We have re-synthesized beta-phase by a better controllable method to clarify this phenomenon. As we expected, the quantum yield for these newly synthesized beta-phase UCNPs is indeed higher than their alpha-phase counterparts (6.82±0.50 versus 5.42±0.43, Table 1). 2) It might be an interesting addition to add a photograph of the main canonical compositions compared with your uniquely structured nanoparticle to showcase improvements. 3) The additional quantum yield data and the utility demonstration are greatly appreciated additions; however, the restructuring of the figures leaves the last one a bit lacking.

Response
Response: We have redrawn the last figure ( Figure 5 in the revised version) and supplemented schematic diagram for loading HMME and producing 1 O 2 on the UCNPs.

4) The proof of concept is a welcome addition to the manuscript, but it would
probably be better served if it was expanded or if there was another figure more comprehensively capturing the scope of your work. For example, a more tailored version of the original figure 4 where figure 4a is replaced by emission color photographs and certain key data points or perhaps a schematic is added to emphasize the results. Either figure 4 should be expanded or another figure should be added. As it stands, the figures as a whole are a bit lacking compared to other recent publications in Nature comm. This includes publications on similar upconversion nanoparticle-based topics.
Response: We have redrawn all figures to improve the quality and highlight the results. In particular, a schematic comparison of the canonical UCNPs with newly structured UCNPs has been added to emphasize the topic or the main scope of our work. Since only green emission varies slightly during laser irradiation, no significant variation in emission color photographs could be observed. A schematic diagram for loading HMME and producing 1 O 2 on the UCNPs has been supplemented in figure 4 (Fiugre 5 in the revised version) providing experimental details and mechanism. There are totally five figures in the revised manuscript.  (2019)] also demonstrate a core-shell structure designed to minimize back energy transfer to provide emissions from low power density excitations. Can you comment on these recent advances and clarify the novelty of your structure? The authors should update and amend the introduction to reflect the current status of this field and the impact of their contribution and findings.

5) While the authors mainly compare their structured
Response: We have synthesized the structures appeared in above mentioned papers and compared the PL spectra of these structures. The results are attached below and supplemented in Supplementary Fig. 32 Figure S4 in the published paper). Even though we increase Yb content to 100% (C Yb @S Y : 10% Er@S Y ), this structure still emits weak UCL and no increased multiphoton upconversion can be observed. As NaYF 4 :20%Yb@NaYF 4 :2%Er@NaYF 4 :20%Yb structure is concerned, the authors did not observe increased contribution from the violet and red emission to the overall emission either (Fig. 3b in the paper). The NaYF 4 @NaYbF 4 :2%Er@NaYF 4 nanoparticles [Nature Photonics 12, 5488-553 (2018)] emit comparable green light and weaker violet and red light as compared with our nanostructures. Considering only one third of Er 3+ ions exist in our trilayered structure (Er/Yb ≈ 1/133 for C Y :10% Er@S Yb @S Y , Er/Yb = 1/49 for C Y @S Yb :2% Er@S Y ) while the amounts of Yb 3+ ions are nearly the same for both structures, we believe UC energy transfer efficiency of our structure is higher than the structure reported in Nature Photonics. In fact, the different variation tendencies of the lifetimes for the upconverted Er 3+ emission bands with increasing Er 3+ concentration in the two structures also validate higher efficiency of our trilayered structure (line 10-13 on page 10, Supplementary Fig.  16). Figure 1. A comparison of α-NaYF 4 :10%Er@NaYbF 4 @NaYF 4 nanoparticles and the recently reported new nanostructures. (a-h) TEM images of α-NaYF 4 @NaYb 0.98 Er 0.02 F 4 @NaYF 4 (C Y @S Yb : 2%Er@S Y ) (a-c for the core, core-shell and core-shell-shell nanostructures), α-NaYbF 4 @NaY 0.9 Er 0.1 F 4 @NaYbF 4 (C Yb @S Y : 10%Er@S Yb ) (d-f for the core, core-shell and core-shell-shell nanostructures) and α-NaYb 0.8 Er 0.2 F 4 @NaYF 4 (C Yb : 20%Er@S Y ) (g, h for the core and core-shell nanostructures). (i) Upconversion emission spectra for these UCNPs upon 980-nm excitation (24.0 W/cm -2 ).

In the meantime, we have amended the introduction (line 7-8 on page 3 and line 3-4 on page 4) and supplemented the corresponding references (17-18).
6) Some of the authors' names are not properly written or spelled. Please proofread the text for these minor errors.