Metathesis Cascade-Triggered Depolymerization of Enyne Self-Immolative Polymers**

Article abstract of DOI:10.1002/anie.202108239

A novel class of enyne self-immolative polymers (SIPs) capable of metathesis cascade-triggered depolymerization is reported. Studies on model compounds established 1,6-enyne structures for efficient metathesis cascade reactions. SIPs incorporating the optimized 1,6-enyne motif were prepared via both polycondensation and iterative exponential growth approaches. These SIPs demonstrated excellent stability in strong acid, base, nucleophiles, or at elevated temperatures, and can undergo efficient and complete depolymerization once triggered by a metathesis catalyst. Further studies revealed that introducing a terminal alkene to the chain end of the enyne SIPs improved the depolymerization efficiency, and established their potential as stimuli-responsive materials.

Full text of DOI:10.1002/anie.202108239

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Metathesis Cascade-Triggered Depolymerization of Enyne Self-
Immolative Polymers
Authors: Jingsong Yuan, Gavin J. Giardino, and Jia Niu
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.202108239
Link to VoR: https://doi.org/10.1002/anie.202108239

10.1002/anie.202108239
Angewandte Chemie International Edition
Metathesis Cascade-Triggered Depolymerization of Enyne Self-Im-
molative Polymers
Jingsong Yuan, Gavin J. Giardino, and Jia Niu*
[a]
Dr. Jingsong Yuan, Gavin J. Giardino, Prof. Dr. Jia Niu
Department of Chemistry
Boston College
Chestnut Hill, Massachusetts 02467, United States,
E-mail: jia.niu@bc.edu
Supporting information for this article is given via a link at the end of the document.
Abstract: A novel class of enyne self-immolative polymers (SIPs) ca-
pable of metathesis cascade-triggered depolymerization is reported.
Studies on model compounds established 1,6-enyne structures for ef-
ficient metathesis cascade reactions. SIPs incorporating the optimized
1,6-enyne motif were prepared via both polycondensation and iterative
exponential growth approaches. These SIPs demonstrated excellent
stability in strong acid, base, nucleophiles, or at elevated temperatures,
and can undergo efficient and complete depolymerization once trig-
gered by a metathesis catalyst. Further studies revealed that introduc-
ing a terminal alkene to the chain end of the enyne SIPs improved the
depolymerization efficiency, and established their potential as stimuli-
responsive materials.
the depolymerization of enyne SIPs is irreversible, as the endocy-
clic trisubstituted alkenes of the depolymerization products are
less-reactive to Ru catalysts.[16]
Scheme 1. Development of the Enyne Self-Immolative Polymer
Self-immolative polymers (SIPs) that can undergo head-to-tail
depolymerization in response to external stimuli are promising ma-
terials for a wide range of emerging applications such as recyclable
plastics, drug delivery, and biosensing.[1] Existing SIPs, such as
2]
polycarbamates,[1a,
polycarbonates,[3] poly(benzyl ether)s,[4]
polyphthalaldehydes,[5] and polyglyoxylates,[6] mainly rely on nu-
cleophilic/anionic depolymerization mechanisms including quinone
methide elimination, cyclization and elimination of five-membered
heterocycles, and hemiacetal fragmentation (Scheme 1A).[7] How-
ever, with a few notable exceptions,[4a, 8] these nucleophilic/anionic
depolymerization mechanisms often result in the high sensitivity of
these SIPs to even mild acidic/basic conditions, as well as slow
and incomplete depolymerization.[9] Furthermore, the highly reac-
tive intermediates generated during the ionic/nucleophilic depoly-
merization, such as quinone methides, can lead to undesirable
side reactions with nucleophiles.[1b]
Our investigation began by studying a series of model com-
pounds 1–7 capable of the enyne metathesis cascade reaction
(Scheme 2). These model compounds consist of two fused 1,6-
enyne structures, such that the adjacent alkenes and alkynes are
separated by a –CH2–X–CH2– spacer, where X is a variable func-
tional group that provides the steric buttressing effect[17] for the me-
tathesis cascade cyclization triggered by third-generation Grubbs
catalyst (G3). The variable X group also helps to improve the sol-
ubility of the model compounds. Except 5, the reactions of model
compounds 1–6 were found to proceed efficiently in the presence
of 3 mol% G3 in dichloromethane (DCM) or tetrahydrofuran (THF)
at room temperature. The reactions reached quantitative conver-
sions to afford the corresponding 1,1’-bicyclopentene derivatives
8–13 in high yields over 30 minutes (Scheme 2A). Likely due to the
weak buttressing effect of the acetyl group, compound 5 only
achieved 47% conversion under the same conditions. These ob-
served reactivities of the enyne metathesis cascade reactions are
consistent with previous reports by Grubbs et al.[18] and Nolan et
al.[19] In contrast, although terminal alkynes are also known to initi-
ate enyne metathesis, only 10% conversion was achieved to gen-
erate 14 when 7 was subjected to the same reaction conditions
(Scheme 2B). The slow reaction of 7 was attributed to the lower
reactivity of the vinyl carbene compared to the alkylidenes formed
in 1–6 (Scheme S1).[16] Taken together, the model compound stud-
ies suggested that enyne structures with bulky substitutions could
undergo efficient metathesis cascade reactions, and that terminal
alkenes offer better reactivity compared to terminal alkynes.
Next, we attempted to prepare enyne SIPs via polycondensa-
tion. We chose X = p-pentyl-N-benzenesulfonamide as the spacer
Since first reported by Katz et al,[10] enyne metathesis cascade
reactions have emerged as powerful approaches to construct com-
plex cyclic structures from linear precursors in synthetic organic
chemistry.[11] Earlier works by Choi,[12] Gutekunst,[13] Hawker,[13a]
and Kilblinger,[14] have demonstrated the high efficiency of ruthe-
nium (Ru)-catalyzed enyne metathesis cascades in chain-growth
polymerization, polymer chain-end modification, and in situ initiator
generation for ring-opening metathesis polymerization (ROMP,
Scheme 1B). However, to the best of our knowledge, enyne me-
tathesis cascade reactions have not previously been used to ena-
ble the depolymerization of synthetic polymers. Herein, we envi-
sion that a novel class of enyne SIPs can be designed to enable
consecutive metathesis cascade-triggered depolymerization with
high efficiency. Our SIP design entails a backbone structure con-
sisting of repeating 1,6-enynes, such that the propagating Ru-car-
bene intermediate can undergo consecutive 5-exo-dig and 5-exo-
trig cyclizations to eliminate a 1,1’-bicyclopentene derivative and
regenerate the Ru carbene (Scheme 1C). The propagation of this
enyne metathesis cascade reaction along the SIP backbone would
lead to head-to-tail depolymerization. Fundamentally distinct from
existing SIPs, our design involves no nucleophilically labile func-
tional groups. Also, unlike the depolymerization of polycycloolefins
(e.g., polycyclopentene),[15] which is the reverse reaction of ROMP,
This article is protected by copyright. All rights reserved.

10.1002/anie.202108239
Angewandte Chemie International Edition
Table 1. Enyne SIPs via Polycondensationa
Scheme 2. Enyne Metathesis of Model Compounds
Entry
1c
Monomer
16+17a
16+17a
16+17a
16+17a
16+17a
16+17a
15
Base
DBU
NaH
Mn(kg∙mol-1)b
1.3
Ðb
1.45
1.50
1.20
1.50
1.33
1.56
1.46
1.21
1.47
2c
1.2
1.2
1.8
1.9
34.2
1.5
10.5
28.0
3
4
5c,d
6
7e
8f,g
9f,g
TMG
K2CO3
K2CO3
Cs2CO3
Cs2CO3
Cs2CO3
Cs2CO3
16+17a
16+17b
aConditions: under air for 20 h, 4 eq. base. All polymerizations
reached over 95% conversion as determined by crude 1H
within the 1,6-enyne structure due to the high reactivity and excel-
lent solubility of the model compound 1 that contains this spacer.
Two distinct polycondensation approaches, AB-type and AABB-
type, were explored. Retrosynthetic analysis of the target SIP gave
rise to a bromosulfonamide enyne monomer 15 for the AB-type
polycondensation, or a disulfonamide alkyne monomer 16 and (E)-
1,4-dibromo-2-butene 17a or (Z)-1,4-dibromo-2-butene 17b for the
AABB-type polycondensation (Table 1; see Supporting Information
for detailed procedures for monomer synthesis). Various inorganic
and organic bases were investigated to optimize the reaction. All
AABB polycondensations of 16 and 17a mediated by strong bases
such as 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) and sodium hy-
dride (NaH) resulted in a dark-colored complex mixture (Table 1,
entry 1–2). In contrast, reactions mediated by 1,1,3,3-tetramethyl-
guanidine (TMG) or potassium carbonate (K2CO3) proceeded with
minimal color change; however, size exclusion chromatography
(SEC) analysis showed that only products with low number aver-
age molecular weights (Mn) were obtained (Table 1, entry 3–4).
Increasing the temperature to 80 C did not improve Mn (Table 1,
entry 5). Polymers with higher Mn were obtained when cesium car-
bonate (Cs2CO3) was used (Table 1, entry 6 and Table S1), which
was attributed to the higher basicity and solubility of Cs2CO3 com-
pared to K2CO3 in N, N-dimethylformamide (DMF). The low Mn cy-
clic oligomers fraction could be efficiently removed from the crude
product by a simple precipitation in diethyl ether, yielding the linear
polymer P1 (Figure S1–S2). In contrast, AB polycondensation of
monomer 15 only afforded low Mn oligomers (Table 1, entry 7),
which was attributed to the high cyclization tendency of 15. To our
surprise, despite lacking terminal alkene chain ends, P1 was able
to undergo depolymerization, yielding a 3,3’-bidihydropyrrole de-
rivative 8 exclusively (Figure 1). However, a relatively high catalyst
loading of 7 mol% G3 (hereafter, catalyst loading is calculated with
respect to the polymer repeating units) was necessary to achieve
a complete conversion.
NMR. Initial monomer concentration [M]0 = 1.0 mol∙L-1. Mo-
b
lecular weight (Mn) and dispersity (Đ) were determined by
SEC analysis calibrated to polystyrene standards. For entries
6, 8, and 9, Mn and Đ are reported after precipitation in diethyl
d
ether. cReaction turned dark within 10 minutes. The reaction
was carried out at 80 oC. eTwo equivalents of base were used.
fMonomer feed ratio: 1/0.975. gReaction time: 3.5 h. Polymer
was capped by 1 equiv. allyl bromide with respect to 16.
produced two orthogonally monoprotected monomers, which were
subsequently coupled via the Mitsunobu reaction to give the pro-
tected dimer 20. Two more iterations of the deprotection-coupling
cycle afforded an octamer (Scheme 3B). Studies on the terminal
alkene-capped dimers derived from 20 indicated that the oligomer
capped at the alcohol end underwent more efficient depolymeriza-
tion compared to the one capped at the sulfonamide end. Both
end-capped dimers were depolymerized more efficiently than the
uncapped 20 (Scheme S2 and Figure S3–S5). In light of this find-
ing, the octamer was capped by a terminal alkene at the alcohol
end to yield a discrete octameric enyne SIP P2, the structure of
which was confirmed by NMR and mass spectrometry (Scheme 3C
and Figure S6).
The surprising depolymerization of P1 prompted us to further in-
vestigate the role of the terminal alkene chain-end group in depol-
ymerization. To this end, we devised an iterative exponential
growth (IEG) route to synthesize discrete oligomers with well-de-
fined terminal chain-end groups.[20] First, a bifunctional IEG mono-
mer 18 carrying orthogonal acetyl (Ac) and t-butyloxycarbonyl (Boc)
protecting groups was conveniently prepared by the Mitsunobu
coupling of a monoBoc disulfonamide 19 and monoacetyl 1,4-bu-
tynediol (Scheme 3A). Selective deprotection of 18 at either ends
Figure 1. 1H NMR spectra indicates that the depolymerization of P1 (A) cleanly
produced 8 (B).
This article is protected by copyright. All rights reserved.

10.1002/anie.202108239
Angewandte Chemie International Edition
starting to decrease (Figure 3B). We reason that the short chains
in P3 were both more efficiently capped by the terminal alkene and
may require fewer cascade steps to fully depolymerize than the
long chains, thus resulting in a more rapid consumption of these
chains in the population (Figure S11). To investigate if the alkene
configuration influences the reactivity of the depolymerization, the
(Z)-monomer 17b was polymerized with 16, and the resulting pol-
ymer was further capped with allyl bromide following the same syn-
thetic procedure of P3 (Table 1, entry 9). The enyne SIP P4 gen-
erated from this reaction (79% of the chains capped) exhibited a
reactivity of depolymerization similar to P3, reaching complete con-
version over 30 minutes with 3% G3 (Figure S12).
Scheme 3. IEG synthesis of discrete SIP oligomers
P3 demonstrated excellent stability against spontaneous, un-
controlled degradation in strong acid, base, and nucleophiles. No
structural changes were observed after P3 was treated by trifluoro-
acetic acid, sodium hydroxide, benzyl amine, or 1-dodecanethiol.
After these treatments, P3 maintained the same reactivity toward
depolymerization (Figure S13–S16). P3 also exhibited an excellent
o
thermal stability with a decomposition temperature (Td) of 265 C
determined by thermogravimetric analysis (Figure S17). Further-
more, we reason that the integration of the enyne SIPs with stimuli-
responsive metathesis catalysts[21] would expand the utility of this
technology in materials science. To this end, we incubated P3 with
a temperature-dependent catalytic system consisting of 3 mol%
second-generation Grubbs catalyst (G2) and 3 mol% tributyl phos-
phite inhibitor.[22] P3 was found to be stable when G2 was inhibited
at room temperature (Figure S18). Upon heating the system to 100
°C, the inhibition was reversed and the depolymerization pro-
ceeded with >95% conversion over just 95 seconds (Figure 3C and
Figure S19). This result not only showcased the potential of the
enyne SIPs as stimuli-responsive materials, but also suggested
that efficiency of depolymerization could increase significantly at
elevated temperatures.
In conclusion, a novel class of enyne SIPs capable of metathesis
cascade-triggered depolymerization was developed. Studies on
model compounds identified a 1,6-enyne motif with the p-pentyl-N-
benzenesulfonamide spacer as the optimal backbone structure,
which was readily incorporated into the SIPs via either polyconden-
sation or IEG approaches. The resulting SIPs were stable in strong
acid, base, nucleophiles, or at elevated temperatures, and can un-
dergo efficient depolymerization when triggered by metathesis cat-
alysts. Further investigation proved the importance of the terminal
alkene for improving the efficiency of depolymerization. This work
established a general approach to stimuli-responsive materials ca-
pable of controlled release of heterocyclic reagents with high effi-
ciency, laying the foundation for future applications in biosensing
and drug delivery.
Depolymerization of P2 was notably more efficient than P1,
reaching full conversion over 40 minutes with 1.5 mol% G3 (Figure
S6). Monitoring the depolymerization of P1 and P2 by SEC re-
vealed different modes of depolymerization. Notably, Mn of P1
gradually decreased as conversion increased (Figure 2A and Fig-
ure S7). In contrast, Mn of P2 remained constant over the course
of depolymerization (Figure 2B and Figure S8). We reasoned that
the differences between P1 and P2 in the depolymerization could
be explained by how the reactions were initiated. The depolymeri-
zation of P1 is initiated when G3 reacts with an internal alkene
within the polymer backbone, generating a “leading” segment with
a Ru-alkylidene chain end and a “lagging” segment with a 1,2-di-
substituted alkene chain end (Figure 2C). While the leading seg-
ment depolymerizes in a head-to-tail fashion, the lagging segment
must react with a second catalyst molecule to reinitiate. As a result,
the depolymerization of the lagging segment is less efficient, re-
quires higher catalyst loading, and leads to a gradual decrease of
Mn as the conversion increases. In contrast, the depolymerization
of P2 always initiates from the terminal alkene chain end (Figure
2D). These results highlight the importance of the terminal alkene
chain end in improving the efficiency of depolymerization for the
enyne SIPs.
To improve the depolymerization efficiency of the enyne SIPs
generated by polycondensation, we first prepared a polymer with
predominantly sulfonamide chain ends by employing a feeding ra-
tio of 16:17a = 1:0.975 (Table 1, entry 8). It should be noted that
although 16 was in excess, the resulting polymer is unlikely to pos-
sess 100% sulfonamide chain ends due to the intrinsic lack of con-
trol over polycondensation. Indeed, further reacting this polymer
with allyl bromide (1 equiv. with respect to 16) produced an enyne
SIP P3 with 74% of the chains capped by at least one terminal
alkene (Figure 3A and Figure S9). This synthetic procedure was
readily scalable to prepare gram quantities of P3, indicative of the
potential for further scaling up. Despite P3 being incompletely
capped, its depolymerization was notably more efficient than the
uncapped P1 (Figure S10), achieving complete depolymerization
over 30 minutes in the presence of 3 mol% G3. Interestingly, alt-
hough the amount of full-length P3 rapidly decreased after initia-
tion, Mn of the remaining polymer slightly increased initially before
Acknowledgements
We thank T. Jayasundera and M. Domin for their assistance in
NMR and MALDI-TOF. This work is supported by a CAREER
award from the National Science Foundation (CHE-1944512) and
a Beckman Young Investigator Award from the Arnold and Mabel
Beckman Foundation.
Conflict of interest
The authors declare no competing financial interest.
This article is protected by copyright. All rights reserved.

10.1002/anie.202108239
Angewandte Chemie International Edition
Figure 2. Depolymerization of P1 and P2. (AB) Evolution of the SEC traces of P1 (A) and P2 (B). Insets: Mn with respect to conversion. (CD) Proposed depolymer-
ization mechanisms of P1 (C) and P2 (D).
Figure 3. (A) Synthesis and 1H NMR of P3. (B) Evolution of the SEC traces of the depolymerization of P3. Insets: Mn with respect to conversion. (C) The conversion-
time plot of thermally responsive depolymerization of P3.
This article is protected by copyright. All rights reserved.

10.1002/anie.202108239
Angewandte Chemie International Edition
Keywords: Self immolative polymer • Enyne metathesis • Cas-
cade reaction • Depolymerization • Stimuli-responsive materials
[1]
a) A. Sagi, R. Weinstain, N. Karton, D. Shabat, J. Am. Chem. Soc.
2008, 130, 5434-5435; b) R. E. Yardley, A. R. Kenaree, E. R.
Gillies, Macromolecules 2019, 52, 6342-6360; c) Y. Xiao, X. Tan,
Z. Li, K. Zhang, J. Mater. Chem. B 2020, 8, 6697-6709; d) G. I.
Peterson, M. B. Larsen, A. J. Boydston, Macromolecules 2012, 45,
7317-7328.
[2]
[3]
A. P. Esser-Kahn, N. R. Sottos, S. R. White, J. S. Moore, J. Am.
Chem. Soc. 2010, 132, 10266-10268.
a) S. Gnaim, D. Shabat, J. Am. Chem. Soc. 2017, 139, 10002-
10008; b) Y. Xiao, Y. Li, B. Zhang, H. Li, Z. Cheng, J. Shi, J. Xiong,
Y. Bai, K. Zhang, ACS Macro Lett. 2019, 8, 399-402.
a) M. G. Olah, J. S. Robbins, M. S. Baker, S. T. Phillips,
Macromolecules 2013, 46, 5924-5928; b) K. Yeung, H. Kim, H.
Mohapatra, S. T. Phillips, J. Am. Chem. Soc. 2015, 137, 5324-
5327.
a) C. Aso, S. Tagami, J. Polym. Sci., Part B: Polym. Lett. 1967, 5,
217-220; b) J. A. Kaitz, C. E. Diesendruck, J. S. Moore, J. Am.
Chem. Soc. 2013, 135, 12755-12761; c) C. E. Diesendruck, G. I.
Peterson, H. J. Kulik, J. A. Kaitz, B. D. Mar, P. A. May, S. R. White,
T. J. Martinez, A. J. Boydston, J. S. Moore, Nat. Chem. 2014, 6,
623-628.
[4]
[5]
[6]
[7]
a) B. Fan, J. F. Trant, A. D. Wong, E. R. Gillies, J. Am. Chem. Soc.
2014, 136, 10116-10123; b) B. Fan, J. F. Trant, E. R. Gillies,
Macromolecules 2016, 49, 9309-9319.
A. Alouane, R. Labruere, T. Le Saux, F. Schmidt, L. Jullien, Angew.
Chem. Int. Ed. 2015, 54, 7492-7509.
[8]
[9]
[10]
A. M. DiLauro, S. T. Phillips, Polym. Chem. 2015, 6, 3252-3258.
S. T. Phillips, A. M. DiLauro, ACS Macro Lett. 2014, 3, 298-304.
a) T. J. Katz, S. J. Lee, M. Nair, E. B. Savage, J. Am. Chem. Soc.
1980, 102, 7940-7942; b) T. J. Katz, E. B. Savage, S. J. Lee, M.
Nair, J. Am. Chem. Soc. 1980, 102, 7942-7944.
[11]
[12]
M. Serra, E. Bernardi, L. Colombo, Synthesis 2020, 53, 785-804.
a) H. Park, T. L. Choi, J. Am. Chem. Soc. 2012, 134, 7270-7273;
b) H. Park, H. K. Lee, T. L. Choi, J. Am. Chem. Soc. 2013, 135,
10769-10775; c) A. Bhaumik, G. I. Peterson, C. Kang, T. L. Choi,
J. Am. Chem. Soc. 2019, 141, 12207-12211.
[13]
[14]
a) W. R. Gutekunst, C. J. Hawker, J. Am. Chem. Soc. 2015, 137,
8038-8041; b) T. Q. Zhang, L. B. Fu, W. R. Gutekunst,
Macromolecules 2018, 51, 6497-6503; c) X. Sui, T. Zhang, A. B.
Pabarue, L. Fu, W. R. Gutekunst, J. Am. Chem. Soc. 2020.
a) S. Pal, F. Lucarini, A. Ruggi, A. F. M. Kilbinger, J. Am. Chem.
Soc. 2018, 140, 3181-3185; b) M. Yasir, P. Liu, I. K. Tennie, A. F.
M. Kilbinger, Nat. Chem. 2019, 11, 488-494.
[15]
[16]
[17]
W. J. Neary, J. G. Kennemur, ACS Macro Lett. 2018, 8, 46-56.
S. T. Diver, A. J. Giessert, Chem. Rev. 2004, 104, 1317-1382.
N. Choony, N. Kuhnert, P. G. Sammes, G. Smith, R. W. Ward, J.
Chem. Soc., Perkin Trans. 2002, 1999-2005.
[18]
[19]
[20]
S. H. Kim, N. Bowden, R. H. Grubbs, J. Am. Chem. Soc. 1994,
116, 10801-10802.
H. Clavier, A. Correa, E. C. Escudero-Adan, J. Benet-Buchholz, L.
Cavallo, S. P. Nolan, Chem-Eur. J. 2009, 15, 10244-10254.
a) J. C. Barnes, D. J. Ehrlich, A. X. Gao, F. A. Leibfarth, Y. Jiang,
E. Zhou, T. F. Jamison, J. A. Johnson, Nat. Chem. 2015, 7, 810-
815; b) F. A. Leibfarth, J. A. Johnson, T. F. Jamison, Proc. Natl.
Acad. Sci. U.S.A. 2015, 112, 10617-10622.
[21]
[22]
A. J. Teator, C. W. Bielawski, J. Polym. Sci. Pol. Chem. 2017, 55,
2949-2960.
I. D. Robertson, L. M. Dean, G. E. Rudebusch, N. R. Sottos, S. R.
White, J. S. Moore, ACS Macro Lett. 2017, 6, 609-612.
5
This article is protected by copyright. All rights reserved.

10.1002/anie.202108239
Angewandte Chemie International Edition
Entry for the Table of Contents
A novel class of enyne self-immolative polymers (SIPs) are de-
veloped to enable cascade depolymerization triggered by me-
tathesis catalysts. Stable in strong acid, base, nucleophiles, and
at elevated temperatures, these enyne SIPs can undergo effi-
cient depolymerization in response to external stimuli. This work
established a general approach to stimuli-responsive materials
capable of controlled release of heterocyclic reagents.
Researcher Twitter username: @niu_group
5
This article is protected by copyright. All rights reserved.