Identifying and explaining the regioselectivity of alkylation of 1,2,4-triazole-3-thiones using NMR, GIAO and DFT methods

Article abstract of DOI:10.1016/j.molstruc.2020.128973

The regioselectivity of the alkylation of four 1,2,4-triazole-3-thiones with eight different organic halides was determined by the comparison of experimentally observed NMR chemical shifts of the product molecules to those predicted by density functional

Full text of DOI:10.1016/j.molstruc.2020.128973

Journal Pre-proof
Identifying and explaining the regioselectivity of alkylation of
1,2,4-triazole-3-thiones using NMR, GIAO and DFT methods
Maksym Fizer , Mikhailo Slivka , Nataliya Korol , Oksana Fizer
PII:
S0022-2860(20)31297-7
DOI:
Reference:
https://doi.org/10.1016/j.molstruc.2020.128973
MOLSTR 128973
To appear in:
Journal of Molecular Structure
Received date:
Revised date:
Accepted date:
17 June 2020
10 July 2020
24 July 2020
Please cite this article as: Maksym Fizer , Mikhailo Slivka , Nataliya Korol , Oksana Fizer ,
Identifying and explaining the regioselectivity of alkylation of 1,2,4-triazole-3-thiones
using NMR, GIAO and DFT methods, Journal of Molecular Structure (2020), doi:
https://doi.org/10.1016/j.molstruc.2020.128973
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier B.V.

Highlights
 alkylation of four disubstituted 1,2,4-triazole-3-thiones with eight alkyl
halides
1
13
 H and C NMR experimental spectra were compared with GIAO-
computed ones

13C NMR experimental and GIAO spectra indicate the alkylation of the
sulfur atom
 HOMO, ELF, electrophilic Fukui function were analyzed to explain the
regioselectivity
 different partial charges were studied for the correct regioselectivity
prediction

Identifying and explaining the regioselectivity of alkylation of 1,2,4-triazole-3-
thiones using NMR, GIAO and DFT methods
MaksymFizer*, Mikhailo Slivka, Nataliya Korol, Oksana Fizer
Department of Organic chemistry, Faculty of chemistry, Uzhhorod national
university, Pidgirna Str. 46, 88000 Uzhgorod, Ukraine
*correspondence author, e-mail: max.fizer@uzhnu.edu.ua
Abstract
The regioselectivity of the alkylation of four 1,2,4-triazole-3-thiones with eight
different organic halides was determined by the comparison of experimentally
observed NMR chemical shifts of the product molecules to those predicted by
density functional theory (DFT) calculations via gauge independent atomic orbital
(GIAO) method. The combination of the employed reactants resulted in ten
different model alkylated triazoles, seven of which are new and not previously
described. The reaction was performed in neutral and alkaline medium with the
observation that S-alkylation occurs selectively and with slightly lower yields
under neutral conditions. Highest occupied molecular orbitals, electron
localization function, electrophilic Fukui function, and different types of partial
charges were considered as reactivity descriptors to reveal the observed
regioselectivity and to explain the structure of synthesized products. In
1
13
conclusion, the comparison of the chemical shifts of H and C NMR of the α-
methylene group of the products with those calculated by incorporating the GIAO
DFT-computed isotropic chemical shielding gives an approach for the correct and
reliable determination of the site of alkylation as the S-atom of the synthesized S-
alkylated 1,2,4-triazoles.

Keywords: Alkylation; Electron localization function; Fukui function; GIAO;
Regioselectivity; Triazole.
1. Introduction
S- and N-alkylated 1,2,4-triazole-containing systems have become
increasingly interesting due to their pharmacological potential. Two widely used
medical drugs, namely lesinurad [1] and thiotriazolinone [2] contain a
carboxymethylsulfonyl fragment, and are synthetically available through the S-
alkylation of corresponding 1,2,4-triazole-3-thiones.
In recognition of the relationship of structure and pharmacological action,
the determination of the molecular structure of azoles is an important subject of
organic chemistry. The search for new methods for the derivatization of 1,2,4-
triazole-containing systems via the alkylation reaction, as well as reliable routes to
solve their structure are important tasks due to the increasing interest in the
obtained S(N)-alkylated products and the possibility for their further
transformation [3-6].
The most recent approaches of S-alkylation of 1,2,4-triazole-3-thiol are
described as the interaction between the mentioned heterocyclic system and
corresponding alkylating agents: alkyl/alkenyl/alkynyl/benzyl halides [3-8],
halogen-substituted carboxylic acids [9-11], halogenacetamides [10, 11],
haloalkanoic acid esters [10-12] and chloroacetophenones [13]. The alkylation
protocols differ in the solvents (alcohol, acetone, water) and bases (KOH, K2CO3,
pyridine) used and influence the regioselectivity in the performed reactions. In
most cases, 1,2,4-triazole thioethers will be obtained as target products if the
sulfur atom is not yet alkylated [4-8]. However, together with the formation of S-
alkylated product some authors described the formation of N-alkylated
derivatives as well. Thus, Samvelyan et al [12] described that alkylation of S-
unprotected 1,2,4-triazole-3-thiol with methyl 3-bromopropanoate gave an

inseparable mixture of S- and 2-N-alkylation products. According to Boraei et al
[7] the 2-N-allylated product was obtained after fusing of allylsulfanyl-4-amino-
1,2,4-triazole in the absence of either solvent or catalyst. The alkylation of S-
substituted 1,2,4-triazoles in the presence of K2CO3 gave a mixture of two
products of which the S-, 2-N-alkylated triazole was formed in higher yields than
S-, 1-N-alkylated triazoles [8].
Investigations on some biological properties of S-(N)-alkylated 1,2,4-
triazole-3-thiols were also reported [6, 10, 11, 13]. The antiproliferative assay
against hepatic and breast cancer cell lines preliminary showed promising
antitumor activity of S- and 1-N-alkylated derivatives of 1,2,4-triazole-3-thiol
compared to the standard drug, doxorubicin [6]. The results of the effects of the
compounds on the DNA methylation level revealed that S-alkylated analogues
possessed a greater demethylating activity compared to the low activity of 1-N-
substituted compounds [10]. According to preliminary screening results, five of
the six tested S-alkylated compounds exhibited antiulcer activity at the level of
the reference drug ranitidine [13]. S-alkylated triazoles also showed a moderate
to good activity against gram-positive (Bacillus cereus, Staphylococcus aureus) and
gram-negative (Escherichia coli, Pseudomonas aeruginosa) bacteria [11].
NMR technique is one of the most powerful methods for resolving the
molecular structure of organic compounds and may be successfully used for the
investigation of the spatial structure of peptides [14], for studying small organic
guest molecules physisorbed on different mesoporous silicas [15], for the
exploration of functionalized N-phenylbenzamides [16], and for enantiomer
discrimination of cathinones [17]. Moreover, combined NMR and DFT GIAO
investigations were performed for the determination of structures of galantamine
derivatives [18] and organosulfur compounds [19]. Such combined studies have
also been carried out for compounds containing the 1,2,4-triazole ring [20, 21].

A description of various methods for the prediction of NMR parameters is
given by Toukach and Ananikov [22]. Among many available methods, the
theoretical approaches based on the DFT calculations might be considered as the
most affordable and accurate, especially in cases of new systems for which
methods that use structure-based parameters can be applied only with limited
accuracy due to the absence of appropriate parameterization. One of the most
routinely used methods to calculate chemical shifts is the gauge independent
atomic orbital (GIAO) method [23], which is satisfactory in the prediction of
chemical shifts values due to the adequate calculation of isotropic chemical
shielding for different nuclei [24-26].
1
13
In the present investigation, the DFT computed H, and C chemical shifts
of various possible isomers of alkylated 1,2,4-triazole-3-thiones are compared
with experimental values to reliably determine the structure of the obtained
alkylated products and to understand the regioselectivity of the performed
alkylation reactions. The numerous reactivity descriptors, namely HOMO, electron
localization function, electrophilic Fukui function, and various atomic partial
charges have been computed to explain the selectivity of the electrophilic attack
on the 1,2,4-triazole-3-thione system. In the course of the investigation, the
13
inclusion of C NMR analysis was found necessary for the correct resolving of the
structure of the alkylated products.
2. Experimental
The melting points were measured on the Stuart SMP30 instrument. Elemental
1
composition determination was performed on Elementar Vario MICRO cube. H
13
NMR (400 MHz) and C NMR (100 MHz) spectra were recorded on Varian VXR
400. Samples were studied in deuterated dimethylsulfoxide (DMSO) solutions.
Tetramethylsilane was used as a standard in both experimental and
computational studies.

2.1. Synthesis
The synthetic procedures for the starting materials, 4-allyl-5-allylamino-2,4-
dihydro-3H-1,2,4-triazole-3-thione a [27], 4-phenyl-5-phenylamino-2,4-dihydro-
3H-1,2,4-triazole-3-thione b [28], 4-phenyl-5-amino-2,4-dihydro-3H-1,2,4-triazole-
3-thione c [29], and 4,5-diphenyl-2,4-dihydro-3H-1,2,4-triazole-3-thione d [30] are
described in the corresponding literature.
Different halogen-containing hydrocarbons were used as alkylation agents
in this study, namely: benzyl chloride (α-chlorotoluene, CAS# 100-44-7), propargyl
bromide (3-bromo-1-propyne, CAS# 106-96-7), allyl bromide (3-bromo-1-
propene, CAS# 106-95-6), cinnamyl chloride ((3-chloropropenyl)benzene, CAS#
2687-12-9), methyl iodide (iodomethane, CAS# 74-88-4), methallylchloride (3-
chloro-2-methyl-1-propene, CAS# 563-47-3), crotonylbromide (trans-1-bromo-2-
butene, CAS# 29576-14-5), and 4-bromo-1-butene (CAS# 5162-44-7). All used
reagents were purchased from commercial suppliers and were of "reagent grade"
purity.
General alkylation of 1,2,4-triazole-3-thione. Method 1
4,5-Disubstituted-2,4-dihydro-3H-1,2,4-triazole-3-thione a-d (10.0 mmol)
and potassium hydroxide (12.0 mmol) were added to 20 mL ethanol and heated
to give a clear solution. The halogen-containing alkylation agent (12.0 mmol) was
added in 5 mL ethanol to the cooled solution of the triazole. The mixture was
refluxed for 1-2 h. After cooling, the precipitated product was filtered, washed
with deionized water and dried in vacuo. In the case of a small quantity of the
precipitated product, the solvent was removed under reduced pressure to form
the target precipitate. Substances were purified by crystallization from ethanol.
General alkylation of 1,2,4-triazole-3-thione. Method 2

4,5-Disubstituted-2,4-dihydro-3H-1,2,4-triazole-3-thione a-c (10.0 mmol)
and halogen-containing alkylation agent (12.0 mmol) were added to 50 mL
ethanol, and the reaction mixture was refluxed for 3 h. After cooling, a solution of
potassium hydroxide (30.0 mmol) in 30 ml water was added to the homogenous
reaction mixture which caused the immediate precipitation of the target product
3. The precipitate was filtered, washed with deionized water, and finally
recrystallized from ethanol to obtain the pure substance.
5-(Benzylsulfanyl)-N,4-diallyl-4H-1,2,4-triazol-3-amine (3a).
Synthesized from 4-allyl-5-allylamino-4H-1,2,4-triazol-3-thione a and benzyl
chloride. White powder with mp 127-128 °C (from ethanol). The yield is 2.69g
(94%) and 2.32g (81%) according to method 1 and method 2, respectively. The
elemental analysis: found: C, 63.1; H, 6.6; N, 19.5; S 11.0%; calc. for C15H18N4S: C,
62.9; H, 6.3; N, 19.6; S, 11.2%.
1H NMR (400 MHz, DMSO-d6) δ (ppm): 7.25-7.37 (m, 5H, C6H5), 6.50 (t, 1H,
J = 6.2 Hz, NH), 5.75-5.90 (m, 2H, 2-CH=), 5.05-5.29 (ddd, 4H, J =54.0, 41.4, 17.2
Hz, 2CH2=), 4.65 (d, 2H, J =3.3 Hz, CH2NH-), 4.39 (s, 2H, SCH2), 4.05 (s, 2H,
CH2Ntriazole).
13C NMR (125 MHz, DMSO-d6) δ (ppm): 152.3, 146.0, 136.6, 133.4, 130.6,
129.5, 129.0, 128.2, 118.7, 117.0, 45.7, 45.2, 37.0.
5-(Benzylsulfanyl)-N,4-diphenyl-4H-1,2,4-triazol-3-amine (3b)
Synthesized from 4-phenyl-5-phenylamino-4H-1,2,4-triazol-3-thione b and
benzyl chloride. White powder with mp 207-208 °C (from ethanol). The yield is
3.44g (96%) and 3.26g (91%) according to method 1 and method 2, respectively.
The elemental analysis: found: C, 70.3; H, 5.4; N, 15.4; S 8.8%; calc. for C21H18N4S:
C, 70.4; H, 5.1; N, 15.6; S, 9.0%.

1H NMR (400 MHz, DMSO-d6) δ (ppm): 8.28 (s, 1H, NH), 6.85-7.53 (m, 15H,
3C6H5), 4.22 (s, 2H, SCH2).
13C NMR (125 MHz, DMSO-d6) δ (ppm): 152.5, 145.9, 141.8, 137.5, 132.7,
130.2, 129.4, 129.0, 128.9, 128.3, 127.8, 120.9, 117.3, 37.2.
N,4-Diphenyl-5-(propargylsulfanyl)-4H-1,2,4-triazol-3-amine (3c)
Synthesized from 4-phenyl-5-phenylamino-4H-1,2,4-triazol-3-thione b and
propargyl bromide. White powder with mp 181 °C (from ethanol). The yield is
2.57g (84%) according to method 1. The elemental analysis: found: C, 66.7; H, 4.9;
N, 18.0; S 10.2%; calc. for C17H14N4S: C, 66.6; H, 4.6; N, 18.3; S, 10.5%.
1H NMR (400 MHz, DMSO-d6) δ (ppm): 8.36 (s, 1H, NH), 6.85-7.58 (m, 10H,
2C6H5), 3.79 (s, 2H, SCH2), 3.31 (s, 1H, HC≡C).
13C NMR (125 MHz, DMSO-d6) δ (ppm): 152.8, 144.8, 141.7, 132.7, 130.3,
129.0, 128.4, 121.0, 117.5, 79.8, 75.2, 21.7.
5-(Allylsulfanyl)-N,4-diphenyl-4H-1,2,4-triazol-3-amine (3d)
Synthesized from 4-phenyl-5-phenylamino-4H-1,2,4-triazol-3-thione b and
allylbromide. White powder with mp166-167 °C (from ethanol). The yield is 2.74g
(89%) according to method 1. The elemental analysis: found: C, 66.4; H, 5.5; N,
18.1; S 10.2%; calc. for C17H16N4S: C, 66.2; H, 5.2; N, 18.2; S, 10.4%.
1H NMR (400 MHz, DMSO-d6) δ (ppm): 8.26 (s, 1H, NH), 6.56-7.80 (m, 10H,
2C6H5), 5.58-6.04 (m, 1H, -CH=), 5.08 (dd, 2H, J = 38.3, 13.3 Hz, CH2=), 3.56 (d, 2H,
J = 5.9 Hz).
13C NMR (125 MHz, DMSO-d6) δ (ppm): 152.5, 145.5, 141.8, 133.8, 132.8,
130.3, 129.0, 128.4, 120.9, 118.8, 117.3, 35.7.
5-(Methylsulfanyl)-N,4-diphenyl-4H-1,2,4-triazol-3-amine (3e)

Synthesized from 4-phenyl-5-phenylamino-4H-1,2,4-triazol-3-thione b and
methyl iodide. White powder with mp 224-226 °C (from ethanol), lit. 225-229°C
[31]. The yield is 2.60 g (92%) according to method 1. The elemental analysis:
found: C, 63.7; H, 5.1; N, 19.8; S 11.3%; calc. for C15H14N4S: C, 63.8; H, 5.0; N, 19.8;
S, 11.4%.
1H NMR (400 MHz, DMSO-d6) δ (ppm): 8.33 (s, 1H, NH), 6.84-7.58 (m, 10H,
2C6H5), 2.50 (s, 3H, SCH3).
13C NMR (100 MHz, DMSO-d6) δ (ppm): 152.5, 147.5, 142.0, 132.8, 130.4,
130.3, 129.1, 128.2, 120.9, 117.3, 14.9.
5-(Cinnamylsulfanyl)-N,4-diphenyl-4H-1,2,4-triazol-3-amine (3f)
Synthesized from 4-phenyl-5-phenylamino-4H-1,2,4-triazol-3-thione b and
cinnamyl chloride. White powder with mp 200-201 °C (decompose) (from
ethanol). The yield is 3.65 g (95%) and 3.15g (82%) according to method 1 and
method 2, respectively. The elemental analysis: found: C, 72.1; H, 5.5; N, 14.4; S
8.1%; calc. for C23H20N4S: C, 71.9; H, 5.2; N, 14.6; S, 8.3%.
1H NMR (400 MHz, DMSO-d6) δ (ppm): 8.22 (s, 1H, NH), 6.82-7.55 (m, 15H,
3C6H5), 6.51 (d, 1H, J = 15.7 Hz, Ph-CH=), 6.16-6.33 (m, 1H, =CH-CH2S), 3.79 (d, 2H,
J = 7.0 Hz, SCH2).
13C NMR (125 MHz, DMSO-d6) δ (ppm): 152.6, 145.8, 141.9, 136.7, 133.3,
132.9, 130.2, 130.2, 129.1, 129.1, 128.5, 128.3, 126.8, 125.2, 120.9, 117.4, 36.0.
5-(Methallylsulfanyl)-N,4-diphenyl-4H-1,2,4-triazol-3-amine (3g)
Synthesized from 4-phenyl-5-phenylamino-4H-1,2,4-triazol-3-thione b and
methallyl chloride. White powder with mp 192-194 °C (from ethanol). The yield is
2.90 g (90%) according to method 1. The elemental analysis: found: C, 67.3; H,
5.9; N, 17.4; S 9.6%; calc. for C18H18N4S: C, 67.1; H, 5.6; N, 17.4; S, 9.9%.

1H NMR (400 MHz, DMSO-d6) δ (ppm): 8.15 (s, 1H, NH), 6.65-7.77 (m, 10H,
2C6H5), 4.81 (d, 2H, J = 13.8 Hz, CH2=), 3.56 (s, 2H, CH2S), 1.67 (s, 3H, CH3).
13C NMR (125 MHz, DMSO-d6) δ (ppm): 152.6, 145.6, 141.9, 140.8, 132.9,
130.3, 130.2, 129.1, 128.5, 121.0, 117.4, 115.1, 40.7, 21.1.
5-(Methylsulfanyl)-4-phenyl-4H-1,2,4-triazol-3-amine (3h)
Synthesized from 4-phenyl-4H-1,2,4-triazol-3-thione c and methyl iodide.
White powder with mp 144-146 °C (from ethanol). The yield is 2.90 g (90%)
according to method 1. The elemental analysis: found: C, 54.6; H, 5.8; N, 25.2; S
14.3%; calc. for C10H12N4S: C, 54.5; H, 5.5; N, 25.4; S, 14.6%.
1H NMR (400 MHz, DMSO-d6) δ (ppm): 7.35-7.57 (m, 5H, C6H5), 5.71 (s, 2H,
NH2), 2.38 (s, 3H, SCH3).
13C NMR (100 MHz, DMSO-d6) δ (ppm): 156.3, 144.8, 133.2, 130.3, 129.8,
127.8, 15.5.
3-{[(2E)-But-2-en-1-yl]sulfanyl}-4,5-diphenyl-4H-1,2,4-triazole (3i)
Synthesized from 4,5-diphenyl-4H-1,2,4-triazol-3-thione d and (2E)-1-
bromobut-2-ene. White powder with mp 137-139 °C (from ethanol), lit. 137-139
°C [6]. The yield is 2.82 g (92%) according to method 1. The elemental analysis:
found: C, 70.5; H, 5.9; N, 13.6; S 10.2%; calc. for C18H17N3S: C, 70.3; H, 5.6; N, 13.7;
S, 10.4%.
1H NMR (400 MHz, DMSO-d6) δ (ppm): 7.29-7.60 (m, 10H, 2C6H5), 5.64-5.70
(m, 1H, –CH=), 5.43-5.59 (m, 1H, –CH=), 3.77 (dd, 2H, J = 28.6, 7.1 Hz, SCH2), 1.61
(s, 3H, CH3).
13C NMR (100 MHz, DMSO-d6) δ (ppm): 154.7, 151.9, 134.3, 130.4, 130.1,
129.8, 129.2, 129.0, 128.8, 128.2, 127.1, 126.2, 34.8, 18.0.
3-[(But-3-en-1-yl)sulfanyl]-4,5-diphenyl-4H-1,2,4-triazole (3j)

Synthesized from 4,5-diphenyl-4H-1,2,4-triazol-3-thione d and 4-bromo-1-
butene. White powder with mp133-135 °C (from ethanol), lit. 133-135 °C [6]. The
yield is 2.86 g (93%) according to method 1. The elemental analysis: found: C,
70.2; H, 5.8; N, 13.4; S 10.1%; calc. for C18H17N3S: C, 70.3; H, 5.6; N, 13.7; S, 10.4%.
1H NMR (400 MHz, DMSO-d6) δ (ppm): 7.44-7.58 (m, 10H, 2C6H5), 5.79 (m,
1H, –CH=), 4.98-5.14 (m, 2H, =CH2), 3.22 (t, 2H, J = 7.2 Hz, SCH2), 2.45 (d, 2H, J =
7.1, CH2).
13C NMR (100 MHz, DMSO-d6) δ (ppm): 154.8, 152.3, 136.6, 134.4, 130.5,
130.4, 130.2, 129.0, 128.3, 128.2, 127.1, 117.1, 33.5, 31.7.
2.2. Computational methods
The choice of the energetically preferred rotamers of all studied
compounds, except those resolved with the XRD technique, was based on
conformers’ search through the systematic rotor engine in the Avogadro [32]
program. The MMFF94 force field [33] was applied in these calculations. Then,
the most stable rotamer was re-optimized with DFT method.
The gas phase geometry optimizations were performed at the PBE/6-
31G(d) level in the PRIRODA program [34-36]. Geometry optimization was
followed by a Hessian calculation to verify that the obtained geometries
correspond to the true energy minima, as there were no imaginary frequencies
present.
The GIAO approach was used to calculate the NMR chemical shifts using the
PBE/6-31G(d) and PBE/λ3 methods [37] in the PRIRODA program and with the
B3LYP/6-31G(d) and B3LYP/6-31G(d,p) methods [38] in the ORCA 4.2.1 program
[39]. In case of the PBE functional, gas phase NMR calculations were performed,
whereas in the case of GIAO-B3LYP calculations, the DMSO solvent influence was
modelled via the CPCM solvation model [40].

Because the PBE0 hybrid functional is known to be suitable for the
modelling of electron densities [41, 42], the electron density of the starting
substrate species was calculated at the PBE0/6-311++G(d,p) level of theory.
Electrophilic Fukui functions [43], as well as electron localization functions [44]
were generated with the Multiwfn 3.6 program [45]. The capability of the correct
regioselectivity prediction was calculated for numerous types of partial charges,
namely: Mulliken [46], Löwdin [47], CHELPG [48], and Hirshfeld [49] partial
charges were computed in the ORCA program, whereas the extensions of Hishfeld
population analysis, like CM5 [50] and ADCH [51] partial charges were calculated
in CM5charges [52] and in Multiwfn programs, respectively. Natural population
analysis [53] and calculation of respective partial charges have been carried out in
JANPA code [54]. Visualization of structures and isosurfaces, as well as figures
rendering was accomplished using the UCSF Chimera software [55]. Density fitting
[56] and resolution of identity with chain-of-spheres techniques [57, 58] were
enabled where possible to speed-up the calculations.
3. Results and Discussion
3.1. Synthesis and NMR study
Alkylation was performed in two principally different manners. In the first
method, the thiones a-d were dissolved in alkaline medium, which causes a
proton elimination from the nitrogen in 2-position of the triazole ring with further
transformation into thiolate anion (Scheme 1) as the nucleophilic reactant for the
subsequent alkylation step. In the second method, the alkylation agent reacts
with the neutral 1,2,4-triazole species, which, due to tautomerism, can exist as a
mixture of the thione and thiol form, whereby the thione form exhibits a higher
stability than the thiol form in neutral medium according to numerous theoretical
and experimental works [59-62]. The hydrogen halide, which was formed during

the alkylation reaction, was then neutralized with potassium hydroxide in the
next step.
In theory, alkylation can take place at different reaction centers of the
model 1,2,4-triazole-3-thions a-d (see Scheme 1). In the case of alkylation of the
nitrogen in 1-position of the triazole ring, the zwitterionic structures 1a-j can be
formed. In addition, 5-R-amino substituted triazole-3-thiones a-c can react to the
alternative structures 4a-h because of the existing tautomerization in the
guanidine fragment. Due to the two anionic mesomeric structures of triazole-1-
ide (anion-1) and triazole-3-thiolate (anion-2), the electrophilic attack of alkyl
halides can lead to the formation of compounds 2a-j or 3a-j, respectively. Finally,
the presence of the exocyclic NH-fragment bound to the 5-position of the triazole
ring, unlocks the possibility of attachment of the alkyl group to the exocyclic
nitrogen and formation of structures 5a-h.
While acknowledging the possible thione-thiol tautomerism for the
structures 4a-h and 5a-h, due to lower stability of thiols, only thionic forms were
considered in the NMR calculations.

Scheme 1. Alkylation of model thiones a-d with all possible alternative products.
To correctly resolve the structure of the obtained products, and to
1
13
determine the regiodirection of the alkylation reaction, the H and C spectra
were recorded. For the assignment of the experimentally observed peaks, it was
decided to compare these with GIAO DFT-based calculated NMR chemical shifts
of all proposed alternative structures 1-5. During analysis of the experimental and
theoretical NMR spectra, we have focused mainly on the signals of the α-
methylene, or methyl in the case of methyl iodide as the alkylating agent. These
groups are highlighted in blue in Scheme 1.
As the first step, a geometry optimization of all proposed structures was
carried out. For geometry optimization and related vibrational frequencies
calculations, the GGA PBE functional was used, which gave relatively good
performance/accuracy ratio in our previous works [63-65]. Here we also have to
mention a good performance of PBE functional for distinguishing conformers'
relative stability [66, 67]. Considering that 46 possible products and 12 structures
of the triazole substrates (three for each triazole a-d) needed to be optimized, the
smallest possible but still sufficiently accurate basis set was used. Therefore, we
have benchmarked triple-zeta 6-311G(d,p) and double-zeta 6-31G(d), 6-31G(d,p)
and 6-31+G(d,p) basis sets over two triazole structures resolved with single crystal
X-ray diffraction. The 4-phenyl-5-phenylamino-1,2,4-triazole-3-thione [28], which
is marked as thione b in this study, and 3-(methylsulfanyl)-4-phenyl-5-
(phenylamino)-4H-1,2,4-triazole-1-ium cation, which is protonated form of
structure 3e, as its hexabromotellurate salt [31], were used as the reference
structures.
The correlation between PBE DFT calculated and experimental values of
interatomic distances and bond angles is shown in Table 1. For the calculated
interatomic distances, in the case of the gas-phase calculations, the correlation

coefficient, R2, systematically marginally increases and equals 0.9553, 0.9557,
0.9571, 0.9586 for the tested 6-31G(d), 6-31G(d,p), 6-31+G(d,p), 6-311G(d,p)
basis set, respectively. An analogous increase of R2 can be observed in the case of
the implicit CPCM solvation model. A similar observation is made for the bond
angles, where in the case of gas-phase computations, the correlation coefficients
equal 0.9658, 0.9660, 0.9663, and 0.9688 for the 6-31G(d), 6-31G(d,p), 6-
31+G(d,p), 6-311G(d,p) basis sets, respectively. It must be noticed the slight
systematic increase of R2 in the case of CPCM interatomic distances in comparison
to the gas-phase data. Oppositely, in the case of bond angles, the correlation
coefficient is higher for the gas-phase calculations.
As a result, only marginal improvement of bond length and bond angles can
be obtained through shifting from the double-zeta 6-31G(d) basis set to more
computationally expensive basis sets. Although the use of the CPCM solvation
model leads to an improvement of interatomic distances, it simultaneously entails
deterioration of bond angles values and is more time-consuming, especially for
calculation of Hessian. For this reason, the optimization of all geometries was
done in a gas phase with more computationally affordable 6-31G(d) basis set.
Moreover, using gas-phase geometries computed at a lower level of theory for
further calculation NMR chemical shifts at a higher level is a well-validated
strategy [68-71].
Table 1. Comparison of PBE DFT calculated and XRD experimental geometrical
parameters: interatomic distances and bond angles. S – slope, I – intercept, R2 –
correlation coefficient.
Interatomic distances
Bond angles
I
Correlation
Basis set
S
I
1.0333 -0.0740
1.0342 -0.0750
1.0386 -0.0828
R2
S
R2
6-31G(d)
0.9553 0.9571 4.9703 0.9658
0.9557 0.9583 4.8275 0.9660
0.9571 0.9633 4.2268 0.9663
XRD
versus
6-31G(d,p)
Gas phase
6-31+G(d,p)

6-311G(d,p)
6-31G(d)
1.0286 -0.0648
1.0006 -0.0279
1.0012 -0.0286
1.0059 -0.0361
1.0008 -0.0273
0.9644 0.0502
0.9642 0.0504
0.9652 0.0496
0.9693 0.0416
0.9586 0.9642 4.1210 0.9688
0.9603 0.9567 5.0039 0.9647
0.9605 0.9575 4.9103 0.9646
0.9607 0.9570 4.9580 0.9619
0.9620 0.9552 5.1725 0.9632
0.9970 0.9976 0.2700 0.9949
0.9969 0.9973 0.3009 0.9950
0.9971 0.9926 0.8600 0.9938
0.9961 0.9897 1.2097 0.9922
XRD
6-31G(d,p)
6-31+G(d,p)
6-311G(d,p)
6-31G(d)
versus
CPCM
DMSO
Gas phase
versus
6-31G(d,p)
6-31+G(d,p)
6-311G(d,p)
CPCM
DMSO
Isotropic chemical shielding values were calculated at different levels of
theory to determine which level can be routinely used for resolving triazole
structures. Gas-phase calculations with PBE functional were performed in
combination with a minimal 6-31G(d) basis set and a large quadruple-zeta basis
set λ3, which is analog to the cc-pVQZ basis. Since in the PRIRODA quantum
chemistry code, calculations with the PBE functional are extremely fast, this
method is computationally attractive even in combination with relatively heavy
basis sets. Moreover, it was shown that gas-phase PBE calculations can be
1
accurate in GIAO modelling of H and 13C NMR chemical shifts [72, 73]. It is worth
mentioning that in the case of the presence of solute-solvent strong hydrogen
bonds, the inclusion of explicit solvent molecules is essential [12, 74, 75].
However, this is beyond the scope of the present work, and only the implicit
solvent model was applied.
B3LYP calculations were combined with 6-31G(d) and 6-31G(d,p) basis sets.
The CPCM implicit solvation model was enabled for the inclusion of the DMSO
solvent effects. The choice of hybrid B3LYP functional was influenced by previous
investigations, where authors found a good prediction capability of this method
for triazole systems [21]. In addition, it was demonstrated that hybrid functionals
even when combined with relatively small basis sets like 6-31G(d,p) or with lighter

6-31G(d) may provide superior correlation between experimental and GIAO
computed 1H NMR isotropic chemical shielding [76].
Table 2.Correlation between experimental and GIAO computed 1HNMR chemical
shifts (in ppm) of the protons attached to the α-carbon of the alkyl group
introduced to the triazole core. S – slope, I – intercept, R2 – correlation
coefficient, MAE –mean absolute error, RMSD – root-mean-square deviation
a
b
c
d
e
f
g
h
i
j
S
I
R2 MAE RMSD
4.39 4.22 3.79 3.6 2.5 3.79 3.56 2.38 3.77 3.22
-
-
-
-
-
EXP
1
4.59 4.57 4.41 4.15 3.15 4.28 4.00 3.18 4.30 3.70 1.227 -1.425 0.970 0.511 0.535
4.95 5.06 4.80 4.62 3.50 4.74 4.51 2.71 4.56 4.01 0.853 -0.185 0.912 0.824 0.851
4.45 4.50 4.12 4.32 2.50 4.49 4.20 2.88 4.25 3.37 0.845 0.221 0.868 0.387 0.461
2
3
3.90 4.23 3.56 3.46 2.64 3.60 3.68 2.87 -
4.15 4.24 3.36 3.77 2.46 4.05 3.72 2.17 -
- 1.330 1.116 0.892 0.226 0.278
- 0.892 0.418 0.905 0.192 0.228
4
5
4.93 4.95 4.85 4.42 3.31 4.59 4.29 3.39 4.72 4.03 1.065 -1.108 0.951 0.826 0.838
5.38 5.53 5.41 5.08 3.84 5.22 4.87 3.02 4.93 4.28 0.771 -0.143 0.896 1.235 1.264
4.65 4.70 4.46 4.59 2.64 4.78 4.48 3.23 4.48 3.47 0.790 0.244 0.814 0.626 0.698
1
2
3
4.30 4.61 3.99 3.81 2.85 3.98 4.08 3.05 -
4.50 4.51 3.66 3.93 2.56 4.32 3.93 2.25
- 1.168 -0.950 0.916 0.328 0.375
- - 0.822 0.480 0.935 0.244 0.288
4
5
4.92 5.96 4.61 4.39 3.46 4.51 4.31 3.47 4.58 4.01 0.810 0.057 0.792 0.900 0.953
5.04 5.17 4.91 4.69 3.68 4.82 4.60 3.55 4.67 4.14 1.160 1.720 0.960 1.005 1.016
4.42 4.49 4.07 4.24 2.59 4.42 4.16 2.53 4.18 3.30 0.831 0.331 0.905 0.319 0.391
1
2
3
4.20 4.38 3.72 3.62 2.81 3.77 3.80 3.05 -
4.21 4.40 3.44 3.90 2.65 4.19 3.84 2.25 -
- 1.331 1.357 0.921 0.209 0.289
- 0.882 0.345 0.883 0.245 0.262
4
5
4.96 4.87 4.74 4.44 3.50 4.02 4.34 3.53 4.67 4.09 1.169 -1.522 0.865 0.794 0.831
5.12 5.25 5.11 4.82 3.79 4.95 4.67 3.36 4.76 4.20 1.015 -1.150 0.929 1.080 1.092
4.44 4.50 4.16 4.33 2.64 4.52 4.25 3.17 4.25 3.29 0.892 -0.006 0.816 0.433 0.514
1
2
3
4.22 4.48 3.87 3.71 2.87 3.86 3.91 3.08 -
4.25 4.44 3.53 3.92 2.67 4.21 3.85 2.26 -
- 1.300 -1.345 0.923 0.263 0.328
- 0.882 0.315 0.899 0242 0.260
4
5
1
The comparison of experimentally observed and computed H NMR
chemical shifts is summarized in Table 2. A direct comparison of linear correlation

between the experimental and computed values in terms of correlation
coefficient R2, slope, and intercept would be the wrong way to proceed, as, at this
point, the structure of obtained products is unknown. The correct approach is to
consider the deviations of computed versus experimental values. For this
strategy, the mean absolute errors (MAEs), and root-mean-square deviations
(RMSDs) were calculated with the result that for all computational methods,
structures 4 and 5 gave the best match with the experiment.
In the next step, the experimentally observed chemical shifts of other
protons need to be included. In the case of structures 4, the proton attached to
the second nitrogen atom must give a broad peak, which is absent in the
experimental spectrum of alkylated products. Structures 5a-g do not contain a
proton near the exocyclic nitrogen, whereas in all experimental spectra of the
obtained products, the signals of these protons are found in the range of 6.50-
8.36 ppm. Obviously, at this point, the predicted 1H NMR spectra are not useful in
resolving the structures of the alkylated products. Which is due to very close
values of chemical shifts calculated for structures 3, 4, 5 (see Table 2).
Nevertheless, a few trends can be noted. In the case of PBE functional, increasing
the basis set size decreases the correlation in general. The MAE and RMSD values
are higher for PBE/λ3 than for PBE/6-31G(d), whereas the values of R2 are lower.
Similar, but less pronounced differences are present in the case of B3LYP
functional. In general, a slightly better correlation was obtained with the 6-31G(d)
basis set.
13
In a similar manner the C NMR chemical shifts were analyzed (see Table
3). The comparison of computed versus experimental chemical shift values show
that for all chosen theoretical levels, MAE and RMSD values are lowest in the case
of structures 3. In addition, the absolute values of intercepts are also the smallest
for this type of geometries. These computations clearly testify the formation of
thioethers 3a-h via the alkylation of the exocyclic sulfur atom. This result is in full

agreement with previous studies, where alkylation of the exocyclic thione sulfur
was proven by XRD technique [10, 31, 77].
Table 3. Correlation between experimental and GIAO computed 13C NMR
chemical shifts (in ppm) of the protons attached to the α-carbon of the alkyl
group introduced to the triazole core. S – slope, I – intercept, R2 – correlation
coefficient, MAE – mean absolute error, RMSD – root-mean-square deviation
a
b
c
d
e
f
g
h
i
j
S
I
R2 MAE RMSD
37.0 37.2 21.7 35.7 14.9 36.0 40.6 15.5 34.8 33.5
-
-
-
-
EXP
1
54.0 55.3 42.5 54.7 36.0 54.4 56.2 33.2 53.7 50.4 1.096 -23.07 0.976 18.3 16.6
53.1 52.6 38.3 52.2 34.2 52.4 54.4 34.0 52.1 50.4 1.139 -23.27 0.989 16.7 14.9
2
43.6 43.5 27.9 40.7 18.7 41.7 42.8 18.9 41.6 38.4 0.940 -2.95 0.978 5.1
4.6
- 1.249 -30.61 0.940 18.6 18.8
- 1.010 -20.15 0.934 19.6 19.8
3
60.0 50.3 41.8 54.0 37.8 54.2 55.2 37.0 -
53.7 55.5 45.6 55.1 36.0 59.0 57.9 32.9 -
4
5
51.4 53.0 39.6 52.6 30.8 52.1 54.5 27.7 51.8 48.5 0.952 -13.30 0.971 15.5 13.8
50.2 49.8 34.9 50.2 29.1 50.1 52.2 28.8 49.6 48.4 1.006 -13.90 0.988 13.6 12.0
1
2
40.0 39.9 23.2 37.3 11.7 38.2 39.2 11.9 38.2 35.0 0.821 4.87 0.974 2.4
2.3
- 1.091 -20.40 0.947 16.2 16.4
- 0.879 -11.23 0.926 16.9 17.1
3
55.2 48.4 39.1 52.7 33.9 52.6 54.1 32.3 -
51.2 53.4 42.7 53.3 30.8 57.9 56.3 28.0 -
4
5
53.5 55.0 42.0 54.7 36.6 54.4 56.3 34.1 54.0 50.9 1.133 -25.01 0.979 18.5 16.6
52.7 52.5 37.9 52.1 34.7 52.2 54.6 34.4 52.3 50.6 1.154 -23.98 0.986 16.7 14.9
1
2
42.3 42.2 26.0 39.5 18.4 40.5 42.0 18.6 40.5 38.1 0.962 -2.82 0.985 4.1
3.6
- 1.305 -32.69 0.950 18.1 18.4
- 1.056 -21.8 0.939 19.1 19.2
3
56.0 50.2 41.0 53.2 37.8 53.2 54.8 37.1 -
53.0 54.7 44.8 54.3 36.3 57.9 57.3 32.9 -
4
5
54.1 55.6 42.6 55.3 36.8 56.0 57.0 34.3 54.6 51.4 1.101 -24.10 0.977 19.1 17.2
53.2 53.0 38.5 52.6 34.8 52.8 55.2 34.5 52.7 51.0 1.138 -23.77 0.988 17.1 15.3
1
2
43.1 43.1 26.9 40.4 18.7 41.3 42.8 18.8 41.3 38.8 0.943 -2.81 0.984 4.8
4.2
- 1.282 -32.3 0.955 18.6 18.9
- 1.033 -21.2 0.939 19.6 19.7
3
56.6 50.9 41.6 53.8 38.1 53.8 55.5 37.3 -
53.7 55.3 45.5 54.9 36.4 58.5 58.0 33.0 -
4
5
In the case of PBE calculations, a great improvement when using of λ3 basis
set instead of 6-31G(d) is noteworthy. For λ3, the MAE and RSMD values are only

2.4 and 2.3 ppm, respectively; whereas for 6-31G(d), the MAE equals 5.1 ppm and
RMSD equals 4.6 ppm. However, the correlation coefficients R2 are very similar
for both the 6-31G(d) and the λ3 basis set. In the case of B3LYP functional, the
quality of observed correlations slightly decreases when the polarization function
is added to the hydrogen atoms. Applying the 6-31G(d) basis, the MAE and RMSD
values equal 4.1 and 3.6 ppm, respectively; whereas, when using the 6-31G(d,p)
basis set, the MAE and RMSD values increase to 4.8 and 4.2 ppm, respectively.
Knowing that the alkylation of model trizole-2-thione leads to the
formation of structures 3, we can improve the GIAO DFT-based method for the
prediction of NMR chemical shifts of the hydrocarbon group attached to the
sulfur atom in alkylsulfanyl-substituted 1,2,4-trizoles. The prediction capability
increases when considering a linear correlation between DFT-computed and
experimental chemical shifts:
δ(predicted) = δ(DFT)*S – I
where δ(predicted) is the predicted value of NMR chemical shift, δ(DFT) is the
chemical shift calculated with DFT methods, S and I are corresponding slopes and
intercepts, respectively, from Tables 2 or 3. These calculations reduce the MAE
and RMSD values, whereas the R2 values remain the same (see Table 4). The
B3LYP/6-31G(d) method is found to be the most preferable as it delivers good
1
results even with relatively small basis set. Thus, in both H and 13C NMR
calculations, this method results in MAE of 0.157 and 0.773 ppm, respectively,
and RMSD values of 0.192 and 1.123 ppm, respectively.
Table 4. Errors in accuracy of chosen GIAO DFT methods for prediction of NMR
chemical shifts of R-CH2-S group in 1,2,4-triazoles
1H NMR
RMSD
0.227
13C NMR
RMSD
1.330
Method
MAE
0.205
0.233
R2
MAE
0.997
1.021
R2
PBE/6-31G(d)
PBE/λ3
0.868
0.814
0.978
0.974
0.269
1.466

B3LYP/6-31G(d)
0.157
0.235
0.192
0.268
0.905
0.816
0.773
0.815
1.123
1.157
0.985
0.984
B3LYP/6-31G(d,p)
3.2. Reactivity of triazoles a-d
To explain the observed regioselective electrophilic attack of the exocyclic
sulfur atom in 1,2,4-triazol-3-thiones, several reactivity descriptors were
computed at the PBE0/6311++G(d,p) level of theory. Geometry optimization of
forms of a-d was performed with the PBE/6-31G(d) method and optimized
structures can be found in Supplementary materials (Tables S1-4). As the
alkylation was performed in alkaline and in neutral mediums, the different forms
of triazoles a-d must be considered. While the anionic states of a-d are the
predominant forms in an alkaline solution, a mixture of thione (major) and thiol
(minor) isomers are the major forms in a neutral solution. As the experiments
were performed in ethanol, the CPCM with appropriate parameters was included
in all computations. For clarity, the numbering in the triazole ring in this section is
consistent with Scheme 1.
At first, the highest occupied molecular orbital (HOMO) must be considered
as the isosurface determining the reactivity of the considered triazoles (see Figure
1). In all cases, HOMO is located over the whole 1,2,4-triazole-3-thione system for
all triazoles. The contribution of the pyrrole-type nitrogen in position 4 of the ring
in the case of an attached phenyl group to the HOMO isosurface is negligible,
which can be explained by the participation of the nitrogen atom’s lone electron
pair in the conjugation of the triazole aromatic system. However, an exocyclic
nitrogen atom – if present – contributes to the HOMO isosurface. In addition, we
have noticed a significant influence of the phenyl and phenylamino groups in the
HOMO formation of thiols b and d. While the contribution of the 5-phenyl group
of thione d into the HOMO is minimal, the 5-phenylamino group of thione b has a
high impact on the HOMO. In general, the analysis of the HOMO isosurfaces
reveals that the 5-R(amino)-1,2,4-triazole-3-thione system, except for the pyrrole-

type nitrogen atom, is the most reactive compound of the molecules a-d.
However, these calculations do not allow for grading the influence of the different
atoms on HOMO and for determining the most reactive atom.
Figure 1. HOMOs of anionic (A), thiol (B), and thione (C) forms of triazoles a-d.
Electron localization function (ELF) allows the mapping of the probability of
the location of electron pairs. In Figure 2, the ELFs of the anionic, thiol, and thione
forms of triazoles a-d are shown. The large green domains correspond to
hydrogen atoms, and the blue blobs represent lone pair domains. Relatively small
yellow and purple areas correspond to covalent bond domains. Yellow and purple
colors were used to grade the blobs area size. Larger purple areas are observed in
phenyl rings and double bonds of allyl fragments, while single bonds and bonds in
the triazole ring are characterized by much smaller yellow areas. A low ELF in the
triazole ring indicates a lower delocalization of electrons over the ring and,
consequently, its lower aromatic character in comparison to phenyl rings. The
lone pair electrons can be donated to the unoccupied orbitals of electrophiles
(alkylation agents). Considering the areas of lone pair domains, the reactivity of
the heteroatoms can be graded as follows: (a) the sulfur atom is the most reactive
centre; (b) the nitrogen atoms in positions 1 and 2 of the triazole ring are slightly
less reactive, while in the case of thione forms (Figure 2C), the second nitrogen
atom is not reactive at all; (c) the exocyclic nitrogen atom in triazoles a-c exhibits
the lowest reactivity. In the case of thiols, a noticeable decrease of the ELF near

the sulfur can be explained by the sharing of some electron density with the
attached hydrogen atom.
Figure 2. ELFs of anionic (A), thiol (B), and thione (C) forms of triazoles a-d.
The Fukui function for an electron elimination process (f–), the so-called
electrophilic Fukui function, determines the preferable site of the attack of an
electrophile. This function was used to explain the regioselectivity of the
alkylation of triazoles a-d. The high prediction capability of Fukui functions in
electrophilic reactions, like protonation [78] and alkylation [79, 80] has been
reported. In Figure 3, the electrophilic Fukui functions of the anionic, thiol, and
thione forms of a-d are presented. Light blue areas represent positive and purple
areas negative values. The highest values of the Fukui function determine the
most favorable site for electrophilic attack. In the case of anionic and thione
species, the largest isosurface areas are located on the sulfur atom, which concurs
with the experimentally observed S-alkylation. The reactivity of the sulfur
significantly decreases in thiols c and d, and becomes negligible in thiols a and b.
Oppositely, the reactivity of the exocyclic nitrogen increases. However, these
results do not contradict the experiment, as the thiol forms are less stable then
thione, and consequently, thiols are present in reaction mixture only at a trace
level. To support this statement, the total energies, total enthalpies, relative
Gibbs free energies, and isomers distribution values of thione and thiol forms of
a-d are presented in Table 5.

Figure 3.Electrophilic Fukui functions of anionic (A), thiol (B), and thione (C) forms
of triazoles a-d.
All data presented in Table 5 were calculated at the CPCM-PBE0/6-
311++G(d,p)//PBE/6-31G(d) level of theory for the gas phase, ethanol and DMSO
CPCM solutions. Comparing the relative Gibbs free energies at 25 °C, it can be
observed that in all cases, the thione form is about 10.9-14.8 kcal/mol more
stable than the corresponding thiol. Therefore, the major fraction of triazoles a-d
exists in the thione form. Furthermore, increasing the electron donating ability of
the substituent in position 5 of the triazole ring leads to the increase of the
stability of thione form. The calculated distributions of respective thione/thiol
forms indicate the presence of only traces of thiols in solution, which do not
influence the regioselectivity of the alkylation reaction.
Table 5.Total energies (E, Ha), total enthalpies (H, Ha), relative Gibbs free
energies at 25 °C (G, kcal/mol), and isomers distribution values (D, %) of thione
and thiol forms of a-d in gas phase (Gas), ethanol (Ethanol) and DMSO (DMSO)
solutions calculated at the PBE0/6-311++G(d,p)//PBE/6-31G(d) (PBE0) level of
theory
Level of
theory
a
b
c
d
thione
thiol
thione
thiol
thione
thiol
thione
thiol
E -928.632 -928.606 -1157.111 -1157.088 -926.288 -926.264 -1101.799 -1101.777
H -928.420 -928.398 -1156.860 -1156.840 -926.120 -926.100 -1101.565 -1101.546
Ethanol

G
D
0.00
100.00 1.81×10-8
13.29
0.00
11.81
0.00
12.51
0.00
10.88
100.00 2.22×10-7 100.00 6.74×10-8
100.00 1.06×10-6
E -928.609 -928.581 -1157.088 -1157.064 -926.264 -926.240 -1101.778 -1101.756
H -928.397 -928.373 -1156.836 -1156.816 -926.096 -926.075 -1101.544 -1101.526
Gas
G
D
0.00
100.00 1.55×10-9
14.75
0.00
12.21
0.00
12.58
0.00
10.86
100.00 1.11×10-7 100.00 5.95×10-8
100.00 1.09×10-6
E -928.632 -928.607 -1157.112 -1157.089 -926.289 -926.265 -1101.799 -1101.777
H -928.420 -928.398 -1156.861 -1156.841 -926.121 -926.100 -1101.565 -1101.547
DMSO
G
D
0.00
100.00 2.02×10-8
13.23
0.00
11.78
0.00
12.50
0.00
10.87
100.00 2.30×10-7 100.00 6.81×10-8
100.00 1.07×10-6
The reactive sites of electrophile-nucleophile interactions can be described
through the analysis of atomic partial charges. The different types of atomic
partial charges of exocyclic nitrogen (N5), nitrogen atoms in position 1 (N1) and 2
(N2) of the triazole ring, and exocyclic sulfur (S3) are presented in Table 6. To
support the observed regioselectivity of alkylation at the sulfur atom, the
calculated partial charges should be most negative at the S3 sulfur atom. All types
of partial charges were calculated at the CPCM(Ethanol)-PBE0/6-311++G(d,p)
level of theory, except the Mulliken and Löwdin charges; since these two
population schemes are not consistent with basis sets containing diffuse
functions, the CPCM(Ethanol)-PBE0/6-311G(d,p) method was used for the
calculation of these types of partial charges.
According to Table 6, none of the studied types of partial charges predicts
the sulfur as the preferred reaction center in thiol forms. However, it cannot be
considered as a limitation of the approach since the thiols rearrange to their
thione form and experimental evidences of regioselectivity of the thiol cannot be
obtained. In the case of the anionic and thione forms, the Hirshfeld population
scheme and its successor ADCH produce partial charges that are in agreement
with the observed reaction path. This result is expected, as a good prediction
capability of Hirshfeld partial charges was reported previously [81-84]. The CM5

partial charges are very similar. In the case of thiones a and c, the N5 atom
indicated as more negative and therefore as the preferred reaction center, which
is not in accordance with the experimental results.
The CHELPG and NPA partial charges were found inadequate for this type of
analysis. This finding is surprising, considering that the CHELPG scheme was
created to reproduce electrostatic potential in the most optimal way and that
these charges were expected to be suitable for the description of strong polar
interactions. The failure of the NPA scheme is also not explainable, especially
considering its good performance in the prediction of physico-chemical
properties, like LogP [85]. Mulliken population analysis results in correct
predictions in the case of anionic forms, but fails for the model thione species.
The Löwdin partial charges are found very suitable for the prediction of S-
alkylation of anionic and thione forms, and thus, could be used for the modelling
of analogues systems.
Table 6.DFT-computed partial atomic charges of N1, N2, S3, and N5 atoms in
anionic, thione, and thiol forms of triazoles a-d. The match of calculated partial
charges with observed regioselectivity highlighted boldly.
anion
thione
thiol
a
b
c
d
a
b
c
d
a
b
c
d
N1 -0.513 -0.432 -0.443 -0.359 -0.397 -0.285 -0.211 -0.413 -0.445 -0.550 -0.539 -1.034
N2 -0.409 -0.670 -0.759 -0.564 -0.079 -0.431 -0.616 -0.052 -0.176 -0.151 -0.189
0.186
-0.124 -0.057 -0.038 -0.014
S3
-0.693 -0.885 -0.924 -0.691 -0.475 -0.513 -0.521 -0.471
N5 -0.432 -0.396 -0.657 -0.376 -0.379 -0.635
-
-
-0.364 -0.450 -0.633
-
N1 -0.378 -0.419 -0.392 -0.334 -0.317 -0.296 -0.324 -0.275 -0.344 -0.324 -0.359 -0.310
N2 -0.410 -0.476 -0.401 -0.380 -0.296 -0.292 -0.294 -0.279 -0.334 -0.341 -0.344 -0.330
S3
-0.437
-0.429
-0.115 -0.099 -0.101 -0.094
-0.419
-0.626 -0.733 -0.614 -0.594
-0.427
N5 -0.485 -0.518 -0.613
-
-0.465 -0.421 -0.592
-
-0.471 -0.425 -0.601
-
N1 -0.283 -0.251 -0.298 -0.239 -0.210 -0.186 -0.220 -0.166 -0.249 -0.351 -0.265 -0.213
N2 -0.322 -0.308 -0.312 -0.295 -0.029 -0.024 -0.026 -0.013 -0.240 -0.523 -0.247 -0.234
S3
-0.046 -0.226 -0.032 -0.025
-0.617 -0.603 -0.606 -0.586 -0.425 -0.416 -0.418 -0.409
N5 -0.157 -0.101 -0.194 -0.129 -0.090 -0.164
N1 -0.543 -0.306 -0.523 -0.421 -0.631 -0.495 -0.589 -0.529 -0.496 -0.225 -0.524 -0.451
-
-
-0.137 -0.917 -0.176
-

N2 -0.617 -0.690 -0.558 -0.597 0.106 -0.016 0.124 0.191 -0.523 -0.243 -0.405 -0.430
S3 -0.827 -0.798 -0.803 -0.792 -0.498 -0.482 -0.475 -0.469 -0.263 -0.030 -0.211 -0.228
N5 -0.880 -0.894 -0.890
-
-0.736 -0.813 -0.865
-
-0.823 -0.094 -0.879
-
N1 -0.449 -0.442 -0.454 -0.359 -0.386 -0.382 -0.385 -0.298 -0.428 -0.425 -0.427 -0.346
N2 -0.482 -0.464 -0.471 -0.450 -0.398 -0.392 -0.395 -0.383 -0.369 -0.397 -0.404 -0.392
S3 -0.542 -0.512 -0.520 -0.488 -0.338 -0.312 -0.317 -0.301 0.012 0.040 0.036
0.048
-
N5 -0.677 -0.617 -0.849
-
-0.654 -0.607 -0.824
-
-0.665 -0.611 -0.836
N1 -0.386 -0.396 -0.383 -0.328 -0.315 -0.329 -0.308 -0.257 -0.371 -0.388 -0.364 -0.323
N2 -0.379 -0.358 -0.368 -0.345 -0.287 -0.276 -0.282 -0.265 -0.291 -0.305 -0.318 -0.300
S3
-0.376 -0.347 -0.358 -0.339 -0.004 0.022 0.014
-0.619 -0.585 -0.600 -0.567
0.028
-
N5 -0.509 -0.561 -0.538
-
-0.506 -0.565 -0.523
-
-0.513 -0.568 -0.532
N1 -0.175 -0.157 -0.179 -0.115 -0.110 -0.098 -0.111 -0.045 -0.153 -0.139 -0.155 -0.095
N2 -0.205 -0.191 -0.196 -0.187 0.152 0.157 0.154 0.163 -0.111 -0.129 -0.135 -0.128
S3
0.243 0.263 0.257
0.037 0.111 0.036
-0.480 -0.456 -0.465 -0.433 -0.257 -0.237 -0.243 -0.227
0.270
-
N5 0.015 0.104 0.011
-
0.050 0.117 0.054
-
Conclusions
Ten S-alkylated triazoles were synthesized through the alkylation of four
1,2,4-triazole-3-thiones with eight different alkyl halides; seven of the obtained 3-
alkylsulfanyl-1,2,4-triazoles are new and have not been previously described in
the literature.
1
The sole comparison of H NMR spectra with GIAO DFT-computed chemical
shifts is not a reliable approach to reveal the structure of alkylation products. The
presence of several reactive centers in the different possible structural forms of
the considered triazoles determines the subsequent formation of the different
alkylation products, which are characterized by relatively close isotropic chemical
shielding values of characteristic groups. The analysis of experimentally observed
and GIAO calculated 13C NMR chemical shifts of the α-methylene(methyl) group of
the introduced hydrocarbon fragment allows for the reliable assignment of the
experimental peaks to S-alkyl groups. A strong correlation (MAE = 0.773 ppm,
RMSD = 1.123 ppm, R2 = 0.985) was obtained between experimental and
calculated values of 13C NMR chemical shifts of the characteristic carbon.

The anionic, thione and thiol forms of the triazoles a-d were studied in
terms of DFT reactivity descriptors to provide insight into the observed alkylation
regioselectivity. The anionic form of the used 1,2,4-triazoles is predominant in
alkaline medium, whereas the thione and thiol forms were considered as the
major species in a neutral solution. However, due to the very low quantity of
thiols, the impact of these forms can be considered as negligible. Thus, PBE0/6-
311++G(d,p) computations reveal that thiones are about 10.9-13.3 kcal/mol more
stable than the corresponding thiols, and that the thione stability increases with
the increasing the electron donating ability of the substituent in position 5 of the
triazole ring.
ELF and Fukui function isosurfaces can correctly predict the observed
preferential reactivity of the exocyclic sulfur atom in the anionic and thione
forms. Oppositely, HOMOs are found not to be suitable for this prediction. Among
several tested types of atomic partial charges, only Hirshfeld, ADCH, and Löwdin
population schemes can correctly predict the preferential attack on the S-atom.
Acknowledgments
This work was partially financially supported by the Ministry of Education and
Science of Ukraine (Projects GR-0119U100232 and GR-0120U100431).
AUTHOR STATEMENT
Maksym Fizer: Conceptualization, Supervision, Investigation, Writing - Original
Draft, Writing - Review & Editing. Mikhailo Slivka: Conceptualization,
Investigation, Writing - Original Draft, Writing - Review & Editing. Nataliya
Korol: Investigation, Writing - Original Draft. Oksana Fizer: Investigation,
Writing - Original Draft, Writing - Review & Editing.

Declaration of interests
☒ The authors declare that they have no known competing financial interests or personal relationships
that could have appeared to influence the work reported in this paper.
☐The authors declare the following financial interests/personal relationships which may be considered
as potential competing interests:
References
[1] V. Shah, Ch. Yang, Z. Shen, B.M. Kerr, K. Tieu, D.M. Wilson, J. Hall, M. Gillen,
C.A. Lee, Metabolism and disposition of lesinurad, a uric acid reabsorption
inhibitor, in humans, Xenobiotica 49 (2019) 811-822,
https://doi.org/10.1080/00498254.2018.1504257.
[2] M.M. Regeda-Furdychko, The influence of thiotriazoline on the indicators of
nitric oxide under the condition of experimental contact dermatitis and
experimental pneumonia, J. Edu. Health Sport 10 (2020) 35-40,
http://doi.org/10.12775/JEHS.2020.10.02.004.
[3] M.C.F.C.B. Damião, R. Galaverna, A.P. Kozikowski, J. Eubanks, J.C. Pastre,
Telescoped continuous flow generation of a library of highly substituted 3-thio-
1,2,4-triazoles, React. Chem. Eng. 2 (2017) 896-907,
https://doi.org/10.1039/C7RE00125H.
[4] A.T.A. Boraei, A new direct synthetic access to 4-amino-2-N-(glycosyl/propyl)-
1,2,4-triazole-3-thiones via hydrazinolysis of 3-N-((acylatedglycosyl)/allyl)-1,3,4-
oxadiazole-2-thiones,Arkivoc, 3 (2016) 71-81,
https://doi.org/10.3998/ark.5550190.p009.399.
[5] A.T.A. Boraei, E.S.H. El Ashry, A. Duerkop, Regioselectivity of the alkylation of
S-substituted 1,2,4-triazoles with dihaloalkanes, Chem. Cent. J. 10 (2016) 22-
35,https://doi.org/10.1186/s13065-016-0165-0.
[6] A.T.A. Boraei, M.S. Gomaa, E.S.H.E. Ashry, A. Duerkop, Design, selective
alkylation and X-ray crystal structure determination of dihydro-indolyl-1,2,4-
triazole-3-thione and Its 3-benzylsulfanyl analogue as potent anticancer agents,