Push-pull unsymmetrical substitution in nickel(II) complexes with tetradentate N2O2 Schiff base ligands: Synthesis, structures and linear-nonlinear optical studies

Article abstract of DOI:10.1039/c9dt01216h

New push-pull (A-D) nickel(ii) compounds of general formula [Ni(5-A-5′-D-saltn)] (3a-3l) with unsymmetrically-substituted N₂O₂ tetradentate Schiff base ligands are reported here. The ligands 5-A-5′-D-saltn^2- (H₂saltn = N,N′-bis(salicylidene)diaminopropane) possess differently-substituted salicylaldehyde (^A/Dsal) moieties condensed to 1,3-diaminopropane (tn), and carry either an electron acceptor (A = H, Br, or NO₂) or donor (D = H, Me, or OMe) group in the para position with respect to the coordinated phenoxido oxygen atoms. These compounds could be obtained by template synthesis involving derivatives [Ni(^Gsal)₂(H₂O)₂], 1a-e (G = NO₂, Br, H, Me and OMe, respectively) and [Ni(^GL)₂], 2a-d (^GL^- = (E)-2-((3-aminopropylimino)methyl)-4-G-phenolate, G = NO₂, Br, H and Me, respectively). Scrambling of the ligands and condensation to compounds 3 was suitably achieved by refluxing compounds 1 and 2 that carry the G groups suitable for the desired final A-D combination. Dinuclear intermediates [Ni₂(μ-^GL)₂(^G′sal)₂] (4a,b,e,f,g) were also detected and isolated. The single-crystal X-ray diffraction structures of [Ni(5′-OMe-saltn)]·CHCl₃ (3c·CHCl₃), [Ni(5-Br-5′-OMe-saltn)]·EtOH (3g·EtOH) and [Ni(5-NO₂-saltn)] (3j) show different degrees of distortion around the central core, leading to saddle-like (3c), planar (3g) and step-like (3j) molecular conformations. DFT geometrical optimization of the compounds showed that, for isolated molecules, the saddle-like conformation is slightly more stable with respect to the other conformations. UV-visible absorption spectra showed structured absorption profiles at about 320-440 nm, whose intensity was amplified by the presence of the nitro group, and this was assigned to a convolution of one metal-to-ligand charge transfer and two intra-ligand charge transfer transitions by TDDFT computations. Surprisingly, UV-visible spectra of the derivatives with Br were comparable to the derivatives with Me, suggesting that in this case the behaviour of the halogen was as a weak electron donor group. The experimental investigation (through electric-field-induced second-harmonic and solvatochromic measurements) of the second-order NLO responses of compounds 3, in conjunction with the theoretical calculations, revealed that the observed NLO efficiency is determined by the combined effect of two almost orthogonal charge transfer directions within the molecules, one along the axis approximately bisecting the donor and the accepting moieties and the other along the A-D axis.

Full text of DOI:10.1039/c9dt01216h

View Article Online
View Journal
Dalton
Transactions
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: L. Rigamonti, A.
Forni, S. Righetto and A. Pasini, Dalton Trans., 2019, DOI: 10.1039/C9DT01216H.
This is an Accepted Manuscript, which has been through the
Volume 45 Number
1 7 January 2016 Pages 1–398
Royal Society of Chemistry peer review process and has been
accepted for publication.
Dalton
Transactions
Accepted Manuscripts are published online shortly after
An international journal of inorganic chemistry
www.rsc.org/dalton
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
ISSN 1477-9226
PAPER
Joseph T. Hupp, Omar K. Farha et al.
Efficient extraction of sulfate from water using
framework
a Zr-metal–organic
rsc.li/dalton

Page 1 of 35
Dalton Transactions
Push-pull unsymmetrical substitution in nickel(II) complexes with tetradentate
View Article Online
DOI: 10.1039/C9DT01216H
N2O2 Schiff base ligands: synthesis, structures and linear-nonlinear optical studies†
Luca Rigamonti,*a,c Alessandra Forni,*b,c Stefania Righettoc and Alessandro Pasinic
a
Dipartimento di Scienze Chimiche e Geologiche, Università degli Studi di Modena e Reggio
Emilia, via G. Campi 103, 41125 Modena, Italy
Email: luca.rigamonti@unimore.it, luca.rigamonti@yahoo.com
b Istituto di Scienze e Tecnologie Molecolari, Consiglio Nazionale delle Ricerche (ISTM-CNR), via
Golgi 19, 20133 Milano, Italy
Email: a.forni@istm.cnr.it
c Dipartimento di Chimica, Università degli Studi di Milano, via C. Golgi 19, 20133 Milano, Italy
† Electronic supplementary information (ESI) available. CCDC 1903483–1903485 for 3cCHCl3,
3gEtOH and 3j, respectively; Experimental Section (cont.); Crystallographic and data collection
parameters for 3c·CHCl3, 3g·EtOH and 3j (Table S1); Computed excitation energies, oscillator
strengths and analysis of the most important contributions to the transitions for compounds 3 and
selected analogue copper(II) complexes (Table S2); Isodensity surface plot of the PBE0/6-
311++G(d,p) frontier orbitals of 3c and 3j mainly involved in the computed transitions (Fig. S1);
UV-visible absorption spectra of 3c: dilution studies from 10–3 down to 10–5 mol L–1 CHCl3
solutions, solvatochromism at 5  10–5 mol L–1 solutions in solvents from low polar toluene to high
polar methanol, and addition of increasing amount of DMSO to the 10–4 mol L–1 CHCl3 solution
(Fig. S2); UV-visible absorption spectra of 3j: solvatochromism at 5  10–5 mol L–1 solutions in
solvents from low polar toluene to high polar methanol, and addition of increasing amount of
DMSO to the 10–4 mol L–1 CHCl3 solution (Fig. S3). For ESI and crystallographic data in CIF or
other electronic format see DOI: 10.1039/***.
Abstract
New push-pull (A–D) nickel(II) compounds of general formula [Ni(5-A-5’-D-saltn)] (3a–3l) with
unsymmetrically-substituted N2O2 tetradentate Schiff base ligands are here reported. The ligands 5-
A-5’-D-saltn2– (H2saltn = N,N’-bis(salicylidene)diaminopropane) possess differently-substituted
salicylaldehyde (A/Dsal) moieties condensed to 1,3-diaminopropane (tn), carrying either an electron
acceptor (A = H, Br, NO2) or donor (D = H, Me, OMe) group in para position with respect to the
coordinated phenoxido oxygen atoms. These compounds could be obtained by template synthesis
involving derivatives [Ni(Gsal)2(H2O)2], 1a–e (G = NO2, Br, H, Me and OMe, respectively) and
[Ni(GL)2], 2a–d (GL– = (E)-2-((3-aminopropylimino)methyl)-4-G-phenolate, G = NO2, Br, H
1

Dalton Transactions
Page 2 of 35
View Article Online
and Me, respectively). Refluxing compounds 1 and 2 carrying G groups suitable for the
DOI: 10.1039/C9DT01216H
desired final A–D combination, scrambling of the ligands occurred and condensation to
compounds 3 was suitably achieved. Dinuclear intermediates [Ni2(-GL)2(G’sal)2] (4a,b,e,f,g)
were also detected and isolated. The single-crystal X-ray diffraction structures of [Ni(5’-OMe-
saltn)]CHCl3 (3cCHCl3), [Ni(5-Br-5’-OMe-saltn)]EtOH (3gEtOH) and [Ni(5-NO2-saltn)] (3j)
show different degrees of distortion around the central core, leading to saddle-like (3c), planar (3g)
and step-like (3j) molecular conformations. DFT geometry optimization of compounds 3 shows
that, for isolated molecules, the saddle-like conformation is slightly more stable with respect to the
other ones. UV-visible absorption spectra show structured absorption profiles at about 320–440 nm,
whose intensity is amplified by the presence of the nitro group, assigned to a convolution of one
metal-to-ligand charge transfer and two intra-ligand charge transfer transitions by TDDFT
computations. Surprisingly, UV-visible spectra of derivatives with Br are comparable with the ones
with Me, suggesting in this case a behaviour of the halogen as a weak electron donor group. The
experimental investigation (through Electric-Field-Induced Second-Harmonic and solvatochromic
measurements) of the second-order NLO responses of compounds 3, in conjunction with theoretical
calculations, reveals that the observed NLO efficiency is determined by the combined effect of two
almost orthogonal charge transfer directions within the molecules, one along the axis approximately
bisecting the donor and the accepting moieties and the other along the A–D axis.
Introduction
Nonlinear optics (NLO)1–8 started to be explored by Franken et al in 1961 with the discovery of
second-harmonic generation (SHG), or frequency doubling,9 right after the construction of the first
laser by Maiman in 1960.10,11 In fact, only when the time-varying electric component E of the
electromagnetic light wave, and at intensities as observed in a laser light beam, is applied, a
molecule (material) can respond nonlinearly to E, and the microscopic (macroscopic) polarization p
(P) can be expressed by the power series in the field strength: p =  E +  EE +  EEE + ... (P =
(1) E + (2) EE + (3) EEE + ...), where  is the linear polarizability and  and  are the quadratic
and cubic hyperpolarizabilities, respectively ((i) are the electric susceptibility coefficients of i
order),5 and the  tensor (or (2)) is the responsible for the second-order NLO response.12 Since the
discovery, NLO has then started finding different applications with growing impact on today’s
daily life.13
The interest in this field, and in particular in the second-order behaviour,14 has led to the
synthesis of several compounds during years bearing a dipolar acceptor--bridge-donor (A--D)
2

Page 3 of 35
Dalton Transactions
molecular architecture,15–18 which gives rise to a non-zero  value due to polarizable electrons in a
View Article Online
DOI: 10.1039/C9DT01216H
non-centrosymmetric environment. Organic derivatives have shown to possess large and very fast
quadratic NLO responses16,18 that can be enhanced by increasing donor or acceptor strength,19,20 or
by molecular skeleton modifications.17,21–23 Organometallic and coordination compounds have been
also actively investigated for NLO applications since they offer a larger variety of molecular
structures compared to pure organic systems.24–34 In particular, the electronic properties are
modulated by the d- and f-electrons26 of the metal ions,35 since the metal-containing moiety of the
molecule can act as either donor30,36,37 or acceptor group,38,39 and the existence of different
oxidation states for the same metal can act as a switch of the NLO properties by redox reactions.40–
44 Supramolecular assemblies of coordination units into metal-organic frameworks have also shown
to be competitive materials for NLO applications.8,45,46
Metal complexes with tetradentate N2O2 salen-type Schiff bases are particularly interesting
for their NLO features25,27,29,47,48 because of the high thermal and chemical stability49 and the non-
innocent role of the metal centre as bridge between the acceptor A and the donor D groups,
positioned on the organic skeleton of the compounds, and this is due to the presence of charge
transfer (CT) transitions at low energies that involve the metal ion.38,50–62 Moreover, they show the
possibility to be incorporated into polymers and create NLO-active thin films.63,64 The second-order
NLO properties of these compounds have been studied in complexes with D groups (OMe, NEt2) in
position 4 of the salicylaldehyde (sal) moieties and A groups (NO2, Cl) on the o-phenylendiamine
(o-phen), where this substitution increases the charge asymmetry of the final [M(salophen)] (M =
Co, Ni, Cu, Zn) structure and enhances their NLO responses.54,55,57 As a matter of fact, nickel(II) is
among the most studied transition metals in this class of compounds49,50,53–55,57–59,61,62,65 thanks to its
role in the NLO efficiency and its inclination for the square planar coordination environment.66
With the scope of increasing the knowledge on salen-type complexes, few years ago we
started studying salen analogues where the two sal moieties bear different substituents in position 5,
i.e. para to the coordinated phenoxido oxygen atom. Our first results regarded the copper(II)
complexes of general formula [Cu(5-A-5’-D-salen)] and [Cu(5-A-5’-D-saltn)],38 in which the
acceptor group was mainly NO2 while the D groups were H, Me, OMe and NMe2 (tn = 1,3-
diaminopropane). In these compounds, the induced charge asymmetry generated by the A–D
unsymmetrical substitution leads to the concomitant modulation of the g (ground state electric
dipole moment) and tot (derived from the full  hyperpolarizability tensor) vector directions. More
recently, unsymmetrically substituted zinc(II) and nickel(II) salophen complexes were also reported
and studied for their third-order (two-photon absorption) NLO properties.67
3

Dalton Transactions
Page 4 of 35
View Article Online
The present work aims to a further step in the electronic features and linear and
DOI: 10.1039/C9DT01216H
second-order NLO behaviours of salen-type derivatives with the flexible aliphatic diamine
tn, studying the effect of the metal change from copper(II) to nickel(II) in the molecular
core of saltn ligands. In fact, the modulating role of the metal centre in the NLO response
of these class of complexes with tetradentate Schiff base ligands51,53 is worthy to be
investigated, in addition to the possibility for the coordination geometry to move away from
a strictly square planar geometry (as for salophen ligands) thanks to the propylene bridge,
which allows the access to more low-lying excited states other than the only metal-centred
charge transfer transition with concomitant enhancement of the NLO responses.53,68 The
template synthesis69 of the compounds [Ni(5-A-5’-D-saltn)] (3a–l, see Table 1 and Scheme
1 for A–D substitutions) is here exploited by reaction of derivatives [Ni(Gsal)2(H2O)2], 1a–e
(G = NO2, Br, H, Me and OMe, respectively) with [Ni(GL)2], 2a–d (GL– = (E)-2-((3-
aminopropylimino) methyl)-4-G-phenolate, G = NO2, Br, H and Me, respectively). The
isolation of some reaction intermediates 4 with probable dinuclear structure70 and peculiar
reactivity depending on the substituents will be also presented (see Scheme 1). The
experimental and theoretical characterizations of the linear and nonlinear optical features
and the comparison with the analogue copper(II) compounds38 will be then described, in
order to clarify the role of the metal, and with the extension in the use of bromine as
further electron-active substituent.
4

Page 5 of 35
Dalton Transactions
O
View Article Online
DOI: 10.1039/C9DT01216H
.
+ Ni(AcO) 4H O
2
G
OH
2
2
(G = NO2, Br, H, Me, OMe)
NH NH
2
2
2
EtOH, reflux, 1 h
EtOH, reflux, 4 h
OH2
NH2
O
O
O
G
O
G
N
2 NH2 NH2
Ni
Ni
N
G
O
G
O
OH2
EtOH, reflux, 4 h
NH2
(G = NO2, Br, H, Me)
1a e
2a d
-
-
(G = NO2, Br, H, Me, OMe)
G'
EtOH, reflux
O
O
N
N
O
N
O
N
O
N
Ni
Ni
O
H2
Ni
G
G
H2
N
or
A
D
O
O
3a l
-
EtOH, reflux
4
intermediates
G'
Scheme 1 Synthetic scheme with reaction conditions of all synthetic steps from precursors 1 and 2
to intermediates 4 and final products 3.
Table 1 Combinations in the reaction between derivatives 1 and 2 and final A–D complexes 3,
together with the isolated intermediates 4 (see Scheme 1).
reaction
–
compound
3a
A
D
yield %
73
–
intermediate
G
H
G’
H
yield %
35
H
H
4a
4b
1d + 2c
1c + 2d
1e + 2c
–
H
Me
H
42
3b
H
Me
39
68
70
–
4b’
Me
47
3c
H
OMe
Br
not detected
3d
Br
–
Br
1c + 2b
1b + 2c
1d + 2b
1b + 2d
1e + 2b
–
4e
4e’
4f
H
80
49
75
3e
3f
Br
Br
H
54
67
82
37
80
82
64
H
Br
Br
not detected
Br OMe
Me
Me
3g
3h
Br
OMe
NO2
4g
67
NO2
–
1a + 2b
1b + 2a
not detected
not detected
3i
NO2
Br
5

Dalton Transactions
Page 6 of 35
View Article Online
DOI: 10.1039/C9DT01216H
1c + 2a
1a + 2c
1d + 2a
1a + 2d
1e + 2a
66
55
50
85
–
not detected
not detected
not detected
not detected
not detected
3j
NO2
NO2
H
3k
3l
Me
NO2 OMe
Experimental section
General information
All used chemicals were reagent grade. Solvents were used as received. Elemental analyses were
performed at the Microanalytical Laboratory at the Università degli Studi di Milano. ESI-MS
spectra were recorded in MeOH solutions with a LCQ Advantage Thermofluxional instrument.
Infrared spectra were recorded as KBr disks using a JASCO FT-IR 410 spectrophotometer with a 2
cm–1 resolution. UV-visible spectra for all soluble compounds 1, 2 and 3 of 10–4 mol L–1 CHCl3
solution at 298 K were recorded with a Hewlett Packard 8453 spectrophotometer; dilution studies
for 3c and 3j from 10–3 down to 5  10–6 mol L–1 CHCl3 solutions at 298 K were performed with a
Jasco V-570 UV/Vis/NIR spectrophotometer, which was also employed for solvatochromism
studies of 3c and 3j at 5  10–5 mol L–1 solutions in the following solvents (Reichardt polarity
parameters ET(30)71,72 in kcal mol–1 and relative polarity to water73 are reported in brackets):
toluene (33.9, 0.099), THF (37.4, 0.207), AcOEt (38.1, 0.228), CHCl3 (39.1, 0.259), CH2Cl2 (41.1,
0.309), acetone (42.2, 0.355), DMF (43.8, 0.386), DMSO (45.0, 0.444), CH3CN (46.0, 0.460),
EtOH (51.9, 0.654) and MeOH (55.5, 0.762); λ values are accurate to ± 1 nm and spectra are
mainly reported as wavenumbers, cm–1 (ε, L mol–1 cm–1). 1H NMR spectra were recorded on a 400
MHz Bruker FT-NMR Advance400 spectrometer at 298 K in CDCl3 as solvent; chemical shifts are
given in parts per million (ppm) versus external TMS and were determined by reference to the
solvent residual signals (7.26 ppm for CHCl3).
Synthetic procedures for compounds 1
[Ni(NO2sal)2(H2O)2] (1a).74–76 5-NO2-salH (517.0 mg, 3.10 mmol) was dissolved in ethanolic KOH
(6.2 mL of a 0.5 mol L–1 solution, 3.10 mmol) and solid Ni(AcO)2·4H2O (384.5 mg, 1.55 mmol)
was added under stirring. The mixture was refluxed for 1 h and then cooled down, yielding the
product as green solid, filtered, washed with EtOH and diisopropyl ether (iPr2O) and dried under
vacuum (597.9 mg, 98%). Anal. Calcd (%) for C14H12N2NiO10 (426.95): C, 39.38; H, 2.83; N, 6.56.
Found: C, 39.08; H, 3.00; N, 6.34. IR (KBr): 3437 (O–H), 1633 (C=O), 1306 (NO2) cm–1. UV-vis:
not soluble in CHCl3. Syntheses of 1b–1e and their characterization can be found in the Electronic
Supplementary Information (ESI).
6

Page 7 of 35
Dalton Transactions
View Article Online
DOI: 10.1039/C9DT01216H
Synthetic procedures for compounds 2
[Ni(NO2L)2] (2a). 5-NO2-salH (840.0 mg, 5.03 mmol) was dissolved in ethanolic KOH (50.0 mL of
a 0.1 mol L–1 solution, 5.00 mmol), and Ni(AcO)2·4H2O (624.5 mg, 2.51 mmol) was added under
stirring. The light yellow mixture was heated at 70 °C for 1 h, then tn (630 L, 7.55 mmol) was
added and the mixture was refluxed for 3 h. After cooling, 2a was recovered by filtration as yellow
solid, washed with EtOH, acetone and dried under vacuum (1214.1 mg, 96%). Anal. Calcd for
C20H24N6NiO6 (503.14): C, 47.74; H, 4.81; N, 16.70. Found: C, 47.81; H, 4.90; N, 16.61. MS
(ESI): m/z 525 ([M + Na]+, 80%), 1027 ([2M + Na]+, 100). IR (KBr): 3313, 3257 (NH2), 1640,
1595 (C=N), 1312, 1289 (NO2) cm–1. UV-vis (CHCl3): 25710 (12550). Syntheses of 2b–2d and
their characterization can be found in the ESI.
Synthetic procedures for compounds 3 and 4
[Ni(saltn)] (3a). First method: This compound was synthesized with a modification of literature
procedures.77,78 salH (116.8 mg, 0.96 mmol) and tn (40 L, 0.48 mmol) were dissolved in EtOH
(15 mL) and the yellow mixture was refluxed for 30 minutes. NiCl26H2O (116.8 mg, 0.49 mmol)
and Et3N (1 mL) were then added and the brownish mixture was refluxed for 4 h. After one night at
i
room temperature, the compound precipitated as green crystals, filtered, washed with EtOH, Pr2O
and dried under vacuum (127.5 mg, 73%). Anal. Calcd (%) for C17H16N2NiO2·1.5H2O (366.04): C,
55.78; H, 5.23; N, 7.65. Found: C, 55.37; H, 5.22; N, 7.71. MS (ESI): m/z 339 ([M + 1]+, 100%). IR
(KBr): 3454 (O–H), 1621 (C=N) cm–1. Second method:79 isolation of the intermediate [Ni2(-
HL)2(Hsal)2] (4a). tn (61 L, 0.73 mmol) was added to a suspension of 1c (245.2 mg, 0.73 mmol) in
EtOH (15 mL) and the mixture was heated to reflux, which causes the complete dissolution and the
formation of a green solution. After few minutes a green crystalline solid precipitated and the
mixture was further refluxed for 2 h. After cooling, the intermediate 4a was filtered, washed with
i
EtOH, Pr2O and dried under vacuum (89.8 mg, 35%). Anal. Calcd (%) for C34H36N4Ni2O6
(714.06): C, 57.19; H, 5.08; N, 7.85. Found: C, 56.75; H, 5.31; N, 7.65. IR (KBr): 3344, 3290
(NH2), 1633 (C=N) cm–1. The green compound 4a (85.0 mg, 0.13 mmol) was suspended again in
EtOH (50 mL) and refluxed for other 2 h, till the solid completely dissolved. The brown solution
obtained was left standing at room temperature for two days till brown needles of 3a precipitated
that were filtered, washed with EtOH and dried under vacuum (53.2 mg, 66%). Anal. Calcd (%) for
C17H16N2NiO2 (339.03): C, 60.23; H, 4.76; N, 8.26. Found: C, 60.71; H, 5.04; N, 7.72. IR (KBr):
1621 (C=N) cm–1.
7

Dalton Transactions
Page 8 of 35
[Ni(5’-Me-saltn)] (3b) and isolation of the intermediates [Ni2(-HL)2(Mesal)2] (4b) and V[ieNw iArt(icle-Online
2
DOI: 10.1039/C9DT01216H
MeL)2(Hsal)2] (4b’). First method 1d + 2c: 1d (123.9 mg, 0.30 mmol) and 2c (109.5 mg, 0.30
mmol) were suspended in EtOH (30 mL) and left under reflux for 5 h. The intermediate 4b was
i
filtered as light green solid, washed with EtOH, Pr2O and dried in vacuo (107.9 mg, 42%). Anal.
Calcd (%) for C36H40N4Ni2O6·EtOH·4H2O (860.24): C, 53.06; H, 6.33; N, 6.51. Found: C, 53.09;
H, 6.33; N, 6.45. MS (ESI): m/z 235 ([Ni(HL)]+, 40%), 253 ([Ni(HL)(H2O)]+, 100). IR (KBr): 3452
(O–H), 3348, 3288 (NH2), 1633 (C=N). Further refluxing in EtOH or drying under vacuum of 4b
did not yield 3b. Second method 1c + 2d: 1c (111.5 mg, 0.33 mmol) was dissolved in EtOH (20
mL) and 2d (157.5 mg, 0.31 mmol) was added, yielding a green suspension that was refluxed for 5
i
h. The intermediate 4b’ was recovered by filtration as light green solid, washed with Pr2O and
dried under vacuum (107.7 mg, 47%). Anal. Calcd (%) for C36H40N4Ni2O6 (742.11): C, 58.26; H,
5.43; N, 7.55. Found: C, 58.50; H, 5.74; N, 7.39. MS (ESI): m/z 249 ([Ni(MeL)]+, 25%), 267
([Ni(MeL)(H2O)]+, 100). IR (KBr): 3346, 3287 (NH2), 1648 sh, 1634 (C=N) cm–1. 4b’ and the left
reaction mixture were further mixed together and refluxed for other 5 h, obtaining 3b as light green
solid, which was filtered, washed with iPr2O and dried under vacuum (85.1 mg, 39%). Anal. Calcd
(%) for C18H18N2NiO2H2O (371.06): C, 58.26; H, 5.43; N, 7.55. Found: C, 58.28; H, 5.51; N, 7.59.
MS (ESI): m/z 353 ([M + 1]+, 100%), 707 ([2M + 1]+, 60), 727 ([2M + Na]+, 70). IR (KBr): 3438
1
(O–H), 1627 (C=N) cm–1. H NMR (CDCl3, 298 K, 400 MHz):  1.32 (2H, H2O), 1.90 (2H, tn
central CH2), 2.23 (3H, Me), 3.56 (4H, tn lateral CH2), 6.5–7.1 (9H, aromatic and N=CH) ppm.
Syntheses of all other derivatives 3 and 4 and their characterization can be found in the ESI.
X-ray crystal structure determinations
Single crystals suitable for X-ray diffraction of 3c (H,OMe) and 3j (NO2,H) were obtained from
i
slow diffusion of Pr2O into CHCl3 solutions, while for 3g (Br,OMe) were obtained from slow
diffusion of n-hexane into an EtOH solution. Crystal data and details of data collection and
refinement are given in Table S1 in the ESI. Intensity data for all compounds were collected on a
Bruker Apex CCD area detector using graphite monochromatic Mo-K radiation. During data
collections, no crystal decay was observed, so that no time-decay correction was needed. Data
reductions were performed with SAINT, and absorption corrections based on multiscan were
obtained with SADABS.80 All the structures were solved by direct methods and refined with
SHELXL-2014/781 implemented in WinGX–2014.1 system.82 The program ORTEPIII was used for
graphics.83 Anisotropic thermal parameters were used for all non-hydrogen atoms. The isotropic
thermal parameters of H atoms were fixed at 1.2 (1.5 for methyl groups) times those of the atom to
which they were attached. All H atoms were placed in calculated positions and refined by a riding
model. Complete crystallographic data were deposited at the Cambridge Crystallographic Data
8

Page 9 of 35
Dalton Transactions
Centre (CCDC 1903483–1903485 for 3c·CHCl3, 3g·EtOH and 3j, respectively). These data can be
View Article Online
DOI: 10.1039/C9DT01216H
obtained free of charge from the CCDC via www.ccdc.cam.ac.uk/data_request/cif.
Second-order NLO measurements
EFISH measurements of derivatives 3, with the exception of 3d (Br,Br) and 3h (NO2,NO2) due to
their too low solubility, were carried out in CHCl3 solutions (concentrations ranging from 5  10–4
to 7  10–5 mol L–1) working at non-resonant incident wavelength  = 1.907 m, using a Q-
switched, mode-locked Nd3+:YAG laser (Atalaser), equipped with a Raman shifter; the apparatus
for the EFISH measurements was made by SOPRA (France). The NLO molecular responses were
measured by the solution-phase direct current EFISH generation method. It provides EFISH, the
direct information on the molecular NLO properties, through the following equation:

EFISH = (gvec / 5kT) + (–2;,,0)
(1)
where gvec/5kT is the dipolar orientational contribution, with vec = (5/3)|| (see note of Table 8)
being the projection along the dipole moment axis of the vectorial component of the quadratic
hyperpolarizability tensor , and (2;,,0) is the cubic electronic contribution to EFISH, which
can be considered negligible for the kind of push-pull molecules here investigated. Such hypothesis
was verified by performing THG(–3;,,) measurements for 3c and 3j derivatives in 2  10–4
mol L–1 CHCl3 solutions, using the same SOPRA instrument, and comparing the ‘purely’ cubic
contribution to hyperpolarizability, THG, with the ‘mixed’ quadratic/cubic one, EFISH. Since the
measured THG, 0.07 and 0.02  10–33 esu for 3c and 3j respectively, were lower than the 5% of

EFISH, falling in the 3–6  10–33 esu for all derivatives, the third order contribution of Eq. (1) can be
confidently considered as negligible.
Computational details
Benchmark DFT and TDDFT calculations were performed on 3c (H,OMe) in order to fix the
protocol to be used for all compounds 3. The X-ray molecular structure of 3c (H,OMe), rather than
that of 3g (Br,OMe) or 3j (NO2,H), was in fact deemed the most appropriate one to be used as
starting point for geometry optimization of all complexes, after substitution of OMe with the proper
D group and of H on the opposite side with the desired A group. On one side, in fact, the presence
of strongly interacting dimers in 3g (see below, Crystal structures description) was presumably
responsible for an otherwise unexpected almost planar conformation of the complex. To verify such
hypothesis, we performed a geometry optimization of a head-to-tail dimer of 3c, starting from the
corresponding experimental one where molecules are laterally shifted giving rise to weak
intradimer interaction. In the optimized 3c dimer, molecules are almost overlapped and
9

Dalton Transactions
Page 10 of 35
significantly more planar with respect to the ‘saddled-shaped’ geometry as obtained for the isolated
View Article Online
DOI: 10.1039/C9DT01216H
molecule, which is close to the experimental one, confirming that strong - interactions in solid
state, as observed in 3g, induce a greater planarity with respect to that expected in solution. On the
other side, the molecular structure of 3j is characterized by a quite distorted ‘step-like’
conformation, which was not deemed to be representative for the whole series of complexes. We
individuated the origin of such distortion in the conformation of the tn (C1–C2–C3) bridge. In the
case of 3c, C1 and C2 are placed on the same side with respect to the N2O2 coordination plane, with
C3 lying on this plane, while in 3j C1 and C3 are on either sides with respect to the N2O2 plane,
which contains also C2. Geometry optimization of 3c starting from the latter conformation leaded
in fact to a ‘step-like’ structure, quite similar to that of 3j, which was 0.29 kcal mol–1 less stable
than the saddle-shaped one. Therefore, the form with tn bridge on the same side with respect to the
N2O2 plane was deemed statistically more probable in solution than the other one, though its
presence should not be excluded. It is to be noted that this conformation correspond to the “up”
isomer as defined in our previous theoretical analysis of [Cu(5-A-5’-D-saltn)] complexes.38
To assess the most appropriate protocol for DFT and TDDFT calculations, we first
evaluated the better performance between the B3LYP84–86 and the M0687 functionals in reproducing
the experimental geometry of 3c, using the 6-311++G(d,p) basis set. The former functional was
used owing to its widely recognized performance, while the latter was tested because it was
properly developed to treat organometallic complexes. On average, a slightly better agreement with
the experimental structure was obtained with the M06 functional, which was then adopted for all
geometry optimizations. Two functionals were then tested to compute the excitation energies, i.e.
M06-2X87 and PBE0,88,89 both of them having been suggested for simulations of absorption spectra
of organic and organometallic molecular compounds.90–93 While both functionals well reproduced
the overall features of the observed UV-vis spectrum of 3c, the M06-2X excitation energies were
significantly shifted towards high energies while a good numerical agreement was obtained using
PBE0, which was then adopted for TDDFT calculations. SHG hyperpolarizabilities, i.e. the (–
2;,) tensors, were computed within the Coupled Perturbed Kohn–Sham (CPKS) approach at
the same frequency as used as in the EFISH experiments. To further confirming the hypothesis of
negligible third order contribution coming from the EFISH measurements (Eq. 1), additional
calculations were performed on 3c and 3j to evaluate, by Finite Field technique, the SHG second
hyperpolarizabilities, i.e. the (–2;,,0) tensors. The resulting || was 15.6 and 8.4% with respect
to gvec/5kT (T = 298 K) for 3c and 3j, respectively, confirming the predominant contribution of
the second-order term in the EFISH measurements for this class of compounds. The CAM-B3LYP
functional,94 which has been recently recommended for hyperpolarizability calculations of mid-size
organic chromophores,90 was adopted. A pruned (99,590) grid was selected for computation and
10

Page 11 of 35
Dalton Transactions
use of two-electron integrals and their derivatives. The solvent (CHCl ) effect on the computed ’s
View Article Online
DOI: 10.1039/C9DT01216H
3
was taken into account through the PCM approach based on the integral equation formalism model
(IEFPCM).95 All calculations were performed with the Gaussian16 suite of programs.96
Results and discussion
Template synthesis and spectroscopic characterization
Reaction of Ni(AcO)24H2O with two moles of substituted 5-G-salH yields the compounds 1a–e of
general formula [Ni(Gsal)2(H2O)2] (G = NO2,74–76 Br,74,75 H,74,75,97,98 Me, OMe, respectively) in
which the nickel(II) ion brings two coordinated deprotonated salicylaldehydato anions occupying
the square planar plane and two apical water molecules.74–76,97–99 Subsequent addition of  2.5
moles (excess) of tn to derivatives 1, or direct synthesis without isolation of 1 starting from 5-G-
salH:Ni(AcO)24H2O:tn in a 2:1:2.5 ratio, led to obtain compounds 2a–d of general formula
[Ni(GL)2] (G = NO2, Br,100,101 H,102 Me, respectively) in excellent yields. The hypothetic product 2e
starting from 5-OMe-salH could not be isolated, maybe due to the high solubility given by the
methoxy group; any attempt of precipitation or isolation led to decomposition to undesired
products. In compounds 2 the nickel(II) ion is hexa-coordinated in an octahedral fashion, with two
tridentate NNO Schiff base ligands GL–, obtained from the mono-condensation of one mole of 5-G-
salH with tn. There are some examples of the synthesis100,102–104 and crystal structures101,105–108 of
these kind of compounds in the literature, as well as similar complexes with different carbonyl
derivatives109,110 or diamines.111–113 By the way, this synthetic methodology cannot be applied using
1,2-ethanediamine (en), since it leads to the symmetric bicondensed product [Ni(salen)], probably
due to the strained 5-membered metallacycle. Main infrared peaks and UV-visible absorption bands
of compounds 1 and 2 are reported in Table 2.
The synthesis of the unsymmetrically-substituted [Ni(5-A-5’-D-saltn)] compounds, 3b, 3c,
3e, 3f, 3g, 3i, 3j, 3k and 3l (A = H, Br, NO2; D = H, Me, OMe, or Br only with the nitro group as
A), was achieved using the template path reported in Scheme 1 from the reaction under reflux in
ethanol of the appropriate derivatives 1 and 2 in a 1:1 ratio. This method revealed to be strictly
necessary, and it was described for the first time by Elder102 and subsequently successfully applied
by Gomes et al.69 In fact, the synthesis of the free ligands 5-A-5’-D-H2saltn should proceed in two
subsequent mono-condensation steps of 5-A-salH and 5’-D-salH to tn. However, our attempts to
obtain such mono-condensation always failed and similar difficulties have been also reported by
others:69 reaction of equimolar amounts of salH (or substituted ones) and tn gave invariably the
double condensation products, H2saltn, even when a 4-fold excess of the diamine was used, making
this synthetic method useless. Less-reactive ketones (acetophenone, acetylacetone, etc.) instead of
11

Dalton Transactions
Page 12 of 35
salH makes possible the isolation of the tridentate proligands HL and then the synthesis of the
View Article Online
DOI: 10.1039/C9DT01216H
unsymmetrically-substituted tetradentate ligands and their metal complexes.59,114–117 SalH bearing
t
bulky substituents (i.e. Bu) with diamines like 1,2-diaminocyclohexane or o-phenylenediamine
also is reported to react to give the monocondensation product.118
Table 2 Main infrared peaks as KBr disks and absorption bands in the UV-vis region as chloroform
solution (10–4 mol L–1) for compounds 1 and 2.
Infrared peaks (cm–1)
UV-visible absorption bands b
(NH2)
(C=X) a
(NO2)
max
max

max
max

1a (NO2)
1b (Br)
–
–
–
–
–
1633
1631
1653
1633
1659
1640
1627
1629
1639
1306
not soluble
–
399
386
25060
25910
25060
23700
25710
26110
27250
26180
4130
3280
2300
2550
12550
7490
7450
7240
349
28650
4620
5510
5620
6830
1c (H)
–
331
345
369
30210
1d (Me)
1e (OMe)
–
399
28990
–
1312
–
422 (sh)
389
27100
2a (NO2) 3313, 3257
not present
not present
not present
not present
2b (Br)
2c (H)
3326, 3257
3326, 3255
3317, 3276
383
–
367
2d (Me)
–
382
a X = O for compounds 1 and N for compounds 2; b max in nm, max in cm–1,  in L mol–1 cm–1
As stated above, in the case of salH and tn the monocondensation can be achieved by
template effect of nickel(II) in compounds 2 or with other metal ions like copper(II).119–121 At this
stage, it would be possible to proceed with the sequestration of nickel as [Ni(dmg)2] (Hdmg =
dimethylglioxime) obtaining the free demetallated proligands HL.122–125 Anyway, the following
reaction with a second differently-substituted salH gave in our hands always small amounts of
symmetrically-substituted derivatives, as detected my mass spectrometry, which in our case
hampers obtaining pure compounds 3 after reaction with nickel(II) salts.
Due to the unsymmetrical substitution of compounds 3, their synthesis following Scheme 1
was achieved with two different combinations, depending on whether the substituents A (and D)
were carried by derivative 1 or 2. Then, each compound 3 has two synthetic methods, except for 3c
and 3g, since 2e could not be obtained. Also the hypothetic compound with A = NO2 and D = OMe
had only one possible combination, 1e + 2a, but even after one months under reflux, we could not
obtain the desired compound [Ni(5-NO2-5’-OMe-saltn)] (3l) but always recover a mixture of the
two starting materials. The reaction times for the different compounds 3 are very different,
depending on the substituents and their positions (See Experimental Section). For example, 1a + 2c
12

Page 13 of 35
Dalton Transactions
yielded 3j in only 5 h, while the mixture 1c + 2a needed to be refluxed for 12 h before total
View Article Online
DOI: 10.1039/C9DT01216H
conversion in the final product.
Compounds 3a, 3d and 3h with symmetrically-substituted ligands, [Ni(5,5’-G2-saltn)], G =
H, Br, NO2, respectively, were synthesized by a modification of literature procedures,77,78 using
nickel(II) chloride and the addition of Et3N to deprotonate the proligands 5,5’-G2-H2saltn, obtained
in situ without isolation (see Experimental Section). The use of the chloride salt instead of
nickel(II) acetate revealed to be necessary since the reaction of 5,5’-G2-H2saltn (either previously
isolated or in situ) and Ni(AcO)24H2O in a 1:1 ratio led to the trinuclear derivatives [Ni3(-5,5’-
G2-saltn)2(-AcO)2] (G = NO2, Br, H).78,126–129
Table 3 Infrared stretching bands (cm–1) and ESI+ mass peaks (m/z) for compounds 4.
compound
reaction
1c + tn
1d + 2c
1c + 2d
1c + 2b
1b + 2c
1d + 2b
1e + 2b
G
H
G’
H
m/z
(NH2)
(C=N/C=O)
1633
4a
4b
4b’
4e
3344, 3290
3348, 3288
3346, 3287
3342, 3293
3349, 3293
3339, 3290
3341, 3290
–
H
Me
H
1633
235 [Ni(HL)]+
249 [Ni(MeL)]+
313 [Ni(BrL)]+
235 [Ni(HL)]+
–
Me
Br
H
1648, 1634
1652, 1631
1638
H
4e’
4f
Br
Br
Br
Me
OMe
1631
4g
1637
–
In some cases, after few hours under reflux of the derivatives 1 and 2, it was possible to
isolate light green compounds whose elemental analyses were close to the ones for the desired
products 3, but still show two NH2 stretching bands in their infrared spectra. Those bands were at
higher wavenumbers compared to complexes 2, and the C=O or C=N band was also shifted (Table
3), meaning that the starting materials reacted to a different compound, which is not the final
product 3. Furthermore, peaks due to the [Ni(GL)]+ fragments are still visible in the mass spectra
(Table 3), meaning that the condensation has not yet occurred. These intermediates 4 are most
probably dinuclear species (Scheme 1) in analogy to what recently reported by Ghosh et al,70
G
formed by scrambling of the ligands G’sal– and L– between the two nickel ions of compounds 1
and 2. Interestingly, from the synthesis of 3a with the Holm’s method79 by reaction of
[Ni(Hsal)2(H2O)2] (1c) with tn in a 1:1 ratio under reflux in ethanol for 2 h, we could also isolate the
intermediate [Ni2(-HL)2(Hsal)2] (4a), visible in the infrared spectrum by the appearance of the NH2
stretching bands at higher wavenumbers compared to 2c, as observed for the other intermediates 4.
It is noteworthy to observe how no intermediates with NO2 group could be isolated, but compounds
1 and 2 directly scramble their ligands, and condensation to the unsymmetrically-substituted
13

Dalton Transactions
Page 14 of 35
derivatives 3 occurred on nickel(II) sites immediately after, suggesting an effect due to the electron-
View Article Online
DOI: 10.1039/C9DT01216H
withdrawing nature of the substituent.
X-ray crystal structures of 3cCHCl3, 3gEtOH and 3j
i
Crystals suitable for X-ray structure determination were obtained by slow diffusion of Pr2O into
CHCl3 (3c and 3j), or n-hexane into EtOH (3g). Selected bond distances and angles are provided in
Table 4, and an ORTEP diagram of their structures with atom numbering scheme is shown in Figs.
1–3. All compounds crystallize in head-to-tail stacked couples to form centrosymmetric pairs.
In 3cCHCl3, the two molecules of the centrosymmetric pair are slightly shifted one with
respect to the other, leading to a relatively long NiNi1-x,1-y,1-z distance, 3.7035(4) Å. The shorter
interatomic contact within a pair is C5C171-x,1-y,1-z, 3.3100(25) Å, while the NiN21-x,1-y,1-z
distance, 3.5162(16) Å, is too long to consider the Ni ion in a square-pyramidal coordination. Each
centrosymmetric pair interacts with an adjacent one through weaker  interactions (shorter
contact, C14C172-x,2-y,1-z, 3.3681(26) Å), affording a zig-zag motif (Fig. 1). The molecule assumes
an almost planar conformation, with a slight deformation towards a saddle shape where the dihedral
angle between the least-squares (l.s.) planes through the aromatic rings is 13.6(1)°. The nickel(II)
ion resides in a slightly distorted square planar environment, with a maximum deviation from the
N2O2 l.s. plane equal to 0.003(1) Å. The six-membered chelate ring NiN1C1C2C3N2 assumes
approximately a screw-boat conformation ( = 79.0(2)°,  = 66.9(1)°), with a total puckering
amplitude Q = 0.617(2) Å.130,131 The structure is stabilized by a co-crystallized chloroform
molecule, hydrogen bonded with both the phenoxido oxygen atoms O1 and O2 (HO distances,
2.41(1) and 2.19(1) Å, C–HO angles, 143.1(1) and 155.5(1)°, respectively, Fig. 1).
Table 4 Experimental coordination distances (Å) and angles (°) for 3cCHCl3, 3gEtOH and 3j.
3cCHCl3
3gEtOH
3j
Ni–O1
Ni–O2
Ni–N1
Ni–N2
NiNi a
1.8489(12)
1.8501(11)
1.8971(14)
1.9115(13)
1.859(3)
1.855(3)
1.906(4)
1.908(4)
1.8573(9)
1.8446(10)
1.8757(11)
1.8679(11)
3.4662(3)
3.7035(4)
3.3632(11)
O1–Ni–O2
O1–Ni–N1
O2–Ni–N2
N1Ni–N2
O2Ni–N1
78.57(5)
92.74(6)
91.81(5)
96.89(6)
171.26(6)
78.59(13)
92.18(16)
92.55(14)
96.70(16)
170.73(16)
83.94(4)
92.11(4)
92.86(5)
91.46(5)
172.13(5)
14

Page 15 of 35
Dalton Transactions
170.36(6)
0.003(1)
170.97(15)
0.007(2)
175.36(5)
0.028(1)
View Article Online
DOI: 10.1039/C9DT01216H
O1Ni–N2
b
NiN2O2
a
symmetry operations are: 1–x,1–y,1–z (3c), –x,2–y,–z (3g) and 1–x,–y,–z
(3j); b metal distance from the N2O2 l.s. coordination plane.
Fig. 1 Stacking of centrosymmetric pairs of molecules in 3cCHCl3 with short intermolecular
contacts. Thermal ellipsoids are drawn at 30% probability level. Colour code: Ni = green, Cl =
yellow-green, O = red, N = purple, C = blue, H = white.
In compound 3gEtOH, the two molecules of the centrosymmetric pair are almost
overlapped, giving rise to a relatively short NiNi-x,2-y,-z distance, 3.3632(11) Å. This packing could
be considered responsible for an almost planar conformation of the complex, where the dihedral
angle between the l.s. planes through the aromatic rings is 3.3(1)°. The only atoms significantly
deviating from the molecular plane are the carbon atoms of the tn chain, which is partially
disordered over two positions, A and B (in Fig. 2 only the most populated site, A, is reported). Both
the disordered chelate rings assume approximately a twisted conformation ( = –99.6(5) and
79.2(9)°,  = 83.1(4) and 98.7(6)° for A and B rings, respectively) and are more puckered than the
chelate ring of 3cCHCl3 (Q = 1.78(5) and 1.93(9) Å, respectively). Adjacent dimers interact
through weak C–HO interactions. The structure is stabilized by a co-crystallized EtOH molecule,
whose hydroxyl group forms a bifurcated hydrogen bond with the phenoxido oxygen atoms O1 and
O2 (HO distances, 2.17(1) and 2.55(1) Å, O–HO angles, 161.3(1) and 112.1(1)°, respectively).
15

Dalton Transactions
Page 16 of 35
View Article Online
DOI: 10.1039/C9DT01216H
Fig. 2 ORTEP plot of a centrosymmetric pair of 3gEtOH with short intermolecular contacts.
Thermal ellipsoids are drawn at 30% probability level. Colour code: Br = brown, Ni = green, O =
red, N = purple, C = light blue, H = white.
In 3j (see (Fig. 3), the two molecules of the centrosymmetric pair are slightly shifted one
with respect to the other and the shortest intradimeric distance is C9C171-x,-y,-z, 3.215(2) Å. The
complex assumes a step-like shape, probably as a consequence of the conformation of the tn bridge
placing C1 and C3 on either sides with respect to the N2O2 plane. The l.s. planes through the two
aromatic rings are almost parallel (the dihedral angle between them is equal to 5.0(1)°) but strongly
deviate from the N2O2 l.s. plane (the dihedral angles formed by the latter with the C5–C10 and the
C11–C16 l.s. planes measure 25.16(3) and 27.87(4)°, respectively). In such a quite distorted
conformation, the nickel(II) ion shows a slightly larger deviation from the N2O2 l.s. plane (0.028(1)
Å) with respect to 3c and 3g, towards the oxygen atom O2 of the pair stacked molecule (NiO21-x,-
= 3.339(1) Å). It is worth noting that O2 belongs to the donor sal moiety of the
y,-z
unsymmetrically-substituted saltn ligand. The chelate ring assumes a twisted conformation ( =
88.4(1)°,  = 90.2(1)°) with total puckering amplitude Q = 0.770(1) Å, comparable with that of
3cCHCl3.
16

Page 17 of 35
Dalton Transactions
View Article Online
DOI: 10.1039/C9DT01216H
Fig. 3 ORTEP plot of a centrosymmetric pair of 3j with short intermolecular contacts. Thermal
ellipsoids are drawn at 30% probability level. Colour code: Ni = green, O = red, N = purple, C =
light blue, H = white.
Optimized Geometries
The coordination bond lengths for all the optimized compounds are reported in Table 5. By looking
at the two Ni–O and Ni–N distances obtained for the symmetrically-substituted 3a (H,H), 3d
(Br,Br) and 3h (NO2,NO2), differing by about 0.010 Å, it appears that these complexes can be
intrinsically considered non symmetric from a structural point of view, a feature already pointed out
in our previous report on the analogues [Cu(5-A-5’D-saltn)] derivatives38 and observed in the X-ray
molecular structures of [Cu(saltn)]132,133 but also in some [Ni(saltn)] ones134 with unoccupied fifth
coordination position. We ascribed such asymmetry to the strain arising from the presence of the
three fused six-membered rings, which is mainly relaxed through a widening of the N–Ni–N angle
with respect to the O–Ni–O one (clearly observed in both X-ray and optimized structures), but also
by the observed non-equivalence of the coordination bond lengths.
Substitution of the A group on the saltn skeleton (compare e.g. 3a with 3e and 3j, or 3b with
3f and 3k, or even 3c with 3g and 3l) influences in a greater extent the Ni–O distances with respect
to the Ni–N ones, with almost negligible variations going from H to Br but more important ones
from H to NO2. In particular, addition of the NO2 group results in an increase of the Ni–O1 and Ni–
N1 bond lengths by about 0.010 and 0.006 Å (acceptor side), respectively, and a decrease of the
Ni–O2 and Ni–N2 bond lengths by about 0.009 and 0.005 Å (donor side), reflecting the tendency of
the nickel(II) ion to move closer to the donating sal moiety. Comparison with the X-ray structures
of 3c, 3g and 3j indicates that the coordination bond lengths are satisfactorily reproduced within
0.01 Å with the only exception of the Ni1–N distances of 3j, which are significantly shorter (by
about 0.04 Å) in the experimental geometry with respect to the optimized one. Such discrepancy
could be however ascribed to the severely distorted conformation of the X-ray structure of 3j, as
also suggested by comparison with 3cCHCl3 and 3gEtOH. In the latter structures not only the
17

Dalton Transactions
Page 18 of 35
Ni1–N bond lengths are 0.02-0.04 Å longer than those of 3j, but the O1–Ni–O2 and N1–Ni–N2
View Article Online
DOI: 10.1039/C9DT01216H
angles are respectively greater and smaller by about 5° than those observed in 3j.
Table 5 Coordination bonds (Å) and angles (°) for the optimized geometries of 3a–3l.a
compd.
A
H
D
H
Ni–O1
1.854
Ni–O2
1.864
Ni–N1
1.906
Ni–N2
1.917
O1–Ni–O2 N1–Ni–N2
81.94
81.98
82.04
81.98
81.95
81.95
81.99
81.75
81.86
81.81
81.81
81.84
94.53
94.51
94.54
94.49
94.55
94.54
94.58
94.57
94.57
94.63
94.61
94.66
3a
H
H
Me
OMe
Br
1.854
1.854
1.853
1.863
1.862
1.864
1.907
1.906
1.906
1.908
1.908
1.908
1.906
1.910
1.912
1.912
1.912
1.917
1.915
1.917
1.915
1.915
1.914
1.915
1.913
1.912
1.911
1.910
3b
3c
3d
3e
3f
Br
Br
H
1.856
1.861
Br
Me
1.856
1.860
Br
OMe
1.857
1.855
1.861
1.864
1.864
1.865
1.859
1.866
1.858
1.855
1.855
1.853
3g
3h
3i
NO2 NO2
NO2
NO2
NO2
Br
H
3j
Me
3k
3l
NO2 OMe
a Calculations performed at M06/6-311++G(d,p) level of theory; O1 and N1 belong to the sal
moiety with A group, while O2 and N2 to the sal moiety with D group.
The NiN2O2 distances, going from 0.026 to 0.031 Å for all optimized structures, are
slightly larger than those observed in 3cCHCl3 and 3gEtOH, further suggesting some role of the
head-to-tail intradimeric stacking on the molecular conformation in the solid state. Such effect is
also shown by a more pronounced saddle shape of the optimized geometries. The dihedral angles
between the l.s. planes through the aromatic rings of the two sal moieties are in the range 31.33–
33.73°, higher than the experimental ones due to the absence of head-to-tail pairs. A modulation
given by the donor group side can be observed in the three series 3a–3c (32.42, 32.32 and 32.22°),
3d–3g (33.21, 32.40, 32.34 and 31.86°) and 3h–3l (33.73, 33.19, 32.33, 31.84 and 31.33°) on going
from NO2 to Br, H, Me and OMe.
Experimental and simulated electronic absorption spectra
The electronic spectra of compounds 3 in CHCl3 at 10–4 mol L–1 solutions are reported in Fig. 4a in
the 10000–35000 cm–1 (1000–285 nm) region, where it can be clearly discerned the distinctive
18

Page 19 of 35
Dalton Transactions
behaviour of complexes without (3a, 3b, 3c, 3e, 3f and 3g) and with NO2 (3i, 3j and 3k). VIienw AfraticcletO, nline
DOI: 10.1039/C9DT01216H
the presence of the nitro as acceptor group causes the threefold increase in the intensity of the broad
absorption from 20000 to 33300 cm–1 (500–300 nm). This effect can be also observed comparing
[Ni(salophen)]51 with [Ni(4,4’-bis-diethylaminosalo-4,5-dinitro-phen)].55 The list of the observed
maxima with corresponding molar extinction coefficients, , are reported in Table 6, where it is
possible to recognise mainly two or three absorptions together with shoulders, assignable to the
convolution of MLCT and ILCT transitions (see below for further details). The d-d transition falls
around 16700 cm–1 (600 nm) with very low  ( 80 L mol–1 cm–1),79 which makes very difficult to
see it in the absorption spectrum. It has also to be noted that Br as substituent seems to behave as a
weak donor group, similar to Me (compare the spectra of 3b with 3e, 3c with 3g and 3i with 3k),
rather than a weak acceptor group: this can be clearly seen on going from 3b to 3f, where the
introduction of Br in the combination (Br,Me) causes only very subtle modification in the
absorption spectrum with respect to (H,Me). The same feature can be recognised in derivatives 1
and 2 in Table 2.
Table 6 Absorption bands of compounds 3 in the UV-visible region in CHCl3 solution (10–4 mol L–
1, optical path = 0.5 cm).a,b
MLCT
max
ILCT *
ILCT *
A
D
max

max
350
351
357
353
354
357
max

max
316
321
max

3a
3b
3c
3e
3f
H
H
419
23850 7460
23640 5500
23200 5520
23700 5240
23310 5610
22990 5740
28600 8960
28490 6530
28010 6750
28330 6570
28250 6240
28010 6030
27860 15690
27930 17770
27700 15890
31650 8030
31150 7030
H
Me
423
H
OMe 431
330sh 30300 5920
322sh 31060 5650
327sh 30580 5730
Br
Br
Br
H
422
429
Me
3g
3i
OMe 435
331
30210 5920
NO2 Br
NO2
NO2 Me
383 c
382 c
381 c
26110 14580 359
26110 17960 358
26240 15080 361
3j
H
3k
a 3d and 3h not soluble; b max in nm, max in cm–1,  in L mol–1 cm–1; c convolution of MLCT+ILCT
Excitation energy calculations on compounds 3 (see Table S2 in ESI for the transitions with
oscillator strength f > 0.01) reproduce satisfactorily the trend observed in the absorption spectra, in
particular: i) it is recognized the presence of three main CT bands (besides the low energy d-d
transitions); ii) the presence of the D group implies systematically a red-shift of the three transitions
(compare e.g. 3a with 3b and 3c) while no clear trend could be associated with the presence of the
A group; iii) introduction of NO2 leads to an increase of the oscillator strengths for the highest
19

Dalton Transactions
Page 20 of 35
energy transitions, though not so marked as observed in the experimental spectra. By analysing the
View Article Online
DOI: 10.1039/C9DT01216H
major contributions to the transitions, it is confirmed the attribution of the three bands to MLCT
(for the lowest energy one) and ILCT (for the two highest energy ones) character, though the
contribution of the metal’s d orbitals to the HOMO is generally quite low (d population  0.11), as
also evident by looking at the isodensity surface plots of the frontier orbitals mainly involved in the
computed transitions for 3c and 3j (see Fig. S1 in the ESI). The MLCT nature of the lowest-lying
transition is in agreement with what previously reported for N2O2 Schiff base nickel(II)
complexes.50,54 Nevertheless, the accessibility to ILCT charge transfer transitions with similar
energies is in this case present, especially because of the blue-shift of the MLCT band compared to
[Ni(salophen)].51 This fact does confirm the expected effect of the distortion from perfectly square
planar geometry around the metal centre induced by the flexibility of the tn bridge between the two
sal moieties, as observed from the X-ray molecular structures and confirmed in solution from DFT
calculations. For comparison purposes, excitation energy calculations were computed also for two
selected analogue copper(II) complexes, namely Cu3a and Cu3l (see Table S2 in ESI), where the
substitution is the same as in the nickel(II) complexes 3a and 3l. Interestingly, the complexes Cu3a
and Cu3l reveal the opposite LMCT character of the lower energy band with an enhanced
contribution of the metal’s d orbitals (d population about 0.60), in agreement with what previously
reported by us with a different DFT approach.38
UV-visible studies at increasing dilutions from 10–3 down to 5  10–6 mol L–1 were also
performed for 3c (H,OMe) and 3j (NO2,H) in the 20000–35000 cm–1 (500–285 nm) region in order
to inspect the presence of aggregation bands due to dimeric units and unambiguously assign the
absorption maxima reported in Table 6. In fact, as observed in their X-ray crystal structures, these
dipolar molecules tend to couple in head-to-tail pairs, which might survive even at a concentration
as low as 10–4 mol L–1. Furthermore, this aggregation phenomenon is well known for salen-type
zinc(II) derivatives,135–139 which show dimeric blue-shifted absorption bands that decrease in
intensity upon high dilution of the solution down to 10–6 mol L–1 with the formation of
batochromic-shifted absorptions of the monomeric units.135 The results obtained for 3j and 3c are
reported in Fig 4b and Fig. S2a in the ESI, respectively, and it can be observed that there are only
very minor changes in the absorption profile for both complexes, the most notable one being the
increase of intensity of the maximum at 358 nm with respect to the one at 383 nm for 3j (Fig. 4b).
This allows us to exclude the presence of aggregation in solution, if not in really minor percentage.
20

Page 21 of 35
Dalton Transactions
View Article Online
DOI: 10.1039/C9DT01216H
Fig. 4 UV-visible absorption spectra of a) compounds 3 in 10–4 mol L–1 CHCl3 solutions, and b)
dilution studies of 3j from 10–3 down to 5  10–6 mol L–1 CHCl3 solutions.
An explanation for the observed difference between zinc(II) and nickel(II) complexes can
arise from the fact that ZnO intermolecular interactions are much stronger and more specific with
the former metal ion (i.e. 2.05 Å for [Zn(salen)]140,141) than in the case of our nickel(II) derivatives,
in which we could observe again short but much longer NiO intermolecular contacts only in 3j
(3.339(1) Å) while in the other cases there are non-specific contacts.
Compounds 3c and 3j as representatives of the two categories without and with NO2 group,
respectively, were also subjected to qualitative solvatochromism studies. For the former compound
without nitro group, the MLCT band shows a hypsochromic behaviour on going from 23100 cm–1
(433 nm) in the less polar toluene to 24270 cm–1 (412 nm) in the most polar MeOH (Fig. S2b in the
ESI), with a small but gradual shift with the polarity of the solvents from one extreme to the other.
This is in agreement with the computed small decrease (1.41 D) of the dipole moment for its
HOMO  LUMO transition (see Table S2), which also justifies the absence of hypsochromic shift
of the CT band upon addition of increasing amount of DMSO in the CHCl3 solution (Fig. S2c in the
ESI). In fact, the MLCT band in pure DMSO is at 23590 cm–1 (424 nm), only 7 nm away from the
value in pure CHCl3, 23200 cm–1 (431 nm).
21

Dalton Transactions
Page 22 of 35
In the case of 3j, moving from low-medium polar (toluene, AcOEt, CHCl3, CH2Cl2) to
View Article Online
highly polar (CH3CN, EtOH and MeOH) solvents with the exception of DMSO andDDOI:M10F.10,3t9h/Ce9rDeT0i1s216H
a decrease in intensity of the maximum at 27900 cm–1 (358 nm) with a concomitant increase in
intensity of the signal at 26100 cm–1 (382 nm) in CHCl3, together with its batochromic shift to
25840 cm–1 (387 nm) in MeOH, as it can be observed in Fig. S3a in the ESI. This solvatochromism
is opposite to what observed for 3c and to what highlighted in the previous dilution study, so
confirming the absence of stacked pairs in solution. Our data suggest that 3j should possess an
excited state with a higher dipole moment with respect to the ground state, and this is fully
consistent with what observed for the analogue copper(II) complexes in the presence of the nitro
group.38 Though such band is a convolution of both MLCT (or LMCT in copper(II) complexes) and
ILCT bands, its observed red-shift is confirmed by the computed increase (equal to 5.39 D for 3j)
of the dipole moment from the ground to the excited state of the HOMO  LUMO transition.
The behaviour observed in DMF and DMSO, where the MLCT band shows the highest
hyperchromic shift (397 and 404 nm, respectively) even if those solvents are not the most polar in
the Reichardt scale, could be also caused by the apical interaction of the solvent molecules. To
prove this point, we added increasing amount of DMSO to the 10–4 mol L–1 CHCl3 solution of 3j
and the results are reported in Fig. S3b in the ESI. The addition of gradually increasing amount of
DMSO from 5 up to 200 L (2800 mol, 14,000-fold excess) leads the spectrum to an increase in
the molar absorptivity especially for the band at higher wavelength, which moves from 382 to 388
nm. The spectrum becomes more similar to the one in pure DMSO, which then suggests that we are
observing the combined effect of the increasing polarity of the solvent mixture together with the
interaction of the solvent with the nickel(II) centre. The difference observed with 3c can be
reasonably justified by the presence of the electron withdrawing NO2 group in 3j, which makes the
metal more eager of further donation from surrounding molecules, while the electron donating OMe
is sufficient to avoid such an interaction.
NLO properties
The NLO properties of selected [Ni(5-A-5’-D-saltn)] complexes as obtained by EFISH
technique are reported in Table 7. EFISH experiments provide the second-order NLO response
(vec) along the g direction, which coincides, in the case of linear or pseudo-linear push-pull
molecules, with that of the intramolecular CT transition. In complexes 3, as previously reported for
the analogue copper(II) complexes,38 two main CT directions can be identified, one from the A
towards the D groups (approximately, the x-axis) and the other from the N2 towards the O2 atoms of
the donor set (approximately, the y-axis), as indeed present in the unsubstituted [Ni(saltn)] (see
below). The z component of CT, when present, may be neglected on a first approximation because
22

Page 23 of 35
Dalton Transactions
it derives from the distortion of complexes from planarity and it is therefore strictly conneVcietwedArt,iclienOnline
DOI: 10.1039/C9DT01216H
both direction and magnitude, with their conformational flexibility in solution. Therefore, vec
values might not be fully informative about the multiple origin of the NLO response in complexes
3. Aware of this problem, we are here reporting the EFISH measurements in combination with
theoretical calculations, in order to distinguish the different components and compare the NLO
response of nickel(II) derivatives 3 with that of the analogue copper(II) ones38 and of all the other
nickel(II) compounds with similar N2O2 tetradentate Schiff base ligands reported in the literature,49–
51,53–55,57,58,65,142 mainly analysed by means of EFISH measurements.
Table 7 Second-order EFISH data at 1907 nm of compounds 3 in CHCl3 solution (error on the
experimental gvec data is about 10%).
a
concentration
(mol L–1)
gvec
vec
compd.
A
H
D
H
( 10–30 D cm5 esu–1) ( 10–30 cm5 esu–1)
5  10–4, 10–4
5  10–4, 10–4
5  10–4, 10–4
5  10-4, 10-4
5  10–4, 10–4
5  10–4, 10–4
10–4 b
73, 98
48, 54
73, 108
52, 57
63, 72
56, 65
55 b
3a
3b
3c
3e
3f
610, 820
390, 440
710, 1050
480, 520
570, 660
580, 670
630 b
H
H
Me
OMe
Br
H
Br
Me
OMe
Br
Br
3g
3i
NO2
NO2
NO2
5  10–4, 10–4
2.5  10–4, 10–4
H
42, 66
54, 61
3j
520, 820
680, 770
Me
3k
a
Extracted using the g values computed on the M06/6-311++G(d,p) optimized geometry at
CAM-B3LYP/6-311++G(d,p)/IEFPCM(CHCl3) level of theory, see Table 8 for complete
data; b 3i not soluble enough for the highest concentration.
The measured gvec quantities are in the same range as that of the corresponding [Cu(5-A-
5’-D-saltn)] complexes previously reported38 at the same concentration (400–700 vs. 300–600 
10–30 D cm5 esu–1, respectively, at 5  10–4 mol L–1). Measurements on 10–4 mol L–1 solutions
usually provided slightly larger gvec values than those obtained at 5  10–4 mol L–1. By the way,
the observed increase is only about 8–14%, i.e. comparable to the experimental error (10%), with
the exception of 3a, 3c and 3j (25–37%). At working concentrations lower than 10–4 mol L–1 the
EFISH response started to be very weak: measurements at 7  10–5 mol L–1 CHCl3 solution of 3c
and 3j gave gvec values equal to 1240 and 900  10–30 D cm5 esu–1 but with higher error (about
25%). This means that the observed increase of gvec values does not reflect a real effect of
23

Dalton Transactions
Page 24 of 35
dilution, in support of the UV-visible spectroscopic studies which suggest discarding the pVrieewsAernticcleeOnline
DOI: 10.1039/C9DT01216H
of aggregated pairs in solution.
Using the dipole moments computed in CHCl3 (see Table 8), the vec values reported in the
last column of Table 7 are obtained. Comparison with the values of corresponding copper(II)
complexes,38 for the few cases where they are available, i.e. (H,H), (H,OMe) and (NO2,H),
indicates that the NLO response of the nickel(II) complexes is slightly higher than that of copper(II)
ones. On the other hand, any systematic trend of the NLO response due to the different A–D
substitutions cannot be undoubtedly discerned in either series of complexes.
The dynamical hyperpolarizability tensors 1.907 of all nickel(II) complexes were also
computed in CHCl3 by means of Coupled-Perturbed (i.e. analytical) techniques in order to shed
light on the molecular origin of the NLO response. The results are reported in Table 8 together with
the ground state dipole moment g and its Cartesian components (see note a of Table 8 and Fig. 5
for representative examples with the adopted reference system). Owing to the presence of two CT
directions, we also report both quantities vec and tot, the latter being the modulus of a “vector”
derived from the full  tensor (see note a of Table 8 for its definition), together with the Cartesian
“components” of tot. vec and tot will coincide (in magnitude) when the CT is unidirectional and
oriented along the dipole moment. This situation is depicted by the symmetrically-substituted
compounds 3a (H,H) and 3d (Br,Br), for which only the y-components of both g and tot are
significantly different from zero. Interestingly, for 3h (NO2,NO2), while the y component is
predominant over the x and z ones, the major component of tot is along the z-axis, even if the
distortion degree from planarity of the optimized geometry toward a saddle conformation is similar
to that of the other derivatives. This is due to the strong NLO effect played by the NO2 groups (see
also below), which are on the same side with respect to the mean molecular plane, providing a z-
contribution to tot. Unfortunately, 3h is practically insoluble in any organic solvent preventing
experimental assessment to this prediction. It is also to be noted that, on account of the different
electronic properties of the H, Br and NO2 substituents, for 3a and 3d the two vectors point toward
opposite directions, giving rise to a negative vec projection, while for 3h they point in the same
direction, resulting into a positive vec projection.
Unsymmetrical substitution on either side of the complexes ‘activates’ in a variable extent
the x component of g and tot vectors, tilting their direction with respect to the y axis, and with the
major effect associated with the NO2 group. In particular, the presence of this group generates a
large x component which prevails on the y one generating a positive vec projection, which is
otherwise negative (see Table 8 and Fig. 5 for the value of the angle  between the two vectors g
and tot). Moreover, introduction of Me or OMe donor group does not affect significantly the x
24

Page 25 of 35
Dalton Transactions
component but strongly increase x (compare 3b, 3c with 3a or 3f, 3g with 3e or even 3k,V3iewl AwrticiltehOnline
DOI: 10.1039/C9DT01216H
3j).
Fig. 5 Directions of the computed g and tot vectors in the xy Cartesian plane (z axis going inside
to complete the right-handed coordinate system), and projection of the latter along the direction of
the former, vec, together with the plot of the optimised geometries for 3c, 3g, 3j and 3l (colour
code: Br = brown, Ni = green, O = red, N = purple, C = blue, H = white).
For comparison purposes, the g and tot vectors were also computed for four selected
analogue copper(II) complexes, namely Cu3a, Cu3c, Cu3j and Cu3l, where the A–D substitution
is the same as in the nickel(II) complexes 3a, 3c, 3j and 3l, at the same level of theory as for the
nickel(II) ones. It should be first noted that the reduced z-component of both g and tot vectors in
the copper(II) complexes is due to their lower distortion from planarity. This is the result of the
increased Cu–O(N) bond lengths compared to the Ni–O(N) ones and the consequent reduced strain
associated with the presence of the three six-membered fused rings. As for the x and y components,
it is observed that those of the dipole moments do not undergo significant variations going from the
corresponding nickel(II) to copper(II) complexes. On the other hand, the y component
systematically increases from Ni to Cu, while the x one is dependent on the nature of the
25

Dalton Transactions
Page 26 of 35
substituents, increasing for the (H,OMe) complex, decreasing for the (NO2,H) one and remaining
View Article Online
DOI: 10.1039/C9DT01216H
essentially unchanged for the (H,H) and (NO2,OMe) derivatives.
Table 8 Computed dipole moments (g, x, y and z, D), SHG ( = 1907 nm) hyperpolarizabilities
(tot, x, y and z, 10–30 cm5 esu–1) in CHCl3 and angle (, °) between g and tot vectors for 3a–l
and selected analogue copper(II) complexes Cu3x (x = a, c, j, l).a
compd.

A
D
g
x
y
z
tot
x
y
z
vec
H
H
8.36
0.23
8.16
1.81
8.6
0.9
–8.0 –3.1
–8.7 –3.0
–8.4 167.6
3a
H
H
Me
8.12
0.32
0.32
0.27
4.03
4.20
7.84
9.37
8.94
8.06
7.81
9.12
2.10
2.72
1.81
1.76
2.13
2.48
10.0
13.5
10.3
9.3
3.9
8.2
–9.1 155.5
3b
OMe 9.76
–10.5 –2.4 –10.4 140.4
3c
Br
Br
H
9.13
9.18
9.12
1.2
–9.3
–8.6
–4.1
–3.4
–9.9 164.0
–8.7 159.3
–8.3 142.2
–8.1 128.2
3d
Br
–1.0
1.5
3e
Br
Me
10.5
13.1
–49.2 –3.3
–11.3 –2.3
3f
Br
OMe 10.34 4.19
6.1
3g
NO2
NO2
NO2
NO2
NO2 11.28 0.42 11.24 –0.74 20.5
2.9
7.5
0.5
–18.9
8.8
64.6
65.1
53.2
51.2
55.3
3h
Br
H
11.39 4.98 10.23 0.67
45.9
44.6
48.5
53.3
12.7
44.6
43.5
47.6
52.4
1.0
–10.9 19.3
3i
12.48 9.07
12.66 9.16
8.51
8.65
1.02
1.21
–6.3
–4.7
–5.2
–12.6
–7.5
–8.4
–8.5
0.0
26.7
30.4
30.3
3j
Me
3k
NO2 OMe 13.65 9.12 10.06 1.39
3l
H
H
H
8.41
0.00
0.06
8.41
0.27
–12.6 172.8
Cu3a
Cu3c
Cu3j
OMe 9.42
9.40 –0.71 18.6
8.76 0.35 41.1
11.7 –14.4 –3.7 –14.3 140.2
NO2
H
12.65 9.12
39.4 –11.8
1.1
20.3
26.4
60.4
61.2
NO2 OMe 13.82 9.22 10.28 –0.65 54.8
53.4 –12.4 –0.7
Cu3l
a
Calculations performed on the M06/6-311++G(d,p) optimized geometry at (CP)CAM-B3LYP/6-
311++G(d,p)/IEFPCM(CHCl3) level of theory. Unrestricted formalism used for copper(II)
complexes. The i and i components are defined according to the following convention, based on
the Gaussian ‘standard orientation’ and uniformed for comparison purposes: the x and y axes lie
on the molecular plane, with x placed in the direction of the largest extension of the molecule,
pointing from the acceptor to the donor group, and y pointing from the O2 to the N2 atoms of the
donor set; the z axis is approximately normal to the N2O2 plane, completing the right-handed
2
2
2 1/2
reference system). Hyperpolarizability is defined as tot = (x + y + z ) and vec = i(ii)/||
26

Page 27 of 35
Dalton Transactions
where i = (1/3)j(ijj + jij + jji).15 In particular tot = (5/3)|| and vec = (5/3)||(z) where || and ||(z)
View Article Online
DOI: 10.1039/C9DT01216H
are the quantities printed in the Gaussian16 output according to the Taylor convention.149
Comparison between computed and measured vec values shows a disagreement in the sign
for the less polar compounds (3a–g) for which the EFISH values are always positive. This result
was already reported for the copper(II) derivatives for which another DFT approach (ROB3LYP/6-
311++G** in vacuo calculations using the finite-field method) was adopted.38 Since the sign of vec
is determined by the relative predominance of the x (positive vec) over the y component (negative
vec) of CT, it can be understood that the final result is ultimately determined by a fine calibration
of both terms. It can be therefore hypothesized that, on one side, the adopted theoretical approaches
suffer from neglecting or incorrectly treating some important contributions to hyperpolarizability,
such as vibrations, dispersion and solvation.143–146 On the other side, since CT values derived from
solvatochromic measurements of 3c (negative CT) and 3j (positive CT), besides those previously
obtained for copper(II) derivatives,38 follow the same trend as the computed values, it is recognized
that hyperpolarizability values strongly relies on the adopted experimental technique. It is also to
be mentioned that the recently reported hypsochromic shift of the CT transition in 3a with
increasing polarity of the solvent148 confirms a negative CT component for the less-polar
derivatives, as here determined. Moreover, even if the two-state model147 could not be applied for
these compounds, showing several CT transitions, it is evident that for 3c most transitions have
small or negative variations of the dipole moment, while for 3j only positive and large variations
are observed.
Literature data available on nickel(II) complexes with similar tetradentate N2O2 Schiff base
ligands can be roughly divided into two categories: i) compounds in which the donor and acceptor
groups have a metal-free nature (i.e. NO2, CN, Cl, OMe, NEt2, etc.) and the NLO properties in
solution were mainly measured by EFISH technique,49–51,53–55,57,142 and ii) compounds in which the
donor group is mainly ferrocene while the acceptor group ranges from the NO2 group to the
positively-charged ruthenocene-like fragment, and the NLO measurements were performed with
Harmonic Light Scattering (HLS) technique.59,61,62,114,115,150–152 The present work can be then
considered more related to the first group of compounds, for which the reported second-order
efficiencies are in line with our data. Nevertheless, the introduction of organometallic push-pull
fragments has widened the modulation possibilities, especially from a switching point-of-view.29
Conclusions
27

Dalton Transactions
Page 28 of 35
Nickel(II) complexes 3a–l with unsymmetrically-substituted A–D saltn ligands could be suitably
View Article Online
DOI: 10.1039/C9DT01216H
obtained by selecting and applying the proper template synthesis,69,102 which avoids the
contamination with the symmetrical derivatives. The electronic nature of the substituents revealed
also to be able to promote or impede the reaction conditions. The flexibility of the propylene bridge
between the two salicylaldimine units in the final compounds 3 shows itself in the structural
diversity of the molecular structures obtained through single-crystal X-ray diffraction studies of 3c
(H,OMe), 3g (Br,OMe) and 3j (NO2,H), as well as in the DFT-optimized structures, for which the
choice of the appropriate starting point revealed to be critical in the final geometry. In particular,
distortion around the nickel(II) centre is able to partially relief the strain arising from the presence
of the three fused six-membered metallacycles, and such saddle-shaped geometry is expected to be
the most stable in solution.
Changing the metal centre from open-shell d9 copper(II) to closed-shell d8 nickel(II) led to a
switch of the charge transfer transition from a pronounced ligand-to-metal character for the
former38 to a metal-to-ligand band. Distortion of the coordination, higher in the case of nickel(II)
with respect to copper(II), causes the presence of energetically low-lying states close to the MLCT
transition, which can contribute to the NLO response. By the way, the effect of the substituents, and
especially the presence of NO2, is the noteworthy increase of the CT and hence the NLO efficiency,
as depicted by theoretical calculations. Furthermore, the reciprocal arrangement of the electron
acceptor A and donor D groups on the organic skeleton modulates both the intensity and the
directions of g and tot vectors, as observed for copper(II) analogues.38

vec values through EFISH measurements set around 50–110  10–30 cm5 esu–1. These values
represent a fraction of the total hyperpolarizability in some instances, especially when NO2 is
present, due to the non-coincidence of the directions of g and tot. Taking into account the  angle
between the two vectors (Table 8) the experimental vec value of 3j (NO2,H), for example, would
derive from a tot of about 110  10–30 cm5 esu–1 (vec = tot cos), an almost double value
compared to the fraction sampled with the EFISH technique. A similar estimation of tot can be
extended to all derivatives, which seems to balance the SHG efficiency to about 100–150  10–30
cm5 esu–1. As stated above, this can derive from the use of the EFISH technique, which might
flatten the NLO responses, regardless the A–D substitution. Anyway, due to the fine tuning given
by the wide range and position of the substituents as highlighted by computed values, the second-
order NLO efficiency of both copper(II) and nickel(II) salen-type complexes can be considered
promising, and this will be object of our future research, especially by exploring further
experimental measurements of the NLO features in order to ascertain the full potentiality of these
compounds.
28

Page 29 of 35
Dalton Transactions
Conflicts of interest
View Article Online
DOI: 10.1039/C9DT01216H
There are no conflicts to declare.
Acknowledgements
This work has been supported by the Italian Ministero dell’Istruzione, dell’Università e della
Ricerca (MIUR). The authors thank Mr. Paolo Mercandalli for some experimental work.
References
1
2
3
P. N. Prasad and D. J. Williams, Introduction to nonlinear optical effects in molecules and
polymers, Wiley, New York, 1991.
S. R. Marder, J. E. Sohn and G. D. Stucky, Eds., Materials for nonlinear optics: chemical
perspectives, American Chemical Society, Washington, DC, 1991.
M. G. Kuzyk, M. Eich, R. A. Norwood and Society of Photo-optical Instrumentation
Engineers, Eds., Linear and nonlinear optics of organic materials III: 4-6 August, 2003, San
Diego, California, USA, SPIE, Bellingham, Wash, 2003.
4
M. G. Papadopoulos, A. J. Sadlej and J. Leszczynski, Eds., Non-linear optical properties of
matter: from molecules to condensed phases, Springer, Dordrecht, 2006.
R. W. Boyd, Nonlinear optics, Academic Press, Amsterdamꢀ; Boston, 3rd ed., 2008.
N. Kamanina, Ed., Nonlinear Optics, InTech, 2012.
M. Wakaki, Optical materials and applications, CRC Press, Boca Raton, FL, 2013.
B. Chen and G. Qian, Eds., Metal-Organic Frameworks for Photonics Applications,
Springer Berlin Heidelberg, Berlin, Heidelberg, 2014, vol. 157.
P. A. Franken, A. E. Hill, C. W. Peters and G. Weinreich, Phys. Rev. Lett., 1961, 7, 118–
119.
5
6
7
8
9
10
11
T. H. Maiman, Nature, 1960, 187, 493–494.
R. J. Collins, D. F. Nelson, A. L. Schawlow, W. Bond, C. G. B. Garrett and W. Kaiser,
Phys. Rev. Lett., 1960, 5, 303–305.
12
13
14
15
16
17
M. G. Kuzyk, K. D. Singer and G. I. Stegeman, Adv. Optics Photon., 2013, 5, 4.
E. Garmire, Optics Express, 2013, 21, 30532.
M. G. Kuzyk, J. Mater. Chem., 2009, 19, 7444.
D. R. Kanis, M. A. Ratner and T. J. Marks, Chem. Rev., 1994, 94, 195–242.
J. A. Delaire and K. Nakatani, Chem. Rev., 2000, 100, 1817–1846.
M. J. Cho, D. H. Choi, P. A. Sullivan, A. J. P. Akelaitis and L. R. Dalton, Progr. Polym.
Sci., 2008, 33, 1013–1058.
18
19
20
21
L. R. Dalton, P. A. Sullivan and D. H. Bale, Chem. Rev., 2010, 110, 25–55.
L. Beverina and G. A. Pagani, Acc. Chem. Res., 2014, 47, 319–329.
A. Qin and B. Z. Tang, Nat. Mater., 2014, 13, 917–918.
H. Kang, A. Facchetti, H. Jiang, E. Cariati, S. Righetto, R. Ugo, C. Zuccaccia, A.
Macchioni, C. L. Stern, Z. Liu, S.-T. Ho, E. C. Brown, M. A. Ratner and T. J. Marks, J. Am.
Chem. Soc., 2007, 129, 3267–3286.
22
O. Mongin, L. Porrès, M. Charlot, C. Katan and M. Blanchard-Desce, Chem. Eur. J., 2007,
13, 1481–1498.
29