A catalytic and DFT study of selective ethylene oligomerization by nickel(II) oxime-based complexes

Article abstract of DOI:10.1016/j.molcata.2012.09.029

The reactivity of nickel(II) thiophenealdoxime complex (3) toward oligomerization of ethylene in the presence of an alkylaluminum co-catalyst has been studied. The complex was found to be a selective ethylene dimerization catalyst in the presence of co-catalysts such as methylalumoxane (MAO) and diethylaluminum chloride (DEAC). With DEAC, the productivity was considerably higher than with MAO. Under optimum conditions the productivity reaches 388 kg/mol catalyst/h/bar with DEAC whereas for MAO this value was 119 kg/mol catalyst/h/bar. Complex 3 displays good ethylene conversions of up to a maximum of 90% with exceptionally high α-selectivity for 1-butene (>99.5%) amongst C₄ products. Computational studies using density functional theory (DFT) were also carried out to ascertain the decomposition pathway for 3 as well as that for Ni(II) complex of the pyridine ketoxime ligand 2. The results suggest that loss of one of the two bidentate oxime ligands attached to the metal center through reaction with DEAC is likely for both 2 and 3. Further, calculations indicate that the subsequent decomposition step was significantly more probable for 3 than for 2 thus explaining why the pyridine ketoxime ligand bound nickel complex 2 was experimentally found to be more stable than the thiophene aldoxime bound nickel complex 3. Calculations also show that the proton of the -OH group (oxime) plays a major role in the stability of the molecules. This was confirmed experimentally by synthesizing the Ni(II) dichloro complex of Pyridine-2-carbaldehyde O-methyloxime 5 and reacting it with ethylene under similar conditions. 5 was found to be highly active even at low co-catalyst concentrations.

Full text of DOI:10.1016/j.molcata.2012.09.029

Journal of Molecular Catalysis A: Chemical 366 (2013) 238–246
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
A catalytic and DFT study of selective ethylene oligomerization by nickel(II)
oxime-based complexes
Kamlesh N. Tayadea, Manoj V. Maneb, Suman Sena, C.N. Murthya, Gopal L. Tembec,∗
,
S. Muthukumaru Pillaic,∗, Kumar Vankab, Soumen Mukherjeec,1
a Applied Chemistry Department, Faculty of Technology & Engineering, The Maharaja Sayajirao University of Baroda, Vadodara 390001, Gujarat, India
b Physical Chemistry Division, National Chemical Laboratory, Pune 411008, Maharashtra, India
c Reliance Technology Group, Reliance Industries Limited, Vadodara Manufacturing Division, Baroda 391346, Gujarat, India
a r t i c l e i n f o
a b s t r a c t
Article history:
The reactivity of nickel(II) thiophenealdoxime complex (3) toward oligomerization of ethylene in the
presence of an alkylaluminum co-catalyst has been studied. The complex was found to be a selective
ethylene dimerization catalyst in the presence of co-catalysts such as methylalumoxane (MAO) and
diethylaluminum chloride (DEAC). With DEAC, the productivity was considerably higher than with MAO.
Under optimum conditions the productivity reaches 388 kg/mol catalyst/h/bar with DEAC whereas for
MAO this value was 119 kg/mol catalyst/h/bar. Complex 3 displays good ethylene conversions of up to
a maximum of 90% with exceptionally high ␣-selectivity for 1-butene (>99.5%) amongst C4 products.
Computational studies using density functional theory (DFT) were also carried out to ascertain the decom-
position pathway for 3 as well as that for Ni(II) complex of the pyridine ketoxime ligand 2. The results
suggest that loss of one of the two bidentate oxime ligands attached to the metal center through reaction
with DEAC is likely for both 2 and 3. Further, calculations indicate that the subsequent decomposition step
was significantly more probable for 3 than for 2 thus explaining why the pyridine ketoxime ligand bound
nickel complex 2 was experimentally found to be more stable than the thiophene aldoxime bound nickel
complex 3. Calculations also show that the proton of the -OH group (oxime) plays a major role in the
stability of the molecules. This was confirmed experimentally by synthesizing the Ni(II) dichloro complex
of Pyridine-2-carbaldehyde O-methyloxime 5 and reacting it with ethylene under similar conditions. 5
was found to be highly active even at low co-catalyst concentrations.
Received 22 May 2012
Received in revised form
21 September 2012
Accepted 29 September 2012
Available online 9 October 2012
Keywords:
Ni(II) complex
Thiophene aldoxime
Ethylene dimerization
Butene-1
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
complexes that have been described as active catalysts in selective
ethylene oligomerization to butenes [6]. However, only a limited
For the past few decades a large number of catalysts have been
reported for selective ethylene oligomerization to give mostly the
C4–C20 range of ␣-olefins [1]. These ␣-olefins are used primarily in
the production of detergents, plasticizers and linear low-density
polyethylene (LLDPE) [1–3]. The nickel (II) based Shell Higher
Olefins Process (SHOP) is one of the early known methods that
gives a wide range of ␣-olefins [4]. Shift in market demand toward
a ‘selective’ olefin rather than a range of linear alpha olefins (LAO)
has lead to the discovery of a number of catalysts that selectively
oligomerize ethylene to the dimer (butene), trimer (hexene) or
tetramer (octene) [5]. Currently, there are a number of nickel(II)
number of catalysts give very high ␣-selectivity.
In our earlier work [7] we had synthesized and characterized
oxime-based nickel(II) complexes (Fig. 1a). The ligands pyridine
aldoxime and ketoxime [8] were employed to prepare Ni(II) com-
plexes 1 and 2 which were demonstrated to be efficient catalysts,
particularly for olefin oligomerization when activated with alky-
laluminums like DEAC and MAO. They were found to be highly
efficient toward ethylene oligomerization with high selectivity for
1-butene [7].
In this study Ni(II) based oligomerization catalysts have been
designed with the help of mechanistic insights gained from den-
sity functional theory calculations. The strategies adopted for this
purpose are: (i) Ni(II) complexes with S-containing oxime ligands
were synthesized and their activity toward olefin oligomerization
tested. The easily synthesizable thiophene aldoxime was the
chosen ligand (Fig. 1b) for two reasons: (a) the nickel(II) complex
3 (Fig. 2) is not reported earlier as a catalyst; (b) the effect of a
different hetero-atom (here sulfur) in catalysis could be studied.
∗
Corresponding authors. Tel.: +91 0265 6696066; fax: +91 0265 6693934.
E-mail addresses: gopal.tembe@ril.com (G.L. Tembe),
muthukumaru.pillai@ril.com (S.M. Pillai).
1
Current address: Gulbrandsen Chemicals Private Limited., Baroda 391440,
Gujarat, India.
1381-1169/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcata.2012.09.029

K.N. Tayade et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 238–246
239
and metal complexes were recorded on a Nicolet 6700 instrument
from 400 to 4000 cm−1. Mass spectra were taken on Thermofisher
PolarisQ MSD in DIP mode and 1H NMR was obtained on a Brüker
400 MHz spectrometer. Polymer grade ethylene (>99.95% purity)
was used directly from a commercial plant, the pressure of which
was adjusted with two-stage regulator. MAO (Chemtura, 5.2 wt.%
Al solution in toluene) and DEAC (11 wt.% solution in toluene) were
used as received. For high pressure reactions, 600 ml Parr reactors
equipped with a heater, a mechanical stirrer and a continuous feed
of ethylene was used. The reactors were activated before use by
heating at 140 ◦C under an atmosphere of nitrogen with subse-
quent flushing with nitrogen. Gas chromatographic analysis of the
reaction products was performed using Shimadzu GC-2010 instru-
ment with a RTX-1 capillary column (30 M length, 0.53 mm i.d. and
film thickness 6.3 mm) working at 40 ◦C for 5 min and then heat-
ing at 15 ◦C min−1 until 250 ◦C. n-Heptane was used as an internal
standard for GC.
Fig. 1. (a) Oxime-based Ni(II) complexes used for selective ethylene oligomerization
and (b) ligand used in this work.
(ii) Calculations were carried out using density functional theory
(DFT) to gain insight into the mechanism by which the Ni(II) oxime
complexes 2 and 3 decompose so as to ascertain their long-term
catalytic efficiencies. The mechanism proposed is based on the
experimental observation of the catalytic activity with complex 2
(as previously reported) and 3 (current work).
Based on the insights gained from the DFT calculations, a modi-
fication was made and a new nickel based catalyst 5 was designed
and tested for ethylene dimerization in presence of alkyl aluminum
co-catalyst. As discussed below, the modified catalyst was seen to
provide significantly better results.
2.1.1. Synthesis of Ni(2-thiophene aldoxime)2 dichloride 3 [9]
To a hot ethanolic solution (10 ml) containing 2-thiophene
aldoxime (0.64 g, 5 mmol) was added a solution of NiCl2·6H2O
(0.60 g, 2.5 mmol) slowly. The contents were heated at 50 ◦C for
10 min and allowed to cool. A greenish yellow microcrystalline
compound separated. This was washed twice with cold ethanol and
dried (0.72 g, 75%). IR (KBr, cm−1): 3310, 1652, 1426, 962, 806, 709,
585. Anal. Calcd for C10H10N2S2O2NiCl2: C, 31.29%; H, 2.63%; N,
7.30% Found: C 31.04, H 2.60, N 7.18. MS (TOF): m/z 383.83 (M+),
312 (M+−2Cl).
The following sections discuss this approach in more detail and
highlight the importance of a tandem experimental and computa-
tional approach toward designing new catalysts.
2.1.2. Synthesis of Ni(pyridine-2-carbaldehyde O-methyloxime)2
dichloride 5
2. Experimental
To 10 ml hot ethanolic solution of NiCl2·6H2O (1.20 g, 5 mmol), 4
(1.40 g, 10 mmol) was added slowly. The solution was refluxed for
1 h. The light green solid product obtained was filtered hot, recrys-
tallized from hot ethanol and dried. (1.43 g, 71%) ꢀeff = 3.18BM. IR
(KBr, cm−1): 3091, 3008, 2963, 2926, 1610, 1472, 1357, 1046, 22,
902, 803. Anal. Calcd for C14H16Cl2N4NiO2: H, 4.01%; C, 41.84%;
N,13.94% Found: H 3.98 C 42.10 N 14.05. MS (TOF): m/z 331.01
(M+−2Cl).
2.1. Materials
All manipulations were performed using standard vacuum line,
Schlenk techniques and Glove Box under a purified nitrogen atmo-
sphere. Toluene was distilled from sodium benzophenone ketyl
and methanol was distilled from magnesium. Solvents were stored
˚
under nitrogen in 4 A molecular sieves. 2-Thiophene carboxal-
dehyde, pyridine-2-carbaldehyde, methoxyamine hydrochloride,
hydrazine hydrochloride and NiCl2·6H2O were purchased from
Aldrich and used as received. The ligands 2-thiophene aldoxime
[9] and pyridine-2-carbaldehyde O-methyl oxime 4 (Fig. 3) [10]
were prepared by a literature procedure. Micro analysis was car-
ried out on a Perkin Elmer C, H and N analyzer. IR spectra of ligands
2.1.3. Synthesis of Ni(pyridine-2-carbaldehyde O-methyloxime)2
diiodide 6 [10]
NiCl2·6H2O (0.24 g, 1 mmol) was dissolved in minimum amount
of hot water (2 ml) and pyridine-2-carbaldehyde O-methyl oxime
(4, 0.27 g, 2 mmol) was added. The green solution was cooled and
Fig. 2. Synthetic protocol for proposed Ni(II) complex 3.

240
K.N. Tayade et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 238–246
Fig. 3. Synthetic protocol for ligand 4.
four obtained transition states are: (i) for 2 2, −1019.64 cm−1; (ii)
for 2 5, −998.61 cm−1; (iii) for 3 2, −1047.72 cm−1; and (iv) for 3 5,
−697.79 cm−1 (for understanding the numbering of the transition
state structures, please see Figs. 4 and 5). Solvent effects were incor-
porated using the COSMO [17] model, with toluene (ε = 2.38) as the
solvent.
a saturated solution of an equivalent amount of potassium iodide
(0.332 g, 2 mmol) was slowly added. The color of the solution turns
brown. Slow evaporation of water gave crystalline material after
several days. The dark brown colored crystals were washed with
methanol and dried (0.49 g, 84%) ꢀeff = 3.26 BM. IR (KBr, cm−1):
3112, 3000, 2929, 1609, 1439, 1157, 1038, 928, 902, 783. Anal. Calcd
for C14H16I2N4NiO2: H, 2.76%; C, 28.75%; N, 9.58% Found: H 2.82 C
28.22 N 9.74. MS (TOF): m/z 331.32 (M+−2I).
In order to test the efficacy of the basis set (TZVP) employed,
single point calculations for a specific pathway – the conver-
sion of 2 1 to 2 4, has been done with the higher, TZVPP basis
set (which is at the basis set limit for DFT) using the optimized
geometries that had been obtained at the TZVP level [18] (see
Table 1 in Appendix A: Supplementary Data). The values indicate
that the results from the TZVPP calculations are quite similar to the
values obtained from the TZVP calculations, thereby indicating the
soundness of the basis set employed). A further check was done
to test the reliability of the functional: BP-86 that was employed
for the calculations. Single point calculations were done with the
geometries optimized at the TZVP/BP-86 level using the B3-LYP
functional, that is, at the TZVP/B3LYP level, for all the reaction path-
ways reported in this work (Appendix A). It is seen that the values
are uniformly higher for every case as compared to the TZVP/BP-
86 level. However, what is important to note is that the trends
observed for every reaction pathway discussed in Section 3 remains
the same for the B3LYP calculations. This means that all the con-
clusions made at the TZVP/BP-86 level remain valid and reliable.
2.1.4. Synthesis of thiophene-2-carbaldehyde O-methyloxime 7
The procedure is similar to that of 4 [10]. (6.1 g, 86%), 1H NMR
(400 MHz, CDCl3). ı = 3.98 (s, 3H), 7.05 (dd, 1H), 7.08 (dd, 1H),
7.51(dd, 1H), 8.21 (s, 1H). 13C NMR (400 MHz, CDCl3): 68.0 (1C,
OCH3), 132.1 (1C, CH), 136.2 (1C, CH), 140 (1C, CH), 145.2 (1C, CH),
155.4 (1C, HC N). IR (KBr, cm−1) 2937, 1674, 1605, 1714, 1307,
1209, 1048, 912, 756. EI-MS: m/z 141.2
2.2. Ethylene oligomerization
Catalytic ethylene oligomerization was conducted in a 600 ml
SS 316 high pressure stirred autoclave (Model 4563 Parr, USA) con-
nected to a process controller (model 4842). Prior to the run, the
reactor was baked to 150 ◦C under a nitrogen pressure (10 kg/cm2)
for 2 h and cooled to ambient temperature. In a typical experiment
a suspension of catalyst 3 (15 mg, 4 × 10−5 mol) in toluene (50 ml)
and DEAC (0.52 ml, 4 × 10−3 mol) in toluene (150 ml) were carefully
charged into the reactor under a nitrogen atmosphere. The Al/Ni
was kept at 100 (mol/mol). The reactor temperature was gradu-
ally raised to 45 ◦C and then pressurized with ethylene to 20 bar
(290 psi). The reaction was continued for 1 h and at the end of this
period the reactor was cooled to 0 ◦C. The gaseous product and
excess ethylene were collected over a calibrated aspirator bottle. To
the rest of the solution in the reactor a 10% HCl solution in methanol
was added to quench the catalyst solution. Both the gaseous and liq-
uid samples were analyzed by GC and the productivity values were
calculated.
3. Results and discussion
3.1. Synthesis and characterization of 3
The ligand 2-thiophenealdoxime was easily prepared by con-
densation of 2-thiophene carboxaldehyde and hydroxylamine as
reported earlier [9]. On treatment of 2 equiv. of this ligand with
1 equiv. of NiCl2·6H2O in hot ethanol, an yellowish green solid was
obtained. Attempts to obtain X-ray quality single crystals of this
solid failed due to its poor solubility in most of the organic solvents.
The IR spectra of the complex shows a sharp peak at 3310 cm−1 cor-
responding to the free OH of the oxime part. Elemental analysis
and mass spectroscopy indicated a ligand to metal stoichiom-
etry of 2:1. A tentative structure of this complex is shown in
Fig. 2.
2.3. Quantum chemical calculations
All the calculations were carried out with density functional
theory using the Turbomole suite of programs [11]. Geometry opti-
mizations were performed using the BP-86 density functional [12].
The electronic configuration of the atoms was described by a triple-
␨ basis set augmented by a polarization function (TURBOMOLE basis
set TZVP) [13]. The resolution of identity (RI) [14] along with the
multipole accelerated resolution of identity (marij) [15] approx-
imations were employed for an accurate and efficient treatment
of the electronic Coulomb term in the density functional calcula-
tions. Frequency calculations were done for the species involved in
order to complete the free energy calculations. In order to account
for the fact that all the species are in solution, the translational
entropy term in all the calculated structures was corrected through
a free volume correction introduced by Mammen et al. [16] Care
was taken to ensure that the obtained transition state structures
possessed only one negative frequency corresponding to the cor-
rect normal mode. The value of the sole negative frequency for the
3.2. Oligomerization of ethylene
The activity of 3 was evaluated for ethylene oligomerization
in presence of alkylaluminum co-catalysts. The effect of reaction
conditions on productivity and selectivity for dimerization were
examined (Table 1). A comparison was also made between the Ni(II)
complex with pyridine oxime ligand (catalysts 1 and 2, Fig. 1) and 3
in terms of their catalytic activities [7]. As studied earlier, two dif-
ferent co-catalysts MAO and DEAC were used. In general, 3 showed
much higher activity with DEAC as co-catalyst though MAO also dis-
played good productivity. In contrast 1 and 2 showed much lower
activity for MAO than DEAC. Other co-catalysts such as ethylalu-
minum dichloride (EADC) and triethylaluminum (TEA) were also
tested; however, highest productivity was obtained with DEAC as

K.N. Tayade et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 238–246
241
Fig. 4. The free energy profile for the decomposition of the nickel–N-S-oxime complex: loss of one oxime ligand; the values in parenthesis are the gas phase values, the
values outside the parenthesis are solvent corrected values; all values are in kcal/mol.
Table 1
Selective ethylene oligomerization data with different Ni(II) oxime catalysts and MAO/DEAC as co-catalysts.
Entrya
Cat.
Co-cat
Cat (mmol)
Co-cat. (Equiv.)
Press. (bar)
%Butene b
%Hexene
Yield (g)
Productivityc
1
2
3
4
5
6
7
8
9
10
11
12b
13d
14e
15e
3
3
3
3
3
3
3
3
3
3
3
1
2
3
2
MAO
MAO
MAO
DEAC
DEAC
DEAC
DEAC
DEAC
DEAC
DEAC
DEAC
DEAC
DEAC
DEAC
DEAC
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.03
0.05
0.04
0.04
0.04
0.04
150
300
450
100
150
200
250
100
100
100
100
100
100
100
100
20
20
20
20
20
20
20
10
15
20
20
20
20
20
20
82
81
78
81
74
77
76
83
87
82
79
85
79
70
82
18
19
22
19
26
23
24
17
13
18
21
15
21
30
18
18
46
47
45(2.4)
114(6.1)
119(6.3)
219(11.6)
388(20.7)
327(17.5)
301 (16.0)
262(6.9)
249(10.0)
214(11.2)
219(11.5)
145(7.7)
123(6.1)
148(7.9)
129(6.4)
87
155
131
120
52
75
64
110
58
49
178
155
a
General Conditions 650 ml Parr Reactor, Toluene(Solvent) = 200 ml, Temperature 45◦ C, Reaction time 1hr
1-Butene selectivity > 99.5%.
b
c
In (kg of product) (mol of catalyst)−1 h−1 bar−1 (values in parenthesis are in kg product/g catalyst/h).
Values of entry 12 and 13 taken from Ref. [7].
d
e
Entries 14 and 15 for 90 min.

242
K.N. Tayade et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 238–246
Fig. 5. The free energy profile for the decomposition of the nickel–N-N-oxime complex: loss of the first oxime ligand; the values in parenthesis are the gas phase values, the
values outside the parenthesis are solvent corrected values; all values are in kcal/mol.
cocatalyst. It was observed that after co-catalyst addition, the insol-
uble catalyst 3 dissolves in toluene and forms a clear red solution.
When this solution was reacted with ethylene a rapid increase in
temperature was noted. The temperature was maintained at 45 ◦C
throughout the reaction with an ice-water mixture. Interestingly,
the selectivity of 1-butene was nearly 100% (in the C4 fraction)
with no solid product formation. Such exceptionally high selectiv-
ity to 1-C4 is surprising considering the simplicity of the ligand
system.
with selectivity toward ethylene dimerization remaining constant
at ∼80%.
3.2.2. Ethylene oligomerization with DEAC as co-catalyst
Reaction of 3 in combination with DEAC (Ni/Al molar ratio = 100)
at an ethylene pressure of 20 bar displayed an activity of
219 kg/mol catalyst/h/bar (entry 4, Table 1). The main products
were butene and hexenes. The amount of butene varied from 70%
to 87% with 1-butene being the exclusive product in the C4 frac-
tion (>99.5% alpha sel.). The C6 fraction consists of hexene isomers
similar to that observed with MAO as co-catalyst.
3.2.1. Ethylene oligomerization with MAO as co-catalyst
When activated with MAO, 3 reacts with ethylene to give pre-
dominantly butene and hexene. The hexene fraction contained
1-hexene and a mixture of other isomers. The distribution of hex-
ene isomers observed with 3 was nearly similar to that obtained
with 1 and 2 under different conditions. These isomeric hexenes
are produced by the co-dimerization of ethylene with 1-butene as
suggested by Braunstein et al. [19]. Significantly no high molecu-
lar weight polyethylene was observed in this reaction indicating
that the rate of chain termination is much faster than the rate
of propagation (kt ꢀ kp). Over two-fold increase in productivity
was observed on increasing the MAO concentration (entries 1–3,
Table 1) from 150 to 300 equiv. (cf. 45 vs.114 kg/mol catalyst/h/bar).
Beyond 300 equiv. of MAO little change in productivity was noted
(i) Effect of catalyst concentration: The concentration of 3 was var-
ied from 0.03 to 0.05 mmol with the catalyst to DEAC ratio of
1:100 at 45 ◦C and 20 bar ethylene pressure (entries 4, 10 and
11, Table 1). An increase in yield was observed on increasing the
catalyst concentration. This trend was similar to that observed
with 1 and 2. However, the overall productivity of the latter
catalysts was lower than 3. Thus, the average productivity for
3 was 219 kg/mol catalyst/h/bar while that for 1 and 2 was 146
and 137 kg/mol catalyst/h/bar, respectively.
(ii) Effect of co-catalyst concentration: The DEAC concentration was
varied from 4 mmol (100 equiv.) to 10 mmol (250 equiv.). It was
observed that the productivity value reaches a maximum at

K.N. Tayade et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 238–246
243
150 equiv. (388 kg/mol catalyst/h/bar) but shows a decreasing
trend thereafter when the amount was increased to 250 equiv.
(301 kg/mol catalyst/h/bar) (entries 4–7, Table 1). This indi-
cates that a catalyst deactivation might be occurring at higher
DEAC concentrations. On the other hand such deactivation
with 1 and 2 was much slower as observed from the produc-
tivity values at similar catalyst to co-catalyst ratios.
(toluene, ε = 2.38) modeled is non-polar, solvent corrections to the
free energy values were found to be quite small.
The mechanism, as illustrated in Fig. 4, involves the coordination
of Et2AlCl (DEAC) to 3 1 followed by the abstraction of hydrogen
from the oxime OH group leading to elimination of ethane from
the complex (Fig. 4a). As can be seen the barrier to this process –
going from 3 1 to 3 3 via transition state 3 2 is 32.1 kcal/mol. This
barrier is quite reasonable considering the reaction to be entrop-
ically highly disfavored. It is to be noted that for all the reaction
pathways considered in this calculation, the reactants have been
assumed to be totally separated species. It is possible that the bar-
riers would be lower if adduct species were considered instead,
which would reduce the entropic cost of the reaction. Nevertheless,
in a relative sense, when comparing different reaction cycles, the
consistent use of this approach of only taking separated reactants
when calculating barriers and products should provide a reliable
route to comparing the pathways. It is quite likely, from the barrier
obtained, that the reaction would be feasible at 45 ◦C, as observed
experimentally. It is also quite likely that the process would lead
further to the formation of the species 3 4, with only one oxime
ligand coordinated to the nickel center (Fig. 4). This is because,
for the species 3 3, the forward pathway leading to 3 4 has a
much smaller barrier (1.3 kcal/mol) than the reverse pathway (bar-
rier = 21.8 kcal/mol) leading to the formation of the complex 3 1.
That the deprotonation of the oxime hydrogen is a major factor for
destabilization of these catalysts is also supported by experimen-
tal observations with the synthesis of the Ni(II) dichloro complex
of pyridine-2-carbaldehyde O-methyloxime 5 and reacting it with
ethylene under similar conditions (see Section 3.4.2). An important
consideration has to be addressed regarding the alkylation of the
nickel center as is the case in the mechanism shown in Fig. 4 before
such an ethane elimination reaction could take place – what if the
ethane elimination was equally facile directly from 3? Calculations
were done to investigate this possibility. For the specific case of the
removal of ethane from 3, only the ꢁE values have been considered
(see Supplementary Data). The ꢁE barrier to ethane formation was
found to be much higher {30.4 kcal/mol (gas phase value)} as com-
pared to the corresponding ꢁE of 20.1 kcal/mol for the conversion
of 3 1 to 3 3. Therefore, it is likely that the decomposition route for
3 is followed only after it has been alkylated. This implies that had
the decomposition of the metal complex been more favorable, no
alkylation could have taken place and the catalyst would not show
any activity.
(iii) Effect of ethylene pressure: The catalytic activity of 3 was
examined at different ethylene pressures at a Ni/Al ratio of
1:100. Interestingly, an increase in ethylene pressure lead
to a decrease in the yield of products (entries 4, 8 and 9,
Table 1). For instance the corresponding yields at 10 bar,
15 bar and 20 bar ethylene pressures were 52, 75 and 87 g,
respectively, while the productivity values were 262, 249 and
219 kg/mol catalyst/h/bar, respectively. This trend in loss of
productivity at higher pressures indicates possible catalyst
degradation. The dimerization selectivity remains between
81% and 87%. Comparing the catalytic performances of 1 and 2
under identical conditions, it was found that the average pro-
ductivity for 3 as catalyst was higher at 20 bar pressure. On the
other hand the productivity for both 1 and 2 was constant even
on increasing the pressure.
(iv) Effect of time: All ethylene oligomerization runs with 3 were
performed for 30 min. To compare the catalyst performance
and stability for a longer period of time, a reaction was
carried out for 90 min. It was observed that the catalyst
does not remain active for longer duration and consequently
the productivity decreases substantially. Thus, the yields for
30 and 90 min runs were 87 and 178 g, respectively (cf.
entries 4 and 14, Table 1), with productivity values of 219
and 148 kg/mol catalyst/h/bar, respectively. In comparison, the
productivity with 2 remains constant suggesting that this cata-
lyst solution remains stable for a longer period of time (entries
13 and 15, Table 1). The higher productivity demonstrated by
catalyst 3 compared to 1 or 2 may be attributed to the bond-
ing nature of central Ni(II) ion to the ring hetero atom and to
the availability of free sites for co-ordination. To understand
the probable active species for 2 and 3 and gain some insight
on the instability of 3, a computational study, employing DFT
was undertaken to elucidate the possible structures and their
corresponding energies, as well as the mechanism for deacti-
vation.
After the formation of 3 3, calculations indicated that the system
needs only 1.3 kcal/mol further amount of energy to decompose
further to form the mono-oxime coordinated 3 4 (Fig. 4). This sug-
gests that 3 4 may be very prevalent in solution and could be the
active catalytic species involved in the oligomerization process. 3 4
can undergo further decomposition (Fig. 4b) through the process of
hydrogen transfer from the ␤ carbon of the ethyl chain to the OH
group of the remaining oxime ligand on the complex to eliminate
water and produce 3 6, which is essentially a molecule that has
lost an alkyl chain through hydrogen transfer. This is likely to be
inactive because it cannot be re-alkylated by DEAC as it does not
have any chlorine atoms that it can exchange, and the absence of
electron donating oxime ligands on this decomposed system would
make it unlikely for entropically unfavored alkene coordination to
take place at the nickel site.
The discussion, so far, of the potential energy surface shown in
Fig. 4 illustrate the complete pathway for the conversion of the
alkylated nickel complex 3 1 to 3 6 which is essentially a decom-
posed and inactive molecule. However, a pertinent point that can be
raised is with regard to the more successful Ni-pyridine-ketoxime
catalyst (2 in Fig. 1a) [7] which is catalytically active for much
longer periods of time without showing appreciable decomposi-
tion: what are the energetics for the corresponding mechanistic
3.3. DFT calculations
Calculations were done to study the possible pathway by which
decomposition of 3 could occur which is observed to be an active
catalyst for the oligomerization of ethylene, but has a major limi-
tation in that it decomposes faster than 1 or 2.
Calculations were first done to determine the lowest energy
structure for methylated nickel oxime precursor complex 3 1 (not
isolated). Out of the several conformations studied it was found that
the lowest energy was the one that had only nitrogen atoms of the
two oxime ligands attached to the nickel center, while the sulfur
atom and the rest of the oxime ring preferred to remain unbounded
to the nickel center (see Fig. 4). This is an interesting result because
both computational and crystal structure data indicate that the pre-
ferred lowest energy conformation for the previously studied nickel
pyridine ketoxime complex 2 1 (Fig. 5) has a bidentate linkage of
the oxime to the central metal [7].
The proposed mechanism for the decomposition of 2 and 3
are shown in Figs. 4–5 with the solvent corrected free energies
(ꢁG) determined for the entire reaction cycle (given in Section
3.2.2). It is to be noted here that since all the reactants, transition
state, intermediates and product species are neutral and the solvent

244
K.N. Tayade et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 238–246
pathway for this catalyst system? Is ethane elimination and subse-
quent alkene formation a less likely process for this system? This
point was addressed by calculations with the Ni-pyridine-ketoxime
system indicated as 2 1 in Fig. 5. Calculations reveal that decompo-
sition of 2 1 to yield ethane and 2 3 is actually even more favorable
(barrier 21.2 kcal/mol) than for 3 1 (barrier 32.1 kcal/mol)! Indeed
experiments with 2 1 did indicate the instantaneous bubbling of
gas upon the addition of DEAC to the solution, which suggests that
a similar decomposition mechanism is operational for this catalyst
system as well. Again, as for 3, the smallness of the barrier points
to the high feasibility of this process occurring during the catalysis.
The subsequent conversion of 2 3 to the species 2 4 is also quite
facile and exothermic (Fig. 5a) and therefore the species 2 5, like
3 5 discussed earlier, may be the actual active catalyst species in
solution.
decomposition mechanism proposed here for both the complexes
2 and 3 is based on the addition of only one DEAC molecule to the
complexes 3 1 and 2 1, (Figs. 4 and 5). One could conceivably con-
sider an alternate mechanism where two molecules of DEAC would
simultaneously approach 2 1 or 3 1, leading to the loss of two
ethane molecules from the two oxime ligands attached to the metal
center.
Such a mechanism is highly unlikely due to entropic reasons:
the more ordered structure resulting from the addition of two DEAC
molecules to the complexes is likely to be highly disfavored. More-
over, it is also unlikely that DEAC could play a part in the subsequent
decomposition step outlined in Figs. 5 and 3, where water is formed
by the intramolecular reaction between the OH of the oxime ligand
and a ␤-hydrogen of the attached ethyl group. This again is because
of entropic reasons: an intramolecular reaction is much more likely
than an intermolecular one because the former is unimolecular
while the latter is bimolecular.
Overall, the results of the calculations discussed here suggest
that a mono-coordinated oxime complex having two alkyl chains
is quite likely to be the active species in solution, for both the
pyridine and the thiophene oxime ligand cases, and that the real
decomposition step is the elimination of the alkyl chain through
transfer of a hydrogen from the ␤-carbon of the chain to the OH
group of the oxime ligand. The importance of the oxime hydrogen
in the decomposition process can be substantiated by substituting
the hydrogen with an alkyl group and carrying out the reaction in
similar conditions.
However, as mentioned earlier, the pyridine ketoxime ligand
bound nickel catalyst system is significantly more long lasting
than the Ni-thiophene-oxime catalyst system and the subsequent
decomposition step may provide the explanation for this. As Fig. 5b
illustrates, the conversion of mono-oxime coordinated nickel com-
plex 2 4 to the water eliminated species 2 6 via the transition state
2 5 has a barrier of 38.8 kcal/mol, significantly higher than the bar-
rier (25.8 kcal/mol) for the corresponding decomposition step for
the Ni-thiophene-oxime coordinated complex. The reason for this
is made clear in Fig. 6 showing the structures of complexes 3 4
and 2 4. In structure 2 4, the pyridine ketoxime ligand has a biden-
tate linkage, binding through both the nitrogen atoms to the nickel
center while in 3 4, the thiophene oxime ligand has a monoden-
tate linkage, binding though only the single nitrogen atom, with
the sulfur atom rotated away from the nickel center. This reduced
coordination at the nickel center in 3 4 leads to the formation of
an agostic interaction between the ␤-hydrogen of one of the ethyl
groups and the nickel center. As Fig. 6 indicates, the bond dis-
tance between the ␤-hydrogen on the ethyl group and the nickel
3.4. Synthesis and characterization of 4–7
DFT calculation with 3 and 2 indicates the abstraction of hydro-
gen from oxime group for both the complexes plays an important
role in the decomposition pathway (Figs. 4 and 5). To verify this
experimentally, and therefore, to improve the stability and activ-
ity of the catalyst, the pyridine-2-carbaldehyde O-methyloxime
ligand 4 was synthesized. In 4, the oxime proton has been sub-
stituted by a methyl group (Fig. 3) [10]. Addition of 1 equiv. of
nickel(II) chloride to 2 equiv. of this ligand in hot ethanol gave a
light green colored solid (5) which was insoluble in most of the
organic solvents. It was thus not possible to obtain crystal struc-
ture of 5 due to its highly insoluble nature. Spectroscopic and
elemental analysis confirms the presence of two ligands and one
nickel(II) in this complex. However, when 2 equiv. of an aqueous
˚
is 1.78 A in 3 4, while the closest hydrogens of the ethyl chains
˚
in 2 4 are more than 2.4 A away, indicating the absence of agos-
tic interactions in 2 4. Since the decomposition step illustrated
in Figs. 4 and 5 involve the transfer of the ␤-hydrogen from the
ethyl group to the OH group of the ligand it is easier to extract
the hydrogen from 3 4 than from 2 4 because of the weakened
interaction of the ␤-hydrogen with the carbon in 3 4, due to its
agostic interaction with the metal center, and hence the observed
differences in the decomposition barriers. It is also noted that the
Fig. 6. A comparison of the structures, “3 4” and “2 4”: the complexes that will be formed after the loss of one oxime ligand from the metal center.

K.N. Tayade et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 238–246
245
Fig. 7. Diamond view of highly disordered
(deg): Ni(1) N(2 A) = 2.063(5), Ni(1) N(1) = 2.073(2), Ni(1) I(1) = 2.7862(2), N(2 A) Ni(1) N(1) = 103.91(16), N(2 A) Ni(1) I(1) = 89.89(14), N(1) Ni(1) I(1) = 88.22(6),
I(1)#1 Ni(1) I(1) = 180.
6 and its chemical structure: ortep cannot be plotted due to disorder. Selected bond lengths (Å) and angles
˚
˚
˚
solution of KI was added to an aqueous mixture of 4 and NiCl2·6H2O
the color changes to brown. Slow evaporation of this solution
gave dark brown crystalline material. This complex was charac-
terized by single crystal X-ray crystallography and was found to
be the nickel(II) di-iodo complex 6 [10]. The sulfur analog of 4,
thiophene-2-carbaldehyde O-methyloxime 7 was prepared by sim-
ilar methods. Several attempts to isolate the Ni(II) complex of 7
were unsuccessful. Possibly in 3, hydrogen bonding between the
chloride atom and the oxime hydrogen might help in stabiliz-
ing the complex. However, in case of 7, the absence of any such
hydrogen bonding probably hinders the formation of the nickel(II)
complex.
180◦. This indicates that the Ni(II) atom is in a nearly perfect octa-
hedral geometry. This is also supported by the I Ni N (1-pyridine)
and I Ni N (2 A) angles which are nearly 90◦. Details of bond
lengths and angles for 6 are provided in Appendix A Supplementary
Data.
3.4.2. Reactivity with 5 and 6 as catalyst
The insight gained from the DFT calculations was that the
oxime hydrogen plays a crucial role in the decomposition of
the catalyst systems, since the first step of the proposed and
studied decomposition mechanism involved the extraction of
the oxime hydrogen from the catalyst. In 5 and 6 this hydrogen
is replaced by the methyl group, the bulkier methyl group is
not as easy to extract from the oxygen as the hydrogen. The
reactivity of 5 toward ethylene oligomerization was thus inves-
tigated. After workup, the productivity value was calculated to
be 612 kg/mol catalyst/h/bar (entry 2, Table 2) which was nearly
three times than that of 3 and five times than 2 under identical
conditions. Reactions for longer time were also carried out using
5 as catalyst (entries 4 and 5, Table 2). The productivity value
remained unchanged even for longer duration thus showing the
stability of the catalyst system. These results thus effectively cor-
roborate and validate the conclusions from the DFT calculations.
The complex 6 was studied next. 6 was suspended in toluene
and DEAC was added, but the suspension did not fully dissolve in
toluene. Nevertheless, a productivity of 200 kg/mol catalyst/h/bar
(entry 6, Table 2) was recorded which was lower
than 5.
3.4.1. X-ray structure of 6
The molecular structure of 6 along with selected bond distances
and angles is shown in Fig. 7. The structure consists of two lig-
ands and two iodo atoms attached to the metal center. As evident
from the crystal structure, there is a high degree of disorder present
in the molecule probably due to the cis- and trans-orientation of
the methyl group. The complex displays a near perfect octahe-
dral geometry with the two iodo atoms positioned trans to the
metal center due to the absence of hydrogen bonding which is
in contrast to 1 or 2 where the two chloride molecule are in cis-
position [7]. The two nitrogen atoms from each of the oxime ligand
coordinate to the nickel atom. For complex 6, the average Ni
N
(pyridine), Ni N (imine) and Ni I bond lengths of 2.073, 2.063
˚
and 2.786 A, respectively, are similar to those in other octahedral
Ni (II) complexes [8c]. The bond angles are also interesting with
the I Ni I and N (1-pyridine) Ni N (1-pyridine) bond angle being
Table 2
Selective ethylene oligomerization by catalysts 5 and 6 with DEAC as co-catalyst.
Entrya
Cat.
Cat (mmol)
Co-cat. (equiv.)
Time (min)
%Butene b
%Hexene
Yield (g)
Productivityc
1
2
3
4
5
6
5
5
5
5
5
6
0.02
0.02
0.02
0.01
0.01
0.02
50
100
150
50
50
100
30
30
30
60
90
30
78
81
67
79
82
90
12
19
33
21
18
10
82
122
127
95
142
40
445
612
638
475
473
200
a
General conditions: 450 ml Parr reactor; toluene (solvent) = 150 ml; pressure = 20 bar; co-catalyst = DEAC; and temperature 45 ◦C.
1-Butene selectivity > 99.5%.
b
c
In (kg of product) (mol of catalyst)−1 h−1 bar−1
.

246
K.N. Tayade et al. / Journal of Molecular Catalysis A: Chemical 366 (2013) 238–246
4. Conclusions
(b) F. Speiser, P. Braunstein, L. Saussine, Organometallics 23 (2004) 2633.
[4] (a) W. Keim, F.H. Kowaldt, R. Goddard, C. Krüger, Angew. Chem. Int. Ed. Engl.
17 (1978) 466;
2-Thiophene aldoxime Ni(II) complex 3 has been shown to be
a selective catalyst for the dimerization of ethylene to 1-butene
in the presence of a co-catalyst like DEAC or MAO. The activity of
this system was higher than the earlier reported Ni(II) oxime com-
plexes 1 and 2. The possible mechanism for decomposition of 2 and
3 and their observed experimental stability was correlated using
density functional theory (DFT). The insight gained from mecha-
nistic studies regarding the role of the oxime hydrogen allowed
us to improve the stability and activity of the catalyst by substitu-
tion of the hydrogen atom with a methyl group as in 5. Catalyst 5
was found to be highly reactive and stable for selective ethylene
oligomerization under similar reaction conditions than either 1, 2
or 3, effectively proving that the removal of hydrogen in presence of
DEAC triggers the decomposition of these oxime based complexes.
The results discussed, therefore, represent an example of how a
symbiotic interplay between experiment and calculations can lead
to design of better catalysts.
(b) W. Keim, A. Behr, B. Limbäcker, C. Krüger, Angew. Chem. Int. Ed. Engl. 22
(1983) 503;
(c) M. Peuckert, W. Keim, J. Mol. Catal. 22 (1984) 289.
[5] (a) J.T. Dixon, M.J. Green, F.M. Hess, D.H.J. Morgan, J. Organomet. Chem. 689
(2004) 3641;
(b) J.W. Freeman, J.L. Buster, R.D. Knudsen, U.S. Patent 5,856,257 (1999);
(c) A. Jabri, C.B. Mason, Y. Sim, S. Gambarotta, T.J. Burchell, R. Duchateau, Angew.
Chem. Int. Ed. 47 (2008) 9717;
(d) A. Bollmann, K. Blann, J.T. Dixon, F.M.M. Hess, E. Killian, H. Maumela, D.S.
McGuinnes, D.H. Morgan, A. Neveling, S. Otto, M. Overett, A.M.Z. Slawin, P.
Wasserscheid, S. Kuhlmann, J. Am. Chem. Soc. 126 (2004) 14712.
[6] (a) P. Braunstein, Chem. Rev. 106 (2006) 134;
(b) S. Mecking, Coord. Chem. Rev. 203 (2000) 325;
(c) K.P. Tellmann, V.C. Gibson, A.J.P. White, D.J. Williams, Organometallics 24
(2005) 280;
(d) G.J.P. Britovsek, V.C. Gibson, O.D. Hoarau, S.K. Spitzmesser, A.J.P. White, D.J.
Williams, Inorg. Chem. 42 (2003) 3454;
(e) M.E. Bluhm, C. Folli, O. Walter, M. Döring, J. Mol. Cat. A: Chem. 229 (2005)
177;
(f) W.H. Sun, K. Wang, K. Wedeking, D. Zhang, S. Zhang, J. Cai, Y. Li,
Organometallics 26 (2007) 4781;
(g) R. Gao, M. Zhang, T. Liang, F. Wang, W.H. Sun, Organometallics 27 (2008)
5641;
(h) M. Brasse, J. Campora, P. Palma, E. lvarez, V. Cruz, J. Ramos, M.L. Reyes,
Organometallics 27 (2008) 4711;
Acknowledgments
(i) E.F. Younkin, T.R. Connor, J.I. Henderson, S.K. Friedrich, R.H. Grubbs, D.A.
Bansleben, Science 287 (2000) 460;
We are thankful to Dr. R.V. Jasra, RTG, Vadodara for providing
support. We are grateful to Reliance Industries Limited for allow-
ing us to publish this work. KV and MM wish to thank the Center
of Excellence in Scientific Computing, in NCL Pune, for providing
computational facilities.
(j) C. Bianchini, D. Gatteschi, G. Giambastiani, R.I. Guerrero, A. Ienco, F. Laschi,
C. Mealli, A. Meli, L. Sorace, A. Toti, F. Vizza, Organometallics 26 (2007) 726.
[7] S. Mukherjee, B.A. Patel, S. Bhaduri, Organometallics 28 (2009) 3074.
[8] (a) S. Khanra, T. Weyhermüller, P. Chaudhuri, Dalton Trans. (2007) 4675;
(b) T. Weyhermüller, R. Wagner, S. Khanra, P. Chaudhuri, Dalton Trans. (2005)
2539;
(c) P. Chaudhuri, T. Weyhermüller, R. Wagner, S. Khanra, B. Biswas, E. Bothe, E.
Bill, Inorg. Chem. 46 (2007) 9005, and references therein.
[9] M.P. Coakley, M.E. Casey, J. Inorg, Nucl. Chem. 34 (1972) 1937.
[10] G. Lees, F. Holmes, A.E. Underhill, D.B. Powell, J. Chem. Soc. A (1971) 337.
[11] (a) R. Ahlrichs, M. Bär, H.P. Baron, R. Bauernschmitt, S. Böcker, M. Ehrig, K.
Eichkorn, S. Elliott, F. Furche, F. Haase, M. Häser, H. Horn, C. Huber, U. Huniar,
M. Kattannek, C. Kölmel, M. Kollwitz, K. May, C. Ochsenfeld, H. Öhm, A. Schäfer,
U. Schneider, O. Treutler, M. von Arnim, F. Weigend, P. Weis, H. Weiss, TURBO-
MOLE (Version 5. 3);, Universität Karlsruhe, Karlruhe, Germany, 2000;
(b) A. Schäfer, C. Huber, R. Ahlrichs, J. Chem. Phys. 100 (1994) 5829;
(c) D. Andrae, U. Haeussermann, M. Dolg, H. Stoll, H. Preuss, Theor. Chim. Acta
77 (1990) 123.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/
j.molcata.2012.09.029.
References
[1] (a) B. Cornils, W.A. Herrmann (Eds.), Applied Homogeneous Catalysis
with Organometallic Compounds, Wiley-VCH, Weinheim, Germany, 1996,
p. 245;
[12] (a) A.D. Becke, Phys. Rev. A 38 (1988) 3098;
(b) J.P. Perdew, Phys. Rev. B 33 (1986) 8822.
[13] A. Schäfer, H. Horn, R. Ahlrichs, J. Chem. Phys. 97 (1992) 2571.
[14] K. Eichkorn, O. Treutler, H. Öhm, M. Häser, R. Ahlrichs, Chem. Phys. Lett. 240
(1995) 283.
(b) G.W. Parshall, S.D. Ittel, Homogeneous Catalysis: The Applications and
Chemistry of Catalysis by Soluble Transition Metal Complexes, Wiley, New
York, 1992, p. 68;
[15] M. Sierka, A. Hogekamp, R. Ahlrichs, J. Chem. Phys. 118 (2003) 9136.
[16] M. Mammen, E.I. Shakhnovich, J.M. Deutch, G.M. Whitesides, J. Org. Chem. 63
(1998) 3821.
[17] A. Klamt, G. Schuurmann, J. Chem. Soc., Perkin Trans. 2 (1993) 7.
[18] F. Wiegend, R. Ahrlichs, Phys. Chem. Chem. Phys. 7 (2005) 3297.
[19] Q.Z. Yang, A. Kermagoret, M. Agostinho, O. Siri, P. Braunstein, Organometallics
25 (2006) 5518.
(c) C. Bianchini, G. Giambastiani, I.G. Rios, G. Mantovani, A. Meli, A.M. Segarra,
Coord. Chem. Rev 250 (2006) 1391.
[2] (a) V.C. Gibson, S.K. Spitzmesser, Chem. Rev. 103 (2003) 283;
(b) S.D. Ittel, L.K. Johnson, M. Brookhart, Chem. Rev. 100 (2000) 1169;
(c) S. Mecking, Angew. Chem. Int. Ed 40 (2001) 534.
[3] (a) P. Barbaro, C. Bianchini, G. Giambastiani, I.G. Rios, A. Meli, W. Oberhauser,
A.M. Segarra, Organometallics 26 (2007) 4639;