Mineral Oil/Methanol: A Cheap Biphasic Reaction Medium with Thermomorphic Properties and Its Application to the Palladium-Catalyzed Carbonylation of Nitrobenzene to Methyl Phenylcarbamate

Article abstract of DOI:10.1002/cctc.201500452

We developed the methanol/mineral oil system as a new thermomorphic solvent mixture that allows homogeneous catalyst recovery without the use of water or fluorinated solvents. Mineral oil (nujol) is cheap and shows a very limited miscibility with methanol at room temperature but becomes completely miscible at high temperatures. Here we report the synthesis of some phenanthrolines substituted with long-chain alkyl groups, which render them soluble in hydrocarbons and insoluble in methanol, and their application to the palladium-catalyzed carbonylation of nitrobenzene to methyl phenylcarbamate. Both the reagents and products of the catalytic reaction are much more soluble in methanol than in hydrocarbons, so an easy separation of the catalyst is possible.

Full text of DOI:10.1002/cctc.201500452

DOI: 10.1002/cctc.201500452
Full Papers
Mineral Oil/Methanol: A Cheap Biphasic Reaction Medium
with Thermomorphic Properties and Its Application to the
Palladium-Catalyzed Carbonylation of Nitrobenzene to
Methyl Phenylcarbamate
Francesco Ferretti, Emma Gallo, and Fabio Ragaini*[a]
We developed the methanol/mineral oil system as a new ther-
momorphic solvent mixture that allows homogeneous catalyst
recovery without the use of water or fluorinated solvents. Min-
eral oil (nujol) is cheap and shows a very limited miscibility
with methanol at room temperature but becomes completely
miscible at high temperatures. Here we report the synthesis of
some phenanthrolines substituted with long-chain alkyl
groups, which render them soluble in hydrocarbons and in-
soluble in methanol, and their application to the palladium-
catalyzed carbonylation of nitrobenzene to methyl phenylcar-
bamate. Both the reagents and products of the catalytic reac-
tion are much more soluble in methanol than in hydrocarbons,
so an easy separation of the catalyst is possible.
Introduction
Difficulties in catalyst recovery and recycling are a major obsta-
cle to the commercialization of many homogeneously cata-
lyzed processes. Among the different possible strategies, bi-
phasic conditions are promising and have already been ap-
plied in industry to large-scale processes. This field is dominat-
ed by aqueous mixtures, but water is not compatible with
many reactions. In recent years, many “thermomorphic” sys-
tems have been developed in which the catalyst may change
its solubility in a given solvent or in which two solvents may
change their miscibility as an effect of a change in tempera-
ture.[1]
gene. As a result of the high toxicity of phosgene, the com-
mercialization of a new process that does not involve its use is
very important. The transition-metal-catalyzed reductive car-
bonylation of nitroarenes is one of the most promising and
studied alternatives.[3] The direct synthesis of aromatic isocya-
nates from nitroarenes is possible but it is usually a difficult
and poorly selective reaction, and research has focused on the
synthesis of carbamates, which can be produced more effi-
ciently and thermally cracked to the isocyanate in a separate
step. A technology investigated currently for the production of
MDI is shown in Scheme 1.
We became interested in the field of thermomorphic sys-
tems as a possible solution to catalyst recovery and recycling
in a synthesis of isocyanates that is greener than that em-
ployed presently.
The most active catalytic systems are based on the use of
Pd phenanthroline complexes[4] as catalysts and carboxylic[5] or
phosphorus acids[6] as cocatalysts. Recently, we further im-
proved the catalytic performance of the system by using elec-
tronically nonsymmetric phenanthrolines as ligands.[7] The re-
covery and recycling of the catalyst is especially difficult in the
case of methyl N-phenylcarbamate, a possible intermediate in
the synthesis of MDI, because all the reagents, products, and
even byproducts of the reaction are very soluble in methanol,
which needs to be used as both a reagent and solvent. An ef-
fective solution to this problem has not yet been found.
Isocyanates, both aromatic and aliphatic, are chemical inter-
mediates ubiquitous in many industrial applications and their
annual production, which accounts for several million metric
tons, is increasing gradually. Toluenediisocyanate (TDI) and
4,4’-methylenediphenyldiisocyanate (MDI) are by far the most
important.[2] The industrial synthesis of aromatic isocyanates
employed currently involves the reaction of anilines, obtained
by the reduction of the corresponding nitroarenes, with phos-
[a] Dr. F. Ferretti, Prof. Dr. E. Gallo, Prof. Dr. F. Ragaini
Department of Chemistry
Università degli Studi di Milano
Via C. Golgi 19, I-20133 Milan (Italy)
E-mail: fabio.ragaini@unimi.it
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201500452.
This publication is part of a Special Issue on Palladium Catalysis. To view
the complete issue, visit:
http://onlinelibrary.wiley.com/doi/10.1002/cctc.v7.14/issuetoc
Scheme 1. Synthetic strategy for the production of MDI.
ChemCatChem 2015, 7, 2241 – 2247
2241
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
As the addition of water at any stage is not compatible with
the investigated reaction, we decided to pursue two related
strategies: 1) the use of modified phenanthrolines that are in-
soluble in methanol at room temperature but soluble at the
reaction temperature (typically 150–1708C) and 2) the use of
modified phenanthrolines that are preferentially soluble in
a low-polarity phase immiscible with methanol at room tem-
perature but miscible at higher temperatures. The introduction
of long aliphatic chains to the ligands was considered to be
the best approach to reach both goals, and herein, we report
the synthesis of so-substituted phenanthrolines and their use
as thermomorphic ligands for the Pd-catalyzed carbonylation
of nitrobenzene to methyl N-phenylcarbamate. To the best of
our knowledge, this is the first time that phenanthrolines have
been used as ligands in a thermomorphic system.
tadecane, 9-(iodomethyl)nonadecane (2), and 19-iodoheptatria-
contane (3), respectively. The two linear iodoalkanes are com-
mercially available. Compound
2 was prepared by the
iodination of 2-octyldodecan-1-ol using a procedure reported
in the literature for different primary alcohols (Scheme 3).[11]
Compound 3 was instead synthesized by forming the Grignard
reagent of 1-bromoctadecane and reacting it with methyl for-
mate. The so-obtained alcohol was then iodinated using a stan-
dard halogenation procedure (Scheme 4).
Scheme 3. Iodination of 2-octyldodecan-1-ol to 2.
Results and Discussion
Synthesis of the ligands
To design a good ligand for a catalyst recovery strategy, the re-
searcher must be sure of the nature of the resting state. In the
case of the carbonylation of nitroarenes, the catalyst resting
state is a phenanthroline palladium bisalkoxycarbonyl complex
[Pd(Phen)(COOR)2], which forms the carbonylated active spe-
cies only under a CO atmosphere.[8] This complex is neutral
and thus recyclable in a nonpolar phase if tagged with the ap-
propriate phase selector.
Scheme 4. Synthesis of 3.
Catalyst recycling
If we take into account the results obtained previously with
nonsymmetric phenanthrolines,[7] the ligand of choice for a re-
covery strategy should be a 4-alkoxy-substituted phenanthro-
line. However, the exchange of the alkoxy moiety in aryl ethers
is easy under the catalytic reaction conditions (1708C in meth-
anol in the presence of a phosphorus acid), and only one long
alkyl chain may not be sufficient to render the catalyst insolu-
ble in methanol or selectively soluble in a nonpolar phase.
Thus we synthesized four different phenanthrolines substituted
symmetrically with long alkyl chains in the 4- and 7-positions.
Among the synthesized ligands, only the two substituted with
linear alkyl chains were reported previously in the literature.[9]
The syntheses were performed following a procedure report-
ed previously.[10] The commercially available 4,7-dimethylphe-
nanthroline was lithiated and then reacted with an iodoalkane
(Scheme 2). 4,7-Ditridecylphenathroline (C13-Phen), 4,7-dinona-
decylphenanthroline (C19-Phen), 4,7-bis(3-octyltridecyl)phenan-
throline (C21-Phen), and 4,7-bis(2-octadecylicosyl)phenanthro-
line (C38-Phen) were obtained from 1-iodododecane, 1-iodooc-
One of the main methodological errors present in many
papers that deal with catalyst recovery is to reach total conver-
sions of reagents or report just yields and take them as a good
index of the system recyclability. In this way much information
is lost and a nonrecyclable system could be misleadingly con-
sidered a recyclable one. Thus all the reactions were conduct-
ed to avoid conversions higher than 80%, and turnover fre-
quencies (TOFs) were taken as an index of the system recycla-
bility as suggested by Gladysz in his catalyst recovery guide-
lines.[12] In the present case, the TOF is indeed related directly
to the reaction rate as the kinetics of the reaction is zero-order
in nitrobenzene and no induction time is observed if a small
amount of aniline is added initially, as demonstrated in a previ-
ous study.[8]
In our studies on the carbonylation of nitroarenes, we usual-
ly employ [Pd(Phen)2][BF4]2 as a catalyst precursor. However,
the presence of two equivalents of phenanthroline, which
competes with the thermomorphic ligands for metal coordina-
tion, could result in catalyst loss after the first recycle. Several
attempts were made to synthesize an analogous catalyst pre-
cursor with long-chain ligands, which either started from palla-
dium acetate[13] or used ligand exchange with [Pd(Bipy)2][BF4]2
(Bipy=bipyridine). However, the obtained complexes were in
a gummy form and were inseparable from the excess ligand.
Thus the active complexes were generated in situ by ligand ex-
change with [Pd(Bipy)2][BF4]2. Bipyridine is a much weaker
ligand with respect to phenanthroline; thus it does not bind
Pd competitively. H3PO4 (85 wt%) was used as the acidic pro-
Scheme 2. General synthesis of alkyl-substituted phenanthrolines.
ChemCatChem 2015, 7, 2241 – 2247
www.chemcatchem.org
2242
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
moter. In accord with previous studies, dimethoxypropane was
added as an internal drying agent, and a small amount of ani-
line was added initially to avoid any induction time.[6a,b,8] Con-
centrations of the ligands were unoptimized and the optimal
ligand ratio identified previously for 4,7-dimethylphenanthro-
line was employed.[7a]
with methanol, and higher linear alkanes are too expensive.
Decalin was chosen at first because of its low miscibility with
MeOH and relatively low cost. Among the other potential sol-
vents, we identified mineral oil as a possible choice because
of: 1) its almost complete immiscibility with methanol at room
temperature, 2) its partial or complete miscibility at the reac-
tion temperature, which depends on the ratio with methanol,
and 3) its low cost. However, mineral oil is not always the same
mixture of hydrocarbons, thus its density and miscibility varies
with the composition, and the miscibility of the available mate-
rial should be checked before use.[14] To the best of our knowl-
edge, this is the first report of the use of mineral oil as a cosol-
vent for thermomorphic biphasic systems. Recently, Bergbreiter
et al. reported the replacement of standard alkane solvents in
a biphasic system by a low-molecular-weight polyethylene
wax.[15] However, this system differs from ours in that the poly-
ethylene wax was solid at room temperature and only melted
if heated. Moreover, the addition of some water was necessary
at the end of the reaction to force the catalyst to stay in the
resolidified wax.
Before starting catalytic experiments, the solubility of the
synthesized ligand was tested in methanol. To this aim, the li-
gands were suspended in methanol in a pressure tube and the
suspension was heated until complete solubilization was ob-
served. Then the solution was cooled to room temperature,
the precipitated ligand was separated by centrifugation, and
the solution was evaporated to check the possible presence of
dissolved ligand. The results evidenced that C13-Phen was too
soluble even at ambient temperature (1.5 mgmLÀ1), thus it
was not used in catalytic tests. C19-Phen and C38-Phen were
completely soluble if the temperature was raised, and no ap-
preciable amount of ligand remained in solution after cooling
to room temperature. C21-Phen is an oil at room temperature;
thus its solubility in methanol was not tested and it was used
only in liquid–liquid separation strategies (vide infra).
The catalytic results (Table 2) showed that ligand C21-Phen is
not suitable for catalyst recycling. A decrease of more than
56% in activity in the second run was observed if decalin was
employed as the cosolvent, and the decrease was more evi-
dent (87%) if mineral oil was used. This means that the com-
plexes formed by these two ligands with Pd have a higher af-
finity with the polar phase than the apolar one.
At first, a liquid–solid separation strategy was studied. After
the separation of the methanol phase, by centrifugation or de-
cantation depending on the system, all the reagents with the
exclusion of the ligand and the catalyst were refilled into the
reactor for the subsequent run. As a result of the air-sensitive
nature of the resting state of the Pd catalyst, all the operations
performed on the recycled phase were conducted under an
inert atmosphere.
If the decalin/methanol (1.5:1 in volume) solvent mixture
was tested with C19-Phen as the ligand, it became evident that
the nonpolar decalin phase diminished visibly at every recycle,
which led to the low recyclability of the catalyst. Thus decalin/
MeOH cannot be considered as a suitable solvent mixture, and
C38-Phen was not tested in it. Notably, although pure methanol
and decalin are almost immiscible, the reaction mixture also
contains non-negligible amounts of aromatic substances that,
although they are markedly more soluble in methanol than in
decalin, have an intermediate polarity and increase the misci-
bility of the two phases.
Although the ligands are virtually insoluble in pure metha-
nol, the solubility increases in the reaction mixture because of
the presence of the other reactants. Good recyclability was ob-
tained over the first three cycles for the linear C19-Phen with
a decrease in the fourth cycle, whereas a more gradual de-
crease was noted for C38-Phen (Table 1).
The thermomorphic liquid–liquid separation strategy is more
interesting from an industrial point of view as it involves the
separation of two liquid phases and avoids the filtration or mi-
crofiltration of small amounts of catalyst and free ligand that
could cause technical problems on a large scale. The choice of
the nonpolar cosolvent for this separation was challenging.
Indeed hexane, heptane, and cyclohexane are too miscible
If we employed a biphasic system of mineral oil/methanol
(2:5 in volume), both the C19-Phen and C38-Phen systems
showed a good recyclability. However, although the C19-Phen
system showed no loss of activity over five cycles, with C38-
Phen the activity decreased gradually.
A surprising feature of the experiments with C19-Phen in
MeOH/mineral oil is that the catalytic activity increases upon
recycling. This effect cannot be caused by an induction period
for the generation of the active catalyst from the Pd precata-
lyst because we have shown previously that in the presence of
a small amount of aniline such an induction period should be
extremely short.[8] If we consider that the kinetics is first order
in aniline, a more likely explanation is that a small amount of
aniline accumulates gradually in the mineral oil phase and con-
tributes to accelerate the following reactions.
Table 1. Catalyst recycling in nitrobenzene carbonylation under thermo-
morphic conditions: liquid–solid separation.[a]
Cycle no.
C19-Phen
C38-Phen
[b]
TOF [hÀ1
]
1
2
3
4
780
750
690
500
820
620
460
390
[a] Experimental conditions: [Pd(Bipy)2][BF4]2 =1.3”10À3 mmol, molar
ratios PhNO2/PhNH2/H3PO4/Ligand/[Pd(Bipy)2][BF4]2 =3800:200:750:120:1
for C19-Phen and 4000:200:750:120:1 for C38-Phen, in MeOH (10 mL)+2,2-
dimethoxypropane (0.5 mL), PCO =60 bar, at 1708C for 3 h. [b] TOF=mol
PhNO2 reacted/(mol Pd”t).
To quantify any Pd loss during recycling, the best series of
reactions (C19-Phen/mineral oil) was repeated and the Pd con-
tent in the methanol phase was determined by atomic absorp-
tion spectroscopy (AAS). A Pd loss between 1.1 and 1.3% was
ChemCatChem 2015, 7, 2241 – 2247
www.chemcatchem.org
2243
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Table 2. Catalyst recycling in nitrobenzene carbonylation under thermomorphic conditions: liquid–liquid separation.[a]
Cycle no.
C21-Phen/decalin[b,e]
C21-Phen/mineral oil[b,e]
C19-Phen/decalin[b]
C19-Phen/mineral oil
]
C19-Phen/mineral oil[c]
C38-Phen/mineral oil[d]
[f]
TOF [hÀ1
1
2
3
4
5
860
380
–
–
–
970
140
–
–
–
980
790
550
320
240
310
340
400
400
430
220
380
320
50
630
460
340
280
160
–
[a] Experimental conditions: molar ratios PhNO2/PhNH2/H3PO4/Ligand/[Pd(Bipy)2][BF4]2 =2000:100:375:60:1, [Pd(Bipy)2][BF4]2 =2.7”10À3 mmol, in MeOH
(10 mL)+2,2-dimethoxypropane (0.5 mL), PCO =60 bar, at 1708C for 3 h. Solvent mixture volume ratio: decalin/MeOH=1.5:1; mineral oil/MeOH=2:5.
[b] The amount of catalyst was halved and the concentrations of the other reagents were maintained. [c] Octadecylphosphonic acid was added instead of
H3PO4. The addition was made only in the first cycle. [d] The reaction was stopped after 2.5 h to avoid complete conversion. [e] A higher amount of nitro-
benzene was employed. PhNO2/Pd molar ratio=7000. [f] TOF= mol PhNO2 reacted/(mol Pd”t).
measured over five reactions (1.1, 1.2, 1.3, 1.2, and 1.1% with
respect to the initial amount of Pd). The TOF values were 320,
350, 390, 400, and 410, respectively, which are in line with
those measured for the original series (Table 2). These results
confirm a good recovery of Pd, although there is clearly room
for improvement.
cylphosphonic acid as the acidic promoter and C19-Phen as the
ligand. A high amount of azoxybenzene was obtained in this
case (25%). This is not surprising as most of the acid was lost
during the first three cycles and thus its concentration was
very low. We have shown previously that the acid not only af-
fects the rate of the reaction but also impacts the selectivity
by decreasing the amount of azoxybenzene formed significant-
ly. This result confirms the need for an acidic promoter to
ensure the good selectivity and activity of the system.
The results obtained with C19-Phen in MeOH/mineral oil indi-
cate that a good recovery and recycling of Pd and the ligand
can be achieved, but the acidic promoter (H3PO4) is clearly lost
at every recycle. In an attempt to take a step further and recy-
cle even the acidic promoter, we ran a series of experiments in
which octadecylphosphonic acid was employed as the acidic
promoter in place of phosphoric acid and the addition of new
acid was eliminated after each recycle (Table 2). However, this
strategy was not fully successful. Although the catalytic activity
increases between the first two runs, which indicates that a sig-
nificant amount of octadecylphosphonic acid remains in the
mineral oil phase, the TOF of the reaction then decreases
quickly. This is clearly because of acid loss in the methanol
phase. Given the structure of the acid, this may be caused by
micelle formation in methanol, but we have not investigated
this aspect in detail.
Conclusions
We implemented a thermomorphic strategy for the recovery of
the Pd catalyst in the reductive carbonylation of nitroarenes.
The results indicate that catalyst recycling is possible by recov-
ering the catalyst either as a solid or in solution, but the latter
approach gave better results. The increased length of the alkyl
chains on the phenanthrolines has, as expected, a positive
effect. However, the introduction of a branch on the chain af-
fects recyclability negatively even if the total number of
carbon atoms is increased, which was unexpected. We intro-
duced branched chains to increase the number of CH2 groups
on the phenanthroline without requiring the use of linear io-
dides or bromides longer than C20, which are not commercially
available. However, this is clearly not the right strategy.
So far we have focused on TOF as a good index of the recy-
clability of the system, but the selectivity of the reaction
should also be taken into account. The selectivity to methyl N-
phenylcarbamate is not a good probe for the total selectivity
of the reaction as N,N’-diphenylurea, which cannot be detected
by GC, is formed as an intermediate. It is only later alcoholyzed
to yield methyl N-phenylcarbamate and aniline, which re-enter
the catalytic cycle. The alcoholysis reaction is not complete at
low conversions and short reaction times, which thus leads to
an underestimation of the selectivity of the system. Good in-
dexes are instead the azo- and azoxybenzene selectivities as
these are the only real byproducts formed in significant
amounts. For azoxybenzene, the major byproduct, the selectiv-
ity values were almost unvaried and were always lower than
6%. These values and those obtained for the other byproduct
are in accord with the corresponding values obtained in a ho-
mogeneous environment[7a] (Tables S1 and S2), which indicate
that the mixed solvent conditions do not affect the selectivity
of the reaction negatively. The only exception to this is repre-
sented by the last run of the reaction conducted using octade-
The main challenge to find a solvent that could be used in
the thermomorphic conditions coupled with methanol was
overcome by using a nonconventional solvent present in most
chemical research laboratories: mineral oil. A limit to its use
with methanol is the high temperature at which the two sol-
vents become fully miscible. However, many reactions are run
on an industrial scale at high temperature to maximize their
rate. Moreover, higher alcohols and a variety of polar solvents
exist that become miscible with mineral oil at lower tempera-
tures and are immiscible at room temperature even without
the addition of water to the system. We have not investigated
these alternatives in the present work because methanol was
a prerequisite for our reaction, but many possibilities exist that
are worth investigation. In this context, it is important to evi-
dence that purified mineral oil is nontoxic and thus thermo-
morphic catalyst recovery based on its use could find applica-
tion in pharmaceutical synthesis.
ChemCatChem 2015, 7, 2241 – 2247
www.chemcatchem.org
2244
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
was also added (see Table 2 for amounts). At the end of the reac-
tion, the autoclave was cooled with an ice bath, vented, and the
glass tube transferred rapidly to a Schlenk tube with a wide mouth
under dinitrogen. Then, if mineral oil was employed, it was neces-
sary to stir the reaction mixture for some minutes to allow the lib-
eration of dissolved CO. The separation of the two phases occurred
rapidly in most of the cycles. However, sometimes it was necessary
to wait up to 1 h to achieve a complete separation. The methanol
upper phase was separated and the nonpolar phase was washed
twice with methanol with stirring for 5 min each time under a dini-
trogen atmosphere. The mother liquor and the washings were
combined, and the products were analyzed by GC (Dani 8620 gas
chromatograph, equipped with a Supelco SLBꢁ-5 ms column;
naphthalene as an internal standard). 2,2-Dimethoxypropane,
PhNH2, PhNO2, H3PO4, and methanol were then added to the tube
that still contained the nonpolar phase, and the procedure repeat-
ed for the subsequent cycles.
Experimental Section
General procedures
Unless otherwise stated all manipulations and reactions were con-
ducted under a dinitrogen atmosphere. All solvents were dried by
standard procedures and distilled under dinitrogen immediately
before use. All glassware and magnetic stirring bars used in catalyt-
ic reactions were kept in an oven at 1208C for at least 2 h and al-
lowed to cool under vacuum before use. Phenanthroline ligands
were dried before use by dissolving them in CH2Cl2, drying the re-
sulting solution with Na2SO4, and evaporating the filtered solution
in vacuo. They were then stored under dinitrogen. They can be
weighed in the air without problem, but must be stored under an
inert atmosphere if water uptake is to be avoided. Nitrobenzene
and aniline were distilled and stored under dinitrogen before use.
[Pd(Bipy)2][BF4]2 was prepared according to a reported proce-
dure.[13] The mineral oil employed in the catalytic reactions was
Sigma–Aldrich (M3516) Mineral oil for infrared spectroscopy (light
oil), d=0.84 gmLÀ1. Immediately before use it was degassed under
vacuum for 10 min with stirring, but no other purification was
made. All other chemicals were purchased from Sigma–Aldrich,
Acros Organics, or Alfa Aesar. NMR spectra were recorded by using
a Bruker Avance 300-DRX or Avance 400-DRX spectrometer. Ele-
mental analyses were recorded by using a PerkinElmer 2400 CHN
elemental analyzer.
Phenanthrolines synthesis
General procedure
The synthesis was adapted from a procedure reported previous-
ly.[10] In an oven-dried Schlenk flask, 4,7-dimethyl-1,10-phenanthro-
line (1.0 g, 4.8 mmol) was dissolved in dry THF (100 mL). Lithium
diisopropylamide solution in THF/heptane/ethylbenzene (2m,
6.0 mL) was added slowly at À788C with stirring. After 1 h, the
alkyl iodide (12 mmol) dissolved in dry THF (30 mL) was added,
and the mixture was stirred for 5 h at À788C and then overnight
at RT. After the evaporation of the solvent under vacuum, CH2Cl2
(100 mL) was added, and the mixture was washed with H2O (3”
25 mL) and dried over Na2SO4. The solvent was evaporated, and
the crude product dissolved in the minimum amount of CH2Cl2
and filtered through an alumina pad (washed with CH2Cl2 until
excess iodoalkane was eluted completely and then CH2Cl2/1%
MeOH to elute the phenanthroline). Evaporation of the solvent af-
forded the product as a white solid.
Catalytic reactions
Liquid–solid separation
In a typical catalytic reaction, alkyl-substituted ligand and PhNO2
were weighed into a 30 mL round-bottomed centrifuge tube. The
tube was placed in a Schlenk tube with a wide mouth under dini-
trogen and was frozen with liquid N2. It was evacuated and filled
with dinitrogen three times. The appropriate amount of solvent,
PhNH2, and 85 wt% H3PO4 were added by volume (see Table 1 for
amounts). Finally, a solution of the catalyst in methanol was added
by volume to avoid large errors in weighing small amounts of cata-
lyst. Then the liner was closed with a screw cap that had a glass-
wool-filled open mouth to allow gaseous reagents to exchange
and transferred rapidly to a 200 mL stainless-steel autoclave with
magnetic stirring. The autoclave was then evacuated and filled
with dinitrogen three times. CO was then charged at RT at 60 bar,
and the autoclave was immersed in an oil bath preheated at
1708C. Other experimental conditions are reported in the captions
to the tables. At the end of the reaction, the autoclave was cooled
with an ice bath, vented, and the glass tube was transferred rapid-
ly to a Schlenk tube with a wide mouth under dinitrogen. It was
then closed with a screw cap and centrifuged for 10 min at
2750 rpm. The liquid phase was separated, and the solid was
washed twice with methanol under a dinitrogen atmosphere. The
mother liquor and the washings were combined, and the products
were analyzed by GC (Dani 8610 gas chromatograph, equipped
with a Supelco SLB-5 ms column; naphthalene as an internal stan-
dard). 2,2-Dimethoxypropane, PhNH2, PhNO2, H3PO4, and methanol
were then added to the tube that still contained the solid catalyst
and excess ligand, and the procedure was repeated for the subse-
quent cycles.
C13-Phen: White powder. 57% yield. The 1H NMR spectrum con-
forms to that reported previously.[9b] 1H NMR (300 MHz, CDCl3,
3
300 K): d=9.13 (d, 3JH,H =4.1 Hz, 2H), 8.11 (s, 2H), 7.54 (d, JH,H
=
3
3
4.5 Hz, 2H), 3.18 (t, JH,H =7.5 Hz, 4H), 1.90 À1.70 (m, JH,H =7.5 Hz,
4H), 1.54–1.10 (m, 36H), 0.87 ppm (t, 3JH,H =6.6 Hz, 6H); 13C NMR
(100 MHz, CDCl3) d=149.7 (CH), 149.1 (C), 146.5 (C), 127.2 (C), 123.1
(CH), 122,0 (CH), 32.6 (CH2), 32.0 (CH2), 30.5 (CH2), 29.7 (CH2), 29.6
(CH2), 29.5 (CH2), 29.4 (CH2), 22.8 (CH2), 14.2 ppm (CH3); elemental
analysis calcd (%) for C38H60N2 (544.9): C 83.76, H 11.10, N 5.14;
found: C 83.56, H 11.23, N 5.12.
C19-Phen: White powder. 61% yield. 1H NMR (300 MHz, CDCl3,
3
300 K): d=9.05 (d, 3JH,H =4.5 Hz, 2H), 8.04 (s, 2H), 7.43 (d, JH,H
=
3
4.5 Hz, 2H), 3.13 (t, JH,H =7.6 Hz, 4H), 1.90–1.68(m, 4H), 1.53–1.09
(m, 64H), 0.90 ppm (t, JH,H =6.3 Hz, 6H); 13C NMR (75 MHz, CDCl3,
3
300 K): d=150.31 (CH), 149.00 (C), 147.27 (C), 127.52 (C), 123.32
(CH), 122.23 (CH), 32.92 (CH2), 32.35 (CH2), 30.90 (CH2), 30.12 (CH2),
30.00 (CH2), 29.90 (CH2), 29.78 (CH2), 23.11 (CH2), 14.53 ppm (CH3);
elemental analysis calcd (%) for C50H84N2 (713.2): C 84.20, H 11.87,
N, 3.93; found: C 84.10, H 11.84, N 3.79.
9-(Iodomethyl)nonadecane: The synthesis was adapted from
a procedure reported previously in the literature for different pri-
mary alcohols.[11] 2-Octyldodecan-1-ol (5.0 mL, 14.0 mmol), HI (57%,
5.0 mL), and H3PO4 (85%, 5.0 mL) were added to a round-bot-
tomed flask. The resulting biphasic mixture was stirred vigorously
at 1308C. After 4 h, the reaction mixture was allowed to cool,
Liquid–liquid separation
The procedure for the preparation of the reaction was identical to
that used for liquid–solid separation except that a nonpolar solvent
ChemCatChem 2015, 7, 2241 – 2247
www.chemcatchem.org
2245
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
poured into ice-water (10 mL), and extracted with CH2Cl2 (3”
10 mL). The combined organic layers were washed with a saturated
solution of sodium thiosulfate (2”10 mL) and then again with
water (10 mL). The organic phase was dried over MgSO4, and the
solvent was evaporated to afford a light brown oil (5.22 g,
and H are because of the presence of hexane, which remains in
the product after several hours in vacuo.
C38-Phen: Pure product was obtained by column chromatography
(silica gel, hexane/3% triethylamine/1% isopropanol). White
1
powder (31% yield); H NMR (400 MHz, CDCl3, 300 K): d=9.07 (d,
1
12.8 mmol, 91% yield). The H NMR spectrum conforms to that re-
3
3JH,H =4.5 Hz, 2H), 8.05 (s, 2H), 7.42 (d, JH,H =4.5 Hz, 2H), 3.06 (d,
ported previously.[16] 1H NMR (300 MHz, CDCl3, 300 K): d=3.27 (d,
3JH,H =4.49 Hz, 4H), 1.35–1.00 (m, 64H), 0.88 ppm (t, 3JH,H =7.0 Hz,
6H); 13C NMR (75 MHz, CDCl3, 300 K): d=38.8 (CH), 34.5 (CH2), 32.0
(CH2), 29.9 (CH2), 29.8 (CH2), 29.5 (CH2), 26.6 (CH2), 22.8 (CH2), 16.9
(CH2), 14.2 ppm (CH3); EI-MS: m/z (%): 281 [MÀI]+ (4), 183, 169, 155,
141, 127, 113, 99, 85, 71, 57 (100), 43, 41.
3JH,H =6.8 Hz, 4H), 1.88 (m, 2H), 1.47–1.06 (m, 136H), 0.90 ppm (t,
3JH,H =7.1 Hz, 12H); 13C NMR (100 MHz, CDCl3, 300 K): d=149.5
(CH), 147.8 (C), 147.0 (C), 127.5 (C), 124.2 (CH), 121.9 (CH), 39.2 (CH),
37.5 (CH2), 33.5 (CH2), 31.9 (CH2), 30.0 (CH2), 29.69 (CH2), 29.65 (CH2),
29.3 (CH2), 26.5 (CH2), 22.7 (CH2), 14.1 ppm (CH3); elemental analysis
calcd (%) for C88H160N2 (1246.2): C 84.81, H 12.94, N 2.25; found: C
84.63, H 13.31, N 1.93.
C21-Phen: Light brown oil. 42% yield. 1H NMR (400 MHz, CDCl3,
3
300 K): d=9.03 (d, 3JH,H =4.5 Hz, 2H), 8.01 (s, 2H), 7.41 (d, JH,H
=
4.5 Hz, 2H), 3.11–3.02 (m, 4H), 1.74–1.64 (m, 4H), 1.54–0.95 (m,
66H) 0.92–0.75 ppm (m, 12H); 13C NMR (100 MHz, CDCl3): d=149.8
(CH), 149.2 (C) 146.7 (C), 127.0 (C), 122.8 (CH), 121.7 (CH), 37.7 (CH),
34.7 (CH2), 33.5 (CH2), 31.9 (CH2), 30.1 (CH2), 29.8 (CH2), 29.7 (CH2),
29.6 (CH2), 29.3 (CH2), 26.7 (CH2), 22.6 (CH2), 14.1 ppm (CH3); ele-
mental analysis calcd (%) for C54H92N2 (769.3): C 84.31, H 12.05, N
3.64; found: C 84.11, H 12.01, N 3.67.
AAS
The leaching of Pd was determined by AAS by using a PerkinElmer
PinAAcle 900T Atomic Absorption Spectrometer. After the separa-
tion of the polar phase from mineral oil, the volatiles were re-
moved under vacuum. The sample was dissolved in 65% nitric
acid (4 mL) by heating to reflux for 12 h. The solution was cooled
to RT, and 65% perchloric acid (1 mL) was added. This solution was
heated to reflux for 8 h. The solution was concentrated to ꢀ2 mL
by distillation of the acid. MilliQ water (2 mL) was added and re-
moved by distillation three times to afford ꢀ1 mL of solution. The
sample was then diluted with 4% HNO3 and analyzed in an air-
acetylene flame at the 244.79 nm line to determine the amount of
Pd.
Heptatriacontan-19-ol: In an oven-dried Schlenk flask, Mg turnings
(0.74 g, 30.4 mmol) were suspended in dry and deoxygenated THF
(15 mL). A small crystal of iodine was added (to obtain a light
yellow coloration), and the solution was heated until decoloration
occurred. 1-Bromooctadecane (6.65 g, 19.9 mmol) was dissolved
separately in THF (15 mL) and the obtained solution was added
dropwise with stirring at such a rate to keep the mixture warm.
The suspension was left for 1 h at RT and for a further hour at
408C. After cooling to RT, dry degassed methyl formate (410 mL,
6.67 mmol) was added, and the flask was cooled with water (the
reaction is very exothermic but ice-cooling of the flask could result
in Grignard reagent precipitation). The mixture was stirred over-
night and then quenched with H2O (50 mL), and 1m H2SO4 was
added to consume the excess Mg. CH2Cl2 was added, and most of
the solid passed into the organic phase to form a suspension. The
organic layer was washed with water (50 mL), and the solvent was
evaporated. The obtained solid was suspended in acetone
(100 mL) and heated to reflux for 1 h. The mixture was allowed to
cool and filtered to
Acknowledgements
We thank the MiUR for financial support and F.F. thanks the Uni-
versity of Milan for a fellowship.
Keywords: biphasic catalysis
chemistry · homogeneous catalysis · palladium
·
carbonylation
·
green
[1] D. E. Bergbreiter in Recoverable and Recyclable Catalysts (Ed.: M. Bena-
glia), Wiley, Chichester, 2009, pp. 117–153.
[2] H.-W. Engels, H.-G. Pirkl, R. Albers, R. W. Albach, J. Krause, A. Hoffmann,
H. Casselmann, J. Dormish, Angew. Chem. Int. Ed. 2013, 52, 9422–9441;
Angew. Chem. 2013, 125, 9596–9616.
[3] a) O. Kreye, H. Mutlu, M. A. R. Meier, Green Chem. 2013, 15, 1431–1455;
b) F. Ragaini, Dalton Trans. 2009, 6251–6266; c) F. Paul, Coord. Chem.
Rev. 2000, 203, 269–323; d) S. Cenini, F. Ragaini, Catalytic Reductive Car-
bonylation of Organic Nitro Compounds, Kluwer Academic Publishers,
Dordrecht, 1996.
[4] a) A. Bontempi, E. Alessio, G. Chanos, G. Mestroni, J. Mol. Catal. 1987,
42, 67–80; b) E. Alessio, G. Mestroni, J. Organomet. Chem. 1985, 291,
117–127; c) E. Alessio, G. Mestroni, J. Mol. Catal. 1984, 26, 337–340;
d) E. Drent, P. W. N. M. van Leeuwen, EP 0086281A1 (US 4474978), 1983;
e) T. J. Mooibroek, E. Bouwman, E. Drent, Organometallics 2012, 31,
4142–4156.
[5] a) S. Cenini, F. Ragaini, M. Pizzotti, F. Porta, G. Mestroni, E. Alessio, J. Mol.
Catal. 1991, 64, 179–190; b) P. Wehman, V. E. Kaasjager, F. Hartl, P. C. J.
Kamer, P. W. N. M. van Leeuwen, J. Fraanje, K. Goubitz, Organometallics
1995, 14, 3751–3761; c) P. Wehman, L. Borst, P. C. J. Kamer, P. W. N. M.
van Leeuwen, J. Mol. Catal. A 1996, 112, 23–36; d) P. Wehman, P. C. J.
Kamer, P. W. N. M. van Leeuwen, Chem. Commun. 1996, 217–218; e) M.
Gasperini, F. Ragaini, S. Cenini, E. Gallo, J. Mol. Catal. A 2003, 204–205,
107–114.
afford a white product that was only slightly soluble in all
common solvents (4.12 g, 7.67 mmol, 77% yield). This was used in
the subsequent step without further characterization except for el-
emental analysis. Elemental analysis calcd (%) for C37H76O (537.0): C
82.76, H 14.27; found: C 82.38, H 14.34.
19-Iodoheptatriacontane: The synthesis was adapted from a pro-
cedure reported previously for the synthesis of bromides.[17] In
a two-necked round-bottomed flask, imidazole (0.57 g, 8.47 mmol),
PPh3 (2.23 g, 8.48 mmol), and I2 (2.22 g, 8.7 mmol) were added to
19-heptatriacontanol (4.12 g, 7.67 mmol) suspended in dry THF
(150 mL). After 24 h, methanol (1.5 mL) was added, and the solvent
was removed under reduced pressure. Purification by filtration
through a pad of silica gel with hexane as eluent afforded a white
powder (4.61 g, 7.13 mmol, 93% yield). 1H NMR (400 MHz, CDCl3,
300 K): d=4.22–4.09 (m, 1H), 1.98–1.81 (m, 2H), 1.72 (m, 2H),
1.63–1.22 (m, 64H), 0.90 ppm (t, 3JH,H =6.8 Hz, 6H); 13C NMR
(75 MHz, CDCl3, 300 K): d=41.2 (CH), 40.8 (CH2), 32.1 (CH2), 29.8
(CH2), 29.7 (CH2), 29.5 (CH2), 29.0 (CH2), 22.9 (CH2), 14.3 ppm (CH3);
EI-MS: m/z (%): 519 [MÀI]+ (3), 278, 264, 250, 236, 222, 209, 195,
181, 167, 153, 139, 127, 125, 111, 99, 97, 85, 83, 71, 69, 57 (100), 43,
41; elemental analysis calcd (%) for C37H75I (646.9): C 68.70, H 11.69;
found: C 71.0, H 11.90. The slightly higher values obtained for C
[6] a) F. Ragaini, C. Cognolato, M. Gasperini, S. Cenini, Angew. Chem. Int. Ed.
2003, 42, 2886–2889; Angew. Chem. 2003, 115, 2992–2995; b) F. Ragai-
ChemCatChem 2015, 7, 2241 – 2247
www.chemcatchem.org
2246
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
ni, M. Gasperini, S. Cenini, Adv. Synth. Catal. 2004, 346, 63–71; c) M.
Gasperini, F. Ragaini, C. Cazzaniga, S. Cenini, Adv. Synth. Catal. 2005,
347, 105–120; d) M. Gasperini, F. Ragaini, C. Remondini, A. Caselli, S.
Cenini, J. Organomet. Chem. 2005, 690, 4517–4529; e) M. Gasperini, F.
Ragaini, S. Cenini, E. Gallo, S. Fantauzzi, Appl. Organomet. Chem. 2007,
21, 782–787.
[13] B. Milani, A. Anzilutti, L. Vicentini, A. Sessanta o Santi, E. Zangrando, S.
Geremia, G. Mestroni, Organometallics 1997, 16, 5064–5075.
[14] The miscibility of two different mineral oils with methanol (2:5 in
volume) was tested in
a pressure tube; Acros-Mineral oil, pure
(4150800): d=0.88 gmLÀ1, is only partially miscible at 1708C; Sigma–Al-
drich-(M3516) Mineral oil for infrared spectroscopy (light oil): d=
0.84 gmLÀ1, is completely miscible over 1608C.
[7] a) F. Ferretti, F. Ragaini, R. Lariccia, E. Gallo, S. Cenini, Organometallics
2010, 29, 1465–1471; b) F. Ferretti, E. Gallo, F. Ragaini, J. Organomet.
Chem. 2014, 771, 59–67.
[8] F. Ragaini, M. Gasperini, S. Cenini, L. Arnera, A. Caselli, P. Macchi, N.
Casati, Chem. Eur. J. 2009, 15, 8064–8077.
[9] a) B. W.-K. Chu, V. W.-W. Yam, Inorg. Chem. 2001, 40, 3324–3329; b) A. R.
Katritzky, Q.-H. Long, N. Malhotra, T. A. Ramanarayanan, H. Vedage, Syn-
thesis 1992, 911 –913.
[15] a) Y. Yang, N. Priyadarshani, T. Khamatnurova, J. Suriboot, D. E. Berg-
breiter, J. Am. Chem. Soc. 2012, 134, 14714–14717; b) N. Priyadarshani,
J. Suriboot, D. E. Bergbreiter, Green Chem. 2013, 15, 1361–1367.
[16] D. Chen, Y. Zhao, C. Zhong, S. Gao, G. Yu, Y. Liu, J. Qin, J. Mater. Chem.
2012, 22, 14639–14644.
[17] C. L. Claypool, F. Faglioni, A. J. Matzger, W. A. Goddard, N. S. Lewis, J.
Phys. Chem. B 1999, 103, 9690–9699.
[10] F. M. Menger, J. J. Lee, J. Org. Chem. 1993, 58, 1909–1916.
[11] S. M. Klein, C. Zhang, Y. L. Jiang, Tetrahedron Lett. 2008, 49, 2638–2641.
[12] J. A. Gladysz, in Recoverable and Recyclable Catalysts (Ed.: M. Benaglia),
John Wiley & Sons, Chichester, 2009, pp. 1–14.
Received: April 24, 2015
Published online on June 24, 2015
ChemCatChem 2015, 7, 2241 – 2247
www.chemcatchem.org
2247
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim