This is the peer reviewed version of the following article: Adv.Synth. Catal. 2015, 357, 3538 –3548, which has
been published in final form at http://onlinelibrary.wiley.com/doi/10.1002/adsc.201500562/abstract. This article
may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Self-Archiving."

Phosphino-amine (PN) ligands for rapid catalyst discovery in
ruthenium-catalyzed hydrogen-borrowing alkylation of
anilines: a proof of principle
Lewis Marc Broomfield, Yichen Wu, Eddy Martin, Alexandr Shafir*
a

Institute of Chemical Research of Catalonia (ICIQ), Av. Països Catalans 16, 43007, Tarragona, Spain.
Fax: (+34) 977-920-222
e-mail: ashafir@iciq.es

Received: ((will be filled in by the editorial staff))
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201######.
Abstract.
A general synthetic protocol for the synthesis of simple phosphinoamine (PN) ligands is described with 19 ligands being isolated in
good yields. High-throughput ligand screening uncovered the
success of two of these ligands for aromatic-amine alkylations via
ruthenium-catalyzed hydrogen borrowing reactions. The
combination
of
N,N’-bis(diphenylphosphino)-N,N’dimethylpropylenediamine with a ruthenium(II) source and
potassium hydroxide (15 mol%) is the optimal system for selective
monobenzylations of aromatic amines (method A). Over 70%
isolated yields have been achieved for the formation of 14
secondary aromatic amines under mild reaction conditions (120 oC
and 1.05 equivalents of benzyl alcohol).

Introduction
A variety of catalytic processes utilize phosphine
ligands to moderate the reactivity of catalytic metal
centers.[1] Despite doing this job very effectively, the
synthesis of these ligands however can be work
consuming and often delicate, due to the sensitive
nature of the reagents applied (Scheme 1, a). Indeed
this aspect causes a bottle neck when requiring
speedy structural modifications for rapid reaction
discovery and optimization. In this context,
phosphino amines, referred here as PN ligands, offer
a valuable scaffold for generating structural ligand
diversity for process optimization. Contrary to their
triorganophosphine
counterparts,
these
PN
derivatives are obtainable in a relatively simple
single-step procedure by condensing an amine with a
chlorophosphine (Scheme 1, reaction B).[2] This
synthetic method is versatile and the abundance of
commercial amines allows ample ligand sets to be
swiftly generated for rapid parallel screening.[3,4]
Indeed, ligands based on the PN scaffold have
been widely used in catalytic processes, including
nitrile hydrations,[5] transfer hydrogenations of
acetophenone derivatives[6] and cross-coupling
reactions with gold.[7] Many enantioselective catalytic
processes also benefit from the modular chiral P-N
systems pioneered by Alexakis and Feringa.[8] In bulk
processes, the use of PNP ligands is the basis for

On the other hand, N,N-bis(diphenylphosphino)-isopropylamine
was the ligand utilized for both selective monomethylation and
monoethylation reactions of aromatic amines (method B). Here
the alcohol is charged as both the reaction medium and substrate
and 9 examples are disclosed with all isolated yields exceeding
70%. These methods have been applied to the synthesis of
important synthetic building blocks based on aminoferrocene.

Keywords: alkylation; hydrogen borrowing; ruthenium
catalysis; PN ligands; high throughput screening

industrial chromium-based production of 1-hexene
and, more recently, 1-octene from ethene.[2a-h]

Scheme 1. A comparison of synthetic approach towards
bisphosphines with bis-phosphino-amines.[2]

Wide-scope screening of PN ligand sets by taking
advantage of their simplistic synthesis however is
scarce. Yet, this feature has been effectively exploited
by Wasserscheid at al.[2d] in ethylene oligomerization.
This study found that the selectivity (in particular 1octene vs. 1-hexene) could be tuned by altering the
steric bulk on the central nitrogen atom of a PNP
ligand. Thus, we wondered to what extent the facile
generation of phosphino amines can be used to
develop and optimize other catalytic processes. In
light of the structural (albeit not electronic) similarity
between linear disphosphines and bis-phosphino
amines built with linear aliphatic diamines, we chose

1

Ru(II)-catalyzed H2-borrowing reactions, more
specifically alkylations of amines by alcohols, as a
possible candidate for the rapid process optimization.
This choice of process was strongly influenced by the
fact that bisphosphines are commonly utilized to
modulate metal centers in these processes (vide infra).
In addition, such catalytic protocols are in many ways
preferable to the more classical nucleophilic
substitution approaches, which typically use alkyl
halides as alkylating agents.[9] Two clear advantages
of using H2-borrowing catalysts to perform aminealkylation reactions are superior reaction control and
the fact that the toxicity of alcohols, employed as
alkylating agents, tends to be lower than of the
corresponding alkyl halide.[10-12] Therefore, amine
alkylation reactions by alcohols complement
perfectly the reductive amination of aldehydes, with
both methods affording the same product. A basic
sequence of this reaction is outlined in Scheme 2 and
operates by metal-catalyzed dehydrogenation of an
alcohol A to form an aldehyde B. Once the imine C is
formed through the condensation with an amine, the
borrowed hydrogen is then returned at the C=N bond
to yield the target secondary amine D.[13]
Significantly water is the only reaction byproduct in
this process.

methylene unit to a pentylene bridge inflicts a drastic
drop in product yield to just 26%. This difference
highlights the importance of testing many ligands
with slight structural modifications in catalysis.
Finally, Wass et al.[16] used a catalytic system
comprising of [RuCl2(p-cymene)]2/DPPM/NaOEt in
selective ethanol upgrade (Guebert process). An
impressive 23% yield with 89% selectivity for nbutanol was achieved by the in-situ system. It was
suggested the small bite angle (βn) of DPPM aided
the superlative selectivity achieved by this
homogeneous Ru(II) catalyst.

Figure 1. Selected examples of commercial bisphosphines
used as ligands in Ru(II)-catalyzed H2-borrowing reactions.

Considering these factors we decided to acquire a
robust synthetic procedure for PN ligands in order to
rapidly generate an initial ligand library. With this
ligand library in hand, we have tested its aptitude in
modulating
Ru(II)-catalyzed
H2-borrowing
alkylations of aromatic amines. Herein, we report our
findings concerning this research.

Scheme 2. Hydrogen borrowing strategy in the alkylation
of amines by alcohols.

Regarding effective catalytic systems for such
transformations, Williams has developed various by
regulating the activity of [RuCl2(p-cymene)]2 with
bisphosphine ligands (DPPF and DPEphos are two
prominent ligands used, Figure 1). The resulting
catalytic systems were found to effectively catalyze a
number of amine alkylation reactions under relatively
mild reaction conditions.[14] Additionally important
pharmaceuticals have been isolated in good yields
using
this
combination
(Piribedil,
87%,
Tripelennamine, 75% and Chlorpheniramine, 81%
are just 3 examples).[14b,14c] Along the same lines,
Deutsch et al.[15] recently applied Ru(II) complexes of
DPEphos for alkylation protocols of ammonia by
alcohols. Yields of above 90% to the desired primary
amines were achievable in some cases. Importantly,
in addition to DPEphos, several commercial
bisphosphines, bridged by linear alkyl chains (Figure
1), were also operative ligands for this reaction. Of
these, particularly effective was DPPB, which gave
cyclohexanamine in 94% yield. Worth mentioning is
that increasing this carbon backbone by one

Results and Discussion
Ligand syntheses

At the outset, three simple families of PN ligands
would be rapidly generated, allowing us to cover a
range of complex structures. The first two families
would derive from primary amines that, depending on
the number of equivalents of chlorophosphine
charged, would give either a “PNP” or a “PNH”
system (type 1 and 2). In addition, the third PNR
family is obtainable using secondary amine scaffolds
(type 3). Furthermore, within each family, both
mono- and diamines can be employed.
Thus, all ligands (1-3) were synthesized by
condensing a chlorophosphine R2PCl (R = Ph or iPr)
with the corresponding amine in the presence of Et3N
(Scheme 3). Given the size of the intended ligand
pool, the published condensation procedure[2g] was
revised in order to minimize the quantities of
chlorophosphine and base required, therefore
simplifying the time-intensive purification protocol.
The 31P NMR proved an invaluable handle for
monitoring progress of the reaction. Under these
revised conditions, the pure products are produced in
good yields (66-91%) upon extraction with toluene
and a subsequent washing of the solids with

2

acetonitrile. Alternatively, for oily products (as for 2c,
3b, 3e, 3f) a ligand of sufficient purity is obtained by
extraction with hexane. For 3b, prepared from N,N’dimethylpropylenediamine, analytically pure samples
were achieved by an additional washing with
degassed ethanol. While both 1,2-ethylenediamine
and 1,3-propylenediamine reacted readily with 4
equivalents of diphenylchlorophosphine, to afford the
corresponding bis-PNP systems (1e and 1f), in the
case of the more constrained trans-1,2cyclohexanediamine the reaction stopped at three
phosphine groups. The solid-state structure of the
resulting 2b placed the remaining NH deep within a
pocket formed by the surrounding residues; attempts
to activate this NH group, including the use of nBuLi,
were futile.

Catalytic activity

The PN ligands generated were then tested in the
model Ru-catalyzed N-benzylation of aniline
(PhNH2) with benzyl alcohol (BnOH). The initial
reaction conditions, chosen based on recent work by
Williams et al.[14b,14c,19] and our own preliminary tests,
consisted in exposing PhNH2 and a slight excess of
BnOH (1.05 equiv) to [RuCl2(p-cymene)]2 (at 1.25
mol% Ru) and KOH (15 mol%) in toluene at 150 oC.
As a benchmark, the coupling in the absence of an
additional ligand led to an 87% consumption of
PhNH2 after 12 h and the formation of a roughly
equimolar mixture of the targeted benzylaniline
(49%) and the corresponding imine (34%).
While no benzylamide (another potential side
product) could be detected, this outcome signaled a
difficulty in the last imine rehydrogenation step under
these “ligand-free” conditions. As seen in Figure 3,
the introduction of any of the PN ligands (at 2P/Ru
loading) from the preliminary library increased the
selectivity towards N-benzylaniline. Nevertheless, for
the PNP and PNH ligand families (1a-f, 2a-c), this
improvement came at a cost of the reaction rate.
However, competitive rates were recovered using the
bridged PNR ligands 3a-e. In particular, to our
delight the use of 3b, comprising two Ph2PN(Me)
moieties bridged by a propylene linker, afforded Nbenzylaniline in a 94% yield, with only traces of the
imine being detected by GC. The privileged size of
the three-carbon tether in 3b is evidenced by the
somewhat inferior catalytic performance of the twoand four-carbon bridged analogs 3a and 3c. Moreover,
the poor performance achieved using the
monophosphine 3e (44% yield, 57% selectivity)
indicates the importance of a tethered system.

Scheme 3. The generated PN-ligand library.

For 2b, the 31P-NMR spectrum contains a singlet
and two doublets, the latter for the diastereotopic
PNP phosphorous atoms.[17] Interestingly, while each
of the N-bound ring C atoms was expected to couple
to all three P atoms, the two 13C-NMR resonances
appear as doublet of doublets, indicating coupling to
only two P in each case. Selective decoupling
experiments (see Supporting Info) reveal that one of
the two P of the PNP unit does not couple with these
C nuclei, likely due to the P lone pair (lp) oriented
nearly antiperiplanar (168o in solid state) to the N-C
bond in the lp-P-N-C system (Figure 2).[18]
Figure 3. Screening of the PN library in the Ru(II)catalyzed N-benzylation of PhNH2 by BnOH. %GC yields
vs. mesitylene (average from at least two separate runs).

Figure 2. “Isolation” of one of the PNP phosphori from the
C1 and C2 spin systems; section of the corresponding 13C
NMR spectrum is shown.

Additional structure-activity insights were gained
by testing further modifications of the “winning”
propylene-bridged system 3b (Figure 4). Replacing
the Ph2P groups with more electron-donating iPr2P
(L1) led to a drop in catalyst selectivity as evidenced

3

by the formation of 25% of imine side product.
Replacing the central CH2 group with a CMe2 unit
(L2 and L3) led to relatively poor catalytic
performance, lending evidence to the key role played
by the hydrogens of the central propylene CH2 group
in 3b (vide infra). Finally, the superior catalyst
performance of 3b appears to be strongly correlated
with the presence of the NMe linkers as replacing
these with either oxygen (L4) or CH2 units (DPPPE,
Figure 6) led to systems with lower catalytic
activities.
Additional exploration (Table 1) showed that the
process could be further improved by lowering the
temperature to 120 oC, whereby the use of 1.25 mol%
of the 1:1 Ru:3b catalyst system now led to
quantitative formation of N-benzylaniline (compare
entries 1 and 2; see Supporting Info for the GC trace).
Nevertheless, inferior results were obtained at 110 oC
(entry 3) or at other ligand-to-metal ratios (entries 4
and 5). Notably, none of the benzyl benzoate side
product is formed at these lower temperatures.

Figure
4.
Performance
survey
of
additional
propylenediamine-bridged systems; conditions as in Fig. 3.

number of commercially available bisphosphine
ligands (Figure 6).
Table 1. Selected optimization experiments of the Ru(II)catalyzed benzylation of aniline.

entry

L:M

1
2
3
4
5
6[c]
7[d]

1:1
1:1
1:1
0.5:1
2:1
1:1
1:1

temp (oC)
150
120
110
120
120
120
120

%conv[b]
100
100
89
92
90
100
100

%amine[b]
94
>99
86
89
66
58
29

%imine[b]
1
0
0
0
24
42
71

[a]

In 15 mL closed tubes under inert atm. [b] % conv. and %
yields (average from at least two separate runs) by GC vs.
mesitylene as int. standard. [c] Septum was pierced with a
needle. [d] Open reactor under reflux.

Figure 5. Catalytic activity of 3b/Ru(II) vs %KOH
(coupling PhNH2 + BnOH). Rest of conditions as in run 4
in Table 1.

The efficiency of the reaction also depended on the
seal created by the septum, as the use of older septa
sometimes led to the appearance of the imine sideproduct. Indeed, piercing the septum or removing it
altogether (open reflux) led to the formation of 42%
and 71% of the imine, respectively (Table 1, entries 6
and 7). These observations indicate the importance of
the equilibrium associated with the ruthenium
dihydride species (eq 1), given its role in the [H2]
drop in non-sealed system.

Of the bases tested, potassium hydroxide led to the
most effective catalyst system (see Supporting Info
for base screening). In this case, the catalytic activity
showed a sharp dependency on KOH loading;
dipping at approx. 2 mol% and then rising sharply at
base loadings of 5-10% (Figure 5). Although 6 h was
sufficient to obtain full conversion, a 12-hour process
was chosen for robustness. We note that under these
optimized reaction conditions, 3b outperforms a

Figure 6. Comparison of 3b with a selection of
commercially available bis-phosphines.

4

Substrate scope

Given the synthetic potential of N-alkylations of
amines using H2-autotransfer in the pharmaceutical
industry,[12j,k,14b,c,20] the substrate scope for this
process was explored using the propylenediaminebased ligand 3b under the optimized conditions.
Indeed, several para- and meta-functionalized
anilines were successfully N-benzylated using BnOH.
For example, N-benzyl-4-chloroaniline was isolated
in 92% yield after 12 h (Table 2, Ae). In contrast,
anilines with either CN or CF3 in para proved
problematic (Table 2, Af and Ag).21 The coupling of
ortho-substituent anilines and alkyl amines were also
more sluggish as exemplified by the N-benzylation of
o-toluidine (42% after 12 h) and n-pentylamine (13%
after 12 h).
In terms of the alkylating agent, a range of
substituted benzyl alcohols were successfully coupled
with aniline (Table 2, products Aa and Aj-An). An
interesting case was found for p-bromobenzyl alcohol,
which initially afforded only a modest yield (40%) of
the target N-(4-bromobenzyl)aniline (Am). However,
a series of control experiments led to a modified
protocol, under which the p-Cl and p-Brbenzylalcohols were used as substrates in the
presence of 10 mol% of BnOH as a booster. The
finding stems from efforts to establish whether the
poor performance of the halogen substituted benzyl
alcohols is due to catalyst deactivation by these
substrates. Thus, for the control aniline benzylation
with BnOH in the presence of p-bromobenzyl alcohol,
not only was the presence of this additive well
tolerated, but the latter was also found to undergo
efficient coupling with aniline. Thus it would appear
that BnOH either prevents the catalyst deactivation
event, or is able to restore the catalytic activity,
particularly for the imine hydrogenation step. This
modification led to respectable yields of the target Nbenzylated products Al (74%) and Am (71%).
Conversely, although the coupling of non-benzylic
alcohols with aniline was possible, (Table 2),
products Ao and Ap), the yields did not exceed 60%.
A solvent screen, conducted as part of the reaction
optimization with 3b, revealed that the use of MeOH
as solvent led to selective aniline mono-methylation,
with N-methylaniline detected in 57% yield.
Encouraged by this result, the initial PN-ligand
library was re-examined seeking to further improve
this highly-selective methylation reaction. Heating a
MeOH solution of aniline (1 mmol in 1 mL) in the
presence of [RuCl2(p-cymene)]2 (Ru(II) = 1.25
mol%) and K2CO3 (15 mol%) at 120 oC for 12 h
under “ligand-free” conditions only afforded 35% of
N-methylaniline (at 44% conversion). The addition of
any bidentate PN ligand from the ligand library
resulted in improved catalytic performance (Figure 7).
Notably the addition of either the PNP-type 1a or the
cyclohexanediamine-based 2b (at 1:1 L:M ratio, that
is 2P:Ru) led to N-methylaniline in >80% yield. In
contrast, the monophosphine tested (3f) proved
ineffective.

Table 2. Ru(II)-catalyzed benzylations of amines using
method A: reaction scope.[a,b]

[a]

In 15 mL tubes; [b] Yields of isolated material. [c]
Corrected %GC yields; [d] %GC yield of imine; [e] at 135 oC.

Although the reaction tubes were heated by an Al
block kept at 120 oC, a thermometer placed inside the
MeOH solution in the closed reaction tube showed
the internal temperature to be 76 oC, i. e.
approximately 11 oC above the boiling point of
MeOH at 1 atm. This “overheating” proved crucial,
as reducing the temperature of the heating block to
100 oC resulted in a drop in internal temperature to 68
o
C and in yield to 63 % (Table 3, also see Supporting
Info).

Figure 7. Screening of PN ligands in standard PhNH2methylation reaction in neat MeOH. %GC yield vs.
mesitylene (average from at least two separate runs).

Considering the ease of preparation of 1a, this
ligand was applied to subsequent aniline N-alkylation
reactions using aliphatic alcohols (method B). Indeed,
by using 1.25 mol% of the Ru(II)/1a catalyst system
several anilines bearing electron-donating and

5

electron-withdrawing groups were successfully
methylated (Table 3, Bb-Bd). For example, 4-chloroN-methylaniline, Bd, is isolated in 75% yield using
these conditions. Furthermore, the use of EtOH in
place of MeOH produced the corresponding Nethylaniline derivatives Be-Bh.
As a test of catalyst utility, the 3b/Ru(II) system
was
applied
to
the
synthesis
of
Nalkylaminoferrocenes. Our interest stemmed from the
promise shown by aminoferrocenes as prodrug
candidates, including their toxicity in human
promyelocytic leukemia (HL-60) and human
glioblastoma-astrocytoma (U373).[22]
Table 3. Ru(II)-catalyzed alkylations of anilines by
method B.[a,b]

[a]

In closed tubes [b] Yields of isolated product. [c] GC
yields vs. mesitylene as int. standard (average from at least
two separate runs). [d] External heating to 100 oC.

As a first step, the parent aminoferrocene was
synthesized from iodoferrocene[23] and NH3(aq) in the
presence of a CuI-Fe2O3 catalyst, as described
recently by Gasser et al.[24] While the original
protocol employed NaOH (2.3 equiv), in our hands
omitting this base was beneficial, leading to partial
suppression of the proto-deiodination side reaction
and resulting in a 77% yield of aminoferrocene on a
3.5 mmol scale (see Supporting Info). This primary
amine underwent smooth coupling with both
electron-rich and electron-poor benzyl alcohols
(Table 4, Ca-Cc). In particular, even the less reactive
para-trifluoromethyl benzyl alcohol is an effective
benzylating agent in this case, leading to 70% yield
of the target monobenzylation product Cc. Moreover,
both the N-methylation and N-ethylation of
aminoferrocene, using this time the Ru(II)/1a-based
system, took place with >80% yield. The high
reactivity of aminoferrocene is consistent with the
trend observed earlier, whereby the N-alkylation is
favored for electron-rich aromatic amines.
Table 4. Alkylation of aminoferrocene.

Complex chemistry

In light of the potent catalytic N-benzylation
activity conferred by 3b, the Ru(II) coordination
chemistry of this ligand was studied. Despite the
potential of 3b to act as a wide-angle bidentate ligand,
exposing this bis-aminophosphine to [RuCl2(pcymene)]2 in toluene inevitably led to the formation
of the 1:2 3b:Ru complex 4. In the solid state,[25] each
phosphorous is bound to a separate (cymene)RuCl2
unit (Scheme 4) thus completing the normal threelegged piano stool arrangement of an 18-e Ru(II)
center.[26] Indeed, while a 1:1 3b:Ru ratio led to a
mixture of 4 and free 3b, at a 1:2 3b:Ru ratio the
bimetallic 4 formed quantitatively within just 30 min
(δ 75.6 ppm by 31P-NMR) and crystallized from the
reaction mixture in an 83% yield. In fact, even a new
monometallic complex 5, obtained (along with 4) by
switching to a toluene:tBuOH solvent mixture, was
shown by both 31P NMR and X-ray diffraction to
contain the 3b unit with only one of the two
phosphorous atoms bound to a (cymene)RuCl2 center.
A glimpse into the potential of 3b as a bidentate
ligand was gained by conducting the same reaction in
MeOH. Within 15 min reaction time, a yellow
solution was observed and found to contain two new
Ru-H species, 6a and 6b. The hydridic 1H-NMR
resonances (apparent dq) are observed at -5.72 and 7.11 ppm respectively (see Supporting Info). The fact
that these signals were still observed in CD3OD
identifies 3b as the hydride source. For 6a, the
apparent dq signals observed at -5.72 ppm (along
with a matching set of four 31P-NMR resonances) was
interpreted as an octahedral Ru(II)-H center bound to
four non-equivalent P atoms, likely from two
chelating 3b units, one CH-activated. Structural
elucidation of 6b led to similar conclusions, and
therefore we tentatively propose that these two
intermediates are isomers. Overtime, as the intensity
of these hydridic resonances diminished and the
solution lightens, giving rise to a new distorted
octahedral Ru(II)-H species 7. The X-ray structure of
this 18-e species reveals a trans-Cl-Ru-H center
supported by a dehydrogenated 3b unit, which now
acts as a three-coordinate mer ligand through the two
P atoms and the η2-olefin moiety. A similar structure
has already been reported by Gusev et al.[27] The
coordination sphere in 7 is completed by a
P(OMe)Ph2 group, likely from the P-N bond
methanolysis[28] of a second 3b molecule. The
hydridic resonance in 7 is found at -13.58 ppm
(apparent td), and the olefinic signals appear at δ 5.38
(dd, 3JHP = 13.6, 3J = 8.5 Hz) and 4.91 ppm (ddd, 3J =
8.5, 4.7 and 1.5 Hz). Through a sequence of
decoupling experiments, the 31P-NMR resonances at
δ 35.5 and 103.7 ppm were assigned to Pa and Pb
respectively.

6

Scheme 4. Synthesis of complexes 4, 5 and 7, along with the corresponding solid state structures

The isolated 3b/Ru(II) complexes 4 and 7 were tested
in the model benzylation reaction of aniline. As seen
in Table 5, while the Ru-H species 7 proved to be a
poor catalyst (entry 2), the performance of the
bimetallic 4 (entry 3, 84%) is in line with that of a 1:2
ligand-to-metal mixture previously tested (see Table
1, 4) Finally, the mixture of 4 (ca. 60 %) with 5 (ca.
40 %), as obtained from the reaction outlined in
scheme 4, was as efficient as the 1:1 catalyst formed
in situ (compare entry 1 and entry 4).[29]
Importantly, no hydridic resonances were detected
when [RuCl2(p-cymene)]2 was exposed to ligands 3a,
3c, 3f and L3 in CD3OD. Thus, although 7 is clearly
a catalyst decomposition product, we believe that its
formation, along with that of intermediates 6a and 6b,
signals that CH activation of the propylene backbone
could play an important role in the catalytic cycle for
the 3b/Ru(II) system (Scheme 5). Indeed, as seen
above (Figure 4) poor catalytic performance was
observed for L3 featuring a CMe2 at this key position.

As a final note, the ability of 3b to form wide bite
angle chelates was confirmed through the synthesis of
a 3b-Cr(CO)4 complex prepared in 86% yield from
Cr(CO)6. Its solid-state structure revealed a distorted
octahedral geometry with the P-Cr-P chelate angle of
107o[1b]. For this species, at 298 K the lateral NCH2
groups of the propylene bridge are unobservable in
the 1H-NMR spectrum due to a fluxional process
(ΔG‡ = 13.3 kcal mol-1). At 328 K these appear as a
single broad resonance at δ 3.40 ppm, which is
consistent with an average C2V symmetry. We also
note that heating two equiv. of Cr(CO)6 with 3b in
toluene (120 oC for 50 h) gives the bimetallic
complex 9 in 51% yield.

Figure 8. X-ray structure of 8. Selected bond lengths (Å)
and angles (˚): Cr(1)-P(1) 2.41, Cr(1)-P(2) 2.41 (1); P(1)Cr(1)-P(2) 107.0.
Scheme 5. A possible evolution of the Ru-3b system.
Table 5. Testing of the isolated Ru(II) complexes of 3b in
the benzylation of PhNH2 by BnOH.[a,b]
entry
1
2
3
4
[a]

Ru complex
3b/[RuCl2](cymene)]2
7
4
4/5 mix

% amine
>99
40
84
>99

% imine
0
35
0
0

In 15 mL closed tubes at 120 oC; [b] %GC yield vs.
mesitylene (average from at least two separate runs)

Conclusions
In summary, by screening our PN-ligand library in
Ru(II)-catalyzed H2-borrowing alkylation reactions of
aromatic amines, we were able to discover and
optimize two new high-yielding methods for
synthesizing secondary aromatic amines.
Method A utilizes bidentate PN ligands from the
PNR family with 3b giving the superlative selectivity
and product yields for this process. Indeed the
optimal combination of [RuCl2(p-cymene)]2/3b/KOH
successfully catalyzes benzylations of a variety of
aromatic
amines,
including
aminoferrocene.

7

Nonetheless this method does have its limitations and
in fact only works effectively for benzylation
reactions. Thus a second method (method B) was
developed to overcome this initial restriction.
Method B takes advantage of ligands from a
different subfamily with the PNP derivatives giving
the best performances. More specifically ligands 1a
and 2b produce the highest yields of desired
secondary amines. Moreover the alcohol serves not
only as the alkylating agent but also as the reaction
solvent in this methodology.
Noteworthy is that the two methods profit from
different ligand structures. This fact demonstrates the
importance of rapid ligand synthesis for such
screening methodologies. In a more general sense
taking advantage of facile PN bond formations for
parallel screening of ligand sets is a powerful tool for
quick discover and optimization of new reaction
methodologies. In this context we are currently
studying other processes in which this simplistic
connection can be an asset.

μmol, 1.25 mol%), potassium hydroxide (4.2 mg, 0.08
mmol, 15 mol%) and toluene (0.5 mL). Column
chromatography: silica gel, Rf = 0.44 for Hex:EtOAc,
10:1). Red solid, yield: 124 mg, 0.35 mmol, 70%. 1H NMR
(400 MHz, CDCl3) δ 7.62 – 7.59 (m, 2H, Ar), 7.53 – 7.50
(m, 2H, Ar), 4.23 (d, 3J = 5.8 Hz, 2H, CH2), 4.17 (s, 5H,
C5H5), 3.88 (s, 4H, Cp CH), 2.73 (t, 3J = 5.0 Hz, 1H, NH).
13
C NMR (101 MHz, CDCl3) δ 144.1 (Ar, ipso-CCH2),
129.6 (q, 2JCF = 32.4 Hz, ipso-CF3), 128.0 (Ar CH), 125.6
(q, 3JCF = 3.8 Hz, Ar CH), 124.3 (q, JCF = 272 Hz, CF3),
110.5 (ipso-CpN), 68.2 (Cp, C5H5), 63.3 (Cp CH), 56.2
(Cp CH), 51.7 (CH2). HRMS (ESI+) m/z calcd for
C18H16F3FeN [M]+ 359.0579, found: 359.0582.

Experimental Section

References

Synthesis of ligand 3b. A solution of N,N’dimethylpropylenediamine (1.00 mL, 8.00 mmol) in
toluene (30 mL) was treated with triethylamine (2.50 mL,
17.92 mmol) and the flask was cooled to 0 oC.
Chlorodiphenylphosphine (2.95 mL, 15.95 mmol) was
slowly added. Formation of a white salt was immediately
observed. Once the addition was completed, the cold bath
was removed. After stirring for 6 h, the mixture was
concentrated to 20 mL and filtered to separate the
ammonium salt. On removal of volatiles under reduced
pressure, 3b was recovered as a colorless oil (2.53 g, 5.38
mmol, 67%). A higher purity was achieved by extracting
the residue into hexane (50 mL), filtering, evaporating the
solvent and further washing with degassed ethanol. 1H
NMR (400 MHz, CD2Cl2) δ: 7.45–7.34 (m, 20H, Ar), 3.08
–3.02 (m, 4H, NCH2), 2.49 (d, 3JHP = 6.1 Hz, 6H, NCH3),
1.73 (p, 3J = 8.0 Hz, 2H, CH2). 13C NMR (100 MHz,
CDCl3) δ: 139.5 (d, 1JCP = 14.7 Hz, ipso-Ar CP), 132.1 (d,
2
JCP=19.5 Hz, o-Ar CH), 128.4 (p-Ar CH), 128.2 (d, 3JCP =
5.7 Hz, m-Ar CH), 54.4 (d, 2JCP = 28.2 Hz, NCH2), 37.2 (d,
2
JCP = 1.2 Hz, NCH3), 28.6 (t, 3JCP = 5.9 Hz, CH2). 31P
NMR (161 MHz, CDCl3) δ: 67.33.
Catalysis by Ru(II)-3b: synthesis of N-benzylaniline. In
an argon-filled glovebox, an oven-dried
screw-top reaction tube was charged with
[RuCl2(p-cymene)]2 (3.8 mg, 6.3 μmol,
1.25 mol% Ru), the ligand 3b (5.9 mg,
12.5 μmol, 1.25 mol%), potassium
hydroxide (8.4 mg, 0.15 mmol, 15 mol%) and toluene (1
mL). Benzylalcohol (109 µL, 1.05 mmol) and aniline (92
µL, 1.0 mmol) were then added, the tube was sealed with a
screw-cap septum and heated at 120 oC for 12 h with
agitation. Column chromatography: silica gel, Rf = 0.45
Hex:EtOAc, 9:1. Colorless oil (solidifies), yield: 169 mg,
0.92 mmol, 92%. 1H NMR (400 MHz, CDCl3) δ 7.47-7.24
(m, 7H, Ar), 6.83-6.70 (m, 3H, Ar), 4.40 (s, 2H, CH 2), 4.08
(s, 1H, NH).
Catalysis
by
Ru(II)-3b:
synthesis
of
N-(4triflouromethyl)benzyl aminoferrocene. The compound
was prepared following the
procedure described for aniline.
Here,
a
mixture
of
aminoferrocene (100 mg, 0.5
mmol),
ptrifluoromethylbenzylalcohol
(92 mg, 0.525) was exposed to [RuCl2(p-cymene)]2 (1.9
mg, 3.2 μmol, 1.25 mol% Ru), the ligand 3b (3.0 mg, 6.3

[1] See the following list for some examples: a) M. E van
der Boom, D. Milstein, Chem. Rev. 2003, 103, 1759–
1792; b) P. W. N. M. van Leeuwen, P. C. J. Kamer, J.
N. H. Reek, P. Dierkes, Chem. Rev. 2000, 100, 2741–
2769; c) K. H. Shaughnessy, P. Kim, J. F. Hartwig, J.
Am. Chem. Soc. 1999, 121, 2123–2132, d) C. A.
Tollman, Chem. Rev. 1977, 77, 313–348.
[2] a) A. Carter, S. A. Cohen, N. A. Cooley, A. Murphy, J.
Scutt, D. F. Wass, Chem Commun., 2002, 858–859; b)
D. F. Wass, BP Chemicals Ltd, WO 02/04119, 2002; c)
A. Bollmann, K. Blann, J. T. Dixon, F. M. Hess, E.
Killian, H. Maumela, D. S. McGuinness, D. H.
Morgon, A. Neveling, S. Otto, M. Overett, A. M. Z.
Slawin, P. Wassercheid, S. Kuhlmann, J. Am. Chem.
Soc. 2004, 126, 14712–14713; (d) S. Kuhlmann, K.
Blann, A. Bollmann, J. T. Dixon, E. Killian, M. C.
Maumela, H. Maumela, D. H. Morgan, M. Pretorius, N.
Taccardi, P. Wasserscheid, J. Catal., 2007, 245, 279−
284; e) T. Agapie, Coord. Chem. Rev. 2011, 255, 861–
880 and references therein; f) D. S. McGuinness,
Chem.Rev. 2011, 111, 2321 – 2341 and references
therein; g) Y. Shaikh, K. Albahily, M. Sutcliffe, V.
Fomitcheva, S. Gambarotta, I. Korobkov, R.
Duchateau, Angew. Chem. Int. Ed. 2012, 51, 1366–
1369.
[3] a) D. J. Jones, V. C. Gibson, S. M. Green, P. J. Maddox,
A. J. P. White, D. J. Williams, J. Am. Chem. Soc. 2005,
127, 11037–11046; b) T. R. Boussie, G. M. Diamond,
C. Goh, K. A. Hall, A. M. Lapointe, M. K. Leclerc, C.
Lund, V. Murphy, Symyx Technologies, Inc., USA,
W2002038628, 2002; c) D. J. Jones, V. C. Gibson, S.
M. Green, P. J. Maddox, Chem. Commun. 2002, 1038–
1039 d) J. Tian, G. W. Coates, Angew. Chem., Int. Ed.
2000, 39, 3626-3629.
[4] a) S. M. Aucott, M. L. Clarke, A. M. Z. Slawina, J. D.
Woollins, J. Chem. Soc., Dalton Trans. 2001, 972-976;
b) for a recent review on aminophosphines as synthons

Acknowledgements
We thank Dr. Andrew Chapman for his generous help at the early
stages of this project. L. M. Broomfield thanks the CELLEX
Foundation for a post-doctoral contract. This work was financed
through grants from ICIQ, MINECO (CTQ2013-46705-R and
2014-2018 Severo Ochoa Excellence Accreditation SEV-20130319) and the Generalitat de Catalunya (2014 SGR 1192).
Financial support from CELLEX Foundation through the
CELLEX-ICIQ High Throughput Experimentation platform is
gratefully acknowledged.

8

and ligands, see: J. Gopalakrishnan, Appl. Organomet.
Chem. 2009, 23, 291–318.
[5] a) E. Tomás-Mendivil, R. García-Alvarez, C. Vidal, P.
Crochet, V. Cadierno, ACS Catal. 2014, 4, 1901–1910;
b) R. García-Alvarez, P. Crochet, V. Cadierno, Green
Chem. 2013, 15, 46-66; c) R. García-Alvarez, J. Díez, P.
Crochet, V. Cadierno, Organometallics 2011, 30,
3908–3910.
[6] M. Aydemir, A. Baysal, Polyhedron 2010, 29, 1219–
1224.
[7] M. D. Levin, F. D. Toste, Angew. Chem. Int. Ed. 2014,
53, 6211–6215.
[8] a) S. R. Harutyunyan, T. den Hartog, K. Geurts, A. J.
Minnaard, B. L. Feringa, Chem. Rev., 2008, 108, 2824–
2852; b) A. Alexakis, J. E. Backvall, N. Krause, O.
Pàmies, M. Diéguez, Chem. Rev., 2008, 108, 2796–
2823.
[9] a) Y. Ju, R. S. Varma, Green Chem. 2004, 6, 219–221;
b) J. Clayden, N. Greeves, S. Warren, P. Wothers,
Organic Chemistry, Oxford University Press: USA,
2001; c) R. N. Salvatore, C. H. Yoon, K. W. Jung,
Tetrahedron 2001, 57, 7785–7811; d) C. Grimm, O.
Hamed, T. L. Vigo, N. R. Bertoniere, J. Org. Chem.
2000, 65, 9234–9237; e) K. Mohri, K. Suzuki, M. Usui,
K. Isobe, Y. Tsuda, Chem. Pharm. Bull. 1995, 43, 159–
161; f) A. Pelter, K. Smith, H. C. Brown, Borane
Reagents, Academic Press: London, 1988; g) H.
Yamamoto, K. Maruoka, J. Org. Chem. 1980, 45,
2739–2740.
[10] a) F. Valot, F. Fache, R. Jacquot, M. Spagnol, M.
Lemaire, Tetrahedron Lett. 1999, 40, 3689–3692; b) S.
Narayanan, B. P. Prasad, J. Chem. Soc. Chem. Commun.
1992, 1204–1205; c) H. Matsuhashi, K. Arata, Bull.
Chem. Soc. Jpn. 1991, 64, 2605–2606.
[11] a) C. Chiappe, D. Pieraccini, Green Chem. 2003, 5,
193–197; b) R. N. Salvatore, A. S. Nagle, K. W. Jung,
J. Org. Chem. 2002, 67, 674–683.
[12] a) G. E. Dobereiner, R. H. Crabtree, Chem. Rev. 2010,
110, 681–703; b) G. Guillena, D. J. Ramon, M. Yus,
Chem. Rev. 2010, 110, 1611–1641; c) T. D. Nixon, M.
K. Whittlesey, J. M. J. Williams, Dalton Trans. 2009,
753–762; d) S. Bähn, A. Tillack, S. Imm, K. Mevius, D.
Michalik, D. Hollmann, L. Neubert, M. Beller,
ChemSusChem. 2009, 2, 551–557; e) B. Blank, S.
Michlik, R. Kempe, Adv. Synth. Catal. 2009, 351,
2903–2911; f) S. Liu, M. Rebros, G. Stephens, A. C.
Marr, Chem. Commun. 2009, 2308–2310; g) B. Blank,
M. Madalska, R. Kempe, Adv. Synth. Catal. 2008, 350,
749–758; h) K.-I. Fujita, Y. Enoki, R. Yamaguchi,
Tetrahedron 2008, 64, 1943–1954; i) M. H. S. A.
Hamid, P. A. Slatford, J. M. J. Williams, Adv. Synth.
Catal. 2007, 349, 1555–1575; j) A. Tillack, D.
Hollmann, D. Michalik, M. Beller, Tetrahedron Lett.
2006, 47, 8881–8885; k) K-I. Fujita, T. Fujii, R.
Yamaguchi, Org. Lett. 2004, 6, 3525–3528; l) C.
Gunanathan, D. Milstein, Science 2013, 341, 249-260.
[13] a) W. M. J. Ma, T. D. James, J. M. J. Williams, Org.
Lett. 2013, 15, 4850-4853; b) A. J. A. Watson, J. M. J.
Williams, Science 2010, 329, 635–636.
[14] a) A. J. A. Watson, A. C. Maxwell, J. M. J. Williams,
J. Org. Chem. 2011, 76, 2328–2331; b) M. H. S. A.

Hamid, C. L. Allen, G. W. Lamb, A. C. Maxwell, H. C.
Maytum, A. J. A. Watson, J. M. J. Williams, J. Am.
Chem. Soc. 2009, 131, 1766–1774; c) M. H. S. A.
Hamid, J. M. J. Williams, Chem. Commun. 2007, 725–
727.
[15] W. Baumann, A. Spannenberg, J. Pfeffer, T. Haas, A.
Köckritz, A. Martin, J. Deutsch, Chem. Eur. J., 2013,
19, 17702–17706.
[16] G. R. M. Dowson, M. F. Haddow, J. Lee, R. L.
Wingad, D. F. Wass, Angew. Chem. Int. Ed. 2013, 52,
9005–9008.
[17] a) R. H. Contreras, J. E. Peralta, Prog. Nucl. Mag. Res.
SP. 2000, 37, 321–425 b) R. J. Cross, T. H. Green, R.
Keat, J. Chem. Soc., Dalton Trans. 1976, 1424–1428.
[18] a) W. H. Hersh, S. T. Lam, D. J. Moskovic, A. J.
Panagiotakis. J. Org. Chem. 2012, 77, 4968–4979; b) R.
Ballesteros-Garrido, L. Bonnafoux, F. Blanco, R.
Ballesteros, F. R. Lerroux, B. Abarca, F. Colobert, I.
Alkorta, J. Elguero, Dalton Trans. 2011, 40, 1387–
1395; c) L. D. Quin, K. C. Caster, J. C. Kisalus, K. A.
Mesch, J. Am. Chem. Soc. 1984, 106, 7021–7032; d) L.
D. Quin, M. J. Gallagher, G. T. Cunkle, D. B. Chesnut,
J. Am. Chem. Soc. 1980, 102, 3136 – 3143.
[19] S. Agrawal, M. Lenormand, B. Martín-Matute, Org.
Lett. 2012, 14, 1456–1459.
[20] K. Fujita, Z. Z. Li, N. Ozeki, R. Yamaguchi,
Tetrahedron Lett. 2003, 44, 2687–2690.
[21] Interestingly para-CN does not deactivate the catalyst
and total product yields formed from the standard
aniline are high (> 90 %). Nonetheless, the reduction of
the cyano group and its subsequent alkylation reduce
the efficiency of rehydrogenation processes. On the
other hand, 4-triflouromethylaniline does deactivate the
catalyst.
[22] a) P. Marzenell, H. Hagen, L. Sellner, T. Zenz, R.
Grinyte, V. Pavlov, S. Daum, A. Mokhir, J. Med. Chem.
2013, 56, 6935–6944; b) H. Hagen, P. Marzenell, E.
Jentzsch, F. Wenz, M. R. Veldwijk, A. Mokhir, J. Med.
Chem. 2012, 55, 924– 934.
[23] a) M. S. Inkpen, S. Du, M. Driver. T. Albrecht, N. J.
Long, Dalton Trans. 2013, 42, 2813–2816; b) J. C.
Goeltz, C. P. Kubiak, Organometallics 2011, 30, 3908–
3910.
[24] a) A. Leonidova, T. Joshi, D. Nipkow. A. Frei, J.-E.
Penner, S. Konatschnig, M. Patra, G. Gasser,
Organometallics 2013, 32, 2037–2040.
[25] CCDC-1428556, CCDC-1428557, CCDC-1428558,
CCDC-1428559 and CCDC-1428560 contain the
supplementary crystallographic data for this paper.
These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[26] a) J. H. Nelson, H. D. Hansen, Organometallics 2000,
19, 4740–4755; b) C. Moise, P. Le Gendre, P. J.
Richard, J. Organomet. Chem. 2000, 605, 151-156; c) I.
Moldes, E. Encarnación, J. Ros, A. Alvarez-Larena, J.
F. Piniella, J. Organomet. Chem., 1998, 566, 165–174;
d) M. R. J. Elsegood, D. A. Tocher, Polyhedron 1995,
14, 3147–3156

9

[27] a) V. F. Kuznetsov, A. R. Kamaluddin, A. J. Lough, D.
G. Gusev, J. Am. Chem. Soc. 2006, 128, 14388–14396;
b) D. G. Gusev, A. J. Lough, Organometallics 2002, 21,
5091–5099; c) D. G. Gusev, A. J. Lough,
Organometallics 2002, 21, 2601–2603.
[28] a) K. G. Gaw, M. B. Smith, J. B. Wright, A. M. Z.
Slawin, S. J. Coles, M. B. Hursthouse, G. J. Tizzard, J.
Organomet. Chem. 2012, 699, 39–47; b) J.
Gopalakrishnan. Appl. Organometal. Chem. 2009, 23,
291–318; c) W. Ong, A. B. Beeler, S. Kesavan, J. S.
Panek, J. A. Porco, Angew. Chem. Int. Ed. 2007, 46,
7470–7472; d) S. Priya, M. S. Balakrishna, J. T. Mague,
S. M. Mobin, Inorg. Chem. 2003, 39, 1272–1281; e) P.

W. Dyer, J. Fawcett, M. J. Hanton, R. D. W. Kemmitt,
R. Padda, N. Singh, Dalton Trans. 2003, 104–113; f) V.
Chandrasekhar, S. Kingsley, A. Vij, K. C. Lam, A. L.
Rheingold, Inorg. Chem. 2000, 39, 3238–3242; g) S.
Kingsley, A. Vij, V. Chandrasekhar, Inorg. Chem. 2001,
40, 6057–6060.
[29] As suggested by a referee, an attempt was made to
carry out the reaction using a mixture 6a+6b. Some
activity was observed, but the results have been
inconclusive, giving a reasonable activity (80%) but a
selectivity only slightly better than for complex 7.

10

Graphical abstract

Aiming to accelerate catalyst discovery, we show that large families of bis- and monodentate
aminophosphine ligands can be generated rapidly in one step from commercial amines. One such family is
shown to produce a Ru catalyst, based on N,N´-dimethyl- N,N´-bis(diphenylphosphine) propylenediamine,
highly active in alcohol amination. The best system was further investigated in terms of substrate scope and
several relevant Ru complexes were isolated and characterized. The flexibility of the approach is further
seen in the identification of a different catalyst system specific for the MeOH and EtOH electrophiles.

11