Sodium Butylated Hydroxytoluene (NaBHT) as a New and Efficient Hydride Source for Pd-Catalysed Reduction Reactions

Article abstract of DOI:10.1002/chem.201902876

NaBHT (sodium butylated hydroxytoluene), a hindered and soluble base for the efficient arylation of various base-sensitive amines and (hetero)aryl halides has been found to have an unanticipated role as a hydride donor to reduce (hetero)aryl halides and allylic acetates. Mechanistic studies have uncovered that NaBHT, but not BHT, can deliver multiple hydrides through oxidation of the benzylic methyl group in NaBHT to the aldehyde. Further, performing the reduction with NaBHT-d₂₀ has revealed that the redox-active benzylic position is not the only hydride donor site from NaBHT with one hydride in three coming, presumably, from the tert-butyl groups. The reduction works well under mild conditions and, incredibly, only consumes 20 percent of the NaBHT in the process; the remaining 80 percent can be readily recovered in pure form and reused. This, combined with the low cost of the material in ton-scale quantity, makes it practical and attractive for wider use in industry at scale.

Full text of DOI:10.1002/chem.201902876

DOI: 10.1002/chem.201902876
Communication
&
Catalysis |Very Important Paper|
Sodium Butylated Hydroxytoluene (NaBHT) as a New and Efficient
Hydride Source for Pd-Catalysed Reduction Reactions
Sepideh Sharif,[a] Michael J. Rodriguez,[b] Yu Lu,[b] Michael E. Kopach,[b] David Mitchell,[b]
Howard N. Hunter,[a] and Michael G. Organ*[a, c]
challenging substrates.[4] We have developed new bases specif-
Abstract: NaBHT (sodium butylated hydroxytoluene), a
ically for amine arylation in order to curb these compatibility
hindered and soluble base for the efficient arylation of
problems leading to chromanoxide[3] and alkaline butylated hy-
various base-sensitive amines and (hetero)aryl halides has
droxy toluene (BHT) salts[5] (see Table 1 for structures). These
been found to have an unanticipated role as a hydride
soluble bases are strong enough (pKa of electron-rich phenols
donor to reduce (hetero)aryl halides and allylic acetates.
ꢀ11.3 in water) to greatly accelerate rates, yet are hindered to
Mechanistic studies have uncovered that NaBHT, but not
prevent their untoward activity at base-sensitive groups, such
BHT, can deliver multiple hydrides through oxidation of
as esters and nitriles.[6] In particular, NaBHT (sodium butylated
the benzylic methyl group in NaBHT to the aldehyde. Fur-
hydroxytoluene), which is available either as a preformed solid
ther, performing the reduction with NaBHT-d20 has re-
or can be generated in situ from widely available and inexpen-
vealed that the redox-active benzylic position is not the
sive BHT, has proven efficacious in amine arylation, including
only hydride donor site from NaBHT with one hydride in
at scale in industry.[7]
three coming, presumably, from the tert-butyl groups. The
reduction works well under mild conditions and, incredi-
bly, only consumes 20 percent of the NaBHT in the pro-
cess; the remaining 80 percent can be readily recovered
in pure form and reused. This, combined with the low
cost of the material in ton-scale quantity, makes it practi-
cal and attractive for wider use in industry at scale.
Table 1. Amide arylation by using NaBHT as base and the precatalyst 5.[a]
Pd-catalysed arylation of amines is a prized reaction to estab-
lish CÀN bonds for a wide variety of applications.[1] The process
requires a base to deprotonate the metal ammonium complex
formed in the catalytic cycle in order to facilitate the reductive
elimination step.[2] Some bases that have been used to facili-
tate this step are quite strong, such a tert-butoxide (pKa(tert-
butanol)ꢀ19 in water) and they suffer from poor compatibility
as they attack and destroy useful functional groups.[3] Mild in-
organic bases, such as carbonate, are generally compatible
with most, if not all, labile functional groups. However, their
weak basicity (pKa(bicarbonate)ꢀ10.5 in water) and complete
insolubility in organic solvents means that reactions are very
slow, require high temperatures and often do not work with
Entry
NaBHT [equiv.]
Solvent
3 [%]
4 [%]
1
2
3
4[b]
1.5
1.5
3.0
3.0
1,4-dioxane
toluene
toluene
81
83
0
15
15
100
70
1,4-dioxane
30
[a] Reactions were performed at 0.5 molar level with respect to 1. [b] Re-
action was performed in the presence of 0.3 equivalents of 2,2,6,6-tetra-
methylpiperidine N-oxyl (TEMPO).
We are studying the arylation of amides[8] and in evaluating
different bases we uncovered an interesting and unexpected
result that appears to be unique to NaBHT. The reaction of 1
and 2 in the presence of 1.5 equivalents of NaBHT led to the
desired amide 3 and a small amount of the reduced product 4
(Table 1, entries 1 and 2). Undesired side reactions in cross-cou-
pling, such as homocoupling and reduction, are moderately
understood processes that are commonplace.[9] However,
when they are relatively minor in the face of otherwise good
results, such as we see here, they are ignored as the desired
product can be obtained readily in pure form and in high
[a] S. Sharif, Dr. H. N. Hunter, Prof. M. G. Organ
Department of Chemistry, York University
4700 Keele Street, Toronto, Ontario, M3J1P3 (Canada)
E-mail: organ@uottawa.ca
[b] Dr. M. J. Rodriguez, Dr. Y. Lu, Dr. M. E. Kopach, Dr. D. Mitchell
Lilly Research Laboratories, Indianapolis, IN 46285 (USA)
[c] Prof. M. G. Organ
Centre for Catalysis Research and Innovation (CCRI) and
Department of Chemistry and Biomolecular Sciences
University of Ottawa, Ottawa, Ontario, K1N6N5 (Canada)
Supporting information and the ORCID identification number(s) for the
author(s) of this article can be found under:
https://doi.org/10.1002/chem.201902876.
Chem. Eur. J. 2019, 25, 13099 –13103
13099
ꢀ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
yield. Shockingly, when we merely doubled the base load,
none of the coupled product 3 was formed and we only ob-
tained the reduced product 4 (Table 1, entry 3).
strong control over the extent of reduction, which only pro-
ceeded half way with 1.5 equivalents but completed when at
least 2.5 equivalents were present (Scheme 2). We further dem-
onstrated that it is the salt of BHT, and not BHT itself, that is
necessary for the reduction; when we replaced 3 equivalents
of NaBHT with the acid (i.e., BHT), zero reduction occurred.
When the reaction was carried out in the absence of the
starting amide, complete reduction was still observed eliminat-
ing the amide as the source of hydride (Scheme 1). Control ex-
periments without Pd confirmed that the process is metal
mediated, and the steric bulk of the catalyst (i.e., precatalysts
5–9)[10] does not play a significant role in the transformation.
To ensure there is no activation of the NaBHT (or destruction
of it) in the presence of the Pd catalyst, we stirred NaBHT and
5 in 1,4-dioxane at 908C for 24 h and, upon workup and purifi-
cation by column chromatography, ꢀ99% of BHT were recov-
ered, confirming that the oxidative addition partner (e.g., 1) is
necessary to react the NaBHT.
Scheme 2. Impact of the number of equivalents of NaBHT on the percent re-
duction of 1.
We next examined, which structural elements of NaBHT
were necessary for the reduction (Scheme 3). The presence,
and larger size, of two ortho substituents appear to be key;
anything smaller (11–13) than isopropyl (14), which has full
conversion, shows no sign of activity. Removal of one of the
bulky ortho substituents (16) saw a conversion drop from
quantitative (15) to only 10% illustrating the importance of
the 2,6-disubstitution pattern. Removing the methyl group
from the para position of BHT (15) had no impact on the re-
duction. It is interesting that 12 failed to provide any reduced
product, yet 14 and 15 fully reduce the chloride. It would be
anticipated that the para methyl group is strongly redoxac-
tive[11,12] and a likely candidate for the hydride source. Yet,
clearly and surprisingly, a full equivalent of hydride can be de-
livered without the benzylic methyl substituent (see below).
To better understand what is happening in this reduction,
the reaction was performed in THF-d8 in a sealed NMR tube
Scheme 1. Reduction of 1 by using NaBHT and Pd-PEPPSI precatalysts.
Investigation into the role of solvent led to two observations
(Table 2). First, the nature of the solvent seems to have little
effect on the reduction as it proceeded to completion in all
solvents attempted (Table 2, entries 1–4). Second, presuming
there is no impactful kinetic isotope effect (kie) operating, the
hydride is not coming from the solvent, thus further implicat-
ing NaBHT was the hydride source (Table 2, entries 5 and 6).
It is interesting in the case of amide arylation that doubling
the equivalence of NaBHT profoundly directed the reaction
pathway from deprotonation to reduction (Table 1). We won-
dered if the reduction pathway, in the absence of the amide,
was itself impacted by the concentration of NaBHT. Increasing
the salt from 1.5 to 3.0 equivalents indeed demonstrated
Table 2. Impact of the solvent on the reduction of 1.
Entry
Solvent
Temp. [8C]
4 [%]
10 [%]
1
2
3
4
5
6
DME
THF
toluene
benzene
benzene-d6
THF-d8
80
60
90
80
80
60
100
100
100
100
100
100
–
–
–
–
0
Scheme 3. Impact of the phenolate structure on the ability to serve as a hy-
dride donor.
0
Chem. Eur. J. 2019, 25, 13099 –13103
www.chemeurj.org
13100
ꢀ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
and followed by proton NMR spectroscopy (Figure 1). Trace di-
ethyl ether present from the in situ formation of NaBHT from
BHT and NaH serves as a useful internal standard to calibrate
and validate small shifts in the spectra. Conversion proceeded
smoothly from 1 to 4 whereas changes clearly occurred to the
bulk of the NaBHT present. The most visible changes took
place to the aromatic signal at d=6.65 ppm, which shifts
downfield as time progresses, and the appearance of a phenol
OH peak after 3 h that becomes significant after 24 h. This
peak was removed by the addition of D2O supporting the
notion that it is the phenol resonance. Further, when NaBHT
was simply warmed in the same NMR tube setup (i.e., without
1 and 5), no such phenol peak was observed over 24 h. This
suggests that the formation of BHT is not simply quenching
the anion from moisture permeating into the NMR tube over
time— it is a consequence of interaction with Pd and the re-
duction.
It could be that 17 simply is unstable and degrades into some-
thing that we have not yet been able to isolate and identify, or
that this is not the only source of hydride on NaBHT (see
below).
To examine the above question about the source of hydride
from NaBHT we conducted the reaction by using NaBHT-d20
(20) (Scheme 4). We chose substrate 19 that would lead to
high-boiling-point products to make the product recovery
highly reliable to lend greater confidence to the results. The re-
The unmistakable appearance of an aldehyde resonance (d
ꢀ9.2 ppm) encouraged us to try to isolate the trace by-prod-
ucts that formed in the reduction. Two oxidation by-products,
17 and 18,[11,12] were identified suggesting that at least some
of the hydride is coming from the methyl position of NaBHT
and that electron-transfer/radical processes are active in the re-
action mixture, respectively. Although one can develop a
mechanism for how not only one, but two hydrides could be
delivered from the benzylic position (so a double oxidation), it
is not clear why there is so little of 17 present in the mixture.
Scheme 4. Reduction of 19 by using NaBHT-d20.
action proceeded to completion and the product was found to
be 34% deuterated (21). Any kie aside, this means that,
indeed, a significant amount of hydride (relative to deuteride)
is in fact delivered. We again isolated the aldehyde 23 and the
dimer 24, but we could not find any other by-product arising
from the transfer of deuterium from 20. As a control, in the ab-
sence of 19, 100% of BHT-d20 was recovered confirming that
there is no Pd-mediated movement or scrambling of deuteri-
um without the oxidative addition partner being present.
The presence of 18[13] and 24 suggests the involvement of
radicals, or at least some form of electron transfer is taking
place in the reaction flask. To examine this, we performed a
series of experiments with TEMPO, a known quencher of radi-
cal-mediated processes.[14] As can be seen in Table 3, a progres-
sive decline in the reduction was observed when TEMPO was
added until, at half an equivalent, the reaction was fully sup-
pressed.
Table 3. Impact of TEMPO on the reduction of 1.
Entry
TEMPO [equiv.]
Conversion [%][a]
1
2
3
4
5
0
100
26
15
8
0.1
0.2
0.3
0.5
0
[a] Percent conversion of 1 to 4 was determined by 1H NMR spectroscopic
analysis of the crude reaction mixture and averaged over two runs.
Figure 1. 1H NMR spectra of the reaction mixture shown above after
a) 2 min, b) 1 h, c) 3 h, d) 24 h.
Chem. Eur. J. 2019, 25, 13099 –13103
www.chemeurj.org
13101
ꢀ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
In a related experiment, we repeated the amide arylation ex-
periment with 3.0 equivalents of NaBHT, but in the presence of
0.3 equivalents of TEMPO (Table 1). Whereas no arylation oc-
curred without TEMPO (Table 1, entry 3), its presence led to
30% conversion to 3 (Table 1, entry 4). In addition to illustrat-
ing this “recovery” in the arylation activity, this experiment also
shows that TEMPO does not shut down the activity of the Pd
catalyst, which is required both for cross-coupling and for re-
duction. It could be interpreted that reduction is a free-radical
process and shutting it down “prolongs the life” of NaBHT in
the reaction mixture to allow arylation, presumed to be an
ionic process (i.e., deprotonation)[2] a competitive opportunity
to occur.
Table 4. Substrate scope of the NaBHT-mediated reduction of (hetero)ar-
yl halides.
Combining the reaction NMR results (Figure 1), the outcome
of the deuterium study in Scheme 4, and the need for at least
2.5 equivalents of NaBHT (Scheme 1), the fate of all the NaBHT
was called into question. The salt appeared to vanish at the
end of the reaction in Figure 1, yet we know that we can re-
cover the protonate acid (i.e., BHT). When we performed a
careful workup and purification by column chromatography of
the reaction given in Table 1, entry 3 with 3.0 equivalents of
NaBHT, we recovered 2.4 of the 3.0 equivalents of the phenol
(i.e., BHT), so 80% of the total NaBHT used in the transforma-
tion could be reclaimed. The appearance of 17, and the ab-
sence of any sign of the corresponding benzyl alcohol inter-
mediate, itself a stable molecule, suggests that the second oxi-
dation is faster than the first. Based on the results in Scheme 4,
we propose that some hydride comes from one of the tert-
butyl groups, although we have not isolated such a by-product
perhaps owing to its instability. This seems more logical than
the transfer of one of the two potential hydrides from the aro-
matic ring. Further, the fact that the benzylic position (i.e., a
hydrogen) is transferred in a ratio of 2:1 relative to deuterium,
also helps to account for why less than one full equivalent of
the reducer is consumed. Accounting for the mass balance of
BHT, approximately 0.6 equivalent of NaBHT is supplying hy-
dride to the reduction and trace amounts are giving rise to
side products such as 18. So, remarkably, whereas about
3.0 equivalents are necessary to achieve full reduction of the
aryl halide at a good rate, 80% of the initial NaBHT salt is re-
covered as BHT.
Scheme 5. Pd-mediated NaBHT reduction of an allylic acetate.
partner, still 2.5 equivalents of NaBHT are required to push
halide reduction to completion. Remarkably, when 3.0 equiva-
lents of NaBHT are used in the reduction, 80% of it is recov-
ered in the acid form (i.e., BHT), which means that only
0.6 equivalents are consumed per equivalent of the oxidative
addition partner. This means that multiple hydrides are coming
from some NaBHT molecules, which is consistent with the ob-
servation of the aldehyde by-product 17 in the product mix-
ture. There is precedent in the literature for the oxidation of
this methyl group to produce 17.[11,12]
The reduction would appear to involve radicals as the addi-
tion of TEMPO suppresses the reduction until it ceases entirely
at 0.5 equivalents of the radical scavenger. The reduction
works broadly across a wide variety of (hetero)aryl halides and
also for allylic acetates. The majority of the BHT can be recov-
ered, suggesting that sustainable reduction protocols are pos-
sible by using NaBHT given its multi ton-scale availability and
very low cost (pennies on the gram).[7]
We examined the generality of the reduction and the trans-
formation worked smoothly across a selection of (hetero)aryl
substrates in high yield (Table 4).
We wondered if other Pd-catalysed reductions would also
work in the presence of NaBHT. Indeed, treatment of the allylic
substrate 30 under our standard reduction conditions led to
complete reduction producing olefin isomers 31 and 32, con-
firming the existence of a Pd–p-allyl intermediate (Scheme 5).
To summarise, we have uncovered a previously unknown hy-
dride-delivering role for the sodium salt of BHT. When barely
enough (e.g., 1.5 equiv.) NaBHT is used in Pd-catalysed amide
arylation reactions, the cross-coupled product is obtained in
high recovery with trace halide reduction. However, surprising-
ly, when the amount of the salt is increased to only 3.0 equiva-
lents, arylation no longer occurs and rather halide reduction
takes over exclusively. In the absence of the amide coupling
Acknowledgements
This work was supported by NSERC Canada in the form of a
CRD grant and by the Eli Lilly Research Award Program (LRAP).
Conflict of interest
The authors declare no conflict of interest.
Chem. Eur. J. 2019, 25, 13099 –13103
www.chemeurj.org
13102
ꢀ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Communication
[9] a) C. Adamo, C. Amatore, I. Ciofini, A. Jutand, H. Lakmini, J. Am. Chem.
Soc. 2006, 128, 6829–6836; b) Q. Liu, Y. Lan, J. Liu, G. Li, Y. D. Wu, A. Lei,
J. Am. Chem. Soc. 2009, 131, 10201–10210.
Keywords: amination · butylated hydroxy toluene · cross-
coupling · hydride reduction · PEPPSI
[10] a) A. Khadra, S. Mayer, M. G. Organ, Chem. Eur. J. 2017, 23, 3206–3212;
b) C. Lombardi, J. Day, N. Chandrasoma, D. Mitchell, M. J. Rodriguez, J. L.
Farmer, M. G. Organ, Organometallics 2017, 36, 251–254; c) S. Sharif, D.
Mitchell, M. Rodriguez, J. L. Farmer, M. G. Organ, Chem. Eur. J. 2016, 22,
14860–14863; d) B. Atwater, N. Chandrasoma, D. Mitchell, M. J. Rodri-
guez, M. G. Organ, Chem. Eur. J. 2016, 22, 14531–14534.
[1] P. Ruiz-Castillo, S. L. Buchwald, Chem. Rev. 2016, 116, 12564–12649.
[2] M. M. Driver, J. F. Hartwig, J. Am. Chem. Soc. 1997, 119, 8232–8245.
[3] K. H. Hoi, M. G. Organ, Chem. Eur. J. 2012, 18, 804–807.
[4] a) K. H. Hoi, S. C¸alimsiz, R. D. J. Froese, A. C. Hopkinson, M. G. Organ,
Chem. Eur. J. 2012, 18, 145–151; b) K. H. Hoi, S. C¸alimsiz, R. D. J. Froese,
A. C. Hopkinson, M. G. Organ, Chem. Eur. J. 2011, 17, 3086–3090.
[5] S. Sharif, R. P. Rucker, N. Chandrasoma, D. Mitchell, M. Rodriguez, M.
Pompeo, R. D. J. Froese, M. G. Organ, Angew. Chem. Int. Ed. 2015, 54,
9507–9511; Angew. Chem. 2015, 127, 9643–9647.
[11] K. Omura, J. Org. Chem. 1984, 49, 3046–3050.
[12] For an example of oxidised BHT isolated from environmental samples
leading to 17, see: E. Fries, W. Püttmann, Water Res. 2002, 36, 2319–
2327.
[13] For oxidations of BHT that lead to coalescence of two molecules of BHT
into one, see references [11] and [12].
[6] R. D. J. Froese, C. Lombardi, M. Pompeo, R. P. Rucker, M. G. Organ, Acc.
Chem. Res. 2017, 50, 2244–2253.
[14] M. S. Oderinde, R. D. J. Froese, M. G. Organ, Angew. Chem. Int. Ed. 2013,
52, 11334–11338; Angew. Chem. 2013, 125, 11544–11548.
[7] From personal accounts of M.G.O. in consulting trips to process groups
of several major pharmaceutical companies. BHT is available on a multi-
ton scale from Alfa Aesar for $0.02 per gram (quoted price).
[8] a) For seminal work in Pd-catalysed amide arylation, see: J. Yin, S. L.
Buchwald, Org. Lett. 2000, 2, 1101–1104; b) For published accounts
from our group, see: S. Sharif, J. Day, H. Hunter, Y. Lu, D. Mitchell, M. J.
Rodriguez, M. G. Organ, J. Am. Chem. Soc. 2017, 139, 18436–18439.
Manuscript received: June 21, 2019
Revised manuscript received: July 4, 2019
Version of record online: September 19, 2019
Chem. Eur. J. 2019, 25, 13099 –13103
www.chemeurj.org
13103
ꢀ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim