Catalytic behavior of hexaphenyldisiloxane in the synthesis of pyrite FeS2

Article abstract of DOI:10.1039/d0cc03397a

Functional small molecules afford opportunities to direct solid-state inorganic reactions at low temperatures. Here, we use catalytic amounts of organosilicon molecules to influence the metathesis reaction: FeCl2 + Na2S2 → 2NaCl + FeS2. Specifically, hexa

Full text of DOI:10.1039/d0cc03397a

ChemComm
View Article Online
View Journal
COMMUNICATION
Catalytic behavior of hexaphenyldisiloxane
in the synthesis of pyrite FeS2†
Cite this:DOI: 10.1039/d0cc03397a
Paul K. Todd, * Andrew J. Martinolich
and James R. Neilson
Received 11th May 2020,
Accepted 6th July 2020
DOI: 10.1039/d0cc03397a
rsc.li/chemcomm
Functional small molecules afford opportunities to direct solid-state these molecules do not decompose in order to take advantage of
inorganic reactions at low temperatures. Here, we use catalytic amounts their targeted functional groups.
of organosilicon molecules to influence the metathesis reaction: FeCl
Na - 2NaCl + FeS . Specifically, hexaphenyldisiloxane ((C Si O) is temperature solid-state syntheses where reaction kinetics can be
shown to increase pyrite yields in metathesis reactions performed at controlled using small molecules to direct product selectivity. For
50 8C. In situ synchrotron X-ray diffraction (SXRD) paired with differ- example, the addition of molecules like H O and (C P has
ential scanning calorimetry (DSC) reveals that diffusion-limited inter- been shown to aid in atom transfer and intermediate conversion
mediates are circumvented in the presence of (C Si O. Control in solid-state metathesis reactions, which permits the stabilization
reactions suggest that the observed change in the reaction pathway is of metastable phases like superconducting CuSe with no applied
2
+
Exchange reactions, including metathesis, provide low-
2
S
2
2
6 5 6 2
H )
1
2
6 5 3
H )
6 5 6 2
H )
2
1
5,6
imparted by the Si–O functional group. H NMR supports catalytic pressure. Exploring more functional molecules and understand-
behavior, as (C Si O is unchanged ex post facto. Taken together, ing their role in changing reaction kinetics will permit more
we hypothesize that the polar Si–O functional group coordinates to iron predictive synthesis of inorganic materials.
chloride species when NaCl and Na form, forming an unidentified, In this work, we study a metathesis reaction (Na
transient intermediate. Further exploration of targeted small molecules - FeS + 2NaCl) that produces pyrite FeS and attempt to
in these metathesis reaction provides new strategies in controlling influence the reaction pathway at low temperatures. The neat
6 5 6 2
H )
2
S
4
2 2 2
S + FeCl
2
2
inorganic materials synthesis at low-temperatures.
reaction, proceeds to completion at 350 1C for 24 h, whereas
temperatures lower than 350 1C result in diffusion-limited products
7
The use of small functional molecules to direct the outcome of a (Na S , Na S , NaFeS , and Fe S), as we have previously showed.
2
4
2 5
2
1Àx
chemical reaction is foundational to solution-based synthetic In contrast, temporary exposure to H O(g) allows the reaction to
2
chemistry, where catalysts and additives change activation barriers proceed to completion at temperatures as low as 150 1C, but it is not
along the reaction pathway. Due to the high thermal energy barriers known how this is accomplished or if any selectivity is imparted.
to diffusion, the reaction conditions of solid-state materials synth-
Here, organosilicon molecules are considered for their physical
esis exceed the temperatures where most molecules are stable, or properties and role as a mediator for ion transport (e.g., halogen
8
9
can impart selectivity, thus limiting the degree of kinetic control transfer, ferrosiloxane adducts, Table S1, ESI†) when changing
1
attainable in these reactions. There are examples of inorganic the Si–X (Si–O–Si, Si–OH, Si–Cl) functional group. We demon-
materials syntheses that take advantage of small molecules in strate that the addition of (C H ) Si O to inorganic precursors
6
5 6
2
solution, such as incorporation of weak Lewis acid or base additives selectively yields crystalline NaCl and FeS at low concentrations
2
to assist atom transfer in synthesizing transition metal phosphides and reaction times. Hexaphenyldisiloxane, a weak Lewis base,
2
and chalcogenides. These solution-based syntheses can be appears to catalytically lower the energy barrier to formation of
2 6 5 6 2
expanded through step-wise cation or anion exchange reactions FeS . Varying the concentration of (C H ) Si O changes the reac-
that use molecules like (CH Si–(S/Se) and trioctylphosphine to tion kinetics, in which a narrow concentration range (2–5 mol%)
3 3
)
3,4
1
coordinate and transfer atoms into specific lattice sites. For solid- of the molecule produces the most pyrite. H NMR reveals that the
state synthesis, reactions must proceed at temperatures where molecule remains unchanged ex post facto. Analysis of the reac-
tion pathways using in situ synchrotron X-ray diffraction (SXRD)
and differential scanning calorimetry reveals that (C H ) Si O
6
5 6 2
Colorado State University, Department of Chemistry, Fort Collins,
Colorado 80523-1872, USA. E-mail: james.neilson@colostate.edu
changes the reaction pathway by reducing the observed number
of diffusion limited intermediates. Changing the concentration of
the molecule tunes the reaction kinetics from vigorous propagation
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
d0cc03397a
This journal is © The Royal Society of Chemistry 2020
Chem. Commun.

View Article Online
Communication
events to a more mild pathway that increases the yield of FeS
ChemComm
2
.
6 5 6 2
perspective, 8.5 mol% (C H ) Si O represents B80% by volume of
Pairing solid-state precursors with functional small molecules the total precursor mixture. In previous metathetical preparations
provides new avenues for synthesizing functional inorganic mate- of transition-metal borides and nitrides where the reaction is
rials at low temperatures.
highly exothermic, additional alkali-metal halide salts are added
11
The presence of (C H ) Si O yields FeS in metathesis reactions to precursors to act as a heat sink. Hexaphenyldisiloxane also
6
5 6
2
2
performed at temperatures that are significantly lower than for the appears to function similarly. At molecular loadings where the
neat reaction. PXRD data in Fig. S1 (ESI†) of the neat metathesis reaction proceeds to iron sulfide products (0.5 o x r 4.5 mol%),
6 5 6 2
reaction show formation of crystalline NaCl, small amounts of the (C H ) Si O molecule is hypothesized to provide a heat sink to
pyrite FeS
Table S2, ESI†). As compared to the metathesis reaction without a (Fig. 1). Thus, reactions with 4.5 mol% (C
molecular additive, the PXRD results using 4.5 mol% (C H ) Si O propagate, which results in higher yields of FeS2.
2
, along with high amounts of unreacted Na
2
S
2
precursor dampen the observed propagation kinetics observed at 0.5 mol%
(
6
H
5
)
6
2
Si O do not
6
5 6 2
(MP: 225 1C) show an increase in the crystallinity and phase
Temperature-dependent in situ synchrotron X-ray diffraction
fraction of FeS and NaCl (B4:1 NaCl:FeS in Fig. S1, ESI†). After experiments reveal that (C H ) Si O reduces the number of crys-
2
2
6
5 6
2
24 h at 150 1C, the reaction has not proceeded to completion talline intermediates observed along the reaction pathway. Fig. 2
(
i.e., 2:1 NaCl:FeS
yield more FeS
reaction at 150 1C. Nonetheless, as the (C
2
in Fig. 1), yet longer reaction times (424 h) presents combined in situ SXRD and differential scanning calori-
À1
2
, suggesting that growth of FeS
2
is still limiting the metry (DSC) experiments heated at 10 1C min
Si O concentration Metathesis reactions with 4.5 mol% (C Si
to 300 1C.
6
H
5
)
6
2
H )
6 5 6
2
O (Fig. 2(a)) show
increases from 0.5 mol% to 4.5 mol%, the mol% phase fraction of a dramatic change from reactants to products below 150 1C.
the FeS in the PXRD results increases (Table S2, ESI† and Fig. 1). At Denoted by dotted lines, the precursors quickly react to yield
2
higher mol% concentrations (x Z 4.5 mol%), the products are NaCl and Na S as the only observed crystalline intermediates
2
4
comprised of Na2S2 and Na S . At 0.5 mol% (C H ) Si O, the before FeS crystallizes at high temperatures. In the DSC trace in
2 4
6
5 6
2
2
reaction self-propagates rapidly, as indicated by complete pellet Fig. 2(a), two exotherms are observed: the crystallization of NaCl at
deformation and the presence of sulfur condensed on the ampule 150 1C and the crystallization of FeS near 250 1C. The melting of
Fig. 1). PXRD reveals NaCl and Fe but no FeS supporting (C Si O is observed as the small endothermic inflection near
previous studies of highly exergonic metathesis reactions, such as 225 1C. The SXRD and DSC results in Fig. 2(b) present the neat
MoCl + 5/2Na S, where similar disproportionation of alkali sulfides metathesis reaction. Here, we observe a pathway that proceeds
results in decomposition of sulfide anions to sulfur.
2
(
7
S
8
2
H )
6 5 6
2
5
2
10,11
Taken through diffusion-limited intermediates (Na S , Na S , NaFeS ,
2
4
2
5
2
7
together, these observations indicate that (C H ) Si O directly influ- Fe S), which reproduces our prior work. Compared to Fig. 2(a),
6
5 6
2
1Àx
ences the pathway and the reaction rate.
intermediate reactions take place over a broad temperature range
The decrease in reactivity at higher molecular loadings (x Z as compared to the abrupt change in Fig. 2(a). In the DSC trace in
.5 mol%) is explained by the relative volume fractions, which Fig. 2(b), the same exotherms are observed for NaCl and FeS , yet
4
2
results in a decrease in physical contact between precursors. For an additional broad exotherm appears (175 1C to 225 1C) which we
attribute to the crystallization of diffusion-limited intermediates.
The absence of this exotherm paired with Na
identified crystalline intermediate in Fig. 2(a) suggests that
C H ) Si O directly influences the pathway by circumventing
2 4
S as the only
(
6
5 6
2
the crystallization of diffusion-limited intermediates.
1
6 5 6 2
Analysis of the (C H ) Si O molecule ex post facto using H NMR
spectroscopy (Fig. S3, ESI†) reveals that the molecule remains
unchanged. The similarities between the washed 4.5 mol%
(
C
6
H
H
5
5
)
)
6
6
Si
Si
2
O assisted metathesis products as compared to a
O standard qualitatively supports catalytic behavior.
(C
6
2
Considering the physical properties of the molecule, the
C H ) Si O is hypothesized to influence the reaction pathway
(
6
5 6
2
through three mechanisms. First, the heat released upon crystal-
lization of NaCl has the potential to melt the molecular additive,
where previous studies have observed transient temperatures of
11
metathesis reactions in excess of 1300 1C, whereas the melting
point of hexaphenyldisiloxane is 225 1C. Second, the (C Si
6 5 6 2
H ) O
could form adducts with iron or sulfur as a Lewis base or aryl-based
ligand to shuttle atoms within the reaction, similar to previous
12
Fig. 1 Mole fraction of crystalline products from Rietveld refinements of
PXRD data when varying the mol% concentration of (C Si O in
metathesis reactions at 150 1C for 24 h. The inset shows an intact reaction
pellet when using 4.5 mol% (C Si O as compared to the reaction with
.5 mol% (C Si O. At 0.5 mol% (C Si O, the pellet disintegrated
studies with triphenylphosphine. Finally, the (C H ) Si O could
6
5 6 2
6
H
5
)
6
2
form some transient, metastable intermediate that is directly
related to the polarizability of the Si–O–Si moiety. In order to test
these hypotheses, control reactions were performed using other
small molecules with targeted functional groups or properties.
6
H
)
5 6
2
0
6
H
)
5 6
2
6
H
)
5 6
2
from a rapid propagation event leaving condensed sulfur.
Chem. Commun.
This journal is © The Royal Society of Chemistry 2020

View Article Online
ChemComm
Communication
À1
Fig. 2 In situ SXRD results of the reaction Na
2
S
2
+ FeCl
2
using (a) 4.5 mol% (C
6
H
5
)
6
Si
2
O and (b) with no molecular additive heating at 10 1C min to 300 1C. DSC
plots are appended to (a) and (b) to compare crystallization exotherms. Dotted lines denote the start and stop of reactive zones in the diffraction data. The reaction
proceeds from precursors, reactive intermediates, and finally products. The colored arrows denote the exotherm, or lack there of, for crystalline intermediates.
Control reactions highlight the molecular role of (C H ) Si O in
6
5 6 2
reactions at 150 1C for 24 hours. Eicosane (C20H42 MP: 42 1C) was
used as a control to identify if the metathesis reaction could proceed
by increasing diffusion through a molten, non-specific alkane;
however, PXRD results in Fig. 3 show limited reactivity as only NaCl,
Na
metathetical preparation of CuSe
C H ) P promotes kinetic control in the formation of the metastable
2
S
2
, and Na
2
S
4
crystallize from the melt. Previous work on the
2
have shown that the Lewis base
(
6
5 3
pyrite polymorph by acting as a molecular shuttle for copper and
12
selenium. Therefore, triphenylphosphine ((C
the precursor reaction mixture at 4.5 mol% loading. The PXRD
results (Fig. 3) show NaCl formation with broad, poorly resolved FeS
reflections. Compared to the (C Si O addition, these results
6 5 3
H ) P) was added to
2
6 5 6 2
H )
suggest Lewis base addition is not general to product formation.
The lack of crystalline Fe–S phases in the diffraction pattern
suggests the (C H ) P prevents FeS nucleation and growth,
6
5 3
2
perhaps by the same atom-transfer mechanism suggested in
the CuSe
2
system. However, the phosphorus–sulfur bonds
À1
(BDE: 105.6 kcal mol ) are stronger than the phosphorus–sele-
Fig. 3 PXRD results from the reaction: Na
with various small molecule additives. The reflections for Na
NaCl are shown as tick marks above and reflections are highlighted by
2
S
2
+ FeCl
2
at 150 1C for 24 h
À1 13
nium bonds (BDE: 86.92 kcal mol ). As such, (C
tested, yet the PXRD results yielded similar features as (C
Fig. 3), which suggests that species that react primarily with sulfur
do not yield FeS . Introduction of a less basic molecule, (C N,
also yields poorly crystalline FeS , suggesting that the strength of
6 5 3
H ) PS was
2
S
2
, FeS , and
2
6 5 3
H ) P
colored rectangles for comparison. (*) denote reflections for Na
concentrations are at 4.5 mol% of each molecule.
2 4
S . All
(
2
6 5 3
H )
2
Lewis base may not be the determining factor, either. Instead, the Na S , but not with FeCl (Fig. S4, ESI†), thus yielding NaCl and
2
2
2
function of the (C
functional group in these metathesis reactions.
Changing the Si–R functional group changes the outcome of the are similar to those without a molecule. Therefore, the formation of
reaction. Control reactions with (C SiCl show reaction with a thiosilane product impedes the nucleation of FeS by extracting
6
H
5
)
6
Si
2
O seems directly related to the Si–R poorly crystalline FeS
2 6 5 3
(Fig. 3). Furthermore, when (C H ) CCl is
used in these metathesis reactions (Fig. 3), the reaction products
6
H
5
)
3
2
This journal is © The Royal Society of Chemistry 2020
Chem. Commun.

View Article Online
Communication
ChemComm
sulfur. Reactions with 13 mol% (C
pyrite and NaCl, similar to the results using (C
6
H
5
)
3
SiOH reveals crystalline directly at low temperatures. The proposed pseudo-elementary
Si O (Fig. 3). steps in eqn (1a) and (1b) suggest complete conversion to FeS
6
H
5
)
6
2
2
,
However, myriad control reactions show that this results from the yet the results in Fig. 1 suggest a portion of the products may be
formation of H O and (C H ) Si O from self-condensation reac- amorphous or poorly crystalline. Thus, we cannot eliminate the
2
6 5 6 2
tions, as detailed in the ESI.†
The collective results show that (C
tically in reducing the activation barriers in these metathesis observed in a related system.
reactions. Currently, our knowledge of how the (C Si
interacts with species present along the pathway is not well cules change the reaction pathway in the metathetical preparation
defined, yet we can hypothesize the pseudo-elementary steps to of FeS at low temperatures (150 1C). Addition of (C Si O has a
hypothesis that some of the iron and sulfur is consumed in a
6
H
5
)
6
Si
2
O behaves cataly- pathway that involves an amorphous Fe–S phase as previously
7
6
H
5
)
6
2
O
This communication demonstrates that organosilicon mole-
2
6
H
5
)
6
2
product formation based on related observations:
Na + 2FeCl + (C Si O - 2NaCl + {int} + Na
int} + Na - 2NaCl + FeS + (C Si
catalytic effect on the formation of FeS over a narrow range in
concentration (2–5 mol%). Analysis of the reaction pathway with
2
2
2
S
2
2
6
H
5
)
6
2
2
S
4
(1a)
(1b)
(
C H ) Si O reveals a decrease in the number of diffusion-limited
6 5 6 2
{
2
S
4
2
6
H
5
)
6
2
O
6 5 6 2
crystalline intermediates. We hypothesize that (C H ) Si O cata-
lyzes the reaction pathway by stabilizing reactive iron chloride
species either through a cleavage of the Si–O bond or through a
haptic coordination between iron and phenyl R-groups. This
contribution identifies strategies for rationally designing solid-
state reactions at low-temperature by coordinating functional
molecules with specific reactant compositions.
where an unknown intermediate, {int}, transiently stabilizes reac-
tive iron chloride species to avoid reactions yielding diffusion
limited intermediates; this intermediate is consistent with an
average stoichiometry: [(C
studies of Na + FeCl
reactivity is initiated by the breaking of Fe–Cl bonds. Thus, an
essential attribute of our hypotheses is that the (C Si O aids in
6 5 6 2 2
H ) O–(FeCl) ]. Previous in situ PXRD
Si
2
S
2
2
under air-free conditions show that
7
This work was supported by the National Science Foundation
6 5 6 2
H )
(DMR-1653863). We thank Dr Andrey Yakovenko and Dr Wenqian
stabilizing the Fe–Cl species as NaCl forms. The Si–O–Si moiety
possesses an unusually high bond angle (1441 Æ 0.91) and short
Si–O bond length (1.64 Æ 0.3 Å) which has been attributed to the
Xu of Argonne National Laboratory APS 17-BM for their assistance
with experiments. The authors thank Dr R. G. Finke for insightful
discussions and Dr C. Rithner for assistance with NMR spectroscopy.
14
ionic character of the Si–O bond in comparison to C–O analogs.
This polarizability could result in cleavage of the disiloxane bonds
by transition-metal halides to form transition-metal silanolates
Conflicts of interest
6 5 3
e.g., (C H ) SiO–[FeCl] ). Ferrosiloxanes and ferrosiliconate com-
15
(
plexed anions have also been reported that can form salts such There are no conflicts to declare.
16
9
as Na[Fe(OSiMe ) ], as well as cage-like complexes with iron.
3
4
Siloxanes have also been shown to stabilize cubane-type clusters Notes and references
+
17
of [Fe S ] , which have a local structure reminiscent of the
3
4
1
2
A. J. Martinolich and J. R. Neilson, Chem. Mater., 2017, 29, 479–489.
S. Seefeld, M. Limpinsel, Y. Liu, N. Farhi, A. Weber, Y. Zhang,
N. Berry, Y. J. Kwon, C. L. Perkins, J. C. Hemminger, R. Wu and
M. Law, J. Am. Chem. Soc., 2013, 135, 4412–4424.
amorphous Fe–S intermediate we previously reported in water-
promoted metathesis reactions. Another plausible hypothesis is
that (C H ) Si O may complex the iron(II) chloride through cation–p
6 5 6 2
interactions. As the siloxane functional group is quite sterically
hindered, we may be observing a haptic-type interaction where the
intermediate iron-containing species is stabilized through inter-
action with phenyl groups as chloride anions are removed to form
NaCl. Interestingly, the phenyl groups in these disiloxane mole-
7
3
A. E. Powell, J. M. Hodges and R. E. Schaak, J. Am. Chem. Soc., 2016,
138, 471–474.
4 J. L. Fenton and R. E. Schaak, Angew. Chem., Int. Ed., 2017, 1–5.
5
6
7
8
9
A. J. Martinolich and J. R. Neilson, J. Am. Chem. Soc., 2014, 136,
5654–15659.
1
A. J. Martinolich, J. A. Kurzman and J. R. Neilson, J. Am. Chem. Soc.,
2015, 137, 3827–3833.
A. J. Martinolich, J. A. Kurzman and J. R. Neilson, J. Am. Chem. Soc.,
cules can exhibit significant dipole moments along the (C
bond ((p - d)
6
H
5
)–Si–O
p
). The electron withdrawal from the phenyl groups
2016, 138, 11031–11037.
18
H. Gilman, H. N. Benedict and H. Hartzfeld, J. Org. Chem., 1954, 19,
419–427.
U. N. Nehete, G. Anantharaman, V. Chandrasekhar, R. Murugavel,
M. G. Walawalkar, H. W. Roesky, D. Vidovic, J. Magull, K. Samwer
and B. Sass, Angew. Chem., Int. Ed., 2004, 43, 3832–3835.
could hypothetically provide a haptic bonding environment, where
a half-sandwich type complex could form. The Si–O bond polariza-
tion would explain why (C H ) SiCl or (C H ) CCl are not as
6 5 3 6 5 3
reactive. This type of coordination complex would explain the 10 P. R. Bonneau, R. F. Jarvis and R. B. Kaner, Nature, 1991, 349, 510–512.
1
1
1 E. G. Gillan and R. B. Kaner, Chem. Mater., 1996, 8, 333–343.
2 A. J. Martinolich, R. F. Higgins, M. P. Shores and J. R. Neilson, Chem.
Mater., 2016, 28, 1854–1860.
absence of crystalline inorganic intermediates containing iron or
molecular intermediates. As occurs oftentimes in molecular cata-
lysis, we cannot identify the true intermediate due to its short 13 T. Cottrell, The Strengths of Chemical Bonds, Butterworths Scientific
Publications, 1958.
4 C. Eaborn, J. Chem. Soc., 1950, 3077–3089.
5 L. Jenkins and G. Willey, J. Chem. Soc., Dalton Trans., 1979, 1697–1700.
lifetime or metastability at 150 1C. In reactions without (C
iron chloride species will react to yield diffusion-limited intermedi-
ates (e.g., NaFeS
or intermediates to a reaction interface is slower than the lifetime
of the iron chloride species. By coordinating these iron chloride
species, their lifetime is extended to permit nucleation of FeS2
6 5 6 2
H ) Si O,
1
1
2
, Fe1ÀxS), which suggests diffusion of reactants 16 H. Schmidbaur and M. Schmidt, J. Am. Chem. Soc., 1962, 84, 3600–3601.
1
1
7 L. Deng, A. Majumdar, W. Lo and R. H. Holm, Inorg. Chem., 2010, 49,
1118–11126.
1
8 R. Varma, A. G. MacDiarmid and J. G. Miller, J. Organomet. Chem.,
1967, 9, 77–81.
Chem. Commun.
This journal is © The Royal Society of Chemistry 2020