Homoleptic trimethylsilylchalcogenolato zincates [Zn(ESiMe3)3]- and stannanides [Sn(ESiMe3)3]- (E = S, Se): precursors in solution-based low-temperature binary metal chalcogenide and Cu2ZnSnS4 (CZTS) synthesis

Article abstract of DOI:10.1039/c9dt04144c

Organic cation salts with homoleptic zincate and stannanide anions, Ph₄P [Zn(ESiMe₃)₃] (E = S (1a), Se (1b)) and Cat [Sn(ESiMe₃)₃] (Cat = Ph₄P⁺ (E = S (2a-Ph₄P); Cat =

Full text of DOI:10.1039/c9dt04144c

Dalton
Transactions
View Article Online
View Journal
PAPER
Homoleptic trimethylsilylchalcogenolato zincates
[Zn(ESiMe3)3]− and stannanides [Sn(ESiMe3)3]−
(E = S, Se): precursors in solution-based
low-temperature binary metal chalcogenide and
Cu2ZnSnS4 (CZTS) synthesis†
Cite this: DOI: 10.1039/c9dt04144c
Jannick Guschlbauer, Tobias Vollgraff and Jörg Sundermeyer
*
Organic cation salts with homoleptic zincate and stannanide anions, Ph4P [Zn(ESiMe3)3] (E = S (1a), Se
(1b)) and Cat [Sn(ESiMe3)3] (Cat = Ph4P+ (E = S (2a-Ph4P); Cat = PPN+ (E = S 2a-Ph4P, Se (2a)) are pre-
sented and structurally characterized. Efforts to isolate neutral thermally metastable stannane precursors
[Sn(ESiMe3)4] (E = S (3), Se(4)) are reported as well. The thermal decomposition of the presented precur-
sors to yield binary sulfides and selenides of zinc and tin is investigated. Furthermore, the potential of
using the title anions as precursors in solution-based low-temperature synthesis of Cu2ZnSnS4 (CZTS) by
coprecipitation with [Cu(tmtu)3]PF6 and subsequent annealing is discussed.
Received 24th October 2019,
Accepted 29th January 2020
DOI: 10.1039/c9dt04144c
rsc.li/dalton
them to prepare materials of the CIGS family.6b Herein we
present the corresponding zincates [Zn(E-SiMe3)3]− and stan-
Introduction
Metal complexes containing trimethylsilyl chalcogenolato nanides [Sn(E-SiMe3)3]− (E = S, Se) as organic cation salts. We
ligands ESiMe3 (E = S, Se) are an emerging class of metastable also present an approach to uncharged thermally metastable
molecular compounds. These may serve as precursors for tin(IV) derivatives [Sn(E-SiMe3)4]. It will be demonstrated, that
molecular clusters or metal chalcogenide materials via low- such silylated molecular metal chalcogenide precursors ther-
temperature desilylation and condensation steps.1–3 So far, mally decompose to corresponding binary sulfides ZnS and
focus was put on uncharged heteroleptic complexes with SnS or selenides ZnSe, SnSe, and SnSe2. As representative
strongly binding coligands such as in [(tmeda)Zn(SSiMe3)2], study, for E = S, we investigate their potential as building
which decomposes thermally to ZnS nanoparticles,4 or in blocks for quaternary material Cu2ZnSnS4 (CZTS), a photovol-
[(Mes2NHC)AuS-SiMe3], which serves as starting material for taic absorber with kesterite structure that is made up by earth
the synthesis of multinary chalcogenide metal clusters with abundant and nontoxic elements only.7 CZTS has gained inter-
variable optical properties.5 Despite these perspectives there est in cadmium-free thin film solar cells with efficiencies more
was only one report on a charged coordination compounds than 10%.8 The performance can be enhanced by partial re-
with cheap and easily accessible trimethylsilylchalcogenolate placement of sulphur by selenium to form CZTSSe materials
ligands ESiMe3 (E = S, Se) – namely Li2[Mn(SSiMe3)4].6a We are with an efficiency of up to 12.6%.9 The direct band gap can be
interested to develop a new class of simple homoleptic meta- tuned from 1.04 eV to 1.48 eV by varying the chalcogen ratio of
10
late complexes with metals typically used in metal chalcogen- the solid solution system Cu2ZnSnSxSe4−x
.
An atom efficient
ide semiconductor and solar absorber materials. As new and solution based low cost access to CZTS is of interest for
examples, we introduced organic cation gallates and indates the sake of its benefits regarding ecological aspects of sustain-
[M(E-SiMe3)4]− (M = Ga, In; E = S, Se) and successfully used ability that might be incompatible with expensive vacuum
technology and high-temperature gas phase sulfuration or
selenation processes incorporating excess H2E. Another
problem is the challenging preparation of phase pure CZTS, as
the presence of binary impurities like CuxS, ZnS, or CuxSnSy
are common problems,11 that also arise in molecular precursor
decomposition strategies.12,13 As CZTS is a quaternary com-
pound the preparation is usually performed in a four com-
ponent system, in which every required element is introduced
Fachbereich Chemie and Materials Science Center, Philipps-Universität,
Hans-Meerwein-Str. 4, 35032 Marburg, Germany. E-mail: JSU@staff.uni-marburg.de
†Electronic supplementary information (ESI) available: Synthetic procedures,
experimental detail, comments on mechanisms, NMR spectra and XRD data,
TGA measurements. The CIF files of the presented structures. CCDC
1940534–1940536. For ESI and crystallographic data in CIF or other electronic
format see DOI: 10.1039/c9dt04144c
This journal is © The Royal Society of Chemistry 2020
Dalton Trans.

View Article Online
Paper
Dalton Transactions
with a distinct precursor. By using a binary precursor contain- used. Ph4P+ representatives are only accessible in minor yields.
ing the chalcogen atoms in the desired amount, the complex- They decompose slowly under inert gas at ambient tempera-
ity of the system can be reduced to a three component ture which is indicated by a pronounced darkening of the col-
system.12 Despite of such interesting application perspectives ourless powders after a few days (inert gas, 25 °C).
there are very few metastable precursors known containing two
Ph4P[Zn(SSiMe3)3] (1a) crystallizes in the monoclinic space
or more of the desired elements in one formula unit.12,13 group P21/n with four formula units and two equivalents of
Herein we show how the new title compounds act as templates diethyl ether per unit cell (Fig. 1). No intermolecular bonding
for binary intermediates [MxEy]− (E = S, Se; M = Zn, Sn) that interactions could be detected. The zinc atom is trigonal
are useful building blocks for solution based preparations of planar coordinated by three –SSiMe3 ligands. The Zn–S bond
CZTS. This approach avoids vacuum techniques and sub- lengths range from 2.2534(8) Å to 2.2543(7) Å and the S–Zn–S
sequent treatment with large excess of toxic chalcogen sources angles from 117.65(2)° to 121.72(2)°. The sum of all
such as H2E at higher temperatures.
three S–Zn–S angles is 360°, indicating the planarity of the
anion. Though trigonal planar coordinated neutral zinc
complexes14,15 and supersilyl stabilised zincates like [LxLi]{Zn
[SSi(SiMe3)3]3}16 are known, 1a is the first metastable zincate
with organic cation and cheap, technologically available and
reactive SSiMe3 ligands, an important aspect, if applications in
sustainable materials synthesis are envisaged. It is a com-
monly observed fact, that compounds with ESiR3 (E = S, Se)
substituents of decreasing steric demand of R display an
increasing lability and decreasing thermostability.1 This
makes the lightest representatives based on ESiMe3 exception-
ally attractive for thermally metastable precursors in low temp-
erature solvothermal synthesis of chalcogenide materials.
The charged nature of the [Zn(SSiMe3)3]− anion prevents
fast condensation at ambient temperature. In contrast to the
corresponding neutral zinc complexes [LxZn(SSiMe3)2],3,17 1a
can be stored without cooling for some days.
Results and discussion
Syntheses and structures of Cat [M(E-SiMe3)3] (M = Zn, Sn; E =
S, Se)
The zincates Ph4P[Zn(SSiMe3)3] (1a) and Ph4P[Zn(SeSiMe3)3]
(1b), and the stannanides Cat[Sn(SSiMe3)3] (2a-Cat), and PPN
[Sn(SeSiMe3)3] (2b) (Cat = PPN+ = [Ph3PNPPh3]+) were prepared
by treating in situ prepared chlorometalates Cat[MCl3] (M = Zn,
Sn(II)) with three equivalents of LiE-SiMe3 (E = S, Se) in tetra-
hydrofuran and allowing the reaction mixture slowly to reach
ambient temperature (Scheme 1). Cooling of the reaction
mixture and a slow warm up are important, as partly substi-
tuted intermediates [MClx(ESiMe3)3−x
methylsilylchloride and to condense with precipitation of the
corresponding metal chalcogenides at ambient or slightly
higher temperatures.
−
]
tend to eliminate tri-
The selenium homologue Ph4P[Zn(SeSiMe3)3] (1b) crystal-
lizes as crystallographically imposed centrosymmetric dianio-
2−
nic dimer [(Me3SiSe)2Zn(µ-SeSiMe3)]2 in the triclinic space
In order to separate the desired product from lithium chlor-
ide, thf is removed from the reaction mixture and the lipophi-
lic title compounds are extracted into diethyl ether. The highly
reactive anions are stable towards hydrocarbons or ethers but
they decompose in chlorinated or protic solvents, even in
dipolar aprotic MeCN and DMSO. Therefore, the yield in the
extraction and separation process can be optimized using lipo-
philic cations typically applied in ionic liquid technology. In
this fundamental study however, we intended to use higher
symmetry organic cations Ph4P+ and PPN+ in order to induce
higher melting points and crystallinity for XRD analytical
reasons. On the other hand, the cation has an influence on the
stability of these salts: stannanides 2a,b are prepared in much
better yields and can be stored at ambient temperature for
long periods of time if the PPN+ cation instead of Ph4P+ is
ˉ
group P1 with two formula units per unit cell. Surprisingly, sel-
enium with its larger atomic radius and higher polarisability
compared to sulfur leads to tetracoordinate zinc atoms and
the formation of a four membered ring structure with bridging
µ-SeSiMe3 ligands. Both zincate centres display a distorted
tetrahedral configuration. As expected, the two terminal
SeSiMe3 substituents reveal shorter Zn–Se bond distances of
Fig. 1 Molecular structure of Ph4P[Zn(SSiMe3)3] (1a). Atoms are shown
as ellipsoids with 50% probability. Hydrogen atoms and half a molecule
of diethyl ether are omitted for clarity. Selected bond lengths (Å) and
angles (°) for 1a: Zn1–S1 2.2543(7); Zn1–S2 2.2536(8); Zn1–S3 2.2534(7);
S1–Si1 2.1240(8); S2–Si2 2.1176(7); S3–Si3 2.1055(9); S1–Zn1–S2 120.64
Scheme 1 General procedure for the preparation of the title com- (2); S2–Zn1–S3 117.65(2); S3–Zn1–S1 121.72(2); Zn1–S1–Si1 104.42(3);
pounds Cat[M(ESiMe3)3]− (M = Zn, Sn; E = S, Se).
Zn1–S2–Si2 105.12(2); Zn1–S3–Si3 108.95(2).
Dalton Trans.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Dalton Transactions
Paper
Fig. 3 Molecular structure of Ph4P[Sn(SSiMe3)3] (2a). Atoms are shown
as ellipsoids with 50% probability. Hydrogen atoms and disordered
atoms are omitted for clarity. Selected bond lengths (Å) and angles (°)
for 2a: Sn1–S1 2.5521(9); Sn1–S2 2.5372(1); Sn1–S3 2.5366(6); S1–Si1
2.116(1); S2–Si2 2.113(1); S3–Si3 2.114(2); S1–Sn1–S2 94.24(3); S2–Sn1–
S3 93.01(3); S1–Sn1–S3 95.38(3); Sn1–S1–Si1 97.55(4); Sn1–S2–Si2
100.69(5); Sn1–S3–Si3 100.32(5).
Fig. 2 Molecular structure of Ph4P[Zn(SeSiMe3)3] (1b). Shown is the
dianion [(Me3SiSe)2Zn(µ-SeSiMe3)]2 with one cation. Atoms are shown
2−
as ellipsoids with 50% probability. Hydrogen atoms, disordered atoms
and one cation are omitted for clarity. Selected bond lengths (Å) and
angles (°) for 1b: Zn1–Se1 2.414(1); Zn1–Se2 2.528(1); Zn1–Se3 2.447(1);
Zn1–Se2’ 2.571(1); Se1–Si1 2.243(2); Se2–Si2 2.273(1); Se3–Si3 2.243(1);
Se1–Zn1–Se2 115.68(3); Se2–Zn1–Se3 114.42(3); Se3–Zn1–Se1 108.72
(3): Se1–Zn1–Se2’ 119.28(3); Zn1–Se1–Si1 109.14(3); Zn1–Se2–Si2
117.03(4); Zn1–Se3–Si3 103.96(4); Se2–Zn1–Se2’ 89.66(3); Zn1–Se2–
Zn1’ 90.34(3). Symmetry operation: I: 2 − x, 1 − y, 2 − z.
eties with Sn–S bond lengths ranging from 2.543(1) Å to 2.554
(9) Å and S–Sn–S angles ranging from 90.34(3)° to 96.57(3)°.
This is the first structurally characterised example of a
ESiMe3 (E = S, Se) moiety attached to tin. Sn–S–E linkages were
however isolated in sterically highly congested [Sn(µ2-S)(SSi
20a
(OtBu)3)2]2
and supersilyl derivatives {Sn[ESi(SiMe3)3]2}2
2.414(1) Å to 2.447(1) Å and an Se–Zn–Se angle of 108.72(3)°,
while the two bridging µ-SeSiMe3 ligands show Zn–Se bond
distances of 2.528(1) Å, and 2.571(1) Å, and an Se–Zn–Se′ angle
of 89.66(3)°, while the Zn–Se–Zn′ angle is 90.34(3)° (Fig. 2).
The –SiMe3 moieties of the bridging ligands are oriented in
anti fashion with an angle of 49.50(4)° towards the Zn2Se2-ring
plane. No non-ionic bonding interaction between cation and
anion can be observed (the shortest C–C distance between
anion cation is 3.984(6) Å). The dinuclear nature of the
and {Me3P-Sn[ESi(SiMe3)3]}.20b As 2a-Ph4P reveals a rather low
solubility even in thf and at the same time low thermal stabi-
lity, we neither could get acceptable NMR spectra nor an accep-
table elemental analysis of this particular compound.
Therefore we synthesised better soluble PPN[Sn(SSiMe3)3]
(2a-PPN) in order to provide the non-structural analytical and
spectroscopic data. The 29Si NMR spectra in d8-thf show a
singlet at 8.3 ppm which is superimposed by a doublet with a
2
coupling constant of JSiSn = 41.8 Hz arising from coupling of
2−
dianion [(Me3SiSe)2Zn(µ-SeSiMe3)]2 is most likely not pre-
the I = 1/2 spin isotope 29Si with the I = 1/2 spin nuclei 119Sn
and 117Sn. In the 119Sn NMR spectra a singlet at 228.5 ppm is
superimposed by a doublet with the corresponding coupling
constant 2JSnSi = 41.4 Hz.
served in thf solution, as the 1H/13C/29Si NMR spectra show
just one sharp signal for the SeSiMe3 groups. This might be
explained by dynamic ligand exchange or dimer dissociation/
association in thf solution. Another obvious explanation is,
that the affinity of Zn2+ towards thf as donor ligand is high
enough to prevent dimer formation in thf solutions. The
[Zn(ESiMe3)3]− (E = S, Se) anion is one of the very few systems
in which the selenium homologue shows a different structural
motif as the corresponding sulfur representative. The pre-
sented compounds are closing the gap of knowledge between
polynuclear silylchalcogenolato complexes of zinc14,15 and
structurally comparable alkylthiolatozincates.18,19 To the best
of our knowledge, dianionic dinuclear compounds with the
central [(RSe)2Zn(µ-SeR)2Zn(SeR)2]2− core have not yet been
Despite tremendous effort we were not able to grow XRD
diffractive crystals of the selenium homologue PPN[Sn
(SeSiMe3)3] (2b). In d8-thf only one signal was detected in mul-
tinuclear NMR spectra: 119Sn NMR (297.4 ppm, superimposed
singlet and doublet, 1JSeSn = 808.6 Hz), 77Se NMR (−263.0 ppm,
singlet) and 29Si NMR (4.3 ppm, superimposed singlet and
2
doublet, JSiSn = 44.2 Hz). Therefore, we assume a monomeric
form of 2b to be present in tetrahydrofuran solution.
Syntheses of [Sn(ESiMe3)4] (E = S, Se)
observed: the Coulomb repulsion of two approaching monoa- As tin in target materials such as SnS2 and Cu2ZnSnS4 (CZTS)
nions is probably higher for more electron rich R = alkyl occurs in oxidation state four, we were interested to develop tri-
derivatives than for R = trimethylsilyl derivatives. In nearly all methylsilylchalcogenide precursors containing Sn(IV), ideally
of such structurally characterised four-membered ring com- the unknown neutral homoleptic stannanes [Sn(ESiMe3)4] with
pounds, the substituents at the bridging chalcogen atoms E = S (3) and Se(4). Two examples of homoleptic supersilylchal-
show anti configuration. There is one example for syn cogenolates Sn{ESi(SiMe3)3}4 (E = S, Se) were mentioned in a
configuration.19
monograph.23 However they were not fully characterized by
Ph4P[Sn(SSiMe3)3] (2a-Ph4P) crystallizes in the triclinic elemental analyses and MS or XRD methods. We reacted SnCl4
ˉ
space group P1 with four formula units per unit cell (Fig. 3). In in toluene with 4 eq. of LiSSiMe3 and LiSeSiMe3 applying a
accord with VSEPR model the tin(II) centre is trigonal pyramid- 18 h temperature gradient starting at −80 °C towards +20 °C.
ally (or pseudo tetrahedrally) coordinated by three SSiMe3 moi- After filtration at 0 °C and evaporation of toluene at room
This journal is © The Royal Society of Chemistry 2020
Dalton Trans.

View Article Online
Paper
Dalton Transactions
temperature or below, we obtained pentane soluble waxy solids Furthermore, we believe, that step 2 having only a share of 6%
in about 89% and 83% yield assumed to be spectroscopically referring to the starting point of measurement (not of pure 3)
pure and analytically nearly pure [Sn(SSiMe3)4] (3) (colourless) is too small because thermal decay of 3 does not follow a
and [Sn(SeSiMe3)4] (4) (pale yellow).
purely redox neutral elimination of (Me3Si)2S but parallel a
In 29Si NMR spectra of [Sn(ESiMe3)4] in C6D6 singlets with reductive elimination of disulfide (Me3Si)2S2. In fact, PXRD of
2
superimposed doublets at 21.6 ppm with JSiSn = 32.8 Hz (E = the substance recovered from the first plateau is revealing a
2
S, 3) and at 17.7 ppm with JSiSn = 33.8 Hz (E = Se, 4) are mix of SnS2 and SnS (ESI†). For this reason, the mass loss is
observed. The 119Sn NMR spectrum of [Sn(SSiMe3)4] (3) shows not matching with theory, starting from pure SnS2 (14%,
a singlet with the corresponding superimposed doublet at Scheme 2), but is lower (only 10% relative to the plateau).
2
50.5 ppm with JSnSi = 33.6 Hz. Under standard conditions,
Another indication for a parallel occurring reductive elimin-
this coupling cannot be observed in the 119Sn NMR spectrum ation mechanism is gained, when [Sn(SSiMe3)4] (3) is ther-
of [Sn(SeSiMe3)4] (4). Instead a singlet at −307.5 ppm with a mally decomposed not at 1 bar with slow heating rate (TGA)
superimposed doublet (1JSnSe = 1570.3 Hz) is observed, which but under reduced pressure (0.01 mbar) and a very fast heating
is caused by the coupling of the I = 1/2 spin isotope 119Sn with to 180 °C (dumping a flask in a preheated oil bath). Such
the I = 1/2 spin nucleus 77Se. This assignment is confirmed by annealing of the obtained residue under inert gas at 180 °C for
a corresponding signal in the 77Se NMR spectrum. Besides the three days leads to a black powder identified via PXRD as SnS
singlet at −114.6 ppm, two superimposing doublets are with only traces of SnS2 as contamination (ESI Fig. S2†). This
1
observed with JSeSn coupling constants of 1565.1 Hz and can be explained by reductive elimination of (Me3Si)2S2 from
1496.9 Hz to both abundant 119Sn and 117Sn nuclei.
molecular Sn(IV) → Sn(II) intermediates as suggested in
Under inert conditions, compounds 3 and 4 are metastable Scheme 2.
at temperatures higher than 0 °C and start to lose (Me3Si)2E.
Certainly, a more elaborate study with focus on DFT calcu-
This is confirmed by TGA and elemental analysis of lations analysing barriers of different decay mechanisms and
[Sn(SSiMe3)4] (3) showing a slightly too low mass content for S, on TGA-MS experiments under inert gas coupled with Rietveld
C and H. Deviations from calculated mass contents of 3 are analyses to determine the fractions of phases from PXRD data
about one percent. In order to prove that, even on their trans- will have to follow to fully understand these preliminary experi-
port to the CHN analyser combustion chamber, there is a mental results.
certain loss of (Me3Si)2S at ambient temperature we followed
[Sn(SeSiMe3)4] (4) has an even higher thermal, moisture
that loss of mass by TGA at 25 °C in a glovebox. It starts and oxygen sensitivity than sulfur analogue 3. Therefore, this
without heating. With a heating rate of 10 K min−1 we observe species was only characterized by elemental analysis and mul-
an accelerated decay and two-step process. This behaviour tinuclear NMR so far. Contrastingly, when decomposing
makes it impossible to get a reliable offset temperature [Sn(SeSiMe3)4] (4) at 200 °C for two days under reduced
without active cooling. At 121 °C a plateau is reached, roughly pressure, only the expected solid SnSe2 could be identified by
referring to the formation of SnS2 (calc. −66%/obs. −41%). PXRD (refer to ESI† for details). Unfortunately, even very soft
Further heating leads to a second thermal decay step referring MS ionisation techniques such as liquid injection field desorp-
to loss of S and formation of SnS (calc. −6% relative to starting tion (LIFDI) did not allow to observe the molecular ions of
point of measurement, Fig. 4). This thermal conversion of both precursor molecules 3 and 4 or of defined higher mole-
SnS2 to SnS, is well documented in literature.24 We explain the cular tin chalcogenide clusters as intermediates of the conden-
mismatch between calc. and lower observed mass loss in step sation process, so far: it is characteristic, that silyl groups are
1 by the fact, that the measurement was started at room temp- too labile upon ionisation. As a matter of fact, such neutral
erature, when already part of
3
had lost (Me3Si)2S. pentane soluble compounds [Sn(ESiMe3)4] are difficult to crys-
tallize even at −80 °C. They show a much higher tendency to
condense to higher nuclear clusters and, driven by the high
lattice energy, finally to SnE2 than the four isoelectronic
anions in salts Cat [M(ESiMe3)4] (M = Ga, In).6b The thermal
decay of 3 and 4 in pentane is observable, as clear colourless
Scheme 2 Parallel thermal decomposition paths discussed: (1) elimin-
ation of (Me3Si)2S and (2) reductive elimination of (Me3Si)2S2 towards Sn
(II) species (PXRD in the ESI†).
Fig. 4 TGA of [Sn(SSiMe3)4] (3) (heating rate 10 K min−1
.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Dalton Transactions
Paper
solutions become cloudy and ochre coloured with time at
ambient temperature under inert gas. Addition of non-dried
aprotic or protic solvents such as MeOH leads to quantitative
desilylation, formation of Me3SiOMe, H2E and rapid precipi-
tation of canary yellow SnS2 or brownish-black SnSe2 particles
from purely organic solvents. This is another perspective of
using these highly reactive, highly lipophilic and thermally
metastable Sn/S/Se precursors in coating techniques.
Thermal decomposition of Cat[M(E-SiMe3)3] (M = Zn, Sn; E =
S, Se)
Fig. 6 TGA measurements (25 °C–600 °C, 10 K min−1) of PPN[Sn
(SSiMe3)3] (2a-PPN, black) and PPN[Sn(SeSiMe3)3] (2b, red).
In order to investigate the thermal decomposition of the more
stable metalates 1a–2b, TGA measurements were conducted in
a nitrogen filled glovebox. The residues obtained were investi-
gated via PXRD. The TGA diagrams of the zincates 1a and 1b
are shown in Fig. 5. For both compounds, decomposition
starts below 160 °C. At 500 °C no more mass loss is detected.
It is assumed, that the rather broad temperature range of
decomposition is due to a dissociation step of the metalate
complexes into a neutral [M(ESiMe3)2] and ionic Ph4P[ESiMe3]
species as shown in Scheme S2.† An analogue decomposition
path was spectroscopically detected in solutions of related triel
metallates such as Ph4P[In(SSiMe3)4].6b In the ESI, Fig. S9,† it
is demonstrated, that lack of a sharp decay is due to the broad
multistep decay of the formed ionic species Ph4P[ESiMe3]
rather than the expected spontaneous decay of unstable
[M(E-SiMe3)2] (M = Zn, Sn; E = S, Se). It is worth to mention,
that both salts Ph4P[ESiMe3] degrade to volatile species includ-
ing Ph3PE without substantial residue at >400 °C (Fig. S9†).
Much sharper decomposition points at lower temperatures are
observed in N,N′-dialkyl imidazolium salts [RR′Im][ESiMe3]
offering a different decay path of N-dealkylation by the highly
nucleophilic [ESiMe3] anion.34
process starts below 160 °C. At 500 °C no more mass loss can
be detected. The mass loss is even less rapid compared to the
zincates 1a and 1b which is probably connected to the higher
thermal stability of the PPN+ compared to Ph4P+ cations. The
total mass losses were found to be 83% for PPN[Sn(SSiMe3)3]
(2a-PPN) and 77% for PPN[Sn(SeSiMe3)3] (1b), which is close
to the theoretical 85% (2a-PPN to SnS) and 83% (2b to SnSe).
Both residues could be identified as SnS and SnSe, respect-
ively, by PXRD (ESI Fig. S8†).
Preparation of CZTS
As the precursors [Sn(ESiMe3)4] are too labile for long time
storage in a technical printed electronics application, a
different protocol was developed involving oxidation of stanna-
nide anion [Sn(SSiMe3)3]− (2) by one atom equivalent of
elemental sulfur. The latter step leads instantaneously to a pre-
1
cipitate, not even soluble enough to be examined via H NMR
spectroscopy, unfortunately. Based on the high condensation
tendency, we assume, that sulfur is oxidatively added to tin(II)
and tin(IV) intermediates are formed. At this stage, it cannot be
fully excluded, that sulfur might insert into a Sn(II)–S bond
forming a disulfido species [Sn(SS-SiMe3)(S-SiMe3)2]− first. To
the best of our knowledge, Sn(II)disulfido species have never
been isolated which might be due to their tendency to form
Sn(IV)sulfido species at elevated temperatures. If existent, a
[Sn(SS-SiMe3)(S-SiMe3)2]− species should be of molecular
nature and soluble, which is incompatible with the instantly
starting condensation and precipitation of a cocktail of species
referred here as in situ formed “[Sn(S)(SSiMe3)3]−”. Planned
studies will be revealing, whether a replacement of purely
organic cations by alkali cryptand cations will allow to crystal-
lise intermediates of this condensation and cluster forming
process. In this study, the focus was put on gaining first proof
of concept, that such soluble and metastable sources for the
elements Sn/Zn/S/Se can in fact be used in a solution based
approach towards colloids and microcrystalline co-precipitates
of CZTS and its building blocks. This would be a precondition
for further developing a printed electronics process.
Total mass losses presented in Fig. 5 were found to be 77%
for Ph4P[Zn(SSiMe3)3] (1a) and 83% for Ph4P[Zn(SeSiMe3)3]
(1b), which is close to theoretical mass losses 86% (1a to ZnS)
and 83% (1b to ZnSe). Accuracy is influenced by residues of
the Ph4P[ESiMe3] decay. The microcrystalline residue of the
TGA measurements of 1a and 1b were identified as ZnS and
ZnSe respectively by PXRD (ESI Fig. S7†).
The TGA diagram of the stannanides 2a-PPN and 2b is
shown in Fig. 6. For both compounds the decomposition
In a preliminary experiment, we observed, that thermal pre-
treatment of a finely ground 1 : 1 mixture of “Cat [Sn(S)
(SSiMe3)3]” and Cat [Zn(SSiMe3)3] (160 °C, 10 min) under
Fig. 5 TGA measurements (25 °C–600 °C, 10 K min−1) of Ph4P[Zn
(SSiMe3)3] (1a, black) and Ph4P[Zn(SeSiMe3)3] (1b, red).
This journal is © The Royal Society of Chemistry 2020
Dalton Trans.

View Article Online
Paper
Dalton Transactions
reduced pressure leads to partial dissociation of the metalates only on optimising the CZTS particle co-precipitation process
into Cat [SSiMe3] (detected by NMR in an extract). The plaus- and determining the phase fractions via Rietveld analyses of
ible neutral species “[Sn(S)(SSiMe3)2]n”, and “[Zn(SSiMe3)2]n” the powder XRD diffraction data obtained by different precipi-
formally formed upon dissociation are very labile, prone to tation methods and correlating them with EDX data is needed
eliminate (Me3Si)2E at these conditions and to form metal to further develop the application of our molecular building
chalcogenide nano- and microparticles. This characteristic blocks described here. Some suggestions for simplifying the
condensation tendency is reduced or even inhibited at room process are found in the ESI, Schemes S1 and S2.†
temperature by the presence of Cat[ESiMe3] via formation of
more stable metalate complexes. In contrast to the starting
materials, the residue of this thermal pre-annealing process
can be suspended as thermally stable beige, nearly coloureless
Conclusions
dispersion in MeCN. To this dispersion two equivalents of Novel trimethylsilylthio- and seleno zincates [Zn(ESiMe3)3]−
[Cu(tmtu)3]PF6 (tmtu = N,N,N′,N′-tetramethylthiourea), dis- (E = S, Se) and corresponding stannanides [Sn(ESiMe3)3]− (E =
solved in MeCN, were added dropwise. The suspension S, S), were isolated and fully characterized as their organic
immediately turns black. After 18 h at 20 °C, the solvent was cation salts. Interestingly, thiolato zincate Ph4P[Zn(SSiMe3)3]
removed in vacuum, replaced by diglyme and the black dis- (1a) reveals a trigonal planar configuration in the XRD ana-
persion was stirred in diglyme at 160 °C/5 d for equilibrating lysis, selenolato zincate [(Me3SiSe)2Zn(µ-SeSiMe3)]22−, (1b) a
the nanoparticular phases under a high boiling dispersing tetrahedral configuration at zinc forming a dianionic dinuclear
agent, that at the same time dissolves LiCl and PPN[PF6] by- complex. As expected for a main group element, trimethyl-
products. Finally, the matured precipitate was centrifuged, silylthiolato stannanide Ph4P[Sn(SSiMe3)3] (2a) is coordinated
washed, dried and annealed at 300 °C for 18 h. The combi- in a trigonal pyramidal (pseudotetrahedral) fashion. The
nation of PXRD and Raman spectroscopy revealed the presence thermal decomposition of title salts 1a, 1b and 2a-PPN and 2b
of microcrystalline kesterite CZTS next to reflexes at 26.9, 30.5 has been investigated by TGA measurements. The residues
and 39.6, which can be attributed to either wurtzite ZnS or have been identified as corresponding binary chalcogenides
CZTS of wurtzite type. The PXRD patterns of orthorhombic ZnS, ZnSe, SnS and SnSe by PXRD. While these metalates, due
CZTS and cubic ZnS are nearly identical.21 At this stage we did to their anionic nature and large organic cation, are thermally
not have access to EDX analysis, therefore further evidence for rather stable, neutral tin(IV) derivatives [Sn(ESiMe3)4] (E = S (3),
the presence of orthorhombic CZTS was gained from Raman Se (4)) are metastable at room temperature. Nevertheless, they
spectroscopy: The signal at 336 cm−1 is associated with the can be synthesised, NMR spectroscopically fully characterised
totally symmetric M–S vibration in MS4 tetrahedra. This is the and used as sources to precipitate SnS2 and SnSe2 from
by far most dominant first order mode for CZTS.25 Both the organic solution containing water or MeOH. Thermal decay of
wavenumber and line half width perfectly fit to colloidal nano- 3 and 4 does not allow storing them for extended periods of
powders of CZTS obtained by hot injection methods (Fig. 7).22
time as inks at room temperature. Therefore a different solu-
An excellent investigation on the polymorphism of tion-based approach towards CZTS particles starting from tin
Cu2ZnSnS4 and its implication to Raman spectra of these (II) derivative PPN[Sn(SSiMe3)3] 2a-PPN, its oxidation by
materials as well as their characteristic phase impurities ZnS elemental sulfur, subsequent additions of 1 eq. zincate PPN
and SnS has appeared recently.27b A follow-up study focussing [Zn(SSiMe3)3] and 2 eq. of copper source [Cu(tmtu)3]PF6 in
MeCN was investigated. The non-crystalline black precipitate
from this three-component mix gave, after washing and
annealing steps, microcrystalline quaternary chalcogenide
material CZTS. Though further optimization is needed in
order to get phase pure CZTS, this study might be seen as fun-
damental investigation of an optional sol–gel-type conden-
sation approach towards such multinary higher chalcogenide
materials. There are plenty of options by modifying the
organic cations and solvents without changing the novel
anionic Zn/Sn/S/Se sources described here. Our results
demonstrate, that CZTS particles can be prepared by precipi-
tation of non-crystalline particles from organic solvents at low
temperatures.
Fig. 7 PXRD of Cu2ZnSnS4 after annealing at 300 °C for 1 day (PDF of
kesterite CZTS reference 26-0575). The additional reflections at 26.9,
30.5, 39.6 reveal the presence of either wurtzite ZnS or wurtzite-type
CZTS, which has a nearly identical powder diffractogram as wurtzite ZnS
(see ESI Fig. S10†). A Raman spectrum was recorded as an additional
indicator for CZTS. The presence of the 336 cm−1 and absence of the
first order Raman mode at 351 cm−1 is an indicator but no proof for a
low fraction of w-ZnS.27
Materials and methods
All preparative operations were conducted by using standard
Schlenk techniques and freshly dried solvents. Solvents were
Dalton Trans.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Dalton Transactions
Paper
dried according to common procedures26 and passed through Preparation of Ph4P [Zn(SSiMe3)3] (1a)
columns of aluminium oxide, R3-11G-catalyst (BASF) or mole-
A mixture of ZnCl2 (38 mg, 0.28 mmol, 1.0 equiv.) and Ph4P
[Cl] (104 mg, 0.28 mmol, 1.0 equiv.) in thf (10 mL) is stirred
for 18 h at room temperature. To a solution of S(SiMe3)2
(153 mg, 0.86 mmol, 3.1 equiv.) in thf (5 mL) a 2.9 M solution
of n-BuLi in hexane (0.29 mL, 0.83 mmol, 3.0 equiv.) is added
dropwise at 0 °C in a separate reactor. After 30 min at 0 °C and
further 30 min at ambient temperature, all volatiles were
removed under fine vacuum with the help of diethyl ether
(2 mL) and pentane (2 mL). The obtained colourless powder is
dissolved in thf (10 mL) and added dropwise to the suspension
of Ph4P[ZnCl3] in thf (10 mL) at −78 °C. The reaction mixture
is stirred for 18 h within the cooling bath and allowed to
slowly warm up to ambient temperature at the end of this
period. All volatiles are removed in fine vacuum and the
residue is extracted with Et2O (60 mL). After filtration all volatiles
were removed from the obtained filtrate in fine vacuum. The
residue was washed with pentane and dried in fine vacuum. 1a is
obtained as a colourless solid (119 mg, 0.17 mmol, 60%). X-ray
suitable single crystals were grown by layering a saturated solu-
tion of 1a in Et2O under the same volume of pentane at 0 °C.
1H NMR (300.3 MHz, thf[d8]): δ = 7.95–7.81 (m, 20 H, Ph4P+), 0.19
(s, 27 H, [Zn(SSiMe3)3]−) ppm. 13C NMR (75.5 MHz, thf[d8]): δ =
cular sieves (3 Å or 4 Å). Elemental analyses (C, H, N, S) were
carried out by the service department for routine analysis from
the department of chemistry at the university of Marburg with
a vario MICRO cube (Elementar). Samples for elemental ana-
lysis were weighted into tin capsules inside a nitrogen filled
glovebox. 1H and proton decoupled 13C NMR spectra were
recorded in automation with a Bruker Avance II 300 spectro-
meter, 29Si NMR spectra were recorded by the service depart-
ment for NMR analyses with a Bruker Avance II HD 300, DRX
400 or Avance III 500 spectrometer. All spectra were recorded
at ambient temperature. 1H and 13C NMR spectra were cali-
brated using residual proton signals of the solvent (thf[d8]: δH
3.58 & 1.72 ppm, δC 67.21 & 25.31 ppm, C6D6: δH 7.16 ppm, δC
128.06 ppm). 29Si NMR spectra were referenced externally
(SiMe4: δSi 0.00 ppm). For NMR solvents other than thf[d8],
such as DMSO[d6], MeCN[d3], CDCl3, and CD2Cl2, decompo-
sition of the title anions was observed. C6D6 is not polar
enough to dissolve the ionic compounds, while thf[d8] only
slightly dissolves the ionic target compounds. Silicon grease
impurities that in some samples are indicated in the NMR
spectra are not attributed to impure substances, as proven by
elemental analysis, but rather to the low solubility of the com-
pounds in thf[d8].
4
3
136.2 (d, JCP = 3.0 Hz, (HC(CH)2(CH)2C)4P+), 135.6 (d, JCP
=
2
10.4 Hz, (HC(CH)2(CH)2C)4P+), 131.3 (d, JCP = 13.0 Hz, (HC
TGA measurements starting at 25 °C were conducted with a
DSC-TGA 3 (Mettler Toledo) in a glovebox. Decomposition
temperatures were determined using data of the DSC-TGA.
All solvents were dried according to common procedures
and passed through columns of aluminium oxide, 3 Å mole-
cular sieves and R3-11G catalyst (BASF) or stored over mole-
cular sieves (3 Å or 4 Å). The reagents were used as received
1
(CH)2(CH)2C)4P+), 118.9 (d, JCP = 89.5 Hz, (HC(CH)2(CH)2C)4P+),
6.5 (s, [Zn(SSiMe3)3]−) ppm. 29Si NMR (59.7 MHz, thf[d8]): δ =
8.3 (s, [Zn(SSiMe3)3]−) ppm. Elemental analysis exp. (calc.): C 55.1
(55.0), H 6.9 (6.6).
Preparation of Ph4P [Zn(SeSiMe3)3] (1b)
unless stated otherwise. [Cu(tmtu)3][PF6],27 E(SiMe3)2,28 A mixture of ZnCl2 (32 mg, 0.23 mmol, 1.0 equiv.) and Ph4P
29
LiESiMe3 (E = S, Se; tmtu = N,N,N′,N′-tetramethylthiourea) [Cl] (87 mg, 0.23 mmol, 1.0 equiv.) in thf (10 mL) is stirred for
were prepared according to literature known procedures or 18 h at room temperature. To a solution of Se(SiMe3)2 (162 mg,
slightly modified.
0.72 mmol, 3.1 equiv.) in thf (5 mL) a 2.9 M solution of n-BuLi
The data collection for the single-crystal structure determi- in hexane (0.24 mL, 0.70 mmol, 3.0 equiv.) is added dropwise
nation was performed on a Stoe Stadivari diffractometer or a at 0 °C in a separate reactor. The reaction mixture is stirred for
Bruker D8 Quest diffractometer by the X-ray service depart- 30 min at 0 °C and further 30 min at ambient temperature. All
ment of the department of chemistry at the university of volatiles were removed under fine vacuum. The residue was
Marburg. Information concerning the used hardware, and soft- washed with pentane and dried in fine vacuum. The obtained
ware used for data collection, cell refinement and data colourless powder is dissolved in thf (10 mL) and added drop-
reduction as well as structure solution and refinement can be wise to the suspension of Ph4P[ZnCl3] in thf (10 mL) at
reviewed in the attached CIF. After the solution (Shelxt)30 and −78 °C. The reaction mixture is stirred for 18 h and allowed to
refinement process (Shelxl 2017/1)31 the data was validated by slowly warm up to ambient temperature at the end of this
using Platon.32 All graphic representations were created with period. All volatiles are removed in fine vacuum and the residue
Diamond 4.33
is extracted with Et2O (60 mL). After filtration all volatiles were
PXRD measurements were conducted by giving a small removed from the obtained filtrate in fine vacuum. The residue
amount of the sample between two layers of Scotch tape. The was washed with pentane and dried in fine vacuum. 1b·0.5Et2O
diffractograms were recorded on a Stoe Stadi MP using Cu or is obtained as a colourless solid (118 mg, 0.13 mmol, 52%).
Mo-Kα1 radiation that passed a bent focusing monochromator X-ray suitable single crystals were grown by layering a saturated
and a high sensitivity Mythen-detector. The transmission diffr- solution of 1b·0.5Et2O in Et2O under the same volume of
actograms were recorded from 2theta = 5°–105° (5° min−1). pentane and storing the mixture at 0 °C for a few days. 1H NMR
Raman Spectra were recorded using an InVia Raman micro- (300.3 MHz, thf[d8]): δ = 7.95–7.82 (m, 20 H, Ph4P+), 3.39 (q,
scope from Renishaw utilized with a frequency doubled Nd- 3JHH = 7.0 Hz, 1H, 0.5 O(CH2CH3)2), 1.12 (t, 3JHH = 7.0 Hz, 1.5 H, 0.5
YAG laser with a wavelength of 532 nm.
O(CH2CH3)2), 0.33 (s, 27 H, [Zn(SeSiMe3)3]−) ppm. 13C{1H} NMR
This journal is © The Royal Society of Chemistry 2020
Dalton Trans.

View Article Online
Paper
Dalton Transactions
(75.5 MHz, thf[d8]): δ = 136.2 (d, 4JCP = 3.1 Hz, (HC(CH)2(CH)2C)4P+), 2JCN = 1.9 Hz, ((HC(CH)2(CH)2C)3P)2N+), 6.1 (s, [Sn(SSiMe3)3]−)
3
2
135.6 (d, JCP = 10.4 Hz, (HC(CH)2(CH)2C)4P+), 131.3 (d, JCP
=
ppm. 29Si NMR (59.7 MHz, thf[d8]): δ = 8.3 (s with superimposed
12.9 Hz, (HC(CH)2(CH)2C)4P+), 118.9 (d, JCP
=
89.5 Hz, d, JSiSn = 41.8 Hz, [Sn(SSiMe3)3]−) ppm. 119Sn NMR (112.0 MHz,
1
2
(HC(CH)2(CH)2C)4P+), 7.0 (s, [Zn(SeSiMe3)3]−) ppm. 29Si NMR thf[d8]): δ = 228.5 (s with superimposed d, JSnSi = 41.4 Hz,
2
(59.7 MHz, thf[d8]): δ = 4.3 (s, [Zn(SeSiMe3)3]−) ppm. Elemental ana- [Sn(SSiMe3)3]−)
ppm.
Elemental
analysis
exp.
(calc.
lysis exp. (calc. C35H52O0.5PSe3Si3Zn): C 46.9 (46.8), H 5.9 (5.8).
C45H57NP2S3Si3Sn): C 55.1 (55.6), H 5.6 (5.9), N 1.5 (1.4).
Preparation of Ph4P [Sn(SSiMe3)3] (2a-Ph4P)
Preparation of PPN [Sn(SeSiMe3)3] (2b)
A mixture of SnCl2 (49 mg, 0.26 mmol, 1.0 equiv.) and Ph4P A mixture of SnCl2 (85 mg, 0.45 mmol, 1.0 equiv.) and PPN[Cl]
[Cl] (97 mg, 0.26 mmol, 1.0 equiv.) in thf (10 mL) is stirred for (258 mg, 0.45 mmol, 1.0 equiv.) in thf (10 mL) is stirred for
18 h at room temperature. To a solution of S(SiMe3)2 (141 mg, 18 h at room temperature. To a solution of Se(SiMe3)2 (314 mg,
0.79 mmol, 3.1 equiv.) in thf (5 mL) a 2.9 M solution of n-BuLi 1.39 mmol, 3.1 equiv.) in thf (5 mL) a 2.9 M solution of n-BuLi
in hexane (0.28 mL, 0.78 mmol, 3.0 equiv.) is added dropwise in hexane (0.46 mL, 1.35 mmol, 3.0 equiv.) is added dropwise at
at 0 °C in a separate reactor. The reaction mixture is stirred for 0 °C in a separate reactor. The reaction mixture is stirred for
30 min at 0 °C and further 30 min at ambient temperature. All 30 min at 0 °C and further 30 min at ambient temperature. All
volatiles were removed under fine vacuum. The residue was volatiles were removed under fine. The obtained colourless
washed with pentane and dried in fine vacuum. The obtained powder is directly dissolved in thf (10 mL) and added slowly to
colourless powder is directly dissolved in thf (10 mL) and the solution of PPN[SnCl3] in thf (10 mL) at −78 °C. The reaction
slowly added to the solution of Ph4P [SnCl3] in thf (10 mL) at mixture is stirred for 18 h within the cooling bath and allowed
−78 °C. The reaction mixture is stirred for 18 h within the to slowly warm up to ambient temperature at the end of this
cooling bath and allowed to slowly warm up to ambient temp- period. All volatiles are removed in fine vacuum and the residue
erature at the end of this period. All volatiles are removed in is extracted with Et2O (60 mL). After filtration all volatiles were
fine vacuum and the residue is extracted with Et2O (60 mL). removed from the obtained filtrate in fine vacuum. The residue
Half of the filtrate was kept at 0 °C, layered with the same was washed with pentane and dried in fine vacuum. 2b-PPN is
volume of pentane and this mixture was stored at −30 °C for obtained as a yellow-reddish oil (406 mg, 0.36 mmol, 81%).
several days to yield X-ray diffractive single crystals of 2a-Ph4P 1H-NMR (300.3 MHz, thf[d8]): δ = 7.72–7.55 (m, 30 H, PPN+),
(ca. 30 mg). The crystalline product is too insoluble for multi- 0.34 (s, 27 H, [Sn(SSiMe3)3]−) ppm. 13C{1H} NMR (75.5 MHz, thf
nuclar NMR measurements and thermally too unstable at elev- [d8]): δ = 134.5 (s, ((HC(CH)2(CH)2C)3P)2N+), 133.1 (m, ((HC
ated temperature in solvents, that do not decompose it. In (CH)2(CH)2C)3P)2N+), 130.1 (m, ((HC(CH)2(CH)2C)3P)2N+), 128.0
some attempts to force it into solution, spontaneous precipi- (dd, 1JCP = 108.6 Hz, 2JCN = 2.0 Hz, ((HC(CH)2(CH)2C)3P)2N+), 6.0
tation of yellow material was observed. Therefore, no satisfying (s, [Sn(SeSiMe3)3]−) ppm. 29Si NMR (59.7 MHz, thf[d8]): δ = 4.3 (s
2
elemental analyses could be obtained.
with superimposed d, JSiSn = 44.2 Hz, [Sn(SeSiMe3)3]−) ppm.
77Se NMR (57.3 MHz, thf[d8]): δ = −263.0 (s, [Sn(SeSiMe3)3]−)
ppm. 119Sn-NMR (112.0 MHz, thf[d8]): δ = 297.4 (s with superim-
Preparation of PPN[Sn(SSiMe3)3] (2a-PPN)
A mixture of SnCl2 (97 mg, 0.51 mmol, 1.0 equiv.) and PPN[Cl] posed d, 1JSnSe = 808.6 Hz, [Sn(SeSiMe3)3]−) ppm. Elemental ana-
(295 mg, 0.51 mmol, 1.0 equiv.) in thf (10 mL) is stirred for lysis exp. (calc. C45H57NP2Se3Si3Sn): C 49.2 (48.5), H 5.5 (5.2),
18 h at room temperature. To a solution of S(SiMe3)2 (284 mg, N 1.9 (1.3). According to proton NMR results, the slightly too
1.59 mmol, 3.1 equiv.) in thf (5 mL) a 2.9 M solution of n-BuLi high C and H content arises from the presence of silicon grease
in hexane (0.53 mL, 1.54 mmol, 3.0 equiv.) is added dropwise traces, which could not be fully extracted, as 2b-PPN is an oil.
at 0 °C in a separate reactor. The reaction mixture is stirred for
Preparation of [Sn(SSiMe3)4] (3)
30 min at 0 °C and further 30 min at ambient temperature. All
volatiles were removed under fine vacuum with the help of A 2.43 M solution of n-BuLi in hexane (0.61 mL, 1.48 mmol,
diethyl ether and pentane. The obtained colourless powder is 4.0 equiv.) is given dropwise into a solution of S(SiMe3)2
directly dissolved in thf (10 mL) and slowly added to the solution (271 mg, 1.52 mmol, 4.1 equiv.) in thf (5 mL) at 0 °C. The solu-
of PPN [SnCl3] in thf (10 mL) at −78 °C. The reaction mixture is tion is stirred for 30 min at 0 °C and another 30 min at
stirred for 18 h within the cooling bath and allowed to slowly ambient temperature. All volatiles were removed at 40 °C/10−3
warm up to ambient temperature at the end of this period. All mbar. After digesting the residue in diethyl ether/pentane and
volatiles are removed in fine vacuum and the residue is extracted removing all volatiles, the obtained colourless powder is
with Et2O (60 mL). After filtration all volatiles were removed from directly dissolved in toluene (10 mL) and added dropwise to a
the obtained filtrate in fine vacuum. The residue was washed solution of SnCl4 (97 mg, 0.37 mmol, 1.0 equiv.) in toluene
with pentane and dried in fine vacuum. 2a-PPN is obtained as a (10 mL) at −78 °C. The reaction mixture is stirred within the
colourless solid (320 mg, 0.32 mmol, 64%). 1H NMR (300.3 MHz, cooling bath for 18 h, thereby allowed to slowly reach ambient
thf[d8]): δ = 7.73–7.60 (m, 30 H, PPN+), 0.24 (s, 27 H, [Sn temperature. The suspension is filtered and all volatiles are
(SSiMe3)3]−) ppm. 13C{1H} NMR (75.5 MHz, thf[d8]): δ = 134.5 (s, removed from the filtrate at 15 °C/10−4 mbar. 3 is obtained as
((HC(CH)2(CH)2C)3P)2N+), 133.1 (m, ((HC(CH)2(CH)2C)3P)2N+), colourless waxy solid (178 mg, 0.33 mmol, 89%) of remarkable
1
130.1 (m, ((HC(CH)2(CH)2C)3P)2N+), 128.0 (dd, JCP = 107.2 Hz, sensitivity towards oxygen, moisture and slightly elevated
Dalton Trans.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Dalton Transactions
Paper
temperatures. 1H NMR (300.3 MHz, C6D6): δ = 0.55 (s, 36 H, Sn
(SSiMe3)4) ppm. 13C{1H} NMR (75.5 MHz, C6D6): δ = 4.8 (t,
1JCSi = 8.3 Hz, Sn(SSiMe3)4) ppm. 29Si NMR (59.7 MHz, C6D6):
Notes and references
1 M. W. DeGroot and J. F. Corrigan, Z. Anorg. Allg. Chem.,
2006, 632, 19–29.
2
δ = 21.6 (s with superimposed d, JSiSn = 32.8 Hz, Sn(SSiMe3)4)
ppm. 119Sn NMR (112.0 MHz, C6D6): δ = 50.5 (s with superim-
posed d, 2JSnSi = 33.6 Hz, Sn(SSiMe3)4) ppm. Elemental analysis
exp. (calc. C12H36S4Si4Sn): C 25.7 (26.7), H 6.2 (6.7), S 21.4
(23.8). Low values of C, H and S are due to beginning partial
loss of (Me3Si)2S even at room temperature.
2 (a) M. W. DeGroot and J. F. Corrigan, Angew. Chem., Int.
Ed., 2004, 43, 5355–5357; (b) M. W. DeGroot, N. J. Taylor
and J. F. Corrigan, J. Am. Chem. Soc., 2003, 125, 864–865.
3 C. B. Khadka, A. Eichhöfer, F. Weigend and J. F. Corrigan,
Inorg. Chem., 2012, 51, 2747–2756.
4 M. W. DeGroot, C. Khadka, H. Rösner and J. F. Corrigan,
J. Cluster Sci., 2006, 17, 97–110.
Preparation of [Sn(SeSiMe3)4] (4)
5 A. M. Polgar, F. Weigend, A. Zhang, M. J. Stillman and
J. F. Corrigan, J. Am. Chem. Soc., 2017, 139, 14045–14048.
6 (a) C. B. Khadka, D. G. Macdonald, Y. Lan, A. K. Powell,
D. Fenske and J. F. Corrigan, Inorg. Chem., 2010, 49, 7289–
7297; (b) J. Guschlbauer, T. Vollgraff and J. Sundermeyer,
Inorg. Chem., 2019, 58, 15385–15392.
7 X. Liu, Y. Feng, H. Cui, F. Liu, X. Hao, G. Conibeer,
D. B. Mitzi and M. Green, Prog. Photovolt.: Res. Appl., 2016,
24, 879–898.
8 X. Cui, K. Sun, J. Huang, J. S. Yun, C.-Y. Lee, C. Yan,
H. Sun, Y. Zhang, C. Xue, K. Eder, L. Yang, J. M. Cairney,
J. Seidel, N. J. Ekins-Daukes, M. Green, B. Hoex and
X. Hao, Energy Environ. Sci., 2019, 12, 2751–2764.
9 W. Wang, M. T. Winkler, O. Gunawan, T. Gokmen,
T. K. Todorov, Y. Zhu and D. B. Mitzi, Adv. Energy Mater.,
2014, 4, 1301465.
10 W. Sun, X. Geng, J. C. Armstrong, J. Cui and T.-p. Chen, in
2014 IEEE 40th Photovoltaic Specialist Conference (PVSC),
IEEE, 2014, pp. 421–424.
A 2.43 M solution of n-BuLi in hexane (0.45 mL, 1.10 mmol,
4.0 equiv.) is given dropwise into a solution of Se(SiMe3)2
(254 mg, 1.12 mmol, 4.1 equiv.) in thf (5 mL) at 0 °C. The solu-
tion is stirred for 30 min at 0 °C and another 30 min at
ambient temperature. All volatiles were removed at 40 °C/10−3
mbar. After digesting the residue in diethyl ether/pentane and
removing all volatiles, the obtained colourless powder is
directly dissolved in toluene (10 mL) and added dropwise to a
solution of SnCl4 (72 mg, 0.27 mmol, 1.0 equiv.) in toluene
(10 mL) at −78 °C. The reaction mixture is stirred within the
cooling bath for 18 h, thereby allowed to slowly reach ambient
temperature. The suspension is filtered and all volatiles were
removed from the filtrate at 15 °C/10−4 mbar. 4 is obtained as
pale yellowish waxy solid (166 mg, 0.23 mmol, 83%) of remark-
able sensitivity towards oxygen, moisture and slightly elevated
temperatures. 1H-NMR (500.2 MHz, C6D6): δ = 0.63 (s, 36 H, Sn
(SeSiMe3)4) ppm. 13C{1H} NMR (125.8 MHz, C6D6): δ = 5.0 (s
1
with superimposed d, JCSi = 14.0 Hz, Sn(SeSiMe3)4) ppm. 29Si-
NMR (99.3 MHz, C6D6): δ = 17.7 (s with superimposed d,
2JSiSn = 33.8 Hz, Sn(SeSiMe3)4) ppm. 77Se NMR (95.4 MHz,
11 M. Kumar, A. Dubey, N. Adhikari, S. Venkatesan and
Q. Qiao, Energy Environ. Sci., 2015, 8, 3134–3159.
12 D. Fuhrmann, S. Dietrich and H. Krautscheid, Chem. – Eur.
J., 2017, 23, 3338–3346.
1
C6D6): δ = −114.6 (s with two superimposed d, JSeSn
=
1496.9 Hz & 1565.1 Hz, Sn(SeSiMe3)4) ppm. 119Sn NMR
(186.5 MHz, C6D6): δ = 50.5 (s with superimposed d, JSnSi
1
=
13 D. Fuhrmann, S. Dietrich and H. Krautscheid, Inorg.
Chem., 2017, 56, 13123–13131.
14 I. E. Medina-Ramirez, M. J. Fink and J. P. Donahue, Acta
Crystallogr., Sect. C: Cryst. Struct. Commun., 2009, 65,
m475–m477.
1570.3 Hz, Sn(SeSiMe3)4) ppm. Due to its even more meta-
stable and waxy nature compared to 3, no satisfying elemental
analysis could be obtained.
Preparation of Cu2ZnSnS4 and reference PXRD data
15 F. Meyer-Wegner, M. Bolte and H.-W. Lerner, Inorg. Chem.
Commun., 2013, 29, 134–137.
See ESI.†
16 M. Kotsch, C. Gienger, C. Schrenk and A. Schnepf, Z.
Anorg. Allg. Chem., 2016, 642, 670–675.
17 M. W. DeGroot and J. F. Corrigan, Organometallics, 2005,
24, 3378–3385.
Conflicts of interest
There are no conflicts to declare.
18 (a) I. L. Abrahams, C. D. Garner and W. Clegg, J. Chem. Soc.,
Dalton Trans., 1987, 6, 1577–1579; (b) W. P. Chung, J. C. Dewan
and M. A. Walters, Inorg. Chem., 1991, 30, 4280–4282;
(c) M. Gelinsky and H. Vahrenkamp, Z. Anorg. Allg. Chem.,
2002, 628, 1017–1021; (d) Y. Matsunaga, K. Fujisawa, N. Ibi,
N. Amir, Y. Miyashita and K.-I. Okamoto, Bull. Chem. Soc. Jpn.,
2005, 78, 1285–1287; (e) A. D. Watson, C. P. Rao, J. R. Dorfman
and R. H. Holm, Inorg. Chem., 1985, 24, 2820–2826.
Acknowledgements
Financial support of the German Research Foundation DFG
and its priority program SPP 1708: “Material Synthesis near
Room Temperature” is gratefully acknowledged. We thank
Melanie Förster for recording the Raman spectra and the
service facilities of the chemistry department of the Philipps 19 X.-Y. Tang, R.-X. Yuan, J.-X. Chen, W. Zhao, A.-X. Zheng,
Universität Marburg for NMR spectroscopy, elemental analysis
and XRD for measurements and fruitful discussion.
M. Yu, H.-X. Li, Z.-G. Ren and J.-P. Lang, Dalton Trans.,
2012, 41, 6162–6172.
This journal is © The Royal Society of Chemistry 2020
Dalton Trans.

View Article Online
Paper
Dalton Transactions
20 (a) M. Kloskowska, A. Konitz, W. Wojnowski and B. Becker,
Z. Anorg. Allg. Chem., 2006, 632, 2424–2428;
(b) A. L. Seligson and J. Arnold, J. Am. Chem. Soc., 1993,
115, 8214–8220.
(b) M. Dimitrievska, F. Boero, A. P. Litvinchuk, S. Delsante,
G. Borzone, A. Perez-Rodriguez and V. Izquierdo-Roca,
Inorg. Chem., 2017, 56, 3467–3474.
28 T. J. Curphey, Phosphorus, Sulfur Silicon Relat. Elem., 2001,
173, 123–142.
21 X. Yu, A. Ren, F. Wang, C. Wang, J. Zhang, W. Wang, L. Wu,
W. Li, G. Zeng and L. Feng, Int. J. Photoenergy, 2014, 29 D. Taher, A. I. Wallbank, E. A. Turner, H. L. Cuthbert and
861249. J. F. Corrigan, Eur. J. Inorg. Chem., 2006, 22, 4616–4620.
22 S. Kumar, V. Kumar, V. Mikli, T. Varema, M. Altosaar and 30 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Adv., 2015,
M. Grossberg, Energy Procedia, 2016, 102, 136–143. 71, 3–8.
23 H. Lange, U. Herzog and G. Roewer, Organosilicon 31 G. M. Sheldrick, ta Crystallogr., Sect. C: Struct. Chem., 2015,
Chemistry V, Wiley-VCH GmbH & Co. KGaA, Weinheim,
2003, pp. 288–293.
24 T. Shimada, F. S. Ohuchi and B. A. Parkinson, J. Vac. Sci.
Technol., A, 1992, 10, 539–542.
25 M. Himmrich and H. Haeuseler, Spectrochim. Acta, 1991,
47A, 933–942.
26 W. L. F. Armarego and D. D. Perrin, Purification of labora-
tory chemicals, Butterworth-Heinemann, Oxford, 4th edn,
paperback edn, 1997.
71, 3–8.
32 A. L. Spek, Acta Crystallogr., Sect. D: Biol. Crystallogr., 2009,
65, 148–155.
33 H. Putz and K. Brandenburg, Diamond, Crystal Impact
GbR, Bonn, 2012.
34 (a) L. H. Finger, F. Wohde, E. I. Grigoryev, A.-K. Hansmann,
R. Berger, B. Roling and J. Sundermeyer, Chem. Commun.,
2015, 51, 16169–16172; (b) L. H. Finger and
J. Sundermeyer, Chem. – Eur. J., 2016, 22, 4218–4230;
(c) J. Guschlbauer, T. Vollgraff and J. Sundermeyer, Dalton
Trans., 2019, 48, 10971–10978.
27 (a) M. S. Weininger, G. W. Hunt and E. L. Amma, J. Chem.
Soc.,
Chem.
Commun.,
1972,
20,
1140–1141;
Dalton Trans.
This journal is © The Royal Society of Chemistry 2020