Synthesis and characterization of diacylgermanes: persistent derivatives with superior photoreactivity

Article abstract of DOI:10.1039/d1dt02091a

Acylgermanes are known as highly efficient photoinitiators. In this contribution, we present the synthesis of new diacylgermanes4a-eviaa multiple silyl abstraction methodology. The method outperforms the state-of-the-art approach (Corey-Seebach reaction)

Full text of DOI:10.1039/d1dt02091a

Dalton
Transactions
View Article Online
View Journal
PAPER
Synthesis and characterization of diacylgermanes:
persistent derivatives with superior
photoreactivity†
Cite this: DOI: 10.1039/d1dt02091a
Sabrina D. Püschmann,a Philipp Frühwirt,b Stefanie M. Müller,b Stefan H. Wagner,c
Ana Torvisco,a Roland C. Fischer, a Anne-Marie Kelterer, b Thomas Griesser,c
a
Georg Gescheidt b and Michael Haas
*
Acylgermanes are known as highly efficient photoinitiators. In this contribution, we present the synthesis
of new diacylgermanes 4a–e via a multiple silyl abstraction methodology. The method outperforms the
state-of-the-art approach (Corey–Seebach reaction) towards diacylgermanes in terms of group tolerance
and toxicity of reagents. Moreover, these compounds are decorated with bulky mesityl groups in order to
improve their storage stability. The isolated diacylgermanes were characterized by multinuclear NMR-,
UV-Vis spectroscopy and X-ray crystallography, as well as photolysis experiments (photobleaching) and
photo-DSC measurements (photopolymerization behavior). Upon irradiation with an LED emitting at
385 nm, all compounds except for 4a and 4c bleach efficiently with quantum yields above 0.6. Due to
their broad absorption bands, the compounds can be also bleached with blue light (470 nm), where
especially 4e bleaches more efficiently than Ivocerin®.
Received 23rd June 2021,
Accepted 5th August 2021
DOI: 10.1039/d1dt02091a
rsc.li/dalton
this PI (Chart 1).11 Another disadvantage is the inefficient curing
Introduction
depth at wavelengths above 500 nm.7 Our group has introduced
State-of-the-art visible light initiators for photo-induced free a new one-pot synthetic protocol providing tetraacylgermanes Ge
radical polymerization are used for a variety of different appli- [C(O)R]4 (R = aryl) in high yields (Chart 1).8 However, most
cations such as 3D-printing,1 tissue engineering,2 coatings3 implemented acylgermanes as initiators have stability issues in
and dental applications.4 The vast majority of these state-of- various solvents limiting the field of applications.
the-art initiators decompose into radical species via Norrish
Here, we introduce an easy to perform synthetic pathway
type I (α-cleavage). Modern Norrish type I initiators absorb towards diacylgermanes 4a–f with bulky mesityl groups, avoid-
light above 400 nm e.g. acylphosphines,5 acylstannanes6 and ing the Corey–Seebach reaction. The introduction of these
acylgermanes.7–10
mesityl groups leads to an increased stability in comparison to
Especially the latter were intensively investigated in the last the state-of-the-art germanium based photoinitiators. Moreover,
two decades, which boosted the number of available synthetic the broad absorption band of these compounds lead to a good
pathways towards acylgermanes significantly. However, di(4- applicability in the field of photopolymerization.
methoxybenzoyl)diethylgermane (Ivocerin®),
a commercially
available diacylgermane, is synthesized via a complex procedure,
which relies on a Corey–Seebach reaction followed by column
chromatography and consequently results in the high costs of
Results and discussion
Synthetic procedures
aInstitute of Inorganic Chemistry, Technical University Graz, Stremayrgasse 9/IV,
8010 Graz, Austria. E-mail: michael.haas@tugraz.at
The entry into this chemistry is provided by the easy to
perform synthetic protocol towards dimesityldi(trimethylsilyl)
bInstitute of Physical and Theoretical Chemistry, Technical University Graz,
Stremayrgasse 9/II, 8010 Graz, Austria
cInstitute of Chemistry of Polymeric Materials, Montanuniversitaet Leoben, Otto-
Gloeckelstrasse 2, A-8700 Leoben, Austria
†Electronic supplementary information (ESI) available: Experimental procedures
and characterization data for all new compounds, full details of computational
studies. Crystal data, details of data collections and refinements. CCDC
2078935–2078939. For ESI and crystallographic data in CIF or other electronic
format see DOI: 10.1039/d1dt02091a
Chart 1 The state-of-the-art germanium-based photoinitiators.
This journal is © The Royal Society of Chemistry 2021
Dalton Trans.

View Article Online
Paper
Dalton Transactions
germane 2. Subsequently 2 was reacted with equimolar
amounts of KOtBu and 18-crown-6 generating the germanide 3
as crucial intermediate. On the one hand the addition of
18-crown-6 is necessary to stabilize the anion and on the other
hand the usage of crown ether circumvents the introduction of
sulfur protecting groups and therefore is more sustainable.
Analytical and spectroscopic data clearly support the structure
of 3 (for details consult the Experimental section and the
ESI†). This germanide was added to the respective twofold
excess of acid fluoride in situ yielding the desired diacylger-
manes 4a–e in good yields (see Scheme 1). Analytical and spec-
troscopic data that support the structural assignment are given
in the Experimental section, together with experimental
details. Particularly striking is the possible synthesis and iso-
lation of compound 4e, where a benzothiophene group is sub-
stituted at the carbonyl moiety, as the introduction of hetero-
cyclic groups is not possible with the Corey–Seebach reaction.
Compound 4f was also formed with this protocol, unfortu-
nately due to the instability with silica gel it was not isolable.
Fig. 1 Absorption spectrum of synthesized compounds 4a–e com-
pared with the commercially available Ivocerin®, tetra(o-toluoyl)
germane (5) and tetra(mesitoyl)-germane (6) with a concentration of 1 ×
10−3 M in chloroform.
UV-Vis spectroscopy
The broad absorption band of compounds 4a–e are centered at
around 410 nm (see Fig. 1), which can be allocated to the n–π*
transition. This band is responsible for the photo-induced
cleavage of the Ge–C bond. In comparison to Ivocerin® these
new diacylgermanes show broader absorption bands, which
results in a bathochromic shift of their absorption edge.
Consequently, 4b–e absorb light above 450 nm whereas the
commercial PI, Ivocerin®, has a weak performance at this wave-
length. In comparison to tetra(o-toluoyl)germane 5 and tetra
(mesitoyl)germane 6 as representative examples of tetraacylger-
manes, 4a–e have similar absorption edges, however lower
extinction coefficients, which is due to the presence of only two
chromophore groups at the diacylgermanes, compared to the
four chromophores in the case of tetraacylgermanes.
DFT calculation
Based on the X-ray data, we have optimized the geometries of
4a–4e by DFT. Basically, the calculated geometries correspond
with the experimental data. For 4a, an additional, more stable
conformer (Erel = 8.2 kJ mol−1) was found. However, the elec-
tronic transitions calculated for the latter isomer were essen-
tially identical with those computed for that corresponding to
the X-ray structure. The vertical excitation energies (TDDFT)
and are presented in the ESI (Tables S3 and S4†).
The TDDFT calculated excitation spectra of 4a–e agree well
with the experimental UV-Vis spectra (Fig. 2). The first absorp-
tion band at ca. 420 nm consists of two vertical excitations,
HOMO → LUMO and HOMO → LUMO+1. Both transitions are
n–π* transitions. The HOMO is localized on both oxygen lone
pairs of the CvO groups. LUMO and LUMO+1 are both linear
combinations of the CvO π-orbitals with contributions of the
neighboring aromatic system. The influence of substitution on
orbital energies and localization can be explained by consider-
ing the electronic donating (EDG) effects of the substituents.
The methyl group is inductive electronic donating and the
methoxy group is mesomeric electronic donating.
Consequently, the presence of EDG groups (Me in 4a and 4c
and OMe in 4d) increase the electron density of the HOMO
orbital on the aromatic systems. Additionally, both EDG
Scheme 1 Synthetic protocol towards diacylgermanes 4a–f. a Yield
determined by 1H NMR analysis. b Isolated yield.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
Fig. 3 (a) Relevant orbitals for the first two vertical excitations of 4b. (b)
Relevant orbitals for the first two vertical excitations of 4e.
In particular, the dihedral angle between the aromatic sub-
stituent and the adjacent acyl group directs the electronic pro-
perties of the diacylgermanes. The bulky mesityl group in 4a
causes an almost perpendicular arrangement of the CvO
group and the π-plane of the mesityl substituent (dihedral
angle = 75.8°) whereas the π-system and the CvO group are
almost coplanar in 4e (3.28°) and 4c (21.3°) > 4d (16.36°) > 4b
(12.4°) being “intermediate” cases.
Consequently, in 4a, the electron density is almost entirely
localized at the acyl group (Fig. S34†). Solely, HOMO−1 indi-
cates distinguishable coefficients at one mesityl group. For 4c–
4e with smaller dihedral angles delocalization is more pro-
nounced rationalizing the red shifts in the UV-VIS spectra.
Photobleaching
Efficient curing of photoinitiator/monomer mixtures is crucial
for applications like dental restoration, therefore steady-state
photolysis experiments were performed to assess the photo-
bleaching behavior of the compounds. To that end, degassed
solutions of 4a–e in a 1/1 (v/v) mixture of toluene/methyl meth-
acrylate (with absorbance of about 0.7 at 385 nm) were irra-
diated with two different low power LEDs with emission
Fig. 2 Simulated spectrum of 4a–e together with the vertical tran-
sitions (vertical lines).
groups reduce the HOMO–LUMO-gap slightly compared to the maxima at about 385 and 470 nm (LED385 and LED470;
unsubstituted 4b (Table S4†), and, thus, cause a bathochromic described in the Experimental section of the ESI† in more
shift (see Table S3† and Fig. 2).
detail) – emission wavelengths, which are also found in com-
The extension of the π system in 4e by the annelated thio- mercial dental lamps.12
phene also leads to a red-shift compared with 4a–d. For 4e the
Fig. 4 shows the time traces of the normalized absorbances
four calculated transitions between 411 and 447 nm are in when irradiating with LED385 or LED470. It can be seen that
good consent with the rather broad absorptions between ca. 4a and 4c bleach least efficiently, whereas 4d is almost com-
400 and 470 nm. Whereas the lines at 411, 431, and 447 nm parable to Ivocerin® or to 6. From an exponential fit of the
possess two dominating components contributions, that at concentration traces, the quantum yields of decomposition
443 is dominated by the HOMO–LUMO transition (Table S3†).
can be determined using the procedure by Stadler et al.13 The
Fig. 3 shows the relevant orbitals for the first two vertical compounds 4a and 4c feature the smallest quantum yields
excitations of 4b and 4e. For all other compounds the relevant (0.20 and 0.49, respectively), while the quantum yields of the
orbitals can be found in the ESI (Fig. S34†).
other compounds are above 0.62 (see Table 1). For compound
This journal is © The Royal Society of Chemistry 2021
Dalton Trans.

View Article Online
Paper
Dalton Transactions
4d a higher quantum yield comparable to the one of Ivocerin®
(ϕ = 0.8314) is obtained. We ascribe this to the electronic pro-
perties of the methoxy groups having a compatible effect when
residing in either o- or p-position of the benzoyl substituents.
Despite the fact, that 4c and 4e show nearly identical extinc-
tions above 450 nm (see Fig. 1), photobleaching of 4e with the
blue LED emitting at 470 nm is more efficient due to its
higher quantum yield of decomposition and even faster than
Ivocerin® (see Fig. 4b).
NMR spectroscopy
NMR spectra and detailed characterization of 4a–f are provided
in the ESI.† All compounds show very similar 13C chemical
shifts for the carbonyl carbon between 219.47 ppm and
240.67 ppm, which is characteristic for the carbonyl groups
directly linked to the germanium atom.
Storage stability tests
Low long-term stability in various solvents (such as chloro-
form, benzene or toluene) is a drawback of Ivocerin® and ter-
aacylgermanes limiting the field of applications. By introdu-
cing bulky mesityl groups at the germanium atom the storage
stability was drastically improved. UV-Vis spectroscopy as well
as NMR spectroscopy were used to determine the storage stabi-
lity of Ivocerin®, tetra(o-toluoyl)germane 5 and compounds
4a–e. For the UV-Vis method, all acylgermanes were measured
in chloroform with a concentration of 1 × 10−3 M and the
spectra were recorded once a week over a period of 21 days. To
confirm the stability in MMA, UV-Vis spectra of Ivocerin®,
reference compound 5, compound 4c and 4d were recorded
once a week over a period of 21 days with a concentration of 1
× 10−3 M. In case of the NMR method, the measurements were
performed in degassed CDCl3 and C6D6 (0.01 M in 0.5 mL).
Once a week over a period of 21 days a spectrum of those were
recorded. During these long-term tests, the NMR tubes, as well
as the brown glass volumetric flask for the UV-Vis measure-
ments were sealed with parafilm and stored in the dark. As
representative example, compound 4e, as well as the spectra of
Ivocerin® and tetra(o-toluoyl)germane for comparison, are
shown in Fig. 5 (UV-Vis data) and Fig. 8 (NMR data, left: in
C6D6, right: in CDCl3). The measurements in MMA are shown
in Fig. 6. All other spectra can be found in the ESI.†
Fig. 4 Steady-state photolysis of 4a–e with (a) LED385, (b) LED470 in
toluene/MMA (1/1 v/v). The absorbance traces are normalized to the
initial absorptions at the observation wavelengths (maxima of n/σ–π*
transitions; 4a: 387 nm, 4b: 413 nm, 4c: 413 nm, 4d: 399 nm, 4e:
419 nm). Data for Ivocerin® and tetramesitoylgermane (6) were taken
from ref. 10.
Table 1 Wavelength of n/σ–π* absorption maxima and extinction
coefficients (in toluene/MMA 1/1 (v/v)) and determined quantum yields
of 4a–e
Compound
λmax,exp [nm]
ε [M−1 cm−1] at λmax,exp
Φ (385 mm)
4a
4b
4c
4d
4e
387
413
413
399
419
903
510
624
495
606
0.20 0.01
0.68 0.02
0.49 0.02
0.83 0.02
0.62 0.01
Fig. 5 Comparison of the long-term stability in solution (1 × 10−3 M in chloroform) detected by UV-Vis spectroscopy. Zoomed in part of the UV-Vis
spectra of (a) compound 4e, (b) Ivocerin® and (c) 5.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
Fig. 6 Comparison of the long-term stability in solution (1 × 10−3 M in MMA) detected by UV-Vis spectroscopy. (a) 4c, (b) 4d and (c) Ivocerin®.
As shown in Fig. 5, the newly synthesized compound 4e is X-Ray crystallography
stable and even after 21 days no degradation is observed. In
Crystals suitable for single-crystal XRD were obtained for com-
contrast to that the absorption of Ivocerin® and 5 slightly
pounds 4a–e. Compound 4e is shown as representative
decrease (see Fig. 5b and c). Moreover, we could confirm by
example (Fig. 7), the other structures can be found in the ESI.†
The torsion angle between the CvO bond and the aromatic
plane vary a lot among the different substituents. The evalu-
our stability tests that all investigated diacylgermanes are
stable in MMA (Fig. 6a–c).
To conclude, the commercial PIs degrade in solution, in
ated data of the bond lengths are slightly elongated compared
particular in chloroform as well as in benzene, whereas the
to the average Ge–C bond (1.97 Å)15 and the average CvO
new synthesized compounds 4a–e are significantly more stable
and show no degradation. In monomers, like in MMA, diacyl-
bond (1.19 Å) (Table 3).16
germanes are stable over time.
Photo-DSC measurements
The same statement can be made after measuring the NMR
Photo-DSC is a versatile method to evaluate the performance
spectra of all compounds. The newly synthesized compounds
of PIs in polymerizable resins. One single measurement can
give information about the reaction kinetics (time to reach the
(e.g. 4e) do not show any degradation. Even after 21 days, the
spectra remain unaffected (see Fig. 8a and b). On the opposite,
maximum heat flow (tmax), maximum rate of polymerization
Ivocerin® and 5 already show degradation after 7 days. During
(Rp,max), time to reach 95% of final conversion (t95%)) and the
the course of this experiments the degradation product, the
aldehyde, get more and more dominant, until only the degra-
dation product is present (Table 2 and Fig. 8). Although the
acute toxicity of aldehydes is low, the exposure to aldehydes
causes irritation of the skin, eyes and mucous membranes of
the respiratory passage. As our new presented initiators 4a–e
do not show any formation of these degradation product, we
think that this is an important step forward in this research
field.
Table 2 Formed aldehyde [%] as degradation product during the stabi-
lity test monitored by 1H analysis of Ivocerin® and compound 5 after 7,
14 and 21 days in benzene and chloroform. In case of 4e 0% aldehyde is
formed over the measured time
Fig. 7 ORTEP representation of 4e. Thermal ellipsoids are depicted at
the 50% probability level. Hydrogen atoms are omitted for clarity. The
torsion angle (mean value) between the CvO group and the aromatic
Compound
Ivocerin®
Solvent
C6D6
Days
Aldehyde formation [%]
ring plane of the thiofuran group is 1.72°.
7
15
28
50
18
32
100
14
21
7
14
21
Table 3 Mean bond lengths d [Å] and torsion angles between the CvO
group and the aromatic ring plane [°] of compounds 4a–e
CDCl3
Compound
dGe–C
dCvO
∠OvC–R
5
C6D6
7
25
14
21
7
14
21
62
100
44
73
100
4a
4b
4c
4d
4e
2.060
2.041
2.044
2.033
2.032
1.213
1.219
1.217
1.217
1.223
68.45
7.95
25.54
13.19
1.72
CDCl3
This journal is © The Royal Society of Chemistry 2021
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Fig. 8 Representation of the stacked 1H NMR spectra (300 MHz) of the respective compound (0.01 M in 0.5 mL) in the corresponding degassed
solvent, left: in C6D6 (61.7 ppm H2O) right: in CDCl3 (26 ppm H2O). (a) shows compound 4e in C6D6, (b) compound 4e in CDCl3, (c) Ivocerin® in
C6D6, (d) Ivocerin® in CDCl3, (e) 5 in C6D6 and (f) 5 in CDCl3.
double bond conversion (DBC, calculated from the overall reac- However, measured at similar PCG concentration, they show a
tion enthalpy ΔH (peak area) and the theoretical heat of slightly slower turnover than Ivocerin®, but higher double
polymerization (ΔH0,p)).
bond conversion than both Ivocerin® and 5. This correlates to
The photopolymerization experiments were conducted in the quantum yields measured at 385 nm, where 4b and 4d
1,6-hexanediol diacrylate (HDDA) as model monomer system show the highest values (Fig. 9).
(for further details consult the ESI†). Besides the synthesized
PIs (i.e. 4a–4e), the polymerization behavior of Ivocerin® and 5
was determined as reference.
The PI performance was analyzed at (1) equal molar PI
(0.30 mol%) as well as (2) equal photo-cleavable group (PCG)
Experimental
General considerations
concentration (0.15 mol% for 5; 0.30 mol% for 4a–4e and All synthetic steps were performed under inert conditions
Ivocerin®), respectively. In general, the synthesized PIs provide using standard Schlenk techniques. Solvents were dried with a
kinetics and a DBC comparable to the reference compounds. column purification system.17 Commercial tetrachlorogermane
In more detail, 4b and 4d show a slightly faster polymerization (GeCl4), tert-butyllithium (t-BuLi), mesitylbromide (MesBr),
initiation than the other synthesized PIs and can be compared chloro(trimethyl)silane (SiMe3Cl), and KOtBu were used
in their reactivity with 5 at the same PI concentration. without further purification. The used acid fluorides were pro-
Dalton Trans.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
31.1 mmol, 1.00 eq.) was dissolved in THF (90 mL) and was
added over one hour. The reaction was stirred for further 2 h
at −70 °C. The cooling was removed and the reaction was
stirred over night at room temperature. After aqueous workup
with saturated NH4Cl, the phases were separated and the
aqueous phase was extracted three times with diethyl ether.
The organic layer was dried over Na2SO4, filtered and the
solvent was removed under reduced pressure. The crude
product was recrystallized from acetone at −30 °C to yield
11.5 g (81%).
1H NMR (300 MHz, CDCl3, ppm): δ 6.76 (s, 4H, Mes–H),
2.23 (s, 6H, Mes–pCH3), 2.17 (s, 12H, Mes–oCH3), 0.21 (s, 18H,
Si–(CH3)3).
Synthesis of dimesityldimesitoylgermane (4a)
Mes2Ge(SiMe3)2 2 (2.00 g, 4.37 mmol, 1.00 eq.), 18-crown-6
(1.27 g, 4.81 mmol, 1.10 eq.) and KOtBu (0.54 g, 4.81 mmol,
1.10 eq.) were dissolved 25 mL benzene and stirred for 1 h at
room temperature. A second flask was charged with 2,4,6-tri-
methylbenzoyl fluoride (1.45 g, 8.75 mmol, 2.00 eq.) in toluene
(10 mL) and cooled to 0 °C. The germanide 3 was added drop-
wise to the fluoride solution. The cooling was removed, and
the reaction mixture was allowed to warm to room temperature
and stirred overnight. After aqueous workup with saturated
NH4Cl, the phases were separated and the aqueous phase was
extracted three times with DCM. The organic layer was dried
over Na2SO4, filtered with silica gel to remove the crown ether,
and the solvent was removed under reduced pressure. The
crude product was recrystallized from pentane at −70 °C to
yield 1.51 g (57%) of pure slightly yellow crystals.
M.p. 183–186 °C; UV-Vis (chloroform): λ = 385 and 399 nm,
ε = 924 and 902 L mol−1 cm−1; IR: v[cm−1] = 1636, 1603 (m,
νCvO); elemental analysis (%) calcd for C38H44GeO2: C 75.39,
H 7.33; found: C 75.31, H 7.34; 1H NMR (300 MHz, CDCl3,
ppm): δ 6.59 (s, 4H, C(O)Mes–H), 6.51 (s, 4H, Mes–H), 2.12 (s,
6H, C(O)Mes–pCH3), 2.12 (s, 6H, Mes–pCH3), 1.98 (s, 12H, C(O)
Mes–oCH3), 1.91 (s, 12H, Mes–oCH3). 13C NMR (75 MHz,
CDCl3, ppm): δ 240.67 (GeCOMes), 143.99, 142.35, 138.88,
138.73, 135.25, 134.15, 129.32, 128.75, 128.69 (Aryl–C), 25.14
(Mes–oCH3), 21.17 (Mes–pCH3), 21.05 (Mes–pCH3), 19.68
(Mes–oCH3). HRMS: calcd for [C38H44GeO2]+ (M+): 606.2562.
Found: 606.3292.
Fig. 9 Photo-DSC (left) and conversion plots (right) for the photo-
polymerization of HDDA with 0.3 mol% PI (top) and equal PCG concen-
tration (bottom).
duced according to the corresponding literature.18 1H
(299.95 MHz) and 13C (75.43 MHz) NMR spectra were recorded
with a Varian INOVA 300 spectrometer in CDCl3 or C6D6 solu-
2
tion and were referenced versus TMS by using the internal H
lock signal of the solvent. UV-Vis spectra were recorded with
an Agilent Cary 60 UV-Vis spectrometer. Mass spectra were
acquired with a Q-TOF Premier from Waters, Manchester,
England. The original ESI source of the instrument was
replaced by a standard LIFDI source from Linden CMS, Weyhe,
Germany. IR spectra were recorded with a Brucker ALPHA.
Melting points were determined by a Stuart automatic melting
point (SMP50).
Synthesis of dichlorodimesitylgermane (1)
A 3.0 L flask was charged with MesBr (52 mL, 339 mmol, 2.00
eq.) and THF (800 mL). The solution was cooled to −78 °C.
400 mL of t-Buli (1.70 M, 678 mmol, 4.00 eq.) were added with
a dropping funnel and stirred for additional 30 minutes.
Afterwards GeCl4 (36.4 g, 170 mmol, 1.00 eq.) was dissolved in
small amounts of THF and were added to the reaction. After
warming up to room temperature, the reaction was stirred
overnight. The reaction mixture was transferred into a flask
and the solvent was removed under reduced pressure. The
solid was extracted with pentane and the solvent was again
removed under reduced pressure leading to white crystals
(24.4 g, 38%).
Synthesis of dimesityldibenzoylgermane (4b)
Compound 2 (4.00 g, 8.75 mmol, 1.00 eq.), 18-crown-6 (2.54 g,
9.62 mmol, 1.10 eq.) and KOtBu (1.08 g, 9.62 mmol, 1.10 eq.)
were dissolved in benzene (60 mL) and stirred for 1 h at room
temperature. The germanide 3 was added dropwise to a solu-
tion of benzoyl fluoride (1.90 mL, 17.5 mmol, 2.00 eq.) in
toluene (10 mL) at 0 °C. After heating up to room temperature,
the orange-red reaction solution was stirred overnight.
1H NMR (300 MHz, CDCl3, ppm): δ 6.81 (s, 4H, Mes–H),
2.28 (s, 12H, Mes–oCH3), 2.26 (s, 6H, Mes–pCH3).
Synthesis of dimesityldi(trimethylsilyl)germane (2)
A 500 mL 3-neck flask with a dropping funnel and a reflux con- Followed by aqueous workup with NH4Cl solution, extraction
denser was charged with THF (90 mL) and Li (1.70 g, with DCM, drying over Na2SO4 and filtration over silica gel to
249 mmol, 4.00 eq.), which was cooled to −70 °C. remove the excess of crown ether. The solvent was removed
Trimethylchlorosilane (13.3 mL, 104 mmol, 3.35 eq.) was under reduced pressure. The crude product was further puri-
added to the solution and afterwards compound 1 (11.9 g, fied by column chromatography (pentane/toluene 1 : 2) and
This journal is © The Royal Society of Chemistry 2021
Dalton Trans.

View Article Online
Paper
Dalton Transactions
recrystallized from pentane at −70 °C, yielding 2.50 g (55%) of from pentane at −70 °C, 4d was isolated as yellow crystals
yellow crystals.
(2.64 g, 52%).
M.p. 155–157 °C; UV-Vis (chloroform): λ = 411 and 437 (sh)
M.p. 201–203 °C; UV-Vis (chloroform): λ = 396, 418 (sh) and
nm, ε = 369 and 241 (sh) L mol−1 cm−1; IR: v[cm−1] = 1620, 440 (sh) nm, ε = 516, 429 (sh) and 265 (sh) L mol−1 cm−1; IR: v
1590, 1574 (m, νCvO); elemental analysis (%) calcd for [cm−1] = 1617, 1590 (m, νCvO); elemental analysis (%) calcd
1
C32H32GeO2: C 73.74, H 6.19; found: C 73.31, H 6.17; H NMR for C34H36GeO4: C 70.25, H 6.24; found: C 70.02, H 6.10; 1H
(300 MHz, CDCl3, ppm): δ 7.80–7.77 (d, J = 7.3 Hz, 4H, Ph–H), NMR (300 MHz, CDCl3, ppm): δ 7.58–7.55 (d, J = 7.7 Hz, 2H,
7.44–7.39 (t, J = 7.3 Hz, 2H, Ph–H), 7.31–7.28 (d, J = 7.7 Hz, 4H, Ph–H), 7.15–7.10 (t, J = 7.8 Hz, 2H, Ph–H), 6.82 (s, 4H, Mes–H),
Ph–H), 6.87 (s, 4H, Mes–H), 2.28 (s, 6H, Mes–pCH3), 2.20 (s, 6.79 (d, J = 7.7 Hz, 2H, Ph–H), 6.35–6.62 (d, J = 8.3 Hz, 2H, Ph–
12H, Mes–oCH3). 13C NMR (76 MHz, CDCl3, ppm): δ 228.15 H), 3.16 (s, 6H, OCH3), 2.26 (s, 12H, Mes–oCH3), 2.24 (s, 6H,
(GeCOPh), 143.59, 140.61, 139.45, 134.12, 133.04, 129.64, Mes–pCH3). 13C NMR (76 MHz, CDCl3): δ 222.52 (GeCO),
128.75, 128.35 (Aryl–C), 25.11 (Mes–oCH3), 21.15 (Mes–pCH3). 158.49, 144.05, 137.81, 135.71, 134.04, 128.72, 126.96, 120.43,
HRMS: calcd for [C32H32GeO2]+ (M+): 522.1622. Found: 110.29 (Aryl–C), 52.25 (OCH3), 24.28 (Mes–oCH3), 21.09 (Mes–
522.2216.
pCH3). HRMS: calcd for [C34H36GeO4]+ (M+): 582.1833. Found:
582.2579.
Synthesis of dimesityldi(o-toluoyl)germane (4c)
Synthesis of dimesityldibenzothiophenegermane (4e)
Compound 2 (4.00 g, 8.75 mmol, 1.00 eq.), 18-crown-6 (2.54 g,
9.62 mmol, 1.10 eq.) and KOtBu (1.08 g, 9.62 mmol, 1.10 eq.)
were dissolved in benzene (60 mL) and stirred for 1 h at room
temperature. A second flask was charged with o-toluoyl fluor-
ide (2.43 g, 17.5 mmol, 2.00 eq.) and dissolved in 10 mL
toluene. The solution was cooled to 0 °C. After complete
addition of the germanide 3 to this solution, aqueous workup
with 10% H2SO4 followed. The aqueous phase was extracted
three times with DCM, dried over Na2SO4, filtrated over silica
gel and the solvent was removed under reduced pressure. The
crude product was further purified by column chromatography
(pentane/toluene 2 : 1) and recrystallization from pentane at
−70 °C, pure yellow crystals were isolated (2.59 g, 54%).
M.p. 142 °C; UV-Vis (chloroform): λ = 415 nm, ε = 658 L
mol−1 cm−1; IR: v[cm−1] = 1630, 1598, 1563 (m, νCvO);
elemental analysis (%) calcd for C34H36GeO2: C 74.35, H 6.61;
found: C 74.19, H 6.45; 1H NMR (300 MHz, CDCl3, ppm): δ
7.68–7.66 (d, J = 7.6 Hz, 2H, Ph–H), 7.20–7.15 (t, J = 7.4 Hz, 2H,
Ph–H), 7.06–7.01 (t, J = 7.1 Hz, 4H, Ph–H), 6.86 (s, 4H, Mes–H),
2.31 (s, 6H, Ph–oCH3), 2.27 (s, 6H, Mes–pCH3), 2.25 (s, 12H,
Mes–oCH3). 13C NMR (76 MHz, CDCl3, ppm): δ 231.25 (GeCO),
143.66, 140.04, 139.22, 137.19, 134.59, 132.60, 132.44, 131.99,
129.61, 129.57, 125.20 (Aryl–C), 25.15 (Mes–oCH3), 25.02 (oTol–
CH3), 21.49 (Mes–pCH3). HRMS: calcd for [C34H36GeO2]+ (M+):
550.1927. Found: 550.2034.
A flask was charged with Mes2Ge(SiMe3)2 2 (1.00 g, 2.18 mmol,
1.00 eq.), 18-crown-6 (0.64 g, 2.40 mmol, 1.10 eq.) and KOtBu
(0.27 g, 2.40 mmol, 1.10 eq.). The compounds were dissolved
in benzene (20 mL) and stirred for 1 h at room temperature. In
a second flask, thiofuran fluoride (0.99 g, 5.47 mmol, 2.50 eq.)
was dissolved in 10 mL toluene and cooled to 0 °C. The germa-
nide 3 was added dropwise to the acid fluoride and was
allowed to warm up to room temperature. At this temperature,
the reaction was stirred overnight. Followed by aqueous
workup with saturated NH4Cl solution and the aqueous layer
was extracted with DCM. The combined organic phases were
dried over Na2SO4, filtered over silica gel and the solvent was
removed under reduced pressure. The crude product was
further purified by column chromatography (pentane/toluene
1 + 2) and recrystallization from pentane at −70 °C, yielding in
0.25 g (23%) of yellow crystals.
M.p. 171–174 °C; UV-Vis (chloroform): λ = 371 (sh), 414
(sh), 441 (sh) nm, ε = 1990 (sh), 636 (sh) and 403 (sh) L mol−1
cm−1; IR: v[cm−1] = 1602, 1591 (m, νCvO); elemental analysis
(%) calcd for C36H32GeO2S2: C 68.27, H 5.09, S 10.12; found: C
1
68.31, H 5.08, S 10.09; H NMR (300 MHz, CDCl3) δ 7.82–7.80
(d, J = 8.1 Hz, 2H, Aryl–H), 7.75 (s, 2H, Aryl–H), 7.71–7.68 (d, J
= 7.8 Hz, 2H, Aryl–H), 7.44–7.38 (t, J = 7.6 Hz, 2H, Aryl–H),
7.34–7.29 (t, J = 7.5 Hz, 2H, Aryl–H), 6.88 (s, 4H, Mes–H), 2.29
(s, 6H, Mes–pCH3), 2.23 (s, 12H, Mes–oCH3). 13C NMR
(76 MHz, CDCl3) δ 219.47 (GeCO), 148.34, 143.98, 142.22,
140.05, 139.37, 133.13, 129.89, 127.84, 126.89, 124.78, 123.14
Synthesis of dimesityldi(o-methoxy)germane (4d)
Compound 2 (4.00 g, 8.75 mmol, 1.00 eq.), 18-crown-6 (2.54 g, (Aryl–C), 25.21 (Mes–oCH3), 21.22 (Mes–pCH3). HRMS: calcd
9.62 mmol, 1.10 eq.) and KOtBu (1.08 g, 9.62 mmol, 1.10 eq.) for [C34H36GeO4]+ (M+): 634.1062. Found: 634.1927.
were dissolved in benzene (60 mL) and stirred for 1 h at room
Synthesis of dimesityldibenzofurangermane (4f)
temperature. In a second flask, o-methoxybenzoyl fluoride
(2.71 g, 17.5 mmol, 2.00 eq.) was dissolved in toluene (10 mL) Compound 2 (0.50 g, 1.09 mmol, 1.00 eq.) was dissolved in
and cooled to 0 °C. The germanide 3 was added dropwise to benzene (15 mL). KOtBu (0.14 g, 1.20 mmol, 1.10 eq.) and
the fluoride and after warming up to room temperature stirred 18-crown-6 (0.32 g, 1.20 mmol, 1.10 eq.) were added and the
overnight. Followed by aqueous workup with 10% H2SO4, reaction was stirred for 1 h at room temperature. A second
extraction with dichloromethane, drying over Na2SO4, filtration flask was charged with benzofuran fluoride (0.45 g,
over silica gel and removal of the solvent under reduced 2.73 mmol, 2.50 eq.) and was dissolved in toluene (9 mL). The
pressure. The crude product was further purified by column germanide 3 was added to the fluoride solution at 0 °C and
chromatography (pentane/toluene 1 : 3) and recrystallization after full addition the solution was allowed to warm up to
Dalton Trans.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
room temperature. At this temperature, the reaction was
stirred overnight. A portion was separated and the solvent was
Acknowledgements
removed, which was then dissolved in CDCl3. A NMR spectrum We gratefully acknowledge financial support from NAWI Graz
was measured. Followed by aqueous workup with saturated and FWF (Vienna, Austria) (project number P 32606-N). The
NH4Cl solution and the aqueous layer was extracted with authors thank Linden CMS GmbH, Germany for measuring
DCM. The combined organic phases were dried over Na2SO4, the mass spectra.
filtered and the solvent was removed under reduced pressure.
The purification by either crystallization (normal atmosphere
and inert atmosphere) or by column chromatography was not
Notes and references
successful. Therefore, clean 4f could not be isolated.
1
1H NMR (300 MHz, CDCl3) δ H NMR (300 MHz, CDCl3) δ
7.57–7.50 (dd, J = 17.7, 7.2 Hz, 4H, Aryl–H), 7.34–7.33 (m, 4H,
Aryl–H), 7.23–7.21 (d, J = 5.9 Hz, 2H, Aryl–H), 6.86 (s, 4H, Mes–
H), 2.34 (s, 6H, Mes–pCH3), 2.25 (s, 12H, Mes–oCH3).
1 (a) A. Bagheri and J. Jin, ACS Appl. Polym. Mater., 2019, 1,
593–611; (b) H.-B. Sun and S. Kawata, in NMR • 3D Analysis
• Photopolymerization, Springer Berlin Heidelberg, Berlin,
Heidelberg, 2004, vol. 170, pp. 169–273.
2 (a) T. Billiet, M. Vandenhaute, J. Schelfhout, S. van
Vlierberghe and P. Dubruel, Biomaterials, 2012, 33, 6020;
(b) K. T. Nguyen and J. L. West, Biomaterials, 2002, 23,
4307–4314.
Conclusions
3 I. V. Khudyakov, M. B. Purvis and N. J. Turro, in
Photoinitiated Polymerization, Am. Chem. Soc, 2003, vol. 847,
pp. 113–126.
4 (a) N. Moszner and U. Salz, Prog. Polym. Sci., 2001, 26, 535–
576; (b) A. Santini, I. T. Gallegos and C. M. Felix, Prim.
Dent. J., 2013, 2, 30–33.
5 (a) H. Grützmacher, J. Geier, D. Stein, T. Ott, H. Schönberg,
R. H. Sommerlade, S. Boulmaaz, J.-P. Wolf, P. Murer and
T. Ulrich, Chimia, 2008, 62, 18–22; (b) L. Gonsalvi and
M. Peruzzini, Angew. Chem., Int. Ed., 2012, 51, 7895–7897;
(c) W. Feuerstein, S. Höfener, W. Klopper, I. Lamparth,
N. Moszner, C. Barner-Kowollik and A.-N. Unterreiner,
ChemPhysChem, 2016, 17, 3460–3469; (d) R. Appel,
G. Haubrich and F. Knoch, Chem. Ber., 1984, 117, 2063–
2075.
To conclude, we were able to synthesize a variety of substituted
diacylgermanes 4a–e with bulky mesityl groups by the multiple
silyl abstraction methodology, avoiding the Corey–Seebach
reaction. Moreover, these compounds show excellent stabilities
in solvents, which was determined by long-term stability tests
in chloroform, benzene and MMA via UV-Vis spectroscopy and
in chloroform and benzene via NMR spectroscopy. The new
compounds show a broadening in their absorption bands with
absorption above 450 nm, which makes photobleaching with
blue light (470 nm) quite efficient. With 4e a tailing up to
490 nm was observed. All compounds show high quantum
yields above 0.5, the only exception is compound 4a. DSC-
measurements show good photopolymerization behavior for
all synthesized compounds with double bond conversions
comparable to or higher than Ivocerin®. Therefore, these new
derivatives can be implemented in a broad field of appli-
cations. Further studies to probe the scope of these initiators
are currently in progress.
6 (a) M. Haas, J. Radebner, A. Eibel, G. Gescheidt and
H. Stueger, Chem.
– Eur. J., 2018, 24, 8258–8267;
(b) M. Mitterbauer, P. Knaack, S. Naumov, M. Markovic,
A. Ovsianikov, N. Moszner and R. Liska, Angew. Chem., Int.
Ed., 2018, 57, 12146–12150.
7 B. Ganster, U. K. Fischer, N. Moszner and R. Liska,
Macromolecules, 2008, 41, 2394.
8 J. Radebner, A. Eibel, M. Leypold, C. Gorsche, L. Schuh,
R. Fischer, A. Torvisco, D. Neshchadin, R. Geier,
N. Moszner, R. Liska, G. Gescheidt, M. Haas and
H. Stueger, Angew. Chem., Int. Ed., 2017, 56, 3103.
9 (a) J. Radebner, M. Leypold, A. Eibel, J. Maier, L. Schuh,
A. Torvisco, R. Fischer, N. Moszner, G. Gescheidt,
H. Stueger and M. Haas, Organometallics, 2017, 36, 3624;
(b) J. Lalevée, X. Allonas and J. P. Fouassier, Chem. Phys.
Lett., 2009, 469, 298; (c) D. Neshchadin, A. Rosspeintner,
M. Griesser, B. Lang, S. Mosquera-Vazquez, E. Vauthey,
V. Gorelik, R. Liska, C. Hametner, B. Ganster, R. Saf,
N. Moszner and G. Gescheidt, J. Am. Chem. Soc., 2013, 135,
17314; (d) S. D. Püschmann, P. Frühwirt, M. Pillinger,
A. Knöchl, M. Mikusch, J. Radebner, A. Torvisco,
R. C. Fischer, N. Moszner, G. Gescheidt and M. Haas,
Chem. – Eur. J., 2021, 27, 3338–3347.
Author contributions
S. D. P. performed the synthesis towards the
diacylgermanes. P. F. performed the Steady-State Photolysis
and Determination of Quantum Yields. S. M. M. did the
Photo-DSC measurements. S. H. W. and A-M. K. performed the
DFT computations and analysed the results. A. T. and R. C. F.
measured the X-ray structures. S. D. P., S. M. M. and
P. F. jointly wrote the manuscript with help from T. G.,
G. G. and M. H. All authors discussed the results and com-
mented on the manuscript. M. H. provided supervision and
wrote the final version of the manuscript.
Conflicts of interest
There are no conflicts to declare.
This journal is © The Royal Society of Chemistry 2021
Dalton Trans.

View Article Online
Paper
Dalton Transactions
10 P. Frühwirt, A. Knoechl, M. Pillinger, S. M. Müller, 14 A. Eibel, J. Radebner, M. Haas, D. E. Fast, H. Freißmuth,
P.
T.
Wasdin,
R.
C.
Fischer,
J.
Radebner,
E. Stadler, P. Faschauner, A. Torvisco, I. Lamparth, N. Moszner,
A. Torvisco, N. Moszner, A.-M. Kelterer, T. Griesser,
H. Stueger and G. Gescheidt, Polym. Chem., 2018, 9, 38–47.
G. Gescheidt and M. Haas, Inorg. Chem., 2020, 59, 15204– 15 K. M. Baines and W. G. Stibbs, Coord. Chem. Rev., 1995,
15217. 145, 157–200.
11 N. Moszner, F. Zeuner, I. Lamparth and U. K. Fischer, 16 M. Palusiak, S. Simon and M. Solà, J. Org. Chem., 2006, 71,
Macromol. Mater. Eng., 2009, 294, 877. 5241–5248.
12 T. Völkel, Scientific Documentation Bluephase® Family – LED 17 A. B. Pangborn, M. A. Giardello, R. H. Grubbs, R. K. Rosen
for Every Use, Ivoclar Vivadent AG, Schaan, Liechtenstein,
2009.
13 E. Stadler, A. Eibel, D. Fast, H. Freißmuth, C. Holly,
M. Wiech, N. Moszner and G. Gescheidt, Photochem.
Photobiol. Sci., 2018, 17, 660–669.
and F. J. Timmers, Organometallics, 1996, 15, 1518–1520.
18 Methoden der organischen Chemie, ed. B. Baasner,
J. L. Adcock, J. Houben, E. Müller, T. Weyl, H. Kropf,
K. H. Büchel and O. Bayer, Thieme, Stuttgart, 4th edn,
1998, vol 10a, pp. 552–554.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2021