Synthesis of 1,2-Dioxetanes as Thermochemiluminescent Labels for Ultrasensitive Bioassays: Rational Prediction of Olefin Photooxygenation Outcome by Using a Chemometric Approach

Article abstract of DOI:10.1002/chem.201603765

Great interest in new thermochemiluminescent (TCL) molecules, for example, in bioanalytical assays, has prompted the design and synthesis of a small library of more than 30 olefins to be subjected to photooxygenation, with the aim of obtaining new 1,2-dioxetane-based TCL labels with optimized properties. Fluorine atoms on the acridan system remarkably stabilize 1,2-dioxetanes when they are located in the 3- and/or 6-position (4 h and 4 i). On the other hand, 2,7-difluorinated acridan dioxetane (4 j) showed a significantly enhanced fluorescence quantum yield with respect to the unsubstituted dioxetane (4 a). Some of the synthesized olefins did not undergo singlet oxygen addition and a rationale was sought to ease the photooxygenation step, leading to the TCL dioxetanes. A chemometric approach has been adopted to exploit principal component analysis and linear discriminant analysis of the structural and electronic molecular descriptors obtained by DFT optimizations of olefins 3. This approach allows the steric and electronic parameters that govern dioxetane formation to be revealed.

Full text of DOI:10.1002/chem.201603765

DOI: 10.1002/chem.201603765
Full Paper
&
Substituent Effects
Synthesis of 1,2-Dioxetanes as Thermochemiluminescent Labels
for Ultrasensitive Bioassays: Rational Prediction of Olefin
Photooxygenation Outcome by Using a Chemometric Approach
Luca A. Andronico, Arianna Quintavalla,* Marco Lombardo,* Mara Mirasoli,
Massimo Guardigli, Claudio Trombini, and Aldo Roda[a]
Abstract: Great interest in new thermochemiluminescent
(TCL) molecules, for example, in bioanalytical assays, has
prompted the design and synthesis of a small library of
more than 30 olefins to be subjected to photooxygenation,
with the aim of obtaining new 1,2-dioxetane-based TCL
labels with optimized properties. Fluorine atoms on the acri-
dan system remarkably stabilize 1,2-dioxetanes when they
are located in the 3- and/or 6-position (4h and 4i). On the
other hand, 2,7-difluorinated acridan dioxetane (4j) showed
a significantly enhanced fluorescence quantum yield with re-
spect to the unsubstituted dioxetane (4a). Some of the syn-
thesized olefins did not undergo singlet oxygen addition
and a rationale was sought to ease the photooxygenation
step, leading to the TCL dioxetanes. A chemometric ap-
proach has been adopted to exploit principal component
analysis and linear discriminant analysis of the structural and
electronic molecular descriptors obtained by DFT optimiza-
tions of olefins 3. This approach allows the steric and elec-
tronic parameters that govern dioxetane formation to be re-
vealed.
Introduction
vantage of a thermochemiluminescence (TCL) label is that its
light emission is simply produced by heating, and nothing
else. In fact, the TCL phenomenon originates from heating
a suitable thermodynamically relatively unstable molecule,
which decomposes to yield a product mainly in the singlet ex-
cited state that decays with photon emission (see Scheme 1).
Nowadays, TCL[9] offers challenging and unexplored opportuni-
ties for the development of reagentless and ultrasensitive de-
tection methods exploitable, for example, in simple portable
biosensors with TCL molecules as labels, as either single mole-
cules or included in functionalized silica nanoparticles (SiN-
Ps).[10a,b]
From the 1970s to date, the 1,2-dioxetane system[1] has contin-
ued to attract great interest due to its key role in chemilumi-
nescence[2] (CL) and bioluminescence[3] (BL) reactions. Its pecu-
liar properties have stimulated and enabled the development
of a wide array of applications. For example, recently, 1,2-diox-
etanes have been exploited as luminescent mechanophores,
namely, responsive units able to transduce mechanical stress
into an optical response.[4] Primarily, dioxetane analogues have
proved to be potent tools in clinical diagnostics as chemilumi-
nescent substrates for enzymes, such as alkaline phosphatase,
used as labels in immunoassays and gene assays.[5] Indeed,
CL,[6] BL,[7] and electrogenerated chemiluminescence (ECL)[8] de-
tection techniques are particularly well suited in applications in
which high sensitivity is required, offering high detectability,
even in low volumes, a wide linear range of responses, and
a high signal to noise ratio. Nevertheless, CL, BL, and ECL de-
tection requires the addition of reagents to induce light emis-
sion, thus decreasing assay rapidity and simplicity, which are
crucial for on-field biosensor applications. The remarkable ad-
Scheme 1. Acridan-containing 1,2-dioxetane 4a, which was employed by us
as a TCL label.
[a] L. A. Andronico, Dr. A. Quintavalla, Prof. M. Lombardo, Prof. M. Mirasoli,
Prof. M. Guardigli, Prof. C. Trombini, Prof. A. Roda
Department of Chemistry “G. Ciamician”
Alma Mater Studiorum, University of Bologna
Via Selmi 2, 40126 Bologna (Italy)
Thermochemiluminescent 1,2-dioxetanes, described for the
first time in 1963, were proposed for analytical applications in
the late 1980s with limited success due to the relatively high
temperature of decomposition and very poor light emission ef-
ficiency.[11] Over the last four years, we have reinvestigated acri-
dan-based 1,2-dioxetanes (for example 4a;[12] Scheme 1) as TCL
compounds. As an application, we incorporated 4a into SiNPs
E-mail: arianna.quintavalla@unibo.it
marco.lombardo@unibo.it
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201603765.
Chem. Eur. J. 2016, 22, 1 – 14
1
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
with the aim of generating TCL probes for ultrasensitive immu-
noassays.[10] SiNPs doped with our acridan-based 1,2-dioxetane
displayed remarkable advantages, compared to TCL labels pro-
posed in the past,[11] such as a lower trigger temperature
(below 1008C) and highly improved detectability, comparable
with that obtained with enzyme-based CL detection.
1,2-dioxetane formation through photooxygenation of the cor-
responding alkene strongly depended on the substitution pat-
tern of the acridan moiety.
Intrigued by the considerable effects of the acridan substitu-
ents on the properties of thermochemiluminescent 1,2-dioxe-
tanes, we planned substantial structural modifications to the
fluorophore moiety of the dioxetane with the purpose of
1) improving and modulating the photophysical characteristics
and activation parameters of the TCL labels, and 2) studying
the impact of several structural features on a series of proper-
ties of the 1,2-dioxetanes, including the photooxygenation rate
of the parent olefin with singlet oxygen (1O2). The complete
list of thermochemiluminescent 1,2-dioxetanes synthesized
and investigated is shown in Figure 1. All members of this li-
brary were fully characterized in terms of TCL properties.
The synthetic approach to thermochemiluminescent 1,2-di-
oxetanes 4 is depicted in Scheme 2. It consisted of two steps:
a) reductive coupling of two properly selected ketones (1 and
2) under McMurry conditions,[13] and b) photooxygenation[14] of
the obtained tetrasubstituted alkene 3 to provide the desired
1,2-dioxetane 4.
In a preliminary stage, we studied the effect of weak elec-
tron-donating groups (EDGs), such as methyl substituents, on
the acridan moiety of 4a (4b–4g; Figure 1).[10c] For example,
we found that, with acridones as the emitting species, tri- and
tetramethyl-substituted acridones showed the highest fluores-
cence quantum yields (fF), in the range of 0.48–0.52. Accord-
ingly, the corresponding 1,2-dioxetanes 4 f and 4g (Figure 1)
presented limit of detection (LOD) values more than one order
of magnitude lower than that of the unsubstituted derivative
4a, as determined by TCL imaging experiments. Moreover, we
noticed that the impact of the substituent was greater in the
2-position (and/or 7-position). Methyl groups also caused
a clear decrease in the activation energy (Ea) of the thermoche-
miluminescent reaction. Lastly, we observed that the rate of
Scheme 2. Two-step synthetic approach to TCL 1,2-dioxetanes 4.
In a number of cases examined, however, photooxygenation
did not occur. Thus, we decided to perform a chemometric in-
vestigation with the aim of correlating selected representative
structural and electronic features of the olefin reagent with the
success of the photooxygenation reaction.
Results and Discussion
Synthesis
At the outset of this study, we optimized the reaction condi-
tions of the two key steps that led to parent 1,2-dioxetane
4a,[10,12] which was selected as a model compound (see
Table S1 in the Supporting Information). The optimized proce-
dure is detailed in Table 1. The most relevant modifications
concerning the photooxygenation step were the replacement
of polymer-bound Bengal Rose with Methylene Blue as the
sensitizer and an increased temperature; these conditions al-
lowed us to achieve better yields in a shorter time.
By using the optimized synthetic protocol, we accomplished
a series of modifications on the fluorophoric moiety of 4a
(Table 1), preserving unaltered the adamantyl unit that acts as
a stabilizing framework.[15]
Three different structural modifications were examined on
parent molecule 4a. First, we decorated the acridone ring with
fluorine substituents, as electron-withdrawing groups (EWGs),
Figure 1. Library of 1,2-dioxetanes studied herein.
&
&
Chem. Eur. J. 2016, 22, 1 – 14
www.chemeurj.org
2
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Table 1. Synthesis of differently substituted 1,2-dioxetanes 4a, 4h–t.[a]
Entry
1
Ketone 2
Olefin 3 (yield [%])[b]
3a (94)
1,2-Dioxetane 4 (time [h], yield [%])[b]
4a (2, 92)
2
3
4
5
6
7
8
3h (86)[c]
4h (2, 95)
4i (7, 55)[d]
4j (4, 89)
3i (88)
3j (76)
3k (48)[e]
4k (4, 89)
4l (2, 80)
3l (89)
3m (95)[c]
3n (95)
4m (12, 80)[f]
4n (2, 91)[f]
9
3o (92)[c]
4o (2, 85)
10
3p (93)[c]
4p (2, 87)[f]
11
3q (91)[c]
3r (90)
3
4q (2, 92)
12
4r (5, 91)
Chem. Eur. J. 2016, 22, 1 – 14
www.chemeurj.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Table 1. (Continued)
Entry
Ketone 2
Olefin 3 (yield [%])[b]
3s (77)
1,2-Dioxetane 4 (time [h], yield [%])[b]
4s (7, 82)[f]
13
14
3t (58)
4t (10, 50)
[a] Reaction conditions: a) McMurry coupling: adamantanone 1 (0.28 mmol), ketone 2 (1 equiv), TiCl4 (6.1 equiv), Zn (13.5 equiv), THF (6 mL), reflux, 45 min;
b) photooxygenation: Methylene Blue (6.5 mol%), UV cutoff filter (l=550 nm), À208C, CH2Cl2 (1 mL for 10 mg of olefin 3), 500 W halogen lamp, O2 (1 atm,
balloon). [b] Determined after purification by flash chromatography. [c] Ketone 2 (3.2 equiv), 1.5 h. [d] À308C. [e] Poor solubility of the starting ketone 2k.
[f] À408C.
to obtain difluoro- or tetrafluoroacridan derivatives, as well as
products containing both fluorine and a methyl substituent as
EDG. McMurry olefination proceeded smoothly on acridones
2h–l, regardless of the substitution pattern, to provide the cor-
responding alkenes 3h–l in high yields (up to 89%).
Also, the photooxygenation step provided, in most cases,
good results, but the reactivity of various olefins showed some
differences. First, we observed that the formal [2+2] cycload-
dition of singlet oxygen to acridane-derived alkenes strongly
depended on steric hindrance. In fact, when a fluorine atom
occupied the 1-position on the aromatic system (Table 1,
entry 3), the reaction was significantly slower. On the other
hand, the presence of EWGs (two or four fluorine atoms;
Table 1, entries 2–4 and 5, respectively) or a combination of an
EWG and an EDG (Table 1, entry 6) did not affect 1,2-dioxetane
formation.
ting species in the TCL process (Scheme 1).[18] For comparison,
we also show data for previously reported 1,2-dioxetane 4a
and acridone 2a (Table 2, entry 1).[10]
When we measured the fluorescence quantum yields (fF) of
ketones 2a and 2h–t (Table 2), it appeared clear that the emis-
sion performances of the acridones (Table 2, entries 1–11) were
strongly affected by the substitution pattern. In particular,
compared to reference compound 2a (Table 2, entry 1), fluo-
rine atoms on the acridone aromatic rings significantly de-
creased the fF values when present in the 3- and/or 6-posi-
tions (Table 2, entry 2) or in the 1-position (Table 2, entry 3).
Conversely, the insertion of fluorines in 2,7-positions enhanced
the fluorescence quantum yield (2j cf. 2a, 2k cf. 2h). These
unexpected and intriguing results suggested that the emitting
performances of F-substituted acridane-based ketones were
mostly affected by the position rather than the electronic fea-
tures of the substituents on the aromatic rings. As a further
confirmation, by replacing a fluorine atom (EWG) in the 6-posi-
tion with a methyl group (weak EDG), we obtained only
a slight increase in the fF value (2l, entry 6, cf. 2h, entry 2,
Table 2).[19] The observed behavior also proved to be peculiar
in comparison with the fluorescence quantum yield trend pre-
viously recorded for (poly)methylated acridones.[10c] In the pres-
ence of EDGs, the fF values showed an increase for substitu-
tion at both the 2- and 3-positions. On the contrary, the inser-
tion of fluorine atoms generated the opposite effect, depend-
ing on their location on the aromatic system.[21] To the best of
our knowledge, very few examples are present in the literature
that describe the behavior of EWG-substituted 1,2-dioxe-
tanes,[22] and no studies have been reported on fluorine-bear-
ing aryl dioxetanes or concerning the dependence of 1,2-diox-
etanes properties on the EWG distribution on the acridan
system. In particular, opposite effects on fluorescence quantum
yield as a function of the substituent position have never been
observed.
In the second stage, we replaced the ethyl acetate on the ni-
trogen with alkyl or aryl substituents (Table 1, entries 7–11).
The corresponding 1,2-dioxetanes 4m–q were always accessi-
ble, but when the electron-donating effect was more pro-
nounced (Table 1, entries 7, 8, and 10) we were forced to de-
crease the photooxygenation temperature to À408C to obtain
good results, since at À208C partial degradation of the dioxe-
tane 4 to the corresponding ketones 1 and 2 occurred.
We also focused our attention on the nature of the endocy-
clic heteroatom and replaced the nitrogen atom with oxygen
(Table 1, entries 12 and 13). The oxygen-bearing xanthyl deriva-
tives 2r and 2s smoothly underwent McMurry olefination
(3r,s) and subsequent photooxygenation; this allowed us to
isolate the corresponding 1,2-dioxetanes 4r and 4s in high
yields after chromatographic purification.[16] The reaction with
singlet oxygen was also acceptable (Table 1, entry 14) when
a direct linkage between the two aromatic rings was created,
as in fluorenone-derived olefin 3t.[17]
After examining the photophysical properties of ketones
2h–l, we turned our attention to the activation parameters of
the thermal decomposition of the fluorinated 1,2-dioxetanes
4h–l.[23] The trend shown by acridane 1,2-dioxetanes contain-
ing EDGs consists of lowered activation energies (Ea) and pre-
Photophysical and TCL properties
The TCL properties of this series of 1,2-dioxetanes 4h–t were
investigated (Table 2), together with the photophysical proper-
ties of the corresponding ketones 2h–t, which were the emit-
&
&
Chem. Eur. J. 2016, 22, 1 – 14
www.chemeurj.org
4
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Table 2. Photophysical properties of ketones 2a and 2h–t and activation parameters for the thermal decomposition of the corresponding 1,2-dioxetanes
4a and 4h–t.
[b]
[b]
Entry
2
fF[a]
4
Ea [kcalmolÀ1
]
lnA [sÀ1
]
t1/2 [months][c]
1
2
3
4
5
6
7
8
2a
2h
2i
2j
2k
2l
2m
2n
2o
2p
2q
2r
0.11
0.01
4ꢁ10À3
0.65
0.11
0.04
0.61
0.44
0.28
0.43
4ꢁ10À4
1ꢁ10À4
2ꢁ10À4
0.025
4a
4h
4i
4j
4k
4l
4m
4n
4o
4p
4q
4r
31.5Æ0.8
41.7Æ4.4
36.1Æ4.3
27.9Æ1.2
25.8Æ1.1
35.1Æ2.7
30.9Æ1.5
23.7Æ1.6
27.7Æ0.7
26.2Æ2.4
32.0Æ1.4
46.4Æ3.4
27.0Æ1.1
33.8Æ1.4
35.6Æ1.0
48.0Æ5.6
41.8Æ5.6
30.6Æ1.5
28.6Æ1.4
42.9Æ3.5
37.3Æ2.0
27.2Æ2.2
31.0Æ0.9
29.5Æ3.3
37.4Æ1.9
58.0Æ4.7
29.5Æ1.5
40.8Æ1.9
11
1400 (117 years)
55
3.9
1.0
3.4
0.76
0.097
1.9
0.66
4.4
180 (15 years)
2.6
3.1
9
10
11
12
13
14
2s
2t
4s
4t
[a] Determined in acetonitrile by using quinine sulfate as standard (fF =0.53 in H2SO4 0.05 molLÀ1). [b] Mean ÆSD of three independent measurements.
[c] Calculated for the solid compound at 258C.[20]
exponential coefficients (A) and increased fluorescence quan-
tum yields (fF).[10c] We reasoned that the introduction of fluo-
rine atoms as EWGs should give more stable and easy to
handle 1,2-dioxetanes, without causing an excessive impair-
ment of the fluorescence quantum yield. The activation param-
eters recorded for the fluorinated 1,2-dioxetanes confirmed
our hypothesis, but they also revealed some unexpected re-
sults. The 3,6-difluorodioxetane 4h (Table 2, entry 2) was char-
(Table 2, entries 7 and 8, respectively). This finding further con-
firmed the stabilizing effect postulated for the acetate moiety,
due to its electron-withdrawing character.[10d,12a] Then, we in-
vestigated N-aryl derivatives (2o–q, 4o–q; Table 2, entries 9–
11) and observed, in general, a lower thermal stability of the
dioxetanes[27] with respect to the reference compound 4a,
even if the decrease was more pronounced for the EDG-substi-
tuted compound 4p than that for the para-fluorophenyl deriv-
ative 4q.
acterized by the highest activation energy (Ea =41.7 kcalmolÀ1
)
and showed an outstanding and unprecedented calculated
half-life (t1/2) of 117 years, if stored as a solid at 258C.[24] In-
creased stability compared with unsubstituted dioxetane 4a
was also recorded for compound 4i (Table 2, entry 3; Ea =
36.1 kcalmolÀ1, t1/2 =55 months), whereas the remaining fluori-
nated dioxetanes 4j–l (Table 2, entries 4–6) proved to be less
thermally stable. The results obtained from this family of com-
pounds confirmed that there was a correlation between acri-
done fluorescence quantum yield and thermal stability of the
corresponding 1,2-dioxetane. The substituents on the aromatic
system that enhance the fF value, such as fluorine atoms in
the 2,7-positions (2j–k), facilitate dioxetane thermal decompo-
sition (4j–k) at the same time. Conversely, 3,6-difluoro- (2h)
and 1,6-difluoro-substituted (2i) acridones showed decreased
fluorescence quantum yields, but the corresponding dioxe-
tanes 4h–i displayed significantly longer half-life values. A
comparison of these data with those recorded for methyl-sub-
stituted acridan-based dioxetanes[10c] confirmed that the intro-
duction of weak EDGs increased the ketone fluorescence quan-
tum yield[25] and lowered the dioxetane activation energy, re-
gardless of the position on the aromatic ring. On the contrary,
the impact of fluorine atoms depended dramatically on the
substituent position.
Thus, for acridane-based 1,2-dioxetanes, we conclude that
1) the introduction of fluorine atoms onto the aromatic rings
represents a remarkable tool to stabilize the dioxetane deriva-
tives, in particular, when they are located at the 3- and/or 6-
position, or to enhance their emitting performances, when lo-
cated at the 2,7-positions; 2) the acetate moiety proves to be
the best performing N-substituent, ensuring a good balance
between ketone fluorescence quantum yield and dioxetane
stability.
As far as the activation parameters for the thermal decom-
position of xanthyl-derived 1,2-dioxetanes (4r,s) and the pho-
tophysical properties of the corresponding emitting species 2r
and 2s are concerned, we observed that the replacement of
the endocyclic nitrogen with an oxygen provided a very stable
dioxetane (4r), but also strongly decreased the fluorescence
quantum yield (cf. Table 2, entries 1 and 12). Moreover, EDGs
on the xanthyl scaffold (3,6-positions) dramatically decreased
the activation parameters without a perceptible benefit in the
fF value (cf. Table 2, entries 12 and 13). Lastly, fluorenyl dioxe-
tane 4t, which featured a direct linkage between the aromatic
rings, afforded acceptable emitting properties (fF =0.025) ac-
companied by moderate thermal stability (Table 2, entry 14).
The second structural feature we studied was the nature of
the substituent on the acridan nitrogen (Table 2, entries 7–
11).[26] Removing (2m, 4m) or elongating (2n, 4n) the ester
group present in the parent compounds (2a, 4a) provided the
expected enhancement of fF and decrease in Ea values
Chemometric analyses
To widen our library of TCL 1,2-dioxetanes and to learn more
about the structure–property relationships of this class of com-
Chem. Eur. J. 2016, 22, 1 – 14
www.chemeurj.org
5
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Scheme 3. A representation of the various hypothesized mechanisms for the
formation of 1,2-dioxetanes through singlet oxygen addition to olefins.
sively on very simple systems.[31] Moreover, the prediction of
the mechanism of singlet oxygen reactions has proven a chal-
lenging task for many reasons,[32] including because different
possible reaction pathways (Scheme 3) are often predicted de-
pending on the level of theory employed.[33]
To rationalize all of our results, we decided to avoid an ex-
tremely demanding and time-consuming quantum mechanical
approach, considering that both excited and radical states are
involved in the reaction. Rather, we envisioned a chemometric
approach as a more practical tool to identify which structural
and electronic properties of starting olefins, or combination of
them, determine the success of the photooxygenation step.
At the outset, we adopted principal component analysis
(PCA)[34] of the structural and electronic molecular descriptors
obtained by DFT optimizations of olefins 3.[35] After an initial
conformational screening by using a molecular mechanics
(MM) force field on the data set of molecules, the conformers
obtained in a 8 kcalmolÀ1 window were optimized by using
DFT at the B3LYP/6-31G(d) level of theory (see the Supporting
Information for details). Among the different calculated elec-
tronic descriptors, we selected the energy of the HOMO, the
energy of the LUMO, the Mulliken charges on the olefin
carbon terminals Ca-Mul and C9-Mul (Figure 3) and the coeffi-
cients of the atomic contribution of the alkene carbon atoms
to the HOMO and LUMO orbitals, calculated by using the Mul-
liken decomposition method (Ca-HOMO, Ca-LUMO, C9-HOMO,
and C9-LUMO). Moreover, we selected a few structural parame-
ters that reflect the geometry of the molecules in close prox-
imity to the double-bond reaction site. In particular, we chose
Figure 2. Differently substituted alkenes unreactive under photooxygenation
conditions (the yields of olefin formation are given in parentheses; for syn-
thetic details, see the Supporting Information).
pounds, we also synthesized a series of olefins that, unexpect-
edly, did not undergo the photooxygenation step (Figure 2).
Previously investigated acridane-based alkenes 3u–x[10c] and
the 1-Me, 6-F derivative 3y are all characterized by the pres-
ence of a methyl group at the 1-position of the aromatic
system. Such results suggest that steric crowding in this region
disfavors the addition of singlet oxygen. The slow photooxyge-
nation rate observed for 1,6-difluoro compound 3i (Table 1,
entry 3) supports this hypothesis.
The sulfur-containing olefins 3z–B did not react with singlet
oxygen, regardless of the sulfur oxidation state.[28] We tested
other reaction temperatures (namely, 0 and À208C) and ex-
tended the reaction times up to 7 h, but only the starting
olefin was recovered. Finally, when we replaced the endocyclic
heteroatom with a sp3- or sp2-hybridized carbon (3C–E), in all
cases, photooxygenation did not occur.
These findings reveal that the reactivity of an olefin in the
photooxygenation process strongly depends on its substitu-
tion pattern. It is known that the formal [2+2] cycloaddition
of singlet oxygen to alkenes takes place efficiently on electron-
rich systems.[29] However, the mechanism of this process has
been extensively debated in the literature (Scheme 3),[30] and
the most reliable one can vary case by case, depending on the
molecular structure of the substrate,[30a,d–i,k] the solvation ener-
gy,[30a,e,i] and also the sensitizer employed.[30b,c,j,l] As a conse-
quence, the nature of the olefin substituents (electron-donat-
ing ability, presence of hydrogen-bond acceptors and/or
donors, geometry, and steric hindrance) not only affects the
success of singlet oxygen addition, but it also determines the
reaction pathway.
The mechanism of singlet oxygen addition to double bonds
has been studied computationally several times, but the rigor-
ous approach to this diradical system requires expensive treat-
ments of multireference states and it has been applied exclu-
Figure 3. DFT-optimized structure of 3a and structural molecular descriptors
employed in the PCA.
&
&
Chem. Eur. J. 2016, 22, 1 – 14
www.chemeurj.org
6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
the two dihedral angles, f1 and f2, which represent the devia-
tion from planarity of the aromatic rings with respect to the
plane defined by the carbon–carbon double bond, and the
four interatomic distances d1–d4 as an index of steric crowding
of the aromatic ring substituents around the CaÀC9 double
bond (Figure 3). Specifically, when a substituent was present in
the C1 position of the aromatic system, d2 and d4 were taken
as the interatomic distances between the double-bond carbon
atoms Ca and C9 and the closest atom of the substituent
group, as shown in Figure 3.
the plot (blue triangles) and the unreactive molecules clus-
tered on the left (red dots), mainly discriminated by the first
PC.
The loading plot, reported in Figure 5B with the variables
colored by relative contributions, shows that the variables
mainly composing the first PC are the Mulliken charge C9-Mul
(cos2 =0.834), the dihedral angles f1 and f2 (cos2 =0.808 and
0.807, respectively), the interatomic distances d2 and d4 (cos2 =
0.815 and 0.669, respectively), and, to a lesser extent, the in-
teratomic distance d1 (cos2 =0.541) and the Mulliken charge
Ca-Mul (cos2 =0.582). The only variables significantly contribu-
ting to the second PC are the interatomic distance d3 (cos2 =
0.641) and, to a lesser extent, the interatomic distance d1
(cos2 =0.413) and the energies of the HOMO and LUMO orbi-
tals (cos2 =0.461 and 0.483, respectively). PCA results seem to
suggest that the difference in reactivity in the photooxygena-
tion reaction between the two groups of alkenes is mainly due
to steric effects and not to electronic ones. In particular, the
presence of a bulky methyl group in the 1-position of the acri-
dan skeleton results in a significantly larger deviation from pla-
narity for compounds 3u–y, with respect to all other reactive
alkenes (see Table S3 in the Supporting Information). This geo-
metrical distortion has the overall effect of making the CaÀC9
double bond more sterically crowded and less accessible to
singlet oxygen, as also evidenced by the shorter interatomic
distances d1 and d2 in the unreactive compounds.
We analyzed first a set of 23 similar molecules that pos-
sessed a common acridane skeleton (3a–q, 3u–y, and 3F; Fig-
ure 4),[10c,12d] the molecular descriptors of which are collected
The same analysis was repeated on the expanded data set
comprising the oxygen-bearing xanthyl derivatives 3r–s (XAN
family), the sulfur-containing olefins 3z, 3A, and 3B (THI
family), and the anthrone-derived alkenes 3C–E (ANT family),
in addition to the previously analyzed acridane derivatives 3a–
q, 3u–y, and 3F (ACR family). The eigenvalues greater than
1 derived from PCA of the as-obtained 31ꢁ14 data matrix are
reported in Table 4, while the corresponding molecular descrip-
tors employed are collected in Table S4 in the Supporting In-
formation. By using the first three PCs, the model can be used
to explain 73.8% of the total variance, whereas 56% of the var-
iance is accounted for by using only the first two PCs.
Figure 4. Initial set of acridan-containing alkenes analyzed with the chemo-
metric approach.
in Table S3 in the Supporting Information. Data were mean-
centered before analysis and were autoscaled by dividing
them by the standard deviation of each column sample varia-
ble. The first four eigenvalues greater than 1 obtained by PCA
of the 23ꢁ14 data matrix and the explained variances relative
to each component are reported in Table 3. The model ob-
tained is able to explain 84.9% of the total variance by using
only three PCs, whereas two PCs explain 64.7% of the var-
iance.
The score plot representing the projections of the data
points in the plane defined by the two first PCs is reported in
Figure 6A and the corresponding loading plot with the varia-
bles colored by relative contributions is shown in Figure 6B.
Again, a quite marked separation in the score plot is appar-
ent between the reactive and unreactive olefins, which in this
expanded data set is discriminated by both the first and
second PCs.
The score plot representing the projections of the data
points in the plane defined by the two first PCs is reported in
Figure 5A. A clustering of data is apparent, with the alkenes
that undergo photooxygenation grouped in the right side of
Table 4. Eigenvalues and percentage of explained variance relative to the
first five PCs of the model used for the description of olefins 3a–s, 3u–z,
and 3A–F.
Table 3. Eigenvalues and percentage of explained variance relative to the
first four principal components (PCs) of the model used for the descrip-
tion of olefins 3a–q, 3u–y, and 3F.
PC
Eigenvalue
Variance
[%]
Cumulative
Eigenvalue
Cumulative
variance [%]
PC
Eigenvalue
Variance
[%]
Cumulative
Eigenvalue
Cumulative
variance [%]
PC1
PC2
PC3
PC4
PC5
4.45
3.39
2.49
1.42
1.09
31.8
24.2
17.8
10.1
7.8
4.45
7.84
10.33
11.75
12.84
31.8
56.0
73.8
83.9
91.6
PC1
PC2
PC3
PC4
5.96
3.11
2.82
1.16
42.5
22.2
20.1
8.3
5.96
9.07
11.89
13.05
42.5
64.7
84.9
93.1
Chem. Eur. J. 2016, 22, 1 – 14
www.chemeurj.org
7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Figure 5. PCA results for the data matrix descriptors of olefins 3a–q, 3u–y, and 3F: A) PCA score plot relative to the first two PCs (blue triangles: reactive
compounds; red dots: unreactive compounds). B) PCA loading plot relative to the first two PCs (variables are colored according to their relative contribution).
The relative loading plot (Figure 6B) reveals that the varia-
bles contributing to the first PC are mainly the steric parame-
ters, similarly to what is obtained for the previously analyzed
ACR family. In particular, the highest contributing variables are
mainly the dihedral angles f1 and f2 (cos2 =0.906 and 0.915,
respectively) and, to a lesser extent, the interatomic distances
d2 and d4 (cos2 =0.632 and 0.542, respectively) and the Mullik-
en charge Ca-Mul (cos2 =0.654). On the other hand, the varia-
bles significantly contributing to the second PC are mainly
electronic parameters, namely, the energies of the HOMO and
LUMO (cos2 =0.592 and 0.759, respectively) and, to a lesser
extent, the Mulliken charge C9-Mul (cos2 =0.475).
For this expanded data set, the PCA results suggest that
both steric and electronic effects play a role in determining
the reactivity under photooxygenation conditions; both factors
are necessary, but not sufficient to trigger reactivity. The score
plot of the data projections (Figure 6A) is apparently divided
into four quadrants, with the reactive olefins in the top-right
sector, where both steric and electronic parameters cooperate
in the same positive direction. The acridane-containing unreac-
tive olefins (3u–y) are positioned in the top-left quadrant, in
a region where electronic effects alone would allow the photo-
oxygention reaction, which is hampered solely by steric hin-
drance around the CaÀC9 double bond exerted by the methyl
Figure 6. PCA results for the data matrix descriptors of olefins 3a–s, 3u–z, and 3A–F: A) PCA score plot relative to the first two PCs (blue triangles: reactive
compounds; red dots: unreactive compounds); B) PCA loading plot relative to the first two PCs (variables are colored according to their relative contribution).
&
&
Chem. Eur. J. 2016, 22, 1 – 14
www.chemeurj.org
8
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
group present at the 1-position (121–1248 versus 1308 mean
f1/f2 values for all compounds). On the other hand, the an-
throne-derived alkenes 3C–E are grouped in the bottom-right
sector of the score plot, with the same coordinates of the reac-
tive olefins along the first PC. Although photooxygenation
should not be disallowed for steric reasons for these alkenes,
the overall linear combination of their electronic descriptors
renders them unreactive. The sulfur-containing olefins 3z and
3A are located in a borderline central region of the plot,
where both steric and electronic factors are not ideal for pro-
moting the photooxygention reaction, whereas the sulfone an-
alogue 3B is projected farther away along the negative PC2
axis. Finally, the reactive xanthyl derivatives 3r and 3s are cor-
rectly projected in the same top-right quadrant of the previ-
ously examined reactive acridane-based olefins, which possess
both steric and electronic properties that allow the photooxy-
genation reaction to proceed smoothly. The score plot of the
expanded data set relative to the first two PCs is again report-
ed in Figure 7, with the projected data points now colored ac-
cording to the common structural features of the employed al-
kenes (XAN: 3r,s; THI: 3z, 3A,B; ANT: 3C–E, and ACR: 3a–q,
3u–y, 3F). The marked clustering obtained in this plot clearly
shows the capability of the PCA model to discriminate be-
tween different alkenes on the basis of their common structur-
al features.
(mean of canonical variables=À11.8). The corresponding raw
and standardized coefficients of the obtained linear discrimi-
nant are reported in Table 5, whereas the canonical scores are
plotted in Figure 8 for all compounds.
It is interesting to note that, when LDA was performed by
using a stepwise forward variable selection, only 8 of the 14
variables were used to obtain a statistically significant model
(Wilks’ lambda=0.0191). Moreover, the most contributing vari-
ables were almost exactly the same as those that composed
the main PCs of the corresponding PCA, namely, the dihedral
angle f1, the interatomic distances d2 and d4, and the Mulliken
charges Ca-Mul and C9-Mul.
To verify the capability of the LDA model to predict the reac-
tivity of new compounds with singlet oxygen, the fluorenyl
Table 5. Raw and standardized coefficients of the linear discriminant ob-
tained in the LDA model used for the description of olefins 3a–s, 3u–z,
and 3A–F.
Variable
Raw coefficient
Standardized coefficient
HOMO
LUMO
Ca-Mul
C9-Mul
f1
f2
d1
d2
9.549
À4.564
865.7
3.176
À1.216
6.167
1.631
1.664
À3.620
À2.000
2.630
83.51
0.8260
À1.319
À84.28
32.63
To further validate our chemometric approach, a supervised
linear discriminant analysis (LDA)[34c,d] was performed on the
expanded 31ꢁ14 data set. Using the previously chosen param-
eters, we obtained a highly statistically significant model
(Wilks’ lambda=0.0121, canonical correlation=0.9939), show-
ing perfect discrimination between the two classes of reactive
(mean of canonical variables=6.48) and unreactive olefins
d3
d4
À3.254
À6.801
31.76
À40.83
23.65
154.1
À194.9
À0.05056
À0.3888
1.629
À1.979
0.7734
2.684
Ca-HOMO
C9-HOMO
Ca-LUMO
C9-LUMO
constant
–
Figure 7. PCA results for the data matrix descriptors of olefins 3a–q, 3u–y,
and 3F (ACR); 3r-s (XAN); 3z, 3A–B (THI); and 3C–E (ANT). PCA score plot
relative to the first two PCs, colored according to the structural features of
the analyzed alkenes.
Figure 8. Canonical scores relative to the LDA model for alkenes 3a–s, 3u–z
and 3A–F.
Chem. Eur. J. 2016, 22, 1 – 14
www.chemeurj.org
9
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
alkene 3t (central five-membered ring) and the specially de-
signed new olefin 3G (central seven-membered ring; Figure 9)
were projected onto the linear discriminant.
tained by DFT optimizations of olefins 3. This approach al-
lowed us to determine steric and electronic parameters that
govern dioxetane formation. Great interest in new thermoche-
milumiscent molecules, for example, in bioanalytical assays,
was the driving force of this study, which allowed us to discov-
er that fluorine atoms on the acridan system remarkably stabi-
lized 1,2-dioxetanes when located in the 3- and/or 6-position
(4h and 4i). On the other hand, 2,7-difluorinated acridane di-
oxetane 4j showed a significantly enhanced fluorescence
quantum yield with respect to the unsubstituted dioxetane
4a. Our investigations for the development of original and en-
hanced dioxetane-based TCL labels are still ongoing. For in-
stance, flavone-[36] and coumarin-containing[37] derivatives 4H
and 4I (Scheme 4) are of interest. Indeed, even if their synthe-
sis still requires some optimization, 1,2-dioxetanes 4H and 4I
have been isolated and characterized as TCL compounds.
Figure 9. Olefins selected to verify the capability of the model to predict re-
activity with singlet oxygen.
The fluorenyl-derived alkene 3t was predicted to be reactive
with a raw canonical score of 30.8; a particularly high value
due to an almost completely planar arrangement of the aro-
matic rings with respect to the plane defined by the carbon–
carbon double bond (f1, f2 =159.88). As already mentioned in
Table 1, entry 14, when 3t was subjected to the photooxyge-
nation conditions, the corresponding 1,2-dioxetane 4t was iso-
lated in 50% yield overall. On the contrary, 5-dibenzosubere-
none-derived alkene 3G was predicted to be unreactive with
a raw canonical score of À39.9, again a particularly low value,
mainly due to its significantly smaller dihedral angles f1 and f2
(113.38) and to the greater negative Mulliken charge on C9
(À0.165 versus À0.09 mean C9-Mul charge). When subjected
to the photooxygenation conditions, olefin 3G did not react
and only the starting material was observed in the reaction
mixture.
Scheme 4. Preliminary results concerning flavone- (4H) and coumarin-based
(4I) 1,2-dioxetanes.
The experimental findings obtained for the supplementary
alkenes 3t and 3G validate the developed chemometric
model, which is able to anticipate the outcome of a photooxy-
genation reaction carried out on structurally different tetrasub-
stituted olefins, as characterized by an adamantyl unit coupled
to a tricyclic aromatic scaffold (regardless of the nature of the
central ring).
Unexpectedly high values of the activation parameters for
thermal decomposition were recorded for flavone-derived 1,2-
dioxetane 4H, which resulted in a remarkable calculated half-
life at 258C (more than 6000 years); this offers the opportunity
of useful applications in diagnostics and bioanalysis. On the
other hand, coumarin-based 1,2-dioxetane 4I showed limited
thermal stability. An analogous chemometric approach will be
reported in due course for flavones and coumarins once a sig-
nificant number of compounds has been synthesized.
Conclusion
A small library of more than 30 olefins of general structure 3
have been subjected to photooxygenation, with the aim of
synthesizing new dioxetane-based thermochemiluminescent
labels with optimized light emission efficiency and CL tempera-
ture triggering. Among them, 20 starting olefins provided the
corresponding 1,2-dioxetanes, whereas no traces of product
were observed with the other members of the series. To antici-
pate the feasibility of the photooxygenation step leading to
new potentially thermochemiluminescent dioxetanes, we
sought a rationale. A complete DFT analysis of the transition
states for the formation of all dioxetanes was not possible for
us, given the heavy and time-consuming computational effort
required by the multiconfigurational treatment of both closed-
shell and free-radical excited states at a rigorous level. Thus,
we adopted a chemometric approach by exploiting PCA and
LDA of the structural and electronic molecular descriptors ob-
Experimental Section
Representative procedure for the synthesis of alkenes 3
Under a nitrogen atmosphere, TiCl4 (1m in dichloromethane,
6.1 equiv) was added to a suspension of zinc powder (13.5 equiv)
in anhydrous THF (8 mL/0.5 mmol of 1) at 08C, and the suspension
was stirred for 10 min under reflux. A solution of ketone 2
(1 equiv) and 2-adamantanone 1 (1 equiv) in dry THF (2 mL/
0.5 mmol of 1) was added dropwise over a period of 30 min. The
reaction mixture was heated at reflux for 45 min. Then, it was
cooled to room temperature, quenched with water, and extracted
with AcOEt (3ꢁ10 mL). The combined organic layers were dried
over sodium sulfate and evaporated under vacuum. The crude
product was purified by flash chromatography on silica gel.
&
&
Chem. Eur. J. 2016, 22, 1 – 14
www.chemeurj.org
10
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
h) R. C. Buxton, B. Edwards, R. R. Juo, J. C. Voyta, R. Bethell, Anal. Bio-
chem. 2000, 280, 291–300; i) S. Sabelle, P.-Y. Renard, K. Pecorella, S. de
Suzzoni-Dꢂzard, C. Crꢂminon, J. Grassi, C. Mioskowski, J. Am. Chem. Soc.
2002, 124, 4874–4880; j) B. Edwards, A. Sparks, J. C. Voyta, I. Bronstein,
in Bioluminescence and Chemiluminescence, Fundamentals and Applied
Aspects, (Eds.: A. K. Campbell, L. J. Kricka, P. E. Stanley), Wiley, Chichester,
1994, pp. 56–59; k) I. Bronstein, J. C. Voyta, G. H. G. Thorpe, L. J. Kricka,
G. Armstrong, Clin. Chem. 1989, 35, 1441–1446; l) G. H. G. Thorpe, I.
Bronstein, L. J. Kricka, B. Edwards, J. C. Voyta, Clin. Chem. 1989, 35,
2319–2321; m) I. Bronstein, J. C. Voyta, B. Edwards, Anal. Biochem.
1989, 180, 95–98; n) R. Tizard, R. L. Cate, K. L. Ramachandran, M. Wysk,
J. C. Voyta, O. J. Murphy, I. Bronstein, Proc. Natl. Acad. Sci. USA 1990, 87,
4514–4518.
Representative procedure for the synthesis of 1,2-dioxe-
tanes 4
Alkene 3 (16 equiv) and Methylene Blue (1 equiv) were dissolved in
CH2Cl2 (1 mL/10 mg of alkene 3). The solution was cooled (usually
at À208C) and subjected to an oxygen atmosphere (1 atm, bal-
loon). The solution was stirred at the same temperature under irra-
diation by using a 500 W halogen lamp equipped with an UV
cutoff filter (0.5% transmission at l=550 nm). Irradiation was con-
tinued until the starting material disappeared (usually 2 h of irradi-
ation), and the conversion was monitored by 1H NMR spectroscopy.
Product 4 was purified by rapid filtration on a 5 mm layer of silica
gel, by using cooled (À408C) CH2Cl2 as the eluent, and the filtered
solution was evaporated under vacuum at 08C.
[6] For selected recent reviews on CL applications, see: a) F. J. Lara, D.
Airado-Rodriguez, D. Moreno-Gonzalez, J. F. Huertas-Perez, A. M. Garcia-
Campana, Anal. Chim. Acta 2016, 913, 22–40; b) Y. X. Liu, X. Y. Huang,
J. C. Ren, Electrophoresis 2016, 37, 2–18; c) N. Li, D. Liu, R. Niessner,
Anal. Bioanal. Chem. 2014, 406, 5589–5612; d) N. Li, X. Su, Y. Lu, Analyst
2015, 140, 2916–2943; e) M. Iranifam, TrAC Trends Anal. Chem. 2014, 59,
156–183; f) J. F. Huertas-Pꢂrez, D. Moreno-Gonzalez, D. Airado-Rodri-
guez, F. J. Lara, A. M. Garcia-Campana, TrAC Trends Anal. Chem. 2016, 75,
35–48; g) A. Roda, M. Mirasoli, E. Michelini, M. Di Fusco, M. Zangheri, L.
Cevenini, B. Roda, P. Simoni, Biosens. Bioelectron. 2016, 76, 164–179;
h) J. Y. Park, L. J. Kricka, Anal. Bioanal. Chem. 2014, 406, 5631–5637; i) Y.
Dong, J. Cai, X. You, Y. Chi, Analyst 2015, 140, 7468–7486; j) M. Mirasoli,
M. Guardigli, E. Michelini, A. Roda, J. Pharm. Biomed. Anal. 2014, 87, 36–
52.
Acknowledgements
Acknowledgement for financial support is made to MIUR
(Rome), University of Bologna, and FARB Project (RFBO128121,
“Advanced ultrasensitive multiplex diagnostic systems based
on luminescence techniques”).
Keywords: fluorine
·
olefination
·
photooxygenation
·
[7] For selected recent reviews on BL applications, see: a) A. Romieu, Org.
Biomol. Chem. 2015, 13, 1294–1306; b) S. M. Coleman, A. McGregor,
Future Virol. 2015, 10, 169–183; c) T. T. Xu, D. Close, W. Handagama, E.
Marr, G. Sayler, S. Ripp, Front. Oncol. 2016, 6, 150; d) N. Wu, T. Rathnaya-
ka, Y. Kuroda, Biochimica et Biophysica Acta 2015, 1854, 1392–1399;
e) G. Y. Lomakina, Y. A. Modestova, N. N. Ugarova, Biochemistry (Moscow)
2015, 80, 701–713; f) M. Hattori, T. Ozawa, RSC Adv. 2015, 5, 12655–
12663; g) S. V. Markova, E. S. Vysotski, Biochemistry (Moscow) 2015, 80,
714–732; h) A. Roda, M. Guardigli, E. Michelini, M. Mirasoli, TrAC Trends
Anal. Chem. 2009, 29, 307–322; i) E. P. Coutant, Y. L. Janin, Chem. Eur. J.
2015, 21, 17158–17171. See also Refs [6e,j].
[8] For selected recent reviews on ECL applications, see: a) P. Bertoncello,
A. J. Stewart, L. Dennany, Anal. Bioanal. Chem. 2014, 406, 5573–5587;
b) E. H. Doeven, G. J. Barbante, C. F. Hogan, P. S. Francis, ChemPlusChem
2015, 80, 456–470; c) M. Hesari, Z. Mahdi, J. Electrochem. Soc. 2016,
163, H3116–H3131; d) S. E. K. Kirschbaum, A. J. Baeumner, Anal. Bioanal.
Chem. 2015, 407, 3911–3926; e) W.-W. Zhao, J. Wang, Y.-C. Zhu, J.-J. Xu,
H.-Y. Chen, Anal. Chem. 2015, 87, 9520–9531; f) Z. Liu, W. Qi, G. Xu,
Chem. Soc. Rev. 2015, 44, 3117–3142; g) H. Zhou, J. Liu, S. Zhang, TrAC
Trends Anal. Chem. 2015, 67, 56–73; h) C. Ding, W. Zhang, W. Wang, Y.
Chen, X. Li, TrAC Trends Anal. Chem. 2015, 65, 137–150; i) Y. H. Xu, J. Q.
Liu, C. L. Gao, E. K. Wang, Electrochem. Commun. 2014, 48, 151–154;
j) S. Zhang, Y. Ding, H. Wei, Molecules 2014, 19, 11933–11987. See also
Refs [6e,i,j].
[9] For examples of TCL applications, see: a) M. Gilbert, V. Zakharova, A. Ra-
menda, C. Jebsen, B. Schulze, C. Wilhelm, Tetrahedron 2012, 68, 6765–
6771; b) H. Kong, H. Wang, S. Zhang, X. Zhang, Analyst 2011, 136,
3643–3648; c) Z. Wiener-Megnazi, H. Shiloh, L. Avraham, S. Lahav-
Baratz, M. Koifman, A. Z. Reznick, R. Auslender, M. Dirnfeld, Fertil. Steril.
2011, 95, 979–984; d) H. Kong, D. Liu, S. Zhang, X. Zhang, Anal. Chem.
2011, 83, 1867–1870; e) O. Amir, C. Yamin, M. Sagiv, N. Eynon, S. Shniz-
er, T. Kagan, A. Z. Reznick, M. Sagiv, R. E. Amir, J. Sports Med. Phys. Fitness
2009, 49, 105–111; f) O. Amir, H. Paz, O. Rogowski, M. Barshai, M. Sagiv,
S. Shnizer, A. Z. Reznick, R. E. Amir, Clin. Cardiol. 2009, 32, 199–203; g) I.
Sukhotnik, V. Brod, M. Lurie, M. A. Rahat, S. Shnizer, N. Lahat, J. G. Mogil-
ner, H. Bitterman, Crit. Care Med. 2009, 37, 1054–1061; h) Z. Wiener-
Megnazi, L. Vardi, A. Lissak, S. Shnizer, A. Z. Reznick, D. Ishai, S. Lahav-
Baratz, H. Shiloh, M. Koifman, M. Dirnfeld, Fertil. Steril. 2004, 82, 1171–
1176; i) S. Shnizer, T. Kagan, A. Lanir, I. Maor, A. A. Reznick, Luminescence
2003, 18, 90–96; j) S. Shnizer, A. Resnik, A. Lanir, V. Piuk, PCT Int. Appl.
1999, WO 9919728A1 19990422; k) J. L. Kiel, J. L. Alls, E. A. Holwitt,
L. J. V. Stribling, J. E. Parker, Bioelectrochem. Bioenerg. 1998, 47, 253–
257.
substituent effects · thermochemiluminescence
[1] For reviews on 1,2-dioxetanes, see: a) P. D. Bartlett, M. E. Landis, in Sin-
glet Oxygen, (Eds.: H. H. Wasserman, R. W. Murray), Academic Press, New
York, 1979, pp. 243–286; b) W. Adam, in The Chemistry of Peroxide, (Ed.:
S. Patai), Wiley, New York, 1983, pp. 829–920; c) W. Adam, in Small Ring
Heterocycles, (Ed.: A. Hassner), Wiley, New York, 1986, pp. 351–429;
d) W. Adam, M. Heil, T. Mosandl, C. R. Saha-Mçller, in Organic Peroxides,
(Ed.: W. Ando), Wiley, New York, 1992, pp. 221–254; e) C. R. Saha-
Mçller, W. Adam, in Comprehensive Heterocyclic Chemistry II A Review of
the Literature 1982–1995, (Ed.: A. Padwa), Pergamon, New York, 1996,
pp. 1041–1082; f) W. Adam, A. V. Trofimov, in The Chemistry of Perox-
ides, (Ed.: Z. Rappoport), Wiley, New York, 2006, pp. 1171–1209.
[2] For reviews on 1,2-dioxetane-based CL, see: a) G. B. Schuster, Acc.
Chem. Res. 1979, 12, 366–373; b) M. Matsumoto, J. Photochem. Photo-
biol. C 2004, 5, 27–53; c) W. J. Baader, C. V. Stevani, E. L. Bastos, in The
Chemistry of Peroxides, (Ed.: Z. Rappoport), Wiley & Sons, Chichester,
2006, pp. 1211–1278; d) L. F. M. L. Ciscato, D. Weiss, R. Beckert, E. L.
Bastos, F. H. Bartoloni, W. J. Baader, New J. Chem. 2011, 35, 773–775;
e) A. P. Schaap, S. D. Gagnon, J. Am. Chem. Soc. 1982, 104, 3504–3506;
f) M. Matsumoto, N. Watanabe, Bull. Chem. Soc. Jpn. 2005, 78, 1899–
1920; g) N. J. Turro, P. Lechtken, N. E. Shore, G. Schuster, H.-C. Steinmet-
zer, A. Yekta, Acc. Chem. Res. 1974, 7, 97–105.
[3] For information on 1,2-dioxetanes in BL, see: a) E. S. Vysotski, J. Lee,
Acc. Chem. Res. 2004, 37, 405–415; b) O. Shimomura, T. Goto, F. H.
Johnson, Proc. Natl. Acad. Sci. USA 1977, 74, 2799–2802; c) J.-Y. Koo,
S. P. Schmidt, G. B. Schuster, Proc. Natl. Acad. Sci. USA 1978, 75, 30–33.
[4] a) Y. Chen, A. J. H. Spiering, S. Karthikeyan, G. W. M. Peters, E. W. Meijer,
R. P. Sijbesma, Nat. Chem. 2012, 4, 559–562; b) Z. S. Kean, J. L. Hawk, S.
Lin, X. Zhao, R. P. Sijbesma, S. L. Craig, Adv. Mater. 2014, 26, 6013–6018;
c) Y. Chen, R. P. Sijbesma, Macromolecules 2014, 47, 3797–3805.
[5] For chemiluminescent 1,2-dioxetanes as probes for bioanalysis, in par-
ticular, CL immunoassays, see: a) A. Mayer, S. Neuenhoefer, Angew.
Chem. Int. Ed. Engl. 1994, 33, 1044–1072; Angew. Chem. 1994, 106,
1097–1126; b) W. Adam, D. Reinhardt, C. R. Saha-Mçller, Analyst 1996,
121, 1527–1531; c) S. Beck, H. Koster, Anal. Chem. 1990, 62, 2258–2270;
d) J. C. Hummelen, T. M. Luider, H. Wynberg, in Complementary Immuno-
assays, (Ed.: W. P. Collins), John Wiley & Sons, Chichester, UK, 1988;
e) T. M. Luider, J. C. Hummelen, J. N. Koek, D. Oudman, H. Wynberg, in
Luminescence Immunoassay and Molecular Applications, (Ed.: K. Van
Dyke), CRC Press, Boca Raton, 1990; f) I. Bronstein, B. Edwards, J. C.
Voyta, J. Biolumin. Chemilumin. 1989, 4, 99–111; g) B. Edwards, A.
Sparks, J. C. Voyta, I. Bronstein, J. Biolumin. Chemilumin. 1990, 5, 1–4;
Chem. Eur. J. 2016, 22, 1 – 14
www.chemeurj.org
11
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
[10] a) M. Di Fusco, A. Quintavalla, M. Lombardo, M. Guardigli, M. Mirasoli, C.
Trombini, A. Roda, Anal. Bioanal. Chem. 2015, 407, 1567–1576; b) M. Di
Fusco, M. Guardigli, M. Lombardo, M. Mirasoli, A. Quintavalla, A. Roda,
C. Trombini, PCT Int. Appl. 2014, WO 2014024106A1 20140213; c) M. Di
Fusco, A. Quintavalla, C. Trombini, M. Lombardo, A. Roda, M. Guardigli,
M. Mirasoli, J. Org. Chem. 2013, 78, 11238–11246; d) A. Roda, M. Di Fus-
co, A. Quintavalla, M. Guardigli, M. Mirasoli, M. Lombardo, C. Trombini,
Anal. Chem. 2012, 84, 9913–9919; see also Ref. [6g].
[11] For 1,2-dioxetanes applied in TCL, see: a) T. Luider, J. C. Hummelen, J. N.
Koek, H. Wynberg, Eur. Pat. Appl., 1988, EP 261719A2 19880330; b) J. C.
Hummelen, T. M. Luider, H. Wynberg, Pure Appl. Chem. 1987, 59, 639–
650; c) J. C. Hummelen, T. M. Luider, H. Wynberg, Methods Enzymol.
1986, 133, 531–557; d) J. C. Hummelen, H. Wynberg, PCT Int. Appl.
1985, WO 8504958A1 19851107; e) T. M. Luider, J. C. Hummelen, J. N.
Koek, D. Oudman, H. Wynberg, in Luminescence immunoassay and mo-
lecular applications, (Eds.: K. van Dyke, R. van Dyke), CRC Press, Boca
Raton, 1990.
fs. [10c,12e,g]; for examples of other substituents employed as stabiliz-
ing subunits in 1,2-dioxetanes, see: e) M. Matsumoto, H. Suganuma, Y.
Katao, H. Mutoh, J. Chem. Soc. Chem. Commun. 1995, 431–432;
f) L. F. M. L. Ciscato, D. Weiss, R. Beckert, W. J. Baader, J. Photochem. Pho-
tobiol. A 2011, 218, 41–47; for a review concerning the medicinal
chemistry of adamantane derivatives, see: g) L. Wanka, K. Iqbal, P. R.
Schreiner, Chem. Rev. 2013, 113, 3516–3604.
[16] For xanthyl-containing 1,2-dioxetanes, see: a) A. C. Brouwer, J. C. Hum-
melen, T. M. Luider, F. van Bolhuis, H. Wynberg, Tetrahedron Lett. 1988,
29, 3137–3140; b) A. P. Schaap, M. D. Sandison, R. S. Handley, Tetrahe-
dron Lett. 1987, 28, 1159–1162.
[17] For fluorenyl-containing 1,2-dioxetanes, see: a) W. H. Richardson, S. A.
Thomson, J. Org. Chem. 1985, 50, 1803–1810; b) W. Adam, L. A. A. En-
carnaciꢃn, Chem. Ber. 1982, 115, 2592–2605.
[18] We confirmed the nature of the emitting species for methyl-substituted
acridan-1,2-dioxetanes (Ref. [10c]) by comparing the TCL spectra of the
1,2-dioxetanes with the fluorescence spectra of the corresponding acri-
dones. In all cases, the TCL spectra closely matched the fluorescence
spectra of the corresponding acridones; thus demonstrating that the
TCL emission was primarily due to the singlet excited state of acridone.
The singlet excited state of 2-adamantanone could also be produced in
the TCL process. However, its possible emission could not be detected
due to overlap with the fluorescence emission of acridones and the
weak intensity. For 2-adamantanone: a) lmax =425 nm; G. B. Schuster,
N. J. Turro, H.-C. Steinmetzer, A. P. Schaap, G. Faler, W. Adam, J. C. Liu, J.
Am. Chem. Soc. 1975, 97, 7110–7118; b) fF =0.015; V. P. Kazakov, A. I.
Voloshin, S. S. Ostakhov, N. S. Ableeva, High Energy Chem. 1997, 31,
109–112.
[19] The ability of a para-fluorine to decrease the fluorescence quantum
yield of the corresponding acridanone was further confirmed by com-
paring 3-F,6-Me derivative 2l (fF =0.04) with 3-Me derivative (fF =0.21,
reported in ref. [10c]).
[20] The variations of activation energy due to the presence of different
substituents are very often paralleled by an analogous change in the
pre-exponential coefficient. As a result, the trend in the activation ener-
gies often does not reflect the behavior of the kinetic rate constants of
the TCL reaction, and thus, the t1/2 values.
[21] Although a significant dependence of the CL quantum yield on the
substitution pattern was already established for aryl-substituted 1,2-di-
oxetanes (defined as odd/even rationale), this rationalization refers to
chemically triggered dioxetane decomposition, yielding phenolate-like
emitting species (strong EDG substitution). In particular, for chemically
initiated electron exchange luminescence (CIEEL)-active spiroadaman-
tane 1,2-dioxetanes, such as acetoxynaphthyl spiroadamantyl dioxe-
tanes, it was observed empirically that extended conjugated (para sub-
stituted) carbonyl chromophores derived from dioxetanes triggered de-
composition gave rise to low chemiexcitation efficiencies, whereas
cross-conjugated (meta substituted) carbonyl compounds showed
higher quantum yields. The odd/even rationale was proposed to ex-
plain this phenomenon: charge transfer from the donor (phenolate) to
the acceptor (carbonyl group) occurs more effectively when the two
groups are cross-conjugated (odd number of carbons between the in-
teracting groups), and, presumably, the charge-transfer enhances the
excited-state formation and ensures high chemiexcitation efficiency. On
the contrary, extended conjugation (even number of carbons between
the interacting groups) stabilizes the ground state through dipolar res-
onance, which disfavors excited-state formation and, consequently, pro-
vides low efficiency; see Ref. [5g] and: a) M. Matsumoto, T. Hiroshima, S.
Chiba, R. Isobe, N. Watanabe, H. Kobayashi, Luminescence 1999, 14,
345–348; b) Y. Takano, T. Tsunesada, H. Isobe, Y. Yoshioka, K. Yamaguchi,
I. Saito, Bull. Chem. Soc. Jap. 1999, 72, 213–225; c) H. Isobe, Y. Takano,
M. Okumura, S. Kuramitsu, K. Yamaguchi, J. Am. Chem. Soc. 2005, 127,
8667–8679; d) N. Hoshiya, N. Fukuda, H. Maeda, N. Watanabe, M. Mat-
sumoto, Tetrahedron 2006, 62, 5808–5820; e) L. Yue, Y.-J. Liu, J. Chem.
Theory Comput. 2013, 9, 2300–2312; some authors revealed that siloxy-
and acetoxy-substituted adamantylideneacridane dioxetanes did not
follow the odd/even rationale established for related oxy-substituted
benzoates and congeners; see Refs. [11e,21e].
[12] The acridan-containing 1,2-dioxetane 4a was synthesized for the first
time in 1997 and studied as a chemiluminescent (base-induced) label:
a) T. Imanishi, Y. Ueda, R. Tainaka, K. Miyashita, Tetrahedron Lett. 1997,
38, 841–844; for other acridan-containing 1,2-dioxetanes, see: b) E. H.
White, N. Suzuki, W. H. Hendrickson, Chem. Lett. 1979, 8, 1491–1494;
c) N. Suzuki, Y. Kazui, Y. Izawa, Tetrahedron Lett. 1982, 23, 95–96; d) W.
Adam, D. Reinhardt, Liebigs Ann. 1997, 7, 1359–1364; e) W. Adam, D.
Reinhardt, J. Chem. Soc. Perkin Trans. 2 1997, 1453–1463; f) T. Imanishi,
Y. Ueda, R. Tainaka, N. Kuni, M. Minagawa, N. Hoshino, K. Miyashita, Het-
erocycles 1998, 47, 829–838; g) L. F. M. L. Ciscato, F. H. Bartoloni, D.
Weiss, R. Beckert, W. J. Baader, J. Org. Chem. 2010, 75, 6574–6580.
[13] For selected reviews on the McMurry reaction, see: a) P. Choudhary, V.
Sharma, H. Om, P. Devi, Am. J. Chem. 2013, 3, 115–125; b) T. Takeda, A.
Tsubouchi, Org. React. 2013, 82, 1–470; c) M. M. Heravi, Z. Faghihi, Curr.
Org. Chem. 2012, 16, 2097–2123; d) G. J. Bodwell, P. R. Nandaluru, Isr. J.
Chem. 2012, 52, 105–138; e) M. Iyoda, Pure Appl. Chem. 2010, 82, 831–
841; f) T. Takeda, A. Tsubouchi, Sci. Synth. 2010, 47a, 247–325; g) B.
Goess, in Name Reactions for Homologations, McMurry coupling, (Ed.:
J. J. Li), 2009, pp. 268–283; h) M. M. Krayushkin, M. A. Kalik, V. A. Migu-
lin, Russ. Chem. Rev. 2009, 78, 329–336; i) F. T. Ladipo, Curr. Org. Chem.
2006, 10, 965–980; j) M. Ephritikhine, C. Villiers, in The McMurry cou-
pling and related reactions, Modern Carbonyl Olefination, (Ed.: T. Takeda),
2004, pp. 223–285; k) A. Fuerstner, in The McMurry reaction and related
transformations, Transition Metals for Organic Synthesis (Eds.: M. Beller,
C. Bolm), 2nd Edition, 2004, 1, pp. 449–468; l) J. E. McMurry, Chem. Rev.
1989, 89, 1513–1524.
[14] For selected reviews on photooxygenation reaction, see: a) D. Ashen-
Garry, M. Selke, J. Photochem. Photobiol. A 2014, 90, 257–274; b) K. K.
Banerji, Org. React. Mech. 2012, 79–154; c) S. Fukuzumi, K. Ohkubo, in
Photoinduced reactions of radical ions via charge separation, Encyclope-
dia of Radicals in Chemistry, Biology and Materials, (Eds.: C. Chatgilialo-
glu, A. Studer), 2012, 1, 365–393; d) M. R. Iesce, F. Cermola, in Photoox-
ygenation [2+2] and [4+2], CRC Handbook of Organic Photochemistry
and Photobiology, (Eds.: A. Griesbeck, M. Oelgemoller, F. Ghetti), 3rd Edi-
tion, 2012, 1, pp. 727–764; e) S. Lacombe, T. Pigot, Photochemistry
2010, 38, 307–329; f) M. Zamadar, A. Greer, in Singlet oxygen as a re-
agent in organic synthesis, Handbook of Synthetic Photochemistry, (Eds.:
A. Albini, M. Fagnoni), 2010, pp. 353–386; g) M. Rosaria Iesce, F. Cermo-
la, M. Rubino, Curr. Org. Chem. 2007, 11, 1053–1075; h) E. L. Clennan, A.
Pace, Tetrahedron 2005, 61, 6665–6691; i) M. Stratakis, Curr. Org. Synth.
2005, 2, 281–299; j) Active Oxygen in Chemistry, (Eds.: C. S. Foote, J. S.
Valentine, A. Greenberg, J. F. Liebman), Blackie Academic and Professio-
nal, New York, 1995, Vol. 2; k) C. W. Jefford, Chem. Soc. Rev. 1993, 22,
59–66; l) Z. Jiang, Res. Chem. Intermed. 1990, 14, 185–207; m) L. Lopez,
Top. Curr. Chem. 1990, 156, 117–166; n) A. A. Frimer, in Singlet Oxygen,
Volume 2: Reaction Modes and Products, Part I, CRC Press, Boca Raton,
1985.
[15] The adamantyl group is widely recognized as a rigid and sterically de-
manding framework able to significantly stabilize 1,2-dioxetanes: a) F.
McCapra, I. Beheshti, A. Burford, R. A. Hann, K. A. Zaklika, J. Chem. Soc.
Chem. Commun. 1977, 944–946; b) C. Lee, L. A. Singer, J. Am. Chem.
Soc. 1980, 102, 3823–3829; c) P. Lechtken, G. Reissenweber, P. Grubmu-
eller, Tetrahedron Lett. 1977, 18, 2881–2884; d) W. Adam, L. A. A. Encar-
naciꢃn, K. Zinner, Chem. Ber. 1983, 116, 839–846; see also Re-
[22] For chemically or enzymatically triggerable EDG-substituted aryl dioxe-
tanes, on which an EWG was inserted, see: a) I. Bronstein, B. Edwards,
A. Sparks, US5538847, 1996; b) B. P. Giri, WO0014092, 1998; c) M. Mat-
&
&
Chem. Eur. J. 2016, 22, 1 – 14
www.chemeurj.org
12
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
sumoto, T. Mizuno, N. Watanabe, Chem. Commun. 2003, 482–483;
d) B. P. Giri, D. Dagli, WO03054506, 2003; e) N. Watanabe, T. Mizuno, M.
Matsumoto, Tetrahedron 2005, 61, 9569–9585; f) B. Edwards, R.-R. Juo,
A. L. Sparks, US20060079699, 2006. See also: g) Z. Huang, X. Liang, Y.
Chan, Chin. J. Chem. 2010, 9, 506–511; h) I. Y. Bronstein, B. Edwards,
J. C. Voyta, L. J. Kricka, WO8906650, 1989; for light-triggered 1,2-dioxe-
tanes containing the unsubstituted acridan aromatic system, see
Ref. [12g].
A. Griesbeck, M. Oelgemoller, F. Ghetti), CRC Handbook of Organic Photo-
chemistry and Photobiology, 3rd Edition, 2012, 1, 789–808; b) E. L.
Clennan, S. E. Hightower, J. Org. Chem. 2006, 71, 1247–1250; c) W.
Adam, A. S. Kumar, C. R. Saha-Mçller, Tetrahedron Lett. 1995, 36, 7853–
7854; d) F. Cermola, M. R. Iesce, J. Org. Chem. 2002, 67, 4937–4944. See
also Refs [14h,24a] and cited herein.
[29] a) R. Gompper, Angew. Chem. Int. Ed. Engl. 1969, 8, 312–327; Angew.
Chem. 1969, 81, 348–363; b) R. W. Hoffmann, Angew. Chem. Int. Ed.
Engl. 1968, 7, 754–765; Angew. Chem. 1968, 80, 823–835; c) C. S.
Foote, A. A. Dzakpasu, J. H.-P. Lin, Tetrahedron Lett. 1975, 16, 1247–
1250; d) P. D. Bartlett, A. P. Schaap, J. Am. Chem. Soc. 1970, 92, 3223–
3225; e) S. Mazur, C. S. Foote, J. Am. Chem. Soc. 1970, 92, 3225–3226;
f) A. P. Schaap, P. D. Bartlett, J. Am. Chem. Soc. 1970, 92, 6055–6057; see
also Ref. [14d] and references cited therein.
[23] The activation parameters for the thermal decomposition of 1,2-dioxe-
tanes 4a and 4h–t were determined by a standard isothermal kinetic
method, which measured the emission decay kinetics at different tem-
peratures and gave the kinetic constants of the TCL process (see the
Supporting Information for details).
[24] For examples of particularly stable 1,2-dioxetanes, see: a) N. Watanabe,
M. Kikuchi, Y. Maniwa, H. K. Ijuin, M, Matsumoto, J. Org. Chem. 2010,
75, 879–884; b) N. Watanabe, Y. Sano, H. Suzuki, M. Tanimura, H. K.
Ijuin, M. Matsumoto, J. Org. Chem. 2010, 75, 5920–5926; c) M. Matsu-
moto, S. Nasu, M. Takeda, H. Murakami, N. Watanabe, Chem. Commun.
2000, 821–822; d) M. Matsumoto, N. Watanabe, N. C. Kasuga, F.
Hamada, K. Tadokoro, Tetrahedron Lett. 1997, 38, 2863–2866; see also
Ref. [18a]; for a theoretical study on the thermal stability of alkyl-substi-
tuted 1,2-dioxetanes, see: e) E. L. Bastos, W. J. Baader, ARKIVOC (Gaines-
ville, FL, U.S.) 2007, 8, 257–272.
[25] The significant increase in the fluorescence quantum yields of acridones
containing EDGs could be due to a decrease in the intersystem crossing
rate constant, and thus, to a lower efficiency of nonradiative deactiva-
tion processes, as reported for aromatic compounds, such as anthra-
cene (N. Nijegorodov, V. Vasilenko, P. Monowe, M. Masale, Spectrochim.
Acta Part A 2009, 74, 188–194).
[26] Concerning the effects of the N-substituent on the fluorescence quan-
tum yield of acridones, see also Ref. [12]. In particular, regarding the ef-
fects of the acridan substitution pattern on the fluorescence quantum
yields of the emitting species, see Ref. [12e] and references cited there-
in: in the excitation step of the acridone system, charge is transferred
from the donor N atom to the carbonyl group [Eq. (1)].
[30] For mechanistic studies on photooxygenation of olefins to provide 1,2-
dioxetanes, see: a) E. L. Clennan, K. Nagraba, J. Am. Chem. Soc. 1988,
110, 4312–4318; b) K. Ohkubo, T. Nanjo, S. Fukuzumi, Org. Lett. 2005, 7,
4265–4268; c) H. Kotani, K. Ohkubo, S. Fukuzumi, J. Am. Chem. Soc.
2004, 126, 15999–16006; d) Y. Yoshioka, S. Yamada, T. Kawakami, M.
Nishino, K. Yamaguchi, I. Saito, Bull. Chem. Soc. Jpn. 1996, 69, 2683–
2699 and cited herein; e) W. Adam, C. R. Saha-Mçller, S. B. Schambony,
J. Am. Chem. Soc. 1999, 121, 1834–1838; f) M. Matsumoto, H. Kobaya-
shi, J. Matsubara, N. Watanabe, S. Yamashita, D. Oguma, Y. Kitano, H.
Ikawa, Tetrahedron Lett. 1996, 37, 397–400; g) M. Matsumoto, Y. Kitano,
H. Kobayashi, H. Ikawa, Tetrahedron Lett. 1996, 37, 8191–8194; h) W.
Adam, S. G. Bosio, N. J. Turro, J. Am. Chem. Soc. 2002, 124, 8814–8815;
i) T. Poon, J. Sivaguru, R. Franz, S. Jockusch, C. Martinez, I. Washington,
W. Adam, Y. Inoue, N. J. Turro, J. Am. Chem. Soc. 2004, 126, 10498–
10499; j) W. Wu, F. An, Z. Geng, R. Zhang, Z. Jiang, Lett. Org. Chem.
2013, 10, 223–227; k) K. Yamaguchi, S. Yabushita, T. Fueno, K. N. Houk,
J. Am. Chem. Soc. 1981, 103, 5043–5046; l) M. Kojima, M. Nakajoh, C.
Matsubara, S. Hashimoto, J. Chem. Soc. Perkin Trans. 2 2002, 1894–
1901; see also Ref. [14h] and references cited therein; for a selected ex-
ample of a mechanistic study on photooxygenation of acenes, see:
m) W. Fudickar, T. Linker, J. Am. Chem. Soc. 2012, 134, 15071–15082.
[31] a) A. Maranzana, G. Ghigo, G. Tonachini, J. Am. Chem. Soc. 2000, 122,
1414–1423; b) M. Bobrowski, A. Liwo, S. Oldziej, D. Jeziorek, T. Ossow-
ski, J. Am. Chem. Soc. 2000, 122, 8112–8119; c) F. Sevin, M. L. McKee, J.
Am. Chem. Soc. 2001, 123, 4591–4600.
[32] a) N. Ferrꢂ, N. Guihꢂry, J.-P. Malrieu, Phys. Chem. Chem. Phys. 2015, 17,
14375–14382; b) K. Yamaguchi, S. Yamanaka, J. Shimada, H. Isobe, T.
Saito, M. Shoji, Y. Kitagawa, M. Okumura, Int. J. Quantum Chem. 2009,
109, 3745–3766.
The charge-transfer character of the excited acridone chromophore, as
expected for a vinylogous amide, is its main electronic characteristic.
Thus, it is reasonable that modifications on the nitrogen substituent
affect the fluorescence quantum yield much more than the introduc-
tion of substituents on the aromatic rings. Furthermore, the position of
the aromatic rings substituents perturbs this charge-transfer transition
significantly less in acridan-based systems than in benzoate-based sys-
tems. For the odd/even rationale established for oxy-substituted ben-
zoates and naphthoates, see: a) A. P. Schaap, R. S. Handley, B. P. Giri, Tet-
rahedron Lett. 1987, 28, 935–938; b) A. P. Schaap, T.-S. Chen, R. S. Han-
dley, R. DeSilva, B. P. Giri, Tetrahedron Lett. 1987, 28, 1155–1158; see
also Ref. [21].
[33] R. Rajeev, R. B. Sunoj, J. Org. Chem. 2012, 77, 2474–2485.
[34] a) R. Broa, A. K. Smilde, Anal. Methods 2014, 6, 2812–2831; b) H. Abdi,
L. J. Williams, WIREs Comp. Stat. 2010, 2, 433–459; c) A. Afifi, S. May,
V. A. Clark, in Practical Multivariate Analysis, CRC Press, 2011; d) K. Var-
muza, P. Filzmoser, in Introduction to Multivariate Statistical Analysis in
Chemometrics, CRC Press, 2009.
[35] For a similar approach, see: L. Eriksson, H. J. M. Verhaar, L. M. Hermens,
Environ. Toxicol. Chem. 1994, 13, 683–691.
[36] a) M. Mousavi-Kamazani, M. Salavati-Niasari, M. Goudarzi, A. Gharehbaii,
J. Inorg. Organomet. Polym. Mater. 2016, 26, 259–263; b) W. Chen, H.
Luo, X. Liu, J. W. Foley, X. Song, Anal. Chem. 2016, 88, 3638–3646;
c) V. D. Vandekar, J. Nat. Prod. Plant Res. 2015, 5, 43–52; d) M. Megala,
B. J. M. Rajkumar, J. Comput. Electron. 2016, 15, 557–568.
[37] a) V. F. Traven, A. V. Manaev, A. Y. Bochkov, T. A. Chibisova, I. V. Ivanov,
Russ. Chem. Bull. 2012, 61, 1342–1362; b) S. Nath, M. Kumbhakar, H.
Pal, in Hydrogen Bonding and Transfer in the Excited State, (Eds.: K.- L.
Han, G.- J. Zhao), 2011, 1, pp. 419–431; c) S. Fery-Forgues, C. Fournier-
Noel, in Nanofibers, (Ed.: A. Kumar), 2010, pp. 383–404; d) M. A. Musa,
J. S. Cooperwood, M. O. F. Khan, Curr. Med. Chem. 2008, 15, 2664–2679;
e) N. Sekar, Colourage 2005, 52, 71–73; f) J.-P. Goddard, J.-L. Reymond,
Trends Biotechnol. 2004, 22, 363–370; g) B. R. Hirani, M. P. Singh, V. K.
Koul, PharmaChem 2003, 2, 50–52. See also Ref. [22h].
[27] This observation is in agreement with the data reported by other au-
thors on N-aryl acridan 1,2-dioxetanes (see Ref. [12f]). Considering the
possible conformations of N-aryl acridan 1,2-dioxetanes, a steric repul-
sion between the peri-hydrogen atoms of the acridan system and the
aryl group might be generated. Therefore, the two rings would not be
able to conjugate each other sufficiently and the electronic effects of
the aryl group would be largely reduced.
Received: August 5, 2016
Revised: September 7, 2016
Published online on && &&, 0000
[28] Concerning the photo-oxygenation of sulfur-containing compounds,
see: a) E. L. Clennan, in Photooxygenations of sulfur compounds, (Eds.:
Chem. Eur. J. 2016, 22, 1 – 14
www.chemeurj.org
13
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
FULL PAPER
Pick and choose: New high-per-
&
Substituent Effects
formance thermochemiluminescent 1,2-
dioxetanes for later use in ultrasensitive
bioanalytical assays have been synthe-
sized. A chemometric approach enables
dioxetane formation to be rationally
predicted by defining essential steric
and electronic parameters of the olefin
precursor (see figure).
L. A. Andronico, A. Quintavalla,*
M. Lombardo,* M. Mirasoli, M. Guardigli,
C. Trombini, A. Roda
&& – &&
Synthesis of 1,2-Dioxetanes as
Thermochemiluminescent Labels for
Ultrasensitive Bioassays: Rational
Prediction of Olefin Photooxygenation
Outcome by Using a Chemometric
Approach
&
&
Chem. Eur. J. 2016, 22, 1 – 14
www.chemeurj.org
14
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!