Fluorescent 1:2 demultiplexer and half-subtractor based on the hydrolysis of N-salicylidene-3-aminopyridine

Article abstract of DOI:10.1016/j.saa.2013.08.009

In moist chloroform solution, the ultraviolet light irradiation would cause N-salicylidene-3-aminopyridine (L) to hydrolyze into salicylaldehyde and 3-aminopyridine. To consider an optical signal (UV light) and a chemical signal (Zn²⁺) as inputs, the luminescence signals as outputs, the logic behavior of compound L was investigated. Interestingly, excited by two different wavelengths lights, two sharp distinct fluorescent spectra were collected. Consequently, a fluorescent 1:2 demultiplexer and a fluorescent half-subtractor were respectively expressed.

Full text of DOI:10.1016/j.saa.2013.08.009

Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 117 (2014) 397–401
Contents lists available at ScienceDirect
Spectrochimica Acta Part A: Molecular and
Biomolecular Spectroscopy
journal homepage: www.elsevier.com/locate/saa
Fluorescent 1:2 demultiplexer and half-subtractor based on the
hydrolysis of N-salicylidene-3-aminopyridine
⇑
Liyan Zhao, Wujiong Xia , Chao Yang
The State Key Lab of Urban Water Resource and Environment & Academy of Fundamental and Interdisciplinary Science, Harbin Institute of Technology, P.O. Box 3026, 2 Yikuang Street,
Harbin, Heilongjiang 150080, PR China
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Based on the hydrolysis, a simple molecule has implemented the important demultiplexer and half-
subtractor.
ꢀ
ꢀ
ꢀ
A fluorescent 1:2 demultiplexer was
expressed.
A fluorescent half-subtractor were
expressed.
The spectrum changes were
originated from the photoinduced
electron transfer.
UV
In
A
N
H
N
O
In1
In1
2+
Zn
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 4 April 2013
Received in revised form 21 July 2013
Accepted 2 August 2013
Available online 13 August 2013
In moist chloroform solution, the ultraviolet light irradiation would cause N-salicylidene-3-aminopyri-
dine (L) to hydrolyze into salicylaldehyde and 3-aminopyridine. To consider an optical signal (UV light)
and a chemical signal (Zn2+) as inputs, the luminescence signals as outputs, the logic behavior of com-
pound L was investigated. Interestingly, excited by two different wavelengths lights, two sharp distinct
fluorescent spectra were collected. Consequently, a fluorescent 1:2 demultiplexer and a fluorescent half-
subtractor were respectively expressed.
Keywords:
Schiff bases
Ó 2013 Elsevier B.V. All rights reserved.
Fluorescent devices
Logic gate
Half-subtractor
Demultiplexer
Introduction
were exploited [3–11]. Among the logic gates reported, the ones
using fluorescence as outputs are ideal because of its ease of detec-
Following the concept of molecular logic proposed by Aviram in
1988 [1], de Silva firstly contributed an AND logic gate at molecular
level [2]. Since then, a large number of significant logic functions
tion. Salicylidene Schiff base is a kind of active material for its
excellent luminescence and remarkable coordination properties
[12,13]. Therefore, salicylidene Schiff base is very suitable to devel-
op the fluorescent logic operations. In our previous work, we have
reported several logic functions of salicylidene Schiff base [14–19].
Herein, we will present what we have achieved on developing
fluorescent logic function based on salicylidene Schiff base.
⇑
Corresponding author. Tel./fax: +86 451 86403193.
E-mail address: xiawj@hit.edu.cn (W. Xia).
1386-1425/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.saa.2013.08.009

398
L. Zhao et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 117 (2014) 397–401
Scheme 1. The structure, preparation and hydrolysis of compound L.
Recently, we found that, in moist chloroform solution, UV light
lamp (500 W) was used as light source for irradiation reaction in
NMR measurement. The interval time was 1 h.
irradiation could lead salicylidene Schiff bases to hydrolyze into
the corresponding salicylaldehyde and amine [20]. This result
aroused our interest. It can be envisioned that, in moist chloroform
solution, UV light and transition metal ion would likely stimulate
salicylidene Schiff base to act some interesting logic functions. In
order to efficiently utilize the fluorescence signals as outputs, we
introduced pyridine group on salicylidene Schiff base to increase
the coordination ability. Hereby, one salicylidene Schiff base L
was prepared (shown in Scheme 1). We tracked the irradiation
experiment of L by time-dependent absorption spectroscopy and
time-dependent NMR. The results demonstrated that, upon the
irradiation of UV light, L was hydrolyzed into salicylaldehyde
(SA) and 3-aminopyridine (3AP). The structure, preparation and
hydrolysis of L are shown in Scheme 1.
Instrumentation
The UV–Vis absorption spectra were recorded in a Hitachi
(Model U-4100) UV–Vis–NIR spectrophotometer. The fluorescence
spectra were determined with a FluoroMax-4 (JobinYvon) fluorim-
eter. 1H NMR spectra were recorded at 400 MHz on a Bruker AV-
400 instrument. 13C NMR spectra were recorded at 100 MHz.
Low-resolution mass spectra were measured on an Agilent
7890A-5975C GC–MS using electron impact (EI) ionization at
70 eV. High-resolution mass spectra were taken on an Agilent
1200–6520 Q-TOF electro spray mass spectrometer.
UV light and Zn2+ were selected as inputs to stimulate the re-
sponse molecule L. The fluorescence spectroscopy was employed
to investigate the response process. The actions of inputs had in-
duced dramatic fluorescence changes. Interestingly, when different
excitation lights were used (370 nm and 320 nm), two widely dis-
tinct fluorescent spectra were obtained. Consequentially, fluores-
cent 1:2 demultiplexer and fluorescent half-subtractor were
expressed. Demultiplexer [21–26] and half-subtractor [27–33]
are essential logic operations in the information processing and
calculation field respectively.
Results and discussion
Photochemical experiments
First of all, we studied the photochemical behavior of com-
pound L by the time-dependent absorption spectra as shown in
Fig. 1. With irradiation, the original absorption band at 370 nm
gradually decreased; meanwhile, a new absorption at 252 nm
appeared and intensified. In addition, the absorption at 295 nm
gradually decreased and simultaneously red-shifted to 320 nm.
Above spectral changes were similar to the experimental result re-
ported in our previous work [20]. Such changes have been testified
to be due to the hydrolysis reaction of salicylidene Schiff bases
[20]. Therefore, it was speculated that compound L might have
been hydrolyzed into SA and 3AP as shown in Scheme 1.
Experimental
Material
And then, we employed the time-dependent 1H NMR to verify
this hypothesis. Considering that the concentration for NMR mea-
surement must be higher, we increased the concentration of L from
6.06 ꢁ 10ꢂ5 mol/L to 2.02 ꢁ 10ꢂ2 mol/L. Consequently, the irradia-
tion time in time-dependent 1H NMR spectra was longer. The
Salicylaldehyde and 3-aminopyridine were purchased from Al-
drich. Compound L was synthesized by the condensation of the
3-aminopyridine with salicylaldehyde according to the previous
procedures [34]. Pure L was obtained after recrystallization twice
from absolute ethanol. The 1H NMR (DMSO, 25 °C, TMS): d 12.58
(s, 1H), 9.02 (s, 1H), 8.64 (s, 1H), 8.63 (s, 1H), 8.50 (d, 1H), 7.86
(d, 1H), 7.70–7.43 (2H), 7.00 (t, 2H). 13C NMR (CDCl3): d 164.54,
161.13, 148.00, 144.68, 143.10, 133.88, 132.65, 127.98, 123.88,
119.36, 118.94, 117.42. GC–MS (EI): 198 (M+), 181, 165, 147, 120,
104, 78, 51. HRMS (ESI) [M+H]+, 199.0876, calcd for: 199.0866.
0.8
0.6
Experimental details for spectral studies
A
0.4
All of the spectral analysis were accomplished in chloroform.
The concentration of
L solution for spectral analysis was
6.06 ꢁ 10ꢂ5 mol/L. A low-pressure mercury (16 W) lamp was used
as light source for irradiation stimulations in the spectral measure-
ment. The interval irradiation time was 1 min. Solutions for spec-
trophotometric titration were titrated directly in 1 cm absorption
cells by successive additions of corresponding chemical reagent
using a microliter syringe. After addition of each aliquot, the cell
was capped and mixed by inversion. And then, the spectrum was
retaken. The time-dependent 1H NMR experiments were accom-
plished in deuterochloroform. The concentration of L solution for
spectral analysis was 2.02 ꢁ 10ꢂ2 mol/L. A high-pressure mercury
0.2
0.0
250
300
350
\ nm
400
450
λ
Fig. 1. UV–Vis absorption spectral changes of L with 254 nm UV light irradiation.
Interval irradiation time is 1 min. q is the original state, H is the final state.
Conditions: 6.06 ꢁ 10ꢂ5 mol/L of chloroform solution.

L. Zhao et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 117 (2014) 397–401
399
UV light had not caused any changes within the 1H NMR spectra.
Above results suggested that the hydrolysis might be relevant to
the moist chloroform.
According to literature retrieval, it has been confirmed that
chloroform is photolabile. Irradiated by UV light, chloroform will
produce HCl [35,36]. It is well known that HCl can catalyze the
hydrolysis process of Schiff bases. The possible hydrolysis mecha-
nism of N-salicylidenaniline Schiff bases was proposed in the
Scheme 2 [20]. Furthermore, we investigated the hydrolysis in
chloroform solution containing different amount of water. When
the sample L and water were in proportion of 1:1, the hydrolysis
reaction could be carried out completely. When the amount of
water was continuously increased, the velocity of hydrolysis reac-
tion would speed up.
The fluorescent spectral analysis
Fluorescence spectroscopy was employed to investigate the
processes of irradiation and coordination with Zn2+. The concentra-
tion of L was 6.06 ꢁ 10ꢂ5 mol/L and the total irradiation time was
13 min. The characteristic absorptions before and after irradiation
(370 nm and 320 nm shown in Fig. 1) were respectively selected as
the excitation light to investigate the fluorescent properties. Inter-
estingly, two different excitation lights had induced two widely
distinct fluorescent spectra as shown in Figs. 3a and 4a, which
would be particularly discussed in the following sections.
Fig. 2. Partial 1H NMR (400 MHz) spectral changes of compound L in CD3Cl upon
irradiation of high-pressure UV light. Interval time is 1 h. Concentration:
2.02 ꢁ 10ꢂ2 mol/L.
experimental result of time-dependent 1H NMR spectra was shown
in Fig. 2. It could be observed that the primary proton Ha and Hb
gradually decreased. Meanwhile, two new chemical shifts (around
11 ppm and 10 ppm) appeared and intensified with the irradiation.
These two new chemical shifts belonged to the characteristic 1H
NMR signals of reactant SA (Hc and He, the 1H NMR of SA can be
seen in the Electronic Supplementary Information file). Above 1H
NMR spectral changes were consistent with the experimental re-
sults reported in our previous work [20]. Thus, it could be con-
cluded that compounds L had been hydrolyzed into SA and 3AP
as shown in Scheme 1.
Moist chloroform was dried with anhydrous CaCl2. And then,
the filtrate was distilled from P2O5. Consequently, dried chloroform
was obtained. When the photochemical behavior of compound L
was studied in dried chloroform, no hydrolysis reaction was ob-
served. In addition, the hydrolysis reaction did not occur in other
solvents such as benzene, acetonitrile, cyclohexane. Moreover, in
deuterated benzene or deuterated acetonitrile, the irradiation of
Excited with 320 nm light
When 320 nm light was used as the excitation light, the actions
of UV light and Zn2+ had led to sharp changes in the fluorescence
spectra (Fig. 3a) The detailed irradiation and titration spectra are
shown in Supplementary data file: Figs. S1–S3. As seen in Scheme 1,
the hydrolysis product was a mixture of SA and3AP. Which compo-
nent would be responsible for the emissions related to the actions
of irradiation? We tried our best to grow crystals, but failed. Using
320 nm light as excitation light, we measured the fluorescent spec-
tra of SA, 3AP and their complexes. The fluorescence spectra were
collected in Fig. 5. Comparing Fig. 3a with Fig. 5, it can be found
that the shapes, positions and relative intensities of the fluorescent
spectra labeled with q and 4 (Fig. 3a) are consistent with the ones
labeled with 3-aminopyridine and 3-aminopyridine + Zn2+
respectively (Fig. 5). Therefore, 3AP is responsible for the emission
Scheme 2. The hydrolysis mechanisms for salicylidene Schiff bases.

400
L. Zhao et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 117 (2014) 397–401
AND
INH
1.0
(a)
INH
(a)
1.0
0.8
0.6
0.4
0.2
0.0
0.8
INH XOR
0.6
I
I
0.4
0.2
0.0
350
400
450
500
550
600
400
450
500
550
600
650
700
λ\ nm
λ
\ nm
(b)
(c)
(b)
(c)
Fig. 3. (a): Fluorescence spectra of L (6.06 ꢁ 10–5 mol/L of chloroform solution,
kex = 320 nm) upon the actions of inputs: H L alone; q after irradiation by UV light
(total irradiation time is 13 min); N with 5 equiv of Zn2+; 4 irradiation by UV
light + Zn2+
. (b): The logic diagram and (c): The truth table for the 1:2
demultiplexer.
Fig. 4. (a): Fluorescence spectra of L (6.06 ꢁ 10ꢂ5 mol/L of chloroform solution,
kex = 370 nm) upon the actions of inputs: H L alone; q after irradiation by UV light
(total irradiation time is 13 min); N with 5 equiv of Zn2+; 4 irradiation by UV
light + Zn2+. (b): The logic diagram and (c): The truth table for the half-subtractors.
of irradiation products and the complex of 3AP with Zn2+ is
responsible for the emission after the actions of UV light plus
Zn2+. It must be noted that there is some slight difference in above
two pairs of fluorescent spectra respectively. The difference may be
due to the fact that hydrolysis products is a mixture, consequently,
another component SA will affect the correlative fluorescent spec-
tra to some extent. In addition, 1H NMR titration experiments for
the irradiation sample had provided further evidence. As can be
seen from Fig. S8 (shown in the Electronic Supplementary Informa-
tion file), after titration with Zn2+, the signals around 8.60 ppm and
7.10 assigned to the H atoms of 3AP, were shifted upfield. These
differences in the 1H NMR indicated that the change of the electron
density owning to the through-bond electronic effects, in other
words, 3AP had coordinated with Zn2+
.
Coordination will reduce the photoinduced electron transfer
(PET). Therefore, the complexes before and after irradiation all pre-
sented higher emission than the ones of ligands L and 3AP [37].
Since 320 nm was not the optimal excitation wavelength for both
L and LZn2+, they showed relatively weaker fluorescence (Fig. 3a)
than the case of 370 nm light as excitation light (Fig. 4a). It could
be observed that only complex LZn2+ exhibited emission at longer
wavelength region. This red-shifted emission probably due to the
charge transfer transition (CT) presenting in the process of co-ordi-
nation [38,39].
Fig. 5. Fluorescence spectra of salicyladehyde (solid), salicyladehyde + Zn2+ (dash),
3-aminopydine (dash dot) and 3-aminopydine + Zn2+ (dot). Chloroform solution,
kex = 320 nm.

L. Zhao et al. / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 117 (2014) 397–401
401
Excited with 370 nm light
2012043), SKLUWR of HIT (No. ES201010), Project Supported by
Development Program for Outstanding Young Teachers of HIT
(HITQNJS 2009.068), program of excellent Team in HIT, China Post-
doctoral Science Foundation (20080440867), and Heilongjiang
Postdoctoral Fund (LBH-Z08199).
In sharp contrast, when the 370 nm light was employed to ex-
cite the response system, considerable different spectra were ob-
tained (Fig. 4a) The detailed irradiation and titration spectra are
shown in Supplementary data file: Figs. S4–S6. In this case, a more
significant emission at kmax = 520 nm was observed. This emission
also was original from the PET and CT processes of LZn2+ men-
tioned in Section 3.2.1. According to the literature, pyridine ring
in molecule L would undergo a complete rotation after coordina-
tion with metal ion [40]. Consequently, the molecular packing
showed a ‘‘close’’ structure with short distances between ligands.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.saa.2013.08.009.
Therefore, the p–p stacking between salicyl rings and imine double
References
bonds was formatted, which would induce enhanced fluorescence
and red-shift. We tried our best to grow crystal of LZn2+, but failed.
Since 370 nm was not the optimal excitation wavelength, complex
[1] A. Aviram, J. Am. Chem. Soc. 110 (1988) 5687–5692.
[2] A.P. de Silva, N.H.Q. Gunaratne, C.P. McCoy, Nature 364 (1993) 42–44.
[3] K. Szaciłowski, Chem. Rev. 108 (2008) 3481–3548.
[4] E. Katz, V. Privman, Chem. Soc. Rev. 39 (2010) 1835–1857.
[5] J. Andréasson, U. Pischel, Chem. Soc. Rev. 39 (2010) 174–188.
[6] M. Amelia, L. Zou, A. Credi, Coord. Chem. Rev. 254 (2010) 2267–2280.
[7] G. Bergamini, E. Marchi, P. Ceroni, Coord. Chem. Rev. 255 (2011) 2458–2468.
[8] J. Wang, C.S. Ha, Sens. Actuators B 146 (2010) 373–380.
[9] P. Singh, J. Kaur, W. Holzer, Sens. Actuators B 150 (2010) 50–56.
[10] S. Kou, H.N. Lee, D. Van Noort, K.M.K. Swamy, S.H. Kim, J.H. Soh, K.-M. Lee, S.-
W. Nam, J. Yoon, S. Park, Angew. Chem. Int. Ed. 47 (2008) 872–876.
[11] N. Kaur, N. Singh, B. McCaughan, J.F. Callan, Sens. Actuators B 144 (2010) 88–
91.
[12] K.C. Gupta, A.K. Sutar, Coord. Chem. Rev. 252 (2008) 1420–1450.
[13] J.A. Riddle, S.P. Lathrop, J.C. Bollinger, D. Lee, J. Am. Chem. Soc. 128 (2006)
10986–10987.
[14] L. Zhao, S. Wang, Y. Wu, Q. Hou, Y. Wang, S. Jiang, J. Phys. Chem. C. 111 (2007)
18387–18391.
[15] L. Zhao, D. Sui, J. Chai, Y. Wang, S. Jiang, J. Phys. Chem. B. 110 (2006) 24299–
24304.
[16] S. Wang, L. Zang, L. Zhao, X. Wang, Q. Hou, S. Jiang, Spectrochim. Acta A 77
(2010) 226–231.
[17] S. Wang, G. Men, L. Zhao, Y. Wu, Y. Wang, S. Jiang, IEEE Sens. J. 1–3 (2009) 277–
280.
[18] S. Wang, G. Men, L. Zhao, Q. Hou, S. Jiang, Sens. Actuators B 145 (2010) 826–
831.
[19] S. Wang, G. Men, Y. Wang, L. Zhao, Q. Hou, S. Jiang, IEEE Sens. J. 11 (2011) 137–
141.
[20] L. Zhao, W. Xia, B. Gou, Y. Lu, C. Yang, D. Li, Spectrochim. Acta A 95 (2012) 199–
203.
[21] M. Amelia, M. Baroncini, A. Credi, Angew. Chem. Int. Ed. 47 (2008) 6240–6243.
[22] E. Perez-Inestrosa, J.M. Montenegro, D. Collado, R. Suau, Chem. Commun.
(2008) 1085–1087.
[23] J. Andréasson, S.D. Straight, S. Bandyopadhyay, R.H. Mitchell, T.A. Moore, A.L.
Moore, D. Gust, J. Phys. Chem. C 111 (2007) 14274–14278.
[24] S. Gaweda, G. Stochel, K. Szacilowski, Chem. Asian J. 2 (2007) 580–590.
[25] M.F. Budyka, N.I. Potashova, T.N. Gavrishova, V.M. Lee, J. Mater. Chem. 19
(2009) 7721–7724.
of 3AP with Zn2+ showed a relatively weaker emission than LZn2+
.
The logic analysis under actions of UV light and Zn2+
The logic nature derived from Fig. 3a was collected in the truth
table as shown in Fig. 3c. As it could be seen, the fluorescence
intensity at 520 nm (Out1) was high only in the absence of UV light,
but in the presence of Zn2+. This truth table justly expressed an IN-
HIBIT logic gate. The intensity of another fluorescence channel at
368 nm (Out2) was high only in the presence of both UV light
and Zn2+, which worked as an AND logic gate. Therefore, using
UV light as the address input, Zn2+ as the data input, and the fluo-
rescence emission at 368 nm and 520 nm as the dual data outputs,
compound L could act as 1:2 digital demultiplexer. The logic dia-
gram was shown in Fig. 3b.
Now we turned to the discussion of the logic operations as
shown in Fig. 4a. When the emissions were followed at three dif-
ferent channels, 445 nm (Out1), 470 nm (Out2) and 520 nm
(Out3), the clear digital nature of the inputs and outputs became
legible as described in Fig. 4c. In the channels of Out1 and Out3,
compound L could act as two INHIBIT gates. In addition, the chan-
nel of Out2 could express a XOR gate. Therefore, concatenation
half-subtractors were obtained (Fig. 4b). It should be pointed out
that when the Out3 combined with Out2, the a–b numeration
was operated. However, the Out1 and Out2 would perform the b–
a logic function.
[26] M.A. Arugula, V. Bocharova, J. Halamek, M. Pita, E. Katz, J. Phys. Chem. B 114
(2010) 5222–5226.
Conclusions
[27] Y. Liu, W. Jiang, H. Zhang, C. Li, J. Phys. Chem. B 110 (2006) 14231–14235.
[28] Z. Guo, P. Zhao, W. Zhu, X. Huang, Y. Xie, H. Tian, J. Phys. Chem. C 112 (2008)
7047–7053.
[29] V. Luxami, S. Kumar, New J. Chem. 32 (2008) 2074–2079.
[30] S.J. Langford, T. Yann, J. Am. Chem. Soc. 125 (2003) 11198–11199.
[31] A. Coskun, E. Deniz, E.U. Akkaya, Org. Lett. 7 (2005) 5187–5189.
[32] R. Baron, O. Lioubashevski, E. Katz, T. Niazov, I. Willner, Angew. Chem. Int. Ed.
45 (2006) 1572–1576.
[33] M. Suresh, D.A. Jose, A. Das, Org. Lett. 9 (2007) 441–444.
[34] H.E. Smith, S.L. Cook, M.E. Warren, J. Org. Chem. 29 (1964) 2265–2267.
[35] K.S. Georgiou, Appl. Phys. A 92 (2008) 1013–1017.
[36] C. Romanzin, B. Gans, S. Douin, S. Boyé-Péronne, D. Gauyacq, Chem. Phys. 351
(2008) 77–82.
In summary, we further developed the salicylidene Schiff base
as both a fluorescent 1:2 demultiplexer and a fluorescent half-sub-
tractor based on a hydrolysis reaction. The arithmetic and demul-
tiplexer functions constructed in this work may be helpful to
future application of salicylidene Schiff bases in the molecular de-
vices. Furthermore, it is of interest to point out that two different
logic functions presented here mainly depend on two different
wavelength of excitation lights. Work on the crystal growth and
exploring detailed photoinduced proton transfer is underway.
[37] L. Gao, Y. Wang, J.Q. Wang, L. Huang, L.Y. Shi, X.X. Fan, Z.G. Zou, T. Yu, M. Zhu,
Z.S. Li, Inorg. Chem. 45 (2006) 6844–6850.
[38] P.F. Wang, Z. Hong, Z.Y. Xie, S.W. Tong, O.Y. Wong, C.S. Lee, N.B. Wong, L.S.
Hung, S.T. Lee, Chem. Commun. (2003) 1664–1665.
Acknowledgements
This work is supported by the China NSFC (Nos. 20802013 and
21072038), the New Century Excellent Talent Program of Chinese
Ministry of Education (No. NCET-07-0242), the Fundamental Re-
search Funds for the Central Universities (Grant No. HIT. NSRIF.
[39] P.G. Lacroix, S.D. Bella, I. Ledoux, Chem. Mater. 8 (1996) 541–549.
[40] F. Robert, A.D. Naik, B. Tinant, R. Robiette, Y. Garcia, Chem. Euro. J. 15 (2009)
4327–4342.