Radical cation formation in characterization of novel C3- symmetric disks and their precursors by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry

Article abstract of DOI:10.1002/jms.1025

Four C₃-symmetrical tris(dipeptide) disks and their precursors were characterized using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI TOF MS). The C₃-symmetrical disks were based on a benzene-1,3

Full text of DOI:10.1002/jms.1025

JOURNAL OF MASS SPECTROMETRY
J. Mass Spectrom. 2006; 41: 659–669
Published online 10 April 2006 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jms.1025
Radical cation formation in characterization of novel
C3-symmetric disks and their precursors by
matrix-assisted laser desorption/ionization
time-of-flight mass spectrometry
Xianwen Lou,∗ Kelly P. van den Hout, Michel H. C. J. van Houtem, Joost L. J. van Dongen,
Jef A. J. M. Vekemans and E. W. Meijer
Laboratory of Macromolecular and Organic Chemistry, Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven, The Netherlands
Received 17 November 2005; Accepted 16 February 2006
Four C3-symmetrical tris(dipeptide) disks and their precursors were characterized using matrix-assisted
laser desorption/ionization time-of-flight mass spectrometry (MALDI TOF MS). The C3-symmetrical disks
were based on a benzene-1,3,5-triscarboxamide core extended by oligopeptides with trialkoxyanilide
tails. The results indicate that MALDI TOF MS is a powerful and straightforward analytical technique
for characterizing C3-symmetrical disks and their precursors. Clear (pseudo)-molecular ion peaks could
readily be identified. It is remarkable that strong radical ion signals were observed for all the compounds,
including the anilines that were expected to be protonated prior to laser irradiation using acidic MALDI
matrixes. Possible mechanisms for radical ion formation were investigated with the employment of
radical scavengers, with various matrixes and with direct laser desorption/ionization (LDI). Most likely
the radicals are formed by losing one electron from the aniline nitrogen and stabilized by conjugation
through the phenyl ring. It appears that direct photo/thermal ionization of analytes is an important route
for the radical ion formation of the compounds with trialkoxy aniline/anilide groups. Copyright  2006
John Wiley & Sons, Ltd.
KEYWORDS: matrix-assisted laser desorption/ionization time-of-flight mass spectrometry; radical cation; C3-symmetric
disk; aniline; photo/thermal ionization
introduction in the late 1980s, matrix-assisted laser desorp-
tion/ionization time-of-flight mass spectrometry (MALDI
TOF MS) has found numerous applications in many
areas ranging from biomedical compounds to synthetic
polymers.13–15 In this work, four novel C3-symmetrical
disks together with their precursors, all with trialkoxy ani-
line/anilide tails, were analyzed using MALDI TOF MS.
One intriguing phenomenon observed in our experi-
ments is the emergence of strong radical ion peaks of analytes
that should be protonated in the condensed phase prior to
laser irradiation (see below). In MALDI TOF MS, usually not
radical ions, but protonated and/or metal ion adducts were
detected for a great majority of samples. Actually, only a
limited number of compounds with low ionization potential
(IP) values could yield radical ions in MALDI.16–21
It is well known that ionization in MALDI is a very
complicated process happening in the timescale of a few
nanoseconds. Many parameters could affect the outcome
of this process and, hence, the final ion species and their
abundance in an MS spectrum. A number of ionization
models have been developed aiming to improve the
knowledge of how the analyte ions are formed.18,22–25
In most of the models, it is assumed that photoionized
INTRODUCTION
Several C3-symmetrical disks synthesized in our group are
based on a benzene triscarboxamide core and are capable of
associating in supramolecular multivalent helical stacks.1–5
The dynamically ordered helix, which is one of the basic
structural motifs in proteins, has found new applications
in the fields of biology, pharmacy and material sciences.6–8
In the self-assembly process of the disk-shaped molecules,
hydrogen bonding, ꢀ ꢀ ꢀ interactions as well as solvophobic
effects play important roles. An appealing property of these
C3-symmetrical disks is that they can reversibly form helical
stacks consisting of hundreds of molecules both in the solid
state and in solution.
In many chemistry laboratories, characterization using
mass spectrometry (MS) has become imperative. This is
especially the case after the advent of the soft ionization
techniques of matrix-assisted laser desorption/ionization
(MALDI) and electrospray ionization (ESI).9–12 Since its
ŁCorrespondence to: Xianwen Lou, Laboratory of Macromolecular
and Organic Chemistry, Eindhoven University of Technology,
P.O. Box 513, 5600 MB Eindhoven, The Netherlands.
E-mail: x.w.lou@tue.nl
Copyright  2006 John Wiley & Sons, Ltd.

660
X. Lou et al.
matrix radical ions are profuse in the MALDI plume and
a key to ion generation. Interestingly though, most of the
models focused mainly on the generation of protonated
analyte ions or metal ion adducts. Fewer investigations
have been undertaken on the analyte radical ions, probably
because they are less frequently observed in MALDI. For
the analyte radicals, McCarley et al.17 proposed an electron-
transfer ionization mechanism, which could satisfactorily
explain radical ionization of analytes that cannot logically
undergo proton transfer reactions. In this mechanism,
they argued that matrixes absorbed the laser radiation
and became photoionized. Analyte radical ions were then
produced by molecule–ion reactions of neutral analytes with
matrix radical ions. Unfortunately, this mechanism could not
explain most of our experimental results.
The aim of this paper is twofold: Firstly, MALDI TOF
MS was used to analyze the C3-symmetrical disks and their
precursors. It is remarkable that clear radical ions were
observed for all the compounds analyzed. Secondly, possible
mechanisms for the radical ion generation were investigated
by a series of experiments, including the addition of radical
scavengers, the employment of an inorganic particle matrix,
the use of direct laser desorption ionization (LDI) and the
introduction of suppression effects of KC with the addition of
excessive amount of K2CO3. On the basis of the experimental
results, possible mechanisms for the radical ion formation
are discussed.
equipped with a 337-nm nitrogen laser, capable of executing
both linear and reflector modes. The accelerating voltage
was held at 20 kV for all experiments. Mass spectra were
acquired in the reflector mode by summing spectra from 100
selected laser shots and calibrated with oligothiophenes of
known masses. For the experiments with organic matrixes,
matrix solutions were freshly prepared in tetrahydrofuran
(THF) at a concentration of about 20 mg/ml, and sample
solutions at about 2 mg/ml. The sample solution and matrix
solution were mixed 1/1 in an Eppendorf tube, and about
0.5 µl of the mixed solution was pipetted onto a stainless
steel MALDI target plate. Samples prepared in this way
yielded a molar ratio of analyte to matrix of around 1 : 50.
Unless noted otherwise, this ratio is used throughout this
article. This is because similar results were observed for
radical cation formation in a wide range of molar ratios
from 1 : 10 to 1 : 1000 and a ratio of 1 : 50 was convenient
for sample preparation. When a radical scavenger was used,
the scavenger solution (also in THF at about 20 mg/ml) was
combined with the mixed solution of sample and matrix prior
to pipetting. The mole ratios of matrix : scavenger : analyte
were at about 50 : 50 : 1 for BHT and 50 : 25 : 1 for NTPBDA. In
the experiments with the ZnO matrix, the inorganic particle
was dispersed with methanol (20 mg/ml) in an ultrasonic
bath for 30 min. The solution was then spotted on the plate.
After evaporation of methanol, about 0.5 µl of the sample
solution was deposited on top of the inorganic matrix.
ESI MS measurements were carried out with an LCQ
Deca XP MAX instrument (Thermo Finnigan, San Jose, CA)
equipped with an ion trap. A MALDI solution (a mixed
solution of sample and the CHCA matrix, as above) was
first diluted by about 100 times with THF. The diluted
solution (5 µl) was then directly introduced into the ESI MS
via flow injection at a mobile phase flow rate of 20 µl/min.
Water/acetonitrile (1 : 1) was used as the mobile phase. ESI
MS was performed in the positive mode with a spray voltage
of 4.5 kV. The sheath gas flow rate was 40 ml/min and the
EXPERIMENTAL
The structures of compounds studied are depicted in
Fig. 1. The disk-shaped molecules were synthesized via
a convergent approach.1,26,27 First, a trialkoxy amine was
incorporated into an oligopeptide fragment. In case an
achiral glycine is adjacent to the trialkoxy moiety, the Boc-
protected dipeptide was directly linked to the trialkoxyamine
using O-(benzotriazol-1-yl)-N,N,N0,N0-tetramethyluronium
hexafluorophosphate (HBTU) as a coupling agent. When a
chiral phenylalanine is present next to the trialkoxy moiety,
step-by-step coupling of the Boc-protected amino acids is
required to prevent racemization. The amine-terminated
conjugates were obtained after removal of the protecting Boc-
group with TFA followed by a basic work-up. Finally, disk-
shaped molecules were obtained via coupling of the amine-
terminated dipeptide anilines to 1,3,5-benzenetricarbonyl
trichloride using triethylamine as the base.
Four organic and one inorganic matrixes were used in
this work. ˛-Cyano-4-hydroxy cinnamic acid (CHCA) and
2,5-dihydroxybenzoic acid (DHB) were obtained from Fluka
(Buchs, Switzerland), and dithranol and ZnO (nanopowder,
particle size <1 µm) from Aldrich (Steinheim, Germany). 2-
[(2Eꢁ-3-(4-tert-Butylphenyl)-2-methylprop-2-enylidene] mal-
ononitrile (DCTB) was synthesized according to Ulmer
et al.,28 who used it for the MALDI TOF MS analysis
of some substituted fullerenes. Butylated hydroxytoluene
(BHT) was received from Aldrich, and the other radi-
cal scavenger, N, N, N0,N0-tetraphenyl-1,4-benzenediamine
(NTPBDA), was kindly provided by Dr M. M. Wienk.
°
capillary temperature 275 C.
RESULTS AND DISCUSSION
MALDI TOF MS analysis of C3-symmetrical disks
and their precursors
In this section, two matrixes (CHCA and DCTB) were
employed for all the compounds. CHCA is one of the
most widely used MALDI matrixes, and DCTB28,29 is a
nonprotic matrix that has demonstrated good performance
for some labile samples. According to their structures (Fig. 1)
and stages in the synthesis, the compounds studied here
were arbitrarily divided into four groups: trialkoxy anilines
(TAAs), oligopeptide anilides (OPAs), Boc-protection OPAs
(Boc-OPAs) and C3-symmetrical disks (SDs). The values of
IP, proton affinity (PA) and pKa of CHCA are listed in
Table 1. Unfortunately, these values are not available for
DCTB and for the SDs and their precursors. In Table 1, the
values of reference compound, aniline, are also given. Some
qualitative information for the analytes could be deduced by
comparing their structures with that of aniline. For example,
the IP values of the TAAs should be lower than that of aniline
because of the conjugation through the substitution groups.
The MALDI TOF MS measurements were performed
with a Voyager-DE Pro (PerSeptive Biosystems) instrument
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 659–669

MALDI TOF MS of C3-symmetric disks and precursors
661
Trialkoxy anilines
1. R = O6C12H25
2. R = O6C11H23
=
=
O
O
O
O
O
O
O
O
O
O
O
O
O
R
R
H2N
R
3. R = OC10H21
=
TAA1-4
4. R = OC8H17
=
O
R
C -symmetrical disks
Oligo-peptide anilides
3
R
R
R
R
R
R1
O
HN
H
N
H
N
O
R1
N
O
H
R2
O
O
O
NH
R
Boc-OPA1-3
R
R2
R2
O
O
HN
H
N
N
H
O
R
R
N
O
O
R1
O
H
H
R1
N
R
H2N
N
H
R2
O
R
O
NH
SD1-4
R2
OPA1-3
HN
O
R1
1
2
1. R = CH Ph; R = H; R = OC H
2
8
17
NH
1
2
2. R = CH Ph; R = CH Ph; R = OC H
2
2
8
17
1
2
3. R = H; R = H; R = OC
H
10 21
R
R
1
2
4. R = H; R = CH Ph; R = OC H
8 17
2
R
Reference compounds
H2N
O6C12H25
O(CH2)11CH3
H
N
O
O
H
N
O
OH
O6C12H25
O6C12H25
O(CH2)11CH3
O(CH2)11CH3
N
H
O
O
O
O
D1
D2
D3
Figure 1. Structures of the C3-symmetrical disks and their precursors.
Table 1. Values of ionization potential (IP), proton
affinity (PA) and pK of CHCA and anilinea
For the TAAs good MALDI TOF MS spectra were
obtained with both matrixes tested. Figure 2 shows some
typical spectra using CHCA as the matrix. In Fig. 2,
signals of the radical and alkali ion adducts can readily
be identified. CHCA is an acidic matrix. With this matrix,
the anilines should be protonated prior to laser irradiation.
One would expect that the protonated species should be the
major peaks in the MS spectra. In contrast, however, the
protonated ion peaks were found to be much smaller than
their corresponding radical ion peaks, which is evident by
comparing the recorded spectra with the simulated ones.
The reason for this will be discussed later in this article.
Some typical MS spectra with DCTB are shown in Fig. 3.
IP (eV)
PA (kJ/mol)
pK
CHCA
Aniline
9.41
7.70
841
884
2.6 (pKaꢁ
9.3 (pKbꢁ
a The IP values were taken from Refs 30 and 33; the PA
values from Ref. 31; and pK values from Refs 32 and 34,
respectively.
It should also be pointed out here that the accuracy and
applicability of the values might be open to question.22
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 659–669

662
X. Lou et al.
Exp.
Exp.
+
.+
.+
[TAA2+Na]
[TAA2]
[TAA1]
B
A
884
884
886
886
888
890
842
842
844
844
846
848
m/z
m/z
+
[TAA1+Na]
.+
.+
[TAA2] Calc.
[TAA1] Calc.
+
[TAA2+K]
846
848
888
890
m/z
m/z
800
850
900
950
1000
750
800
850
900
950
m/z
m/z
Exp.
Exp.
[TAA4].+
[TAA3].+
560
560
562
m/z
564
.+
C
D
476
476
478
480
m/z
[TAA3] Calc.
.+
[TAA4] Calc.
562
m/z
564
478
m/z
480
500
550
600
650
700
400
450
500
550
600
m/z
m/z
Figure 2. MALDI TOF MS spectra of the trialkoxy anilines using the CHCA matrix together with simulated spectra. (A) TAA1;
(B) TAA2; (C) TAA3; (D) TAA4. The analyte structures are shown in Fig. 1.
.+
.+
[TAA1]
[TAA2]
A
B
+
[TAA1+Na]
+
[TAA2+Na]
800
850
900
950
1000
750
800
850
900
950
m/z
m/z
.+
[TAA3]
.+
[TAA4]
C
D
500
550
600
650
700
400
450
500
550
600
m/z
m/z
Figure 3. MALDI TOF MS spectra of the trialkoxy anilines with the DCTB matrix. (A) TAA1; (B) TAA2; (C) TAA3; (D) TAA4. The analyte
structures are shown in Fig. 1.
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 659–669

MALDI TOF MS of C3-symmetric disks and precursors
663
Table 2. Intensity ratios between the analyte radical cation
and the protonated ion ([A]C/[A + H]C) with the CHCA and
DCTB matrixes
As expected, with the nonacidic matrix, the protonated ion
signals became much weaker or even negligible compared
with the radical signals.
The (Boc)-OPAs are based on TAAs and different
combinations of glycine (G) and phenylalanine (F) with and
without Boc-protection of the amino group. Good MALDI
TOF MS spectra were obtained with both CHCA and DCTB
matrixes. Some typical spectra of the (Boc)-OPAs using the
CHCA matrix are given in Fig. 4. Like for the TAAs, strong
radical signals were also observed even for the OPAs, which
contain free aliphatic amines that were not protected by the
Boc-group.
For the SDs, three identical oligopeptides were attached
onto a benzene-1,3,5-tricarbonyl core. Here again, good
MALDI TOF MS spectra were obtained with both the
matrixes tested. Some typical MS spectra of the disks using
the CHCA matrix are given in Fig. 5. Clear radical ion peaks
were observed for all the SDs.
Trialkoxy anilines (TAAs)
TAA1 TAA2 TAA3
TAA4
6.9
–
CHCA 3.2
2.5
–
6.8
–
a
DCTB
ꢀ
Oligopeptide anilides (OPAs and Boc-OPAs)
OPA1 OPA2 OPA3 BocOPA1 BocOPA2 BocOPA3
CHCA 6.1
DCTB
5.7
–
1.2
–
–
–
–
–
–
–
–
C3-symmetrical disks (SDs)
SD1
SD2
SD3
SD4
–
–
CHCA
DCTB
–
–
–
–
–
–
a Strong analyte radical cations and negligible protonated
analyte ions.
In summary, good MS spectra were obtained for all
the compounds tested. It can be concluded that MALDI
TOF MS is a useful and straightforward method for
the characterization of these compounds. One striking
observation is that even with an acidic matrix, the free amines
display prominent radical ion peaks. Protonated species, in
contrast, were found to be more scarce or even negligible in
the MALDI spectra. The peak intensities of radicals relative
to their corresponding protonated molecules are listed in
Table 2. The ratios were calculated on the basis of the
theoretical isotopic distributions of the given molecules. By
comparing the relative peak intensities of the mono-isotopic
mass peak of [A]C and [A C 1]C in the experimental spectra
with those of the theoretical values, the ratio of radicals to
protonated ions can easily be calculated. The values listed
in Table 2 must be regarded, however, as tentative since
the signal intensities are not always very reproducible in
MALDI. Nevertheless, it can clearly be seen that (1) with the
acidic matrix (CHCA), protonated ions were observed only
for compounds containing free amines (TAAs 1–4 and OPAs
1–3), and even under these circumstances the radical ions
were more abundant than their protonated counterparts;
(2) with the nonprotic matrix (DCTB), protonated ions were
found to be negligible for all the compounds; and (3) for
+
[OPA1+Na]
+
[OPA3+Na]
A
B
.+
[OPA3]
.+
[OPA1]
600
650
700
750
800
600
650
700
750
800
m/z
m/z
+
.+
.+
[BocOPA3+Na]
[BocOPA1]
[BocOPA3]
+
[BocOPA1+Na]
C
D
+
+
[BocOPA1+K]
[BocOPA3+K]
700
750
800
850
900
700
750
800
850
900
m/z
m/z
Figure 4. MALDI TOF MS spectra of oligopeptides with trialkoxy anilide end group using the CHCA matrix. (A) OPA1; (B) OPA3;
(C) BocOPA1; (D) BocOPA3. The analyte structures are shown in Fig. 1.
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 659–669

664
X. Lou et al.
+
[SD2+Na]
+
A
B
[SD1+Na]
.+
[SD1]
+
.+
[SD1+K]
[SD2]
+
[SD2+K]
2100
2200
2300
2400
2300
2400
2500
2600
2700
m/z
m/z
+
+
C
D
[SD4+Na]
[SD3+Na]
+
.+
[SD4+K]
[SD4]
.+
+
[SD3]
[SD3+K]
2000
2100
2200
2300
2400
2100
2200
2300
2400
m/z
m/z
Figure 5. MALDI TOF MS spectra of the C3-symmetrical disks with the CHCA matrix. (A) SD1; (B) SD2; (C) SD3; (D) SD4. The
analyte structures are shown in Fig. 1.
compounds without a free amine (Boc-OPAs 1–3 and SDs
1–4), no significant amount of protonated ions were observed
with either matrix tested. Contrary to the protonated ions,
clear radical cation peaks were recorded in all experiments.
Some possible mechanisms for the formation of radical ions
will be discussed in the following section.
compounds D1, D2 and D3 with the CHCA matrix are shown
in Fig. 6.
Although both aniline and anilide groups could lead to
the generation of radical ions, they exhibit significantly dif-
ferent ability to yield protonated ions, which is evident by
comparing the ionization behaviors of TAAs and Boc-OPAs.
When CHCA was used as the matrix, clear protonated ions,
though considerably weaker than their corresponding radi-
cal ions, were observed for the TAAs. In contrast, however,
protonated ions were found to be negligible for the Boc-
OPAs (Table 2). The dramatic difference between these two
types of compounds can be explained by studying the sta-
bility of the radical ions formed and the basicity of the
molecules. Compared with the TAA radicals, the Boc-OPA
radicals are more stable. This is because in addition to con-
jugation through the aromatic ring, the Boc-OPA radicals
initiated by losing one electron from the anilide nitrogen
are further stabilized by conjugation through the nearby car-
bonyl. Furthermore, the basicity of the Boc-OPAs is much
weaker than that of the TAAs. The combined effects of stable
radical and weak basicity result in the lack of protonated
ions for the Boc-OPAs. With similar reasoning, the absence
or presence of protonated ions for the other two series
of samples studied (SDs and OPAs) could also be under-
stood. Obviously, no protonated ions were observed for the
SDs, which contain three oligopeptide anilides attached to a
benzene-1,3,5-triscarboxamide core. The SD radical ions can
be stabilized at least as well as the Boc-OPAs. For the OPAs,
on the other hand, protonated ions were detected because
of the presence of free aliphatic amines in the molecule. In
the following experiments, efforts were mainly focused on
POSSIBLE MECHANISMS OF RADICAL
CATION FORMATION
In the previous section, the observation of strong radical
cation peaks in MALDI TOF MS of the C3-symmetrical
disks and their precursors has been discussed. Apart
from the trialkoxy aniline/anilide moiety, the structures
of the compounds are quite different (Fig. 1). Therefore,
the formation of radical ions must, at least partly, be
related with the aniline/anilide group. This inference was
further supported by the MALDI TOF MS analysis of FFBoc
(Fig. 1, D1), which is one of the small peptides used to
make our OPAs by combining with trialkoxyaniline. It has
been demonstrated above that OPAs showed strong radical
signals. However, without the anilide group, FFBoc gave no
radical ions. This suggests that radicals are indeed formed via
the aniline/anilide group. Moreover, for the compounds in
which the NH2 group was replaced by, for example, carbonyl
derivatives (compounds D2 and D3 in Fig. 1), no clear radical
ions were detected either. Therefore, it can be concluded that
the aniline/anilide group is essential for the radical ion
formation. The radical ions are most likely developed by the
loss of one electron from the N-atom and then stabilized by
delocalization through the benzene ring. The MS spectra of
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 659–669

MALDI TOF MS of C3-symmetric disks and precursors
665
[TAA1+H]+
886.6
+
[D1+Na]
A
[TAA1+Na]+
908.6
+
[D1+K]
[TAA1+K]+
924.6
0
400
420
440
460
480
500
m/z
870 880 890 900 910 920 930
m/z
B
+
Figure 7. ESI MS spectrum of compound TAA1. The analyte
[D2+Na]
structure is shown in Fig. 1.
Table 3. Effects of radical scavenger addition on the ion
intensity ratios between analyte radical cation and protonated
ž
ion ([A] C/[A C H]Cꢁ with the CHCA matrix
+
[D2+K]
Compounds
TAA1
TAA2
TAA3
TAA4
No scavenger
With BHT
3.2
3.4
1.9
2.5
2.8
0.75
6.8
6.7
2.0
6.9
6.9
1.6
900
920
940
960
+
980
1000
m/z
With NTPBDAa
C
a N, N, N0,N0-tetraphenyl-1,4-benzenediamine.
[D3+H]
analytes and matrix radical ions, the MALDI samples were
mixed with a concentrated radical scavenger solution. The
addition of excessive amounts of radical scavenger would
quench the electron-transfer reaction between the analyte
molecules and the matrix radicals. As a result, the radical
ion peaks should disappear completely or at least become
much weaker. In this study, BHT and NTPBDA were
selected as the scavengers. BHT is a widely used antioxidant
radical scavenger,35,36 and NTPBDA is a compound with
an extremely low IP value.37 Surprisingly, no considerable
influence was observed on the formation of analyte radicals
with the addition of a far excessive amount of BHT, e.g. at a
mole ratio of about 50 : 50 : 1 (matrix : BHT : analyte). For all
the compounds tested, strong radical ion signals could still be
observed (Fig. 8). Moreover, the [A]žC/[A C H]C ratio of the
anilines remained literally unchanged (Table 3). Somewhat
different results were obtained with NTPBDA. As expected,
the analyte ion signals, including the alkali ion adducts,
were significantly suppressed by the excessive amount of
NTPBDA, which is very easy to be ionized. However, analyte
radical ions could still be clearly observed. Comparison of the
data listed in Table 3 shows that the [A]žC/[A C H]C ratios
are slightly lower with the addition of NTPBDA. However,
this should not be interpreted as a support for the electron-
transfer mechanism because, otherwise, the ratio should
be much lower or the aniline radicals should even have
disappeared completely. The lower ratios might be explained
by electron-transfer between the analyte radical and the
NTPBDA scavenger with a lower IP value. Furthermore,
on the basis of the electron-transfer mechanism alone, it is
difficult to explain why only small peaks were observed for
protonated ions that were already preformed with an acidic
+
[D3+Na]
700
720
740
760
780
800
m/z
Figure 6. MALDI TOF MS spectra of D1, D2 and D3 with the
CHCA matrix. (A) D1; (B) D2; (C) D3. The analyte structures are
shown in Fig. 1.
the TAAs. This is because TAAs have the simplest structures
among the four groups of compounds (Fig. 1), and could
yield clear ions of both protonated and radical ion species.
Protonated ions were found to be useful and could serve as
references in the study of effects of the various experimental
parameters on the generation of radical ions.
In the condensed phase with an acidic matrix, the
TAAs should be protonated prior to laser irradiation. Direct
measurements of some MALDI solutions (sample mixed
together with the CHCA matrix) with ESI MS clearly
show strong protonated ions. An ESI MS spectrum of
TAA1 is shown in Fig. 7. From the ESI measurements, the
following two conclusions can be drawn: (1) protonated
trialkoxyaniline ions are stable under MS conditions; and
(2) the formation of trialkoxyaniline radical ions is due to the
MALDI process.
In order to find out whether the radical ions are
formed by the electron-transfer reaction between neutral
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 659–669

666
X. Lou et al.
[TAA1+Na]+
[TAA1+K]+
B
[TAA4].+
A
[TAA1].+
800
850
900
950
1000
400
450
500
550
600
m/z
m/z
[TAA1+Na]+
[TAA1+K]+
D
C
[TAA1].+
[TAA4].+
800
850
900
950
1000
400
450
500
550
600
m/z
m/z
Figure 8. MALDI TOF MS spectra of TAA1 and TAA4 with the addition of radical scavengers. CHCA was used as the matrix. The
mole ratio of matrix/scavenger/analyte is at about 50/50/1. (A) TAA1 with BHT; (B) TAA4 with BHT; (C) TAA1 with NTPBDA; (D) TAA2
with NTPBDA. The analyte structures are shown in Fig. 1.
matrix. From the results above, it appears that there are some
other mechanisms responsible for the radical ion formation.
A possible route for the analyte radical ion formation
is the direct photo-ionization upon laser irradiation. The
anilines should have lower IP values than the matrixes used.
Otherwise, no strong radical ion peaks would be detected
since radicals of an analyte with high IP could not survive
the collision with the neutral matrix molecules in the gas-
phase plume. However, direct LDI of the anilines yielded no
ions when conditions identical to those employed in MALDI
were used. Much higher laser power is required in LDI to
yield detectable ions. It is arguable that the analytes might
be photoionized but difficult to be desorbed without matrix
assistance. A matrix might provide a good environment
facilitating the radical ion desorption. Anyway, on the basis
of the dramatic difference between the results obtained by
LDI and MALDI, it is evident that the matrix plays a crucial
role for the radical ion formation/desorption of the TAAs.
In MALDI, a matrix absorbs laser energy and transfers
most of it into thermal energy. Because of the low thermal
conductivity of the organic solids, a huge temperature
increase can be attained, which may play a key role
in the various processes following the laser adsorption.38
The surface peak temperature of MALDI matrixes under
laser irradiation is normally around 1000 K.38,39 For some
compounds, the shortage of photon energy for direct photo-
ionization might be compensated for by the thermal energy
of the hot MALDI plume. Allwood et al.40,41 developed a
model based on thermal emission of electrons from photo-
excited molecules for the formation of radical matrix ions.
They claimed that this process might also be a plausible route
for matrix ionization in MALDI.
H
2[Matrix]Ł ꢀꢀꢀ! [Matrix]žC C [Matrix]žꢀ
ꢂ1ꢁ
The anilines studied here have low IP values.33,42,43 It is
reasonable to assume that not only the matrix molecules but
also the anilines could be ionized through the photo/thermal
mechanism.
H
[Analyte]Ł ꢀꢀꢀ! [Analyte]žC C e
ꢂ2ꢁ
For a given compound, whether a radical ion could
be formed through photo- or photo/thermal ionization
depends on its IP value. Unfortunately, the actual IP values
in solid crystals or large matrix aggregates are still largely
unknown.22
Supporting evidence for the photo-ionization mechanism
was also found by two more series of independent experi-
ments. In the first series, the trialkoxyanilines were studied
with an inorganic particle matrix (ZnO) and with direct LDI.
Clear radical ions could be observed in both cases. Some typ-
ical MS spectra are shown in Fig. 9. It should be noted that
significantly higher laser intensities were required with ZnO
or in LDI than that with a normal organic matrix. Evidently,
with the inorganic matrix or in LDI, the radical ion formation
could not occur through electron-transfer by molecule–ion
reactions between neutral analytes and matrix radicals.
In the second series of experiments, excessive amount of
K2CO3 was added as the cationization reagent (K2CO3/ana-
lyte at about 5/1 (mole/mole)). Figure 10 shows some
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 659–669

MALDI TOF MS of C3-symmetric disks and precursors
667
[TAA3].+
A
A
[TAA2+K]+
[TAA3+Na]+
[TAA3+K]+
[TAA2].+
750
800
850
900
950
500
550
600
650
700
m/z
m/z
[BocOPA3+K]+
B
B
[TAA3].+
[BocOPA3].+
[TAA3+Na]+
700
750
800
850
900
500
550
600
650
700
m/z
m/z
Figure 9. MS spectra of TAA3 with ZnO matrix and LDI. (A) LDI;
(B) with ZnO matrix. The analyte structure is shown in Fig. 1.
C
[SD2+K]+
typical MALDI TOF MS spectra with the addition of K2CO3.
Although radical ions, and sodium and potassium adducts
could be observed without the addition of any cationization
reagent, the main goal of these experiments was to see
how the generation of the radical ion and sodium adduct
would be affected by adding K2CO3. In the literature,
the formation of cation adducts was generally assumed
and also demonstrated to be mainly through gas-phase
reactions.22–25,44 Because of the excessive amount of KC
introduced, gas-phase reactions other than KC adduction
would be suppressed. Indeed, by comparing Fig. 10 with
Figs 2, 4 and 5, it is clear that the sodium adduct cation
peaks, which were originally much more intensive than their
corresponding potassium adduct peaks, almost completely
disappeared. If the radical ions had also been formed in the
gas-phase through molecule–ion reactions, their formation
should be greatly suppressed too. Contrarily, however,
radical cation signals could still be observed clearly. In
compound SD2, for example, the NaC adduct peak was far
stronger than the radical ion peak without the addition of KC
(Fig. 5(B)). After the addition of KC, the sodium adduct peak
had completely disappeared while the radical cation could
still be observed clearly (Fig. 10(C)). The results suggest that
radical ions might already be formed through photo- or
photo/thermal ionization prior to the plume reactions.
By using the photo- or photo/thermal ionization mecha-
nism, the weakening of protonated ion peaks could also be
reasonably explained. It is widely recognized that generation
of ions in MALDI is a competing process. For the anilines with
an acidic matrix, the protonated species were already pre-
formed prior to laser irradiation, and intensive protonated
[SD2].+
2400
2300
2500
2600
2700
m/z
Figure 10. Typical MALDI TOF MS spectra with the addition of
potassium carbonate. CHCA was used as the matrix. The mole
ratio of matrix/KC/analyte is at about 50/5/1. (A) TAA2;
(B) BocOPA3; (C) SD2. The analyte structures are shown in
Fig. 1.
ion peaks should be expected as in the case of many com-
pounds containing an amino group. Contrarily, however, the
protonated ion peaks were found to be quite small. This is
probably due to the radical ion formation through photo- or
photo/thermal ionization. Obviously, releasing of anilines
as protonated ions straight from the condensed phase would
be significantly suppressed when the analyte molecules had
already been photo- or photo/thermally ionized as radical
cations.
In addition to CHCA and DCTB, two more matrixes
were studied. These were DHB and dithranol. CHCA and
DHB are acidic, while DCTB and dithranol are more or less
neutral. In order to neutralize the effects of matrix acidity,
a large amount of succinic acid was added in the MALDI
sample. Typical mole ratios of matrix : succinic acid : analyte
were around 50 : 100 : 1. After the addition of succinic acid,
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 659–669

668
X. Lou et al.
the final pH values of the MALDI solutions with different
matrixes should be similar. Figure 11 shows some typical
MS spectra of TAA2 with these matrixes. Similar results
were obtained with other TAAs. Although the MALDI spot
became less homogeneous after the addition of succinic acid,
for a given compound the ratio of [A]žC/[A C H]C can
be quite different with a different matrix. CHCA gave the
highest ratio, presumably due to its ‘hotter’ nature compared
with DHB and dithranol.45,46 The results indicate that a
hot matrix might favor the formation/desorption of analyte
radical ions.
Judged from the results illustrated above, it seems that
photo-ionization or photo/thermal ionization is a vital
route for the radical ion formation of the compounds
studied. This mechanism might also be applicable to other
molecules with low IP values. However, it should be
emphasized here that besides the radical ions, protonated
and cationized pseudomolecular ions were also observed.
The isotopic patterns of NaC and KC adducts were found
to be fully compatible with their corresponding theoretical
distributions. Apparently, ionization in MALDI is a very
complicated process that might involve several different
ionization mechanisms.
CONCLUSIONS
MALDI TOF MS is a powerful and straightforward tech-
nique for the characterization of novel C3-symmetrical
tris(dipeptide) disks and their precursors. For all the com-
pounds tested, radical ion peaks and cationized pseudo-
molecular ion peaks could readily be identified. The radicals
were probably formed by losing one electron from the ani-
line/anilide nitrogen and stabilized by conjugation through
the phenyl ring. It appears that direct ionization of the
analytes by a photo- or photo/thermal mechanism is an
important route for the radical ion formation. Important
roles of the matrix for the generation of radical ions might
include providing a good environment for desorption and
making up the energy difference between photon excitation
and analyte IP.
[TAA2+Na]+
[TAA2+K]+
A
[TAA2].+
Acknowledgements
The authors are grateful to Dr M. M. Wienk and Prof. R. A. J. Janssen
for their helpful discussions. Dr M. M. Wienk is further acknowl-
edged for kindly providing NTPBDA. We also thank Dr S. Langereis
for transferring all the figures into the EPS format.
750
750
750
800
850
900
950
m/z
B
REFERENCES
1. van Gorp JJ, Vekemans JAJM, Meijer EW. C3-symmetric
supramolecular architectures: fibers and organic gels from
discotic trisamides and trisureas. J. Am. Chem. Soc. 2002; 124:
14759.
2. Brunsveld L, Schenning APHJ, Broeren MAC, Janssen HM,
Vekemans JAJM, Meijer EW. Chiral amplification in columns
of self-assembled N,N’,N’’-tris[(S)-3,7-dimethyloctyl]benzene-
1,3,5-tricarboxamide in dilute solution. Chem. Lett. 2000; 292.
3. Palmans ARA, Vekemans JAJM, Hikmet RA, Fisher H,
Meijer EW. Lyotropic liquid-crystalline behavior in disc-shaped
compounds incorporating the 3,30-di(acylamino)-2,20-bipyridine
unit. Adv. Mater. 1998; 10: 873.
4. Palmans ARA, Vekemans JAJM, Havinga EE, Meijer EW.
Sergeants-and-soldiers principle in chiral columnar stacks of
disk-shaped molecules with C3 symmetry. Angew. Chem., Int.
Ed. Engl. 1997; 36: 2648.
5. Brunsveld L, Zhang H, Glasbeek M, Vekemans JAJM, Meijer EW.
Hierarchical growth of chiral self-assembled structures in protic
media. J. Am. Chem. Soc. 2000; 122: 6175.
[TAA2+Na]+
[TAA2+K]+
[TAA2].+
800
850
900
950
m/z
C
[TAA2+Na]+
[TAA2+K]+
[TAA2].+
6. Brunsveld L, Folmer BJB, Meijer EW, Sijbesma RP. Supramolec-
ular polymers. Chem. Rev. 2001; 101: 4071.
7. Mateos-Timoneda MA, Crego-Calama M, Reinhoudt DN.
Supramolecular chirality of self-assembled systems in solution.
Chem. Soc. Rev. 2004; 33: 363.
800
850
900
950
8. van Bommel KJC, van der Pol C, Muizebelt I, Friggeri A,
Heeres A, Meetsma A, Feringa BL, van Esch J. Responsive
cyclohexane-based low-molecular-weight hydrogelators with
modular architecture. Angew. Chem., Int. Ed. Engl. 2004; 43:
1663.
9. Cole RB (ed). Electrospray Ionization Mass Spectrometry:
Fundamentals, Instrumentation, and Applications. John Wiley and
Sons: New York, 1997.
m/z
Figure 11. Comparison of MALDI TOF MS spectra of TAA2
using different matrixes with the addition of succinic acid. The
mole ratio of matrix/succinic acid/analyte is at about 50/100/1.
(A) CHCA; (B) DHB; (C) dithranol. The analyte structure is
shown in Fig. 1.
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 659–669

MALDI TOF MS of C3-symmetric disks and precursors
669
10. Yamashita M, Fenn JB. Electrospray ion source. Another
variation on the free-jet theme. J. Phys. Chem. 1984; 88: 4451.
11. Karas M, Bachmann D, Bahr U, Hillenkamp F. Matrix-assisted
ultraviolet laser desorption of non-volatile compounds. Int. J.
Mass Spectrom. Ion Processes 1987; 78: 53.
29. Lou X, van Buijtenen J, Bastiaansen JJAM, de Waal BFM,
Langeveld BMW, van Dongen JLJ. Characterization of some
synthetic Ru and Ir complexes by matrix-assisted laser
desorption/ionization time-of-flight mass spectrometry. J. Mass
Spectrom. 2005; 40: 654.
12. Tanaka K, Waki H, Ido Y, Akita S, Yoshida Y, Yoshida T. Protein
and polymer analyses up to m/z 100 000 by laser ionization time-
of-flight mass spectrometry. Rapid Commun. Mass Spectrom. 1988;
2: 151.
13. Burlingame AL, Boyd RK, Gaskell SJ. Mass spectrometry. Anal.
Chem. 1998; 70: 647R.
14. Nielen MWF. MALDI time-of-flight mass spectrometry of
synthetic polymers. Mass Spectrom. Rev. 1999; 18: 309.
15. Farmer TB, Caprioli RM. Determination of protein-protein
interactions by matrix-assisted laser desorption/ionization mass
spectrometry. J. Mass Spectrom. 1998; 33: 697.
30. Sumner LW, Huang Y, Russell DH. MALDI internal energy
transfer studies utilizing molecular modeling. Proceedings of the
44th ASMS Conference on Mass Spectrometry, Portland, 1996; 271.
31. Jorgensen TJD, Bojesen G, Rahbek-Nielsen H. The proton
affinities of seven matrix-assisted laser desorption/ionization
matrices correlated with the formation of multiply charged ions.
Eur. Mass Spectrom. 1998; 4: 39.
32. Sipila J, Nurmi H, Kaukonen AM, Hirvonen J, Taskinen J,
Yli-Kauhaluoma J. A modification of the Hammett equation for
predicting ionization constants of p-vinyl phenols. Eur. J. Pharm.
Sci. 2005; 25: 417.
16. Juhasz P, Costello CE. Generation of large radical ions
from oligometallocenes by matrix-assisted laser desorp-
tion/ionization. Rapid Commun. Mass Spectrom. 1993; 7: 343.
17. McCarley TD, McCarley RL, Limbach PA. Electron-transfer
ionization in matrix-assisted laser desorption/ionization mass
spectrometry. Anal. Chem. 1998; 70: 4376.
18. Ehring H, Karas M, Hillenkamp F. Role of photoionization and
photochemistry in ionization processes of organic molecules and
relevance for matrix-assisted laser desorption ionization mass
spectrometry. Org. Mass Spectrom. 1992; 27: 472.
33. Farrell PG, Newton J. Ionization potentials of aromatic amines.
J. Phys. Chem. 1965; 10: 3506.
34. Budavari S, O’Neill MJ, Smith A, Heckelman PE (eds). The Merck
Index, 8th ed. Merck and Company: Rahway, NJ, 1989; 104.
35. Fujisawa S, Kadoma Y, Yokoe I. Radical scavenging activity of
butylated hydroxytoluene (BHT) and its metabolites. Chem. Phys.
Lipids 2004; 130: 189.
36. Green TR, Bennett SR, Nelson VM. Specificity and properties of
propofol as an antioxidant free radical scavenger. Toxicol. Appl.
Pharmacol. 1994; 129: 163.
19. Lidgard RO, McConnell DB, Black DSC, Kumar N, Duncan MW.
Fragmentation observed in continuous extraction linear MALDI:
a cautionary note. J. Mass Spectrom. 1996; 31: 1443.
37. Wienk MM. High-Spin molecules of phosphorus and nitrogen-
centered radicals, Ph. D. Thesis, Technical University of
Eindhoven, Eindhoven, 1997.
20. Rader HJ, Spickermann J, Kreyenschmidt M, Mullen K. MALDI-
TOF mass spectrometry in polymer analysis, molecular weight
analysis of rigid rod polymers. Macromol. Chem. Phys. 1996; 197:
3285.
21. Green MK, Medforth CJ, Muzzi CM, Nurco DJ, Shea KM,
Smith KM, Lebrilla CB, Shelnutt JA. Application of matrix-
assisted laser desorption fourier transfer mass spectrometry
to the analysis of planar porphyrins and highly substituted
nonplanar porphyrins. Eur. Mass Spectrom. 1997; 3: 439.
22. Zenobi R, Knochenmuss R. Ion formation in MALDI mass
spectrometry. Mass Spectrom. Rev. 1998; 17: 337.
23. Karas M, Gluckmann M, Schafer J. Ionization in matrix-assisted
laser desorption/ionization: singly charged molecule ions are
the lucky survivors. J. Mass Spectrom. 2000; 35: 1.
24. Knochenmuss R, Zenobi R. MALDI ionization: the role of in-
plume processes. Chem. Rev. 2003; 103: 441.
38. Koubenakis A, Frankevich V, Zhang J, Zenobi R. Time-resolved
surface temperature measurement of MALDI matrices under
pulsed UV laser irradiation. J. Phys. Chem. A 2004; 108: 2405.
39. Mowry CD, Johnston MV. Internal energy of neutral molecules
ejected by matrix-assisted laser desorption. J. Phys. Chem. 1994;
98: 1904.
40. Allwood DA, Dyer PE, Dreyfus RW. Ionization modeling
of matrix molecules in ultraviolet matrix-assisted laser
desorption/ionization. Rapid Commun. Mass Spectrom. 1997; 11:
499.
41. Allwood DA, Dyer PE, Dreyfus RW, Perera IK. Plasma
modeling of UV matrix assisted laser desorption ionization
(MALDI). Appl. Surf. Sci. 1997; 109/110: 616.
42. Birks JB. Photophysics of Aromatic Molecules. Wiley: New York,
1970.
43. Ogawa T, Ogawa T, Nakashima K. Ionization of anilines in
nonpolar hydrocarbon solvents as studied by two-photon
ionization spectra. J. Phys. Chem. 1998; 102: 10608.
25. Karas M, Kruger R. Ion formation in MALDI: the cluster
ionization mechanism. Chem. Rev. 2003; 103: 427.
26. Miller TM, Kwock EW, Neenan TX. Synthesis of four
generations of monodisperse aryl ester dendrimers based on
1,3,5-benzenetricarboxylic acid. Macromolecules 1992; 25: 3143.
27. Sinkeldam RW, van Houtem MHCJ, Vekemans JAJM, Meijer
EW. Synthesis of 3,6-diaminophthalimides for ureidoph-
thalimide-based foldamers. Org. Lett. 2006; 8: 383.
28. Ulmer L, Torres-Gacia HG, Mattay J, Luftmann H. The
use of 2-[(2E)-3-(4-tert-butylphenyl)-2-methylprop-2-enylidene]
malononitrile as matrix for matrix-assisted laser desorp-
tion/ionization mass spectrometry. Eur. J. Mass Spectrom. 2000;
6: 49.
44. Rashidezadeh H, Guo B. Investigation of metal attachment to
polystyrenes in matrix-assisted laser desorption ionization.
J. Am. Soc. Mass Spectrom. 1998; 9: 724.
45. Juhasz P, Vestal ML, Martin SA. On the initial velocity of ions
generated by matrix-assisted laser desorption ionization and
its effects on the calibration of delayed extraction time-of-flight
mass spectrometry. J. Am. Soc. Mass Spectrom. 1997; 8: 209.
46. Gluckmann M, Karas M. The initial ion velocity and its
dependence on matrix, analyte and preparation method in
ultraviolet matrix-assisted laser desorption/ionization. J. Mass
Spectrom. 1999; 34: 467.
Copyright  2006 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2006; 41: 659–669