Differentiating between capillary and counter electrode processes during electrospray ionization by opening the short circuit at the collector

Article abstract of DOI:10.1021/ac980799l

The chemical reactivity of an electroactive neutral compound in the conditions of electrospray mass spectrometry was studied by carrying out preparative experiments. The residue obtained by spraying a micromolar methanolic solution of 9-nitroanthracene onto a metallic plate in negative mode (cathodic capillary) was collected and then analyzed by GC/MS, showing that both reduction and oxidation products were formed. To distinguish between the two processes, the two functions of the metallic plate, definition of the electric potential and storage of products, were separated by using a copper-bronze grid as counter electrode and a silica TLC plate, located a few millimeters behind, as collector. The collected products were thus stored away from the counter electrode, avoiding their possible reoxidation at this site. The amount of anthracene, a reduction product of 9-nitroanthracene, which was detected only in trace amounts with the metallic plate collector, increased significantly with the new device. Thus, anthracene can be viewed as the result of a capillary process whereas 9,10-anthraquinone can be estimated as mainly due to further oxidation at the collector. Furthermore, in deuterated solvents no incorporation was detected, suggesting that some reaction steps take place at the collector.

Full text of DOI:10.1021/ac980799l

Anal. Chem. 1999, 71, 1585-1591
Diffe re n t ia t in g b e t w e e n Ca p illa ry a n d Co u n t e r
Ele c t ro d e P ro c e s s e s d u rin g Ele c t ro s p ra y
Io n iza t io n b y Op e n in g t h e S h o rt Circ u it a t t h e
Co lle c t o r
Flore nc e Cha rbonnie r,*,†,‡ Ludovic Be rthe lot, a nd Chris tia n Rola ndo*
†
,§
D e´ partement de Chimie, Ecole Normale Sup e´ rieure, URA 1679 du CNRS, Processus d’Activation Mol e´ culaire,
4 rue Lhomond, 75231 Paris Cedex 05, France, and UFR de Chimie, Universit e´ des Sciences et
Technologies de Lille (Lille 1), B aˆ timent C4, URA 351 du CNRS, Chimie Organique et Macromol e´ culaire,
9655 Villeneuve d’Ascq Cedex, France
2
5
The chemical reactivity of an electroactive neutral com-
pound in the conditions of electrospray mass spectrom-
etry was studied by carrying out pr epa r a tive experi-
ments. The residue obtained by spraying a micromolar
methanolic solution of 9 -nitroanthracene onto a metallic
plate in negative mode (cathodic capillary) was collected
and then analyzed by GC/ MS, showing that both reduction
and oxidation products were formed. To distinguish
between the two processes, the two functions of the
metallic plate, definition of the electric potential and
storage of products, were separated by using a copper-
bronze grid as counter electrode and a silica TLC plate,
located a few millimeters behind, as collector. The col-
lected products were thus stored away from the counter
electrode, avoiding their possible reoxidation at this site.
The amount of anthracene, a reduction product of 9 -ni-
troanthracene, which was detected only in trace amounts
with the metallic plate collector, increased significantly
with the new device. Thus, anthracene can be viewed as
the result of a capillary process whereas 9,10-anthraquino-
ne can be estimated as mainly due to further oxidation at
the collector. Furthermore, in deuterated solvents no
incorporation was detected, suggesting that some reaction
steps take place at the collector.
involves reduction at the capillary and oxidation at the counter
electrode (Figure 1). For neutral compounds, which are not
ionized in the starting solution, electron transfers either hetero-
geneous (at the capillary tip) or homogeneous (inside the capillary
tip or within droplets) must be responsible for the ions observed.
On this basis, a description of ESI-MS as a controlled-current
electrolytic cell has been comprehensively developed by Van
Berkel3 and Cole.6
-5
The ions formerly produced are partly consumed in reactions
leading to neutrals and partly lost during the transfer from
atmospheric pressure to collector. Finally, the ratio of the ions
detected by the mass spectrometer per electrons consumed at
7
the capillary is approximately only 1 per 10 000. Thus, ESI-MS
analysis requires that neutral compounds be sufficiently electro-
active to get ionized in a significant amount by electron transfer
at the capillary tip and that the created ions be stable enough in
order to reach the collector before subsequent reaction. Very few
experiments have been devoted to studying the hidden electro-
chemistry occurring within the spray. Van Berkel and co-workers
have been able to detect redox processes by fluorescence
spectroscopy analysis of the electrospray mist.8
With the aim of unraveling hidden parts of the electrochemistry
occurring during ESI-MS analysis, we chose to study the fate of
neutral electroactive compounds during electrospray ionization.
In a typical experiment, a solution of the electroactive neutral
compound is sprayed onto a metallic plate in ESI conditions and
the deposit is extracted with a solvent and analyzed by GC/ MS.
In a preliminary report, we demonstrated that such preparative
electrospray experiments can be performed on halogenonitroaro-
Despite an explosive development of electrospray mass spec-
trometry (ESI-MS) as an analytical tool, the basic principles that
govern the droplet charging process remain mostly unknown.1
As pointed out by Kebarle,2 the existence of a spray current
9
matic compounds. The global behavior of the analyte can then
implies electron transfers involving electroactive species both at
the capillary and at the collector. Negative ionization for instance
(
3) Van Berkel, G. J.; McLuckey, S. A.; Glish, G. L. Anal. Chem. 1 9 9 2 , 64,
586-1593.
1
*
Corresponding author: (phone) 33-4-76-51-41-86 or -79; (e-mail) Florence.
(4) Van Berkel, G. J.; Zhou, F. Anal. Chem. 1 9 9 5 , 67, 2916-2923.
(5) Van Berkel, G. J.; Zhou, F. Anal. Chem. 1 9 9 5 , 67, 3958-3964.
(6) Xu, X.; Nolan, S. P.; Cole R. B. Anal. Chem. 1 9 9 4 , 66, 119-125.
(7) Smith, R. D.; Loo, J. A.; Loo, R. R. O.; Busman, M.; Udseth, H. R. Mass
Spectrom. Rev. 1 9 9 1 , 10, 359-451.
(8) Chillier, X. Fr. D.; Monnier, A.; Bill, H.; Ula c¸ ar, F. O.; Bush, A.; McLuckey,
F. A.; Van Berkel, G. J. Rapid Commun. Mass Spectrom. 1 9 9 6 , 10, 299-
304.
Charbonnier@ujf-grenoble.fr.
†
Ecole Normale Sup e´ rieure.
‡
Current address: Universit e´ Joseph-Fourier (Grenoble I), EA 582, BP 53,
3
8041 Grenoble Cedex 9, France.
§
Universit e´ des Sciences et Technologies de Lille (Lille 1).
(
(
1) Cole, R. B., Ed. Electrospray Ionization Mass Spectrometry, Fundamentals,
Instrumentation & Applications; John Wiley & Sons: New York, 1997.
2) Blades, A. T.; Ikonomou, M. G.; Kebarle, P. Anal. Chem. 1 9 9 1 , 63, 2109-
(9) Charbonnier, F.; Nicolas, J.-P.; Eveleigh, L.; Hapiot, P.; Pinson, J.; Rolando,
C. C. R. Acad. Sci. Paris, Ser. IIc 1 9 9 8 , 449-456.
2
114.
1
0.1021/ac980799l CCC: $18.00 © 1999 American Chemical Society
Analytical Chemistry, Vol. 71, No. 8, April 15, 1999 1585
Published on Web 03/16/1999

The capillary was a stainless steel 22s gauge (o.d., 0.72 mm;
i.d., 0.15 mm) needle (Hamilton, Reno, NV) adapted by a Teflon
tube to a 5-mL glass syringe driven by a syringe pump (model
A-99 Razel Scientific Instrument, Stanford, CT). The collector was
a 5 × 5 cm copper plate (for printed board). The distance between
capillary tip and collector was fixed to 6 mm by using an optical
bench (Microcontrole, Evry, France). The whole device was
placed into a glass enclosure where the temperature (∼35 °C)
was controlled by means of a warming belt and measured with a
thermocouple. The capillary was always connected to the grounded
side of the high-voltage supply (positive high voltage (HV) in
reduction mode, i.e., cathodic capillary). The solution was sprayed
under ambient atmosphere at a flow rate of 1.25 mL‚h-1. The time
required for infusing 5 nmol of analyte (5 mL of a 1 µM solution)
was ∼4 h. After spraying, the plate was washed with dichlo-
romethane (50 µL) and 9-cyanoanthracene (2 µL of a 1 mM
methanolic solution) was added as an internal standard to the
collected solution before GC/ MS analysis.
Fig u re 1 . Schematic description of the ESI-MS process: negative
ion mode.
be inferred on the basis of the components identified in the
2
collected residue. We report here on 9-nitroanthracene (HArNO )
submitted to negative ionization in methanol. First, we used a
metallic plate as the collecting surface. In such conditions, an
oxidized derivative of 9-nitroanthracene, 9,10-anthraquinone, was
isolated as the major product. This result led us to consider the
anodic activity of the collector. Since the collecting plate also
worked as the counter electrode in these experiments, it was likely
to involve the compounds formerly produced within the spray in
further oxidation reactions. To study selectively the sequence of
electrolytic processes initiated at the capillary, we designed a two-
piece collector set, made of a metallic grid connected to the high-
voltage supply and a silica trapping surface independent from the
electrospray circuit. With this device, we were able to isolate
anthracene, a typical reduction product of the analyte. We propose
a mechanism for its formation within the electrospray mist.
In the new device, the metallic plate was replaced by a 5 × 5
cm copper-bronze grid (2-mm mesh) positioned at the same
place. A 5 × 5 cm TLC silica plate (Silicagel 60, Merck, Darmstadt,
Germany) separated by a Teflon spacer was placed 1.6 mm away.
After spraying, the silica plate was observed under UV light and
the absorbing zone was collected, extracted with dichloromethane
(
1 mL), and filtered on glass wool. The resulting solution was
concentrated to 50 µL under argon and GC/ MS analysis was
performed as before.
GC/ MS Analysis. GC/ MS analyses were performed on a JMS-
700 MStation mass spectrometer (JEOL, Tokyo, Japan), equipped
with a HP 6890 (Hewlett-Packard, Palo Alto, CA) gas chromato-
graph. The chromatograph was fitted with a CPSIL-5 CB (Chrom-
Pack, Middelburg, The Netherlands) low-bleed column (30 m ×
EXPERIMENTAL SECTION
Reagents and Materials. Reagent grade dichloromethane,
methanol, 9-cyanoanthracene, 9-nitroanthracene, 9-H-anthrone,
0
.25 mm, 0.25-µm film thickness). Injection was performed in
splitless mode. The temperature varied from 80 (15 min) to 280
9
,10-anthraquinone, trifluoroacetic acid, and 25% aqueous tri-
1
°
C, at a 5 °C‚min- rate. Ionization mode was electron impact (70
methylamine (TMA) were purchased from the Aldrich Chemical
Co. (Milwaukee, WI) and used without further purification. Due
to the low concentrations used (in the micromolar range), great
care was taken to avoid contamination (by phthalates, fingerprints,
etc.) throughout the process.
eV) and the mass/ charge range from m/ z 50 to 550 was scanned
in 2 s.
ESI/ MS Analysis. ESI/ MS was performed on a API 100
instrument (Perkin-Elmer, Norwalk, CT), with the analyte solution
infused by means of a Hamilton syringe pump (Hamilton).
Electrospray Device. The high voltage was provided by a
(
10-kV, 1-mA power supply (SDS, Paris, France). It was connected
to the collecting plate through the secondary of a transformer
obtained from the high-voltage unit of a Dupont mass spectrom-
RESULTS AND DISCUSSION
P reparative Electrospray with a Copper P late Collector.
(
2
We selected HArNO as a probe for electron transfers involved
eter). The primary of the transformer was connected to a rf
generator with a maximum amplitude of 10 V peak to peak. With
this device, an alternative voltage (voltage, 0-300 V peak to peak;
frequency, 100-3000 Hz) was added for securing the synchroniza-
tion of the spray. The instantaneous electrospray current was
measured at the capillary with an AD843JN current transducer
at different stages of the ESI process in negative ionization mode
as its electrochemical behavior as well as its photochemical
reactivity has been described comprehensively in reducing
conditions.1
2-16
On the other hand, we had observed during
previous preparative electrospray experiments with 9-bromo-10-
(
Analog Device, Norwood, MA), with a rising time less than 1 µs
(11) Beaugrand, C.; Charbonnier, F.; Rolando, C.; Sablier, M.; Saru, F. Proceed-
ings of the 43th ASMS Conference on Mass Spectroscopy and Allied Topics,
Atlanta, May 21-26, 1995; p 668. Details concerning the synchronization of
the spray will be described in a forthcoming publication.
6
and a 10 amplification ratio. The image of the synchronization
voltage was detected by a capacitor plate located behind the
collector and subtracted from the electrospray current. Data were
recorded using a TDS 320 digital oscilloscope (Tektronix, Bea-
verton, OR) with 1000 data points and an 8-bit dynamic, in the
peak detect mode.10,11
(12) M’Halla, F.; Pinson, J.; Sav e´ ant, J. M. J. Am. Chem. Soc. 1 9 8 0 , 102, 2, 4120-
4127.
(
13) Chapman, O. L.; Heckert, D. C.; Reasoner, J. W.; Thackaberry, S. P. J. Am.
Chem. Soc. 1 9 6 6 , 88, 5550-5554.
(14) Hamanoue, K.; Nakayama, T.; Kajiwara, K.; Yamanaka, S.; Ushida, K. J.
Chem. Soc., Faraday Trans. 1 9 9 2 , 88, 3145-3151.
(
10) Charbonnier, F.; Rolando, C.; Saru, F.; Hapiot, P.; Pinson, J. Rapid Commun.
Mass Spectrom. 1 9 9 3 , 7, 707-710.
(15) Cheng, E.; Sun, T. C.; Su, Y. O. J. Chin. Chem. Soc. 1 9 9 3 , 40, 551-555.
(16) Hammerich, O.; Parker, V. D. Acta Chem. Scand. 1 9 8 1 , B35, 341-347.
1586 Analytical Chemistry, Vol. 71, No. 8, April 15, 1999

Fig u re 2 . Schematic overview of the preparative ESI device.
nitroanthracene (BNA) a progressive increase of the conversion
yield with increasing dilution in the bulk, reaching a 50% conver-
sion yield for micromolar solutions and a near complete trans-
formation for submicromolar solutions.9 Thus, we used here
micromolar solutions for all experiments, expecting in this
concentration range conversion yields will be both fair and
sensitive to experimental conditions.
During a typical experiment, a micromolar methanolic solution
of HArNO
2
containing trifluoroacetic acid (TFA) or 25% aqueous
OH),
trimethylamine (TMA) as the pH modifier (10-3 (v/ v) in CH
3
was sprayed onto a copper plate at 35 °C under atmospheric
pressure at a flow rate of 1.25 mL‚h-1. Figure 2 shows the three
parts of the device, designed for spraying the analyte solution,
measuring the instantaneous current, and stabilizing the spray.
In its free mode, the spray instantaneous current observed on
Fig u re 3 . Time dependence of the instantaneous electrospray
-
1
current for a solution of 9-nitroanthracene (1 µmol‚L ; TFA 2.5 ×
-
4
10
(v/v) in CH3OH; 3 kV). The synchronization voltage is drawn in
thin lines.
the oscilloscope appears as a pseudoperiodic phenomenon.10
A
typical steady-state pattern shows sharp peaks, of which the
periodicity can be measured, allowing one to define a given
spraying mode by a mean pulsation frequency of the instantaneous
spray current. Since we scanned a very narrow range of concen-
trations, as close as possible to routine ESI-MS conditions, to
correlate analytical results with functioning modes, we had to
ensure the stability of spraying modes during several hours. By
superimposing to the main HV an ac voltage of which the
frequency is chosen close to that of the free stable spraying mode,
the spray is assisted and forced to trigger on the ac voltage
maximums.11 The resulting current proves perfectly periodic. A
typical pattern characterizing the time dependence of the elec-
trospray current in such conditions is shown in Figure 3. In
comparison to neutral solutions, the spray was rather difficult to
stabilize for both acidic and basic conditions. For acidic solutions,
spraying conditions were optimized with a high voltage of ∼5 kV
the plateau is no longer detected, as peak succeeds peak at high
frequency, and this simpler pattern is also observed in the absence
of spray (without feeding the capillary). Since, at the highest HV
values, corona discharges can be considered as mainly responsible
for the electrospray current, we assigned the peak preceding the
plateau to a corona discharge, also at lower HV values. The
discharges probably involve the solvent molecules (gaseous
methanol) at the lowest HV values and, at the highest HV values
(“all-corona” mode), the atmospheric components of the inter-
electrode space (oxygen and nitrogen). On the contrary, the half-
period at steady intensity is likely to originate in electron transfers
occurring between electroactive components of the solution and
the metallic surface at the capillary tip. These processes allow
recovery of electroneutrality in the liquid phase. The height of a
plateau is rather independent of the nature of analyte and additives
and we observed almost no change in current, either peak
intensities or mean current, while changing the methanolic
solution of analyte for “pure” solvent. This general result leads to
the conclusion that most redox processes mainly involve the
solvent or impurities. On the basis of the easy reduction of
methanol into methylate anion and hydrogen, the main part of
(
interelectrode distance, 6 mm) at a pulsation frequency of ∼1
kHz.
A pulsation period proceeds in two stages, a sharp peak,
followed by a plateau at a much lower intensity (Figure 4), and
for the usual high-voltage range (3-5 kV), no current is observed
without feeding the capillary. For higher values of the high voltage,
Analytical Chemistry, Vol. 71, No. 8, April 15, 1999 1587

Fig u re 5 . GC/MS analysis of the residue from an acidic methanolic
-
1
-3
Fig u re 4 . Time dependence of the synchronized spray current:
enlargement showing the detailed sequence of events during a
pulsation period, for the spraying conditions of Figure 3.
solution of 9-nitroanthracene (1 µmol‚L ; TFA 10 (v/v) in CH3OH)
collected onto a copper plate.
possible anodic origin of this compound (thus produced at the
counter electrode as a result of oxidation processes) as an
alternative pathway.
the current can be assigned to the solvent electroactivity. More-
over, the resulting methylate anion, as an oxygenated nucleophile,
is possibly responsible for initiating the transformation of the anion
radical derived from the analyte into varied oxygenated products.
Throughout the experiment, the regularity of the spraying
mode is controlled by observing and measuring the instantaneous
current frequency on the oscilloscope. Immediately after the end
of spraying, the residue deposited on the collector is extracted
with dichloromethane, filtered, and concentrated under argon.
Then the internal standard (9-cyanoanthracene, HArCN) is added
before GC/ MS analysis is performed. With this collecting device,
the major components containing the three condensed rings of
anthracene were residual 9-nitroanthracene (molecular peak, m/ z
Opening the Short Circuit at the Collector. Assuming that
some steps could occur at the collector, we designed a new
experimental device allowing prevention of further oxidations at
this site. The metallic plate was replaced by a two-piece set: a
metallic grid with large mesh acting as the counter electrode and
a silica plate acting as the collector (Figure 6). The components
of the electrospray mist were supposed to be sorted out, ionic
components being mainly stopped at the metallic grid and neutral
products being mainly carried out through the grid onto the silica
surface. On the other hand, silica gel trapping was expected to
improve the recovery yield by reducing sublimation of the stored
products. Such processes could previously be favored during the
2
23; retention time, 48.2 min) and 9,10-anthraquinone (molecular
peak, m/ z 208; retention time, 45.1 min) in approximately equal
amounts. 9-H-Anthrone (molecular peak, m/ z 194; retention time,
4
-h stay of the residue on the metallic collector in the warmed
enclosure.
In the context of a study of the pH dependence of charge-
4
9.4 min) and anthracene (molecular peak, m/ z 178; retention
time, 41.1 min) were also detected, both in trace amounts. The
products were identified by their mass spectra and by comparing
the retention times with those of authentic samples. A typical
chromatogram consisting of reconstructed ionic currents at m/ z
state distributions of amino acids, strongly acidic conditions
have been shown to produce intense [M - H]- ions in negative
17
ESI-MS, in the so-called “wrong-way-round” processes. Similarly,
on the basis of our previous results (Table 1), we chose TFA as
1
78 (anthracene), 180 (9,10-anthraquinone; fragment ion, M -
-
4
the pH modifier but a lower concentration (2.5 × 10 v/ v), to
favor reduction processes and at the same time to avoid excess
corona discharges. In such conditions, the spray was stabilized
at a pulsation frequency of 200 Hz for a high voltage of 3 kV.
Coupled Copper-Bronze Grid Counter Electrode and
Silica P late Collector. No difference was observed between the
two collecting systems concerning the nature of the collected
products. Though, with the new device, the overall recovery yield
almost doubled (∼6%) and the proportion of reduction product,
anthracene, increased by a factor 3 (from ∼10% to ∼30%). The
global conversion yield remained around 30-35%. These effects
are illustrated in Figure 7, where anthracene is made visible with
a magnifying ratio of only 10 whereas a ratio of 100 was necessary
in the case of the copper plate collector (Figure 5). A significant
percentage of analyte is still transformed into 9,10-anthraquinone,
supporting the hypothesis of a dual origin for oxygenated
products. Such products would be issued both from electrochemi-
•
+
CO ), 203 (HArCN), and 223 (HArNO
for acidic conditions.
2
) is shown in Figure 5,
Comparative results of GC/ MS analyses in acidic and basic
conditions are reported in Table 1. The overall recovery yield,
calculated by comparison with the internal standard (HArCN:
molecular peak, m/ z 203; retention time, 48.2 min) was ∼3%.
Based on the recovered amount of the four components, the global
conversion yield of 9-nitroanthracene in acidic (respectively basic)
conditions was 35% (respectively 70%), with a selectivity in favor
of oxygenated derivatives (9,10-anthraquinone and 9-H-anthrone)
of 91% (respectively 94%).
The increased conversion yield (oxidation) in basic conditions
can be accounted for by proposing for the production of 9,10-
anthraquinone and 9-H-anthrone SRN-type reactions of the anion
radical with oxygenated nucleophilic species, the concentration
of which is largely increased in basic conditions. Furthermore, it
must be noticed that such processes are catalytic in electrons.
On the other hand, the very high selectivity in favor of 9,10-
anthraquinone even in acidic conditions led us to consider the
(
17) Mansoori, B. A.; Volmer, D. A.; Boyd, R. K. Rapid Commun. Mass Spectrom.
1 9 9 7 , 11, 1120-1130.
1588 Analytical Chemistry, Vol. 71, No. 8, April 15, 1999

Ta b le 1 . P re p a ra t ive ES I o f 9 -Nit ro a n t h ra c e n e in Me t h a n o l
a
TFA 10-3 (v/ v). b TMA 10-3 (v/ v). c Global recovery yield calculated by using 9-cyanoanthracene as the internal standard. d Global conver-
e
sion yield of 9-nitroanthracene into 9,10-anthraquinone, 9-H-anthrone, and anthracene. Product ratio calculated from reconstructed ionic
currents.
on the other side the specific behavior of anthracenic compounds,
where the internal cycle has a rather weak aromatic character,
favoring oxidation processes at this site.18 Thus, for this substrate,
in reducing conditions, in addition to reactions of the anion radical
with hydrogen and proton donors or nucleophilic species, leading
in particular to SRN reaction products, monomeric or dimeric oxime
derivatives have also been identified, giving rise to oxygenated
products.15,16
Under classical electrochemical conditions, 9-nitroanthracene
reacts at a cathode by a reversible one-electron transfer giving
•-
9-nitroanthracen-10-yl-9-ide anion radical HArNO
2
at a half-wave
potential of ∼-1 V (vs SCE), followed by a second irreversible
one-electron transfer at a half-wave potential of ∼-1.5 V (vs SCE)
leading to the dianion.1 In negative ESI ionization only neutral
9,20
Fig u re 6 . Schematic description of the open short circuit device.
compounds with half-wave potentials more positive than -0.8 V
1,22
(
vs SCE) are observed.2
Thus, only the first one-electron
transfer at lower potential can be expected to occur in the ESI of
-nitroanthracene. We analyzed the behavior of 9-nitroanthracene
9
in negative ESI-MS in order to compare detected ions with neutrals
from preparative ESI.
The detection of the anion radical in ESI-MS would require
its transfer with limited decomposition from capillary tip to
droplets and within the electrospray mist to the collector. Neither
•
-
the anion radical HArNO
2
nor any other anionic organic
fragment could be detected. Changing methanol for a less reactive
solvent such as acetonitrile did not allow us to observe a signal.
In the same conditions, the ESI mass spectrum of 9-bromo-10-
nitroanthracene showed no peak corresponding to an anthracenyl
fragment but exhibited an intense bromide ion peak, in agreement
with the known reactivity of 9-bromo-10-nitroanthracene anion
radical.9
Fig u re 7 . GC/MS analysis of the residue from a methanolic solution
-
4
of 9-nitroanthracene (CF3COOH 2.5 × 10 M (v/v) in CH3OD; 1
-1
µmol‚L ), collected onto a silica plate through a copper-bronze grid.
With the separated counter electrode and collector set, the
yield in isolated anthracene was largely improved, confirming that
9,10-anthraquinone results at least in part from oxidation taking
cal processes at the counter electrode and from chemical reactions
with reactive species formed and transported within the mist.
The results of GC/ MS analyses, including experiments in
deuterated solvents, are reported in Table 2. No incorporation
(
18) Sainsbury, M. In Rodd’s Chemistry of Carbon Compounds; A ModernCom-
prehensive Treatise, 2nd ed.; Coffey, S., Ed.; Elsevier: Amsterdam, 1979;
Vol. III, Part H, Chapter 28, pp 5-93.
could be detected in CH
deuterated anthracene (molecular peak, m/ z 179; retention time,
41.1 min) could not be observed.
3 3
OD or in CD OD. In particular, mono-
(
(
19) Kitagawa, T.; Ichimura, A Bull. Chem. Soc. Jpn. 1 9 7 3 , 46, 3792-3795
20) Verniette, M.; Pouillen, P.; Martinet, P. Bull. Soc. Chim. Fr. Ser. I 1 9 8 4 ,
141-144.
∼
(
(
21) Dupont, A.; Gisselbrecht, J.-P.; Leize E.; Wagner, L.; Van Dorsselaer, A.
Tetrahedron Lett. 1 9 9 4 , 35, 6083-6086.
DISCUSSION
22) Van Berkel, G. J. In Electrospray Ionization Mass Spectrometry, Fundamentals,
Instrumentation & Applications: The Electrolytic Nature of Electrospray; Cole,
R. B., Ed.; John Wiley & Sons: New York, 1997; pp 65-105.
As a nitroanthracenic derivative, 9-nitroanthracene exhibits on
one side the typical reactivity of nitroaromatic compounds and
Analytical Chemistry, Vol. 71, No. 8, April 15, 1999 1589

-
4
Ta b le 2 . P re p a ra t ive ES I o f 9 -Nit ro a n t h ra c e n e in Me t h a n o l w it h TFA (2 .5 × 1 0 v/v) a s t h e p H Mo d ifie r (P u ls a t io n
Fre q u e n c y, 2 0 0 Hz)
a
Global recovery yield calculated by using 9-cyanoanthracene as the internal standard. b Global conversion yield of 9-nitroanthracene into 9,10-
c
anthraquinone, 9-H-anthrone, and anthracene. Product ratio calculated from reconstructed ionic currents.
S c h e m e 1 . P la u s ib le Me c h a n is m fo r Re d u c t io n
o f 9 -Nit ro a n t h ra c e n e in t o An t h ra c e n e in
Ne g a t ive ES I
Fig u re 8 . Rough estimate of ESI parameters.
place at the counter electrode and that this competitive process
is efficiently avoided by securing the redox inertness of the
collecting surface. On the other hand, the absence of incorporation
when deuterated solvents were used (neither in anthracene nor
in remaining 9-nitroanthracene) corroborates that 9-nitroan-
thracene dianion is not formed and that reduction stops after the
first one-electron transfer at lower potential, the anion radical being
less prone to protonation.
thracene by anodic oxidation in the case of solution chemistry.
The formation of anthracene implies breakdown of the anion
radical HArNO2•- into nitrite anion and anthracen-9-yl radical,
which must occur mainly within droplets and not inside the
capillary, where the anthracenyl radical would be further reduced
into anthracenyl anion and then protonated or deuterated. On the
other hand, the formation of antracenyl radical in the gas phase
may be driven by the better solvation of the nitrite anion in the
Assuming that the electroactive volume inside the capillary is
roughly equal to the disturbing volume formed at the capillary
tip at each pulsation, we used the spray cone length as an estimate
of the electrolysis electrode length. The calculated residence time
of the solution during an electroactive period is then ∼50 ms.
Assuming a typical value for the speed of droplets, the time of
flight of droplets from capillary tip to collector can be estimated
at 0.6 ms (Figure 8).23 On the other hand, the average current
density can be calculated by assuming that the spray current
remains roughly constant during an electroactive period and equal
to the steady-state value of the plateau (Figure 4). The estimated
value is roughly 0.43 mA‚cm-2, which is in the common range
for electrochemical synthesis processes.
methanolic droplets compared to that of the hydrophobic anion
radical.2
4-27
By comparison with other aromatic radicals, the
anthracenyl radical is known to have a moderate reactivity in
hydrogen abstraction processes.28 Thus, it is supposed to be
transported as a gaseous component of the electrospray mist to
the silica surface where it abstracts an hydrogen. Finally the
results are globally consistent with the following sequence: (i)
formation of the anion radical HArNO2•- at the capillary; (ii)
breakdown of this intermediate within droplets into nitrite ion in
•
condensed phase and anthracen-9-yl radical HAr , which desorbs
in the gas phase; (iii) hydrogen abstraction by the anthracenyl
radical from silica gel at the collector. (Scheme 1).
•-
In condensed phase, the lifetime of the anion radical HArNO
2
(
24) Kebarle, P.; Ho, Y. In Electrospray Ionization Mass Spectrometry, Funda-
mentals, Instrumentation & Applications: On the Mechanism of Electrospray
Mass Spectrometry; Cole, R. B., Ed.; John Wiley & Sons: New York, 1997;
pp 3-63.
is greater by at least 1 order of magnitude than the residence
time in the electroactive capillary tip.19,20 As the time of flight of
droplets to the collector is even shorter, the reversible character
of the formation of the anion radical will lead back to 9-nitroan-
(
(
(
25) Kebarle, P.; Tang, L. Anal. Chem. 1 9 9 3 , 65, 972A-986A.
26) Iribarne, J. V.; Thomson, B. A. J. Chem. Phys. 1 9 7 6 , 64, 2287-2294.
27) Thomson, B. A.; Iribarne, J. V. J. Chem. Phys. 1 9 7 9 , 71, 4451-4463.
(
23) Kozhenkov, V. I.; Fuks, N. A. Russ. Chem. Rev. 1 9 7 6 , 45, 1179-1184.
Translated from Usp. Khim. 1 9 7 6 , 45, 2274-2284).
(28) Chen, R. H.; Kafafai, S. A.; Stein, S. E. J. Am. Chem. Soc. 1 9 8 9 , 111, 1418-
(
1423.
1590 Analytical Chemistry, Vol. 71, No. 8, April 15, 1999

Instead of the expected nitro or anthracenyl reduction products
of 9-nitroanthracene (hydroxyamino and amino or dihydro deriva-
tives), we observed mainly the elimination of the nitro group. Such
nonclassical chemical transformations within the methanolic elec-
trospray mist can be related to novel compound syntheses de-
signed by extension of gas-phase electrical discharge reactions
to the liquid phase, which has been described elsewhere. For in-
stance, the formation of perchlorinated fullerene fragments at cop-
per electrodes immersed into liquid chloroform or carbon tetra-
chloride (10 kV-20 kHz ac voltage, 1 mm interelectrode distance,
routine concentrations of ESI-MS.11 We describe here a new device
where the spraying residue is trapped on a silica plate placed
behind a metallic grid acting as the counter electrode instead of
being collected directly onto a metallic plate. Thus, electron trans-
fers between the collected products and the counter electrode
are avoided. With 9-nitroanthracene as the analyte, whereas using
the simple metallic plate collector most of the detected products
were oxidation derivatives, this new device led to a dramatic
increase of the proportion of isolated reduction product. Thus,
the chemical inertness of the new trapping surface and the
possibility of getting by this means a more straightforward access
to the specific redox processes initiated at the capillary during
negative electrospray ionization were confirmed. The existence
of a redox chemistry even when no ion from the analyte is
detected in ESI-MS was also demonstrated.
2
0-50 mA arc current) was found surprisingly highly selective.29
CONCLUSION
We had previously shown that synchronization allows large
durations and secures the reproducibility of results of preparative
electrospray experiments, with yields becoming quantitative in
the micromolar range, which is 10-100 times lower than the
Received for review July 21, 1998. Accepted December
1
6, 1998.
(
29) Huang, R.; Huang, W.; Wang, Y.; Tang, Z.; Zheng, L. J. Am. Chem. Soc.
9 9 7 , 119, 5954-5955.
1
AC980799L
Analytical Chemistry, Vol. 71, No. 8, April 15, 1999 1591