Electrochemical reduction of phthalide at carbon cathodes in dimethylformamide: Effects of supporting electrolyte and gas chromatographic injector-port chemistry on the product distribution

Article abstract of DOI:10.1016/j.electacta.2013.09.124

Cyclic voltammetry and controlled-potential (bulk) electrolysis have been used to investigate the direct reduction of phthalide at carbon electrodes in dimethylformamide (DMF) containing 0.10 M tetramethylammonium perchlorate (TMAP) or tetra-n-butylammonium perchlorate (TBAP). Cyclic voltammograms recorded with a glassy carbon electrode exhibit a single cathodic peak and a corresponding anodic peak that arise, respectively, from one-electron reduction of phthalide to generate a radical-anion intermediate and from reoxidation of the intermediate. At a scan rate of 100 mV s^-1, quasi-reversible behavior is observed (due to ring-opening of the radical-anion), whereas fully reversible behavior is seen at 5 V s^-1 or higher. Digital simulation of cyclic voltammograms indicates that the lifetime of the radical-anion is 3.5 s. Bulk electrolysis of phthalide at a reticulated vitreous carbon cathode affords products that depend on the procedure used to analyze the catholyte. Direct injection of catholyte into a gas chromatograph shows phthalide and a 2-methylbenzoate ester bearing an alkyl moiety from the supporting-electrolyte cation. However, if the catholyte is partitioned between diethyl ether and aqueous hydrochloric acid before gas chromatographic analysis, phthalide and 2-methylbenzoic acid are observed. Thermally induced reactions that occur in the injector port of the gas chromatograph are responsible for the formation of the 2-methylbenzoate ester as well as for the phthalide found in all electrolyzed solutions.

Full text of DOI:10.1016/j.electacta.2013.09.124

Electrochimica Acta 113 (2013) 557–563
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Electrochemical reduction of phthalide at carbon cathodes in
dimethylformamide: Effects of supporting electrolyte and gas
chromatographic injector-port chemistry on the product distribution
Erick M. Pasciaka, Spencer E. Hochstetlerb, Mohammad S. Mubarakc, Dennis H. Evansd,
Dennis G. Petersa,∗
a Department of Chemistry, Indiana University, Bloomington, IN 47405, USA
b Eastman Chemical Company, Kingsport, TN 37664, USA
c Department of Chemistry, The University of Jordan, Amman 11942, Jordan
d Department of Chemistry, Purdue University, West Lafayette, IN 47907, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Cyclic voltammetry and controlled-potential (bulk) electrolysis have been used to investigate the direct
reduction of phthalide at carbon electrodes in dimethylformamide (DMF) containing 0.10 M tetramethyl-
ammonium perchlorate (TMAP) or tetra-n-butylammonium perchlorate (TBAP). Cyclic voltammograms
recorded with a glassy carbon electrode exhibit a single cathodic peak and a corresponding anodic peak
that arise, respectively, from one-electron reduction of phthalide to generate a radical-anion intermedi-
ate and from reoxidation of the intermediate. At a scan rate of 100 mV s−1, quasi-reversible behavior is
observed (due to ring-opening of the radical-anion), whereas fully reversible behavior is seen at 5 V s−1 or
higher. Digital simulation of cyclic voltammograms indicates that the lifetime of the radical-anion is 3.5 s.
Bulk electrolysis of phthalide at a reticulated vitreous carbon cathode affords products that depend on
the procedure used to analyze the catholyte. Direct injection of catholyte into a gas chromatograph shows
phthalide and a 2-methylbenzoate ester bearing an alkyl moiety from the supporting-electrolyte cation.
However, if the catholyte is partitioned between diethyl ether and aqueous hydrochloric acid before gas
chromatographic analysis, phthalide and 2-methylbenzoic acid are observed. Thermally induced reac-
tions that occur in the injector port of the gas chromatograph are responsible for the formation of the
2-methylbenzoate ester as well as for the phthalide found in all electrolyzed solutions.
Received 6 September 2013
Received in revised form
20 September 2013
Accepted 21 September 2013
Available online 12 October 2013
Keywords:
Phthalide
Digital simulation
Ring-opening reaction
Supporting-electrolyte effects
GC injector-port chemistry
© 2013 Elsevier Ltd. All rights reserved.
1. Introduction
by Benˇo et al. [6], who examined the correlation between half-wave
potentials and ␴ substituent constants.
In the past, most research devoted to the electrochemical reduc-
tion of esters has focused on the behavior of simple, linear species
[1–4]. On the other hand, few reports [5–7] of the electrochemical
reduction of phthalides (cyclic esters or lactones) have appeared in
the literature. As part of an early study of the electrochemistry of
phthalimides, Leedy and Muck [5] observed that reduction of either
phthalimide or N-phenylphthalimide at a mercury pool cathode in
DMF containing tetraethylammoniuim perchlorate leads to the for-
mation of phthalide in substantial yield and that phthalide itself
exhibits a single, quasireversible cyclic voltammetric response at
a mercury-plated platinum electrode. Polarographic reduction of
several substituted 3-arylmethylenephthalides has been described
In our laboratory, the electrochemical reduction of phthalide
(1) at mercury cathodes in DMF containing 0.10 M tetra-n-
butylammonium perchlorate (TBAP) was investigated more than
two decades ago by means of cyclic voltammetry and controlled-
potential electrolysis [7].
O
O
1
At a hanging mercury drop electrode, reduction of phthalide
manifests itself by a single cathodic peak that is quasireversible
at low scan rates (50 mV s−1) and essentially fully reversible at
∗
Corresponding author. Tel.: +1 812 855 9671; fax: +1 812 855 8300.
E-mail address: peters@indiana.edu (D.G. Peters).
0013-4686/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.electacta.2013.09.124

558
E.M. Pasciak et al. / Electrochimica Acta 113 (2013) 557–563
5 V s−1 or higher. This scan-rate dependence provides insight into
the stability of the proposed radical-anion intermediate formed
during the one-electron cathodic process; double-potential-step
chronoamperometry revealed that this radical-anion undergoes
ring-opening and has a half-life of approximately 0.7 s. When
controlled-potential (bulk) electrolyses of phthalide were carried
out at a mercury pool cathode held at −1.60 V (very close to the
cathodic peak potential), the coulometric n value was found to
be 1. After the catholyte was partitioned between 10% aqueous
hydrochloric acid and diethyl ether, a subsequent gas chromato-
graphic analysis revealed the presence of 2-methylbenzoic acid
(50%) and phthalide (44%). In addition, bulk electrolysis of phthalide
at a mercury pool cathode in DMF–0.10 M TBAP containing a tenfold
excess of phenol as a proton donor proved to be a four-electron pro-
cess that affords a mixture of products [2-methylbenzoic acid (28%),
cis-hexahydrophthalide (30%), trans-hexahydrophthalide (6%), and
cis-1,2,3,6-tetrahydrophthalide (32%)] [7].
In later work [8], it was discovered that, if one performs a
bulk electrolysis of phthalide at a mercury pool in DMF–0.10 M
TBAP, under conditions similar to those just described, and then
injects an aliquot of the catholyte directly into a gas chromato-
graph, the products are found to be n-butyl 2-methylbenzoate and
phthalide. Furthermore, if a DMF–0.10 M tetra-n-hexylammonium
perchlorate medium is used for the bulk electrolysis of phthalide,
the products seen after direct injection of catholyte into a gas
chromatograph consist essentially of a 1:1 mixture of n-hexyl 2-
methylbenzoate and phthalide.
In light of our previous research [7,8] concerning the electro-
chemical behavior of phthalide, particularly the peculiar pattern
of products seen as experimental conditions have been varied, we
have employed cyclic voltammetry and controlled-potential (bulk)
electrolysis in the present investigation, first, to probe more deeply
the electrochemical reduction of phthalide at a carbon cathode in
DMF containing either 0.10 M tetramethylammonium perchlorate
(TMAP) or tetra-n-butylammonium perchlorate (TBAP) as suppor-
ting electrolyte and, second, to compare our new results with
those obtained earlier for the reduction of phthalide at a mercury
cathode. In this report, some important questions are addressed.
First, how can one explain the presence of phthalide at the end
of an exhaustive electrolysis? Second, what process is responsi-
ble for the various 2-methylbenzoate esters found at the end of
an electrolysis? Third, what intermediate controls the exclusive
appearance of either 2-methylbenzoic acid or 2-methylbenzoate
ester as an electrolysis product? Fourth, what additional infor-
mation can be obtained from an electrolysis of phthalide in the
presence of deuterium oxide? Answers to these questions have
been deduced with the aid of additional experiments entailing
the use of (i) sodium 2-(hydroxymethyl)benzoate to account for
the presence of phthalide at the end of a bulk electrolysis, (ii) 1-
iodooctane as a trapping agent for 2-methylbenzoate, (iii) pure
sodium 2-methylbenzoate to elucidate the ester-forming process,
and (iv) D2O to track protonation–deprotonation reactions involv-
ing transient anions.
Chemicals), and hydrochloric acid (38%, Mallinckrodt Chemicals).
Deuterium oxide (99.9%) was purchased from Cambridge Isotope
Laboratories and was used for deuterium-labeling studies. Zero-
grade argon (Airgas) was used to deaerate all solutions prior
to electrochemical experiments. Tetramethylammonium perchlo-
rate (TMAP, >99%, GFS Chemicals) and tetra-n-butylammonium
perchlorate (TBAP, >99%, GFS Chemicals) were employed as sup-
porting electrolytes; each compound was recrystallized from ethyl
acetate–hexanes and stored in a vacuum oven at 70–80 ◦C prior to
use.
2.2. Cells, electrodes, procedures, and instrumentation
A description of the one-compartment cell used for cyclic
voltammetry has been previously reported [9]. We constructed
a circular, planar carbon cathode (geometric area of 0.071 cm2)
by press-fitting a short piece of glassy carbon rod (Grade GC-
20, 3.0-mm-diameter, Tokai Electrode Manufacturing Company,
Tokyo, Japan) into a machined Teflon tube; electrical connec-
tion was made by a thin stainless-steel pole that contacted the
glassy carbon and extended upward through the axis of the tube.
A coil of platinum wire was used as the auxiliary electrode for
cyclic voltammetry. Prior to each scan, the cathode was cleaned
with an aqueous suspension of 0.05-␮m alumina on a Master-Tex
(Buehler) polishing pad followed by a rinse with distilled water
in an ultrasonic bath. All potentials are reported with respect
to a reference electrode that consisted of a cadmium-saturated
mercury amalgam in contact with DMF saturated with both cad-
mium chloride and sodium chloride [10–12]; this electrode has a
potential of −0.76 V versus an aqueous saturated calomel electrode
(SCE) at 25 ◦C. Cyclic voltammetry experiments were performed
with a Princeton Applied Research Corporation (PARC) model 273A
potentiostat–galvanostat operated by PowerSuite® software.
Instrumentation used for controlled-potential (bulk) electrol-
ysis is described elsewhere [13,14]. Working electrodes in the
form of circular disks were cut from reticulated vitreous carbon
rods (RVC 2X1-100S, ERG Aerospace Corporation, Oakland, CA)
to have approximately a 2.4-cm diameter, a 0.4-cm thickness,
and a 200-cm2 geometric area. Preparing, cleaning, and storing of
RVC electrodes are described in a previous publication [15]. For
bulk electrolyses, the aforementioned cadmium-saturated mercury
amalgam reference electrode was utilized, and the auxiliary elec-
trode was a carbon rod immersed in a DMF-electrolyte solution
separated from the cathode compartment by a sintered-glass disk
backed by a methyl cellulose–DMF–0.10 M tetraalkylammonium
perchlorate plug. To perform the bulk electrolyses, we employed a
PARC model 173 potentiostat that was controlled with the aid of
LabVIEW software.
2.3. Digital simulations
Simulations of cyclic voltammograms were carried out with the
aid of DigiElch 4.0, a software package for Digital simulation of
common Electrochemical experiments (http://www.elchsoft.com)
[16]. These simulations employed planar diffusion and included
the effects of uncompensated resistance; however, the resistance
was not compensated electronically by the potentiostat. Voltam-
mograms recorded at seven scan rates (0.05, 0.1, 0.2, 0.5, 1, 2,
and 5 V s−1) were simulated according to an ECE-type scheme
(DISP1 mechanism [17]), which comprises an electrode reaction
(one-electron reduction of phthalide to its radical-anion, E-step),
ring-opening of the radical-anion to give the 2-carboxylatobenzyl
radical (C-step), and one-electron reduction of the ring-opened
radical-anion, either at the electrode (E-step) or reduction in solu-
tion by reaction with the phthalide radical-anion (DISP) with the
final reaction being sufficiently rapid to cause the C-step to be
2. Experimental
2.1. Reagents
Each of the following compounds was used, as received, with-
out further purification: 2-methylbenzoic acid (99%, Aldrich),
methyl 2-methylbenzoate (99%, Alfa Aesar), phthalide (98%,
Aldrich), dimethylformamide (DMF, 99.9%, EMD Chemicals), anhy-
drous diethyl ether (absolute, EMD Chemicals), n-hexadecane
(99%, Sigma), 1-iodooctane (98%, Sigma), 1-butanol (99%, J. T.
Baker), chloroform-d (99.8%, Aldrich), sodium hydroxide (97%, EMD

E.M. Pasciak et al. / Electrochimica Acta 113 (2013) 557–563
559
rate-limiting. For the simulations, the objective was to find a set of
simulation parameters that provided the best agreement between
simulated and experimental cyclic voltammograms for all of the
scan rates. This last statement means that agreement between sim-
ulation and experiment for any given scan rate might be improved
by a change in one or more of the parameters but these changes,
when applied to all scan rates, would result in overall poorer agree-
ment.
further confirmed with the aid of electrospray ionization–time-of-
flight (ESI–TOF) high-resolution mass spectrometry (HRMS) m/z:
calcd for C8H7O3Na2 [M–H + 2 Na]+ 197.0191, found 197.0194.
3. Results and discussion
3.1. Cyclic voltammetric behavior of phthalide
Shown in Fig. 1 are cyclic voltammograms recorded for direct
reduction of a 3.0 mM solution of phthalide at a glassy carbon
electrode in DMF containing 0.10 M TBAP and at scan rates of
100 mV s−1 and 5 V s−1. For both scan rates, only a single cathodic
peak is observed, which we attribute to one-electron reduction of
phthalide to form a radical-anion intermediate. As is evident from
Fig. 1, curve A, the cyclic voltammogram shows a small degree of
reversibility at 100 mV s−1, whereas the cyclic voltammogram for
phthalide recorded at 5 V s−1 (Fig. 1, curve B) reveals essentially
perfectly reversible behavior, which provides evidence that the
intermediate is in fact a relatively stable radical-anion. Although we
have not shown cyclic voltammograms for reduction of phthalide
in DMF–0.10 M TMAP, the results are quite similar to those depicted
in Fig. 1; regardless of whether TBAP or TMAP is employed as
supporting electrolyte, the cathodic peak potential for reduction
of phthalide does not differ for a chosen scan rate. It is interest-
ing to compare the cathodic (Epc) and anodic (Epa) peak potentials
for reduction of phthalide in DMF–0.10 M TBAP at a hanging mer-
2.4. Separation, identification, and quantitation of electrolysis
products
At the end of each controlled-potential (bulk) electrolysis, a
1.5-mL aliquot of the catholyte was set aside for direct analy-
sis by means of gas chromatography (GC), and the remaining
catholyte was partitioned between diethyl ether and 10% aque-
ous hydrochloric acid [7]. After being washed twice more with 10%
aqueous hydrochloric acid, the ether phase was dried over anhy-
drous sodium sulfate, filtered, concentrated with the aid of rotary
evaporation, and analyzed by means of GC. An Agilent 7890A gas
chromatograph equipped with a flame-ionization detector and a
30 m × 0.25 mm capillary column (J & W Scientific) with a DB-5
stationary phase was utilized. Because this paper deals with chem-
istry that occurs within the heated injector port of our GC system,
we provide here some additional information; for all GC measure-
ments, a split injector with a glass liner packed with glass wool
(Restek Part No. 20994, Restek Corp., Bellefonte, PA) was used with
a split ratio of either 30:1 or 15:1. Details concerning the quanti-
tation procedure are described in a previous publication [18]. In
addition, we separated and identified the products with the aid
of GC and gas chromatography–mass spectrometry (GC–MS) by
comparing retention times and mass spectra with those of com-
mercially available or chemically synthesized authentic samples.
Peak areas for the different products were determined with respect
to an internal standard (n-hexadecane), added in known amount to
a solution before the start of each electrolysis.
cury drop [7] and at glassy carbon: (i) at a scan rate of 50 mV s−1
,
Epc = −1.47 V and Epa = −1.40 V for mercury, whereas Epc = −1.45 V
and Epa = −1.36 V for glassy carbon at a scan rate of 100 mV s−1; (ii)
at a scan rate of 5 V s−1, Epc = −1.51 V and Epa = −1.43 V for mercury
and Epc = −1.58 V and Epa = −1.27 V for glassy carbon.
To acquire more information about the electrochemistry of
phthalide, we compared simulated with experimental cyclic
voltammograms that were recorded at seven different scan rates
ranging from 0.1 to 5 V s−1 for reduction of 3.0 mM phthalide in
DMF–0.10 M TBAP at a glassy carbon electrode (area = 0.071 cm2);
three of these comparisons are shown in Fig. 2. For each of the
cyclic voltammograms, there is one cathodic peak, with a single
anodic peak being seen on the reverse scan. At the lowest scan rate
(0.1 V s−1), the anodic peak is relatively smaller, due to loss of the
phthalide radical-anion via its ring-opening and further reduction.
At the highest scan rate (5 V s−1), the extent of ring-opening is negli-
gible, and the pair of cathodic and anodic peaks corresponds simply
to the first electron-transfer step, namely the reversible reduction
of phthalide to its radical-anion. Electronic compensation of the
solution resistance was not applied with the potentiostat used for
these experiments, so all of this resistance (300 ꢀ) was incorpo-
rated into the simulations, thus accounting well for the greater
cathodic–anodic peak separation as the scan rate is increased.
Peak separation is also affected by the value of the standard
electron-transfer rate constant (ks,1) for the one-electron reduction
2.5. Synthesis of n-butyl 2-methylbenzoate
Esterification of 2-methylbenzoic acid was conducted by means
of an acid-catalyzed reaction in the presence of 1-butanol. To a
stirred solution of 10 mL of 1-butanol and 1 g of 2-methylbenzoic
acid (7.3 mmol), 1 mL of concentrated (12 M) hydrochloric acid was
added slowly, and the mixture was heated to reflux and allowed
to stir for an additional 2 h. Remaining alcohol and hydrochloric
acid were removed under reduced pressure to afford n-butyl 2-
methylbenzoate: 1H NMR (400 MHz, CDCl3) ı 0.98 (t, 3H), 1.48 (m,
2H), 1.75 (m, 2H), 2.6 (s, 3H), 4.3 (t, 2H), 7.23 (m, 2H), 7.38 (t, 1H),
7.91 (d, 1H); high-resolution (electron impact) mass spectrometry
(HRMS) m/z: calcd for C12H16O2 [M+] 192.1145, found 192.1142.
2.6. Synthesis of sodium 2-methylbenzoate
of phthalide; the value used for the simulations was 0.03 cm s−1
.
Furthermore, the electron-transfer rate constant and the uncom-
pensated resistance affect the cyclic voltammograms similarly,
such that small increases in resistance and small increases in the
rate constant provide similar agreement to that seen in Fig. 2. Thus,
there is some uncertainty in these values. Those parameters chosen
tended to give the best agreement for the shapes of the simulated
voltammograms at all seven scan rates. In contrast, the value of the
rate constant (kf,2) for irreversible ring-opening of the phthalide
radical-anion (C-step) was found to be well determined at 0.20 s−1
(with an uncertainty of about 20%), which corresponds to a half-
life for the radical-anion of 3.5 s; this value can be compared with
that (0.7 s) obtained in our earlier study [7], although the reason
for the difference is unknown. Other simulation parameters ulti-
mately used for this work will be presented later when a complete
To 100 mL of 1% aqueous sodium hydroxide solution, 3.4 g
(25.0 mmol) of 2-methylbenzoic acid was added and the mixture
was stirred for 30 min. Then the solution was filtered and allowed
to cool to near 0 ◦C to crystallize the newly formed salt, which was
subsequently filtered and dried under reduced pressure.
2.7. Synthesis of 2-(hydroxymethyl)benzoic acid
We prepared 2-(hydroxymethyl)benzoic acid according to a
procedure published by La Spisa and co-workers [19], which
involves base hydrolysis of phthalide, followed by acidification
with hydrochloric acid; the 1H NMR spectrum was identical to that
reported in the just-cited paper, and the identity of the product was

560
E.M. Pasciak et al. / Electrochimica Acta 113 (2013) 557–563
Fig. 1. Cyclic voltammograms recorded for reduction of 3.0 mM phthalide at a glassy carbon electrode (area = 0.071 cm2) in oxygen-free DMF containing 0.10 M TBAP at scan
rates of 0.1 and 5 V s−1. Scans go from −1.0 to −1.8 to −1.0 V.
overview of all of the mechanistic steps associated with the elec-
trochemical reduction of phthalide is discussed in Section 3.7.
electrolytes, the coulometric n value was found to be 1.26 0.02,
which indicates that the predominant process is an overall one-
electron reduction. Compiled in Table 1 are product distributions
for all of these experiments; each entry is the average of three
separate electrolyses. As a prelude to the final separation, identi-
fication, and quantitation of the electrolysis products (as well as of
phthalide itself) by means of GC and GC–MS, each post-electrolysis
solution (catholyte) was subjected to two different analytical pro-
cedures. First (denoted “direct analysis” in Table 1), a small aliquot
of the catholyte was injected directly into a gas chromatograph
for quantitation of products. Second (denoted “acid extraction” in
Table 1), the remaining catholyte was partitioned between diethyl
ether and 10% aqueous hydrochloric acid solution; the organic
3.2. Controlled-potential (bulk) electrolyses of phthalide (1) at
reticulated vitreous carbon cathodes in DMF containing different
tetraalkylammonium perchlorates as electrolytes
Bulk electrolyses of 5.0 mM solutions of phthalide were carried
out at reticulated vitreous carbon cathodes held at −1.60 V in
DMF containing 0.10 M electrolyte (TMAP or TBAP). Our choice
of the electrolysis potential for these experiments was intended
to emulate the conditions employed in our earlier investiga-
tions performed with mercury electrodes [7,8]. For both of these
Fig. 2. Comparison of experimentally acquired (solid curves) cyclic voltammograms for reduction of 3.0 mM phthalide in DMF–0.10 M TBAP with digitally simulated cyclic
(open circles) voltammograms at scan rates of 0.1, 1, and 5 V s−1. Scans go from −1.0 to −1.8 to −1.0 V. Simulation parameters are included as part of Scheme 2 which presents
the complete mechanistic picture. See the text for further discussion.

E.M. Pasciak et al. / Electrochimica Acta 113 (2013) 557–563
561
Table 1
Product distributions obtained by means of two different post-electrolysis procedures after direct reduction of 5.0 mM phthalide (1) at reticulated vitreous carbon cathodes
held at −1.60 V in DMF containing 0.10 M TMAP or TBAP.
Entry
Cationb
Post-electrolysis procedure
Product distribution (%)a
Phthalide (1)
2-Methylbenzoate ester (7)
2-Methylbenzoic acid (8)
Total
1
2
3
4
TMA+
TMA+
TBA+
TBA+
Direct analysis
Acid extraction
Direct analysis
Acid extraction
26
28
15
18
28c
–
–
57
–
54
85
57
42d
–
84
102
a
Yield expressed as percentage of starting material incorporated into each product.
b
c
Cation of supporting electrolyte: TMA+, tetramethylammonium; TBA+, tetra-n-butylammonium.
Methyl 2-methylbenzoate.
n-Butyl 2-methylbenzoate.
d
layer was washed twice with 10% aqueous hydrochloric acid, dried
over anhydrous sodium sulfate, and concentrated by means of
rotary evaporation, before being analyzed by means of GC.
corresponding to the end of a controlled-potential electrolysis of
phthalide. When this solution was subjected to our diethyl ether
−10% aqueous hydrochloric acid extraction procedure described
earlier, GC–MS analyses revealed the presence of only phthalide,
and the same result was obtained when the sodium salt of 2-
(hydroxymethyl)benzoic acid was tested. To place this finding in
more precise terms, a 9.0 mM solution of the acid in DMF–0.10 M
TBAP was prepared, and an aliquot of this solution was injected into
our GC system; it should be noted that the sodium salt of the acid is
not sufficiently soluble in DMF to permit a quantitative experiment.
By means of a calibration step with a known amount of phthalide,
we determined that conversion of 2-(hydroxymethyl)benzoic acid
to phthalide is quantitative (within experimental error). Therefore,
we conclude that, under the conditions of our procedure for the
identification of electrolysis products, the phthalide observed at the
end of an electrolysis arises from an efficient thermal conversion
of 2-(hydroxymethyl)benzoate (or its conjugate acid) to phthalide
in the injector port of a gas chromatograph. Precedent for the facile
lactonization of 2-(hydroxymethyl)benzoic acid can be found in the
literature [20].
For both supporting electrolytes and for both post-electrolysis
analytical procedures, phthalide (1) was always observed, even
though the current always decreased to the background level of
a blank experiment. Direct analysis (entries 1 and 3 of Table 1)
revealed the presence of a 2-methylbenzoate ester (7), although
a total of only 50–60% of the starting material can be accounted
for in these product distributions. To explain this low recovery
of products, we propose the formation of a stable intermediate
(2-methylbenzoate anion) that cannot be detected directly with
the aid of GC. We suggest that a fraction of this intermediate
can interact with the tetraalkylammonium cation to form a 2-
methylbenzoate ester (7), as an aliquot of the catholyte passes
through the GC injector port at a temperature of 285 ◦C. On the
other hand, when the acid extraction procedure (entries 2 and 4
of Table 1) is employed, the 2-methylbenzoate anion is converted
quantitatively to 2-methylbenzoic acid (8), which can be efficiently
detected by means of GC; thus, more of the starting material can
be accounted for in the post-electrolysis procedure. These findings
are considered in greater detail in the discussions below.
3.4. Use of sodium 2-methylbenzoate to confirm that esters are
formed via injector-port chemistry
Because of the possibility that the presence of phthalide among
the electrolysis products is due to the fact that the electrolysis
potential used to collect data compiled in Table 1 was only −1.60 V,
we explored the consequences of bulk electrolyses conducted at
both −1.80 and −2.00 V, potentials that are well beyond the value
needed theoretically to reduce all of the starting material. Solutions
containing 5.0, 10.0, and 20.0 mM phthalide in DMF–0.10 M TBAP
were electrolyzed at −1.80 V, and a 20.0 mM solution of phthalide
in the same solvent-supporting electrolyte was reduced at −2.00 V.
For all of these electrolyses, the coulometric n value was essentially
the same as that seen for the experiments performed at −1.60 V.
At the conclusion of each electrolysis, a sample of the solution was
subjected to both of the procedures (direct analysis and acid extrac-
tion) used to identify and quantitate the products. For the 5.0 mM
solution, no phthalide was detected; however, for the experiments
involving electrolysis of 10.0 and 20.0 mM phthalide, starting mate-
rial was always seen in yields quite comparable to those listed in
Table 1. Thus, even for electrolyses carried out at significantly neg-
ative potentials, phthalide is found among the products. In the next
section of this paper, we reveal that the presence of phthalide at the
end of an electrolysis can be attributed to the thermal conversion of
2-(hydroxymethyl)benzoate back to phthalide in the injector port
of a GC or GC–MS system.
To discover why esters appear as electrolysis products, we
carried out experiments involving the use of a model com-
pound, sodium 2-methylbenzoate, to examine further the results
described above. A solution of sodium 2-methylbenzoate was pre-
pared in DMF–0.10 M TBAP. When an aliquot of this solution was
injected into the GC–MS system, we observed that n-butyl 2-
methylbenzoate had been formed. When a 1H NMR spectrum of an
authentic sample of n-butyl 2-methylbenzoate was compared with
that of the solution of sodium 2-methylbenzoate in DMF–0.10 M
TBAP, the latter sample showed no resonance signal (triplet) at
4.3 ppm due to ˛ hydrogen atoms of a butyl moiety of the ester.
This result provides more evidence that injector-port chemistry
is responsible for the formation of the observed esters; as indi-
cated in the experimental section of this paper, we employed a split
injector with a glass liner that provides a large surface upon which
2-methylbenzoate and a tetraalkylammonium cation can interact.
Finally, the solution of sodium 2-methylbenzoate was subjected
to the acidic extraction procedure, and only 2-methylbenzoic acid
(4) was detected by means of GC. Thus, our chemically synthe-
sized 2-methylbenzoate anion shows behavior identical to that of
a proposed anionic product in a post-electrolysis solution.
3.3. Conversion of 2-(hydroxymethyl)benzoic acid or its
conjugate base to phthalide in the injector port of a gas
chromatograph
3.5. Trapping of the 2-methylbenzoate anionic product by
addition of 1-iodooctane to a post-electrolysis solution
To explore further the ability of the 2-methylbenzoate anion
to participate in an SN2 reaction, a 5.0 mM solution of phthalide
in DMF–0.050 M TMAP was electrolyzed at a reticulated vitreous
A
solution containing 15 mg of 2-(hydroxymethyl)benzoic
acid in DMF–0.10 M TBAP was prepared to mimic conditions

562
E.M. Pasciak et al. / Electrochimica Acta 113 (2013) 557–563
Scheme 1. Mechanistic features of the electrochemical reduction of phthalide in the presence of D2O. See the text for further discussion.
Scheme 2. Complete mechanistic picture (including parameters used for digital simulation of cyclic voltammograms) for the electrochemical reduction of phthalide. Simu-
lation parameters used to prepare Fig. 2 are included alongside the pertinent mechanistic steps; in particular, the parameters for reactions (3) and (4) were selected to assure
rapid reduction at the electrode or in solution in the range of potentials covered in the cyclic voltammograms. See the text for further discussion.

E.M. Pasciak et al. / Electrochimica Acta 113 (2013) 557–563
563
carbon cathode. At the end of the electrolysis, pure 1-iodooctane
was added to the catholyte to obtain a 10 mM concentration of the
alkyl iodide, and the resulting mixture was stirred for 60 min. A
portion of this solution was taken for direct analysis by GC–MS,
which revealed the presence of n-octyl 2-methylbenzoate (along
with the phthalide arising from 2-(hydroxymethyl)benzoate, as
described above). Then the remaining solution was subjected to the
acid-extraction procedure followed by GC–MS analysis, and n-octyl
2-methylbenzoate and phthalide were the only products, meaning
that trapping of the 2-methylbenzoate anion by the alkyl iodide
precluded the formation of 2-methylbenzoic acid.
n-octyl 2-methylbenzoate, and (iii) gain a proton, during the acid-
extraction procedure, to give 2-methylbenzoic acid. Finally, we
suggest that, in the presence of an electrogenerated base (includ-
ing hydroxide ion arising from adventitious water), phthalide is
converted to 2-(hydroxymethyl)benzoate (6) (Scheme 2, reac-
tion (6)), which undergoes thermal conversion back to phthalide
in the injector port of the gas chromatograph. Thus, one of the
remarkable facts about the electrochemical reduction of phthalide
appears to be that, under the conditions of our experiments, the
final products in the catholyte are 2-methylbenzoate (5) and 2-
(hydroxymethyl)benzoate (6).
3.6. Controlled-potential (bulk) electrolysis of phthalide (1) in
Acknowledgments
DMF–0.10 M TBAP containing D2O
High-resolution mass analyses were performed in the Indiana
University Mass Spectrometry Facility; the HRMS system was pur-
chased with funds provided by the National Institutes of Health
Grant No. 1S10RR016657-01.
A 5.0 mM solution of phthalide in DMF containing 0.10 M TBAP
was electrolyzed at −1.60 V in the presence of 200 mM D2O. We
used the two post-electrolysis procedures (direct analysis and
acid extraction) to determine the products. As expected on the
basis of Table 1, direct analysis of the catholyte revealed the pres-
ence of two compounds (n-butyl 2-methylbenzoate and phthalide),
whereas the acid-extraction protocol identified 2-methylbenzoic
acid and phthalide as the products. All species arising from these
experiments displayed masses (m/z) that ranged from nondeuter-
ated compounds to those with double deuteration, and some of
the 2-methylbenzoic acid was triply deuterated; even phthalide
found at the end of an electrolysis showed deuterium incorpo-
ration. To confirm that phthalide does not undergo deuterium
exchange under nonelectrochemical conditions, a 5.0 mM solution
of phthalide in DMF was stirred in the presence of 200 mM D2O
for the same amount of time (45 min) that an electrolysis would
normally require; no incorporation of deuterium was observed.
Scheme 1 shows a proposed sequence of reactions that cause
deuteration of phthalide during the course of an electrolysis. Since
phthalide itself can acquire as many as two deuterons, pathways to
the formation of deuterated products are opened. Because its acidic
proton can undergo a facile exchange with a deuteron from D2O,
2-methylbenzoic acid can appear with as many as three deuterium
atoms.
References
[1] L. Eberson, K. Nyberg, in: A.J. Bard, H. Lund (Eds.), Encyclopedia of Electrochem-
istry of the Elements, vol. 12, Marcel Dekker, New York, 1978, p. 261.
[2] J.H. Wagenknecht, L. Eberson, J.H.P. Utley, in: H. Lund, O. Hammerich (Eds.),
Organic Electrochemistry, 4th ed., Marcel Dekker, New York, 2001, p. 453.
[3] S. Kashimura, M. Ishifune, in: H.J. Schäfer (Ed.), Encyclopedia of Electrochem-
istry, vol. 8, Wiley-VCH, Weinheim, 2004, p. 199.
[4] S. Torii, Electroorganic Reduction Synthesis, Kodansha/Wiley-VCH,
Tokyo/Weinheim, 2006, p. 101.
[5] D.W. Leedy, D.L. Muck, Cathodic reduction of phthalimide systems in nonaque-
ous solutions, Journal of the American Chemical Society 93 (1971) 4264.
[6] A. Benˇo, P. Hrncˇiar, M. Lácová, Phthalides and indandiones. XLVIII. Polaro-
graphic reduction and the transmission of electronic effects in substituted
3-arylmethylenephthalides, Collection of Czechoslovak Chemical Communi-
cations 37 (1972) 3295.
[7] M.L. Vincent, D.G. Peters, Electrolytic cleavage of acyclic and cyclic aromatic
esters. Case studies of the reductions of benzyl benzoate and phthalide, Journal
of Electroanalytical Chemistry 327 (1992) 121.
[8] S.E. Hochstetler, Indiana University, Bloomington, Indiana, USA, 1994 (B.S. the-
sis).
[9] K.L. Vieira, D.G. Peters, Voltammetric behavior of tertiary butyl bromide at mer-
cury electrodes in dimethylformamide, Journal of Electroanalytical Chemistry
196 (1985) 93.
[10] L.W. Marple, Reference electrode for anhydrous dimethylformamide, Analyti-
cal Chemistry 39 (1967) 844.
[11] C.W. Manning, W.C. Purdy, Reference electrode for electrochemical studies in
dimethylformamide, Analytica Chimica Acta 51 (1970) 124.
[12] J.L. Hall, P.W. Jennings, Modification of the preparation of a cadmium amalgam
reference electrode for use in N,N-dimethylformamide, Analytical Chemistry
48 (1976) 2026.
[13] K.N. Knust, M.P. Foley, M.S. Mubarak, S. Skljarevski, K. Raghavachari, D.G.
Peters, Electrochemical reduction of 5-chloro-2-(2,4-dichlorophenoxy)phenol
(triclosan) in dimethylformamide, Journal of Electroanalytical Chemistry 638
(2010) 100.
[14] A.A. Peverly, J.A. Karty, D.G. Peters, Electrochemical reduction of (1R,2r,
3S,4R,5r,6S)-hexachlorocyclohexane (Lindane) at silver cathodes in organic
and aqueous–organic media, Journal of Electroanalytical Chemistry 692 (2013)
66.
[15] J.A. Cleary, M.S. Mubarak, K.L. Vieira, M.R. Anderson, D.G. Peters, Elec-
trochemical reduction of alkyl halides at vitreous carbon cathodes in
dimethylformamide, Journal of Electroanalytical Chemistry 198 (1986) 107.
[16] M. Rudolph, Digital simulations on unequally spaced grids. Part 2. Using the
box method by discretisation on a transformed equally spaced grid, Journal of
Electroanalytical Chemistry 543 (2003) 23.
[17] J.-M. Savéant, Elements of Molecular and Biomolecular Electrochemistry. An
Electrochemical Approach to Electron Transfer Chemistry, Wiley, New York,
2006, p. 96.
[18] W.A. Pritts, K.L. Vieira, D.G. Peters, Quantitative determination of volatile prod-
ucts formed in electrolyses of organic compounds, Analytical Chemistry 65
(1993) 2145.
[19] F. La Spisa, A. Feo, R. Mossetti, G.C. Tron, An efficient synthesis of symmetric and
unsymmetric bis-(␤-aminoamides) via Ugi multicomponent reaction, Organic
Letters 14 (2012) 6044.
[20] N. Tomoto, W.J. Boyle Jr., J.F. Bunnett, Confirmation of concurrent general acid,
general base catalysis in the lactonization of 2-(hydroxymethyl)benzoic acid,
Journal of Organic Chemistry 37 (1972) 4315.
3.7. Mechanistic aspects of the reduction of phthalide
On the basis of the various experiments described above, we
conclude that the mechanism proposed in an earlier paper [7]
for the reduction of phthalide depicted in Scheme 2 is opera-
tive. Included in this figure are the parameters used to obtain
the simulated cyclic voltammograms shown in Fig. 2. Accordingly,
we propose that the first step in the reduction of phthalide (1)
is a reversible one-electron process to form a radical-anion (2)
(Scheme 2, reaction (1)). Once produced, the radical-anion (2)
undergoes ring-opening to give 3 (Scheme 2, reaction (2)), with
a half-life of 3.5 s. In reaction (3) of Scheme 2, radical-anion 3
accepts an electron from the carbon cathode to give dianion 4. A key
process involves a solution-phase electron-transfer (DISP1) pro-
cess to afford 4 and to regenerate 1 (Scheme 2, reaction (4)); any
1 regenerated near the electrode surface should undergo reduc-
tion and thereby contribute to the slightly elevated n value of
1.26 mentioned in Section 3.2. Reaction (5) of Scheme 2 shows
the reversible conversion of 4 to 2-methylbenzoate (5); the latter
species can (i) interact with the cation of the supporting elec-
trolyte via an SN2 pathway to form an ester in the injector port of
a gas chromatograph, (ii) react nucleophilically with 1-iodooctane,
introduced at the end of a bulk electrolysis of phthalide, to yield