Effects of surface monolayers on the electron-transfer kinetics and adsorption of methyl viologen and phenothiazine derivatives on glassy carbon electrodes

Article abstract of DOI:10.1021/ac9902392

Five organic redox systems were examined in aqueous electrolytes on polished and chemically modified glassy carbon (GC), to evaluate the effects of surface structure on the heterogeneous transfer rate constant, k°. Methyl viologen reduction to its cation radical exhibited a voltammetric peak potential difference which was insensitive to surface modification, with k°decreasing by only 50% when a chemisorbed monolayer was present. Methylene blue and three other phenothiazines adsorbed to polished GC, but the adsorption was suppressed by surface modification. For all four phenothiazines, chemisorbed or physisorbed monolayers of electroinactive species had minor effects on k°, with a compact nitrophenyl monolayer decreasing k°by 50%. This minor change in k°was accompanied by a major decrease in adsorption, apparently due to inhibition of dipole-dipole or π- π interactions between the phenothiazine and GC. Chlorpromazine oxidation to its cation radical was studied in more detail, under conditions where adsorption was suppressed. A plot of the natural log of the observed rate constant vs the monolayer thickness for a variety of chemisorbed monolayers was linear, with a slope of -0.22 A^-1. The observations are consistent with a through-bond electron-tunneling mechanism for electron transfer to all five redox systems studied. The tunneling constant for CPZ of 0.22 A^-1 is between that reported for electron tunneling through conjugated polyene spacers (0.14 A^-1) and that reported for phenyl-methylene spacers (0.57 A^-1), on the basis of long-range electron transfer in rigid molecules.

Full text of DOI:10.1021/ac9902392

Anal. Chem. 1999, 71, 4081-4087
Effe c t s o f S u rfa c e Mo n o la ye rs o n t h e
Ele c t ro n -Tra n s fe r Kin e t ic s a n d Ad s o rp t io n o f
Me t h yl Vio lo g e n a n d P h e n o t h ia zin e De riva t ive s o n
Gla s s y Ca rb o n Ele c t ro d e s
Hs e uh-Hui Ya ng a nd R. L. Mc Cre e ry*
Department of Chemistry, The Ohio State University, 100 West 18th Avenue Columbus, OH 43210
Five organic redox systems were examined in aqueous
electrolytes on polished and chemically modified glassy
carbon (GC), to evaluate the effects of surface structure
on the heterogeneous transfer rate constant, k°. Methyl
viologen reduction to its cation radical exhibited a volta-
mmetric peak potential difference which was insensitive
to surface modification, with k° decreasing by only 5 0 %
when a chemisorbed monolayer was present. Methylene
blue and three other phenothiazines adsorbed to polished
GC, but the adsorption was suppressed by surface modi-
fication. For all four phenothiazines, chemisorbed or
physisorbed monolayers of electroinactive species had
minor effects on k°, with a compact nitrophenyl monolayer
decreasing k° by 5 0 %. This minor change in k° was
accompanied by a major decrease in adsorption, appar-
ently due to inhibition of dipole-dipole or π-π interactions
between the phenothiazine and GC. Chlorpromazine
oxidation to its cation radical was studied in more detail,
under conditions where adsorption was suppressed. A
plot of the natural log of the observed rate constant vs the
monolayer thickness for a variety of chemisorbed mono-
degradation via oxidation and impurity adsorption, our under-
standing of the behavior of carbon electrodes has lagged that of
metal electrodes, particularly mercury ones. This situation im-
proved dramatically after more attention was paid to surface
preparation and the number of uncontrolled surface variables was
reduced. In particular, several landmarks indicating reproducible
2
performance of sp carbon electrodes, mainly glassy carbon (GC),
have been achieved:
1. Determination of the rapid heterogeneous electron-transfer
rate constants (k°), for outer-sphere systems (e.g., Ru(NH3)6+3/ +2
k° > 0.2 cm/ s),8,9 comparable to those observed on Au and Pt.10
2. Preparation of low-oxide (O/ C < 2%) carbon surfaces which
1
1,12
retain their low oxide levels for at least one month in air.
3. Structural characterization of organic monolayers and
submonolayers on carbon with Raman spectroscopy.13-15
4. Correlation of specific surface sites with electrocatalytic
activity for various redox systems, including ascorbic acid, NADH,
Fe3+/ 2+, etc.8,16-19
5. Systematic classification of redox systems according to their
sensitivity to surface chemistry. Classes include outer sphere
+3/ +2, Co(en) +3/ +2, etc.), systems catalyzed
3
systems (e.g., Ru(NH
3
)
6
-1
layers was linear, with a slope of -0.22 Å . The observa-
tions are consistent with a through-bond electron-tunnel-
ing mechanism for electron transfer to all five redox
systems studied. The tunneling constant for CP Z of 0 .2 2
by specific surface oxides (Feaq+3/ +2, Vaq+3/ +2, Euaq+3/ +2), and
-3/ -4, ascor-
systems requiring a non-oxide surface site (Fe(CN)
bate).19,20
6
With reproducible carbon electrodes in-hand, it is possible to
systematically examine the surface structural factors which control
-
1
Å
is between that reported for electron tunneling
-
1
through conjugated polyene spacers (0 .1 4 Å ) and that
reported for phenyl-methylene spacers (0 .5 7 Å ), on the
-1
(
5) Wightman, R. M.; Deakin, M. R.; Kovach, P. M.; Kuhr, P. M.; Stutts, K. J.
J. Electrochem. Soc. 1 9 8 4 , 131, 1578.
basis of long-range electron transfer in rigid molecules.
(
(
(
6) Nowall, W. B.; Kuhr, W. G. Anal. Chem. 1 9 9 5 , 67, 3583.
7) Xu, J.; Chen, Q.; Swain, G. M. Anal. Chem. 1 9 9 8 , 70, 3146.
8) Chen, P.; Fryling, M. A.; McCreery, R. L. Anal. Chem. 1 9 9 5 , 67, 3115.
Through a variety of efforts from many laboratories, significant
progress has been made toward understanding the electrochemi-
(9) Jaworski, R. K.; McCreery, R. L. J. Electroanal. Chem. 1 9 9 4 , 369, 175.
10) Iwasita, T.; Schmickler, W.; Schultze, J. W. Ber. Bunsen-Ges. Phys. Chem.
(
cal behavior of widely used carbon electrodes.1 Since sp carbon
-7
2
1
9 8 5 , 89, 138.
surfaces are difficult to prepare reproducibly and are prone to
(
(
11) Fagan, D. T.; Hu, I. F.; Kuwana, T. Anal. Chem. 1 9 8 5 , 57, 2759.
12) Kuo, T.-C.; McCreery, R. L., Anal. Chem. 1 9 9 9 , 71, 1553.
*
Corresponding author. Tel.: 614-292-2021. Fax: 614-292-1685. E-mail:
(13) Kagan, M. R.; McCreery, R. L. Langmuir 1 9 9 5 , 11, 4041.
(14) Liu, Y.-C.; McCreery, R. L. J. Am. Chem. Soc. 1 9 9 5 , 117, 11254.
(15) Liu, Y.-C.; McCreery, R. L. Anal. Chem. 1 9 9 7 , 69, 2091.
(16) McDermott, C. A.; Kneten, K. R.; McCreery, R. L. J. Electrochem. Soc. 1 9 9 3 ,
140, 2593.
mccreery.2@osu.edu.
(
1) McCreery, R. L. Carbon Electrodes: Structural Effects on Electron-Transfer
Kinetics. In Electroanalytical Chemistry; A. Bard, Ed.; Marcel Dekker: New
York, 1991; Vol. 17.
(
(
(
2) Kinoshita, K. Carbon: Electrochemical and Physicochemical Properties;
Wiley: New York, 1988.
3) McCreery, R. L.; Cline, K. K.; McDermott, C. A.; McDermott, M. T. Colloids
and Surf. 1 9 9 4 , 93, 211.
(17) Deakin, M.; Stutts, K.; Wightman, M. J. Electroanal. Chem. 1 9 8 5 , 182, 113.
(18) Tse, D. C. S.; Kuwana, T. K. Anal. Chem. 1 9 7 8 , 50, 1315.
(19) Chen, P.; McCreery, R. L. Anal. Chem. 1 9 9 6 , 68, 3958.
(20) McCreery, R. L. Electrochemical Behavior of Carbon Surface. In Interfacial
Electrochemistry; A. Wieckowski, Ed., Marcel Dekker: N.Y., 1999.
4) Hu, I. F.; Karweik, D. H.; Kuwana, T. J. Electroanal. Chem. 1 9 8 5 , 188, 59.
1
0.1021/ac9902392 CCC: $18.00 © 1999 American Chemical Society
Analytical Chemistry, Vol. 71, No. 18, September 15, 1999 4081
Published on Web 08/03/1999

electron transfer reactivity, adsorption, capacitance, etc. In previ-
ously reported experiments, we used several electrode modifica-
tions to classify a collection of mainly inorganic redox systems
according to their sensitivity to surface structure.19 Outer-sphere
6
Benzenediazonium fluoborate, NMR (d -DMSO, 250 MHZ): δ
) 8.00, 8.27, 8.86 (m, 5H), MS (FAB) calcd for C
6
H
5
N
2
m/ z 105.12,
+
found m/ z 104.98 (M-BF
4
) .
4-Methylphenyl fluoborate, NMR (d -DMSO, 250 MHZ): δ )
6
+
3/ +2
systems such as Ru(NH
3
)
6
showed minor kinetic effects of
2.59 (s, 3H), 7.81, 8.56 (d, 4H, J ) 8.45 Hz), MS (FAB) calcd for
+
surface modification, even when a compact organic monolayer
C
7
H
7
N
2
m/ z 119.15, found m/ z 119.07 (M-BF
4
) .
was chemisorbed to the surface before kinetic measurements. In
6
Trifluoromethylbenzenediazonium fluoborate, NMR (d -DM-
contrast, rate constants for Fe3
+/ 2+
and related systems were
SO, 250 MHZ): δ ) 8.30, 8.92 (d, 4H, J ) 8.89 Hz), MS (FAB)
+
dramatically affected by surface preparation, due to electrocatalysis
calcd for C
Ethylbenzenediazonium fluoborate, NMR (d
MHZ): δ ) 1.25 (t, 3H, J ) 7.5 Hz), 2.88 (q, 2H, J ) 7.5 Hz),
7
H
4
N
2
F
3
m/ z 173.12, found m/ z 173.05 (M-BF
4
) .
8
by surface carbonyl groups. For these systems, the outer-sphere
6
-DMSO, 250
rate in the absence of carbonyl groups was comparable to that
observed on metals, and carbonyl groups increased the observed
rate by electrocatalysis.
7.85, 8.58 (d, 4H, J ) 8.6 Hz), MS (FAB) calcd for C
8
H
9
N
2
m/ z
+
133.17, found m/ z 133.36 (M-BF
4
) .
While a framework relating surface modification to kinetic
effects on GC is quite useful for the ∼15 redox systems classified
to date, its generality is unknown. Ascorbic acid was the only
organic redox system examined under the same procedure, and
organic redox reactions are generally more complex. The current
work was undertaken to identify carbon surface structural effects
on electron-transfer kinetics for several organic redox systems,
some with biological importance. Chlorpromazine, promazine,
triflupromazine, methylene blue, and methyl viologen were
subjected to the systematic analysis used previously for inorganic
redox reactions. The results reveal correlations between redox
mechanism and surface effects and provide unexpected new
information on the effect of surface structure on reactant adsorp-
tion. In addition, the approach addresses the issue of long-range
electron transfer between carbon surfaces and organic redox
systems in solution.
Phenylbenzenediazonium fluoborate, NMR (d
MHZ): δ ) 8.33, 8.74 (d, 4H, J ) 8.81 Hz), δ ) 7.69 (m, 5H), MS
6
-DMSO, 250
+
(
FAB) calcd for C12
H
9
N
2
m/ z 181.22, found m/ z 181.09 (M-BF
4
) .
Nitroazobenzene diazonium fluoborate, MS(FAB) calcd for
+
C
12
H
8
N
5
O m/ z 254.3, found m/ z 254.11 (M-BF
4
) .
Electrode Materials and P olishing P rocedure. Commercial
glassy carbon (GC 20) electrodes from Bioanalytical Systems Inc.
MF2070) were used in this work. Before any modification
procedures, electrodes were polished successively in 1 µm, 0.3
µm, and 0.05 µm alumina powder (Buehler) slurries with Nan-
opure water (Barnstead) on microcloth polishing cloth (Buehler)
and subsequently washed and sonicated in Nanopure water for
about 10 min. A low-oxide GC surface was prepared by polishing
in cyclohexane/ alumina slurries instead of Nanopure water/
alumina slurries. Cyclohexane was first saturated with argon for
(
1
5-20 min. One, 0.3, 0.05 µm alumina slurried with cyclohexane
was used successively. Electrodes were polished on bare glass
plates and were sonicated in cyclohexane for 3 min and then in
Nanopure water for another 10 min.
Nonspecific Adsorption. AQDS, methylene blue, and BMB
were adsorbed onto glassy carbon according to the previously
described procedures.13,19 GC electrodes preadsorbed with 2,6-
AQDS were prepared by placing a polished electrode in 10 mM
EXPERIMENTAL SECTION
Reagents. Tetrabutylammonium tetrafluoroborate (NBu
4 4
BF ),
50% fluoboric acid, disodium 2,6-anthraquinonedisulfonate (AQDS),
methylene blue, 1,4-bis(2-methylstyryl)benzene (BMB), aniline,
p-toluidine, 4-ethylaniline, 4-(trifluoromethyl)aniline, 4-nitroaniline,
4-aminobiphenyl, and disperse orange 3 (4-nitro, 4′-amino azoben-
2,6-AQDS aqueous solution for 10 min and rinsing with Nanopure
zene) were purchased from Aldrich Chemical Co. and were used
as received, except AQDS was recrystallized from water. Chlor-
promazine (CPZ), promazine (PMZ), triflupromazine (TPZ),
methyl viologen, and sodium nitrite were purchased from Sigma
Chemical Company and used as received. Ether and cyclohexane
were purchased from J. T. Baker and used as received. Methylene
blue and methyl viologen occur as chloride salts and will be
referred to herein as MB+ and MV+2 in solution.
water three times. The electrodes were then transferred into
electrochemical cells for measurement of voltammetric data. GC
electrodes preadsorbed with methylene blue were prepared by
dipping a polished electrode in 0.1 mM methylene blue in water
for 10 min and rinsing with Nanopure water three times. GC
electrodes with preadsorbed BMB were prepared by placing a
polished electrode in a 1 mM BMB/ acetone solution for 10 min
and rinsing with acetone three times.
Synthesis of Diazonium Tetrafluoroborate Salts. Diazonium
tetrafluoroborate salts were synthesized according to the proce-
dure described by Starkey et al.21 One-tenth mole of the corre-
sponding amine precursor was dissolved in 44 mL of 50% fluoboric
acid. The solution was placed in an ice bath and stirred with an
efficient stirrer, then a cold solution of 0.1 mol of sodium nitrite
in 14 mL of water was added dropwise. When the addition was
complete, the mixture was stirred for several more minutes and
then suction filtered on a sintered-glass filter. The solid diazonium
tetrafluoroborate was washed with cold fluoboric acid, ethanol,
and ether. The products, with their NMR and mass spectroscopy
data, are as follows:
Specific Adsorption. Chemisorption of aryl radicals on GC
surfaces was accomplished by the procedure developed by Saveant
et al.22-24 Electrodes were polished successively in 1, 0.3, and 0.05
µm alumina (Buehler) slurries with Nanopure water (Barnstead)
on microcloth polishing cloth (Buehler) and subsequently washed
and sonicated in Nanopure water for about 10 min, then rinsed
with acetonitrile (ACN). These polished GC surfaces were used
as cathodes for 10 min in the electrolysis of the solution containing
(
(
(
22) Allongue, P.; Delamar, M.; Desbat, B.; Fagebaume, O.; Hitmi, R.; Pinson,
J.; Saveant, J.-M. J. Am. Chem. Soc. 1 9 9 7 , 119, 201.
23) Delamar, M.; Hitmi, R.; Pinson, J.; Saveant, J.-M. J. Am. Chem. Soc. 1 9 9 2 ,
1
14, 5883.
24) Andrieux, C.; Gonzalez, F.; Saveant, J.-M. J. Am. Chem. Soc. 1 9 9 7 , 119,
(
21) Starkey, E. B. Org. Synth. 1 9 3 9 , 19, 40.
082 Analytical Chemistry, Vol. 71, No. 18, September 15, 1999
4292.
4

Ta b le 1 . P e a k P o t e n t ia ls fo r Dia zo n iu m Re d u c t io n ,
Am in e P re c u rs o rs , a n d De p o s it io n P o t e n t ia ls
diazonium salta
amine precursorb
EPc
Emodd
phenyl diazonium
aniline
-0.36
-0.36
-0.24
-0.56
-0.57
-0.44
4
-methyl phenyl diazonium
-trifluoromethylphenyl
diazonium
4-methyl aniline
4-trifluoromethyl
aniline
4
4-nitrophenyl diazonium
4-ethyl phenyl diazonium
4-biphenyl diazonium
4-nitro azobenzene-4′
diazonium
4-nitro aniline
4-ethyl aniline
4-phenyl aniline
4-nitro, 4′-amino
+0.04
-0.42
-0.15
+0.01
-0.16
-0.62
-0.35
-0.19
a
b
All are BF4 salts. Amine from which diazonium salt was synthe-
c
+
sized. Voltammetric peak potential vs Ag/ Ag for reduction of 1 mM
d
diazonium salt in 0.1 M NBu4BF4 in acetonitrile, 0.2 V/ s. Applied
potential during formation of monolayer.
Fig u re 2 . Voltammetry of 10 µM methylene blue in 0.1 M H2SO4,
V/s, no background subtraction. Solid line is polished GC; dotted
2
line is same electrode after cleaning with 2-propanol and activated
carbon; dashed line is GC following nitrophenyl derivatization.
Fig u re 1 . Cyclic voltammetry of 1 mM MV+ in 0.1 M NaCl on
polished GC (solid line) and GC following chemisorption of a
nitrophenyl monolayer (dashed line). 0.2 V/s scan rate, no background
subtraction.
2
Fig u re 3 . Voltammetry of 1 mM methylene blue in 0.1 M H2SO4, 2
V/s, no background subtraction. Solid line is polished GC, dashed is
after nitrophenyl derivatization.
1
mM of the appropriate diazonium salt and 0.1 M NBu
4 4
BF in
acetonitrile at a potential 200 mV negative of E for diazonium
p
3 6
)
+3/ +2, the nitrophenyl monolayer
was the case with Ru(NH
reduced k° by a factor of 2, implying that both the MV+2/ + and
reduction. After derivatization, the GC electrodes were rinsed with
electrolyte and sonicated in acetonitrile for 3 min and then in
Nanopure water for 10 min. The peak potentials for diazonium
reduction are listed in Table 1, along with the amine precursors,
and deposition potentials.
+3/ +2
Ru(NH
3
)
6
systems are acting as outer-sphere redox reactions.
+
Methylene blue (MB ) is generally considered to undergo a
reversible 2 e- reduction, but unlike MV / MV , MB is prone
to adsorption on the GC surface.19 Figure 2 shows voltammograms
for MB+ at low concentration (10 µM), at which the response is
+2
+
+
RESULTS
+
dominated by adsorbed MB . The voltammetric peak area
The simplest redox system studied in the current effort was
2
+
+2
corresponds to 213 pmol/ cm of MB adsorption, based on the
geometric electrode area. Pretreatment of the polished GC surface
with 2-propanol and activated carbon increases the coverage of
the one-electron reduction of methyl viologen (MV ) to its cation
radical. This widely studied reaction is believed to be outer-sphere
on carbon and diamond electrodes, forming a stable cation radical
product.25 A voltammogram for MV+2 on polished GC is shown
+
2
MB on polished GC to 265 pmol/ cm , presumably due to surface
cleaning.26 However, a nitrophenyl monolayer completely sup-
P
in Figure 1, with the small ∆E indicating fairly rapid electron
transfer. Covalent modification of the GC with a nitrophenyl
+
+
presses observable MB adsorption. Increasing the MB solution
concentration to 1 mM permits observation of diffusing as well
monolayer has little effect on the voltammetry, other than a slight
+
+
2
+
as adsorbed MB . As shown in Figure 3, the diffusion wave
decrease in background current. The k° values for MV / MV ,
centered at 0.2 V vs Ag/ AgCl is distinguishable from the
adsorption wave at ∼0.08 V. The nitrophenyl derivatization had
little effect on the diffusion wave, but it suppressed adsorption.
determined from several repetitions of the experiment in Figure
1, are 0.094 ( 0.025 cm/ s (N ) 3) on polished GC and 0.047 (
.013 cm/ s (N ) 3) on the nitrophenyl modified GC surface. As
0
+
The voltammetry indicates that MB is also behaving as an outer-
(
25) Alehashem, S.; Chambers, S.; Strojek, J.; Swain, G.; Rajeshuni, R. Anal. Chem.
9 9 5 , 67, 2812.
1
(26) Ranganathan, S.; Kuo, T.-C.; McCreery, R. L., Anal. Chem., in press.
Analytical Chemistry, Vol. 71, No. 18, September 15, 1999 4083

Fig u re 4 . Voltammetry of 1 mM CPZ in 0.2 M NaCl containing 0.01
M HCl on polished GC in the absence (solid) and presence of 40%
methanol. Two volts per second, no background subtraction.
Fig u re 5 . Adsorption isotherms for CPZ in 0.2 M NaCl, 0.01 M HCl,
40% MeOH. Abscissa is solution CPZ concentration, ΓCPZ was
1/2
determined from the Y intercept of chronocoulometric Q vs t plots,
after subtraction of double-layer charge. Circles are polished GC
surface, single triangle is from a nitrophenyl-modified surface. Line
sphere system, with a minor decrease in electron-transfer kinetics
2
is a fit of polished data to a Langmuir isotherm (r ) 0.9980).
+2
+
caused by a chemisorbed monolayer. Unlike MV , however, MB
exhibits adsorption which is strongly reduced by the presence of
an otherwise inert monolayer. The asymmetry in the voltammetric
peak heights for MB+ is presumably due to the more complex
reaction mechanism, involving 2 e- and 2 H+.
2
9
3 pmol/ cm (based on geometric area) is reached at about 1
mM solution concentration. When the GC was derivatized with a
nitrophenyl monolayer, the adsorption in 40% MeOH became
negligible as measured by chronocoulometry even for 10 mM CPZ
in solution (shown as triangle in Figure 5).
Unlike MB , CPZ adsorbed to the GC surface even when a
nitrophenyl monolayer was present. With methanol absent, the
Chlorpromazine (CPZ) and two related phenothiazines, pro-
mazine (PMZ) and trifluopromazine (TPZ), are more complex
+
+
+2
redox systems than MB and MV and were examined in much
more detail. As described elsewhere, the initial oxidation product
p
slope of a log(i ) vs log(ν) plot (for 0.02 to 20 V/ s) was 0.63 for
of CPZ, PMZ, or TPZ is a cation radical which reacts with water
CPZ on polished GC, well above the value of 0.50 expected in the
absence of adsorption. A nitrophenyl monolayer reduced the slope
to 0.56. However, the combination of 40% MeOH and the
nitrophenyl monolayer decreased the slope to 0.509, implying
greatly reduced adsorption. These results indicate that neither
0% MeOH nor nitrophenyl derivatization can completely suppress
CPZ adsorption by themselves, but the combination can. The
chronocoulometric results (Figure 5), the semi-integrals, and the
or buffer components to yield a sulfoxide.2
7-29
To avoid these
complicating reactions, the pH was adjusted to 2.0 with HCl. The
cation radicals are stable in this medium on the time scale of a
voltammogram, and the phenothiazines undergo a chemically
reversible one-electron oxidation. A second complication for the
phenothiazines is their tendency to adsorb on GC from aqueous
solutions. This property was identified when CPZ electrochemistry
was first examined in detail27 and has been exploited for elec-
troanalytical applications.30,31 The adsorption appears to arise
largely from the hydrophobic nature of the phenothiazine ring
system and can largely be suppressed by adding 40% methanol
to the electrolyte. Figure 4 shows voltammograms of CPZ at
polished GC with and without 40% methanol. Methanol removes
the obvious adsorption features but does not completely eliminate
adsorption. Semi-integration of the CPZ voltammograms shows
the slight peak on the semi-integral plateau which indicates
adsorption,32 even in the presence of 40% methanol. This adsorp-
tion is increased by pretreatment with 2-propanol and activated
4
slopes of log(i
p
) vs log(ν) plots all indicate that observable CPZ
adsorption is suppressed for nitrophenyl derivatized surfaces in
4
0% MeOH.
The effects of various surface modifications on phenothiazine
voltammetry were studied in the absence of methanol, initially,
since these modifications had been characterized previously
without methanol present.19 CPZ, PMZ, and TPZ voltammograms
were acquired at four scan rates (0.02, 0.2, 2, 20 V/ s) on polished
GC following several pretreatments. Anaerobic polishing with
cyclohexane/ alumina produced a GC surface with low surface
oxide coverage.19 AQDS (an anion), BMB, (neutral), and MB+ (a
+
carbon, as was observed for MB . Furthermore, the adsorption
cation) were preadsorbed to polished GC before phenothiazine
voltammetry. Previous results demonstrated that these adsorbates
remain on GC after transfer to aqueous electrolyte from the
adsorption solution.13,19 The rate constants from these experiments
are tabulated in Table 2, along with those for the nitrophenyl
modified surface. In all cases, k° did not show a significant trend
with scan rate, and the standard deviations listed include the entire
0.02-20 V/ sec range. Although some perturbation of these results
is expected as a result of phenothiazine adsorption, Table 2 does
demonstrate that the effects of surface derivatization are minor
for all three phenothiazines. Surface-oxide coverage and the pres-
of CPZ on polished GC adheres to a Langmuir isotherm deter-
mined with chronocoulometry (Figure 5). Saturation coverage of
(
27) Cheng, H. Y.; Sackett, P. H.; McCreery, R. L. J. Am. Chem. Soc. 1 9 7 8 , 100,
62.
9
(
(
28) Cheng, H. Y.; Sackett, P. H.; McCreery, R. L. J. Med. Chem. 1 9 7 8 , 21, 948.
29) Chapter 19. Mayausky, J. S.; Cheng, H. Y.; Sackett, P. H.; McCreery, R. L.
ACS Advances in Chemistry Series 201; American Chemical Society:
Washington, DC, 1982.
(
(
30) Jarbawi, T. B.; Heineman, W. R. Anal. Chim. Acta 1 9 8 2 , 135, 57.
31) Jarbawi, T. B.; Heineman, W. R.; Patriarche, G. Anal. Chim. Acta 1 9 8 1 ,
1
26, 57.
(
32) Bowling, R.; McCreery, R. L. Anal. Chem. 1 9 8 8 , 60, 605.
4084 Analytical Chemistry, Vol. 71, No. 18, September 15, 1999

Fig u re 6 . Voltammograms of CPZ on polished GC and six modified GC surfaces in 40% MeOH, containing 1 M NaCl, without background
subtraction. Solid line in a and c is polished GC before derivatization, remaining curves in a-d are for the seven chemisorbed surfaces listed
in Table 3. The dashed curve in d is the slowest surface observed (nitroazobenzene-modified GC).
Ta b le 2 . k ° o f Ch lo rp ro m a zin e , P ro m a zin e , a n d
Triflu p ro m a zin e fo r Diffe re n t GC S u rfa c e s (0 .2 M Na Cl,
0
.0 1 M HCl, No Me OH)
k° (cm/ s)a
chlorpromazine promazine triflupromazine
conventional polish 0.033 ( 0.003 0.026 ( 0.003 0.023 ( 0.003
b
(
44) (12) (20)
cyclohexane polish 0.034 ( 0.004 0.025 ( 0.003 0.025 ( 0.003
24) (12) (12)
AQDS preadsorbed 0.034 ( 0.002 0.026 ( 0.003 0.023 ( 0.003
24) (12) (12)
0.035 ( 0.002 0.027 ( 0.003 0.025 ( 0.001
24) (12) (12)
0.038 ( 0.002 0.027 ( 0.002 0.023 ( 0.004
(24) (12) (12)
0.020 ( 0.003 0.014 ( 0.003 0.013 ( 0.002
(61) (12) (12)
(
(
BMB preadsorbed
(
Methylene Blue
preadsorbed
nitrophenyl
derivatized
Fig u re 7 . Plot of ln(k°) vs monolayer thickness (d, Å) listed in Table
3. Solid line includes underivatized surface (d ) 1.6 Å) in a linear
least-squares fit, dashed line does not.
a
Obtained from the mean of k° values determined at different scan
b
rates (20 mV/ s to 20 V/ s). Mean ( standard deviation. Parentheses
indicate total number of measurements for four scan rates (0.02, 0.2,
2
, and 20 V/ s).
Ta b le 3 . CP Z Kin e t ic Re s u lt s o n Mo d ifie d S u rfa c e s
∆
EP for CPZ,
monolayer
d, Å
mV, 0.2 V/ s
k°CPZ
ence of anionic, neutral, or cationic adsorbers had little effect on
the observed rates. Only a compact, chemisorbed monolayer of
nitrophenyl groups has a significant kinetic effect, decreasing the
rate constant by about 50%.
The covalently derivatized GC surfaces were examined in more
detail, under conditions where adsorption was suppressed. Vol-
tammetry for CPZ on a series of derivatized surfaces in 40%
methanol is shown in Figure 7. Recall that in these conditions,
the nitrophenyl surface showed no CPZ adsorption, and the CPZ
electron-transfer rate constant was invariant with scan rate
between 0.02 and 20 V/ s. The polished, underivatized surface does
none
phenyl
methylphenyl
trifluoromethylphenyl
nitrophenyl
ethylphenyl
biphenyl
nitroazobenzene
1.6
5.9
6.8
6.8
6.8
8.1
10.2
14.0
63.3 ( 0.6
71.0 ( 1
0.081 ( 0.007
0.029 ( 0.003
0.020 ( 0.001
0.020 ( 0.001
0.020 ( 0.001
0.015 ( 0.001
0.010 ( 0.001
0.005 ( 0.001
b
76.3 ( 1
77.0 ( 1
76.7 ( 1
81.3 ( 1.5
88.7 ( 1.1
98.7 ( 1.1
a
Tunneling distance calculated by Spartan software, structure
optimized with molecular mechanics based on SYBYL force field. Van
der Waals radius yielded d for polished GC of 1.6 Å. b Mean ( standard
deviation for six determinations, all with a scan rate of 0.20 V/ s.
p
show a larger i at 20 V/ s, due to a contribution from adsorbed
CPZ, but this effect is greatly reduced for low scan rate (0.2 V/ s).
The derivatization reagents were chosen to yield monolayers of
increasing thickness, listed in Table 3. The distance between the
GC surfaces and the monolayer/ solution interface (d) was
calculated using molecular mechanics as described in Table 3,
and represents the minimum distance between the surface and a
redox system for a perfect monolayer. The observed rate constants
and calculated d values are listed in Table 3. A plot of ln(k°CPZ) vs
d is shown in Figure 8. The slope of the least-squares line for the
derivatized surfaces is -0.20 Å- (r ) 0.9797) or -0.22 Å (r2
1
2
-1
) 0.9776) if the point for the underivatized surface is included.
Analytical Chemistry, Vol. 71, No. 18, September 15, 1999 4085

DISCUSSION
to adsorption, but the error was minimized by using a low scan
rate. For the polished surface, the addition of methanol increases
the apparent k° from 0.0333 (Table 2) to 0.081 (Table 3). The slope
of the ln k° vs d plot of -0.22 Å- (or -0.20 if polished GC is not
Approximately 50 years of investigations into organic electro-
chemistry has established that electron transfer between solid
electrodes and organic redox systems can involve complex
interactions with the electrode surface. The many examples of
adsorption and electrocatalysis accompanying organic redox
reactions, combined with multiple electron and proton transfers
present formidable hurdles to understanding organic redox
kinetics and mechanisms. For carbon electrodes, one has the
added complications of variable surface condition and a propensity
to adsorb organic molecules from solution. Much of the motivation
for the current effort is the objective of examining organic redox
reactions on reproducibly prepared GC electrodes. Using the
approach developed for inorganic redox systems,19 surface modi-
fication of the GC can provide insight into surface variables
affecting kinetics and adsorption. The redox systems chosen for
this initial study are relatively simple, involving only one electron
and no protons for MV2+/ 1+, CPZ, PMZ, and TPZ, with MB+
following a more complex 2 e / 2 H+ route.
1
included) can be compared to the “tunneling parameter” (â)
discussed in studies of long-range electron transfer.3
3-37
Equation
1
has been invoked to explain the dependence of the observed
rate constant (k°app) for metal electrodes modified with a self-
assembled monolayer of thickness d:
k°app ) k° exp (-âd)
(1)
â depends on the tunneling mechanism and the chemical
-1
structure of the spacer, but values near 1.0 Å have been
proposed for through-bond tunneling in aliphatic chains and near
-1
1
.3-1.8 Å for through-space tunneling on modified gold elec-
trodes.34 â values for through-bond tunneling in unsaturated
spacers in rigid molecules are substantially lower, ranging from
0.14 Å-1 for a conjugated polyene spacer to 0.57 Å-1 for phenyl-
methylene spacers. The closest electrochemical analogy from the
literature examined the distance dependence of electron transfer
for phenylene/ ethynyl spacers adsorbed to gold as thiolate
monolayers.38 For these conjugated spacers, the â value was 0.57
( 0.02 Å-1. The adsorbates used to construct Figure 7 all have
aromatic backbones and are conjugated when more than one ring
-
Other than an apparently minor perturbation from MB+
adsorption, MV2
+/ 1+
and MB are remarkable in their insensitivity
+
to surface modification. A nitrophenyl monolayer decreases the
+
apparent k°s by ∼50% and completely blocks MB adsorption.
+
Clearly, MB adsorption is not required for fast kinetics, and
diffusing MB+ undergoes fast electron transfer even when the
GC surface is occupied by nitrophenyl groups. We concluded
previously, for inorganic redox systems, that the electron tunnels
through the nitrophenyl layer, causing a relatively small decrease
in observed k°.19 Further evidence for the tunneling mechanism
for the case of CPZ is discussed below.
-
1
is present. The observed â value of 0.20 Å is within the range
of 0.14-0.57 Å-1, expected for conjugated, unsaturated spacers.
-
1
Under the assumption of through-bond tunneling, a â of 0.20 Å
is not unreasonable for the series of adsorbates considered here.
However, it is surprising that Figure 7 exhibits the linear ln
k° vs d dependence expected for a tunneling mechanism in the
first place. The chemisorbed monolayers are not expected to self-
assemble on GC, since they are generated rapidly and their
covalent bonds to the surface do not permit annealing. The initial
GC surface is microscopically rough, so an ordered, regular
monolayer is unlikely. However, the monolayers are likely to be
low in pinholes, since any electroactive regions should have been
covered by additional diazonium reduction during derivatization.
The observed rate constants are reproducible to better than (
10% for six determinations, in contrast to the large variation
expected with a variable density of pinholes. Spectroscopic
evidence indicates that diazonium reduction leads to compact
monolayers with the aromatic ring axes perpendicular to the
surface,14 but it is conceivable that the monolayer films are thinner
than predicted geometrically. Such a situation would yield an
erroneously low value of â. The linearity of Figure 7 is consistent
with a compact monolayer which tracks the shape of the rough
GC surface and results in reasonably constant spacing between
the GC substrate and the CPZ at the point of closest approach.
However, the rather small â value of 0.22 Å- is considered
Table 2 indicates the CPZ, PMZ, and TPZ electron-transfer
kinetics are also insensitive to surface modification, even when
adsorption is partly responsible for the observed current. The rate
constants in Table 2 are suspect, due to adsorption, but the k°
p
values and the ∆E results from which they were determined show
no major changes with surface modifications. Adsorption and
electron transfer are uncorrelated and apparently independent, a
conclusion which is inconsistent with an electron-transfer mech-
anism involving electrocatalysis via a chemisorption or mediation
+
+2
mechanism. Like MB and MV , the phenothiazines are behaving
as outer-sphere systems with respect to electron transfer.
The uncertainties about the effects of adsorption on kinetic
observations are greatly reduced for the covalently modified
surfaces in the presence of 40% methanol. The adsorption of CPZ
is presumably driven by dipole-dipole or π-π interactions between
the graphitic surface and the phenothiazine ring system, as well
as hydrophobic effects. The nitrophenyl-modified surfaces show
significantly weaker adsorption than a bare polished surface,
apparently because direct interactions between the carbon surface
and the phenothiazine ring are blocked. There is still some
residual adsorption to the nitrophenyl/ GC surface in the absence
of methanol, presumably because hydrophobic effects are still
present. The combination of nitrophenyl derivatization and 40%
methanol reduces CPZ adsorption below levels detectable by
chronocoulometry (Figure 5) or semi-integration.
1
(
33) Chidsey, C. E. D. Science (Washington, D.C.) 1 9 9 1 , 251, 919.
(34) Finklea, H. O.; Hanshen, D. D. J. Am. Chem. Soc. 1 9 9 2 , 114, 3173.
(
(
(
35) Weber, K.; Creager, S. Anal.Chem. 1 9 9 4 , 66, 3164.
36) Tender, L.; Carter, M.; Murray, R. Anal. Chem. 1 9 9 4 , 66, 3173.
37) Bowler, B. E.; Raphael, A. L.; Gray, H. B. Electron Transfer in Molecules
and Proteins. In Progress in Inorganic Chemistry: Bioinorganic Chemistry;
Lippard, S. J., Ed.; Wiley: N. Y. 1990; Vol. 38, pp 259-322.
38) Sachs, S. B.; Dudek, S. P.; Hsung, R. P.; Sita, L. R.; Smalley, J. F.; Newton,
M. D.; Feldberg, S. W.; Chidsey, C. E. D. J. Am. Chem. Soc. 1 9 9 7 , 119,
10563.
The observed rate constants for CPZ obtained under conditions
of negligible adsorption exhibit linear behavior in a plot of ln k°
vs d (Figure 7). The point for polished GC at 1.6 Å is suspect due
(
4086 Analytical Chemistry, Vol. 71, No. 18, September 15, 1999

semiquantitative at present, given the uncertainties about surface-
roughness effects and monolayer thickness.
Molecular-mechanics calculations43 were used to estimate bond
length changes for Fc/ Fc•+ and CPZ/ CPZ . The maximum
change in bond length calculated for Fc/ Fc•+ was 0.05 Å, while
that for CPZ/ CPZ•+ was 0.04 Å. The small reorganization energy
associated with small changes in bond length is likely to contribute
to the fast kinetics observed for the phenothiazines, compared
•+
There are several possible reasons why the phenothiazine
electron-transfer reactions are fast compared with those of many
organic redox systems, but a strong candidate for the major factor
involves inner-sphere reorganization energy. It has long been
recognized that redox reactions requiring significant changes in
bond distances are often slow, due to the high reorganization
with those observed for many other organic redox systems.
In summary, the phenothiazines and MV2
+/ 1+
behave as outer-
+3/ +2
energy accompanying electron transfer. For example, Fe(H
2
O)
6
sphere redox systems on GC and modified GC, with little kinetic
sensitivity to surface modification. Chemisorbed monolayers
greatly reduce phenothiazine adsorption, but have minor effects
(∼50%) on electron-transfer rates. The dependence of CPZ kinetics
on monolayer thickness is consistent with a through-bond tun-
neling mechanism. More complex organic redox systems believed
to involve chemisorption or proton transfer are currently under
investigation on modified GC electrodes.
has a slow electron-exchange rate39 and a small k° on oxide-free
carbon18 because the Fe-O bond distance must change by 0.13
Å between Fe2 and Fe . Ferrocene (Fc) and anthracene are
much faster, because delocalization of the positive charge upon
oxidation leads to small changes in bond distances. For Fc/ Fc+,
the bond lengths change by at most 0.04 Å upon oxidation.4
The cation in CPZ•+ is quite delocalized, and one would not expect
large bond-length changes during the CPZ/ CPZ•+ redox reaction,
or those of the related phenothiazines, PMZ, TPZ, and MB+.
+
3+
0-42
ACKNOWLEDGMENT
This work was supported by the National Science Foundation,
Division of Analytical and Surface Chemistry.
(
39) Sutin, N. In Inorganic Reactions and Mechanisms; Zackerman, J. J., Ed.;
VCH: Deerfield Beach, FL, 1986; Vol. 15; p 49ff.
(
(
(
40) Seiler, P.; Dunitz, J. D. Acta Crystallogr. 1 9 7 9 , B35, 1068.
41) Lehmann, R. E.; Kochi, J. K. J. Am. Chem. Soc. 1 9 9 1 , 113, 501.
42) Webb, R. J.; Lowery, M. D.; Shiomi, Y.; Sorai, M.; Wittebort, R. J.;
Hendrickson, D. N. Inorg. Chem. 1 9 9 2 , 31, 5211.
43) Spartan; software for ab initio geometry optimization; UHF-3TO-3G basis
set.
Received for review March 2, 1999. Accepted June 28,
1999.
(
AC9902392
Analytical Chemistry, Vol. 71, No. 18, September 15, 1999 4087