Iron(II) catalysis in oxidation of hydrocarbons with ozone in acetonitrile

Article abstract of DOI:10.1021/cs501962m

Oxidation of alcohols, ethers, and sulfoxides by ozone in acetonitrile is catalyzed by submillimolar concentrations of Fe(CH₃CN)₆²⁺. The catalyst provides both rate acceleration and greater selectivity toward the less oxidized products. For example, Fe(CH₃CN)₆²⁺-catalyzed oxidation of benzyl alcohol yields benzaldehyde almost exclusively (>95%), whereas the uncatalyzed reaction generates a 1:1 mixture of benzaldehyde and benzoic acid. Similarly, aliphatic alcohols are oxidized to aldehydes/ketones, cyclobutanol to cyclobutanone, and diethyl ether to a 1:1 mixture of ethanol and acetaldehyde. The kinetics of oxidation of alcohols and diethyl ether are first-order in [Fe(CH₃CN)₆²⁺] and [O₃] and independent of [substrate] at concentrations greater than ～5 mM. In this regime, the rate constant for all of the alcohols is approximately the same, k_cat = (8 ± 1) × 10⁴ M^-1 s^-1, and that for (C₂H₅)₂O is (5 ± 0.5) × 10⁴ M^-1 s^-1. In the absence of substrate, Fe(CH₃CN)₆²⁺ reacts with O₃ with k_Fe = (9.3 ± 0.3) × 10⁴ M^-1 s^-1. The similarity between the rate constants k_Fe and k_cat strongly argues for Fe(CH₃CN)₆²⁺/O₃ reaction as rate-determining in catalytic oxidation. The active oxidant produced in Fe(CH₃CN)₆²⁺/O₃ reaction is suggested to be an Fe(IV) species in analogy with a related intermediate in aqueous solutions. This assignment is supported by the similarity in kinetic isotope effects and relative reactivities of the two species toward substrates.

Full text of DOI:10.1021/cs501962m

Subscriber access provided by KINGSTON UNIV
Article
Iron(II) Catalysis in Oxidation of Hydrocarbons with Ozone in Acetonitrile
Hajem Bataineh, Oleg Pestovsky, and Andreja Bakac
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/cs501962m • Publication Date (Web): 29 Jan 2015
Downloaded from http://pubs.acs.org on February 9, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street
N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 32
ACS Catalysis
1
2
3
4
5
6
Table of Contents Graphic
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O2
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 32
1
2
3
4
5
Iron(II) Catalysis in Oxidation of Hydrocarbons with Ozone in
6
Acetonitrile
7
8
9
Hajem Bataineh, Oleg Pestovsky* and Andreja Bakac*
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Ames Laboratory and Chemistry Department, Iowa State University, Ames, IA 50011
Email: bakac@iastate.edu, pvp@iastate.edu
Abstract Oxidation of alcohols, ethers, and sulfoxides by ozone in acetonitrile is catalyzed by
subꢀmillimolar concentrations of Fe(CH3CN)62+. The catalyst provides both rate acceleration
and greater selectivity toward the less oxidized products. For example, Fe(CH3CN)62+ꢀcatalyzed
oxidation of benzyl alcohol yields benzaldehyde almost exclusively (>95%) whereas uncatalyzed
reaction generates a 1:1 mixture of benzaldehyde and benzoic acid. Similarly, aliphatic alcohols
are oxidized to aldehydes/ketones, cyclobutanol to cyclobutanone, and diethyl ether to a 1:1
mixture of ethanol and acetaldehyde. The kinetics of oxidation of alcohols and diethyl ether are
first order in [Fe(CH3CN)62+] and [O3], and independent of [Substrate] at concentrations greater
than ~5 mM. In this regime, the rate constant for all of the alcohols is approximately the same,
kcat = (8±1) × 104 Mꢀ1 sꢀ1, and that for (C2H5)2O is (5±0.5) × 104 Mꢀ1 sꢀ1. In the absence of
2+
substrate, Fe(CH3CN)6 reacts with O3 with kFe = (9.3±0.3) × 104 Mꢀ1 sꢀ1. The similarity
between the rate constants kFe and kcat strongly argues for Fe(CH3CN)62+/O3 reaction as rate
determining in catalytic oxidation. The active oxidant produced in Fe(CH3CN)62+/O3 reaction is
suggested to be an Fe(IV) species in analogy with a related intermediate in aqueous solutions.
This assignment is supported by the similarity in kinetic isotope effects and relative reactivities
of the two species toward substrates.
1
ACS Paragon Plus Environment

Page 3 of 32
ACS Catalysis
1
2
3
4
Key words
5
6
Iron, ozone, oxidation, catalysis, alcohol, kinetics, mechanism
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Introduction
Previous studies from this group1,2 and others3,4 have established that the reaction of
2+
Fe(H2O)6 with ozone generates an iron(IV) species best described as FeIV(H2O)5O2+ (hereafter
FeaqO2+) on the basis of spectroscopic evidence, conductivity measurements, chemical reactivity
and DFT calculations.1,2,5 Oxidations with this highꢀspin Fe(IV) complex take place by oxygen
atom transfer to e. g. sulfoxides and phosphines, and by hydride and hydrogen atom abstraction
from CꢀH bonds.1 In reactions with alcohols, aldehydes and ethers the latter two mechanisms
operate in parallel, Scheme 1. The hydride path is catalytic as it generates Fe(H2O)62+ which can
be reoxidized to FeaqO2+. Overall, however, the catalytic efficiency is poor because of the loss of
iron as Fe(H2O)63+ in the parallel oneꢀelectron (hydrogen atom transfer) path.
3+
Fe(H2O)6
H+
FeIIIaqOH2+ + CH3C HOH
2+
FeIV
O
+ CH3CH2OH
aq
FeIIaqOH+ + CH3CHO +H+
H+
O3
2+
Fe(H2O)6
Scheme 1
2+
In the reaction between Fe(H2O)6 and H2O2 (Fenton reaction) the reactive intermediate
changes from hydroxyl radicals in acidic solutions to an iron(IV) species at near neutral pH.6
Such a major mechanistic change caused by a modest change in reaction conditions led us to
2
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 32
1
2
3
4
5
6
7
8
9
consider the effect of other parameters, including solvent, on reactions involving solvento iron
species in oxidation states 2+ to 4+. Specifically, the much higher reduction potential of the
Fe(III)/Fe(II) couple in acetonitrile7,8 as compared to that in water suggests that the preference
for twoꢀelectron pathways of a hypothetical iron(IV) species might be greater in acetonitrile. To
explore this possibility and its potential consequences for ironꢀcatalyzed oxidations, we initiated
a study of the reaction of Fe(II) with O3 in acetonitrile in the presence of oxidizable substrates.9
The results are described herein.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Experimental
The following chemicals were obtained commercially and used as received: iron(II)
perchlorate hydrate Fe(ClO4)2.xH2O (98%), deuterium oxide D2O (99.9 atom %D), 1,10ꢀ
phenanthroline (99+ %), benzyl alcohol anhydrous (99.8%), cyclobutanol (99+%) (all Aldrich),
dimethyl sulfoxide (≥ 99.9%, A.C.S spectrophotometric grade) and cyclopentanol (99%) (both
SigmaꢀAldrich), 2ꢀpropanol (99.9% certified ACS), tetrahydrofuran (99.9% HPLC grade), and
ethyl ether anhydrous (99.9% certified ACS) (all Fisher scientific), acetonitrileꢀd3 (99.8 atom %
D) (Cambridge Isotope Laboratories, Inc), acetonitrile (low water content (~10 ppm) for HPLC,
GC and spectrophotometry, Honeywell ꢀ Burdick & Jackson), 2ꢀpropanolꢀ2ꢀd1 (99.8 atom % D)
and 2ꢀpropanolꢀd8 (99.9 atom % D) (both CDN Isotopes), benzylꢀα,αꢀd2 alcohol (98 atom % D,
ISOTEC).
Iron(II) bis(acetonitrile)bis(triflate), Fe(CF3SO3)2(CH3CN)2, was synthesized
according to a literature procedure.10 The solution species in acetonitrile is assumed to be
Fe(CH3CN)62+, and this formula is used throughout the paper although small amounts of mixed
acetonitrileꢀwater complexes cannot be ruled out. This is especially true for experiments with
added water.
3
ACS Paragon Plus Environment

Page 5 of 32
ACS Catalysis
1
2
3
4
5
6
7
8
9
In experiments designed to explore the effect of water on products and kinetics, anhydrous
iron(II) triflate was used instead of hydrated iron(II) perchlorate. Deuterated acetonitrile was
dried over 4A molecular sieves until the HDO/H2O signal disappeared in the 1H NMR spectrum.
The water content at that point is estimated at <40 micromolar, the minimum amount required
for an observable NMR signal under our conditions. More rigorous efforts at achieving strictly
anhydrous conditions were not pursued given that most of the reactions in this work generate
water as one of the products in amounts much greater (up to several millimolar) than those
potentially introduced with our solvents and reagents. Moreover, as shown later, up to 100 mM
of externally added water has no effect on product yields or catalyst recovery, and kinetics are
affected only mildly at > 50 mM water.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
UVꢀVis absorbance measurements and kinetic studies used a Shimadzu UVꢀ3101 PC
1
spectrophotometer and Olis RSMꢀ1000 stoppedꢀflow at 24.9 ± 0.1 °C. H NMR spectra were
recorded with a 400 MHz Bruker DRXꢀ400 or 600 MHz Bruker Avance III spectrometer at room
temperature. Waters GCT accurate mass timeꢀofꢀflight mass spectrometer in positive EI mode
(70 EV) with a scan rate of 0.3 seconds per scan and a mass range of 10–200 Daltons was used
to qualitatively detect some of the products. Waters MassLynx 4.0 software was used to acquire
and process GCꢀMS data. Ozone was generated in an Ozonology Lꢀ100 ozone generator.
Oxygen concentration was measured using Hanna Edge dissolved oxygen meter.
Procedures. Stock solutions of iron(II) perchlorate in CH3CN or CD3CN were prepared fresh
before each set of experiments and standardized with phenanthroline after dilution with H2O and
3+
using ε = 1.14 ×104 Μ−1 cmꢀ1. No Fe(phen)3 was detected in these solutions. Determinations
of Fe(II) concentrations in spent reaction solutions utilized a correction for Fe(III) as previously
described. 1 Ozone solutions were prepared by continuous bubbling of ozone through CH3CN or
4
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 32
1
2
3
4
5
6
7
8
9
CD3CN for >5 min at room temperature and diluted to the desired concentration. The
concentration of ozone in stock solutions was typically 5.6±0.1 mM as determined
spectrophotometrically at 260 nm, ε260 = 3350 Mꢀ1 cmꢀ1. These solutions always contained
residual oxygen, typically about 5.9 mM, as described below.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2+
To determine the amount of oxygen generated in the Fe(CH3CN)6 /O3 reaction in the
presence and absence of substrates, the reactants were mixed rapidly in an airꢀfree, tightly sealed
vial, leaving only minimal head space to avoid equilibration between the solution and gas phases.
A sample (0.5ꢀ1.0 mL) was withdrawn and injected into another sealed vial containing a
dissolved oxygen electrode immersed in 18 mL of airꢀfree water. The measurement was
completed in about 40 seconds after injection. The measured value was corrected for the
concentration of residual oxygen, typically around 5.9 mM in ozone stock solutions as
determined after removal of O3 with excess fumaric or maleic acid.11 The same procedure was
used to determine the concentration of O2 in O2ꢀ saturated acetonitrile. The value obtained, 11.3
mM, is in acceptable agreement with the value reported for airꢀsaturated acetonitrile, 2.42 mM.12
Except in experiments specifically designed to explore the effect of O2 on kinetics and
products, solutions of iron(II) and substrates were prepared and handled anaerobically.
However, since some O2 was present in stock solutions of ozone, see above, and since the
Fe(CH3CN)62+/O3 reaction itself produces O2, none of the reaction solutions were completely O2ꢀ
free.
Competition experiments and product analysis. A solution containing known concentrations of
Fe(II) and two or three substrates were mixed with ozone in a UV cell. After the disappearance
1
of ozone (absorbance at 260 nm), the products were quantified by H NMR. In all of the
experiments the substrate concentrations were sufficiently large to make the kinetics of each
5
ACS Paragon Plus Environment

Page 7 of 32
ACS Catalysis
1
2
3
4
5
6
7
8
9
individual reaction fall into the plateau region, see Results. Product yields for benzyl alcohol,
which absorbs too strongly in the UV for direct kinetic measurements, were shown
independently to remain unchanged at [PhCH2OH]0 ≥4 mM. Similar experiments were
conducted on mixtures of protiated and fully or partially deuterated substrates to determine
kinetic isotope effects. Product yields derived from fully deuterated substrates (diethyl etherꢀd10
and 2ꢀpropanolꢀd8) were estimated as a difference between the total amount of products for the
same competition observed with protiated compounds and the amount of product derived from
the competing protiated substrate.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Catalyst concentrations in product analysis experiments (Tables 1ꢀ7) were chosen so as to
minimize the contribution from uncatalyzed O3/substrate reaction. Typically, >95% of the
reaction proceeded by the catalyzed path except in experiments with 2ꢀpropanol (Tables 1 and 3)
where the uncatalyzed path contributed 16% and 10%, respectively.
Kinetic data were obtained by monitoring the disappearance of ozone at 260 nm (Shimadzu)
or in the 260ꢀ280 nm spectral range (Olis RSMꢀ1000 Rapid Scan). In stopped flow experiments,
2+
a mixture of Fe(CH3CN)6 and substrate was placed in one syringe, and ozone in the other.
Experiments designed to study the effect of [substrate] used 0.06ꢀ0.15 mM ozone, 0.025 mM
Fe(CH3CN)62+ and 1ꢀ50 mM substrate. The effect of [Fe(CH3CN)62+] was explored at 0.08ꢀ0.12
mM ozone, 0.005ꢀ0.1 mM Fe(CH3CN)62+ and 2ꢀ50 mM substrate.
Kinetic traces were fitted to an expression for first order kinetics with Kaleidagraph v4.0 or
with OLIS GlobalꢀWorks v2.0.190. 1H NMR and GCꢀMS analyses were initiated within 5ꢀ15
min after completion of the reaction. 10% D2O (v/v) was added to some NMR solutions to shift
the interfering water peak.
Results
6
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 32
1
2
3
4
5
6
7
8
9
The UV spectrum after completion of the reaction between 1.1 mM benzyl alcohol and 0.22
mM ozone in acetonitrile exhibits a double feature in the 230ꢀ250 nm range, Figure 1, consistent
with a mixture of benzaldehyde (λmax 244 nm) and benzoic acid (λmax 227 nm). The individual
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
spectra are shown in Fig S1. This assignment was confirmed by H NMR, Figure S2. The
reaction with ozone also produced hydrogen peroxide, as shown by 1H NMR signal at 8.56 ppm.
2+
When the same reaction was conducted in the presence of 0.011 mM Fe(CH3CN)6
,
1
benzaldehyde was the major product detected by UV (Fig 1), H NMR (Fig S2), and GCꢀMS.
Small amounts of benzoic acid (~10%) were also observed, some of it possibly generated by
oxidation of benzaldehyde with O2 during sample manipulation. The combined yield of PhCHO
and PhCOOH, based on initial ozone concentration, was 85%.
2
b
1.5
1
a
0.5
0
220
240
260
280
300
Wavelength /nm
Figure 1. UV spectra of the products of the reaction between (a) 1.1 mM PhCH2OH and 0.22
2+
mM O3, and (b) 1.1 mM PhCH2OH and 0.3 mM O3/0.011 mM Fe(CH3CN)6
.
Fe(CH3CN)62+ꢀcatalyzed oxidation of cyclobutanol by O3 (1.8 mM) produced 1.55 mM
1
cyclobutanone, Figure S3. No ringꢀopened products were observed by either H NMR or GCꢀ
MS, ruling out a significant contribution from a path involving cyclobutanol radicals.1 The latter
7
ACS Paragon Plus Environment

Page 9 of 32
ACS Catalysis
1
2
3
4
5
6
are subject to rapid ring opening that ultimately yields an aldehyde(s). At the end of the reaction,
about 80% of iron was still present as Fe(II).
7
8
9
Similarly, ethanol was oxidized to acetaldehyde, Figure S4, 2ꢀpropanol to acetone, and
cyclopentanol to cyclopentanone. The results are summarized in Table 1. Product yields varied
from 70% (acetaldehyde) to 85% (cyclopentanone). 1H NMR of the products of ethanol
oxidation exhibits additional signals at 8.03 and 4.64 ppm, consistent with small amounts of
formic acid and acetal which are common overoxidation products of ethanol.13 The yields of
these products increase somewhat with increasing [O3]/[EtOH] ratio.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. Product Yields in Fe(CH3CN)62+ꢀCatalyzed Oxidation of Alcohols, DMSO and Et2O
by Ozone
Substrate (mM)
dimethyl sulfoxide (9.6)
diethyl ether (8.5)
cyclopentanol (9.8)
cyclobutanol (9.6)
2ꢀpropanol (32)
[O3]/mM
1.9
[Fe(CH3CN)62+]/mM
Major Product
dimethyl sulfone
(ethanol + acetaldehyde)
cyclopentanone
cyclobutanone
acetone
%
Yield
100
0.028
0.052
0.024
0.025
0.025
0.055
0.048
1.2
100
85
85
80
70
70
1.4
1.8
0.1
ethanol (21.1)
0.83
1.8
acetaldehyde
benzyl alcohol (10)
benzaldehyde
The reaction with diethyl ether produced a 1:1 mixture of C2H5OH and CH3CHO in 100%
yield, Figure S5. Dimethyl sulfoxide (DMSO) was oxidized to the sulfone, also in 100% yield,
2+
Figure S6. At an initial Fe(CH3CN)6 concentration of >0.020 mM, the majority (70ꢀ90%) of
iron was still present as Fe(II) at the end of the reactions listed in Table 1 provided the
concentration of substrate exceeded ~5 mM. At much lower initial concentrations of
Fe(CH3CN)62+ and substrate, up to 40ꢀ60 % of Fe(CH3CN)62+ was oxidized to Fe(III).
8
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 32
1
2
Fe(CH3CN)62+ꢀcatalyzed oxidation of THF by O3 yielded several products as shown by GCꢀ
3
4
5
6
1
MS and H NMR, Figures 2 and S7. On the basis of mass spectra, the GC peak at 4.42 min is
7
8
9
assigned to an equilibrated mixture14 of 2ꢀhydroxytetrahydrofuran (2ꢀOHꢀTHF) and 4ꢀ
hydroxybutanal, and that at 6.18 min to γꢀbutyrolactone. All three species were clearly identified
and quantified by 1H NMR, Figure S7. The combined yield is 75% (in 4:2:1 ratio, respectively).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
Also observed in H NMR are small amounts of formic acid which was also found in previous
studies of THF oxidation.15 Several other small NMR peaks and the GC peak at 4.95 min in
Figure 2 were not identified. The corresponding GC peak was also observed and remained
unidentified in an earlier study of THF oxidation.15 The products eluting at 9ꢀ10 min in Figure 2
are attributed to THF dimers and condensation products as deduced from mass spectral data.
These products are also formed upon electrochemical oxidation of THF in aqueous sulfuric
acid.15
Figure 2. Gas chromatogram of products obtained by oxidation of 5.6 mM THF by 1.35 mM
O3/0.1 mM Fe(II).
9
ACS Paragon Plus Environment

Page 11 of 32
ACS Catalysis
1
2
3
4
5
6
7
8
9
The overall picture of THF oxidation changes dramatically when an alcohol is added as
coꢀsubstrate. As shown on the example of THF/ benzyl alcohol mixture, Figures S8ꢀS10, the
main products (>90%) are the acetal 2ꢀORꢀTHF and aldehyde/ketone derived from the alcohol.
THF oxidation products, i. e. hydroxytetrahydrofuran/4ꢀhydroxybutanal and γꢀbutyrolactone,
accounted for only 5% of products.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In search of the source of 2ꢀORꢀTHF we considered the known reaction16 between alcohols
and 4ꢀhydroxybutanal, the latter being one of THF oxidation products. This reaction generates
2ꢀORꢀTHF in the presence of an acid catalyst at 20ꢀ100o C,16 but is extremely slow (about 17
hours) under our experimental conditions. Also, no new products were generated upon mixing
2+
alcohols with product solutions of Fe(CH3CN)6 /O3/THF reaction. A slow overnight reaction
2+
between alcohols and THF in the presence of Fe(CH3CN)6 (0.5 mM) did produce 2ꢀORꢀTHF
when the concentrations of alcohols (60 mM) and THF (80 mM) were about tenꢀfold higher than
is typical in our work. Clearly, the rapid (several seconds) formation of 2ꢀORꢀTHF under our
standard catalytic conditions utilizes a different path and must involve an intermediate(s)
generated in the course of the Fe(CH3CN)62+/O3 oxidation of THF and/or alcohol. Since close to
2+
100% of iron was still present as Fe(II) at the end of the Fe(CH3CN)6 /O3/ THF/alcohol, the
products were either formed in a series of 2ꢀe steps or Fe(III), if involved, was reꢀreduced to
Fe(II) by reaction intermediate(s).
Kinetics. Substrates (5ꢀ50 mM) were used in large excess over ozone (0.06ꢀ0.15 mM) and
Fe(CH3CN)62+. The loss of ozone was monitored at 260 nm. Kinetic traces in the plateau
region, see below, were exponential and yielded pseudoꢀfirstꢀorder rate constants kobs.
Ozone oxidation of organic substrates used in this work is slow but not negligible in
comparison with the Fe(CH3CN)62+ꢀcatalyzed reaction. The rate law for the disappearance of
10
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 32
1
2
3
4
5
6
7
8
9
ozone is thus given by eq 1, where kcat represents the rate constant for the catalytic reaction of eq
2, kO3 is the independentlyꢀdetermined rate constant for the direct O3/substrate reaction, Table
S1, and S is substrate. The contribution from direct reaction to kobs was typically <10%, but
increased as substrate concentrations increased and Fe(II) concentrations decreased. In the least
favorable case (30 mM 2ꢀPrOH at 0.025 mM Fe(II)), this contribution was 16%. The use of
higher concentrations of the catalyst, which would benefit catalytic reaction, was not feasible
because the kinetics became too fast and signalꢀtoꢀnoise ratio poor.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢀd[O3]/dt = kO3[O3][S] + kcat[O3] [Fe(CH3CN)62+]m [S]n = kobs [O3]
(1)
[Fe(II)
O3 + S → Products
kcat
(2)
The experimentally determined kobs was corrected for the direct path to give kcorr, eq 3.
kcorr = kobs ꢀ kO3[S] = kcat [Fe(CH3CN)62+]m [S]n
(3)
The reaction is first order in [Fe(CH3CN)62+] (m = 1) as shown by linear dependence of kcorr
on [Fe(CH3CN)62+] at two different concentrations of 2ꢀPrOH in Figure 3. First order
dependence on [Fe(CH3CN)62+] holds for all of the substrates examined. Identical results were
2+
obtained with Fe(ClO4)2.xH2O and Fe(CF3SO3)2(CH3CN)2 as the source of Fe(CH3CN)6
.
8
6
4
2
0
0
40
80
120
[Fe(II)] /ꢀM
11
ACS Paragon Plus Environment

Page 13 of 32
ACS Catalysis
1
2
2+
3
4
5
6
7
8
9
Figure 3. Plot of kcorr vs concentration of Fe(CH3CN)6 for the catalytic oxidation of 2ꢀPrOH
with ozone (0.08ꢀ0.12 mM.). Concentrations of 2ꢀPrOH are 2 mM (circles) and 50 mM
(squares).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The dependence on [2ꢀPrOH], on the other hand, is quite modest as shown by the small
difference in slopes of the two lines in Figure 3, i. e. 3.6 × 104 Mꢀ1sꢀ1 and 7.7× 104 at [2ꢀPrOH] =
2 mM and 50 mM, respectively. This general picture also holds for other substrates as shown in
Table S2 and illustrated by the plot of kcorr vs [Substrate] in Figure 4.
2.0
1.5
1.0
0.5
0.0
0
20
40
60
[Substrate] /mM
Figure 4. Plot of kcorr against substrate concentration for Fe(CH3CN)62+ꢀcatalyzed oxidations
with ozone of 2ꢀpropanol (circles), ethanol (halfꢀfilled squares), THF (squares), cyclobutanol
(triangles) and diethyl ether (diamonds). All experiments have [Fe(CH3CN)62+]0 = 0.025 mM,
[O3] = 0.06ꢀ0.15 mM.
After the sharp initial rise, the rate constants in Figure 4 reach an approximately constant
value of 1.5 ± 0.2 sꢀ1 for most substrates, and 1.1 sꢀ1 for diethyl ether. The initial concentration
2+
of Fe(CH3CN)6 in these experiments was approximately constant at 0.025 ± 0.002 mM. After
2+
the reaction, ~80% of Fe(CH3CN)6 was recovered in the plateau region in Figure 4, but only
12
ACS Paragon Plus Environment

ACS Catalysis
Page 14 of 32
1
2
3
4
5
6
7
8
9
about 50% in the rising portion at low substrate concentrations. Also, the fit to exponential
kinetics at low [substrate] was poor, and only initial 50% of the reaction was used to evaluate the
rate constants.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2+
In the plateau region the reaction is clearly catalytic in Fe(CH3CN)6 and the rate law is
reasonably well approximated by eq 4 (i. e. n of eq 3 is zero), yielding kcat
=
kcorr/[Fe(CH3CN)62+] = (5±0.5) × 104 Mꢀ1 sꢀ1 for Et2O and (8±1) × 104 Mꢀ1 sꢀ1 for the remaining
substrates.
ꢀd[O3]/dt = d[Product]/dt = kcat[Fe(CH3CN)62+][O3] = kcorr [O3]
(4)
The results for 2ꢀpropanol deviate somewhat from this picture in that the rate constant continues
to increase slowly with increasing [2ꢀpropanol]. Of all the aliphatic alcohols studied, 2ꢀpropanol
is the most reactive in the direct reaction with O3, Table S1, so that the correction for this term
becomes significant at higher concentrations of 2ꢀPrOH as already commented. Under these
conditions the two pathways may not remain completely independent, i. e. intermediates from
one may cross over to the other and lead to the observed increase in rate constant. The
proportion of the direct pathway can be minimized experimentally by increasing catalyst
concentration and enhancing the catalytic path. Such conditions were used in product analysis,
but for kinetic studies this option is not feasible as the increased rate of catalytic component
made the overall reaction too fast to monitor.
Kinetic behavior of DMSO is qualitatively similar to that of alcohols and ethers, but the rate
constant is much larger, reaching an (extrapolated) saturation value of 39 ± 1 sꢀ1 at 0.025 mM
2+
Fe(II), Figure 5. This result implies much greater reactivity of Fe(DMSO)n(CH3CN)6ꢀn
complex(es)17,18 compared to Fe(CH3CN)62+. The support for Fe(DMSO)n(CH3CN)6ꢀn in this
2+
1
work comes from the observation of broadened H NMR methyl resonances of DMSO in
13
ACS Paragon Plus Environment

Page 15 of 32
ACS Catalysis
1
2
3
4
5
6
7
8
9
CD3CN in the presence of Fe(II), consistent with an exchange between free and complexed
DMSO. As expected, the signal sharpens upon addition of D2O which leads to dissociation of
DMSO.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
30
20
10
0
0.000
0.005
0.010
0.015
[DMSO] / M
2+
Figure 5. Plot of kobs vs [DMSO] for the reaction with O3 (0.1 mM) / Fe(CH3CN)6 (0.025
mM).
2+
Kinetic measurements for the Fe(CH3CN)6 /O3 reaction in the absence of substrates and
with Fe(CH3CN)62+ in large excess yielded kobs = 20 sꢀ1 at [Fe(CH3CN)62+]0 = 0.21 mM, and 29 sꢀ
at [Fe(CH3CN)6 ]0 = 0.30 mM, which results in kFe = (9.3±0.3) ×104 Mꢀ1 sꢀ1, eq 5. The
1
2+
similarity between the rate constants kFe and kcat strongly argues that they apply to the same
reaction, i. e. formation of an intermediate, presumably FeIV(CH3CN)5O2+ (hereafter FeIV O2+),
AN
see later, or a related species in analogy with FeIV
O
that is produced in Fe(H2O)62+/O3
2+
aq
reaction in acidic aqueous solutions.1 In this scenario, FeIV O2+ rapidly oxidizes substrates as in
AN
eq 6, thereby regenerating Fe(CH3CN)62+ which reꢀenters eq 5.
Fe(CH3CN)62+ + O3 → FeIV(CH3CN)5O2+ + O2 + CH3CN
kFe
(5)
CH CN, fast
FeIV(CH3CN)5O2+ + (C2H5)2O → Fe(CH3CN)62+ + {C2H5OH + CH3CHO} (6)
3
14
ACS Paragon Plus Environment

ACS Catalysis
Page 16 of 32
1
2
3
4
5
6
7
8
9
Runs with excess ozone exhibited a rapid initial step followed by slower disappearance of
large, nonstoichiometric amounts of ozone in a reaction apparently catalyzed by iron. The fast
initial step took place on a time scale appropriate for kFe that was determined with excess
Fe(CH3CN)62+, but a reliable rate constant could not be extracted under those conditions.
In several experiments the concentration of O2 was determined at the end of reaction with use
of dissolved oxygen electrode as described in Experimental. Under standard catalytic conditions
(0.050 mM Fe(CH3CN)62+, 0.8 mM O3, 20 mM substrate), the reactions with DMSO and with
(C2H5)2O generated 0.9 equivalents of O2 per O3, Table S3. This result, combined with
quantitative product yields in Table 1, leads to the approximate stoichiometry in eq 7.
Fe(II)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O3 + (C2H5)2O (or Me2SO) → {CH3CHO + C2H5OH} (or Me2SO2) + O2
(7)
For the remaining substrates in Table 1, the net increase in O2 content was lower, typically 0.6
equivalents per O3, suggesting some O2 consumption in parallel 1ꢀe processes, see later. When
no substrates were added, the increase in O2 was only about 0.2 equivalents per mole of O3
2+
regardless of whether Fe(CH3CN)6 was used in catalytic amounts or in concentrations
comparable to those of O3 (~ 0.8 mM). In both cases ozone was consumed completely, although
at low iron concentrations the reaction took about 15 minutes, much longer than in the presence
of added substrates. Given that O3 persists in acetonitrile for hours in the absence of
Fe(CH3CN)62+, it is clear that Fe(CH3CN)62+/acetonitrile combination leads to catalytic O3
consumption, although acetonitrile is less reactive than the substrates in Table 1. Moreover,
small amounts of Fe(CH3CN)62+ remained after completion of the reaction even when ozone was
in excess. In experiments with equimolar amounts of Fe(CH3CN)62+ and O3, about 60% of Fe(II)
remained after completion of the reaction in CH3CN, but only traces (<5%) in CD3CN
demonstrating a large solvent kie. Unfortunately, no oxidation products of CH3CN could be
15
ACS Paragon Plus Environment

Page 17 of 32
ACS Catalysis
1
2
1
3
4
detected by H NMR or GCꢀMS owing to interference by the large solvent peaks. No
5
6
formaldehyde was detected with chromotropic acid.
7
8
9
Effect of Fe(III), O2 and water. There is a mild increase in product yields under oxygenꢀrich
conditions, as shown for ethanol in Table 2. At approximately constant concentrations of
Fe(CH3CN)62+, O3, and EtOH, an increase in oxygen concentration from 1 mM to 7 mM led to
an increase in acetaldehyde yield from 82% to 94%, and an increase in the recovery of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2+
Fe(CH3CN)6 from 91% to 98%. Oxygen also appears to have a mild inhibiting effect on the
kinetics. The rate constant in the presence of excess O2 (≥1.3 mM) is about 15% smaller than
that obtained in the experiments that had only a small background concentration of O2 (ca 0.2
mM, comparable to that of ozone).
Externally added Fe(ClO4)3 also improves product yields. As shown in the last entry in
Table 2, the yields of acetaldehyde become quantitative in the presence of 0.24 mM Fe(III).
Replacing the Fe(II) catalyst with Fe(III) results in a slow initial decrease in ozone
concentration, but the reaction accelerates with time suggesting a buildup of Fe(II) and onset of
Fe(II) catalysis. Experiments with Fe(III)/EtOH/acetonitrile confirmed that Fe(II) was indeed
produced.
Up to 100 mM of added water has no effect on product yields or catalyst recovery as shown
for ethanol and 2ꢀpropanol in Table 3, but the rate constant shows a small systematic increase
with increasing [H2O]. At larger concentrations of H2O, product yields and catalyst recovery
both decrease and the rate constant increases. All catalytic activity ceases when water content
reaches 3% (~1.5 M). The presence of water in the coordination sphere of iron and in the solvent
apparently changes the Fe(III)/Fe(II) potentials to an extent sufficient to restore the chemistry to
that characteristic of aqueous solution.1
16
ACS Paragon Plus Environment

ACS Catalysis
Page 18 of 32
1
2
Table 2. Effect of Fe(III) and O2 on Ethanol Oxidationa
3
4
5
6
7
8
9
b
c
[O2]/ mM
[Fe(III)]0
% [CH3CHO]∞
% [Fe(II)]∞
/mM
1
82
82
91
93
1d
7
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
94
98
1
0.052
0.24
85
105
130
1
100
[Fe(CH3CN)62+]0 = 0.047ꢀ0.060 mM, [O3] = 0.93ꢀ1.0 mM, [C2H5OH] = 43ꢀ55 mM. Percent
a
b
yield of CH3CHO. Percent Fe(II) recovered after reaction. d Added [H2O] = 56 mM.
c
Table 3. Effect of H2O on the Kinetics and Catalyst Recoverya
Added [H2O]/
kcorr/sꢀ1
1.7
% [Fe(II)]∞
96
b
[O3]/ mM
0.054
Substrate
mM
2ꢀpropanol
0
0.064
0.056
0.050
0.072
0.063
0.060
2ꢀpropanol
2ꢀpropanol
2ꢀpropanol
ethanol
50
99
2.0
2.6
3.0
1.4
2.3
4.4
92
96
80
92
76
64
198
0
ethanol
149
489
ethanol
a
b
Conditions: [Substrate] = 20 mM, [Fe(CH3CN)62+]0 = 0.025 mM. Percentage of Fe(II)
recovered after reaction
Competition Experiments. Since the Fe(CH3CN)62+/O3 reaction is rate determining, direct kinetic
measurements do not provide information on the reactivity of catalytic intermediate(s). To
obtain some insight into relative reactivity of FeIV
O
AN
toward various substrates and to
2+
determine kinetic isotope effects, competition experiments were performed, see Experimental.
17
ACS Paragon Plus Environment

Page 19 of 32
ACS Catalysis
1
2
3
4
5
6
7
8
9
The results (Figures S11ꢀS16) are summarized in Tables 4ꢀ7. The ratios of rate constants k1/k2
for various substrates were calculated from the expression k1/k2 = [P1][S2]/[P2][S1], where S1 and
S2 are two competing substrates, and P1 and P2 their respective products. In the experiment with
three competing substrates the listed ratios are [P1][S2]/[P2][S1] and [P2][S3]/[P3][S2], where S3
and P3 stand for (CH3)2CHOH and (CH3)2CO, respectively. No corrections were applied for
different numbers of abstractable hydrogens in different substrates.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 4. Results of Competition Experimentsa
b
O3 /mM
Substrate (mM)
benzyl alcohol (4.9) +
ethanol (19)
Product (mM)
benzaldehyde (0.27) +
acetaldehyde (0.25)
benzaldehyde (0.29) +
cyclobutanone (0.31)
cyclobutanone (0.62) +
acetaldehyde (0.76)
cyclobutanone (0.40) +
acetone (0.34)
k1/k2
4.2
0.66
0.73
1.7
benzyl alcohol (4.9) +
cyclobutanol (10)
cyclobutanol (10) +
ethanol (30)
1.9
2.4
0.86
1.0
cyclobutanol (10) +
2ꢀpropanol (11)
1.3
benzyl alcohol (4.7) +
ethanol (10) +
benzaldehyde (0.32) +
acetaldehyde (0.18) +
acetone (0.32)
3.8, 0.54c
2ꢀpropanol (9.6)
1.1
benzyl alcohol (4.8) +
diethyl ether (4.7)
benzaldehyde (0.49) +
acetaldehyde / ethanol (0.37)
1.3
a[Fe(CH3CN)6] = 0.05 – 0.06 mM. b Ratio of rate constants for competing substrates S1 and S2 in
c
the order listed in each set. Ratio of rate constants for ethanol and 2ꢀpropanol.
18
ACS Paragon Plus Environment

ACS Catalysis
Page 20 of 32
1
2
3
4
5
6
7
8
9
All of the rate constants were normalized to kEtOH = 1.0 in Table 5. Similar experiments with
deuterated substrates (Figures S17ꢀS21) yielded the results in Table 6 from which kinetic isotope
effects in Table 7 were calculated.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2+
Table 5. Relative Rate Constants for Oxidations with FeIV
O
AN
Substrate
Average krel
kH2O/Mꢀ1sꢀ1a
ethanol
[1.0]
2.0
2.51 × 103
3.22 × 103
3.13 × 103
4.74 × 103
14.2 × 103
2ꢀpropanol
cyclobutanol
diethyl ether
benzyl alcohol
2.2
3.1
4.0
a Directly measured rate constants for reactions of Fe(H2O)5O2+ in 0.1 M aqueous HClO4
Table 6. Products Obtained in Competition Between Protiated and Deuterated Substratesa
O3 /mM Substrate (mM)
Product (mM)
1.15
diethyl etherꢀd10 (5.6)
acetaldehyde / ethanol (0.20)b
benzaldehyde (0.66)
acetaldehyde (0.53)
benzaldehyde (0.18)
acetone (0.19)
benzyl alcohol (4.8)
ethanol (16.6)
1.1
benzyl alcoholꢀd2 (4.7)
2ꢀpropanolꢀd1 (10.1)
benzyl alcohol (4.8)
cyclobutanol (9.7)
benzyl alcoholꢀd2 (9.1)
0.85
1.39
benzaldehyde (0.46)
cyclobutanone (0.81)
benzaldehyde (0.34)
19
ACS Paragon Plus Environment

Page 21 of 32
ACS Catalysis
1
2
3
4
1
2ꢀpropanolꢀd8 (10.3)
acetone (0.22)b
5
6
benzyl alcohol (4.8)
benzaldehyde (0.54)
7
8
9
b
a By 1H NMR. Estimated from experimentally determined amount of PhCHO and assuming a
75% cumulative yield of all products (as found with protiated substrates).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2+
Table 7. Kinetic Isotope Effects for Reactions of FeIV
O
AN
Substrate
kH/kD
diethyl ether (d10)
benzyl alcohol (d2)
2ꢀpropanol (d1, d8)
2.3
3.8
2.5
Discussion
The oxidation of alcohols, ethers and sulfoxides by ozone in acetonitrile is catalyzed by
Fe(CH3CN)62+. Concentrations of Fe(CH3CN)6 as low as 0.02 mM are sufficient for the
2+
catalytic reaction to dominate over the uncatalyzed O3/substrate reaction at substrate
concentrations lower than about 50 mM. The catalyst not only provides rate acceleration, but
also increases the selectivity toward the less oxidized product. This is illustrated in Figure 1 and
Figure S2 on the example of benzyl alcohol which is oxidized to benzaldehyde in the
Fe(CH3CN)62+ꢀcatalyzed path, and to a 1:1 mixture of benzaldehyde and benzoic acid in direct
oxidation with ozone.
Saturation kinetics are observed at [substrate] >5 mM (Figure 4). The rate constants reach an
approximate limit of kcat = (8±1) × 104 Mꢀ1 sꢀ1 for all of the substrates except diethyl ether which
reacts somewhat more slowly, kcat = (5±0.5) × 104 Mꢀ1 sꢀ1. The observed variations in kcat can be
rationalized by variations in Fe(II)ꢀsubstrate binding constants and perhaps different
20
ACS Paragon Plus Environment

ACS Catalysis
Page 22 of 32
1
2
3
4
5
6
7
8
9
contributions from 1ꢀe and 2ꢀe paths, see later. The role of substrate binding is clearly seen in
the reaction with DMSO which interacts strongly with Fe(CH3CN)62+ and reaches kcat = 1.5 × 106
Mꢀ1 sꢀ1, Figure 5.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
As the concentration of substrate drops below 5 mM, the rate constant decreases sharply,
Figure 4. In this regime up to 50% of iron is oxidized to Fe(III) in the course of the reaction
2+
resulting in slower kinetics. Side reactions of Fe(IV) with Fe(CH3CN)6 and with the solvent,
see later, are also most severe at low substrate concentrations, which further reduces the
efficiency of the catalytic reaction. The remainder of the discussion will focus on the saturation
regime.
Most efficient are the oxidations of diethyl ether and dimethyl sulfoxide. Both generate 2ꢀ
electron oxidation products quantitatively, Figures S5 and S6, with only small losses (≤20%) of
the catalyst over 20ꢀ70 catalytic cycles, Table 1. These data are most easily explained by a
singleꢀstep twoꢀelectron oxidation of substrates by FeIV O2+, eq 6, followed by regeneration of
AN
FeIV O2+ in reaction 5.
AN
Product yields are somewhat lower, 70ꢀ85%, in the reactions with alcohols, Table 1. We
attribute these results to a contribution from a oneꢀelectron path of eq 7 which leads to oxygen
•ꢀ
•ꢀ
radicals (HO•/O•ꢀ, O3 , O2 and others) known to be involved in chain decomposition of O3 in
aqueous solutions.19ꢀ24 Some of the key reactions believed responsible for the loss of O3 in this
work are shown in eq 8ꢀ18, written in analogy with the chemistry in aqueous solutions and in the
gas phase and supported by limited information on the reactivity of ozone and oxygen radicals in
nonꢀaqueous solvents.25ꢀ30 The reaction of Fe(III) with hydroxyalkyl radicals was not considered
in light of the much larger concentrations of the more reactive O2 and O3. Also, there is no
21
ACS Paragon Plus Environment

Page 23 of 32
ACS Catalysis
1
2
2+
3
4
reaction between Fe(CH3CN)6 with O2 or with traces of H2O2 on the time scale of our
5
6
experiments.
7
8
9
FeIV O2+ + RCH2OH → Fe(III) + RC•HOH
(7)
(8)
(9)
AN
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RC•HOH + O2 → RCH(OH)OO•
•
•ꢀ
RCH(OH)OO• → RCHO + HO2 /O2
Fe(IV)
2RCH(OH)OO• →→ {2 RCHO + O2 + H2O2} → HO2 + Fe(III) (10)
•
RC•HOH + O3 → → RC(O)OH + HO2 /O2
(11)
(12)
•
•ꢀ
•
•ꢀ
→
HO2
O2 + H+
←
•ꢀ
O2 + Fe(III) → O2 + Fe(II)
(13)
(14)
•ꢀ
•ꢀ
O2 + O3 → O2 + O3
•ꢀ
•ꢀ
→
O3
O + O
(15)
(16)
←
2
O•ꢀ + H+
HO•
→
←
•ꢀ
(HO•, O3 , O•ꢀ) + CH3CN (+ O2, O3) → Products
(17)
(18)
(HO•, O3 , O•ꢀ) + RCH2OH → RCH•OH + (O2, H2O, HOꢀ)
•ꢀ
Hydroxyalkyl radicals generated in eq 7 react with both O2 (eq 8) and O3 (eq11) and produce
superoxide, a powerful reductant and nucleophile.30 The reduction of Fe(III) by O2 , eq 13, is
•ꢀ 24
•ꢀ 31
the key step that regenerates Fe(II). The competing reaction between O3 and O2 , the latter a
wellꢀrecognized chain carrier in the decomposition of ozone,23,32 generates O3 followed by
•ꢀ
dissociation33 to give O•ꢀ, eq 14ꢀ15. The latter may be protonated (pKa of HO• in H2O = 11.9)34 if
sufficient amount of water is present in the solvent, but protonation is not required for the next
22
ACS Paragon Plus Environment

ACS Catalysis
Page 24 of 32
1
2
step since both HO• and O•ꢀ will oxidize the solvent and/or substrate by hydrogen atom
abstraction,33 eq 17 and 18. Even though the rate constant for the reaction of HO• with
acetonitrile is smaller (k = 1.0 × 106 Mꢀ1 sꢀ1 in acetonitrile)35 than the rate constants for the
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
reactions with alcohols (e. g. kEtOH = 8.3 × 107 Mꢀ1 sꢀ1),26 the concentration advantage makes the
reaction with CH3CN about 5ꢀ10 fold faster at 20ꢀ50 mM ethanol that is typical in this work.
Presumably, other radicals in eq 17ꢀ18 exhibit similar reactivity pattern and together with HO•
lead to a loss of oxidizing equivalents and less than quantitative yields of substrateꢀderived
products. Reaction 18 regenerates RC•HOH which reenters the scheme.
According to the above mechanism, the beneficial effect of added Fe(III) arises mainly from
its efficient scavenging of O2•ꢀ in eq 13.24 This step both regenerates the catalyst and minimizes
the importance of reactions 14ꢀ17 which lead to the loss of O3.
Increased product yields and somewhat slower kinetics of O3 loss under O2ꢀrich conditions
are also consistent with known reactivity of radicals with O3 and O2. At high [O2], most of the
radicals react with O2 as in eq 8, followed by reactions 9ꢀ10 and 12ꢀ18. In the absence of
externally added O2, the concentrations of O2 and O3 are comparable (see Experimental), and
reaction 11 becomes competitive with reaction 8 which increases the rate of ozone consumption
and yields of doubly oxidized products.36 Moreover, alkylperoxyl radicals produced in eq 8 also
react with O3 to generate alkoxyl radicals RCH(OH)O•, eq 19, followed by rearrangement and/or
further reactions with O2, O3 and substrates.37,38
−O
O , O , substrate
2
RCH(OH)OO• + O3 → RCH(OH)O• 3→ products
(19)
2
In agreement with the above scheme, the concentration of O2 found after completion of the
reactions with alcohols is significantly smaller than one would calculate by adding the amount
produced from ozone in reaction 5 to the [O2] initially present. Clearly, some O2 is consumed in
23
ACS Paragon Plus Environment

Page 25 of 32
ACS Catalysis
1
2
3
4
5
6
7
8
9
the course of alcohol oxidation. On the other hand, the concentration of O2 found after the
oxidation of DMSO and (C2H5)2O is close to that calculated, supporting the notion that 1ꢀe
oxidation of these substrates is negligible. DMSO is probably oxidized by OAT, similar to the
reaction in water.1 Quantitative product yields and measurable hydrogen kie for diethyl ether
suggests hydride transfer.1 Additional support for hydride transfer from ethers, and by inference
from alcohols as well, comes from the effect of alcohols on oxidation products of THF. The
formation of acetals is most reasonably explained by hydride transfer that generates an oxonium
ion followed by the reaction with alcohols as shown in Scheme 2. The oxonium ion shown was
proposed earlier as an intermediate in cationic polymerization of THF in the presence of
triphenylmethyl cation salts.39
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Mechanism for THF oxidation by hydride transfer to Fe(CH3CN)5O2+
The alternative rebound mechanism that begins with hydrogen atom transfer would generate
the hemiacetal as shown in Scheme 3. This mechanism is ruled out by our observation that the
hemiacetal does not react with alcohols under our conditions. In contrast to ethers, the
intermediates generated in the process of alcohol oxidation by hydride transfer generate
aldehydes/ketones by rapid deprotonation at oxygen.
24
ACS Paragon Plus Environment

ACS Catalysis
Page 26 of 32
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 3. Mechanism for THF oxidation by hydrogen atom transfer to Fe(CH3CN)5O2+.
The induction period observed in the O3/substrate reaction when iron is initially present as
Fe(III) is most easily explained by the need to reduce Fe(III) to Fe(II), presumably through a
scheme involving oneꢀelectron reduction of O3 by substrate19,40,41 followed by eq 15ꢀ18. The
•ꢀ
complex, multistep chemistry in eq 17ꢀ18 is envisioned to generate some O2 which reduces
Fe(III) to Fe(II) and thus leads to the production of FeIV O2+ via reaction 5. Another possibility
AN
is a slow, direct oxidation of Fe(III) by ozone to generate high oxidation state iron species42 that
would be rapidly reduced to Fe(II).
2+
The disappearance of ozone in the presence of catalytic amounts of Fe(CH3CN)6 in
acetonitrile even in the absence of more reducing substrates shows that the solvent itself can be
2+
catalytically oxidized. It is not clear whether FeIV
O
AN
oxidizes CH3CN in 1ꢀe or 2ꢀe steps.
2+
Measurable amounts of Fe(CH3CN)6 found in such solutions after all of O3 disappeared
support a 2ꢀe catalytic reaction that might take place by oxygen atom transfer or hydride transfer.
On the other hand, as shown above in the reaction with alcohols, 1ꢀe chemistry can also bring
about the disappearance of ozone and formation of Fe(CH3CN)62+. In support of oneꢀelectron
2+
route we note that aqueous FeIV
O
reacts with CH3CN by hydrogen abstraction and does not
aq
regenerate Feaq2+.1 Also, the much smaller amount of recovered Fe(II) after completion of
25
ACS Paragon Plus Environment

Page 27 of 32
ACS Catalysis
1
2
3
4
5
6
Fe(II)/O3 reaction in CD3CN indicates a large solvent isotope effect, again consistent with HAT,
although hydride abstraction cannot be entirely ruled out.
7
8
9
Throughout this discussion it has been assumed that the oxidizing intermediate is an
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Fe(IV) species, FeIV O2+, although so far we have not been able to observe or characterize it
AN
spectroscopically. This assignment is made in analogy to the results in aqueous solution, where
the product of Fe(H2O)62+/O3 reaction has been identified as FeaqO2+,1,2 and at interfaces with
ꢀ
chlorideꢀcontaining solutions where O=FeIVCl3 is generated.43 The relative reactivity toward
substrates in Table 5 appears consistent with this assignment in that the trend in acetonitrile
follows closely that observed for FeaqO2+ in aqueous solutions. In that work it was possible to
carry out direct kinetic measurements of substrate oxidation by preꢀformed FeaqO2+. As shown
in Table 5, benzyl alcohol is the most reactive among alcohols in both solvents, but the yield of
2ꢀe oxidation product in acetonitrile is among the lowest. This result may suggest a significant
contribution from the 1ꢀe path. Alternatively, the reaction may involve an attack by FeIV O2+ at
AN
the aromatic ring to generate multiple products, similar to the reaction of O3 with PhCH2OH,44 or
reactions of other Fe(IV)ꢀoxo complexes with aromatic compounds.45,46
The kinetic isotope effect for the reaction with 2ꢀpropanol is also similar in the two
solvents. The value of kH/kD for the methine CꢀH is 2.5 in acetonitrile (Table 7) and 2.1 in H2O,1
consistent with hydride transfer proposed previously. These results however do not rigorously
rule out other potential oxidizing intermediates, such as an ozonide or Fe(III)ꢀ(CH2CN) radical
that may also react by hydride or hydrogen atom transfer.
Conclusions
Perchlorate and trifluoromethane sulfonate salts of iron(II) efficiently catalyze the oxidation of
alcohols and ethers with ozone in acetonitrile. This result stands in stark contrast with that
26
ACS Paragon Plus Environment

ACS Catalysis
Page 28 of 32
1
2
2+
3
4
obtained in acidic aqueous solutions where, under comparable conditions, all of Fe(H2O)6 is
5
6
3+
quickly oxidized to the unreactive Fe(H2O)6
.
7
8
9
The difference between the two solvents can be rationalized by changes in redox
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
•
•ꢀ
thermodynamics of iron and acidꢀbase chemistry of superoxide, HO2 /O2 , as follows. In both
solvents the reaction between the substrate and active oxidant, an iron(IV) species, takes place in
parallel oneꢀelectron (hydrogenꢀatom abstraction) and twoꢀelectron (hydride transfer) paths. The
twoꢀelectron path is much more prominent in acetonitrile, presumably because it avoids the
strongly oxidizing Fe(III) (E = 1.6 V vs. NHE).8 This path regenerates the active catalyst,
Fe(CH3CN)62+, directly. The minor parallel hydrogen atom transfer path produces carbon
radicals and Fe(III) followed by radical/O2 reaction that generates superoxide (E0 (O2/O2 ) = ꢀ
•ꢀ
0.80 V vs. NHE)30. The rapid reaction of O2•ꢀ with Fe(III) regenerates the catalyst and removes
•ꢀ
O2 , the key intermediate involved in chain decomposition of ozone.
The two paths are of comparable importance in acidic aqueous solutions1 so that a substantial
2+
3+
portion of Fe(H2O)6 is oxidized to Fe(H2O)6 in a single cycle. Similar to the reaction in
acetonitrile, the radical/O2 chemistry generates superoxide, but this intermediate is rapidly
protonated under the acidic conditions employed (pKa (HO2 / O2 ) = 4.69).30 The protonation
•
•ꢀ
prevents the superoxide from reducing Fe(H2O)63+ and regenerating the catalyst.
Supporting Information Available: Figures S1ꢀS21 and Tables S1ꢀS3. This information is
available free of charge via the Internet at http://pubs.acs.org/.
Acknowledgment. We are grateful to Dr. Jana for help with the synthesis of iron(II)
bis(acetonitrile) complex. This research is supported by the U.S. Department of Energy, Office
27
ACS Paragon Plus Environment

Page 29 of 32
ACS Catalysis
1
2
3
4
5
6
7
8
9
of Science, Basic Energy Sciences, Division of Chemical Sciences, Geosciences, and
Biosciences through the Ames Laboratory. The Ames Laboratory is operated for the U.S.
Department of Energy by Iowa State University under Contract DEꢀAC02ꢀ07CH11358.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
References
(1) Pestovsky, O.; Bakac, A. J. Am. Chem. Soc. 2004, 126, 13757ꢀ13764.
(2) Pestovsky, O.; Stoian, S.; Bominaar, E. L.; Shan, X.; Münck, E.; Que, L. J.; Bakac, A.
Angew. Chem., Int. Ed. 2005, 44, 6871ꢀ6874.
(3) Jacobsen, F.; Holcman, J.; Sehested, K. Int. J. Chem. Kin. 1998, 30, 215ꢀ221.
(4) Logager, T.; Holcman, J.; Sehested, K.; Pedersen, T. Inorg. Chem. 1992, 31, 3523ꢀ3529.
(5) Pestovsky, O.; Bakac, A. Inorg. Chem. 2006, 45, 814ꢀ820.
(6) Bataineh, H.; Pestovsky, O.; Bakac, A. Chem. Sci. 2012, 3, 1594ꢀ1599.
(7) Kratochvil, B.; Long, R. Anal. Chem. 1970, 42, 43ꢀ46.
(8) Sugimoto, H.; Sawyer, D. T. J. Am. Chem. Soc. 1985, 107, 5712ꢀ5716.
(9) Bakac, A.; Pestovsky, O.; U.S. Patent 8507730 (2013).
(10) Hagen, K. S. Inorg. Chem. 2000, 39, 5867ꢀ5869.
(11) Leitzke, A.; von Sonntag, C. Ozone: Sci. Eng. 2009, 31, 301ꢀ308.
(12) Franco, C.; Olmsted J., III, Talanta 1990, 37, 905ꢀ909.
(13) Nimlos, M. R.; Wolfrum, E. J.; Brewer, M. L.; Fennell, J. A.; Bintner, G. Env. Sci. Tech.
1996, 30, 3102ꢀ3110.
(14) Hay, M. T.; Geib, S. J.; Pettner, D. A. Polyhedron 2009, 28, 2183ꢀ2186.
(15) Avgousti, C.; Georgolios, N.; Kyriacou, G.; Ritzoulis, G. Electrochim. Acta 1999, 44, 3295ꢀ
3301.
(16) Klang, J. A.; Lawson, A. P.: United States, 1993; Vol. US005254702A, p 1ꢀ4.
28
ACS Paragon Plus Environment