Homogeneous catalysis under ultradilute conditions: TAML/NaClO oxidation of persistent metaldehyde

Article abstract of DOI:10.1021/jacs.6b11145

TAML activators enable homogeneous oxidation catalysis where the catalyst and substrate (S) are ultradilute (pM-low μM) and the oxidant is very dilute (high nM-low mM). Water contamination by exceptionally persistent micropollutants (MPs), including metaldehyde (Met), provides an ideal space for determining the characteristics and utilitarian limits of this ultradilute catalysis. The low MP concentrations decrease throughout catalysis with S oxidation (k_II) and catalyst inactivation (k_i) competing for the active catalyst. The percentage of substrate converted (%Cvn) can be increased by discovering methods to increase k_II/k_i. Here we show that NaClO extends catalyst lifetime to increase the Met turnover number (TON) 3-fold compared with H₂O₂, highlighting the importance of oxidant choice as a design tool in TAML systems. Met oxidation studies (pH 7, D₂O, 0.01 M phosphate, 25 °C) monitored by 'H NMR spectroscopy show benign acetic acid as the only significant product. Analysis of TAML/NaClO treated Met solutions employing successive identical catalyst doses revealed that the processes can be modeled by the recently published relationship between the initial and final [S] (S₀ and S_∞, respectively), the initial [catalyst] (Fe_Tot) and k_II/k_i. Consequently, this study establishes that ΔS is proportional to S₀ and that the %Cvn is conserved across all catalyst doses in multicatalyst-dose processes because the rate of the k_II process depends on [S] while that of the k_i process does not. A general tool for determining the Fe_Tot required to effect a desired %Cvn is presented. Examination of the dependence of TON on k_II/k_i and Fe_Tot at a fixed S₀ indicates that for any TAML process employing Fe_Tot < 1 X 10_-6 M, small catalyst doses are not more efficient than one large dose.

Full text of DOI:10.1021/jacs.6b11145

Article
Journal of
Homogeneous
the American
Chemical
Catalysis
Society is
published by
Under
the American
Chemical
Subscriber access
Society. 1155
provided by NEW YORK
Sixteenth
UNIV
Street N.W.,

Ultra-Dilute
CondiJtoiuornanl osf :
TAML/tNChheeaAmerican
m
CicallO
Society is
published by
Oxidation of
the American
Chemical
Subscriber access
Society. 1155
provided by NEW YORK
Sixteenth
UNIV
Street N.W.,

Persistent
Metaldehyde
Journal of
the American
Liang LC.hTemaincagl ,
Society is
Matthew A.
published by
the American
DeNardo,
Chemical
Subscriber access
ChristSoopcihetey.r1155
provided by NEW YORK
Sixteenth
UNIV
Street N.W.,

J. Schuler,
Matthew R. Mills,
Journal of
Chakicherla
the American
Chemical
Gayathri, Roberto
Society is
R. Gil, Rakesh
published by
the American
Chemical
Subscriber access
Society. 1155
provided by NEW YORK
Sixteenth
UNIV
Street N.W.,

Kanda, and
Terrence J. Collins
Journal of
J. Am. Chem. Soc.,
the American
Chemical
Just Accepted
Society is
Manuscript • DOI:
published by
10.1021/jacs.6b11145
the American
Chemical
Subscriber access
Society. 1155
provided by NEW YORK
Sixteenth
UNIV
Street N.W.,

• Publication Date
(Web): 06 Dec 2016
Journal of
Downloaded from
the American
http://pubCsh.aemcsic.aol rg
on December 6, 2016
Society is
published by
the American
Chemical
Subscriber access
Society. 1155
provided by NEW YORK
Sixteenth
UNIV
Street N.W.,

Just Accepted
Journal of
“Just Accepted” manuscripts
the American
online prior to technical editi
Chemical
Society is
Society provides “Just Acc
published by
dissemination of scientific m
the American
Chemical
appear in full in PDF format a
Subscriber access
Society. 1155
provided by NEW YORK
Sixteenth
UNIV
Street N.W.,

fully peer reviewed, but shou
readers and citable by the D
Journal of
to authors. Therefore, the “J
the American
in the journal. After a manus
Chemical
Society is
Accepted” Web site and pub
published by
changes to the manuscript
the American
Chemical
Subscriber access
Society. 1155
provided by NEW YORK
Sixteenth
UNIV
Street N.W.,

and ethical guidelines that
or consequences arising fro
Journal of
the American
Chemical
Society is
published by
the American
Chemical
Subscriber access
Society. 1155
provided by NEW YORK
Sixteenth
UNIV
Street N.W.,

urnPaalgoef t1hoefA2m7erican Chemical Soci
1
2
3
4
ACS Paragon Plus Environment
5
6
7

Journal of the American Chemical Society
Page 2 of 27
1
2
3
4
5
6
7
8
9
Homogeneous Catalysis Under Ultra-Dilute Conditions:
TAML/NaClO Oxidation of Persistent Metaldehyde
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
Liang L. Tang,† Matthew A. DeNardo,† Christopher J. Schuler,† Matthew R. Mills,† Chakicherla Gaꢀ
yathri,† Roberto R. Gil,† Rakesh Kanda,‡ and Terrence J. Collins*,†
†Institute for Green Science, Department of Chemistry, Carnegie Mellon University, 4400 Fifth Avenue, Pittsburgh, Pennꢀ
sylvania 15213, United States
‡Institute for the Environment, Brunel University, Halsbury Building (130), Kingston Lane, Uxbridge, Middlesex, UB8 3PH,
United Kingdom
ABSTRACT: TAML activators enable homogenous oxidation catalysis where the catalyst and substrate (S) are ultraꢀdilute (pM–
low ꢀM) and the oxidant is very dilute (high nM–low mM). Water contamination by exceptionally persistent micropollutants
(MPs), including metaldehyde (Met), provides an ideal space for determining the characteristics and utilitarian limits of this ultraꢀ
dilute catalysis. The low MP concentrations decrease throughout catalysis with S oxidation (kII) and catalyst inactivation (ki) comꢀ
peting for the active catalyst. The percentage of substrate converted (%Cvn) can be increased by discovering methods to increase
kII/ki. Here we show that NaClO extends catalyst lifetime to increase the Met turnover number (TON) threefold compared with
H2O2, highlighting the importance of oxidant choice as a design tool in TAML systems. Met oxidation studies (pH 7, D2O, 0.01 M
1
phosphate, 25 °C) monitored by H NMR spectroscopy show benign acetic acid as the only significant product. Analysis of
TAML/NaClO treated Met solutions employing successive identical catalyst doses revealed that the processes can be modeled by
the recently published relationship between the initial and final [S] (S0 and S∞, respectively), the initial [catalyst] (FeTot) and kII/ki.
Consequently, this study establishes that ꢁS is proportional to S0 and that the %Cvn is conserved across all catalyst doses in multiꢀ
catalystꢀdose processes because the rate of the kII process depends on [S] while that of the ki process does not. A general tool for
determining the FeTot required to effect a desired %Cvn is presented. Examination of the dependence of TON on kII/ki and FeTot at a
fixed S0 indicates that for any TAML process employing FeTot < 1 ꢀ 10ꢀ6 M, small catalyst doses are not more efficient than one
large dose.
Introduction
Chart 1. Structures of the M I diastereomer of metaldeꢀ
hyde and TAMLa compounds 1.
The development of homogeneous oxidation catalysis that is
effective under ultraꢀdilute conditions promises to enable a
new branch of chemistry with implications for sustainability.
For example, micropollutants (MPs) produce adverse effects
in aquatic organisms at environmentally relevant low parts per
billion (ppb) to low parts per trillion (ppt) concentrations.1ꢀ2
Currently, Switzerland is installing endꢀofꢀplant ozone and
activated carbon (AC) technologies for reducing MP concenꢀ
trations in municipal wastewater, but both are expensive.3
TAML catalyzed peroxide oxidations of MPs function so well
under ultraꢀdilute conditions (low ppb, low nM) that novel and
cheaper processes for the removal of MPs are now within
reach.4ꢀ12 TAML/H2O2 has been shown to remove low ppb–
low ppt (low nM–low pM) concentrations of MPs of highest
concern from London municipal wastewater with comparable
technical performance to ozone and projected superior cost
performance.13 Certain MPs are not effectively removed from
water by ozone or AC.14 For example, metaldehyde (Met,
Chart 1) is so persistent that very expensive advanced oxidaꢀ
tion processes (e.g., UV/H2O2) need to be applied when drinkꢀ
ing water is unacceptably contaminated.15ꢀ17 A discussion of
the challenges Met poses to water treatment and available
removal methods has been published14 and a more thorough
list of the available Met removal methods can be found in the
aTAML is a registered trademark covering tetraꢀamidoꢀN macroꢀ
cyclic ligand complexes.18
Supplementary Information of this work (Tables S1–3). Thereꢀ
fore, we have are exploring TAML catalysis of Met oxidation.
Met oxidation exacts the most demanding performance of
TAML/H2O2 systems studied to date.14 We have shown that
1a/H2O2 (Chart 1) is capable of oxidizing Met. However, in
contrast with most MPs, both the 1a/H2O2 rates of Met oxidaꢀ
tion and the balance between these and the rates of catalyst
inactivation were too unfavorable to promise a realꢀworld soꢀ
lution. Therefore, we have identified Met as a boundary case,
one for which effective oxidation will be either just possible or
just not possible. By working to improve the technical perforꢀ
mance in this extreme case, we have encountered a significant
ACS Paragon Plus Environment

Page 3 of 27
Journal of the American Chemical Society
1cmꢀ1),28 respectively. The specified [NaClO] values are those
advantage resulting from changing the oxidant from H2O2 to
1
2
3
4
5
6
7
hypochlorite. Hypochlorite is widely employed in US water
disinfection processes1 where it is generated in situ from chloꢀ
rine gas at a low cost.19 We have long known that TAML actiꢀ
vators catalyze hypochlorite reactions.20 NaClO activates
TAML catalysts to give FeV=O complexes from ꢀ40 °C21 to
room temperature,22 highlighting a similarity with the most
reactive intermediates in TAML/H2O2 systems.23
of the added reagent—at pH 7 NaClO is ca 70% HOCl.
Instrumental. IR measurements were performed on a Mattson
ATI Affinity 60 AR FTIR spectrometer using KBr pellet. UVꢀ
Vis measurements were performed on an Agilent 8453 UVꢀ
Vis spectrophotometer equipped with an 8ꢀcell transporter and
thermostatic temperature controller. Solution temperatures
were maintained at 25 °C in capped quartz cuvettes (1.0 cm).
TAML catalysis of metaldehyde oxidation was followed by 1H
8
9
Here we examine the 1a (0.4–6.4 ꢀM) catalyzed degradation
of Met (0.33 mM) by NaClO (0.008–0.02 M) at 25 °C and pH
7 (0.01 M phosphate) in D2O. The shift from 1a/H2O2 to
1a/NaClO results in a threefold improvement in process techꢀ
nical performance, establishing that oxidant choice is an imꢀ
portant process design parameter for these systems. Effective
oxidation of the very resistant Met required the addition of
multiple catalyst doses over which a consistent relationship
was found between [Met]0 (initial [Met]) and the ꢁ[Met]
which held for all 18 sequential catalyst treatments. These
results have served as an entry point for augmenting the mathꢀ
ematical logic which models TAML catalysis and for developꢀ
ing a 3ꢀD map that will typically allow prediction of the perꢀ
cent conversion (%Cvn) of any concentration of a substrate by
any TAML catalyst under broadly specified conditions. The
map holds considerable promise as a tool for TAMLꢀbased
water purification. It predicts the amount of catalyst needed to
achieve a targeted substrate removal and identifies boundary
pollutants that are exceptionally difficult to treat with the curꢀ
rent set of TAML catalysts. Finally, we show that when
[TAML] < 1 ꢀ 10ꢀ6 M, the efficiency of oxidation of easily
oxidized pollutants is inversely related to [TAML] because the
substrate undergoes dilution, while in the transformation of
boundary pollutants TON is independent of [TAML].
1
NMR. 1D H spectra were recorded at 300 K on a Bruker
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
Avance III 500 NMR spectrometer operating at 500.13 MHz.
The water signal was suppressed using the presaturation exꢀ
periment (zgpr) from the Bruker pulse programs library.
Chemical shifts are reported in parts per million relative to
TMSP (an internal standard for aqueous solutions, see Abbreꢀ
viations). However, since TAML catalyzed oxidation of an
internal standard could interfere with measurements, none was
added.14 Each sample was scanned 128 times over 16.5
minutes. The Bruker TopSpin 3.0 software was used to proꢀ
cess the NMR data. The Met (CH3CHO)4 and acetic acid
CH3COOH absolute integrals were used for quantification.
Each data point is the average of three measurements and the
associated error is the standard deviation of the set.
Results and discussion
The choice of oxidant is a process design tool for advancing
TAML catalyst lifetimes. To assess the relative performances
1
of 1a/NaClO and 1a/H2O2, H NMR measurements were perꢀ
formed on Met solutions in D2O under identical conditions.
As shown in Fig. 1A, the [Met] is unaffected by either NaClO
or H2O2 alone. The NaClO kinetic trace shows that a single 1a
dose giving a [1a] of 3.98 × 10ꢀ7 M effects a 14.4% conversion
(14.4 %Cvn) as indicated by a reduction in the absolute inteꢀ
gral of the 1.34 ppm Met CH3 signal over the approximate
duration of catalysis of 80 hours (Table 1). For comparison, in
the corresponding 1a/H2O2 experiment, a 4.3 %Cvn is obꢀ
served over the approximate duration of catalysis, which this
time lasted only 10 hours. In both cases, the relationship beꢀ
tween [Met] and time is initially linear allowing calculation of
the initial oxidation rates (v = d[S]/dt) by fitting a linear form.
While the initial rates of Met consumption are identical, a
turnover number (TON) of 106 is observed in the NaClO sysꢀ
tem compared with 32 in the H2O2 system, a threeꢀfold imꢀ
provement. These reaction features are also visible in the kiꢀ
netic traces for acetic acid production (Fig. 1B), with ca. 5
times more acetic acid being produced by 1a/NaClO. Under
these conditions 1a/H2O2 generates an acetic acꢀ
id:acetaldehyde product ratio of 3:1. The respective acetic acid
and acetaldehyde rat LD50 values are 3,310 and 661 mg kgꢀ1
suggesting that selectivity for acetic acid is desirable.29 The
near exclusive production of acetic acid by 1a/NaClO leads to
an even more benign product mixture. Because chlorination of
water containing organic matter generates hazardous disinfecꢀ
tion byproducts (DBPs), including carcinogenic chloroform,30ꢀ
Materials and methods
Materials. All reagents, components of buffered solutions, and
solvents were of at least ACS reagent grade and were used as
received. Met (Acros, 99%) was recrystallized in ethanol24 and
stored at 4 °C. Since three isomers of Met, each having a disꢀ
tinct infrared (IR) spectrum are known, the IR spectrum of the
recrystallized sample was recorded (Fig. S1). The spectrum is
nearly identical to that previously assigned25 to the metaldeꢀ
hyde I (M I) diastereomer, the most abundant form of Met, the
structure of which was previously determined by Xꢀray crysꢀ
tallography (Chart 1).26 Met stock solutions (0.3 mM) were
prepared by sonicating the appropriate amount of Met in buffꢀ
ered D2O (99.9%, Cambridge Isotope Laboratories, Inc.) at
room temperature for 3 h. Buffers of the desired pHs (6.2ꢀ7.0)
were prepared with 0.01 M phosphate and monitored by an
Accumet AB15 pH meter at room temperature. A solution of
DCl in D2O was added to adjust the reaction mixture pH to 7
after NaClO addition. The reported values are the uncorrected
pH meter readings. TAML activator 1a was provided by
GreenOx Catalysts Inc. Stock solutions of 1a (2 × 10ꢀ4 M)
were prepared in D2O and stored at 4 °C. H2O2 and NaClO
solutions were standardized daily by measuring the absorbꢀ
ances at 230 nm (ε = 72.4 Mꢀ1cmꢀ1)27 and 293 nm (ε = 350 Mꢀ
31
we have monitored for chloroform production—none was
detected within the limits of the NMR technique.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 27
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
Figure 1. Oxidation of Met (2.95 × 10ꢀ4 M) by NaClO (3.8 × 10ꢀ3 M, 13 eq) or H2O2 (3.61 × 10ꢀ3 M, 12 eq) at pH 7. Kinetic traces of (A):
metaldehyde consumption and (B): acetic acid formation. Circles: metaldehyde plus oxidant control experiments (● NaClO and ○ H2O2).
Triangles: 1a (3.98 × 10ꢀ7 M) catalyzed oxidations (▲ NaClO and △ H2O2). Conditions: pH 7 D2O (0.01 M, phosphate), reactions were
allowed to proceed at room temperature. Concentrations were calculated from the absolute integrals of the metaldehyde and acetic acid
CH3 signals at 1.34 and 1.92 ppm, respectively. The initial 1H NMR integrals for the trace amounts of acetic acid that formed on sonication
of metaldehyde were subtracted from those of the total acetic acid.
Table 1. Comparison of 1a/NaClO and 1a/H2O2 systems in catalysis of metaldehyde degradation at pH 7. Conditions: [Met]
= 2.95 × 10ꢀ4 M, [1a] = 3.98 × 10ꢀ7 M, [NaClO] = 3.8 × 10ꢀ3 M, [H2O2] = 3.6 × 10ꢀ3 M.
Oxidant
NaClO
v × 108 / M minꢀ1a Removal / %
Major Products
acetic acid
TON
Functioning Time
80 h
2.4 ± 0.4
14.4 ± 0.9
106 ± 6
acetaldehyde and
acetic acid
H2O2
2.2 ± 0.3
4.3 ± 0.7
32 ± 5
10 h
aThe rate (v) is calculated from the slope of the line of best fit to the first three [Met] measurements (v = d[Met]/dt).
The identical initial rates of Met oxidation observed for
NaClO and H2O2 are informative. TAML catalysts function
via the stoichiometric mechanism of Scheme 1. The resting
catalysts (FeIII) undergo activation by an oxidant (Ox, kI) to
form active catalysts (Ac) which then oxidize a substrate (S) to
give FeIII and product(s) (kII) or undergo inactivation (ki or
k2i)—the reverse of catalyst activation, kꢀI, is kinetically negliꢀ
gible.
Substrate oxidation and catalyst inactivation compete for Ac.
Recently, we have developed eq 3,32 a mathematical form that
models the final result of this competition, S∞ (S0 and S∞ are
the initial and final [S], respectively). Equation 3 applies only
when there is an excess of the primary oxidant and substrate
consumption is incomplete (S∞>0), i.e. conditions are set such
that the catalyst is the limiting species and is completely inacꢀ
tivated before all of the substrate has been oxidized. These
conditions were employed in the Met oxidation experiments
described herein.
Scheme 1. General mechanism of TAML catalysis.
ꢉ
ꢃ
ꢑ
ln ꢉ
ꢂ
ꢄꢄ Feꢌꢍꢎ
(3)
ꢃ
ꢒ
ꢓ
For most TAML processes, kI[Ox]<kII[S] and catalyst activaꢀ
tion is rate determining. However, Met oxidation by Ac is
extremely slow. Of tested TAML/H2O2 processes,14 the 1a
system is known to deliver the highest measurable v and
%Cvn. The eq 3 estimated kII value for 1a/H2O2 is 120 ± 30 Mꢀ
Equation 1 models the initial rate of substrate oxidation (FeTot
is the total concentration of catalyst).35 For particularly diffiꢀ
cult to oxidize substrates, catalyst activation outpaces substrate
oxidation (kI[Ox]>kII[S]) and eq 1 simplifies to eq 2.
1
sꢀ1, ca. 340 times lower than the corresponding value for Orꢀ
ange II of 41,000 ± 1,000 Mꢀ1 sꢀ1, noting of course that the
Orange II kII and ki were measured in H2O while the Met data
were recorded in D2O.33 Given this value, the 1a/H2O2 (D2O)
Met process exhibits the lowest kII (and %Cvn, see later) at pH
7 of any MP oxidation studied to date. The eq 3 estimated
1a/H2O2 (D2O) Met kII is comparable to the 1a/H2O2 (D2O) kI
of 180 ± 8 Mꢀ1 sꢀ1. Since kI ~ kII for this Met system, and
[H2O2] > 10 × [Met], the oxidation of this very difficult subꢀ
ꢅ
ꢈꢅ ꢈ
ꢉ
ꢃ ꢃ ꢆꢇ
ꢄ
ꢄꢄ
ꢁ ꢂ
ꢅ ꢈ Feꢌꢍꢎ
(1)
(2)
ꢅ
ꢈ
ꢃ
ꢋꢃ ꢆꢇ ꢋꢃ
ꢉ
ꢄꢄ
ꢊꢄ
ꢄ
ꢅ ꢈ
ꢁ ꢂ ꢏꢐꢐ S Feꢌꢍꢎ
ACS Paragon Plus Environment

Page 5 of 27
Journal of the American Chemical Society
strate by Ac is rateꢀdetermining and eq 2 applies. This is furꢀ
when all of the catalyst is inactivated. Catalysis resumes upon
introduction of a fresh TAML dose.14, 32ꢀ33, 36 In the 1a/NaClO
experiment with [1a] = 1.66 × 10ꢀ6 M, Met consumption did
not resume after introducing an additional dose of 1a or of
NaClO (3.51 × 10ꢀ4 M) at 50 h (Fig. S4), giving evidence for
both irreversible catalyst inactivation and complete oxidant
consumption. This behavior is unusual and is being examined
further.
1
2
3
4
5
6
7
8
ther supported by the similarity of the rates for both oxidants
which also provides strong evidence for rate determining subꢀ
strate oxidation by a common Ac. Noting again that acetaldeꢀ
hyde is observed in the 1a/H2O2 (D2O) study, it follows that
the known oxidation of acetaldehyde by NaClO34 probably
accounts for the virtual absence of acetaldehyde in the
1a/NaClO product mixture. Since, as indicated by eq 3,
lnS0/S∞ is fixed by kIIFeIIITot/ki, and both the NaClO and H2O2
processes share a common rate determining step with a comꢀ
mon Ac, the NaClO performance advantage must derive from
a decrease in ki which is reflected in the greater reduction in
[Met] and longer operating time. The 1a/NaClO ki value
matches that of 1b/H2O2.33 Since the 1a Ac oxidizes substrates
ca. one order of magnitude faster than 1b, this amounts to a
kII/ki gain of ca. one order of magnitude on going from the
1b/H2O2 to 1a/NaClO systems. Considering that kII and ki have
been found to track linearly for all TAML catalysts, this gain
in operational stability with preservation of oxidative activity
is remarkable.33
In both the H2O2 and NaClO systems, catalyst lifetime and
TON diminished significantly on increasing [1a] from 1.99–
3.98 × 10ꢀ7 M to 1.66 × 10ꢀ6 M. This suggested that higher
order catalyst degradation processes in [1a] became kinetically
relevant. We have previously considered the role of inactivaꢀ
tion pathways that are bimolecular in catalyst, labeled k2i proꢀ
cesses (Scheme 1), in TAML catalysis, and estimated that
contributions from such pathways are negligible at catalyst
concentrations < 1 × 10ꢀ6 M.32ꢀ33, 36ꢀ37 These experimental reꢀ
sults support the prior estimates. The detection of these higher
order processes here suggests a k2i pathway in both systems.
However, increasing [1a] impacted the H2O2 and NaClO proꢀ
cesses differently. In the NaClO system, increasing [1a] from
3.98 × 10ꢀ7 M to 1.66 × 10ꢀ6 M gave a 6.6% increase in the
percent of Met oxidized (Table 1). In the H2O2 system, this
same [1a] increase reduced Met oxidation from 4.3% to effecꢀ
tively zero within experimental error (Figs. 2 and S5). We
have interpreted this loss of catalysis in the following way.
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
The inactivation of TAML catalysts has been found to follow
both intramolecular suicidal35ꢀ37 and intermolecular H2O2ꢀ
dependent pathways36, 38 which are unimolecular in catalyst.
To minimize the performance reducing impacts of the latter in
ultraꢀdilute catalysis, a low [H2O2] is generally employed. In
one case, a system in which H2O2 is generated enzymatically
in situ has been devised.39 By using NaClO, we have eliminatꢀ
ed the H2O2ꢀdependent catalyst inactivation pathways altoꢀ
gether. The TON increases by 70% on changing from H2O2 to
NaClO. The recently published 1a/H2O2 ki33 of (1.1 ± 0.3) ×
10ꢀ3 sꢀ1 is ca 70% greater than the eq 3 estimated NaClO ki
value of (3.0 ± 0.8) × 10ꢀ4 sꢀ1. These comparisons indicate that
in D2O, at least 70% of the ki processes are attributable to
H2O2 dependent catalyst inactivation. The lifetime extension
observed for the 1a/NaClO catalysis further establishes the
importance of understanding the nature of the H2O2 dependent
inactivation pathway in TAML/H2O2 catalysis.
Effects of different catalyst concentrations. The identical iniꢀ
tial rates of 1a/H2O2 and 1a/NaClO Met consumption are slow
(Table 1). The effect of [1a] on the oxidation was examined
with the aim of increasing both v and the %Cvn. Since the
H2O2 %Cvn with [1a] = 3.98 × 10ꢀ7 M was small, only higher
[1a] experiments were performed in the H2O2 system (Fig. 1).
Surprisingly, the anticipated increases in v and %Cvn with
increasing [1a] to 1.66 × 10ꢀ6 M were not observed. Instead,
the kinetic trace is largely indistinguishable from that of the
control experiment (Figs. 2 and S5).
Kinetic traces of the oxidation of Met (2.95 × 10ꢀ4 M) by
NaClO (3.8 × 10ꢀ3 M) were recorded for three different [1a]
values: 1.99 × 10ꢀ7 M, 3.98 × 10ꢀ7 M, and 1.66 × 10ꢀ6 M (Fig.
2). As shown in Table 2 and Fig. S2, the initial rates of Met
consumption increased linearly with [1a] but the TON did not
remain constant. At [1a] of 1.99 × 10ꢀ7 M and 3.98 × 10ꢀ7 M,
TONs of ~100 are observed. However, at [1a] of 1.66 × 10ꢀ6
M, the TON decreased by ~60%—the corresponding %Cvn
values are 7.8, 14.4 and 21 %. In each case, approximately
half of the theoretical amount of acetic acid, the only observaꢀ
ble major product, is produced (kinetic traces of acetic acid
generation are shown in Fig. S3). At the lower two [1a] values
the NaClO catalysis is observed to function for about 90 h, but
at the highest [1a] the catalysis lasts only 30 h. Typically,
TAML catalysis is conducted with excess oxidant and stops
Figure 2. Kinetic traces of 1a catalyzed metaldehyde oxidation.
[1a] values are shown on the figure next to the corresponding
kinetic traces. Hollow hexagons: [H2O2] = 3.61 × 10ꢀ3 M (12.2
equivalents); Black symbols: [NaClO] = 3.8 × 10ꢀ3 M (12.8
equivalents). Other conditions: pH 7 D2O (0.01 M, phosphate),
[Met]0 = 2.95 × 10ꢀ4 M.
If for the H2O2 case with [1a] = 1.66 × 10ꢀ6 M, eq 3 is applied
using the estimated kII of 120 ± 30 Mꢀ1sꢀ1 and setting an apꢀ
proximation for the oxidation percentage of 2% (Fig. S5; to
avoid a zero denominator), the estimated total ki value is ca.
1.0 × 10ꢀ2 sꢀ1. This is an order of magnitude greater than the ki
for 1a/H2O2 measured at [1a] ≈ 10ꢀ8 M where inactivation is
exclusively unimolecular in catalyst.33 These results suggest
the operation of inactivation processes that are bimolecular in
catalyst at [1a] ≥ ca. 1 × 10ꢀ6 M which greatly outpace those
that are unimolecular in catalyst. Equation 3 is very sensitive
to the estimated oxidation percentage. If values <2% are choꢀ
sen instead, even higher ki values result. Application of eq 3 to
the analogous NaClO process with [1a] = 1.66 × 10ꢀ6 M gives
a ki of 8.5 × 10ꢀ4 sꢀ1. This is at least one order of magnitude
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 27
Table 2. Summary of metaldehyde degradation at different [1a] values; [Met] = 2.95 × 10ꢀ4 M, [NaClO] ≈ 3.8 × 10ꢀ3 M.
1
2
3
4
5
6
7
8
[1a] × 10ꢀ7 M v × 108 / M minꢀ1a
TON
116 ± 6
106 ± 6
37 ± 5
%Cvn
7.8 ± 0.4
14.4 ± 0.9
21 ± 3
Acetic acid formed / %b
4.1 ± 0.1
t∞c / h
90
1.99
3.98
16.6
1.2 ± 0.2
2.0 ± 0.2
5.4 ± 0.1
9.0 ± 0.7
90
10 ± 1
30
b
aThe rate v was calculated from the slope of the line of best fit to the first five [Met] measurements (v = d[Met]/dt); The percentage of
acetic acid (AA) formation was calculated from the CH3 absolute integral relative to the initial metaldehyde signal (AA% = (Abꢀ
sInt1.92)/(AbsInt1.34)0 × 100); cReaction time.
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
less than the corresponding H2O2 value and is comparable to
the ki value observed for the lower [1a] H2O2 systems. Thus,
one can deduce that at high [1a], the use of NaClO results in
productive catalysis by avoiding k2i processes observed in the
H2O2 system. However, the eq 3 calculated ki for the [1a] =
1.66 × 10ꢀ6 M NaClO system is ca. threefold greater than that
of the lower [1a] NaClO experiments, indicating that k2i proꢀ
cesses operate in the NaClO system. These results highlight
the importance of developing kinetic approaches for determinꢀ
ing the mechanism(s) and rate constant(s) of the k2i proꢀ
cess(es). Such studies are ongoing, and the understanding
gained through them is expected to aide in the design of supeꢀ
rior catalysts for high [TAML] processes.
a TON of 66. This is significantly less than the Experiment 1
dose 1 TON of 112 ± 6. Both ꢁ[Met] and TON decline steadiꢀ
ly with [Met]0 (Table S4) for all Experiment 1 and 2 catalyst
doses. The average ki value of (3 ± 1) × 10ꢀ4 M is identical to
the single catalyst dose value calculated from the data in Table
1. The combined data set has the following broad implications
for ultraꢀdilute TAML catalysis with Feꢌꢍꢎ < 1 × 10ꢀ6
(where no k2i processes are observed).
M
More complete metaldehyde removal with continuous addi-
tions of catalyst and oxidant. As a result of k2i processes,
deeper Met removals are more efficiently accomplished by
using multiple catalyst doses to ensure that [1a] < ca. 1 × 10ꢀ6
M. The optimized catalyst dose was that giving a [1a] of 4 ×
10ꢀ7 M. The data shown in Fig. 3 represent two separate experꢀ
iments demonstrating the efficacy of multiple catalyst doses
each giving a [1a] of 4 × 10ꢀ7 M. Experiment 1 employed a
[Met]0 of 3.32 × 10ꢀ4 M (Fig. 3 solid circles and Fig. S6).
Since Met mineralization requires 20 equivalents of NaClO, a
slight excess (7.61 × 10ꢀ3 M) was added. The first catalyst dose
consumed Met for 72 h before the reaction ceased. Seven adꢀ
ditional 1a doses were added leading to a slow oxidation of
75.3% over 576 h. After the addition of the fifth catalyst dose,
the position of the CH3COOꢀ 1H NMR signal at 1.956 ppm
Figure 3. Kinetic traces of metaldehyde degradation by NaClO
catalyzed by 1a in pH 7.0 D2O (0.01 M, phosphate). Two experiꢀ
ments are shown. Experiment 1: solid circles (●) and top time
axis, [Met]0 = 3.32 × 10ꢀ4 M, [1a] = 4.0 × 10ꢀ7 M, [NaClO] = 7.61
× 10ꢀ3 M. Experiment 2: hollow squares (□) and bottom time
axis, [Met]0 = 1.12 × 10ꢀ4 M, [1a] = 4.0 × 10ꢀ7 M, [NaClO] = 4.71
× 10ꢀ3 M. For both experiments: *indicates the addition of one 4.0
× 10ꢀ7 M catalyst dose, **indicates the addition of NaClO suffiꢀ
cient to raise the [NaClO] by 3.44 × 10ꢀ3 M, ***indicates the adꢀ
dition of one catalyst dose giving a [1a] of 4.0 × 10ꢀ7 M and
NaClO sufficient to raise the [NaClO] by 4.79 × 10ꢀ3 M NaClO.
indicated a solution pH of 5.2, 1.8 units less than the initial pH
40
of 7 (Fig. S7).14,
Therefore at 360 h, enough additional
NaClO was added to raise the [NaClO] by 4.79 × 10ꢀ3 M. This
addition returned the CH3COOꢀ signal to its initial value while
providing additional oxidant. The initial metaldehyde concenꢀ
tration ([Met]0), change in metaldehyde concentration
(ꢁ[Met]), TON, and %Cvn for each catalyst dose are summaꢀ
rized in Table S4. For all catalyst doses, ꢁ[Met] and TON
declined with [Met]0.
First, dilution of the substrate poses a formidable challenge to
ultradilute catalysis by diminishing the rate of substrate transꢀ
formation without altering the rate of catalyst inactivation. By
adapting eq 3, a relation previously derived for the determinaꢀ
tion of ki values32 and recently used to determine the ki values
for 15 TAML catalysts in the pH 7 oxidation of Orange II
(OrII),33 new forms which model the change in substrate conꢀ
centration (∆S, eq 4), % conversion (%Cvn, eq 5) and the
TON (eq 6) can be obtained and fit to the Met data where S0 =
[Met]0, S∞ = [Met]∞, and ∆S = ∆Met. These new forms turn
out to be very useful in developing the map and in better unꢀ
derstanding the meaning of TON where the catalyst is degradꢀ
ing. In the oxidation of any substrate under any one set of conꢀ
ditions, both kII and ki are fixed. For a fixed catalyst dose (Feꢀ
Tot), the ratio of these dictates the change in substrate concenꢀ
tration (ꢁS) and the TON, both of which decline with the iniꢀ
tial substrate concentration (S0) as shown in eqs 4 and 6. The
Met data confirm this (Fig. 4). The mathematical treatment
In Experiment 2, the oxidation of a fresh 1.12 × 10ꢀ4 M Met
solution was studied (Fig. 3 hollow squares and Fig. S8) to
examine if the accumulating acetate or inactivated catalyst
products were contributing to the deterioration in ꢁ[Met] and
TON.41 For each catalyst dose of Experiment 2, the [Met]0,
ꢁ[Met], TON, and %Cvn are also summarized in Table S4.
There is considerable agreement between the data sets. If the
concentration of Met is followed along the black dotted curve
of Experiment 1 through the white squares of Experiment 2,
the result is representative of one continuous degradation proꢀ
cess. Here, a ca. one order of magnitude reduction in [Met]
(330–30 µM) by less than 0.02 M NaClO can be achieved in
~47 days through treatment with 16 doses of 1a each giving a
solution [1a] of 4.0 × 10ꢀ7 M. The combined process would
consume less than 60 equivalents of NaClO and proceed with
ACS Paragon Plus Environment

Page 7 of 27
Journal of the American Chemical Society
that resulted in eq 3 considered all of the events shown in or independent processes, or both. Since contributions from
Scheme 1 to arrive at the most general expression.32 Contribuꢀ
tions from the catalyst activation process kI are absent from
this form indicating that S∞, and thus ꢁS and TON, are deterꢀ
mined by the competition for Ac between the substrate oxidaꢀ
tion (ꢀd[S]/dt = kII[S][Ac]) and catalyst inactivation (d[Ic]/dt =
ki[Ac]) processes alone. The rate of the former depends on the
[S] while that of the latter does not. As a result, ꢁS declines
with S0. This dynamic is especially important when the subꢀ
strate is ultraꢀdilute as is the usual case with MPs in municipal
wastewater.
NaClO dependent inactivation are negligible, the mandatory
changes in ki must originate from NaClO independent processꢀ
es which operate from pH 5.2 to 7. Because the NaClO and
H2O2 Ac are similar and the observed dependence on Lewis
acidity is independent of oxidant, it is therefore likely that
similar processes are operating in the H2O2 systems.
1
2
3
4
5
6
7
8
Third, while most MPs can be effectively treated by existing
TAML systems operating under ultraꢀdilute aqueous condiꢀ
tions, those having kII/ki ratios ≤ 1 × 106 Mꢀ1 cannot. As notꢀ
ed, when [1a] > 1 × 10ꢀ6 M, contributions to ki from decompoꢀ
sition processes that are second order in [1a] (k2i processes)
become significant. Therefore, we consider an FeTot of 1 × 10ꢀ6
M to be an upper bound for effective ultraꢀdilute catalysis. In
principle, below this, eqs 3–6 model the So/S∞ ratio, %Cvn,
and TON in TAML catalysis for the oxidation of any substrate
by any oxidant with any catalyst under any conditions providꢀ
ed inactivation also occurs from Ac and neither the substrate
nor the products perturb the catalytic cycle (e.g. substrate
binding to the catalyst). However, it should be noted that the
critical dimension of time is not captured by eqs 3–6 and thus
cannot be included in the following discussion—we are further
considering how to address the time dimension.
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
ꢙ
ꢙ
ꢄꢄ
ꢓ
ꢚꢛ
∆S ꢂ Sꢔꢕ1 ꢖ ꢗꢘ
ꢜꢝꢞꢟ
(4)
(5)
ꢙ
ꢙ
ꢄꢄ
ꢓ
ꢚꢛ
%ꢠꢁꢡ ꢂ ∆ꢉ ꢀ 100% ꢂ ꢕ1 ꢖ ꢗꢘ
ꢜꢝꢞꢟ ꢀ 100%
ꢉ
ꢑ
ꢙ
ꢙ
ꢄꢄ
ꢓ
ꢊ
ꢥꢦ
ꢜꢝꢞ
ꢢꢣꢘꢤ
ꢧ
TON ꢂ
Sꢔ
(6)
ꢚꢛ
ꢜꢝꢞ
Using the known kII and ki values, eq 5 predicts within ±3%
the observed experimental 5 %Cvn of Met found for 1a/H2O2
treatment ([1a] = 4 × 10ꢀ7 M) in pH 7 buffered D2O (0.01 M
phosphate), the (14.4 ± 0.9) %Cvn of Met for 1a/NaClO
treatment ([1a] = 4 × 10ꢀ7 M) in pH 7 buffered D2O (0.01 M
phosphate), the 18 %Cvn of the persistent SSRI Sertraline
(Ser) for 1a/H2O2 treatment ([1a] = 3 × 10ꢀ7 M) in pH 7 buffꢀ
ered H2O (0.01 M phosphate), the 27 %Cvn of the azo dye
Orange II (OrII) for 1b/H2O2 treatment ([1b] = 2 × 10ꢀ8 M) in
pH 7 buffered H2O (0.01 M phosphate), the >98 %Cvn of
17αꢀethynylestradiol (EE2) for 1b/H2O2 treatment ([1b] = 8 ×
10ꢀ8 M) in pH 7 buffered H2O (0.01 M phosphate),43 and the
17 %Cvn of the dye Safranin O (SO) in pH 11 buffered H2O
([1a] = 7.5 × 10ꢀ8 M).36 Figure 5 is a helpful graphical repreꢀ
sentation of eq 5 onto which these processes can be plotted.
The kII value for 1a oxidation Met in both the H2O2 and
NaClO systems, 120 Mꢀ1sꢀ1, is the lowest of any MP we have
ever tested. The kII value for 1b/H2O2 oxidation of EE2 at pH
7, 8.6 × 105 Mꢀ1sꢀ1, is the highest.12 The value for 1a/H2O2
oxidation of Ser, 740 Mꢀ1sꢀ1, is closer to the Met than the EE2
value as expected. The value for 1b/H2O2 oxidation OrII,
4,950 Mꢀ1sꢀ1, is closer to the EE2 than the Met value as exꢀ
pected. Thus eqs. 3ꢀ6 hold over a ca. 104 Mꢀ1 range of kII/ki for
two catalysts and oxidants and at pH 7 and 11, adding confiꢀ
dence to both the validity and generality of the analytical
model for both academic and realꢀworld applications.
Figure 4. Correlation between the ꢁ[Μet] and the [Met]0. Each
data point represents the ꢁ[Μet] for each Experiment 1 or 2 cataꢀ
lyst dose. The line gives the eq 5 predicted ꢁ[Μet] values with ki
= 3.0 × 10ꢀ4 sꢀ1, kII = 120 Mꢀ1sꢀ1, and [1a] = 4.0 × 10ꢀ7 M. The othꢀ
er conditions are as indicated in the Fig. 3 caption and Table S4.
Second, from pH 5.2–7, the oxidant independent 1a inactivaꢀ
tion process operates via a mechanism that is sensitive to the
Lewis acidity at Fe. The linearity found in Fig. 4 indicates that
a constant kII/ki ratio holds for all of the catalyst doses of both
experiments. The data were collected at varying pH, [NaClO],
[product], and [Ic] such that the linear relationship of Fig. 4
also establishes that these parameters do not alter kII/ki implyꢀ
ing that inhibition by product or inactive catalyst species and
NaClO dependent inactivation processes are negligible. The
data in Fig. 4 were collected at pH values over the range of
5.2–7. TAML kII values have been observed to increase with
pH by more than an order of magnitude from pH 6–7.42
Therefore, the 1a/NaClO kII and ki must track each other closeꢀ
ly over this range as was found in catalysis by 15 TAML actiꢀ
vators with H2O2 at pH 7.33 We have interpreted these fixed
kII/ki ratios as evidence that the Lewis acidity at Fe exerts a
common influence over both kII and ki processes at pH 7, but
for the H2O2 systems we were unable to determine whether the
ratio mandated changes in ki derive from the H2O2 dependent
Typically, water treatment plants are driven by a need to
achieve regulated levels rather than complete removal of polꢀ
lutants. Equation 5 could be used to predict the optimal cataꢀ
lyst loading, FeTot (below 1 × 10ꢀ6 M), needed to effect a given
%Cvn of any MP under any one set of conditions from data
obtained by measuring one incomplete kinetic trace at a
known FeTot which gives kII/ki. This will always hold provided
the process is rapid enough to be deployable. If ki is known
under these conditions, kII is also known—an extensive series
of pH 7 ki values for TAMLs have been determined.33 We are
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 27
would require roughly a doubling of the 1a kII value or halving
1
2
3
4
5
6
7
8
of the 1a/NaClO ki value while keeping the other constant. We
are working towards achieving and exceeding this entirely
realistic goal via iterative catalyst design.
Fourth, when FeTot
≤
ca. 1 × 10-6 M, successive small catalyst
doses are not more effective than larger catalyst doses. As
indicated by eq 6 and Fig. 6, the dependence of TON on the
kII/ki ratio and FeTot is complex. We have interpreted the relaꢀ
tionship in the following way. For any one catalyst operating
under any one set of conditions, ki is fixed. The value of the
kII/ki ratio then varies with kII. Thus, the value of kII determines
the dependence between TON and FeTot and places the oxidaꢀ
tion process on a spectrum between two limits. At the low kII
limit, when kIIFeTot/ki < 1, eq 6 can be approximated by the
first term of a Maclaurin series, TON ~ kIIS0/ki (contributions
from the higher order terms of the series are negligible). Conꢀ
sequently, in the 1a/NaClO conversion of slowly transformed
Met, for which kII/ki = 4 × 105 Mꢀ1, the TON is largely indeꢀ
pendent of FeTot as shown in Table 2 and along the surface of
Fig. 6 at the low kII/ki limit. In these oxidations, every catalyst
molecule of any catalyst dose does about the same amount of
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
work, and the %Cvn scales linearly with FeTot
.
Figure 5. The eq 5 predicted %Cvn as a function of
log(FeIIITot) and either log(kII/ki) for treatment with any suitable
catalyst (top axis), log(kII) for pH 7 treatment with 1a/NaClO,
log(kII) for pH 7 treatment with 1b/H2O2, log(kII) for pH 7
treatment with 1a/H2O2, or log(kII) for pH 11 treatment with
1b/H2O2 (bottom axes which shift with different oxidants and
pHs). The points shown indicate experimental data for converꢀ
sion of various substrates having the identities indicated in the
text. The kII, ki, and %Cvn values for each data point were
obtained individually.
treating a larger set of MPs and examining the behavior of
additional catalysts over a wider pH range to obtain kII, ki, and
%Cvn under different conditions to further cement the usefulꢀ
ness of this predictive tool.
The upper bound of the log(kII/ki) range shown in Fig. 5 (top
horizontal axis) is set at the hypothetical diffusion controlled
limit for 1a/NaClO treatment at pH 7 (kII of ca. 1010 Mꢀ1sꢀ1,
topmost axis below figure).44 In applying TAML processes to
remove micropollutants, useful processes deliver %Cvn ≥ 50.
The equation predicts that subꢀpicomolar [1a] can effect 50
%Cvn for very rapidly oxidizing substrates. However, practiꢀ
cal concerns such as reaction time will likely limit utility
here—we are working to discover a similar tool for reaction
time. The lower kII/ki limit of the green band (50 %Cvn) is
found at ca. 106 Mꢀ1. The kII/ki ratios for TAML catalyzed oxiꢀ
dation of many MPs lie above this. However, this sets the
lower kII bound for the most useful 1a/H2O2 treatment with an
FeTot of 1 × 10ꢀ6 M at ca. 1,100 Mꢀ1sꢀ1 near pH 7. Substrates
more slowly oxidized than this are not well treated with a sinꢀ
gle catalyst dose of 1a/H2O2. The lower ki of 1a/NaClO treatꢀ
ment shifts this bound to ca. 300 Mꢀ1sꢀ1. However, Met remains
just out of reach with a kII of 120 ± 30 Mꢀ1sꢀ1. As the 1a/NaClO
data demonstrate, the key to extending ultraꢀdilute catalysis to
more slowly transformed substrates is the development of
systems that maximize the kII/ki ratio by disproportionately
increasing kII, decreasing ki, or both. An effective Met solution
Figure 6. The eq 6 calculated dependence of the ΤΟΝ on kII/ki
and FeTot for a fixed S0 of 1 × 10ꢀ6 M. From [FeTot] 10ꢀ8–10ꢀ6 M
(not shown), the surface shape changes little from that shown
at [FeTot] = 10ꢀ8 M.
ꢙ
ꢙ
ꢄꢄ
ꢓ
ꢚꢛ
At the high kII limit where ꢗꢘ
≤ 0.1, the numerator of
ꢜꢝꢞ
eq 6 approaches 1 and TON ~ S0/FeTot. For example, for the
1b/H2O2 treatment of EE2 introduced above, kII/ki = 2.9 × 109
ꢙ
ꢙ
ꢄꢄ
ꢓ
ꢚꢛ
Mꢀ1 and ꢗꢘ
ꢂ 2.5 × 10ꢀ100. For treatment of such rapidꢀ
ꢜꢝꢞ
ly oxidized substrates, as FeTot increases, so does ꢁ[S] and
with it the dilution of the substrate during functioning catalyꢀ
sis. This decreases the rate of substrate oxidation relative to
that of catalyst inactivation causing average TONs to drop.
This could lead one to suspect that treatment with multiple
small catalyst doses would lead to improved performance.
However, the use of multiple small catalyst doses is not more
efficient at achieving a set %Cvn than adding the equivalent
Fetot in a single dose provided the sum of all doses is less than
ca. 1 × 10ꢀ6 M because dilution lowers the average TON reꢀ
ACS Paragon Plus Environment

Page 9 of 27
Journal of the American Chemical Society
MPs, micropollutants; Met, metaldehyde; AC, activated carbon;
gardless of how the catalyst is administered (See the SI for a
1
2
3
4
5
6
7
8
ppb, parts per billion; ppt, parts per trillion; Rc, resting catalysts;
Ac, active catalysts; S, substrate; DBPs, disinfection byproducts;
mathematical explanation of this). When eq 3 holds, a single
TAML catalyst dose is optimal for most rapidly achieving a
desired %Cvn.
EE2,
17αꢀethynylestradiol;
TMSP,
sodiumꢀ3ꢀ
trimethylsilylpropionate; NADH, βꢀnicotinamide adenine dinuꢀ
cleotide reduced.
Conclusion
REFERENCES
This study explores the behavior of TAML catalysis at the
performance boundary prescribed by the exceptionally recalciꢀ
trant pollutant, Met. The findings communicated herein have
broad significance for TAML catalysis. The rates of 1a cataꢀ
lyzed Met transformation by NaClO and H2O2 are identical
giving a strong indication that the same reactive intermediate
attacks Met in both systems. However, TAML/NaClO delivers
a threeꢀfold kII/ki enhancement by lowering the rate of catalyst
inactivation, which we attribute to elimination of contributions
from H2O2 dependent inactivation processes. At [1a] > 1 × 10ꢀ
6 M, inactivation pathways that are bimolecular in catalyst (k2i)
and outpace those unimolecular in catalyst (ki) render 1a/H2O2
transformation of Met completely ineffective—this establishes
an upper bound for [TAML] in H2O2 systems of 1 ꢀM. The
use of NaClO allows processes with [1a] > 1 ꢀM by restoring
functioning catalysis indicating that both ki and k2i are lower in
the NaClO system. In both the H2O2 and NaClO systems, k2i
processes outpace ki processes. Further investigations of the
mechanisms of the k2i processes and their dependences on the
identity of the oxidant are underway. Application of eq 3 and
forms derived from it to the Met data indicates that while ∆S
decreases with S0, %Cvn remains constant. The expression for
TON provides a more detailed understanding of the underlying
chemical processes as well as the effect of substrate dilution
on this venerable metric for evaluating catalyst performance.
This work demonstrates the advances in understanding which
can be achieved through detailed analysis of the mechanisms
of both catalyst operation and inactivation—the latter is a critꢀ
ical and underexplored territory.
(1)
Shane A. Snyder; Eric C. Wert; Hongxia Lei; Paul Westerhoff;
Yoon, Y. Removal of EDCs and Pharmaceuticals in Drinking and Reuse
Treatment Processes, 1st ed.; AwwaRF: Denver, 2007.
9
(2)
Schwarzenbach, R. P.; Escher, B. I.; Fenner, K.; Hofstetter, T.
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
B.; Johnson, C. A.; von Gunten, U.; Wehrli, B. Science 2006, 313, 1072ꢀ
1077.
(3)
(4)
McArdell, C. S. Norman Bulletin 2015, 36ꢀ37.
Kundu, S.; Chanda, A.; Thompson, J. V. K.; Diabes, G.;
Khetan, S. K.; Ryabov, A. D.; Collins, T. J. Catal. Sci. Technol. 2015, 5,
1775ꢀ1782.
(5)
T. J. Environ. Sci. Technol. 2013, 47, 5319ꢀ5326.
(6) Kundu, S.; Chanda, A.; EspinosaꢀMarvan, L.; Khetan, S. K.;
Collins, T. J. Catal. Sci. Technol. 2012, 2, 1165ꢀ1172.
(7) Shen, L. Q.; Beach, E. S.; Xiang, Y.; Tshudy, D. J.; Khanina,
Kundu, S.; Chanda, A.; Khetan, S. K.; Ryabov, A. D.; Collins,
N.; Horwitz, C. P.; Bier, M. E.; Collins, T. J. Environ. Sci. Technol. 2011,
45, 7882ꢀ7887.
(8)
Collins, T. J. Catal. Sci. Technol. 2011, 1, 437ꢀ443.
(9) Beach, E. S.; Duran, J. L.; Horwitz, C. P.; Collins, T. J. Ind.
Eng. Chem. Res. 2009, 48, 7072ꢀ7076.
Beach, E. S.; Malecky, R. T.; Gil, R. R.; Horwitz, C. P.;
(10)
Chanda, A.; Khetan, S. K.; Banerjee, D.; Ghosh, A.; Collins, T.
J. J. Am. Chem. Soc. 2006, 128, 12058ꢀ12059.
(11)
Banerjee, D.; Markley, A. L.; Yano, T.; Ghosh, A.; Berget, P.
B.; Minkley, E. G.; Khetan, S. K.; Collins, T. J. Angew. Chem. Int. Edit.
2006, 45, 3974ꢀ3977.
(12)
Mills, M. R.; AriasꢀSalazar, K.; Baynes, A.; Shen, L. Q.;
Churchley, J.; Beresford, N.; Gayathri, C.; Gil, R. R.; Kanda, R.; Jobling,
S.; Collins, T. J. Sci. Rep. 2015, 5, article number: 10511.
(13)
Pharmaceutical Compounds in Wastewater Effluents; London, UK, 2011.
(14) Tang, L. L.; DeNardo, M. A.; Gayathri, C.; Gil, R. R.; Kanda,
R.; Collins, T. J. Environ. Sci. Technol. 2016, 50, 5261ꢀ5268.
(15) Autin, O.; Hart, J.; Jarvis, P.; MacAdam, J.; Parsons, S. A.;
Jefferson, B. Appl. Catal. B-Environ. 2013, 138, 268ꢀ275.
Churchley, J.; Collins, T. J.; Jobling, S. Catalytic Oxidation of
(16)
127, 77ꢀ83.
(17)
James, C. P.; Germain, E.; Judd, S. Sep. Purif. Technol. 2014,
ASSOCIATED CONTENT
Scheideler, J.; Bosmith, A. Aqua Gas 2014, 94, 52ꢀ57.
Supporting Information. Details of metaldehyde degradation
(Figs. S1ꢀ8); metaldehyde degradation summary (Tables S1–4).
This material is available free of charge via the Internet at
http://pubs.acs.org.
(18)
Collins, T. J.; GordonꢀWylie, S. W.; Horwitz, C. P. U.S.
Patent 6054580, 2000.
(19)
(20)
Deborde, M.; von Gunten, U. Water Res. 2008, 42, 13ꢀ51.
Collins, T. J.; GordonꢀWylie, S. W. U.S. Patent 5,847,120,
1998.
(21)
Mills, M. R.; Burton, A. E.; Mori, D. I.; Ryabov, A. D.;
AUTHOR INFORMATION
Collins, T. J. J. Coord. Chem. 2015, 68, 3046ꢀ3057.
(22) Ghosh, M.; Singh, K. K.; Panda, C.; Weitz, A.; Hendrich, M.
P.; Collins, T. J.; Dhar, B. B.; Sen Gupta, S. J. Am. Chem. Soc. 2014, 136,
Corresponding author: Phone: +1 412 268 6335; fax: +1 412
268 1061; E-mail: tc1u@andrew.cmu.edu (T.J.C.)
9524ꢀ9527.
(23)
Kundu, S.; Annavajhala, M.; Kurnikov, I. V.; Ryabov, A. D.;
Author Contributions
Collins, T. J. Chem. Eur. J. 2012, 18, 10244ꢀ10249.
All authors have given approval to the final version of the manuꢀ
script.
(24)
(25)
Fukuta, N. Nature 1963, 199, 475ꢀ476.
Craven, E. C.; Jowitt, H.; Ward, W. R. J. Appl. Chem. 1962,
12, 526ꢀ535.
Funding Sources
(26)
1274ꢀ1278.
(27)
Pauling, L.; Carpenter, D. C. J. Am. Chem. Soc. 1936, 58,
T.J.C thanks the Heinz Endowments for funding. NMR instruꢀ
mentation at CMU was partially supported by NSF (CHEꢀ
0130903 and CHEꢀ1039870).
George, P. Biochem. J. 1953, 55, 220ꢀ230.
(28)
Kelm, M.; Pashalidis, I.; Kim, J. I. Appl. Radiat. Isotopes 1999,
51, 637ꢀ642.
(29)
Edward M. Arnett; W. Emmett Barkley; Peter Beak; Edwin D.
ACKNOWLEDGMENT
Becker; Henry E. Bryndza; Imogene L. Chang; Carol Creutz; Rick L.
Danheiser; Eric M. Gordon; Robert J. Lackmeyer; Lee Magid; Thomas F.
McBride; Ann M. Norberg; Edward W. Petrillo; Stanley H. Pine; Fay M.
Thompson; Tamae Maeda Wong; Kasandra Gowen; Sarah W. Plimpton;
Butera., J. F. Prudent Practices in the Laboratory: Handling and Disposal
of Chemicals, National Academy Press: Washington, D.C., 1995.
T.J.C. thanks the Heinz Endowments for support. M.R.M thanks
the Steinbrenner Institute and the R. K. Mellon Foundation for
Doctoral Fellowships.
ABBREVIATIONS
(30)
Sedlak, D. L.; von Gunten, U. Science 2011, 331, 42ꢀ43.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 27
(31)
5001.
(32)
Yang, X.; Shang, C. Environ. Sci. Technol. 2004, 38, 4995ꢀ
Emelianenko, M.; Torrejon, D.; DeNardo, M.; Socolofsky, A.;
(39)
Miller, J. A.; Alexander, L.; Mori, D. I.; Ryabov, A. D.;
Collins, T. J. New J. Chem. 2013, 37, 3488ꢀ3495.
(40)
Anal. Chim. Acta. 2009, 648, 105ꢀ12.
(41)
(42)
C.; Panda, T. K.; Sen Gupta, S.; Ryabov, A. D.; Collins, T. J. Chem. Eur.
J. 2015, 21, 6226ꢀ6233.
(43)
Properties of a Novel, Completely Aliphatic TAML Activator and
Applications of TAML Catalysis for Oxygen Evolution and Pollution
Remediation. Carnegie Mellon University, Pittsburgh, PA, 2016.
1
2
3
4
5
6
7
8
Tynkkynen, T.; Tiainen, M.; Soininen, P.; Laatikainen, R.
Ryabov, A.; Collins, T. J. Math. Chem. 2014, 52, 1460ꢀ1476.
(33) DeNardo, M. A.; Mills, M. R.; Ryabov, A. D.; Collins, T. J. J.
Am. Chem. Soc. 2016, 138, 2933ꢀ2936.
(34)
1980, 15, 21ꢀ4.
Crabtree, R. H. Chem. Rev. 2015, 115, 127ꢀ150.
Warner, G. R.; Mills, M. R.; Enslin, C.; Pattanayak, S.; Panda,
Abe, K.; Hagiwara, J.; Machida, W. Taiki Osen Gakkaishi
(35)
Bartos, M. J.; GordonꢀWylie, S. W.; Fox, B. G.; Wright, L. J.;
Mills, M. R. Synthesis, Characterization, and Unique
Weintraub, S. T.; Kauffmann, K. E.; Munck, E.; Kostka, K. L.; Uffelman,
E. S.; Rickard, C. E. F.; Noon, K. R.; Collins, T. J. Coordin. Chem. Rev.
1998, 174, 361ꢀ390.
9
(36)
Chanda, A.; Ryabov, A. D.; Mondal, S.; Alexandrova, L.;
(44)
Koenig, S. H.; Brown, R. D. Proc. Nat. Acad. Sci. USA 1972,
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
Ghosh, A.; HangunꢀBalkir, Y.; Horwitz, C. P.; Collins, T. J. Chem. Eur. J.
2006, 12, 9336ꢀ9345.
69, 2422ꢀ2425.
(37)
Horwitz, C. P.; Fooksman, D. R.; Vuocolo, L. D.; Gordonꢀ
Wylie, S. W.; Cox, N. J.; Collins, T. J. J. Am. Chem. Soc. 1998, 120,
4867ꢀ4868.
(38)
Sen Gupta, S.; Stadler, M.; Noser, C. A.; Ghosh, A.; Steinhoff,
B.; Lenoir, D.; Horwitz, C. P.; Schramm, K. W.; Collins, T. J. Science
2002, 296, 326ꢀ328.
ACS Paragon Plus Environment

Page 11 of 27
Journal of the American Chemical Society
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
10
ACS Paragon Plus Environment

Page 13 of 27
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
ACS Paragon Plus Environment
10
11
12

PaJgoeu1r5naolfo2f7the American Chemical Society
1
2
3
4
5
6
7
8
9
ACS Paragon Plus Environment
10
11

Journal of the American ChemicPaalgSeo1c6ieotyf 27
1
2
3
4
5
6
7
8
9
10
ACS Paragon Plus Environment

PaJgoeu1r7naolfo2f7the American Chemical Society
1
2
3
4
5
6
7
8
ACS Paragon Plus Environment
9
10
11

PaJgoeu1r9naolfo2f7the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
ACS Paragon Plus Environment
14
15

Journal of the American Chemical Society
Page 20 of 27
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
ACS Paragon Plus Environment

Page 21 of 27 Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
ACS Paragon Plus Environment
17
18
19