Facile Quantum Yield Determination via NMR Actinometry

Article abstract of DOI:10.1021/acs.orglett.8b00391

A simplified approach to quantum yield (φ) measurement using in situ LED NMR spectroscopy has been developed. The utility and performance of NMR actinometry has been demonstrated for the well-known chemical actinometers potassium ferrioxalate and o-nitrobenzaldehyde. A novel NMR-friendly actinometer, 2,4-dinitrobenzaldehyde, has been introduced for both 365 and 440 nm wavelengths. The method has been utilized successfully to measure the quantum yield of several recently published photochemical reactions.

Full text of DOI:10.1021/acs.orglett.8b00391

Letter
Cite This: Org. Lett. XXXX, XXX, XXX−XXX
pubs.acs.org/OrgLett
Facile Quantum Yield Determination via NMR Actinometry
,†
†
†,‡
§
,†
Yining Ji,* Daniel A. DiRocco, Cynthia M. Hong,
Michael K. Wismer, and Mikhail Reibarkh*
†
Process Research & Development, Merck & Co., Inc., Rahway, New Jersey 07065 United States
‡
Chemical Sciences Division, Lawrence Berkeley National Laboratory and Department of Chemistry, University of California,
Berkeley, California 94720 United States
§
Scientific Engineering and Design, Merck & Co., Inc., Rahway, New Jersey 07065 United States
S Supporting Information
*
ABSTRACT: A simplified approach to quantum yield ( ) measurement using in situ
ϕ
LED NMR spectroscopy has been developed. The utility and performance of NMR
actinometry has been demonstrated for the well-known chemical actinometers potassium
ferrioxalate and o-nitrobenzaldehyde. A novel NMR-friendly actinometer, 2,4-
dinitrobenzaldehyde, has been introduced for both 365 and 440 nm wavelengths. The
method has been utilized successfully to measure the quantum yield of several recently
published photochemical reactions.
n recent years, photochemistry has experienced a so-called
“golden age”, rapidly advancing the field of organic synthesis.
Derived from the Beer−Lambert law, the rate of disappear-
I
ance of an actinometric compound (Act) under incident light is
6
The exploration of photochemical reactivity has opened up
uncharted chemical space with the invention of cutting-edge
chemical transformations facilitating access to novel chemical
governed by eq 1.
d[Act]
dt
εb[Act]
−
= I ϕ(1 − 10−
)
0
1
(1)
entities. Despite great advances in the field, detailed
mechanistic investigations have been sparse. LED-based NMR
illumination devices have been introduced as a simple and
powerful tool to obtain valuable and rich kinetic and structural
insights into photochemical reactions.
One of the most fundamental parameters in photochemistry
With sufficiently high molecular absorptivity ε (M−1 cm−1)
and light path length b (cm), eq 1 simplifies to eq 2.
2
k0
ϕ
I0 =
(2)
is quantum yield ( ). Quantum yield is a measure of efficiency
ϕ
with which absorption of a photon results in a desired chemical
transformation. These data can be used to identify whether a
radical-chain mechanism may be operable, in which case
absorption of a single photon may result in the production of
In this case, the expression becomes independent of the
concentration of the actinometric compound, and the light
intensity is simply measured as the fraction of the zeroth order
−1 −1
kinetic constant k (mol L s ) and the quantum yield. While
0
3
more than one molecule of product. One widely utilized
the assumption above is true for a typical photochemistry setup,
that is not the case for NMR actinometry: the light path length
in the LED-illuminated NMR device is 1.1 mm, which could be
10−20 times shorter than a conventional photochemical cell.
Furthermore, high concentrations of the actinometer can result
in an increase in the solution viscosity, potentially resulting in
line broadening of signals in the NMR spectrum. For this
reason, concentration should be considered when designing
NMR actinometry experiments.
actinometry model system is based on photodecomposition of
4
ferrioxalate ion. However, the protocol is very tedious, requires
complete darkness or the use of deep red light at all stages, and
is associated with inherent sample manipulation errors and
5
analyst variabilities. Direct NMR detection would offer a much
simpler protocol with minimal sample manipulation since all
data points for an analysis are taken directly on a single sample.
Herein, we describe a comprehensive methodology for
quantum yield determination via in situ LED NMR analysis.
The developed method is demonstrated on several chemical
actinometers, including well-known ferrioxalate and o-nitro-
The use of o-NBA as a chemical actinometer is well-known,
and a quantum yield of 0.5 has been reported over its
6
absorption spectral region of 300−410 nm (Scheme 1). For
6
benzaldehyde (o-NBA), as well as 2,4-dinitrobenzaldehyde
(
2,4-DNBA), a novel actinometer developed in the course of
Received: February 2, 2018
this work.
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.8b00391
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Scheme 1. Photochemical Transformation of o-
Nitrobenzaldehyde to o-Nitrosobenzoic Acid in Ultraviolet
Light
ment with 90 mM or higher concentration of o-NBA is
sufficient since its reaction rate plateaus at these concentrations
(Figure 1). Simple recalibration of the light intensity before and
after the actual reaction can be performed (see the Supporting
Information for details). The same simplified calibration
strategy can be applied for any new LED source.
Unfortunately, while convenient for NMR measurements, o-
NBA is limited to a 300−410 nm spectral range. We sought to
extend the scope of NMR actinometry to include a 440 nm
wavelength, widely used for visible-light photoredox catalysis.
o-NBA does not have any appreciable absorption at this
our use, we chose an UV LED source with a peak emission at
11
7
3
65 nm. Our strategy required the measurement of initial rates
for the reaction of o-NBA at increasing concentrations, until the
maximum rate is reached and increasing concentrations no
longer change reaction rates (Figure 1). These data were then
wavelength, but the addition of an electron-withdrawing group
12
could produce the desired red-shift in absorption. For such
purpose, we investigated 2,4-dinitrobenzaldehyde (2,4-DNBA)
as a commercially available chemical actinometer suitable for
4
40 nm photochemical reactions (Scheme 3).
Scheme 3. Photochemistry of 2,4-Dinitrobenzaldehyde
LED−NMR studies showed that 2,4-DNBA undergoes the
same chemical transformation when subjected to 365 or 440
nm light. Despite the presence of two nitro groups, a single 2-
nitroso product was formed at both wavelengths. The structure
of the 2-nitroso product has been confirmed using 2D NMR
data (see the Supporting Information for details). The
concentration dependence of initial reaction rates for 2,4-
DNBA was measured for both wavelengths (Figures 2 and 3).
As expected, the molecular absorptivity at 365 nm was
Figure 1. Experimental (blue circles) and fitted (red line) plot of initial
rates (mM/min) versus different [o-NBA] (mM) for the reaction of
0
Scheme 1.
fitted explicitly into eq 1, allowing extraction of both unknowns,
−
1
−1
namely, the molecular absorptivity (ε, 265 M cm ) and the
−1
−1
light intensity (I , 23.4 μeinstein L s ) (see Supporting
0
Information for details). Our fitted molecular absorptivity
−1
−1
−1
matched the published value for o-NBA (260 M cm ), thus
significantly higher than that at 440 nm: 611.5 and 23.4 M
−1
providing an independent validation for the method.
cm , respectively.
With the light intensity (I ) of the 365 nm LED source
0
measured, we began investigating the applicability of NMR
actinometry for quantum yield determination. For this purpose,
we chose the direct photocatalytic C−H fluorination of leucine
8
methyl ester (Scheme 2). To apply eq 2, the zeroth order
Scheme 2. Direct Photocatalytic C−H Fluorination of
Leucine Methyl Ester
Figure 2. Experimental (blue circles) and fitted (red line) plot of initial
rates (mM/min) versus different [2,4-DNBA] (mM) for the reaction
0
kinetic constant (k ) must be determined. Photoredox catalysts
of Scheme 3 at 365 nm.
0
often possess high molar absorptivities (for example, ε = 1.35
325
4
−1
−1
4−
9
×
10 M cm for [W O ] ). Therefore, we are usually
The light intensity of the 365 nm LED source calibrated with
o-NBA was used to calculate the quantum yield for 2,4-DNBA
(eq 2), which turned out to be 1.00 (see the Supporting
Information for details). Determination of the quantum yield
for 2,4-DNBA at 440 nm required measuring the light intensity
of the 440 nm LED source. Ferrioxalate, a well-characterized
and broadly utilized actinometer with a range of wavelengths
1
0
32
operating close to the maximum achievable rate limit and a
change of the concentration of these photocatalysts would be
sufficient to confirm whether we are in the rate plateau region.
In this case, the same initial rate was obtained using either 2.5
or 10 mol % of the catalyst and a quantum yield of 0.22 was
10
calculated (see the Supporting Information for details).
13
The convenience and practicality of NMR actinometry were
highlighted when we noticed that light intensity of the 365 nm
LED deteriorated somewhat after prolonged usage. Having
characterized the concentration dependence, a single measure-
250−500 nm, was utilized for this purpose. Using classical
14
ferrioxalate actinometry protocol, the light intensity of the
440 nm LED source was determined to be 75.3 μeinstein
−1 −1
L s (see the Supporting Information for details). The
B
DOI: 10.1021/acs.orglett.8b00391
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
concept, a 13C NMR spectrum was acquired in 8.5 min with
8
192 transients for a CO2 enriched sample prepared by
dissolving dry ice in the standard ferrioxalate solution (Figure
). The presence of dissolved CO is evidenced by a
4
2
13
characteristic C chemical shift of ∼124 ppm.
Figure 3. Experimental (blue circles) and fitted (red line) plot of initial
rates (mM/min) versus different [2,4-DNBA] (mM) for the reaction
0
of Scheme 3 at 440 nm.
quantum yield for 2,4-DNBA at 440 nm was thereby calculated
to be 0.08.
With its quantum yield and molecular absorptivity
determined, 2,4-DNBA can be used as an NMR-friendly
chemical actinometer for both 365 and 440 nm wavelengths.
To demonstrate its utility, we repeated the recently published
Figure 4. 13C NMR spectrum of a solution of K [Fe(C O ) ] (0.15
3
2
4 3
M) in 0.05 M H SO in D O, with added 1% v/v MeCN (as a
2
4
2
reference) and CO (from dry ice). D1 = 1 ms; AQ = 49 ms, DS =
2
14
512, NS = 8192; total acquisition time is 8.5 min.
photocatalytic dehydrogenation of indoline (Scheme 4) using
Using LED−NMR to irradiate potassium ferrioxalate
Scheme 4. Photoredox Indoline Oxidation
solution at 440 nm, we confirmed that CO formed during
2
the reaction is detectable as a dissolved gas (Figure 5). CO2
4
40 nm LED−NMR. The quantum yield of this reaction
calculated with 2,4-DNBA as an NMR actinometer (see the
Supporting Information for details) was in a good agreement
with the previously reported value obtained using the classical
15
ferrioxalate method: 0.38 and 0.40, respectively.
The tedious nature and impracticality of the traditional
ferrioxalate protocol prompted the development of the NMR-
friendly 2,4-DNBA actinometer described in this work.
However, ferrioxalate is still the most broadly utilized
actinometer with a superior range of wavelengths (250−500
nm). Classical ferrioxalate actinometry (Scheme 5) requires
Figure 5. 13C NMR spectra tracking the photodecomposition at 440
Scheme 5. Photodecomposition of Ferrioxalate Ion Solution
nm of a solution of K [Fe(C O ) ] (0.15 M) in 0.05 M H SO in
3
2
4
3
2
4
D O. Each spectrum is acquired with D1 = 1 ms; AQ = 49 ms, DS =
2
5
12, NS = 8192 for a total acquisition time of 8.5 min.
further derivatization of the weakly absorbing photolysis
4
products to strongly absorbing compounds, which is both
accumulates in solution until the concentration limit is reached
and bubbles begin to form, leading to deterioration of both lock
level and a line shape. Direct measurement of the initial CO2
formation rate by NMR was conducted by using data points
before the solubility limit was reached. Although finding a
5
cumbersome and leads to inherent measurement errors. Thus,
we envisioned that direct NMR detection would considerably
streamline the experimental setup and lower experimental
errors.
As is evident from Scheme 5, this system does not contain
quantitative 13C NMR reference for CO under the exact
2
1
16
any H nuclei suitable for NMR detection. Furthermore, the
reaction environment was not feasible, one can convert CO2
13
presence of iron ions introduces a significant degree of
integrals in C NMR spectra into concentrations using known
paramagnetism, which leads to line-broadening due to faster
reaction rates.
relaxation rates. Initial NMR studies showed that CO gas
Potassium ferrioxalate photochemistry was conducted using
2
generated in the reaction is dissolved in the reaction mixture in
LED−NMR irradiation at 365 and 440 nm. Data acquired at
1
3
13
detectable concentrations, suggesting the use of C NMR for
365 nm were used to reference C NMR integrals of the CO
2
1
3
the actinometry. For C NMR detection, we were able to use
paramagnetism to our advantage, capitalizing on very short
relaxation delays typical for paramagnetic NMR to detect high
quality 13C NMR spectra in a matter of minutes. As a proof of
resonance utilizing the known light intensity of the 365 nm
LED source (via o-NBA) and known ferrioxalate quantum yield
4
at 365 nm (see the Supporting Information for details). This
referencing was subsequently used to calculate the I of the 440
0
C
DOI: 10.1021/acs.orglett.8b00391
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
nm LED light source (see the Supporting Information for
details). The results were in good agreement with the previous
measurement conducted using traditional ferrioxalate protocol:
(5) Kirk, A. D.; Namasivayam, C. Anal. Chem. 1983, 55 (14), 2428.
(
(
6) Willett, K. L.; Hites, R. A. J. Chem. Educ. 2000, 77, 900.
7) An important consideration for accurate and reproducible NMR
−1
−1
rate measurement is achieving a steady state (see the Supporting
Information for details).
8
0 and 75 μeinstein L s , respectively.
To summarize, we have demonstrated the viability of direct
(
8) Halperin, S. D.; Kwon, D.; Holmes, M.; Regalado, E. L.;
Campeau, L.-C.; DiRocco, D. A.; Britton, R. Org. Lett. 2015, 17, 5200.
9) Yamase, T.; Takabayashi, N.; Kaji, M. J. Chem. Soc., Dalton Trans.
984, 793.
LED−NMR measurements for facile and accurate quantum
yield determination. We have developed a simple and
convenient NMR actinometry protocol, which has been
validated with two well-studied chemical actinometers, o-
nitrobenzaldehyde and ferrioxalate. We have introduced an
NMR-friendly chemical actinometer, 2,4-dinitrobenzaldehyde,
with a superior range compared to the o-nitrobenzaldehyde,
that enabled NMR actinomtery at 440 nm. The applicability of
this methodology has been demonstrated using two recently
reported photochemical transformations. Finally, we have
developed a direct LED−NMR approach for ferrioxalate
actinometry, which significantly streamlines and simplifies
experimental procedures compared to the classical ferrioxalate
actinometry protocol. We believe that the NMR actinometry
framework developed in this work will become a useful tool in
the field of photochemistry.
(
1
(10) Quantum yield is equal to the number of molecules of product
divided by the number of photons absorbed by the system: ϕ =
(molecules product)/(photons absorbed). Since not all photons are
absorbed productively, the typical quantum yield will be less than 1.
(
11) Douglas, J. J.; Sevrin, M. J.; Stephenson, C. R. J. Org. Process Res.
Dev. 2016, 20, 1134.
12) (a) Sarmah, N.; Bhattacharyya, P. Kr.; Bania, K. K. J. Phys. Chem.
(
A 2014, 118, 3760. (b) Zhang, T.-T.; Jia, J.-F.; Wu, H.-S. J. Phys. Chem.
A 2010, 114, 12251.
(
2
(
13) Kuhn, H. J.; Braslavsky, S. E.; Schmidt, R. Pure Appl. Chem.
004, 76, 2105.
14) Yayla, H. G.; Peng, F.; Mangion, I. K.; McLaughlin, M. M.;
Campeau, L.-C.; Davies, I. W.; DiRocco, D. A.; Knowles, R. R. Chem.
Sci. 2016, 7, 2066.
(15) An LED light source at 440 nm has been found to be relatively
ASSOCIATED CONTENT
Supporting Information
stable over time; therefore, the recalibration of the light intensity
before and after the reaction is recommended but not required.
■
*
S
(16) Relaxation time of nuclei used for quantification is a very
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.8b00391.
important yet very difficult to control parameter. Very short recycling
13
delays used in paramagnetic C NMR experiments allow high
sensitivity but exacerbate the differences in relaxation times if one tries
using external concentration reference. Unfortunately, in the presence
of a paramagnetic agent NMR relaxation times of different molecules
are strongly influenced by a degree of its interaction with such
paramagnetic agent. Matching relaxation times of the concentration
reference and of the analyte was found to be practically impossible.
Experimental procedures, data analysis, and spectro-
scopic data of new compounds (PDF)
AUTHOR INFORMATION
Corresponding Authors
■
*
E-mail: yining.jichen@merck.com.
*E-mail: mikhail_reibarkh@merck.com.
ORCID
Yining Ji: 0000-0002-9650-6844
Cynthia M. Hong: 0000-0002-2563-219X
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors would like to acknowledge Dr. Rebecca T. Ruck,
Dr. R. Thomas Williamson, Dr. Gary E. Martin (MRL, Process
Research and Development), and Professor David McMillan
(Princeton) for very useful discussions.
REFERENCES
■
(
1) (a) Zimmerman, H. E. Pure Appl. Chem. 2006, 78 (12), 2193.
b) Hoffmann, N. Chem. Rev. 2008, 108 (3), 1052. (c) Prier, C. K.;
(
Rankic, D. A.; MacMillan, D. W. Chem. Rev. 2013, 113 (7), 5322.
2) (a) Feldmeier, C.; Bartling, H.; Riedle, E.; Gschwind, R. M. J.
(
Magn. Reson. 2013, 232, 39. (b) Feldmeier, C.; Bartling, H.; Magerl,
K.; Gschwind, R. M. Angew. Chem., Int. Ed. 2015, 54, 1347. (c) Kind,
J.; Kaltschnee, L.; Leyendecker, M.; Thiele, C. M. Chem. Commun.
2
016, 52, 12506. (d) Schultz, D. M.; Levesque, F.; DiRocco, D. A.;
Reibarkh, M.; Ji, Y.; Joyce, L. A.; Dropinski, J. F.; Sheng, H.; Sherry, B.
D.; Davies, I. W. Angew. Chem., Int. Ed. 2017, 56, 15274. (e) Renzi, P.;
Hioe, J.; Gschwind, R. M. J. Am. Chem. Soc. 2017, 139, 6752.
(
3) Cismesia, M. A.; Yoon, T. P. Chem. Sci. 2015, 6, 5426.
(
4) (a) Parker, C. A. Proc. R. Soc. London, Ser. A 1953, 220, 104.
(
b) Hatchard, C. G.; Parker, C. A. Proc. R. Soc. London, Ser. A 1956,
235, 518.
D
DOI: 10.1021/acs.orglett.8b00391
Org. Lett. XXXX, XXX, XXX−XXX