Radical cage effects in the photochemical degradation of polymers: Effect of radical size and mass on the cage recombination efficiency of radical cage pairs generated photochemically from the (CpCH2CH 2N(CH3)C(O)(CH2)nCH 3)2Mo2(CO)6 (n = 3, 8, 18) complexes

Article abstract of DOI:10.1021/ja035030r

This study explored the effect of radical size, chain length, and mass on the cage recombination efficiency of photochemically generated radical cage pairs. Radical cage pairs containing long-chain radicals of the type [(CpCH ₂CH₂N(CH₃)C(O)(CH₂) _nCH₃)(CO)3Mo?, ?Mo(CO)₃(CpCH ₂CH₂(CH₃)NC(O)(CH₂) _nCH₃)] were generated in hexanes/squalane solution by photolysis (λ = 546 nm) of the Mo-Mo bonds in (CpCH₂CH ₂N-(CH₃)C(O)(CH₂)_nCH ₃)₂Mo₂(CO)₆ (n = 3, 8, 18). The cage recombination efficiencies (denoted as F_cP, where F _cP = k_cP/(k_cp + k_dP), k _dP is the diffusion rate constant, and k_cP is the radical recombination rate constant) for the radical cage pairs were obtained by extracting them from quantum yield measurements for the photoreactions with CCl₄ (a metal-radical trap) as a function of solvent system viscosity. The results show that F_cP increases as the length of the chain on a radical center increases. This finding likely provides at least one of the reasons why the quantum yields for photolytic polymer degradation (and long-chain molecules, in general) decrease as the polymer chains get longer. In quantitative terms, plots of k_dP/k_cP were linearly proportional to mass^1/2/radius², in agreement with the prediction of Noyes' cage effect theory. The "radius" of a long-chain radical, such as those studied herein, is rather vague, and for that reason a less ambiguous structural parameter was sought to replace the r² term in the Noyes expression. Plots of k_dP/k_cP vs mass^1/2/surface area suggest that surface area can be used in place of the radius² term in the Noyes expression. The significance of being able to use a particle's surface area in the Noyes expression is that the expression becomes useful for nonspherical particles. The new expression allows the approximate prediction of F_cP values for radicals of different sizes and masses.

Full text of DOI:10.1021/ja035030r

Published on Web 07/31/2003
Radical Cage Effects in the Photochemical Degradation of
Polymers: Effect of Radical Size and Mass on the Cage
Recombination Efficiency of Radical Cage Pairs Generated
Photochemically from the
(CpCH2CH2N(CH3)C(O)(CH2)nCH3)2Mo2(CO)6 (n ) 3, 8, 18)
Complexes
Erick Schutte, T. J. R. Weakley, and David R. Tyler*
Contribution from the Department of Chemistry, UniVersity of Oregon, Eugene, Oregon 97403
Received March 6, 2003; E-mail: dtyler@darkwing.uoregon.edu
Abstract: This study explored the effect of radical size, chain length, and mass on the cage recombination
efficiency of photochemically generated radical cage pairs. Radical cage pairs containing long-chain radicals
of the type [(CpCH2CH2N(CH3)C(O)(CH2)nCH3)(CO)3Mo•, •Mo(CO)3(CpCH2CH2(CH3)NC(O)(CH2)nCH3)] were
generated in hexanes/squalane solution by photolysis (λ ) 546 nm) of the Mo-Mo bonds in (CpCH2CH2N-
(CH3)C(O)(CH2)nCH3)2Mo2(CO)6 (n ) 3, 8, 18). The cage recombination efficiencies (denoted as FcP, where
FcP ) kcP/(kcP + kdP), kdP is the diffusion rate constant, and kcP is the radical recombination rate constant)
for the radical cage pairs were obtained by extracting them from quantum yield measurements for the
photoreactions with CCl4 (a metal-radical trap) as a function of solvent system viscosity. The results show
that FcP increases as the length of the chain on a radical center increases. This finding likely provides at
least one of the reasons why the quantum yields for photolytic polymer degradation (and long-chain
molecules, in general) decrease as the polymer chains get longer. In quantitative terms, plots of kdP/kcP
were linearly proportional to mass1/2/radius2, in agreement with the prediction of Noyes’ cage effect theory.
The “radius” of a long-chain radical, such as those studied herein, is rather vague, and for that reason a
less ambiguous structural parameter was sought to replace the r2 term in the Noyes expression. Plots of
kdP/kcP vs mass1/2/surface area suggest that surface area can be used in place of the radius2 term in the
Noyes expression. The significance of being able to use a particle’s surface area in the Noyes expression
is that the expression becomes useful for nonspherical particles. The new expression allows the approximate
prediction of FcP values for radicals of different sizes and masses.
Introduction
degradable polymers, we have been developing a new class of
polymers that contain metal-metal-bonded organometallic
Photochemically reactive polymers1-3 are of considerable
interest because they are useful as degradable plastics,4
photoresists,5-9 biomedical materials, and as precursors to
ceramic materials.5-7,10-13 To expand the repertoire of photo-
dimers interspersed along the polymer backbone.14-16 These
polymers are photodegradable because the metal-metal bonds
can be cleaved with visible light and the resulting metal radicals
captured with oxygen or other traps (Scheme 1).17
In a recent paper, we showed that the overall quantum yields
for the degradation of these polymers and their model complexes
varied as a function of molecular weight and size.15 Reasons
for the dependence of the quantum yields on the chain length
have been a matter of considerable interest and speculation.18,19
(1) (a) Grassie, N.; Scott, G. Polymer Degradation and Stabilization; Cambridge
University Press: New York, 1985. (b) Guillet, J. Polymer Photophysics
and Photochemistry; Cambridge University Press: Cambridge, U.K., 1985.
(2) Rabek, J. F. Mechanisms of Photophysical Processes and Photochemical
Reactions in Polymers; Wiley: New York, 1987.
(3) Geuskens, G. ComprehensiVe Chemical Kinetics; Bamford, C. H., Tipper,
C. F. H., Eds.; Elsevier: New York, 1975; Vol. 14, pp 333-424.
(4) Guillet, J. E. Degradable Materials; Barenberg, S. A., Brash, J. G., Narayan,
R., Redpath, A. E., Eds.; CRC Press: Boston, MA, 1990; pp 55-97.
(5) West, R.; Maxka, J. Inorganic and Organometallic Polymers; Zeldin, M.,
Wynne, K. J., Allcock, H. R., Eds.; ACS Symposium Series 360; American
Chemical Society: Washington, DC, 1988; Chapter 2.
(6) West, R. Actual. Chim. 1984, 64.
(7) West, R. J. Organomet. Chem. 1986, 300, 327.
(8) Hoger, D. C.; Miller, R. D.; Wilson, C G.; Neureuther, A. R. Proc. SPIE-
Int. Soc. Opt. Eng. 1984, 469, 108.
(9) Ishikawa, M.; Nate, K. Inorganic and Organometallic Polymers; Zeldin,
M., Wynne, K. J., Allcock, H. R., Eds.; ACS Symposium Series 360;
American Chemical Society: Washington, DC, 1988; Chapter 16.
(10) Yajima, S.; Hayashi, J.; Omori, M. Chem. Lett. 1975, 931.
(11) Yajima, S. Ceram. Bull. 1983, 62, 893.
(12) Laine, R. M.; Blum, Y. D.; Tse, D.; Glaser, R. Inorganic and Organo-
metallic Polymers; Zeldin, M., Wynne, K. J., Allcock, H. R., Eds.; ACS
Symposium Series 360; American Chemical Society: Washington, DC,
1988; Chapter 10.
(13) Gilead, D. Degradable Materials; Barenberg, S. A., Brash, J. G., Narayan,
R., Redpath, A. E., Eds.; CRC Press: Boston, MA, 1990; pp 191-207.
(14) (a) Tenhaeff, S. C.; Tyler, D. R. Organometallics 1991, 10, 473-482. (b)
Tenhaeff, S. C.; Tyler, D. R. Organometallics 1991, 10, 1116-1123.
(15) Tenhaeff, S. C.; Tyler, D. R. Organometallics 1992, 11, 1466-1473.
(16) Niekarz, G. F.; Tyler, D. R. Inorg. Chim. Acta 1996, 242, 303-310.
(17) (a) Meyer, T. J.; Caspar, J. V. Chem. ReV. 1985, 85, 187. (b) Geoffroy, G.
L.; Wrighton, M. S. Organometallic Photochemistry; Academic Press: New
York, 1979.
9
10.1021/ja035030r CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 10319-10326
10319

A R T I C L E S
Schutte et al.
Scheme 3. In-Cage Trapping of the Mo-Centered Radical
Scheme 1. Photochemical Degradation of a Polymer with
Metal-Metal Bonds along Its Backbone
Scheme 2. Reaction Scheme for Metal-Metal Bond Photolysis
One hypothesis speculates that the dependence is attributable
to the radical cage effect, specifically to changes in the “cage
recombination efficiency” as the chain length is varied (Scheme
2).20 The “cage recombination efficiency” (denoted as Fc and
colloquially known as the “cage effect”) is defined as the ratio
of the rate constant for cage recombination to the sum of the
rate constants for all competing cage processes. (The Fc value
for a photochemically formed cage pair does not necessarily
equal Fc for the same cage pair formed by thermolysis or by
diffusional collision of two free radicals.21 To differentiate these
cases, the photochemical cage efficiency will be denoted FcP
and the associated rate constants as kcP and kdP. In the photolysis
reaction in Scheme 2, FcP ) kcP/(kcP + kdP).)
to facilitate discussion of their cage pairs. Thus, the radical cage
pair for dimer 1-1 becomes [1•, •1], etc.)
To our initial surprise, the FcP values for the four radical cage
pairs formed by photolysis of these complexes were essentially
identical; i.e., there was no dependence of FcP on the value of
n. To explain this result, we hypothesized that, while still in
the cage, the Mo radicals interacted with an H atom on the side
chain to form a six-membered ring (Scheme 3). In effect, the
metal radicals were captured in the cage. The cage recombina-
tion efficiencies were thus independent of n because the
segmental motion of the Mo radical to form the agostic
interaction25 is independent of the chain length.26
To gain greater insight into the role that FcP plays in con-
trolling the degradation of polymers, we synthesized and studied
the series of model complexes (CpCH2CH2OSiR3)2Mo2(CO)6
(R ) Me, i-Pr, n-Pr, n-Hx) and determined that the cage effect
was indeed dependent on the size and mass of the radical
fragments.22 Specifically, kd/kc was linearly proportional to m1/2
/
r2 (where m is the mass of the radical and r its radius). This
result represented the first experimental verification of Noyes’s
prediction concerning the relationship of particle mass and size
to the cage effect.23
(20) (a) The concept of the “cage effect” was introduced by Frank and
Rabinowitch20b-d in 1934 to explain why the efficiency of I2 photodisso-
ciation was smaller in solution than in the gas phase. It was proposed that
the solvent temporarily encapsulates the reactive I• atoms in a “solvent
cage”, causing them to remain as colliding neighbors before they either
recombine or diffuse apart. This concept is illustrated by the reaction in
Scheme 2. The point to note is that the formation of free radicals is preceded
by the formation of a caged radical pair. (b) Frank, J.; Rabinowitch, E.
Trans. Faraday Soc. 1934, 30, 120-131. (c) Rabinowitch, E.; Wood, W.
C. Trans Faraday Soc. 1936, 32, 1381-1387. (d) Rabinowitch, E. Trans
Faraday Soc. 1937, 33, 1225-1233.
(21) (a) Koenig, T.; Fischer, H. Free Radicals; Kochi, J., Ed.; John Wiley: New
York, 1973; Vol. 1, Chapter 4. (b) Koenig, T. Organic Free Radicals; Pryor,
W. A., Ed.; ACS Symposium Series 69; American Chemical Society:
Washington, DC, 1978; Chapter 9.
To study the generality of this result, we next synthesized
and studied the cage effects in the model complexes
(CH3(CH2)nC(O)NHCH2CH2Cp)2Mo2(CO)6 (n ) 3, 8, 13, 18)
(1-1 to 4-4).24 (The numbers assigned to these dimers are used
(22) (a) Male, J. L.; Lindfors, B. E.; Covert, K. J.; Tyler, D. R. Macromolecules
1997, 30, 6404-6406. (b) Male, J. L.; Lindfors, B. E.; Covert, K. J.; Tyler,
D. R. J. Am. Chem. Soc. 1998, 120, 13176-13186.
(23) (a) Noyes, R. M. J. Chem. Phys. 1954, 22, 1349-1359. (b) Noyes, R. M.
Prog. React. Kinet. 1961, 1, 129-160. (c) Noyes, R. M. J. Am. Chem.
Soc. 1955, 77, 2042-2045. (d) Noyes, R. M. J. Am. Chem. Soc. 1956, 78,
5486-5490.
(18) (a) Guillet, J. Polymer Photophysics and Photochemistry; Cambridge
University Press: Cambridge, U.K., 1985; p 274. (b) The overall quantum
yields decreased until a minimum value was reached, at which point a
further increase in chain length did not lead to a decrease in quantum yield.
At this point, segmental motion presumably dominates molecular motion
rather than center-of-mass diffusion, so an increase in chain length will
have little effect on the efficiency of the reaction.
(24) Male, J. L.; Yoon, M.; Glenn, A. G.; Weakley, T. J. R.; Tyler, D. R.
Macromolecules 1999, 32, 3898-3906.
(25) (a) Brookhart, M.; Green, M. L. H. J. Organomet. Chem. 1983, 250, 395.
(b) Brookhart, M.; Green, M L. H.; Wong, L L. Prog. Inorg. Chem. 1988,
36, 1-124. (c) Yao, W.; Eisenstein, O.; Crabtree, R. H. Inorg. Chim. Acta
1997, 254, 105-111. (d) Crabtree, R. H.; Eisenstein, O.; Sini, G.; Peris,
E. J. Organomet. Chem. 1998, 567, 7-11.
(19) Guillet, J. AdV. Photochem. 1988, 14, 91-133.
9
10320 J. AM. CHEM. SOC. VOL. 125, NO. 34, 2003

Photochemical Degradation of Polymers
A R T I C L E S
Scheme 4. Synthesis of [BrCH2CH2NH2CH3][Br]
Scheme 5. Synthesis of C5H5CH2CH2N(H)CH3
To test the in-cage trapping hypothesis and to further explore
the dependence of FcP on mass and chain length, we synthesized
molecules 6-6 to 8-8 in which methyl groups replace the H
atoms of 1-1 to 4-4. No Mo‚‚‚H agostic interactions are
possible with these new complexes. This paper reports the results
of our study.
Scheme 6. Synthesis of
[(η5-C5H4CH2CH2N(H)2CH3)2Mo2(CO)6][NO3]2
Results and Discussion
Syntheses of 6-6 to 8-8. Complexes 6-6 to 8-8 and their
precursors were synthesized by modifying the synthetic route
previously published for complexes 1-1 to 4-4 (Schemes
4-7).24,27-29 Scheme 4 shows the synthesis of the methyl-
substituted amine ligand that was used in Scheme 5. The amine
was produced on the 100 g scale, in high purity, and with yields
>80%. Reaction of the amine with sodium cyclopentadienide
(Scheme 5) resulted in the substituted cyclopentadienyl ligand
(40%), which was then reacted with Mo(CO)6 (Scheme 6) to
yield the indicated molybdenum dimer in yields over 50%.
Molecules 6-6 to 8-8 were then completed by the route in
Scheme 7. The complexes were purified by dissolving each in
THF and then gravity filtering them through slow filter paper
in the dark in a glovebox.
Scheme 7. Synthesis of Molecules 6-6 to 8-8 from
[(η5-C5H4CH2CH2N(H)2CH3)2Mo2(CO)6][NO3]2
The -CH2CH2- group present between the amide group and
the Cp rings in these molecules is noteworthy. As previously
discussed,30 this spacer isolates the Mo-Mo chromophore from
any electronic changes caused by varying the side chains on
the Cp rings. By doing this, the resulting electronic spectra of
the four complexes were virtually identical (see Experimental
Section), suggesting that the changes in the photophysical
parameters will be caused only by differences in the lengths of
the side chains and not by electronic differences in the Mo-
Mo bond chromophores.
of solvent viscosity. (Solvent viscosity was modified by adding
varying amounts of squalane to the hexanes/CCl4 solvent.32)
hν (546 nm); CCl4
(RCp)2Mo2(CO)6 hexanes/squalane 8
2(RCp)Mo(CO)3Cl + [2CCl3] (1)
Measurement of FcP. FcP was obtained using the method
previously described.31 In brief, the values for FcP were extracted
from quantum yield measurements of reaction 1 as a function
The quantum yields as a function of viscosity were then fit
to eq 2, where c is a fitting parameter that contains kcP, φpair is
the quantum yield for formation of the radical cage pair, and
φx is an offset parameter that is necessary to account for an
(26) While reversibility in the formation of the agostic species is suggested once
the radicals are out of the solvent cage, the process need only be irreversible
on the time scale of the cage processes, i.e., recombination and diffusion
apart, to significantly diminish differences in FcP among the four dimers.
(27) (a) Cortese, F. Organic Syntheses; Wiley: New York, 1943; Collect. Vol.
2, p 91. (b) Martin, S. F.; Puckette, T. A.; Colapret, J. A. J. Org. Chem.
1979, 44, 3391-3396. (c) Fries, K. M.; Joswig, C.; Borch, R. F. J. Med.
Chem. 1995, 38, 2672-2680.
(32) In prior work, it was shown that in a sufficiently high concentration of
CCl4 (>0.1 M) all radicals that escape the solvent cage (“free radicals”)
were captured. Thus, formation of a thermal (collisional) cage pair and
recombination of escaped (free) radicals becomes negligible. Under such
conditions, the observed quantum yield (Φobsd) for the disappearance of
the dimer is given by the following equation: Φobsd ) φpair[kdP/(kcP + kdP)]
) φpair[1 - FcP], where φpair is the quantum yield for the formation of the
cage pair (φpair ) kP/(kP + ΣkR)) and 1 - FcP is the fraction of radical
pairs that escape the cage and subsequently are trapped by CCl4.
Rearrangement of this equation yields the following: 1/Φobsd ) [1/φpair][1
+ kcP/kdP]. This equation shows that kcP/kdP (and in turn FcP) can be
calculated if φpair and Φobsd are known. Because Φobsd can be measured,
the problem of determining FcP is reduced to determining φpair, which can
be found by measuring the quantum yields for the disappearance of the
dimer as a function of viscosity. When the usual assumption is made that
φpair and kcP are independent of viscosity for a particular solvent system,
and therefore that the only viscosity dependence arises from the diffusion
of radicals out of the solvent cage, (kdP), then according to the Stokes-
(28) (a) Hughes, A. K.; Meetsma, A.; Teuben, J. H. Organometallics 1993, 12,
1936-1945. Nieckarz, G. F.; Litty, J. J.; Tyler, D. R. J. Organomet. Chem.
1998, 554, 19-28.
(29) (a) Nieckarz, G. F.; Tyler, D. R. Inorg. Chim. Acta 1996, 242, 303-310.
(30) The -CH2-CH2- spacer can viewed as an electronic insulator that reduces
any electronic variations caused by differences in n, the number of -CH2-
groups. For a similar use of this strategy see: Hughes, R. P.; Trujillo, H.
A. Organometallics 1996, 15, 286-294. Note that each molecule has an
intense band at ≈390 nm in THF (ꢀ ≈ 20 000 cm-1 M-1), assigned to the
σ f σ* transition, and a weaker band at ≈505 nm in THF (ꢀ ≈ 2000 cm-1
M-1), assigned to a dπ f σ* transition. See ref 17a for a more in depth
discussion of the electronic structures of these metal-metal-bonded
molecules.
(31) (a) Braden, D. A.; Parrack, E. E.; Tyler, D. R. Photochem. Photobiol. Sci.
2002, 1, 418-420. (b) Braden, D. A.; Parrack, E. E., Tyler, D. R. Coord.
Chem. ReV. 2001, 21, 279-294.
Einstein-Smoluchowski equation kdP
D
1/η and the previous equation
thus becomes the following: Φobsd ) [φpair/(1 + η/c)] + φx.
9
J. AM. CHEM. SOC. VOL. 125, NO. 34, 2003 10321

A R T I C L E S
Schutte et al.
Figure 2. Plots of FcP vs viscosity for cage pairs [1• •1] to [4• •4] (b), [5•
•
•
•
•5] (O), [6• 6] (1), [7• 7] (3), and [9• 9] (0) at 23 ( 1 °C in hexanes/
squalane. The previous FcP vs viscosity data for complexes 1-1 to 4-4
were so similar that to reduce clutter they have been reduced to a single set
of data points, made up of the statistical averages of the FcP values for
those molecules at the viscosity in question.
Figure 1. Plots of Φobsd vs viscosity for the photochemical reaction (λ )
546 nm) of molecules 1-1 to 4-4 (b), 5-5 (O), 6-6 (1), 7-7 (3), and
8-8 (0) with CCl4 (≈2 M) at 23 ( 1 °C in hexanes/squalane. The previous
Φobsd vs viscosity data for complexes 1-1 to 4-4 were so similar that to
reduce clutter they have been reduced to a single set of data points, made
up of the statistical averages of the Φobsd values for those molecules at the
viscosity in question.
increases. An obvious point is that molecules 6-6 to 8-8
behave differently from molecules 1-1 to 4-4, their H-
containing analogues, in that the quantum yields and FcP values
for molecules 6-6 to 8-8 are not all the same. Because the
quantum yields and FcP values for 6-6 to 8-8 are different, it
is concluded that no substantial “in-cage” trapping occurs with
cage pairs [6•, •6], [7•, •7], and [8•, •8], as was expected from
the replacement of H on the amide group by methyl. These
results also support the “agostic interaction hypothesis” that was
invoked to explain the fact that molecules 1-1 to 4-4 all have
the same FcP values at any selected viscosity.
additional reaction that occurs in these systems.33,34
Φobsd ) [φpair/(1 + η/c)] + φx
(2)
Chain Length Effects. Quantum yields for the photochemical
reactions (λ ) 546 nm) of 6-6 to 8-8 with CCl4 as a function
of solvent system viscosity are shown in Figure 1. Note that, at
any selected viscosity, the quantum yields decrease as the chain
length of the side chain on the Cp ligands increase in length.
These results are in contrast to the quantum yields for the
analogous photochemical reactions of complexes 1-1 to 4-4.
With these latter molecules, the quantum yields at any selected
viscosity were independent of chain length, a result attributed
to the in-cage trapping reaction described in the Introduction.
Quantitative Interpretations. The relationship between FcP
and chain length can be quantified, at least approximately. As
discussed in the Introduction, the ratio kdP/kcP (which is equal
to FcP - 1) is linearly proportional to m1/2/r2 for cage pairs
-1
formed by photolysis of the (CpCH2CH2OSiR3)2Mo2(CO)6
molecules (R ) Me, i-Pr, n-Pr, n-Hx). Figure 3 shows that a
similar relationship holds for the [6•, •6], [7•, •7], and [8•, •8]
cage pairs. Although this is just the second experimental
verification of Noyes’s prediction, the Noyes relationship is
emerging as a rather general expression for predicting (or at
least comparing) FcP values.35 It is important to not overinterpret
the plots in Figure 3. Although Noyes predicted a linear
relationship, the plot in Figure 3 may or may not be linear; the
lack of more data points makes it difficult to say for certain.
However, regardless of whether the relationship to m1/2/r2 is
exactly linear, the important point is that kdP/kcP correlates in at
FcP and φpair values for molecules 6-6 to 8-8 were extracted
from the quantum yield data in Figure 1 using the procedure
described in the preceding section. The resulting values are
shown in Figure 2 and in Table S1 (Supporting Information).
Note that, for any selected viscosity, FcP follows the order 6-6
< 7-7 < 8-8; i.e., FcP increases as the side-chain length
(33) As discussed in a prior paper,31a the φx parameter indicates that an additional
reaction occurs, likely by an electron-transfer mechanism, in systems
containing CCl4. This additional mechanism results in a net reaction of the
dimer starting material, even at very high viscosities where diffusion of
the reactants is severely restricted.
(34) Note that in-cage trapping of the radical pair by CCl4 is not likely. A
comparison of rate constants for the reaction of CpMo(CO)3 with CCl4
(104 M-1 s-1) with that for cage recombination (kc) (g109 s-1) and with
that for diffusional separation (kd) (≈109-1010 s-1) shows that the free
radical trapping reaction cannot compete with the other two processes.
(35) A qualitative interpretation of the mass dependence is that heavier particles
can shove aside solvent molecules more easily than lighter particles; hence,
diffusion is favored for heavier particles. The dependence on r2 is discussed
in the text.
9
10322 J. AM. CHEM. SOC. VOL. 125, NO. 34, 2003

Photochemical Degradation of Polymers
A R T I C L E S
Figure 3. Plot of kcP/kdP vs m1/2/r2 (m ) mass of the radical; r ) the radius
of a sphere with the same volume as the static volume of the radical) for
radical cage pairs formed from molecules 6-6 to 8-8 at the measured
viscosities of 0.376, 0.507, 0.704, 1.03, 1.28, 3.23, 5.95, 7.85, 9.28, and
11.6 cP (top to bottom).
Figure 5. Plot of kcP/kdP vs m1/2/(surface area) for radical cage pairs formed
from molecules 1-1 to 5-5 at the measured viscosities of 0.47, 0.72, 0.90,
2.2, and 3.6 cP (top to bottom). Data are from ref 22b.
least an approximate way with the particles’ mass1/2 and radius2.
By knowing the (at least approximate) relationship between kdP/
kcP and the mass and radius, one can reasonably predict the
relative magnitudes of the cage effects for a series of radical
cage pairs.
The interesting feature about the reasonably good linear fit
between kdP/kcP and m1/2/r2 in Figure 3 is that Noyes derived
his expression for spherical particles yet the radicals used in
this study are decidedly not spherical. The explanation for the
good fit lies in the reason for the dependence of FcP on r2. The
parameter r2 is proportional to the surface area of a particle,
which in turn is proportional to the number of interactions a
particle has with solvent molecules. The more interactions a
particle has with the solvent, the greater the viscous drag and
consequently the slower the rate of diffusion (and hence the
inverse dependence of kdP/kcP on r2). This analysis suggests that
Noyes’s expression can be modified by replacing r2 with the
particle’s surface area, i.e., kdP/kcP m1/2/radical surface area.
To test this suggestion, a plot of kdP/kcP vs m1/2/surface area is
shown in Figure 4. Note the reasonable linearity of the plots at
the various viscosities, a result which suggests that using surface
area instead of r2 is acceptable. As a further test, kdP/kcP was
plotted vs m1/2/surface area for the (CpCH2CH2OSiR3)2Mo2-
(CO)6 molecules (R ) Me, i-Pr, n-Pr, n-Hx; data from ref 22b).
The results are shown in Figure 5; again, note that the plots are
reasonably linear, which provides further support for the
modified Noyes expression. The significance of being able to
use a particle’s surface area in the Noyes expression is that the
expression becomes useful for nonspherical particles; i.e., it
becomes possible to use the Noyes expression for virtually all
Figure 4. Plot of kcP/kdP vs m1/2/(surface area) for radical cage pairs formed
from molecules 6-6 to 8-8 at the measured viscosities of 0.376,
0.507, 0.704, 1.03, 1.28, 3.23, 5.95, 7.85, 9.28, and 11.6 cP (top to
bottom).
9
J. AM. CHEM. SOC. VOL. 125, NO. 34, 2003 10323

A R T I C L E S
Schutte et al.
for the recombination efficiency were determined with an Oriel Merlin
system equipped with an Oriel 200 W high-pressure mercury arc lamp
as previously described.22b,31a The experimental method for obtaining
FcP from the quantum yields as a function of viscosity has been
previously described31a but is repeated in the Supporting Information
for convenience.
radicals and not just spherical ones. Further tests of the modified
Noyes expression are proceeding in our laboratory.
Key Conclusions and Insights. The results in this study
provide a second example showing that FcP increases as the
length of the chain on a radical center increases. This result
provides at least one of the reasons why the quantum yields for
photolytic polymer degradation (and long-chain molecules, in
general) decrease as the polymer chains get longer. Quantita-
tively, the results herein show that the ratio kdP/kcP is ap-
proximately linearly proportional to mass1/2/surface area. This
relationship allows the approximate prediction of FcP values for
radicals of different sizes and masses. Undoubtedly, after a
certain chain length and/or mass is reached, no changes in FcP
will be observed due to the vast inertia of the radicals, which
forces their movements to be local and to not involve movement
of the entire center of mass of the chain. We are currently
working on the synthesis of the n ) 13, 23, and 28 complexes,
with the goal of determining at what chain length the FcP values
become independent of chain length.
Photochemical Reactions of Complexes 6-6, 7-7, and 8-8. The
three sets of 30, 3 mL, mixed-viscosity solutions used to obtain quantum
yields for the reactions of 6-6 to 8-8 were prepared in airtight UV-
visible cells in a darkened glovebox. The desired 2.4 mL ratio of both
hexanes and squalane was first injected into each cell followed by
preparation of a CCl4 stock solution containing the dimer of study, of
which each cell received 0.6 mL. (A new stock solution was made up
for each new set of samples and was concentrated enough that when
each 0.6 mL sample was diluted to 3.0 mL the resultant solution still
gave an absorbance between 1.0 and 1.5 at 504 nm.) Each sample was
then irradiated (λ ) 546 nm) over a period of 20 min during which
time 401 intensity observations were collected, of which the middle
200 observations between 5 and 15 min were used for the determination
of quantum yields.
Synthesis of BrCH2CH2N(H)CH3‚HBr. The following is a more
detailed preparation of a rather vague procedure that has been reported
in the literature multiple times.27 A 250 mL round-bottom flask was
charged with a stirbar and 100 mL (0.883 mol) of HBr (48% w/w).
The solution was cooled to 4 °C in a ice bath, and 35 mL (0.438 mol)
of ice cold 2-(methylamino)ethanol was added dropwise with stirring.
Because of the extreme exothermic nature of this reaction, special care
was taken not to add the amine solution too quickly, reducing the
chances of an explosive boil-over. The resulting mixture was then
attached to a distillation apparatus that had a 2 in, fractionating column
built into the distillation head, and the solution was brought to reflux
by heating in an oil bath. As the bath temperature neared 150 °C, the
head temperature reached 100 °C and distillation of H2O began. The
distillation was allowed to continue slowly (9-11 h) until ≈60 mL of
distillate was collected. During this time, distillation was performed
for ≈1 h, and then the oil bath was turned down to the point that reflux
continued but distillation ceased for ≈30 min. (The distillation was
stopped (but reflux continued) whenever the head temperature rose
above 105 °C or dropped significantly below 100 °C, indicating an
absence of water.) In our hands, the distillation/reflux process was
repeated four times. When the solution no longer produced any distillate
at a head temperature of 100 °C, the bath temperature was increased,
the head temperature rose to ≈123 °C, and distillation of crude HBr
began. Failure to distill off the remaining crude HBr at the conclusion
of the synthesis inhibited the precipitation of the final product upon
addition of cold acetone. The HBr distillation was continued for
approximately 2 h, after which time the solution was allowed to cool.
It is important to note that the appearance of a brown distillate is
indicative of decomposition and the distillation should be stopped
immediately and the solution cooled. After the solution had cooled to
60 °C, it was slowly poured with stirring into a 500 mL beaker
containing 300 mL of ice cold acetone, whereupon the desired white
product precipitated from solution. This heterogeneous solution was
capped and placed in the freezer (-20 °C) overnight. The white
precipitate was then vacuum filtered through a 60 mL medium-porosity
sintered glass funnel, washed three times with 100 mL aliquots of ice
cold acetone, and dried in the funnel, yielding 75.2 g (78%) of crude
product. Reconcentration of the mother liquor followed by addition of
another 200 mL of ice cold acetone yielded another 7.2 g of product
bringing the crude yield to 82.4 g (86%). Further purification was
achieved by using a mortar and pestle to crush the crude product,
followed by rinses with 50 mL aliquots of ice cold acetone until the
filtrate ran clear, yielding 80.3 g (83%). 1H NMR (CDCl3): δ 9.23 (t,
2H, BrCH2CH2N(H)2CH3), 3.83 (t, 2H, BrCH2CH2N(H)2CH3), 3.47 (t,
2H, BrCH2CH2N(H)2CH3), 2.82 (s, 3H, BrCH2CH2N(H)2CH3). An
Experimental Section
All manipulations were carried out in the absence of water and
atmospheric oxygen using standard Schlenk line and glovebox tech-
niques.
Materials and Reagents. Molybdenum hexacarbonyl, 2-(methyl-
amino)ethanol, hydrobromic acid, n-butyllithium in cyclohexane,
acetone (99.5%, spectrophotometric grade), sodium cyclopentadienide
in THF (2.0 M), sodium sulfate, iron(III) nitrate nonahydrate, decanoic
acid, and eicosanoic acid were obtained from Aldrich and used with
further purification. Pentane (98%, Mallinckrodt) and methyl alcohol
(anhydrous, Mallinckrodt) were used without further purification. THF
(Burdick and Jackson) and hexanes (Fisher) were distilled from po-
tassium benzophenone ketyl under nitrogen. Ethyl ether (Mallinckrodt)
was distilled from sodium benzophenone ketyl under nitrogen. Ethyl
chloroformate (Eastman), triethylamine (Merck), valeric acid (Aldrich),
and squalane (Aldrich) were degassed with a nitrogen purge for 2 h
prior to use. CDCl3 and DMSO-d6 (Cambridge Isotope) were distilled
from CaH2 under reduced pressure. CDCl3 and CD3OD (Cambridge
Isotope) were used without further purification. CCl4 (Fisher) was
distilled twice from P2O5 and passed through a column of dry basic
alumina prior to use. All solvents for quantum yield measurements were
further degassed by three freeze-pump-thaw cycles and kept sealed
in storage flasks in a darkened glovebox prior to use. Filter papers
were either Fisherbrand P8 Qualitative or Whatman 5 Qualitative.
Instrumentation and Procedures. Infrared spectra were recorded
on a Nicolet Magna 550 FT-IR spectrometer with OMNIC software.
Samples were prepared as either KBr pellets or as solutions in CaF2
cells (path length 0.110 mm). Electronic absorption spectra were
recorded with a Hewlett-Packard 8453 UV-vis spectrophotometer.
Mass spectral data were collected using both the electrospray and APCI
heads in an Agilent 1100 series LC/MSD. Elemental analyses were
performed by Atlantic Microlab, Inc., Norcross GA. The static
molecular volumes for molecules 6-6 to 8-8 were estimated on the
basis of the results obtained previously22b for 1-1 to 4-4 using the
Steric36 computer program. Maximum spherical volume calculations
and maximum surface area estimates were made possible by creating
minimized model complexes in the Spartan program. From these
models, measurements of the longest molecular axis for the radical
(i.e. the diameter of the sphere) were made and the surface area
calculated using that specific option available in the program. The
quantum yields (λ ) 546 nm) for the formation of the cage pair and
(36) The computer program Steric was written by B. Craig Taverner, Department
of Chemistry, Witwatersrand, Private Bag 3, WITS 2050, Johannesburg,
South Africa.
9
10324 J. AM. CHEM. SOC. VOL. 125, NO. 34, 2003

Photochemical Degradation of Polymers
A R T I C L E S
X-ray crystal structure also confirmed the structural identity of this
molecule (Figure S1; see Supporting Information).
a stirbar. The middle neck of the flask was capped with a water cooled
condenser, which was itself capped with an N2 adapter. The two side-
necks were capped by a tight-fitting rubber septum and another N2
adapter, respectively. The apparatus was then evacuated and filled with
N2. Freshly distilled and degassed THF (150 mL) was then cannulated
into the three-neck flask, and the resulting mixture was stirred. The
amine/n-BuLi/THF solution, which contained a white suspension after
stirring for 1 h, was slowly cannulated into the three-neck flask
containing the Mo(CO)6. During cannulation, the solution changed from
clear to yellow to orange/yellow. The rubber septum was then replaced
by a clear glass stopper under a strong N2 counter-flow, and the resulting
solution was brought to reflux under N2, during which time the color
changed from orange/yellow to dark red with a bright yellow precipitate
evident along the edges of the solution. After 48 h, the solution was
cooled to room temperature and concentrated to ≈100 mL in vacuo.
From this point on, all steps were completed under a red light, with
great care being taken to reduce any and all exposure to white light.
The glass stopper was replaced by another tight-fitting rubber septum,
and a solution of Fe(NO)3‚9H2O (24.5 g in 150 mL of deoxygenated
H2O) was slowly cannulated into the solution with stirring. A dark red
solid precipitated immediately upon addition of the iron nitrate solution,
and the flask was placed in an ice bath overnight. The solid was filtered
in the air through a 60 mL sintered glass funnel and washed with three
100 mL aliquots of ice cold, deoxygenated distilled water, followed
by two 25 mL aliquots of ice cold deoxygenated methanol and one
100 mL aliquot of ice cold deoxygenated pentane. The product was
placed in a 250 mL Schlenk flask and dried in vacuo for 24 h. Any
excess Mo(CO)6 was then removed by sublimation onto a liquid-
nitrogen-cooled coldfinger, yielding 11.3 g (51%) of crude product.
1H NMR (DMSO-d6) showed the product to be a mixture of at least
two rotamers:38 δ 8.39 (s br, 4H, C5H4CH2CH2N(H)2CH3), 5.71 (br s,
4H, C5H4CH2CH2N(H)2CH3), 5.63 (br s, 8H, C5H4CH2CH2N(H)2CH3),
5.31 (br s, 4H, C5H4CH2CH2N(H)2CH3), 3.03 (m, 4H, C5H4CH2CH2-
N(H)2CH3), 2.71 (m, 4H, C5H4CH2CH2N(H)2CH3), 2.59 (s, 6H, C5H4-
CH2CH2N(H)2CH3). UV-vis (CH3OH; λmax, nm): 504, 389. MS
(electrospray): m/z 606 (P - 2(NO3)). IR (KBr, cm-1): ν(CO) 2011
(w), 1954 (s), 1914 (s), 1881 (s); ν(NO3-) 1384 (s). Anal. Calcd for
Mo2C22H26N4O12: C, 36.18; H, 3.58. Found: C, 35.99; H, 3.57.
Synthesis of C5H5CH2CH2N(H)CH3. A 500 mL, dry Schlenk flask
was charged with a stirbar and 17.5 g (0.080 mol) of BrCH2CH2N-
(H)CH3‚HBr. The flask was sealed with a rubber septum, and 250 mL
of freshly distilled, deoxygenated THF was added by cannula. The white
heterogeneous solution was cooled to 4 °C in an ice bath, and sodium
cyclopentadienide (80.0 mL, 0.160 mol) was added dropwise through
a syringe. The milky, purple solution was covered to minimize exposure
to light and allowed to slowly come to room temperature. After the
solution was stirred for 48 h, it was noted that the solution had retained
its milky purple appearance. (A change to brown or orange at this point
in the reaction indicated significant decomposition, most likely from
the presence of oxygen.) Deoxygenated distilled water (≈100 mL) was
then cannulated into the solution until it became homogeneous. The
resultant mixture was then cannulated into a 1 L separatory funnel,
which had been capped with a rubber septum and purged with N2 prior
to use. The aqueous layer was removed and freshly distilled ethyl ether
(200 mL) was added by cannula. Deoxygenated distilled water (100
mL) was then cannulated into the solution, and the combined mixture
was shaken vigorously with venting for 1 min. The water layer was
then removed and the process repeated with another 100 mL aliquot
of water. After the water layer was removed a second time, the
separatory funnel held the organic layer and a insoluble layer of dark
brown solids. These solids were also drained off, leaving behind the
clear, green/orange organic layer. This solution was cannulated into a
500 mL round-bottom flask, which had been charged with sodium
sulfate (≈30 g). The heterogeneous mixture was shaken lightly to
promote the drying process and left to sit in the dark for 1 h, after
which time the organic mixture was gravity filtered in the air through
fast filter paper into a 1 L round-bottom flask. The solvent was removed
by rotary evaporation, leaving a viscous orange/brown oil. The oil was
transferred via pipet into a 100 mL round-bottom flask that had been
charged with a stirbar, and the desired product, a clear, colorless oil,
was obtained by distillation at 35-40 °C under reduced pressure,
yielding 3.94 g (0.032 mol, 40%). Because of the instability of the
product, (a yellow tint is indicative of decomposition), it is suggested
that it be used immediately or stored under N2 in the dark at -20°C.
1H NMR (CDCl3) showed the purified product to be a 1:1 mixture of
two ring isomers with the alkyl group attached to the cyclopentadiene
ring in the 1 and 2 positions:37 δ 6.41 (m, 3H, 2 × CH of C5H5 ring
isomer 1, 1 × CH of C5H5), 6.26 (m, 1H, CH of C5H5 ring isomer 2),
6.23 (t, 1H, CH of C5H5 ring isomer 2), 6.06 (t, 1H, CH of C5H5 ring
isomer 1), 2.95 (d, 2H, CH2 of C5H5 ring isomer 1), 2.88 (d, 2H, CH2
of C5H5 ring isomer 2), 2.75 (q, 2H, C5H5CH2CH2N(H)CH3), 2.57 (m,
2H, C5H5CH2CH2N(H)CH3), 2.41 (s, 3H, C5H5CH2CH2N(H)CH3), 1.83
(b s, 2H, 2 × C5H5CH2CH2N(H)CH3).
Synthesis of [η5-C5H4CH2CH2N(CH3)CO(CH2)nCH3]2Mo2(CO)6
(6-6, 7-7, 8-8). These complexes were synthesized from [η5-
C5H4CH2CH2N(H)2CH3]2Mo2(CO)6[NO3]2 under red light using the
procedure previously described for complexes 1-1 to 4-4.24 Purifica-
tion was carried out in the glovebox by dissolving the crude product
in 20-25 mL of freshly distilled freeze-pump-thawed THF and then
gravity filtering the solution through slow filter paper. The resulting
filtrate was concentrated in vacuo (≈5 mL), layered with hexanes (≈100
mL), capped with a glass stopper, and placed in the freezer overnight
to facilitate precipitation of product. The complexes were vacuum
filtered through 15 mL medium sintered glass funnels in the glovebox
and dried under vacuum for 12 h. The precipitated complexes contained
variable amounts of water molecules that could not be removed by
Synthesis of [(η5-C5H4CH2CH2N(H)2CH3)2Mo2(CO)6][NO3]2. The
[η5-C5H4CH2CH2N(H)2CH3]2Mo2(CO)6[NO3]2 precursor complex was
synthesized with several key modifications to the previously reported
procedure.24 Freshly distilled THF (50 mL) was combined with
C5H5CH2CH2N(H)CH3 (7.5 g, 60.4 mmol) in a 100 mL Schlenk flask,
which had been charged with a stirbar. The flask was capped with a
tight-fitting rubber septum and degassed by two freeze-pump-thaw
cycles. While the reaction mixture was still very cold (still in the liquid
state), it was placed in a dry ice/acetone bath (-77 °C) and 6 mL of
n-butyllithium solution (10 M in hexanes, 60.6 mmol) was added
dropwise via syringe with stirring. The resulting solution was allowed
to warm to room temperature and then stirred under N2 for 1 h. In
previous syntheses, the Mo(CO)6 was added to the ligand/THF/n-
butyllithium solution through a sidearm under a counter-flow of N2.
To better maintain an air-free environment, the Mo(CO)6 (15.9 g, 60.4
mmol) was placed in a separate 500 mL three-neck flask charged with
1
further drying in vacuo. H NMR spectra showed all three purified
products to be mixtures of at least two rotamers.38
[η5-C5H4CH2CH2N(CH3)CO(CH2)3CH3]2Mo2(CO)6. Yield: 45%.
IR (KBr; cm-1): ν(CO) 2007 (w), 1947 (br, s), 1904 (sh), 1890 (s),
1
1870 (sh); ν(N(CdO)CH3-): 1640 (m). H NMR (CDCl3): δ 5.25
(m, 8H, C5H4CH2CH2N(CH3)-), 3.49 (m, 4H, C5H4CH2CH2N-
(CH3)-), 2.95 (m, 6H, C5H4CH2CH2N(CH3)-), 2.64 (m, 4H, C5H4CH2-
CH2N(CH3)-), 2.32 (t, 4H, C5H4CH2CH2N(CH3)COCH2CH2CH2CH3),
2.19 (t, 4H, C5H4CH2CH2N(CH3)COCH2CH2CH2CH3), 1.61 (m, 4H,
C5H4CH2CH2N(CH3)COCH2CH2CH2CH3), 1.33 (m, 4H, C5H4CH2CH2N-
(CH3)COCH2CH2CH2CH3), 0.96 (m, 4H, C5H4CH2CH2N(CH3)COCH2-
(37) (a) Kaul, B. B.; Noll, S.; Renshaw, S.; Rakowski Dubois, M. Organo-
metallics 1997, 16, 1604-1611. (b) Hughes, A. K.; Meetsma, A.; Teuben,
J. H. Organometallics 1993, 12, 1936.
(38) (a) Adams, R. D.; Collins, D. M.; Cotton, F. A. Inorg. Chem. 1974, 13,
1086-1090. (b) Adams, R. D.; Cotton, F. A. Inorg Chim. Acta 1973, 7,
153-156.
9
J. AM. CHEM. SOC. VOL. 125, NO. 34, 2003 10325

A R T I C L E S
Schutte et al.
CH2CH2CH3). Anal. Calcd for Mo2C32H40N2O8‚1/2H2O: C, 49.18; H,
5.29; N, 3.58. Found: C, 49.00; H, 5.15; N, 3.59. MS (APCI): m/z
773 (P+).
1.29 (m, 4H, C5H4CH2CH2N(CH3)COCH2CH2CH2CH3), 0.91 (m, 4H,
C5H4CH2CH2N(CH3)COCH2CH2CH2CH3). IR (KBr; cm-1): ν(CO)
2008 (w), 1947 (s), 1914 (s), 1906 (sh), 1895 (s), 1888 (sh), 1870 (sh);
ν(N(CdO)CH3-) 1634 (m). Anal. Calcd for Mo2C62H100N2O8‚
2.2H2O: C, 60.39; H, 8.53; N, 2.27. Found: C, 59.95; H, 7.88; N,
2.46. MS (APCI): m/z 1138 (P - 2(CO)).
[η5-C5H4CH2CH2N(CH3)CO(CH2)8CH3]2Mo2(CO)6. Yield: 38%.
1H NMR (CDCl3): δ 5.21 (m, 8H, C5H4CH2CH2N(CH3)-), 3.46 (m,
4H, C5H4CH2CH2N(CH3)-), 2.95 (m, 6H, C5H4CH2CH2N(CH3)-),
2.62 (m, 4H, C5H4CH2CH2N(CH3)-), 2.28 (t, 4H, C5H4CH2CH2N-
(CH3)COCH2CH2CH2CH3), 2.16 (t, 4H, C5H4CH2CH2N(CH3)COCH2-
CH2CH2CH3), 1.58 (m, 4H, C5H4CH2CH2N(CH3)COCH2CH2CH2CH3),
1.27 (m, 4H, C5H4CH2CH2N(CH3)COCH2CH2CH2CH3), 0.89 (m, 4H,
C5H4CH2CH2N(CH3)COCH2CH2CH2CH3). IR (KBr; cm-1): ν(CO)
2007 (w), 1946 (s), 1916 (s), 1903 (sh), 1894 (s), 1870 (sh); ν(N(CdO)-
CH3-) 1634 (m). Anal. Calcd for Mo2C42H60N2O8‚1/2H2O: C, 54.72;
H, 6.67; N, 3.04. Found: C, 54.66; H, 6.55; N, 3.16. MS (APCI): m/z
858 (P - 2(CO)).
Acknowledgment. The National Science Foundation is
acknowledged for the support of this work.
Supporting Information Available: Tables of selected mass
and photochemical data, mass and volume data for the radicals
and solvents, crystallographic information for the BrCH2CH2N-
(H)CH3‚HBr molecule, atomic coordinates and equivalent
isotropic thermal parameters, bond lengths and angles, and
anisotropic thermal parameters, an ORTEP drawing of the
BrCH2CH2N(H)CH3‚HBr molecule, and a description of the
crystal structure analysis. This material is available free of
charge via the Internet at http://pubs.acs.org.
[η5-C5H4CH2CH2N(CH3)CO(CH2)18CH3]2Mo2(CO)6. Yield: 47%.
1H NMR (CDCl3): δ 5.24 (m, 8H, C5H4CH2CH2N(CH3)-), 3.49 (m,
4H, C5H4CH2CH2N(CH3)-), 2.95 (m, 6H, C5H4CH2CH2N(CH3)-),
2.64 (m, 4H, C5H4CH2CH2N(CH3)-) 2.31 (t, 4H, C5H4CH2CH2N-
(CH3)COCH2CH2CH2CH3), 2.17 (t, 4H, C5H4CH2CH2N(CH3)COCH2-
CH2CH2CH3), 1.62 (m, 4H, C5H4CH2CH2N(CH3)COCH2CH2CH2CH3),
JA035030R
9
10326 J. AM. CHEM. SOC. VOL. 125, NO. 34, 2003