Substituted 2-nitrobenzyltrichloroacetate esters for photodirected oligonucleotide detritylation in solid films

Article abstract of DOI:10.1039/b806902f

Oligonucleotide microarray fabrication by chemical synthesis using photoacid generators in solid films could have advantages over existing methods, but has not matched the accuracy of conventional synthesis where detritylation is performed with acid solut

Full text of DOI:10.1039/b806902f

View Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Substituted 2-nitrobenzyltrichloroacetate esters for photodirected
oligonucleotide detritylation in solid films†
Pawel J. Serafinowski*a and Peter B. Garlandb
Received 23rd April 2008, Accepted 3rd July 2008
First published as an Advance Article on the web 23rd July 2008
DOI: 10.1039/b806902f
Oligonucleotide microarray fabrication by chemical synthesis using photoacid generators in solid films
could have advantages over existing methods, but has not matched the accuracy of conventional
synthesis where detritylation is performed with acid solutions. To address this problem, we explored the
kinetics and equilibria of nucleoside detritylation in solid films, using trichloroacetic acid (TCA)
generated by photolysis from its esters with substituted 2-nitrobenzyl alcohols. We synthesised 25 such
esters, all a-phenyl substituted, and assessed their potential as solid film photoacid generators. They
included sets with (i) mono- or dimethoxy-, (ii) 5-halo-, (iii) alkyl- or aryl-substituted 5-amino-, or (iv)
5-aryl-substituents in the 2-nitro- or 2,6-dinitrobenzyl ring. Absorption maxima of their UV spectra
ranged from 230 to 410 nm, with quantum yields at 365 nm from < 0.01 to nearly 1.0. The esters
formed optically clear solid films on glass slides without added polymer. Kinetics of intrafilm photoacid
generation, proton activity changes and detritylation were measured in situ. The most effective esters
for light sensitivity and detritylation were 5-chloro-, 5-bromo-, 4,5-dimethoxy-, and 4- or
5-aryl-substituted 2,6-dinitrobenzyl esters. Photoacid-induced increases in proton activity and
detritylation were severely inhibited by polymers containing electronegative heteroatoms, but not by
polymers lacking them. In solid films, intrafilm detritylation with photogenerated TCA was fast, but
stopped at an equilibrium well short of completion. Both experiment and theory emphasise the
inadequacy of attempting to force detritylation with high intrafilm acid activity.
solution. This requirement can be met by a system of microwells7
Introduction
and microfluidics.8 Such options for DMT+ removal are unavail-
able if photoacid generation is within a diffusion-restricting solid
film applied to the array surface.9 Although the detritylation
equilibrium position can be shifted further towards completion
by use of stronger photoacids, conflict occurs between the needs
for high stepwise yield and absent depurination. Stepwise yields
in these circumstances have not exceeded 94% in the absence of
unacceptable depurination,9 giving overall yields of ≤ 0.94N for an
N-mer, or ≤ 29% for a 20-mer. Such yields have attracted critical
comment.10
The trityl group is widely used for acid-labile protection of hy-
droxyl groups during organic syntheses.1 It was introduced as the
dimethoxytrityl form (DMT2) for protection of the 5ꢀ-OH group
of nucleoside monomers in oligonucleotide synthesis,3–5 normally
carried out on a solid phase of glass or plastic beads. Removal
of the DMT group prior to each successive round of oligonu-
cleotide chain lengthening is achieved with acid: it is a reversible
reaction and goes to completion if released DMT+ is removed
by the fluid phase. Detritylation can also approach completion
if strong acids are used, but their ability to cause depurination
is unacceptable. High synthetic yields and oligonucleotide purity
require high deprotection yields, ideally 100%. This requirement
is critical when oligonucleotides are synthesised in situ at an array
surface.
Photosensitive protecting groups were introduced in place of
acid-removable DMT groups for fabrication of oligonucleotide
arrays on planar glass surfaces using patterned illumination.6
These had disadvantages of high cost and low light sensitivity, but
nevertheless successfully launched the use of photolithography in
this field.
Nevertheless the use of solid films of photoacid generators for
oligonucleotide array fabrication offers potential advantages of
very high density of array elements and reduced costs. We had
demonstrated11 that TCA photogenerated from its substituted
2-nitrobenzyl ester in solution could achieve >99% stepwise yields
for oligonucleotide synthesis on glass beads, and decided to
explore the possibility of achieving similar yields when TCA is
generated in a solid film.
Our approach as described here was two-fold. We synthesised
and characterised a wider range of PAGs based on TCA esters of
substituted 2-nitrobenzyl alcohols. Their photochemical proper-
ties can be tailored by changes of substituents.12 Other advantages
include the ease with which they form solid films, and the absence
of chemically aggressive photoreaction intermediates. In parallel
we developed UV–vis spectrophotometric methods for directly
measuring within solid films the kinetics of photo-induced TCA
generation, acidification and detritylation. We were then able to
characterise some main factors affecting the equilibrium position
of detritylation within a solid film.
Photolithographically generated acid-patterns must not be
degraded by subsequent diffusion if the acid is photogenerated in
aCancer Research UK Centre for Cancer Therapeutics, Sutton, Surrey, UK
SM2 5NG. E-mail: pawel.serafinowski@icr.ac.uk
bSection of Molecular Carcinogenesis, The Institute of Cancer Research,
Sutton, Surrey, UK SM2 5NG. E-mail: peter@garland.org.uk
† Electronic supplementary information (ESI) available: Experimental. See
DOI: 10.1039/b806902f
3284 | Org. Biomol. Chem., 2008, 6, 3284–3291
This journal is
The Royal Society of Chemistry 2008
©

View Online
Substituted amino groups at the 5-position of the 2-nitrobenzyl
ring gave large increases of height and position of the main
absorption peak (esters 11–19). They could be photolysed at 365
or 405 nm to their nitroso-photoproducts.
Results and discussion
Strategy for synthesis of esters
Fig. 1 shows the general structure for the 25 photosensitive TCA
esters of substituted 2-nitrobenzyl alcohol that we have synthe-
sised. In 4 other related PAGs the acid was p-toluenesulfonic,
acetic, HCl or HBr.
The 5-phenylazo substituted ester 19 had a low absorption
shoulder extending to nearly 500 nm. On illumination at 365 nm
the main peak at 335 nm fell by 40% within a few seconds without
change in its position or shape. The quantum yield for this process
was 0.3, and was probably due to trans–cis isomerism at the
diazo double bond.16 A higher absorption with a maximum at
365 nm and anticipated to be typical of the expected photoproduct,
appeared on prolonged illumination, with a quantum yield of 0.02.
Compounds 20–25 possessed an additional aryl group. Photo-
chemical properties depended strongly on substituents of both
rings (Table 1). Because the bases of oligonucleotides can be
directly damaged by UV light <310 nm, preferred PAGs have
high extinction coefficients at longer wavelengths such as 365 nm
or 405 nm. However, substitutions that increased absorption at
higher wavelengths often lowered the quantum yields, and vice
versa. Esters 2, 15, 16 and 21–24 were the most promising at
365 nm.
Fig. 1 Photosensitive esters and precursors. R = OCOCCl3 or R = OH.
Table 1 and ESI describe R4–R6.† In all cases, R2 is NO2 and R3 is H. The
a-phenyl substituent was present in all esters listed in Table 1 below, except
for 26–29.
The TCA esters were all a-phenyl substituted to improve quan-
tum yields.12 They fall into four families defined by substituents
of the 2-nitro or 2,6-dinitrobenzyl ring. As shown in Table 1,
the families are (i) methoxy-substituted (1–4), (ii) halo-substituted
(5–10), (iii) alkylamino, arylamino, or phenylazo-substituted (11–
19) and (iv) aryl-substituted (20–25). Their general synthetic
pathway started with 5-bromo-, 5-iodo- 5-chloro- or 5-fluoro-2-
nitro-benzylaldehyde or their 2,6-dinitro versions, which are either
available commercially or made as described in the ESI.†
Introduction of the a-phenyl group was achieved by con-
densation of the aldehydes with phenyl magnesium bromide.
Various benzyl ring substitutions were explored as a means of
enhancing quantum yields and light absorption properties. A
second nitro-group at the 6-position enhances quantum yield.13,14
5-Halo-substituted precursors were used in further synthetic
transformations to introduce amino, aryl and phenylazo groups.
We anticipated that these conjugation-extending substituents
would usefully increase extinction coefficients and shift ab-
sorption maxima to higher wavelengths. We also prepared two
2,6-dinitrobenzyl esters with 4-aryl substitution (24, 28). The
precursors (Fig. 1, R = OH) were finally converted into the
corresponding trichloroacetate esters (Fig. 1, R = OCOCCl3) by
reaction with trichloroacetic anhydride.
We also synthesised four esters of other acids. The 4,5-
dimethoxy-2-nitrobenzyl tosylate 26 was stable and has already
been used as a PAG.14 Our attempts to make a-phenyl or a-
methyl substituted versions failed because of spontaneous decay
of the products: the former evolved to 4,5-dimethoxy-2-
nitrosobenzophenone and the latter to a cyclic product, 5,6-
dimethoxy-3-methyl-benzo[c]isoxazol-1(3H)-olate (ESI†). Insta-
bility of 2,6-dinitrobenzyl-p-toluenesulfonate has been described
previously.14 The 2,6-dinitrobenzyl chloride (27), bromide (28),
and acetate (29), like the TCA esters in Table 1, are solids that can
be stored in the dark at −20 ◦C for many months.
2,6-Dinitroesters: which nitro group becomes nitroso?
This question arises for unsymmetrical esters. We answered it by
characterisation of the nitroso-photoproducts of some relevant
esters and showed that the 2-nitro group was the one converted to
a nitroso-group (ESI†).
Photochemical properties in solid films
Solution studies were used to identify potentially useful PAGs.
Most of the esters listed in Table 1 formed optically clear films
when their 2% (w/v) solutions in DCM were allowed to spread
and dry on a flat glass surface. The exceptions were solutions of
the 4- or 5-aryl-substituted esters 20–25, which tended to round up
rather than spread, unless mechanically assisted or in solution with
a polymer. The ester films were not as hard as films of polymers
such as PMMA or PS. Esters 8–10 formed the softest films.
The ester films, although thin, have absorption bands in the
300–420 nm region, allowing measurement of their photolysis,
a reaction that, as in solution, followed first order kinetics. No
significant primary photolytic reactions other than release of the
acid and an equimolar amount of nitroso-photoproduct have been
described. The nitroso-photoproducts have distinctive absorption
spectra in the near UV or higher wavelengths, with maxima at
higher wavelengths than the parent esters. Their absorption was
used to calculate the parallel release of photoacid.
Photochemical properties in solution
The structure, absorption spectra, and quantum yields of
trichloroacetate esters 1–25 in dilute solution (10–50 lM) in DCM
at ambient temperatures are summarised in Table 1. In all cases
photolysis, as measured by time dependent changes in the UV and
near visible absorption spectra during exposure to 365 nm light,
proceeded as a first order reaction with a rate constant dependent
on illumination intensity.
Compounds 1 and 2 are both 4,5-dimethoxy substituted 2-
nitrobenzyl esters: 2 has a 6-nitro substitution.15 They have
reasonable extinction coefficients and quantum yields at 365 nm.
The 5-halo-substituted esters 5–10 have their major absorption
peaks well below 365 nm, but higher quantum yields partially
compensated for lowered light sensitivity at 365 nm.
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 3284–3291 | 3285
©

View Online
Table 1 Photochemical properties of substituted 2-nitrobenzyl esters in solution. Substituents R4–R6 are positioned as in Fig. 1. The esterifying acids
were trichloroacetic for 1–25, p-toluenesulfonic for 26 and acetic for 29. Also included are substituted 2-nitrobenzyl halides 27 and 28, which on photolysis
yield HCl and HBr respectively. Esters 1–25 were a-phenyl substituted
Atom or group at:
R4
emM/mM−1 cm−1 at:
Estera
R5
R6
kmax/nm
kmax
365 nm
405 nm
Ub
111
211
3
OCH3
OCH3
OCH3
OCH3
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
OCH3
OCH3
H
H
F
Cl
Cl
Br
H
NO2
H
NO2
H
H
NO2
H
NO2
H
NO2
NO2
NO2
H
NO2
NO2
NO2
NO2
H
345
323
329
314
232
270
256
276
255
296
376
379
361
403
380
380
370
370
335
454
410
280
324
347
267
334
347
317
298
247
5.3
4.1
4.1
3.2
15.6
6.2
7.5
6.6
8.9
6.5
11.0
12.0
7.6
22.7
14.6
13.1
8,2
6.6
24.0
1.0
7.1
7.6
4.4
4.8
3.5
2.0
nd
0.84
0.4
0.32
0.23
0.31
0.43
0.7
10.4
10.1
7.2
14.6
12.8
12.1
8.1
6.6
15.0
n.a
5.0
0.7
3.0
4.1
0.35
2.3
6.5
0.45
2.2
0.26
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
0.22
6.9
8.1
3.1
22.5
8.2
5.7
5.5
3.7
1.0.
n.a
6.9
<0.1
0.5
1.2
<0.1
1.0
0.42
<0.1
<0.1
<0.1
0.1–0.2
0.4
nd
0.3
0.1
0.2
0.6–0.9
0.1
0.3
0.3
<0.01
<0.01
0.08
<0.02
0.10
0.08
4
5
6
7
8
9
Br
I
(Me)2N
(Et2)N
Morpholin-1-yl
Pyrrolidin-1-yl
Pyrrolidin-1-yl
Piperidin-1-yl
p-Methoxyphenyl-amino
Phenylamino
Phenylazo
10
11
12
13
14
15
16
17
18
19
<0.01
0.05
0.02c
0.3c
20
21
22
23
24
25
26
27d
28
29d
H
H
H
H
p-(Me)2N-phenyl
Phenyl
4-Methoxyphenyl
3,4-Dimethoxyphenyl
H
5-Ethoxynaphthyl
OCH3 (tosylate ester)
H (benzyl chloride)
H (benzyl bromide)
H (acetate ester)
NO2
NO2
NO2
NO2
NO2
NO2
H
NO2
NO2
NO2
0.01
0.5–0.6
0.3
0.1
0.8
<0.01
0.02
0.4
0.3
<0.01
4-Methoxyphenyl
H
OCH3
Br
4-Methoxyphenyl
Br
2.6
4.1
8.5
6.6
2.2
7.0
◦
Notes: Spectral and photochemical values are for solutions in dichloromethane at 20–25 C. a 1 and 2 were referred to previously11 as esters A and D
respectively. Substituted 2-nitrobenzyl halides 27 and 28 are photoacid generators of inorganic acids. b Quantum yields U for photolysis at 365 nm based
on appearance of the absorption band of the substituted 2-nitrosobenzophenones. The values are the mean of at least 3 observations, which varied by
≤20%. c The entry for 19 describes both the main peak (335 nm) and the long wavelength shoulder (445 nm). The higher quantum yield is assumed to
be for cis–trans isomerism, and the lower value for formation of the nitroso-photoproduct. d Use of 2-nitrobenzylhalides to generate hydrohaloacids has
been previously proposed by others.17 Abbreviation “n.a.” is for “not applicable”, “nd” for “not determined”.
Acid-induced increases of intrafilm proton activity were fol-
lowed at 640 nm by inclusion of a pH indicator18 (brilliant
green or BG, pKa in aqueous solution ca. 0.7). To avoid possible
complexities arising from restriction of DMT-nucleotide to a thin
layer attached to the underlying glass surface, we used DMT-T
incorporated in the film as a simple target for detritylation. The
DMT+ carbocation was detected at 510 nm, red-shifted from the
solution maximum3 at 505 nm.
Because of slight peak broadening and red shifts of ca. 5 nm
in the solid state, we did not have precise values for extinction
coefficients for the compounds measured in solid films. So we
used the solution values to calculate approximate quantum yields
in films: in general they were the same or slightly higher than in
solution.
An example of spectra before and after photolysis for a film of
1 containing DMT-T and BG is given in Fig. 2. The absorption
changes due to photolysis of the ester (appearance of a shoulder at
ca. 400 nm due to the nitrosobenzophenone), increased intrafilm
proton activity (fall of the BG peak at 640 nm) and the resulting
detritylation of DMT-T (emergence of the DMT+ peak at 510 nm)
are all apparent. Replacement of ester 1 with 2 resulted in a 90%
fall of the BG peak and ca. 80% conversion of DMT-T to DMT+.
When tested in solid films, photolysis of the 5-alkylamino sub-
stituted esters 15 or 16 failed to increase intrafilm proton activity
to levels detected by BG, nor was DMT+ production observed.
Presumably the heterocyclic nitrogen-containing substituents were
sufficiently basic to bind protons. By contrast 2, 7 and 21–24
were highly effective, 25% photolysis causing almost complete
flattening of the BG peak at 640 nm, and extensive detritylation.
Ester 24 was also effective with 405 nm as the photolysing
wavelength.
Despite the several-fold lower extinction coefficient of 7 at
365 nm when compared to 2, it was no less effective in detritylation
of DMT-T on illumination. It also had the advantage of negligible
interference with DMT+ detection at 510 nm from its nitroso-
photoproduct. Faced with several effective esters, it was necessary
to elect one that would allow us to carry out a wide range of further
experiments, described below, that were primarily concerned with
the performance of TCA in relation to detritylation, rather than
the PAG itself. We chose ester 2.
3286 | Org. Biomol. Chem., 2008, 6, 3284–3291
This journal is
The Royal Society of Chemistry 2008
©

View Online
PAA (4-phenylazoaniline: pKa value ≈ 2.3 in aqueous media).18
The inclusion of PMMA converted intra-film TCA into a weaker
acid, able to protonate PAA but hardly able to protonate BG at all.
This effect of PMMA on the apparent pKa value of TCA can also
be demonstrated in DCM solution simply by titration of the acid
responses of BG and PAA to TCA with or without the polymer.
We concluded that the effect of PMMA on the strength of
photogenerated TCA was responsible for inhibition of detrity-
lation. Carboxylic acids readily form homodimers by hydrogen
bonding,19 and the acidic hydrogen of monomeric TCA could
bond with electronegative atoms of different molecules such as
PMMA, in which case a polymer deficient in electronegative
heteroatoms would have little if any effect on the strength of
TCA. This prediction was fulfilled when PSMS replaced PMMA
(Fig. 4).
Fig. 2 Simultaneous measurement of intrafilm photoacid generation,
photoacid activity and detritylation. Absorption spectra before and after
illumination of a film of 1 containing DMT-T and BG. The casting
solution contained 40 mM ester 1, 0.74 mM BG and 0.7 mM DMT-T
in DCM. Photolysis was at 365 nm (390 mJ cm−2). Under these conditions
detritylation was ca. 40% complete, and ca. 80% if 2 replaced 1.
Effects of including polymers on intrafilm acidification and de-
tritylation. Others have used PMMA9 as a film-forming matrix.
If an experiment as in Fig. 2 was repeated with PMMA present
in the film casting solution at a 1 : 1 or 2 : 1 ratio (w/w) to
either 1 or 2, then although production of TCA as measured by
accumulation of the nitroso-photoproduct was unaffected, very
little rise of intrafilm proton activity (640 nm) or detritylation
(510 nm) was seen.
We explored the ability of PMMA to prevent photogenerated
TCA from increasing intrafilm acidity and effecting detritylation
by carrying out what in effect were intrafilm titrations of proton
activity in response to increasing photoacid concentration. Fig. 3
shows results using a pair of incorporated pH indicators, BG and
Fig. 4 Effect of PMMA or PSMS on acid strength in films. Panel (a)
absorption spectra before and after photolysis (1.1 J cm−2 at 365 nm) of
a film cast from 2% PSMS, 0.2 mM BG, 6.3 mM ester 2 and 0.2 mM
DMT-T. Panel (b) as (a) but PSMS replaced by PMMA.
To provide further evidence for weakening of photogenerated
TCA by hydrogen bonding to electronegative heteroatoms, we
examined a range of film-forming polymers with or without such
atoms. The number of possibilities is large: we considered ca.
3000.20 In addition to PMMA and PSMS already tested (Fig. 3),
we selected a further 41 that were easily dissolved in DCM, formed
suitable films, and covered a range of compositions.
Of the 43 polymers tested for their ability to support intrafilm
acidification and detritylation, eleven deficient in electronegative
heteroatoms supported detritylation and intrafilm acidification to
no less an extent than a film of 2 without polymer. A further
six either lacking heteroatoms, or containing aryl halides or a
trisilane group, gave more moderate support. The remaining 26
were polymers of subunits containing electronegative heteroatoms
(O, N, Cl): they supported acidification and detritylation only
poorly or not at all (ESI†).
Fig. 3 Effects of PMMA on intrafilm pH changes in response to
photogeneration of TCA. The surface area densities for ester 2, BG (pKa =
0.7) and PAA (pKa = 2.3) in the films were 105, 2, and 1.2 nmol cm−2
respectively. Illumination at 365 nm (10 mW cm−2) was continuous, via a
dichroic filter. Absorption spectra from 420 to 750 nm were periodically
recorded during photolysis. Absorption values for the protonated PAA
(510 nm) and non-protonated BG (640 nm) at each time interval were
used to construct the figure. The BG values were normalised to their
pre-photolysis values, and the PAA values to the maximum observed in
the absence of PMMA. The ratio of PMMA to 2 in the films was either
zero or 2 : 1 (w/w).
Hydrogen bonding by photoacid generators. PMMA and other
oxygen-containing polymers weakened intrafilm TCA strength,
whereas compounds 1–10 and 20–25 did not, despite their
high oxygen contents. This difference between oxygen-containing
polymers and PAGs can be rationalised by reference to the detailed
studies of Laurence et al.21 on H-bonding strengths. They point to
the following conclusions: (a) the strongest H-bonding group of
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 3284–3291 | 3287
©

View Online
1–10, 20–25 and 29 is the ester carbonyl, (b) TCA esters have
lowered H-bonding because of the strong electron withdrawing
effect of the trichloromethyl group (c) the 2-nitro group is a modest
H-bonder.
Loss of the H-bonding ester group on photolysis and concomi-
tant conversion of 2-nitro to a 2-nitroso group with possibly
weaker H-bonding may explain the sigmoid relation between
photoacid concentration and extent of detritylation (Fig. 6).
Variations of H-bonding may underlie the varying ability of PAGs,
although lacking nitrogenous substitutions as in 11–19, to effect
the same extent of detritylation for the same extent of photolysis.
Thus 2 is more effective than 1 but less effective than 7.
Fig. 6 Stepwise detritylation. The larger graph has a set of spectra from
a film, 1.5 lm thick, cast from 1% PMS (w/v), 1% ester 2 (6.3 mM) and
DMT-T (0.3 mM) in DCM, and exposed to 13 mW cm−2 at 365 nm for
10 successive exposures of 5 s each followed by 4 at 20 s each and finally 1
at 80 s. A spectrum was recorded 10 s after each exposure: their positions
in the graph are in ascending duration of cumulative exposure. All spectra
were referenced to the pre-photolysis spectrum to give difference spectra
and a flat zero-time baseline. The emergent peak at 384 nm is due to
nitroso-photoproduct accumulation. The smaller graph is a plot of the
area density of DMT+, calculated from the peak at 510 nm, against TCA
area density. The latter was derived from the assumption that the area
density of TCA is equal to that of the substituted 2-nitrosobenzophenone
as calculated from the light-induced absorption peak at 384 nm.
Kinetics of detritylation
Fig. 5 shows detritylation kinetics for a film of polymer and 2 (8 :
1, w/w) containing DMT-T during continuous or interrupted illu-
mination at 365 nm. In the main part of Fig. 5, DMT+ production
rapidly reached a plateau despite continuing photolysis. The brief
lag phase at the onset of illumination is a constant feature of these
experiments. The insert graph shows that detritylation in response
to interrupted illumination stopped promptly within 5–10 s of the
end of each illumination period. Similar behaviour was observed
with other heteroatom-deficient polymers and various ratios of
polymer to ester 2.
2 containing DMT-T resulted in photoacid accumulation and
detritylation. The increments of DMT+ release became smaller
with increasing number of illumination periods, finally becoming
almost undetectable even though further photoacid was generated.
This behaviour is that of a system being repeatedly progressed from
one equilibrium position to the next, towards but never achieving
complete detritylation.
The relationship between the area densities of photoacid and
DMT+ is shown as an insert within Fig. 6. Release of DMT+
commenced after a brief lag-phase, and proceeded with increasing
photoacid generation to a plateau that did not entirely flatten.
Part, at least, of the continuing but small increase at 510 nm was
due to a contribution from the long-wavelength absorption tail
of the nitroso-photoproduct. The extent of detritylation in the
plateau region was 80–90%.
Fig. 5 Kinetics of detritylation. Films cast from a solution of PMS (2%
w/v), ester 2 (0.25% w/v), 0.2 mM DMT-T. Appearance of DMT+ in films
measured in dual wavelength mode at 510–550 nm. The larger recording
(a) is for continuous 365 nm illumination (13 mW cm−2) starting at the
arrow. The smaller panel (b) is for 3 separate illumination periods of 5 s
each, again at 13 mW cm−2 and commencing at the arrows.
Keq as defined in eqn (1) is dimensionless, and area densities
can be used for its calculation in place of concentrations. The
calculated values for several experiments lay between 0.02 and 0.1.
We ignored the lag in DMT+ formation following commencement
of illumination, so our calculated values are underestimates.
Because total acid concentration is used in place of proton activity,
the effect of acid strength is buried within the value of Keq, which
is therefore specific to a given photoacid.
Inhomogeneous distribution of photoacid generator and DMT-
T in films composed of both polymer and photoacid genera-
tor would create photoacid-deficient microdomains of DMT-T
resulting in an equilibrium position that was lowered due to
failure of photoacid to contact all DMT-T molecules, unless
inter-domain diffusion equilibrated concentrations. We have not
observed significant differences between the apparent equilibrium
constants for films of photoacid generator with or without added
polymer. So if microdomains did exist, either their effect was
negated by diffusion or contributed too little to be detected in these
Equilibrium constant for detritylation
An apparent equilibrium constant for acid-induced detritylation
of DMT-T can be defined as:
(1)
where DMT+A− is the ion pair anticipated in an aprotic medium
and DMT-T is 5ꢀ-O-dimethoxytritylthymidine. [HA] is the total
concentration of a protic acid and includes both undissociated
acid and the ion pair H+A−. Keq is dimensionless, and when
all reaction participants are dispersed throughout the same film
volume, concentrations can be replaced with area densities. Use of
[HA] rather than [H+] in eqn (1) makes Keq dependent on the pKa
of the acid.
Data for calculation of Keq were obtained in experiments as
in Fig. 6, where repeated brief illuminations of a film of ester
3288 | Org. Biomol. Chem., 2008, 6, 3284–3291
This journal is
The Royal Society of Chemistry 2008
©

View Online
experiments where the intrafilm diffusion coefficient for TCA was
sufficiently high (50 nm−2 s−1)22 to allow movement over several
hundred nm during the experiment.
We found that PMMA and numerous other polymers con-
taining electronegative heteroatoms weaken TCA at ambient
temperature (Fig. 3 and 4, and ESI†). A similar phenomenon
also attributed to hydrogen bonding has been reported be-
tween dichloroacetic acid and the solvent acetonitrile when used
for detritylation in solid-phase oligonucleotide synthesis.24 We
identified several polymers that are suitable as inert matrices
for photogenerated TCA. Examples are polystyrene, poly(a-
methylstyrene, poly(styrene-co-a-methylstyrene), polylimonene
and poly-(indene-co-coumarone). The last of these contains
oxygen but the content is low, < 2% by wt, and does not have
a significant effect on TCA strength (Supporting Information).
The acidic hydrogen of other acids can also participate in
hydrogen bonds. We observed partial inhibition by PMMA, but
not PSMS, of the ability of photogenerated p-toluenesulfonic acid
to effect intra-film detritylation. Heteroatom-deficient polymers
may therefore be the preferred matrix for other photoacids such
as alkyl- or aryl-sulfonic acids.
Esters of other acids as photoacid generators
Photolysis of 26 releases p-toluenesulfonic acid. The quantum
yield in solution was low, ca. 0.02, rising to 0.06 in films of
PMS or PSMS. The extent of photolysis required to cause
comparable detritylation of intrafilm DMT-T was lower than
that required by photolysis of 2, in keeping with the greater
strength of p-toluenesulfonic acid. Photolysis of 26 in films of
PMMA compared to PSMS was not inhibited, but a three-fold
increase of intrafilm acid concentration was required to obtain
comparable detritylation of intrafilm DMT-T. Thus the effects
of electronegative atoms in a polymer apply to p-toluenesulfonic
acid, although less severely than they do to TCA.
Ester 28 releases HBr on photolysis, and although lacking
an a-phenyl group, it performed similarly to 2 when tested
for detritylation in films of PS. The HCl-releasing 27 was less
effective, suffering from low light absorption at 365 nm. Ester
29 is a photogenerator for acetic acid, which was several-fold less
effective at detritylation than comparable intrafilm concentrations
of TCA. None of these alternative acids appeared to offer any
advantages over TCA.
Not all photoacid generators can themselves form films when
cast from solution but even when they do, as with most of those
described in Table 1, there are advantages in using a polymer
matrix: e.g. control of the concentration of PAG and intrafilm
viscosity. On the other hand a film composed solely of PAG cannot
form PAG-deficient microdomains.
Photoacid performance in solid films. The detritylation target
in our experiments was DMT-T incorporated within diffusion-
restricting solid films. Detritylation is reversible, and its equilibra-
tion following photoacid generation was attained within ca. 10 s.
The apparent Keq value using ester 2 in a PMS film was ca. 0.02–
0.1, and even at intrafilm TCA concentrations of 1.9 M, the extent
of detritylation did not exceed 80–90%.
The relation between photoacid concentration and extent of
detritylation at equilibrium for a given value for Keq and initial
DMT-T concentration can be calculated, with results as in Fig. 7.
The important conclusion is that the approach of detritylation
towards completion is asymptotic, becoming highly insensitive
to increasing acidity once detritylation reaches ca. 96%. Unless
photoacid is at least 50-fold stronger than TCA, the achievable
photoacid level in a film would be unable to achieve >99%
Design of TCA photogenerators based on substituted
2-nitrobenzylalcohols
Electron withdrawing substitutions of the 2-nitrobenzyl ring such
as halides or NO2 groups typically increase the quantum yield and
lower both the height and wavelength of the absorption maximum,
whereas amino groups have the opposite effects. The quantum
yield enhancing effects of a-methyl or a-phenyl substitution are
well known, as is the similar effect of a second nitro-group at the 6-
position. That leaves positions 3–5 for substitutions that may shift
absorption bands from >300 nm to higher wavelengths. Several
useful 4- and 5-substituted 2,6-dinitrobenzyl esters of TCA, all
a-phenyl substituted, were identified (2, 7, 21–24). Esters 24 and
28 were the only 2,6-dinitrobenzyl esters not substituted at either
the 3- or 5- position. Comparison of the isomeric pair 22 (U ≈
0.3) and 24 (U ≈ 0.8) shows that substitution at the 4-position
to create a symmetrically substituted 2-nitrobenzyl ring favours a
high quantum yield.
A further consideration is that neither the ester nor its pho-
toproduct should bind protons at the intrafilm proton activity
required for extensive detritylation. For this reason, the alkyl-
and aryl-amino substituted esters 11–18 were ineffective, despite
excellent absorption properties and an adequate quantum yield
in 15. The remaining ester of this class, 19, had a quantum yield
too low to be useful. But the substituted 2-nitrobenzyl alcohol
precursor of ester 15 might be useful as a starting point for a
protecting group for other applications (e.g. photogenerators of
bases).23
Fig. 7 Effects of photoacid concentration and Keq on extent of detrityla-
tion at equilibrium. Eqn (1) was rearranged to yield [HO-oligonucleotide]
as the variable in a quadratic equation which was then solved for a range
of HA concentrations for given values of initial DMT-T concentration
and Keq. The plots are for calculated % detritylation in response to varied
concentration of HA over a range of Keq values from 0.05 to 2.5. The initial
value for DMT-O-oligonucleotide concentration was 10 mM.
Polymer matrices. Previous reports9 of photoacid generation
in diffusion-restricting solid films as part of the fabrication of
oligonucleotide arrays described mainly PMMA as a film forming
support polymer for the PAG.
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 3284–3291 | 3289
©

View Online
completion of detritylation (Fig. 7, compare Keq = 2.5 with Keq
=
Apparatus for photolysis and spectrophotometry of thin films on
glass slides
0.05).
Experimentally determined detritylation also showed asymp-
totic behaviour as it reached higher levels in response to increasing
photoacid concentrations (Fig. 6).
This laboratory-constructed instrument is described in the ESI.†
Sensitivity varied with the detector, but for the experiments
described here, sensitivity was not an issue because DMT-T was
present throughout the films to give area densities several-fold
higher than those obtainable by its attachment as a monolayer to
the underlying glass.
Conclusions
We have provided effective photogenerators of TCA based on
esters of substituted 2-nitrobenzyl alcohols. They form solid,
uniform and optically clear films without addition of polymers.
The presence of electronegative heteroatoms in polymers when
included in the films renders TCA ineffective as a DMT-T
detritylation agent. Equilibrium of detritylation reactants is
reached within 5–10 s. Both experiments and simulations show that
higher levels of detritylation are approached only asymptotically.
Pushing the reaction towards completion with strong acids or high
concentrations is an unsatisfactory approach. Future emphasis
should be on removal of DMT+ despite the presence of solid
films.
Measurement of proton activity in films
We used a pH indicator in the casting solutions to yield films
incorporating a spectrophotometric measure for intra-film proton
activity. Amongst a variety of pH indicators18 with solution pKa
values around 0.5–1.0, the triphenylmethane cationic indicator
brilliant green was the most suitable. It dissolves in DCM,
where it has a pH-sensitive absorption peak at 630 nm (EmM
=
96 mM−1 cm−1), and relatively low absorption in the green–blue–
UV spectral region. Other indicators with higher pKa values and
suitably placed absorption bands can be used in combination with
BG to extend the response range to include weaker acids. The
absorption spectrum of BG responded predictably to changes of
photoacid acid concentration in solid films.
Experimental
Synthesis of photoacid generators
Measurement of photoacid acid production in films
Esters 1–25 described in the results section table are based
on a-phenyl-2-nitrobenzylalcohol. Full synthetic details and
spectroscopic characterisation of the esters and examples of
nitroso-photoproducts are given in the ESI.† We have previously
described11 the synthesis of esters 1 and 2.
Photolysis of substituted 2-nitrobenzyl esters produces the cor-
responding acid and substituted 2-nitrosobenzophenones. Al-
ternative reaction routes are negligible, and there is a 1 : 1
stoichiometry between acid release and formation of the 2-
nitrosobenzophenone photo-product. The absorption difference
spectrum of after minus before photolysis is due to generation of
the 2-nitrosobenzophenone, as shown in Fig. 6. The area density of
the 2-nitrosobenzophenone photoproduct can be calculated from
the peak height, and is assumed to equal that of photogenerated
acid. For ester 2, the difference spectrum between after and before
photolysis has a peak at 384 nm, with emM = 4.1 mM−1 cm.−1
Thin solid films on glass of esters or esters with polymer
We used glass microscope slides, either intact or cut down to
ca. 25 × 6 mm slices. Film-casting solutions, typically 0.5–2%
(w/v) in DCM, were used in one of two ways: either dip-
coating or spreading. Dip-coating makes more uniform films,
on both sides of a slide. Unwanted film on one side can be
removed with tissue moistened with solvent. Alternatively, the
two films can remain and both are photolysed and measured
together in series, doubling the measured absorption changes.
Another alternative is to first coat one side of the slide with an
anti-wetting agent (Repelcote, a 2% v/v solution of dimethyl-
dichloro-silane in octamethylcyclotetrasiloxane, from Merck Ltd,
Poole, Dorset, UK). The thickness achieved by dip-coating was
varied by changing the concentration of the casting solution, or
varying the speed at which the slide was withdrawn from the
solution.
Spreading used glass diagnostic slides, 75 × 25 × 1 mm (VWR
International, Lutterworth, Leicestershire, UK, or www.vwr.com)
that carry on one face a hydrophobic mask of black Teflon that
exposes one or more circular areas of the glass surface, each
typically 10–15 mm in diameter. These areas form shallow wells,
and 10–15 lL of casting solution placed with a micro-syringe into
one of them on a horizontal slide rapidly spreads over the exposed
glass area and evaporates to l◦eave a circular film. If wished, pre-
heating of the slide to 45–55 C accelerates drying, and protects
the film area against evaporative cooling that could attract water
condensation from the atmosphere.
Detritylation of intrafilm DMT-T
Measurement of the extent of detritylation requires that the value
of pre-photolysis intrafilm DMT-T density is known. To obtain
this value, we first measured in a sample of the film casting
solution, diluted 100–200 fold in 2.5 ml DCM, the value of the
324 nm shoulder due to the presence of ester 2. Addition of 2 M
TCA to obtain a final concentration of 0.05 M gave a 505 nm
peak due to essentially complete conversion of DMT-T to DMT+.
The ratio of the two peak heights was used as a factor to convert
the 324 nm reading in films to the anticipated 510 nm peak if all
DMT-T were to be detritylated. The observed 510 nm reading in
response to photolysis was then used to calculate the extent of
DMT-T detritylation.
Measurement of film thickness
At the onset of this work, we made measurements by interfer-
ometry, using the spectrophotometer with a reflectance probe.
The intensity of reflected light as a function of wavelength was
measured over the range from 400 or 450 nm to 870 nm. Peaks
and troughs were observed, their number increasing with film
3290 | Org. Biomol. Chem., 2008, 6, 3284–3291
This journal is
The Royal Society of Chemistry 2008
©

View Online
thickness. A glass slide without a film was used as the reference.
For calculation of film thickness, we used a model with normal
incidence of the light beam on the film, and interference between
the back reflections from the first and second interfaces, namely
air/film and film/glass. The effects of multiple reflections are
negligible and can be ignored. But otherwise we obtained the
film thickness from the absorption spectra. The area density
(nmol cm−2) of PAG was calculated from the characteristic pre-
photolysis UV-absorption of the PAG, using emM values from
Table 1. The area density of material (lg cm−2) was then calculated
from the PAG density (and the casting solution composition if
polymer was also included), and converted to a film thickness on
the assumption that the film density was 1 g cm−3.
6 S. P. A. Fodor, L. J. Read, M. C. Pirrung, L. Stryer, A. T. Lu and D.
Solas, Science, 1991, 251, 773–776.
7 X. Gao, P. Yu, E. LeProust, L. Sonigo, J. P. Pellois and H. J. Zhang,
J. Am. Chem. Soc., 1998, 12698–12699; X. Gao, E. LeProust, H. Zhang,
O. Srivannavit, E. Gulari, P. Yu, C. Nishiguchi, Q. Xiang and X. Zhou,
Nucleic Acids Res., 2001, 29, 4744–4750.
8 X. Zhou, S. Cai, A. Hong, Q. You, P. Yu, N. Sheng, O. Srivannavit, S.
Muranjan, J. M. Rouillard, Y. Xia, X. Zhang, Q. Xiang, R. Ganesh, Q.
Zhu, A. Matejko, E. Gulari and X. Gao, Nucleic Acids Res., 2004, 32,
5409–5417.
9 G. Wallraff, J. Labadie, P. Brock, R. DiPietro, T. Nyugen, T. Huynh, W.
Hinsberg and G. McGall, CHEMTECH, 1997, 22–32; J. E. Beecher,
G. H. McGall and M. J. Goldberg, Polym. Mater. Sci. Eng., 1997, 76,
597–598; J. E. Beecher, M. J. Goldberg, G. H. McGall, US Pat., 6
083 697, 2000; M. J. Goldberg, R. G. Kulmeilis, G. H. McGall, N. A.
Parker, G. Xu, US Pat. Appl. 0 164 258 A, 2005.
10 A. B. Sierzchala, D. J. Dellinger, J. R. Betley, T. K. Wyrzykiewicz,
C. M. Yamada and M. H. Caruthers, J. Am. Chem. Soc., 2005, 125,
13427–13441.
Acknowledgements
11 P. J. Serafinowski and P. B. Garland, J. Am. Chem. Soc., 2003, 125,
962–966.
This work was supported by The Institute for Cancer Research
and grant no. EGM/16064 provided jointly by the Biotechnology
and Biological Sciences Research Council (BBSRC) and the
Engineering and Physical Sciences Research Council (EPSRC).
Early stages of the work were assisted by an Emeritus Fellowship
to PBG from the Wolfson Foundation. We are indebted to Dr
Amin Mirza for performing MS analyses and to Professor Colin
Cooper for his support and encouragement.
12 F. M. Houhilan, T. X. Neenan, E. Reichmanis, J. M. Kometani and T.
Chin, Chem. Mater., 1991, 3, 462–471.
13 E. Reichmanis, B. C. Smith and R. Gooden, J. Polym. Sci., Polym.
Chem. Ed., 1985, 23, 1–8.
14 T. X. Neenan, F. M. Houlihan, E. Reichmanis, J. M. Kometani,
B. J. Bachman and L. F. Thompson, Macromolecules, 1990, 23, 145–
150.
15 Referred to as esters A and D rather than 1 and 2 in P. J. Serafinowski
and P. B. Garland, J. Am. Chem. Soc., 2003, 125, 962–965.
16 M. Irie, Y. Hirano, S. Hashimoto and K. Hayashi, Macromolecules,
1981, 14, 262–267.
17 H. Barzynski and D. Sanger, Angew. Makromol. Chem., 1981, 93, 131–
References
141.
1 P. J. Kocien´ski, Protecting Groups, Georg Thieme Verlag, Stuttgart, 2nd
edn, 2000, pp. 269–274.
18 F. J. Green, The Sigma-Aldrich Handbook of Stains, Dyes and Indicators,
Aldrich Chemical Company Inc., Milwaukee, USA, 2nd edn, 1991.
19 J. D. Roberts and M. C. Caserio, Basic Principles of Organic Chemistry,
W. A. Benjamin Inc., London, 2nd edn, 1977, pp. 791–793.
20 R. L. Miller, in Polymer Handbook, ed. J. Branderup, E.H. Immergut
and E. A. Grulke, J. Wiley & Sons Inc., Chichester, U.K., 4th edn,
1999, pp. V1/113-VI/124; Aldrich Polymer Products CD-Catalogue
and Reference Guide, 2000, available at www.sigmaaldrich.com.
21 C. Laurence, M. Berthelot, M. Lucon and D. G. Morris, J. Chem. Soc.,
Perkin Trans. 2, 1994, 491–493; C. Laurence, M. Berthelot, M. Herbert
2 Abbreviations: A− and HA, acid anion and acid respec-
tively; BG, brilliant green; DCM, dichloromethane; DMT, 4,4ꢀ-
dimethoxytrityl group; DMT+, 4,4ꢀ-dimethoxytrityl cation; DMT-T,
5ꢀ-O-dimethoxytritylthymidine; ester 1, trichloroacetate ester of a-
phenyl-4,5-dimethoxy-2-nitrobenzyl alcohol; ester 1 photoproduct,
4,5-dimethoxy-2-nitrosobenzophenone; ester 2, trichloroacetate ester
of a-phenyl-4,5-dimethoxy-2,6-dinitrobenzyl alcohol; ester 2 pho-
toproduct, 4,5-dimethoxy-6-nitro-2-nitrosobenzophenone; PAA, 4-
phenylazoaniline; PAG, photoacid generator; PMMA, poly(methyl-
methacrylate); PMS, poly(a-methyl-styrene); PS, polystyrene; PSMS,
poly(styrene-co-a-methyl-styrene); TCA, trichloroacetic acid.
3 M. Smith, D. H. Rammler, I. H. Goldberg and H. G. Khorana, J. Am.
Chem. Soc., 1962, 84, 430–440.
and K. Sraıdi, J. Phys. Chem., 1989, 93, 3799–3802; F. Besseau, C.
Laurence and M. J. Berthelot, J. Chem. Soc., Perkin Trans. 2, 1994,
485–489.
¨
22 P. B. Garland and P. J. Serafinowski, unpublished work.
23 J. F. Cameron and J. F. Frechet, J. Am. Chem. Soc., 1991, 113, 4303–
´
4 S. L. Beaucage and R. P. Iyer, Tetrahedron, 1992, 48, 2223–2311.
5 M. H. Caruthers, Acc. Chem. Res., 1991, 24, 278–284.
4313.
24 C. H. Paul and A. T. Royappa, Nucleic Acids Res., 1996, 24, 3048–3052.
This journal is
The Royal Society of Chemistry 2008
Org. Biomol. Chem., 2008, 6, 3284–3291 | 3291
©