High yield detritylation of surface-attached nucleosides with photoacid generated in an overlying solid film: Roles of translational diffusion and scavenging

Article abstract of DOI:10.1039/b813319k

Conventional solid-phase oligonucleotide synthesis overcomes the reversibility of acid-dependent detritylation by washing away the released dimethoxytrityl cations (DMT⁺) with acid. This option is unavailable if the acid is photogenerated in an overlying solid film, as in the photolithographic fabrication of oligonucleotide arrays on planar surfaces. To overcome the resulting reversibility problem we developed methods of achieving ≥98% detritylation of glass-attached 5′-O-DMT-thymidine, a model for 5′-O-DMT-protected oligonucleotides, by the photogeneration of trichloroacetic acid in a solid film. Enhanced intrafilm diffusion, insufficient to degrade the photolithographic resolution but enabling DMT⁺ to move from its plane of release into the overlying photoacid-generating film, increased detritylation from ≤30% to ≥98%. Inclusion of an intrafilm carbocation scavenger such as a triarylsilane hydride converted the detritylation into a time-dependent irreversible process proceeding to ≥99% detritylation within 60 s following brief photoacid generation. Light sensitivity is high, exceeding direct photodeprotection methods by 15-100 fold. The Royal Society of Chemistry 2009.

Full text of DOI:10.1039/b813319k

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
High yield detritylation of surface-attached nucleosides with photoacid
generated in an overlying solid film: roles of translational diffusion and
scavenging†
Peter B. Garland*a and Pawel J. Serafinowskib
Received 31st July 2008, Accepted 28th October 2008
First published as an Advance Article on the web 28th November 2008
DOI: 10.1039/b813319k
Conventional solid-phase oligonucleotide synthesis overcomes the reversibility of acid-dependent
detritylation by washing away the released dimethoxytrityl cations (DMT+) with acid. This option is
unavailable if the acid is photogenerated in an overlying solid film, as in the photolithographic
fabrication of oligonucleotide arrays on planar surfaces. To overcome the resulting reversibility problem
we developed methods of achieving ≥98% detritylation of glass-attached 5¢-O-DMT-thymidine, a model
for 5¢-O-DMT-protected oligonucleotides, by the photogeneration of trichloroacetic acid in a solid film.
Enhanced intrafilm diffusion, insufficient to degrade the photolithographic resolution but enabling
DMT+ to move from its plane of release into the overlying photoacid-generating film, increased
detritylation from ≤30% to ≥98%. Inclusion of an intrafilm carbocation scavenger such as a
triarylsilane hydride converted the detritylation into a time-dependent irreversible process proceeding
to ≥99% detritylation within 60 s following brief photoacid generation. Light sensitivity is high,
exceeding direct photodeprotection methods by 15–100 fold.
Scheme 1. It depends on the removal of DMT+, either by diffusion
or scavenging.
Introduction
The early fabrication1,2 of oligonucleotide microarrays by synthesis
in situ was on planar glass surfaces, using the same chemistry as
for much larger scale conventional solid phase oligonucleotide
synthesis3 on glass or plastic beads. Physical barriers directed
critical reagent solutions to specified array elements. Array density
was low, with individual elements of mm dimensions. Much higher
array densities were obtained by the use of a photocleavable
blocking group in place of DMT,4 and exposure of the array
surface to patterned illumination to define the array elements and
their base sequences.5
An alternative photolithographic approach retained DMT as
the protecting group, and used patterned illumination of a solid
polymer film containing a photoacid generator applied to the array
surface for the detritylation step.6,7 The film prevents diffusion of
the photoacid from illuminated to non-illuminated array elements,
but also prevents the diffusion of DMT+ away from its site
of formation. Because acid induced detritylation is reversible it
Scheme 1 Vertical section through a photoacid generating film and
its solid support. Shading of the film region adjacent to the support
surface represents the layer of attached oligonucleotides, a thin virtual
compartment communicating by diffusion with the rest of the film.
Numbered processes 1 and 4 are irreversible, whereas 2 and 3 are reversible,
reaching equilibrium in a few s in the case of 2 but very much longer with
3, depending on intrafilm viscosity and film thickness.
proceeds no further than the equilibrium position with a given
acid activity.8 The need for >98% completion is not met unless the
photoacid is strong, but this causes depurination.9
We set out to resolve, in solid films, the conflict between the
need for virtually complete detritylation and the avoidance of
photoacid-induced depurination. Our solution is described in
aSection of Molecular Carcinogenesis, The Institute of Cancer Research,
Sutton, Surrey, UK SM2 5NG. E-mail: peter@garland.org.uk
Previous estimates10 of detritylation efficiency for photoacids in
solid films depended on the synthetic yield of oligonucleotides: in
its place we have used in situ spectrophotometric measurements of
the intrafilm photoacid-induced release of DMT+ from DMT-T
covalently attached by 3¢-O linkage to a planar glass surface. The
area densities of surface-attached DMT-T were high, 5–15 fold
higher than the 40 pmol cm-2 normally used for oligonucleotides
bCancer Research UK Centre for Cancer Therapeutics, The Institute of Can-
cer Research, Sutton, Surrey, UK SM2 5NG. E-mail: Pawel.serafinowski@
icr.ac.uk
† Electronic supplementary information (ESI) available: Further infor-
mation on experimental; detritylation equilibrium equations; computer
simulation details; kinetics of DMT+ scavenging; diffusion and film
composition. See DOI: 10.1039/b813319k
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 451–459 | 451
©

View Article Online
in arrays.1 These high densities were chosen to improve the
accuracy of detritylation determinations. Their drawback was a
lower detritylation extent at equilibrium with photoacid.
are considerable: a 20-fold reduction from 1.0 to 0.05 can increase
detritylation from <80% to >98%.
DMT+ diffusion. A film overlaying a layer of oligonucleotides
covalently attached to a planar support surface can be regarded as
two compartments: the film, typically several hundred nm thick,
and the surface-attached oligonucleotide layer which has a depth
of only 10–20 nm or less depending on chain length. Diffusion of
DMT+A- from the oligonucleotide layer into the overlying film is
therefore a possible means of enhancing detritylation.
One-dimensional molecular diffusion can be simulated by using
equidistant jumps d in length occurring at constant frequency.
The Einstein-Smoluchowski equation12 applies:
Results
Computer modeling (ESI†)
Detritylation equilibrium. The ratio of tritylated to non-
tritylated nucleotide concentrations in equilibrium with an acid
is given by:
+
-
È
˘
DMT A
DMT-O - oligonucleotide
[
]
=
Î
˚
(1)
HO-oligonucleotide
HA K
[
]
[
]
eq
< d2 > = 2Dt
(2)
where Keq is the apparent equilibrium constant, DMT+A- is the
ion pair of DMT+ carbocation and acid anion, and HA is the
total acid which is assumed to be in many-fold excess over the
other reactants. Concentrations are used in place of activities.
Detritylation approaches completion when the left hand term
in equation (1) approaches zero. This relationship gives three
independent options for increasing the extent of detritylation at
equilibrium: increase the strength of the acid; replace DMT with
a more acid-sensitive group; remove DMT+ from its site of origin.
The first of these three options is limited: Keq as defined by
equation (1) increases with acid strength, but so does depurination
risk. The second suffers from loss of a reagent used in many
well characterized and commercially available compounds; the
required increase in acid-sensitivity might also be accompanied
by instability.11 The third option, removing DMT+, has two
possibilities for realisation: (a) provide for intrafilm diffusion of
DMT+ away from its site of formation on the surface of the array
support, and (b) include within the film a scavenging agent for
DMT+.
where < d2 > is the mean square of the distance diffused, D the
diffusion coefficient and t the jump period.
Fig. 2 shows results from simulating the effects of diffusion
on the intrafilm profile of DMT+ concentration along the z-axis.
DMT+ is released into the oligonucleotide layer at the underlying
support surface, and its diffusion into the overlying film generates
new concentration profiles. The calculated concentration of DMT+
in the thin surface zone of attached oligonucleotide fell by over
10-fold after 104 diffusional jumps. This fall would significantly
increase detritylation (Fig. 1).
Fig. 1 shows the calculated effect on the detritylation equilib-
rium of reducing the activity of DMT+ by 1–50 fold. The effects
Fig. 2 Calculated changes of intrafilm concentration profiles of DMT+
in the z-axis due to diffusion from an immobile 10 nm thick layer of
5¢-O-DMT-oligonucleotide following photoacid generation. At zero time
all DMT+ is within the oligonucleotide layer. Initial values used for calcula-
tions were Keq = 0.1, [HA] = 250 mM, and [5¢-O-DMT-oligonucleotide] =
10 mM (corresponding to 100 pmol cm-2 before photoacid generation).
Detritylation after 0, 101, 102, 103 and 104 diffusional jumps was 76, 79,
83, 92 and 97% respectively.
But enhancement of diffusion to achieve greater detritylation
raises a question: would comparable rates of photoacid diffusion
in the plane of the film cause loss of photolithographic resolution?
Diffusion constants are only weakly dependent on molecular
weight. We carried out simulations of TCA diffusion across the
boundary between illuminated and non-illuminated areas, again
using a repetitive jump model. After 104 jumps of 1 nm significant
concentrations of photoacid did not extend more than 200–250 nm
beyond the light/dark boundary (ESI†), a distance unlikely to
compromise resolution of 10 ¥ 10 mm array elements.
Fig. 1 Calculated effects of DMT+ removal on the extent of detritylation
at equilibrium. The number by each curve is an activity coefficient applied
to [DMT+A-] in equation (1). The value for Keq was 0.05, and the
initial concentration of surface-attached DMT-groups of 25 mM was
calculated from a surface density of 250 pmol cm-2 occupying a thickness
of 10 nm. The value used for Keq is from measurements of detritylation by
photogenerated TCA for DMT-T distributed throughout a solid film of
photoacid generator.8
452 | Org. Biomol. Chem., 2009, 7, 451–459
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
Intrafilm diffusion and film composition
Surface loss of TCA. Both increased intrafilm acidity (sensed
by BG, a pH sensitive dye) and DMT+ formation following
photoacid generation can be measured spectrophotometrically
in situ.8 In thin films (<100 nm) of ester 2 as the film-forming
component these absorbance changes slowly reverse with time
(Fig. 3). Reversal was not seen if a cover slip overlaid the film
surface, or the if the film thickness was a few-fold greater.
Fig. 4 Patterned photo-acidification of a film of 2. The film was made
by rapidly dip-coating a 25 ¥ 6 mm piece of glass slide with a solution
of 2% (w/v) ester 2 (ca. 40 mM) and 5 mM BG and allowing it to dry.
Film was removed from one surface with solvent. The remaining film
was illuminated (365 nm, 20 s, 40 mW cm-2) through a contacted Ronchi
ruling of 50 line pairs mm-1, then examined by transmitted white light
with a digital microscope. The white and green bands, 10 mm wide, are the
exposed and non-exposed areas respectively, corresponding to protonated
and non-protonated BG.
Detritylation of surface-attached DMT-T
We used 5¢-O-dimethoxytritylthymidine as a model for acid-
induced oligonucleotide detritylation. Rates and equilibria of de-
tritylation may vary between nucleosides, nucleotides or oligonu-
cleotides. But the reaction itself, in the solid environment of
a diffusion-restricting photoacid-generating film, is essentially
unchanged, as is the task of achieving ≥98% completion of
detritylation.
Fig. 3 Time-dependent absorption changes following photolysis (15 s
at 32 mW cm-2) of a thin (80–90 nm) film of ester 2 (18 nmol cm-2)
containing DMT-T (0.2 nm cm-2) and BG (0.2 nmol cm-2), measured at
510 nm (DMT+) and 640 nm (BG). Absorbance values are normalized to
those immediately prior to photolysis for BG, and 10 s after photolysis for
DMT+.
Extent. Using glass slides with surface-attached DMT-T (120–
700 pmol cm-2) and coated with thin films (0.3–1.5 mm) of ester
2 without added polymers, we measured detritylation as shown
by the formation of DMT+ in response to photolytic illumination
(typically 20–25 mW cm-2 for 30 s).16 To determine the amount of
DMT-T remaining on the glass the exposed film was removed with
solvent, replaced with a fresh film and photolysed as before. Under
these conditions little if any DMT+ was detected after photolysis
of the second film. We concluded that the first photolysis had
achieved ≥98% detritylation. These high levels of detritylation
of surface-attached DMT-T were independent of the attachment
chemistry.
Solid TCA vaporizes, and its continued loss from a film is
dependent on intrafilm diffusion to reach the air/film interface
The intrafilm concentration of DMT+ is not a linear representation
of intrafilm TCA concentration, but its time-dependent decay
as seen in Fig. 3 may be compared with computer simulated
data (ESI†) to obtain an approximate value of ca. 50 nm2 s-1
for DTCA, the intrafilm diffusion coefficient for TCA in films
of ester 2. Decays as in Fig. 3 were not observed if the film
contained PMS4 at weight ratios to ester 2 ranging from 1:1
to 4:1, where we estimated DTCA < 0.1 nm2 s-1. These values
are semi-quantitative, but the large differences between them do
indicate differences in intrafilm viscosity as experienced by TCA.
Diffusion coefficients show only a weak dependence on mol wt,
and we anticipate that similar values would apply to the ion pair
DMT+A-.
Kinetics. We had previously observed8 that release of DMT+
from DMT-T incorporated within a film of ester 2, with or without
added PMS, remained approximately in phase with repeated short
periods of illumination, equilibrating with photoacid within 5–10 s
after each period.
Different kinetics were observed when using surface-attached
DMT-T. Fig. 5 shows the time course of detritylation following
a single brief photolytic exposure. The initial post-photolysis
height of the 510 nm absorbance peak corresponded to ca. 40%
conversion of DMT-T to DMT+. Thereafter the peak slowly grew
in height to reach a plateau corresponding to >90% conversion. In
this experiment the area densities of photoacid and initial DMT-T
were respectively 14.6 and 0.4 nmol cm-1. The acid is distributed
throughout the 200 nm thick film, whereas DMT-T is restricted
to the surface where it forms a layer of local concentration
comparable to that of photoacid. The post-illumination progress
of detritylation could depend on diffusion of acid from the
overlying film to replace that consumed in the DMT-T layer at
the glass surface, and of DMT+ into the film. DMT+ movement
is the more important in this experiment because without it high
levels of detritylation could not be achieved by TCA alone.8
Patterned photoactivation. Diffusion in the x or y axes in a
plane can be measured by a patterned photoactivation method.13,14
Diffusion is observed as time-dependent decay of the pattern.
Using photolysis through a Rhonchi ruling15 with 50 line pairs
per mm contacted with the surface of a film of ester 2 containing
BG, we generated a pattern of alternating green and colorless lines
(Fig. 4). Unfortunately the relatively large width of the lines made
relaxation of the pattern too slow for useful spectrophotometric
measurement on films where the ruling was left in place following
photolysis (ESI†). But it was useful to demonstrate the imaging
potential of such films at a resolution appropriate for array
fabrication.
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 451–459 | 453
©

View Article Online
Table 1 Reactivity of DMT+ with carbocation scavengers in solution
2nd order rate constanta
(M-1 s-1)
Reagent
Tri-n-butyltinhydride
Triphenyltinhydride (TPTH)
Tris(trimethylsilyl)silane
Poly(carbomethylsilane)b
Tetramethyldisiloxanec
Bis(4-methoxyphenyl)phenylsilane
(4-Methoxyphenyl)diphenylsilane (MPDPS)
5-Methoxyindole
1.6 ¥ 103
8.7 ¥ 101
1.9 ¥ 101
4.3 ¥ 10-1
1.6 ¥ 10-1
1.1 ¥ 10-1
1.0 ¥ 10-1
8.8 ¥ 10-2
5.0 ¥ 10-2
2.5 ¥ 10-2
1.3 ¥ 10-2
1.2 ¥ 10-2
6.0 ¥ 10-3
4.0 ¥ 10-3
3.3 ¥ 10-3
<10-4
Triethylsilane
Poly(methylhydrosiloxane)d
Triphenylsilane (TPS)
Indole
Tri-n-propylsilane
Pyrrole
Diphenylsilane
Phenylsilane
Triethoxysilane
Tri-isopropylsilane
<10-4
<10-4
Notes:a The values in the Table for each compound are the averages
for at least 3 observations which were consistent to within a few %. In
view of the million-fold range of rate constants, no attempt was made
to calculate standard deviations. b Polymer, average mol wt 2,000 or 900
Da. The repeating subunit is–CH2-SiH(CH3)–. c Two silylhydride groups
per molecule. d Oligomer, average mol wt 390 Da, with an average of
three silylhydride groups per molecule. All measurements were made
spectrophotometrically in 3.2 mL of DCM containing DMT-T (6–10 mM)
and TCA (0.1–0.15 M) followed by a few mL of scavenger solution at a
concentration that gave suitable rates of absorbance change at 505 nm for
absorption measurements over 30–180 s.
Fig. 5 Kinetics of detritylation by a 200 nm thick film of ester 2 coating
a glass slide surface carrying DMT-T at ca. 400 pmol cm-2. Photolysis
(365 nm, 5 s, 165 mJ cm-2) was immediately followed by measurement of
absorbance spectra at 5 s intervals. The peaks at 510 nm and 385 nm are due
to DMT+ (emM ª 80 mM-1 cm-1) and the substituted 2-nitrosobenzophe-
none photoproduct (emM ª 4.1 mM-1 cm-1). The anticipated absorbance at
510 nm for a DMT+ density of 400 pmol cm-2 is 0.032, superimposed on a
baseline increase due to the nitroso-photoproduct of 2. The baseline noise
is that of a CCD detector.
Scavenging agents for DMT+
Intrafilm polymer. Detritylation of surface-attached DMT-T
caused by photolysis of overlying ester 2 films containing an
equal or greater amount of PMS17 gave <30% detritylation,
even though photolysis was continued to give similar intrafilm
TCA concentrations to those used in the absence of polymer.
A protracted dark phase of detritylation as in Fig. 5 was not
detected. Parallel experiments with the incorporated pH indicator
BG showed that the presence of an appropriate polymer (i.e.
lacking electronegative heteroatoms8) did not lower proton activity
at comparable intrafilm TCA concentrations. Our results for the
extent and kinetics of detritylation for surface-attached DMT-
T differ markedly from those8 for DMT-T incorporated within
the volume of the photoacid generating film. They are consistent
with the diffusional model used for Fig. 2 and visualized in
Scheme 1.
Potential scavengers. Scavengers of DMT+ have been included
in both solution phase19 and conventional solid-phase synthesis.20
In the latter case the scavenger was in the acidic detritylating
solution.
We examined the performance of compounds from four classes
of candidates: pyrroles, indoles, trialkyl- or triarylsilanes and
trialkyl- or triaryltin hydrides. The first two classes react by
electrophilic substitution with DMT+, the latter two by hy-
dride transfer. We extended the range of triarylsilanes by syn-
thesizing (4-methoxyphenyl)-diphenylsilane and bis(4-methoxy-
phenyl)-phenylsilane as described under Experimental.
The reactivity of the candidate scavengers with DMT+ in
solution provided a simple first screen (Table 1). In all cases
the scavenger was present in a several- to many-fold excess
over DMT-T. The decay of acid-generated DMT+ concentrations
on addition of scavengers followed pseudo-first order kinetics.
The calculated second-order reaction rate constants varied >106-
fold with scavenger structure. The most reactive were tri-n-
butyltinhydride and triphenyltinhydride.
Plasticizers. So-called plasticizers18 incorporated within solid
polymers lower intrafilm viscosity and increase thermoplasticity
by reducing binding between polymer molecules. We screened a
range of 16 plasticizers, all lacking electronegative heteroatoms to
avoid weakening of TCA,8 and containing aryl groups in antic-
ipation that they would confer compatibility with polystyrene-
based polymers (ESI†). Further selection criteria were high
solubility in DCM and no effect on the film-forming ability of
solutions of the host polymer (PMS) at plasticizer to polymer
ratios ≤1.0 by weight. Triphenylmethane (TPM) was the most
promising: unlike the other candidates it formed an optically
clear solid film when cast from DCM solution without added
polymer.
Performance. When tested in solid films (Fig. 6) the rates
of DMT+ reaction with TPS or MPDPS followed first order
kinetics, in keeping with the excess of scavenger over DMT+.
The reaction of TPTH with DMT+ differed: despite a very fast
initial rate ca. 10% of the DMT+ remained unreacted. This
unreactive fraction may represent scavenger-less microdomains
of DMT+. The inclusion of TPM (ca. 10% by weight) as a
plasticizer in TPTH-containing films resulted in ≥98% removal
454 | Org. Biomol. Chem., 2009, 7, 451–459
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
DMT+A- + Ar3SiH = DMTH + Ar3SiA
(3)
We identified DMTH (i.e. 4,4¢-(phenylmethylene)-bismethoxy-
benzene)) and Ar3SiA (trisilanetrichloroacetate) as products of
the reaction of triphenylsilane in solution with DMT+ generated
from DMT-T with TCA. We found no evidence of 5¢-O-tri-
phenylsilylthymidine formation.
Pyrrole, indole and 5-methoxyindole were reactive with the
DMT+ in films but there were complex changes in the UV-vis
absorbance spectra. Following photolysis the films were not easily
washed from the glass slide with solvents. These changes, not
seen in solutions where acid was provided by addition rather than
photolysis, may be due to photo-induced polymerization of ring
systems.
Fig. 6 Reaction kinetics of hydride donors with DMT+ in a solid film. The
inserted graph is semilogarithmic. Films were cast from a DCM solution,
1 mL of which contained 0.75 mg each of ester 2 and PMS, 100 nmol
of DMT-T, and one of TPS (0.75 mg) or DMTPS (0.75 mg) or TPTH
(0.1 mg). Photolysis (250–320 mJ cm-2) generated TCA to release DMT+,
measured at 510 nm from 5 s onwards after photolysis.
Intrafilm scavenging for surface-attached DMT-T. Experi-
ments were with slides carrying DMT-T at a surface density of
ca. 400 pmol cm-1. Any DMT-T remaining after the irradiation
of a slide coated with a film of ester 2 with added polymer
and scavenger was measured by removing the spent film and
replacing it with a second film of ester 2 alone before a second
irradiation (see Experimental). For example, part of a slide with
surface attached DMT-T was coated with a first film composed
of PMS, ester 2 and the scavenger MPDPS (1:1:1 by weight), and
then briefly photolysed (60 mJ cm-2). The film was removed 60 s
later with dichloromethane and then replaced by a second film
consisting of ester 2 alone, before a second photolysis (180 mJ
cm-2). Spectrophotometric scans of the film following the second
photolysis would have detected, at a sensitivity of 1.25 pmol
cm-2, any DMT+ originating from surface-attached DMT-T that
had escaped detritylation by the first exposure to photoacid. None
was detected: thus the first illumination when the film contained
a scavenger had been sufficient to remove >99% of DMT-groups
within 60 s.
of DMT+. The less-reactive scavengers TPS and MPDPS are
required at several-fold higher concentrations than TPTH, and
may themselves act as plasticizers in achieving >98% removal of
DMT+.
We explored incomplete access of scavengers to DMT+ within
films, of poly(carbomethylsilane)21 acting as both scavenger and
film-forming polymer. Ester 2 and DMT-T were included in
the films, and DMT+ removal was measured following brief but
intense photolysis to generate TCA (Fig. 7). Reaction kinetics
were biphasic, and incomplete. Inclusion of TPM (10% by weight)
increased the rate of DMT+ scavenging by 3–5 fold, and changed
the complex kinetics towards a single monotonic decay. The rates
were ca. two-fold greater with poly(carbomethylsilane) of lower
molecular weight.
The high surface densities of DMT-T and spectrophotometric
sensitivity for DMT+ required for these experiments are unusual,
but without them detritylation values of 98–99% could not have
been differentiated from lower ones such as 95–97%. On the
other hand the high initial concentration of DMT-T makes
it more difficult to achieve high percentage detritylation at
equilibrium.
Film composition and detritylation of surface-attached DMT-T
Intrafilm viscosity. Table 2 summarizes the extent of detrityla-
tion of surface-attached DMT-T following photolysis of overlying
films of ester 2 with or without added PMS or plasticizer TPS.
Major observations are that (i) photolysis of ester 2 films supports
extensive detritylation, up to at least 98%, (ii) inclusion of PMS
reduces detritylation to a maximum of 30%, and (iii) this effect
of PMS is overcome by further inclusion of a plasticizer. These
observations are consistent with lowered intrafilm viscosity being
a cause of higher levels of detritylation.
Fig. 7 Scavenging of DMT+ in solid films of poly(carbomethylsilane)
and the effects of added plasticizer. Average mol wt of the polymer sample
was either 800 or 2000 Da. Films were cast from a DCM solution, 1 mL
of which contained poly(carbomethylsilane) (12 mg), ester 2 (3 mg), and
DMT-T (24 nmol) with or without triphenylmethane (1.7 mg). Photolysis
and timing of absorbance measurements were as in Fig. 6.
For comparison we include in Table 2 results for films of ester 2
with incorporated DMT-T,8 where no amount of diffusion allows
DMT+ to escape from proximity to detritylated thymidine. In those
conditions only incomplete detritylation of 70–80% was achieved
despite photoacid generation for 3–4 photolytic half-lives.
Reaction product(s). Reaction of hydride donors with DMT+
is irreversible, the C-H bond being stronger than the Si-H or Sn-H
bonds.22 The reaction can be written as:
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 451–459 | 455
©

View Article Online
Table 2 Characteristics of detritylation of DMT-T by photogenerated TCA in solid filmsa,b
Property or process
Film compositionb
DMT-T distributed in film
PAG alone
DMT-T attached to glass support
PAG + PMS
PAG alone
PAG + PMS
PAG + PMS + TPM
DTCA (nm2 s-1)c
ca. 50
< 0.1
ca. 50
< 0.1
100–200
Plus dark phasef
≥ 98%
≤ 2%
Detritylationd,e
Tracks photoacid generation
70–80%
Plus dark phasef
≥ 98%
No dark phase
≤ 30%
Extent of detritylationg
Second detritylation steph
→
→
Not applicable
≤ 2%
≥ 30%
Notes:a Detritylation experiments at ambient temperature (20–23 ◦C), 1–1.5 mm thick films were made by spread-coating. Ester 2 was used throughout.
Casting solutions were in DCM with 2% solids by weight. Ratios by weight were 1:4 for PAG:PMS and 1:4:0.5 for PAG:PMS TPM. For DMT-T attached
to glass the surface density was 200–700 pmol cm-2. b The values given for % detritylation and diffusion coefficients are for ranges identified by at least three
separate experiments for each set of conditions. c Diffusion coefficients for TCA were calculated from separate experiments, from Thalf for decay of BG
protonation following photolytic illumination at 365 nm, as shown in Fig. 3 with thin films and also the thicker films (0.5 mm) of PAG + PMS + plasticizer
(TPM). d As determined by DMT+ accumulation, measured by absorbance changes at 510 nm. e Following exposure at 365 nm, 60–100 mJ cm-2 in 5 sec.
f Continued detritylation after cessation of photoacid generation, as in Fig. 5. g Photolysis at 300–400 J cm-2, and exposure plus post-exposure time of
100s. % Detritylation calculated either by reference to pre-photolysis BG or PAG concentration as internal standards for DMT-T incorporated within
films, or to the total DMT+ released from surface-attached DMT-T by repeated exposure to photoacid or elution with TCA solution. Relatively high
incident energies were used to minimize any variations in photoacid formation. h Following removal by solvent of the film used for the first detritylation
step, and then repeating film casting with ester 2 alone followed by illumination as a second detritylation step to detect remaining DMT-T.
The role of intrafilm diffusion in increasing the detritylation of
surface-attached compound overlaid by a photoacid-generating
Discussion
Computer simulations
solid film has not been previously reported, although diffusional
enhancement may have unknowingly occurred in experimental
protocols using post-exposure heating either with7 or without6
chemical amplification of the photoacid. High intrafilm con-
centrations of reagents for chemical amplification7,24 or photo-
sensitization25 may also act as plasticizers.
Analytical solutions to diffusion equations are complex, but digital
simulations are not. We found it instructive, as in Fig. 2, to
simulate some of the situations involving diffusion, detritylation
and scavenging (see ESI†).
Scavengers. Removal of DMT+ by carbocation scavengers has
the attraction of converting detritylation following photoacid
generation into a time-dependent irreversible process. But mi-
crodomain formation is possible, such that not all DMT+ is
accessed by the scavenger. Examples of restricted removal of
DMT+ by scavengers are illustrated in Fig. 6 and 7, as is the
ability of added plasticizers to overcome them.
Triarylsilanes have desirable properties for intrafilm scavengers,
namely no interference with film formation, no unwanted pho-
tochemical reactions, detritylation of DMT-T to below detection
limits, monophasic DMT+ removal kinetics proceeding to com-
pletion, and reactivity giving rapid DMT+ removal. It is likely that
the usefulness of triarylsilanes is enhanced by an ability to act also
as plasticizers. Triphenyltinhydride is ca. 100-fold more reactive
than triarylsilanes and used at lower concentrations, in which case
any need for increased intrafilm diffusion to abolish microdomains
can be met by the inclusion of a plasticizer.
Solid films, polymers, and plasticizers. These topics are inter-
related. Solid films lower intrafilm diffusion rates, with two
outcomes: bimolecular reactions become diffusion-limited and
movement of reaction products away from a localized site of
production is impaired. These effects are not marked in solid films
of monomeric photoacid generators, and it was fortunate that our
preferred class, substituted 2-nitrobenzyl esters of TCA, formed
optically clear and uniform films when cast from solution. But
inclusion of polymer may be required to provide a more adjustable
and less expensive solid environment.
The inclusion of polymers brings potential complications.
Diffusion of small molecules in a solid film of linear polymers
involves thermal movement of polymer chains to create holes
(“free volumes”) into which adjacent small molecules may jump.23
The rates of hole formation are greater at the ends of polymer
chains then at their central regions, creating locally different
diffusion rates. Polymers can immobilize and isolate smaller
molecules that can be trapped between the rigid segments of
polymers. Furthermore, supra-molecular microdomains may arise
by phase-separation of low molecular weight compounds within a
polymer matrix. The inclusion of plasticizers loosens the polymer
structure and diminishes the diffusional isolation of smaller
molecules within the film.
Because DMT+ scavenging coupled to acid-induced detrity-
lation is irreversible, the use of weaker acids than TCA can
be considered, possibly extending the range of useful photoacid
generators.
Light sensitivity
A further complication arises if components of the photoacid-
generating film segregate into microdomains. This cannot occur
if the film is composed solely of PAG, as with film-forming
substituted 2-nitrobenzyl esters of TCA.8 Such films have relatively
low internal viscosity and included molecules of low molecular
weight, such as photosensitizers or scavengers, are less likely to
form diffusionally isolated domains.
Light sensitivity is of considerable practical importance for
photodirected array synthesis. It is influenced not only by PAG
properties but also the necessary intrafilm acid concentration for
extensive detritylation. Our use of a triarylsilane for scavenging
achieved >99% detritylation at 60 mJ cm-2 (365 nm) within 60 s,
a light sensitivity that is 15–100 fold higher than for direct photo-
deprotection.26,27 Higher light sensitivities could be achieved
456 | Org. Biomol. Chem., 2009, 7, 451–459
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
by increasing the post-illumination scavenging period, because
the combination of detritylation with scavenging is irreversible.
Scavenging is the rate-limiting step: the more reactive the scavenger
the faster the detritylation.
Polymeric photoacid generators. The use of enhanced diffusion
and its associated movement of photoacid would require careful
control to be used in fabrication of very high density arrays with
elements of say 4 ¥ 4 mm or less, otherwise photolithographic
resolution would be degraded. But a polymeric PAG that on
illumination yielded a polymeric acid28 could overcome this
restraint. A co-polymer of styrene and 4-styrene-sulphonic acid
esterified with a substituted 2-nitrobenzyl alcohol29 is an example
but the polymeric sulphonium salts described by Kim et al.30 are
not, because the photoacid is monomeric.
Fig. 8 Structure of trichloroacetic acid photogenerators. We have either
described or referred to the synthesis of 25 photosensitive esters of
Cl3CCOOH with variously substituted 2-nitrobenzyl alcohols.8 Ester 2
(R4 = R5 = OCH3, R6 = NO2) was used for virtually all experiments
described here, but mention is made below to ester 1 (R4 = R5 = OCH3,
R6 = H), ester 6 (R4 = H, R5 = Cl, R6 = H) and ester 7 (R4 = H, R5 = Cl,
R6 = NO2). Each of these four esters formed solid and transparent thin
films on glass when cast from DCM solution.
Diffusion of H+ or DMT+ within a solid film of polymeric
photoacid could occur by electroneutral hopping of these cations
between immobile anionic groups. Illuminated areas lack other
mobile anions, non-illuminated areas lack ions of any sort, so
diffusion of H+ from the former areas to the latter would be
electrogenic, self-limiting and probably not extending beyond the
light/dark boundary by more than a few nm. These circumstances
would not be altered by the inclusion of a plasticizer such
as TPM to enhance collision rates between surface-attached
oligonucleotides and acidic groups on the polymeric photoacid,
or between DMT+ and scavenger molecules.
(4-Methoxyphenyl)diphenylsilane
Our synthetic route differs from those previously reported.31,32
4-Methoxyphenyl magnesium bromide (1M solution in THF
1.8 mL, 1.8 mmol), was added dropwise by syringe over 10 minutes,
at 0–5 ◦C, under argon, to a solution of diphenylchlorosilane
(0.435 g, 1.5 mmol) in THF (6 mL). The resulting mixture was
◦
stirred at rt for 16 hours. The mixture was cooled to 0–5 C and
2% aqueous hydrochloric acid (25 mL) was added dropwise over
20 minutes followed by dichloromethane (50 mL). The organic
layer was washed with water (20 mL), 3% aqueous sodium
bicarbonate (3 ¥ 20 mL), water (20 mL), brine (20 mL), dried
(sodium sulphate) and concentrated in vacuo. The residue was
chromatographed on a silica gel column eluting with hexane-ethyl
acetate (24:1, v/v) to give the product as a colourless liquid; (0.17
Conclusion
These concepts and methods for chemical scavenging and the
control of intrafilm viscosity within photoacid-generating solid
films were developed for the fabrication of oligonucleotide
arrays, to pull detritylation to completion rather than force
it with strong acid. Novel phosphoramidite building blocks
with photolabile protecting groups are not required, and light
sensitivity is high. Our own approach may be of wider appli-
cability for the photolithographic miniaturization of chemical
syntheses on solid surfaces in cases where a photogenerated
compound is a reagent for a reaction that would not proceed
to the required degree of completion unless a reaction prod-
uct within the film is removed either by scavenging or z-axis
diffusion.
1
g, 41%); Rf (A) 0.61; H NMR 3.78 (s, 3H, OCH3), 5.38 (s, 1H,
SiH), 7.02 (d, 2H, J= 5.67 Hz, H-3¢, H-5¢), 7.53 (m, 12H, H-2¢,
H-6¢, 2 x (H-2-H-6)); 13C NMR 54.95 (OCH3), 114.17 (C-3¢, C-5¢),
160.86 (C-4¢); GC MS for C19H17OSi, [M - H], found 289.3.
Bis(4-methoxyphenyl)phenylsilane
4-Methoxyphenyl magnesium bromide (1 M solution in THF
10 mL, 10 mmol), was added dropwise by syringe over 15 minutes,
at -10 to 0 ◦C, under argon, to a solution of dichlorophenyl-
silane (1.8 g, 1.5 mL, 10 mmol) in THF (6 mL). The resulting
mixture was stirred at -10 to 0 ◦C, under argon, for 2 hours. More
4-methoxyphenyl magnesium bromide (1M solution in THF
10 mL, 10 mmol), was added dropwise by syringe over 15 minutes.
◦
The mixture was kept at -10 to 0 C, for 1 hr and then at rt for
16 hours. The mixture was cooled again to 0 ◦C, and 2% aqueous
hydrochloric acid (50 mL) was added dropwise over 30 minutes
followed by dichloromethane (100 mL). The organic layer was
washed with water (20 mL), 3% aqueous sodium bicarbonate (3 ¥
20 mL), water (20 mL), brine (20 mL), dried (sodium sulphate)
and concentrated in vacuo. The residue was chromatographed on
a silicagel column eluting with hexane-ethyl acetate (95.5:4.5) to
give the product as a colourless liquid; (0.99 g, 32%); Rf (A) 0.29;
1H NMR 3.76 (s, 6H, OCH3), 5.50 (s, 1H, SiH) 6.97 (d, J = 5.52
Hz, 4H, 2x H-3¢, 2x H-5¢), 7.43 (m, 9 H, 2x H-2¢, 2x H-6¢, H-2-
H-6); 13C NMR (DMSO-d6) 54.96 (OCH3), 114.01 (C-3¢, C-5¢),
161.16 (C-4¢); GC MS for C20H19O2Si, [M - H]-, found 319.19;
LRMS for C20H20O2SiNa, [M + Na]+, found 342.17.
Experimental
Analytical methods
1H and 13C NMR spectra were recorded using Bruker Avance DPX
600 MHz spectrometer in DMSO-d6 solutions and referenced to
1
the residual DMSO signal. All H and 13C signals are given in
ppm (d scale). The LRMS (LCMS) in ES mode were run using
Waters LCT with ESI Lockspray source and Alliance 2795 LC
system. GC MS spectra were recorded on ThermoFinnigan Polaris
Q instrument. Thin layer chromatography was run on HPTLC
(high performance TLC), Merck Kieselgel 60F254 analytical plates
in hexane/EtOAc (9:1) (A). Coarse ICN silica gel was used for
short column chromatography.
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 451–459 | 457
©

View Article Online
Attachment of DMT-T to glass slides
Film preparation
We followed literature methods to attach linkers and mononu-
cleotides to glass. Slides derivatized with 3-aminopropyl-
triethoxysilane were either purchased from Sigma-Aldrich or
prepared in our laboratory. We followed literature procedures
to prepare (a) 5¢-O-DMT-thymidine-3¢-O-succinate by reacting
commercially available 5¢-O-DMT-thymidine with succinic anhy-
dride in the presence of dimethylaminopyridine in dry pyridine,
and (b) 5¢-O-DMT-thymidine-3¢-O-[(p-nitrophenyl)-succinate] by
the condensation of 5¢-O-DMT-thymidine-3¢-O-succinate with p-
nitrophenol in the presence of dicyclohexylcarbodiimide (DCC)
in dioxane.33 As an alternative glass slides were derivatized with
3-glycidoxypropyltrimethoxysilane followed by ethylene glycol or
hexaethylene glycol to provide a terminal primary hydroxyl group.3
In both preparations we used either whole slides in a staining jar
or slides that had been cut into rectangular pieces roughly 20 ¥
6–7 mm. These pieces can be treated in a round-bottomed flask
requiring smaller volumes of reagents.
Thin solid films (ca. 0.1–1.5 mm) were cast from DCM solutions
onto microscope slides or pieces thereof, as previously described.8
The presence in the casting solution of some but not all of the
tested silanes diminishes the ability of the solutions to wet and
spread on a glass surface. This effect became more marked as the
concentration or hydrophobicity of the silanes increased. Tri-n-
butyltinhydride also interfered with film casting. Film-formation
was not impaired by triarylsilanes or triphenyltinhydride in film-
casting solutions at up to 33% of the solute mass.
Apparatus for photolysis and absorbance measurements of films on
glass slides
This was as previously described.8 However, sensitivity became a
major issue when ≥99% detritylation in films left residual DMT-T
area densities of ≤2–8 pmol cm-2 attached to the underlying glass
surface, remaining from pre-detritylation densities of respectively
200–800 pmol cm-2. For these extreme sensitivities we used a
deep-well spectrometer with a signal-to-noise ratio in the 480–
530 nm region of 2000:1, which with signal averaging reduced the
baseline noise to ca. 0.0005 absorbance units, corresponding to
a DMT+ area density (measured at 510 nm) of 0.6 pmol cm-2.
This sensitivity was in itself satisfactory, but not if the DMT+
absorption peak was lost on the side of the much larger peak of
the nitroso-photoproduct formed by photolysis of a substituted 2-
nitrobenzyl ester (Fig. 8). Amongst such esters8 the photoproduct
of ester 1 gave severe interference, that of ester 2 was satisfactory
provided that photolysis was not extensive, and those of esters 6
and 7 gave insignificant interference.
Coupling of mononucleotides to glass slides derivatized with
3-aminopropyltriethoxysilane
Glass microscope slides, 75 ¥ 25 ¥ 1 mm, were placed in a staining
jar and immersed, with stirring, for 120 h at rt in 1 M DCC solution
in DCM (45 mL) containing 5¢-O-DMT-thymidine-3¢-O-succinate
(0.50 g). The slides were then washed thoroughly with dry DMF,
dioxane, methanol, acetone, acetonitrile, and diethyl ether and
were finally dried in vacuo over phosphorus pentoxide overnight.
A small fragment (ca. 2–3 cm2) was treated with 1 mL of 3% TCA in
DCM (w/v) and the concentration of trityl cation in the resulting
solution determined from absorbance measurement at 503 nm.
Using commercially available derivatized slides we observed a
surface density for 5¢-O-DMT-thymidine of 50–60 pmol cm-2, and
twice that value using slides derivatized in our laboratory.
Measurement of photoacid acid production in films
Photolysis of substituted 2-nitrobenzyl esters produces the acid
and substituted 2-nitrosobenzophenones. Alternative reaction
routes are negligible, and there is a 1:1 stoichiometry between
acid release and formation of the 2-nitrosobenzophenone product.
The difference spectrum between before and after photolysis is
due to the 2-nitrosobenzophenone, and its measurement gives
the amount of acid released. For ester 2 the difference spectrum
between before and after photolysis in DCM has a peak at 384
nm, with emM = 4.1 mmol-1 cm-1.
Coupling of mononucleotides to glass slides derivatized with
3-glycidoxypropyltrimethoxysilane
Rectangular pieces of a derivatized slide were placed in a
round bottom flask and treated with 5¢-O-DMT-thymidine-3¢-
O-phosphoramidite (0.25 g) in dry acetonitrile (5 mL) and
activator solution (5 mL of 0.25 M tetrazole in acetonitrile, from
Applied Biosystems) added under argon. The reaction was stirred
vigorously at rt for 10 minutes and exposed to ultrasound for
5 min. The slides were washed thoroughly with dry acetonitrile
and treated with the oxidiser solution (25 mL), from Applied
Biosystems, for 60 sec at rt. After being washed again with dry
acetonitrile the slides were treated with the Cap A and Cap
B solution (1:1 v/v, 25 mL), from Applied Biosystems, and
finally washed thoroughly with dry DMF, dioxane, methanol,
acetonitrile, acetone, and diethyl ether and dried in vacuo over
phosphorus pentoxide overnight. A small fragment (2–3 cm2)
was treated with 3% TCA in DCM (1 mL). Absorbance of the
resulting solution at 503 nm was used to calculate the loading with
5¢-O-DMT-thymidine, which was typically 700 pmol cm-2 for this
method of derivatization.
Measurement of film thickness
At the onset of this work we made measurements by interfer-
ometry, using the spectrophotometer8 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
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 absorbance spectra. The area density (nmol
cm-2) of PAG was calculated from the characteristic pre-photolysis
UV-absorbance of the PAG. The area density of material (mg cm-2)
458 | Org. Biomol. Chem., 2009, 7, 451–459
This journal is
The Royal Society of Chemistry 2009
©

View Article Online
was then calculated from the PAG density and the casting solution
composition, and converted to a film thickness on the assumption
that the film density was 1 gm mL-1.
12 P. R. Bergethon, E. R. Simons, E. Biophysical, Chemistry, (Springer-
Verlag, Berlin), 1989, 233-234.
13 B. A. Smith and H. M. McConnell, Proc. Natl. Acad. Sci. USA, 1978,
75, 2759–2763.
14 J. Davoust, P. F. Devaux and L. Leger, EMBO J, 1982, 10, 1233–1238.
15 The ruling consists of evenly spaced parallel opaque chrome lines on
clear glass. The line-to-space ratio is unity.
Acknowledgements
16 Photolysis was at 365 nm for all experiments.
This work was supported by grants from the Leverhulme Trust,
the BBSRC and EPSRC jointly through Grant No. EGM16064,
and the Institute of Cancer Research. We also thank Professor
Colin Cooper for his encouragement and support.
17 We have used polymer preparations of relatively low mol wt, in
anticipation that bimolecular reactions in the solid state would be less
restrained if intrafilm viscosity is lowered. PMS preparations (Aldrich)
were obtained with average Mn values 750 or 1300.
18 Aldrich Chemical Company Polymer Products, CD-Catalog and
Reference Guide, 2000, 446–448.
19 C. B. Reese, H. T. Serafinowska and G. Zappia, Tetrahedron Lett, 1986,
References
37, 2291–2294.
20 V. Ravikumar, M. Andrade, D. Mulvey, D. J. Cole, US Patent, 1996,
5,510,476.
1 U. Maskos and E. M. Southern, Nucleic Acids Res, 1992, 20, 1679–1684.
2 U. Maskos and E. M. Southern, Nucleic Acids Res., 1993, 21, 2267–
2268.
21 Polycarbomethylsilane (CH3(C2H6Si)nCH3)was available at Mav values
of 800 (mp ca. 82 ◦C) or 2000 (mp > 300 ◦C). Their films have a slight
weakening effect on the ability of photogenerated TCA to detritylate
DMT-T incorporated in the film, or to protonate intrafilm BG. This,
may be due to bonding of the silyl hydride (-Hd-) with carboxylic
hydrogen (-Hd+).
22 N. Wang, J. R. Hwu and E. H. White, J. Org. Chem., 1991, 56, 471–474;
J. B. Lambert, S. Zhang and S. M. Ciro, Organometallics, 1994, 13,
2430–2443.
3 M. H. Caruthers, Acc. Chem. Res., 1991, 24, 278–284.
4 Abbreviations. BG, Brilliant Green; BGH+, protonated BG; BMPS,
bis(4-methoxyphenyl)phenylsilane; DCM, dichloromethane; DMT,
4.4¢-dimethoxytrityl group; DMT+; 4,4¢-dimethoxytrityl carboca-
tion; DMT-T, 5¢-O-DMT-thymidine; ester 1, a-phenyl-4,5-dimethoxy-
2-nitrobenzyltrichloroacetate; ester 2, a-phenyl-4,5-dimethoxy-2,6-
dinitrobenzyltrichloroacetate; ester 6, a-phenyl-5-chloro-2-nitro-
benzyltrichloroacetate; ester 7, a-phenyl-5-chloro-2,6-dinitrobenzyl-
trichloroacetate; MPDPS, 4-(methoxyphenyl)-diphenyl-silane; PAG,
photoacid generator; PMS, poly(a-methylstyrene); TCA, trichloro-
acetic acid; TPM, triphenylmethane; TPS, triphenylsilane; TPTH,
triphenyltinhydride.
5 A. C. Pease, D. Solas, E. J. Sullivan, M. T. Cronin, C. P. Holmes and S.
P. A. Fodor, Proc. Natl. Acad. Sci. USA, 1994, 91, 5022–5026.
6 G. Wallraff, J. Labadie, P. Brock, R. DiPietro, T. Nyugen, T. Huynh,
W. Hinsberg and G. McGall, CHEMTECH, 1997, 22–32.
7 J. E. Beecher, G. H. McGall and M. J. Goldberg, Polym. Mat. Sci. Eng,
1997, 76, 597–598.
23 J. L. Duda, J. M. Zielinski, Diffusion in Polymers, 1996, Ed. P. Neogi,
143-171. Marcel Dekker Inc, New York.
24 J. E. Beecher, M. J. Goldberg, G. H. McGall, 2000, US Patent 6,083,697.
25 R. G. Kuimelis, G. H. McGall, M. J. Goldberg, G. Xu, 2007, US Patent
Application 20070161778.
26 S. Bu¨hler, I. Lagoja, H. Giegrich, K.-P. Stengele and W. Pfleiderer, Helv.
Chim Acta, 2004, 87, 620–659.
27 S. P. A. Fodor, J. L. Read, M. C. Pirrung, L. Stryer, A. T. Lu and D.
Solas, Science, 1991, 251, 767–773.
28 Throughout this article “acid” refers to Brønsted acid.
29 J. E. Hanson, E. Reichmanis, F. M. Houlihan and T. X. Neenan, Chem.
Mat., 1992, 4, 837–842.
30 M.-H. Kim, D.-Y. Kim, B.-K. Moon, J.-C. Park, Y.-H. Kim, Y.-H., S.-J.
Seo, US Patent 6,660,479.
31 H. Gilman and G. E. Dunn, J. Amer. Chem. Soc, 1951, 73, 3404–3407.
32 N. Lachance and M. Gallant, Tetrahedron Lett., 1998, 39, 171–174.
33 T. M. Atkinson, M. Smith, Oligonucleotide Synthesis: a Practical
Approach. 1984, Ed. M. J. Gait, IRL Press, Oxford & Washington
DC, pp. 47–48.
8 P. J. Serafinowski and P. B. Garland, Org. Biomol. Chem, 2008, 6, 3284–
3291.
9 A. B. Sierzchala, D. J. Dellinger, J. R. Betley, T. K. Wyrzykiewicz, C.
M. Yamada and M. H. Caruthers, J. Amer. Chem. Soc., 2005, 125,
13427–13441.
10 M. C. Pirrung, L. Fallon and G. McGall, J. Org. Chem, 1998, 63,
241–246.
11 E. H. Fisher and M. H. Caruthers, Nucleic Acids Res., 1983, 11, 1589–
1599.
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 451–459 | 459
©