Characterization of the non-uniform reaction in chemically amplified calix[4]resorcinarene molecular resist thin films

Article abstract of DOI:10.1071/CH11242

The ccc stereoisomer-purified tert-butoxycarbonyloxy-protected calix[4]resorcinarene molecular resists blended with photoacid generator exhibit a non-uniform photoacid-catalyzed reaction in thin films. The surface displays a reduced reaction extent, compared with the bulk, with average surface-layer thickness 7.0±1.8nm determined by neutron reflectivity with deuterium-labelled tert-butoxycarbonyloxy groups. Ambient impurities (amines and organic bases) are known to quench surface reactions and contribute, but grazing-incidence X-ray diffraction shows an additional effect that the protected molecular resists are preferentially oriented at the surface, whereas the bulk of the film displays diffuse scattering representative of amorphous packing. The surface deprotection reaction and presence of photoacid were quantified by near-edge X-ray absorption fine-structure measurements.

Full text of DOI:10.1071/CH11242

RESEARCH FRONT
Full Paper
CSIRO PUBLISHING
www.publish.csiro.au/journals/ajc
Aust. J. Chem. 2011, 64, 1065–1073
Characterization of the Non-uniform Reaction in
Chemically Amplified Calix[4]resorcinarene Molecular
Resist Thin Films1
A F
,
A
A
Vivek M. Prabhu, Shuhui Kang, R. Joseph Kline,
A
B
A
Dean M. DeLongchamp, Daniel A. Fischer, Wen-li Wu,
C
D
E
Sushil K. Satija, Peter V. Bonnesen, Jing Sha, and
E
Christopher K. Ober
A
Polymers Division, National Institute of Standards and Technology, Gaithersburg,
MD 20899, USA.
B
Ceramics Division, National Institute of Standards and Technology, Gaithersburg,
MD 20899, USA.
C
Center for Neutron Research, National Institute of Standards and Technology,
Gaithersburg, MD 20899, USA.
D
Center for Nanophase Materials Sciences, Oak Ridge National Laboratory, Oak Ridge,
TN 37831, USA.
E
Cornell University, Materials Science and Engineering, Ithaca, NY 14853, USA.
Corresponding author. Email: vprabhu@nist.gov
F
The ccc stereoisomer-purified tert-butoxycarbonyloxy-protected calix[4]resorcinarene molecular resists blended with
photoacid generator exhibit a non-uniform photoacid-catalyzed reaction in thin films. The surface displays a reduced
reaction extent, compared with the bulk, with average surface-layer thickness 7.0 ꢀ 1.8 nm determined by neutron
reflectivity with deuterium-labelled tert-butoxycarbonyloxy groups. Ambient impurities (amines and organic bases) are
known to quench surface reactions and contribute, but grazing-incidence X-ray diffraction shows an additional effect that
the protected molecular resists are preferentially oriented at the surface, whereas the bulk of the film displays diffuse
scattering representative of amorphous packing. The surface deprotection reaction and presence of photoacid were
quantified by near-edge X-ray absorption fine-structure measurements.
Manuscript received: 15 June 2011.
Manuscript accepted: 19 July 2011.
Introduction
alternative resists architectures are sought to maintain Ld , CD
and Rg , CD. The Ld may be reduced by increasing the
molecular mass of the photoacid,[15–19] but this can lead to
disadvantageous surface segregation and phase separation.[20,21]
More recently, covalent bonding of the PAG to the polymer has
shown markedly reduced Ld, as the acidic proton is restricted
by the charge-neutralizing counter-ion covalently bound to
the glassy polymer.[21–24] Alternatively, molecular resists are
attractive owing to the smaller Rg, miscibility with photoacid
generators[25] and well-defined molecular mass,[26] such
that monodisperse distribution of protecting groups led to sub-
30-nm features.[27] Measurements of the reaction-diffusion
kinetics of molecular resists show that the deprotection reaction
couples to the photoacid transport, hence Ld.[28,29] Measure-
ments of the film reaction kinetics are necessary to develop
predictable models to aid in resist design.[30–33]
Photoresist thin films are the enabling technology to fabricate
nanoscale features for the semiconductor industry with photo-
lithography.[1] However, imaging critical dimensions (CD)
smaller than 22 nm with low line-edge roughness (LER) with
193-nm immersion and extreme-ultraviolet (EUV) lithography
remains a challenge with chemically amplified resist materials
that use blends of photoacid generator (PAG) with polymer.
Alternative strategies with non-chemically amplified photo-
resists are also considered for next-generation EUV lithogra-
phy.[2,3] For chemically amplified resists, the quality of the
patterned structures depends on sequential process steps that
influence the spatial distribution of photoacids and the photo-
acid-catalyzed reaction-diffusion process.[1,4–9] Controlling the
reaction-diffusion process remains the predominant material
strategy to achieve smaller CD and LER for the ultimate reso-
lution of a printed feature.[10–14] However, because the photo-
acid diffusion length (Ld) and the size of the resist polymer (Rg)
have approached the length scales of the CD and LER,
Measurements that determine reaction kinetics with infrared
spectroscopy average throughout the film thickness. Such
approaches may not have the resolution or contrast to measure
1Official contribution of the National Institute of Standards and Technology; not subject to copyright in the United States
Ó CSIRO 2011
10.1071/CH11242
0004-9425/11/081065

RESEARCH FRONT
1066
V. M. Prabhu et al.
(a)
non-uniformities in sub-100-nm thin films. Deviations from
bulk reactions at the interfaces become more important and can
lead to surface (T-topping and closure) and substrate (undercut-
ting and footing) defects on the final lithographic features.[34]
Surface segregation of PAG was observed in polymer resists by
near-edge X-ray absorption fine structure (NEXAFS) spectros-
copy.[35–37] Time-of-flight secondary-ion mass spectrometry
was used to show uniform depth profiles in optimized resist
formulations under controlled-atmosphere clean-room condi-
tions.[38] In addition to the non-uniform distribution of PAG,
amine moieties in the ambient conditions quench the reaction at
the free surface. Such effects are typically not considered by
reaction kinetics studies because the film thickness to determine
kinetics constants is usually larger, such that the surface does not
contribute substantially to the change in bulk signal.
The reaction kinetics of chemically amplified molecular
resists for next-generation lithography were previously quanti-
fied by time-resolved Fourier-transform infrared reflectance
absorption spectroscopy.[28,29] This approach uses a bilayer film
whereby a PAG is blended with resist and transferred on a PAG-
free resist film on a silicon substrate[28,29] by the aid of a poly
(dimethylsiloxane) (PDMS) stamp. This bilayer is then exposed
to UV light and post-exposure baked (PEB) to provide a one-
dimensional photoacid gradient that mimics a step-exposure line
edge. Using the film-average composition change in films
thicker than 100 nm, reaction kinetics constants and photoacid
diffusion constants are estimated. However, on measuring the
reaction front bilayers with high-resolution neutron reflectivity,
the surface reaction was found to be inhibited, leaving a thin
layer of highly protected resist that comprised less than 10 % of
the 100-nm thin film (Accessory Publication, Fig. S1). The
surface layer eventually completely deprotects with longer
reaction time. Two experimental origins were suspected; the
first was the use of the PDMS stamp, which may absorb and
retain PAG, leaving the surface depleted of PAG. The second
was the role of ambient amines that quench the photoacid at the
surface to reduce the reactivity.[39,40]
Here, we demonstrate that this surface layer was not an
artefact of the PDMS stamping, but a property of the purified ccc
stereoisomer tert-butoxycarbonyloxy calix[4]resorcinarene (t-
Boc-CM4R) molecular resist that preferentially orders at the
free surface to kinetically limit the photoacid-catalyzed reac-
tion. Neutron reflectivity provides a nanometre-resolution
depth profile of the fraction of deprotection, whereas grazing-
incidence X-ray diffraction (GIXD) quantifies the molecular
order at the surface and bulk. The surface chemistry, character-
ized by NEXAFS, verifies the presence of photoacid as well as
the change in surface reaction with reaction time and tempera-
ture within the top 1 to 6 nm. Although PAG segregation and
surface-versus-bulk reactions are known in polymer
resists,[35–37,41] such effects are less understood for well-defined
molecular resists with low polydispersity.[42,43] Further, the
alternative topology found with molecular resists may aid in
the design of resists with order–disorder transitions. Such
materials may find particular promise for thermal-induced
scanning-probe lithography.[44]
O
O
O
O
O
O
HO
OH
O
O
O
O
O
O
O
O
CD3
O
O
O
O
HO
HO
OH
OH
CD2
CD3
Photoacid
+
CO2
O
O
HO
OH
O
O
O
O
(b)
Calculated
NR & FTIR (exp. data)
2.4
2.2
2.0
1.8
1.6
1.4
1.2
1.0
8
6
4
2
0
Average number of D-t-Boc groups per CM4R
Fig. 1. (a) The photoacid-catalyzed deprotection of deuterated tert-
butoxycarbonyloxy calix[4]resorcinarene (D-t-Boc-CM4R) is shown with
main volatile products (carbon dioxide and deuteron-isobutene). The extent
of reaction is dependent on reaction kinetics conditions. (b) Calculated and
measured neutron scattering length density as a function of number of D-
t-BocgroupsperCM4R.Uncertaintiesareestimatedtobelessthan5 %based
on propagation of angular and wavelength divergence used to fit the neutron
scattering length density profiles within a given chi-squared statistic.
substrates. After a post-apply bake, the films were UV-exposed
and PEB for 60 s at temperatures ranging from 608 to 808C. This
elevated temperature is required to increase the photoacid-cat-
alyzed deprotection reaction rate of D-t-Boc-CM4R. On post-
exposure baking, the D-t-Boc group leads to volatile deprotec-
tion products (carbon dioxide and deutero-isobutene) as shown
in Fig. 1. The loss of deuterium protecting groups causes a
decrease in film thickness proportional to the deprotection
extent. The average film t-Boc protection level was character-
ized by infrared spectroscopy. The change in film D-t-Boc
content also provides neutron reflectivity contrast, because
deprotected CM4R and 100 % D-t-Boc-protected CM4R have a
large scattering length density difference (Fig. 1). Neutron
reflectivity methods can differentiate films that are deprotected
uniformly from structures with buried interfaces.[45]
Fig. 2 shows neutron reflectivity data for five PEB tempera-
tures as Fresnel-reduced reflectivity Q4R versus Q, where R is
the absolute reflectance and Q the wavevector. The reflectivity
data (symbols) were fitted using the Parrat model (solid line)
with the fit quality x2 statistic shown.[46] The interference of the
reflected neutrons from the substrate and film interfaces leads to
Kiessig fringes that are inversely proportional to the character-
istic length (thickness) within the film plane. The total film
thickness may be estimated directly from the fringe periodicity
whereas fringe persistence provides evidence of the low surface
roughness even after deprotection. The scattering length density
(SLD ¼ b/v) profiles that are results of the model slab fits are
shown as Q2c ¼ 16pb=v distance (Z) from the silicon substrate,
silicon oxide, organic resist film, then air (Q2c ¼ 0). Where b is
the total scattering length over all atomic elements per molecule
within molecular volume v. This depth profiling in units of Qc2
shows that for each film, the surface maintains a higher SLD
Results and Discussion
Thin film blends of triphenylsulfonium perfluoro-
butanesulfonate (TPS-PFBS) PAG and CM4R that was fully
protected with deuterated tert-butoxycarbonyloxy groups (D-t-
Boc-CM4R) were prepared as thin films on silicon wafer

RESEARCH FRONT
Structure of calix[4]resorcinarene resist thin films
1067
10Ϫ7
10Ϫ8
80 ЊC
χ2 ϭ 2.21
2
1
0
2
1
0
2
1
0
2
1
0
2
1
0
2
1
0
10Ϫ9
10Ϫ7
75 ЊC
χ2 ϭ 2.42
10Ϫ8
10Ϫ9
10Ϫ7
70 ЊC
χ2 ϭ 3.67
10Ϫ8
10Ϫ9
10Ϫ7
65 ЊC
χ2 ϭ 3.0
10Ϫ8
10Ϫ9
10Ϫ7
60 ЊC
χ2 ϭ 4.8
10Ϫ8
10Ϫ9
10Ϫ7
no PEB
χ2 ϭ 8.29
10Ϫ8
10Ϫ9
0.00 0.02 0.04 0.06 0.08 0.10
Q [ÅϪ1
0
200 400 600 800 1000
Z from Silicon [ÅϪ1
]
]
Fig. 2. (left panel) Fresnel-normalized neutron reflectivity as a function of post-exposure bake temperature
after DUV exposure for single-layer films of D-t-Boc-CM4R blended with TPS-PFBS with Parratt model fits
(solid lines); and (right panel) corresponding scattering length density profiles showing bulk reaction and
surface-inhibited reaction extent.
than the bulk. The expected change in Q2c for 100 % substituted
to 0 % D-t-Boc-substituted CM4R was calculated as shown in
experimental values and the calculated values based on mass
density and chemical composition is excellent. Therefore, the
ꢃ2
˚
A )
Fig. 1. In Fig. 2, the surface remains highly protected, whereas
the bulk of the film approaches the limiting value of 0 %
˚
surface maintains ,100 % D-t-Boc-CM4R (,2.5 ꢂ 10ꢃ4
at lower reaction temperatures until the surface layer is depro-
tected to 45 % D-t-Boc-CM4R at 808C. These data are only
observations after fixed reaction times, but the observations of
such a discrete layer at the surface with an interfacial width were
unexpected.
protection (Q2c ꢁ 1 ꢂ 10ꢃ4
A
ꢃ2) at the highest PEB temperature.
As the film reaction extent increases, the fraction of surface
layer (slab thickness) to total film thickness increases with PEB
temperature: 608, 658, 708C, and 758 as 9, 12, 17 and 18 %
respectively. Between 608 and 758C after 60-s reaction time, the
The earliest analysis of t-Boc-protected polymer resists con-
cludedthatairbornecontaminantsquenchthephotoacid-catalyzed
reaction at the near-surface, leading to T-topping and closure[39]
defects. This was exacerbated by a prolonged post-exposure delay
as found in poly(t-butoxycarbonyloxystyrene-sulfone)[47] and
poly(4-t-butoxycarbonyloxystyrene) (PBOCSt)[39,48] resists that
use the same protecting group as in this study and are also
phenolic-based, as in CM4R. However, PAG surface segregation
may also enhance the surface reaction, relative to the bulk, owing
to higher concentration of photoacid at the surface in PBOCSt.[35]
PAGsegregationmaybeobservedbyX-rayreflectivityrelyingon
the mass density contrast using iodonium-containing ionic PAGs
atloadingsupto30%bymass.[49] However,theneutronscattering
surface-layerQ2c didnotchange,withaverageQ2c (2.49 ꢀ 0.025) ꢂ
10ꢃ4
A
ꢃ2 of thickness 7.5 ꢀ 1.2nm with uncertainties shown as
˚
one standard deviation. However, at 808C, the surface-layer Qc2
reduced to 2.11 ꢂ 10ꢃ4
A
ꢃ2 with thickness 4.5 nm, suggestive of
significant deprotection and 13 % fraction of surface layer. The
˚
corresponding bulk of the film had Q2c values of 2.39 ꢂ 10ꢃ4
,
for
2.3 ꢂ 10ꢃ4, 2.1 ꢂ 10ꢃ4, 1.97 ꢂ 10ꢃ4 and 1.23 ꢂ 10ꢃ4
A
ꢃ2
˚
PEB temperatures 608, 658, 708, 758 and 808C respectively.
Therefore, the bulk of the film proceeds with increasing reaction
extent at higher temperatures. Fig. 1 shows the experimental
average neutron Q2c for the film versus the average degree of
protection quantified by FT-IR. The agreement between the

RESEARCH FRONT
1068
V. M. Prabhu et al.
(a) 0.008
(b)
π*
C ϭ O
no PAG
PEB 70ЊC for 60 s
PEB 75ЊC for 60 s
PEB 80ЊC for 60 s
3.0
2.5
2.0
1.5
1.0
0.5
0.0
0.006
0.004
0.002
0.000
π*
C ϭ C
noPAG
PEB 65 ЊC for 60 s
PEB 70 ЊC for 60 s
PEB 75 ЊC for 60 s
PEB 80 ЊC for 60 s
680
690
700
710
720
730
740
750
280
285
290
295
300
305
Photon energy (eV)
Photon energy (eV)
Fig. 3. Near-edge X-ray absorption fine structure (NEXAFS) (a) F K-edge showing presence of F–C bonds (,693 eV) content due to photoacid generator
(PAG) with increase in relative concentration with post-exposure bake (PEB) temperature. (b) C K-edge showing loss of C¼O (,290.3 eV) from t-Boc
protecting groups with increasing PEB temperature and concurrent increase in C¼C (,285 eV) due to increase in local concentration of aromatic species from
calix[4]resorcinarene.
length density contrast between protected and partially depro-
tected D-t-Boc-CM4R makes the surface layer sensitive to the
protection level rather than PAG. On UV exposure, the hydroge-
nous triphenyl groups of the PAG leave the film, with perfluor-
obutanesulfonic acid remaining. Characterization by NEXAFS
at ꢃ150V entrance-grid bias (EGB) on unexposed films shows
that PAG is present at the surface in levels up to 10.6% volume
fraction via a linear least-squares fit of the C K-edge spectrum
(Accessory Publication, Fig. S3), based on spectra normalized in
the post-resonantregion. Thequality offit isthusbasedon relative
peak intensities and notthe absoluteintensity; the fit uncertainty is
nevertheless extremely small because the spectral shapes of the
components are quite different. This result proves that PAG is
initially within the top 1 to 6 nm.
NEXAFS measurements at the fluorine K-edge show that
photoacid is present at the surface after the post-exposure baking
(Fig. 3a). The control film without any PAG of D-t-Boc-CM4R
shows no F bonds. Therefore, the PAG-containing films, the
same as those measured by neutron reflectivity, show F content
(F–C bonds E 693 eV) at the surface after UV exposure and
post-exposure baking. The aromatic groups of the PAG leave the
film after UV exposure with PFBS remaining in the film. As the
post-exposure baking temperature increased, a relative increase
in partial electron yield (PEY) was observed due to the increase
in surface PFBS concentration caused by the volume loss by
deprotection of D-t-Boc-CM4R. As the reaction occurs, the
volatile D-t-Boc deprotection products increase the local PFBS
concentration, as shown by the carbon K-edge data of Fig. 3b.
The C K-edge shows a loss of 1s - p*C¼O (,290.3 eV) with
increasing post-exposure baking temperature, with concurrent
increase in C 1s - p*C¼C transition (,285 eV) due to an
increase in concentration of aromatic species from CM4R. The
largest decrease in C¼O content occurs between 758 and 808C,
consistent with the observations of Fig. 2. The NEXAFS data
shown in Fig. 3b are sensitive to the top 6 nm that includes the
surface layer observed by neutron reflectivity. These NEXAFS
data were used as independent data to refine the neutron
reflectivity model and served as a boundary condition to aid in
the model fitting.
A difference between these D-t-Boc-CM4R materials and
polymer resists is the ability for the 100 %-protected CM4R to
crystallize because of the purification of the ccc stereoisomer
form. On deprotection, a distribution of ccc and ctt stereoi-
somers may disrupt crystallization. GIXD provides surface
sensitivity to the crystallinity and orientational order. When a
highly collimated beam of X-rays are incident on a thin film
below the film critical angle, an evanescent wave propagates
along the film with depth that depends on the incidence angle a.
Therefore, the near-surface structure may be measured with
sensitivity distinct from the bulk depending on the grazing
angle. The measurement of the diffracted intensity by an area
detector provides reciprocal space information about the struc-
ture along the film thickness as well as the lateral order through
intensity as a function of scattering vectors qz and qxy. The same
thin films were examined by GIXD in Fig. 4 for the lowest
(a ¼ 0.068, surfaces-sensitive) and largest (a ¼ 0.128, bulk-sen-
sitive) grazing angle.
A control film with no PAG (no post-exposure bake) does not
show any surface order, but a weak diffuse diffraction ring due
to the amorphous packing of D-t-Boc-CM4R (Fig. 4) that is also
prominent in the bulk scan. At 658C, the bulk reaction occurs;
however, GIXD only shows a weak diffuse ring and does not
show surface structure. Only at temperatures greater than 658C
is the surface order observed, as seen by the weak diffraction arc
in the 2D images along with the diffuse ring as summarized by
the line-scan averages in Fig. 5. At 708 to 808C, diffraction
peaks appear as arcs centred at a well-defined qz, as shown for
the 758C data. The fact that the diffraction in the surface-
sensitive scans (a E 0.068) appears as arcs and not rings means
the crystals near the surface are preferentially oriented. The
diffraction peaks appear only along qz and not as distinct spots
within qxy, consistent with a layered structure lacking in-plane
order. Incident-angle-dependent NEXAFS spectra are unable to
show significant molecular orientation at the surface (Accessory

RESEARCH FRONT
Structure of calix[4]resorcinarene resist thin films
1069
α ϭ 0.06Њ
α ϭ 0.12Њ
75 ЊC
1.5
1.5
1.0
0.5
0.0
1.0
0.5
0.0
Ϫ2
Ϫ2
Ϫ2
Ϫ1
Ϫ1
Ϫ1
0
1
1
1
2
2
2
Ϫ2
Ϫ2
Ϫ2
Ϫ1
Ϫ1
Ϫ1
0
1
1
1
2
2
2
qxy (ÅϪ1
)
)
)
qxy (ÅϪ1
)
)
)
65 ЊC
1.5
1.0
0.5
0.0
1.5
1.0
0.5
0.0
0
0
qxy (ÅϪ1
qxy (ÅϪ1
No PEB
1.5
1.0
0.5
0.0
1.5
1.0
0.5
0.0
0
0
qxy (ÅϪ1
qxy (ÅϪ1
Fig. 4. Grazing-incidence X-ray diffraction for different post-exposure bake temperatures and a control film without photoacid generator and no post-
exposure bake (PEB). Area detector shows sensitivity to the surface a E 0.068 and the bulk a E 0.128.
Publication, Fig. S4), owing to the lack of a common ring
plane or linear backbone, which eliminates any significant
aggregate transition dipole moment for the prominent reso-
nances. The surface diffraction peak intensity increases with
the post-exposure baking temperature (Fig. 5a), even though
there is not substantial surface deprotection occurring at the
surface below 808C. The surface diffraction peaks are relatively
sharp and have a vertical coherence length of 27, 20 and 37 nm
for the 708, 758 and 808C post-exposure baking based on the
Scherrer equation. This coherence length is larger than the
surface-layer thickness measured in neutron reflectivity. It
should be noted that the scattering intensity of the surface
crystals is relatively weak and that concentrations of a few per
cent would not affect the neutron reflectivity profiles. At lower
reaction temperatures, the lack of surface order by GIXD may be
due to too short reaction times. For instance, a longer reaction
time, while leading to low deprotection extent, may provide
sufficient kinetics of crystallization that become faster at higher
temperatures.
At the highest grazing angle (0.128), the X-rays penetrate
through the entire film and represent a film-average structure as
the signal from the surface is reduced. At all post-exposure
baking temperatures, diffuse rings are observed indicative of
amorphous films with substantially reduced diffraction from the
crystals, consistent with a greater contribution of the bulk to the
scattering signal. The low-q diffuse ring appears at q* E 0.54
and 1.3 Aꢃ1, which correspond to characteristic lengths (2p/q*)
˚
of 1.16 and 0.48 nm respectively. As the bulk reaction occurs,
the low-q amorphous ring disappears, suggesting that this
characteristic length is associated with the D-t-Boc-protected
upper rim of CM4R, as at 808C, the reaction is near complete
within the bulk, whereas the high-q peak broadens with peak
position relatively unchanged, suggestive of structural origin
within the aromatic crown of the CM4R with little dependence
on degree of D-t-Boc substitution. The decrease in intensity
could be due to the loss in contrast via loss in O content from the
volatile protecting group products (CO2). The mean distance
between methyl-substituted bridging carbons was ,0.51 nm,[50]

RESEARCH FRONT
1070
V. M. Prabhu et al.
(a)
60
(b)
(a) 1.8
α ϭ 0.06Њ
No Exp
PEB 75 ЊC for 60s
PEB 80 ЊC for 60s
65 °C
70 °C
75 °C
80 °C
1.6
1.4
1.2
1.0
0.8
0.6
50
40
30
20
10
0
Highly-protected surface
layer with preferential
orientation
Surface layer
partially
deprotected with
preferentially orientation
Bulk reaction
extent,
amorphous
Bulk reaction
extent,
amorphous
0.2
0.4
0.6
0.8
q (ÅϪ1
1.0
1.2
1.4
1.6
)
(b)
α ϭ 0.12°
No Exp
65 °C
70 °C
75 °C
80 °C
6
5
4
3
2
Non-uniform
deprotection near
substrate
Completely deprotected
as substrate
0
2
4
6
8
0
2
4
6
8
Average D-t-Boc protection
Fig. 6. Composite hypothesis using the neutron reflectivity depth profile
plotted as average number of D-t-Boc-substituted CM4R (degree of protec-
tion) for two different post-exposure bake (PEB) temperatures.
dependence of the D-t-Boc protection. The surface of the films
forms a layer that contains a mixture of preferentially ordered
and oriented D-t-Boc-CM4R, but in the presence of photoacid
and amorphous D-t-Boc-CM4R, as shown by NEXAFS. Be-
tween 758 and 808C, bulk deprotection occurs, enabled by
dispersed photoacid, as well as surface deprotection such that
the average surface deprotection level decreases, but maintains
a fraction of the surface ordered 100 % D-t-Boc-CM4R. At
this point, we cannot quantify the degree of crystallinity at the
surface, but these data are the first to show that the surface
order persists as a discrete layer.
The structure of the ordered phase appears to have order
primarily along the film thickness (qz). The molecular crystal
structure is not known for these materials and comparison with
the unprotected CM4R may be erroneous because the protection
groups could change the unit cell dimensions that contain
a different number of D-t-Boc-CM4R and possibly be
co-crystallized with solvent. The data do imply a first-order
(hidden by the beam stop) d-spacing of 4.7 nm owing to the
observed second, third (highest-intensity) and fifth order peaks.
The non-monotonic decay of intensity versus peak order is
consistent with a complex structure in the unit cell, which would
be consistent with multiple molecules per unit cell. A higher-
order structure such as a lamellar phase cannot be ruled out, if
the repeat is larger than that of the molecular unit-cell dimen-
sion, considering the one-dimensional order with lack of in-
plane structure. However, the persistence of the diffraction at
the surface with reduced presence in the bulk along with the
0.2
0.4
0.6
0.8
q (ÅϪ1
1.0
1.2
1.4
1.6
)
Fig. 5. Grazing incidence X-ray diffraction (GIXD) line scan averages
along qz cut for the (a) surface, and (b) bulk incidence angles. Note that in the
GIXD geometry, the specular conditions are not met and the axis labeled qz
is several degrees away from the sample normal dependent on the q.
as found in CM4R co-crystals. It is hypothesized that the
presence of the surface order may exclude photoacid generator,
such that deprotection is kinetically limited by the less accessi-
ble t-Boc groups within the ordered domains and reduced
average diffusivity of photoacid that remains blended with
amorphous D-t-Boc-CM4R material at the surface. On average,
the kinetics of deprotection are suppressed by the presence of the
ordered and oriented D-t-Boc-CM4R.
The depth profiles of neutron reflectivity combined with
the GIXD data allow us to develop a refined hypothesis for the
lack of surface deprotection in these purified D-t-Boc-CM4R
resist PAG blends. Fig. 6 shows a composite plot for the
film thickness dependence on average D-t-Boc protection
level. The solid lines show the surface, bulk and substrate

RESEARCH FRONT
Structure of calix[4]resorcinarene resist thin films
1071
estimate of the crystal size are suggestive of a population of
crystals that extend from the surface layer into the bulk of the
film. The relatively large crystals could contribute to the
observed interfacial width between the bulk of the film and
surface layer via a heterogeneous interface of extended crystals
of varying length within a deprotected matrix, rather than a
smooth compositional gradient of protection level. Off-specular
reflectivity could probe such a buried interface.
as the dashed lines in Fig. 1 for per-deuterated t-Boc-protected
CM4R. The mass density was determined by X-ray reflectivity
on pure-component thin film control samples without PAG. The
box symbols are the experimentally determined average SLD for
partially reacted thin films with an average deprotection extent
determined by infrared spectroscopy. The experiment and model
agreement allows a mapping of SLD to average fraction of
D-t-Boc-CM4R protection level. Uncertainties are calculated as
theestimatedstandarddeviationfromthemean. Inthecasewhere
the limits are smaller than the plotted symbols, the limits are
removed for clarity. Estimation of the total error in the extracted
fit parameters (SLD and film profiles) is difficult as the precision
of the fit parameters, determined by the inverse of the curvature
matrix is, in general, less than the significant figures reported.[51]
Experimental
Materials and Film Processing
tert-Butoxycarbonyloxy-protected calix[4]resorcinarene mo-
lecular resists were synthesized with deuterium-substituted
tert-butoxycarbonyl (D-t-Boc) protecting groups (Fig. 1)
on each of the eight CM4R hydroxyl moieties. The PAG used
was TPS-PFBS (Sigma–Aldrich). Solutions of D-t-Boc-CM4R/
TPS-PFBS were prepared for spin-coating with 5 % PAG by
mass of solids in toluene at a concentration of 2 % by mass. This
solution was spin-coated onto clean silicon wafers (76 mm di-
ameter, 800 mm thick, Virginia Semiconductor Inc.) followed by
a post-apply bake (PAB) for 60 s at 808C with a Brewer Science
CEE hotplate with vacuum contact. The initial film layer
thickness before reaction was ,100 nm. Subsequently, a dose of
,150 mJ cmꢃ2 from an Oriel UV exposure system with 248-nm
broadband radiation activated all the PAG. Post-exposure bake
was performed at temperatures from 608 to 808C for 60 s.
After the post-exposure bake, a noticeable film thickness change
was observed by the change in colour of the reflected ambient
light. The average deprotection level of D-t-Boc-CM4R was
quantified by the loss in C¼O stretch (nC¼O E 1760 cmꢃ1) by
FT-IR using a Bomem FTLA 2000 instrument in reflection
mode at 16-cmꢃ1 resolution as described in detail elsewhere.[28]
Near-edge X-ray Absorption Fine Structure Spectroscopy
Near edge X-ray absorption fine structure spectroscopy mea-
surements were conducted at the NIST/Dow soft X-ray material
characterization facility, beamline U7A of the National Syn-
chrotron Light Source at Brookhaven National Laboratory. In a
NEXAFS experiment, tunable soft X-rays are preferentially
absorbed by the sample (characteristic depth of ,200 nm) when
the incident radiation is at the appropriate energy to allow the
excitation of a core-shell electron of a specific element (C, N, O
or F) to a chemical bond-specific unoccupied molecular orbit-
al.[52] Owing to the well-defined energy gap associated with a
core shell to unoccupied orbital transition, NEXAFS is sensitive
to the bonding characteristics of the element, giving a discrete
peak for each chemical bond. Auger electrons and fluorescence
photons are emitted when the excited core hole from the irra-
diated sample relaxes. Auger electrons emitted from deep
(.10 nm) within the film undergo multiple inelastic scattering
and lose their kinetic energy within the film and hence cannot
escape the surface potential to reach the detector. In contrast,
electrons originating from near the top (1 to 6 nm for carbon K-
edge electron yield spectra) of the film surface have sufficient
kinetic energy to escape the surface potential. Electrons that
escape the surface potential will have different final kinetic
energies on detection depending on their inelastic energy loss
(depth of creation). By applying a negative voltage EGB at the
PEY detector, electrons of low kinetic energy can be rejected.
As the negative EGB is gradually increased, lower-kinetic-
energy electrons are discriminated against and the effective
electron yield sampling depth gets closer to the film surface.
This scheme uses EGB from ꢃ50 to ꢃ250 V, probing the top 6 to
1 nm respectively. The spectra were collected with the incident
beam (Io) at the magic angle (54.78) relative to the sample to
remove any polarization dependence of the NEXAFS intensi-
ties. All PEY spectra in this paper were Io-normalized for beam
instabilities and monochromator abortion features using the
total yield of clean gold Io mesh placed in the incident beam
before the sample. Carbon K-edge spectra were collected in PEY
mode with a grid bias of ꢃ50 V. Spectra fitted to a composition
ratio were pre-edge normalized via subtraction to make the
average intensity in the range of 280 to 283 eV zero. The stan-
dard uncertainty in PEY is ꢀ2 % and photon energy is ꢀ0.2 eV.
Neutron Reflectivity
Specular neutron reflectivity was performed on the NG7 hori-
zontal cold neutron reflectometer at the NIST Center for Neu-
tron Research. The reflectivity was normalized by the incident
beam intensity and measured as a function of the wave vector
(Q) normal to the film, Q ¼ 4p/lsiny, where l is the incident
˚
neutron wavelength of 4.75 A and y is the specular angle
of reflection. The specular reflected intensity provides a
nanometre-resolution depth profile of the film due to neutron
scattering length density variations (contrast). The scattering
length density (SLD ¼ b/v) is quantified by the total scatter-
P
ing length (b ¼ biover all atomic elements per molecule)
2
i
within molecular volume v and often reported as Qc ¼ 16pb=v.
Although the atomic scattering length (bi) varies from
element to element, a large scattering-length difference occurs
between hydrogen (bH ¼ ꢃ0.374 ꢂ 10ꢃ12 cm) and deuterium
(bD ¼ 0.667 ꢂ 10ꢃ12 cm) isotopes. Therefore, the contrast in a
neutron reflectivity experiment may be enhanced by deuterium
substitution to measure composition profiles with nanometre
resolution. The reflectivity data were fitted to the results cal-
culated from the modelled depth profiles using the Parratt
algorithm in units of SLD using the NIST Reflpak software.[46]
In general, this approach uses successive slab layers of constant
SLD with interfaces smeared by a Gaussian function. By vary-
ing the slab fit parameters (SLD, absorption, thickness and
roughness), a multilayered model can be established to deter-
mine the best fit with the x2 statistic. The SLD may be calculated
knowing the chemical composition and mass density. Calcula-
tions for the SLD versus number of t-Boc per CM4R are shown
Grazing-incidence X-ray Diffraction
Grazing incidence X-ray diffraction was performed on beam
line 11–3 at Stanford Synchrotron Radiation Lightsource
(SSRL) with photon energy of 12.73 keV. The diffraction pat-
terns were recorded using a two-dimensional plate detector
(MAR-345) with a spatial resolution of 150 mm (2300 ꢂ 2300)

RESEARCH FRONT
1072
V. M. Prabhu et al.
pixels that was located at a distance of 400.8 mm from the
sample centre. The incidence angle was optimized such that
signal-to-background ratio was maximized and typically, the
angle between the incident beam and the sample surface was
varied between 0.068, 0.088, 0.108 and 0.128 to provide depth
sensitivity. The surface sensitivity is highest at 0.068.
[4] W. Hinsberg, F. A. Houle, J. Hoffnagle, M. Sanchez, G. Wallraff,
M. Morrison, S. Frank, J. Vac. Sci. Technol. B 1998, 16, 3689.
doi:10.1116/1.590392
[5] A. Pawloski, A. Acheta, I. Lalovic, B. LaFontaine, H. Levinson,
Proceedings of the SPIE, Advances in Resist Technology and Proces-
sing XXI 2004, 5376, 414.
[6] H. Ito, J. Polym. Sci. A Polym. Chem. 2003, 41, 3863. doi:10.1002/
POLA.10963
[7] C. Willson, H. Ito, J. Frechet, T. Tessier, F. Houliham, J. Electrochem.
Conclusions
Soc. 1986, 133, 181. doi:10.1149/1.2108519
Advances to novel photosensitive-materials architectures are
required to extend photolithography to feature sizes below
22 nm. Molecular resists offer one possibility by reducing the
resist size while maintaining processability. Model purified ccc
stereoisomer D-t-Boc-CM4R exhibits thin film structure not
observed by polymer resists. A reduced surface extent of reac-
tion appears in spin-cast PAG-containing films not caused by
PAG depletion. The reduced reaction extent was aided by an
ordered surface phase. It is not understood what determines the
length scale of the surface layer of between 4 to 7 nm, rather than
the entire film. However, it is a kinetic transient that appears for
short reaction times. Further, the role of spin-coating solvent and
plasticization by volatile reaction products at elevated tem-
peratures cannot be ruled out, as the degree of surface-order
diffraction intensity increased with average reaction extent.
Although typical photoresist formulations do not contain such
highly purified isomers, these monodisperse materials are
model systems, especially as polydispersity was shown to affect
patternability[42,43] and is a design criterion. The applications of
pure stereoisomer CM4R are also important as additives to
polymer resists formulations.[53] Therefore, the structure of
blends of polymer and molecular resists may offer important
practical remedies in photolithography.
[8] F. A. Houle, W. D. Hinsberg, M. I. Sanchez, J. A. Hoffnagle, J. Vac.
Sci. Technol. B 2002, 20, 924. doi:10.1116/1.1475985
[9] G. M. Wallraff, W. D. Hinsberg, Chem. Rev. 1999, 99, 1801.
doi:10.1021/CR980003I
[10] J. H. Kim, Y. H. Kim, S. M. Chon, T. Nagai, M. Noda, Y. Yamaguchi,
Y. Makita, H. Nemoto, J. Photopolym. Sci. Technol. 2004, 17, 379.
doi:10.2494/PHOTOPOLYMER.17.379
[11] G. M. Schmid, M. D. Stewart, V. K. Singh, C. G. Willson, J. Vac. Sci.
Technol. B 2002, 20, 185. doi:10.1116/1.1431954
[12] M. D. Stewart, H. V. Tran, G. M. Schmid, T. B. Stachowiak, D. J.
Becker, C. G. Willson, J. Vac. Sci. Technol. B 2002, 20, 2946.
doi:10.1116/1.1523027
[13] J. A. Hoffnagle, W. D. Hinsberg, M. I. Sanchez, F. A. Houle, Opt. Lett.
2002, 27, 1776. doi:10.1364/OL.27.001776
[14] International Technology Roadmap for Semiconductors 2009 Edition
2009, Lithography Section, p. 9. Available at: http://www.itrs.net,
verified July 2011.
[15] T. Itani, H. Yoshino, M. Fujimoto, K. Kasama, J. Vac. Sci. Technol. B
1995, 13, 3026. doi:10.1116/1.588315
[16] W. D. Hinsberg, F. A. Houle, M. Sanchez, M. Morrison, G. M.
Wallraff, C. Larson, J. Hoffnagle, P. Brock, G. Breyta, Proceedings
of the SPIE, Advances in Resist Technology and Processing XVII 2000,
3999, 148.
[17] S. L. Ablaza, J. F. Cameron, G. Y. Xu, W. Yueh, J. Vac. Sci. Technol. B
2000, 18, 2543. doi:10.1116/1.1288136
[18] H. Iwasaki, T. Itani, M. Fujimoto, K. Kasama, Proc. of SPIE 1994,
2195, 164.
Accessory Publication
[19] B. D. Vogt, S. Kang, V. M. Prabhu, E. K. Lin, S. K. Satija,
K. Turnquest, W. Wu, Macromol 2006, 39, 8311. doi:10.1021/
MA061209L
[20] N. Sundararajan, C. Keimel, N. Bhargava, C. Ober, J. Opitz, R. D.
Allen, G. Barclay, G. Xu, J. Photopolym. Sci. Technol. 1999, 12, 457.
doi:10.2494/PHOTOPOLYMER.12.457
Accessory neutron reflectivitity and near-edge X-ray absorption
fine structure data on bilayers and single layer films for this
article are available on the Journal’s website.
Acknowledgements
[21] T. H. Fedynyshyn, D. K. Astolfi, A. Cabral, J. Roberts, Proc. SPIE
We acknowledge Michael Toney (SSRL) for assistance in GIXD experi-
mentation. This work was supported by a cooperative research and devel-
opment agreement (CRADA 1893) between Intel Corporation and NIST. A
portion of this research was carried out at Oak Ridge National Laboratory’s
Center for Nanophase Materials Sciences under User Proposal 2008–286,
and was sponsored by the Scientific User Facilities Division, Office of Basic
Energy Sciences, US Department of Energy. Cornell Nanoscale Science and
Technology Facility (CNF), Cornell Center for Materials Research (CCMR)
and a grant from the National Science Foundation (DMR-0518785) are
acknowledged for partial support of this work.
2007, 6519, X5190.
[22] Y. Fukushima, T. Watanabe, R. Ohnishi, H. Kinoshita, H. Shiotani,
S. Suzuki, M. Hayakawa, Y. Endo, T. Yamanaka, S. Yusaa, J.
Photopolym. Sci. Technol. 2007, 20, 419. doi:10.2494/PHOTOPOLY
MER.20.419
[23] R. D. Allen, P. J. Brock, Y. H. Na, M. H. Sherwood, H. D. Truong,
G. M. Wallraff, M. Fujiwara, K. Maeda, J. Photopolym. Sci. Technol.
2009, 22, 25. doi:10.2494/PHOTOPOLYMER.22.25
[24] R. Gronheid, A. V. Pret, B. Rathsack, J. Hooge, S. Scheer, K. Nafus,
H. Shite, J. Kitano, Journal of Micro-Nanolithography, MEMS and
MOEMS 2011, 10, 013017.
[25] D. L. VanderHart, V. M. Prabhu, A. De Silva, N. M. Felix, C. K. Ober,
J. Mater. Chem. 2009, 19, 2683. doi:10.1039/B816290E
[26] H. Oizumi, Y. Tanaka, T. Kumise, D. Shiono, T. Hirayama, H. Hada,
J. Onodera, A. Yamaguchi, I. Nishiyama, J. Photopolym. Sci. Technol.
2007, 20, 403. doi:10.2494/PHOTOPOLYMER.20.403
[27] T. Owada, H. Shiotani, K. Aoyama, T. Kashiwamura, M. Shibata,
T. Takeya, H. Oizumi, T. Itani, Proc. SPIE 2010, 7636, 32.
doi:10.1117/12.846069
[28] S. H. Kang, W. L. Wu, K. W. Choi, A. De Silva, C. K. Ober, V. M.
Prabhu, Macromol 2010, 43, 4275. doi:10.1021/MA902548A
[29] J. Sha, J. K. Lee, S. H. Kang, V. M. Prabhu, C. L. Soles, P. V.
Bonnesen, C. K. Ober, Chem. Mater. 2010, 22, 3093. doi:10.1021/
CM9038939
Certain commercial equipment and materials are identified in this paper
in order to specify adequately the experimental procedure. In no case does
such identification imply recommendations by the National Institute of
Standards and Technology nor does it imply that the material or equipment
identified is necessarily the best available for this purpose.
References
[1] H. Ito, Adv. Polym. Sci. 2005, 172, 37.
[2] A. G. Yu, H. P. Liu, J. P. Blinco, K. S. Jack, M. Leeson, T. R. Younkin,
A. K. Whittaker, I. Blakey, Macromol. Rapid Commun. 2010, 31,
1449. doi:10.1002/MARC.201000117
[3] K. J. Lawrie, I. Blakey, J. P. Blinco, H. H. Cheng, R. Gronheid, K. S.
Jack, I. Pollentier, M. J. Leeson, T. R. Younkin, A. K. Whittaker,
J. Mater. Chem. 2011, 21, 5629. doi:10.1039/C0JM03288C

RESEARCH FRONT
Structure of calix[4]resorcinarene resist thin films
1073
[30] F. A. Houle, W. D. Hinsberg, M. Morrison, M. I. Sanchez, G. Wallraff,
C. Larson, J. Hoffnagle, J. Vac. Sci. Technol. B 2000, 18, 1874.
doi:10.1116/1.1303753
[42] D. Shiono, H. Hada, H. Yukawa, H. Oizumi, L. Nishiyama, K. Kojima,
H. Fukuda, Advances in Resist Materials and Processing Technology
XXIV 2007, 6519, U5193.
[31] B. Rathsack, K. Nafus, S. Hatakeyama, Y. Kuwahara, J. Kitano,
R. Gronheid, A. V. Pret, Proc. SPIE 2009, 7273, 47. doi:10.1117/
12.814287
[43] D. Shiono, H. Hada, H. Yukawa, H. Oizumi, I. Nishiyama, K. Kojima,
H. Fukuda, Microelectron. Eng. 2007, 84, 1084. doi:10.1016/J.MEE.
2007.01.146
[32] M. D. Smith, C. A. Mack, Proc. SPIE 2003, 5040, 1509. doi:10.1117/
12.485541
[33] M. D. Smith, J. D. Byers, C. A. Mack, Proc. SPIE 2004, 5376, 322.
doi:10.1117/12.537581
[34] R. Gronheid, B. Rathsack, S. Bernard, A. V. Pret, K. Nafus,
S. Hatakeyama, J. Photopolym. Sci. Technol. 2009, 22, 97.
doi:10.2494/PHOTOPOLYMER.22.97
[35] J. L. Lenhart, R. L. Jones, E. K. Lin, C. L. Soles, W. L. Wu, D. A.
Fischer, S. Sambasivan, D. L. Goldfarb, M. Angelopoulos, J. Vac. Sci.
Technol. B 2002, 20, 2920. doi:10.1116/1.1524970
[36] J. L. Lenhart, D. A. Fischer, S. Sambasivan, E. K. Lin, R. L. Jones, C.
L. Soles, W. L. Wu, D. L. Goldfarb, M. Angelopoulos, Langmuir 2005,
21, 4007. doi:10.1021/LA047160Z
[37] V. M. Prabhu, S. Sambasivan, D. Fischer, L. K. Sundberg, R. D. Allen,
Appl. Surf. Sci. 2006, 253, 1010. doi:10.1016/J.APSUSC.2006.03.089
[38] T. Hirayama, D. Shiono, J. Onodera, A. Yamaguchi, H. Fukuda,
Polym. Adv. Technol. 2006, 17, 116. doi:10.1002/PAT.670
[39] W. D. Hinsberg, S. A. Macdonald, N. J. Clecak, C. D. Snyder, Chem.
Mater. 1994, 6, 481. doi:10.1021/CM00040A024
[44] D. Pires, J. L. Hedrick, A. De Silva, J. Frommer, B. Gotsmann,
H. Wolf, M. Despont, U. Duerig, A. W. Knoll, Science 2010, 328,
732. doi:10.1126/SCIENCE.1187851
[45] V. M. Prabhu, S. H. Kang, D. L. VanderHart, S. K. Satija, E. K. Lin,
W. L. Wu, Adv. Mater. 2011, 23, 388. doi:10.1002/ADMA.201001762
[46] P. A. Kienzle, K. V. O’Donovan, J. F. Ankner, N. F. Berk, C. F.
Majkrzak, NCNR Reflectometry Software 2006 Available at: http://
www.ncnr.nist.gov/reflpak, verified July 2011.
[47] O. Nalamasu, E. Reichmanis, J. E. Hanson, R. S. Kanga, L. A.
Heimbrook, A. B. Emerson, F. A. Baiocchi, S. Vaidya, Polym. Eng.
Sci. 1992, 32, 1565. doi:10.1002/PEN.760322104
[48] S. A. MacDonald, W. D. Hinsberg, H. R. Wendt, N. J. Clecak, C. G.
Willson, C. D. Snyder, Chem. Mater. 1993, 5, 348. doi:10.1021/
CM00027A018
[49] T. Fukuyama, T. Kozawa, S. Tagawa, R. Takasu, H. Yukawa, M. Sato,
J. Onodera, I. Hirosawa, T. Koganesawa, K. Horie, Applied Physics
Express 2008, 1, 065004.
[50] R. Kuzmicz, V. Kowalska, S. Domagala, M. Stachowicz, K. Wozniak,
W. Kolodziejski, J. Phys. Chem. B 2010, 114, 10311. doi:10.1021/
JP1015565
[40] F. A. Houle, W. D. Hinsberg, M. I. Sanchez, J. Vac. Sci. Technol. B
2004, 22, 747. doi:10.1116/1.1688351
[51] P. R. Bevington, Data Reduction and Error Analysis for the Physical
[41] E. L. Jablonski, V. M. Prabhu, S. Sambasivan, D. A. Fischer, E. K. Lin,
D. L. Goldfarb, M. Angelopoulos, H. Ito, Proceedings of the SPIE,
Advances in Resist Technology and Processing XXI 2004, 5376,
302–11.
Sciences 1969 (McGraw-Hill: New York).
[52] J. Sto¨hr, NEXAFS Spectroscopy 1992 (Springer-Verlag: New York).
[53] H. Ito, T. Nakayama, M. Sherwood, D. Miller, M. Ueda, Chem. Mater.
2008, 20, 341. doi:10.1021/CM7021483