of 11
1
Supporting Information
Epitaxy: Programmable Atom Equivalents
V
e
rsus
Atoms
Mary X. Wang,
&,†,‡
Soyoung E. Seo,
&,§,‡
Paul A. Gabrys,
Dagny Fleischman,
Byeongdu Lee,
#
Youngeun Kim,
,‡
Harry A. Atwater,
Robert J. Macfarlane,
*,
and Chad A. Mirkin
*,†,§,
,‡
2
Table S1
. DNA sequences used for functionalizing AuNPs and the substrate. Thiolated strands (X
-
SH
)
that had a 3’ propyl thiol
-
modifier were functionalized onto AuNPs. These strands consisted of two of six
ethylene glycol units (denoted as
(EG6)
2
) close to 3’ propyl thiol to increase the flexibility of the DNA.
HS
-
A DNA strands used for nanoparticles and the substrate are identical, and so are Linker A strands.
DNA Type
Sequence (5’
-
3’)
Nanoparticles
HS
-
A
TCA ACT ATT CCT ACC TAC (EG6)
2
-
SH
HS
-
B
TCC ACT CAT ACT CAG CAA (EG6)
2
-
SH
Linker A
GTA GGT AGG AAT AGT TGA A TTT AGT CAC GAC GAG TCA
TT A TTT AGT CAC GAC GAG TCA TT A TTCCTT
Linker B
TTG CTG AGT ATG AGT GGA A TTT AGT CAC GAC GAG TCA
TT A TTT AGT CAC GAC GAG TCA TT A AAGGAA
Duplexer
AAT
GAC TCG TCG TGA CTA AA
Substrate
HS
-
A
TCA ACT ATT CCT ACC TAC (EG6)
2
-
SH
Linker A
GTA GGT AGG AAT AGT TGA A TTT AGT CAC GAC GAG TCA
TT A TTT AGT CAC GAC GAG TCA TT A TTCCTT
Duplexer
AAT GAC TCG TCG TGA CTA AA
Oligonucleotides were synthesized on an
ABI 394 (Applied Biosystems) automated oligonucleotide
synthesizer using standard phosphoramidite chemistry on controlled pore glass (CPG) beads.
Phosphoramidite reagents and CPG beads (GlenUny Support) were purchased from Glen Research
(Sterling, VA). Aft
er synthesis, oligonucleotides were deprotected using a fast deprotection method,
where 1 μmole of synthesized oligonucleotides was mixed with 1 mL of a 1:1 mixture (
v
) of 40
%
aqueous methylamine and 30
% ammonium hydroxide solution, then was allowed to sit
at 25 °C for 2
hours. After deprotection, the solvent containing the DNA was evaporated with nitrogen, and the sample
was filtered through a 0.2 μm pore syringe filter to remove the CPG beads and impurities. To remove
failure strands from the success stra
nds, the DNA was purified using reverse
-
phase high performance
liquid chromatography (Varian RP
-
HPLC) on an Agilent C18 column. After the purification step
via
RP
-
HPLC, the DNA was lyophilized overnight. To cleave the acid
-
labile 4,4’
-
dimethoxytrityl (DMT)
protecting group off the DNA, 1
-
2 mL of 20% acetic acid was added per 1
-
2 μmole columns of dry DNA.
After the solution was allowed to sit for approximately 1 hour, 2 mL of water and 3
-
4 mL of ethyl acetate
were added to remove hydrophobic DMT groups and e
xtract the purified DNA from the solution. All
oligonucleotides were characterized and confirmed by matrix
-
assisted laser desorption/ionization time
-
of
3
flight (MALDI
-
TOF) mass spectrometry to ensure that all molecular weights corresponded to the
theoretica
l masses. The MALDI
-
TOF mass spectrometry matrix was prepared by dissolving 30 mg of 3
-
hydroxypicolinic acid (Fluka) in 1 mL of a 1:1 MeCN:H
2
O solution and mixed with 10 mg of ammonium
citrate dibasic (Sigma
-
Aldrich). 1.5 μL of diluted oligonucleotide was
aliquoted and mixed with 1.5 μL of
this matrix on a steel plate to crystallize prior to MALDI
-
TOF analysis. The absorbance of
oligonucleotides was measured on a Cary 5000 UV
-
Vis
-
NIR spectrophotometer (Agilent) using
calculated extinction coefficients from
Integrated DNA Technologies (IDT) website.
Figure S
1
:
Analysis of SAXS Data
.
a)
Sector averaging to determine degree of epitaxy. The background
signal from the diffuse ring is determined from azimuthally averaging arc A. The signal corresponding to
the
epitaxial NPs is determined from the azimuthally averaging arc B. b) SAXS scattering pattern of a
blank patterned substrate. The scattering intensity comes from the gold posts on the pattern.
The 1D
averaged data from the 2D pattern is shown for the 10
-
lay
er thin film samples assembled at c) 25 °C and
d)
(
T
m
-
4
)
°C and annealed.
Since epitaxial PAEs are oriented in the bcc [100] direction, their in
-
plane scattering produces
strong
signal intensity present as a spot. On the other hand, PAEs that are not epitaxial are randomly oriented in
the z
-
direction, even if they possess bcc symmetry, and their scattering produces a circular ring.
Therefore, by comparing the integrated signal in
tensity from the (110) peak produced by scattering of the
4
epitaxial PAEs to that of the diffuse ring produced by scattering from the non
-
oriented PAEs, we can
determine the degree of epitaxy.
To do this, the center and axis of each
SAXS 2D
scattering
patte
rn
were
aligned
so that each
sample
is averaged in the same way
(Figure S1a)
.
The background signal from the
diffuse ring is determined
by
azimuthally averaging
the 2D scattering pattern
over
arc A
and then fitting
the peak using a Voigt profile
to get
I
arc
A,peak
.
Azimuthally averaging
over
arc B gives the signal intensity
from a combination of 1) PAEs epitaxial with the pattern, 2) the pattern itself (Figures 1
middle
and S1b)
,
and 3)
the background signal. Th
ese
1D data (examples given in Figure S1c an
d d) obtained for the region
corresponding to the bcc (110) peak (Figure 1) w
ere
then fit to a Voigt profile
, giving I
arc B,
(110)peak
.
The
r
elative
degree of epitaxy was then calculated
from the ratio of
signal from the bcc (110) peak and the
corresponding background signal
according to equation S1.
For each sample, the relative contribution
from the lithographically defined pattern remains the same, since
the signal for the diffraction peak that
overlaps
with the region of interest, the (110),
is
entirely encompassed by arc B
for all diffraction
patterns
.
Equation (
S1
)
=
푎푟푐퐵
,
(
110
)
푝푒푎푘
푎푟푐
,
푝푒푎푘
푎푟푐퐵
,
(
110
)
푝푒푎푘
5
Figure S2
: Quantita
tiv
e analysis of FIB
-
SEM
cross
-
section.
Degree
of epitaxy for 10
-
layer thin films as a
function of l
ayer distance from the template, calculated from analysis shown in Figure S5.
Table S2
.
Mean height and roughness for films of varying layer number grown using different depositi
on
conditions.
Layer #
Mean Height (nm)
Roughness (nm)
Roughness (nm)
1. 25 °C deposition
Arithmetic Average
Root Mean Squared
2
130.56
28.46
34.34
5
311.49
38.46
47.91
10
775.95
63.87
82.66
2. 25 °C deposition, annealed at (T
m
-
2) °C
5
321.78
46.96
58.82
10
883.43
79.58
100.09
3. (T
m
-
4) °C deposition, annealed at (T
m
-
2) °C
2
97.22
20.47
25.76
5
170.38
36.44
47.52
10
257.56
29.80
40.31
4. (T
m
-
4) °C deposition, annealed at (T
m
-
2) °C, and intercalated
2
87.36
23.99
30.42
10
220.86
46.35
66.95
Unpatterned substrate
5
226.34
40.93
53.60
6
Table S3.
Growth Conditions for Templated DNA
-
NP Superlattice Thin Films
.
Condition
1
Thin films were grown at 25 °C up to 2, 5, and 10 layers
Condition
2
Thin films were
grown to their full thickness at 25 °C, followed by annealing at (
T
m
-
2)
°C for 15 minutes in Buffer A. Annealing temperatures varied depending on layer
number. Typically, 2
-
layer thin films were annealed at (
T
m
-
6) °C, and 5
-
and 10
-
layer
thin films were an
nealed at (
T
m
-
2) °C due to melting temperature depression observed
for thin films.
Condition
3
The first 2 layers were grown at 25 °C, in which the sample was then annealed at (
T
m
-
2)
°C. For
third and fourth
layers, substrates were immersed in particle solutions at (
T
m
-
6)
°C and annealed at (
T
m
-
4) °C. For the rest of the layers, the thin films were grown at
(
T
m
-
4)
°C and annealed at (
T
m
-
2)
°C.
Condition
4
The growth protocol was same as condition
3
, with the addition of a
two
hour
incubation step in 80
μM
intercalator after each annealing step to ensure complete
intercalation.
Figure S3
: Melting point depression of the templated thin film superlattice as a functio
n of nanoparticle
layer number, as determined by SAXS.
7
Figure S
4
: GISAXS of a) 5
-
layer and b) 10
-
layer thin films grown at
equilibrium conditions
.
On the right
hand side,
the scattering patterns were indexed to bcc crystals with (100) orientation
corresponding to
space group I4/mmm (#139). The higher order peaks evident in the scattering patterns are indicative of
long
-
range order. The high levels of diffuse scattering in the 10
-
layer film are hypothesized to be due to
the thickness of the film, ma
king it difficult for X
-
ray penetration.
Scheme S1
: Demonstration of FIB
-
SEM cut in different plane orientations.
8
Figure S5
: Degree of epitaxy analysis for FIB
-
SEM cross
-
sections of 10
-
layer films grown at a) 25 °C, b)
25 °C and annealed, c) (
T
m
-
4) °C and annealed, and d) (
T
m
-
4) °C, annealed, and in
tercalated. Scale bar is
200 nm.
Photoshop and Matlab were used to track the positions of internal PAEs relative to the positions of the
templated posts. After correcting for tilt, vectors were calcu
lated between adjacent particles in the [001]
direction and plotted atop the cross
-
section of the SEM image.
P
erfectly epitaxial superlattices would
display vectors completely vertical from the posts up throughout the layers; this direction was taken to be
0° (and displayed as white arrows on the overlaid image). Any deviation in z
-
direction was calculated in
terms of degrees where the limits are therefore
-
90 to 90° (with arrows becoming increasingly blue or red,
respectively, as they deviate); note that t
his vector exists as the projection onto the plane of the cross
-
section, not as a 3D vector. The first row of vectors was then aggregated and the standard deviation (σ) of
their angles was calculated using Matlab. If the vectors were all vertically aligned
,
i.e.
the superlattices
were perfectly epitaxial, σ = 0. However, if the vectors were completely random,
i.e.
the superlattices
were completely disordered, the maximum standard deviation for
this system would be = √ ((
-
90
-
90)
2
/4)
= 90. Therefore, to cal
culate X
A
or “Degree of Epitaxy” for that row of vectors such that 1 is perfect
epitaxy and 0 is completely disordered, X
A
= (90
-
σ)/90. This process was repeated for each row of
vectors in the superlattices and plotted in Figure S2. These data corroborat
e the SAXS results on effects
of deposition protocol on epitaxy. While SAXS data provides an averaged information on degree of
epitaxy, the analysis on FIB
-
SEM highlights the waning force of epitaxy the template exhibits over the
PAEs as a function of laye
r number. PAEs within the bulk crystal are more tightly bound and networked
with neighboring particles, limiting their vibrational motion. PAEs near the surface of the thin film have
fewer neighboring particles, thus have many near
-
equilibrium positions to
easily oscillate between. This
9
indicates the importance of annealing each layer so that as few defects as possible in the surface layer
exist upon the deposition of the subsequent layer to avoid trapping defects in the bulk crystal.
Figure S6
:
Epitaxial growth of DNA
-
functionalized nanoparticle thin films at 2 and 5 layers is observed
when they are assembled at (
T
m
-
4) °C and annealed. SEM, SAXS, and FIB
-
SEM show crystalline,
epitaxial thin films at 10 layers of nanoparticles. Scale bars for SEM
and FIB
-
SEM are 500 nm and 200
nm, respectively.
10
Figure S7
: FIB
-
SEM cross
-
sections of a) 10
-
layer film grown on a patterned substrate at (
T
m
-
4) °C, b)
10
-
layer film grown on a patterned substrate at 25 °C, and c) a 5
-
layer film assembled on an
unpatterned
substrate and annealed at (
T
m
-
2) °C. Scale bars are 200 nm.
Figure S8
: Intercalated 10
-
layer thin film presenting roughened surface morphology (SEM) and defect
propagation along the z
-
axis (FIB
-
SEM). SAXS was used to determine the degree of e
pitaxy (X
A
= 0.65).
Scale bars for SEM and FIB
-
SEM are 500 nm and 200 nm, respectively.
11
Figure S9
: Polycrystalline 5
-
layer thin film grown on an unpatterned substrate. Scale bar is 1 μm.