Ultraviolet surface plasmon-mediated low temperature hydrazine decomposition
Siying Peng, Matthew T. Sheldon, Wei-Guang Liu, Andres Jaramillo-Botero, William Andrew Goddard III, and
Harry A. Atwater
Citation: Applied Physics Letters
106
, 023102 (2015); doi: 10.1063/1.4905593
View online: http://dx.doi.org/10.1063/1.4905593
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/106/2?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
High-performance AlGaN metal–semiconductor–metal solar-blind ultraviolet photodetectors by localized surface
plasmon enhancement
Appl. Phys. Lett.
106
, 021112 (2015); 10.1063/1.4905929
Strong coupling between Rhodamine 6G and localized surface plasmon resonance of immobile Ag nanoclusters
fabricated by direct current sputtering
Appl. Phys. Lett.
102
, 143112 (2013); 10.1063/1.4801633
Sensing properties of infrared nanostructured plasmonic crystals fabricated by electron beam lithography and
argon ion milling
J. Vac. Sci. Technol. B
30
, 06FE02 (2012); 10.1116/1.4767274
Tailoring plasmon resonances in the deep-ultraviolet by size-tunable fabrication of aluminum nanostructures
Appl. Phys. Lett.
101
, 081110 (2012); 10.1063/1.4747489
Molded plasmonic crystals for detecting and spatially imaging surface bound species by surface-enhanced
Raman scattering
Appl. Phys. Lett.
94
, 243109 (2009); 10.1063/1.3155198
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.215.70.231 On: Tue, 20 Jan 2015 15:53:04
Ultraviolet surface plasmon-mediated low temperature hydrazine
decomposition
Siying Peng (
彭
斯
颖
),
1
Matthew T. Sheldon,
1
Wei-Guang Liu,
2
Andres Jaramillo-Botero,
2
William Andrew Goddard III,
2
and Harry A. Atwater
1
1
Thomas J. Watson Laboratories of Applied Physics, California Institute of Technology, Pasadena, California
91125, USA
2
Materials and Process Simulation Center, California Institute of Technology, Pasadena, California 91125,
USA
(Received 27 August 2014; accepted 26 December 2014; published online 12 January 2015)
Conventional methods require elevated temperatures in order to dissociate high-energy nitrogen
bonds in precursor molecules such as ammonia or hydrazine used for nitride film growth. We report
enhanced photodissociation of surface-absorbed hydrazine (N
2
H
4
) molecules at low temperature by
using ultraviolet surface plasmons to concentrate the exciting radiation. Plasmonic nanostructured
aluminum substrates were designed to provide resonant near field concentration at
k
¼
248 nm
(5 eV), corresponding to the maximum optical cross section for hydrogen abstraction from N
2
H
4
.
We employed nanoimprint lithography to fabricate 1 mm
1 mm arrays of the resonant plasmonic
structures, and ultraviolet reflectance spectroscopy confirmed resonant extinction at 248 nm.
Hydrazine was cryogenically adsorbed to the plasmonic substrate in a low-pressure ambient, and
5 eV surface plasmons were resonantly excited using a pulsed KrF laser. Mass spectrometry was
used to characterize the photodissociation products and indicated a 6.2
overall enhancement in
photodissociation yield for hydrazine adsorbed on plasmonic substrates compared with control sub-
strates. The ultraviolet surface plasmon enhanced photodissociation demonstrated here may pro-
vide a valuable method to generate reactive precursors for deposition of nitride thin film materials
at low temperatures.
V
C
2015 AIP Publishing LLC
.[
http://dx.doi.org/10.1063/1.4905593
]
Gallium nitride (GaN) and indium gallium nitride
(InGaN) are tunable wide band gap semiconductor materials
with important applications in optoelectronic devices such as
LEDs and photovoltaics.
1
,
2
Conventional growth methods
for these materials include chemical vapor deposition (CVD)
and molecular beam epitaxy (MBE), which both require high
temperatures (
>
500 K) to create atomic nitrogen growth spe-
cies by thermal decomposition of nitride precursor mole-
cules.
3
Alternatively, magnetron sputtering can be used to
grow nitride semiconductors at room temperature.
4
However, the sputtering process utilizes high-energy plasma
ions to bombard nitride precursors, resulting in a high kinetic
energy distribution of atomic nitrogen that limits the crystal-
line quality of the growing film. By contrast, ultraviolet
(UV) radiation can resonantly dissociate nitrogen bonds,
through a pathway that does not produce species with high
kinetic energy or require high temperatures. By generating a
kinetic energy distribution of the dissociated nitrogen species
similar to that achieved in molecular beam epitaxy or chemi-
cal vapor deposition processes, resonant ultraviolet dissocia-
tion of nitrogen precursors may enable low temperature film
growth with superior crystalline quality. Additionally,
because film growth could occur at lower temperatures, such
a process may open doors for synthesis of difficult-to-make
semiconductor materials such as indium rich InGaN, which
is normally hindered due to phase separation into InN and
GaN at elevated growth temperatures.
5
,
6
We utilize the optical confinement properties of surface
plasmons to localize ultraviolet radiation in a surface-adsorbed
hydrazine layer and thus generate a high flux of reactive
nitrogen species athermally, with the aim of overcoming the
limitations of conventional high temperature nitride film
growth, a method of nitride precursor generation that has not
been demonstrated before. Surface plasmons are coherent
oscillations of charge density at the interface of a metal and a
dielectric that confine incident radiation within evanescent
fields of subwavelength extent in the direction perpendicular
to the surface, thereby providing concentration of optical in-
tensity at resonance up to 1000
greater than the free space
intensity in some cases.
7
This remarkable property of surface
plasmons has been studied extensively for various applications
such as surface-enhanced Raman scattering (SERS),
8
surface
enhanced photo-chemistry,
9
–
15
optical tweezing,
16
and nano-
particle therapeutics.
17
In our experiment, the enhancement of
radiation fields via surface plasmon excitation is employed to
maximize resonant ultraviolet decomposition of nitride growth
precursor molecules that are cryogenically condensed at high
concentration on plasmonically active substrates.
We identify hydrazine (N
2
H
4
) as a promising nitride
growth precursor molecule for low temperature film growth.
N
2
H
4
molecules have a dissociation threshold at
k
¼
376.2 nm
and very large optical absorption cross section at 5 eV
(
k
¼
248 nm),
18
,
19
which is compatible with an KrF excimer
laser excitation at 248 nm. Figure
1
shows the potential
energy diagram along the N-H bond in a N
2
H
4
molecule cal-
culated at the level of GVB-RCI (generalized valence bond -
restricted configuration interaction) (restricted double excita-
tion)
20
/cc-pVTZ//B3LYP/6–311
þþ
G**. Optical excitation at
5 eV resonantly excites the N-H bond in the singlet electronic
ground state (S
0
) which we calculated to have an energy bar-
rier of 4.0 eV after including the zero point energy correction,
from a lower roto-vibrational state to a higher roto-vibrational
0003-6951/2015/106(2)/023102/4/$30.00
V
C
2015 AIP Publishing LLC
106
, 023102-1
APPLIED PHYSICS LETTERS
106
, 023102 (2015)
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.215.70.231 On: Tue, 20 Jan 2015 15:53:04
state. The higher roto-vibratio
nal state is sufficiently energetic
to overcome the potential well of the S
0
state, resulting in the
stretch of the N-H bond and ultimately, bond dissociation, and
energizing other degree of freedoms in the molecule with the
excess energy from the surface plasmon (1.0 eV in this case).
We note that the energy gap between the ground state S
0
and
the first excited state T
1
is 8.6 eV, which is too large to be
excited by a 5 eV surface plasmon. We utilize surface plas-
mons to enhance the ultraviolet radiation field that excites the
5 eV dissociation pathway in N
2
H
4
. Ultraviolet surface plas-
mons have been studied in the past to enhance surface photo-
chemical reactions,
9
increase sensitivity of SERS,
8
and
generate photoelectron emission.
21
In the above context, alumi-
num has been identified as a promising ultraviolet plasmonic
material for two reasons: its high bulk plasmon energy in the
deep ultraviolet and low optical loss
21
–
23
entails that it provides
strong ultraviolet field enhancement compared to more con-
ventional plasmonic materials such as gold and silver.
To optimize plasmonic concentration of optical energy at
5 eV and facilitate dissociation of N
2
H
4
, we performed full
wave electromagnetic simulations of surface plasmon modes
in aluminum nanostructures. We identified a periodic alumi-
num grating structure as a simple and promising geometry,
because it enables momentum matching between incident
photons and surface plasmon modes on the metal. Shown in
Figure
2(a)
is the modeled reflectance over energy and wave
number k
g
¼
(
2
p
grating pitch
Þ
, of an aluminum grating where a
4 nm native oxide on the surface (shown in Figure
2(b)
)is
included in the model. The width of the ridges (W in Figure
2
(b)
) of the grating increases as the grating pitch increases.
Both the photonic dispersion and surface plasmon dispersion
are manifested as a valley in reflectance spectrum (blue color
in Figure
2(a)
), indicated by the dashed lines. At lower wave-
number (larger pitch), the surface plasmon approaches the
light line. As the wavenumber increases, the surface plasmon
diverges from the light line and approaches an asymptotic
limit, indicating localization of the optical field on the surface
of the aluminum. In Figure
2(a)
, the reflectance valley indi-
cated by the circles arises from the localized surface plasmon
mode at the ridges of the grating. Similar to the localized
plasmonic mode of a metal nanoparticle, the energy of this
localized surface plasmon mode decreases as W increases
with the grating pitch. The broader reflectance valley at small
wavenumber indicated by the stars corresponds to photonic
modes and reveals the diffraction orders of the aluminum gra-
ting. We performed a systematic parameter variation chang-
ing W, the grating height (H in Figure
2(b)
), and pitch of the
grating to determine the geometry with minimum reflectance
at 248 nm, corresponding to maximum field concentration in
the plasmonic mode. We observe the strongest surface plas-
mon resonance at this wavelength for a grating period of
205 nm, width of 24 nm, and height of 50 nm. The electric
field magnitude of the optimal geometry (Figure
2(b)
) indi-
cates that the optical energy is clearly confined close to the
surface of the metal. By integrating of the electric field mag-
nitude, we observe a near-field enhancement factor of 25
in
the optimized design.
FIG. 1. N
2
H
4
energy curve: Solid black curve (S
0
) is the singlet ground state
for the N-H bond. Inside the potential well of S
0
, the horizontal lines repre-
sent the roto-vibrational states. Optical excitation at 5 eV (free space wave-
length 248 nm) excites the ground state to higher roto-vibrational states and
dissociates the N-H bond. The calculated curves depict the triplet first
excited state (T
1
, solid red), the singlet first excited state (S
1
, dash black),
the triplet second excited state (T
2
, dashed red), the singlet second excited
state (S
2
, dotted black), and the triplet third excited state (T
3
, dotted red).
FIG. 2. (a) Full wave (finite-difference time-domain) simulation of the re-
flectance versus wavenumber, k
g
¼
(
2
p
gra
ti
ng pitch
Þ
, and energy of the aluminum
grating with 4 nm aluminum oxide on top. The surface plasmon dispersion
and the photon dispersion are manifested as the reflectance minimums
(dashed white) in the density plot. Localized surface plasmon modes
(circles) and photonic modes (stars) are also indicated. (b) The electric field
magnitude for two grating periods of the optimized design, with 248 nm
incident photons, shows strong field confinement near the surface. The alu-
minum grating and aluminum oxide layers are marked by white dotted lines.
(c) Full wave simulation (FDTD) of the reflectance spectrum at normal inci-
dence is shown (black). The red curve is the experimentally measured reflec-
tance. The dip in reflectance in both the simulated and measured spectra
indicates a surface plasmon resonance at 248 nm. Inset: AFM image of the
fabricated grating.
023102-2 Peng
etal.
Appl. Phys. Lett.
106
, 023102 (2015)
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.215.70.231 On: Tue, 20 Jan 2015 15:53:04
A large area aluminum grating structure with the opti-
mized design identified above was fabricated using a nano-
imprint lithography process in which electron beam
lithography was used to fabricate the nanoimprint stamp.
24
Briefly, electron beam lithography was used to pattern a sili-
con wafer coated with negative resist. A grating pattern was
etched into the silicon wafer via reactive plasma etching
(C
4
F
8
and SF
6
). The pattern on the silicon wafer was then
transferred into a stamp of high modulus polymethyl di-
siloxane (PDMS). The transferred pattern was stamped into a
sol-gel silica film spun onto a sputtered aluminum film via
soft contact with the high modulus PDMS stamp. A reactive
plasma etch (Cl
2
and HBr) was then performed to transfer
patterns from the silica into the aluminum film. An atomic
force microscope (AFM) image of the fabricated aluminum
grating is shown as the inset of Figure
2(c)
. The total area of
the patterned structure is 1 mm
1 mm, and the average pe-
riod is 205 nm, with an average grating width of 24 nm.
We measured the ultraviolet reflectance (Figure
2(c)
)of
the fabricated aluminum grating using an integrating sphere
purged with nitrogen, in order to minimize ultraviolet
absorption by atmospheric oxygen. The ultraviolet source of
the integrating sphere illuminates the sample at an angle of
8
relative to normal incidence. In experiments that utilize
the grating to dissociate N
2
H
4
, described below (Figure
3
), a
248 nm excimer laser excites the grating at normal incidence.
However, the surface plasmon resonance of the grating is
highly dependent on incident angle, so the measured reflec-
tance spectrum, taken at 8
incidence, has been transformed
mathematically to depict the reflectance at normal incidence
in order to better compare with experimental conditions. The
measured spectrum after this transformation is shown as the
red curve in Figure
2(c)
. In both the measured and simulated
(black) reflectance spectrum, a surface plasmon resonance at
248 nm is clearly visible. The difference in reflectance
between the experimentally measured spectrum and the
simulated spectrum is attributed to defects in the fabricated
aluminum grating.
To experimentally characterize the efficiency of the ultra-
violet surface plasmon mediated dissociation of N
2
H
4
,weuti-
lize mass spectrometry measurements of surface desorbed
species, as shown in Figure
3
. The aluminum plasmonic gra-
ting is placed inside a vacuum system at 10
6
Torr pressure,
mounted on a copper stage that is thermally coupled to a liq-
uid nitrogen bath outside the chamber. When the stage is
cooled to 77 K, N
2
H
4
vapor is introduced into the vacuum
system. The chamber pressure is raised to 10
5
Torr by intro-
duction of N
2
H
4
,andN
2
H
4
molecules are adsorbed on the
aluminum grating surface. An unpolarized, pulsed KrF laser
operating at 10 W m
2
at 248nm and 1 Hz pulse rate is
directed through a viewport into the vacuum chamber at nor-
mal incidence on the grating. The plasmonically enhanced op-
tical field resonantly dissociates N
2
H
4
molecules into
dissociation products such as N
2
H
3
. The dissociation products
are monitored via mass spectrometry with a filament at an
ionization energy of 70 eV.
The mass spectrometry data from a typical experiment
are shown in Table
I
. The left column lists the dissociation
products of N
2
H
4
that are detected by the mass spectrometer.
In the middle two columns, we compare the partial pressure
and quantum yield of the dissociation products from the alu-
minum grating substrate and a control substrate consisting of
a quartz slide exposed to hydrazine under the same experi-
mental conditions. The signal from each species,
S
i
, in the
measured mass spectrum is a convolution of ions generated
by resonant ultraviolet dissociation of N
2
H
4
, labeled
S
Li
,as
well as species that are generated by gas phase N
2
H
4
mole-
cules that are dissociated by the mass spectrometer filament
directly, labeled
S
Di
. To correct for mass signals that are not
due to ultraviolet dissociation of hydrazine, the partial pres-
sure and quantum yield are calculated according to the fol-
lowing procedure. First, the relative ratio of dissociated
species is measured in the dark, defining
S
Di
. The corrected
signal,
S
Li
, listed in Table
I
is obtained by subtracting the
partial pressure of each dissociation product,
S
i
, from the
partial pressure measured in the dark,
S
Di
, normalized for the
total amount of N
2
H
4
molecules that were ionized and
detected
S
Li
¼
S
i
S
Di
P
i
S
Di
X
i
S
i
:
(1)
We assume that the relative ratio of species generated only
by the spectrometer filament,
S
Di
, does not change when the
laser is on versus in the dark. Our mathematical procedure
ensures that condensed N
2
H
4
species on the plasmonic sub-
strate that are released without undergoing ultraviolet disso-
ciation, e.g., from heating or ablation, are not counted in the
corrected yield. For each of the dissociated species, the quan-
tum yield,
QY
i
, is determined from the partial pressure and
calculated to be
QY
i
¼
g
S
Li
number of photons
;
(2)
FIG. 3. Schematic illustration of ultraviolet surface plasmon mediated athe-
rmal resonant dissociation of N
2
H
4
. A 248 nm laser at normal incidence
excites ultraviolet surface plasmons on the surface of a nanostructured Al
grating which has been cooled to 77 K inside a low pressure atmosphere of
N
2
H
4
(10
5
Torr). The localized optical field resonantly dissociates N
2
H
4
molecules adsorbed to the grating surface, and the dissociated species are
detected via mass spectrometry.
023102-3 Peng
etal.
Appl. Phys. Lett.
106
, 023102 (2015)
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.215.70.231 On: Tue, 20 Jan 2015 15:53:04
where
g
is a geometric factor relating partial pressure to
absolute number of particles based on our chamber configu-
ration. The right column of Table
I
shows the enhancement
ratio,
R
i
, of each detected hydrazine decomposition product
on the plasmonic substrate as compared with the products
from the control substrate. Significantly, the plasmonic sub-
strate enhances the dissociation of every species. The overall
enhancement ratio R is calculated by taking the average of
all enhancement factors for each dissociation product, where
i and j are independent parameters
R
¼
P
i
R
i
S
NHi
2
þ
X
j
R
j
S
N
2
Hj
P
i
S
NHi
2
þ
S
N
2
Hi
:
(3)
The total enhancement factor is 6.2
for all N
2
H
4
dissocia-
tion products when using the plasmonic aluminum substrate
compared with the quartz substrate.
In conclusion, we demonstrate a plasmonic nanostruc-
ture to concentrate ultraviolet radiation and enhance the
generation of reactive nitrogen species athermally via reso-
nant photodissociation by an overall factor of 6.2
.Surface
plasmon-enhanced athermal dissociation processes, like
those demonstrated here, may potentially open alternative
pathways for the growth of nitride semiconductor films, by
generating reactive nitrogen-containing precursors for film
growth at low temperatures.
This work was supported by DARPA under Grant No.
W911NF-13-1-0040 and utilized facilities of the Kavli
Nanoscience Institute at Caltech.
1
Y. K. Kuo, T. H. Wang, J. Y. Chang, and M. C. Tsai,
Appl. Phys. Lett.
99
,
091107 (2011).
2
O. Jani, I. Ferguson, C. Honsberg, and S. Kurtz,
Appl. Phys. Lett.
91
,
132117 (2007).
3
S. Strite and H. Morkoc,
J. Vac. Sci. Technol., B
10
, 1237 (1992).
4
T. Miyazaki, T. Fujimaki, and S. Adachi,
J. Appl. Phys.
89
, 8316
(2001).
5
F. K. Yam and Z. Hassan,
Superlattices Microstruct.
43
, 1 (2008).
6
M. A. Hoffbauer, T. L. Williamson, J. J. Williams, J. L. Fordham, K. M.
Yu, W. Walukiewicz, and L. A. Reichertz,
J. Vac. Sci. Technol., B
31
,
03C114 (2013).
7
A. Kinkhabwala, Z. Yu, S. Fan, Y. Avlasevich, K. Mullen, and W. E.
Moerner,
Nat. Photonics
3
, 654 (2009).
8
L. Brus,
Acc. Chem. Res.
41
, 1742 (2008).
9
C. J. Chen and R. M. Osgood,
Phys. Rev. Lett.
50
, 1705 (1983).
10
R. T. Kidd, D. Lennon, and S. R. Meech,
J. Chem. Phys.
113
, 8276
(2000).
11
S. Linic, P. Christopher, and D. B. Ingram,
Nat. Mater.
10
, 911 (2011).
12
S. M. Morton and L. Jensen,
J. Chem. Phys.
135
, 134103 (2011).
13
H. Nabika, M. Takase, F. Nagasawa, and K. Murakoshi,
J. Phys. Chem.
Lett.
1
, 2470 (2010).
14
A. Nitzan and L. E. Brus,
J. Chem. Phys.
75
, 2205 (1981).
15
R. A. Pala, K. T. Shimizu, N. A. Melosh, and M. L. Brongersma,
Nano
Lett.
8
, 1506 (2008).
16
M. Juan, M. Righini, and R. Quidant,
Nat. Photonics
5
, 349 (2011).
17
R. Bardhan, S. Lal, A. Joshi, and A. Halas,
Acc. Chem. Res.
44
, 936
(2011).
18
G. L. Vaghjiani,
J. Chem. Phys.
98
, 2123 (1993).
19
E. L. Woodbridge, M. N. R. Ashfold, and S. R. Leone,
J. Chem. Phys.
94
,
4195 (1991).
20
W.-G. Liu and W. A. Goddard,
J. Am. Chem. Soc.
134
, 12970 (2012).
21
Y. Watanabe, W. Inami, and Y. Kawata,
J. Appl. Phys.
109
, 023112
(2011).
22
M. W. Knight, L. Liu, Y. Wang, L. Brown, S. Mukherjee, N. S. King,
H. O. Everitt, P. Nordlander, and N. J. Halas,
Nano Lett.
12
, 6000
(2012).
23
J. Hu, L. Chen, Z. Lian, M. Cao, H. Li, W. Sun, N. Tong, and H. Zeng,
J.
Phys. Chem. C
116
, 15584 (2012).
24
M. A. Verschuuren, Ph.D. dissertation, Utrecht university, 2010.
TABLE I. The corrected partial pressure and quantum yield of the dissociated N
2
H
4
species based on mass spectrometry.
Plasmonic substrate
Control substrate
Dissociated
species
Corrected partial
pressure (10
9
Torr),
S
Li
Quantum
yield (10
4
),
QY
i
Corrected partial
pressure (10
9
Torr)
Quantum
yield (10
4
)
Enhancement
ratio,
R
i
N
0.3
4
0.008
1
3.6
NH
2
4.5
70
0.1
2
32.6
NH
3
4.1
60
0.2
4
17.6
N
2
3.0
40
1.4
20
2.1
N
2
H
4.3
70
0.2
3
20.3
N
2
H
3
1.9
30
0.4
6
5.0
Total
18.0
200
2.4
30
6.2 (R)
023102-4 Peng
etal.
Appl. Phys. Lett.
106
, 023102 (2015)
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
131.215.70.231 On: Tue, 20 Jan 2015 15:53:04