Photoionization mass spectrometric study of neutral species from pulsed laser ablation of SnO2

Article abstract of DOI:10.1016/S0009-2614(99)00062-7

We report the application of vacuum ultraviolet (VUV) photoionization mass spectrometry (PIMS) to probe neutral species generated in the 532 nm laser ablation of sintered SnO₂ targets. The major (>90%) Sn containing species are of composition (SnO)_x (x=1,2,3), with near-natural abundance isotopic distributions. The translational energy distribution was determined for each product and compared to a Maxwellian velocity distribution. The utility of VUV PIMS as a universal probe of neutral species produced in laser ablation is discussed.

Full text of DOI:10.1016/S0009-2614(99)00062-7

5
March 1999
Chemical Physics Letters 301 Ž1999. 517–523
Photoionization mass spectrometric study of neutral species from
pulsed laser ablation of SnO2
S.A. Reid )
Department of Chemistry and Center for Sensing Technology, Marquette UniÕersity, P.O. Box 1881, Milwaukee, WI 53201-1881, USA
Received 7 October 1998; in final form 1 December 1998
Abstract
We report the application of vacuum ultraviolet ŽVUV. photoionization mass spectrometry ŽPIMS. to probe neutral
species generated in the 532 nm laser ablation of sintered SnO targets. The major Ž)90%. Sn containing species are of
2
composition ŽSnO. Ž xs1,2,3., with near-natural abundance isotopic distributions. The translational energy distribution was
x
determined for each product and compared to a Maxwellian velocity distribution. The utility of VUV PIMS as a universal
probe of neutral species produced in laser ablation is discussed. q 1999 Elsevier Science B.V. All rights reserved.
1
. Introduction
Many studies have used PLA for thin-film deposi-
tion w19x, and various techniques have been applied
Thin films of semiconductor materials such as
to probe both neutral and charged species in the
ablated plume. These include optical spectroscopies
such as optical emission spectroscopy w20x, laser-in-
duced fluorescence ŽLIF. w21x, and resonance-en-
hanced multiphoton ionization ŽREMPI. w22x, and
mass spectrometry w23–29x. Optical spectroscopies
are species- and state-selective but require advance
knowledge of the electronic states of the species of
interest and are difficult to implement for transient
species with poorly known or dissociative electronic
states. Mass spectrometry has been extensively ap-
plied to detect charged species w23–27x; however, at
fluences typical of deposition conditions the ion
product fraction is small Žtypically -1%. w22,25x,
and neutral detection necessitates post-ablation ion-
ization. Single-photon ŽVUV. photoionization mass
spectrometry ŽPIMS. is an attractive method of de-
SnO have important applications in, e.g., gas sen-
sors, display systems, and solar cells w1–5x. Thin
2
SnO2 films have been grown using a variety of
techniques including organometallic chemical vapor
deposition ŽCVD. w6–9x, laser-assisted CVD w10,11x,
spray pyrolysis w12,13x, reactive-ion sputtering w14x,
and pulsed-laser ablation ŽPLA. w15–18x. The latter
is particularly useful for growth of high-quality ori-
ented films due to the facile production of highly
energetic gas-phase reactants, which facilitates film
growth and adhesion to the substrate Žsee, e.g., Ref.
w19x.. An important aspect of film deposition is the
need to monitor the chemical composition and ener-
getics of the ablated material in order to understand
film growth processes.
)
tection for transient neutral species which to date has
received little attention in PLA studies w30,31x.
Corresponding author. Fax: q1 414 288 7066; e-mail:
reids@mu.edu
0
009-2614r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.
PII: S0009-2614Ž99.00062-7

5
18
S.A. ReidrChemical Physics Letters 301 (1999) 517–523
In this study we apply VUV Ž118.2 nm. PIMS
less-steel vacuum chamber and passed through the
interaction region of a commercial ŽR.M. Jordan.
Wiley–McLaren-type linear TOF mass spectrometer
ŽTOFMS. and a 3 mm diameter circular aperture
using time-of-flight ŽTOF. mass analysis to analyze
the composition and energetics of Sn-containing
species produced in the 532 nm ablation of sintered
SnO targets at typical deposition fluences. The de-
before striking a cylindrical SnO target. The beam
2
2
position of SnO films using PLA of both SnO and
Sn targets has been reported w15–18x; however, to
was focussed onto the target using a 1 m CaF lens,
2
2
2
to a typical spot size of 0.5 mm Žaverage fluence
2
date no plume characterization studies have been
reported. Moreover, previous PLA studies using mass
spectrometric techniques have largely focused on
charged species w23–27x. This study demonstrates
that VUV PIMS is a powerful tool for the investiga-
tion of neutral species produced in PLA, and the
high photon energy Ž10.49 eV. used here provides a
f0.5–10 Jrcm .. The backscattered material flew a
distance of 70"2 mm, passing through the aperture
and into the interaction region of the TOFMS, and
was ionized by 118.2 nm photons generated by
non-resonant tripling of ;10 mJ of the 355 nm
output of a second Nd:YAG system ŽContinuum
Surelite II. in a 15.2 cm long cell filled with 4 Torr
universal means of detecting bare and oxidized Sn
molecules and clusters.
of Xe. The beam was focused into the cell using a
fused silica lens of 7.5 cm focal length, and the
generated VUV light focussed into the interaction
region using a 7.5 cm focal length LiF lens which
served as the exit window. The cell was inserted into
vacuum through a compression fitting.
2
. Experimental section
A schematic of the experimental apparatus is
The ions were extracted, accelerated and flew a
distance of 118 cm prior to striking a dual mi-
crochannel plate detector. The microchannel plate
signal was amplified Ž=25. using a fast preamplifier
ŽStanford Research 445. and fed into a multichannel
scaler ŽSR 430.. The ablation laser Q-switch was
shown in Fig. 1. A pulsed Nd:YAG laser ŽCon-
tinuum NY-61. was used for ablation. Typical pulse
energies at 532 nm were 1–20 mJ in a 8 ns pulse
ŽFWHM.. The ablation beam was sent into a stain-
triggered at a repetition rate of 5 Hz, or one-half that
of the ionization laser, using a frequency divider
ŽAvtech., and the scaler operated in a toggle or
background subtraction mode. The delays between
laser pulses were adjusted using a digital delay gen-
erator ŽSR DG535.. The TOFMS was calibrated
using PLA of OFHC copper, and the observed Cux
Ž xs1,2,3. mass peaks fit to the following expres-
sion:
'm satqb ,
Ž .
1
where m is the mass in amu and t the flight time in
ms. The constants determined from this fit were:
as0.74606, bs0.012865.
The SnO target was prepared by pressing 10.4 g
2
of SnO powder ŽAldrich, )99.995%. into a pellet
2
Y
Y
Žlength ;0.75 ; diameter 0.5 .. This pellet was
then fired in an oven for 70 h at 12008C. The
sintered target was mounted on the shaft of a
motorized, computer-controlled linear-rotary motion
feedthrough. Translation and rotation were both
repetitively cycled, ensuring that a fresh surface was
Fig. 1. Schematic diagram of the experimental apparatus. TP,
turbomolecular pump; DP, diffusion pump; MCP, microchannel
plate; MCS, multichannel scaler; PrDG, pulserdelay generator.

S.A. ReidrChemical Physics Letters 301 (1999) 517–523
519
exposed to each laser shot and producing a helical
ablation pattern over the full extent of the target.
Žionization. laser shots, at an ablation–ionization
delay of 65 ms. The ablation pulse energy was 5 mJ
Žf2.5 Jrcm .. The broad background arises from
ions produced directly in the ablation which can be
2
During experiments the main chamber was kept at
y7
;
2=10
Torr using a diffusion pump system,
while the flight tube was evacuated by a turbomolec-
ular pump to ;10 Torr.
Two types of experiments were conducted. In the
first, TOF spectra accumulated over typically 10000
Žionization. laser shots were obtained at a fixed
neutral arrival time Žablation–ionization laser delay..
deflected from the beam using biased plates w25x.
y7
Note that the ion signal is integrated over all arrival
times, species, and states, and thus is not indicative
of the true ion fraction w22x. Assignments of the
observed mass peaks are given following Eq. Ž1.. A
persistent signal due to the alkali metals Na, K, and
Ca was observed, with this and other targets ŽAl, Cu.
we have ablated. These impurities have been ob-
served in resonant ablation of Cu and attributed to
their high volatility, low ionization potentials ŽIP.,
and presence on the surface in ionic compounds w32x.
In the second, the delay between ablation and ioniza-
tion laser was varied and complete TOF spectra
accumulated over 5000 laser shots at each delay.
Long-term drift in signal Ž;20–30% over a span of
8
h. due to laser-induced surface modification was
corrected for by selecting a reference delay near the
peak of the distribution and repetitively accumulat-
ing spectra at this every 2–3 delay points.
More important are the strong signals observed
for SnO and Sn O . Smaller signals are observed for
2
2
Sn, Sn O and the hypermetalated w33x compound
3
3
Sn O, while no SnO is detected. The observation of
2
2
Sn O here follows recent experimental and theoreti-
2
3
. Results and discussion
cal studies of other hypermetalated compounds w33x.
The IP of both O Ž13.64 eV. w34x and O Ž12.071
2
Fig. 2 displays a mass spectrum obtaining by
eV. w34x lie above 10.49 eV and the ground-state
species cannot be ionized by a single VUV photon.
Small signals at the respective arrival times for O
averaging four scans, each accumulated over 10000
and O are observed, which may arise from electron
2
impact ionization. The fate of the ‘missing’ O, i.e.,
whether atomic or diatomic oxygen is preferentially
produced, is an open question. Free O Ž xs1,2.
x
species must be produced under our conditions, and
the detection and characterization of these species
will be important in further characterizing PLA of
SnO targets, and in the use of PLA for SnO film
2
x
deposition.
Fig. 3a–c displays expanded views of the spec-
trum of Fig. 2 in the region of Sn, SnO, and
Sn OrSn O mass peaks, together with fits using
2
2
2
natural abundance isotopic distributions. The ion
background was first fit using an exponential func-
tion, as discussed below, and subtracted from the
data to achieve a flat baseline. Sn and SnO peaks
were fit to the sum of 10 Gaussian distributions Žall
stable Sn isotopes. with the peak center flight time
determined using Eq. Ž1. and the width of each
distribution and overall intensity used as fit parame-
ters. The spectra were fit using non-linear least-
squares fitting based on the Marquardt–Levenberg
algorithm. The Sn OrSn O fit is the best visual fit
Fig. 2. VUV PIMS spectrum of the neutral products of 532 nm
pulsed-laser ablation of a solid SnO2 target, obtained as described
in the text.
2
2
2

5
20
S.A. ReidrChemical Physics Letters 301 (1999) 517–523
well if the time window chosen was not too large.
The zero baseline spectra were then integrated over
all channels for each species and this signal normal-
ized to the appropriate reference scan. The Sn O
2
2
integration limits were chosen to exclude contribu-
tions from Sn O, which represented ;5% of the
2
total signal in the 22 ms region.
Shown as the solid line in each panel of Fig. 4 is
a fit to a unshifted time transformed Maxwellian
velocity distribution w32x:
4
2
t
peak
A
tpeak
expy2 ž /
S
Ž
t
.
s
.
2
Ž .
t
ž
3
/
ž /
d
t
In Eq. Ž2. ds70"2 mm is the distance between
target and ionization region, t
is the time corre-
peak
sponding to the maximum in the arrival time distri-
bution, and A is a scaling constant. Each fit used
Fig. 3. TOF spectra Žpoints. and fits Žlines. for Ža. Sn, Žb. SnO,
and Žc. Sn O rSn O mass peaks. The fits demonstrate near-nat-
2
2
2
ural abundance isotopic distributions.
obtained to the sum of 44 Gaussian functions using
flight times determined from Eq. Ž1.. The fits
demonstrate near-natural abundance isotopic distri-
butions.
The translational energy distribution of the pri-
mary Sn containing products was also investigated.
Fig. 4 displays the integrated mass counts Žpoints. as
a function of neutral arrival time Žablation–ioniza-
tion laser delay. in the range 40–160 ms for Sn,
SnO, and Sn O . These scans were normalized to a
2
2
reference delay of 65 ms as described above. The
background ion signal exhibited a strong dependence
on delay time, hindering us from acquiring data for
neutral flight times -40 ms, and gradually de-
creased over the course of the experiment. Individual
scans were thus analyzed by first fitting the back-
ground signal to an exponential function and sub-
tracting this to get a zero baseline, which worked
Fig. 4. Arrival time distributions Žpoints. for: Ža. Sn, Žb. SnO, and
Žc. Sn O . The lines in each plot are fits to a time-transformed
Maxwellian velocity distribution.
2
2

S.A. ReidrChemical Physics Letters 301 (1999) 517–523
521
non-linear least-squares fitting based on the Mar-
quardt–Levenberg algorithm, with tpeak and A used
as fit parameters. For the fluences and wavelength
used here we expect a dominant contribution from
thermal vaporization w32x, and with the assumption
from parent ŽSnO, Sn O . fragmentation in the ion-
2 2
ization step.
We used the fits shown in Fig. 4 integrated over
time to estimate relative species abundances in the
plume, with the modification that the entire inte-
grated signal in the 22 ms region was used. The
Sn O rSn O abundance was estimated from the fit
of a thermal desorption process derive temperatures
4
3
of: 1.1=10 K ŽSn., 7.8=10 K ŽSnO., and 1.1=
2
2
2
4
1
0 K ŽSn O . from the fits in Fig. 4. The estimated
uncertainty is "2=10 K.
to the 65 ms spectra with the assumption of similar
Sn O rSn O translational energy distributions. Rel-
2
2
3
2
2
2
An important issue in the application of mass
spectrometry is extraction of the neutral distribution
from the observed ion intensities. Fragmentation in
the ionization step can be important, particularly for
electron impact ŽEI. ionization. The advantage of
single-photon VUV laser photoionization compared
to EI is the smaller ionization energy Ž10.49 eV here
vs. 20–100 eV for EI., which reduces fragmentation.
ative SnrSnOrSn O rSn O abundances of
2
2
2
0.22:0.65:1.0:0.05 are obtained with the assumption
of similar photoionization efficiencies. This is proba-
bly a poor assumption for Sn and SnO, because the
IPs of ground-state Sn Ž7.344 eV. w34x and SnO
Ž10.5"0.5 eV. w34,35x are very different, and that of
SnO lies close to 10.49 eV. The precise calibration
of intensities in VUV PIMS requires knowledge of
VUV absorption cross-sections and ionization quan-
tum yields, which is a weakness of the PIMS tech-
nique. Nonetheless, important relatiÕe trends may
still be identified. For example, Fig. 5 displays a
PIMS spectrum from 355 nm ablation of the same
target under similar conditions. Marked changes in
species composition are observed compared to the
For example, using 10.49 eV photons the channel:
q
y
SnOqhn™Sn qOqe
3
4
0
=
D0
Ž
Sn–O
.
qIP
Ž
Sn
.
s12.83 eV
,
3
Ž .
is closed to ground-state SnO species, and requires
an internal energy of ;2.34 eV. Moreover, Sn ions
produced via this channel would have an arrival time
distribution identical to SnO, yet these distributions
are clearly different ŽtpeakŽSnO.s51.4 ms, tpeakŽSn.
s41.4 ms. and the Sn distribution shows no obvious
5
32 nm spectrum ŽFig. 2., which reflects the depen-
dence of PLA on ablation wavelength. A detailed
comparison of PLA of SnO2 targets at different
wavelengths will be forthcoming, and future work
will also incorporate an EI source for comparison as
well as apply pulsed extraction methods to mass
resolve ions produced directly in ablation.
bimodality.
Similar arguments hold for SnO production from
Sn O . Using 10.49 eV photons, the channel:
2
2
q
y
Sn O qhn SnO qSnOqe
2
2
3
4.35
0
=
D0
Ž
OSn–SnO
.
qIP
Ž
SnO
.
s13.39 eV
,
4
Ž .
is closed to ground-state Sn O species, and requires
2
2
an internal energy of ;2.9 eV. The SnO and Sn O2
2
arrival time distributions are different ŽtpeakŽSnO.s
5
1.4 ms, tpeakŽSn O .s59.0 ms., which is not as
2
2
obvious from a visual inspection of Fig. 4.
To estimate the degree of fragmentation, we fit
the Sn and SnO arrival time distribution to the sum
of two distributions, each of the form shown in Eq.
Ž2., with tpeak of one distribution fixed at the fit
value for SnO Žfor Sn. and Sn O Žfor SnO.. From
2
2
this procedure we estimate in each case an upper
limit of 5% on the daughter ŽSn, SnO. signal arising
Fig. 5. VUV PIMS spectrum of the neutral products of 355 nm
PLA of a solid SnO2 target, under similar conditions as for Fig. 2.

5
22
S.A. ReidrChemical Physics Letters 301 (1999) 517–523
4
. Conclusions
calibration of the experimental apparatus, and Prof.
Frank Lamelas ŽMU Physics Department. for the
We have applied VUV PIMS to probe neutral
X-ray diffraction data.
species generated in the 532 nm laser ablation of
sintered SnO targets at typical deposition fluences.
2
References
The major Ž)85%. Sn containing species observed
are of composition ŽSnO. Ž xs1,2,3., characterized
x
1
w x
H. Iida, N. Shibata, T. Mishuka, A. Ito, H. Karasawa, Y.
Hayashi, IEEE Electron Device Lett. 4 Ž1983. 157.
by near-natural abundance isotopic distributions. In
addition, atomic Sn and the hypermetalated com-
w2x D.K. Schroder, IEEE Trans. Electron Devices 25 Ž1978. 90.
w3x W. G o¨ pel, Sensors Actuators B 18r19 Ž1994. 1.
pound Sn O were observed. The neutral translational
2
w4x K. Takahata, in: T. Seiyama ŽEd.., Development and Appli-
cations in Chemical Sensor Technology, 1988, p. 39.
energy distribution was determined for each primary
product and found to be well represented by a time-
transformed Maxwellian velocity distribution. Fits
w x
5
G. Heiland, Sensors Actuators 2 1982 343.
w6x J. Ogawa, M. Nixhikawa, A. Ake, J. Appl. Phys. 53 Ž1982.
448.
Ž
.
4
4
yield temperatures of ;1=10 K, and the inte-
w7x N.S. Murty, G.K. Bhagawat, S.R. Jawalekar, Thin Solid
grated fits yield relative abundances of SnrSnOr
Films 92 Ž1982. 347.
J.C. Lou, M.S. Lin, J. Chyi, Y. Tar, Thin Solid Films 106
Sn O rSn Os0.22:0.65:1.0:0.05, uncorrected for
2
2
2
8
w x
variations in photoionization efficiency. The advan-
tages of VUV PIMS as a probe of neutral species
produced in PLA include the ‘soft’ ionization which
minimizes fragmentation and universal detection. The
need for quantitative photoionization efficiency data
to obtain accurate relative species abundances is a
weakness of this method.
Ž
1
983 163.
.
w9x J. Tabuchi, K. Yamagishi, Y. Tarui, Jpn J. Appl. Phys. 26
Ž1987. L186.
w10x R.R. Kunz, M. Rothschild, D.J. Erlich, Appl. Phys. Lett. 54
Ž1989. 1631.
w11x V.C. Simianu, J.M. Hossenlopp, Appl. Organometal. Chem.
1
1 1997 147.
Ž .
w12x D.R. Acosta, A. Maldonaldo, R. Asomoza, J. Mater. Sci.
Mater. Elec. 4 Ž1983. 187.
Recent PLA deposition studies have shown that
w13x R. Poimmer, C. Gril, J. Marucchi, Thin Solid Films 77
Ž1981. 91.
w14x G. Beensh-Marchwicka, L. Krol-Stepniewska, A. Misink,
room temperature deposition using SnO targets pro-
2
duces largely amorphous SnO films w18x. We have
x
used 532 nm PLA to deposit SnOx films onto
single-crystal Si substrates held at room temperature,
and preliminary X-ray diffraction studies show a
Ž
.
Thin Solid Films 113 1984 215.
w15x R.D. Vispute, V.P. Godbole, S.M. Chaudhari, S.M. Kanetkar,
S.B. Ogale, J. Mater. Res. 3 Ž1988. 1180.
w16x R. Lal, R. Grover, R.D. Vispute, R. Viswanathan, V.P.
Godbole, S.B. Ogale, Thin Solid Films 206 Ž1991. 88.
w17x W.S. Hu, Z.G. Liu, J.G. Zheng, X.B. Hu, X.L. Guo, W.
G o¨ pel, J. Mater. Sci. – Mater. in Elec. 8 Ž1997. 155.
w18x W.S. Hu, Z.G. Liu, Z.C. Wu, D. Feng, Mater. Lett. 28
Ž1996. 369, and references therein.
w19x D.B. Chrisey, G.K. Hubler ŽEds.., Pulsed Laser Deposition
of Thin Films, John Wiley, New York, 1994, and references
therein.
dominant amorphous SnO phase. Interestingly, an-
x
nealing at 3008C produces a polycrystalline SnO
film, which correlates with the dominant 1:1 stoi-
chiometric composition we observe in the plume.
This preliminary work illustrates the importance of
plume characterization studies in understanding the
growth of PLA generated thin film materials. The
characterization of PLA deposited SnO films will
w
20 W.A. Weimer, J. Appl. Phys. 52 1988 2171.
x Ž .
x
be the subject of future reports.
w21x C.E. Otis, R.W. Dreyfuss, Phys. Rev. Lett. 67 Ž1991. 2102.
w22x R.C. Estler, N.S. Nogar, J. Appl. Phys. 69 Ž1991. 1654.
w23x K. Fukushima, Y. Kanke, T. Morishita, J. Appl. Phys. 74
Ž1993. 6948.
Acknowledgements
w24x R.P. van Ingen, J. Appl. Phys. 79 Ž1996. 467.
w25x H. Kang, J.L. Beauchamp, J. Phys. Chem. 89 Ž1985. 3364.
w26x J.C.S. Kools, S.H. Brongersma, E. van de Reit, J. Dieleman,
Appl. Phys. B 53 Ž1991. 125.
Acknowledgment is made to the National Science
Foundation CAREER program, the Arnold and Ma-
bel Beckman Foundation, and the Marquette Center
for Sensing Technology for support of this research.
The author thanks Binyong Liang and Jason P.
Schmidt for assistance in construction and initial
w27x R. Teghil, A. Santagata, V. Marotta, S. Orlando, G. Pizzella,
A. Giardini-Guidoni, A. Mele, Appl. Surf. Sci. 90 Ž1995.
5
05.
w28x T.L. Thiem, L.R. Watson, R.A. Dressler, R.H. Salter, E.
Murad, J. Phys. Chem. 98 Ž1994. 11931.

S.A. ReidrChemical Physics Letters 301 (1999) 517–523
523
w29x R. Timm, P.R. Willmott, J.R. Huber, J. Appl. Phys. 80
Ž1996. 1794.
w30x M. Macler, M.E. Fajardo, Appl. Phys. Lett. 65 Ž1994. 2275.
w31x F. Tokai, M. Tanawaki, Y. Koga, Y. Kakudate, S. Fujiwara,
K. Fukuda, M. Kawaguchi, J. Appl. Phys. 77 Ž1995. 2220.
w32x C.G. Gill, T.M. Allen, J.E. Anderson, T.N. Taylor, P.B.
Kelly, N.S. Nogar, Appl. Opt. 35 Ž1996. 2069.
w33x S.L. Wang, K.W.D. Ledingham, R.P. Singhal, J. Phys. Chem.
100 Ž1996. 11282, and references therein.
w34x A.A. Radzig, B.M. Smirnov, Reference Data on Atoms,
Molecules, and Ions, Springer, Berlin, 1985.
w35x R. Colin, J. Drowart, A.G. Verhaegen, Trans. Faraday Soc.
61 Ž1965. 1364.