Application of mid-infrared tuneable diode laser absorption spectroscopy to plasma diagnostics: A review

Abstract

Within the last decade mid-infrared absorption spectroscopy over a region from 3 to 17µm and based on tuneable lead salt diode lasers, often called tuneable diode laser absorption spectroscopy or TDLAS, has progressed considerably as a powerful diagnostic technique for in situ studies of the fundamental physics and chemistry in molecular plasmas. The increasing interest in processing plasmas containing hydrocarbons, fluorocarbons, organo-silicon and boron compounds has led to further applications of TDLAS because most of these compounds and their decomposition products are infrared active. TDLAS provides a means of determining the absolute concentrations of the ground states of stable and transient molecular species, which is of particular importance for the investigation of reaction kinetic phenomena. Information about gas temperature and population densities can also be derived from TDLAS measurements. A variety of free radicals and molecular ions have been detected by TDLAS. Since plasmas with molecular feed gases are used in many applications such as thin film deposition, semiconductor processing, surface activation and cleaning, and materials and waste treatment, this has stimulated the adaptation of infrared spectroscopic techniques to industrial requirements. The recent development of quantum cascade lasers (QCLs) offers an attractive new option for the monitoring and control of industrial plasma processes. The aim of the present paper is threefold: (i) to review recent achievements in our understanding of molecular phenomena in plasmas, (ii) to report on selected studies of the spectroscopic properties and kinetic behaviour of radicals and (iii) to describe the current status of advanced instrumentation for TDLAS in the mid-infrared.

Full text

INSTITUTE OF PHYSICS PUBLISHING PLASMA SOURCES SCIENCE AND TECHNOLOGY
Plasma Sources Sci. Technol. 15 (2006) S148–S168 doi:10.1088/0963-0252/15/4/S02
Application of mid-infrared tuneable
diode laser absorption spectroscopy to
plasma diagnostics: a review
JR
¨
opcke1,5, G Lombardi2, A Rousseau3andPBDavies4
1INP-Greifswald, 17489 Greifswald, Friedrich-Ludwig-Jahn-Str. 19, Germany
2CNRS LIMHP, Universit´
e Paris XIII, 99, av. J.B. Cl´
ement, 93430 Villetaneuse, France
3Laboratoire de Physique et Technologie des Plasmas, Ecole Polytechnique, CNRS, 91128 Palaiseau,
France
4Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge CB2 1EW, UK
E-mail: roepcke@inp-greifswald.de
Received 11 November 2005
Published 6 October 2006
Online at stacks.iop.org/PSST/15/S148
Abstract
Within the last decade mid-infrared absorption spectroscopy over a region from 3 to
17µm and based on tuneable lead salt diode lasers, often called tuneable diode laser
absorption spectroscopy or TDLAS, has progressed considerably as a powerful
diagnostic technique for in situ studies of the fundamental physics and chemistry in
molecular plasmas. The increasing interest in processing plasmas containing
hydrocarbons, fluorocarbons, organo-silicon and boron compounds has led to further
applications of TDLAS because most of these compounds and their decomposition
products are infrared active. TDLAS provides a means of determining the absolute
concentrations of the ground states of stable and transient molecular species, which is
of particular importance for the investigation of reaction kinetic phenomena.
Information about gas temperature and population densities can also be derived from
TDLAS measurements. A variety of free radicals and molecular ions have been
detected by TDLAS. Since plasmas with molecular feed gases are used in many
applications such as thin film deposition, semiconductor processing, surface activation
and cleaning, and materials and waste treatment, this has stimulated the adaptation of
infrared spectroscopic techniques to industrial requirements. The recent development
of quantum cascade lasers (QCLs) offers an attractive new option for the monitoring
and control of industrial plasma processes. The aim of the present paper is threefold:
(i) to review recent achievements in our understanding of molecular phenomena in
plasmas, (ii) to report on selected studies of the spectroscopic properties and kinetic
behaviour of radicals and (iii) to describe the current status of advanced
instrumentation for TDLAS in the mid-infrared.
(Some figures in this article are in colour only in the electronic version)
1. Introduction
Low-pressure, non-equilibrium molecular plasmas are of
increasing interest not only in fundamental research but also
in plasma processing and technology. Molecular plasmas are
used in a variety of applications such as thin film deposition,
semiconductor processing, surface activation and cleaning, and
5Author to whom any correspondence should be addressed.
in materials and waste treatment. The investigation of plasma
physics and chemistry in situ requires detailed knowledge of
plasma parameters, which can be obtained by appropriate
diagnostic techniques. The need for a better scientific
understanding of plasma physics and chemistry has stimulated
the improvement of established diagnostic techniques and
the introduction of new ones. Methods based on
traditional spectroscopy have now become amongst the most
important.
0963-0252/06/040148+21$30.00 © 2006 IOP Publishing Ltd Printed in the UK S148

TDLAS for plasma diagnostics
distance
ν
intensity
I0I
plasma
exponential
decay
Figure 1. Absorption of external radiation in a plasma according to
the Beer–Lambert Law [1].
Over the past decade, several new techniques have been
successfully introduced for diagnostic studies of chemically
reactive plasmas in which many short-lived and stable species
are produced. It has been possible to determine absolute
concentrations of ground states using spectroscopy thereby
providing a link with chemical modelling of the plasma.
The other essential component needed to reach the objective
of improved knowledge of molecular phenomena is to
determine physical parameters of the plasma by an appropriate
experimental methodology.
The methods of absorption spectroscopy (AS) are of
great importance in plasma diagnostics because they provide
a means of determining the population densities of species
in both ground and excited states. The spectral line
positions provide species identification while line profiles are
often connected with gas temperature and relative intensities
provide information about population densities. An important
advantage of AS over optical emission spectroscopy (OES)
methods is that only relative intensities need to be measured to
determine absolute concentrations, avoiding the problems of
complete instrument calibration inherent in the OES methods.
Absorption spectroscopy has been applied right across the
spectrum from the vacuum ultraviolet (VUV) to the far infra-
red (FIR). Continuously emitting lamps (e.g. the Xe-lamp for
the VIS and NIR and the D2-lamp for the UV) and tuneable
narrow-band light sources (e.g. tuneable dye lasers, diode
lasers) can be used as external light sources.
In the case where an external light source has much
higher intensity than that of the plasma itself, the absorption
of radiation can be described by the Beer–Lambert law which
is,
Iν(l) =Iν(0)exp(−κ(ν)l). (1)
Iv(0)and Iv(l) are the fluxes of the radiation entering
and leaving the plasma, lis the length of the absorbing
(homogeneous) plasma column and κ(v) is the absorption
coefficient. Figure 1illustrates this situation [1].
A wide variety of light sources, dispersive elements,
detectors and data acquisition methods can be used for
absorption spectroscopy [2]. The classic dispersion
experiment for measuring the density of atomic or molecular
states in plasmas by AS relies on continuous light sources and
a spectrograph with a suitable detector for the spectral range
of interest. The Fourier transform infra-red (FTIR) technique
with a Michelson interferometer also uses a continuous light
source for absorption measurements. The transmitted light
intensity depends on the (variable) optical path difference
between the mirrors in the two interferometer arms and yields
the interferogram carrying the spectroscopic information.
In contrast to dispersion techniques the FTIR spectrometer
+
-
p region
n region
p-n junction
metal contact
metal contact
insulator
resonator front
facet
active zone
© U. Haeder
Figure 2. Schematic of the compositional structure of a lead salt
diode laser and inset photograph of the laser in its packaging [1].
records the whole spectrum simultaneously (multiplex or
Fellgett advantage). In principle, the resolution of FTIR can
be as high as 0.002 cm−1, determined by the distance scanned
by the moveable mirror, but at the expense of recording time.
Fractional absorptions as small as 10−4can be measured.
With the development of tuneable, narrow band light
sources such as tuneable dye lasers and infrared diode lasers,
these have been substituted for continuous light sources in
AS experiments. These narrow band laser sources have the
advantage of high spectral intensity, narrow bandwidth and
continuous tuneability over the absorption profile. Figure 2
shows a homostructure lead salt diode laser used for infrared
tuneable diode laser absorption spectroscopy (TDLAS) [1].
Alternatively tuneable infrared radiation can be generated
using the methods of difference frequency mixing or optical
parametric oscillators (OPOs). In the past these systems
however had the disadvantage of rather low radiation power
and were restricted to specific wavelength regions [3,4]. New
technical developments have led to solutions, which provide
up to a mW of single mode power and tuneability of up to
100 cm−1[133,134].
Another highly sensitive novel laser technique is cavity
ring-down (CRD) absorption spectroscopy. This method is
based on the measurement of the intensity decay rate of a laser
pulse injected into an optical cavity formed by two very highly
reflective mirrors which also enclose the plasma. Absorptions
as low as 10−9can be measured with an acceptable signal-
to-noise ratio [5]. Near infrared tuneable diode lasers have
been used as the light sources for CRD spectroscopy [6].
More recently, CRD spectroscopy has been applied for density
measurements of the SiH2radical and of nanometre sized dust
particles in silane plasmas [7] and for detecting N+
2ions in
nitrogen discharges [8].
The increasing interest in processing plasmas containing
hydrocarbons, fluorocarbons or organo-silicon compounds has
led to further applications of infrared AS techniques because
most of these compounds and their decomposition products
are infrared active. FTIR spectroscopy has been used for in
situ studies of methane plasmas for a number of years, but it
is generally insufficiently sensitive for detecting free radicals
or ions in processing plasmas. TDLAS is increasingly being
S149

JR
¨
opcke et al
used in the spectral region between 3 and 20 µm for measuring
the concentrations of free radicals, transient molecules and
stable products in their electronic ground states. TDLAS
can also be used to measure neutral gas temperatures [9]
and to investigate dissociation processes of molecular low
temperature plasmas [10–13]. The main applications of
TDLAS until now have been for investigating molecules and
radicals in fluorocarbon etching plasmas [9,12,14] and in
plasmas containing hydrocarbons [13,15–21]. A wide variety
of low molecular weight free radicals and molecular ions have
been detected by TDLAS in purely spectroscopic studies,
e.g. Si−
2[22] and SiH+
3[23] in silane plasmas. Most of
these spectroscopic results have yet to be applied in plasma
diagnostic studies.
Molecular plasmas are increasingly being used not
only for basic research but also, due to their favourable
properties, for materials processing technology. These fields
of application have stimulated the development of infrared
spectroscopic techniques for industrial requirements. In order
to exploit the capabilities of infrared TDLAS for effective
and reliable on-line plasma diagnostics and process control
in research and industry, compact and transportable tuneable
infrared multi-component acquisition systems (IRMA, TOBI)
have been developed [24,25] (see section 4). These systems
are mainly focused on (i) high speed detection of stable and
transient molecular species in plasmas under non-stationary
excitation conditions and (ii) on sensitive (sub-ppb) trace gas
detection with the aid of multi-pass absorption cells.
The main disadvantage of TDLAS systems, based upon
lead salt diode lasers, is the necessary cryogenic cooling of
the lasers (and also of the detectors), because they operate
at temperatures below 100 K. Systems based upon lead salt
diode lasers are typically large in size and require closed
cycle refrigerators and/or cryogens such as liquid nitrogen.
The recent development and commercial availability of pulsed
quantum cascade lasers (QCLs) offers an attractive new option
for infrared absorption spectroscopy.
The present paper is intended to give an overview of recent
achievements which have led to an improved understanding of
phenomena in non-equilibrium molecular plasmas based on the
application of TDLAS techniques. The paper is divided into
three main sections: in section 2special attention is devoted
to recent studies of plasma chemistry and reaction kinetics in
gas discharges containing hydrocarbons, organo-silicon and
boron compounds. A link is thereby provided with chemical
modelling of the plasmas. Section 3concerns recent results
of spectroscopic properties and kinetic behaviour of selected
radicals, which are of special importance for reaction kinetics
and chemistry in molecular processing plasmas. The current
status of advanced spectroscopic instrumentation is described
in section 4.
2. Plasma chemistry and reaction kinetics
2.1. General considerations
Low temperature plasmas, in particular microwave and radio
frequency (RF) plasmas, have high potential for applications
in plasma technology. In molecular low temperature plasmas,
the species and surface conversion is frequently governed
by high degrees of dissociation of the precursor molecules
and the high amounts of chemically active transient and
stable molecules present. For further insight into plasma
chemistry and kinetics a challenging subject is to study the
mainly electron induced plasma reactions leading to entire
series of different chemical secondary reactions involving
the whole group of substances making up the source gas
molecules. Hydrocarbon precursors are of special importance,
since they are used in a variety of plasma enhanced chemical
vapour deposition (PECVD) processes to deposit thin carbon
films. In all cases, the monitoring of transient or stable
plasma reaction products, in particular the measurement of
their ground state concentrations, is the key to improved
understanding of fundamental phenomena in molecular non-
equilibrium plasmas which can be applied in turn to many other
aspects of plasma processing.
Transient molecular species, in particular radicals,
influence the properties of nearly all molecular plasmas, both
in the laboratory and in nature. They are of special importance
in several areas of reaction kinetics and chemistry. The study of
the behaviour of radicals together with their associated stable
products provides a very effective approach to understand
phenomena in molecular plasmas. Radicals containing carbon
are of special interest for basic studies and for application in
plasma technology.
Although the methyl radical (CH3) is acknowledged to be
one of the most essential intermediates in hydrocarbon plasma
chemistry, only a few methods are available for its detection in
situ. Sugai et al and Zarrabian et al employed the technique of
threshold ionisation mass spectrometry to detect the methyl
radical in electron cyclotron resonance plasmas containing
methane [26–28]. Based on cavity ring-down spectroscopy
with ultraviolet radiation CH3concentration measurements
have been performed in a hot-filament reactor [29–31]. Most
of the measurement techniques for detecting the methyl radical
are based on absorption spectroscopy either with 216 nm
ultraviolet radiation or in the infrared near 606 cm−1.For
example, the ultraviolet absorption of CH3at 216 nm was
used for number density measurements by Child et al and
Menningen et al and in different CVD diamond growth
environments, hot-filament, dc and microwave plasmas [32–
34]. In 2003 Lombardi et al performed a comparative study
to detect methyl radicals using both broadband ultraviolet
absorption and TDLAS [35] (see section 3.1).
The infrared TDLAS technique has proven to be highly
useful because it can also be used to measure the concentrations
of related species provided they are IR active. Already in
1990 Wormhoudt demonstrated this flexibility by measuring
CH3and C2H2inaCH
4–H2RF plasma using a long path
plasma absorption cell [36]. Actually, TDLAS is probably
the best method for detecting the methyl radical for several
reasons. The v2out-of-plane bending mode is not only intense
but has many lines between 600 and 650cm−1. It is then
possible to derive rotational and vibrational temperatures from
their relative intensities. The (J=K)Q-branch lines of
the v2fundamental band near 606 cm−1are particularly useful
because several of them lie within 0.5cm−1of each other, i.e.
within a single laser spectral mode. Rotational temperatures
in the plasma are therefore easily measured from them. For
more than a decade, quantifying the concentrations of methyl
S150

TDLAS for plasma diagnostics
He closed cycle
refrigerator TDL's
HgCdTe
detector
TDL system
monochromator microwave window
microwave appliance
module (2.45 GHz)
discharge vessel with long path cell
plasma region
objective
mirror box
field
mirror box
(a) (b)
Figure 3. (a) Experimental arrangement of the planar microwave plasma reactor (side view) with White cell multiple pass optical
arrangement and TDL infrared source. The laser beam path is indicated by dotted lines [45]. (b) White cell with field mirror and objective
mirrors showing the laser beam path [45].
radicals via the determination of the line strength of the Q(8,8)
line of methyl at 608.3 cm−1by Wormhoudt and McCurdy
has been highly important [37]. In 2005 using the decay of
the methyl radical in the off-phase of pulsed plasmas new,
precise measurements of the transition dipole moment of the υ2
fundamental band have been performed [38] (see section 3.2).
One of the most successful applications of TDLAS is
for studying the decomposition of hydrocarbons in a variety
of PECVD processes. Systematic TDLAS measurements of
several different hydrocarbons, including methyl, in a 20kHz
methane plasma in a parallel plate reactor were reported by
Davies and Martineau [10,39,40]. Goto and co-workers
have published numerous studies of methyl and methanol
concentrations in RF and electron cyclotron resonance (ECR)
plasmas under different conditions, e.g. investigating the
influence of rare gases on the plasma. They havealso combined
IR absorption with emission spectroscopy and investigated the
effect of water vapour on the methyl radical concentration in
argon/methane and argon/methanol RF plasmas using TDLAS
[17,18,41–43]. Kim et al measured CH3,C
2H2and CH3OH
concentrations in methanol/water RF plasmas by TDLAS and
found that methanol was almost completely dissociated even
at medium applied power levels [44]. In 1999 a group of
eleven species, CH4,C
2H2,C
2H4,C
2H6, CO, CO2,CH
3,
H2O, CH2O, CH3OH, HCOOH, were detected in O2–H2–
Ar microwave plasmas with small admixtures of methane or
methanol by TDLAS [13]. Busch et al monitored the densities
and temperatures of CH4,O2,CH
3, CO and CO2and studied
aspects of the chemistry in a capacitively coupled RF discharge
in 2001 [20].
2.2. Molecular microwave plasmas containing hydrocarbons
In recent years several types of microwave discharge
containing hydrocarbons as precursor gases have been at the
centre of interest. The most recent applications of TDLAS
for plasma diagnostic purposes include studies in which many
different species have been monitored under identical plasma
conditions [35,45]. These experimental data have frequently
been used to model plasma chemical phenomena.
2.2.1. Studies in planar microwave reactors. In 2003
Hempel et al studied hydrocarbon plasmas with admixtures
of nitrogen in a planar microwave reactor [46] using a
tuneable diode laser (TDL) spectrometer [45]. The interest
in such plasmas is based on various applications including
deposition of diamond layers [47–49] and of hydrogenated
carbon nitride films [50–53], detoxification of combustion
gases [54], conversion to higher hydrocarbons [55], studies
of astronomical objects such as interstellar clouds and stellar
atmospheres [56,57]. Such types of plasma are also gaining
importance in fusion physics, since they are representative of
the edge discharges observed in the proximity of the carbon
surfaces of the tokamak divertors [58].
Figure 3shows the experimental arrangement. Hempel
and co-workers used TDLAS to detect the methyl radical and
nine stable molecules, CH4,CH
3OH, C2H2,C
2H4,C
2H6,
NH3, HCN, CH2O and C2N2,inH
2–Ar–N2microwave
plasmas containing up to 7% of methane or methanol,
under both flowing and static conditions. The degree of
dissociation of the hydrocarbon precursor molecules varied
between 20% and 97%. The methyl radical concentration
was found to be in the range 1012–1013 molecules cm−3.It
was established by analysing the temporal development of the
molecular concentrations under static conditions that HCN and
NH3are the final products of plasma chemical conversion.
The fragmentation rates of methane and methanol and the
respective conversion rates to methane, hydrogen cyanide and
ammonia were determined for different relative proportions of
hydrogen to nitrogen.
The novel experimental aspect introduced by Hempel et al
was the installation of multiple pass optics directly within the
plasmas reactor to achieve higher sensitivity. Twenty four
passes were realized with the White cell arrangement, leading
to an optical length inside the reactor of about 36 m [59].
Figure 3(b) shows a ray diagram of the alignment within the
White cell [60].
S151

JR
¨
opcke et al
606.1 606.2
0.6
0.7
0.8
0.9
1.0
7
6
45
3
21
wavenumber [cm-1]
intensity [a.u.]
16.500 16.498 16.496
wavelength [µm]
Figure 4. TDL absorption spectra of some methyl and methanol
lines in a H2–Ar–N2–CH3OH microwave discharge (1,3,7—CH3;
2—CH3OH; 4,5,6—N2O). The dotted lines due to N2O are from a
reference gas cell placed in the beam path [45].
Figure 5. Molecular concentrations in a methanol containing
discharge under flowing conditions as a function of the
nitrogen/hydrogen flow rate (—T–CH3OH; ♦—HCN, ◦—NH3,
✳—CH3,—CH4,×—CH2O, —C2H2, +—C2H4,
∇—C2H6)[45].
In fact, it is sometimes possible to detect the IR spectra of
more than one species in a single laser mode using TDLAS, as
shown in figure 4. As an example of the experimental results
figure 5gives an overview of the mass balance and degree
of dissociation, as well as the product concentrations which
range over five orders of magnitude in a methanol containing
discharge under flowing conditions as a function of the nitrogen
flow rate. A key objective of this type of study is to be able to
model the chemistry of the plasma, for which it is necessary
to monitor as many plasma species as possible.
In an earlier paper chemical modelling was successfully
used to predict the concentrations of molecular species
in methane plasmas in the absence of oxygen and the
concentration trends of the major chemical products as oxygen
were added [13,61,62].
In the work of Hempel et al chemical modelling of the
methane plasma with admixtures of nitrogen under static
conditions was performed to predict the concentrations of those
gaseous species which had been detected so far. A total of
Figure 6. Calculated total electron impact dissociation rate
coefficients, kdis, for hydrogen, nitrogen and various hydrocarbons
as a function of the reduced electric field strength, E0/N, (left
panel: —H2,•—N2,—CH, ◦—CH2,—CH3,∇—CH4; right
panel: —C2H, ◦—C2H2,—C2H3,∇—C2H4,♦—C2H5,
—C2H6)[45].
NH3HCN CH3CH4C2H2C2H4C2H6
1E11
1E12
1E13
1E14
1E15
concentration [molecules cm-3]
Figure 7. Comparison of species concentrations in a representative
H2–N2–Ar–CH4-plasma (white—measured by TDLAS,
grey—calculated) [45].
145 reactions for 22 gaseous species were included in the
model leading to relatively close agreement of experimental
and calculated concentrations and to improved knowledge of
the main chemical reaction pathways and plasma chemical
processes [45].
The rate coefficients for the electron collision processes
have been determined by solving the time-dependent
Boltzmann equation for the given values of the reduced electric
field, microwave frequency and mixture composition up to
the establishment of the steady state. This electron kinetic
equation has been solved by means of the multiterm method
described by Loffhagen and Winkler [63]. Respective cross
sections for electron impact collisions for hydrogen, argon,
oxygen and methane were taken from the established literature
[61,64,65]. Figure 6shows the calculated total electron
impact dissociation rate coefficients for hydrogen, nitrogen
and various hydrocarbons as a function of the reduced electric
field strength.
A comparison of modelled and experimental species
concentrations in a representative H2–N2–Ar–CH4-plasma is
presented in figure 7showing good agreement between them.
S152

TDLAS for plasma diagnostics
microwave
waveguide
metallic
cavity
back reflector substrate holder plasma region IRMA system
quartz bell
KBr window
antenna
IR
detector
monochromator
Figure 8. Schematic diagram of the microwave PECVD bell jar
reactor used for nanocrystalline and polycrystalline diamond
deposition and the integrated TDL spectrometer (IRMA) [68].
2.2.2. Studies in bell jar microwave reactors. Microwave
PECVD bell jar reactors are used to produce diamond films
with different morphologies depending on the gas mixture
used, e.g. poly-crystalline diamond (PCD) for CH4strongly
diluted in H2and nano-crystalline diamond (NCD) for
Ar/H2/CH4plasmas. The microwave plasmas generated in
a bell jar reactor present a very complex chemistry, due to
significant thermal gradients between the plasma bulk and the
growing diamond surface. In order to improve the deposition
process and optimize growth rates, a detailed spatially resolved
physico-chemical modelling of the discharge is required.
These models need to be experimentally validated, and TDLAS
is a particularly relevant tool to provide hydrocarbon densities
which can be compared with calculated values.
As an example, one can give a brief comment about studies
performed in discharges used for NCD synthesis. Since the
demonstration of the feasibility of its deposition ten years ago
[66], NCD films have been the subject of increasing interest.
This interest was motivated by the fact that, in addition to some
physical properties similar to those of PCD, NCD possesses a
very low thickness-independent roughness suitable for some
tribological and electronic applications that often require very
smooth films [66,67].
In a recent paper Lombardi and co-workers characterized
Ar/H2/CH4microwave discharges used for nanocrystalline
diamond deposition in a bell jar cavity reactor by both
experimental and modelling investigations [68]. Figure 8
shows a schematic diagram of the microwave PECVD bell
jar reactor used for nanocrystalline diamond deposition and
Ar/H2/CH4microwave discharge diagnostics. The usual feed
gas used for the microwave PECVD process employed for
NCD film synthesis is an Ar/H2/CH4gas mixture characterized
by low H2and CH4concentrations [66].
Discharges containing 1% CH4and H2percentages
ranging between 2% and 7% were analysed as a function of the
input microwave power under a pressure of 200 mbar. TDLAS
was employed in order to measure the mole fractions of carbon-
containing species such as CH4,C
2H2and C2H6present. A
thermo-chemical model was developed and used in order to
estimate the discharge composition, the gas temperature and
the average electron energy assuming a quasi-homogeneous
Figure 9. CH4mole fraction calculated in the Ar/H2/CH4
microwave discharges under 200mbar as a function of the absorbed
MWP and for different %H2[68].
plasma. Experiments and calculations yielded consistent
results with respect to plasma temperature and composition.
The CH4mole fraction calculated in the Ar/H2/CH4microwave
discharges under 200 mbar pressure as a function of the
absorbed MWP and for different %H2is shown in figure 9.
Similar analysis associating modelling and TDLAS
measurements were also performed in the case of PCD
deposition using H2/CH4plasmas [69,70], with special
emphasis on the detection of the methyl radical which plays
a major role in the gas phase chemistry and the deposition
process.
2.2.3. Studies in surface wave discharges. A challenging
subject of plasma technology is the effective conversion of
natural gas, which has methane as the main constituent,
to higher hydrocarbons. In the last decade, a variety
of papers have been published studying the conversion
of hydrocarbons by different experimental and theoretical
approaches. Only some recent examples shall be given here;
a more comprehensive discussion can be found in [13,61].
Bugaev et al investigated the oxidative conversion of a mixture
of natural gas and oxygen in a barrier-discharge by monitoring
stable reaction products using gas chromatography [71]. Hsieh
et al used a RF plasma to convert methane into acetylene,
ethylene and ethane, which were detected by Fourier transform
infrared spectrometry [72].
In contrast to the many studies of RF plasmas systematic
investigations of plasma chemistry and kinetics in molecular
surface wave discharges are rather rare. Due to the working
conditions of the surface wave discharge, the gas flow and
the pressure can be varied over a relatively wide range. Spatial
effects appear mainly only in one dimension, in the direction of
the discharge tube axis, and can be modelled relatively easily.
TDLAS has been used to detect the methyl radical and
four stable molecules, CH4,C
2H2,C
2H4,C
2H6,inaH
2surface
wave discharge (f=2.45 GHz, power density ≈10–
50Wcm
−3)containing up to 10% of methane under different
flows (: 22–385sccm) and pressures (p:0.1–4 Torr).
Figure 10 shows the schematic diagram of the experimental
set-up [62]. For the TDLAS measurements a compact and
S153

JR
¨
opcke et al
gas flow
pump
short circuit
plasma
microwave
quartz tube
IRMA
IR detector
Figure 10. Schematic diagram of the experimental set-up used for
studies in surface wave discharges [62].
0 100 200 300 400
1E14
1E15
conversion rate [molecules J-1]
flow [sccm]
Figure 11. Rate of conversion as a function of flow rate in a
H2–CH4plasma from a surface wave discharge measured by
TDLAS (P=600 W, p=1.3 mbar, 10% CH4),•,C
2H2;,C
2H4;
,C
2H6[62].
transportable infrared multi-component acquisition (IRMA)
system was used [24].
The degree of dissociation of the methane precursor varied
between 20 and 85% and the methyl radical concentration was
found to be of the order of 1012 molecules cm−3. The methyl
radical concentration and the concentrations of the stable C-2
hydrocarbons C2H2,C
2H4,C
2H6, produced in the plasma,
increased with increasing amounts of added CH4as well as
with increasing pressure. For the first time, fragmentation
rates of methane (RF(CH4):1×1015–2.5×1016 molecules
J−1) and conversion rates to the measured C-2 hydrocarbons
(RC(C2Hy):5×1013–3 ×1015 molecules J−1) and their
dependence on flow and pressure in a surface wave discharge
could be estimated. Figure 11 shows the rate of conversion
as a function of flow in a H2–CH4plasma in a surface wave
discharge measured by TDLAS. The influence of diffusion
and convection on the spatial distribution of the hydrocarbon
concentration in the discharge tube was considered using a
simple model [62].
The following plasma chemical phenomena were found
under flow and pressure variation: (i) a lower degree of
dissociation of CH4at higher flows but increasing absolute
rates of fragmentation leading only to slightly reduced
concentrations of the measured C-2 hydrocarbons and (ii) an
increasing degree of dissociation with pressure corresponding
to higher rates of fragmentation [62].
2.3. Fragmentation of hexamethyldisiloxane in RF discharges
Plasmas containing organosilicon precursors are used in
a variety of plasma enhanced chemical vapour deposition
(PECVD) processes to deposit thin films with advantageous
mechanical, electrical or optical properties. For many years
the well-established plasma-assisted polymerization technique
has been based on hexamethyldisiloxane (HMDSO), (CH3)3–
Si–O–Si–(CH3)3, one of the simplest siloxane compounds
[73]. Nevertheless, to open up new applications the deposition
of coatings with a wide range of chemical and physical
properties by varying plasma parameters is a challenging topic
in plasma technology.
The analysis of the properties of various discharges
containing HMDSO and of the chemical structure of deposited
layers has a long and interesting history; for details see [74,75].
As far as it is known, no absolute concentrations of reaction
products in HMDSO plasmas have been ever measured. Only
in his early paper in 1973 Schmidt used mass spectrometry to
determine the fraction of reaction products in HMDSO glow
discharges under static conditions [76]. In 1994 based on
mass spectrometry, the same group published estimates of the
relative concentrations of methyl and methane in HMDSO/Ar
discharges [77].
Recently spectroscopic diagnostic studies of pure
HMDSO plasmas and of admixtures of argon in a low-
pressure, asymmetric, capacitively coupled radio frequency
discharge (f=13.56 MHz) have been described [75]. In
these TDLAS studies, the methyl radical and three stable
molecules, CH4,C
2H2and C2H6, were detected. For the first
time the methyl radical concentration and the concentration
of hydrocarbons, produced in the plasma, have been measured
while the discharge power (P=20–200 W), total gas pressure
(p=0.08–0.6 mbar) and gas flow rate (M=1–10 sccm) were
varied. One of the objectives was to determine conversion
rates into the measured hydrocarbons as a function of flow,
power and pressure. The analysis of gas phase phenomena
was completed by deposition experiments and by mass
spectrometric studies.
The experimental arrangement of the low pressure,
asymmetric RF discharge and the tuneable diode laser (TDL)
system is shown in figure 12. As an example, figure 13 presents
molecular concentrations as a function of power in a pure
HMDSO plasma measured by TDLAS. This figure gives an
overview of the mass balance and of hydrocarbon product
concentrations, ranging over three orders of magnitude, in pure
HMDSO plasmas. An increasing amount of power coupled to
the plasma leads to higher concentrations of the (measured)
stable hydrocarbons reaching saturation values at about 80 W
[75].
The methyl radical concentration was found to be of
the order of 1013 molecules cm−3, while methane and ethane
were the dominant stable hydrocarbons with concentrations
of 1014–1015 molecules cm−3. Conversion rates to the
measured stable hydrocarbons (RC(CxHy):2×1012–2 ×
1016 molecules J−1s−1)could be estimated as a function of
power, flow, mixture and pressure. Under the experimental
conditions used, a maximum polymer deposition rate of about
400 nm min−1was found [75].
S154

TDLAS for plasma diagnostics
Figure 12. Experimental set-up for studies of fragmentation of
hexamethyldisiloxane in RF discharges [75].
0 50 100 150 200
1E12
1E13
1E14
1E15
1E16
concentration [molecules cm-3]
p
ower [W]
Figure 13. Molecular concentrations as a function of power in a
pure HMDSO plasma measured by TDLAS (HMDSO =5 sccm,
p=0.15 mbar), dashed line: added HMDSO; ,C
2H6;+,CH
4;•,
C2H2;,CH
3[75].
2.4. On NOxproduction and volatile organic compound
removal
Recent concerns about air quality have led to increasing
research in the field of pollution abatement from gas exhausts.
Apart from conventional techniques, such as catalysis,
scrubbers and active carbon, the use of electric discharges is
a promising technique for toxic gas removal, especially when
these gases are present in low concentrations [78–83]. There is
of course a wealth of literature dealing with the NOxproblem
and so only a single example dealing with new aspects is
presented here.
One of the key issues when studying plasma processing
for gas treatment is to make sure that no undesirable by-product
results from the process. Among them, NOxcompounds are
readily produced in air plasmas. The production of undesirable
NO and NO2and the removal of 3-pentanone, an example of a
volatile organic compound (VOC), has been studied in a pulsed
microwave discharge in air at near atmospheric pressures. The
influence of changing pulse duration from 25 to 500 µs and
of the pulse repetition rate from 10 to 500 Hz is reported. At
a relatively high pressure of p=800mbar plasma ignition
pressure gauge
pump
plasma
Multipath cell
Gasflow
N2/O2
VOC
diode laser IR
detector
BaTiO3
coupling
device
Figure 14. Experimental set-up: plasma generator, laser device and
the multipass cell where the infrared laser beam is collimated. The
microwave power is injected into the plasma via a ‘surfaguide’. The
plasma is generated in a quartz tube in contact with high dielectric
permittivity BaTiO3pellets. The gas pressure is 800 mbar in the
plasma reactor and 1 mbar in the multipass cell [84].
is achieved by inserting BaTiO3pellets inside the microwave
excitator.
Both NO and NO2could be measured simultaneously
using TDLAS spectroscopy in the infrared region. NOx
densities as high as several thousands ppm were found to be
produced at high specific energies. In contrast to what was
expected, the use of short pulses did not lead to an effective
curtailment of the NOxproduction. It was found that the NOx
formation depended only on the average power injected into
the plasma independent of the pulse duration and repetition
rate.
Furthermore, the efficiency of the pulsed microwave
discharge for VOC oxidation, in this case of 1400ppm of 3-
pentanone in dry air, has been studied. The VOC removal
efficiency was determined using gas chromatography. The
oxidative efficiency of the discharge was found to increase
linearly with the pulse repetition rate as well as with the
pulse duration, the power duty cycle ratio being the key
parameter [84]. Figure 14 shows the experimental set-up:
plasma generator, laser device and the multipass cell where the
infrared laser beam is collimated. In figure 15 the production
of NO and NO2as a function of the pulse duration for different
pulse repetition rates is shown.
2.5. Detecting boron hydrides in microwave plasmas
Non-equilibrium low-pressure plasmas containing boron,
commonly used in combination with other reactive molecular
gases, are widely employed for surface treatment in a variety
of PECVD processes, for example, to deposit thin hard films of
diamond or boron nitride [85–90]. Also this type of plasma is
used for producing semiconductors and thermo-polymers and
is also of interest in astrophysics and for jet engines and rocket
propulsion. Various precursor molecules such as BCl3,BF
3
or B2H6are used to provide boron containing species in the
plasma.
In contrast to the rather widespread application of
plasmas containing boron and hydrogen not much is known
about internal plasma properties and phenomena, including
processes of precursor fragmentation, plasma heating and
related dissociation processes.
S155

JR
¨
opcke et al
0 100 200 300 400 500 600
0
500
1000
1500
2000
2500
500Hz
200Hz 100Hz
50Hz
10Hz
[NO] (ppm)
pulse duration (µs)
0 100 200 300 400 500 600
0
1000
2000
3000
4000
5000
500Hz
200Hz 100Hz
50Hz
10Hz
[NO2] (ppm)
pulse duration (µs)
Figure 15. Production of NO and NO2as a function of the pulse duration for different pulse repetition rates. The gas flow rate was 250 sccm
of dry air (N2:O
2=4 :1) at 800 mbar [84].
2596.0 2596.2 2596.4 2596.6 2596.8
0
50
100
150
transmittance [a.u.]
wavenumbers [cm−1]
2596.05 2596.10 2596.15
140
145
150
BH3
2596.75 2596.80
130
135
140
145
BH3
Figure 16. Part of the absorption spectrum of diborane recorded in
the H2–Ar–B2H6(64 : 33 : 3) gas flow (dots) and in the H2–Ar–B2H6
discharge plasma at p=1.5 mbar, P=1.5 kW (open circles). The
arrows show two spectral lines of the Qbranch of the BH3radical
produced in the plasma, at 2596.077 cm−1Q(3,0)and
2596.764 cm−1Q(1,0)[94].
The structure and molecular constants of the BH3
molecule were determined from the analysis of its ν2and ν3
fundamental bands by TDLAS and FTIR, respectively [91,92].
Chemical aspects of boron nitride deposition in Ar–B2H6–
NH3and Ar–B2H6–N2RF discharges were studied by FTIR
absorption spectroscopy [87]. In the spectral range between
500 and 3500 cm−1several diborane bands were detected.
It was observed that their intensity decreased due to B2H6
dissociation caused by the discharge activity.
The basic spectroscopy and reaction kinetics of diborane
and its fragmentation products, BH3and BH [93], have been
widely studied, but investigations of the behaviour of these
boron hydrides in non-equilibrium plasmas are uncommon.
Only in their recent FTIR study of Ar–B2H6–NH3and Ar–
B2H6–N2RF discharges did Franz et al find, not surprisingly,
a decreased absorption of diborane bands in the plasma
compared with the gas phase. They were able to roughly
estimate the degree of dissociation of the precursor but the
behaviour of the other fragmentation products was not reported
[87].
The TDLAS measurements were performed with the
IRMA system to measure absorption lines of B2H6,BH
3and
2143.32 2143.36 2143.40
0.0
0.3
0.6
0.9
κ(ν) [a.u.]
wavenumbers [cm-1]
Figure 17. The spectral distribution of the absorption coefficient of
N2O(•) and BH (◦) molecules measured in a N2O reference gas
cell and in a H2–Ar–B2H6microwave plasma (P=2.5kW,
p=2.5 mbar). The proportions in the H2–Ar–B2H6gas mixture
were (64 : 33 : 3) [94].
BH. For the detection of the BH radical an optical multiple
pass arrangement was used providing higher sensitivity.
Since the TDLAS method does not provide overview
spectra over a wide range selected spectral windows were
chosen, where absorption features of the species of interest
such as B2H6,BH
3and BH could be expected. An example
of absorption spectra containing lines of B2H6and BH3is
presented in figure 16. Although the infrared spectrum of B2H6
is rather rich in lines in the vicinity of 2596 cm−1,twoBH
3
lines, even at absorptions less than a few per cent, can be clearly
identified.
The analysis of the profile of two BH absorption lines
of the 1–0 band of the BH X1+state, at 2143.39 cm−1and
at 2170.0391 cm−1[94], provided an additional approach to
measure the gas temperature (figure 17). A typical example of
an absorption spectrum of BH together with N2O lines used
for identification and calibration of the wavenumber is shown
in figure 17.
Normalized on the maximum, or in the case of diborane
normalized on its absorption with the plasma off, the
absorption coefficients of B2H6,BH
3and BH lines as a
S156

TDLAS for plasma diagnostics
0.0 0.5 1.0 1.5 2.0 2.5
0.0
0.5
1.0
absorption coefficients [a.u.]
P
[
kW
]
Figure 18. Integral absorption coefficients of B2H6(—line at
position 2596.280 cm−1),BH
3(—line at 2596.077 cm−1)and BH
(◦—line at 2143.39 cm−1)measured in the H2-Ar-B2H6discharge
plasma (64 :33 : 3 a gas mixture) as a function of the microwave
power (p=1.5 mbar, =150 sccm, in the case of diborane
normalized on its absorption in the gas flow without discharge). For
clarity, straight lines are drawn from P=1.4kWtoP=0kW[94].
function of the discharge power are shown in figure 18. The
increase in electron density (due to the increase in microwave
power) leads to a decrease in the density of diborane in
favour of BH3up to a maximum around P=1.6kW. On
further increasing the power the BH3density decreases while
the diatomic molecule boron hydride (BH) shows a rapid
growth in its density. This behaviour should be considered
as direct evidence of the electron impact character of the
dissociation in the present conditions. The straight dotted
lines drawn in figure 18 from 0 to 1.4 kW represent possible
trends of the population densities of those species in the plasma.
Unfortunately, absolute values of the particle densities cannot
be obtained yet because of the lack of assignment of the
absorption lines to certain rovibronic transitions (in the case of
B2H6)and of values of the corresponding Einstein coefficients
for the boron hydrides under study. The determination of line
strengths of the transient molecules, BH3and BH, would be
of great importance for further quantifying measurements (for
details see [94]).
3. Kinetic studies and molecular spectroscopy of
radicals
3.1. Comparative study of CH3detection by IR-TDLAS and
UV absorption techniques
The methyl radical is important not only in carbon film
deposition plasmas but also in combustion and in atmospheric
and interstellar molecular chemistry. Although in recent
years several studies to quantify the methyl concentration in
hydrocarbon plasmas have been performed in the ultraviolet
and infrared spectral range, never have both spectroscopic
approaches been compared directly to verify the applicability
of the absorption cross sections or line strengths for the
conditions under study. This comparison is of particular
importance since the validity of the line parameters is directly
related to the accuracy of calculated methyl concentrations and
in turn to the quality of related plasma chemical modelling.
214 216 218 220
0.990
0.995
1.000
208 210 212 214 216 218 220 222 224
0.975
0.980
0.985
0.990
0.995
Sfn
λ [nm]
transmittance
wavelen
g
th [nm]
fit of CH3 absorption spectrum
P-branch / Q-branch / R-branch
Figure 19. Ultraviolet transmission spectrum of methyl in a
H2–CH4microwave plasma in the bell jar reactor. The measured
spectrum is fitted by a calculated spectrum representing a sum of the
P-, Q- and R-branches of the B(2A
1)←X(2A
2) transition. A
polynomial fit, shown inset, was used to determine the baseline.
(p=25 mbar, P=600 W, MWPD =9Wcm
−3,total =200 sccm,
H2+4%CH
4admixture [35]).
For the purpose of concentration measurements, the
frequencies, absorption cross sections or line strengths and the
pressure broadening coefficients of the lines need to be known
accurately. Unfortunately, the number of quantitative studies
is relatively small and the validity of the reported values is
limited (for details see [35]).
Recently a comparative study has been performed in
plasmas of two different microwave reactors (f=2.45 GHz),
(i) in H2–Ar plasmas with small admixtures of methane or
methanol in a planar microwave reactor, at a pressure of
1.5 mbar, and (ii) in H2–CH4plasmas in a bell jar reactor,
at pressures of 25 and 32 mbar.
The planar microwave plasma reactor used for these
experiments is comparable to that shown in figure 3(a). Details
of the reactor, the infrared tuneable diode laser and broadband
ultraviolet spectrometer can be found elsewhere [95–97]. A
schematic diagram of the bell jar reactor together with the
optical arrangement is shown in figure 8[33,35,98].
For classical absorption measurements in the ultraviolet
spectral range a deuterium lamp is the light source and a
monochromator of medium spectral resolution combined with
(i) an optical multi-channel analyser as detector (method 1)
or (ii) employing chopper modulation and a photo multiplier
detector together with a lock-in amplifier (method 2). In the
current work, the absorbance of the methyl radical in the
ultraviolet spectral region was determined by the best fit of
the measured spectrum to a calculated spectrum representing
the P-, Q- and R-branches of the B(2A
1)←X(2A
2)
transition. Figure 19 shows a typical example of an ultraviolet
transmission spectrum of methyl, together with the related
fits, in a H2–Ar–CH3OH microwave plasma using the planar
reactor.
For infrared absorption measurements in the planar
reactor, a one-channel TDLAS system [13] and, in the bell jar
reactor, the IRMA system were used [24]. The data acquisition
method was an advanced form of sweep integration including
S157

JR
¨
opcke et al
010203040
0
2
4
6
8
10
[CH3][1013molecule cm-3]
CH4 [sccm]
Figure 20. The concentration of the methyl radical as a function of
the methane admixture in the planar microwave reactor measured by
ultraviolet and infrared absorption techniques ( —TDLAS Q(3,3)
line, ×—UV absorption method 1, ◦—UV absorption method 2,
p=1.5 mbar, P=1.5kW,total =555 sccm) [35,101].
the non-linear least-squares fitting of the methyl line shapes
[99,100].
Figure 20 shows the concentration of the methyl radical
as a function of the flow rate of methane measured by
ultraviolet and infrared absorption techniques in the planar
microwave reactor. Taking into account the uncertainties of
the UV absorption cross section and of the line strength,
the data shown in figure 20 demonstrate that the results of
both techniques are in excellent accord. In addition to data
former published by Lombardi et al [35] the influence of
stimulated emission has been considered [101]. In summary,
the present comparative quantitative study of the methyl radical
by broadband ultraviolet and infrared absorption techniques in
the non-equilibrium plasma of a planar microwave reactor has
led to a significant result: the application of the CH3absorption
cross section of the B(2A
1)←X(2A
2)transition at 216 nm,
reported by Davidson et al [96], and of the line strength of
the Q(8,8)line of the ν2fundamental band near to 16.6 µm,
found by Wormhoudt and Mc Curdy [37], gives, within the
errors of the measurements and of the spectroscopic data, the
same concentration. For further details see Lombardi et al [35].
3.2. Line strengths and transition dipole moment of CH3
This section describes a new measurement of µ2for the ν2
fundamental band of the methyl radical in order to resolve the
differences between earlier experimentally measured values
and between experiment and theory. The method used for
determining the absolute methyl radical concentrations was
the same as that used by Yamada and Hirota [15]. However,
integrated intensities and many more methyl radical lines were
used. Furthermore, the kinetic conditions were more precisely
specified and the temperature determined more exactly. The
resulting value of µ2is now in much better agreement with
theory.
The methyl radical has no electric dipole allowed
rotational transitions because of its D3hsymmetry and so
606.0 606.1 606.2 606.3
0.0
0.4
0.8
1.2
Q4,4
Q1,1
Q3,3
Q2,2
[cm-1]
transmittance
Figure 21. Survey spectrum showing several (J=K)Qbranch
lines of the v2fundamental of the CH3free radical around the band
origin. The spectrum represented by the dashed line is a calibration
spectrum from N2O and CO2[38].
IR spectroscopy is one of the few suitable methods for its
detection. The determination of methyl radical concentrations
in terrestrial and astronomical sources using IR spectroscopy
relies on the availability of accurate line strengths and
transition dipole moments. The ν2band of CH3is the strongest
among its IR active fundamentals and particularly useful for
quantitative measurements. The need for a more accurate and
precise value of µ2has been highlighted by the measurements
of CH3in the atmospheres of Saturn [102], Neptune [103] and
in the interstellar medium [104].
The experimental set-up of the planar microwave plasma
reactor with the optical arrangements used for the methyl
transition dipole moment study is comparable to that shown
in figure 3(a). Details of the diode laser spectrometer, IRMA,
and discharge absorption cell have been reported elsewhere
[24,46]. The methyl radical was produced in mixtures of
tertiary butyl peroxide ([(CH3)3CO]2)and argon at a total
pressure of 1 mbar. Two kinds of experiment were performed:
(a) time-dependent measurement of the decay of the absorption
coefficient when the discharge was turned off, to obtain
absolute methyl concentrations and (b) measurements of the
absorption coefficients of different rovibronic lines. In total
ten lines were studied in the fundamental band, seven in the
first hot band and one from the second hot band. A survey
spectrum of the Q-branch region of the ν2fundamental band
is shown in figure 21.
In order to derive accurate line strengths and the
transition dipole moment, it is necessary to obtain the absolute
concentration of the methyl radical and its temperature in the
discharge. The decay method was the experimental approach
for methyl radical concentration measurements. The plasma
was switched on and off for periods of ten seconds and
the decay of the methyl radical signal measured during the
off period with ms time resolution [99,100]. The absolute
concentration was obtained from the decay of the integrated
absorption coefficient and the recombination rate constant. It
is well known from numerous kinetic studies that the main loss
channel under the conditions used here is self-recombination
via a three body reaction. Hence, by measuring the integrated
absorption coefficient as a function of time and knowing
S158

TDLAS for plasma diagnostics
200 400 600 800 1000 1200 1400 160
0
1E-23
1E-22
1E-21
1E-20
1E-19
1E-18
S [cm molecule-1]
temperature [K]
Q(1,1)
Q(2,2)
Q(3,3)
Q(4,4)
Q(5,5)
Q(6,6)
Q(8,8)
Q(10,10)
Q(12,12)
Figure 22. Line strengths, S, for different Qbranch transitions of CH3as a function of temperature. Values calculated from the reference
temperature values at 296 K [38].
the value of the recombination rate constant the absolute
concentration of the methyl radical can be obtained.
The rate constant k1for the self-recombination reaction
of methyl radicals has been extensively investigated in
experimental and theoretical work [105–112]. The selected
value for k1was based on the compilation of Baulch et al
and was appropriate for the specific temperature and argon
concentration [38,113]. The translational, rotational and
vibrational temperatures of the methyl radical were measured.
A near similarity of Ttrans and Trot was observed. Based on
experimental results the vibrational temperature was found
to be in equilibrium with the translational and rotational
temperature within experimental uncertainties, i.e. Tvib =
600 K. For details see [38,101].
Figure 22 shows an example of the temperature
dependence of the line strengths of several transitions from the
lower energy levels. The line strengths have been calculated
from the experimental data. It should be mentioned that the
line strength dependences can be used to determine the gas
temperature when two absorption coefficients are measured.
Measuring the ratio of two absorption coefficients is the
same as the ratio of their line strengths. Taking a measured
absorption coefficient ratio, the temperature for which their line
strength ratio is the same determines the rotational temperature.
In recent measurements (not shown here) which were done
in a hot filament diamond deposition reactor [114] the gas
temperature has been determined from the ratio of Q(6,6)
and Q(12,12). The gas temperature was found to be about
800 K near the deposition substrate, which was essentially the
same as the temperature measured by a thermocouple probe.
The line strengths of the nine Qbranch lines in the
ν2fundamental band of the methyl radical in its ground
electronic state were used to derive a more accurate value of the
transition dipole moment of this band: µ2=0.215(25)Debye.
Improved accuracy over earlier measurements of µwas
obtained by integrating over the complete line profile instead
of measuring the peak absorption and assuming a Doppler line
width to deduce the concentration and the derivation of more
accurate temperatures by examining a large number of lines.
In addition a more precise value for the rate constant for methyl
radical recombination than available earlier was employed.
The new value of µ2is in very good agreement with high
quality ab initio calculations. Furthermore, the ratio of the
transition dipole moments of the ν2and ν3fundamental bands
in the gas phase is now in highly satisfactory agreement with
the ratio determined for the condensed phase. Figure 23 shows
the chronological summary of calculated and measured values
of µ2the transition dipole moment of the ν2fundamental band.
3.3. Molecular spectroscopy of the BO radical
Transient molecular species, to be found in the gas phase
and in plasmas in particular free radicals, are amongst the
most fascinating species in the field of molecular spectroscopy.
They are of special importance for several areas of reaction
kinetics and chemistry. The study of the high resolution spectra
of radicals provides a very effective approach to understand
their properties and dynamics. There is continuing interest
in the boron and oxygen containing intermediates [115] that
may be formed during the oxidation of elemental boron
leading eventually to the final, thermodynamically stable,
oxide product, B2O3. However, relatively little is known
about the spectroscopy of these species in the gas phase.
Only the simple oxides BO and BO2have been investigated
in detail by high-resolution spectroscopy. Both of these
radicals have electronic emission spectra in the visible region
of the spectrum. In the case of BO there are two extensive
band systems, the A2–X2and B2–X2, which have
been studied since the early days of quantum mechanics, for
example, by Mulliken [116]. Most relevant are the recent
analyses of these band systems by Coxon et al [117] and M´
elen
et al [118]. In addition, the microwave spectra of the two
lowest rotational transitions of the radical have been reported
by Tanimoto et al [119].
S159

JR
¨
opcke et al
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.4
5
k1 - new data
J Pacansky, W Koch and M D Miller
theoretical calculations (1990) [128]
G D Stancu, P B Davies, J Röpcke
measurement (2005) [96]
J Wormhoudt and K E McCurdy
measurement (1989) [37]
P Botshwina, J Flesch and W Meyer
wavefunctions calculation (1983) [127]
C Yamada and E Hirota
measurement (1983) [36]
µ(v=1 0) [D]
C Yamada, E Hirota and K Kawaguchi
estimated (1981) [46]
Figure 23. Chronological summary of calculated and measured values of µ2, the transition dipole moment of the ν2fundamental band of
CH3[38].
Recently, a study has reported on the fundamental bands
of the 10B and 11 B isotopomers of BO in their ground 2+
states, detected in natural abundance in the O + BCl3reac-
tion, at Doppler limited resolution using tuneable diode laser
absorption spectroscopy. It extended a preliminary study re-
ported earlier, in which only the spectrum of 11BO was ex-
amined quantitatively [120]. The radicals were made in a 1 m
long White-type absorption cell with 26 passes, following the
method described by Clyne et al and Llewellyn et al [121,122].
Oxygen atoms were produced by partially titrating a flow of ni-
trogen atoms, generated in a 2.45 GHz microwave discharge in
molecular nitrogen, with nitric oxide to just below the end point
so that some nitrogen atoms remained in the gas flow. Boron
trichloride vapour was then added to produce an intense, light
blue chemiluminescence from the emission of BO.
The fundamental band origins for 10B16 O and 11B16 O are
found to be 1915.30674(14) cm−1and 1861.92409(13) cm−1,
respectively. The rotational constants for the lower states
are 1.877680(22) cm−1and 1.773421(10) cm−1, which are
in excellent agreement with the microwave values for B0.
Assuming a linear dependence of Bvon (v+1
2) the derived
value of Reis 1.204552(5) Å. The stick diagram shown in
figure 24 displays the detected lines for both species calculated
at a rotational temperature of 300 K [123].
3.4. Molecular spectroscopy of the CN radical
The CN radical is of fundamental importance in laboratory
spectroscopy and in astrophysics. Electronic emission spectra
arising from the red (A2–X2+) and violet (B2+–X2+)
band systems excited in flames and discharges have been
studied in the laboratory over decades while CN spectra have
been detected in the atmospheres of stars and in the interstellar
medium. Most recently the electronic band systems have been
very extensively measured and analysed in emission [124,125]
using high resolution Fourier transform spectroscopy.
Rotationally resolved spectra of the fundamental band of
the CN free radical in four isotopic forms have been measured
1820 1840 1860 1880 1900 1920 1940 1960
0.0
0.2
0.4
0.6
0.8
1.0
intensity [a.u.]
cm-1
Figure 24. Stick diagram of the measured 10 B16O (bold) and 11 B16 O
lines. The intensities were calculated for 300 K [123].
using tuneable diode laser absorption spectroscopy [126]. The
source of the radical was a microwave discharge in a mixture of
isotopically selected methane and nitrogen diluted with argon.
The lines were measured to an accuracy of 5 ×10−4cm−1and
fitted to the formula for the vibration rotation spectrum of a di-
atomic molecule, including quartic distortion constants. The
band origins of each of the isotopomers from the five parameter
fits were found to be 12C14 N: 2042.42115(38) cm−1,13 C14N:
2000.08479(23) cm−1,12C15 N: 2011.25594(25) cm−1,13 C15N:
1968.22093(33) cm−1with one standard deviation from the fit
given in parenthesis. Some of the lines showed a resolved
splitting due to the spin rotation interaction. This was av-
eraged for fitting purposes. The average equilibrium inter-
nuclear distance derived from the v=0 and v=1 ro-
tational constants of the four isotopomers is 1.171800(6) Å
which is in good agreement with the value determined from
microwave spectroscopy. Figure 25 shows a stick dia-
gram of all the lines measured in the four isotopic forms
and their intensities calculated for a rotational temperature
of 950 K [126].
S160

TDLAS for plasma diagnostics
1900 1950 2000 2050 2100
0
5
10
15
20
intensity [a.u.]
wavenumber [cm-1]
1900 1950 2000 2050 2100
0
5
10
15
intensity [a.u.]
wavenumber [cm-1]
(a) (b)
Figure 25. Stick diagrams showing the detected absorption lines from the fundamental band of CN and their intensities at 950 K: (a)12C14N
and 13C14 N(······) and (b)12C15 N and 13C15 N(······)[126].
Figure 26. Photograph of the IRMA system: left—optical table,
right—data acquisition rack with diode laser controllers and
industrial computers [24].
4. Infrared absorption for plasmas diagnostics and
control
4.1. The IRMA system
In order to exploit the capabilities of infrared tuneable diode
laser absorption spectroscopy for effective and reliable on-
line plasma diagnostics and process control in research and
industry, a compact and transportable tuneable infrared multi-
component acquisition system (IRMA) has been developed
[24]. Figure 26 shows a photograph of the complete IRMA
system including the optical table and the data acquisition rack.
The IRMA system contains four independent laser
stations. The narrowband infrared emission of four lead-
salt diode lasers, which can be temporally multiplexed, is
used to monitor the infrared absorption features of the target
species. The arrangement of the instrument optical table
has the dimensions of 110 ×60 cm2. The four diode lasers
are mounted in individual cold stations, which are thermally
coupled to the cold finger of a closed cycle cryostat. The
temperature of each laser is controlled at milli-Kelvin precision
between 30 and 100 K. A single cold station can be heated up
to room temperature and decoupled from the main vacuum to
allow laser replacement in typically 90 min with the other laser
stations remaining operative. Four grating monochromators
serve as mode filters. The light of the four lasers is converged
into a single beam, which is used for measurement and
reference purpose, respectively. The measurement path can
either be passed through a measurement cell, for example a
plasma, external to the optical table or through an internal
multi-path astigmatic Herriot cell. This cell with a path length
of 36 m is included for exhaust gas detection. In the reference
path, the spectral absorption from small optical cells containing
the target gas at high concentration is monitored as input for
a laser emission wavelength control loop (line-locking). The
two measurement beams and the reference beam are focused on
photoconductive infrared detectors mounted in liquid nitrogen
cooled dewars.
The data acquisition system for the four-channel tuneable
diode system consists of two computers combined with
two high speed (300 kHz) boards and two dual diode laser
controllers housed in a single transportable rack (figure 26).
Based on rapid scan software using direct absorption with
sweep integration, the absolute concentrations of several
molecular species can be measured simultaneously within
milliseconds and used as digital output for on-line process
control. The software of the system is able to fit up to four
species simultaneously using up to 45 individual spectral lines
per species.
The result of time-dependent species density measurement
is useful especially for process control. The burst data
acquisition mode provides the highest possible time-resolution
by transferring fifty measured spectra directly to the extended
memory of the computer following a trigger pulse [24]. The
IRMA system has been applied successfully in several studies
of molecular phenomena in plasmas, e.g. [38,62,68,70,94,
128,129].
4.2. The TOBI system
For further progress in understanding molecular processes,
in particular in plasmas under non-stationary excitation
conditions, there is a need for improved temporal resolution of
TDLAS, i.e. for improved, sensitive high speed diagnostics.
For these purposes a compact and transportable two laser
beam infrared (TOBI) (figure 27) system has been designed
for simultaneous measurement of two gaseous species with
two tuneable lasers operating simultaneously. In comparison
S161

JR
¨
opcke et al
Figure 27. Photographic overview of the TOBI system [25].
with the IRMA system TOBI is mainly focused on high
speed detection. A new generation of rapid scan software
for the Windows operating system (‘TDL Wintel’) allows
investigation of transient plasma conditions on time scales
of tens of microseconds. This program implements direct
absorption with sweep integration to measure the absolute
concentrations of several molecular species and provides
continuous digital output, which can be used for process
control. The TDL Wintel program has many further advanced
features. This includes several convenient frequency locking
options as well as methods for suppressing background
interference. The laser sweep and signal digitization cycle
can be synchronized with a pulsed plasma apparatus. The
program will collect spectra which are the ratios of the spectra
with the discharge on to those with the discharge off. This
provides a highly sensitive method for searching for transient
species. The software also provides spectral simulation,
remote instrument operation and gas concentration calibration.
The optical system can switch between two measurement
modes, either directing the probe beams through the plasma
apparatus or through a 100 m multi-pass cell to measure gas
extracted from an operating plasma. The capabilities of the
TOBI system have been demonstrated in plasmas of pulsed H2–
N2surface wave and pulsed air–CH4DCdischarges (figure 28).
The potential use of the TOBI system can be summarized
as being mainly focused on (i) the high speed detection of
transient molecular species, radicals and molecular ions, in
plasmas under non-stationary excitation conditions using fast
external detectors, and on (ii) the sensitive (sub-ppb) trace gas
detection based on the multi-pass absorption cell [25].
4.3. The Q-MACS system
For reasons of enhanced efficiency, increased stability and
product quality the direct control of plasma applications is
a challenging subject for plasma technology. Therefore,
appropriate diagnostic tools are necessary allowing for on-line
process monitoring in such applications. Pulsed QCLs are able
to emit mid-IR radiation near room temperature. Compared
with lead salt lasers, QCLs allow the realization of very
compact mid-infrared sources characterized by narrow line
width combining single-frequency operation and considerably
0123
1013
1014
300 sccm
100 sccm
30 sccm
[NH3] [molecules cm-3]
time [s]
Figure 28. Example of NH3concentrations as function of time in a
pulsed N2–H2surface wave discharge (f=2.45 GHz,
Ppulse =1 kW, pulse =1 s, of period =5 s, flow conditions,
N2/H2=9/1, p=1.3 mbar) [25].
Figure 29. Q-MACS-Basic system with laser head (left-hand side,
size 700 cm3), supply unit and connection cable [131].
higher power values, i.e. of tens of mW. The output power
is sufficient to combine them with thermoelectrically cooled
infrared detectors, which permits a decrease in the apparatus
size and gives a unique opportunity to design compact liquid
nitrogen-free mid-IR spectroscopic systems. These positive
features of quantum cascade laser absorption spectroscopy
(QCLAS) can open up new fields of application in research
and industry, including studies of gases in atmospheric,
environmental and plasma chemistry but also for in situ control
of industrial plasma processes.
Recently, a compact quantum cascade laser measurement
and control system (Q-MACS) has been developed for time-
resolved plasma diagnostics, process control and trace gas
monitoring [130,131]. The Q-MACS system contains a
tuneable quantum cascade laser which can be directed through
a plasma or into a multi-pass cell for exhaust gas detection.
Rapid scan software with real-time line shape fitting provides a
time resolution up to 1 µs to study kinetic processes of infrared
active compounds in plasmas or gases. The Q-MACS-Basic
system has been designed as a platform for various applications
of QCLAS (figure 29).
Measurements with the Q-MACS-Basic were performed
in an industrial pulsed plasma reactor used to deposit boron
based hard coatings. Figure 30 shows an absorption spectrum
of diborane near 1613 cm−1recorded with 5% admixture of
this molecular precursor gas in hydrogen and argon. Using the
S162

TDLAS for plasma diagnostics
Figure 30. Fitted absorption spectrum of diborane at 1613 cm−1
measured in an industrial pulsed DC reactor with 5% diborane at the
gas inlet (p=200 Pa) [131].
Figure 31. Absorption spectrum of N2O contained in a reference
gas cell (l=15 cm, p=6.6 mbar) at 2235 cm−1and also the etalon
pattern of known fringe spacing (fsr =0.049 cm−1)both recorded
in a single QCL pulse (t=100 ns) [131].
software package ‘TDLWintel’ it was possible to continuously
monitor the concentration change of diborane while driving
a process plasma. The measurement results are not only
displayed but can be transferred via hardware interfaces to
other systems, e.g. for the use in automated process monitoring
and controlling. These measurements prove that QCLAS can
be useful for improving the reliability and effectiveness of
industrial plasma processes.
The scan through an infrared spectrum can be achieved by
a bias dc ramp, in the interpulse mode. Another option is the
scanning in single, longer pulses, using the intrapulse mode.
An example of a 1.7 cm−1long scan in this mode is shown in
figure 31. Since the duty cycle of QCL is about 1 per cent
the temporal resolution in this mode can be as good as a few
µs. Therefore, it fits very well to measurements of rapidly
changing chemical processes.
For high sensitivity trace gas measurements a compact
and transportable measurement system, the Q-MACS-Trace,
was developed. It combines the Q-MACS-Basic with a 56 m
astigmatic multi-pass absorption cell of the Herriot type [132],
a compact optical system and a computer system. The
sensitivity of the Q-MACS-Trace was as high as a few ppb
Figure 32. Photograph of the Q-MACS-Trace system with optical
arrangement and laptop for analysis. (1 channel system with 56 m
long path cell, sensitivity: ppb [131].)
(parts per billion—1/109). Figure 32 shows a photograph
of the Q-MACS-Trace with optical arrangement and laptop
for analysis. The Q-MACS-Trace optical system is designed
around the QCL laser head.
5. Summary and conclusions
During the past few years a variety of phenomena in molecular
non-equilibrium plasmas in which many short-lived and stable
species are produced have been successfully studied based on
diode laser absorption techniques in the mid-infrared spectral
range, with which the present review paper is concerned. It
has been possible to determine absolute concentrations of
ground states using spectroscopy thereby providing a link
with chemical modelling of the plasma, the ultimate objective
being to understand better the chemical and reaction kinetic
processes occurring in the plasma. In table 1the major results
are shown. The other essential component needed to reach this
objective is to determine physical parameters of the plasma, for
example, temperatures, degrees of dissociation and dynamics
of reaction kinetic processes, and the present paper discusses
methods for achieving this. The need for a better scientific
understanding of plasma physics and chemistry has stimulated
the application of TDLAS, which has been proven to be one of
the most versatile techniques for studying molecular plasmas.
Based on the recent development of quantum cascade lasers the
further spread of this method of high resolution mid-infrared
spectroscopy to industrial applications has become a reality.
S163

JR
¨
opcke et al
Table 1. Tabulation of the major results.
Wavenumber Plasma
Molecular region used Concentration Gas chemistry
species cm−1Type of plasma range temperature modelling References
Relative
BH 2143.39 MW concentration 1000 K No [94]
2170.039 up to 8 mbar
BH32596 Ar/H2/B2H6
B2H61583
2596 Relative
10B16 O∼1915 MW post-discharge concentration 300 K No [120,123]
200 mTorr
11B16 O∼1861 N2/O2/BCl3Relative
CF 1298 RF concentration 350–400K No [9,12,14]
CF21096 up to 500mTorr
CF31260 Ar+CF4or CHF3
CF41283
CH3606–608 ac parallel plate reactor ∼1012 cm−3325 K Yes [10,13]
CH4or CH4/H2/O2Relative
606–612 AC glow discharge concentration 400–800K No [15]
[(CH3)3CO]2
CH3I
(CH3)2CO
CH3SH
CH3OH
∼606 RF No [16–18]
CH4diluted — —
in He, Ne, Ar,Kr or Xe
∼606 RF+MW No [41,43]
H2/CH3OH — —
∼606 RF 1 mbar 1011–1012 cm−3∼300–500 K No [20]
CH4
RF ∼1012 cm−3No [44]
CH3OH/H2O 1 Torr — —
606.120 32 RF Ar/HMDSO ∼1013 cm−3500 K No [75]
0.08–0.6 mbar
∼606 ECR ∼1011 cm−3No [42]
H2/CH4or CH3OH
1.3 Pa
606.120 32 MW 1012 –1013 cm−31000 K Yes [35,45]
H2/Ar/N2+CH4
or CH3OH mixture
1 mbar
MW ∼1012 cm−3800–2500 K No [62]
Surface wave discharge
Up to 10 mbar H2/CH4
612.413 44 MW Bell–Jar cavity 1013–1014 cm−32000–3500 K Yes [35,38,68–70,101]
Up to 100 mbar
H2/CH4
CH46067 RF 1 mbar ∼1016 cm−3∼300–500 K No [20]
CH4plasma
3013.7 RF ∼1014 cm−3500 K No [75]
Ar/HMDSO
0.08–0.6 mbar
1347.054 29 MW ∼1015 cm−31000 K Yes [45]
H2/Ar/N2+CH4
1347.195 15 or CH3OH mixture
1 mbar
1302.736 MW 1015–1016 cm−3800–2500 K No [62]
2978.848 Surface wave discharge
Up to 10 mbar
H2/CH4
1302.4515 MW 1015–1016 cm−32000–3500 K Yes [35,68–70]
2989.9814 Bell–Jar cavity
Up to 100 mbar
H2/CH4
S164

TDLAS for plasma diagnostics
Table 1. Continued.
Wavenumber Plasma
Molecular region used Concentration Gas chemistry
species cm−1Type of plasma range temperature modelling References
C2H2760–790 ac parallel plate reactor ∼1013 cm−3325 K Yes [13]
CH4/H2/O2
RF ∼1012 cm−3No [44]
CH3OH/H2O
1 Torr
700.9004 RF 1013–1014 cm−3500 K No [75]
Ar/HMDSO
0.08–0.6 mbar
1347.1631 MW 1013–1014 cm−31000 K Yes [45]
H2/Ar/N2+CH4
or CH3OH mixture
1 mbar
1302.5968 MW 1014–1015 cm−3800–2500 K No [62]
Surface wave discharge
Up to 10 mbar
H2/CH4
1302.5968 MW 1015–1016 cm−32000–3500 K Yes [35,68–70]
Bell–Jar cavity
Up to 100 mbar
H2/CH4
C2H4945–960 AC parallel plate reactor ∼1013 cm−3325K Yes [13]
CH4/H2/O2
947.9829 MW 1011–1012 cm−31000 K Yes [45]
H2/Ar/N2+CH4
948.002 or CH3OH mixture — — —
1 mbar
898.2883 MW 1013–1014 cm−3800–2500 K No [62]
Surface wave discharge
Up to 10 mbar
H2/CH4
800–820 AC parallel plate reactor ∼1013 cm−3325K Yes [13]
CH4/H2/O2
C2H62999.5 RF 1015–1016 cm−3500 K No [75]
Ar/HMDSO
0.08–0.6 mbar
2993.361 MW 1012–1014 cm−31000 K Yes [45]
H2/Ar/N2+CH4
2993.477 or CH3OH mixture — — —
1 mbar
2980.073 MW 1013–1014 cm−3800–2500 K No [62]
Surface wave discharge
Up to 10 mbar
H2/CH4
2993.361 MW 1014–1015 cm−32000–3500 K Yes [35,68–70]
Bell–Jar cavity
Up to 100 mbar
H2/CH4
CH3OH 1020–1045 AC parallel plate reactor ∼1013 cm−3325 K Yes [13]
CH4/H2/O2
1347.572 MW ∼1015 cm−31000 K Yes [45]
H2/Ar/N2+CH4
or CH3OH mixture
1 mbar
CH2O 1720–1750 ac parallel plate reactor ∼1014 cm−3325 K Yes [13]
CH4/H2/O2plasma
1777.3196 MW ∼1012 cm−31000 K Yes [45]
H2/Ar/N2+CH4
or CH3OH mixture
1 mbar
C2N22149.990 48 MW 1011–1012 cm−31000 K Yes [45]
H2/Ar/N2+CH4
or CH3OH mixture
1 mbar
CO 2050–2150 ac parallel plate reactor ∼1014 cm−3325 K Yes [13]
S165

JR
¨
opcke et al
Table 1. Continued.
Wavenumber Plasma
Molecular region used Concentration Gas chemistry
species cm−1Type of plasma range temperature modelling References
CO2600–620 CH4/H2/O2∼1014 cm−3
CO 2060 RF 1 mbar ∼1014–1015 cm−3∼300–500 K No [20]
CO22250 CH4/O2∼1014–1015 cm−3
HCOOH 1720–1750 ac parallel plate reactor ∼1013 cm−3325 K Yes [13]
CH4/H2/O2plasma
HCN 785.3894 MW 1014–1015 cm−31000 K Yes [45]
H2/Ar/N2+CH4
785.576 66 or CH3OH mixture
1 mbar
NH3948.232 06 MW 1013 –1014 cm−31000 K Yes [45]
H2/Ar/N2+CH4
or CH3OH mixture
1 mbar
NO 1880 Pulsed MW 100–2000 ppm 1000 K No [84]
800 mbar
N2/O2/C5H10ORelative
Si−
2740–820 AC glow discharge H2/SiH4concentration — No [22]
SiH+
3730–1015 [23]
Acknowledgments
This work was partly supported by (a) the Deutsche
Forschungsgemeinschaft, Sonderforschungsbereich 198 and
Transferbereich 36, (b) the Bundesministerium f¨
ur Bildung
und Forschung, FKZ 13N7451/8, (c) the Deutscher
Akademischer Austauschdienst and EGIDE as part of the
French–German PROCOPE Collaboration Program (Project
04607QB) and (d) in the framework of the German–
Russian–French Trilateral Cooperation Project of University
of Greifswald, St Petersburg State University and University
of Paris-South. The authors convey their sincere thanks to all
present and former members of the laboratories involved in
Greifswald, Paris and Cambridge for permanent support and
a stimulating scientific climate. In particular, the authors are
indebted to all co-authors of former papers whose contributions
made the present review possible.
References
[1] Mechold L 2000 PhD Thesis University of Greifswald,
Germany (Berlin: Logos)
[2] Demtr¨
oder W 2002 Laser Spectroscopy: Basic Concepts and
Instrumentation (Berlin: Springer)
[3] Duxbury G 2000 Infrared Vibration–Rotation Spectroscopy;
From Free Radicals to the Infrared Sky (Chichester:
Wiley)
[4] Tittel F K, Richte D and Freed A 2003 Mid-infrared laser
applications in spectroscopy Solid State Infrared Sources
ed I T Sorokina andKLVodopyanov p 445 (Berlin
Heidelberg: Springer)
[5] Romanini D and Lehmann K K 1993 J. Chem. Phys. 99 6287
[6] Romanini D, Kachanov A A, Sadeghi N and Stoeckel F 1997
Chem. Phys. Lett. 264 316
[7] Campargue A, Romanini D and Sadeghi N 1998 J. Phys. D:
Appl. Phys. 31 1168
[8] Kotterer M, Conceicao J and Maier J P 1996 Chem. Phys.
Lett. 259 233
[9] Haverlag M, Stoffels E, Stoffels W W, KroesenGMWand
de Hoog F J 1996 J. Vac. Sci. Technol. A14 380
[10] Davies P B and Martineau P M 1992 Adv. Mater. 4729
[11] Naito S, Ito N, Hattori T and Goto T 1995 Japan. J. Appl.
Phys. 34 302
[12] Haverlag M, Stoffels E, Stoffels W W, KroesenGMWand
de Hoog F J 1994 J. Vac. Sci. Technol. A12 3102
[13] R¨
opcke J, Mechold L, K¨
aning M, Fan W Y and Davies P B
1999 Plasma Chem. Plasma Process. 19 395
[14] Haverlag M, Stoffels E, Stoffels W W, KroesenGMWand
de Hoog F J 1996 J. Vac. Sci. Technol. A14 384
[15] Yamada C and Hirota E 1983 J. Chem. Phys. 78 669
[16] Naito S, Ikeda M, Ito N, Hattori T and Goto T 1993 E Japan.
J. Appl. Phys. 32 5721
[17] Naito S, Ito N, Hattori T and Goto T 1994 Japan. J. Appl.
Phys. 33 5967
[18] Ikeda M, Ito N, Hiramatsu M, Hori M and Goto T 1997
J. Appl. Phys. 82 4055
[19] KroesenGMW,denBoerJHWG,Boufendi L, Vivet F,
Khouli K, Bouchoule A and de Hoog FV 1996 J. Vac. Sci.
Technol.A14 546
[20] Busch C, M¨
oller I and Soltwisch H 2001 Plasma Sources Sci.
Technol.10 250
[21] Serdioutchenko A, M¨
oller I and Soltwisch H 2004
Spectrochim. Acta A60 3311
[22] Liu Z and Davies P B 1996 J. Chem. Phys. 105 3443
[23] Davies P B and Smith D M 1994 J. Chem. Phys.100 6166
[24] R¨
opcke J, Mechold L, K¨
aning M, Anders J, Wienhold F G,
Nelson D and Zahniser M 2000 Rev. Sci. Instrum. 71 3706
[25] McManus J B, Nelson D, Zahniser M, Mechold L, Osiac M,
R¨
opcke J and Rousseau A 2003 Rev. Sci. Instrum. 74 2709
[26] Sugai H, Kojima H, Ishida A and Toyoda H 1990 Appl. Phys.
Lett. 56 2616
[27] Sugai H and Toyoda H 1992 J. Vac. Sci Technol. A81193
[28] Zarrabian M, Leteinturier C and Turban G 1998 Plasma
Sources Sci. Technol. 7607
[29] Zalicki P, Ma Y, Zare R N, Wahl, E H, Dadamio J R,
Owano T G and Kruger C H 1995 Chem. Phys. Lett.
234 269
[30] Zalicki P, Ma Y, Zar R N, Wahl E H, Dadamio J R,
Owano T G and Kruger C H 1995 Appl. Phys. Lett. 67 144
[31] Wahl E H, Owano T G, Kruger C H, Zalicki P, Ma Y and
Zare R N 1997 Diamond. Relat. Mater. 6476
S166

TDLAS for plasma diagnostics
[32] Childs M A, Menningen K L, Chevako P, Spellmeyer N W,
Anderson L W and Lawler J E 1992 Phys. Lett. A171 87
[33] Menningen K L, Childs M A, Chevako P, Toyoda H,
Anderson L W and Lawler J 1993 Chem. Phys. Lett.
204 573
[34] Menningen K L, Childs M A, Toyoda H, Ueda Y,
Anderson L W and Lawler J E 1994 Diamond. Relat.
Mater.3422
[35] Lombardi G, Stancu G D, Hempel F, Gicquel A and R¨
opcke J
2004 Plasma Sources Sci. Technol. 13 27
[36] Wormhoudt J 1990 J. Vac. Sci. Technol. A81722
[37] Wormhoudt J and McCurdy K E 1989 Chem. Phys. Lett.
156 47
[38] Stancu G D, R¨
opcke J and Davies P B 2005 J. Chem. Phys.
122 014306
[39] Davies P B and Martineau P M 1990 Appl. Phys. Lett. 57 237
[40] Davies P B and Martineau P M 1992 J. Appl. Phys.
71 6125
[41] Naito S, Ito N, Hattori T and Goto T 1995 Japan. J. Appl.
Phys. 34 302
[42] Ikeda M, Aiso K, Hori M and Goto T 1995. Japan. J. Appl.
Phys.34 3273
[43] Ikeda M, Hori M, Goto T, Inayoshi M, Yamada K,
Hiramatsu M and Nawata M 1995 Japan. J. Appl. Phys.
34 2484
[44] Kim S, Billesbach D P and Dillon R 1997 J. Vac. Sci.
Technol. A15 2247
[45] Hempel F, Davies P B, Loffhagen D, Mechold L and
R¨
opcke J 2003 Plasma Sources Sci. Technol.12 S98
[46] Ohl A 1998 J. Phys. IV 883
[47] Vandevelde T, Nesladek M, Quaeyhaegens C and Stals L
1997 Thin Solid Films 308–309 154
[48] Vandevelde T, Wu T D, Quaeyhaegens C, Vlekken J,
D’Olieslaeger M and Stals L 1999 Thin Solid Films
340 159
[49] Mutsukura N 2001 Plasma Chem. Plasma Process.21 265
[50] Bhattacharyya S, Granier A and Turban G 1999 J. Appl.
Phys. 86 4668
[51] Zhang M, Nakayama Y, Miyazaki T and Kume M 1998
J. Appl. Phys. 85 2904
[52] Dinescu G, De Graaf A, Aldea E, and van de SandenMCM
2001 Plasma Sources Sci. Technol. 10 513
[53] De Graaf A, Aldea E, Dinescu G and van de SandenMCM
2001 Plasma Sources Sci. Technol. 10 524
[54] Penetrante B M, Hsiao M C, Bardsley J N, Merritt B T,
Vogtlin G E, Kuthi A, Burkhart C P and Bayless J R 1997
Plasma Sources Sci. Technol. 6251
[55] Kareev M, Sablier M and Fujii T 2000 J. Phys. Chem.104
7218
[56] Coll P, Coscia D, Gazeau M C, De Vanssay E, Guillemin J C
and Raulin F 1995 Adv. Space Res. 16 93
[57] De Vanssay E, Gazeau M C, Guillemin J C and Raulin F
1995 Planet. Space Sci. 43 25
[58] Tabares F L, Tafalla D, Tanarro I, Herrero V J, Islyaikin A and
Maffiotte C 2002 Plasma Phys. Control. Fusion 44 L37
[59] White J U 1942 J. Opt. Soc. Am. 32 285
[60] Gemini Scientific Instruments 2001 Multi-Pass Optics
Alignment Instructions (Manual) www.gascell.com
[61] Fan W Y, Knewstubb P F, K¨
aning M, Mechold L, R¨
opcke J
and Davies P B 1999 J. Phys. Chem. A103 4118
[62] Mechold L, R¨
opcke J, Duten X and Rousseau A 2001 Plasma
Sources Sci. Technol. 10 52
[63] Loffhagen D and Winkler R 1996 J. Phys. D: Appl. Phys.
29 618
[64] Buckman S J and Phelps A V 1985 J. Chem. Phys. 82 4999
[65] Phelps A V and Pitchford L C 1985 Phys. Rev. 31 2932
[66] Gruen D M, Pan X, Krauss A R, Liu S, Luo J and Foster C M
1994 J. Vac. Sci Technol.A12 1491
[67] B´
en´
edic F, Assouar M B, Mohasseb F, Elmazria O, Alnot P
and Gicquel A 2004 Diamond Relat. Mater. 13 347
[68] Lombardi G, Hassouni K, B´
en´
edic F, Mohasseb F, R¨
opcke J
and Gicquel A 2004 J. Appl. Phys.96 6739
[69] Lombardi G, Hassouni K, Stancu G D, Mechold L,
R¨
opcke J and Gicquel A 2005 Plasma Sources Sci.
Technol. 14 440
[70] Lombardi G, Hassouni K, Stancu G D, Mechold L, R¨
opcke J
and Gicquel A 2005 J. Appl. Phys. 98 053303
[71] Bugaev SP, Kozyrev A V, Kushinov V A, SochugovNSand
Khryapov P A 1998 Plasma Chem. Plasma Process.
18 247
[72] Hsieh L-T, Lee W-J, Chen C-Y, Chang, M-B and Chang H-C
1998 Plasma Chem. Plasma Process. 18 215
[73] Wrobel A M and Wertheimer M R 1990 Plasma-polymerized
organosilicones and organometallics Plasma Deposition,
Treatment and Etching of Polymers ed R d’ Agostino (San
Diego: Academic) p 163
[74] R¨
opcke J, Ohl A and Schmidt M 1993 J. Anal. At. Spectrom.
8803
[75] R¨
opcke J, Revalde G, Osiac M, Li K and Meichsner J 2002
Plasma Chem. Plasma Process. 22 137
[76] Schmidt M 1973 Beitr. z. Plasma Phys. 13 347
[77] Foest R, Schmidt M, Hannemann M and Basner R 1994 On
the ions in an argon-hexamethyldisiloxane
radio-frequency-discharge Gaseous Dielectrics VII ed
L G Christophorou and D R James (New York: Plenum)
p 335
[78] Van Veldhuizen E M (ed) 2000 Electrical Discharges for
Environmental Purposes Fundamentals and Applications
(Huntington, NY: NOVA Science Publishers)
[79] Penetrante B 1993 Non-Thermal Plasma Techniques for
Pollution Control (NATO ASI Series) vol G34, ed
S Schultheis (Berlin: Springer)
[80] Yamamoto T 1997 J. Electrostat. 42 227
[81] Filimonova E, Amirov R, Kim H and Park I 2000 J. Phys. D:
Appl. Phys. 33 1716
[82] Mc Adams J 2001 J. Phys. D: Appl. Phys.34 2810
[83] Filimonova E, Kim Y-H, Hong S, Han S-Y and Song Y-H
2001 Proc. 15th Int. Symp. Plasma Chem. (Orl´
eans,
France) p 3041
[84] Rousseau A, Dantier A, Gatilova L V, Ionikh Y,
R¨
opcke J and Tolmachev Y A 2005 Plasma Sources Sci.
Technol. 14 70
[85] Albella J M, Gomez-Aleixandre C, Sanchez-Garrido O,
Vazquez L and Martinez-Duart J M 1995 Surf. Coat.
Technol. 70 163
[86] Andujar J L, Bertran E and Polo M C 1998 J. Vac. Sci.
Technol. 16 578
[87] Franz D, Hollenstein M and Hollenstein C 2000 Thin Solid
Films 379 37
[88] Ichiki T, Momose T and Yoshida T 1994 J. Appl. Phys.
75 1330
[89] Ichiki T and Yoshida T 1994 Appl. Phys. Lett. 64 851
[90] Gonon P, Deneuville A, Fontaine F, Gheeraert E, Campargue
A, Chenevier M and Rodolphe S 1995 J. Appl. Phys.
78 7404
[91] Kawaguchi K, Butler J E, Yamada C, Bauer S H,
Minowa T, Kanamori H and Hirota E 1987 J. Chem. Phys.
87 2438
[92] Kawaguchi K 1992 J. Chem. Phys. 96 3411
[93] Pianalto F S, O’Brien L C, Keller P C and Bernath P F 1988
J. Mol. Spectrosc. 129 348
[94] Lavrov B P, Osiac M, Pipa A V and R¨
opcke J 2003 Plasma
Sources Sci. Technol. 12 576
[95] Arthur N L 1986 J. Chem. Soc. Faraday Trans.II 82 331
[96] Davidson D F, Chang A Y, Di Rosa M D and Hanson R K
1993 J. Quant. Spectrosc. Radiat. Transfer 49 559
[97] Gicquel A, HassouniK, Owano T, DutenX, and Cappelli M
1999 6th Int. Symp. on Diamond Material (Electrochem.
Society Proc.) (Honolulu) vol 99–32 p 41
[98] Gl¨
anzer K, Quack M and Troe J 1977 16th Int. Symp. on
Combustion (Pittsburgh: The Combustion Institute) p 949
[99] Zahniser M S, Nelson D D and Kolb C E 2002 Applied
Combustion Diagnostics ed K Kohse-Hoinghaus and
J Jeffries (New York: Taylor and Francis) p 648
S167

JR
¨
opcke et al
[100] Nelson D D, Shorter J H, McManus J B and Zahniser M S
2002 Appl. Phys. B75 343
[101] Stancu G D 2004 PhD Thesis University of Greifswald,
Germany
[102] B´
ezard B, Feuchtgruber H, Moses J I and Encrenaz T 1998
Astro. Astrophys. 334 L41
[103] Bezard B, Romani P N, Feuchtgruber H and Encrenaz T 1999
Astrophys. J. 515 868
[104] Feuchtgruber H, Helmich F P, van DishoeckEFand
Wright C M 2000 Astrophys. J. 535 L111
[105] Laguna G A and Baughcum S L 1982 Chem. Phys. Lett.
88 568
[106] Cody R J, Payne Jr W A, Thorn Jr R P, Nesbitt F L,
Iannone M A, Tardy D C and Stief L 2002 J. Phys. Chem.
A106 6060
[107] Walter D, Grotheer H H, Davies J W, PillingMJand
Wagner A F 1990 23rd Symp. on Combustion p 107
[108] Macpherson M T, Pilling M J and SmithMJC1983 Chem.
Phys. Lett. 94 430
[109] Macpherson M T, Pilling M J and SmithMJC1985 J. Phys.
Chem.89 2268
[110] Slagle I R, Gutman D, Davies J W and Pilling M J 1988
J. Phys. Chem.92 2455
[111] Wagner A F and Wardlaw D M 1988 J. Phys. Chem.92 2462
[112] Callear A B and Metcalfe M P 1976 Chem. Phys.14 275
[113] Baulch D L et al 1994 J. Phys. Chem. Ref. Data 23 980
[114] Schwarz S, Rosiwal S M, Frank M, Breidt D and Singer R F
2002 Diamond Relat. Mater. 11 589
[115] Hunt N T, R¨
opcke J and Davies P B 2000 J. Mol. Spectrosc.
204 120
[116] Mulliken R S 1925 Phys. Rev. 25 259
[117] Coxon J A, Foster S C and Naxakis S 1984 J. Mol. Spectrosc.
105 465
[118] M´
elen F, Dubois I and Bredohl H 1985 J. Phys. B: At. Mol.
Phys.18 2423
[119] Tanimoto M, Saito S and Hirota E 1986 J. Chem. Phys.
84 1210
[120] Osiac M, R¨
opcke J and Davies P B 2001 Chem. Phys. Lett.
344 92
[121] ClyneMAAandHeavenMC1980 Chem. Phys.51 299
[122] Llewellyn I P, Fontijn A and ClyneMAA1981 Chem. Phys.
Lett.84 504
[123] Stancu G D, R¨
opcke J and Davies P B 2004 J. Mol.
Spectrosc.223 181
[124] Rehfuss B D, Suh M H, Miller T A and Bondybey V E 1992
J. Mol. Spectrosc. 151 437
[125] PrasadCVV,Bernath P F, Frum C and Engleman R 1992
J. Mol. Spectrosc. 151 459
[126] H¨
ubner M, Castillo M, Davies P B and R¨
opcke J 2005
Spectrochim. Acta A Mol. Spectrosc. 61 57
[127] R¨
opcke J 2000 Acta Phys. Univ. Comen. XLI 87
[128] R¨
opcke J, Mechold L, Duten X and Rousseau A 2001
J. Phys. D: Appl. Phys.34 2336
[129] Schulz-von der Gathen V, R¨
opcke J, Gans T, K¨
aning M,
Lukas C and D¨
obele H F 2001 Plasma Sources Sci.
Technol. 10 530
[130] Hempel F, Glitsch S, R¨
opcke J, Saß S and Zimmermann H
2005 On the measurement of absolute plasma species
densities using quantum cascade laser-absorption
spectroscopy Plasma Polymers and Related Materials ed
M Mutlu (Ankara: Hacettepe University Press) p 142
[131] R¨
opcke J, Glitsch S, Hempel F, Sass S, Schulz K-D,
Weltmann K-D and Zimmermann H 2005 Proc. 48th
Annual SVC Conf. (Denver, USA) p54
[132] McManus J B, Kebabian P L and Zahniser M S 1995 Appl.
Opt. 34 3336
[133] Lancaster D G, Richter D, and Tittel F K 1999 Appl. Phys. B
69 459
[134] Rehle D, Leleux D, Erdely M, Tittel F K, Fraser M and
Friedfeld S 2001 Appl. Phys. B72 947
S168