The gas temperature in non-equilibrium plasmas is often obtained from the plasma-induced emission by measuring the rotational temperature of a diatomic molecule in its excited state. This is motivated by both tradition and the availability of low budget spectrometers. However, non-thermal plasmas do not automatically guarantee that the rotational distribution in the monitored vibrational level of the diatomic molecule is in equilibrium with the translational (gas) temperature. Often non-Boltzmann rotational molecular spectra are found in non-equilibrium plasmas. The deduction of a gas temperature from these non-thermal distributions must be done with care as clearly the equilibrium between translational and rotational degrees of freedom cannot be achieved. In this contribution different methods and approaches to determine the gas temperature are evaluated and discussed. A detailed analysis of the gas temperature determination from rotational spectra is performed. The physical and chemical background of non-equilibrium rotational population distributions in molecular spectra is discussed and a large range of conditions for which non-equilibrium occurs are identified. Fitting procedures which are used to fit (non-equilibrium) rotational distributions are analyzed in detail. Lastly, recommendations concerning the conditions for which the gas temperatures can be obtained from diatomic spectra are formulated.

Dc-excited discharges generated in water at the tip of a tungsten wire which is located at the orifice of a quartz capillary are investigated by time-averaged optical emission spectroscopy.Two distinctive discharge modes are observed. For small conductivities of the liquid the discharge is a streamer-like discharge in the liquid itself (liquid mode). For conductivities above typically 45 µS cm −1 a large vapour bubble is formed and a streamer discharge in this vapour bubble is observed (bubble mode).Plasma temperatures and electron densities are investigated for both modes. The gas temperature is estimated from the rotational temperature of N 2 (C-B) and is 1600 ± 200 K for the bubble mode and 1900 ± 200 K for the liquid mode. The rotational temperature of OH(A-X) is up to 2 times larger and cannot be used as an estimate for the gas temperature. The rotational population distribution of OH(A), ν = 0 is also non-Boltzmann with a large overpopulation of high rotational states. This discrepancy in rotational temperatures is discussed in detail.Electron densities are obtained from the Stark broadening of the hydrogen Balmer beta line. The electron densities in the liquid mode are of the order of 10 21 m −3 . In the bubble mode electron densities are significantly smaller: (3-4) × 10 20 m −3 . These values are compared with the Stark broadening of the hydrogen alpha and gamma lines and with electron densities obtained from current density measurements. The chemical reactivities of the bubble and liquid modes are compared by means of the hydrogen peroxide production rate.

In this contribution, optical emission spectroscopy is evaluated and thoroughly analysed as a diagnostic to characterize plasmas in and in contact with liquids. One of the specific properties of plasmas in and in contact with liquids is the strong emission of OH(A–X) and of hydrogen lines. As an example a 600 ns pulsed dc excited discharge in Ar, He and O2 bubbles in water is investigated by time resolved optical emission spectroscopy. It is shown that the production processes of excited species and the plasma kinetics strongly influence the emission spectrum. This complicates the interpretation of the spectra but provides the opportunity to derive production mechanisms from the time resolved emission. The importance of recombination processes compared with direct electron excitation processes in the production of excited states of the water fragments in plasmas with high electron densities is shown. The OH(A–X) emission spectrum illustrates that even in these highly collisional atmospheric pressure discharges the rotational population distribution deviates from equilibrium. A two-temperature fit of the OH rotational population distribution leads to realistic gas temperatures for the temperature parameter corresponding to small rotational numbers. The Hα and Hβ lines are fitted with two component profiles corresponding to two different electron densities. The obtained electron density is in the range 1021–1023 m−3. Possible complications in the interpretation of obtained temperatures and electron densities are discussed.

Effect of Ti-Al cathode composition on plasma generation and plasma transport in direct current vacuum arc Estimation of electron temperature and density of the decay plasma in a laser-assisted discharge plasma extreme ultraviolet source by using a modified Stark broadening method A magnetized hydrogen plasma beam was generated with a cascaded arc, expanding in a vacuum vessel at an axial magnetic field of up to 1.6 T. Its characteristics were measured at a distance of 4 cm from the nozzle: up to a 2 cm beam diameter, 7.5ϫ 10 20 m −3 electron density, ϳ2 eV electron and ion temperatures, and 3.5 km/ s axial plasma velocity. This gives a 2.6ϫ 10 24 H + m −2 s −1 peak ion flux density, which is unprecedented in linear plasma generators. The high efficiency of the source is obtained by the combined action of the magnetic field and an optimized nozzle geometry. This is interpreted as a cross-field return current that leads to power dissipation in the beam just outside the source.

In this paper the rotational temperature of OH(A–X) and rotational population distribution of OH(A) are investigated for streamer discharges in bubbles and glow discharges with liquid electrodes, both at atmospheric pressure. The influence of the filling gas is investigated in detail and the non‐Boltzmann nature of the rotational population distributions is discussed. It is shown that the rotational population distribution of OH(A) is even at atmospheric pressure an image of the formation process or is at least influenced by it. As a consequence the rotational temperature is in this case not a good estimate of the gas temperature as the rotational population distribution is not an image of a kinetic temperature. In some cases rotational states with small rotational numbers offer a possibility to obtain the gas temperature. The influence of these results on the determination of gas temperatures in the field of liquid plasmas is discussed.

In this paper it is shown that electronic quenching of OH(A) by water prevents thermalization of the rotational population distribution of OH(A). This means that the observed ro-vibrational OH(A-X) emission band is (at least partially) an image of the formation process and is determined not only by the gas temperature. The formation of negative ions and clusters for larger water concentrations can contribute to the non-equilibrium. The above is demonstrated in RF excited atmospheric pressure glow discharges in He-water mixtures in a parallel metal plate reactor by optical emission spectroscopy. For this particular case a significant overpopulation of high rotational states appears around 1000 ppm H 2 O in He. The smallest temperature parameter of a non-Boltzmann (two-temperature) distribution fitted to the experimental spectrum of OH(A-X) gives a good representation of the gas temperature. Only the rotational states with the smallest rotational numbers (J 7) are thermalized and representative for the gas temperature.

The operation of a cascaded arc hydrogen plasma source was experimentally investigated to provide an empirical basis for the scaling of this source to higher plasma fluxes and efficiencies. The flux and efficiency were determined as a function of the input power, discharge channel diameter, and hydrogen gas flow rate. Measurements of the pressure in the arc channel show that the flow is well described by Poiseuille flow and that the effective heavy particle temperature is approximately 0.8 eV. Interpretation of the measured I -V data in terms of a one-parameter model shows that the plasma production is proportional to the input power, to the square root of the hydrogen flow rate, and is independent of the channel diameter. The observed scaling shows that the dominant power loss mechanism inside the arc channel is one that scales with the effective volume of the plasma in the discharge channel. Measurements on the plasma output with Thomson scattering confirm the linear dependence of the plasma production on the input power. Extrapolation of these results shows that ͑without a magnetic field͒ an improvement in the plasma production by a factor of 10 over where it was in van Rooij et al. ͓Appl. Phys. Lett. 90, 121501 ͑2007͔͒ should be possible.

A highly sensitive imaging Thomson scattering system was developed for low temperature (0.1-10 eV) plasma applications at the Pilot-PSI linear plasma generator. The essential parts of the diagnostic are a neodymium doped yttrium aluminum garnet laser operating at the second harmonic (532 nm), a laser beam line with a unique stray light suppression system and a detection branch consisting of a Littrow spectrometer equipped with an efficient detector based on a "Generation III" image intensifier combined with an intensified charged coupled device camera. The system is capable of measuring electron density and temperature profiles of a plasma column of 30 mm in diameter with a spatial resolution of 0.6 mm and an observational error of 3% in the electron density (n(e)) and 6% in the electron temperature (T(e)) at n(e) = 4 x 10(19) m(-3). This is achievable at an accumulated laser input energy of 11 J (from 30 laser pulses at 10 Hz repetition frequency). The stray light contribution is below 9 x 10(17) m(-3) in electron density equivalents by the application of a unique stray light suppression system. The amount of laser energy that is required for a n(e) and T(e) measurement is 7 x 10(20)n(e) J, which means that single shot measurements are possible for n(e)>2 x 10(21) m(-3).