During this first century of plasma physics research, dating from Langmuir’s discoveries in the 1920s of the medium like behavior of a new state of matter, plasma, the Langmuir probe has proved to have been the single most important diagnostic of plasma parameters. This is true in part, because of its extraordinary range of applicability1. In plasma encountered by satellites2,3,4, in semiconductor processing experiments,5,6,7,8 at the edges of plasma confined in tokamaks,9,10,11 and in wide range of basic plasma physics experiments, Langmuir probes have been used to measure plasma densities and temperatures spanning the ranges 108≤ne≤1019m-3, and 10-3≤Te≤102eV , respectively. Simultaneously in the 1920s, he invented the probe now named after him and the emissive probe12. The emissive probe is now primarily used as a diagnostic of plasma potential. Although it cannot measure the breadth of plasma parameters that the Langmuir probe can, it too is a diagnostic of wide utility when it comes to the measurement of plasma potential, or, as it is sometimes called, the electrostatic space potential. For example, the emissive probe can accurately measure space potentials even in a vacuum, where Langmuir probes are incapable of measuring anything.
The basic setup of the Langmuir probe consists of putting an electrode into the plasma and measuring the collected current. The resulting current-voltage (I-V) characteristics can be used to interpret plasma parameters such as electron temperature Te, electron density ne, and plasma potential ϕ13. For a Maxwellian plasma, the relationship between collected electron current Ie (taken to be positive) and probe bias VB can be expressed as14:

where Ie0 is the electron saturation current,

and where S is the collecting area of the probe,
is the bulk electron density, e is the electron charge, Te is the electron temperature, me is the electron mass. The theoretical relation of I-V characteristics for the electron current is illustrated in two ways in Figure 1A and Figure 1B. Note, Eq. (1a,b) only applies to bulk electrons. However, Langmuir probe currents can detect flows of charged particles, and adjustments must be made in the presence of primary electrons, electron beams, or ion beams etc. See Hershkowitz14 for more details.
The discussion here takes up the ideal case of Maxwellian electron energy distribution functions (EEDF). Of course, there are many circumstances in which non-idealities arise, but these are not the subject of this work. For example, in materials processing etching and deposition plasma systems, typically RF generated and sustained, there are molecular gas feed stocks that produce volatile chemical radicals in the plasma, and multiple ion species including negatively charged ions. The plasma becomes electronegative, that is, having a significant fraction of the negative charge in the quasineutral plasma in the form of negative ions. In plasma with molecular neutrals and ions, inelastic collisions between electrons and the molecular species can produce dips15 in the current-voltage characteristics, and the presence of cold negative ions, cold relative to the electrons, can produce significant distortions16 in the vicinity of the plasma potential, all of which of course are non-Maxwellian features. We prosecuted the experiments in the work discussed in this paper in a single ion species noble gas (argon) DC discharge plasma, free of these kinds of non-Maxwellian effects. However, a bi-Maxwellian EEDF is typically found in these discharges, caused by the presence of secondary electron emission17 from the chamber walls. This component of hotter electrons is typically a few multiples of the cold electron temperature, and less than 1% of the density, typically easily distinguished from the bulk electron density and temperature.
As VB becomes more negative than ϕ, electrons are partially repelled by the negative potential of the probe surface, and the slope of the ln(Ie) vs. VB is e/Te, ie. 1/TeV where TeV is the electron temperature in eV, as shown in Figure 1B. After TeV is determined, the plasma density can be derived as:

Ion current is derived differently than electron current. Ions are assumed to be “cold” due to their relatively large mass, Mi >> me, compared to that of the electron, thus, in a weakly ionized plasma, the ions are in fairly good thermal equilibrium with the neutral gas atoms, which are at the wall temperature. Ions are repelled by the probe sheath if VB ≥ ϕ and collected if VB < ϕ. The collected ion current is approximately constant for negatively biased probes, while the electron flux to the probe decreases for probe bias voltages more negative than the plasma potential. Since the electron saturation current is much larger than the ion saturation current, the total current collected by the probe decreases. As the probe bias becomes increasingly negative, the drop in current collected is great or small as the electron temperature is cold or hot, as described above in Eq. (1a). The equation for ion current in this approximation is:

where

and

We note that constant ion flux collected by the probe exceeds the random thermal ion flux due to acceleration along the presheath of the probe and thus ions reach the sheath edge of the probe at the Bohm speed18, uB, rather than the ion thermal speed19. And the ions have a density equal to the electrons since the presheath is quasineutral. Comparing the ion and electron saturation current in Eqn.5 and 2, we observe that the ion contribution to the probe current is smaller than that of electrons by a factor of
. This factor is about 108 in the case of argon plasma.
There is a sharp transition point where the electron current goes from exponential to a constant, known as the ”knee”. The probe bias at the knee can be approximated as the plasma potential. In the real experiment, this knee is never sharp, but rounded due to the space-charge effect of the probe, that is, the expansion of the sheath surrounding the probe, and also to probe contamination, and plasma noise13.
The Langmuir probe technique is based on collection current, whereas the emissive probe technique is based on the emission of current. Emissive probes measure neither temperature nor density. Instead they provide precise plasma potential measurements and can operate under a variety of situations due to the fact that they are insensitive to plasma flows. The theories and usage of emissive probes are fully discussed in the topical review by Sheehan and Hershkowitz20, and references therein.
For plasma density 1011 ≤ ne ≤ 1018 m-3, the inflection point technique in the limit of zero emission is recommended, which means to take a of series of I-V traces, each with different filament heating currents, finding the inflection point bias voltage for each I-V trace, and extrapolate the inflection points to the limit of zero emission to get the plasma potential, as shown in Figure 2.
It is a common assumption that Langmuir and emissive probe techniques agree in quasineutral plasma, but disagree in the sheath, the region of the plasma in contact with the boundary in which space-charge appears. The study focuses on the plasma potential near plasma boundaries, in low temperature, low pressure plasma in an effort to test this common assumption. To compare potential measurements by both Langmuir probe and emissive probe, plasma potential is also determined by applying inflection point technique to Langmuir probe I-V, as shown in Figure 3. It is generally accepted1 that the plasma potential is found by finding the probe bias voltage at which the second derivative of the current collected differentiated with respect to the bias voltage,
, that is, the peak of the dI/dV curve, with respect to the probe bias voltage. Figure 3 demonstrates how this maximum in dI/dV, the inflection point of the current-voltage characteristic, is found.
Langmuir probes (collecting) and emissive probes (emitting) have different I-V characteristics, which also depend on the geometry of the probe tip, as shown in Figure 4. The space-charge effect of the probe must be considered before the probe fabrication. In the experiments, for the planar Langmuir probes, we used a ¼" planar Tantalum disk. We could collect more current and get bigger signals with a larger disk. However, in order for the analyses above to apply, the area of the probe, Ap must be kept smaller than the electron loss area of the chamber, Aw, satisfying21 the inequality
. For the cylindrical Langmuir probe, we used a 0.025 mm thick, 1 cm long Tungsten wire for the cylindrical Langmuir probe and a same thickness for the Tungsten wire for the emissive probe. It is important to note that for cylindrical Langmuir probes, for the plasma parameters of these experiments, the radius of the probe tip, rp, is much smaller than its length, Lp, and smaller than the Debye length, λD; that is,
, and
. In this range of parameters, applying Orbital Motion Limited theory and Laframboise’s development of it22 for the case of thermal electrons and ions, we find that for probe bias voltages equal to or greater than the plasma potential, the electron current collected may be parameterized by a function of the form
, where the exponent
. The important point here is that for values of this exponent less than unity, the inflection point method for determining the plasma potential, as described in the paragraph above, applies to cylindrical Langmuir probes too.