plasmapy.analysis.swept_langmuir.floating_potential.find_floating_potential(voltage: numpy.ndarray, current: numpy.ndarray, threshold: int = 1, min_points: Optional[Union[int, float]] = None, fit_type: str = 'exponential')

Determines the floating potential (\(V_f\)) for a given current-voltage (IV) curve obtained from a swept Langmuir probe. The floating potential is the probe bias where the collected current equals zero \(I = 0\). (For additional details see the Notes section below.)

Aliases: find_vf_

  • voltage (numpy.ndarray) – 1-D numpy array of monotonically ascending/descending probe biases (should be in volts)

  • current (numpy.ndarray) – 1-D numpy array of probe current (should be in amperes) corresponding to the voltage array

  • threshold (positive, non-zero int) – Max allowed index distance between crossing-points before a new crossing-island is formed. That is, if threshold=5 then consecutive crossing-points are considered to be in the same crossing-island if they are within 5 index steps of each other. (Default: 1)

  • min_points (positive int or float) –

    Minimum number of data points required for the fitting to be applied to. See Notes section below for additional details. The following list specifies the optional values:

    • min_points = None (Default) The largest of 5 and factor * array_size is taken, where array_size is the size of voltage and factor = 0.1 for fit_type = "linear" and 0.2 for "exponential".

    • min_points = numpy.inf The entire passed array is fitted.

    • min_points >= 1 Exact minimum number of points.

    • 0 < min_points < 0 The minimum number of points is taken as min_points * array_size.

  • fit_type (str) –

    The type of curve to be fitted to the Langmuir trace, "linear" or "exponential" (Default). This selects which FitFunction class should be applied to the trace.






  • vf (float or numpy.nan) – The calculated floating potential (same units as the voltage array). Returns numpy.nan if the floating potential can not be determined. How \(V_f\) is calculated depends on the fit function. This is described in the root_solve() method of the relevant fit function (e.g. the root_solve() method of ExponentialPlusOffset).

  • vf_err (float or numpy.nan) – The uncertainty associated with the floating potential calculation (units same as vf). Returns numpy.nan if the floating potential can not be determined. Like \(V_f\):, the calculation depends on the applied fit function. The root_solve() method also describes how this is calculated.

  • rsq (float) – The coefficient of determination (r-squared) value of the fit. See the documentation of the rsq property on the associated fit function (e.g. the rsq property of ExponentialPlusOffset).

  • func (sub-class of AbstractFitFunction) – The callable function \(f(x)\) representing the fit and its results.

  • islands (List[slice]) – List of slice objects representing the indices of the identified crossing-islands.

  • indices (slice) – A slice object representing the indices of voltage and current arrays used for the fit.


The internal functionality works like:

  1. The current array current is scanned for all points equal to zero and point pairs that straddle \(I = 0\). This forms an array of “crossing-points.”

  2. The crossing-points are then grouped into “crossing-islands” in based on the threshold keyword.

    • A new island is formed when a successive crossing-point is more (index) steps away from the previous crossing-point than allowed by threshold.

    • If multiple crossing-islands are identified, then the span from the first point in the first island to the last point in the last island is compared to min_points. If the span is less than or equal to min_points, then that span is taken as one larger crossing-island for the fit; otherwise, the function is incapable of identifying \(V_f\) and will return numpy.nan values.

  3. To calculate the floating potential…

    • If the crossing-island contains fewer points than min_points, then each side of the crossing-island is equally padded with the nearest neighbor points until min_points is satisfied.

    • A fit is then performed using scipy.stats.linregress for fit_type="linear" and scipy.optimize.curve_fit for fit_type="exponential".