Using quantum noise correlation analysis to measure ion temperature Daniel J. Den Hartog Department of Physics University of Wisconsin–Madison DOE/SC Program Office: Fusion Energy Sciences DOE Award Number: DE‐SC0021640 Final Technical Report November 2025 I. Background/Introduction In an insightful and pioneering experiment, Hanbury Brown and Twiss (HBT) demonstrated that quantum statistical noise correlations could be used to measure the size of the star Sirius [Hanbury Brown 1956]. This correlation effect relies on quantum interference between possible detection paths for two indistinguishable photons. This technique, often called intensity interferometry, exploits the interference of photons independently emitted from different regions of a non-coherent light source. The basis is the measurement of intensity correlations as a function of time and detector spacing. No beam splitting or preservation of phase information is required. The discovery of the HBT effect was a landmark in the development of the field of quantum optics [Glauber 2006]. The original HBT experiment exploited two-photon correlations manifested in a time-domain signal to measure the size of a distant photon source (a star). We attempted to test the same principle of quantum statistical noise correlation to measure the ion temperature in a laboratory plasma. If eventually successful, the next step will be to build a compact, inexpensive, and robust ion temperature diagnostic suitable for a burning plasma environment. This diagnostic would require only a small optical view of the plasma through a port to record impurity line emission. It would be straightforward to locate the instrumentation remote from the reactor behind the neutron and bio-shielding. This technique relies solely on quantum correlations to measure ion temperature—there is no grating dispersion of a wavelength spectrum or pulse-height analysis of photon energy. There is no need to inject a laser or particle beam into the plasma to stimulate emission. The quantum noise correlation instrumentation tested consisted of two single-photon detectors capable of picosecond resolution, a time-tagging unit, and simple light collection optics with a filter monochromator to pass only photons from a bright spectral line in the visible wavelength range. This project was an application of quantum sensing to enhance diagnostic capability for plasma and fusion science. Quantum sensing is “use of a quantum system, quantum properties, or quantum phenomena to perform a measurement of a physical quantity” [Degen 2017]. This project applied a variation of the HBT technique to a plasma, and is an application of a quantum sensing approach to improve plasma measurement. A significant challenge for a diagnostic based on the HBT effect is maximization of the second-order correlation function. This requires careful characterization of instrumental effects, collection of data ensembles, and development of appropriate analysis algorithms. II. Research Methods Under typical fusion plasma conditions, electron impact excitation is the dominant process through which ions radiate line emission. Doppler broadening typically dominates other line broadening mechanisms, and is determined by the ion temperature. Specifically, for a spatially uniform quasi- monochromatic thermal light source (a plasma), the second-order correlation function associated with the HBT effect takes the form [Boiko 2009]: & 𝑤𝑥$% 𝜏" 𝑔(") "𝑥$% , 𝜏& = 1 + "sinc " .𝜋 3 exp 7−𝜋 " 9, (1) 𝜆𝐿 𝜏' where the second term in the right-hand side takes into account the decorrelation effects due to unpolarized light (the coefficient ½), the zero-delay degree of spatial coherence, and the Gaussian profile of the delayed first-order correlation function for an inhomogeneously-broadened line. To measure ion temperature, we concentrate on measurement of the delayed first-order correlation function. The coherence time 𝜏' = 2√2𝜋ln2⁄∆𝜔 is determined by the inhomogeneous broadening Δω (FWHM) of the 1 emission line. For most magnetically confined laboratory plasmas, thermal Doppler broadening dominates the inhomogeneous broadening of typical emission lines, usually resulting in a coherence time tc in the range of 1 to 10 ps. The spatial coherence term in Eq. 1 contains an argument similar to the dimensionless number in physical optics often called the Fresnel number (FN). The value of FN is related to the number of uncorrelated photon modes contributing to the intensity interference at the detector array. In our setup, the beamsplitter makes xij (the spacing between the two detectors) effectively zero, and thus the value of the sinc function in Eq. 1 is approximately 1. Historically, the temporal correlation function in Eq. 1 presented a major challenge to precise measurement of g(2), particularly in the original stellar measurements made by Hanbury Brown and Twiss. The HBT effect is apparent only if a pair of photons is emitted within a coherence time. The response bandwidth of currently available detectors has improved substantially since the original HBT measurement. For example, the hybrid photodetectors in our setup have an instrument response function (IRF) of approximately 20 ps. For two detectors in a correlation setup, the combined response is √2 𝐼𝑅𝐹 = 28 ps. With an emission line coherence time of 5 ps, we expected the peak of the measured g(2) to be 1 + 5/28, or 18% above the uncorrelated background level. The number of photon counting events needed for an acceptable uncertainty of a measured coherence time (and thus the ion temperature in the plasma) is unacceptably large with two single-photon detectors. The instrumentation assembled to measure the temporal correlation function of an emission line from a plasma is illustrated in Fig. 1. The upper panel is a schematic illustration of the optical layout, the lower panel is a photograph of the actual setup. Space constraints required the entire setup to be contained on 3 × 6 foot optical table. Light was transferred from a small collection telescope via a 600 µm core solarization resistant fiber optic. The f = 100 mm compact monochromator selects the emission line to be measured, with a range of 200 – 650 nm. The 600 µm fiber core is the entrance aperture of the monochromator, and the 200 µm pinhole is the exit aperture, which provides a passband of several nm for a Doppler broadened line. The distance between the 200 and 2000 µm pinholes is 2.7 m, which keeps the spatial coherence term close to 1. The Rochon polarizer passes linearly polarized light of one orientation (while diverting the orthogonal orientation) in order to eliminate decorrelation effects due to unpolarized light. Although not shown in the figure, the light path contains multiple 6.3 mm diameter apertures at various locations in order to eliminate stray light propagation through the system. The photon pulses produced by the hybrid photodetectors are recorded by time-correlated single photon counting modules with time bin width down to 813 fs. These modules are mounted in a PC sitting on a shelf mounted on the optical table frame. Thus control and data acquisition moves with the instrumentation, and the entire setup was portable. If ion temperature measurement is ultimately successful with this technique, the next step will be to design a compact, inexpensive, and robust ion temperature diagnostic suitable for a burning plasma environment. This diagnostic would require only a small optical view of the plasma through a port plug, and would be straightforward to locate remote from the device behind the neutron and bio-shielding (Fig. 2). Suitable visible emission lines for the ion species of interest would need to be identified (even highly charged tungsten produces visible emission, e.g. [Shinohara 2015]). The two detectors shown in the figure would be arrays of SPADs in order to multiplex photon counting and reduce the time needed to record an ion temperature with sufficient accuracy. The filter, beamsplitter, and detectors could be placed in a purpose-built mount, and the electronics integrated into the package to produce a compact diagnostic system occupying a fraction of a cubic meter. Testing could be done on a large-scale plasma experiment since the port space requirement would be small. The technique may also be useful for measurements on plasmas that are spatially small, such as in a Z-pinch device. Furthermore, while SPAD detectors are most convenient for this measurement, there are hybrid and other detector technologies to extend sensitivity into the ultraviolet and even x-ray. 2 Fig. 1. Upper: Schematic illustration of the instrumentation to measure the temporal correlation function of an emission line from a plasma. Lower: Photograph of the instrumentation as installed on a 3 × 6 foot optical table. The photograph shows the stages and positioning equipment needed to properly align the light path. The entire system was enclosed to prevent dust contamination and stray light ingress. 3 Fig. 2. Schematic diagram of an ion temperature diagnostic based on quantum noise correlation analysis, installed on a burning plasma device. The diameter w of the plasma viewing aperture and the distance L from the aperture to the detectors are flexible as long as w/L < 10-2; e.g., for w = 10 mm, L could be tens of meters. Since light phase information does not need to be preserved, the optical path and shielding labyrinth can be laid out according to physical plant requirements. An optical interference filter with a narrow passband isolates the emission line of interest. The beamsplitter ensures that both detectors are viewing the same plasma volume. This diagram is not drawn to scale in order to enhance visibility. With this diagnostic technique, even a single view of the plasma could provide information on the radial profile of the ion temperature. In any given plasma there are multiple ion species, and the emissivity of each ion species peaks at a certain electron temperature. Using a separately measured electron temperature profile (e.g., from Thomson scattering), the approximate location from which the line radiation of each ion species was emitted can be determined. Correlating the measured ion temperature of each species with the plasma minor radius enables production of an approximate ion temperature profile. The main remaining question is how long the signal will have to be collected to make an ion temperature measurement. There is a high probability the HBT quantum optical effect will behave as expected. With a two-detector setup, it will be too long (>>1 sec) for a practical measurement on a fusion plasma. This could be addressed with an array of detectors to multiplex measurement of photon correlation. Megapixel SPAD arrays are available [Morimoto 2020], but readouts necessary for time- correlated single photon counting of such large arrays are not yet available. References Cited D. L. Boiko, N. J. Gunther, N. Brauer, M. Sergio, C. Niclass, G. B. Beretta, and E. Charbon, “On the application of a monolithic array for detecting intensity-correlated photons emitted by different source types,” Opt. Express 17, 15087 (2009). C. L. Degen, F. Reinhard, and P. Cappellaro, “Quantum sensing,” Rev. Mod. Phys. 89, 035002 (2017). R. J. Glauber, “Nobel Lecture: One hundred years of light quanta,” Rev. Mod. Phys. 78, 1267 (2006). R. Hanbury Brown and R. Q. Twiss, “A test of a new type of stellar interferometer on Sirius,” Nature 178, 1046 (1956). Kazuhiro Morimoto, Andrei Ardelean, Ming-Lo Wu, Arin Can Ulku, Ivan Michel Antolovic, Claudio Bruschini, and Edoardo Charbon, “Megapixel time-gated SPAD image sensor for 2D and 3D imaging applications,” Optica 7, 346 (2020). M. Shinohara, K. Fujii, D. Kato, N. Nakamura, M. Goto, S. Morita, M. Hasuo, and LHD Experiment Group, “Visible emission spectroscopy of highly charged tungsten ions in LHD: I. Survey of new visible emission lines,” Physica Scripta 90, 125402 (2015). 4