WFC3/UVIS EPER CTE 2009-2025

TL;DR

The study analyzes the time evolution of charge transfer efficiency (CTE) losses in the HST WFC3/UVIS detector using the Extended Pixel Edge Response (EPER) method over 2009–2025. It models the CTE/CTI relationship with a power-law form $CTE = 1 - m\times S^{\rho}$, finding a stable slope $\rho = -0.72 \pm 0.05$ and an increasing intercept, and compares linear, quadratic, and cubic time fits to the CTE trend. Results show continued CTE decline, most pronounced at the lowest illumination level (160 e−), with a total drop of about $0.0015$ and a rate of $0.0001$ per year; quadratic and cubic fits provide better residuals and goodness-of-fit than a linear model. Lomb-Scargle analysis reveals residual periodicities of approximately $8.1 \pm 0.4$ years (linear) and ~ $9.1 \pm 0.3$ / $9.0 \pm 0.3$ years (quadratic/cubic), suggesting a possible but not conclusive link to solar activity. These findings support ongoing EPER-based monitoring to maintain robust calibration of WFC3/UVIS over long timescales.

Abstract

In this report, we examine the behavior of Charge Transfer Efficiency (CTE) on the WFC3/UVIS detector over time as computed by the Extended Pixel Edge Response (EPER) technique, using internal calibration data acquired from 2009 through 2025. We find that the CTE has continued to decline as expected, with a steeper loss rate for lower signal levels. The lowest signal level (160e-) has continued to decline at a rate of 0.0001 per year, with a total overall decline of 0.0015. Analyses from 2016 and 2020 found that the rate of decline was not well fit by a linear function. This report verifies the rate of decline is instead better fit by a quadratic function (which results in the smallest min. and max. residuals, on average) or a cubic function (which has the best "goodness of fit" $χ^2$ and $R^2$ values). We continue to see periodic oscillations of the residuals around all three fit lines (linear, quadratic, and cubic) on which we perform a Lomb-Scargle periodogram analysis of the residuals. We find a periodicity of about 8 years for the residuals around the linear fit lines and about 9 years for the quadratic and cubic fit lines.

Figures (6)

• Figure 1: Average CTE measurement using the EPER method as a function of time for 5 different illumination levels: 160e-, 400e-, 800e-, 1650e-, and 5000e-. A linear (top), quadratic (middle), and cubic (bottom) function are fitted to the data. The WFC3/UVIS CTE as measured by the EPER method continues to decline with time and declines more steeply at fainter illumination levels.
• Figure 2: Residuals between the data and the linear (top), quadratic (middle), and cubic (bottom) fit functions for all illumination levels.
• Figure 3: (Top) CTI versus the signal level in electrons with a linear fit (overplotted in red) to CTI values for images within each observation. Four different colors represent each of the four amps of the detector. (Middle) $\rho$ (from the power law relationship of CTI to signal level) versus time on a log-log scale. (Bottom) The intercept of the CTI ($log(m)$) as a function of time on a log-log scale.
• Figure 4: Lomb-Scargle periodogram of CTE-loss residuals around a linear fit line for each of the illumination levels. The characteristic best period is marked with a vertical black line.
• Figure 5: Lomb-Scargle periodogram of CTE-loss residuals around a quadratic fit line for each of the illumination levels. The characteristic best period is marked with a vertical black line.