1.1.

Instrument Overview

The WRXR’s large FoV was enabled by a passive focuser that consisted of a series of electroformed nickel plates, each with 185 slits.¹³ The overall FoV allowed by the focuser geometry was $3.32\mathrm{deg}\times 17.6\mathrm{deg}$. However, the angular extent of the Vela SNR, grating diffraction effects, and detector extent at the focal plane limited the total instrument FoV to $\sim 13{\mathrm{deg}}^2$. The widths of the slits and of the supporting frame that separates the slits converged over the 0.9-m length of the focuser system to produce a single line spread function (LSF) at the focal plane. The focuser LSF, therefore, consisted of an overlapping LSF from each of the 185 slits with a $\sim 2\mathrm{mm}$ full width at half maximum (FWHM).

Immediately following the focuser in the spectrometer was an array of x-ray reflection gratings. The grating array covered the full exit aperture of the focuser and thereby intercepted and redirected focused x-rays to a diffraction arc at the focal plane. An x-ray HCD, developed as part of a collaboration between Penn State University (PSU) and Teledyne Imaging Sensors,^14,15 was used to image the diffracted spectral lines. The WRXR HCD was $36.9\mathrm{mm}\times 36.9\mathrm{mm}$, which was sufficient to capture the four target emission lines (CVI, OVIII, and two orders of OVII) but covered only $\sim 20\%$ of the lines’ extent in the cross-dispersion direction. To improve instrument FoV and the effective area, an array of Ni-coated mirrors was installed after the grating array to provide a secondary reflection for diffracted photons. The mirror array was designed to reflect a portion of the photons that would fall outside the detector coverage onto the 36.9-mm active area, increasing instrument throughput by $\sim 20\%$. A computer-aided design (CAD) model of the WRXR science payload is shown in Fig. 1 with the primary spectrometer components labeled.

Fig. 1

A CAD model of the complete WRXR science payload. The components of the spectrometer are labeled.

1.2.

Observation Target

The WRXR spectrometer’s observation target was the diffuse soft x-ray emission from the northern part of the Vela SNR. The northern region was selected to observe a relatively unexplored area of Vela while simultaneously avoiding contamination from the Vela pulsar near the center of the remnant and the two other SNRs near Vela: Puppis A and Vela Jr.

The Vela SNR is a shell-type remnant at a distance of $\sim 250\mathrm{pc}$¹⁶ and with an apparent diameter of $\sim 8\mathrm{deg}$. The morphology of Vela suggests that the local interstellar medium has been swept up by the precursor wind and supernova blast wave into a nearly spherical zone of interaction. Shock-heated plasmas in this zone reach collisional ionization equilibrium on a timescale that is inversely proportional to density; the denser regions, therefore, equilibrate more quickly and cool more efficiently than rarified regions, which remain hot and have emission at higher energies. This scenario is also seen in other type II shell remnants^17,18 and a spectrum of the broad emission in the northern part of the remnant will provide a basis for comparison among various shell-type remnants.

Vela’s large apparent size, high surface brightness, and low intervening hydrogen column density have made the remnant a common target for both sounding rockets^19–21 and space telescopes.^22–26 However, previous studies have yielded some inconsistencies about the prevailing conditions across the remnant; most authors conclude that ionization equilibrium conditions exist (though with different proposed temperature models), but some observations suggest that a second temperature component or nonequilibrium conditions exist to an undetermined extent. Additionally, previous observations suffer from either prohibitively low spectral resolving power (sounding rockets and ROSAT) or a small FoV. The WRXR sought to combine a large FoV and moderate resolving power to perform an observation on the northern part of the remnant.

5.1.

Component-Level Measurements

Prior to instrument buildup, calibration and optimization data were obtained for both the reflection gratings and the HCD; the mechanical focuser had flown on two previous instruments, and data from those instruments were used in place of new calibrations. Two gratings were randomly selected from the population used for the final flight array and tested for diffraction efficiency at Lawrence Berkeley National Laboratory’s Advanced Light Source (ALS). The ALS experimental setup, data collection, and analysis followed the procedures discussed in Ref. 29. Figure 6 shows the absolute total and single-order diffraction efficiency from one of the gratings tested at the ALS.

Fig. 6

The absolute diffraction efficiency of a WRXR grating measured before the flight. The black dashed line shows the reflectivity of nickel and sets the upper limit on possible diffraction efficiency. The blue solid line is the total measured diffraction efficiency (not including zero-order reflection). Each other line represents single-order diffraction efficiency.

The grating test geometries at the ALS differed slightly from the nominal WRXR geometry, which resulted in a small but noticable shift in single-order efficiency as a function of energy; this is evident in Fig. 6, where the intersection between third and fourth orders occurs at $\sim 590\mathrm{eV}$ rather than $\sim 570\mathrm{eV}$ as designed for OVII emission. Despite the slight shift in efficiency curves due to small deviations in test geometry, however, the total diffraction efficiency is representative of grating performance in similar orientations. Total absolute diffraction efficiency exceeds 70% at low energies and near the CVI emission energy ($\sim 375\mathrm{eV}$) and is above 50% at OVII and OVIII ($\sim 570$ and $\sim 650\mathrm{eV}$, repectively). Further, the efficiency curves demonstrate that the grating performance closely matches design; most of the efficiency at the desired energies is contained in individual orders: $\sim 60\%$ efficiency in second-order at CVI, $\sim 50\%$ efficiency split between third and fourth orders at OVII, and $\sim 35\%$ efficiency in fourth order at OVIII. The total diffraction efficiency of both gratings tested before the flight is shown in Fig. 9. Finally, a grating coverage experiment was performed, where the diffraction efficiency was measured at 570 eV (near the OVII emission energy) as a function of position on the grating surface. The total area covered in the experiment was limited by stage travel to $80\mathrm{mm}\times 80\mathrm{mm}$ (out of the $100\mathrm{mm}\times 110\mathrm{mm}$ gratings). The diffraction efficiency of third and fourth orders at this energy (the two orders that are captured by the H2RG in the flight instrument) varied by $\sim 6\%$ from position-to-position over the grating surface, and total diffraction efficiency (all orders other than 0th-order reflection) varied by $<5\%$.

Similarly, the H2RG camera was calibrated independently before being integrated into the full instrument. Calibration runs consisted of several hundred integrations for which the data were processed with a typical x-ray analysis pipeline that included event thresholding and grading.¹⁴ Extensive data were collected at various temperatures to compensate for temperature variations in flight; active cooling of the detector package ceased just prior to rocket liftoff, causing the temperature of both the detector and SIDECAR™ to drift during the course of flight. To simulate flight conditions, calibration data were obtained at a range of temperatures with the temperature held constant (conditions on the ground just prior to liftoff) and also in a series of “drift” runs in which active cooling was turned off and the detector and SIDECAR™ were allowed to warm naturally (conditions throughout the flight). Detector read noise and energy resolution as a function of temperature were obtained from the resulting data using a variety of calibration sources with unique energies in the x-ray bandpass: the ${}^{55}\mathrm{Fe}$ source provided $\mathrm{Mn}\text{-}\mathrm{K}\alpha$ x-rays at 5.9 keV, and a Po-210 alpha particle source fluoresced an Mg target mounted on Al to produce K-shell emission for O (0.53 keV), Mg (1.25 keV), and Al (1.49 keV). At the frame rate used in this instrument, the H2RG dark current became negligible below $\sim 170\mathrm{K}$; preflight calibration, therefore, focused on determining the SIDECAR™ temperature that allowed the most optimal read noise.

Read noise was calculated as the average RMS pixel value of a sequence of dark images using the half of the H2RG that was not illuminated by ${}^{55}\mathrm{Fe}$ x-rays. Results are summarized in Table 2 for several combinations of SIDECAR™ and H2RG temperatures. At SIDECAR™ temperatures in excess of 190 K, the noise introduced by the SIDECAR™ itself became high enough to completely saturate the OVII line (0.57 keV). To ensure noise levels sufficient to detect the target oxygen lines in flight, an optimal SIDECAR™ launch temperature was set at 173 K, corresponding to an H2RG temperature of 130 K.

Table 2

Read noise of the H2RG measured as a function of temperature in preflight component-level testing.

Energy resolution was determined using Gaussian fits to the calibration spectra and was measured using single-pixel events for spectral lines below 2 keV and single- and two-pixel events for the 5.9 keV line; event grading selections were based on the availability of events in the flight data. Table 3 summarizes the FWHM energy resolution at different energies with the SIDECAR™ and H2RG cooled to 173 and 130 K, respectively.

Table 3

Energy resolution of the H2RG during preflight component-level and instrument-level testing. All data were collected at a SIDECAR™ temperature of ∼173  K and H2RG temperature of ∼130  K.

5.2.

Instrument-Level Calibrations

Following component-level optimization, the focuser, gratings, mirrors, detector, and all support equipment were built up into the complete spectrometer. The WRXR spectrometer had a focal length of $\sim 2\mathrm{m}$ and was designed to fit inside a 22-in. diameter cylindrical tube, which also provided the external mechanical structure for integration into the suborbital rocket package. The instrument electronics section and transition section aft of the spectrometer optics tapered from 22 to 17 in. to allow the instrument to fit into a standard suborbital rocket fairing (refer to Fig. 1).

The aft end of the payload, the direction toward which the optics pointed, was fitted with a port to accommodate a bolt-on x-ray calibration source for ground testing. The calibration source used for the WRXR was an electron-impact source that produced x-rays by boiling electrons off of a tungsten filament and accelerating those electrons into an aluminum anode. With this method, K-shell emission was produced from C (0.28 keV), N (0.39 keV), and O (0.53 keV). Without the ability to produce the highly ionized states of oxygen and carbon that are observable in high-energy sources, such as the Vela SNR, K-shell emission of the associated elements was the most convenient way to calibrate the space-based instrument from the ground.

The x-ray calibration source directed a diverging beam of photons through the optics channel, producing a line that was then diffracted by the gratings, reflected by the mirrors (or allowed to pass freely through the mirror array depending on the photon positions), and imaged by the HCD. Placing diffracted orders of multiple energies on the detector during ground calibrations allowed for both positional measurements of optics-to-detector alignment and detector performance characterization. Figure 7 shows a 2-D histogram of x-ray counts from the three spectral lines produced with the electron-impact calibration source. Preflight, full-instrument H2RG energy resolution analysis yielded $\sim 109\mathrm{eV}$ at 0.28 keV, $\sim 90\mathrm{eV}$ at 0.39 keV, $\sim 88\mathrm{eV}$ at 0.52 keV, and $\sim 202\mathrm{eV}$ at 5.9 keV (Table 3). Read noise was measured each time independent instrument tests were performed, with the lowest noise measurement reaching $5.93\pm 0.02{\mathrm{e}}^{-}$ at the conclusion of preflight testing.

Fig. 7

A 2-D histogram of low-energy x-ray events imaged with the H2RG after being generated by an electron-impact source and passing through the full spectrometer. From top to bottom, the three spectral lines visible on the detector are second-order $\mathrm{C}\text{-}\mathrm{K}\alpha$, third-order $\mathrm{O}\text{-}\mathrm{K}\alpha$, and second-order $\mathrm{N}\text{-}\mathrm{K}\alpha$. The calibration lines do not cover the full cross-dispersion extent of the detector due to the nature of the calibration setup; the calibration source was mounted on the payload itself and was unable to fully illuminate the entrance aperture of the focuser.

6.1.

Flight Plan

The nominal flight plan for the WRXR was developed to maximize time on the science target while still allowing for in-flight instrument calibration. Toward that goal, several features were implemented into the detector chamber design and flight control plan to provide flexibility during flight. First, the addition of the aforementioned ${}^{55}\mathrm{Fe}$ calibration source provided a steady flux of x-rays onto the detector. Second, the gate valve that isolated the detector chamber from the main payload was controllable via an experiment command-uplink system, allowing for real-time actuation to either expose or protect the detector before, during, and after the flight. Finally, a pull-away liquid nitrogen cooling system was implemented to actively cool the detector until seconds before launch, providing control of H2RG and SIDECAR™ temperature leading up to flight.

After final rocket motor arming was completed $\sim 2\mathrm{h}$ before launch, the camera was powered on and allowed to operate continuously until after the science observation in flight. At the time of liftoff, the H2RG and SIDECAR™ were at 124 and 177 K, respectively. At the time of camera shut down near the end of the flight, the H2RG had warmed to 133 K and the SIDECAR™ to 178 K, keeping both near the target temperatures discussed in Sec. 5.1. Continuous camera operation on the ground, during rocket ascent, throughout the science observation, and for $\sim 30\mathrm{s}$ after the science observation allowed for data to be collected for final ground calibration information, flight dark frames, and flight calibration frames with the on-board ${}^{55}\mathrm{Fe}$ source. With the presence of a constant calibration source throughout the flight, the ability to collect dark frames before and after the science observation, and the collection of dark sky images as the rocket slewed to the SNR, the WRXR successfully pointed at Vela for the entire duration of the allocated observation time without sacrificing calibration data. The total on-target observation time throughout the flight was 289 s.

6.2.

Flight Data

Driven by an automated exposure cadence, the 289 s of on-target observation yielded $\sim 280\mathrm{s}$ of active exposure on the Vela SNR. A processed image of all events near the OVII emission energy (0.57 keV) and a spectrum of those events are shown in Fig. 8. Analysis of the background in the spatial regions where spectral lines should have diffracted indicated that $\sim 12$ events were necessary in a given spectral line to obtain a $3\sigma$ detection of OVII x-rays from Vela. Though $\sim 150$ events are visible over the entire processed image area, there are no apparent spectral lines or preferential event positions. The spectrum further demonstrates that there are no OVII x-ray events that can be identified above the background with statistical significance. More in-depth analysis, including examinations of CVI and OVIII energies and of the regions on the detector at which the spectral lines should be present, yielded the same conclusion: there was no significant detection of soft x-ray emission from Vela during the WRXR flight. This conclusion motivated postflight instrument testing and calibration to determine if there were any unexpected instrument performance issues during flight or if a potential upper limit could be placed on emission from Vela.

Fig. 8

All events detected during the flight observation that satisfied image processing and OVII-thresholding limitations. (a) A scatter plot of the position on the detector of each event and (b) an energy spectrum of all events. The spectrum shows a low-energy background and is flat near the OVII energy (570 eV) with no indication of significant OVII x-ray events.

6.3.

Flight Anomaly

During the payload’s ascent, the pressure measured by the WRXR pressure transducers increased by over an order of magnitude in the span of a few seconds as the payload ascended through the atmosphere. As designed, the gate valve that isolated the detector chamber should have protected the cold (124 K at launch) detector from this pressure spike. However, weather-induced payload damage experienced in the days leading up to launch necessitated a modified flight plan; a sudden rain shower during launch readiness testing damaged the detector chamber’s ion pumping system (discussed in Sec. 3) beyond repair. As a result, the detector chamber lacked a dedicated vacuum pump. To compensate for the ion pump failure, the payload was launched with the gate valve open, exposing the detector to the main instrument earlier than designed but allowing the low pressure of the main instrument to act as a getter for the detector chamber. However, the pressure spike that the payload experienced as it ascended through the atmosphere subjected the cold detector to higher-than-desirable pressure and likely deposited a condensation layer on the detector surface, a possibility discussed further in Sec. 7.2.

7.1.

Component Verification

The two primary objectives of postflight component verification were to measure diffraction efficiency of the gratings and the performance of the H2RG after both experienced a suborbital rocket flight. At the completion of instrument-level testing, the reflection grating array was uninstalled from the instrument and a grating was removed. The disassembled grating was then taken again to the ALS for a similar diffraction efficiency experiment. The results, summarized in Fig. 9, were consistent with preflight performance to within the variation measured in the grating coverage experiment ($\sim 5\%$). The postflight grating efficiency experiment provided verification that the gratings maintained their fidelity throughout exposure to launch conditions, space, and recovery efforts. Similarly, H2RG read noise and energy resolution were again measured under the same temperature conditions, yielding $6.47\pm 0.06{\mathrm{e}}^{-}$ read noise and $212\pm 4\mathrm{eV}$ energy resolution at 5.9 keV. A comparison of postflight detector measurements to both preflight and flight data is summarized in Table 4.

Fig. 9

The total absolute diffraction efficiency of three WRXR gratings. Two gratings were tested before the flight and a third grating was uninstalled from the flight array and tested as part of postflight calibrations. Though there are slight deviations in total diffraction efficiency from grating to grating, which is partially a product of unique test configurations, the total efficiency as a function of energy is consistent in all gratings measured.

Table 4

A comparison of read noise and energy resolution at the 5.9 keV Mn-Kα line produced with the Fe55 radioactive decay source for preflight testing, flight performance, and postflight testing.

7.2.

Condensation Analysis

After verifying that the instrument alignment held throughout the flight and that the individual components maintained their performance levels, postflight analysis moved to the most critical investigation: the expected significance of the pressure spike experienced during the payload’s ascent to space. Extensive testing was performed in an effort to re-create the flight conditions to directly observe the effects on x-ray transmission. The experiment that was determined to be representative of the flight anomaly began with frozen water inside a tube that was connected to the camera chamber, which was under vacuum to best simulate flight conditions. The ice was then allowed to melt, resulting in pure water vapor that was then allowed to condense onto the detector surface. The water vapor volume and pressure used in postflight testing were the best approximation to the conditions experienced in flight to expose the detector to the same amount of water vapor; the increase in pressure during flight is believed to have been caused by the external skin heating and outgassing water, and the postflight experiment sought to match the pressure of water available to condense onto the detector. After the condensation layer was deposited, H2RG x-ray flux levels were measured and compared with pre-exposure measurements with the same source flux. Table 5 summarizes the results of the test, which demonstrated heavy attenuation of all x-rays and $\sim 85\%$ attenuation at $\mathrm{O}\text{-}\mathrm{K}\alpha$ (the closest line in energy to the dominant emission lines from the Vela SNR), and Fig. 10 shows x-ray flux before and after the condensation layer was added to the detector. An estimate of the condensation layer required to attenuate x-rays to the observed level yielded a thickness of $17.5\pm 1.5\mu \mathrm{m}$. Figure 11 shows the transmission of soft x-rays through a $17.5\mu \mathrm{m}$ thick layer of water, demonstrating that a condensation layer of comparable thickness could have effectively eliminated the expected x-ray counts from Vela during flight and prevented any conclusions about Vela’s soft x-ray emission.

Table 5

A summary of the postflight condensation experiment where the camera was subjected to approximately the same level of water vapor as during the flight. The attenuation of x-rays at each distinct energy is consistent with a layer of condensation of ∼17.5  μm thick.

Fig. 10

Results of the postflight condensation test for O, Mg, and Al x-ray lines. (a) Before condensation and (b) after condensation, with equal exposure times. X-rays of all energies are almost entirely attenuated by the condensation layer.

Fig. 11

The transmission of x-rays through a $17.5\text{-}\mu \mathrm{m}$ thick layer of water.³³ The dominant emission lines from Vela (CVI at $\sim 370\mathrm{eV}$, OVII at $\sim 570\mathrm{eV}$, and OVIII at $\sim 650\mathrm{eV}$) would all be almost entirely occulted by such a layer.