Europa's H₂O₂: Temperature Insensitivity and a Correlation with CO₂

Hydrogen peroxide (H₂O₂) is a product of the important water-ice radiolytic cycle on Europa and was first detected by the Galileo Near-Infrared Mapping Spectrometer (NIMS) via a prominent absorption band at 3.5 μm (R. W. Carlson et al. 1999). The continuous bombardment of Europa's surface by electrons, protons, and ions from the Jovian magnetosphere leads to the dissociation of water molecules and the recombination of the fragments to form H₂O₂, molecular oxygen (O₂), and thus an oxidized surface environment (O. Bain & P. A. Giguère 1955; R. E. Johnson & T. I. Quickenden 1997; R. W. Carlson et al. 1999; P. D. Cooper et al. 2003; R. E. Johnson et al. 2003; M. Loeffler et al. 2006; R. W. Carlson et al. 2009). Surface oxidants such as H₂O₂, if transported through Europa's ice shell, could interact with ocean water that is potentially reduced by hydrothermal processes at the seafloor. Such mixing could create a redox potential, providing a viable chemical energy source for Europa's interior ocean (C. F. Chyba 2000; K. P. Hand et al. 2006; K. P. Hand et al. 2009; M. A. Pasek & R. Greenberg 2012). Understanding what controls H₂O₂ can affect our understanding of its stability and how likely it is to get recycled into the interior. It also has implications for the production and stability of related radiolysis products, such as O₂ (P. D. Cooper et al. 2003), that are produced by the same or similar processes but are harder to observe. H₂O₂, O₂, and H₂ formation from ion- and electron-irradiated water ice is expected to be a common radiation chemistry pathway throughout the outer solar system.

Laboratory spectral measurements and irradiation experiments have been used to understand the radiolysis and photolysis of water ice, including the production and destruction of H₂O₂ (e.g., R. E. Johnson & T. I. Quickenden 1997; M. Moore & R. Hudson 2000; M. Loeffler et al. 2006; T. Cassidy et al. 2010; K. P. Hand & R. W. Carlson 2011). In addition to water ice and sufficient irradiation being essential precursors for H₂O₂ formation, irradiation experiments have also shown an inverse relationship between the equilibrium abundances of H₂O₂ and the ice temperature (M. Moore & R. Hudson 2000; M. Loeffler et al. 2006; W. Zheng et al. 2006; K. P. Hand & R. W. Carlson 2011). Thus, it was expected that Europa should have an enrichment of H₂O₂ in colder, more ice-rich regions, compared to a more minimal presence in warmer, less ice-rich regions (K. P. Hand & R. W. Carlson 2011).

However, contrary to the laboratory predictions, recent Keck adaptive optics (AO) observations of Europa's 3.5 μm H₂O₂ absorption feature across its leading hemisphere showed enhanced H₂O₂ in geologically young, low-latitude chaos terrains characterized by lower water-ice abundance (M. E. Brown & K. P. Hand 2013; P. D. Fischer et al. 2015, 2016), warmer temperatures (J. A. Rathbun et al. 2010; S. K. Trumbo et al. 2018), and a notable enrichment of sodium chloride (NaCl; S. K. Trumbo et al. 2019b, 2022). This observation finds contradictory results on the effects of both water-ice availability and temperature. While the findings indicate that temperature is not the dominant factor, it remains uncertain whether temperature plays a role at all in governing Europa's H₂O₂.

In fact, the distribution suggests that the underlying composition of the chaos terrain exerts a stronger control on H₂O₂ abundance. Thus, the authors hypothesized that carbon dioxide (CO₂) may be enhanced in these same terrains (S. K. Trumbo et al. 2019a), where it may act to enhance H₂O₂ concentrations by inhibiting the breakdown of newly formed H₂O₂ by consuming the destructive electrons produced as the ice continues to be irradiated (M. Moore & R. Hudson 2000)—an effect supported at the time by only limited laboratory data (M. Moore & R. Hudson 2000; G. Strazzulla et al. 2005) and unpublished Galileo NIMS spectra (R. W. Carlson 2001). However, this hypothesis remains untested owing to the lack of reliable CO₂ mapping on Europa and the inability to observe it from the ground, leaving the effects of temperature and composition on Europa's H₂O₂ largely unresolved.

To further investigate the effects of temperature and CO₂ on Europa's H₂O₂ concentrations, we analyze a combination of ground- and space-based observations of Europa's H₂O₂ from the the NASA Infrared Telescope Facility (IRTF), Keck Observatory, and JWST. Using IRTF/SpeX, we take disk-integrated observations of Europa's 3.5 μm H₂O₂ absorption before and after eclipse, which preserves approximately the same geometry across a large temperature change, enabling a largely controlled test of possible temperature effects. To examine the reaction of H₂O₂ to temperature fluctuations across longer, diurnal timescales, we present spatiotemporally resolved observations across Europa's day acquired using the Keck II telescope's Near InfraRed Spectrograph (NIRSPEC) and AO system. Finally, we use spatially resolved JWST Cycle 1 NIRSpec observations of Europa, which simultaneously span wavelengths sensitive to both H₂O₂ and CO₂, to search for the previously hypothesized correlation between these two species.

2.1. Disk-integrated Pre- and Post-eclipse IRTF/SpeX Observations

To disentangle the effect of temperature on Europa's H₂O₂ from that of composition, we designed a controlled temperature experiment using disk-integrated spectra obtained in 1 hr windows across 2 hr pre-eclipse and 6 hr post-eclipse with IRTF/SpeX (J. T. Rayner et al. 2003) across several dates with similar eclipse windows in 2018. The SpeX data obtained span a wavelength range of 1.68–4.23 μm with R ∼ 2000. Dates, time, and geometries are given in Table 1.

Table 1. Table of Observations

Date	Telescope	Hemisphere	Slit Orientation	Air Mass Range	Integration	Central Longitude	Central Latitude	Star Calibrator
(UT)					(minutes)
2018 Mar 29	IRTF/Spex	Sub-Jovian a	Disk-integrated	1.3	32	345 W	4 S	HD 140990
2018 Mar 29	IRTF/Spex	Sub-Jovian b	Disk-integrated	1.3–1.4	48	348 W	4 S	HD 140990
2018 May 11	IRTF/Spex	Sub-Jovian c	Disk-integrated	1.2–1.3	32	11 W	4 S	HD 128596
2018 May 11	IRTF/Spex	Sub-Jovian d	Disk-integrated	1.2	40	14 W	4 S	HD 128596
2018 Jun 30	IRTF/Spex	Sub-Jovian e	Disk-integrated	1.2	8	29 W	4 S	HD 128597
2018 Jun 30	IRTF/Spex	Sub-Jovian f	Disk-integrated	1.2	43	31 W	4 S	HD 128597
2018 Jun 30	IRTF/Spex	Sub-Jovian g	Disk-integrated	1.2–1.3	33	36 W	4 S	HD 128597
2018 Jun 30	IRTF/Spex	Sub-Jovian h	Disk-integrated	1.3–1.5	38	39 W	4 S	HD 128597
2013 Dec 29	Keck AO	Anti-Jovian	E/W	1.0–1.2	40	195 W	3 N	HD 54371
2013 Dec 29	Keck AO	Anti-Jovian	E/W	1.1	20	195 W	25 S	HD 54372
2016 Feb 24	Keck AO	Trailing/sub-Jovian	E/W	1.5–1.7	40	334 W	2 S	HD 98947
2016 Feb 24	Keck AO	Trailing/sub-Jovian	E/W	1.2–1.3	40	338 W	0 N	HD 98947
2016 Feb 24	Keck AO	Trailing/sub-Jovian	E/W	1.0–1.1	40	342 W	1 S	HD 98947
2016 Feb 24	Keck AO	Sub-Jovian	N/S	1	40	350 W	2 S	HD 98947
2016 Feb 24	Keck AO	Sub-Jovian	N/S	1	40	348 W	2 S	HD 98947
2016 Feb 25	Keck AO	Leading	E/W	1.2–1.4	40	79 W	0 S	HD 98947
2016 Feb 25	Keck AO	Leading	E/W	1.1–1.2	40	82 W	12 S	HD 98947
2016 Feb 25	Keck AO	Leading	N/S	1.0–1.1	40	88 W	2 S	HD 98947
2016 Feb 25	Keck AO	Leading	E/W	1.1	40	95 W	0 N	HD 98947
2018 Jun 06	Keck AO	Leading	E/W	1.5	20	115 W	7 S	HD 128596
2018 Jun 06	Keck AO	Leading	N/S	1.4	10	120 W	5 S	HD 128596
2018 Jun 06	Keck AO	Leading	N/S	1.4	10	126 W	5 S	HD 128596
2018 Jun 06	Keck AO	Leading/anti-Jovian	N/S	1.2–1.3	20	171 W	2 S	HD 128596
2018 Jun 06	Keck AO	Leading/anti-Jovian	N/S	1.2–1.3	20	150 W	2 S	HD 128596
2018 Jun 06	Keck AO	Leading/anti-Jovian	N/S	1.3	10	166 W	1 S	HD 128596
2018 Jun 06	Keck AO	Leading	N/S	1.2	15	113 W	4 S	HD 128596
2018 Jun 06	Keck AO	Leading	N/S	1.2–1.3	20	100 W	6 S	HD 128596
2018 Jun 06	Keck AO	Leading	N/S	1.3	20	82 W	3 S	HD 128596
2018 Jun 06	Keck AO	Leading	E/W	1.3–1.4	20	131 W	1 S	HD 128596
2018 Jun 07	Keck AO	Trailing	E/W	1.2–1.3	40	222 W	4 S	HD 128596
2021 Sep 24	Keck AO	Trailing	E/W	1.2–1.4	60	275 W	20 N	HD 203311
2021 Sep 25	Keck AO	Trailing	N/S	1.3–1.4	40	318 W	8 N	HD 203311
2021 Sep 25	Keck AO	Sub-Jovian/Leading	N/S	1.3	40	41 W	2 N	HD 203311
2022 Nov 23	JWST	Leading	⋯	⋯	21	93 W	3 N	GSPC P330-E

Notes.

^a 2 hr pre-eclipse. ^b 1 hr pre-eclipse. ^c 1 hr post-eclipse. ^d 2 hr post-eclipse. ^e 3 hr post-eclipse. ^f 4 hr post-eclipse. ^g 5 hr post-eclipse. ^h 6 hr post-eclipse.

Download table as:  ASCIITypeset image

Depending on the exact thermal inertia of the surface materials, an eclipse leads to a sudden temperature drop of tens of degrees kelvin, while the satellite maintains approximately the same geometry as viewed from Earth (Figure 1(a)). As simple extrapolation from laboratory studies suggests that a ∼20 K temperature drop may increase the equilibrium band area of the H₂O₂ absorption by a factor of ∼4 (K. P. Hand & R. W. Carlson 2011), we assessed changes in Europa's H₂O₂ absorption band immediately before and after eclipse and throughout this recovery period to look for such a response. As Europa is tidally locked to Jupiter and experiences an eclipse of its sub-Jovian hemisphere every orbit, pre- and post-eclipse spectra taken on different dates are still sampling the same temperature drop and recovery (Table 1). Our data cover the wavelength range of ∼1.68–4.2 μm. For pre-eclipse observations, we observed HD 140990, a V = 7.85 G2V star with a 708 separation from Europa, as the telluric calibrator. For post-eclipse observations, we observed HD 128596, a V = 7.48 G2V star with a 245 separation from Europa, as the telluric calibrator.

Zoom In Zoom Out Reset image size

We used Spextool (Spectral Extraction TOOL), an interactive data language (IDL)–based data reduction package (M. Cushing et al. 2004), to reduce the IRTF/SpeX data. We followed the standard methodology of flat-field correction, image pair subtraction, target detection, spectra extraction, wavelength calibration, combination of multiorder spectra, and telluric correction (W. Vacca et al. 2003). Spectra from within each 1 hr window were merged into one single spectrum per window. We clipped out bad pixels and any highly variable telluric lines by comparing data points to the means of a 51-point window with a threshold of 0.01, as determined by manually inspecting the spectra so that only telluric lines and outliers are clipped. To further improve the signal-to-noise ratio, we then smoothed the spectra using a Savitzky–Golay filter with a window length of 5 data points and a polynomial order of 3.

Though the geometry was nearly identical between subsequent time windows, Europa did rotate ∼50° in longitude across the entire 8 hr of observation. To address this caveat, we used spatially resolved Keck AO observations of the sub-Jovian hemisphere to estimate the effects of small changes in observed band strength attributable to the minor amount of rotation in the eclipse experiment.

2.2. Spatially Resolved Keck AO Observations across Europa's Day

2.3. Spatially Resolved JWST Observation

2.4. Band Strength Calculation

To isolate the effect of temperature on Europa's H₂O₂ from those of composition or geography, we measure the 3.5 μm H₂O₂ band on Europa's sub-Jovian hemisphere across a 2 hr period before eclipse and across a 6 hr period following the conclusion of its daily eclipse. Using the thermal model of S. K. Trumbo et al. (2017b) and the best-fit thermal inertia value (60 J/(m² K s^1/2)) found in S. K. Trumbo et al. (2018) from Atacama Large Millimeter/submillimeter Array (ALMA) observations of the sub-Jovian hemisphere, we estimate an expected ∼20 K temperature drop across eclipse and a recovery to pre-eclipse temperatures that span the 6 hr post-eclipse period across which we observed (Figure 1(a)).

K. P. Hand & R. W. Carlson (2011) irradiated pure water ice with 10 keV electrons at temperatures of 80, 100, and 120 K and monitored the change in H₂O₂ band absorption until equilibrium concentrations were achieved at flux levels of ∼10¹⁹ eV cm⁻². The authors' measured equilibrium H₂O₂ band area at 120 K is approximately a factor of 4 weaker than the measured equilibrium H₂O₂ band area at 100 K and a factor of 6 weaker than the measured equilibrium H₂O₂ band area at 80 K. This laboratory result is consistent with the expected increased mobility of OH with increasing temperature (T. I. Quickenden et al. 1991; R. E. Johnson & T. I. Quickenden 1997; M. Loeffler et al. 2006; K. P. Hand & R. W. Carlson 2011). However, it is worth noting that the magnitude of temperature effects seen in laboratory studies varies, with some experiments (e.g., O. Gomis et al. 2004; M. Loeffler et al. 2006; W. Zheng et al. 2006) observing smaller changes in H₂O₂ concentration than K. P. Hand & R. W. Carlson (2011) and some modeling work predicting smaller effects as well (B. D. Teolis et al. 2017). Nevertheless, taking the most simplistic extrapolation from these experiments, we might expect up to a factor of 4 increase in our observed H₂O₂ band area immediately after the ∼20 K eclipse temperature drop and then a slow decrease of the absorption band as the temperature climbs back up toward pre-eclipse values.

In Figure 1(b), we plot H₂O₂ band areas of the averaged spectra from each 1 hr time window surrounding the eclipse, where the error bars are estimated by varying the continuum fits across the wavelength range from 3.33–3.37 μm to 3.715–3.78 μm. The unknown true continuum presents a larger source of error than noise, but our error bars are still far smaller than the large increases in band strength we seek to detect. All of the spectra used to derive the band areas in Figure 1 are included in Figure A1. We find that our data clearly rule out such a dramatic increase in the H₂O₂ band following eclipse, instead exhibiting only a small 22.6% change (factor of 1.23) between the measurements immediately pre-eclipse and immediately post-eclipse and a 38.7% change (factor of 1.53) between the weakest 3.5 μm absorption at 2 hr pre-eclipse and the strongest 3.5 μm absorption at 5 hr post-eclipse.

It is possible that this small change does reflect a delayed response to the eclipse temperature drop resulting from the unknown timescale needed for Europa's surface H₂O₂ concentrations to re-equilibrate. However, the 2 and 3 μm water-ice absorption bands also appear to change following the eclipse and across the recovery period, which suggests that Europa's small degree of rotation from the slightly less icy trailing/sub-Jovian geometry pre-eclipse (central longitude = 348° west) toward a slightly more icy sub-Jovian/leading geometry (central longitude = 36° west) by 5 hr post-eclipse needs to be considered as well. Our spatially resolved Keck observation across the sub-Jovian hemisphere suggests up to a factor of ∼2.1 increase in H₂O₂ band strength as the longitude shifts from 352° west to 33° west. Thus, the small change (factor <2) of H₂O₂ band strength following eclipse could be reasonably attributed to the ∼50° of rotation of Europa surrounding eclipse, and our eclipse experiment reveals no clear evidence for any response of Europa's H₂O₂ to the temperature changes across eclipse, let alone the large factors seen in laboratory experiments. However, it is possible that the unknown time it would take Europa's H₂O₂ to re-equilibrate following such a perturbation is longer than the eclipse experiment allowed.

While we find no clear effect of temperature in our IRTF eclipse experiment, the laboratory data that see strong differences in H₂O₂ with temperature represent a direct comparison of equilibrium concentrations (e.g., K. P. Hand & R. W. Carlson 2011), which may not have been reached across the eclipse timescale. For this reason, we also use Keck NIRSPEC AO observations to look for variability in Europa's H₂O₂ over longer, diurnal timescales. In particular, we examine slits oriented in an east/west direction that cover the same surface regions at different times of the local day. As noted above, K. P. Hand & R. W. Carlson (2011) suggest that equilibrium abundances of H₂O₂ can vary by a factor of 4 across the 100–120 K temperature range relevant to Europa and that H₂O₂ can be destroyed during the day in equatorial regions and produced efficiently at night across the surface, such that higher H₂O₂ concentration might be expected in the early morning as opposed to in the warm afternoon.

During our 2016 and 2018 observations, we purposely aligned slits taken on adjacent nights to investigate the hypothesized diurnal variability of H₂O₂. We examine five east–west slit pairs centered within ±20° of the equator, spanning a timescale of 20–29 hr—approximately one-quarter to one-third of a Europa day—and covering the temperature progression from morning to afternoon. Similar to what we see from the eclipse experiment, the changes in integrated band area from morning to afternoon appear to be much smaller than the equilibrium variations observed in the lab and instead range from −32% to 30% (factor of 0.68–1.3), with no clear or consistent trend attributable to diurnal variability.

As a representative example, in Figure 2 we compare spectra of the same location (186 west, 0.4° north) observed in the morning (cold temperatures) and afternoon (warmer temperatures) of 2016 February. Using the thermal model of S. K. Trumbo et al. (2017a) and a thermal inertia of 60 J/(m² K s^1/2) suggested by ALMA observations of the sub-Jovian hemisphere (S. K. Trumbo et al. 2018) and consistent with thermal inertia values estimated from the Galileo Photopolarimeter-Radiometer (PPR) observations (J. R. Spencer et al. 1999; J. A. Rathbun et al. 2010), we estimate a corresponding ∼12 K temperature increase over a ∼24 hr timescale for this location. We observe only a minimal change (−3.1%) in the integrated H₂O₂ band areas for this location, and, indeed, the entirety of the overlapping portions of the two slits appear consistent. All of the overlapping E/W slit positions we inspect are presented in Figure A2.

Zoom In Zoom Out Reset image size

To expand on this analysis, we also assess two east/west slit pairs observed over a timescale of 4 and 16 hr between 2013 and 2018 (rather than subsequent nights) in our larger data set (Figure A2). We observe only minimal changes in integrated band areas from −0.24% to 0.44% (factor of 0.76–1.44). We also note that there appears to be some evidence for stochastic H₂O₂ variability between years, as noted in S. K. Trumbo et al. (2019a), where the bands observed in 2016 are overall stronger than in later years (Figure A2). We suggest that this could be related to potential variability of the dynamic radiation environment.

The lack of a strong or consistent diurnal trend, along with the minimal changes observed across eclipse, suggest that Europa's H₂O₂ is largely insensitive to the temperature fluctuations that it experiences on a daily basis. Thus, our results indicate that, in addition to temperature not influencing the spatial variations of Europa's H₂O₂, temperature also appears not to drive strong temporal variations. We discuss this apparent discrepancy with laboratory work further in Section 7.

In addition to suggesting that diurnal temperature changes do not strongly control Europa's H₂O₂, the Keck observations provide additional spatial coverage of Europa's surface compared to past maps of H₂O₂ (S. K. Trumbo et al. 2019a). In Figure 3, we present maps of H₂O₂ band area composed of east/west-oriented and north/south-oriented slits from 2013, 2016, 2018, and 2021, which include additional coverage beyond the previous maps. We separated out coverage of the leading and trailing hemispheres for clarity and to highlight the longitudinal variation along the two included anti-Jovian E/W slits. Overall, the updated distribution agrees well with the trends shown by S. K. Trumbo et al. (2019a).

Zoom In Zoom Out Reset image size

Our maps present a clear difference between low-latitude chaos and low-latitude plains terrain in east/west slits crossing the leading chaos regions (Figures 2(b), 3(a), A2(a)–(d)) and a north/south slit just east of Tara Regio (∼85° west; Figure 3(a)) shows overall weaker H₂O₂ absorption and less latitudinal variation in H₂O₂ than do north/south slits crossing Tara Regio. This distribution further confirms the curious enrichment of the 3.5 μm H₂O₂ feature within the large-scale chaos terrains at low to midlatitudes on the leading hemisphere (S. K. Trumbo et al. 2019a), which are understood to contain endogenic non-ice material like salts (P. D. Fischer et al. 2015, 2016; S. K. Trumbo et al. 2019b, 2022).

Our maps also provide additional north/south and east/west coverage on the trailing hemisphere. Given the drastic changes in the spectral continuum shape between the two hemispheres associated with changes in water-ice abundance, we carefully inspected all continuum fitting manually and adjusted fitting parameters accordingly. Consistent with the hemispheric data of K. P. Hand & M. E. Brown (2013) and the spatially resolved data of S. K. Trumbo et al. (2019a), the trailing regions show much weaker H₂O₂ absorption overall. This hemispheric difference in H₂O₂ abundance may stem from the strong contrast in water-ice fraction between the leading and trailing hemispheres (K. P. Hand & M. E. Brown 2013; S. K. Trumbo et al. 2019a). The widespread SO₂ on the trailing hemisphere (e.g., T. M. Becker et al. 2022) may also contribute to the depletion of H₂O₂ via interactions to create sulfate (M. J. Loeffler & R. L. Hudson 2013; S. K. Trumbo et al. 2019a).

Given that the H₂O₂ absorption on the trailing hemisphere is so weak against the background continuum, we cannot robustly distinguish between trailing chaos and non-chaos regions or make a clear detection of any latitudinal trend on this hemisphere. However, we do repeatedly see a depletion of H₂O₂ west of the anti-Jovian point around 210° west, which corresponds to a gap between the large-scale chaos regions of the leading and trailing hemispheres (Figures 3, A2), hinting that similar geological associations could exist below our detection across the trailing side. This depletion around 210° west also corresponds to lower mean reflectances across the L band (3.15–3.95 μm) mapped in P. D. Fischer et al. (2016). Future high signal-to-noise ratio JWST observations of Europa's trailing hemisphere would help further investigate the geographic variation of its weak H₂O₂ bands.

With the additional Keck coverage, we confirm a strong correlation between concentrated H₂O₂ and chaos terrains. This suggests that the endogenic composition of these regions may play an important role in controlling the distribution of H₂O₂ and that a possible correlation with CO₂ remains a key hypothesis (S. K. Trumbo et al. 2019a).

To investigate the hypothesized effect of CO₂ on the unexpected distribution of H₂O₂ (S. K. Trumbo et al. 2019a), we measure the H₂O₂ band in the Cycle 1 JWST data recently used to map Europa's leading-hemisphere CO₂ (G. L. Villanueva et al. 2023; S. K. Trumbo & M. E. Brown 2023) and evaluate whether H₂O₂ and CO₂ are geographically correlated. Both features are observed simultaneously in the same data set, ensuring the best geographic alignment.

In Figure 4(a), we map the integrated band areas of the 3.5 μm H₂O₂ feature across Europa's leading hemisphere and show the corresponding map of Europa's ν₃ CO₂ band, as obtained by S. K. Trumbo & M. E. Brown (2023) (Figure 4(b)). In agreement with the Keck observations (S. K. Trumbo et al. 2019a; this work), we find that the strongest H₂O₂ absorption occurs in the leading-hemisphere chaos terrains Tara Regio (∼85° west) and Powys Regio (∼125° west) (Figure 4(a)), which also exhibit the strongest CO₂ signatures (Figure 4(b)). In fact, a simple linear fit between the H₂O₂ and CO₂ band strengths in the JWST data reveals a strong positive correlation (Pearson correlation coefficient of 0.94), as would be expected for the previously hypothesized effect of endogenic CO₂ on radiolytically produced H₂O₂ (Figure 4(c)).

Zoom In Zoom Out Reset image size

Two distinct peaks are resolved within the ν₃ band of Europa's CO₂ (at 4.25 and 4.27 μm; S. K. Trumbo & M. E. Brown 2023), along with a combination band at 2.7 μm (G. L. Villanueva et al. 2023). As these signatures might indicate different CO₂ host materials or trapping mechanisms, we also assess the correlation of H₂O₂ with the individual depths of each CO₂ peak within the ν₃ band and with the strength of the 2.7 μm feature. The Pearson correlation coefficients (R) calculated with respect to the 4.25 μm peak, the 4.27 μm peak, and the 2.7 μm band—with coefficients of 0.92 (P < 0.001), 0.92 (P < 0.001), and 0.90 (P < 0.001), respectively—are consistent with that calculated with respect to the overall ν₃ band (R = 0.94; P < 0.001). As all three CO₂ signatures are individually the strongest within Tara Regio (∼85° west) and Powys Regio (∼125° west; G. L. Villanueva et al. 2023; S. K. Trumbo & M. E. Brown 2023), it is unsurprising that we cannot separate a correlation between H₂O₂ and a particular CO₂ phase from a simple strong correlation with the overall abundance of CO₂ regardless of its trapping mechanisms or host materials. Nevertheless, the strong pixel-to-pixel correspondence suggests that the presence of CO₂ in some form may be an important factor affecting the distribution of H₂O₂ in the chaos terrains, though the possibility remains that H₂O₂ could be associated with chaos terrains for an independent reason.

Our findings suggest that Europa's H₂O₂ is largely insensitive to the daily temperature changes it experiences. We find no strong evidence to support the idea that H₂O₂ builds up efficiently overnight and then becomes depleted across Europa's day, as hypothesized from laboratory work demonstrating an inverse dependence of equilibrium H₂O₂ concentrations on ice temperature (K. P. Hand & R. W. Carlson 2011). We suggest that this apparent discrepancy might be due to the short timescales of Europa's temperature changes, though we also note that the observed magnitude of the influence of temperature also varies widely across different irradiation experiments and may be part of the explanation as well (e.g., O. Gomis et al. 2004; M. Loeffler et al. 2006; W. Zheng et al. 2006). Indeed, conditions directly analogous to Europa's diurnal cycles have, to our knowledge, not been simulated in the lab, and the timescale for the reaction of the dynamic equilibrium of production and destruction of H₂O₂ in impure ice to temperature perturbations is uncertain. K. P. Hand & R. W. Carlson (2011) infer that, if starting from pristine ice (with no H₂O₂ or other non-ice materials, like CO₂) and irradiating at a constant surface temperature (as was done in their experiments), it would take an estimated ∼48 hr of irradiation on the trailing hemisphere and approximately 2 weeks on the leading hemisphere to reach equilibrium (R. W. Carlson et al. 2009; K. P. Hand & R. W. Carlson 2011). These timescales are extremely short in geologic terms, suggesting that a dynamic equilibrium with respect to H₂O₂ is almost instantly achieved following any resurfacing events, but they do not directly enable precise predictions of diurnal variations. Future irradiation experiments that first reach equilibrium H₂O₂ concentrations and then continuously monitor changes in H₂O₂ absorption, while simulating the temperature fluctuations observed on Europa through controlled heating and cooling, could help further investigate the timescales needed for Europa's H₂O₂ to respond to local temperature changes.

Though our IRTF and Keck results contradict the hypothesized temperature effects, our JWST results showing a strong correlation between H₂O₂ and CO₂ within large-scale, leading-hemisphere chaos regions support the previously proposed hypothesis that CO₂ may be an important factor controlling the geographic distribution of H₂O₂ on Europa (S. K. Trumbo et al. 2019a). Certain laboratory experiments have shown that electron-accepting contaminants, such as O₂ and CO₂, have the potential to enhance H₂O₂ yields (M. Moore & R. Hudson 2000; G. Strazzulla et al. 2005), potentially by inhibiting the breakdown of newly formed H₂O₂ by consuming the destructive electrons produced as the ice continues to be irradiated (M. Moore & R. Hudson 2000). Indeed, recent work continues to build on these prior experiments and appears to confirm the augmenting effects of small amounts of CO₂ more robustly (B. Mamo et al. 2023; U. Raut et al. 2024), which further suggests that the spatial correlation between Europa's CO₂ and H₂O₂ reflects this effect, rather than coincidence.

Though CO₂ may explain the surprising association of radiolytically produced H₂O₂ with endogenic terrains on the leading hemisphere, it remains unclear how important it is for the trailing side. It is possible that the lower water-ice fraction (e.g., M. E. Brown & K. P. Hand 2013; P. D. Fischer et al. 2015; N. Ligier et al. 2016) and warmer temperatures (e.g., J. A. Rathbun et al. 2010; S. K. Trumbo et al. 2018) on the trailing side may help explain its comparative lack of H₂O₂. The potential chemical reaction with SO₂ (M. J. Loeffler & R. L. Hudson 2013) also remains a compelling possibility. As there are few existing constraints on the possibility of CO₂ on Europa's trailing hemisphere (R. W. Carlson et al. 2009; G. B. Hansen & T. B. McCord 2008), the magnitude of its influence relative to these other effects remains an open question. Future JWST observations of the trailing side would help further explore the H₂O₂–CO₂–chaos connection and in the context of the added complications of sulfur radiolysis (e.g., R. W. Carlson et al. 2002; R. W. Carlson et al. 2009).

H₂O₂ was recently discovered on Ganymede and was found to be associated with its cold, icy, and highly irradiated polar regions (S. K. Trumbo et al. 2023). This discovery contrasts with the distribution of Europa's H₂O₂ in warm, ice-poor, large-scale chaos terrains at low to midlatitudes and instead aligns better with laboratory expectations on the effects of water-ice availability, temperature, and irradiation (e.g., M. Moore & R. Hudson 2000; M. Loeffler et al. 2006; W. Zheng et al. 2006; K. P. Hand & R. W. Carlson 2011). Indeed, Ganymede's equatorial latitudes, where little H₂O₂ is observed, are shielded from irradiation by its intrinsic magnetic field (e.g., K. Khurana et al. 2007; A. R. Poppe et al. 2018) and are warmer than Europa's equator (e.g., G. S. Orton et al. 1996; J. R. Spencer et al. 1999; K. de Kleer et al. 2021). Thus, though such factors do not seem to control the geographic distribution of H₂O₂ on Europa, they may still be important on Ganymede. Interestingly, however, Ganymede's surface also features ubiquitous CO₂ (e.g., C. A. Hibbitts et al. 2003), which JWST data suggest exists in multiple trapped phases across the surface (D. Bockelée-Morvan et al. 2024). The CO₂ appears most abundant in the equatorial regions, where it is suggested to be trapped in dark and non-icy substrates (C. A. Hibbitts et al. 2003; D. Bockelée-Morvan et al. 2024), which would most likely separate it from the production of H₂O₂ in water ice. However, in the polar regions the CO₂ may be trapped within amorphous water ice and is spatially coincident with the H₂O₂ (D. Bockelée-Morvan et al. 2024). This opens up the possibility of a similar influence on H₂O₂ as seems likely on Europa and is consistent with the possibility that CO₂, in addition to the low temperatures, more abundant water ice, and preferential radiolysis driven by Jovian magnetospheric particles, might help enhance H₂O₂ abundances at Ganymede's poles.

JWST observations of Callisto, in contrast, reveal no signs of H₂O₂ (R. J. Cartwright et al. 2024), despite its widespread CO₂ (C. A. Hibbitts et al. 2000, 2002; R. J. Cartwright et al. 2024). The lack of H₂O₂ on Callisto might be related to multiple factors. First, Callisto's surface, although featuring localized bright and ice-rich knobs (J. M. Moore et al. 2004), appears to be primarily characterized by a dark, ice-poor lag deposit (J. M. Moore et al. 1999; R. J. Cartwright et al. 2024), which would restrict the availability of water ice for radiolysis. Moreover, Callisto's distance from Jupiter results in a much weaker impact of charged-particle irradiation on its surface, and the flux of charged particles from Jupiter's magnetosphere is hundreds of times lower than at Europa (J. F. Cooper et al. 2001; R. Johnson et al. 2004). Additionally, Callisto (like Ganymede's equatorial latitudes) has a higher surface temperature (representative equatorial daytime temperature ∼155 K) than Europa (representative equatorial daytime temperature of ∼120 K) (R. Hanel et al. 1979; J. R. Spencer 1987; J. R. Spencer et al. 1999; J. A. Rathbun et al. 2010; S. K. Trumbo et al. 2018; M. Camarca et al. 2023). Collectively, these factors may inhibit the production of H₂O₂ at Callisto to the point where the presence of CO₂ is not enough to result in detectable H₂O₂.

Looking at H₂O₂ across the Galilean system may thus help us understand the factors controlling its presence on each moon. CO₂ appears likely important for H₂O₂, on Europa and potentially also on Ganymede. Surface temperature, while not important for spatial or temporal variations of H₂O₂ on Europa, does align with the influences expected from laboratory studies (e.g., M. Moore & R. Hudson 2000; M. Loeffler et al. 2006; W. Zheng et al. 2006; K. P. Hand & R. W. Carlson 2011) when looking at the icy Galilean satellites as a whole. Future JWST observations of the icy satellites of Saturn and icy Kuiper Belt objects—like Charon (Pluto's largest moon), where both CO₂ and H₂O₂ were recently detected (S. Protopapa et al. 2023)—may enhance our understanding of the complex interplay of factors influencing the radiolytic production of H₂O₂, thereby providing deeper insights into this important process in the outer solar system.

In this study, we use a combination of IRTF, Keck, and JWST observations of Europa to assess the influence of temperature and CO₂ on Europa's H₂O₂. We find at most minimal changes in H₂O₂ as a result of local temperature variations on timescales associated with Europa's eclipse by Jupiter and diurnal cycle, suggesting a surprising temperature insensitivity that appears at odds with some laboratory studies. We suggest that this apparent discrepancy might be explained if the diurnal timescales are short compared to those needed for the H₂O₂ concentrations to respond to a temperature perturbation and recommend laboratory experiments to explore this further. Using JWST NIRSpec data, we find a clear and strong correlation between Europa's H₂O₂ and CO₂, supporting the previously suggested hypothesis that endogenic CO₂ may enhance H₂O₂ abundances and thereby explain the unexpected concentration of H₂O₂ in low-latitude chaos terrain. This correlation is also in agreement with limited past and recent laboratory work suggesting that CO₂ can act as an electron-scavenging contaminant that increases H₂O₂ yields by inhibiting its breakdown. Future spatially resolved observations on Europa's trailing hemisphere can further test this correlation and explore the possible influence of sulfur-bearing, trailing-hemisphere components as well.

The IRTF/Spex data presented were obtained at the Infrared Telescope Facility, which is operated by the University of Hawaii under contract 80HQTR24DA010 with the National Aeronautics and Space Administration. The Keck AO data presented were obtained at Keck Observatory, which is a private 501(c)3 nonprofit organization operated as a scientific partnership among the California Institute of Technology, the University of California, and the National Aeronautics and Space Administration. The Observatory was made possible by the generous financial support of the W. M. Keck Foundation. The authors wish to recognize and acknowledge the very significant cultural role and reverence that the summit of Maunakea has always had within the indigenous Hawaiian community. We are most fortunate to have the opportunity to conduct observations from this mountain. The JWST data were obtained from the Mikulski Archive for Space Telescopes (MAST) at the Space Telescope Science Institute, which is operated by the Association of Universities for Research in Astronomy under NASA contract NAS 5-03127 for JWST. These observations are associated with program No. 1250. The authors acknowledge H. Hammel and the GTO team led by PI G. Villanueva for developing their observing program with a zero-exclusive-access period. The specific observations analyzed can be accessed via DOI:10.17909/e269-sm44. S. K. Trumbo acknowledges support from the Heising-Simons Foundation through a 51 Pegasi b Fellowship. We thank Ryleigh Davis for fruitful discussions and Jonathan Lunine for insightful comments and constructive suggestions. We thank Erin Leonard for providing shape files for Europa's geologic units.

Facilities: IRTF/Spex - , Keck II - , JWST - James Webb Space Telescope

Software: Astropy (T. P. Robitaille et al. 2013; A. M. Price-Whelan et al. 2018, 2022), Scikit-image (S. van der Walt et al. 2014), Scipy (P. Virtanen et al. 2020), Cartopy (Met Office 2010-2015), Astroquery (A. Ginsburg et al. 2019), Spextool (M. Cushing et al. 2004)

This appendix includes supplementary figures of averaged IRTF spectra for each 1‐hr time window (Figure A1), additional Keck paired slits to illustrate the lack of a strong diurnal trend (Figure A2), and two example JWST spectra used for Figure 4 (Figure A3).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size