1 3C 249.1, z_abs_=0.24676. In this system, HI Ly{beta} is mildly blended with
 1  an unrelated line. However, most of the Ly{beta} profile is free from the
 1  blend, and the unblended portion of Ly{beta} provides useful constraints
 1  and was included in the fit. Hot pixels are present in the STIS spectrum
 1  on both the blue and red sides of the OVI {lambda}1037.62 line
 1  (see Fig. 5). The OVI identification is secure because the
 1  {lambda}1031.93 and {lambda}1037.62 profiles agree well in the regions
 1  that are not affected by hot pixels. However, the regions affected by
 1  hot pixels were excluded from the fit.
 2 3C 249.1, z_abs_=0.30811. Both lines of the OVI doublet are detected at high
 2  significance at this redshift (see Fig. 7). However, the OVI profiles are
 2  strongly saturated in the component at v=0km/s. In addition, hot pixels are
 2  present in the core of the OVI {lambda}1031.93 line. Consequently, the line
 2  parameters are highly uncertain for the v=0km/s component.
 3 3C 249.1, z_abs_=0.31364. The OVI {lambda}1037.62 line at this redshift is
 3  only detected at the 1.9{sigma} level. However, the strength of the 2{sigma}
 3  feature is in good agreement with the expected strength implied by the
 3  well-detected OVI {lambda}1031.93 line. In addition, the
 3  OVI {lambda}1031.93 line is well-aligned with the HI Ly{alpha}, Ly{beta},
 3  and Ly{gamma} lines detected at this redshift. The metal profiles
 3  marginally suggest the presence of a second component, but better S/N
 3  is required to verify and measure the second component.
 4 3C 273.0, z_abs_=0.00334. The blue side of the Ly{beta} profile is blended
 4  with a Galactic H2 absorption line (Sembach et al., 2001ApJ...561..573S).
 4  Consequently, only the red side of the Ly{beta} line (which is free from
 4  blending) was used to constrain the fit. The OVI {lambda}1037.62 line is
 4  severely blended with Galactic H2 absorption, so only the
 4  OVI {lambda}1031.93 line can be measured. The OVI {lambda}1031.93 line is
 4  affiliated with a well-detected HI absorber at the same redshift
 4  (Sembach et al., 2001ApJ...561..573S). This absorber clearly shows evidence
 4  of multiple components in the HI Ly{alpha} profile (see Fig. 33). The OVI
 4  profile, on the other hand, only shows one clear component. However, the
 4  OVI line is broad and shallow, and the breadth of the OVI feature is
 4  consistent with the velocity range spanned by the HI absorption. The degree
 4  of velocity-centroid alignment of the OVI with HI is ambiguous; the OVI is
 4  aligned, to within the 2{sigma} uncertainty, with both HI components derived
 4  from profile fitting in Table 3. Following the conventions outlined in
 4  Section 2.3.2, we assign the OVI to the broader HI component. In addition,
 4  as discussed in Section 4.2, the breadth of the HI component at v=-8km/s
 4  depends critically on the number of components assumed for the fit.
 5 3C 273.0, z_abs_=0.12003. At this redshift, the OVI doublet is covered by 
 5  both the STIS and the FUSE spectrum of 3C 273.0. Both lines of the OVI 
 5  doublet are clearly detected in the FUSE spectrum of 3C 273.0 (see 
 5  Fig. 3 in Tripp et al., 2006ASPC..348..341T). In the STIS spectrum, the OVI
 5  lines are located in a region of rapidly decreasing S/N. The STIS spectrum
 5  shows the OVI 1037.62 line at 4.0{sigma} significance, but the significance
 5  of the 1031.93 line is <2{sigma}. Consequently, we base our fit on the FUSE
 5  data, but we note that fitting the STIS data yields consistent (but
 5  noisier) results.
 6 3C 273.0, z_abs_=0.15779. The OVI 1031.93 line is detected at 4.5{sigma}
 6  significance, but the weaker OVI {lambda}1037.62 line is not detected. The
 6  OVI identification is favored based on the precise alignment of the
 6  OVI {lambda}1031.93 candidate with an HI Ly{alpha} line at the same
 6  redshift (see Fig. 18).
 7 3C 351.0, z_abs_=0.21811. Figure 38 shows the HI Ly{alpha} and OVI lines that
 7  we detect in this system; the top panels show the absorption profiles and 
 7  the bottom panel compares the OVI N_a_(v) profiles. The OVI {lambda}1031.93,
 7  {lambda}1037.62 lines are detected at the 5.5{sigma} and 3.7{sigma} levels,
 7  respectively, and the N_a_(v) profiles are in good agreement. The OVI
 7  profiles are broad and shallow and, hence, are sensitive to continuum
 7  placement. Thom & Chen (2008ApJ...683...22T) do not agree with this system
 7  identification; the most likely source of this discrepancy is continuum
 7  placement, but differences in data reduction procedures could play a role.
 7  Higher S/N observations with the Cosmic Origins Spectrograph (COS) would be
 7  valuable for confirmation of broad and shallow lines such as these. Similar
 7  component structure is evident in the OVI {lambda}1031.93 and
 7  {lambda}1037.62 profiles, and the similarity of the component structure
 7  favors a multicomponent fit. However, the OVI profiles are moderately noisy.
 7  We flag these measurements with a colon, because while three components are
 7  suggested by the OVI data, better S/N is needed to robustly establish that
 7  three components are present. If we fit the OVI lines with a single
 7  component instead of the three-component fit listed in Table 3, we obtain
 7  b(O^vi^)=82{+/-}13km/s and logN(O^vi^)=14.06{+/-}0.05 for the single line.
 8 3C 351.0, z_abs_=0.22111. An archival FUSE spectrum of 3C 351.0 shows that
 8  this is an optically thick Lyman limit absorber with N(H^i^)>10^17^/cm2. The
 8  Ly{alpha} profile is strongly saturated but shows complex structure at the
 8  edges of the profile. This structure could be partly due to damping wings,
 8  but this profile structure cannot be unambiguously attributed to damping
 8  wings. Consequently, the HI column density is highly uncertain. The
 8  OVI {lambda}1031.93 line is severely blended with the Galactic
 8  SiII {lambda}1260.42 line, and consequently, {lambda}1031.93 cannot be
 8  measured. However, many metal lines are detected at the redshift of this
 8  strong Lyman limit system including transitions of CII, NII, SiII, SiIII,
 8  and SiIV. The OVI {lambda}1037.62 line is identified based on its alignment
 8  with the other metals at this redshift. All available HI lines are strongly
 8  saturated at the velocities of the metal lines, so the degree of alignment
 8  of the OVI and HI lines cannot be evaluated. However, analysis of the low-
 8  and high-ionization metals lines indicates the presence of multiple phases,
 8  so this system is classified as a complex absorber.
 9 H1821+643, z_abs_=0.02438. HI Ly{beta} is mildly blended with a Galactic H2
 9  absorption line (see K. R. Sembach et al. 2008, in preparation), but the
 9  Ly{beta} line is mostly free from the blend. The blended portion of
 9  Ly{beta} was excluded from the fit. OVI {lambda}1037.62 is lost in a
 9  blend with Milky Way FeII and H2 absorption. The OVI identification is
 9  based on the precise alignment of OVI {lambda}1031.93 with Ly{alpha} and
 9  Ly{beta} lines at the same redshift.
10 H1821+643, z_abs_=0.12143. Thom & Chen (2008ApJ...683...22T) challenge this
10  OVI identification, noting that "there is strong absorption at the OVI 1037
10  position, but no OVI 1031, which should be easily detected, given the
10  strength of the weaker supposed OVI 1037 line." However, as discussed in
10  detail by Tripp et al. (2001ApJ...563..724T), the OVI {lambda}1037.62 line
10  is significantly blended with strong HI Ly{delta} absorption from the
10  absorber at z_abs_=0.22496 (Fig. 2 in Tripp et al., 2001ApJ...563..724T),
10  and it appears that Thom & Chen (2008ApJ...683...22T) have not taken this
10  serious blend into account. Moreover, Thom & Chen base their conclusions on
10  the STIS data only, which have S/N<~3/pixel in this wavelength range,
10  whereas the FUSE observations we used have S/N>~13/pixel here. When the FUSE
10  data are employed and the Ly{delta} blend is accounted for, we find
10  compelling evidence supporting this system. Because of the strong blend, our
10  OVI measurements are based on the {lambda}1031.93 line alone. The
10  OVI {lambda}1031.93 line is detected at the 6.9{sigma} level in our data.
10  While the blend hampers confirmation based on the {lambda}1037.63 line, we
10  note that there are no other clear identifications for the 6.9{sigma} line
10  at the {lambda}1031.93 wavelength. This is not an HI Ly{alpha} line because
10  the redshift places the line blueward of the Ly{alpha} region, nor is it a
10  higher Lyman series HI line because corresponding strong HI lines would be
10  obvious in the STIS spectrum but are not evident.
11 H1821+643, z_abs_=0.21331. The OVI {lambda}1037.62 line is blended with weak,
11  high-velocity SII {lambda}1259.52 absorption from Milky Way gas (see Savage
11  et al., 1995ApJ...449..145S, and Tripp et al., 2003AJ....125.3122T for
11  information about the Galactic high-velocity gas toward H1821+643).
11  Comparison of the Galactic SII {lambda}1259.52 and {lambda}1253.81 lines
11  shows that there is excess optical depth in the {lambda}1259.52 line,
11  and the excess is consistent with the expected contribution from the
11  OVI {lambda}1037.62 line at z_abs_=0.21331 (based on the strength of
11  the unblended OVI {lambda}1031.93 line), which supports the identification
11  of OVI at this redshift. In addition, the OVI {lambda}1031.93 line is
11  aligned with HI Ly{alpha}, {beta}, and {gamma} lines at the same zabs.
12 HE 0226-4110, z_abs_=0.01747. The OVI {lambda}1037.62 line is blended with
12  OVI {lambda}787.71 at z_abs_=0.34035 (see Lehner et al., 2006ApJS..164....1L
12  and our Fig. 29). The OVI identification is favored based on the precise
12  alignment of the OVI {lambda}1031.93 candidate with an HI Ly{lamba} line
12  at the same redshift. We note that the comparison of the Na profiles in
12  Figure 18 does not show the absorption in the wings of the Ly{alpha}
12  line very clearly; Figure 29 more clearly shows how the HI line is
12  slightly broader than the OVI lines.
13 HE 0226-4110, z_abs_=0.20701. Detailed analysis of this system has been
13  presented by Savage et al. (2005ApJ...626..776S). The HI Ly{gamma} line is
13  recorded in both the STIS spectrum and the FUSE spectrum of HE 0226-4110.
13  The apparent component structure in the STIS recording of the Ly{gamma}
13  line is incompatible with the FUSE recording of Ly{gamma} and with the
13  other (higher) Lyman series lines (see Savage et al.,
13  2005ApJ...626..776S), and the STIS Ly{gamma} line was excluded from the fit.
14 HE 0226-4110, z_abs_=0.35525. A weaker line offset by +40 km/s is present
14  next to the main component that is clearly detected in the
14  OVI {lambda}1031.93 and {lambda}1037.62 profiles at this redshift. The
14  +40km/s feature does not appear to be OVI, because it is not confirmed by
14  the {lambda}1037.62 line. However, the +40km/s feature would be
14  relatively weak in the {lambda}1037.62 transition, and it could be
14  hidden by noise. Following Lehner et al. (2006ApJS..164....1L), we do
14  not include the 40km/s component in the OVI measurements; higher S/N
14  data are needed to establish the identity of this feature.
15 HE 0226-4110, z_abs_=0.42670. At this redshift, the Ly{alpha} is redshifted
15  beyond the long-wavelength cutoff of our STIS spectrum, and the Ly{beta}
15  line is not detected. The OVI identification is based on the good agreement
15  of the OVI1031.93 and {lambda}1037.62 profiles (see Lehner et al.,
15  2006ApJS..164....1L). The OVI1031.93 line is partially affected by hot
15  pixels that were excluded from the fit. We note that identification of
15  this system depends critically on the STIS warm-hot pixel correction
15  algorithm. If we turn off the hot-pixel repair algorithm, we find
15  that the OVI1037.62 line is largely filled in by warm pixels. It is
15  important to obtain future observations of this system with COS in
15  order to test the reliability of the identification and to expand
15  the utility of this system with additional information (e.g.,
15  better HI absorption constraints).
16 HE 0226-4110, z_abs_=0.49246. This complex, multispecies system has been
16  analyzed in detail by Ganguly et al. (2006ApJ...645..868G). The OVI
16  component at v=0km/s is uncertain due to substantial saturation. The
16  OVI1037.62 line is partially blended with Galactic CIV (see Fox et al.,
16  2005ApJ...630..332F), but the distinctive component structure seen in the
16  OVI1031.93 profile can be clearly recognized in the in the {lambda}1037.62
16  profile as well (see Ganguly et al., 2006ApJ...645..868G), so the
16  identification is secure, and the weaker OVI components are well-constrained
16  by the OVI1031.93 line.
17 HS 0624+6907, z_abs_=0.33979. HI Ly{beta} is mildly blended with an unrelated
17  line. However, most of the Ly{beta} profile is free from the blend, and the
17  unblended portion of Ly{beta} was included in the fit. As shown in the left
17  panel of Figure 39, hot pixel features are present within the OVI1031.93
17  line and at the red edge of the OVI1037.62 profile. Fortunately, in this
17  case the QSO was observed on two different dates (in 2002 January and
17  February; see Table 1), and the position of the spectrum on the detector was
17  shifted between these two dates. Inspection of the data reveals that the hot
17  pixel features are only present in the 2002 February data. As shown in the
17  right panel of Figure 39, by masking and rejecting the affected hot pixels
17  in the February data, we can suppress this problem with a minimal loss of
17  the S/N.
18 HS 0624+6907, z_abs_=0.37053. In this proximate absorber, HI Ly{alpha} is not
18  detected despite good S/N (see Fig. 7). As shown in Figure 7, the OVI
18  identification is quite secure; both lines of the OVI doublet show multiple
18  components and are in excellent agreement.
19 PG 0953+415, z_abs_=0.06807. We have carried out extensive investigations of
19  this absorber in previous papers (Savage et al., 2002ApJ...564..631S; Tripp
19  et al., 2006ApJ...643L..77T). Comparison of the OVI1031.93 and
19  {lambda}1037.62 lines indicates moderate saturation, and application of
19  the method of Jenkins (1996ApJ...471..292J) indicates that N(OVI) could
19  be 0.25dex higher.
20 PG 0953+415, z_abs_=0.14231. The Ly{beta} profile is partially blended with
20  an HI Ly{delta} line from z_abs_=0.23351. The blended part of the Ly{beta}
20  line was not used in the fit. However, the Ly{alpha} line has a complex
20  profile with many components (see Tripp & Savage, 2000ApJ...542...42T),
20  and the unblended portion of the Ly{beta} line provides useful
20  constraints for the fit.
21 PG 0953+415, z_abs_=0.22974. In this proximate absorber of PG 0953+415,
21  HI Ly{alpha} is not detected despite good S/N. The OVI identification is
21  based on the good agreement of the OVI lines over a large portion of both
21  profiles; the {lambda}1031.93 and 1037.62 profiles agree well over ~20pix
21  between v=-30 and 40km/s. However, the OVI profiles are discrepant at
21  v<-30km/s. While this discrepancy has the appearance of a hot pixel feature,
21  comparison of the data from 1998 December 4 and 11 shows the same profile
21  structure. The location of the spectrum on the detector was shifted between
21  1998 December 4 and 11, so this discrepancy cannot be due to hot pixels. We
21  conclude that the OVI1031.93 line is blended with an unrelated Ly{alpha}
21  line on the blue side of the profile. This part of the {lambda}1031.93
21  profile was excluded from the Voigt profile fit.
22 PG 0953+415, z_abs_=0.23351. A Ly{delta} line with approximately correct
22  strength is detected at this redshift, but was not used in the fit due to
22  blending with Ly{beta} from the complex, multicomponent absorber at
22  z_abs_=0.14231
23 PG 1116+215, z_abs_=0.05927. Both lines of the OVI doublet are detected and
23  in excellent agreement at v=0km/s (see Fig. 29). The OVI1031.93 line is
23  also detected at v=-84km/s, but the OVI1037.62 line is not significantly
23  detected at that velocity. A small portion of the detected OVI1031.93
23  component at v=-84km/s is blended with Galactic H2 (see Sembach et al.,
23  2004, Cat. <J/ApJS/155/351>). The v=-84km/s component is also identified as
23  OVI, based on the good agreement of the OVI and HI Ly{alpha} line shapes
23  at v=-84km/s (not including the portion blended with H2, which was also
23  excluded from the fit), as shown in Figure 18.
24 PG 1116+215, z_abs_=0.13849. The OVI doublet at this redshift is detected
24  with both FUSE and STIS (see Sembach et al., 2004, Cat. <J/ApJS/155/351>).
24  The fit reported here is based on the STIS data. This system, which has a
24  high HI column and is detected in many Lyman series lines, has been
24  analyzed in detail by Sembach et al. (2004, Cat. <J/ApJS/155/351>). The
24  OVI lines are aligned with the HI lines, but analysis of the many
24  low-ionization metal absorption lines detected in this system clearly
24  establishes that this is a complex multiphase absorber (see Sembach
24  et al., 2004, Cat. <J/ApJS/155/351>).
25 PG 1116+215, z_abs_=0.16553. The HI components at v=-12, 143, 170, and
25  342km/s are well-constrained by the detected absorption lines. Additional
25  absorption is clearly and significantly detected in other velocity ranges
25  in the Ly{alpha} profile, e.g., at v~70km/s, and components were added to
25  the fit to account for this additional absorption. However, these
25  components are not well-constrained due to blending with the adjacent
25  features.
26 PG 1116+215, z_abs_=0.17340. The OVI1031.93 line is detected at 5.8{sigma}
26  significance and is well-aligned with an HI line at the same redshift. In
26  addition, the OVI1037.62 line is detected with the expected wavelength and
26  relative strength at 2.9{sigma}.
27 PG 1216+069, z_abs_=0.12360. The Ly{alpha} profile has good S/N and shows
27  clear inflections and asymmetries that reveal the complicated component
27  structure including at least eight components; four of the components are
27  clearly evident in the Ly{beta} profile as well (see Fig. 4). However, all
27  of the components are either strong and significantly saturated (in both
27  Ly{alpha} and Ly{beta}) or are highly blended with adjacent strong
27  components. Moreover, some of the saturated components show that there are
27  errors in the flux zero level of the Ly{alpha} line, and the Ly{beta} line
27  is relatively noisy. Ly{gamma} lines are also evident (see Tripp et al.,
27  2005ApJ...619..714T), but the Ly{gamma} data are too noisy to usefully
27  constrain the fits. The HI component parameters are highly uncertain due to
27  these combined problems. Both lines of the OVI doublet are strong and are
27  clearly detected with component structure similar to the Ly{beta}
27  components. The apparent column density profiles of the OVI doublet are in
27  good agreement (see Fig. 4), which suggests that the OVI lines are not badly
27  saturated. However, the OVI profiles are also relatively noisy.
28 PG 1216+069, z_abs_=0.26768. The OVI1037.62 line cannot be measured, because
28  it is lost in a blend with a strongly saturated HI Ly{beta} absorption line
28  from z_abs_=0.28189. As shown in Figure 40, the OVI identification at
28  z_abs_=0.26768 is based on the alignment of OVI1031.93 with Ly{alpha} and
28  Ly{beta} at the same redshift. At this redshift, initial inspection
28  identified a candidate CIII977.02 line that is somewhat blended with
28  NV1238.82 absorption from the Milky Way. The CIII candidate cannot be a
28  second component of Galactic NV absorption, because it is not evident in the
28  profile of the other line of the NV doublet. However, closer inspection
28  reveals that this line is not CIII, but rather is the HI Ly{gamma} line from
28  the strong HI system at z_abs_=0.27353, so we can only place an upper limit
28  on CIII absorption at this redshift.
29 PG 1259+593, z_abs_=0.04637. Only the OVI1031.93 line is detected at this
29  redshift (OVI1037.62 is redshifted into a relatively noisy region of the
29  FUSE spectrum that is only recorded by the SiC channels). Nevertheless,
29  the identification is secure, because the OVI1031.93 profile has a
29  distinctive two-component structure that matches the component structure
29  seen in the CIII and CIV lines detected at the same redshift. An
29  unrelated OIV787.71 line is located near the Ly{gamma} profile; the
29  region affected by the OIV feature was excluded from the fit.
30 PG 1259+593, z_abs_=0.21949. Several components due to Galactic SII1250.58
30  are located on the blue side of the Ly{beta} line (see Richter et al.,
30  2004ApJS..153..165R); the velocity range affected by the Milky Way SII
30  absorption was excluded from the fit.
31 PG 1259+593, z_abs_=0.31972. The OVI1037.62 line is quite weak and mildly
31  blended with high-velocity NiII absorption from the Milky Way (see Richter
31  et al., 2004ApJS..153..165R). As shown in the right panels of Figure 40,
31  the OVI1031.93 line is clearly detected, and an absorption feature with
31  the right relative strength (compared to {lambda}1031.93) is present at
31  the expected wavelength of OVI1037.62, but because it is weak and
31  blended with Galactic NiII, our fit is based on the OVI1031.93 line only.
32 PG 1444+407, z_abs_=0.22032. The HI Ly{alpha} and OVI1031.93, {lambda}1037.62
32  lines for this absorption system are shown in the left panels of Figure 41.
32  The OVI1031.93 line at z_abs_=0.22032 is blended with the Galactic
32  SII1259.52 line. However, comparison of the Galactic SII1253.81 and
32  SII1259.52 profiles shows a clear and significant excess of absorption at
32  the expected velocity of OVI1031.93 at z_abs_=0.22032. Moreover, the
32  excess absorption has precisely the expected strength compared to the
32  (unblended) OVI1037.62 line, as can be seen in the comparison of the
32  OVI^Na(v)^ profiles shown in Figure 41. Only the unblended portion of
32  OVI1031.93 was included in the fit.
33 PG 1444+407, z_abs_=0.26738. The Ly{beta} line is slightly blended with an
33  unrelated weak line; the region affected by the blend was excluded from the
33  fit. More importantly, the Ly{alpha} line is located at the peak of the
33  broad Ly{alpha} emission line, and this introduces significant continuum
33  placement uncertainty. We note that a broad and shallow feature is
33  located just blueward of the Ly{alpha} line. This feature could be due
33  to additional weak HI absorption, but its significance is highly
33  dependent on the uncertain continuum placement. Consequently, we did
33  not include the broad, shallow feature in the fit.
34 PHL 1811, z_abs_=0.07765. Only the stronger OVI1031.93 line is detected
34  at >3{sigma} significance. However, many metals are detected at this
34  redshift including CII1334.53, SiII1260.42, CIII977.02, CIV1548.20,
34  (marginal) SiIV1393.76, and several HI Lyman series lines. The strongest
34  CIV component shows a ~-25km/s offset from the lower ionization metals,
34  but the OVI1031.93 line is aligned with the CIV1548.20 transition.
35 PHL 1811, z_abs_=0.15786. The Ly{alpha} profile at z_abs_=0.15786 is blended
35  with OI1302.17 absorption from the Lyman-limit system at z_abs_=0.08092 (see
35  Jenkins et al., 2005ApJ...623..767J; see also our Fig. 35). The narrow core
35  in this blend is predominantly due to the Lyman-limit OI line. However,
35  close inspection of this profile (see Fig. 38) reveals weak component
35  absorption straddling the narrow core on the short- and long-wavelength
35  sides. We cannot corroborate that the weaker components are also OI;
35  similar component structure is not clearly evident in the other profiles of
35  low-ionization metals in this Lyman-limit absorber, which suggests that
35  these weaker components could be unrelated to the OI and could be Ly{alpha}
35  at z_abs_=0.15786. However, better S/N data are needed to reliably determine
35  the origin of the weak components and to accurately deblend and measure
35  their parameters, so we flag this absorption with a colon in Table 3 to
35  reflect the substantial uncertainty of this Ly{alpha} case.
36 PHL 1811, z_abs_=0.17650. At this redshift, OVI1031.93 falls in the
36  saturated core of the Milky Way damped Ly{alpha} line. The OVI1037.62 line
36  is identified based on its alignment with multiple Lyman series lines, and
36  this is supported by the detection of CIII977.02 and SiIII1206.5 in this
36  absorber. However, we note that the OVI is offset by ~-25km/s compared to
36  the CIII and SiIII lines.
37 PKS 0312-770, z_abs_=0.15890. OVI1031.93 is detected at 5.6{sigma}
37  significance, and the corresponding OVI1037.62 line is also detected, but
37  only at 2.4{sigma} significance. As shown in Figure 41, the N_a_(v)
37  profiles of the two lines of the OVI doublet are in good agreement, and
37  the marginal detection of the {lambda}1037.62 line supports the OVI
37  identification. The OVI identification is also supported by the precise
37  alignment of the OVI1031.93 line with HILy{alpha} at the same redshift.
38 PKS 0312-770, z_abs_=0.19827. OVI1031.93 is detected at 4.7{sigma}
38  significance, but the OVI1037.62 line is not significantly detected in the
38  data at full resolution. However, as shown in the left panels of Figure 42,
38  if we mildly bin the data to 7km/s pixels to improve the S/N, we find a
38  feature in the spectrum at the expected wavelength of the 1037.62 line
38  that is fully consistent with the detected OVI1031.93 line [compare the
38  N_a_(v) profiles shown in the bottom left panel of Fig. 42]. The OVI
38  identification is bolstered by the close alignment of OVI1031.93 with
38  Ly{alpha} and Ly{beta} at the same redshift (see Fig. 42, left).
39 PKS 0312-770, z_abs_=0.20266. All accessible HI lines in the STIS bandpass
39  (Ly{alpha}, Ly{beta}, Ly{gamma}) are strong and highly saturated. Moreover,
39  the Ly{beta} and Ly{gamma} profiles show complex component structure with
39  at least five distinct components. Since most of the HI components are
39  black in the line cores, the HI profile parameters are poorly constrained,
39  and we have not attempted to fit the HI lines. Brief inspection of an
39  archival FUSE spectrum reveals that this is an optically thick Lyman limit
39  absorber. While the HI component properties are poorly constrained,
39  comparison of the low- and high-ionization metal lines reveals that this is
39  a complex multiphase system (see Section 2.4.1 and Fig. 3). In addition, the
39  individual components are spread over a large velocity range with low- and
39  high-ionization components detected at velocities ranging from -204 to
39  +135km/s.
40 PKS 0405-123, z_abs_=0.16692. Many Lyman series lines are available for
40  constraining the HI column density at this redshift, and the absence of
40  strong Lyman limit absorption places a firm upper limit on the total HI
40  column density (Prochaska et al., 2004, Cat. <J/ApJ/617/718>).
40  Nevertheless, the parameters of the individual HI components are highly
40  uncertain in this system. The Lyman series lines clearly require a
40  multicomponent fit, but the close spacing and blending of the
40  components causes the component parameter uncertainties to be substantial.
40  Thus, the degree of alignment of the HI and OVI components is highly
40  uncertain. Nevertheless, comparison of the low- and high-ionization metal
40  lines shows that this is a complex multiphase case (see Chen & Prochaska,
40  2000ApJ...543L...9C).
41 PKS 0405-123, z_abs_=0.36156. The HI Ly{alpha} and OVI1031.93,
41  {lambda}1037.62 absorption profiles and apparent column densities are
41  compared in the right panels of Figure 42. The OVIN_a_(v) profiles are seen
41  to be in reasonable agreement. We note that the OVI1037.62 profile shows a
41  small excess of absorption on the blue side compared to the 1031.93 line;
41  this could be due to blending with an unrelated line (several unrelated
41  lines are readily apparent in the vicinity of {lambda}1037.62), but this
41  could also simply be a noise feature. An inflection is also evident in the
41  Ly{alpha} profile at the velocity of the OVI lines. It is interesting to
41  note that this system has a relatively high N(HI) and is detected in several
41  Lyman series lines, but no OVI absorption is evident at the velocity of the
41  main component where the strong HI lines are found (see Fig. 42). Instead,
41  the OVI is in the wing of the profile near the weak HI inflection component.
41  Several other absorbers show similar offsets between the strong main HI
41  absorption component and the OVI lines (see Fig. 18).
42 PKS 0405-123, z_abs_=0.36335. The OVI1031.93 line is detected at the
42  4.5{sigma} level, but OVI1037.62 is measured at only 1.7{sigma}
42  significance. However, the apparent column density profiles of the two OVI
42  lines agree precisely; the wavelength separation and relative strength of
42  the lines makes the OVI identification compelling. The corresponding HI
42  absorption is relatively weak, and moreover, the HI Ly{alpha} line is
42  strongly blended with Galactic CI and CI* absorption lines from the CI1656
42  multiplet. We can see that HI Ly{alpha} absorption is present at this
42  redshift because the Galactic CI*1657.38 line is clearly too strong compared
42  to other CI* lines in the PKS 0405-123 spectrum. However, the HI Ly{alpha}
42  line is difficult to measure reliably due to this strong blending with
42  Galactic CI*1657.38.
43 PKS 0405-123, z_abs_=0.49501. In this case, the HI Ly{alpha} line is
43  redshifted beyond the long wavelength cutoff of the STIS spectrum, but
43  Ly{beta} and Ly{gamma} are detected at the 4.1{sigma} and 2.0{sigma}
43  levels, respectively. This absorber is detected in a variety of metals
43  (Prochaska et al., 2004, Cat. <J/ApJ/617/718>), and many of the metal
43  profiles, including the OVI lines, show evidence of multiple components.
43  However, it should be noted that there is a discrepancy evident in one
43  of the components of the OVI1031.93,1037.62 lines. To show this, we
43  compare the apparent column density profiles of the OVI1031.93,
43  1037.62 lines in Figure 43. We also compare the OVI1031.93 N_a_(v) profile
43  to those of CIII977.02, OIV787.71, and OV629.73 in Figure 43. The
43  OVI1031.93, 1037.62 profiles agree well in the stronger component at
43  v~0km/s.Looking closely at the strongest component, we can see that the
43  profile is asymmetric with a sharp edge on the red side and a more gradually
43  decreasing apparent column density on the blue side. This asymmetry suggests
43  that the main feature is a blend of two components, and this is corroborated
43  by the CIII and OIV profiles, which also show an extra component on the blue
43  side. A third component is evident at v~70km/s in the CIII and OIV
43  transitions. This component appears to be present in the OV and OVI profiles
43  as well, but at a somewhat lower velocity (v=57km/s). However, there is a
43  2-3 pixel offset between the peak of the OVI1031.93 and {lambda}1037.62
43  profiles in the v=57km/s component (see Fig. 43), in contrast to the v~0km/s
43  component in which the OVI profiles agree well. Initially, this offset
43  appeared to be due to a hot pixel feature falling in the middle of the
43  OVI1031.93 line, but this cannot be the cause, because the PKS 0405-123
43  observations were obtained on two separate occasions (see Table 1), and the
43  spectrum detector position was shifted between the two visits. The same
43  component structure is evident in the OVI profiles extracted separately from
43  the two visits, so this problem is not due to a hot-pixel feature. Apart
43  from this small offset, the N_a_(v) profiles of the OVI lines at v=57km/s
43  appear to be quite consistent; the relative strength and shape of the two
43  lines are in agreement. This suggests that the offset could be caused by an
43  instrumental calibration problem. For example, the STIS geometric distortion
43  can cause offsets of this magnitude if not properly corrected (Walsh et al.,
43  2001, Instrument Science Report STIS 2001-02; Maiz-Apellaniz & Ubeda,
43  2004, Instrument Science Report STIS 2004-01), and some problems with the
43  distortion correction have been noted (Maiz-Apellaniz & Ubeda, 2004,
43  Instrument Science Report STIS 2004-01). Evidence of wavelength calibration
43  problems have also been noted when comparing lines that should have
43  identical component structure (e.g., Jenkins & Tripp, 2001ApJS..137..297J;
43  Tripp et al., , 2005ApJ...619..714T). While we have found these problems to
43  be relatively rare, the stability of the STIS geometric distortion
43  correction has not been studied systematically, and it remains possible
43  that the discrepancy in the v=57km/s component is caused by a calibration
43  problem such as this. Nevertheless, we flag the OVI lines in the v=57km/s
43  component with a colon in Table 3, because of this disagreement, and we
43  treat this component as an insecure identification. With the new COS, it
43  will be possible to reobserve PKS 0405-123 to determine if the offset is
43  due to a STIS instrumental problem. The OVI1037.62 line at this redshift
43  falls close to Galactic CIV1550.78 (which is the source of the extra
43  absorption evident at v<-40km/s in Fig. 43), but from the corresponding
43  Galactic CIV1548.20 line, we can see that the Milky Way C iv has little
43  impact on the redshifted OVI1037.62 profile. We give this system an
43  uncertain classification due to the insecure identification of the OVI
43  component at v=57km/s and the fact that the Ly{alpha} profile, which is
43  needed to detect low-N(HI) components, has not been observed at high
43  resolution.
44 PKS 1302-102, z_abs_=0.19159. The profile-fitting code obtains its best fit
44  to the HI lines with a narrow and deep component superimposed on a broad and
44  shallow component at the same velocity. The absorption that requires the
44  broad component could be nearly as well-fit with multiple narrower
44  components, and higher S/N is required to break this degeneracy. Only
44  OVI1031.93 is detected, but the OVI line is precisely aligned with the HI
44  lines, and CIII977.02 and SiIII1206.50 are also detected with good
44  significance at this redshift.
45 PKS 1302-102, z_abs_=0.22563. Some parts of the Ly{beta} profile are affected
45  by blends; those regions of Ly{beta} were excluded from the fit. The
45  OVI1037.62 line is also located next to an unrelated strong line. However,
45  the OVI1031.93 and 1037.62 N_a_(v) profiles show excellent agreement over
45  most of the velocity range where {lambda}1031.93 is clearly detected, and
45  the velocity range of the OVI1037.62 line that is affected by the adjacent
45  interloper was excluded from the fit.
46 TON 28, z_abs_=0.13783. The OVI1037.62 line is only detected at the
46  2.2{sigma} level, but as shown in Figure 44, its wavelength and relative
46  strength agree with the corresponding OVI1031.93 line. In addition, we can
46  see from this figure that the OVI lines are aligned with a clearly detected
46  component in the corresponding HI Ly{alpha} line.
47 TON 28, z_abs_=0.27340. The HI Ly{alpha} line is strongly blended with Milky
47  Way CIV1548.20. However, the HI at this redshift is securely identified and
47  measured based on the well-detected Ly{beta} and Ly{gamma} lines, and
47  comparison of the Galactic CIV1548.20 and 1550.78 apparent column density
47  profiles verifies that substantial extra optical depth (due to the blended
47  Ly{alpha} line) is present in the Galactic CIV1548.20 profile.