메뉴 건너뛰기




Volumn 110, Issue 21, 2006, Pages 6592-6601

Active thermochemical tables: Accurate enthalpy of formation of hydroperoxyl radical, HO 2

Author keywords

[No Author keywords available]

Indexed keywords

DEPROTONATION; EQUILIBRIUM CONSTANTS; PROTON AFFINITY; THERMOCHEMISTRY;

EID: 33745459479     PISSN: 10895639     EISSN: None     Source Type: Journal    
DOI: 10.1021/jp056311j     Document Type: Article
Times cited : (260)

References (93)
  • 12
    • 0003363438 scopus 로고    scopus 로고
    • Neutral Thermochemical Data
    • NIST Stand. Ref. Database Number 69; Linstrom, P. J., Mallard, W. G., Eds.; National Institute of Standards and Technology: Gaithersburg, MD, June
    • Afeefy, H. Y.; Liebman, J. F.; Stein, S. E. Neutral Thermochemical Data. In NIST Chemistry WebBook; NIST Stand. Ref. Database Number 69; Linstrom, P. J., Mallard, W. G., Eds.; National Institute of Standards and Technology: Gaithersburg, MD, June 2005 Release (http:// webbook.nist.gov).
    • (2005) NIST Chemistry WebBook
    • Afeefy, H.Y.1    Liebman, J.F.2    Stein, S.E.3
  • 16
    • 1642577314 scopus 로고
    • NSRDS-NBS 37; National Bureau of Standards, U.S. Department of Commerce: Washington, DC, including revised tables issued through 1979
    • Stull D. R.; Prophet, H. JANAF Thermochemical Tables, 2nd ed.; NSRDS-NBS 37; National Bureau of Standards, U.S. Department of Commerce: Washington, DC, 1971, including revised tables issued through 1979.
    • (1971) JANAF Thermochemical Tables, 2nd Ed.
    • Stull, D.R.1    Prophet, H.2
  • 28
    • 33745465743 scopus 로고    scopus 로고
    • note
    • 298(ROO-H) are the same; see also refs 32 and 35.
  • 43
    • 33745437577 scopus 로고    scopus 로고
    • Private communication to B. Ruscic
    • Karkach, S. P. Private communication to B. Ruscic, 1997;
    • (1997)
    • Karkach, S.P.1
  • 52
    • 33745436102 scopus 로고    scopus 로고
    • note
    • Expressing uncertainties as 95% confidence intervals is the accepted standard in thermochemistry, originally introduced by Rossini in ref 49 and subsequently embraced by virtually all thermochemical compilations. The requirement that all determinations have (at least roughly) a commensurate level of expression of uncertainties (i.e., coverage factor) is a conditio sine qua non when comparing competing determinations, such as in a network. A similar condition should be met when, for example, enthalpies of formation from various sources are compared. This is, indeed, routinely the case when the enthalpies originate from experimental papers (which tend to adhere to the 95% confidence intervals custom) or compilations but, contrary to expectations, appears not to be generally the case for theoretical results. In many papers, the theoretical results are reported with an explicit or implied uncertainty that is based on MAD (mean absolute deviation), which severely (by a factor of 2-3) underestimates the 95% confidence intervals, thus creating an unfair basis for comparison with experimental data.
  • 80
    • 33745442637 scopus 로고    scopus 로고
    • note
    • -1 is obtained.
  • 81
    • 33745454851 scopus 로고    scopus 로고
    • note
    • If all of the initial uncertainties (as originally reported or as reevaluated during the critical data analysis step that occurs prior to inserting the determination into the TN) were indeed 95% uncertainty intervals, then the ATcT statistical analysis would generally find that ∼95% of the uncertainties are large enough to completely contain the difference between the measured quantity and the same quantity as recalculated from the final ATcT enthalpies (i.e., the "error" of the original determination), while ∼5% of the uncertainties would not be sufficiently large to contain this difference. The ∼5% set of uncertainties would be composed roughly as follows: ∼3% of uncertainties would be able to contain the "terror" if they were augmented by 20% or less, ∼1.6% would need to be augmented by 20-50%, ∼0.4% would need to be augmented between 50 and 100%, and about 0.02% would need to be augmented by a factor between 2 and 4. Corrections larger than a factor of 4 would be exceedingly rare. Such ideal behavior almost never takes place in practice, attesting to the fact that a significant fraction of uncertainties reported in the literature are "optimistic". In fact, from the number of "optimistic" uncertainties uncovered and corrected within the global ATcT analysis of the ∼6400 determinations present in the current version of C(A)TN, one obtains a hint that the average initial uncertainty (many of which have been already expanded during the critical evaluation step above the originally reported size by a significant factor) roughly corresponds only to an ∼86% confidence interval (approximately equivalent to 1.5 std dev, as opposed to the desired coverage factor of ∼2 std dev). Constructing the network using the unconnected uncertainties (as claimed in the original literature) would have resulted in a substantially higher percentage of "optimistic" uncertainties. The picture is even murkier once the underlying distribution is examined in more detail: A substantial fraction of the "optimistic" uncertainties correspond to high-magnitude outliers, far beyond the statistical expectation. Some of these probably correspond to hidden systematic errors. The unusual ability of the TN approach of ATcT to uncover at least some of these systematic errors will be discussed in future publications on ATcT.
  • 92
    • 84858902042 scopus 로고    scopus 로고
    • http://cmcs.org.
  • 93
    • 84858904300 scopus 로고    scopus 로고
    • http://www.iupac.org/projects/2003/2003-024-1-100.html.


* 이 정보는 Elsevier사의 SCOPUS DB에서 KISTI가 분석하여 추출한 것입니다.