Manual/AM1_accuracy.html at main · openmopac/Manual · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
<html>

<head>

<!-- Google tag (gtag.js) -->
<script async src="https://www.googletagmanager.com/gtag/js?id=G-60EDW3Y6ET"></script>
<script>
  window.dataLayer = window.dataLayer || [];
  function gtag(){dataLayer.push(arguments);}
  gtag('js', new Date());

  gtag('config', 'G-60EDW3Y6ET');
</script>


<title>Accuracy of AM1</title>
</head>
<body link="#0000ff" vlink="#800080">


<h1>Accuracy of AM1 </h1>
<p>AM1 is a distinct improvement over MNDO, in that the overall accuracy is considerably improved. Specific improvements are: </p>
<dl>
<dt>1. The strength of the hydrogen bond in the water dimer is 5.5 kcal/mol, in accordance with experiment. </dt>
<dt>2. Activation barriers for reaction are markedly better than those of MNDO. </dt>
<dt>3. Hypervalent phosphorus compounds are considerably improved relative to MNDO. </dt>
<dt>4. In general, errors in <font face="Symbol">D</font>H<sub>f
</sub>obtained using AM1 are about 40% less than those given by MNDO. </dt>
</dl>
<p>Unfortunately, with this improvement a few deficiencies were introduced. The most important of these are: </p>
<dl>
<dt>1. <a name="BM24227">&nbsp;</a> AM1 phosphorus has a spurious and very sharp potential barrier at 3.0 &Aring;ngstroms. The effect of this is to distort otherwise symmetric geometries and to introduce spurious activation barriers. A vivid example is given by P<sub>4</sub>O<sub>6</sub>, in which the notionally equal P-P bonds are predicted by AM1 to differ by 0.4 &Aring;. This is by far the most severe limitation of AM1. </dt>
<dt>2. Alkyl groups have a systematic error due to the heat of formation of the CH<sub>2</sub> fragment being too negative by about 2 kcal/mol. </dt>
<dt>3. Nitro compounds, although considerably improved relative to MNDO, are still systematically too positive in energy. </dt>
<dt>4. The lowest-energy conformer for ethanol is predicted to have a gauche C-C-O-H angle. High level <em>ab initio</em> calculations, and experimental evidence, suggests that the correct geometry is trans. This error was reported by Andreas Jabs, at Martin-Luther-Universitaet Halle-Wittenberg, Halle/Saale, Germany. </dt>
<dt>5. The peroxide bond is still systematically too short by about 0.17&Aring;. </dt>
<dt>6. UHF <font face="Symbol">D</font>H<sub>f</sub> for organic radicals are,
in general, too negative.&nbsp; If&nbsp; <font face="Symbol">D</font>H<sub>f</sub>
are important, use RHF instead of UHF.</dt>
</dl>
</body>
</html>