The pure rotational spectrum of the ground state of NH3/ND3
      and the Stark effect
    NH3
    The data files given here were
      constructed by fitting to the line positions and intensities in
      the HITRAN 2008 (http://www.hitran.com)
      database. The fit was restricted to J < 16, and is sufficient to reproduce the
      line positions to better than 0.001 cm-1 and the
      intensities to better than 10-20  cm2 cm-1 Molecule-1.
      There are two points that are slightly non-standard:
    
      - The point group used is D3h, rather than C3v.
        This is necessary to account for the inversion doubling in the
        ground state, which appears in the data file as two vibrational
        states:
 
      
        - s/0+ with A1'
          symmetry
 
        - a/0- with A2"
          symmetry
 
      
      - To fit the K = 3n
        levels requires three perturbation
        operators:
 
      
        - <0+|J+-^6|0+>
          = J+6 + J-6 
 
        -  <0+|J^2J+-^6|0+>
          = J2(J+6
          + J-6)
 
        - <0+|J+-^12|0+>
          = J+12 + J-12 
 
      
    
    Two separate data files are
      provided, with the required constants. The first is 
nh3x0.pgo which has the two
      inversion doublets in separate manifolds (s and a), with a
      transition moment acting between the two manifolds. This is a good
      starting point for simulating electronic spectra, where the
      transitions will typically be from one inversion doublet or other.
    
 
    The Stark Effect
    
    The second file, 
nh3x0one.pgo, has the same
      constants in, but with the two inversion doublets in the same
      manifold. The resulting energy levels will be the same, but as the
      dipole moment now acts within the manifold, the Stark effect in
      the ground state can be simulated with this file. To give quick
      calculation times Jmax has been set to 5 for this file. The 
energy level plot window can then be
      used to produce a plot of energy against field, as for example:
    
 
    
    This shows the K = 1, M = 1 levels show a strong
      Stark effect, but the other levels show a rather weaker effect.
      The "Summary" button gives the following information: 
       M  Sym   #    g Population Name J K (kl) Sym M         Energy Linear         Dipole     Err  Quadratic       Err  Two Level Delta     C              Dipole2    Err
   0    -   1    4 .660533316 a  0  0 A2"  0             0.7933 -5.20523347e-9 -.0309983 15.8% -1.0409798e-15 .02 %
   0    -   2    2 .036129526 s  1  1 E"  0             16.1730 -1.47243024e-9 -.0087686 15.8% -2.9448374e-16 .002%
   0    -   3    2 .032246486 a  1  1 E'  0             16.9632 -1.53137719e-9 -.0091197 15.8% -3.0627282e-16 .002%
   0    -   4    4 .042329414 s  1  0 A2'  0            19.8899 3.241105275e-9 .01930147 15.8% 6.48153456e-16 .033%
   1    -   1    2 .036129526 s  1  1 E"  1             16.1730 -7.02797997e-8 -.4185312 13.9% -1.3786935e-14 5.18% 16.583449 .82098935 -1.25126674e-7 -.7451561 .013%
   1    -   2    2 .032246486 a  1  1 E'  1             16.9632 6.802667704e-8 .40511341 13.9% 1.33363101e-14 5.36% 16.583784 .75886528 1.190970869e-7 .70924861 .015%
   1    -   3    4 .042329414 s  1  0 A2'  1            19.8899 -1.47290548e-9 -.0087715 15.8% -2.9457878e-16 .002%
    This indicates that the K
    = 1 levels are intermediate between a first and second order Stark
    effect, and the others are essentially showing a second order
    (quadratic) Stark effect. See External
      Fields - The Zeeman and Stark Effects for more information.
    ND3
    A similar set of data files for ND3 is given below. The
    data sources are L Fusina and S N Murzin, J. Molec. Spectrosc. 167 464 (1994), L H Coudert and
    E Roueff, Astron. Astrophys. 449
    855 (2006) and the ND3
      linelist from the Cologne Database for
      Molecular Spectroscopy (CDMS).
    
      - 
        
nd3x0.pgo,
          similar to the nh3x0.pgo file above. This was derived by from
          an unweighted fit to all lines in the CDMS linelist with 
J < 16, which gave an
          average error of 7×10
-5 cm
-1 and a
          maximum error of 0.003 cm
-1. This has the two
          ground state inversion doublets in different manifolds, giving
          a simpler description than in the next file. 
 
       
       
      - 
        
nd3x0one.pgo,
          equivalent to the nh3x0one.pgo file above, with both ground
          state inversion doublets in the same manifold to allow the
          Stark effect to be simulated. This uses exactly the
          Hamiltonian and constants used in the Fusina and Murzin paper.
          This includes perturbations acting between the two inversion
          doublets, though note that the sign of the α
J and α
K constants must
          be changed in 
PGOPHER
          to give consistent matrix elements. The line positions agree
          with the CDMS linelist to < 0.006 cm
-1 (
<
            0.0005 cm-1 for J < 15). The operators mixing states
            are: 
       
      
        - <0-|JzJ+-^3|0+>
          = [J+3 + J-3,Jz]+
          (term in α)
 
        -  <0+|J^2JzJ+-^3|0+>
          = J2[J+3
          + J-3,Jz]+ (-term
          in αJ)
 
        - <0+|J+-^3|0+>
          = J+3 + J-3 (-term
          in αK) 
 
      
      - 
        
nd3x0hyp.pgo
          which has the hyperfine structure due to the nitrogen nucleus
          added to the simulation, giving a linelist consistent with the
          Coudert and Roueff paper.