Skip to main content
Journal of Research of the National Bureau of Standards. Section A, Physics and Chemistry logoLink to Journal of Research of the National Bureau of Standards. Section A, Physics and Chemistry
. 1968 Mar-Apr;72A(2):207–216. doi: 10.6028/jres.072A.020

Thermodynamic Properties of Ammonia as an Ideal Gas

Lester Haar 1
PMCID: PMC6640604  PMID: 31824090

Abstract

Thermodynamic functions for ammonia as an ideal gas at one atmosphere pressure have been evaluated. The contribution of the highly anharmonic out-of-plane vibrational mode, including its large coupling with rotation and its coupling with the other vibrational modes, is considered in detail. Tables of CP°/R, (H°E0°)/RT, (E0°G°)/RT, and S°/R have been calculated at closely spaced intervals from 50 to 5000 °K within an overall uncertainty of less than 0.1 percent at 1000 °K.

Keywords: Ammonia, ideal gas, thermodynamic functions

1. Introduction

This paper is part of a program originating with the Office of Standard Reference Data of the National Bureau of Standards to obtain accurate thermodynamic properties for ammonia. Subsequent papers of this program will include real-gas corrections to the ideal-gas properties contained herein.

In this paper the ideal-gas thermodynamic properties of ammonia in the ground electronic state are calculated at closely spaced temperature intervals from 50 to 5000 °K. The calculations include the contributions of ordinary vibrational anharmonicity and vibrational-rotational coupling together with rotational stretching and rotational quantum effects. Particular attention is given to the treatment of the rotational and vibrational anharmonic effects due to the molecular inversion. To this end a method is developed for calculating this contribution such that an overall accuracy for the thermodynamic properties of ammonia approaching that of the simplest diatomic molecules is obtained.

In the vibrational ground state ammonia is a pyramidal molecule with the nitrogen atom approximately 0.38 Å above the plane of the three hydrogen atoms. The distance between the nitrogen atom and each of the hydrogen atoms is approximately 1.0 Å. Thus the pyramid is relatively flat and the potential barrier to vibration of the nitrogen atom through the plane of the hydrogen atoms so as to invert the pyramid is low (approximately 2000 cm−1 [1]).1

Ammonia has six vibrational degrees of freedom. However, owing to symmetry there are only four normal vibrational modes, two of which have double degeneracies. The vibrational structure may thus be characterized by the six quantum numbers (v1v2v3l3v4l4), where the νi are the principal vibrational quantum numbers associated with the vibrational modes νi and the l characterize the small coupling of the degenerate v3 and ν4 asymmetric vibrational modes with the angular momentum of the molecule. The ν1 and v2 modes are symmetric, nondegenerate vibrations corresponding to stretching the N-H bond and spreading the pyramid, respectively. (The v2 mode is also referred to as the out-of-plane vibration.) The rotational structure for the ground state is that of a symmetric top, so that the rotational energy levels are characterized by the two quantum numbers J and K.

The v2 mode is the vibration associated with the inversion of the pyramid. Each vibrational state of this mode is split into two components, corresponding respectively to symmetric and antisymmetric eigen-functions. (In the nomenclature these are identified by the superscript “s” or “a” to the quantum number v2.) The spacing of these levels is highly anharmonic except for vibrational energies sufficiently higher or lower than the potential barrier to inversion. Thus the vibrational levels for the ground state of the v2 mode are only slightly split, of the order of a fraction of a wavenumber, but the splitting increases rapidly with increasing vibrational quantum number. At energy levels sufficiently elevated above the barrier (greater than about 3000 cm−1) the levels become uniformly spaced with a spacing of approximately half that of the v2 fundamental. Similarly, due to inversion the rotational structure undergoes a transition from that of a pyramid for the low values of the v2 quantum number to that of a planar molecule for the higher values, so that the rotational motion is also strongly coupled to the v2 mode. For a more detailed discussion of these effects, see the review by Herzberg [2] of the pyramidal XY3 molecule and Nielsen [3]. The energy structure of the other vibrational modes is similar to that of typical, slightly anharmonic rotating vibrators.

In section 2 the equations for the calculations are developed. This is followed in section 3 by a discussion of the molecular data, and in section 4 by a discussion of the calculated properties and an evaluation of their accuracy. Finally, a discussion is given in section 5 of some of the earlier work.

2. Partition Function

It is convenient to represent the rotational and vibrational structure by an expression for which the energy levels of the v2 mode and its coupling with the other degrees of freedom are written explicitly. Thus the energy levels in wavenumbers are written

T(v,J,K)=G(v)+F(v,J,K)+G2(v,J,K), (1)

where the primes denote that dependence on v2 is omitted. The first term in eq (1) describes the vibrational energy characterized by (v10v3l3v4l4) and in notation similar to that in Herzberg [2] is approximated by

G(v)=i'νiνi+i>j'xijvivj+i'xii(vi1)vi++g34l3l4+g33(l32v3)+g44(l42v4)+,i,j2, (2)

where the νi are fundamentals and the x and g are relatively small anharmonicities. The second term in eq (1) is approximated by

F(v,J,K)=BvJ(J+1)+(CvBv)K2DvJ'J2(J+1)2DvJK'J(J+1)K2DvK'K4+, (3)

where Cν is the rotational constant corresponding to rotation about the symmetry axis and Bν is the constant for axes perpendicular to the symmetry axis. The quantities DvJ, DvJK, and DvK are small coefficients associated with centrifugal distortion of the molecule. The subscript v to each of the coefficients of eq (3) indicate dependence or coupling with the vibrational motion. The primes indicate that dependence on v2 is omitted; that is, the rotational constants of eq (3) are evaluated for v2 equal to zero. The coupling of the rotational motion to vibrations other than the v2 mode is relatively small and is included here only to first order in the fundamentals.

Bv=B0i'αiBvi+C'v=C0i'αicvi+,  i2

where B0 and C0 are rotational constants, and the α are small corrections. All coordinates involving the v2 mode, including its coupling with the other degrees of freedom are included in G2. Thus G2 includes those effects characterized by

(v1v2v3l3v4l4JK)(v10v3l3v4l4JK).

The partition function for the rotational and vibrational structure may be written

Q=v,J,Kexp{T(v,J,K)c2/T}, (4)

where T(ν, J, K) is discussed in eqs (1), (2), and (3). c2 is the second radiation constant and T is the absolute temperature. The sum extends over all rotational and vibrational degrees of freedom including all values of the ν, J, K quantum numbers. In the evaluation of eq (4) the contributions of the vibrational anharmonicity and vibration-rotation coupling effects contained in eqs (2) and (3) are all to be accounted for to first order; that is, only terms linear in x, g, and α are retained. Though G2 is not conveniently described by the usual power series in the vibrational quantum numbers, its contribution, including coupling effects with the other degrees of freedom, is also to be included to this order.

The contribution of the G2 term is obtained by expanding the partition function about the term-by-term sum of the observed (0v200) energy levels. This summation incorporates the major effect of the highly anharmonic v2 mode, which is the coupling between the symmetric and antisymmetric states, but it does not include the contribution of the v2 coupling with the other vibrational modes. From an examination of the spectroscopic vibrational data (see Benedict et al. [4]), it can be seen that this cross coupling produces a shifting of energy levels of about the same magnitude as that of the other vibrational couplings (those represented by the x and g terms in eq (2)). Thus the contribution to the partition function of this cross coupling should be of the same order as that from the other anharmonicities, and a first order approximation for this contribution should be adequate.

In the expansion that follows expressions for the cross coupling are developed and their contribution for each individual (0ν200) state obtained explicitly. The cross coupling between a pair of vibrational states, e.g., states corresponding to the particular quantum numbers ν1 and v2, is defined as

Gv1v2G(v1v200)G(v1000)G(0v200), (5)

where G(v1v2v3l3v4l4) is the vibrational energy. A useful representation for the coupling involving the highly anharmonic v2 mode is the development of Gv1v2 as a power series of products of the observed energies, that is

Gv1v2=k12G(v1000)G(0v200)+. (6)

The constant k12 is readily determined in terms of the observed coupling. Thus for coupling between the fundamentals the observed coupling x12 is given by

x12=k12G(1000)G(0100)  +=k12ν1ν2+.

Note that eq (6) reduces identically to the power series of eq (2) when the cross coupling is between nearly harmonic modes.

Consider the energy of the nearly harmonic oscillator νm, neglecting all anharmonicities but the cross coupling with v2. The energy levels for such an oscillator are given by

Gm=νmvm+Gvmv2.

Substituting from eq (6) and defining E2G(0v200), this becomes

Gmνmvm+k2mνmvmE2. (7)

The contribution of Gm to the partition function is the factor

Qm={vmexp[(c2/T)(νmvm+k2mE2νmvm)]}dm, (8)

where dm is the degeneracy of the mth mode. The anharmonicity term in the exponential of eq (8) is now expanded and only the linear term retained to yield

Qm={vm{exp[c2Tνmvm]}{1c2Tk2mνmvmE2}}dm. (9)

Using the identity

ddαveαv=vveαv,

the contribution of Qm becomes

Qm=(1ec2Tνm)dm(1f2m), (10)

where

f2m=dmc2Tk2mE2νm/(ec2νm/T1). (11)

The factor (1 − f2m) is the contribution of the interaction of a particular (0v200) energy state with the mth vibrational mode, and it depends explicitly on E2. It is clear then that the effect of this coupling is to modify each of the terms in the term-by-term sum over the (0v200) energy levels by the factor (1 − f2m). The term-by-term sum including cross coupling with the other vibrational modes is thus given by

Qv2(vib)=E2m(1f2m)exp(c2E2/T). (12)

The prime indicates that the m = 2 factor in the product corresponding to self coupling of the v2 mode is omitted.

The coupling of rotation with the v2 mode is included by evaluating the classical rotational partition function for each of the (0v200) energy levels and employing these as separate weighting factors to each of the terms in eq (12). These factors are the usual expression for a classical nonrigid rotator,

QR(v2)=π1/2(σB2σC)1/2(1+ρ(1)T+ρ(2)T2). (13)

The quantities σB and σC are given by,

σB=c2TBv2,
σC=C2TCv2,

where Bv2 and Cv2 are the observed rotational constants for the particular (0v200) energy level. The factor 1 + ρ(1)T + ρ(2)T2 is the contribution of rotational stretching (see, for example (5 and 14)), where the ρ are calculated from the rotational stretching coefficients that apply to the particular (0v200) level.

One further simplification is required in order to evaluate the out-of-plane contribution. This is to obtain an approximation for the contribution of the energy levels corresponding to high values of the v2 quantum number. Note that the v2 mode is nearly harmonic for energies corresponding to the fourth quantum number and higher, so that the contribution of these levels may be replaced by the well-known expression for a slightly anharmonic rotating vibrator. Combining eqs (12) and (13) and including the closed form approximation for the contribution of the states corresponding to quantum numbers given by v2 > 3 yields the expression for G2 (ν, J, K) used in this calculation.

Qv2={v2<4B0BE(C0CE)1/2(1+ρ(1)T+ρ(2)T2). j'(1f2j)ec2TE}+exp(c2E4s/T)1exp(c2ν2/T)B0B4(C0C4)1/2. j(1θ2j)(1+ϕ2)(1+ρ(1)T), (14)

where EE2 and the subscript to E now refers to the particular v2 quantum number. The quantity in the braces in eq (14) includes levels only up to v2 = 3a. Note that the rigid rotational contribution for the ground state has been factored from eq (14). In the second term the factor exp(c2E4s/T)1exp(c2ν2/T) is the harmonic oscillator contribution with E4s as the first level and ν2 the fundamental. The factor 1 − θ2j is the ordinary anharmonicity coupling with the jth vibrational mode, where

θij=C2Txijdidj(exp(c2νi/T)1)(exp(c2νj/T)1),θii=c2T(xiigii/3)di(di+1)(exp(c2νi/T)1)2. (15a)

The factor 1 + ϕ2 is the well-known contribution of rotation-vibration coupling, where

ϕi=di(αiBB+12αiCC). (15b)

Details concerning the derivation of quantities in eqs (15) may be found in Friedman and Haar [6], Woolley [7], and Pennington and Kobe [8], for example.

The contributions to the partition function of the two remaining terms G′ (ν, J, K) and F′ (ν, J, K) are quite straightforward (see [6, 7, 8]). With these the complete partition function is finally obtained.

Q=1Sm=14'(1ec2Tνm)dmij4'(1θij)k=14'(1+ϕk)(ΠσB02σC0)1/2Qv2QtrQqr. (16)

As noted earlier the primes on the product summations indicate that factors containing the vibrational coordinate v2 are omitted. The quantity S is the symmetry number. For the usual symmetric top molecule this has the value three. However, each of the components of the nondegenerate but split (0v200) levels have been counted with a statistical weight of unity, so that the factor Qν2 is larger by a factor of two. This error is corrected by setting S = 6. (Incidentally S = 6 corresponds to the symmetry for a planar configuration for NH3, and eq (16) might be thought of as an expansion about this planar configuration.) The product over m is the harmonic oscillator contribution for the modes other than v2; the products over ij and over k are the corresponding anharmonicity and rotation-vibration coupling contributions, respectively. The degeneracy coefficients d are unity for m = 1, 2, and two for m = 3, 4. The θij and ϕk are given by eq (15). Qtr is the usual contribution of translation. Lastly, the factor Qqr is a correction for rotational quantum effects. It is very small at temperatures for which the vibrations are appreciably excited, so that it is not coupled to the v2 mode. The contribution Qqr is given by Stripp and Kirkwood [9] for an asymmetric rotator, which for a symmetric top rotator reduces to

Qqr=1+σB3(114σBσC)+σB215(134σBσC+732σB2σC2).

3. Molecular Data

The molecular data upon which the calculations are based are listed in tables 1 and 2. Table 1 lists the data sensitive to the inversion coordinates and utilized in the term-by-term sum (the quantity in braces in eq (14)). The vibrational energies corresponding to v2 = 0, 3 are from [4]. The energies corresponding to v2 = 1, 2 are from Garing et al. [10]. The values listed for the rotational constants Bν and Cν corresponding to the v2 = 0 states are from [4]; those corresponding to the v2 = 1, 2 states from [10]; and to the v2 = 3 states from Benedict et al. [11]. The rotational stretching constants DvJ, DvJK, DvK are the average of the symmetric and anti-symmetric values. The values for v2 = 0 are from [4]; for v2 = 1, from [10]; for v2 = 3, from [11]; finally, data for v2 = 2 were taken equal to the v2 = 1 values.

Table 1.

The out-of-plane molecular constants in cm−1, v2 ⩽ 3, (0v200)

State 08 0a 18 1a 28 2a 38 3a
v2 0.0 0.793 932.41 968.04 1597.42 1882.16 2383.46 2895.48
Bv2 9.9443 9.939 10.07 9.89 10.26 9.73 9.50 9.20
Cv2 6.196 6.198 6.05 6.13 5.90 6.13 6.16 6.26
Dv2J 8.1 × 10−4 8.9 × 10−4 8.9 × 10−4 −3.3 × 10−4
Dv2JK −14.5 × 10−4 −17.5 × 10−4 −17.5 × 10−4 15.0 × 10−4
Dv2K 7.5 × 10−4 10.5 × 10−4 10.5 × 10−4 −11.3 × 10−4

Table 2.

Vibrational, rotational and coupling constants in cm−l; see eqs (2), (3)

v1(1 0 0 0) v2' = E4a − E4s v3(0 0 1 0) v4(0 0 0 1)
3323.06 475 3443.381 1627.77
x11 = −30 x22 = + 9 x33 = −18.5 x44 = −8.8
x12 = + 20.6 x23 = + 32.36 x34 = −17.3 g44 = + 2.65
x13 = −92 x24 = −10.7
x14 = −6.7 B4 = 9.35, C4 = 6.21
α1B = 0.135 α2B = + 0.015 α3B = + 0.176 α4B = −0.230
α1C = .078 α2C = + 0.098 α3C = −0.009 α4C = + 0.066

The remaining molecular data are given in table 2. The vibrational fundamental ν1 is from Benedict et al. [12]; v3 is from [11]; v4 is from [10]. The constants ν2 and x22 which correspond to the high energy, nearly-harmonic region of the v2 mode are derived from the reported values for the (04s00) and (05s00) levels at 3330 cm−1 and 4315 cm−1 given by Walsh and Warsop [13] and from theoretical considerations that at high energies the spacing between the levels should become nearly equal to one-half the (0100) value. Also pertaining to this region are the rotational constants B4 and C4 which are taken equal to the averages of the values for v2 = 3s and 3a. The anharmonicity coefficients x12, x13, x14, x23, x24, x33, and x34 are from [4], and the coefficients x44, and g44 are from [10]. The anharmonicity coefficient x11 is estimated from data for similar stretching vibrations. The coupling constants αiN are from [4]. The small coefficient g33 is taken as zero.

The values used for the atomic masses are from Cameron and Wichers [35]. The values used for the other physical constants are given in [36].

4. Thermodynamic Functions

The thermodynamic properties of ammonia as an ideal gas at a pressure of one atmosphere are listed in table 3. These include the Gibbs (free) energy function, enthalpy function, heat capacity at constant pressure, and entropy tabulated in dimensionless units at closely spaced temperature intervals from 50 to 5000 °K. Should it be desired to obtain the properties for a pressure of 1 bar (106 dynes cm−2 or 105 Nm−2) the magnitude of the tabulated free energy and entropy values should be increased by 0.0132. Conversion factors to other units in frequent use are given in table 4. The entries in table 3 are truncated at four decimals (except for CP°/R above 2500 °K which is truncated at three) but at best they are accurate to three decimals. It is felt that the additional significance could be useful in applications involving differences between the entries. The density of entries permits use of linear interpolation throughout the table with a resulting interpolation error smaller than the uncertainty in the tabulated values.

Table 3. Thermodynamic functions for ammonia.

The tabulated values are spaced sufficiently close to permit linear interpolation between the entries with a resulting error less than their overall uncertainty. For convenience the values at 273.15 and 298.15 °K have also been included.

T (E0°G)/RT (HE0°)/RT CP/R S°/R
50. 12.0017 3.9539 4.0029 15.9555
60. 12.7234 3.9620 4.0026 16.6854
70. 13.3346 3.9678 4.0026 17.3024
80. 13.8648 3.9721 4.0027 17.8369
90. 14.3328 3.9756 4.0029 18.3084
100. 14.7519 3.9783 4.0033 18.7302
110. 15.1311 3.9806 4.0039 19.1117
120. 15.4776 3.9826 4.0051 19.4602
130. 15.7964 3.9844 4.0069 19.7808
140. 16.0918 3.9861 4.0097 20.0779
150. 16.3668 3.9878 4.0138 20.3546
160. 16.6243 3.9896 4.0194 20.6138
170. 16.8662 3.9915 4.0267 20.8577
180. 17.0944 3.9937 4.0358 21.0881
190. 17.3104 3.9962 4.0468 21.3066
200. 17.5154 3.9991 4.0599 21.5145
210. 17.7106 4.0023 4.0749 21.7129
220. 17.8969 4.0060 4.0920 21.9029
230. 18.0751 4.0101 4.1110 22.0852
240. 18.2458 4.0147 4.1319 22.2606
250. 18.4098 4.0199 4.1548 22.4297
260. 18.5676 4.0255 4.1794 22.5931
270. 18.7196 4.0317 4.2057 22.7513
273.15 18.7663 4.0338 4.2143 22.8001
280. 18.8664 4.0384 4.2336 22.9048
290. 19.0082 4.0456 4.2630 23.0539
298.15 19.1204 4.0520 4.2882 23.1724
300. 19.1455 4.0534 4.2939 23.1989
310. 19.2785 4.0617 4.3260 23.3402
320. 19.4076 4.0704 4.3593 23.4781
330. 19.5330 4.0797 4.3937 23.6127
340. 19.6550 4.0895 4.4291 23.7444
350. 19.7737 4.0997 4.4654 23.8733
360. 19.8893 4.1104 4.5025 23.9997
370. 20.0021 4.1215 4.5402 24.1235
380. 20.1121 4.1330 4.5785 24.2451
390. 20.2196 4.1449 4.6173 24.3646
400. 20.3247 4.1572 4.6565 24.4819
410. 20.4275 4.1699 4.6961 24.5974
420. 20.5282 4.1829 4.7359 24.7111
430. 20.6268 4.1962 4.7759 24.8230
440. 20.7234 4.2098 4.8161 24.9332
450. 20.8181 4.2237 4.8563 25.0419
460. 20.9111 4.2379 4.8966 25.1491
470. 21.0024 4.2524 4.9368 25.2548
480. 21.0921 4.2671 4.9771 25.3592
490. 21.1802 4.2820 5.0172 25.4622
500. 21.2669 4.2971 5.0572 25.5640
520. 21.4360 4.3278 5.1368 25.7639
540. 21.6000 4.3593 5.2158 25.9592
560. 21.7591 4.3912 5.2939 26.1503
580. 21.9137 4.4237 5.3712 26.3374
600. 22.0643 4.4566 5.4475 26.5208
620. 22.2109 4.4898 5.5230 26.7007
640. 22.3540 4.5232 5.5975 26.8772
660. 22.4937 4.5569 5.6711 27.0506
680. 22.6302 4.5907 5.7438 27.2210
700. 22.7638 4.6247 5.8157 27.3885
720. 22.8946 4.6588 5.8867 27.5533
740. 23.0227 4.6929 5.9568 27.7156
760. 23.1483 4.7271 6.0261 27.8754
780. 23.2715 4.7613 6.0946 28.0328
800. 23.3925 4.7954 6.1622 28.1879
820. 23.5113 4.8296 6.2291 28.3409
840. 23.6281 4.8637 6.2952 28.4918
860. 23.7430 4.8978 6.3604 28.6407
880. 23.8559 4.9317 6.4249 28.7877
900. 23.9672 4.9656 6.4885 28.9328
920. 24.0767 4.9994 6.5513 29.0761
940. 24.1845 5.0331 6.6133 29.2176
960. 24.2909 5.0667 6.6745 29.3575
980. 24.3957 5.1001 6.7348 29.4958
1000. 24.4990 5.1334 6.7943 29.6324
1020. 24.6010 5.1665 6.8530 29.7676
1040. 24.7017 5.1995 6.9108 29.9012
1060. 24.8010 5.2323 6.9678 30.0334
1080. 24.8991 5.2650 7.0239 30.1641
1100. 24.9960 5.2975 7.0792 30.2935
1120. 25.0918 5.3298 7.1335 30.4216
1140. 25.1864 5.3619 7.1870 30.5483
1160. 25.2799 5.3938 7.2397 30.6738
1180. 25.3724 5.4256 7.2914 30.7980
1200. 25.4639 5.4571 7.3424 30.9209
1220. 25.5543 5.4884 7.3924 31.0427
1240. 25.6438 5.5195 7.4416 31.1633
1260. 25.7324 5.5504 7.4900 31.2828
1280. 25.8200 5.5811 7.5375 31.4011
1300. 25.9068 5.6115 7.5841 31.5183
1320. 25.9927 5.6418 7.6299 31.6345
1340. 26.0778 5.6718 7.6749 31.7495
1360. 26.1620 5.7016 7.7190 31.8636
1380. 26.2455 5.7311 7.7624 31.9766
1400. 26.3281 5.7604 7.8049 32.0886
1420. 26.4101 5.7895 7.8466 32.1996
1440. 26.4912 5.8184 7.8875 32.3096
1460. 26.5717 5.8470 7.9277 32.4187
1480. 26.6514 5.8754 7.9671 32.5268
1500. 26.7305 5.9035 8.0058 32.6340
1550. 26.9252 5.9729 8.0992 32.8981
1600. 27.1159 6.0407 8.1881 33.1566
1650. 27.3028 6.1071 8.2728 33.4099
1700. 27.4861 6.1720 8.3532 33.6581
1750. 27.6659 6.2354 8.4298 33.9013
1800. 27.8424 6.2974 8.5027 34.1398
1850. 28.0158 6.3579 8.5719 34.3737
1900. 28.1862 6.4171 8.6378 34.6032
1950. 28.3536 6.4748 8.7004 34.8284
2000. 28.5182 6.5312 8.7599 35.0494
2050. 28.6802 6.5863 8.8164 35.2664
2100. 28.8395 6.6400 8.8702 35.4795
2150. 28.9964 6.6925 8.9214 35.6889
2200. 29.1508 6.7437 8.9701 35.8945
2250. 29.3030 6.7937 9.0162 36.0966
2300. 29.4528 6.8425 9.0602 36.2953
2350. 29.6005 6.8901 9.1021 36.4906
2400. 29.7460 6.9366 9.1418 36.6826
2450. 29.8895 6.9820 9.1797 36.8715
2500. 30.0310 7.0263 9.216 37.0573
2550. 30.1706 7.0696 9.250 37.2402
2600. 30.3083 7.1118 9.282 37.4201
2650. 30.4441 7.1531 9.313 37.5972
2700. 30.5782 7.1933 9.342 37.7716
2750. 30.7106 7.2327 9.370 37.9432
2800. 30.8413 7.2711 9.397 38.1123
2850. 30.9703 7.3086 9.422 38.2789
2900. 31.0977 7.3452 9.445 38.4429
2950. 31.2236 7.3810 9.468 38.6046
3000. 31.3479 7.4160 9.490 38.7639
3100. 31.5922 7.4835 9.529 39.0757
3200. 31.8308 7.5480 9.564 39.3788
3300. 32.0640 7.6096 9.596 39.6736
3400. 32.2921 7.6684 9.623 39.9605
3500. 32.5152 7.7246 9.648 40.2398
3600. 32.7336 7.7783 9.670 40.5119
3700. 32.9474 7.8297 9.688 40.7771
3800. 33.1568 7.8788 9.704 41.0357
3900. 33.3621 7.9258 9.717 41.2879
4000. 33.5633 7.9707 9.728 41.5341
4100. 33.7607 8.0137 9.737 41.7744
4200. 33.9543 8.0548 9.743 42.0091
4300. 34.1443 8.0941 9.748 42.2384
4400. 34.3308 8.1317 9.750 42.4626
4500. 34.5140 8.1677 9.751 42.6817
4600. 34.6939 8.2021 9.749 42.8960
4700. 34.8706 8.2350 9.746 43.1056
4800. 35.0443 8.2664 9.741 43.3107
4900. 35.2151 8.2964 9.734 43.5115
5000. 35.3833 8.3253 9.725 43.7086

Table 4.

Conversion factors to convert tabulated value to quantity having dimensions indicated below Multiply by
Cal mole−1 °K−1 1.98717
Cal g−1 °K−1 0.116682
Joules mole−1 °K−1 8.3143
Joules g−1 °K−1 0.488197

As indicated by the earlier discussion the present calculation, except for the out-of-plane vibrations, is based on an expansion about the harmonic oscillator-rigid rotator model. In this expansion only first order (linear) terms in the coupling between the various degrees of freedom are retained. In addition to the errors inherent in the derivation of the partition function, eq (16), there are also errors due to uncertainties in the molecular constants, errors which arise from certain resonances between several of the low-lying vibrational states, the effects of which are omitted, and lastly errors arising from omission of excited electronic states.

The expansion about the harmonic oscillator limit is essentially a low temperature expansion, and the errors resulting therefrom for a particular vibrational mode are small provided νT is large, and are of the order of about the square of the linear term. The largest error of this type occurs for the v4 mode. Thus for the most sensitive function, the heat capacity, the uncertainty at 2000 °K is less than 10−2 dimensionless units, but increases to the order of 0.1 at 5000 °K. The higher order anharmonicities would contribute an uncertainty perhaps as large. Though the anharmonicity coefficients x11 and x13 have large uncertainties, the errors introduced by these are mitigated by the fact that the fundamentals ν1 and v3 are large. A 50 percent uncertainty in x11 and x13 results in an error of about 0.05 dimensionless units in CP°/R at 5000 °K. Setting g33 equal to zero introduces negligible error since this coefficient introduces a small correction to an already small anharmonic contribution (owing to the large value of the v3 fundamental). The remaining anharmonicities (other than the x22 value in table 2) are probably accurate to better than several percent, and the uncertainties resulting from these are comparatively small. Uncertainties arising from errors in the rotational fundamentals are also comparatively small, as are uncertainties resulting from the use of estimates for some of the rotational stretching constants. The uncertainties that arise from errors in the rotation-vibration coupling constants are perhaps half as large as those arising from errors in the anharmonicity coefficients.

The uncertainties in the values used for the fundamental frequencies are less than several tenths of a wavenumber, and the uncertainty from this source is negligible. Also negligible are the uncertainties due to errors in the observed overtones of the v2 fundamental.

It is somewhat more difficult to determine the uncertainty resulting from estimates used for the molecular constants in the second term in eq (14). The largest uncertainty here arises from possible errors in ν2 and E4s. An uncertainty of several percent in these results in an uncertainty in CP°/R of less than 0.002 dimensionless units at 1000 °K, which increases to a maximum near 2000 °K of less than 0.01 dimensionless units. The errors from estimates of the vibrational coupling constants used in this term are completely negligible below 2000 °K. Even if a 100 percent uncertainty in their contribution is assumed, the resulting uncertainty at 5000 °K still would be less than 0.1 dimensionless units in CP°/R.

The spectrum of ammonia is further complicated by several large resonances that occur between some of the low lying vibrational bands, such as the resonance between the combination bands (2v4 + v3) and (ν1 + v3) and between the bands 2v4 and ν1 as discussed in [4]. Since the resonances tend to displace the respective bands symmetrically about their unperturbed values, it can be shown that the errors are less than the neglected higher order anharmonicities. A similar type of error occurs in the treatment of the coupling of the v2 mode with the other modes. The coupling constants as reported are the average of the symmetric and antisymmetric values. Thus the effect of this approximation is to introduce errors similar to those arising from neglect of the resonances discussed above. The treatment of centrifugal distortion, which includes a quadratic term in the temperature, is consistent with the other approximations and the error from this source at 5000 °K is probably less than 0.05 in CP°/R. At the low temperature end there is a small error arising from the rotational quantum approximation Qqr (eq (17)). At 100 °K this error is less than 10−3 in CP°/R, but is appreciably larger at 50 °K.

Finally, the U.V. absorption spectrum indicates bands extending from ~ 46,000 cm−1 to over 100,000 cm−1, as discussed, for example, in [13]. The error from omission of these even at the very highest temperatures considered here is completely negligible. It is felt, therefore, that the uncertainty in the tables in the temperature region 100 to 1000 °K is probably less than 0.003 dimensionless units in CP°/R. This uncertainty increases to 0.02 at 2000 °K and then increases very rapidly to perhaps 0.3 at 5000 °K. A more complete list of the uncertainty estimates, including those for the enthalpy and entropy, are given in table 5.

Table 5.

Estimates of uncertainty in dimensionless units

T ΔCP°/R Δ(H°E0°)/RT ΔS°/R
100 0.003 0.002 0.003
1000 .003 .002 .003
2000 .02 .01 .02
3000 .05 .02 .03
4000 .1 .05 .07
5000 .3 .1 .2

5. Discussion

Ideal-gas calculations for NH3 have been reported in the literature since the early thirties, but many of these calculations were based on relatively incomplete and inaccurate molecular data and on simplified structural models. Also, for the most part, the temperature ranges were somewhat limited. The references [1727] are representative of these early efforts.

Some of the more recent calculations which have been widely used include those of Harrison and Kobe [28] who calculated the heat capacity, enthalpy, and entropy from 273.16 to 1500 °K. They included the major effects associated with the molecular inversion by summing the lower states of the v2 mode term by term, but omitted all vibration and rotational coupling effects other than rotational stretching. The review by Davies reported in the monograph edited by Din [29] includes the specific heat, entropy, and enthalpy from 200 to 1000 °K. The calculation follows essentially the procedure in [28], except that centrifugal stretching is omitted. Also it appears that the contribution of the states in the term-by-term sum of the v2 mode was omitted above the level 3a. The most detailed of the previous calculations is that reported by Yungman et al. [30], (which also appears in [16]), in which the free energy, entropy, and enthalpy are tabulated at 100 °K intervals from 298.15 to 6000 °K. In addition to the term-by-term sum for the v2 mode, this work also includes several of the first order vibrational and rotational coupling contributions. It omits, however, the contributions due to coupling of the v2 mode with the other vibrational modes, and includes a somewhat simplified coupling of the v2 mode with rotation. These calculations are based on essentially the same molecular data as those reported here, except that the values used for the vibrational coupling constants x13 and x44 are about twice as large. One of the most widely used set of tables is that of JANAF [31]. The reference refers to the latest revision for ammonia, September 30, 1965. This calculation follows the procedures of [28] and [30], though the specific details are not given in the text. Finally, there are several recent calculations based on somewhat simplified molecular models, which are included in references [3234].

The figures 1 and 2 include a comparison of some of the previous calculations with those made here. The ordinates are the deviations from the present calculations, that is the present calculation minus the others. Figure 1 is a plot of entropy differences, and figure 2 is a plot of heat capacity differences. It is most interesting to compare the data of Yungman et al. [30], curve #5 in figure 1. At the low temperatures the agreement is quite good, but as the temperature is increased, a positive deviation occurs until about 3000 °K, when the deviation becomes strongly negative. The difference at the lower temperatures is probably due to their simplified treatment of the coupling of the v2 mode with rotation. At temperatures above 1000 °K, the dominant contribution to the difference is from the x24 term omitted in [30] which yields a positive contribution. At temperatures above 3000 °K, the dominant contribution is from the omitted x12 and x23 terms which yield large negative contributions. In addition to these, the high value they used for x44 (larger by a factor of two) tends to reduce the differences at temperatures below 3000 °K, but magnifies them above. The same general comments apply to the JANAF [31] comparison as seen in curve #1, except that the difference goes negative at somewhat higher temperatures and the positive deviations are considerably larger. Incidentally, the JANAF and Harrison and Kobe [28] (curve #4) calculations are nearly identical for temperatures up to 1000 °K. Curve #3 applies to a modified rigid rotator harmonic oscillator calculation as given in an older (now superseded) JANAF table. The comparison for curve #2 [29] illustrates the sensitivity to the omission of the hierarchy of states corresponding to large v2 quantum numbers (v2 > 3). The differences plotted in figure 2 exhibit essentially the same behavior as those in figure 1 except that they are somewhat larger.

Figure 1.

Figure 1.

Dimensionless entropy differences.

Figure 2.

Figure 2.

Dimensionless heat rapacity differences.

Acknowledgments

It is a pleasure to thank Joseph Hilsenrath for many helpful technical discussions, particularly relating to the use of OMNITAB to program the calculation. In this latter connection, I also wish to thank Robert C. Thompson.

This work was supported in part by the Office of Standard Reference Data, NBS.

Footnotes

1

Figures in brackets indicate the literature references at the end of this paper.

6. References

  • [1].Swalen J. D. and Ibers J. A., J. Chem. Phys. 36, 1914 (1962). [Google Scholar]
  • [2].Herzberg G., Infrared and Raman Spectra of Polyatomic Molecules, D. Van Nostrand & Co., New York: (1945). [Google Scholar]
  • [3].Nielsen H. H., Revs. Modern Phys. 23, 90 (1951). [Google Scholar]
  • [4].Benedict W. S. and Plyler E. K., Can. J. Phys. 35, 1235 (1957). [Google Scholar]
  • [5].McDowell R. S., J. Chem. Phys. 39, 526 (1963). [Google Scholar]
  • [6].Friedman A. S. and Haar L., J. Chem. Phys. 22, 2051 (1954). [Google Scholar]
  • [7].Woolley H. W., Res J.. NBS 56, 105 (1956). [Google Scholar]
  • [8].Pennington R. E. and Kobe K. A., J. Chem. Phys. 22, 1442 (1954). [Google Scholar]
  • [9].Stripp K. F. and Kirkwood J. G., J. Chem. Phys. 19, 1131 (1951). [Google Scholar]
  • [10].Garing J. S., Nielsen H. H., and Rao K. N., J. Mol. Spect. 3, 496 (1959). [Google Scholar]
  • [11].Benedict W. S., Plyler E. K., and Tidwell E. D., J. Chem. Phys. 29, 829 (1958). [Google Scholar]
  • [12].Benedict W. S., Plyler E. K., and Tidwell E. D., J. Chem. Phys. 32, 32 (1960). [Google Scholar]
  • [13].Walsh A. D. and Warsop P. A., Trans. Faraday Soc. 57, 345 (1961). [Google Scholar]
  • [14].Yungman Y. S., Zhur. Fiz. Khim. 35, 319 (1961). [Google Scholar]
  • [15].Godnev I. N., Zhur. Fiz. Khim. 27, 1702 (1953). [Google Scholar]
  • [16].Glushko V. P., Ed., Termodinamicheskie Svoistva Individual’nykh Vestchestv, II, Acad. Sci., USSR, Inst. of Applied Chemistry, Moscow: (1962). [Google Scholar]
  • [17].Villars D. S., Phys. Rev. 38, 1552 (1931); [Google Scholar]; J. Am. Chem. Soc. 53, 2006 (1931). [Google Scholar]
  • [18].Trautz M., Ann. Physik 9, 465 (1931). [Google Scholar]
  • [19].Schmidt F. A. F., Forsch. Arb. Ing. Wes. 8, 91 (1937). [Google Scholar]
  • [20].Overstreet R. and Giauque W. F., J. Am. Chem. Soc. 59, 254 (1937). [Google Scholar]
  • [21].Kelley K. K., U.S. Bur. Mines, Bull. 407 (1937). [Google Scholar]
  • [22].Justi E., Spez-Wärme, Enthalpie und Dissociation Technischer Gase. J. Springer, Berlin: (1938). [Google Scholar]
  • [23].Stephenson C. C. and McMahan H. O., J. Am. Chem. Soc. 61, 437 (1939). [Google Scholar]
  • [24].Pitzer K. S., J. Chem. Phys. 7, 251 (1939). [Google Scholar]
  • [25].Haupt R. F. and Teller E., J. Chem. Phys. 7, 925 (1939). [Google Scholar]
  • [26].Zeise H., Z. Elektrochem.. 47, 380 (1941). [Google Scholar]
  • [27].Thomson H. W., Trans. Faraday Soc. 37, 344 (1941). [Google Scholar]
  • [28].Harrison R. H. and Kobe K. A., Chem. Eng. Progress 49, 351 (1953). [Google Scholar]
  • [29].Din F., Thermodynamic Functions of Gases 1 (Butterworths, London, 1956). [Google Scholar]
  • [30].Yungman Y. S., Gurvich L. V., and Rtischeva N. P., Tr. Gos. Inst. Prikl. Khim. 49, 26 (1962). [Google Scholar]
  • [31].JANAF Thermochemical Tables, Distrib. by Clearinghouse for Federal Scientific Information, PB168370–1, Sept. 30, 1965.
  • [32].Mader C. L., Ideal Gas Thermodynamics of Detonation Products, U.S. A.E.C., AECU–4508 (1959).
  • [33].Khlebnikova V. N. and Morozov V. P., Ukrain. Khim. Zhur. 24, 3 (1958). [Google Scholar]
  • [34].Sundaram S., Suszek F., and Cleveland F. F., J. Chem. Phys. 32, 251 (1960). [Google Scholar]
  • [35].Cameron A. E. and Wichers E., J. Am. Chem. Soc. 84, 4192 (1962). [Google Scholar]
  • [36].NBS Tech. News Bull. 47, No. 10 (Oct. 1963). [Google Scholar]

Articles from Journal of Research of the National Bureau of Standards. Section A, Physics and Chemistry are provided here courtesy of National Institute of Standards and Technology

RESOURCES