Rep:Mod2Proj:Yuko.Isayama3001

Mini Project; Phosphazene (Phosphonitrilic) Compounds

The project involved the investigation of the electronic structure of cyclic phosphazene trimers, (NPX₂)₃. Furthermore, the effects of substituents X (X=F, Cl, Br and NH₂) on the vibrational properties of the phosphazene as well as the effects of increasing the trimer to specifically the cyclic tetramer will be studied.

Structural and IR Analysis of (NPCl₂)₃

Initial investigation specifically involved the structural analysis of hexachlorocyclotriphosphazene, (PNCl₂)₃, which was drawn as a planar six-membered ring structure. A loose optimisation of the compound was performed using the DFT, B3LYP method with LANL2MB pseudo potential and basis set before a second optimisation using the same method with LANL2DZ pseudo potential and basis set, followed by freqency analysis.

The geometric pararmeters as a result of the optimisation based on are given in the following tables with the experimental literature values:

Bond length/± 0.01Å
Bond	Computed	Literature¹
P-N	1.68	1.580
P-Cl	2.21	1.990

Bond angle/± 0.1°
Bond	Computed	Literature¹
Cl-P-Cl	100.8	101.4
N-P-N	117.0	118.4
P-N-P	123.0	121.4

DOI:10042/to-2749

Here is a summary table of the forms i.e. motion, frequency and intensity of each degenerate pair of the P-N-P vibration;

Mode	Form of vibration	Computed frequency/±10cm^-1	Literature frequency²/±10cm^-1	Intensity
1	https://www.ch.ic.ac.uk/wiki/index.php/Image:Phosp_Cl_freq25.gif	615	-	9
2	https://www.ch.ic.ac.uk/wiki/index.php/Image:Phosp_Cl_freq26.gif]	615	-	9
3	https://www.ch.ic.ac.uk/wiki/index.php/Image:Phosp_Cl_freq29.gif	1056	1218	1029
4	https://www.ch.ic.ac.uk/wiki/index.php/Image:Phosp_Cl_freq30.gif	1056	1218	1029

DOI:10042/to-2759

The computed P-N stretching frequencies based on the B3LYP method and LANL2DZ pseudo-potential and basis set of the other compounds of study;

Characteristic P-N stretching frequencies/± 10cm^-1
Compound	Computed	Literature¹
(NPF₂)₃	1114	1297
(NPCl₂)₃	1056	1218
(NPBr₂)₃	1038	1157

(NPF₂)₃ DOI:10042/to-2754 (NPCl₂)₃ DOI:10042/to-2759 (NPBr₂)₃ DOI:10042/to-2752

Comparison of the computed bond lengths and vibrational frequencies of (PNCl₂)₃ with the experimental literature values show that there is a slight disagreement between them. Frequency analysis of other halocyclotriphosphazenes also showed discrepancies with the experimental literature values. Thus, it can be concluded that the LANL2DZ pseudo-potential and the associated basis set was not good enough to give results of high accuracy; in this method, the phosporous still only has a minimal basis set with the valence s and p orbital functions but it is known in theory that the low-lying d-orbitals of the phosphorous atom contributes to its hypervalency. Additionally, the change in the P-N-P vibration frequency in the halophosphazenes is partly due to the contraction of the 3d orbitals. Hence, another set of optimisation and frequency calculation were performed for each compound with the dAO functions added as an extra basis set which gave better agreement of results.

Structural and IR Analysis Based on the Basis Set with the dAO Functions

The geometric parameters as a result of the optimisation of (NPCl₂)₃ are given in the following tables:

Bond length/± 0.01Å
Bond	Computed	Literature¹
P-N	1.60	1.580
P-Cl	2.06	1.990

Bond angle/± 0.1°
Bond angle	Computed	Literature¹
Cl-P-Cl	102	101.4
N-P-N	114	118.4
P-N-P	126	121.4

DOI:10042/to-2758

Here is a summary table of the forms i.e. motion, frequency and intensity of each degenerate pair of the P-N-P vibration based on (NPCl₂)₃:

Mode	Form of vibration	Computed frequency/±10cm^-1	Literature frequency²/±10cm^-1	Intensity
1	https://www.ch.ic.ac.uk/wiki/index.php/Image:Phosp_Cl2_freq26.gif	847	-	16
2	https://www.ch.ic.ac.uk/wiki/index.php/Image:Phosp_Cl2_freq27.gif	847	-	16
3	https://www.ch.ic.ac.uk/wiki/index.php/Image:Phosp_Cl2_freq29.gif	1268	1218	1314
4	https://www.ch.ic.ac.uk/wiki/index.php/Image:Phosp_Cl2_freq30.gif	1268	1218	1314

DOI:10042/to-2757

Using the dAO basis functions as a basis set gives the same form of vibrations as the previous LANL2DZ basis set but in terms of the bond length and frequency, they are in better agreement with those of the experimental literature values than the previous method, suggesting that the d-orbitals have a significant contribution to both properties.

The geometric parameters as a result of the optimisation of the other compounds of study are given in the following table:

Bond length/± 0.01Å
Bond	(NPF₂)₃	(NPCl₂)₃	(NPBr₂)₃	(NPNH₂)₃	(NPCl₂)₄
P-N	1.58	1.59	1.60	1.62	1.57
P-X (X=F, Cl, Br, NH₂)	1.57	2.06	2.27	1.67	2.07

(NPF₂)₃ DOI:10042/to-2763 (NPCl₂)₃ DOI:10042/to-2758 (NPBr₂)₃ DOI:10042/to-2767 (NPNH₂)₃ DOI:10042/to-2768 (NPCl₂)₄ DOI:10042/to-2770

The computed P-N stretching frequencies of the other compounds of study;

Selected characteristic P-N-P frequency/± 10cm^-1
Compound	Computed	Literature²
(NPF₂)₃	1337	1297
(NPCl₂)₃	1268	1218
(NPBr₂)₃	1240	1175
(NPNH₂)₃	1250	1170
(NPCl₂)₄	1457	1315

(NPF₂)₃ DOI:10042/to-2773 (NPCl₂)₃ DOI:10042/to-2757 (NPBr₂)₃ DOI:10042/to-2775 (NPNH₂)₃ DOI:10042/to-2777 (NPCl₂)₄ DOI:10042/to-2778

Frequency analysis of (NPCl₂)₃ and the other cyclotriphosphazenes show a strong characteristic absorption in the 1200/1300cm^-1 region, which corresponds to P-N-P stretching mode, which is degenerate. The other degenerate modes of vibration at 847cm-1 are which was analysed in (NPCl₂)₃ represent the the forbidden symmetric P-N-P stretch.

For the cyclic tetramer, octachlorocyclotetraphozphazene, the 1200cm^-1 absorption has been shifted to a higher frequency by 200cm^-1 approximately with respect to the corresponding trimer. A possible reason for the vibrational shift could be due to the stronger P-N pi-bonding in the tetramer but the higher cyclic analogoues do not exhibit additional shifts² and so do not support this. It is more likely that the shift corresponds to a change in the ring stretching vibrational mode which is due to increased flexibility of the tetramer compared to the rigid trimer¹.

High electronegavtivity of the substituents increases the P-N stretching frequencies with the highest frequency found in (NHF₂)₃, whilst NH₂ which is an electron-donating ligand, lowers it. This effect is attributed to the fact by withdrawing electrons from the phosphorous, the electronegative substituents increasingly permits nitrogen's ability to donate its lone pairs of electrons to the phosphorous. As a result, the P-N bonds appear stronger with higher vibrational frequencies. Another explanation is that the electron withdrawing effect of the ligands from the phosphrous may lead to the contraction of the 3d orbitals so that there is a more effective dpi-ppi bonding around the ring². Either way, both explanations justify that the shortest P-N bond was found in (NHF₂)₃.

Geometric Analysis of the Phosphazenes

Optimisation of the all the cyclotriphosphazes present a planar cyclic arrangement, which could be furthed supported by calculating the dipole moment; a zero or near zero dipole moment would indicate near perfect planarity.

Optimised structure of (NPCl₂)₄ based on LANL2MB method

Optimised structure of (NPCl₂)₄ based on LANL2DZ/dAO method

With regards to the geometry of the cyclic tetramer, Gaussian computes the optimisation as a puckered molecule after a loose optimisation but a planar molecule following the second optimisation. Gaussian computes molecules as if they are in the gas phase and in theory,(NPCl₂)₄ is planar in the vapor state and puckered in the condensed phases². Thus, by using LANL2DZ/dAO which is a more accurate pseudo-potential and basis set than LANL2MB, an accurate geometry of the cylic tetramer was computed.

(NPNH₂)₄
Conformation	LANL2MB	LANL2DZ	Crystallised
Planar		Yes
Puckered	Yes		Yes

The puckered and planar conformations of the tetramer have an energy of -361.15585934 and -364.85859912AU, with an energy difference of 9721.46kJmol^-1 which indicates that the planar conformation is more stable due to minimised non-bonding interaction between the chlorine substituents.

MO analysis of (NPCl₂)₃

Early interpretations of the bonding in phosphazenes were based on Dewar's island model³ which states that delocalization occurs via (P)dpi-(N)ppi overlap, resulting in a sigma-framework built from two sp² hybrid orbitals of nitrogen and four sp³ hybrid orbitals of phosphorous, whilst the d_xy and _dyz orbitals of phosphorous form pi-bonds by interaction with the p_z orbital of nitrogen. However, following the MO calculation of (NPCl₂)₃, the MO representations shown below suggest that the valence d-orbitals participate little to bonding; despite this, dAO functions are required in the calculations to give an accurate representation of the MOs, but they serve as polarization functions only⁴ (explained later).

The MOs show that it is in fact pi'(in-plane) and pi(out-of-plane) bonding systems produced by the p-orbitals which are significant in planar cyclic triphosphazenes.

NBO Analysis of (NPCl₂)₃

Results show that a significant amount of electron density is dominated on the ring nitrogen, particularly on the 2p orbital where the natutral electron configuration is 4.94, which suggests that it is essentially an ionic system; the charge distribution of such system are best described in terms of a zwitterionic form⁵. This further supports the analysis of the MOs; electron density is not accumulated around the phosphorous atoms but rather the nitrogen atoms of the phosphazene. Having electron configurations of 1s², 2s², 2p³ and 1s², 2s², 2p⁶, 3s², 3p³, for nitrogen and phosphorous atoms respectively, we would expect similar computed values for the natural electron configuration. However, the presence of the phosphorous 3d-orbitals lead to the polarisation of the nitrigen 2p orbitals, which explains the accumulation of electron densiy on the nitrogen 2p orbital.

   Atom  No          Natural Electron Configuration
----------------------------------------------------------------------------
    Cl    1      [core]3S( 1.90)3p( 5.31)
    Cl    2      [core]3S( 1.90)3p( 5.31)
    Cl    3      [core]3S( 1.90)3p( 5.31)
    Cl    4      [core]3S( 1.90)3p( 5.31)
    Cl    5      [core]3S( 1.90)3p( 5.31)
    Cl    6      [core]3S( 1.90)3p( 5.31)
     P    7      [core]3S( 0.90)3p( 2.03)3d( 0.12)4p( 0.03)
     P    8      [core]3S( 0.90)3p( 2.03)3d( 0.12)4p( 0.03)
     P    9      [core]3S( 0.90)3p( 2.03)3d( 0.12)4p( 0.03)
     N   10      [core]2S( 1.54)2p( 4.94)3p( 0.01)
     N   11      [core]2S( 1.54)2p( 4.94)3p( 0.01)
     N   12      [core]2S( 1.54)2p( 4.94)3p( 0.01)

Atomic charges of the atoms also support this analysis, where the charge on nitrogen is -0.346 where the phosphorous has a charge of 0.403 as shown below:

The pi and pi' bonding in phosphazene compounds are contributed by the p atomic orbitals, nevertheless computational results show that neglecting the dAO functions on the phosphrous atom shows discrepancies in the structure and geometry with experimental data, particularly the bond lengths and the frequency. It can be concluded that phosphazene frequency, consequently the bond length is not just affected by the electronegativity of the X ligands, but also ring size.

References

G. I. Bullen, J. Chem. Soc. A 1450 (1971)DOI:10.1039/J19710001450
H. R. Allcock, Chem. Rev. 72, 315 (1972) DOI:10.1021/cr60278a002
M. J. S. Dewar, E. A. C. Lucken, M. A. J. Whitehead, Chem. Soc. 2423-2429 (1960)
E. J. Magnusson, Am. Chem. Soc. 115, 1051-1061 (1993)
M. Breza, Polyhedron, 19, 389 (2000)DOI:10.1016/S0277-5387(99)00368-X
R. Jaeger, M. Debowski, I. Manners, G. J. Vancso, Inorg. Chem. 38, 1153 (1999)DOI:10.1021/ic980877n