Mass Spectrometric Study of NF₂, NF₃, N₂F₂, and N₂F₄

Abstract

Appearance potentia’s have been measured for selected ions from NF₂, NF₃, N₂F₂, and N₂F₄. Ionization-dissociation processes are identified and bond dissociation energies are calculated. In addition, the bond dissociation energy, D(F₂N–NF₂), has been directly measured to be 5.14±0.38 kj/mole (21.5± 1.6 kcal/mole). A summary is made of available thermochemical and mass spectrometric data for N–F compounds and some evidence is presented to support the designation of cis and trans structures for the N₂F₂ isomers.

1. Introduction

The synthesis of a new series of compounds containing nitrogen and fluorine atoms has aroused considerable interest in their chemical and physical properties; in particular, heats of formation, bond dissociation energies, and ionization processes. Some of these data have been obtained from mass spectrometric studies [1, 2, 3].1 In general, however, the data are fragmentary and in some cases are based on doubtful assumptions by analogy to N—H compounds. In a previous paper [3], we reported an electron impact study of tetrafluorohydrazine in which a value of 53 kcal/mole for the F₂N–NF₂ bond dissociation energy was calculated from estimated values of the N–F bonds in NF₃ [1]. It was also suggested that the failure of other workers to find ions of m/e greater than that corresponding to NF₂⁺ in the mass spectrum of N₂F₄ was due to decomposition of N₂F₄ into NF₂ radicals in the mass spectrometer ion source. In light of the recently reported [4] value of 19.2 kcal/mole for the dissociation energy of the N–N bond in N₂F₄ this seems quite reasonable. We have made measurements of the effect of temperature on the N₂F₄⁺/NF₂⁺ ratio in the mass spectrum of N₂F₄. In addition, we have made a mass spectrometric study of the thermal dissociation of N₂F₄, and re-examined the ionization-dissociation processes for this molecule. We report appearance potentials of various ions in the related N—F compounds: NF₂, NF₃, and the two available isomers of N₂F₂.

A recent study [5] of the absorption spectra of the N₂F₂ isomers has given rise to a controversy concerning their structure. Although not unequivocal, the data reported here give evidence for the similarity in bond energies and heats of formation of these isomers and hence support the designation of the N₂F₂ isomers as cis and trans.

2. Experimental Procedure

The mass spectrometer used in this research is a first order, direction focusing instrument with a nominal 60° sector field and a 12-in. radius of curvature. The analyser tube and the source and collector housings are fabricated from nonmagnetic stainless steels and made vacuum tight with gold wire gaskets. Separate pumping systems are provided for the source housing and analyser tube. The source housing contains a flanged re-entrant port to admit thermal reactors or electrodeless discharge tubes for the introduction of free radicals or other active species to the ion source with a minimum of wall collisions. In addition, the electron impact source is provided with a conventional gas introduction system.

Carefully regulated power supplies are utilized for the magnet current, the ion accelerating voltage and focusing controls and the electron emission circuit. The latter circuit is designed to permit the precise measurement of appearance potentials of either positive or negative ions and to examine ionization probability curves over the range from zero to 100 ev.

The resolved ion currents are detected by means of a 14-stage electron multiplier. The integrated ion current is measured with a vibrating-reed electrometer and pen recorder. The nominal detection limit for this system was about 10⁻¹⁷ amps.

A simple thermal reactor was attached to the mass spectrometer to study the dissociation of N₂F₄. The reactor, shown schematically in figure 1, was connected to a 2-liter reservoir volume which remained at room temperature. The N₂F₄ at a pressure of about 0.2 mm effused from the reactor through a 1-mil glass leak located at the line-of-sight inlet to the ion source. The temperature of the N₂F₄ vapor was measured by a glass-encased thermocouple located about 1 mm from the leak.

Figure 1.

Thermal reactor for kinetic studies of the dissociation of N₂F₄

The temperature variation of the mass spectrum of N₂F₄ was studied using the technique described by Reese, Dibeler, and Mohler [6]. Briefly, the mass spectrometer filament is turned off and the ion source allowed to cool to room temperature. The N₂F₄ at normal operating pressures is admitted to the ion source through the conventional gas inlet and the filament turned on. Ion currents for the NF₂⁺ and N₂F₄⁺ ions were measured immediately and remeasured at frequent intervals using nominal 70 ev electron energies. The temperature was monitored by means of a thermocouple attached directly to the ion source.

Appearance potentials of NF₂, NF₃, the cis and trans isomers of N₂F₂, and N₂F₄ were measured as described in previous work [7]. For NF₂, measurements were made on the vapors effusing from the reactor containing N₂F₄, at 170 °C.

The NF₃ and N₂F₄ were obtained through D. E. Mann. Their purity has been noted elsewhere [1, 3]. The cis and trans isomers of N₂F₂ were kindly prepared and purified for us by Charles S. Cleaver of the E. I. Du Pont de Nemours Experimental Station, Wilmington, Del. Immediately after separation by gas chromatography, the isomers were placed in Monel cylinders and cooled with solid CO₂. They were transported and maintained at this temperature until introduced to the mass spectrometer. Gas chromatographic analysis reported by Cleaver indicated the following compositions:

• trans–N₂F₂: 0.2% air, <0.1% NF₃, <0.1% N₂O,>99.6% trans–N₂F₂;

• cis–N₂F₂: 0.6% air, 0.2% N₂O, 5.2% trans–N₂F₂, 94.0%. cis–N₂F₂.

These analyses were supported by our mass spectrometric observations.

For conversion from electron volts to joules, 1 ev is taken to be 9.6496×10⁴ joules. For conversion to the thermochemical calories, 1 cal is taken to be 4.1840 joules.

3. Results and Discussion

3.1. Thermal Dissociation of N₂F₄

A typical set of data for the thermal dissociation of N₂F₄ is summarized in table 1. Column 1 gives the absolute temperature of the reactor, and columns 2 and 3 the observed ion currents of the N₂F₄⁺ and NF₂⁺ ions in arbitrary units.

Table 1.

Summary—calculation of the equilibrium constants for the thermal dissociation of N₂F₄

T(°K)	Observed ion currents		${(\text{NF}}_2^{+}{)\mathrm{N}}_2{\mathrm{F}}_4$	${(\text{NF}}_2^{+}{)\text{NF}}_2$	$\frac{K_p}{k}=\frac{{\left({\text{NF}}_2\right)}^2}{{\mathrm{N}}_2{\mathrm{F}}_4}$	${(\text{NF}}_2^{+}{)\mathrm{N}}_2{\mathrm{F}}_4$	${(\text{NF}}_2^{+}{)\text{NF}}_2$	$\frac{K_p}{k}=\frac{{\left({\text{NF}}_2\right)}^2}{{\mathrm{N}}_2{\mathrm{F}}_4}$	${(\text{NF}}_2^{+}{)\mathrm{N}}_2{\mathrm{F}}_4$	${(\text{NF}}_2^{+}{)\text{NF}}_2$	$\frac{K_p}{k}=\frac{{\left({\text{NF}}_2\right)}^2}{{\mathrm{N}}_2{\mathrm{F}}_4}$
	N2F4+	NF2+
450.7	28.0	16320	660	15660	8.76×106	640	15680	8.78×106	630	15690	8.79×106
434.3	68.5	16140	1620	14520	3.09	1260	14580	3.11	1550	14590	3.11
423.9	111.0	16110	2620	13490	1.64	2540	13570	1.66	2510	13600	1.67
412.0	184	16110	4350	11760	7.50×105	4200	11910	7.72×105	4160	11950	7.76×103
401.6	259	16200	6110	10090	3.92	5920	10280	4.08	5900	10300	4.10
382.4	440	17070	10400	6670	1.01	10040	7030	1.12	9950	7120	1.15
362.0	650	18150	15330	2820	1.22×104	14840	3310	1.69×104	14700	3450	1.83×104
343.6	740	18300	17470	830	9.32×102	16900	1400	2.65×103	16720	1580	3.37×103
333.0	750	17700	17700	0	0	17120	580	4.49×102	16920	775	8.01×102

For a first approximation, it is assumed that no NF₂ is formed at the lowest reactor temperature, i.e., 333.0 °K. The ratio of NF₂⁺/N₂F₄⁺ at this temperature was taken as characteristic of the mass spectrum of N₂F₄ and was applied to the data in column 2, table 1 to calculate the contribution to the observed NF₂⁺ peak of NF₂⁺ ions resulting from dissociative ionization of N₂F₄ (column 4). The contribution resulting from the ionization of NF₂ is obtained by difference (column 5). On the further assumption that the observed N₂F₄⁺ ion abundance and the calculated NF₂⁺ ion abundance are measures of the partial pressures of N₂F₄ and NF₂, respectively, an equilibrium constant can be obtained from the relation

$$K_p=k{\left({\text{NF}}_2\right)}^2/{\mathrm{N}}_2{\mathrm{F}}_4$$ (1)

where k is a factor relating measured ion abundances to partial pressures. Values of K_p/k are given in column 6.

From the usual integrated van’t Hoff equation, we plot log K_p versus 1/T to obtain the enthalpy, ΔH, of the reaction. In this case, however, the slope of the plot must be obtained by successive approximation. The data of table 1 are plotted as the open circles of figure 2. The best straight line through these points is extrapolated to the lowest temperature (333.0 °K) and a first estimate made of the ratio NF₂/N₂F₄ from eq (1). This is then used to calculate a more nearly correct set of data. The process is repeated until the indicated constant slope is obtained, shown as solid circles in figure 2. The mean of four such determinations, resulted in a value of ΔH=5.14 ±0.38 kj/mole(21.5± 1.6 kcal/mole). The uncertainty given is the estimated standard deviation. The value of the gas constant used in the calculations was R=8.314 joule/degree mole. This is in good agreement with the previously reported value of 19.2 kcal/mole [4].

Figure 2.

Log K_p versus 1/T for the equilibrium N₂F₄⇌2NF₂.

From the value, ΔH=21.5 ±1.6 kcal/mole for the reaction N₂F₄→2NF₂, and the ΔH_f(N₂F₄) = −2.0±2.5 kcal/mole [8] we calculate ΔH_f(NF₂)=9.8±2.1 kcal/mole. Further, from ΔH_f(NF₃) = −29.7 ±1.8 kcal/mole [9] and ΔH_f(F) = 18.9±0.5 kcal/mole [10], we calculate D(NF₂-F) = 58. 4±4.4 kcal/mole. Similarly, from NF₂→N+2F, we calculate D(N–F) average, in NF₂=70.5±1.6 kcal/mole. Finally, from NF₂→NF+F, we calculate ΔH_f (NF) = 61.4 ±4.2 kcal/mole.

As the average bond energy in NF₃ is 66.3 kcal/mole [9], it would appear that the first N—F bond is the weakest bond in NF₃. This is contrary to the observed bond order in NH₃, in which the first and subsequent N—H bond dissociation energies are reported to be 104, 88, and 88 kcal/mole, respectively [11]. This would negate the assumptions made by Reese and Dibeler [1] in their calculations of the ionization potentials of NF₂ and NF radicals.

3.2 Appearance Potential Data

Two studies of N₂F₄ have been reported [2,3] but the original interpretation of the NF⁺ and NF₂⁺ appearance potentials did not account for the dissociation of N₂F₄ into NF₂ radicals within the ion source.

The effect of ion source temperature on the N₂F₄⁺/NF₂⁺ ratio in the mass spectrum of N₂F₄ is shown in figure 3. Although an extrapolation of the data to lower temperatures is difficult, it seems apparent that the limiting value of the ratio is about 0.08. The change in mass spectrum of N₂F₄ with temperature, due to decomposition of N₂F₄ in the ion source, thus accounts for the differences in the mass spectrum of N₂F₄ reported by different workers [2, 3, 4, 12]. The data of Loughran and Mader [2] have already been reinterpreted assuming the presence of NF₂ [4] in the ion source.

Figure 3.

Effect of ion source temperature on the N₂F₄⁺/NF₂⁺ ratio in the mass specturm of N₂F₄.

A summary of the available appearance potential data for the N—F compounds is shown in table 2. Column 1 identifies the molecule, columns 2 and 3 give the ion and the probable process of formation, column 4 gives the observed appearance potential and column 5 reports the source.

Table 2.

Summary of appearance potential data for N–F compounds

Parent molecule	Ion	Probable process	Appearance potential	Reference
			ev
NF2	$\{\begin{array}{l}{{\text{NF}}_2}^{+}\\ {\text{NF}}^{+}\end{array}$	NF2→NF2+	$\{\begin{array}{l}12.0\pm 0.1\\ 11.8\pm 0.2\end{array}$   11.8±0.2	This work. [2] This work.
		NF2→NF++F−→NF++F	$\{\begin{array}{l}15.5\pm 0.2\\ 15.0\pm 0.2\end{array}$	Do. [2]
NF3	$\{\begin{array}{l}{{\text{NF}}_3}^{+}\\ {{\text{NF}}_2}^{+}\\ {\text{NF}}^{+}\end{array}$	NF3→NF3+	$\{\begin{array}{l}13.2\pm 0.2\\ 13.2\end{array}$a	[1] This work.
		NF3→NF2++F	$\{\begin{array}{l}14.2\pm 0.3\\ 14.6\\ 14.2\to 14.6\end{array}$a	[1] [2] This work.
		NF3→NF++2F	17.9±0.3	[1]
trans-N2F2	$\{\begin{array}{l}{\mathrm{N}}_2{{\mathrm{F}}_2}^{+}\\ {\mathrm{N}}_2{\mathrm{F}}^{+}\\ \\ {\text{NF}}^{+}\end{array}$	N2F2→N2F2+	13.1±0.1	This work.
		N2F2→N2F++F	13.9±0.2	Do.
		N2F2+→N2F++F (metastable)	13.4±0.2	Do.
		N2F2→NF++NF	17.0±0.2	Do.
cis-N2F2	$\{\begin{array}{l}{\mathrm{N}}_2{\mathrm{F}}^{+}\\ {\text{NF}}^{+}\end{array}$	N2F2→N2F++F	14.0±0.2	Do.
		N2F2→NF++NF	16.9±0.2	Do.
N2F4	$\{\begin{array}{l}{\mathrm{N}}_2{{\mathrm{F}}_4}^{+}\\ {\mathrm{N}}_2{{\mathrm{F}}_3}^{+}\\ \\ {{\text{NF}}_2}^{+}\\ {\text{NF}}^{+}\end{array}$	N2F4→N2F4+	12.0±0.1	[3]
		→N2F3++F−	12.0a	This work.
		→N2F3++F−	15.6a	Do.
		See text.
		See text.

^aSingle observation.

NF₂

The ionization potential of NF₂ measured in this work was 12.0 ±0.1 ev in good agreement with that of Loughran and Mader. The average of the two values is 11.9 ±0.2 ev.

Differences in the reported NF⁺ appearance potentials from NF₂ are much greater. We observe two processes leading to the formation of NF⁺. The difference in the appearance potentials of these processes is almost equal to the electron affinity of the fluorine atom (3.6 ev) [13]. This gives considerable support to the present identification.

From

$$\begin{array}{c}{\text{NF}}_2\to {\text{NF}}^{+}+\mathrm{F}\\ \mathrm{A}\left({\text{NF}}^{+}\right)\ge \mathrm{D}\left(\text{NF}-\mathrm{F}\right)+\mathrm{I}\left(\text{NF}\right)\end{array}$$

where the inequality accounts for any excess energy involved in the reaction, we calculate an upper limit for I(NF) = 12.4±0.3 ev, assuming D(NF−F) = D(N−F) average in NF₂. This differs from the previous estimate of I(NF) = 12.0 ev [1]. However, the present value is considered the more reliable for reasons stated in the previous section.

NF₃

The two reported values for the appearance potential of the NF₂⁺ ion from NF₃ differ by 0.4 ev. Different methods of evaluating the appearance potential were used by each investigator. We also find it possible, by using different graphical methods, to interpret our data so as to obtain either limiting value from the same set of measurements. However, the appearance potential is readily calculated from the equation

$${\text{NF}}_3\to {{\text{NF}}_2}^{+}+\mathrm{F}$$

from which

$$\begin{array}{c}\mathrm{A}({{\text{NF}}_2}^{+})\ge \mathrm{D}({\text{NF}}_2-\mathrm{F})+\mathrm{I}({\text{NF}}_2)\\ \ge 14.4\pm 0.4\text{ev}.\end{array}$$

The calculated value lies just between the two limiting experimental values.

The NF⁺ appearance potential has been reported as 17.9±0.3 ev [1], and ascribed to the reaction

$${\text{NF}}_3\to {\text{NF}}^{+}+2\mathrm{F}.$$

From the relation

$${\mathrm{A}(\text{NF}}^{+})\ge \mathrm{D}({\text{NF}}_2-\mathrm{F})+\mathrm{D}(\text{NF}-\mathrm{F})+\mathrm{I}(\text{NF})$$

and the values of D(NF₂–F), D(NF–F), and I(NF) given above, we calculate A(NF⁺)≥18.0 ±0.6 ev, in good agreement with the measured value. Thus there appears to be no evidence for a lower energy process for this reaction which would result in the formation of molecular fluorine.

N₂F₂

The mass spectra of the cis and trans N₂F₂ were similar in most respects to those reported previously [5,13], However, additional very diffuse peaks in the mass spectra at nonintegral m/e ratios were observed and attributed to metastable transitions [15]. These metastable ions were observed only in the mass spectrum of the trans species. This is consistent with the fact that the cis isomer apparently produces no parent ion. The relative abundance of the metastable ion appearing at the nominal m/e=33.5 was 0.22 percent of the largest normal ion peak and was attributed to the transition, N₂F₂⁺→N₂F⁺+F. The ion appearing at m/e=16.5 was 0.02 percent of the maximum peak and was attributed to the transition, N₂F₂⁺→NF⁺+NF. Appearance potential measurements of the ions at m/e=33.5 and 16.5 ruled out the possibility of doubly charged ions.

The relatively large abundance of the m/e=33.5 metastable peak in trans N₂F₂ made it possible to measure the appearance potential of this ion with good precision. As might be expected on the basis of the statistical theory of mass spectra [12], the appearance potential is somewhat lower than that of the same ions collected at m/e=47. However, the magnitude of the difference is unexpectedly large.

The appearance potentials of the normal fragment ions NF⁺ and N₂F⁺ are identical within experimental uncertainty for both cis and trans N₂F₂. The heats of formation of the two isomers are also very similar; thus Armstrong and Marantz [16] report ΔH_f(N₂F₂) cis=16.4 kcal/mole and Δ(N₂F₂) trans=19.4 kcal/mole with an uncertainty of about 1.5 kcal/mole. Thus if there is no excess kinetic or excitational energy involved in the dissociative ionization of either of the isomers, it would appear that they are similar in molecular structure.

This argues in favor of the cis and trans designations for the N₂F₂ isomers contrary to the recent suggestion by Sanborn [5] that the isomer presently designated “cis” actually has the 1,1-difluorodiazine structure as first considered by Bauer [17].

Similarly, these data do not support the recently reported [18] heat of isomerization of 27.5±5.0 kcal/mole for the N₂F₂ isomers. However, we have been unable to calculate this value from the data as given in the reference.

On the basis of nearly equal heats of formation for the cis and trans isomers, we can calculate the N = N bond dissociation energy for either isomer of N₂F₂ from the reaction:

$${\mathrm{N}}_2{\mathrm{F}}_2\to {\text{NF}}^{+}+\text{NF}$$

and the relation D (FN=NF) ≤ A(NF⁺) − I (NF). Using the values A(NF⁺) = 17.0±0.2 ev and I(NF) = 12.4±0.3 ev, we obtain D(FN=NF)≤4.6±0.5 ev, or ≤ 106 ±12 kcal/mole.

A check on this calculation can be made using the measured values for ΔH_f (N₂F₂) and the reaction

$${\mathrm{N}}_2{\mathrm{F}}_2\to 2\text{NF}$$

from which D (FN=NF)=2ΔH_f/NF−ΔH_fN₂F₂. Using the previously calculated value for ΔH_fNF = 64.4 ±4.2, we calculate D(FN=FN) cis=106±10 kcal/mole and D (FN=NF) trans= 103 ± 10 kcal/mole.

These values may be compared with the value of D(HN = NH) = 104±6 kcal/mole in diimide as reported by Foner and Hudson [19]. However, it should be emphasized that both methods used to calculate D(FN=NF) involve a common approximation, i.e., that the bond dissociation energy D(FN−F) = D(N−F) average in NF₂. The uncertainty in these and previous calculations are conservatively estimated from the algebraic sum of uncertainties in the contributing measurements.

A summary of measured and derived thermochemical data for the N−F compounds is given in table 3.

Table 3.

Summary of thermochemical data for N—F compounds

Molecule	ΔHf	Ionization potential	Bond dissociation energy
	kcal/mole	ev	kcal/mole
NF	61.4±4.2	≤12.4±0.3	……………………………
NF2	9.8±2.1	12.0±0.1	D(N−F)av=70.5±1.6
NF3	−29.7±1.8[9]	13.2±0.2[1]	D(F2N−F) = 58.4±4.4
cis N2F2	16.4±1.5[16]		D(FN = NF)=106±10
trans N2F2	19.4±1.5[16]	13.1±0.1	D(FN=NF) =103±10
N2F4	−2.0±2.5[8]	12.0±0.1 [3]	D(F2N−NF2)=21.5±1.6