TECHNICAL REPORT

Prepared for
The Ikon Corporation
Gulfport, MS

Thermal Stabilities of Fluoroiodocarbons (FICs)

Jon Nimitz
Environmental Technology & Education Center
Albuquerque, NM USA

Executive Summary

This report summarizes and compares thermal stability data on products containing fluoroiodocarbons (FICs). These data yield rate constants and Arrhenius equations for thermal decomposition of Ikon-12 refrigerant. The Arrhenius equation allows calculation of time to unacceptable levels of decomposition as a function of temperature. Data are also given on decomposition rates of perfluoro-n-propyl iodide, perfluoro-n-butyl iodide, and perfluoro-n-hexyl iodide.

Decomposition of FICs has been monitored by five methods: ppm fluoride, ppm iodide, total acid number (TAN), CHF₃ in the vapor, and visible light absorption by molecular iodine produced. The most reliable are ppm fluoride, TAN, and CHF₃.

The most stringent criterion for refrigerant purity is the requirement of 500 ppm or less of fluoride, which corresponds to 0.24% decomposition of the CF₃I in Ikon-12 refrigerant. Table 3 shows time to 0.24% decomposition at different temperatures. Data are now required on the temperature-time profiles of the refrigerant in various systems. Once these data are obtained accurate estimates of the operational lifetime of Ikon-12 refrigerant in these systems can be made.

Introduction

The two most recurrent criticisms of fluoroiodocarbons (FICs) are that they are "highly toxic" and "thermally unstable". These criticisms appear to be based on two sources:

1. conventional wisdom in organic chemistry which indicates that iodides are less stable and more toxic than the corresponding chlorides or bromides, and
2. isolated pieces of data not analyzed in comparison to other chemicals.

It is in general true that iodides are less stable and more toxic than the corresponding chlorides or bromides. However, the presence of a large number of fluorine atoms instead of hydrogen counteracts most of these effects.

Toxicity issues have been investigated in studies sponsored by the U.S Air Force and the Ikon Corporation. FICs have been shown to have toxicities similar to those of many widely-used industrial chemicals. The EPA has determined that FICs have acceptably low toxicities for refrigeration and for firefighting in unoccupied areas.

To investigate the issue of thermal stabilities, the Ikon Corporation has sponsored or otherwise arranged testing by four organizations: Spauschus and Associates in Atlanta, GA, the EPA in Research Triangle Park, NC, the Environmental Technology & Education Center in Albuquerque, NM, and another organization which remains confidential.

The purposes of this report are to present, compare, and analyze the thermal stability data from these four sources and to develop equations to predict decomposition vs. time and useful lifetimes of FICs (before reclamation) in various applications.

The percent decomposition (d) of a chemical is a function of temperature (T) and time (t), as shown in Equation (1)¹.

d = f(T,t)

(1)

In principle, it is possible that thermal decomposition may be affected by other materials in the system. However, as a first approximation, the decomposition of FICs is considered to be a first-order reaction independent of other materials in the system. Once the FIC has undergone the initial decomposition step (cleavage of the carbon-to-iodine bond, shown in reaction [1]), the fragments may well react with other materials nearby.

R-I --> R· + I· --> other reactions

[1]

Breakdown Products

The main organic breakdown product observed from thermal decomposition of CF₃I is trifluoromethane, CHF₃, also known as HFC-23 or fluoroform. This observation is consistent with reports in the literature that several syntheses of CF₃I yield CHF₃ as a by-product, particularly under high-temperature conditions. The properties of Ikon- 12 refrigerant in which the CF₃I has undergone 1% decomposition have been calculated and are given in Table 1. These calculations show only very small (4-6%) increases in system pressure.

Table 1. Vapor Pressure of Ikon-12 Refrigerant with up to 1% Decomposition of CF3I to CHF3
Temp °F	VP CF3I PSIA	VP R-152a PSIA	VP R-23	Ikon-12 VP 0% d	Ikon-12 VP 0.24% d	P increase %	Ikon-12 VP 1% d	P increase %
-30	8.53	9.66	125.95	9.1	9.5	4	9.7	6
0	17.86	19.82	217.62	18.9	19.7	4	19.9	5
30	34.46	37.225	351.74	36.0	37.3	4	37.5	4
60	61.69	65.005	543.07	63.5	65.9	4	65.9	4

When copper is in prolonged contact with CF₃I, a trace of white powder forms. This material is copper (I) iodide, CuI. It has not yet been determined whether the formation of CuI poses a potential problem in long-term operation. The amount of copper involved is very small and any powder that breaks loose from the surface is expected to be trapped in the filter drier. It should be determined if this powder poses any threat of clogging capillary tubes or other components.

If it is desired to eliminate formation of CuI, it is likely that a special filter-drier can be designed that contains an active metal such as magnesium or zinc to scavenge iodine before it reacts with copper. Such a filter drier could also contain a relatively fine filter to remove any metal iodide particles circulating.

Methods Of Analysis of Decomposition

Five types of measurements of extent of decomposition of FICs have been used. Molecular iodine concentration (measured colorimetrically by visible spectrometry), concentration of CHF₃ (from CF₃I, measured by gas chromatography), total acid number (TAN, determined by titration), ppm fluoride ion (determined by ion chromatography), and ppm iodide ion (determined by ion chromatography) have all been used. To compare the data, the units of measurement of decomposition products had to be standardized. It was decided to standardize on percent decomposition.

To convert ppm fluoride to percent decomposition the following reasoning was applied. Parts per million corresponds to micrograms per gram, as shown in Equation (2).

500 ppm F- = 500 micrograms F-/g refrigerant

(2)

The density of Ikon-12 refrigerant is 1.71 g/mL. Thus 1% decomposition of 1 mL of refrigerant would yield 0.0171 g (17.1 mg) of breakdown products. The fraction of fluoride from CF₃I in Ikon-12 refrigerant is calculated in Equation (3).

0.718 g CF3I/g Ikon-12 x (3 moles F atoms x 19.0 g/mole F atoms)/195.9 g/mol CF3I = 0.209 = 20.9%

(3)

Thus the weight of fluoride in 1 g of Ikon-12 refrigerant which has undergone 1% decomposition of its CF₃I is obtained by Equation (4).

0.01 decomposition x 1.00 g x 0.209 g F-/g Ikon-12 x 106 micrograms/g = 2,090 micrograms F-

(4)

Thus 500 ppm fluoride corresponds to 500 micrograms F^-/g refrigerant and this corresponds to 500/2090 or 0.24% decomposition. A figure of 1 microgram F^- per g refrigerant corresponds to 1/2090 of 1% or 0.000478% decomposition.

A similar analysis for iodide indicates that Ikon-12 refrigerant is 46.51% iodide, and the weight of iodide in 1 g of refrigerant which has undergone 1% decomposition is 4,651 micrograms. Thus 5000 ppm iodide represents slightly more than 1% decomposition. A figure of 1 microgram I^- per g refrigerant corresponds to 1/4651 of 1% or 0 000215% decomposition.

The total acid number (TAN) equals the milligrams of KOH needed to neutralize one mL of refrigerant. The borderline of acceptability occurs near a TAN of 1.0. The mmol of acid corresponding to a TAN of 1.0 is shown in Equation (5).

1.0mg KOH x 1 mole/56 g = 0.0179 mmol

(5)

Equation (5) indicates that 0.0179 mmol of acid formation (i.e., decomposition) has occurred at a TAN of 1.0. One mL of refrigerant contains 6.27 mmol CF₃I as calculated in Equation (6).

1.71 g x 0.718 x 1 mole/l95.9 g = 0.00627 moles CF3I = 6.27 mmol

(6)

Decomposition of 0.0179 mmol out of 6.27 mmol represents 0.28% decomposition.

Thus among the criteria of 500 ppm fluoride, 5000 ppm iodide, or 1.0 TAN, the most stringent criterion is the 500 ppm fluoride, corresponding to 0.24% decomposition

Reaction Kinetics

A first-order decomposition process can be described by Equation (7),

C = Coe-kt

(7)

where C_o is the initial concentration, C is the concentration remaining after the decomposition, k is the first-order rate constant (in units of inverse time) and t is time. Dividing by C_o, replacing C with C_o-x (where x is the fraction decomposed), and taking natural logarithms of both sides gives Equation (8).

ln(Co/C) = ln(Co/Co-x) = kt

(8)

The rate constant k is itself a function of temperature, according to the Arrhenius equation, Equation (9),

k = Ae-Ea/RT

(9)

where k is the first-order rate constant, A is the pre-exponential term, E_a is the activation energy in units of energy/mole, R is the gas constant, and T is absolute temperature. The pre-exponential term A is a weak function of T and in many cases can be considered constant over the temperature range of interest. In the case of thermal stability studies of FICs, the temperature range of the data is from 90 to 175°C, or 363 to 448 K. The pre-exponential factor A is approximately l0¹³ or 10¹⁴s^-1 for C-X dissociation reactions (Ref). The pre-exponential factor A can be approximated as in Equation (10),

kBT/h= 1010.319T

(10)

where k_B is Boltzmann's constant and h is Planck's constant.

The activation energy for thermal pyrolysis of C-X bonds is approximately equal to the bond dissociation energy (BDE). In the case of CF₃I this bond dissociation energy is 54 kcal/mole. For other FICs the C-I BDE is expected to be similar.

Taking natural logarithms of the Equation (9) gives Equation (11).

ln(k) = ln(A) - Ea/RT

(11)

If 1/T is plotted on the X-axis and ln(k) on the Y-axis, the intercept will be ln(A) and the slope equal to -E_a/R.

The hard-sphere collision theory describes the rate constant k as a function of temperature by Equation (12).

k(T) = B(T)e-Eb/RT

(12)

In Equation (12), the pre-exponential B(T) is proportional to T^1/2, and E_b differs slightly from E_a.

The difference between T^1/2 at 363K(which is 19.05) and at 448 K (which is 21.16) is about 10%, well within the experimental uncertainties of the data. Thus it was decided to use the Arrhenius equation, Equation (9), and treat the pre-exponential factor as a constant.

Arrhenius Equations for Ikon-12 Refrigerant

The Spauschus data (including those using iodine analysis) are highly internally consistent. The data are shown in Table 2 and Figure 1. They yield Equation (13).

log k = -3992.8(1/T) + 0.9797 with R2 = 0.7179 for 24 points

(13)

The implied activation energy is therefore given by Equation (14)

3992.8 x 1.987 cal/mole x2.303 x l kcal/l000cal = 18.3 kcal/mole

(14)

The pre-exponential term is given by Equation (15)

100.9797 = 9.54

(15)

Thus the complete first-order expression for the rate constant k is given by Equation (16).

k=9.54e-18,300/1.987T = 9.54e-9210/T

(16)

The Spauschus data were combined with other confidential data on the rate of decomposition of Ikon-12, with which they were in good agreement.

For the combined data, Equation (17) holds.

log k = 5277(1/T) + 4.3758 with R2 = 0.5609 for 27 points

(17)

The implied activation energy is therefore given by Equation (18)

5277 x 1.987 cal/mole x 2.303 x 1 kcal/l000 cal = 24.1 kcal/mole

(18)

The pre-exponential term is given by Equation (19).

104.38 = 2.40 x 104

(19)

Thus the complete first-order expression for the rate constant k is given in Equation (20).

k = 2.40x104e-24,100/1.987T = 2.40x104e-12130/T

(20)

Table 2. Compilation of Spauschus Thermal Stability Data on Ikon-12 Refrigerant
Temp °C	Time			% Decomp	ln(Co/C)	1/T(K)	log (k)	ln(k)	k
	days	hours	seconds
Analysis Method: ppm F-
105	14	336	1209600	0.01	0.0001	0.0026445	-9.96	-22.92	1.11E-10
105	28	672	2419200	0.03	0.0003	0.0026445	-9.91	-22.82	1.23E-10
105	43	1032	3715200	0.06	0.0006	0.0026445	-9.82	-22.62	1.51E-10
105	56	1344	4838400	0.06	0.0006	0.0026445	-9.90	-22.81	1.25E-10
125	7	168	604800	0.03	0.0003	0.0025116	-9.32	-21.45	4.82E-10
125	15	360	1296000	0.06	0.0006	0.0025116	-9.36	-21.55	4.39E-10
125	21	504	1814400	0.10	0.0010	0.0025116	-9.24	-21.29	5.69E-10
125	28	672	2419200	0.14	0.0014	0.0025116	-9.25	-21.30	5.64E-10
150	3	72	259200	0.03	0.0003	0.0023632	-8.96	-20.62	1.11E-09
150	7	168	604800	0.14	0.0014	0.0023632	-8.65	-19.92	2.24E-09
150	10	240	864000	0.30	0.0030	0.0023632	-8.46	-19.48	3.46E-09
150	14	336	1209600	0.65	0.0065	0.0023632	-8.27	-19.05	5.36E-09
Analysis Method: ppm I-
105	14	336	1209600	0.10	0.0010	0.0026445	-9.08	-20.90	8.36E-10
105	28	672	2419200	0.17	0.0017	0.0026445	-9.15	-21.06	7.12E-10
105	43	1032	3715200	0.15	0.0015	0.0026445	-9.38	-21.60	4.16E-10
105	56	1344	4838400	0.18	0.0018	0.0026445	-9.43	-21.72	3.70E-10
125	7	168	604800	0.20	0.0020	0.0025116	-8.48	-19.52	3.34E-09
125	15	360	1296000	0.17	0.0017	0.0025116	-8.88	-20.45	1.31E-09
125	21	504	1814400	0.18	0.0018	0.0025116	-9.01	-20.76	9.68E-10
125	28	672	2419200	0.34	0.0034	0.0025116	-8.85	-20.38	1.41E-09
150	3	72	259200	0.17	0.0017	0.0023632	-8.18	-18.83	6.62E-09
150	7	168	604800	0.22	0.0022	0.0023632	-8.43	-19.42	3.68E-09
150	10	240	864000	0.43	0.0043	0.0023632	-8.30	-19.12	4.98E-09
150	14	336	1209600	0.49	0.0049	0.0023632	-8.40	-19.33	4.02E-09

Figure 1. Spauschus Data on Thermal Stability of Ikon-12 Refrigerant

Data on Other FICs

It seems a reasonable assumption that all primary FICs (i.e., those containing a -CF₂I group) studied have similar activation energies for decomposition, within the uncertainties of the data. The thermal stability data from ETEC sealed-tube tests of the FIC solvents 1-C₃F₇I, 1-C₄F₉I, and l-C₆F13I were re-examined to see if they were in agreement with the data on Ikon-12 refrigerant. Analysis of these data was complicated by the fact that many of the tubes showed a "leveling-out," a point past which net production of molecular iodine did not increase, at a very low total level of decomposition (less than 1%). It may be that in the tubes some other reactions were occurring. The best kinetic data is obtained when the decomposition data is constant, in other words when the percentage decomposition continues increasing with time. Several tubes exhibited this behavior (Figures 2-6) and their kinetics were analyzed (Table 3 and Figure 7). Data from the FIC solvents in the sealed tubes indicated much slower decomposition than that observed for CF₃I by both Spauschus and Carrier, by approximately one to two orders of magnitude. Because the sealed-tube tests of FIC solvents, monitored colorimetrically for molecular iodine formation, have more uncertainties than the Spauschus data, the solvent data have been set aside until more complete analysis can be performed.

Table 3. Compilation of ETEC Thermal Stability Data on FIC Solvents
Temp °C	Time			% Decomp	ln(Co/C)	1/T(K)	log (k)	ln(k)	k
	days	hours	seconds
FIC Product: 1-C3F7I Analysis Method: visible absorption of I2
150	42	1000	3600000	0.02	0.0002	0.0023632	-10.26	-23.61	5.56E-11
150	83	2000	7200000	0.04	0.0042	0.0023632	-10.26	-23.61	5.56E-11
150	125	3000	10800000	0.06	0.0006	0.0023632	-10.26	-23.61	5.56E-11
FIC Product: 1-C4F9I Analysis Method: visible absorption of I2
175	42	1000	3600000	0.04	0.0004	0.0022314	-9.95	-22.92	1.11E-10
175	83	2000	7200000	0.08	0.0008	0.0022314	-9.95	-22.92	1.11E-10
175	125	3000	10800000	0.12	0.0012	0.0022314	-9.95	-22.92	1.11E-10
175	42	1000	3600000	0.08	0.0008	0.0022314	-9.65	-22.23	2.22E-10
175	83	2000	7200000	0.16	0.0016	0.0022314	-9.65	-22.23	2.22E-10
175	125	3000	10800000	0.24	0.0024	0.0022314	-9.65	-22.23	2.22E-10
FIC Product: 1-C6F13I Analysis Method: visible absorption of I2
175	42	1000	3600000	0.08	0.0008	0.0022314	-9.65	-22.23	2.22E-10
175	83	2000	7200000	0.16	0.0016	0.0022314	-9.65	-22.23	2.22E-10
175	125	3000	10800000	0.24	0.0024	0.0022314	-9.65	-22.23	2.22E-10
175	42	1000	3600000	0.02	0.0002	0.0022314	-10.26	-23.61	5.56E-11
175	83	2000	7200000	0.04	0.0004	0.0022314	-10.26	-23.61	5.56E-11
175	125	3000	10800000	0.06	0.0006	0.0022314	-10.26	-23.61	5.56E-11

Figure 2. Percent Decomposition of 1-C₃F₇I vs. Time at 150°C

Figure 3. Percent Decomposition of 1-C₄F₉I vs. Time at 175°C with Filter-drier Beads

Figure 4. Percent Decomposition of 1-C₄F₉I vs. Time at 175°C with Molecular Sieves

Figure 5. Percent Decomposition of 1-C₆F₁₃I vs. Time at 175°C with Filter-Drier Beads

Figure 6. Percent Decomposition of 1-C₆F₁₃I vs. Time at 175°C with Molecular Sieves

Figure 7. Thermal Stability from ETEC Tests on C3, C4, and C6 FICs

Analysis of Time to Decomposition

Rearranging Equation (8) to solve for time gives Equation (21)

t = ln(Co/C)/k

(21)

Substituting 0.24% decomposition and Equation (20) for k gives the time (t_d to 0.24% decomposition (time to 500 ppm fluoride) as in Equation (22)

td = ln(100/99.76)/2.40x104e-12130/T = 1.00 x 10-7e12130/T

(22)

Table 4 shows time to 0.24% decomposition (t_d) as a function of temperature. Assuming that the refrigerant is effectively at its maximum temperature for about 5% of the cycle, the operational hours until 0.24% decomposition can be calculated. A more accurate treatment would involve integration of the decomposition function over the temperature curve of the cycle.

Table 4. Time to 0.24% Decomposition as a Function of Temperature
Temp °F	Operational Yrs (5% at Tmax)	Operational Hrs (5% at Tmax)	Td Hrs	Temp °C	Temp °k	12130/T	exp(12130/T)	Time to 0.24% decomp (sec)
120	1447	12,675,808	633,790	49	322	37.67	2.28E16	2,281,645,517
130	764	6,692,094	334,605	54	328	37.03	1.20E16	1,204,576,995
140	412	3,609,115	180,456	60	333	36.41	6.50E16	649,640,612
150	227	1,986,260	99,313	66	339	35.81	3.58E15	357,526,885
160	127	1,114,404	55,720	71	344	35.23	2.01E15	200,592,778
170	73	636,827	31,841	77	350	34.68	1.15E15	114,628,894
180	42	370,338	18,517	82	355	34.13	6.67E14	66,660,911
190	25	218,990	10,949	88	361	33.61	3.94E14	39,418,114
200	15	131,573	6,579	93	366	33.10	2.37E14	23,683,087
210	9.2	80,263	4,013	99	372	32.60	1.44E14	14,447,368
220	5.7	49,680	2,484	104	378	32.12	8.94E13	8,942,428
230	3.6	31,181	1,559	110	383	31.66	5.61E13	5,612,595
240	2.3	19,833	992	116	389	31.21	3.57E13	3,569,887
250	1.5	12,776	639	121	394	30.77	2.30E13	2,299,764
260	1.0	8,332	417	127	400	30.34	1.50E13	1,499,751
270	0.6	5,498	275	132	405	29.92	9.90E12	989,565
280	0.4	3,668	183	138	411	29.52	6.60E12	660,316
290	0.3	2,474	124	143	416	29.12	4.45E12	445,396
300	0.2	1,686	84	149	422	28.74	3.04E12	303,559

It is not known if this assumption of 5% of the cycle at maximum temperature is reasonable. Most likely the refrigerant is within 10°F of maximum temperature for only a very small portion of the cycle near the discharge into the condenser. ETEC has been able to obtain almost no data so far about the maximum temperatures and fractions of time that the refrigerant is near the maximum temperature. These data are needed to calculate estimated operational hours for different systems. An important task is now to collect temperature-time charts of the refrigerant as it circulates in the systems from manufacturers of refrigeration systems. Lance Lankford estimated that the maximum temperature experienced by the refrigerant in R-12 condensers is about 220°F, since many polymers will not survive temperatures above 230°F.

If the above assumptions are made, estimates of refrigerant lifetime can be made. For example, the lifetime of an automotive refrigeration system is about 2500 hours of operation. Table 4 shows that, as long as the maximum temperature of the refrigerant in the system is no greater than 290°F and for no more than 5% of the cycle, Ikon-12 refrigerant should survive 2500 hours of operation before accumulating 500 ppm fluoride. In actuality it is believed that the maximum temperature of the refrigerant in an automotive air conditioner remains below 200°F, indicating that the refrigerant should survive over 100,000 hours of operation (more than 40 times the designed life of the system).

It should be kept in mind that automotive and transport refrigeration systems leak a portion of their charges annually and therefore on average require recharging every few years. Thus it is very unlikely that the refrigerant will remain in an operating system continuously for, say, 20 years. Even if it did, the filter drier present would remove small amounts of decomposition products.

In the recharging process used refrigerant is removed and reclaimed, removing accumulated impurities. The operational hours in Table 4 can also be interpreted to mean the times at which the refrigerant in the system (assuming it has not needed recharging) should be passed through a recycling unit. It may be desirable for some types of units to recommend removal and cleaning of refrigerant (and changing the filter-drier) every five years as a standard maintenance procedure.

R-500 is an azeotropic blend of R-152a (26.2% by weight) and R-12 (73.8% by weight). Its properties are very similar to Ikon-12 refrigerant, which indicates that Ikon-12 may be an excellent replacement for R-500 as well as R-134a and R-12.

Lance Lankford has carried out an analysis of the isentropic theoretical cycles for R-12, R-500, and Ikon-12 refrigerant Table 5 shows calculated discharge temperatures assuming that all the work of compression goes to heating the gas (100% efficient compression with no heat losses) Table 5 shows that discharge temperatures in most cases are not expected to exceed 170°F. In reality, under extreme conditions and considering friction losses and non-ideal compression, the temperatures will be raised slightly, perhaps up to 180°F.

Table 5. Estimated Discharge Temperatures
Refrigerant	Evap Temp., °F	Condenser Temp., °F	Discharge Temp., °F
R-12	5	86	101
R-12	40	140	163
R-500	5	86	105
R-500	40	140	165
Ikon-12	5	86	106
Ikon-12	40	140	165

Data collected by Dole Fresh Fruit indicate that the discharge temperature on refrigerated transport units may rise to about 180°F. In any case, there is no evidence that discharge temperatures for R-12 or R-500 systems ever exceed 200°F Elastomers used in construction of compressors cannot withstand extended temperatures of greater than 230°F, so systems are not designed to be exposed to greater than 230°F. If temperatures were higher compressor materials would fail.

Footnotes

¹For the convenience of the reader, equations are numbered in parentheses and reactions are numbered in square brackets.

References

1. Herschbach, D. R., Chemical Kinetics, Butterworths, Baltimore, 1976.

2. Eyring, H, Lin, S H., and Lin, S M, Chemical Kinetics, Wiley Interscience, New York, 1980.

3. Stiller, W, Arrhenius Equation and Non-Equilibrium Kinetics, Teubner, Leipzig, 1989, pp 29-41.

4. Patai, S , ed., The Chemistry of the Carbon-Halogen Bond, parts 1 and 2, John Wiley and Sons, New York, 1973

5. Benson, S W, and O'Neal, H. E, Kinetic Data on Gas-Phase Unimolecular Reactions, NSRDS-NBS 21, U.S. Dept. of Commerce, 1970

6. Nimitz, J S., Development of Nonflammable. Environmentally Compliant Fluoroiodocarbon Solvents: Phase I Final Report, Prepared for U.S A.F Wright Laboratories under contract F33615-94-C-5003, ETEC 95-1, January 1995.