Active substances study in fire extinguishing by water mist with potassium salt additives based on thermoanalysis and thermodynamics

Accepted Manuscript Active Substances Study in Fire Extinguishing by Water Mist with Potassium Salt Additives Based on Thermoanalysis and Thermodynamics Zhang Tianwei, Liu Hao, Han Zhiyue, Du Zhiming, Wang Yong PII: DOI: Reference:

S1359-4311(16)33895-9 http://dx.doi.org/10.1016/j.applthermaleng.2017.05.053 ATE 10364

To appear in:

Applied Thermal Engineering

Received Date: Revised Date: Accepted Date:

5 December 2016 5 May 2017 11 May 2017

Please cite this article as: Z. Tianwei, L. Hao, H. Zhiyue, D. Zhiming, W. Yong, Active Substances Study in Fire Extinguishing by Water Mist with Potassium Salt Additives Based on Thermoanalysis and Thermodynamics, Applied Thermal Engineering (2017), doi: http://dx.doi.org/10.1016/j.applthermaleng.2017.05.053

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Active Substances Study in Fire Extinguishing by Water Mist with Potassium Salt Additives Based on Thermoanalysis and Thermodynamics Zhang Tianweia,b, Liu Haob, Han Zhiyuea1, Du Zhiminga, Wang Yongb State Key Laboratory of Explosion Science and Technology, Beijing Institute of Technology, Beijing 10081, China b The Chinese People’s Armed Police Forces Academy, Hebei, Langfang 065000, China Abstract: The active substances during the process of fire extinguishing by water mist with potassium salt additives was studied. The minimum extinguishing concentration (MEC) experiment results showed that K2CO3 have the most benefit in improving the fire extinguishing efficiency of pure water with the improving rate of 37.6%, 47.2% and 64.8% which the mass percent was 1%, 2% and 5%, respectively. Other potassium salt additives followed the order by: K2C2O4>CH3COOK>KNO3>KCl>KH2PO4, and the reason attributed to the different active substances decomposed from flame temperature by different kinds of potassium salt additives. Thermoanalysis and characterization of the pyrolysis products of potassium salt solutions at flame temperature were analyzed through TG-DSC, XRD, and SEM, the results showed the KOH was the main product in the high extinguishing performance of potassium salts in flame chemical reactions. Thermodynamics analysis by HSC CHEMISTRY showed the K2CO 3 could provide 4.85% KOH in equilibrium substances, which well above other potassium salts, and other active substances good for fire extinguishing benefited from the intermediate product during the process of the KOH reacted with the flame free radicals. Key Words: cup-burner; potassium salt additives; fire extinguishing; water mist; active substance; thermoanalysis; thermodynamics a

1 Introduction In recent years, as the rapid development of the economic development, energy has become an important guarantee for the development of energy issues and even directly restricts the economic development. With the increase of population density and building, as well as the increased fuel consumption, fuel structure has been changed, gaseous fuel such as coal gas, liquefied petroleum gas and natural gas gradually becomes the main fuel for city life[1]. However, the widespread use of natural gas increasing the risk of natural gas leakage and fire and explosion accidents. CH4 fire threatening the life security of the public and causing great damage to the material property. How to solve the CH4 fire better become the research focus of fire science. Water is a main method for the treatment of fire since ancient times. With the continuous development of society, the fire form is constantly updated, and with the progress of science and technology, the form of water is also constantly innovating. The method of using water to treat fire has been greatly improved, and the water mist is the most representative of fire treatment means[2]. The addition of additives in fine water mist can enhance the fire extinguishing efficiency of pure water according to various studies on the fire extinguishing efficiency of additives [3–9]. In 1

Corresponding author. Tel:+86-010-68912765 E-mail address: [email protected]

these researches, the water mist additives include KCl, NaCl, CH3COOK, LiI, NiCl2, CoCl2, NaHCO3, NaOH, CaCl2, MgCl2, FeCl2, and MnCl2. Experiments showed that the additives with the highest efficiency are the compounds of Fe and K [10]. Toxicity and cost problems resulted in the limited application of the Fe compound as an additive [11], which made potassium salts the focus of the present research. The potassium salts decomposed at the flame temperature, and the fire extinguishing occurred with the reaction between the generated substances and flame radical by the chain termination reaction. The detailed methane pyrolysis process was GRI3.0, which contained 53 species and 325 reactions [12]. Warnatz [13] proposed that due to the slow chemical reaction rate of N in the CH 4 oxidation process, some C species has been rapid oxidation of COx before reactive with some N species. Therefore, the production of some N containing species is much smaller in the whole reaction process, which most detailed chemical reaction mechanism containing N species could negligible, while OH、H、O、CH、CH2 were mainly components during methane pyrolysis. Venkata [14] also confirmed COx and H were the main species in methane decomposition. Based on a premixed flame burner, Mitani [15] proved that alkali metal salt additives can induce chemical extinguishment by eliminating free radical came from CH 4 pyrolysis through gas-phase homogeneous reaction. Zhang [16] reported that gaseous KOH could inhibit OH and H free radical in methane pyrolysis, generated intermediate transition KO and K, which combined with OH or H to generate gaseous KOH again. The content of the free radicals in the flame propagation process was decreased by catalytic reaction and indeed inhibited the pyrolysis reaction. Previous studies on active substances in fire extinguishing by potassium salts mainly through the change of OH radical concentration in flame doped with K element, fit the chemical reaction rate function and inferred possible mechanisms by designing different burner. Firedman [17] studied the extinguishing effect of K stream through a clashing flame experiment and demonstrated that KOH is the extinguishing active substance in potassium salts by combining computation of equilibrium products and possible reaction pathways. Based on the chemical equilibrium computation with the updated database, Machale [18] further confirmed that the active product in the extinguishing reaction of potassium salts is gaseous KOH. However, due to the difficulty in measuring the concentration of OH radical in the experiment, the accuracy and repeatability was uncertain. Furthermore, for lack of uniform experimental design, the results obtained from different researchers are quite different. For example, the flame reaction doped with potassium salts, Slack [19] reported the fitting results of the rate constants for the critical reactions

K  OH  M  KOH  M was 5×10-32cm6/molec2·s-1（T=2000K），while Patrick and Golden [20] calculated the same rate constants was 1.9×10-25cm6/molec2·s-1（T=2000K） ；Husain [21] got the fitting results of the rate constants expression for the raction

K  O2  M  KO2  M

was

1.89×10-22T-2.68，while Silver [22] gave the fitting result was 1.19×10-10T-0.56. These studies showed that the extinguishing active substance of potassium salts may theoretically be KOH, which has not been verified by a compelling experiment yet. In the present study, six common potassium salts K2C2O4, CH3COOK, K2CO3, KNO3, KCl, and KH2PO4 were chosen, a reliable and repeatable experimental method based on classic thermoanalysis in proving active substances during extinguishing process by water mist with

potassium salt additives was reported. In addition, the minimum extinguishing concentration (MEC) of these potassium salts were tested by an improved cup burner was also investigated. Furthermore, the relationship between the extinguishing efficiency and the amount of active substances was proposed through thermodynamics by HSC CHEMISTRY6. 2 Experimental Equipment and Method 2.1 Cup-burner Experiment Fig. 1 shows the cup burner device. The experiment method is the same as reference[7]. axial flow fan rotameter chimney

Laser particle size analyzer stainless steel cup

cylinders connect line

circulation pump

water mist generator

Fig. 1. Cup-burner apparatus

2.2 TG-DSC Experiment Thermoanalysis of chemical extinguishing substances was used to measure the dependence of the characteristic parameters of the substances on temperature. Thermoanalysis can provide thermodynamic parameters and chemical reaction kinetic parameters of materials. At present, the most common method of studying thermostability of substances include differential scanning calorimetry (DSC), thermogravimetry (TGA), and differential thermal analysis (DTA). A chemical extinguishing agent only decomposes into extinguishing active substances under burnt temperature and thereby develops its chemical extinguishment capacity after the active substances have already reacted with flame radicals. Thermoanalysis has been highly appreciated by Chinese and foreign scholars in studying the decomposition mechanism of chemical extinguishing substances and possible extinguishing active substances. TGA and DTA of six potassium salts in atmosphere were implemented by TGA-DSC synchronous tester of Switzerland METTELER Company. The reference compound was α-Al2O3, and the air inflow rate was 50.0 mL·min−1. 2.3 XRD and SEM Experiment The cup burner test system was improved to collect the pyrolysis products of water mist after ultrasonic atomization under different heating temperatures. The chimney unit was replaced by a tube type resistance furnace for electric heating. The highest heating temperature, total heating power, and temperature control accuracy were 1100 °C, 20 kW, and ±1 °C, respectively. TEX-C100 temperature controller was used. The tube type resistance furnace has one reaction tube, in which fogdrop gained from ultrasonic atomization can directly induce conversion reaction. This reaction tube was made of materials with high temperature resistance, corrosion resistance, good heat transmission, and temperature change tolerance. An axial flow fan and a product collection bottle were set at the top of the tube type resistance furnace. Collected target powder was analyzed by D8-Advance X-ray diffractometer, and its surface microstructure was observed

by the S4700 cold field emission SEM. 3 Results and Discussion 3.1 MEC of potassium salt solutions Theoretically, the MEC of water mist containing potassium salts is a fixed value. However, in this experiment, the concentration of water mist containing additives was adjusted by changing the working voltage of the atomization modules and their number. Therefore, concentration changed discontinuously, and the upper and lower limits of the MEC were observed. Extinguishing concentration using the upper and lower limits may extinguish fires successfully or unsuccessfully. Test results of MEC of six potassium salt solutions are listed in Table 1. Table 1. MEC of water mist with potassium salt additives Agent

Upper limit of MEC

Lower limit of MEC

(mass percent, %)

(mass percent, %)

pure water

12.11

10.07

1% K2CO3

7.56

6.98

2% K2CO3

6.39

5.08

5% K2CO3

4.26

3.95

1% K2C2O4

8.38

7.78

2% K2C2O4

7.96

7.25

5% K2C2O4

4.51

3.75

1% CH3COOK

8.85

6.76

2% CH3COOK

6.05

5.75

5% CH3COOK

5.04

4.63

1% KNO3

10.8

8.6

2% KNO3

8.34

6.71

5% KNO3

5.05

3.43

1% KCl

9.34

7.13

2% KCl

7.95

6.84

5% KCl

6.97

4.62

1% KH2PO4

9.16

7.68

2% KH2PO4

10.12

8.48

5% KH2PO4

10.63

8.91

In conclusion, potassium salt additives can improve the fire extinguishing efficiency of water mist. The degree of improvement in the fire extinguishing efficiency of the said additives is ranked as follows: K2CO3>K2C2O4>CH3COOK>KNO3>KCl>KH2PO4. Additives influence fire extinguishing efficiency mainly by increasing chemical actions and reducing the competition between the evaporative capacities of pure water. The fire extinguishing efficiency of K2CO3, K2C2O4, CH3COOK, KNO3, and KCl solutions increased with increasing concentration. This effect indicates that the degree of increase in chemical actions is greater than the weakening of the evaporative capacity of pure water. The fire extinguishing efficiency of KH 2PO4 solution decreased with increasing concentration. Hence, the chemical fire extinguishing capacity of the two salts is weak. Increasing concentration dominates the competitive relation for reducing the evaporative capacity of pure water. 3.2 Thermal analysis of potassium salts

TG-DSC results of the six potassium salts at a temperature rate of 10 °C·min−1 are shown in supplementary materials and the results are shown in Table 2. Table 2. Thermal decomposition data of potassium salts in air Potassium salts K2CO3

Temperature range (°C)

theoretical value

actual value

62.54–125.27

K2CO3  2H 2O  K2CO3  2H 2O

20.69

21.15

84.13–110.32

K2C2O4 H 2O  K2C2O4  H 2O

9.783

10.68

510.92–613.34

K2C2O4  K2CO3  CO

16.87

15.70

2CH3COOK  K2C2O4  C2 H 6

15.31

13.78

K2C2O4  K2CO3  CO

16.87

13.01

60.54

66.72

K2C2O4

CH3COOK

Weight loss (%) Thermal decomposition procedure

313.07–606.85

2KNO3  2KNO2  O2

KNO3

602.44–740.82

4KNO2  2K2O  2 N2  3O2

207.98–271.01

2KH 2 PO4  K2 H 2 P2O7  H 2O

6.62

6.9

291.76–358.78

K2 H 2 P2O7  2KPO3  H 2O

7.09

7.64

KH2PO4

,Table 2 shows that the final decomposition products of K2CO3, K2C2O4 and CH3COOK are all K2CO3, and anhydrous K2CO3 within the temperature range of 25–800 °C is very stable and shows almost no loss of weight until the compound reached its 901 °C melting point, in which it decomposed until the K2O boiling point temperature [23]. Kuang [24] showed that K2CO3 had no suppression efficient compare to KHCO3 mostly because of stability. Huttinger [25] concluded that K2CO3 could easily react and produce gaseous KOH only combine with water vapor under high temperature. Based on the above deduction, it is difficult for solid K2CO3 to decompose in the general fire temperature.

DSC curve for K2C2O4 shows the two endothermic peaks and one exothermic peak. The corresponding TG curve shows two weight loss platforms. The first endothermic peak is within 84.13–110.32 °C, and the peak temperature is 102.16 °C. The endothermic peak area or decomposition enthalpy is 228.17 J·g−1. According to the platforms on the TG curve and the weight loss percentage, the deduced weight loss rate is 10.68%. The second endothermic peak is within 374.25–411.34 °C, and the peak temperature is 390.37 °C. The TG curve reveals that no weight loss occurred during this process. According to the test results of the X-4 microscopic melting point apparatus, the peak temperature is not the melting temperature and may be the endothermic peak caused by crystal transfer of K 2C2O4. The exothermic peak is between 510.92–613.34 °C, and the peak temperature is 592.54 °C. The TG curve reflects K2C2O4 decomposed within this range, releasing abundant heat and the enthalpy is 446.90 J·g−1. Ren[26] reported the decomposition of K2C2O4 also had two steps: the first step corresponding the

temperature range of 67.19-123.44 ℃ ,with weightlessness rate of 9.83%; the second step corresponding the temperature range of 565.04-689.84℃,with weightlessness rate of 15.02%. The results quite agreement with the present study testified the decomposition process of K2C2O4, and the experimental error is within the allowable range. DSC curve for CH3COOK shows the endothermic peak and two exothermic peaks. The corresponding TG curve has two weight loss platforms. The endothermic peak is within 285.16–316.74 °C, and the peak temperature is 303.56 °C. The TG curve reveals that no weight loss occurred during this process. According to the test results of the X-4 microscopic melting point apparatus, the peak temperature is the melting point of CH3COOK. The TG curve reflects CH3COOK decomposed within the range of 313.07–606.85 °C with two corresponding exothermic peaks on the DSC curve, which shows that the thermal decomposition of CH3COOK is completed in two steps. The first exothermic peak temperature is 416.77 °C,with the weight loss rate of 13.78%, and the exothermic enthalpy is 1220.60 J·g−1. The second exothermic peak temperature is 488.68 °C, with the weight loss rate of 13.01%, and the exothermic enthalpy is 926.52 J·g−1. Fleming[27] showed that the aqueous of CH3COOH had better performance of extinguishing than pure water illustrated the CH3COOH could decomposed in flame temperature and demonstrated the results of Huttinger.

DSC curve for KNO3 shows the three endothermic peaks and the corresponding TG curve has one weight loss platform. The first endothermic peak is within 128.47–150.16 °C, and the peak temperature is 130.92 °C. The TG curve reveals that no weight loss occurred in this process. According to the test results of the X-4 microscopic melting point apparatus, the peak temperature is not the melting temperature but may be the endothermic peak caused by crystal transfer of KNO3. The second endothermic peak is within 300.14–342.16 °C, and the peak temperature is 326.5 °C. According to the test results of the X-4 microscopic melting point apparatus, the second peak temperature is the melting point of KNO3, and the melting enthalpy is 67.04 J·g−1. The TG curve reflects the decomposition of KNO3 within the range of 602.44–740.82 °C with one corresponding endothermic peak on the DSC curve, a peak temperature of 681.46 °C, and weight loss rate of 66.72%. The reaction absorbs a lot of heat when the endothermic enthalpy is 1867.55 J·g−1. Zhang [28] described the decomposition process of KNO3 in oil flame environment and pointed out that K2O, which generated from the decomposition reaction was the main substance in fire suppressing. While the experiment results of Wang [29] showed that the decomposition process of KNO3 would be activate in the presence of C condition, which demonstrated the decomposition process and products. DSC curve for KCl shows that the endothermic peak is within 758.61–779.18 °C, and the peak temperature is 769.62 °C, indicating that the endothermic peak is the melting point based on the melting point of KCl (771 °C) [23]. The TG curve indicates that KCl is very stable, and almost no weight loss occurred before the melting point. When the KCl reached the melting point, due to the high saturated vapor pressure of KCl, a part of the KCl volatile gas was blown off from the crucible as driven by the purge gas, which led to the loss of weight. However, unlike solid K2CO3, KCl showed a better extinguishing efficiency by Joseph [30]. Rosser [31] explained this chemical

extinguishment effect of some salts by a volatility experiment using alkali metal salts, perfectly explained why the material is stable but also has a certain fire extinguishing efficiency. DSC curve for KH2PO4 shows the two endothermic peaks and the corresponding TG curve has two weight loss platforms, which indicate that the thermal decomposition of KH2PO4 is completed in two steps. The first endothermic peak temperature is within 207.98–271.01 °C, with the peak temperature is 235.85 °C, weight loss rate is 6.9%, and endothermic enthalpy is 274.09 J·g−1. The second endothermic peak temperature is within 291.76–358.78 °C, whose peak temperature is 336.48 °C, weight loss rate is 7.64%, and endothermic enthalpy is 30.68 J·g−1. The TG curve shows that the thermal decomposition of KH2PO4 was completed at 360 °C, and no mass loss was observed at the test temperature of 800 °C. Du [32] also reported the decomposition process of KH2PO4 with two steps, one with the temperature range of 190-270°C, and the other was 285-470°C, which was not too different from the present study. Zhang [2] concluded the suppression efficiency of KH2PO4 by the covering effect of the stable white glassy state KPO3, which generated from the decomposition of KH2PO4. Mojumdar [33] declared that with the increase in temperature rate, the decomposition peak of potassium salts skews to high temperature and becomes much sharper gradually. The decomposition peak temperature of potassium salts under different temperature rates varies. Decomposition activation energy of different potassium salts can be acquired by the Kissinger method. The Kissinger equation is as follows:  A R E 1    ln  '2   ln  k   k ' , i  1, 2, , n  TP   Ek  R TP

In the equation, β is the heating rate, °C·min−1; T'P is the peak temperature, K; Ak is the pre-exponential factor, s−1; Ek is the activation energy, kJ·mol−1; R is the gas constant, J·mol−1·K−1.    1 ln  '2  '2 T T The expression  P  to P was mapped as the proper mechanism function with Ek as the

slope and Ak as the intercept. The decomposition thermodynamic data of potassium salts are listed in Table 3.Table 3. Decomposition thermodynamic data of potassium salts Potassium salts

Fitting curve

r

Ek/ kJ·mol−1

Ak/ s−1

K2CO3

y=8.91777−6.81801x

0.89412

56.67

0.51×105

y=18.91288−10.66307x

0.98206

88.65

1.74×109

y=51.27881−53.90384x

0.87462

448.16

4.05×1026

y=29.32699−27.88944x

0.66278

231.82

1.52×1014

y=8.09588−14.52782x

0.96207

120.76

0.48×105

y=-0.36433−10.83982x

0.70798

90.10

7.53

y=15.26101−12.89462x

0.98187

107.18

5.47×107

y=3.4327−8.37709x

0.86028

69.63

259.35

K2C2O4

CH3COOK

KNO3

KH2PO4

3.3 Decomposition products analysis of potassium salts The temperature of the heating tube was set according to the initial decomposition temperature of different potassium salts and was not higher than the limit temperature of the heating tube, more significantly, as 800℃ represented most of the actual flame temperature，this temperature existed in any set of experiments. Decomposition products were collected and analyzed, the experimental equipment was followed by Supplementary Figure 7. The characterization of decomposition products under different temperature Temperature of the tube type resistance furnace was controlled at 800, 900, and 1000 °C, respectively. Driven by air, 5% (mass concentration) ultrasonically atomized KCl water mist was added into the heating tube, the products are collected and the XRD analysis is shown in Fig. 2. [KCL-800-200-30.raw] 75-0296> Potassium Chloride - KCl

40

(200)

(220)

(111)

x10^3

(311) (222)

(400)

(311) (222)

(400)

(311) (222)

(400)

Intensity(Counts)

[KCL-900-200-30.raw]

40

(200)

(220)

(111)

x10^3 [KCL-1000-200-30.raw]

40

(200)

(220)

(111)

x10^3 10

20

30

40

50

60

Two-Theta (deg)

Fig.2. XRD results of heating product of water mist with 5% KCl

Through contrast with PDF standard card, particles gained from ultrasonic atomization of KCl under different temperatures proved to be the diffraction peaks of KCl. XRD spectra of the product samples at different temperatures are similar with each other. Peaks for other impurity phases were absent, and the main diffraction peak is sharp, indicating the perfect crystal structure of the products. According to XRD spectra in Fig. 16, KCl particles could be gained from ultrasonic atomization of KCl at 1000 °C. SEM images of the product particles are shown in Fig. 3.

800℃

900℃

1000℃

Fig. 3. SEM results of heating product of water mist with 5% KCl

Product particles at 800 °C were typical crystalline polymers with clear edges and corners. Although some particle clusters were present, the disparity remained good, and distinct cubic structure could be detected under 2,500 amplifications. Crystals appeared, and the particle size was about 5 µm. Particles at 900 °C also showed good disparity, but particle size further increased,

and the cubic structure gradually became rounder, indicating the spheroidization of the particles. Particle size at 1000 °C increased continuously, and particle clustering decreased. However, particle size increased to different extents, resulting in the poor particle uniformity. Particle edges and corners were significantly round, and spherical particles could be clearly observed under 8,000 amplifications. Driven by air, 5% (mass concentration) ultrasonically atomized water mist with other potassium salt additives was added into the heating tube, and the products were collected for XRD and SEM analysis.The characterization results of decomposition products by XRD under different temperature are shown in Table 4, XRD and SEM pictures are shown in supplementary materials. Table 4 Characterization results by XRD under different temperature Characterization results by XRD

Potassium salts

300℃

500℃

600℃

K2CO3 K2C2O4

K2C2O4

CH3COOK

CH3COOK

KNO3

K2C2O4, KOH, K2CO3 KOH, K2CO3 KNO3

KCl KH2PO4

KH2PO4, KPO3

KPO3

800℃

900℃

1000℃

KOH, K2CO3

KOH, K2CO3

KOH, K2CO3

KOH, K2CO3 KOH, K2CO3 KO3, KO2,

KO3, KO2,

KNO2,KOH

KNO2,KOH

KCl

KCl

KCl

KPO3

As shown in Table 1 and Table 4，fire suppression efficiency and the behavior of potassium salts at high temperature depending on the anion of potassium salts. The decomposition products contained KOH and potassium oxides(KxOy), e.g. KO2 and KO3 in water mist with potassium salt additives with higher chemical fire extinguishing efficiency. Kong [34] and Liu [35] calculated the stability of alkali metal oxides through thermodynamics and showed that the final stable product in excess air could only be KO2. However, large number of KO3 existed in the decomposition products of water mist with KNO3 additive in current research could be only from further oxidation of KO2, which inferred the KO2 might also be the active substance besides KOH during fire suppression. In addition, KOH has a good stability, that means KOH can not only act as an intermediate product, but also stabilize in the product because of the poor stability of intermediate product during the thermal decomposition of potassium salt. Chen [36] reported that within the range of inorganic compounds containing potassium ions, only KOH and K2CO3 had activity under low temperature which could relatively easy to hydrolyze K xOy compounds in water vapor environment. Decomposition products of water mist with KH2PO4 contains a lot of KPO3, however, the reason for chemical extinguishing efficiency is not high partly due to higher stability of KPO3, which causes it difficult to open up the chemical bond and releases potassium ions in the flame temperature; on the other hand, no KOH or KxOy found in the products, which indirectly proved that KOH is a good active substances during fire extinguishing. There also has no KOH or KxOy found in the products of water mist with KCl, however, due to the low melting point of KCl, K-Cl bond breaks in the form of melting, potassium ions are released and participate in fire extinguishing through chain termination reaction. Characterization results by SEM shows that the particle size of the products increased in

different degrees with the increase of the temperature, which is due to the influence of temperature on the formation and growth of crystal particles. The temperature changes the internal energy of the atom, that means particles with smaller average size can be obtained under low temperature for its beneficial to the formation of crystal particles, which is not conducive to the growth of crystal particles. The increase of temperature is equivalent to supply the kinetic energy of the crystal growth, which increases the mass transfer coefficient of the reaction and accelerates the growth of crystal particles, leading the particle size grow up [37]. The characterization of decomposition products under different concentration The temperature is the main effect factor on decomposition of water mist with potassium salts, while the concentration may be one of the factors on the size and uniformity of decomposition particle, the phenomenon which was found in many nano materials preparation process [38]. Guo [39] shows that the concentration could not change the kind of the products but the particle size. Theoretically, the solute content increases with increasing concentration under the same size of droplet. Once the droplet enters the reaction zone, the solvent evaporations and the solute separated out under the flame temperature. When the solvent evaporates completely, the temperature of the precipitated crystal rises rapidly to the ambient temperature and has a chemical reaction temperature condition. The higher concentration leads to the probability of particle collision and combination greatly increased, which inevitably leads to large particles. Large particles indirectly increase the decomposition time of the chemical particles, which is not conducive to the formation of chemical extinguishing active substances. In addition, the droplet diameter increases with the increase of the viscosity and surface tension of the solution, then, smaller particle size is conducive to product when the solution concentration is low [40]. SEM of the decomposition products of the K2C2O4, CH3COOK and K2CO3 aqueous solution under 800℃ are shown in Fig.4.

1% K2C2O4

1% CH3COOK

1% K2CO3

2% K2C2O4

2% CH3COOK

5% K2C2O4

5% CH3COOK

2% K2CO3

Fig.4 SEM of the decomposition products of the K2C2O4, CH3COOK and K2CO3 aqueous solution under 800℃

5% K2CO3

As shown in Fig.4, the morphology and particle size of the product is affected by concentration, and the general trend is that the particle size increases with the increase of solution concentration with more obvious agglomeration. When the mass fraction of the solution is less than 5%, the interface between particles is clear. The phenomenon of inter particle connection becomes obvious when the concentration increases, and finally agglomerated together to form a larger size particle. The experimental results show that, when the concentration is low, the dispersion and uniformity of the product particles is better, and the average particle size is relatively small. Hamins [41] reported that the time for full chemical reaction of alkali metal salt additives in flame reaction had a competitive relation with the residence time of alkali metal salt particles. If the residence time of alkali metal salt particles is shorter than the reaction time, then the fire-extinguishing agent cannot complete the chemical extinguishment. Small size is beneficial for chemical particle to use the least time to release the active substances. As shown in Table 1, K2CO3 has better suppression efficiency corresponding higher concentration than K2C2O4 and CH3COOK partly due to the smaller particle size under the same concentration. Similarly, the pyrolysis products of KNO3, KCl and KH2PO4 solution at different concentrations have similar rules and no more details here. 3.4 Fire extinguishing mechanism analysis of water mist with potassium salts Mitani [15] reported that when the droplet size is smaller than 5 µm, liquid drops serve as the carrier of powder rather than the extinguishing agent. After quick evaporation of liquid drops at flame temperature, micro solid powders or decomposition products of powders enter into the reaction zone of the diffusion flame. In this experiment, particle sizes of both ultrasonically atomized pure water and water mist with additives were found to be smaller than 5 µm by the laser particle analyzer. Therefore, water was the powder carrier in this experiment. Based on the calculated results of thermoanalysis activation energy, the decomposition reactions of K 2C2O4 and CH3COOK are sensitive to temperature, and the salts decomposed into K2CO3 immediately at flame temperature. The possible chemical reactions and reaction pathways of K 2CO3 in vapor are as follows:

K2CO3  H 2O  2KOH  CO2 (1) K2CO3  K2O  CO2

(2)

2KOH  K2O  H 2O (3) K2O  H 2O  2KOH

(4)

Variations in standard Gibbs free energy of the above four reactions at 298–1400 K are shown in Fig. 5.

Reaction(1) Reaction(2) Reaction(3) Reaction(4)

400 300

ΔGθ (kJ)

200 100 0 -100 -200 200

400

600

800

1000

1200

1400

Temperature (K)

Fig. 5. Standard Gibbs free energy of reaction (1)–(4)

Fig. 5 shows that the chemical bond of K2CO3 is easier to break in the presence of vapor. KOH can be produced directly by Reaction 1 or gained indirectly by producing intermediate product K2O (Reaction 2) first and then through Reaction 4. Although the standard Gibbs free energy of Reaction 4 is smaller than 0, and Reaction 4 can be activated automatically without any external conditions, the reaction must first produce K2O through Reaction 2. Provided that the standard Gibbs free energy of Reaction 2 is relatively large, and external reaction conditions are very strict, the reaction producing KOH is apt to be Reaction 1, indicating the possibility of inducing the hydrolysis reaction of K2CO3 first than the decomposition reaction under vaporization conditions. This condition agrees with the mechanism of carbonate reduction by C of McKee et al [42,43]. On the other hand, the standard Gibbs free energy of Reaction 4 is smaller than 0, which also implies the unstable existence of K 2O. K2O is converted into KOH in the presence of H2O. KOH existence was also verified by the XRD test. The decomposition reaction (5) of KNO3 includes the formation reaction (6) of KNO2 and the decomposition reaction (7) of KNO2:

2KNO3  K2O  NO  NO2  O2 (5) 2KNO3  2KNO2  O2 (6) 2KNO2  K2O  NO  NO2 (7) The standard Gibbs free energy of KNO2 at temperatures higher than 298 K is inapplicable [23]. Therefore, thermodynamic calculation was accomplished according to Reaction 5. Possible reaction of KNO3 in the presence of vapor is as follows:

2KNO3  H 2O  2KOH  NO  NO2  O2 (8) Variations in the standard Gibbs free energy of Reactions 5 and 8 at 298–700 K are shown in Fig. 6.

Reaction(5) Reaction(8)

600 550

ΔGθ (kJ)

500 450 400 350 300 250 300

400

500

600

700

Temperature (K)

Fig. 6. Standard Gibbs free energy of Reactions (5) and (8)

In the presence of vapor, KNO3 easily decomposes into K2O, which can automatically produce KOH through hydrolysis reaction. Activity of KNO 3 results from the decomposition reaction (5), and KNO2 is a possible intermediate product of the reaction. Although thermodynamic calculation shows that Reaction 5 has high ΔGθ, its existence can be proven by thermoanalysis. Based on thermodynamic calculation, Liu [35] pointed out that KO2 is the stable product of K in excessive air and can react with excessive air to form KO3. Hence, the XRD test revealed that the final products of the complicated decomposition reaction of KNO 3 under the existence of vapor are mixtures of KOH, KO2, and KO3. The possible reaction of KCl in the vapor atmosphere is as follows:

KCl  H 2O  KOH  HCl

（9）

KOH was not observed in the XRD analysis of the product, indicating that the reaction conditions are strict. Based on the thermoanalysis, solid KCl melts when the temperature exceeds 1044 K, which can further accelerate the hydrolysis reaction of KCl. On the other hand, provided that alkali halide has high saturated vapor pressure [10], some accelerated hydrolysis reactions may be attributed to gaseous hydrolysis. KOH is the only possible product in KCl activation, which is consistent with the research conclusions of McKee and Verra et al[42,43]. Verra believed that K2O was the product of the first-step KCl reaction under the existence of gaseous vapor, but the compound can only be the intermediate product under the existence of vapor. Hydrolysis of KCl requires high temperature. Given the high saturated vapor pressure of KCl, evaporation of KCl accelerates its hydrolysis reaction and offsets the disadvantage of low burnt temperature. Based on thermoanalysis and XRD test, reaction products of KH2PO4 under the existence of vapor were stable and were composed mainly of glassy KPO3. No KOH was found. Equilibrium calculation of products of CH4/air burning system with 5% potassium salt solutions was implemented by HSC Chemistry 6. Results are shown in Fig. 7and Fig.8. 2.0

mol

2.0

K2CO3

1.5

1.5

1.0

1.0

KOH

0.5

mol KNO3

KOH

0.5

KO2

KNO2 0.0

Temperature

0

200

400

600

800

(a) 5% K2CO3

1000

1200 C

0.0

Temperature

0

200

400

600

800

(b) 5% KNO3

1000

1200 C

2.0

mol

2.0

KCl

1.5

1.5

1.0

1.0

0.5

0.5

0.0

Temperature

0

200

400

600

800

1000

1200 C

0.0

mol KH2PO4

KPO3

Temperature

0

100

(c) 5% KCl

200

300

400

500

600

700 C

(d) 5% KH2PO4

Fig. 7. Results of equilibrium calculation

3.69

1.91

KOH(g)

3.33E-08 1.02E-14 9.96E-15 0.0423 4.75E-15 1.34E-05 6.90E-16 7.41E-06 2.00E-25 2.26E-05 7.87E-24 0.0264 1.56E-06 0.0147 6.41E-11 9.86E-04 3.90E-11 0.0992 7.69E-05 0.0549 2.32E-07 0.00206 9.22E-08 1.40E-04 8.88E-07 K2O2H2(g) 3.08E-07 K2(OH)2(g) K(g) KO(g) K2O(g) K2O2(g) 5%K2CO3

5%KNO3

5%KCl

5%KH2PO4

Fig.8. The amount of equilibrium species by different potassium salts

Fig.7 and Fig.8.show that KOH exists as equilibrium products of K2C2O4, CH3COOK, K2CO3, and KNO3 in the flame reaction zone. We speculated that KOH is the key component in producing extinguishing active substances. At the same time, KNO 3 can reduce the free radical OH concentration significantly. The SEM experiment showed that under high temperature, pyrolysis products of KNO3 follow a spheroidization trend, which is beneficial in developing the extinguishing performance and offsetting the disadvantage of lower KOH percentage than K 2CO3. Moreover, the extinguishing performance of KNO3 observed in the cup burner experiment was basically consistent with those of K2C2O4, CH3COOK, and K2CO3. The chemical extinguishing performance of KCl was caused by the high percentage of KCl in the equilibrium products, indicating that KCl has certain volatility at high temperature, and its chemical extinguishing performance was influenced to a large extent by the few KOH produced by gaseous hydrolysis reaction. In previous studies, Jensen and Jones [44] had proposed the mechanism of potassium salts to inhibit fire radical as follows, which M is a third body participates in chain reactions:

KOH ( g )  H  H 2O( g )  K （10） K  OH  M  KOH ( g )  M （11） Hynes[45] and Slack[19] showed that the suppression kinetic mechanism containing

potassium salts should also contain XO and XO2, revealed the possible reaction path and holed that the importance of the reaction (12) is equivalent to the reaction (11)

K  O2  M  KO2 ( g )  M （12） KO2 ( g )  H  KO( g )  OH （13） Williams and Fleming[46] considered all the researchers and put forward a suppression kinetic mechanism model containing sodium species:

NaOH ( g )  H  H 2O( g )  Na （14） Na  OH  M  NaOH ( g )  M （15）

NaOH ( g )  OH  NaO( g )  H 2O( g ) （16） Considering the similarity of sodium and potassium, the active substances from the decomposition products of water mist with potassium salts for chemical kinetics mechanism should include at least the species KOH, KO2 KO and K, and this conclusion can be expressed by Fig.8.

Fig.8. The path of active substances from decomposition of water mist with potassium additives

Potassium salt with oxygen containing can be used to provide active substances such as KOH and KxOy in fire extinguishing under the condition of water vapor. However, for the KCl-water mist, Zheng [4] pointed out the promotion of suppression might from Cl2 radical which produced by Cl radical from KCl at flame temperature.

Cl  Cl  M  Cl2  M (17) Then, as an inhibitor, Wilson [47] reported the chemical mechanism path by reaction 18-20.

Cl2  H  HCl  Cl (18) HCl  H  H 2  Cl (19) H  OH  Cl  H 2O  Cl (20)

In summary, the extinguishing occurred not only requires the use of active substances of decomposition, but also considering the synchronization of water, the formation of K cations, flame radical, and other anions.

4 Conclusions The current research enriches the database of minimum fire extinguishing concentration of water mist with some potassium salts. The active substances from decomposition in flame temperature of water mist with some potassium salts are creatively obtained by thermoanalysis experimental method which has more reliability and accuracy, and the quantitative analysis of active substances in fire suppression is obtained through thermodynamics method by HSC CHEMISTRY 6. The main conclusions of the present research are: (1) Potassium salts chosen from current research can improve the extinguishing performance of pure water, and K2CO3 has the most benefit with the improving rate of 37.6%, 47.2% and 64.8% which the mass percent is 1%, 2% and 5%, respectively, followed by K2C2O4, CH3COOK, KNO3, KCl and KH2PO4. (2) Active substances that participate in extinguishment under gas vapor were consistent and unrelated with the type of potassium salts used. Therefore, extinguishing performance of different potassium salts is due to the activation behaviors of potassium salts under gasification conditions. Potassium salts, which have high extinguishing performance, produce KOH during flame chemical reactions. As the key component of activation of potassium salts, KOH can further react with free radicals in flames to produce extinguishing active substances. (3) Higher fire extinguishing efficiency corresponds to more active substances. The main potassium salts such as K2CO3, KNO3, KCl and KH2PO4 provides active substances KOH 4.85%, 1.3%, 5.13×10−13 % and 7.16×10−30 % respectively, which indicates the influence of anions on the fire extinguishing efficiency. Furthermore, the MEC of some other kinds of potassium additives should be obtained, and for the engineering application of water mist with potassium salt additives, physical and chemical properties of fire extinguishing agent, mostly the toxicity, for example, the emission of Cl, P and N species whether be harmful to firefighters or environment should be the further work. References [1] D. Das, T. N. Veziroglu, Hydrogen production by biological processes: a survey of literature, International Journal of Hydrogen Energy 1(2001)13-28. [2]T. W. Zhang, Z. Y. Han, Z. M. Du, Application of thermal mechanism to evaluate the effectiveness of the extinguishment of CH4 /air cup-burner flame by water mist with additives, International Journal of Hydrogen Energy 33(2016)15078-15088. [3] H. Shilling, B. Z. Dlugogorski, E. M. Kennedy, Extinction of diffusion flames by ultra fine water mist doped with metal chlorides, Proc. of the Sixth Australasian Heat and Mass Transfer Conf ., Begell House, New York , (1996)275–282. [4] R. Zheng, B. Rogg, K. N. Bray. Effect of sprays of water and NaCl-aqueous solutions on the extinction of laminar premixed methane-air counterflow flames, Combustion Science and Technology 126(1997)389–401. [5] A. K. Lazzarini, R. H. Krauss, H. K. Chelliah, Extinction conditions of non-premixed flames with fine droplets of water and water/NaOH solutions, Proceedings of the Combustion Institute

28(2000)2939–2945. [6] B. Mesli, I. Gokalp, Extinction limits of opposed jet turbulent premixed methane air flames with aerosols of water and NaCl-aqueous solution, Combustion Science and Technology 153(2000)193–211. [7] R. Hirst, K. Booth, Measurement of flame extinguishing concentrations, Fire Technology 13(1977)296–315 . [8] A. Hamins, Flame extinction by sodium bicarbonate powder in a cup burner, Proceedings of the Combustion Institute 27(1998)2857–2864 . [9] G.T. Linteris, V. R. Katta, F. Takahashi, Experimental and numerical evaluation of metallic compounds for suppressing cup-burner flames, Combustion and Flame 138(2004)78–96 . [10] O. P. Korobeinichev, A. G. Shmakov, V. M. Shvartsberg, Testing organophosphorus, organofluorine, and metal-containing compounds and solid-propellant gas-generating composition doped with phosphorus-containing additives as effective flame suppressants, Combustion Explosion and Shock Waves 42(2006) [11] G.T. Linteris, M. D. Rumminger, V. I. Babushok, Inhibition by ferrocene and blends of inert and catalytic agents, Proceedings of the Combustion Institute 28(2000)2965–2972. [12] H. Wang, E. R. Hawkes, J. H. Chen, A direct numerical simulation study of flame structure and stabilization of an experimental high Ka CH4/air premixed jet flame, Combustion & Flame 180(2017)110-123. [13] J. Warnatz, Rate Coefficients in the C/H/O System, Combustion Chemistry. Springer US, (1984)197-360. [14] V. Dasireddy, B. Likozar, Activation and Decomposition of Methane over Co, Cu and Fe-based Heterogeneous Catalysts for COx-free Hydrogen and Multi-walled Carbon Nanotubes Production, Energy Technology (2016). [15] T. Mitani, T. Niioka, Extinction phenomenon of premixed flames with alkali metal compounds, Combustion & Flame 1(1984)13-21. [16] T. W. Zhang, H. Liu, Z. Y. Han, Numerical Model for the Chemical Kinetics of Potassium Species in MethaneAir Cup-Burner Flames, Energy & Fuels 31(2017)4520-4530. [17] R. Friedman, J. B. Levy, Inhibition of opposed jet methane-air diffusion flames. The effects of alkali metal vapours and organic halides, Combustion & Flame 63(1963)195-201. [18] E. T. Mchale Flame inhibition by potassium compounds, Combustion & Flame 75(1975)277–279. [19] M.Slack, J. W. Cox, A. Grillo, Potassium kinetics in heavily seeded atmospheric pressure laminar methane flames, Combustion & Flame 3(1989)311-320. [20] R. Patrick, D. M. Golden ChemInform Abstract: kinetics of the reactions of amidogen radicals with ozone and molecular oxygen, Chemischer Informationsdienst 22(1984)4123-4128. [21] D.Husain, Y. H. Lee, P. Marshall, Temperature dependence of the absolute third-order rate constant for the reaction between K+O2+N2 over the range 680–1010K studied by time-resolved atomic resonance absorption spectroscopy, Journal of the Chemical Society Faraday Transactions 2(1987)143-154. [22] J. A. Silver, M. S. Zahniser, A. C. Stanton, Temperature dependent termolecular reaction rate constants for potassium and sodium superoxide formation, Symposium on Combustion 1(1985)605–612.

[23] J. G. Speight, Lange's handbook of chemistry, 15th ed., New York, McGRAW-HILL Section 5(1999)8-20. [24] K. Kuang, W. K. Chow, X. Ni, Fire suppressing performance of superfine potassium bicarbonate powder, Fire & Materials 6(2011)353–366. [25] K. J. Huttinger, R. Minges, Influence of the catalyst precursor anion in catalysis of water vapour gasification of carbon by potassium : 1. Activation of the catalyst precursors, Fuel 8(1986)1112-1121. [26] N. Ren, J. J. Zhang, Study on thermal decomposition kinetics of the potassium oxalate monohydrate, Journal of Hebei Normal University(Nature Science Edition) 4(2005)376-379. [27] J. W. Fleming, N. A. Jaffa, R. Ananth, Assessment of the Fire Suppression Properties of Mists of Aqueous Potassium Acetate, A Technical Working Conference Orlando, FL (2010)16-19. [28] T. W. Zhang, Z. M. Du, Z. Y. Han, Performance evaluation of water mist with additives in suppressing cooking oil fires based on temperature analysis, Applied Thermal Engineering 102(2016)1069-1074. [29] H. E. Wang, R. Q. Shen, Y. H. Ye, Thermal Behaviour and Decomposition Reaction Kinetics of Graphite/potassium Nitrate, Chinese Journal of Energetic Materials 6(2012)731-734. [30] P. Joseph, E. Nichols, V. Novozhilov, A comparative study of the effects of chemical additives on the suppression efficiency of water mist, Fire Safety Journal 6(2013)1131-1144. [31] W. A. Rosser, S. H. Inami, The effect of metal salts on premixed hydrocarbon-air flames, Combustion & Flame 63(1963)107-119. [32] X. Du, Z. M. Du, L. S. Zhao, Study on the feasibility of common potassium salts applied to superfine powder fire-extinguishing agent, Industrial safety and environmental protection 9(2013)61-64. [33] S. C. Mojumdar, M. Sain, R. C. Prasad, Selected thermoanalytical methods and their applications from medicine to construction, Journal of Thermal Analysis and Calorimetry 3(2007)653-662. [34] C. Y. Kong, Discussion on the stability of alkali metal oxides, Journal of Dezhou Teacher’s College 2(2000)45-46. [35] H. Y. Liu, The Discuss of the stability of alkali metal oxide, Journal of Tangshan Teachers College 2(2001)32-33. [36] F. M. Chen, Study on influence of potassium carbonate to gasification of coal with steam, East China University of Science and Technology, Shang Hai, 2013. [37] L. Soriano, M. Abbate, A. Fernández, Oxidation state and size effects in CoO nanoparticles, Journal of Physical Chemistry B, 32(1999)6676-6679. [38] S. Chen, Study on morphology and size control of mono-dispersed Co3O4 particles produced by hydro-chemistry Method, Central South University, Changsha, 2003. [39] Q. S. Guo, Fundamental and technological study on preparation of ultra-fine Ni&Co-based powders by solution spray methods, Central South University, Changsha, 2010. [40] C. Y. Yu, Spay drying: functionalization of target product, Drying Technology & Equipment 4(2007)165-171. [41] A. Hamins. Flame extinction by sodium bicarbonate powder in a cup burner, Symposium on Combustion 2(1998)2857-2864. [42] D. W. Mckee, Mechanisms of the alkali metal catalysed gasification of carbon, Fuel 2(1983)170-175.

[43] D. W. Mckee, D. Chatterji, The catalyzed reaction of graphite with water vapor, Carbon 1(1978)53-57. [44] D. E. Jensen, G. A. Jones, Kinetics of flame inhibition by sodium, Journal of the Chemical Society Faraday Transactions, 9(1982)2843-2850. [45] A. J. Hynes, M. Steinberg, K. Schofield, The chemical kinetics and thermodynamics of sodium species in oxygen ‐ rich hydrogen flames, Journal of Chemical Physics 6(1984)2585-2597. [46] B. A. Williams, J. W. Fleming, Suppression mechanisms of Alkali Metal Compounds, Halon Options Technical Working Conference (1999)27-29. [47] W. E. Wilson, J. T. O'Donovan, R. M. Fristrom, Flame inhibition by halogen compounds, Symposium on Combustion, 1(1969)929–942. [27]

Highlights 1. The MECs of six potassium salts were tested to enriched the database of water mist with additives. 2. The active substances KOH or KxOy in extinguishing process by water mist with potassium salts are creatively obtained by thermoanalysis experimental method. 3. The quantitative analysis of active substances KOH or KxOy in fire suppression is obtained through thermodynamics method. 4. Extinguishing occurred is the result of the synchronization of water, K cations, flame radical, and other anions.