You are on page 1of 96

Chlorine 1

Chlorine
Peter Schmittinger, CREANOVA GmbH, Werk Lülsdorf, Niederkassel, Federal Republic of Germany
(Chap. 1, 2, 3, 4, 5, 9, 10, 11, 12, 13, 14 and 15)
Thomas Florkiewicz, OxyTech Systems, Chardon, Ohio, United States (Chap. 6 and 9)
L. Calvert Curlin, OxyTech Systems, Chardon, Ohio, United States (Chap. 6 and 9)
Benno Lüke, Uhde, Dortmund, Federal Republic of Germany (Chap. 7 and 9)
Robert Scannell, DeNora Deutschland GmbH, Rodenbach, Federal Republic of Germany (Chap. 8.1)
Thomas Navin, OxyTech Systems Inc., Chardon, Ohio, United States (Chap. 8.2)
Erich Zelfel, Infraserv, Knapsack, Federal Republic of Germany (Chap. 12)
Rüdiger Bartsch, Technische Universität München, München, Federal Republic of Germany (Chap. 16)

1. Introduction . . . . . . . . . . . . . . 2 7.2.1. Brine Purification . . . . . . . . . . . 47


2. Physical Properties . . . . . . . . . . 3 7.2.2. Commercial Membranes . . . . . . . 50
3. Chemical Properties . . . . . . . . . 6 7.2.3. Power Consumption . . . . . . . . . . 51
4. Chlor-Alkali Process . . . . . . . . 8 7.2.4. Product Quality . . . . . . . . . . . . 51
4.1. Brine Supply . . . . . . . . . . . . . . 11 7.3. Membrane Cells . . . . . . . . . . . 51
4.2. Electricity Supply . . . . . . . . . . 13 7.3.1. Monopolar and Bipolar Designs . . 51
5. Mercury Cell Process . . . . . . . . 13 7.3.2. Commercial Electrolyzers . . . . . . 53
5.1. Principles . . . . . . . . . . . . . . . . 14 7.3.3. Comparison of Electrolyzers . . . . 57
5.2. Mercury Cells . . . . . . . . . . . . . 19 7.3.4. Cell Room . . . . . . . . . . . . . . . . 57
5.2.1. Uhde Cell . . . . . . . . . . . . . . . . 19 8. Electrodes . . . . . . . . . . . . . . . 57
5.2.2. De Nora Cell . . . . . . . . . . . . . . 20 8.1. Anodes . . . . . . . . . . . . . . . . . . 57
5.2.3. Olin – Mathieson Cell . . . . . . . . . 21 8.1.1. General Properties of the Anodes . 59
5.2.4. Solvay Cell . . . . . . . . . . . . . . . 21 8.1.2. Anodes for Mercury Cells . . . . . . 60
5.3. Operation . . . . . . . . . . . . . . . . 21 8.1.3. Anodes for Diaphragm Cells . . . . 61
5.3.1. Brine System . . . . . . . . . . . . . . 21 8.1.4. Anodes for Membrane Cells . . . . . 62
5.3.2. Cell Room . . . . . . . . . . . . . . . . 23 8.2. Activated Cathode Coatings . . . . 62
5.3.3. Treatment of the Products . . . . . . 24 9. Comparison of the Processes . . . 63
5.3.4. Measurement . . . . . . . . . . . . . . 24 9.1. Product Quality . . . . . . . . . . . . 64
5.3.5. Mercury Emissions . . . . . . . . . . 25 9.2. Economics . . . . . . . . . . . . . . . 64
6. Diaphragm Process . . . . . . . . . 27 9.2.1. Equipment . . . . . . . . . . . . . . . . 64
6.1. Principles . . . . . . . . . . . . . . . . 28 9.2.2. Operating Costs . . . . . . . . . . . . 66
6.2. Diaphragm Cells . . . . . . . . . . . 31 9.2.3. Summary . . . . . . . . . . . . . . . . 66
6.2.1. Dow Cell . . . . . . . . . . . . . . . . 33 10. Other Production Processes . . . . 67
6.2.2. Glanor Electrolyzer . . . . . . . . . 34 10.1. Electrolysis of Hydrochloric Acid 67
6.2.3. OxyTech “Hooker” Cells . . . . . . 35 10.2. Chemical Processes . . . . . . . . . 69
6.2.4. HU Monopolar Cells . . . . . . . . 35 10.2.1. Catalytic Oxidation of Hydrogen
6.2.5. OxyTech MDC Cells . . . . . . . . 37 Chloride by Oxygen . . . . . . . . . . 69
6.3. Operation . . . . . . . . . . . . . . . . 37 10.2.2. Oxidation of Hydrogen Chloride by
6.3.1. Brine System . . . . . . . . . . . . . . 37 Nitric Acid . . . . . . . . . . . . . . . 71
6.3.2. Cell Room . . . . . . . . . . . . . . . . 39 10.2.3. Production of Chlorine from Chlo-
6.3.3. Diaphragm Aging . . . . . . . . . . . 41 rides . . . . . . . . . . . . . . . . . . . 71
6.3.4. Treatment of the Products . . . . . . 41 11. Chlorine Purification and Lique-
6.3.5. Measurement . . . . . . . . . . . . . . 43 faction . . . . . . . . . . . . . . . . . . 71
7. Membrane Process . . . . . . . . . . 43 11.1. Cooling . . . . . . . . . . . . . . . . . 71
7.1. Principles . . . . . . . . . . . . . . . . 44 11.2. Chlorine Purification . . . . . . . . 72
7.2. Process Specific Aspects . . . . . . 46 11.3. Drying . . . . . . . . . . . . . . . . . . 72

c 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim


10.1002/14356007.a06 399.pub2
2 Chlorine

11.4. Transfer and Compression . . . . 73 12.7. Safety . . . . . . . . . . . . . . . . . . 81


11.5. Liquefaction . . . . . . . . . . . . . . 74 13. Quality Specifications and Analyt-
11.6. Chlorine Recovery . . . . . . . . . . 75 ical Methods . . . . . . . . . . . . . . 82
12. Chlorine Handling . . . . . . . . . . 76 13.1. Quality Specifications . . . . . . . . 82
12.1. Storage Systems . . . . . . . . . . . 76 13.2. Analytical Methods . . . . . . . . . 82
12.2. Transport . . . . . . . . . . . . . . . . 77 14. Uses . . . . . . . . . . . . . . . . . . . 83
12.3. Chlorine Discharge Systems . . . . 78 15. Economic Aspects . . . . . . . . . . 86
12.4. Chlorine Vaporization . . . . . . . 79 16. Toxicology . . . . . . . . . . . . . . . 87
12.5. Treatment of Gaseous Effluents . 80 17. Acknowledgement . . . . . . . . . . 89
12.6. Materials . . . . . . . . . . . . . . . . 80 18. References . . . . . . . . . . . . . . . 89

1. Introduction alkali metal and alkaline earth metal chlorides, in


which the chlorine is a byproduct. Purely chemi-
Although C. W. Scheele reported the formation cal methods of chlorine production are currently
of chlorine gas from the reaction of manganese insignificant.
dioxide with hydrochloric acid in 1774, he did Since 1975, the membrane cell process has
not recognize the gas as an element [37]. H. Davy been developed to a high degree of sophisti-
is usually accepted as the discoverer (1808), cation. It has ecological advantages over the
and he named the gas chlorine from the Greek mercury processes and has become the most
κλω̃oσ (chloros), meaning greenish yellow. economically advantageous process. The mem-
Chlorine for bleaching textiles was first pro- brane cell process has become widely accepted,
duced from manganese dioxide and hydrochlo- and all new plants are using this technology. By
ric acid by a process developed by Weldon, 2000 more than 30 % of the chlorine will be pro-
the yield of chlorine being 35 % of the theo- duced in membrane cells.
retical value. In 1866, Deacon developed a pro- World capacity for chlorine exceeds
cess based on the oxidation of hydrogen chloride 45 × 106 t/a. With an annual energy consump-
gas by atmospheric oxygen in the presence of a tion of about 1.5 × 1011 kWh, the chlor-alkali
copper salt, CuCl2 , as the catalyst and obtained process is one of the largest industrial consumers
yields up to 65 % of the theoretical value. of electrical energy. The chlorine production of
In 1800, Cruickshank was the first to pre- a country is an indicator of the state of develop-
pare chlorine electrochemically [38]; however, ment of its chemical industry.
the process was of little significance until the
development of a suitable generator by Siemens Occurrence and Formation. Chlorine is
and of synthetic graphite for anodes by Ache- the 11th most abundant element in the litho-
son and Castner in 1892. These two develop- sphere. Because it is highly reactive, it is rarely
ments made possible the electrolytic production found in the free state and then mainly in vol-
of chlorine, the chlor-alkali process, on an in- canic gases. It exists mainly in the form of
dustrial scale. About the same time, both the di- chlorides, as in sea water, which contains an av-
aphragm cell process (1885) and the mercury erage of 2.9 wt % sodium chloride and 0.3 wt %
cell process (1892) were introduced. The mem- magnesium chloride. In salt deposits formed by
brane cell process was developed much more evaporation of seas, there are large quantities of
recently (1970). Currently, more than 95 % of rock salt (NaCl) and sylvite (KCl), together with
world chlorine production is obtained by the bischofite (MgCl2 · 6 H2 O), carnallite (KCl ·
chlor-alkali process. Since 1970 graphite anodes MgCl2 · 6 H2 O), tachhydrite (CaCl2 · 2 MgCl2
have been superseded by activated titanium an- · 12 H2 O), kainite (KCl · MgSO4 · 3 H2 O), and
odes in the diaphragm and mercury cell pro- others. Occasionally there are also heavy metal
cesses. The newer membrane cell process uses chlorides, usually in the form of double salts,
only activated titanium anodes. such as atacamite (CuCl2 · 3 Cu(OH)2 ), and
Other electrochemical processes in which compounds of lead, iron, manganese, mercury,
chlorine is produced include the electrolysis of or silver. Chlorates and perchlorates occur to a
hydrochloric acid and the electrolysis of molten small extent in Chile saltpeter. Free hydrochloric
Chlorine 3

acid is occasionally found in gases and springs The density of chlorine gas at 101.3 kPa is a
of volcanic origin. Plants and animals always function of temperature:
contain chlorine in the form of chlorides or free t, ◦ C 0 50 100 150
hydrochloric acid. , kg/m3 3.213 2.700 2.330 2.051
Chlorine is formed by oxidation of hy-
drochloric acid or chlorides by such compounds The density up to 300 ◦ C is higher than that
as manganese dioxide, permanganates, dichro- of an ideal gas because of the existence of more
mates, chlorates, bleaching powder, nitric acid, complex molecules, for example, Cl4 . In the
or nitrogen oxides. Oxygen, including atmo- range 400 – 1450 ◦ C, the density approximates
spheric oxygen, acts as an oxidizing agent in that of an ideal gas, and above 1450 ◦ C ther-
the presence of catalysts. Some metal chlo- mal dissociation takes place, reaching 50 % at
rides produce chlorine when heated, for exam- 2250 ◦ C. The density of chlorine gas as a func-
ple, gold(III) chloride or platinum chloride. tion of temperature and pressure is shown in Fig-
ure 1. The gas state can be described by the van
der Waals equation with
2. Physical Properties
a = 6.580 L2 bar mol−2 , b = 0.05622 L/mol
Chlorine [7782-50-5], EINECS no. 231–959–5,
exists in all three physical states. At STP it is a
greenish-yellow pungent, poisonous gas, which
liquefies to a mobile yellow liquid. Solid chlo-
rine forms pale yellow rhombic crystals. The
principal properties are given below; more de-
tails, including thermodynamic values are given
in [40] and in “New Property Tables of Chlorine
in SI Units” [41].

Atomic number Z 17
Relative atomic mass Ar 35.453
Stable isotopes (abundance) 35 (75.53 %)
37 (24.47 %)
Electronic configuration in [Ne] 3s2 3p5
the ground state
2
Term symbol in the ground P3/2
state
Melting point mp 172.17 K (− 100.98 ◦ C)
Boiling point bp 239.10 K (− 34.05 ◦ C)
Critical density crit 565.00 kg/m3
Critical temperature T crit (t crit ) 417.15 K (144.0 ◦ C) Figure 1. Density of chlorine gas as a function of
Critical pressure pcrit 7.71083 MPa temperature and pressure
Density of gas  3.213 kg/m3
(0 ◦ C, 101.3 kPa)
Density relative to air d 2.48
Enthalpy of fusion ∆H f 90.33 kJ/kg The density of liquid chlorine is given in
Enthalpy of vaporization ∆H v 287.1 kJ/kg Figure 2. The compressibility of liquid chlo-
Standard electrode potential E ◦ 1.359 V rine is the greatest of all the elements. The
Enthalpy of dissociation ∆H diss
volume coefficient per MPa at 20 ◦ C over the
239.44 kJ/mol
(2.481 eV)
Electron affinity A 364.25 kJ/mol (3.77 eV) range 0 – 10 MPa is 0.012 %. The coefficient
Enthalpy of hydration ∆H hyd 405.7 kJ/mol increases rapidly with temperature: 0.023 % at
of Cl− 35 ◦ C, 0.037 % at 64 ◦ C, and 0.064 % at 91 ◦ C.
Ionization energies ∆E i 13.01, 23.80, 39.9, 53.3,
67.8, 96.6, 114.2 eV One liter of liquid chlorine at 0 ◦ C produces
EC No. 017–001–00–7 456.8 dm3 of chlorine gas at STP; 1 kg of liq-
uid produces 311 dm3 of gas.
4 Chlorine

Thermodynamic information is given in Ta-


ble 1, from which the data required for work-
ing with gaseous and liquid chlorine can be ob-
tained [42]. The Joule – Thomson coefficient is
0.0308 K/kPa at STP.
At STP the specific heats of chlorine are
cp = 0.481 kJ kg−1 K−1
cv = 0.357 kJ kg−1 K−1
κ = cp /cv = 1.347
The molar heat capacity at constant volume
cv increases with temperature [43]:
t, ◦ C 0 100 200 500 1000
cv , J/mol 24.9 26.4 28.1 28.9 29.7
Figure 2. Density of liquid chlorine
The heat capacity of liquid chlorine decreases
The vapor-pressure curve for chlorine is over the temperature range − 90 ◦ C to 0 ◦ C:
shown in Figure 3. t, ◦ C − 90 − 70 − 50 − 30 0
−1 −1
c, J kg K 0.9454 0.9404 0.9341 0.9270 0.9169
c, J mol−1 K−1 67.03 66.70 66.23 65.73 65.02

The thermal conductivities of chlorine gas


and liquid are almost linear functions of tem-
perature from − 50 ◦ C to 150 ◦ C:
t, ◦ C − 50 − 25 0 25 50 75 100
λg , 6.08 7.06 7.95 8.82 9.75 10.63 11.50
W m−1 K−1 ×102
λl , 0.17 0.16 0.15 0.135 0.12 0.11 0.09
W m−1 K−1

The viscosities of chlorine gas and liquid are


shown in Figure 4 over the same temperature
range. The surface tension at the liquid – gas in-
terface falls rapidly with temperature:
t, ◦ C − 50 − 25 0 25 50

Figure 3. Vapor pressure of liquid chlorine σ, mJ/m2 29.4 25.2 20.9 16.9 13.4

The vapor pressure can be calculated over the


temperature range 172 – 417 K from the Mar-
tin – Shin – Kapoor equation [41]:
B E (F −T ) ln (F −T )
ln P = A+ +ClnT +DT +
T FT

A = 62.402508
B = − 4343.5240
C = − 7.8661534
D = 1.0666308×10−2
E = 95.248723
F = 424.90 Figure 4. Viscosities of chlorine gas and liquid
Chlorine 5
Table 1. Properties of liquid and gaseous chlorine [41]. Lower values are quoted in more recent literature [38, 39], especially in the region of
the critical points.

Temperature t, Pressure, Specific volumes, dm3 /kg Specific enthalpies, kJ/kg* Specific entropies, kJ kg−1 K−1
p,

C bar liquid vapor liquid vaporization vapor liquid vaporization vapor

−70 0.1513 0.6042 1563 351.11 306.89 658.00 3.9021 1.5106 5.4127
−60 0.2768 0.6135 894.4 360.69 301.58 662.27 3.9481 1.4147 5.3629
−50 0.4762 0.6233 541.8 370.15 296.29 666.41 3.9917 1.3276 5.3193
−40 0.7772 0.6336 344.9 379.70 290.73 670.43 4.0336 1.2468 5.2804
−30 1.212 0.6445 229.0 389.37 284.95 674.33 4.0737 1.1719 5.2456
−20 1.816 0.6560 157.7 399.21 278.84 678.05 4.1131 1.1015 5.2147
−10 2.628 0.6682 112.1 408.88 272.73 681.61 4.1508 1.0362 5.1870
0 3.689 0.6812 81.89 418.68 ** 266.28 684.96 4.1868 ** 0.9747 5.1615
10 5.043 0.6951 61.26 428.43 259.67 688.10 4.2215 0.9169 5.1385
20 6.731 0.7100 46.77 438.19 252.80 690.99 4.2546 0.8625 5.1171
30 8.800 0.7261 36.35 447.90 245.72 693.63 4.2873 0.8106 5.0978
40 11.30 0.7435 28.66 457.66 238.31 695.97 4.3183 0.7612 5.0790
50 14.27 0.7627 22.88 467.45 230.53 697.98 4.3480 0.7134 5.0614
60 17.76 0.7837 18.44 477.50 222.07 699.57 4.3781 0.6665 5.0447
70 21.84 0.8073 14.97 487.76 212.90 700.66 4.4074 0.6205 5.0279
80 26.55 0.8339 12.20 498.56 202.60 701.16 4.4376 0.5736 5.0112
90 31.95 0.8646 9.944 510.25 190.79 701.04 4.4665 0.5254 4.9919
100 38.14 0,9010 8.082 523.35 176.85 700.20 4.5004 0.4739 4.9743
110 45.18 0.9456 6.508 537.88 160.14 698.02 4.5372 0.4178 4.9551
120 53.18 1.0039 5.169 554.62 139.59 694.21 4.5787 0.3550 4.9337
130 62.24 1.0890 4.001 575.10 113.30 688.39 4.6277 0.2809 4.9086
140 72.50 1.2624 2.842 603.74 71.18 674.91 4.6934 0.1725 4.8659
144 77.01 1.7631 1.763 642.30 0 642.30 4.7825 0 4.7825

* These values have been calculated in S.I. units according to DIN 1345.
** The enthalpy of liquid chlorine at 0 ◦ C was taken to be H 0 = 418.66 kJ/kg; the entropy of liquid chlorine at 0 ◦ C was taken to be
0 = 4.1868 kJ kg−1 K−1 .

The specific magnetic susceptibility at 20 ◦ C yellow crystals. Chlorine hydrate is a clathrate,


is − 7.4 × 10−9 m3 /kg. and there is no definite chlorine : water ratio. The
Liquid chlorine has a very low electri- chlorine – water system has a quadruple point at
cal conductivity, the value at − 70 ◦ C being 28.7 ◦ C; the phase diagram has been worked out
10−16 Ω−1 cm−1 . The dielectric constant of the by Ketelaar [44].
liquid for wavelengths greater than 10 m is 2.15 Chlorine is readily soluble in sulfur – chlorine
at − 60 ◦ C, 2.03 at − 20 ◦ C, 1.97 at 0 ◦ C, and compounds, which can be used as industrial sol-
1.54 at 142 ◦ C, near the critical temperature. vents for chlorine. Disulfur dichloride [10025-
Chlorine gas can be absorbed in consider- 67-9], S2 Cl2 , is converted to sulfur dichloride
able quantities onto activated charcoal and silica (SCl2 ) and sulfur tetrachloride (SCl4 ). Some
gel, and this property can be used to concentrate metallic chlorides and oxide chlorides, such as
chlorine from gas mixtures containing it. vanadium oxide chloride, chromyl chloride, ti-
Chlorine is soluble in cold water, usually tanium tetrachloride, and tin(IV)chloride, are
less so in aqueous solutions. In salt solutions, good solvents for chlorine. Many other chlorine-
the solubility decreases with salt concentration containing compounds dissolve chlorine read-
and temperature. In hydrochloric acid, chlorine ily. Examples are phosphoryl chloride, carbon
is more soluble than in water, and the solubil- tetrachloride (Fig. 7), tetrachloroethane, pen-
ity increases with acid concentration (Fig. 5 and tachloroethane, hexachlorobutadiene (Fig. 7),
Fig. 6). In aqueous solutions, chlorine is partially and chlorobenzene. Chlorine also dissolves in
hydrolyzed, and the solubility depends on the glacial acetic acid, dimethylformamide, and ni-
pH of the solution. Below 10 ◦ C chlorine forms trobenzene. The solubility of chlorine in a num-
hydrates, which can be separated as greenish- ber of these solvents is given in Table 2.
6 Chlorine

Figure 7. Solubility of chlorine in hexachlorobutadiene


(——) and carbon tetrachloride (– – –) at 101 kPa as a
function of temperature

Figure 5. Solubility of chlorine in water, hydrochloric acid


(two concentrations), and sodium chloride solutions (three
concentrations)
3. Chemical Properties
All percentages are weight percents.
Inorganic Compounds. Chlorine, fluorine,
bromine, and iodine constitute the halogen
group, which has marked nonmetallic proper-
ties. The valence of chlorine is determined by
the seven electrons in the outer shell. By gaining
one electron, the negatively charged chloride ion
is formed; the chloride ion has a single negative
charge and a complete shell of electrons (the ar-
gon structure). By sharing one to seven electrons
from the outer shell with other elements, the var-
ious chlorine oxidation states can be formed, for
Figure 6. Solubility of chlorine in solutions of KCl, NaCl,
H2 SO4 , and HCl at 25 ◦ C
example, in the oxides of chlorine, hypochlorites
(+ 1), chlorates (+ 5), and perchlorates (+ 7).
The bonds between chlorine and the other
Table 2. Solubility of chlorine in various solvents
halogens are mainly covalent. In the chlorine –
Solvent Temperature, Solubility, fluorine compounds ClF and ClF3 , there is some

C wt % ionic character to the bond, with chlorine the an-
Sulfuryl chloride 0 12.0 ion, and in the chlorine – iodine compounds ICl3
Disulfur dichloride 0 58.5 and ICl, there is some ionic character to the bond,
Phosphoryl chloride 0 19.0 with chlorine the cation. Chlorine is very reac-
Silicon tetrachloride 0 15.6
Titanium tetrachloride 0 11.5
tive, combining directly with most elements but
Benzene 10 24.7 only indirectly with nitrogen, oxygen, and car-
Chloroform 10 20.0 bon. Excess chlorine in the presence of ammonia
Dimethylformamide 0 123 * salts forms the very explosive nitrogen trichlo-
Acetic acid, 99.84 wt % 15 11.6 *
ride, NCl3 . Hypochlorites react with ammonia
* g/100 cm3 to produce the chloramines NH2 Cl and NHCl2 .
Chlorine 7

Oxygen and chlorine form several chlorine ox- sulfuryl chloride, SO2 Cl2 . Under these condi-
ides (→ Chlorine Oxides and Chlorine Oxygen tions carbon monoxide and chlorine react to pro-
Acids). duce the colorless, highly toxic carbonyl chlo-
Chlorine gas does not react with hydrogen ride (phosgene), COCl2 .
gas [1333-74-0] at normal temperatures in the Chlorine reacts with sodium cyanide and
absence of light. In sunlight or artificial light of sodium thiocyanate to produce cyanogen chlo-
wavelength ca. 470 nm or at temperatures over ride and thiocyanogen chloride. The reaction
250 ◦ C, the two gases combine explosively to of chlorine with sodium thiosulfate [7772-98-
form hydrogen chloride. The explosive limits of 7] (Antichlor) is used to remove free chlorine
mixtures of pure gases lie between ca. 8 vol % from solutions.
H2 and ca. 14 vol % Cl2 (the detonation limits).
The limits depend on pressure, and the detona-
Na2 S2 O3 +4 Cl2 +5 H2 O → 2 NaHSO4 +8 HCl
tion range can be reduced by adding inert gases,
such as nitrogen or carbon dioxide (Fig. 8) [45,
46]. Chlorine reacts with carbon disulfide to pro-
duce carbon tetrachloride and disulfur dichlo-
ride.

CS2 +3 Cl2 → CCl4 +S2 Cl2

The reaction of chlorine with phosphorus


produces phosphorus trichloride (PCl3 ) and pen-
tachloride (PCl5 ). Wet chlorine attacks most
metals to form chlorides. Although titanium
[7440-32-6] is resistant to wet chlorine, it is
rapidly attacked by dry chlorine. Tantalum is re-
sistant to both wet and dry chlorine. Most met-
als are resistant to dry chlorine below 100 ◦ C,
but above a specific temperature for each metal,
combustion takes place with a flame. This spe-
Figure 8. Explosive limits of chlorine – hydrogen – other
cific temperature, the ignition temperature, also
gas mixture depends on the particle size of the metal so that
Horizontally hatched area = Explosive region with residue the following values are only approximate: iron
gas from chlorine liquefaction (O2 , N2 , CO2 ) at 140 ◦ C, nickel at 500 ◦ C, copper at 200 ◦ C,
Checkered area = Explosive region with inert gas (N2 , CO2 )
and titanium at 20 ◦ C.
Most metal chlorides are soluble in water [3,
Chlorine reacts vigorously with ammonia p. 668], notable exceptions being those of silver
(AgCl) and mercury (Hg2 Cl2 ). Chlorine liber-
ates bromine and iodine from metallic bromides
3 Cl2 +4 NH3 → NCl3 +3 NH4 Cl and iodides, but is itself liberated from metal
chlorides by fluorine.
In the presence of the catalyst bromine, chlo-
rine reacts with nitric oxide to give nitrosyl chlo- 0.5 Cl2 +KBr → KCl+0.5 Br2
ride

Selenium and tellurium react spontaneously


NO+0.5 Cl2 → NOCl
with liquid chlorine, whereas sulfur begins to
react only at the boiling point. Liquid chlorine
Sulfur dioxide and chlorine in the presence of reacts vigorously with iodine, red phosphorus,
light or an activated carbon catalyst react to form arsenic, antimony, tin, and bismuth. Potassium,
8 Chlorine

sodium, and magnesium are unaffected in liquid


chlorine at temperatures below − 80 ◦ C. Alu-
minum is unattacked until the temperature rises
to − 20 ◦ C, when it ignites. Gold is only slowly
attacked by liquid chlorine to form the trichlo-
ride (AuCl3 ). Cast iron, wrought iron, carbon
steel, phosphor bronze, brass, copper, zinc, and
lead are unaffected by dry liquid chlorine, even
in the presence of concentrated sulfuric acid.

Organic Compounds. The chlorine–carbon


bond is covalent in nature. Chlorine reacts with
hydrocarbons either by substitution or by ad-
dition. In saturated hydrocarbons, chlorine re-
places hydrogen, either completely or partially,
to form chlorinated hydrocarbons and hydrogen
chloride. Methane can be chlorinated in stages
through to carbon tetrachloride.
In the reaction with unsaturated hydro-
carbons, chlorine is added to the double or
triple bond yielding dichloro- or tetrachloro-
hydrocarbons, respectively. In industry the re-
action velocity is increased by light (photochlo-
rination), heat (cracker furnace), or catalysts. In
aromatic hydrocarbons, both addition and sub-
stitution are possible, depending on conditions
(light, temperature, pressure, or catalysts).

Chlorine Compounds in Nature. More


than 1800 chlorine-containing compounds have
been identified in nature at present. These or-
ganic chlorine compounds, which range from
chloromethane to the complex antibiotic van-
comycin, are produced by marine and terrestrial
plants, bacteria, fungi, lichen, insects, marine
animals, some higher animals, and a few mam-
mals. For example, the annual natural produc-
tion of chloromethane is estimated to exceed Figure 9. Flow diagram of the chlor-alkali mercury process
4 × 106 t/a [47].
takes place in two parts, at the anode and at the
cathode. The evolution of chlorine takes place at
4. Chlor-Alkali Process the anode:
In the chlor-alkali electrolysis process, an aque-
ous solution of sodium chloride is decom- 2 Cl− → 2 Cl+2 e− → Cl2 +2 e−
posed electrolytically by direct current, produc-
ing chlorine, hydrogen, and sodium hydroxide
solution. The overall reaction of the process There are three basic processes for the elec-
trolytic production of chlorine, the nature of
the cathode reaction depending on the specific
2 NaCl+2 H2 O → Cl2 +H2 +2 NaOH
Chlorine 9

process. These three processes are (1) the di- These three processes are described in detail
aphragm cell process (Griesheim cell, 1885), in the following three chapters. The basic flow
(2) the mercury cell process (Castner – Kellner sheets of the three processes are shown in Fig-
cell, 1892), and (3) the membrane cell process ures 9, [10]], [11]]. In all three processes, nearly
(1970). saturated, purified brine is introduced into the
Each process represents a different method of electrolysis cell.
keeping the chlorine produced at the anode sep-
arate from the caustic soda and hydrogen pro-
duced, directly or indirectly, at the cathode.

Figure 10. Flow diagram of the chlor-alkali diaphragm


process
Figure 11. Flow diagram of the chlor-alkali membrane pro-
cess
(* optional)
10 Chlorine

Figure 12. Processing of hydrogen gas from the amalgam decomposer


a) Vertical decomposer; b) Individual cell hydrogen cooler; c) Safety seal; d) Demister; e) Blower; f) Final hydrogen cooler;
g) Mercury removal equipment

The hydrogen produced is cooled as it leaves The products are extremely pure. The chlo-
the decomposer or the cathode compartment rine, along with a little oxygen, generally can be
and is carried through electrically insulated used without further purification. The sodium
pipework to a vessel fitted with a water seal hydroxide solution contains little chloride and
(Fig. 12). If a hydrogen – air mixture forms be- leaves the decomposer with a 50 wt % concen-
cause of a shutdown or breakdown, the seal tration.
allows the mixture to escape. A demister en- Of the three processes, the mercury process
sures that the gas is free of spray, whether wa- uses the most electric energy; however, no steam
ter or sodium hydroxide solution. The hydrogen is required to concentrate the caustic solution.
is compressed by Roots-type blowers or recip- The use of large quantities of mercury demands
rocating compressors before it passes through measures to prevent environmental contamina-
coolers on its way to the consuming plants. At tion. In addition, the hydrogen gas and sodium
no stage is the pressure allowed to fall below solution must be freed from mercury. Generally,
ambient pressure. the operation of the cells is not simple.
Electrolytic hydrogen is very pure, > 99.9 In the diaphragm cell process, the anode area
%; however, unwanted traces of oxygen can be is separated from the cathode area by a perme-
removed by reaction with the hydrogen over a able, generally asbestos-based diaphragm. The
platinum catalyst. The hydrogen is used for or- brine is introduced into the anode compart-
ganic hydrogenation, catalytic reductions, am- ment and flows through the diaphragm into the
monia synthesis and to provide hot flames or pro- cathode compartment. Cheaper solution-mined
tective atmospheres in welding technology, met- brine can be used; the brine is purified by pre-
allurgy, or glass manufacture. It is also used in cipitation – filtration.
the manufacture of high-purity hydrogen chlo- A caustic brine leaves the cell, and this brine
ride by combustion with chlorine and as a fuel must be freed from salt in an elaborate evap-
for heating and drying. orative process. Even so, the resultant 50 wt %
In the mercury cell process, sodium amalgam sodium hydroxide solution contains up to 1 wt %
is produced at the cathode. The amalgam is re- NaCl. The salt separated from the caustic brine
acted with water in a separate reactor, called the can be used to saturate dilute brine. The chlo-
decomposer, to produce hydrogen gas and caus- rine contains oxygen and must be purified by
tic soda solution. liquefaction and evaporation.
Because the brine is recirculated, solid salt The consumption of electric energy with the
is required for resaturation. The brine, which diaphragm cell process is ca. 15 % lower than
must be quite pure, is first dechlorinated and for the mercury process, but the total energy con-
then purified by a straightforward precipitation – sumption is higher because of the steam required
filtration process. to concentrate the caustic brine (see Fig. 62). En-
vironmental contamination with asbestos must
Chlorine 11

be avoided. Under constant operating condi- In the United States and Europe, rock salt is
tions, cell operation is relatively simple. most commonly used. The most important im-
In the membrane cell process, the anode and purities are shown in Table 3. The concentrations
cathode are separated by a cation-permeable ion- of these impurities depend on the method of pro-
exchange membrane. Only sodium ions and a duction and on the different grades: crude rock
little water pass through the membrane. salt, prepared rock salt, and evaporated salt. So-
As in the mercury process, the brine is dechlo- lar salt is used in Japan and many other parts
rinated and recirculated, which requires solid of the world, the most important sources being
salt to resaturate the brine. The life of the ex- Australia, Mexico, China, Chile, India, and Pak-
pensive membrane depends on the purity of the istan. The salt produced by solar evaporation is
brine. Therefore, after purification by precipita- usually much less pure than rock salt. In a few
tion – filtration, the brine is also purified with an cases the salt may be obtained from other pro-
ion exchanger. cesses, such as caustic soda evaporation in the
The caustic solution leaves the cell with a diaphragm process.
concentration of 30 – 36 wt % and must be con- A new upgrading process (Salex) has been
centrated. The chloride content of the sodium developed by Krebs Swiss [48]. It removes the
hydroxide solution is almost as low as that from impurities by selective cracking of the salt crys-
the mercury process. The chlorine gas contains tals and a washing process. Salt losses are min-
some oxygen and must be purified by liquefac- imized, and the purity exceeds 99.95 % NaCl.
tion and evaporation.
Table 3. Impurities in rock salt and sea salt, wt %
The consumption of electric energy with the
membrane cell process is the lowest of the three Rock salt Sea salt
processes, ca. 25 % less than for the mercury Insolubles ≤2 0.1 – 0.3
process, and the amount of steam needed for Water ≤3 2.0 – 6.0
concentration of the caustic is relatively small Calcium 0.2 – 0.3 0.1 – 0.3
Magnesium 0.03 – 0.1 0.08 – 0.3
(see Fig. 62). The energy consumption should Sulfate ≤0.8 0.3 – 1.2
be even lower when oxygen-consuming elec- Potassium ≤0.04 0.02 – 0.12
trodes become common. There are no special
environmental problems. The cells are easy to
operate and are relatively insensitive to current
changes, allowing greater use of the cheaper off- Brine Resaturation. In older plants, the
peak-time electric power. open vessels or pits used for storing the salt
are also used as resaturators. The depleted brine
from the cells is sprayed onto the salt and is sat-
4.1. Brine Supply urated, the NaCl concentration reaching 310 –
315 g/L. Modern resaturators are closed vessels,
The brine used in the mercury cell and mem- to reduce environmental pollution [49], which
brane cell processes is normally saturated with could otherwise occur by the emission of a salt
solid salt although there are some installations spray or mist. The weak brine is fed in at the
that use solution-mined brine on a once-through base of the resaturator, and the saturated brine is
basis. The brine supply for diaphragm cells is drawn off at the top. If the flow rates of the brine
always used on a once-through basis, although and the continously added salt are chosen care-
the salt recovered from caustic soda evaporators fully, the differing dissolution rates of NaCl and
may be recycled into the brine supply. CaSO4 result in little calcium sulfate dissolv-
ing within the saturator [50]. Organic additives
Salt. The basic raw material for the mercury also reduce the dissolution rate of calcium sul-
cell and membrane cell processes is usually solid fate [51]. The solubility (g per 100 g of H2 O) of
salt. This may be obtained from three sources: NaCl in water does not increase much with tem-
rock salt, solar salt, or vacuum-evaporated salt perature (t, ◦ C), whereas the solubility of KCl
from purifying and evaporating solution-mined does:
brine.
12 Chlorine

t 0 20 40 60 80 100 moved from the sludge by washing with water.


cNaCl 35.6 35.8 36.4 37.0 38.5 39.2 Barium salts may be recovered by treating with
cKCl 28.2 34.4 40.3 45.6 51.0 56.2 sodium carbonate under pressure [62]. The puri-
fied brine should contain ideally cCa < 2 mg/L,
cMg < 1 mg/L, and cSO4 < 5 g/L.
Brine Purification. In mercury cells, traces In the diaphragm process, the removal of sul-
of heavy metals in the brine give rise to danger- fate is not always necessary because SO2− 4 can
ous operating conditions (see page 15), as does be removed from the cell liquor as pure Na2 SO4
the presence of magnesium and to a lesser ex- during the concentration process. In the mem-
tent calcium [52]. In membrane cells, divalent brane process, the brine must be purified to a
ions such as Ca2+ or Mg2+ are harmful to the much higher degree to avoid the deterioration
membrane. The circulating brine must be rigor- of the membrane. The Ca2+ and Mg2+ concen-
ously purified to avoid any buildup of these sub- tration must be < 0.02 ppm (20 ppb), so a sec-
stances to undesirable levels [7]. Calcium is usu- ond, fine purification step is required (see Sec-
ally precipitated as the carbonate with sodium tion 7.2.1).
carbonate; magnesium and iron, as hydroxides Before the brine enters the electrolysis cells,
with sodium hydroxide; and sulfate, as barium it should be acidified with hydrochloric acid to
sulfate. pH < 6, which increases the life of the titanium
The reagents are usually mixed with weak anode coating, gives a purer chlorine product
brine and added to the brine stream at a con- with higher yield, and reduces the formation of
trolled rate. If solar salt is used, treatment costs hypochlorite and chlorate in the brine.
may be reduced by prewashing the salt [53]. In
order to precipitate calcium at low pH, sodium Brine Dechlorination. In the mercury and
bicarbonate [54] or phosphoric acid [55] can be membrane processes, the depleted brine leav-
added. ing the cells must be dechlorinated before resat-
The sulfate content can be reduced without uration. Further acidification with hydrochloric
the use of expensive barium salts by discharg- acid to pH 2 – 2.5 reduces the solubility of chlo-
ing a part of (purging) the brine [56], by crys- rine by shifting the equilibrium point of hydrol-
tallization of Na2 SO4 · 10 H2 O on cooling the ysis and inhibits the formation of hypochlorite
brine stream [57], by precipitation of the double and chlorate. Chlorine discharged in the anolyte
salt Na2 SO4 · CaSO4 [58], by an ion-exchange tank prior to dechlorination may be fed into the
process, or by membrane nanofiltration [59]. chlorine system. The dissolved chlorine of the
Hoechst [60] has a process for recovering bar- brine then is still 400 – 1000 mg/L, depending on
ium sulfate of pure pigment quality by precipita- pH and temperature. The brine is passed down a
tion under acid conditions. Chlorate buildup can packed column or sprayed into a vacuum of 50
be avoided by addition of sodium metabisulfite – 60 kPa, which reduces the chlorine concentra-
Na2 S2 O5 [61]. tion in the brine to 10 – 30 mg/L. The vacuum
After stirring for 1 – 2 h, the precipitated im- is produced by steam jet or liquid-ring vacuum
purities are removed by filtration alone or by pump. The pure chlorine gas obtained is fed into
sedimentation followed by filtration. Sedimenta- the chlorine stream.
tion is carried out in large circular settling tanks, The water that evaporates from the dechlori-
from which the slurry is removed by mechanical nated brine is condensed in a cooler. The conden-
raking equipment, e.g., Clariflocculator, Cycla- sate, which may be chemically dechlorinated, is
tor, or Dorr thickener. Filtration is carried out returned to the brine circulation system if neces-
with a sand filter, a pressure-leaf filter with fil- sary to maintain the volume of the brine circuit.
ter cloths of chlorine-resistant fabrics, or can- If necessary, the remaining chlorine content can
dle filters automatically cleaned by backflow of be further reduced by blowing with compressed
brine. The filter is cleaned by water jets, vibrat- air, by a second vacuum treatment, by treatment
ing, or shaking. The separated filter cake is con- with activated carbon [63], or by chemical treat-
centrated to 60 – 80 % solids content in rotary ment with hydrogen sulfite, thiosulfate, sulfur
drum vacuum filters or centrifuges before dis- dioxide, or sodium hydrogensulfide.
posal. Any soluble material present may be re-
Chlorine 13

Brine Monitoring. The sodium chloride Each set of rectifiers is connected through
concentration in the brine is determined by den- high-voltage switchgear to the three-phase sup-
sity measured by equipment involving radioac- ply [65]. Smaller units use a 10 – 30 kV supply,
tive isotopes, vibration techniques, hydrometry, but large units can be connected into the high-
or weighing. voltage power system (> 100 kV) [66].
The pH following alkali or acid additions is The unit cost of the d.c. supply decreases
determined with glass electrodes, and the redox with increasing voltage and current. A plant is
potential following chlorine removal is deter- therefore most economical when as many high-
mined with metal electrodes. Excess OH− and current cells as possible are connected in series
CO2−3 ions ensure adequate precipitation of dis- [67]. Total currents of 450 000 A are achieved.
solved calcium, iron, and magnesium. After fil- The switches for short-circuiting the cells are de-
tration, a test sample of 100 mL should require signed for 10 000 – 30 000 A and are operated by
4 – 6 mL of 0.1 N acid to reach the phenolph- compressed air, hydraulically, or by spring ac-
thalein end point and a further 0.5 – 1.5 mL to tion. Erosion of the main contacts is dealt with
reach the methyl orange end point. Inadequate by using replaceable pre-contacts [68]. The con-
filtration is detected by turbidimetry in transmit- tacts are protected from corrosion by installation
ted light or by the Tyndall effect. Calcium and in vacuum housings.
magnesium are determined hourly, and chlorate The current in the bus bars or in anode rods
and sulfate about once per day, all by titration. can be measured by means of iron-free trans-
portable equipment with an accuracy of ca. 1 %
[69].
4.2. Electricity Supply

Since 1960 the direct current for electrolysis has 5. Mercury Cell Process
been provided exclusively by silicon rectifiers.
A set of rectifiers can supply up to 450 000 A. The clean separation of chlorine from the cath-
Voltages up to 4.0 kV per diode are feasible, but ode products is possible because of the high
usually for safety, a peak a.c. voltage of 1500 V, overvoltage of hydrogen at the mercury elec-
corresponding to a d.c. output of 1200 V, is not trode. Hydrogen and sodium hydroxide are not
exceeded. Liquid cooling of the diodes permits produced at the cathode; instead, sodium is pro-
a compact design, and self-contained equipment duced and dissolves in the mercury as an amal-
reduces leakage losses. Modern membrane cell gam. The liquid amalgam is removed from the
plants also use continuously variable thyristor electrolytic cell to a separate reactor, called the
converters in place of silicon diodes [64]. decomposer or denuder, where it reacts with
Rectification equipment is required to pro- water in the presence of a catalyst to form the
vide steady direct current at a voltage determined sodium hydroxide and hydrogen gas. The pro-
by the cell room. The current must remain steady cess may also be used to produce potassium hy-
even though the voltage is varied both by the op- droxide by feeding the cell with potassium chlo-
erating condition of the cells and by the number ride solution, although this is much less com-
of cells operating. The rectifier equipment usu- mon. The sodium hydroxide is produced from
ally consists of the denuder at a concentration of ca. 50 wt %;
the maximum value is 73 wt %. The hydroxide
transformer capable of variable output volt- solution is very pure and almost free from chlo-
age with adequate compensation for changing ride contamination.
input voltage The process was developed in 1892 almost si-
silicon rectifiers or thyristors multaneously by H. Y. Castner and C. Kellner
constant-current control gear and used on an industrial scale, although the
transducers for metering and control amount of chlorine produced remained rela-
control panels tively small until 1930, when the rapid growth
isolators of the rayon (artificial silk) industry, especially
cooling equipment in Germany, increased the demand for pure
ancillary safety and monitoring equipment
14 Chlorine

chloride-free sodium hydroxide solution. At this In 1984, the mercury cell process accounted
time, the horizontal high-current cell was devel- for 45 % of world chlorine production [71].
oped and output increased rapidly. The devel- Since then no new plants have been built. In
opment work in Germany was described in the the coming decades most of the existing mer-
FIAT final reports, published after World War II, cury cell plants will be shut down or con-
and this led to widespread use of the process in verted to membrane cell technology. Only plants
Europe and Japan [70]. In the United States, the with speciality products such as extremely pure
mercury cell process became more widespread, sodium hydroxide, potassium hydroxide, alco-
increasing its share of chlorine production from holates, and dithionites will use the mercury pro-
3 – 4 % in 1945 to 20 % in 1960, reaching a max- cess in future. These plants will meet the highest
imum of 27 % in 1970. emission control standards.
The development of the mercury cell can be
followed in the technical data: the cell current
increased from 3.4 kA in 1895 to ca. 30 kA in 5.1. Principles
1945, 200 kA in 1960, and 450 kA in 1970. The
current density rose from 2 kA/m2 in 1950 to the The cathode reaction
current maximum of 15 kA/m2 . The cell area in-
creased over the same period from ca. 7 m2 to
37.5 m2 , while the k-factor ( specific voltage co- Na+ +e− +Hgx → NaHgx
efficient, see page 16) was reduced by 50 %. forming sodium amalgam, is followed by the
Since 1972 the importance of the mercury decomposition reaction in a separate reactor
cell has decreased. Increasing concern about the
effect of mercury on the environment has led to
a considerable increase in the number and varia- 2 NaHgx +2 H2 O → 2 NaOH+H2 (g) +2 Hgx
tion of statutory regulations that affect the mer-
cury cell process.
In particular, widespread concern about cases Process Description (Fig. 13). Mercury
of mercury poisoning in Japan, which were not flows down the inclined base of the electrolytic
related to the mercury cell process, caused the cell (A). The base of the cell is electrically con-
process to be legally banned since 1972. How- nected to the negative pole of the d.c. supply. On
ever, conversion to the alternative processes was top of the mercury and flowing cocurrently with
delayed because of demand for low-chloride it is a concentrated brine with a sodium chlo-
sodium hydroxide and because of the anticipated ride content of ca. 310 g/L at the inlet. Anodes
advantages of the rapidly developing membrane are placed in the brine so that there is a small
cell process. The last remaining mercury cell in- gap between the anode and the mercury cathode.
stallations for NaCl were closed in 1986. The concentration of the amalgam is maintained
In the other countries, existing mercury cell at 0.2 – 0.4 wt % Na, so that the amalgam flows
plants are still in operation, but official regula- freely (Fig. 14). The chlorine gas and depleted
tions and uncertainty about possible further legal brine (270 g/L) flow out of the cell, either sepa-
restrictions have hindered expansion. rately or as a two-phase mixture separated later
In Europe and the United States great ef- in the process. The amalgam flows out of the
forts are being made to develop methods of pro- cell through a weir and into the decomposer. The
tecting the environment from mercury (see Sec- amalgam may be passed through a water wash
tion 5.3.5). These measures have greatly reduced between the cell and the decomposer to remove
emissions of mercury into the atmosphere and traces of sodium chloride. The amalgam flows
into wastewater to the extent that the present lev- through the decomposer countercurrent to a flow
els of emitted mercury are negligible in compar- of softened or demineralized water in the pres-
ison to those arising from natural sources, such ence of a catalyst to produce sodium hydroxide
as volcanic action, geological erosion, or other solution and hydrogen. Stripped of its sodium,
nonnatural sources such as fuel combustion or the mercury flows out of the lower end of the
metallurgical processes. decomposer and is recirculated through a pump
back into the cell.
Chlorine 15

Figure 13. Schematic view of a mercury cell with decomposers


A) Mercury cell: a) Mercury inlet box; b) Anodes; c) End box; d) Wash box
B) Horizontal decomposer: e) Hydrogen gas cooler; f) Graphite blades; g) Mercury pump
C) Vertical decomposer: e) Hydrogen gas cooler; g) Mercury pump; h) Mercury distributor; i) Packing pressing springs

Cathode Reactions. The standard potential


of the hydrogen-liberating reaction is 0 V, which
is considerably higher than the potential for
the formation of 0.2 wt % sodium amalgam,
− 1.868 V. However, hydrogen is not liberated
at the mercury surface because the reaction is
kinetically inhibited. Mainly sodium ions are
discharged. At the sodium chloride concentra-
tions used, the reversible potential is reduced by
ca. 0.2 V. (Exact values of the discharge poten-
tial are given as a function of the sodium con-
centration in the amalgam, the sodium chloride
concentration in the brine, and the temperature
[72].) Electrochemical side reactions occur: the
reduction of chlorine molecules or hypochlor-
ous acid and the liberation of hydrogen gas. In
addition, sodium in the amalgam can react di-
Figure 14. Freezing point curves of sodium amalgam and
potassium amalgam
rectly with free chlorine, or chlorite and chlorate
ions can be reduced to chloride by the action of
nascent hydrogen at the cathode. All of these
side reactions represent a loss of efficiency, nor-
Anode Reactions. The oxidation of chloride mally ca. 2 – 4 % under good operating condi-
ions to chlorine gas has a standard potential of tions.
1.358 V. In a 300 g/L sodium chloride solution Contamination of the system by heavy metals
at 70 ◦ C, the reversible reaction potential is re- can lead to a reduction of the hydrogen discharge
duced to 1.248 V [15, p. 339]. Some side reac- potential at the mercury cathode, thus increasing
tions occur, such as the oxidation of OH− and hydrogen liberation, and reducing amalgam for-
SO2−4 ions and the electrochemical formation mation [73]. The hydrogen concentration in the
of chlorate ions. Nonelectrochemical reactions chlorine can increase to the point at which the
also take place in the region of the anode, such cell and downstream chlorine handling equip-
as hypochlorite formation (because of hydrol- ment contains explosive mixtures. The proba-
ysis of chlorine) and chlorate formation. All of bility of such problems is estimated by a hazard
these side reactions represent a loss of efficiency. analysis of an existing plant [74 – 76].
16 Chlorine

The cell system is sensitive to trace quanti- camalgam = 0.15 %, and 70 ◦ C, the reversible de-
ties of catalysts in the brine, for example, vana- composition voltage is E = 3.095 V [77].
dium, molybdenum, and chromium at the 0.01 – Factor 2. The overpotential of chlorine de-
0.1 ppm level or iron, cobalt, nickel, and tung- pends on the material and shape of the anodes. At
sten at the parts per million level. Magnesium, the high current densities (10 kA/m2 ) present in
calcium, aluminum, and barium are also active modern cell rooms, the overpotential can reach
at the parts per million level. several hundred millivolts, outweighing the ef-
In addition, relatively high concentrations of fect of concentration changes in the electrolyte
sodium in the amalgam (> 0.5 wt %) can cause (concentration polarization) [78] and the retard-
increased hydrogen evolution in the cells. Potas- ing effect that formation of molecular chlorine
sium chloride electrolysis is considerably more has on the process of ion discharging [79]. Chlo-
sensitive to both catalysts and high concentra- rine gas bubbles cover part of the anode surface,
tion in the amalgam than the sodium chloride thereby increasing the current density at the free
process. surface. The anode is designed so that the gas
bubbles are liberated as quickly as possible. The
Current Efficiency. The theoretical electro- rapid removal of these gas bubbles from the re-
chemical equivalents representing the materi- action zone is one of the advantages of titanium
als produced or consumed in the electrolysis of anodes over graphite anodes [80].
sodium chloride or potassium chloride brines The overpotential of sodium on the amalgam
are given in Table 4. In practice, the yield is cathode is caused by the limited diffusion rate
ca. 95 – 97 % of the theoretical value, owing to of the liberated sodium atoms into the amalgam,
side reactions at the electrodes and in the elec- but it is small compared to the chlorine overvolt-
trolyte. With activated titanium anodes, the yield age.
is largely independent of the distance between Factor 3. The specific conductivity of
the electrodes. sodium chloride solutions increases with con-
centration and temperature (Fig. 15), but is inde-
The decrease in salt concentration ∆c is de- pendent of pH over the range 2 – 11. The brine
termined by the current I, the brine flow rate M, normally enters the cells at 60 – 70 ◦ C and leaves
and the electrochemical equivalent f. the cells at 75 – 85 ◦ C. The conductivity of potas-
sium chloride solutions at 70 ◦ C is 30 % greater
∆c = f I/M than that of sodium chloride solutions. Chlorine
gas bubbles in the electrolyte increase the resis-
tance between anodes and cathode. Better circu-
The usual units are c in g/L, f in kg kA−1 h−1 , lation of the electrolyte in the gap between elec-
I in kA, and M in m3 /h. trodes allows more rapid removal of gas bubbles,
thus reducing the voltage.
Cell Voltage. The d.c. voltage across the cell Factor 4.The voltage losses in the cell room
circuit is determined by five factors: are minimized by compactly arranging the cells,
1) The reversible decomposition voltage of the which shortens the current path. The relatively
salt low conductivity of steel cell bases can be im-
2) The overpotentials of the chlorine and alkali proved by copper or aluminum fittings. These
metal at the electrodes measures also reduce problems caused by mag-
3) The voltage drop in the electrolyte netic fields, which occur in wide cells [81].
4) Voltage losses in the bus bars, switches, elec- Factor 5. In practice, cell current density and
trical conductors, anode materials, and cath- cell voltage have a linear relationship. The slope
ode of the line is termed the specific voltage coeffi-
5) The operating current density of the cells cient or k-factor, a useful measure of the specific
energy requirement of cells produced by differ-
Factor 1.The reversible decomposition po- ent manufacturers.
tential of NaCl under standard conditions
is E ◦ = 3.226 V (KCl E ◦ = 3.234 V). Under
the operating conditions cNaCl = 290 g/L,
Chlorine 17

Table 4. Electrochemical equivalents f, kgkA−1 h−1

Element Element produced Salt required Alkali produced

Na 0.8580 2.1810 (NaCl) 1.4923 (NaOH)


K 1.4586 2.7816 (KCl) 2.0931 (KOH)
Cl2 1.3228 (0.4115 m3 STP)
H2 0.0376 (0.4185 m3 STP)

In addition to the d.c. voltages considered


above, there are energy losses across the trans-
former and rectification equipment. All cell in-
stallations use a.c. power, which is rectified by
silicon diodes in which the energy losses are
minimized by operating at greater than 100 V.
This voltage is achieved by operating at least 25
cells in series.

Energy Consumption. To operate a cell in-


stallation economically, the consumption of d.c.
electrical energy per unit mass of product must
be minimized. The specific energy consumption
w is given by
w = 1000 Ucell /a f

where w = kWh/t, a = yield factor or cur-


rent efficiency, f = electrochemical equivalent,
kg kA−1 h−1
Figure 15. Specific conductivity of sodium chloride
solutions
For example, if the cell voltage U cell is 4.20 V
and the current efficiency is 0.970, then ca.
3275 kWh is required to produce 1 t of chlo-
The cell voltage is given by U cell = 3.15 + kJ, rine + 1.13 t of NaOH. Since af is almost a con-
J = current density, kA/m2 , k = specific voltage stant, the specific energy consumption per tonne
coefficient, V m2 kA−1 . of chlorine w is effectively proportional to the
Computer-controlled cells with activated tita- cell voltage. In that case, w also depends on the
nium anodes are run with k-factors from 0.085 cell current density (see Fig. 16). In the example,
to 0.11. The corresponding cell voltages at w = 3275 kWh corresponds to 10 kA/m2 .
10 kA/m2 are 4.00 – 4.25 V (Fig. 16). The total energy requirement per tonne of Cl2
must also include the transformer and rectifier
losses (30 – 40 kWh/t) and the energy require-
ments of all of the ancillary equipment (120 –
160 kWh/t).
A mathematical model of the cell has been
described [82].

Decomposition of the Amalgam. The


amalgam is decomposed in horizontal decom-
posers, alongside or beneath the cell, or more
often since ca. 1960 in vertical decomposers
Figure 16. Cell voltage and specific energy consumption
per tonne of Cl2 versus cell current density
or denuders. The energy stored in the amalgam
18 Chlorine

has an emf of ca. 0.8 V. The hydrogen over- Attempts to recover some of the energy stored
potential at the amalgam prevents spontaneous in the amalgam by creating an electrical circuit
decomposition in contact with water, and a cat- by using the catalyst as the anode separated from
alyst (depolarizer) must be used. The overall the amalgam or by using the amalgam electrode
decomposition reaction is with an oxygen gas diffusion electrode have so
far had no practical outcome [84].
2 NaHgx +2 H2 O → 2 NaOH+H2 +2 Hgx Horizontal decomposers are ducts with a
rectangular cross section, which are installed
with a 1 – 2.5 % slope near to or underneath the
and takes place in two stages, first as an anode
cells. The amalgam flows in a stream ca. 10 mm
reaction at the surface of the amalgam
in depth, and the sodium content is thereby re-
duced to < 0.02 wt %. The catalyst consists of
2 Na → 2 Na+ +2 e− graphite blades 4 – 6 mm thick, which are im-
mersed in the amalgam in a lengthwise direction
and then as a cathode reaction on the catalyst (Fig. 18, also see Fig. 13 B). The water for the
surface, where the water is decomposed reaction, which is softened or demineralized by
ion exchange, flows in the direction opposite the
2 H2 O+2 e− → 2 OH− +H2 amalgam and is removed as 50 % caustic alkali
solution. The hydrogen gas is cooled as it leaves
the decomposer so that any condensed water and
Industrial decomposers are essentially mercury run back into the decomposer. Advan-
short-circuited electrochemical primary cells tages of the horizontal decomposer are service-
(Fig. 17). The most common catalyst is graphite ability, simple construction, and a pure product
[7782-42-5], usually activated by oxides of iron, that is low in mercury. However, horizontal de-
nickel, or cobalt or by carbides of molybde- composers require a greater mercury inventory
num or tungsten. The hydrogen overpotential than vertical decomposers.
on graphite (0.5 – 0.6 V at 2 kA/m2 and 80 ◦ C)
increases with current density and decreases
with temperature; therefore, the decomposer
should be operated at as high a temperature
as possible [83]. Good catalyst material must
meet many requirements: resistance to alkali
solutions, hydrogen, and mercury; low hydro-
gen overpotential; good electrical conductivity;
long-lasting activity; wettability by amalgam;
and incapability of amalgamation.

Figure 18. Cross section through a horizontal decomposer


a) Amalgam; b) Bolt; c) Graphite blades; d) Hydrogen;
e) Sodium hydroxide solution; f) Decomposer casing;
g) Spacers

Vertical decomposers are designed as towers


[85] containing packings of activated graphite
spheres or other shapes 8 – 20 mm in diame-
ter. The towers are packed 0.6 – 0.8 m high. The
cross section of the tower is 0.35 m2 per 100 kA
of cell current. The amalgam is fed in via an
overhead distributor, and the mercury is pumped
Figure 17. Principle of amalgam decomposition
Chlorine 19

from the base of the tower back to the cell by rubberized fabric. The anodes, generally made
a closed centrifugal pump (see Fig. 13 C). The of activated titanium, hang in groups from car-
water for the reaction is fed into the base of the rying devices that can be varied in height manu-
tower and flows upward counter to the amal- ally, hydraulically, or by motor-operated lifting
gam. The 50 % caustic alkali solution flows out devices. Each cell can be short-circuited exter-
at the top. The smaller volume of the vertical nally by a switch. The cell bus bars are usually
decomposer leads to higher product tempera- copper. The anodes are protected from internal
ture because of the greater energy intensity of short circuits by means of electronic monitor-
the system. Cooling the hydrogen is essential. ing systems. The size of the cells can be varied
Compared with the horizontal decomposer, the within a broad range to give the desired chlorine
amount of space required is small, and the mer- production rate. Computer programs optimize
cury inventory is small, but the caustic alkali the cell size, number of cells, and optimum cur-
contains more mercury. rent density as a function of the electricity cost
In alternative decomposition reactions, other [87] and capital cost.
products may be obtained from the amalgam For comparison, a list of cells manufactured
in place of sodium or potassium hydroxide so- by leading engineering firms and cell character-
lutions [10, p. 518], [86]: sodium sulfide from istics is given in Table 5. Cathode surface areas
sodium polysulfide solution, alcoholates from are ca. 17 – 30 m2 , and nominal currents are ca.
alcohols, sodium dithionite from sodium hydro- 170 – 300 kA [32, p. 204].
gen sulfite, hydrazobenzene or aniline from ni-
trobenzene, adiponitrile from acrylonitrile, and
alkali metals by distillation.
5.2.1. Uhde Cell

5.2. Mercury Cells The Uhde cells (Fig. 19, also see Fig. 22) are
available with a cathode surface area 4 – 30 m2
During the first decades after the rocking cells of for chlorine production rates 10 – 1000 t/d for the
Castner and Kellner were first commissioned, complete cell installation. The brine flows in via
considerable efforts were made to develop suit- an inlet box fitted with two pipes for the removal
able materials for the cells and the anodes. A of chlorine. The weak brine is removed at the
large number of cell configurations were tested, end of the cell. The solid cover is fixed to the
resulting in the development of the continuous side walls by clamps. The anodes are suspended
cell. Since 1950, the cell areas and the specific in groups in carrying frames supported near the
load were increased considerably. cells on transverse girders with lifting gear. The
In 1972, the changeover from graphite to anode rods are raised and lowered within a bel-
metallic anodes began, with a parallel develop- lows seal. Short copper bus bars between the
ment of computer monitoring and control, lead- cells also serve for shunt measurement of the an-
ing to improved short-circuit protection and a ode currents. The electric current is brought in
reduction of the specific energy consumption by above the cell covers via flexible copper straps
computer-controlled anode adjustment, of great that run immediately above the anode rods and
significance in view of the drastic increase in are bolted to them. The compressed-air switches
electricity costs in the late 1970s. In the years are situated under the cells. The cell bottom is
following 1972, producers operating the elec- usually a current conductor when cells are short-
trolysis plants also concentrated on the develop- circuited, but in wide, heavily loaded cells the
ment and installation of devices to reduce mer- cathode current is carried by copper bus bars to
cury emissions. prevent the occurrence of strong magnetic fields,
The cells currently available possess a num- which could interfere with the amalgam flow.
ber of common features. The mercury flows over The automatic equipment for protection and ad-
a steel base that has a slope of 1.0 – 2.5 %. The justment of the anodes depends on the shunt
flanged side walls are lined with rubber. The cell measurement of the currents and is controlled
covers are mostly steel, lined with rubber or ti- by a central computer. In this way, an optimum
tanium on the underside, but they may also be k-factor is selected for each cell. The vertical
20 Chlorine

Table 5. Characteristics of modern mercury cells

Characteristic Manufacturer
Uhde De Nora Olin – Mathiesen Solvay Krebs Paris

Cell type 300 – 100 24M2 E 812 MAT 17 15 KFM


Cathode area, m2 30.74 26.4 28.8 17 15.4
Cathode dimensions, l×b, m2 14.6×2.1 12.6×2.1 14.8×1.94 12.6×1.8 9.6×1.6
Slope of cell base, % 1.5 2.0 1.5 1.7
Rated current, kA 350 270 288 170 160
Max. current density, kA/m2 12.5 13 10 10 10.4
Cell voltage at 10 kA/m2 , V 4.25 3.95 4.24 4.10 4.30
Number of anodes 54 48 96 96 24
Stems per anode 4 4 2 1 4
Number of intercell bus bars 36 32 24 24 12
Quantity of mercury per cell, kg 5000 4550 3800 1650
Energy requirement per tonne of Cl2 , kWh d.c. 3300 3080 3300 3200 3400

Figure 19. Uhde mercury cell


a) Cell base; b) Anode; c) Cover seal; d) Cell cover; e) Group adjusting gear; f) Intercell bus bar; g) Short-circuit switch;
h) Hydrogen cooler; i) Vertical decomposer; j) Mercury pump; k) Anode adjusting gear; l) Inlet box; m) End box

decomposers are provided with hydrogen cool- to electric currents from 45 to 400 kA. The cover
ers and are situated at the end of each cell. The is a flexible multilayer sheet of elastomer spread
amalgam flows into the decomposer under the over the cell trough. This cover is supported by
force of gravity [88]. the anode rods and seals them. The DSA anodes
(see Section 8.1) are held rigidly in strong car-
rying frames, which are automatically adjusted
5.2.2. De Nora Cell by electric motors. Individual anode adjustment
is not provided. The anode rods are individually
The size of the De Nora cell (Fig. 20, also see connected by flexible copper straps to the an-
Fig. 23) varies from 4.5 to 36 m2 , corresponding
Chlorine 21

Figure 20. Cross section through the De Nora mercury cell


a) Cell base (steel); b) Side wall (rubber-lined steel); c) Lifting gear; d) Transverse support; e) Lengthwise support; f) Anode
carrier; g) Anode rod; h) Anode surface; i) Adjusting motor; k) Bus bar; l) Flexible anode current strap; m) Multilayer cell
cover; n) Service walkway; o) Intercell bus bar; p) Switch; q) Insulator; r) Switch drive; s) Support

ode bus bars. The cathode current is carried by 5.2.4. Solvay Cell
copper bus bars. Devices for the improvement
of brine circulation and gas removal within the The bus bars in the Solvay cells are made primar-
cells reduce specific energy consumption. Con- ily of aluminum. Above the cells is a cover that
sequently, the reduction in brine concentration also serves as a convenient walkway, giving ac-
can be increased from the usual 35 – 40 g/L to cess to the anode rods. The titanium anodes are
60 – 70 g/L, and the brine circulating rate can be specially coated and are automatically adjusted
reduced by ca. 40 %. Separate outlets are present by computer. The tall vertical decomposers are
at the inlet box for the normal chlorine gas pro- located under the cells.
duction and the weak chlorine gas produced dur-
ing start-up. The graphite catalyst in the vertical
decomposer is activated with molybdenum. 5.3. Operation

The aspects of the operation of mercury cells that


5.2.3. Olin – Mathieson Cell typically differ from those of the other processes
are the brine circulation system, the cell room,
The special feature of the Olin – Mathieson cell treatment of the products, measurement and con-
lies in the system of mounting and adjusting the trol, and reduction of mercury emissions.
anodes. Above each row of anode rods, a U-
shaped copper or aluminum bus bar also serves
to support the anode lifting gear. The anode rods 5.3.1. Brine System
are bolted to the U-shaped bus bar. The anodes
are adjusted as a group, either manually or by a A typical brine circulation system for the mer-
remote computer with the remote computerized cury cell process is shown in Figure 21. In the
anode adjuster (RCAA) system. The currents are cells the sodium chloride concentration of the
measured independently of the cell potentials by brine is reduced by 35 – 60 g/L to 260 – 280 g/L
means of reed contacts [89].
22 Chlorine

Figure 21. Schematic diagram of a brine circulation system in the mercury cell process
a) Electrolysis cell; b) Anolyte tank; c) Vacuum column dechlorinator; d) Cooler; e) Demister; f) Vacuum pump; g) Seal tank;
h) Final dechlorination; i) Saturator; k) Sodium carbonate tank; l) Barium chloride tank; m) Brine reactor; n) Brine filter;
o) Slurry agitation tank; p) Rotary vacuum filter; q) Vacuum pump; r) Brine storage tank; s) Brine supply tank

Figure 22. Cell room: Uhde mercury cells

at 70 – 85 ◦ C. To avoid mercury emissions into has been developed that allows rapid determi-
the air, the resaturators are generally closed ves- nation of the suitability of any particular salt or
sels. The mercury cathode is very sensitive to brine.
poisoning by heavy metals; therefore, a test [90]
Chlorine 23

5.3.2. Cell Room

The cells are usually situated in a building


(Fig. 22), although sometimes they are erected
in open air (Fig. 23). Figure 24 shows a bird’s
eye view, and Figure 25 shows a cross section of
cell room. The cells are arranged parallel to each
other so that bus bars and supply lines are kept
short. The cells stand on supporting structures
and are insulated to prevent shorting to the earth.
The transformer and rectifiers are situated at one
end of the room, and the cell service and repair
area is at the opposite end. Ancillary equipment
is installed near the cell room in a spillage con-
tainment area.

Figure 24. Mercury cell room (bird’s-eye view, schematic)


a) Cell room; b) Transformer room; c) Rectifier room;
d) Bus bars; e) Turnaround bus bars; f) Service walkways;
g) Ancillary equipment; h) Electrolysis cells; i) Vertical de-
composers; k) Cell assembly and maintenance area

Figure 25. Mercury cell room (cross section, schematic)


a) Basement floor; b) Floor drains; c) Cell supports with
Figure 23. Open-air cell room: De Nora cells insulators; d) Supply pipes; e) Cells; f) Decomposers;
g) Service walkways; h) Crane; i) Ridge ventilator;
j) Ventilation air supply; k) Windows/lighting
Cell floors, gangways, and spillage contain-
ment areas are constructed with smooth, sloping
floors so that any mercury can be easily recov-
ered or wash water can be conveniently collected Occupational Health. Anyone working in
for treatment. The supply pipes run under the the cell area must undergo regular health checks.
cells and are connected to them by flexible, in- Euro Chlor has prepared a Code of Practice
sulating connections. The heat given off by the “Control of Worker Exposure to Mercury in
cells and the decomposers is removed by a ven- the Chlor-Alkali Industry” [91]. The U.S. Chlo-
tilation system. rine Institute has released guidelines “Medical
The plant is operated with continuous 24 h/d Surveillance and Hygiene Monitoring Practices
supervision and control. An additional day-shift of Worker Exposure to Mercury in the Chlor-
team carries out anode changes, repairs, and Alkali Industry” [92]. The U.S. Environmental
cleaning. Protection Agency has established 18 rules re-
24 Chlorine

lating to cleanliness of the cell room [93]. Ad- block, flakes, prills, or powder. For the processes
herence to these rules eliminates any danger to involved and for uses, see → Sodium Hydroxide.
the health of personnel caused by mercury. The
maximum allowable concentration or threshold
limit value (TLV) of mercury in the atmosphere
in Western Europe and in the United States is
between 0.025 and 0.100 mg/m3 .

5.3.3. Treatment of the Products

Chlorine. See Chapter 11.

Hydrogen. The treatment of the hydrogen


gas leaving the decomposer is described in
Chapter 4. It must pass special equipment for
the removal of the traces of mercury before it is
used (see page 25).

Sodium Hydroxide Solution. The great ad-


vantage of the mercury cell process is that very
pure sodium hydroxide solution is produced (see
Table 20) at a suitable concentration. The chlo-
ride content is only 5 – 50 mg/kg.
Figure 27. Freezing and boiling point curves of sodium
Sodium hydroxide from the decomposer usu- hydroxide solutions
ally has a concentration of 50 % and a temper-
ature of 80 – 120 ◦ C. It passes through rubber-
lined steel pipe work to nickel or Incoloy cool- 5.3.4. Measurement
ers, where it is cooled to 40 – 60 ◦ C. Any parti-
cles of graphite from the decomposer or traces of The condition of the brine, the cells, and the
mercury are effectively removed by centrifuges, products must be continuously and carefully
candle filters, or precoated leaf filters (Fig. 26). monitored, since even small deviations from the
correct conditions can increase the hydrogen
concentration in the chlorine. The measuring op-
erations are mostly automatic: critical limits are
chosen and if these limits are exceeded, alarms
are set off.

Cell Operating Conditions. The sodium


concentration in the amalgam is determined at
the cell inlet (max. 0.05 wt %) and outlet (max.
Figure 26. Processing of sodium hydroxide solution from 0.45 wt %).
the amalgam decomposer
a) Vertical decomposer; b) Collection main; c) Collecting
A 20-g sample of the amalgam is reacted with
tank; d) Pump; e) Cooler; f) Mercury removal filter 30 wt % aqueous sulfuric acid in an absorption
pipette. The evolution of 1 cm3 of gas corre-
The freezing-point and boiling-point curves sponds to 0.01 wt % Na.
of sodium hydroxide solutions are shown in Fig- A portable analytical and recording apparatus
ure 27. The phases separating from the solution, is available that works electrochemically [94].
i.e., ice, hydrates, and NaOH, are indicated along If the mercury pump stops, the steel cathode
the freezing-point curve. Sodium hydroxide is base of the cell is exposed to electrolyte, and
supplied to consumers as aqueous solution, solid hydrogen evolves to form an explosive mixture
with the chlorine in the cell.
Chlorine 25

Failure of the mercury pump or mercury flow residual emissions to water and air are ecologi-
automatically short-circuits the cell. Mercury cally acceptable. The mercury in the electrolytic
flow failure is detected by monitoring the mer- cells circulates in a closed system. All materi-
cury level at the lowest part of the mercury cir- als that come into contact with the mercury –
culation system or in the inlet box, by direct flow equipment, products, auxiliary chemicals, wash
measurement [95], or by loss of pressure at the water, waste gases, other waste materials – may
pump delivery. become contaminated with mercury, and must
The motors for the mercury pumps, the chlo- be treated before release to the environment or
rine absorption plant, and the most vital con- must be safely deposited. For the exact measure-
trol equipment are all provided with an emer- ment of these trace amounts and for control of
gency power supply, ensuring safe shutdown of the effectiveness of the measures to reduce the
the plant if a power failure occurs. The instal- emissions, analytical methods have been devel-
lation is protected by a complex system of in- oped with sensitivities in the microgram region
terlocks so that failure of important equipment, [103].
such as the chlorine compressor, shuts down the Many countries have set legal limits on emis-
rectifiers. sions in waste air and water. The limits on the
A large number of systems have been devel- products of a chlor-alkali plant may depend on
oped for the protection of cells from short cir- their end use, e.g., drinking water treatment or
cuits. Titanium anodes are destroyed by short food processing. The sources of contamination
circuits and must be raised before any contact are listed, and means of reducing them are de-
with the amalgam takes place. scribed: Mercury cells are sealed vessels, and the
The operation of the monitoring system de- products are conveyed in closed pipes. The cell
pends on magnetic-field current measurement rooms must have smooth joint-free floors with
for individual anodes [89, 96, 97] or on shunt easily cleaned drainage surfaces and irrigated
measurement of the supply bus bars [98]. Moni- collection gutters (see Fig. 25). Spilled mercury
toring is achieved by comparison of the anode – is immediately washed away with water into col-
cathode voltage of different cell sections [99] or lecting tanks or sucked up with a vacuum system.
by following the conductivity of the brine in the Control of mercury loss is only possible if
electrode gap [100]. the mercury content of all the cells is known ex-
The signals from the instruments are fed into actly. The gravimetric and volumetric methods
central computers or local microprocessors at formerly used were cumbersome and led to ad-
each cell [101] that control the anode lifting gear. ditional mercury emissions, disadvantages that
The mercury inventory in each cell may be mea- are avoided by a radioactive tracer method.
sured by a radioactive tracer technique once a
year without affecting cell operation [102]. Mercury in Products. Hot, moist chlorine
Products. The concentration of the sodium leaving the cell contains small amounts of mer-
hydroxide solution is determined from its den- curic chloride. This is almost completely washed
sity, and the purity its checked by titration to de- out in the subsequent cooling process and may
termine hydroxide, carbonate, and chloride con- be fed back into the brine with the condensate. In
tents. The purity of the water for the decomposer the cooled and dried chlorine gas, there are only
is determined from its conductivity. minute traces of mercury: 0.001 – 0.01 mg/kg.
The oxygen content of the hydrogen gas is de- The equilibrium mercury concentration in hy-
termined from the magnetic susceptibility, oxy- drogen gas is a function of temperature and pres-
gen being paramagnetic. sure. The mercury concentration (mg of Hg per
m3 of H2 at 101.325 MPa) increases rapidly with
temperature:
5.3.5. Mercury Emissions (→ Mercury, t, ◦ C 0 20 40 60 80 100
Mercury Alloys, and Mercury Compounds,
Chap. 5) c, mg/m3 2.36 14.1 66.1 255 836 2404

Any chlor-alkali plant up to modern technical Subjecting the hydrogen gas to pressure low-
standards is not a hazard to the environment. The ers the mercury content of the resulting prod-
26 Chlorine

uct gas at atmospheric pressure. For example, at 2) Cell cleaning operations


5 ◦C 3) Cleaning of floors, tanks, pipes, and disman-
tled apparatus
cHg p = 0.37 Pa kg m−3
The amount of wastewater can be reduced by
separately disposing of the cooling water and
cHg = concentration of mercury in hydrogen process water and by feeding the condensate
gas at atmospheric pressure, mg of Hg back into the brine, provided the water balance
per m3 of H2 allows this. A wastewater rate of 0.3 – 1.0 m3 per
tonne of chlorine is achievable.
p = pressure to which the hydrogen gas is There are various methods of making
subjected, MPa wastewater suitable for discharge:
When the mixture is cooled to 2 – 3 ◦ C, the 1) Chemical removal of mercury by reducing
mercury concentration is reduced to ca. 3 mg/m3 any compounds to the metal with hydrazine or
at standard pressure. This mercury content can sodium borohydride or by precipitating mer-
be reduced by compressing and further cool- curic sulfide with thiourea or sodium sulfide.
ing, adding chlorine to form mercurous chloride The mercury metal or sulfide is then filtered
(calomel), which is collected on rock salt or sim- off.
ilar material in a packed column, washing with 2) Oxidation of the mercury by chlorine,
solutions containing active chlorine, or by ad- hypochlorite, or hydrogen peroxide and ad-
sorption on activated carbon impregnated with sorption on an ion-exchange medium. Elutri-
sulfur or sulfuric acid, leaving a mercury con- ation is done with hydrochloric acid, which is
centration in hydrogen of 0.002 – 0.03 mg/m3 . then used to acidify the brine [104].
The highest purity can be achieved by adsorption The Clean Water Act of 1972 (United States)
on copper/aluminum oxide or silver/zinc oxide, demands the use of the “best available technol-
< 0.001 mg/m3 . ogy economically achievable.” Since 1982 each
Centrifugation or filtration in candle filters plant has been limited to a maximum of 0.1 g
or in disk filters precoated with charcoal gives of Hg per tonne of chlorine averaged over 30 d
sodium hydroxide solutions containing mercury measured at the outlet of the wastewater treat-
concentrations of < 0.05 ppm (mg/kg of 50 % ment plant.
caustic soda). In Western Europe, an EC directive has been
The circulating brine contains mercury con- issued on the subject of the mercury content of
centrations of 2 – 20 mg/L. Mercury emissions wastewater from chlor-alkali plants, following
from the brine system can occur through losses various earlier agreements such as the Rhine pro-
of brine into the wastewater, by brine vapor- tection agreement, the EC guidelines concerning
ization in the resaturators, or by disposal of the protection of natural waters, and the Paris
the residues from the brine purification filter. Convention [105]. This directive requires plants
These emissions are minimal at a chlorine con- with circulating brine systems to have a limit of
centration < 30 mg/L, giving a redox poten- 1.0 g of Hg per tonne of chlorine produced.
tial > 500 mV vs. NHE. Under these conditions
mercury remains dissolved in the brine as a mer- Mercury in Process Air. Air from the pro-
cury chloride complex even if the brine is alka- cess, for example, the cell end box ventilation
line. system, vents from liquid collection tanks (caus-
tic, wastewater), from the vacuum cleaning sys-
Mercury in Wastewater. Mercury- tem, or from the distillation unit for mercury-
containing wastewater has several sources: contaminated wastes can be treated to remove
mercury by the methods used for hydrogen.
1) The process, e.g., condensate and wash liquor
from treatment of chlorine, hydrogen, and Ventilation of the Cell Room. The heat
brine; stuffing-box rinse water from pumps produced during electrolysis requires that the air
and blowers; brine leakages; ion-exchange must be changed 10 – 25 times per hour, depend-
eluate from process-water treatment ing on the type of building. Mercury spillage can
Chlorine 27

occur during essential operations involving cells sure that in 2007 no plant emits more than 1.5 g
or decomposers, for example, opening the cells Hg per tonne of chlorine capacity to air, water,
for anode changing or cleaning, assembling or and products [109].
dismantling equipment, or replacing defective
pipes. Spillage leads to small losses in the ex-
haust air owing to the vapor pressure of mercury.
In addition, products that contain mercury, such
as the sodium hydroxide solution, hydrogen, or
process waste air, can escape via faulty seals
in pipes and equipment, leading to emissions.
Closed cell construction and special care in
handling mercury, i.e., good housekeeping, by
adhering to the EPA rules or the Code of Prac-
tice “Mercury Housekeeping” [106], keep the
mercury concentrations and, hence, emissions
below the allowable work place concentrations
(MAK and TLV) [107].
Purification of large volumes of waste air Figure 28. Mercury emissions from the European chlor-
alkali industry
containing mercury in very low concentrations is
not effective. In the United States, the upper limit
for the emission of mercury in process waste air
and hydrogen is 1 kg per day per facility, and
for ventilation air it is 1.3 kg per day per facil- 6. Diaphragm Process
ity [18, p. 372]. In Western Europe the Parcom
Decision 90/3 [108] requires a standard of 2 g The commercial production of chlorine by elec-
Hg/t of chlorine capacity for emissions to the trolytic processes began in Europe and the
atmosphere from existing plants. United States in the 1890s. Early cells of the
bell-jar type had no diaphragm and relied on the
Mercury in Residues. Mercury-containing flow of anolyte toward the cathode to prevent
residues include brine filter slurry, spent de- the hydroxide ion from back-migrating toward
composer catalyst, discarded cell components, the anode. This method had limited capacity be-
residues from the purification of products, waste cause gas evolution caused mixing and loss of
material from rinsing media, adsorption mate- efficiency. The Griesheim cell, another early de-
rials, ion-exchange media, etc. Mercury can be sign, used porous cement as the diaphragm.
recovered from these materials by distillation E. A. Le Sueur is credited with the design of
in closed retorts. The residues after distillation a cell incorporating a percolating asbestos di-
must be disposed at special sites. Mercury in aphragm, which is the basis for all diaphragm
safely deposited wastes is not considered an chlor-alkali cells currently in use. When brine is
emission to the environment. caused to flow into the anolyte and subsequently
through the diaphragm into the catholyte, con-
Summary. The continued efforts of all pro- tinuous operation with much improved effi-
ducers have led to a steady decrease in mercury ciency is obtained. This Le Sueur cell, and the
emissions over the years; for example, in West- similar Billiter cell developed in Germany, in-
ern Europe from 16 g in 1978 to 2 g Hg per tonne corporated a horizontal asbestos sheet as the di-
of chlorine capacity in 1996, as shown in Fig- aphragm. During the 1920s, the Billiter cell be-
ure 28 [109]. With this low emission level, the came the most widely used cell in the world; a
contribution of the chlor-alkali industry to the to- few are still in operation.
tal natural and anthropogenic mercury emissions Following the invention of synthetic graphite,
is less than 0.1 % [111]. Euro Chlor is develop- numerous cells were developed. These fall into
ing a BAT (best available techniques) for reduc- three basic types:
ing mercury emissions from existing mercury-
based plants, the application of which will en- 1) Rectangular vertical electrode cells
28 Chlorine

2) Cylindrical vertical electrode cells sodium chloride solution (brine) enters the an-
3) The vertical electrode bipolar filter press cell ode compartment and completely covers the an-
developed by Dow Chemical odes and the cathode tubes or fingers. The chlo-
rine leaves the cell through an outlet in the cell
In 1913, C. W. Marsh developed a cell with head. The anolyte flows through the diaphragm
finger cathodes and side-entering anodes and into the cathode compartment because of the
cathodes, which greatly increase the electrode difference in liquid level between the two com-
area per unit of floor space. About 1928, Ken- partments. The catholyte is a solution of sodium
neth Stewart of Hooker Chemical (now Occi- chloride and sodium hydroxide because a por-
dental Chemical Co.) developed a method of de- tion of the water is converted to hydroxide at
positing asbestos fibers onto the cathode by im- the cathode. The hydrogen produced at the same
mersing the cathode in a slurry of asbestos fibers time leaves the cell through an outlet on the cath-
and applying a vacuum. All significant installa- ode. The solution of sodium chloride and sodium
tions of diaphragm cells currently in operation hydroxide overflows the cell through a level con-
are derived from that development [2]. trol pipe on the cathode, and is then commonly
All diaphragm cells produce cell liquor that called cell liquor.
contains ca. 11 wt % caustic soda and 18 wt %
sodium chloride. To market the caustic soda, its
concentration must be increased to 50 %. Dur-
ing the evaporation and cooling processes, the
salt becomes less soluble in the stronger caustic,
and at 50 % NaOH the NaCl concentration is ca.
1 %.

6.1. Principles
The principles needed to understand the efficient
operation of the diaphragm process involve the
current efficiency, cell voltage, power consump-
tion, and optimization of the operating condi-
tions [112, 113].
The reaction at the positively charged anode
Figure 29. Basic chemical reactions within the cell
is the same for all three chlor-alkali processes a) Anode compartment; b) Cathode compartment;
c) Deposited diaphragm on cathode tubes, rims, and end
screens
2 Cl− → Cl2 +2e−

The reaction at the negatively charged cath- Current Efficiency. Current efficiency is de-
ode of the diaphragm cell is fined as the amount of product actually produced
divided by the amount of product that theoret-
2 H2 O+2 e− → H2 +2 OH−
ically should have been produced on the basis
of the amount of direct-current electrical energy
input. The current efficiency is never 100 % be-
Figure 29 is a cutaway view of a diaphragm cause of side reactions. The efficiency of a di-
cell that shows the orientation of the various aphragm cell is usually based on the chlorine
parts of the cell and the various reactions that production.
take place. Figure 29 also shows the location of The side reactions that lower the efficiency
the diaphragm, which is deposited on the out- are a result of chlorine that enters the catholyte
side of the cathode screen and which separates compartment or of hydroxide ions that enter the
the cell into two compartments, one containing anolyte compartment. The amount of chlorine
the anodes and one containing the cathode. The that enters the catholyte compartment is small.
Chlorine 29

The majority of the efficiency losses in a di- dized:


aphragm cell are due to migration of hydroxide
ions from the catholyte through the diaphragm 2 OH− → 1/2 O2 +H2 O+2 e−
and into the anolyte. This back migration takes
place because the negatively charged hydroxide
ions are attracted to the positively charged an- Sodium chlorate in the cell liquor is a result
odes and because of the hydroxide ion concen- of hydroxide ions that migrate through the di-
tration gradient across the diaphragm. This mi- aphragm into the anolyte and react with chlorine
gration of hydroxide ions through the diaphragm before reaching the anode:
is in equilibrium with the opposing flow of brine
through the diaphragm. 3 Cl2 +6 NaOH → NaClO3 +5 NaCl+3 H2 O
Three factors control the migration of the hy-
droxide ions into the anolyte:
Equations. The simplest equations for calcu-
1) Concentrations of hydroxide and chloride
lating cell efficiency are based on the masses
ions at the cathode side of the diaphragm
of products produced per unit of electrical in-
2) Flow rate of brine through the diaphragm
put. Theoretically, 1.492 kg of sodium hydrox-
3) Condition of the diaphragm
ide and 1.323 kg of chlorine are produced per
These three factors in dynamic equilibrium kiloampere-hour. It then follows that
determine the efficiency of the cell.
Factor 1. The higher the concentration of hy- Cathode efficiency,% = (kg of NaOH×100) /
droxide ions in the catholyte, the larger the con- (Q×1.492×the number of cells)
centration gradient across the diaphragm, and Anode efficiency,% = (kg of Cl2 ×100) /
(Q×1.324×the number of cells)
the higher the probability of hydroxide ions
crossing through the diaphragm. As a result,
changing cell liquor strength strongly affects cell where Q is the quantity of electricity in kA h
efficiency. The concentration of chloride ions in Unfortunately, the production of a single cell
the catholyte also affects cell efficiency because cannot be measured with sufficient accuracy to
some of the chloride ions migrate in place of give meaningful results. To get around this prob-
hydroxide ions. lem, the chlorine industry uses an equation based
Factor 2. Decreasing brine flow rate to a cell on the analysis of the chlorine gas, the cell liquor,
increases the conversion of sodium chloride to and the anolyte:
sodium hydroxide and raises the hydroxide con- %CE = [%Cl2 ×100] /
centration in the catholyte because of reduced [%Cl2 +2 (%O2 ) + (%Cl2 ×anox×F ) /cNaOH ]
overflow from the cell. The decreased flow rate
of brine through the diaphragm allows increased
where
migration of hydroxide ions into the anolyte.
These factors decrease cell efficiency. % CE = anode current efficiency, %
Factor 3. The condition of the diaphragm
is extremely important. Nonuniformity in the % Cl2 = percent chlorine in cell gas (air free)
diaphragm results in high flow rates of brine
through thin or loosely compacted areas and low % O2 = percent oxygen in cell gas (air free)
flow rates through thick or compacted areas. In
the areas where there is a low brine flow rate, anox = oxidizing power of anolyte ex-
back migration of hydroxide is increased. pressed as grams of NaClO3 per
The degree of inefficiency in a cell is indi- liter
cated by the two products of the side reactions,
oxygen in the chlorine and sodium chlorate in cNaOH = NaOH concentration in the cell
the cell liquor. Oxygen in the chlorine gas is the liquor, g/L
result of hydroxide ions that migrate through the
diaphragm into the anolyte, where they are oxi- F = conversion factor
30 Chlorine
Table 6. Typical voltage distribution a

Component Current density, kA/m2


voltages, V
1.24 1.55 1.86 2.17 2.48
b
Anode potential 1.30 1.30 1.30 1.30 1.31
Cathode potential b 1.12 1.13 1.15 1.16 1.17
Structure loss c 0.11 0.14 0.17 0.20 0.22
Brine loss 0.11 0.15 0.19 0.23 0.27
Diaphragm loss 0.24 0.31 0.36 0.41 0.47
Intercell bus 0.02 0.02 0.03 0.03 0.03

Total 2.90 3.05 3.20 3.33 3.47


a ◦
OxyTech MDC-55 cell with Modified Diaphragm and expandable anodes. Conditions: anolyte temperature 93 C, anolyte NaCl
concentration 250 g/L, catholyte NaOH concentration 130 g/L.
b
Potential vs. NHE.
c
Includes anode base, anodes, cathode, cathode screens, copper end connectors, and copper side plates.

The denominator is the amount of chlorine quired for nonreversible kinetics. The cell struc-
produced plus the amount of chlorine consumed ture voltage drop includes the voltage losses in
in the side reactions. This is equivalent to the the cathode, anodes, intercell bus, and all other
amount of chlorine that could have been pro- connectors in the cell. The sum of the diaphragm
duced theoretically from the input of current. voltage drop and the anolyte – catholyte (brine)
The conversion factor F is the product of a vol- voltage drop is the potential between the elec-
ume factor, an electric field factor, and a stoi- trodes. All of these voltages are functions of cur-
chiometric factor. In practice, it is a function of rent density. Table 6 shows how cell voltage is
cell liquor strength. strongly affected by cell current. Cell tempera-
The SIX equation is a practical alternative ture, feed brine NaCl concentration, and the cell
to the previous equation and is often used with liquor NaOH concentration also affect cell volt-
computers linked to a gas chromatograph and an age (Table 7), because they affect the conductiv-
automated cell liquor analyzer. The SIX equa- ity of the solutions between the electrodes. Table
tion is 7 clearly shows that current density is the most
important factor.
%CE = [%Cl2 ×100] /
    Table 7. Factors affecting cell voltage: the change in cell voltage
%Cl2 +2 (%O2 ) + %Cl2 ×6×cNaClO3 /cNaOH ∆U cell divided by the change in four important cell factors

Factor Modified Standard


This equation also accounts for chlorine lost cell * cell *
to the anolyte. However, it approximates the ox-
∆U cell /change in current
idizing potential of the anolyte with the concen- density J, mV m2 kA−1 450 450
tration of chlorate in the cell liquor and assumes ∆U cell /change in cell tempera-
a fixed conversion factor from anolyte concen- ture t cell , mV/◦ C −7.7 −10.1
tration to catholyte, namely SIX. The SIX equa- ∆U cell /change in brine
concentration cNaCl , mV L g−1 −0.7 −1.8
tion approximates the standard equation within ∆U cell /change in cell liquor
0.5 %. concentration cNaOH , mV L g−1 0.26 0.6

* OxyTech MDC-55 cell. The modified cell is outfitted with the


Cell Voltage. The voltage of a cell is the sum Modified Diaphragm and expandable anodes, whereas the standard
of five component voltages: anode potential, cell is outfitted with the standard asbestos diaphragm and the
standard DSA anode. Conditions: anolyte temperature of 93 ◦ C,
cathode potential, cell structure voltage drop, di- anolyte NaCl concentration of 250 g/L, and catholyte NaOH
aphragm voltage drop, and anolyte – catholyte concentration of 130 g/L.
voltage drop. The anode and cathode potentials
are sums of the reversible voltages, which are
Excessive brine impurities or other severe op-
the thermodynamic minimum amounts of work
erating problems can adversely affect the voltage
to cause the reactions to take place, and the over-
of the cell.
potentials, which are the additional voltages re-
Chlorine 31

Power Consumption. The power consump- the cathode of the next cell unit, thus minimizing
tion of a cell, kWh per tonne of Cl2 , may be intercell voltage loss. These units are assembled
calculated from the cell voltage by the follow- in series like a filter press, and therefore, the
ing equation: voltage of an electrolyzer is the sum of the in-
dividual cell voltages created by the anode of
Power consumption = Ucell ×756/ε one unit, a diaphragm, and the cathode of the
next unit. Bipolar electrolyzers have high volt-
where ages and relatively low amperage; therefore, the
cost of electrical rectification is lower per unit of
U cell = cell voltage, V production capacity. Bipolar electrolyzers either
must be installed in a large number of electrical
ε = cell efficiency expressed as a decimal circuits or be designed with very large individ-
ual cell components. Developers have chosen the
Optimization. The relationships described option for large components.
in the preceding paragraphs can be used to de- Dow Chemical was the only early devel-
termine the optimum economical cell operating oper to have chlorine production needs large
conditions. The optimizations that must be con- enough to consider the bipolar option [2]. Later,
sidered are the following: following the development of the DSA anode,
PPG Industries and Oronzio De Nora Impianti
1) Higher cell liquor caustic strength and lower Elettrochimici designed, and PPG Industries in-
steam usage in caustic evaporation versus stalled Glanor bipolar electrolyzers in a large
lower cell efficiency and higher power con- complex at Lake Charles, Louisiana [12].
sumption The monopolar electrolyzer is assembled so
2) Lower current density, lower voltage, and that the anodes and cathodes are in parallel.
lower power consumption versus additional Therefore, the potential difference of all cells in
cells and higher capital costs the electrolyzer is the same, and the amperage
3) Lower feed brine temperature, thus decreased at any particular current density only depends
steam usage for brine heating, versus higher on the electrode surface area. A monopolar elec-
cell voltage, lower efficiency, and higher trolyzer has low voltage and high amperage. The
power consumption highest amperage rating of the most common
4) High brine pH and reduced acidification costs modern monopolar cells is ca. 150 kA. Because
versus lower chlorine efficiency, higher power a monopolar electrolyzer has a voltage of only
consumption, and lower product purity 3 – 4 V, circuits of up to 200 electrolyzers have
Each diaphragm cell chlorine plant must de- been constructed, producing 900 t of chlorine
termine its own optimum conditions for the most per day.
economical operations.
Diaphragms. The earliest asbestos di-
aphragms were made of sheets of asbestos paper.
6.2. Diaphragm Cells Asbestos was chosen because of its good chem-
ical stability and its ion-exchange properties.
Electrolyzers for the production of chlorine and Asbestos has been relatively inexpensive, since
sodium hydroxide, including both diaphragm it is a relatively abundant natural material that
and membrane cells, are classified as either was already being mined and processed for other
monopolar or bipolar. The designation does not industrial purposes, such as insulation.
refer to the electrochemical reactions that take The deposited asbestos diaphragm developed
place, which of course require two poles or elec- by Hooker Chemical in 1928 was the most com-
trodes for all cells, but to the electrolyzer con- mon diaphragm until 1971, when what is now
struction or assembly. There are many more OxyTech Systems developed the Modified Di-
chlor-alkali production facilities with monopo- aphragm. The Modified Diaphragm is a mix-
lar cells than with bipolar cells. ture of asbestos and a fibrous fluorocarbon poly-
Bipolar electrolyzers have unit assemblies of mer [114]. The polymer stabilizes the asbestos,
the anode of one cell unit directly connected to
32 Chlorine

which in itself lowers cell voltage and also al- precathode by standard vacuum techniques. To
lows for the use of the expandable DSA anode control the diaphragm porosity, a pore-forming
[115]. In its various formulations, the Modified agent is incorporated in the slurry. Optimized
Diaphragm is the most common diaphragm. The porosity leads to lower cell voltage and higher
Modified Diaphragm still contains a minimum current efficiency.
of 75 % asbestos. The technology is currently evaluated on
Environmental concern over the use of as- OxyTech HC-3B electrolyzers at the Chloralp
bestos has increased. France, Saudi Arabia, and Pont-De-Claix facility. Compared to polymer-
Norway have banned the use of asbestos as a sep- modified diaphragms, Chloralp asbestos-free
arator in chlorine cells. These nations allowed separators provide a 150 mV saving on voltage
local chlorine producers several years to install and a 2 % improvement in current efficiency.
non-asbestos replacement separators in existing Combined benefits from the precathode and the
diaphragm cells or to replace the cells them- diaphragm, as well as an extended lifetime,
selves with membrane cells. Other European na- could lead in the near future to full conversion
tions issue permits for the continued use of as- of the 240 000 t/a Pont-De-Claix plant.
bestos until an agreed upon BAT (best available
technology) for diaphragm cells is established. OxyTech Polyramix Diaphragm. OxyTech
There is also concern in the chlor-alkali- Systems has developed and commercialized a
industry for the future supply of asbestos, as synthetic non-asbestos diaphragm called the
most of the North American mines producing Polyramix diaphragm. The Polyramix fiber is
the grades of asbestos previously used for chlo- composed of a PTFE [poly(tetraflouroethylene)]
rine cell diaphragms have closed. At present the fibrid base with zirconium oxide ceramic parti-
most common source of asbestos for chlorine cles embedded in and on the fiber. The Polyra-
cell diaphragms is Zimbabwe. In addition as- mix diaphragm is vacuum deposited onto the
bestos disposal costs and regulation have contin- cathode and then baked in an oven to fuse
ued to increase. All of theses factors have led the the fibers together. The process is very sim-
chlor-alkali industry to consider non-asbestos ilar to OxyTech’s widely used Modified Di-
diaphragm technology. Three non-asbestos di- aphragms. The Vulcan Chemicals plant in Geis-
aphragm systems are commercially available to- mar Louisiana was fully converted to the use
day. of Polyramix diaphragms in 1993. Most major
diaphragm cell chlorine plants have operated
Chloralp Asbestos-Free Technology. Chlo- from 2 to 40 Polyramix diaphragm cells. The
ralp, a joint venture of Rhône-Poulenc and longest life cell with a Polyramix diaphragms
La Roche, has developed an asbestos-free di- has been in operation for over ten years [116].
aphragm based on the built-in activation con-
cept. The Chloralp separator is made of two PPG Industries Tephram Diaphragm.
vacuum-deposited layers: PPG Industries has developed and commercial-
1) The first layer, known as the precathode, is ized the Tephram diaphragm as its entry into the
a conductive mat of carbon fibers, contain- non-asbestos diaphragm market. Major goals of
ing an electrocatalytic powder to decrease the PPG’s non-asbestos program were to produce
cathode overpotential a diaphragm that deposited and operated sim-
2) The second layer is the diaphragm itself, in ilarly to asbestos, with longer diaphragm life
which PTFE and inorganic materials have re- and lower power consumption. This technology
placed asbestos utilized vacuum deposition to produce a base
diaphragm composed of PTFE and a topcoat
The cathodic activation provides energy savings from a slurry of inorganic particulate materials.
from 50 to 150 mV, depending on the plant oper- Dopants are priodically added to the anolyte dur-
ating conditions, and the catalyst content in the ing cell operation to adjust the diaphragm per-
mat. Other benefits from the precathode are elec- meability to maintain or improve cell operation.
trocatalytic destruction of chlorates and lower This diaphragm has been successfully tested of
hydrogen contents in the chlorine after plant the major monopolar cell technologies, as well
shutdowns. The diagraphm is deposited on the
Chlorine 33

as on PPG’s Glanor bipolar cells. An active pro- The cell employs vertical DSA coated tita-
gram is in place to operate diaphragms on a trial nium anodes, vertical cathodes of woven wire
basis at various producer sites, allowing for site- mesh bolted to a perforated steel backplate,
specific evaluation of the overall economics of and a vacuum-deposited modified asbestos di-
the diaphragm. aphragm. A single bipolar element may have
100 m2 of both anode and cathode active area.
The anode of one element is connected to the
6.2.1. Dow Cell [2, 117, 118] cathode of the next by copper spring clips. This
connection is immersed in the cell liquor dur-
The Dow Chemical Company is the largest ing operation. Figures 30 B and 31 show these
chlor-alkali producer, accounting for one-third internal cell parts.
of the U.S. production and one-fifth of the world
capacity. Because Dow’s production capacity is
large and concentrated in a few sites, Dow’s cell
development followed a different path than other
chlor-alkali technology developers. Dow uses its
own cell design of the filter press bipolar type.
Dow has operated filter press cells for over 90
years. Dow cell development occurred in several
stages, characterized by simple rugged construc-
tion and relatively inexpensive materials.

Figure 31. Dow diaphragm cell, section view


a) Perforated steel backplate; b) Cathode pocket;
c) Asbestos diaphragm; d) DSA anode; e) Copper
backplate; f) Titanium backplate

Dow operates at lower current densities than


others in the chlor-alkali industry. The elec-
trolyzers are normally operated with 50 or more
cells in one unit or series. One electrical cir-
cuit may consist of only two of these electrolyz-
ers. Figure 30 A shows a view of six electrolytic
Figure 30. Dow diaphragm cell cells.
A) Six-cell series
B) Internal cell parts: a) Cathode element; b) Cathode Treated saturated brine is fed to the anolyte
pocket elements; c) Copper spring clips; d) Perforated steel compartment, where it percolates through the
backplate; e) Brine inlet; f) Chlorine outlets; g) Copper diaphragm into the catholyte chamber. The per-
backplate; h) Titanium backplate; i) Anode element colation rate is controlled by maintaining a level
of anolyte to establish a positive, adjustable hy-
34 Chlorine

drostatic head. The optimum rate of brine flow


usually results in the decomposition of ca. 50
% of the incoming NaCl, so that the cell liquor
is a solution containing 8 – 12 wt % NaOH and
12 – 18 wt % NaCl.
The Dow diaphragm cell, optimized for low
current density, consumes less electrical energy
per unit of production than the rest of the indus-
try. The cell voltage at these low current densities
is only 300 – 400 mV above the decomposition
potential of the cell. However, Dow has a larger
investment in the electrolyzers, especially an-
odes.
The electrolyzers are operated at ca. 80 ◦ C,
lower than the 95 ◦ C typical of other types of
cells. This lower operating temperature allows
cell construction with less expensive materi-
als, such as vinyl ester resins and other plastics
[117]. Operating data have not been published.

Figure 32. Glanor bipolar electrolyzer


6.2.2. Glanor Electrolyzer [12, 119 – 121] a) Disengaging tank; b) Chlorine outlet; c) Hydrogen out-
let; d) Bipolar element; e) Brine inlet; f) Cell liquor trough;
g) Cell liquor outlet
Glanor bipolar electrolyzers are a joint de-
velopment of PPG Industries and Oronzio De
Nora Impianti Elettrochimici S.p.A. The Glanor
electrolyzer consists of several bipolar cells
clamped between two end electrode assemblies
by means of tie rods, thereby forming a fil-
ter press type electrolyzer (Fig. 32). The elec-
trolyzer is equipped with DSA titanium anodes.
Each electrolyzer normally consists of 11 or 12
cells. A lower number of cells can, however, be
assembled in one electrolyzer. The Glanor elec-
trolyzer was especially designed for large chlor-
alkali plants.
The current is fed into the electrolyzer by
Figure 33. Glanor bipolar electrolyzer type V-1144
means of anodic and cathodic end elements. The
anodic compartment of each cell is connected
to an independent brine feed tank by means of The V-1144 electrolyzer (Fig. 33) was the
flanged connections. first commercial unit, and eight plants uti-
Chlorine gas leaves each cell from the top lize this model. The second generation is the
through the brine feed tank and then passes to V-1161 electrolyzer, which employs Modified
the cell room collection system. Hydrogen gas Diaphragms, narrower electrode gaps, lower
leaves from the top of the cathodic compartment current density, and DSA anodes to achieve
of each cell, while the catholyte liquor leaves lower power consumption than the V-1144 elec-
from the bottom through an adjustable level con- trolyzer.
nection. The operating characteristics of the Glanor
electrolyzers are shown in Table 8.
Chlorine 35

Table 8. Glanor bipolar diaphragm electrolyzers: design and


with both ends open, extending across the cell,
operating characteristics as the circulation space requirement was satis-
fied by the change from solid graphite anodes to
Item Model Model
V-1144 V-1161 the open DSA anodes (Fig. 34).
Cells per electrolyzer 11 11
Active anode area per cell, m2 35 49
Electrode gap, mm 11 6
Current load, kA 72 72
Current density  at 72 kA, kA/m2 2.05 1.47
Cell voltage, V 3.50 3.08
Current efficiency, % 95 – 96 95 – 96
Power consumption (d.c.), 2500 2200
kWh/t *
Anode gas composition
(alkaline brine)
Cl2 , % 97.3 – 98.0 97.0 – 98.0
O2 , % 1.5 – 2.2 1.5 – 2.2
H2 , % <0.1 <0.1
CO2 , % 0.4 0.4
Figure 34. OxyTEch type H-4 cell
Cell liquor
NaOH, g/L 135 – 145 135 – 145
NaClO3 , % 0.03 – 0.15 0.03 – 0.15
Production per electrolyzer
Chlorine, t/d ** 26.7 26.7
NaOH, t/d 29.8 29.8

* Per short ton of chlorine.


** Short tons.

6.2.3. OxyTech “Hooker” Cells [2, 12, 122,


123]

The first commercialized deposited asbestos di-


aphragm cell was the Hooker type S-1 monopo-
lar cell, introduced in 1929. The basic design fea-
tured vertical graphite anode plates connected to
a copper bus bar and a cathode with woven steel
wire cloth or perforated steel fingers between the Figure 35. OxyTech/Uhde HU-type cells
a) Cell bottom; b) Cathode; c) Anode; d) Cell cover; e) Bus
anodes. The cathode held vacuum-deposited as- bars; f) Brine level gauge; g) Brine flow meter; h) Bypass
bestos fiber diaphragms that separated the anode switch
and cathode compartments. The cathode fingers
did not extend completely across the cell, but Table 9 is a summary of operating character-
left a central circulation space. In the following istics and current densities of the H-series cells
40 years, a family of S series cells with similar currently available for license.
characteristics evolved, with over 12 000 having
been installed in licensed plants.
In 1973, a new H series of monopolar cells
was introduced. They incorporated the use of 6.2.4. HU Monopolar Cells [123]
DSA anodes, which had been developed and
commercialized in the late 1960s. These cells The HU type cells were a joint development
have significant voltage savings over the S series, of Hooker (now OxyTech Systems) and Uhde.
thus allowing increases in cell capacity without The HU-type electrolyzer (Fig. 35) is rectangu-
corresponding increases in rectification capac- lar, not cubic, and is narrow in the direction of
ity. The H series also incorporate cathode tubes current flow, since anodes are arranged in a sin-
36 Chlorine

Table 9. OxyTech Systems Hooker H-series diaphragm cells: design and operating characteristics

H-2A H-4

Operating current, A 80 000 150 000


Anode area, m2 36.16 64.52
in.2 56 050 100 000
Current density, A/m2 2212 2325
A/in.2 1.43 1.50
Cell voltage, V 3.44 3.44
Approximate cell 1.87×2.66 2.58×3.11
dimensions, m
Diaphragm life, days 300 – 500 300 – 500
Anode life, years 5–7 5–7
Operating NaOH
concentration, g/L 140 160 140 160
% 11.35 12.89 11.33 12.87

Current Efficiency, % 96.4 94.6 96.6 94.9


Chlorine output,
metric ton/day 2.45 2.41 4.60 4.52
short ton/day 2.70 2.65 5.07 4.98
Caustic soda output,
metric ton/day 2.76 2.71 5.19 5.10
short ton/day 3.05 2.99 5.72 5.62

Table 10. HU series diaphragm cells: design and operating characteristics

Item Cell type


HU 24 HU 30 HU 36 HU 42 HU 48 HU 54 HU 60

Number of anodes 24 30 36 42 48 54 60
Anode surface area, m2 20.6 25.8 31.0 36.1 41.3 46.4 51.6
Load, kA 30 – 45 40 – 60 50 – 70 55 – 85 60 – 95 70 – 105 80 – 120
Cl2 production, t/d 0.90 – 1.36 1.19 – 1.82 1.49 – 2.12 1.64 – 2.58 1.79 – 2.88 2.09 – 3.18 2.39 – 3.64
NaOH (100 %) production, t/d 1.01 – 1.54 1.35 – 2.05 1.68 – 2.39 1.85 – 2.91 2.02 – 3.25 2.36 – 3.59 2.69 – 4.10
H2 production, kg/d 25 – 39 34 – 52 42 – 60 47 – 73 51 – 82 59 – 91 68 – 103
Cell length, m 2.1 2.6 3.0 3.5 3.9 4.4 4.8
Distance, cell-to-cell, m 1.5 1.5 1.5 1.5 1.5 1.5 1.5

gle row. The cathode is long and narrow; conse- cury cells, creating a second operating floor. The
quently, the current density is lower through the interconnecting bus bars are flexible and are dis-
cathode shell. The long, narrow cathode fabri- tributed over the entire length of the cell. The
cation lends itself to closer anode – cathode tol- HU cell design incorporates a bus bar for each
erances and spacing. Copper on and around the individual anode. This, as well as the elevation
cathode shell has been eliminated. Another ad- of the cell from the floor below, which allows
vantage of the long, narrow design is shorter access, enables connection of facilities for mon-
electrolysis current paths through the cell room, itoring the current flowing through each anode.
resulting in savings in piping and other materi- During operation of the bypass switch, connec-
als. The HU-type cell incorporates a Modified tion is made for each individual anode, and no
Diaphragm. additional contact bus bars are required.
A further novelty of the HU cell system is the The HU-type cells are offered to cover 30 –
design and arrangement of the bypass switch. 150 kA. All of the different cell types are
The HU switch is installed underneath, not next equipped with cathodes and anodes of identi-
to, the circuit of cells. This is accomplished by cal height and width. The only basic difference
raising the cells from the floor, similar to mer- between the various cell models is the number of
Chlorine 37

elements and consequently the length of the cell Copper connectors attached at the ends of the
(Table 10). Cell voltage and power consumption bonded copper side plates complete the encom-
per tonne of chlorine, identical for all cell types, passing of the cathode with copper. Anodes are
are shown in Table 11 for the specific current connected to a copper patented cell base, which
loads of 1.5 and 2.3 kA/m2 . is protected from the anolyte by a rubber cover or
Table 11. HU series diaphragm cells: specific load, cell voltage,
a titanium base cover (TIBAC) [125]. Orienta-
and power consumption tion of the cathode tubes is parallel to the cell cir-
cuit, the opposite of a Hooker-type cell. This ar-
Specific load, kA/m2
1.5 2.3
rangement accommodates thermal expansion of
the cell and circuit without changing the anode-
Cell voltage, V 3.12 3.41 to-cathode alignment.
Power consumption
(d.c., average), kWh/t* 2500 2700 The combination of the Modified Diaphragm
and expandable DSA anodes reduces power con-
* Per tonne of chlorine.
sumption by 10 – 15 % from that of regular as-
bestos diaphragms and standard, fixed DSA an-
odes [126]. Table 12 presents performance data
for the two most common MDC cell sizes [124].
The OxyTech MDC-29 is shown in Fig. 38. The
licensed chlorine capacity of OxyTech cells now
exceeds 20 000 t/d.

6.3. Operation

The process description in this section is in-


tended to provide an overview of typical di-
aphragm cell process areas. A general block di-
agram for a diaphragm cell facility is shown in
Figure 9. Included on the drawing are many pro-
cess areas that may be optional, depending on
the design of the plant and its end products. The
operation of a cell room may be broken down
into six areas: the two incoming systems, brine
and electrical; the cells; and the three outgo-
Figure 36. OxyTech Systems MDC cells ing systems, chlorine, hydrogen, and cell liquor.
a) Brine feed rotometer; b) Head sight glass; c) Cell Some of these are essentially the same for all
head; d) Cathode assembly; e) Tube sheet; f) Grid plate;
g) Cathode tube; h) Grid protector; i) DSA expandable an- three chlor-alkali processes and are described
ode in Chapter 4 — the brine system (general), the
electrical system, and the hydrogen system. The
treatment of the chlorine is the subject of Chap-
6.2.5. OxyTech MDC Cells [12, 124] ter 11. Only aspects that are reasonably specific
to the diaphragm cell process are described in
OxyTech Systems manufactures and licenses this section.
the MDC series of monopolar diaphragm cells
(Fig. 36). The MDC cells feature woven steel
wire cathode screen tubes open at both ends, 6.3.1. Brine System
which are welded into thick steel tube sheets at
each end. The tubes, tube sheets, and the outer Most commonly, diaphragm cells are supplied
steel cathode shell form the catholyte chamber with well brine on a once-through basis. The
of the cell (Fig. 37). Copper is bonded, rather treated well brine flows to the treated brine stor-
than welded, to the rectangular cathode shell on age tanks, which usually have 12-h capacity.
the two long sides parallel to the tube sheets. From there the brine is fed to the cell room. The
38 Chlorine
Table 12. OxyTech Systems MDC cells: operating capacities and characteristics

Item Model number and operating range, kA

MDC-29 MDC-55

35 to 80 75 to 150

Chlorine capacity,
metric ton/day 1.05 2.41 2.33 4.53
short ton/day 1.16 2.66 2.48 5.00

Caustic capacity,
metric ton/day 1.21 2.76 2.59 5.18
short ton/day 1.33 3.04 2.85 5.70

Hydrogen capacity,
m3 /day 335 765 720 1435
cubic feet/day 11 830 27 010 25 420 50 670

Current density,
kA/m2 1.21 2.76 1.37 2.74
A/in.2 0.78 1.78 0.88 1.76

Cell voltage, V a
steel cathode 2.90 3.62 3.00 3.62
activated cathode 2.80 3.51 2.90 3.51

Power consumed
(d.c., steel
cathode) b ,
kW h/t 2310 2876 2390 2870
kW h/short ton 2100 2610 2175 2610

Power consumed
(d.c., activated
cathode) b ,
kW h/t 2230 2786 2310 2780
kW h/short ton 2025 2530 2100 2530

Diaphragm life, years 1–2 0.5 – 1.0 1–2 0.5 – 1.0


Anode life, years 8 – 10 5–8 8 – 10 5–8
Cathode life, years 10 – 15 10 – 15 10 – 15 10 – 15

Distance between
cells c ,
m 1.60 2.13
inches 63 84
a
Cell voltage includes loss in intercell bus.
b
Power consumed per ton (metric or short) of chlorine produced.
c
Distance centerline-to-centerline and side-by-side with bus connecting.

flow to each individual electrolyzer is controlled Brine sludges or out-of-spec brine can be sent
by a rotameter. If the flow of brine to the cells is to the lagoon. Supernatant clear brine can be re-
suddenly disrupted by failure of the brine feed covered from the lagoon.
pump, the rectifiers automatically shut down In most cases, operation with acidic brine is
since an inadequate supply of brine to the cells preferred because of the reduced amount of side-
is potentially dangerous. The specifications for reaction products in the chlorine and the cell
brine for diaphragm cells are given in Table 13. liquor.
A brine recovery lagoon is usually available
to handle any major upsets in the brine system.
Chlorine 39

Figure 37. Exploded view of an OxyTech MDC-55 cathode


a) End plate; b) Rim screen; c) Side screens; d) Tube sheet; e) Full cathode tube; f) Half-cathode or end tube; g) Side
plate; h) Lifting lug; i) Punched and coined stiffener strap; j) Bosses; k) End plate, operating aisle end; l) Hydrogen outlet;
m) Connector bar; n) Caustic outlet; o) Clip angles; p) Grid bar, connector side; q) Side plate

6.3.2. Cell Room

Typical cell rooms are shown in Figures 39


(bipolar cells) and 40 (monopolar cells).

Figure 39. Cell room: bipolar PPG Industries/De Nora


Glanor cells

A cell in normal operation requires little at-


tention. The critical requirement is that the brine
flow rate is sufficient to maintain an anolyte level
above the cathode.
Figure 38. OxyTech MDC-29 with the author (1971)
40 Chlorine

Table 13. Typical brine feed specifications for diaphragm cells


Any change in the anolyte level or brine flow rate
should be investigated.
Parameter Specification As the cell ages, the diaphragm will undergo
NaCl 320 g/L
changes in porosity because of the following:
pH 2.5 – 3.5
Hardness (Ca2+ + Mg2+ ) <5 ppm
1) Electrolysis effect
Magnesium <0.4 ppm 2) Brine impurities
Sodium sulfate (Na2 SO4 ) <5 g/L 3) Upsets in operation
Organics <1 ppm 4) Gradual wear of the diaphragm
Manganese <0.01 ppm
Barium <0.01 ppm A change in porosity may necessitate a
Nickel <0.1 ppm
Iron <0.5 ppm
change in brine flow rate. If the increase in poros-
Silicon <15 ppm ity is severe, the cell may be replaced or doped
Cobalt <0.02 ppm with an asbestos slurry or inorganic salts.
Mercury <1 ppm Impurities in the brine often lead to decreased
Phosphate <1 ppm
porosity. Decreased porosity can be offset to
some extent by increasing the anolyte level and,
if necessary, by lowering the catholyte level.
A cell operated with the anolyte level at the
maximum value and the lowest catholyte level
is called a sleeper. To gain additional diaphragm
life after a cell has entered the sleeper position,
the brine flow rate must be decreased below nor-
mal. This is not normally a recommended prac-
tice because current efficiencies of these cells
are usually low.
For safe operation of diaphragm cells, the
header pressures must be maintained at the
proper values.
The chlorine header should be maintained at
positive pressure to permit detection and correc-
tion of any chlorine piping leaks. The hydrogen
header is also maintained at a positive pressure
to avoid pulling air into the hydrogen, creating a
potentially explosive mixture. The brine header
pressure should be maintained to give the desired
caustic concentration in the cell liquor. Normal
practice is to adjust individual brine feed valves
so that each cell receives the correct brine flow
rate.
Load changes must be smooth to avoid fluc-
tuations in the header pressures and detrimental
effects on the diaphragms.
Figure 40. Cell room: monopolar OxyTech H-4 cells
The brine feed rate to each cell should be in-
creased to the new rate before circuit amperage is
Under no circumstances should a cell be op- increased. The brine feed rate to each cell should
erated with an inadequate or excessive anolyte be decreased immediately after amperage is de-
level. Operation with the anolyte level not visible creased. During any period of operation when
in a sight glass is unsafe. At least one operator brine flow rates are being changed, extra atten-
should be in the cell room at all times. The cell tion should be given to the anolyte levels of the
room operator should inspect the anolyte level cells. Adjustment of the brine feed temperature
and brine flow to each cell at least once per hour. may also be necessary when a load change oc-
curs.
Chlorine 41

Figure 41. Process flow diagram: triple-effect caustic evaporator


a) First-effect vapor body; b) First-effect heat exchanger; c) Second-effect vapor body; d) Second-effect heat exchanger;
e) Second-effect forwarding pumps; f) 50 % caustic transfer pumps; g) Third-effect vapor body; h) Third-effect heat exchanger;
i) Third-effect forwarding pump; j) Barometric condenser; k) First-stage ejectors; l) Intercondenser; m) Second-stage ejector;
n) Liquor flash tank

6.3.3. Diaphragm Aging ing chemical attack on the asbestos. Diaphragm


cell plant operators should strive to minimize
Of all the cell components, the diaphragm usu- these changes.
ally has the shortest life. The ability of a di- The real importance of the equations in Sec-
aphragm to resist the back migration of hydrox- tion 6.1 is as an aid in deciding when the di-
ide slowly becomes impaired with service life. aphragms should be replaced.
The performance of the diaphragm deteriorates
for the following reasons:
1) Chemical attack 6.3.4. Treatment of the Products
2) Brine impurities
3) Unsteady operating conditions Chlorine. See Chapter 11.
The major reason for the deterioration is Hydrogen. See Chapter 4.
chemical attack on the asbestos by the alkaline
catholyte and acidic anolyte. The rate of chemi-
Sodium Hydroxide Solution. The hydrox-
cal attack can be minimized and diaphragm life
ide produced at the cathode is associated with
maximized by careful operation of the cell. The
sodium ions and water to form a 10 – 12 wt %
most important situations to avoid are high con-
sodium hydroxide solution leaving the elec-
centrations of brine impurities and unsteady op-
trolytic cell. This cell liquor also contains
erating conditions. High brine impurities cause
18 wt % unreacted sodium chloride.
plugging of the diaphragm with insoluble hy-
Most large modern chlor-alkali plants have or
droxides, which reduce the diaphragm’s sepa-
will soon have an associated cogeneration power
ration ability. The most common harmful im-
plant. In these facilities, the caustic evaporators
purities are calcium, magnesium, iron, nickel,
are an important use for the byproduct steam.
silicates, aluminum, manganese, and barium.
Modern diaphragm cell plants use triple-
Unsteady operation, such as electrical load
effect evaporators and, in many cases,
changes, cell liquor strength changes, changes
quadruple-effect evaporators.
in brine concentration or pH, gas-pressure fluc-
tuations, and shutdowns, change the pH of the
various regions of the diaphragm, thus accelerat-
42 Chlorine

Figure 42. Caustic purification system


a) 50 % caustic feed tank; b) 50 % caustic feed pumps; c) Caustic feed preheater; d) Ammonia feed pumps; e) Ammonia
feed preheater; f) Extractor; g) Trim heater; h) Ammonia subcooler; i) Stripper condenser; j) Anhydrous ammonia storage
tank; k) Primary flash tank; l) Evaporator reboiler; m) Evaporator; n) Caustic product transfer pumps; o) Purified caustic
product cooler; p) Purified caustic storage tank; q) Ammonia stripper; r) Purified caustic transfer pumps; t) Overheads con-
denser; u) Evaporator; v) Evaporator vacuum pump; w) Aqueous storage ammonia tank; x) Ammonia scrubber; y) Scrubber
condenser; z) Ammonia recirculating pump; aa) Ammonia recycle pump

Caustic Soda Evaporation. A flow diagram concentrate liquor, which flows by gravity to
for a typical triple-effect caustic soda evapora- the pressure filter feed tank, contains ca. 1.0 –
tor is shown in Figure 41. The evaporator is of 1.5 wt % dissolved NaCl and ca. 0.1 wt % crys-
the backward-feed design and concentrates 10 – talline NaCl.
11.3 wt % NaOH cell liquor to 50 wt % NaOH. The liquor is pumped from the pressure filter
Liquor flows from the third to the second to the feed tank into the pressure leaf filters, where the
first effect and from the first effect to the liquor remaining traces of salt crystals are removed.
flash tank. A cyclone is used for each effect to The product caustic flows by gravity to the fil-
utilize the pressure drop across the circulating tered product tank and then is pumped to storage.
pump to clarify the transfer liquor. Salt removed in the pressure filters is reslurried
The salt precipitated in the liquor flash tank with cell liquor and pumped to the evaporator
is isolated from the rest of the salt precipitated feed tank via the filter backwash pump.
in the evaporator and used as seed crystals in The salt discharged from the centrifuges
the cooling system to help diminish coil scaling drops into the leaching tank, where it is reslur-
and supersaturation of the product liquor with ried with condensate and recycled brine from
sodium chloride. The sodium chloride and triple the Glauber’s salt (Na2 SO4 · 10 H2 O) crystal-
salt (NaOH – NaCl – Na2 SO4 ) precipitated in lizer. Concentrate from the pusher centrifuges
the liquor flash tank and cooling system is re- flows by gravity into the evaporator feed tank.
moved from the cooled product liquor with cen- The product salt is discharged from a cyclone
trifuges. The salt precipitated in the three effects into the salt reslurry tank. The overflow from
flows countercurrent to the liquor flow so that all the cyclone is returned to the leaching tank. The
of the salt is discharged from the last effect, the product salt is diluted with brine and pumped to
effect that has the coldest liquor and the lowest the resaturator tank.
caustic soda concentration. Brine containing the dissolved sodium sul-
Two-stage steam-jet air ejectors with a com- fate is separated from the salt crystals in a cy-
mon intercondenser are used to maintain vac- clone. The underflow returns to the leaching
uum in the evaporator. In the caustic cooling tank. The overflow is collected in the feed tank
system, agitated tanks are used to cool the slurry for the Glauber’s salt crystallizer. Sodium sulfate
discharged from the liquor flash tank. The slurry is crystallized from the liquor in a continuous
flows from the cooling system to the centrifuge vacuum cooled crystallizer.
feed tank from where it is pumped into cen- Mother liquor removed from the crystallizer
trifuges. Salt discharged from the centrifuges is pumped under level control to the brine tank.
drops into the evaporator feed tank, where it Slurry discharged from the crystallizer is thick-
is dissolved in cell liquor. The 50 wt % NaOH ened to ca. 50 wt %. The liquor from the thick-
Chlorine 43

ener is collected in the brine tank and pumped ampere load on each circuit
back to the leaching tank. The thickened slurry is voltage for each circuit
redissolved in the Glauber’s salt dissolving tank chlorine header pressure
and pumped to a waste treatment system. hydrogen header pressure
Caustic Purification [127]. Diaphragm-cell brine header pressure or flow rate
chlor-alkali producers requiring higher purity brine temperature
caustic than that produced by the diaphragm pro- brine pH
cess can use caustic purification or DH process cell liquor temperature
(Fig. 42). Salt removal in the purification unit is
Samples of brine should be taken every 4 h
effected by contacting the 50 wt % caustic with
and combined into a daily composite. In addi-
anhydrous liquid ammonia under pressure suffi-
tion, samples of cell liquor should be taken from
ciently high to maintain all materials in the liquid
each cell string, the sodium hydroxide content
state.
analyzed, and the temperature taken every 4 h.
The liquid ammonia absorbs salt, chlorate,
A daily composite should be made and samples
carbonate, water, and some caustic. It is then
should be analyzed by the laboratory for the fol-
stripped, concentrated, and returned to the ex-
lowing:
traction process. The concentrated caustic leav-
ing the extractor is stripped free of ammonia, NaOH content
which is recovered, concentrated, and recircu- NaCl content
lated. Typical purities before and after caustic salt : caustic ratio
purification are shown in Table 20. This pro- NaClO3 content
cess is offered for license by PPG Industries and NaOCl content
OxyTech Systems. Fe content
In addition, producers and users of average temperature
diaphragm-cell caustic may wish to reduce specific gravity at 25 ◦ C
metal impurities by utilizing the porous cath- Chlorine gas from each cell circuit should be
ode cell process (PPG Industries) [128]. The analyzed for chlorine and hydrogen content at
process consists of an electrolysis cell with least twice each 8-h shift. Each day a complete
porous nonmetallic cathodes. The caustic soda analysis of the chlorine header gas should be
(50 wt %) is freed from iron, nickel, lead, and made. Additions or extensions to list may be dic-
copper, which are deposited on the cathode. The tated by plant operation.
cell must be regenerated periodically with water Each plant must develop a procedure for tak-
and hydrochloric acid. Typical feed and product ing individual cell data so that individual cells
analyses based on anhydrous NaOH are may be scheduled for renewal. The following is
a minimum schedule:
Metal Content, ppm
Feed Product weekly voltage and cell liquor composition
Iron 10.0 2.0 (NaOH, NaCl, NaClO3 )
Nickel 3.0 0.2
Lead 4.0 0.4 monthly chlorine composition (Cl2 , O2 , H2 ,
Copper 0.2 0.1
CO2 , N2 )

6.3.5. Measurement 7. Membrane Process

Recorded data is an important tool for determin- In the membrane process, the anolyte and
ing the operating condition of the plant and di- catholyte are separated by a cation-exchange
agnosing problems. membrane that selectively transmits sodium ions
The following should be recorded continu- but supresses the migration of hydroxyl ions
ously or hourly: from the catholyte into the anolyte. A strong
caustic soda solution with a very low sodium
44 Chlorine

chloride content can be obtained as the catholyte The anolyte is discharged from the cell.
efflux. The advantages of the membrane process The electric field causes hydrated sodium ions
are its energy efficiency and its ability to produce to migrate through the membrane into the
caustic soda of high quality, with almost no im- catholyte. In the cathode compartment, hydro-
pact on the environment. Depending on the par- gen is evolved at the cathode, leaving hydroxyl
ticular design, membrane sizes from 0.2 to 5 m2 . ions, which together with permeating sodium
The production capacity of an electrolyzer can ions constitute the caustic soda:
be up to 90 t/d NaOH (100 %).
The process was started in the early 1970s
2 H2 O+2 e− → H2 +2 OH−
with development of the perfluorosulfonate
membrane Nafion by DuPont [129]. In 1975, a
perfluorocarboxylate membrane capable of pro-
Na+ +OH− NaOH
ducing 35 wt % caustic soda became available,
from Asahi Glass in Japan [130]. In 1978 the
first two-layer membrane was developed, with Liquid and gaseous phases anolyte/Cl2 and
low electrical resistance and high current effi- catholyte/H2 can be separated either in the cell
ciency [131]. compartment or downstream of the cell outlet.
The industrial success of the membrane pro- The chlorine-saturated anolyte is then treated in
cess started in Japan, where the abolition of the a dechlorination unit to recover the dissolved
mercury process on environmental grounds had chlorine.
been promoted by the government. Today the
membrane process is the state of the art process Membrane.
for producing chlorine and caustic soda or potas- Structure. The membrane is exposed to chlo-
sium hydroxide. rine and anolyte on one side and strong caustic
The production capacity of chlor-alkali plants solution on the other side at high temperature
using the membrane process reached about 21 % (90 ◦ C). Only ion-exchange membranes made
of total world production capacity in 1995 and of perfluoropolymer can withstand such severe
is predicted to increase to about 28 % in 2001 conditions. The ion-exchange groups of the orig-
[132]. inal polymers are in the fluorosulfonate form,
–SO3 F, or the carboxylate form, –COOR.
7.1. Principles
In a membrane cell a cation-exchange mem-
brane separates the anolyte and catholyte, as
shown in Figure 43. Saturated brine is fed into
the anode compartment, where chlorine gas is
evolved at the anode:
Fluorosulfonate form
2 Cl− → Cl2 +2 e−

Carboxylate form m = 0 – 1 n = 1 – 5 R = alkyl


The first membranes that showed signifi-
cant potential for use in the chlor-alkali pro-
cess were made of a perfluorosulfonate layer.
These proved to be durable in chlor-alkali cells
but were relatively inefficient. The perfluorocar-
boxylate polymers, with a lower water content,
Figure 43. Principle of the membrane cell
Chlorine 45

showed higher selectivity but led to higher elec- gen gas can be depressed by selecting an anode
trical resistance and high electrical power con- coating with suitable characteristics (see Section
sumption [133]. Combining the advantages of 8.1) or by decreasing the pH in the anode com-
high current efficiency and low electrical resis- partment by acidifying the inlet brine (Fig. 45).
tance a composite membrane (Fig. 44) was de-
veloped with a layer containing SO− 3 groups on
the anode side and a layer containing COO−
groups on the cathode side [131, 134]. The per-
fluorosulfonate layer is thicker than the perflu-
orocarboxylate one and is the major constituent
of the membrane [133].

Figure 45. Oxygen content in chlorine [135]

Convection and diffusion determine the flow


of uncharged compounds and ions through the
membrane. Chloride anions in the catholyte are
excluded by the cation-exchange membrane and
repelled by the electric field, so that the trans-
fer rate of chloride anions from the anolyte is
extremly low. As a result, a caustic soda solu-
tion of about 32 – 35 wt % with a salt content
of less than 20 ppm can be obtained. The wa-
ter transport through the membrane is about 3.5
to 4.5 moles of water per mole of sodium ions,
Figure 44. Membrane structure and can be regarded as the hydration sphere of
the migrating sodium ions. Water flux increases
Flux through the Membrane. The total flux with decreasing anolyte concentration [136].
through the membrane can be divided in to three Migration, convection and diffusion influ-
parts [133]: ence each other, and the resulting flux depends
on membrane type, current density, temperature,
1) Migration due to electric field
and composition of anolyte and catholyte.
2) Convection
3) Diffusion due to chemical gradients
Cell Voltage. The cell voltage of a mem-
Migration is the flux of ions through the mem- brane cell is composed of the following terms:
brane, driven by electric field. This includes the
desired transfer of sodium ions to the cathode 1) Decomposition voltage
compartment and the undesired transfer of hy- 2) Membrane potential between anolyte and
droxyl ions to the anode compartment. The ca- catholyte
pacity for selective separation of the cation ex- 3) Electrode overpotentials for chlorine and hy-
change membrane is determined by its repulsive drogen
force for hydroxyl ions. This effect determines 4) Ohmic drop in the membrane
the current efficiency. 5) Ohmic drop in the electrolytes
The backmigration of hydroxyl ions in- 6) Ohmic drop in electrodes and conductors
creases the formation of oxygen, hypochlorite, Term 1. The decomposition voltage of the
and chlorate in the anode compartment and chlor-alkali process is about 2.20 V, depending
causes a loss of current efficiency of 3 – 7 % in on temperature, concentration, and pressure.
caustic soda production. The evolution of oxy-
46 Chlorine

Term 2. This term describes the overpoten- approximately 2 – 4. This leads to a pH gradient
tials at the surfaces of the membrane. Un- across the membrane cross section. The solubil-
der standardized operating conditions (3 kA/m2 , ity of impurities, which are always present in
90 ◦ C, 32 wt % caustic solution), the membrane the pure brine, depends on the pH. Therefore,
potential is approximately 0.08 V. depending on the type of impurities and on the
Term 3. Titanium anodes coated with oxides pH, precipitation inside the membrane can take
of Ir, Ru or Pt are generally used in membrane place. This leads to mechanical destruction of
cells and lead to a chlorine overvoltage of ap- the membrane, which has a irreversible effect
proximately 0.05 V at 3 kA/m2 . Hydrogen over- on current efficiency. In addition, the cell volt-
potentials of about 0.1 V at 3.0 kA/m2 are at- age rises due to the crystals formed inside the
tained with activated cathodes. Mainly nickel membrane.
substrates are coated by painting and thermal
treatment or galvanic deposition [137, 138].
Coating materials include Ni, Co, Ru and others 7.2. Process Specific Aspects
(see Section 8.2).
Term 4. The ohmic drop of advanced com- The performance of a membrane cell depends
mercial membranes under standardized operat- on the following operating conditions:
ing conditions is about 0.25 – 0.30 V at 3 kA/m2 . 1) Concentration of anolyte and catholyte
Term 5. To minimize the ohmic drop of an 2) Current density
electrolyte, the gaps between the membrane and 3) Temperature
the electrodes are minimized in membrane cells. 4) Brine impurities
However, if the gap is very small, a rise in voltage The optimum caustic strength depends on
is observed due to the entrapment of gas bubbles the composition of the membrane polymer. To
between the electrodes and the hydrophobic flu- achieve stable operation with high current ef-
oropolymer membrane. ficiency, fluctuations in operating conditions or
Term 6. The voltage losses in the electroly- upsets must be avoided. Fluctuations in the caus-
sis cell that occur due to unfavorable current tic strength beyond the optimum range influence
paths along the metallic structure are reduced both the current efficiency and the cell voltage,
by an appropriate design. In modern chlor-alkali as shown in Figure 46 [131]. Dilution of the
electrolysis cells the typical ohmic drop is 20 – anolyte caused by an upset in the brine feed also
40 mV at 3 kA/m2 . decreases the current efficiency. The sensitivity
of membrane performance to operating condi-
Current Efficiency. The current efficiency
tions is attributed to changes in the water content
(CE) for caustic soda can be obtained either by
of the membrane.
directly measuring the quantity of caustic soda
produced or by an anodic balance, i.e., the com-
positions of the anode gas and the anolyte with
the following equation:

CE (%,NaOH)
= 100−ηO2 −ηClO3 −ηClO+ηNaOH+ηNa2 CO3

where ηO2 , ηClO, and ηClO3 represent the


loss of current efficiency due to the genera-
tion of oxygen hypochlorite, and chlorate, while
ηNaOH and ηNa2 CO3 take into account NaOH
and Na2 CO3 introduced in the feed brine.
The current efficiency is mainly dependent
on membrane performance. On the cathode side
the membrane is in contact with a concentrated
NaOH solution, while the anode side has a pH of Figure 46. Dependence of cell voltage and current effi-
ciency on NaOH concentration [139]
Chlorine 47

7.2.1. Brine Purification hydroxyl ions from migrating from the cathode
compartment through the membrane to the an-
The introduction of membrane technology into ode compartment [143]. This is usually a re-
chlor-alkali electrolysis has dramatically in- sult of physical damage caused by precipita-
creased the demands on brine purity [140]. The tion and crystallization of impurities inside the
lifetime of chlor-alkali membrane cells is deter- membrane. Impurities precipitate because the
mined by the operating conditions and the qual- environment in the membrane changes from an
ity and purity of the feed into the electrolyzers. acidic salt solution (pH 2 – 4) to a caustic solu-
Good long-term performance of the cells may tion (pH 14 – 15) over the 100 – 300 µm thick-
be obtained if brine impurities are kept within ness of the membrane.
the limits recommended in Table 14. It is important to consider not only the impu-
A major source of performance decline is rities themselves but also their interaction. The
the accumulation of solid material in the mem- presence of one impurity may not be harmful,
brane [141]. Specific impurity levels are depen- but its synergistic combination with others may
dent on membrane design, cell design, operating be [143]. For example, silica itself is not harmful
conditions, the impurity itself and other impu- for membranes. Only in the presence of calcium
rities present. The prerequisite for long mem- and aluminum do precipitates form and damage
brane life is to maintain low levels of, for ex- the membrane irreversibly. The concentration of
ample, Ca2+ , Mg2+ , Sr2+ , Ba2+ , Al3+ , SO2− 4 silica and/or the concentration of aluminum plus
and SiO2 in the brine. Traces of these impuri- calcium can be adjusted to give the optimum op-
ties damage the membrane and/or electrodes and erating conditions. For example, with an effec-
result in irrecoverable decreases in current effi- tive secondary brine purification, higher levels
ciency and/or increased cell voltage. In the case of silica can be tolerated. Similarly, if aluminum
of a closed brine loop with no purge, each impu- concentration is high, calcium or silica concen-
rity brought into or formed in the system must be tration must be reduced to maintain acceptable
removed to keep it below its specification level membrane performance.
and to prevent accumulation. To meet the strict requirements on brine pu-
The contaminants can be brought into the rity outlined in Table 14 brine treatment is gener-
brine system by salt, by chemicals used in brine ally carried out in the following main steps in the
purification steps, by water for dissolving the brine loop: saturation, precipitation, clarifica-
salt, from materials of tanks, pipework, and cell tion, filtration, polishing filtration, ion exchange,
components, or by the process itself [141]. The electrolysis, chlorate decomposition and dechlo-
impurities in the salt depend upon the origin of rination.
the raw material. Rock salt, vacuum salt, sea salt, Calcium and magnesium are precipitated and
brine from well mining, or salt from waste in- separated from the saturated brine with the insol-
cinerators serve as supplies of NaCl. The more uble materials. A ca. 10 wt % sodium carbonate
varied the sources are, the more diverse the im- and barium carbonate (barium chloride) solution
purities. and 32 wt % caustic soda are used as precipi-
Membrane and electrode damage effect cell tants.
performance, i.e., cause lower current efficiency,
increased cell voltage, and, as a result, increased
Ca2+ +Na2 CO3 → CaCO3 +2 Na+
power consumption [142]. Some impurities af-
fect the anode or cathode coating and cause an
increase in overvoltage or simply deposit in the SO2− 2−
4 +BaCO3 → BaSO4 +CO3
membrane, increasing its resistance and thus the
cell voltage. The increase in voltage may in some
cases be partially reversible when the impurity Mg2+ +2 NaOH → Mg(OH)2 +2 Na+
concentration drops to the recommended limits.
Current efficiency declines are strictly related
to the membrane. Impurities lower the current Alternatively:
efficiency by reducing the membrane’s ability to
reject anions, specifically the ability to prevent SO2−
4 +BaCl2 → BaSO4 +2 Cl

Table 14. Impurities and their effects

48
Solubility in Negative effect on performance
Max. limit Voltage increase
Impurity Source (w/w) Reagents brine caustic Mechanism Damage An. Cath. Mem. CE PQu Methods of control

Ca2+ salt Ca2+ +Mg2+ Ca2+ ; OH− X - precipitation near the physical disruption of ++ precipitation with
cathode the membrane Na2 CO3 plus ion
< 20 ppb Ca2+ ; X - side of the membrane, ++ exchange
SO2− formation

Chlorine
3
Ca2+ ; X - of large crystals (blister formation) ++
SO2−
3 ;
AL3+
Ca2+ ; I− - - very fine precipitation -
in the membrane
2+ 2+ 2+ 2+
Mg salt Ca +Mg Mg ; X - fine precipitation near binding of ++ precipitation with
< 20 ppb OH− the anode side of the ion-exchange sites in NaOH plus ion
membrane the membrane exchange
Sr2+ salt < 4 ppm Sr2+ ; OH− X X/- precipitation on the physical disruption of + ++ coprecipitation with
cathode side of the the membrane Na2 CO3 plus ion
membrane, formation exchange
of crystals
Ba2+ salt < 0.5 ppm Ba2+ ; OH− X X purge;
Ba2+ ; I− X - very fine precipitation minor damage on the + + precipitation with
in the membrane membrane, minor NaHSO3 plus ion
interaction with exchange
ion-exchange sites
Ba2+ ; coating of the anode +
SO2−
4
3+
Al salt < 0.1 ppm Al3+ X X - precipitation as
hydroxide at pH 7 – 9,
ion exchange under
acid conditions
Al3+ ; X - formation of crystals disruption of the ++
SiO2−
3 (zeolites, sodalites, membrane
faujacites) near the
cathode side of the
membrane
SiO3− ; X - precipitation near disruption of the ++ Purge
Ca2+ cathode membrane
SiO3− ; X - side of the membrane
Al3+ and crystalization
2+
Hg parallel < 0.2 ppm Hg2+ X X deposition on the partially reversible, ++ precipitation with
operation heavy metals cathode covering of active Na2 S
of an amalgam cathode coating
plant
Table 14. Continued

Solubility in Negative effect on performance


Max. limit Voltage increase
Impurity Source (w/w) Reagents brine caustic Mechanism Damage An. Cath. Mem. CE PQu Methods of control

Fe3+ salt, pipework, < 0.1 ppm Fe3+ X X deposition on the covering of active ++ precipitation with
tank material, cathode (in extreme coating (puncturing NaOH
anti-caking cases: dendritic of the membrane)
agent growth from cathode
toward the anode)
Ni2+ salta, < 0.2 ppm Ni2+ ; OH− X X absorption of Ni in the ++ + ion-exchange, purge
pipework, tank heavy metals membrane, deposition
material, on the cathode
cathode
I− salt < 0.2 ppm I− ; Na+ X - precipitation on the physical disruption of ++ purge
cathode side of the the membrane
membrane, formation
of crystals
I− ; Ca2+ - - very fine precipitation -
in the membrane
I− ; K + X X -
F− salt < 0.5 ppm X destruction of the ++ purge
anode coating
+
SO2−
4 salt, < 4 – 8 g/l SO2−
4 ; Na X - precipitation near the reduction of the OH− ++ purge, precipitation
dechlorination Na2 SO4 cathode surface of the ion rejection with BaCO3 or BaCl2
with NaHSO3 membrane capability plus ion exchange
+
SO2−
4 ; Ba coating of the anode +
CO2−
3 salt, < 0.4 g/l CO2−
3 ; H2 O formation of CO2 ++
precipitation Na2 CO3
with Na2 CO3
or BaCO3
ClO−
3 process, side < 10 g/l chlorination of destruction of chlorate
reactions NaClO3 ion-exchange resin ion-exchaange resin decomposition by
acidification
TOC salt < 1 ppm increased foaming, + + + filtration
overplating
K4 (Fe(CN)6 ) anti- < 0.1 ppm Fe3+ X as Fe3+ ++ oxidation with active
caking-agent Fe3+ chlorine plus
for salt precipitation with
NaOH
Chlorine

evolution of N2 ++

∗ An. = Anode; Cath. = Cathode; Mem. = Membrane; CE = current efficiency; PQu = product quality.
49
50 Chlorine

the anolyte to avoid accumulation. Resaturation


From the precipitation tank, the brine is fed then closes the loop.
into the clarifier, where a defined quantity of
flocculant is added to promote the settling of the
precipitated solids and gels. The brine is then 7.2.2. Commercial Membranes
pumped to a filtration system followed by an
The ion-exchange membrane is the key compo-
ion exchange purification.
nent of the membrane cell. The energy consump-
Additionally, if the brine circulating system
tion and the quality of the products depend on
of an existing mercury cell plant also serves
membrane performance. Requirements for the
membrane cells, all mercury must be removed
membrane are as follows:
in a chemical treatment facility. The brine from
the primary filtration is acidified with hydrochlo- 1) Durability under the conditions of chlor-alkali
ric acid to pH 2.0 – 2.5 and sodium sulfide is electrolysis
added to precipitate mercury sulfide. Subse- 2) High selectivity for sodium ion transport
quently, the brine is filtered, alkalized to pH 9.5 – 3) Low electrical resistance
11 by adding caustic soda, and finally fed to the 4) Sufficient mechanical strength for practical
secondary purification section. use
The content of calcium and magnesium must 5) Low sensitivity to changing operating condi-
not exceed 20 ppb. Such low contents can be tions
achieved by using ion-exchange columns. The The importance of 1 – 3 is described in Sec-
polished brine is pumped at approximately tion 7.1. High mechanical strength is necessary
70 ◦ C to an ion-exchange system with two resins for installing the membrane and during service
beds operating in series according to the lead/lag life, in which the membrane has to cope with de-
principle. When the leading ion exchanger is ex- viations in temperature, concentration, and pres-
hausted it is put to the regeneration and condi- sure.
tioning mode, while the lagging one takes over As the performance of the membrane is the
the lead position. After treatment in this sec- most important element in the economy of a
ondary purification step, the purity limits are membrane cell, many refinements have been
met and the brine is fed to the membrane cells. made in membrane manufacturing. To reduce
In another secondary purification system two the current screening due to fabrics, membranes
columns operate in series while the third is in reinforced with dispersed microfibers and inter-
regeneration mode. When the first column is ex- woven fabrics made of electrolyte-soluble fibers
hausted, the regenerated column is put in second and PTFE have been developed [144, 145].
position. The improvement of hydrophilicity by cov-
Prior to the resaturation, the byproduct chlo- ering the surface on the cathode side or on both
rate and dissolved chlorine must be eliminated sides with a nonconductive inorganic material
from the anolyte. Chlorate concentration is con- brought about a significant reduction in the cell
trolled by acidification of a partial stream of voltage. The surfaces of the membrane are cov-
anolyte with an excess of hydrochloric acid ered with thin layers of a porous inorganic ma-
[143]. Chlorine is removed under vacuum fol- terial. This material is an oxide, hydroxide, or
lowed by addition of sodium bisulfite and hy- carbide of the metals of groups 4, 5 and 6 or the
droxide. iron triad (Fe, Co, Ni) [138, 146].
Figure 47 illustrates the effect of hydrophilic
NaClO3 +6 HCl → NaCl+3 Cl2 +3 H2 O
cathode surface modification. The surface-
modified membrane (Type B) has a lower cell
voltage than the conventional membrane (Type
2 Cl2 +2 NaHSO3 +6NaOH → 4NaCl+4 H2 O+2 Na2 SO4 A). Tthe voltage of the surface-modified mem-
brane decreases linearly with decreasing gap
size. With these advanced membranes, so-called
All other impurities not precipitated, filtered zero-gap cells have been made possible, and the
out, or extracted by the ion exchangers can only ohmic loss in electrolytes has been reduced to a
be controlled by purging a partial stream of minimum.
Chlorine 51

The active life of a membrane is determined where U is the cell voltage (V), F the Faraday
by the economic balance between membrane constant for NaOH (1.4923 kg/kAh), and CE the
cost and energy cost in use [148]. NaOH current efficiency (%).
The specific power consumption is the main
indicator for economic plant operation, and con-
tinuous efforts are made to lower the voltage
and increase the current efficiency. At a thermo-
dynamic minimum the decomposition voltage
of about 2.2 V limits the theoretical minimum
energy requirement to about 1480 kWh/t 100
% NaOH, as shown in Figure 48. At practical
current densities of 3.0 – 5.5 kA/m2 for present-
day commercial cells and membranes, power
consumption measured at the electrolyzer ter-
minals is in the range of 1950 to 2180 kWh/t
100 % NaOH dependent on the selected cur-
rent density (anolyte/catholyte temperature 90

C, NaOH concentration 32 wt %, NaCl concen-
tration in the anolyte 220 g/L). The power con-
sumption rises with increasing operating time
due to aging effects, such as decreasing cur-
rent efficiency and increasing voltage. Invest-
Figure 47. Effect of cathodic surface modification ment costs rise when operating at low current
densities, as more cells are needed to meet pro-
duction. Hence electrolyzers are operated at low
The performance of membranes depends on current densities in countries with high energy
the operating conditions, especially on the caus- prices, and at high current densities in countries
tic strength of the solution (Figure 46). Commer- with low energy prices.
cially available membranes, delivered by Asahi
Chemical, Asahi Glass and Du Pont, are desig-
nated for use in a specific strength of caustic. For 7.2.4. Product Quality
economic production the selection of the appro-
priate membrane is essential. Table 15 gives an The caustic soda solution has a concentration of
overview of the most widely used membranes. up to 32 ± 1 wt % NaOH. If a NaOH concentra-
Most membranes are operated in the narrow- or tion of 50 wt % is required, evaporation can be
zero-gap configuration to minimize power con- used. The typical NaCl content is 20 ppm in a
sumption. 32 wt % caustic solution.
The hydrogen has almost synthesis quality
with a concentration of about 99.9 vol % H2 (dry
basis). The chlorine has an oxygen content of
7.2.3. Power Consumption about 1.5 vol % (dry basis). Chlorine with an
oxygen content below 0.6 vol % (dry basis) can
For monitoring cell performance and comparing be obtained by acidifying the brine with hy-
different electrolyzer designs, the electric power drochloric acid.
required to produce one tonne of NaOH 100 %
is considered. This figure is determined by the
voltage drop over one cell and the NaOH current 7.3. Membrane Cells
efficiency.
7.3.1. Monopolar and Bipolar Designs
DC Energy
= UMIt = U

100% NaOH produced  F ·CE
= t kWh
A commercial membrane plant has multiple cell
elements combined into a single unit, called the
52 Chlorine

Table 15. Commercial membranes

Available Tear strength, Tensile strength, Ohmic drop Caustic strength,


since (at 3 kA/m2 , T = 90 ◦ C,
c = 32 wt %),
kg kg/cm V wt %
a
Asahi Chem F5201 1991 4 5.5 33 – 36
a
Aciplex F4202 1993 4 5.5 30 – 34
a
F4203 1997 4 5.5 30 – 34
F890 1989 4.5 6 0.35 31.5 – 32.5
Asahi Glass F892/old 1990 4.5 5 0.27 30 – 35
Flemion F892/new 1994 4.5 5 0.28 31.5 – 33.5
F893/new 1994 4.5 5 0.26 31.0 – 32.5
N90209 1984 2.5 5.6 0.35 30 – 35
a
DuPont N966 1988 5.5 7.3 30 – 35
a
Nafion N981 1996 3.3 30 – 35
a
Not published.

Figure 48. Specific power consumption

electrolyzer. The electrolyzers follow two basic copper conductors to lower the resistance. Be-
designs: monopolar and bipolar [147]. cause of this basic principle, ohmic losses in the
In a bipolar arrangement the elements are monopolar cells are 80 – 100 kWh per tonne 100
connected in series with resultant low current % NaOH, which is much higher than in equiva-
and high voltage. The cathode of a cell is con- lent bipolar cells. Furthermore, the bipolar safety
nected directly to the anode of the adjacent cell, system is not applicable to the monopolar de-
as shown in Figure 49. The operation of a bipo- sign, since the cell elements are arranged in par-
lar electrolyzer can be easily monitored by mea- allel, which does not permit the monitoring of
surement of element voltages. If element up- deviations in individual cell voltages.
sets occur, a safety interlock system actuates
the breakers (short-circuiting switches) and iso-
lates the electrolyzer from the electric circuit. As
the influx and efflux of electrolytes for the cells
with different electric potential are gathered in
common headers, problems of stray current may
arise.
In the monopolar type all anodes and cath-
odes are connected in parallel, forming an elec-
trolyzer with high current and low voltage (Fig-
ure 50). Due to the long current path, the voltage
drop is high and can only be reduced by mini-
mizing the size of cells or introducing internal
Figure 49. Bipolar electrolyzer
Chlorine 53

ranges from 0.2 to 5.0 m2 . Current density varies


between 1.5 and 7 kA/m2 .
The cells are filled with electrolytes, and gas-
separating means are provided outside the cells.
Many cells generally stacked like a filter press,
constitute one electrolyzer with high production
capacity.
The performance of a plant is determined by
the electrolyzer, the cell voltages, and the cur-
rent efficiency of the membrane. It is essential
to design an electrolyzer with an homogeneous
electrolyte concentration, temperature, and cur-
rent density distribution across the whole area
of the membrane.
Figure 50. Monopolar electrolyzer The construction materials of the cell are se-
lected to withstand the corrosive electrolytes. In
Multiple electrolyzers are employed in a sin- most electrolyzers, titanium and nickel are used
gle d.c. circuit (Fig. 51). Usually bipolar elec- for the anode and cathode compartments of the
trolyzers are connected in parallel with low cur- cell. In older electrolyzers, stainless steel is used
rent and high voltage. Monopolar electrolyzers on the cathode side.
are often connected in series, resulting in a high For economic and environmental reasons,
current circuit and low voltage. Though both mercury and diaphragm plants are increasingly
principles still appear on the market, investment being converted to membrane electrolyzers. The
and operating cost considerations, such as for the existing facilities, such as rectifiers, equipment
rectifier system, the cell room space required, for brine purification, and equipment for product
for piping, valves, instrumentation, busbars and treatment are utilized as much as possible.
switches, significantly favor the bipolar design.
Asahi Kasei ACILYZER-ML/NC Elec-
trolyzer. The Asahi bipolar electrolyzer
(Fig. 52) is of the filter-press type. The bipo-
lar cell frames are suspended in a steel frame
and compressed by a hydraulic device. Each
cell frame consists of an anode and cathode
compartment separated by a partition wall. The
anode compartment is made of titanium, and
the cathode compartment consists of special
stainless steel and nickel. The anode and cath-
ode structures are spot welded onto ribs in each
compartment. Each compartment has an inlet
nozzle for electrolytes at the bottom and an
outlet nozzle for gas and electrolyte on top,
connected to the gas/liquid separation chamber.
Two types of cell frames are available: frames
Figure 51. Electrolyzer architecture with forced circulation of electrolytes by pumps,
and frames with natural circulation in each com-
partment by means of a special arrangement of
integrated ducts. The current is connected to the
7.3.2. Commercial Electrolyzers first and last element by flexible busbars.

Generally, membranes are clamped vertically CEC BITAC 800 Electrolyzer. The Chem-
between the meshlike metal anodes and cath- ical Engineers Corporation (CEC) bipolar
odes. The effective membrane area of a cell BITAC electrolyzer was jointly developed with
54 Chlorine

Figure 52. Cell structure of ACILYZER ML32NC


a) Gasket; b) Nickel; c) Cathode; d) Anode; e) Titanium; f) Partition wall; g) Membrane; h) Rib; i) Reinforcing rib; j) Duct;
k) Gas – liquid separation chamber

Tosoh Corporation. The design follows the filter- are cladded with titanium and stainless steel.
press principle. Up to 80 bipolar electrode Inlet and outlet tubes for liquids and gases and
frames are clamped together by end plates and are made of transparent PTFE.
spring-loaded tie rods. The frames are made of
special titanium alloy for the anode and nickel Uhde BM 2.7 Electrolyzer. The bipolar
for the cathode. The electric current flows along Uhde electrolyzer (Fig. 53) is a single-element
the nickel pans, since the electrical conductivity concept. Each element comprises anode and
of nickel is six times higher than that of titanium. cathode half-shells, electrodes, a membrane,
Gas and electrolytes leave the cell compartment flanges, and the sealing system. This enables
in overflow mode with little pressure fluctuation. long-term storage of pre-assembled and fully
Transparent PTFE tubes are attached at the elec- tested elements. The electrodes are attached
trolyte inlet and outlet nozzles of each element. with continuous laser weld to the current trans-
Anolyte recirculation takes place through an ex- fer and support blades and hence to the half-
ternal loop. shells. The anode is made of titanium and the
cathode of nickel. The individual cell elements
CEC CME DCM 400 Electrolyzer. The of an electrolyzer are suspended in a steel frame
CME monopolar electrolyzer consists of large in which they are lightly pressed together for
elements compressed in a filter-press arrange- electrical contact. Large sealing forces are not
ment. The electric current travels into each an- required in the single-element concept, as each
ode element through conductor rods and current element is a separate, stand-alone electrolysis
distributors. This design achieves uniform cur- cell. The feed and discharge lines of the cell are
rent distribution over the large electrode area. located underneath the cells and connected to the
The current distributors serve the additional catholyte and anolyte headers. The area above
role of a downcomer pipe, which creates a nat- the electrolyzer is free of piping or bus bars,
ural circulation within the cell, providing a uni- simplifying access and eliminating the risk of
formly distributed electrolyte concentration as leakage and associated corrosion problems. The
well as good gas release. The anode frames current is conducted from cell to cell by continu-
are constructed from titanium and the cathode ously laser-welded, explosion-bonded titanium
frames from a special stainless steel. The rods – nickel contact strips on the anode half-shell.
Chlorine 55

Figure 53. Uhde BM 2.7 electrolyzer


a) Single element; b) Contact strip; c) Cell rack; d) Busbars; e) Inlet hoses; f) Outlet hoses; g) Header

The brine and caustic soda feeds enter the cell lar), omitting the copper distributors and simply
at the bottom, and the product streams are dis- pressing the nickel cathode pan onto the nickel-
charged downwards through internal overflow plated back of the anode pan. The integral feed
pipes. The internal baffle plate at the top of the and discharge manifolds are designed to avoid
anode half-shell prevents gas-phase blistering current leakage.
of the membrane. The chlorine gas is effectively
removed from the membrane, preventing con- EL-Tech ExLM Monopolar Electrolyzer.
tact and improving the inherent safety of the The ExLM monopolar electrolyzer is an im-
electrolyzer. Natural circulation around a down- proved version of the MGC electrolyzer, which
comer plate and a distribution pipe for brine and has been in service for more than 15 years.
caustic achieve homogeneous temperature and The elements are sealed with O-rings in a stag-
concentration profiles within the element and gered gasket design. The cathode O-ring is lo-
assist in achieving uniform current distribution. cated closer to the liquid than the anode O-ring.
This protects the anode O-ring from the chlo-
EL-Tech ExLB Bipolar Electrolyzer. The rination degradation, making it a long-life back
ExLB bipolar electrolyzer (Fig. 54) is basically up seal. The elements are pressed together by
of the same design as the EL-Tech ExLM elec- tie rods with copper distributor plates and con-
trolyzer. Instead of the copper distributors with ductive interface material to provide good cur-
interface material, which provide the parallel rent distribution. Electrolytes and gases are fed
(monopolar) arrangement in the electric cir- and discharged to and from the elements through
cuit, the elements are connected in series (bipo- the manifold passage attached to the cell ele-
56 Chlorine

Figure 54. EL-Tech ExLB bipolar electrolyzer

ments. Increased internal electrolyte circulation INEUS FM21-SP Electrolyzer. The FM21-
is achieved by an improved electrode design. SP (Fig. 55) is a monopolar electrolyzer incorpo-
rating a simple pressed electrode structure. The
EL-Tech ExLDP Dense Pak Unit Elec- anode assembly is composed of a 2 mm thick
trolyzer. The EL-Tech ExLDP dense pak elec- titanium panel between compression molded
trolyzer comprises monopolar sections in one joints of a special cross-linked EPDM elastomer.
electrolyzer filter press compression set us- The cathode assembly is composed of a 2 mm
ing standard monopolar cell components. Each thick nickel panel between compression molded
monopolar cell section is separated by an insu- joints, also of EPDM.
lating Inter Pak Spacer. Mostly three monopolar The anodes and cathodes are assembled be-
electrolyzer sections are included, with 2 to 10 tween 2 end plates until the number of elec-
elements per section. The dense pak can be con- trodes required for the desired electrolyzer ca-
figurated to match special rectifier/transformer pacity is reached, up to 60 anodes in the FM21-
configurations of existing plants, making it suit- SP and up to 90 anodes in the larger FM1500.
able for mercury and diaphragm cell conver- A key feature of both designs is the elimination
sion projects. For new plants, the advantage of of any external piping to individual cell com-
the ExLDP electrolyzer is reduced current, in- partments by the use of a simple but effective
creased voltage circuits compared to an equiva- internal header/manifold arrangement.
lent monopolar cell unit.
Chlorine 57

Figure 55. INEUS FM21-SP elelctrolyzer


a) Tie-rod; b) Floating end plate; c) Copper electrical connections; d) Ion exchange membrane; e) Fixed end plate; f) Anode
electrode assembly (titanium panel between compression molded gaskets; g) Cathode electrode assembly (nickel panel be-
tween compression molded gaskets); h) Support rail

The electrolyzer has coated titanium anodes. The media are fed to and discharged from the
The cathodes are pure nickel, also available with electrolyzers by a header system arranged along
a coating to lower the hydrogen overpotential if the walls of the cell room on one side.
necessary. Both electrodes are pressed from inte- From the other side power is supplied either
gral sheets of pure metal, and this makes recoat- from separate transformer/rectifier units for each
ing of the electrodes extremly simple and cost electrolyzer or from one unit for two or more (up
effective. Hence recoated structures can be sent to six) electrolyzers in parallel. The switches are
to site prior to electrolyzer refurbishment from arranged close to the rectifiers. They are actuated
a pool of electrodes available to all customers. automatically and connected to the common in-
Effective electrode area is 2 × 0.21 m2 per terlock system for safety reasons.
electrode, which gives a very compact elec- In the middle space remains available for the
trolyzer. The individual electrodes are readily electrolyzers and their individual feed and dis-
handled without the need for lifting apparatus, charge piping. Only a light crane is required to
which allows the electrolyzer to be rebuilt and handle single electrode frames or elements. Thus
refurbished in the minimum of time. only a light structure for the entire cell house is
used.

7.3.3. Comparison of Electrolyzers


8. Electrodes
Operating parameters of bipolar electrolyzers
are compared in Table 16, and those of monopo- 8.1. Anodes
lar electrolyzers, in Table 17.
The initial anodes used for the electrolytic gen-
eration of chlorine were made of platinum or
magnetite. However, as the plant grew in the size,
7.3.4. Cell Room
the cost of platinum and limitations of the current
density for magnetite led to the wide-scale intro-
Typical bipolar membrane cell rooms are shown
duction of graphite anodes, which were used ex-
in the following Figures 56 and 57.
clusively up to 1970. The graphite of choice was
58 Chlorine
Table 16. Bipolar electrolyzers

Company ASAHI KASEI CEC KRUPP UHDE EL-TECH


Cell ML 32 ML 60 BITAC 800 BM 2.7 ExLB

Effective membrane 2.72 5.05 3.276 2.72 1.5


area, m2
Max. no. of elements 150 150 80 160 80
Current density, up to 6.0 up to 6.0 1.5 – 6.0 1.5 – 6.0 1.5 – 7.0
kA/m2
Max. capacity of 45 90 54 90 29
electrolyzer t/d NaOH
100 %
d.c. power 2100 2100 2150 2130 2100
consumption kWh/t
NaOH
(at current density) (4.0) (4.0) (5.0) (5.0) (5.0)

Table 17. Monopolar electrolyzers

Company CEC INEUS EL-TECH


Cell CME DCM 400 FM 21-SP ExLM
2
Effective membrane area, m 3.03 0.21 1.5
Max. no. of elements 32 120 30
Current density, kA/m2 1.5 – 4.0 1.5 – 4.0 1.5 – 6.0
Max. capacity of electrolyzer, t/d 13 7 9
NaOH 100 %
d.c. power consumption kWh/t 2150 2140 2150
NaOH
(at current density) (3.5) (4.0) (5.0)

Figure 56. Bipolar cell room by Ashai Glass


Chlorine 59

Figure 57. Bipolar cell room by Krupp Uhde

low in ash and vanadium and composed of var- oxide coating containing one or more of the plat-
ious types of particulate coke and pitch binder. inum metal group oxides. The second patent de-
Following extrusion, baking at ca. 1000 ◦ C, and scribed coatings in which mixed metal oxide
graphitization at 2600 – 2800 ◦ C, the final shape crystals contained a non-platinum metal oxide
of the electrode was achieved by machining. The in addition to the platinum metal oxide (includ-
shape of the horizontally suspended anodes with ing Ti, Ta, and Zr oxides).
an initial thickness of 7 – 12 cm for the amalgam Further improvements in the coating and the
process was similar to that of modern titanium anode structures followed rapidly along with the
anodes due to the retrofitting of existing cells. commercialization of anodes by De Nora [150]
The anodes had vertical slits and holes to allow under the trade name Dimensionally Stable An-
the removal of the gaseous chlorine. Due to the ode (DSA). Because of the dimensional stability
cogeneration of oxygen and the resulting forma- and the lifetime of the coating and the ability to
tion of CO and CO2 , electrode wear was high, increase the current densities, rapid introduction
in the range of 1.8 – 2.0 kg graphite per tonne of the activated titanium anodes was possible. At
of chlorine from NaCl and 3 – 4 kg per tonne present only a few plants still use graphite an-
from KCl. Even with a daily adjustment of the odes, largely due to the initial investment costs
anodes to compensate for the changes in dimen- for titanium anodes.
sion a k value of only 0.12 to 0.14 Vm2 kA−1
was achievable.
The initial attemps to replace the graphite 8.1.1. General Properties of the Anodes
anodes with activated titanium anodes began
as early as 1957 with platinized titanium and Coating Properties and Preparation.
Pt/lr-coated anodes. However because of the Comprehensive reviews on preparation and
short lifetimes of the anodes, they were not eco- properties are given in [151, 152].
nomic. The use of mixed metal oxides was first
patented by Beer in 1965 and 1967 [149]. The Chemical Composition. Because of its
initial patent described a coated metal electrode price and performance, Ru is the basic com-
in which the active material was a mixed metal ponent in all commercial coatings at present,
60 Chlorine

along with an oxide of a non-platinum metal electrochemically vary ca. 400 to 1000 times the
(e.g., Ti, Sn, or Zr). In most cases a second geometric surface area [157].
platinum metal oxide is added to increase the
performance of the anode coating. There is an Overpotential and Current – Voltage Rela-
optimum ratio of platinum metal oxides to non- tionship. The observed overpotential for chlo-
platinum metal oxides in terms of overpotential, rine evolution at 2 – 10 kA/m2 is in the range of
wear rate, and costs. The optimum depends 80 – 110 mV [158 – 161], about 70 – 100 mV of
on the operating conditions and the method of which is due to diffusion overpotential effects
preparation of the coating and normally lies in [161]. The overpotential for the generation of
the range of 20 : 80 to 55 : 45 by weight. Some of oxygen under similar pH and temperature con-
these coatings may contain glassy fibers [153] ditions lies is ca. 300 mV more anodic than that
and some contain pre-oxidized material such as of chlorine generation. Other than oxygen evo-
Li0.5 Pt3 O4 [154]. lution, the only other side reaction is formation
of chlorate.
Preparation. The solvent used for the prepa-
ration of the precursors solutions are chosen on Coating Wear and Coating Lifetime. The
the basis of the desired electrochemical prop- coating lifetime is strongly dependent not only
erties and the method of application, which is on the type of cell – membrane, diaphragm, or
mainly determined by the anode structure. Most mercury – but also on a range of process parame-
coating solutions are prepared by dissolving ters, including brine quality, current density, and
salts or organometallic complexes in aqueous, membrane or diaphragm quality. The upper limit
organic, or mixed solvents. The coating can be of the wear rate would seem to be in the region
applied by spraying, brushing, dipping, or other of 500 t Cl2 /m2 anodic area for a standard com-
techniques. Following evaporation of the bulk of mercial loading.
the solvent, the anode is heated to 350 – 600 ◦ C The wear rate mechanism is discussed in de-
to form the oxidic coating then cooled prior to tail in [162, 163]. The effects of various impu-
the next coating cycle. This is repeated until the rities and materials in the brine can be divided
desired coating thickness is applied. Optional into three types.
post-thermal treatment can also be carried out.
The optimal performance of the coating de- 1) Compounds or ions which attack the sub-
pends on the above parameters and on the coat- strate, e.g., fluoride or organic acids such as
ing thickness per coating cycle, which must be formic or oxalic acid.
optimized for each coating and surface pretreat- 2) Materials which built up blocking layers on
ment step. the surface of the anode, e.g., hydraulic oil
or polymer films resulting from delamination
Crystallographic Composition, Morphol- of membranes. The irreversible poisoning of
ogy, and Real Surface Area. A rutile phase is coating is caused by ultrathin aluminum sili-
the electrochemically active phase of the coat- cate layers.
ing, and although it is thermodynamically unsta- 3) Electrochemically active film-forming mate-
ble, it remains even after many years of opera- rials such as MnO2 , which may lead to an
tion. The stable phase – anatase TiO2 in the case increased oxygen content in the chlorine.
of TiO2 – RuO2 coatings – is electrochemically Other examples, such as the insensitivity of
inactive [155]. The degree of crystallinity and the performance of diaphragm anodes to almost
the composition are related to the processing pa- complete surface coverage by iron oxides illus-
rameters [156] and the various degrees of mixed trate the robustness of commercial coatings.
crystals exhibit different stabilities. The real sur-
face area of the coating is a function of both
the titanium pretreatment and the coating com- 8.1.2. Anodes for Mercury Cells
position. The surface of the chlorine-generating
coating is often described as “cracked mud” due Structure. The classical structure of anodes
to its resemblance to a dry river bed. The BET for this process still reflects the retrofitting con-
surface area of the coatings or that determined cept used during the 1970s and the high current
Chlorine 61

operations at ca. 10 kA/m2 . A typical mercury value < 0.5 mm is achieved, mostly by manual
cell anode consists of a number of copper shafts, straightening after manufacturing or recoating.
protected by either a permanently welded or re-
movable titanium outer sleave, from which the Coating Life. The coating life is determined
current is distributed to the active surface over by a wide range of practical aspects and nor-
distributor bars (Fig. 58). mally not directly related to the electrochemical
wear rate of the coating. These include:
– Mechanical damage to the anode caused by
short circuiting [165, 166]
– The need to maintain a recoating schedule,
due to production demands and the labor in-
tensive refitting of a cell.
– Synchronization of recoating with the ex-
change of other consumable parts of the cells
such as covers and gaskets.
The tendency has been to increase the lifetime
from about 180 t Cl2 /m2 to 300 – 400 t Cl2 /m2 .
This has been achieved by the introduction of
Figure 58. Four-stem anode for amalgam cells
a) Active surface; b) Current distributor; c) Riser tube to
better control systems in the cells and the de-
protect the copper bar inside velopment of intermediate layers of plasma-
sprayed conductive TiO2−x between the active
The quick release of gas and the supply of coating and the titanium substrate [159].
fresh brine to the active surface are the major
requirements of an mercury cell anode, and a
wide range of designs have been built. The most 8.1.3. Anodes for Diaphragm Cells
common types are shown in Figure 59. The dif-
ferences are more evident at current densities The predominate determinants in the design of
> 7 kA/m2 . The use of baffles on the back of diaphragm anodes are:
the active surface to enhance the gas lift and aid – The relatively low current density of ca.
the supply of brine to the active surface is also 2.0 kA/m2
common [164]. – The minimization of the anode – diaphragm
Flatness is critical for the optimal perfor- gap
mance of the anodes in the cells. A typical – The need to remove and replace the anode
array from the cathode, hence the use of re-
tractible anodes

Figure 59. Anode designs for quick gas release


A) Flatt profile (channel blades); B) Rod type (3-, 4-, 5-mm diameter); C) 3D side profile anode
62 Chlorine

– The limiting height of the cathode, integrated


with the diaphragm manufacturing technol-
ogy
The conventional anodes shown in Figure 60
[167] have been further developed to optimize
the energy consumption of the cells by replac-
ing the simple flat expanded metal with complex
structures [169]. At present very few plants are
still operating without expandable anodes.
Figure 61. Empirical fit of observed nonlinear wear rate of
coating thickness versus years on line.
L(t) is the loading at time t, L the initial loading, r the wear
rate, t time, A(t) the active surface area at time t, q an em-
pirical factor related to the current density sensitivity of the
wear rate r

Coating Life. At present the second-gen-


eration coatings for membrane cells are showing
lifetimes comparable to those of the diaphragm
process. The actual lifetime of the anodes is de-
pendent on the extent of damage by caustic flow
Figure 60. Anode for monopolar diaphragm cells through holes in the membrane or by contami-
a) Activated (coated) expanded metal; b) Expanding spring; nation with the poisons.
c) Titanium-clad copper bar; d) Copper thread to fix the an-
ode to the cell base
Oxygen Content [151]. The oxygen content
has also been improved by the optimization of
Another type of diaphragm anode is used in
internal circulation of the brine within the cells.
the bipolar Ganor cells [168].

Coating Life and Mechanism of Deactiva-


tion. The coating lifetime of DSA coatings ex- 8.2. Activated Cathode Coatings
ceeds 12 years, and production of chlorine ex-
ceeds 240 t Cl2 /m2 . The wear is caused by the Since 1910 diaphragm brine electrolyzers have
relatively high oxygen content in a diaphragm used carbon steel cathodes and continue to use
cell [170] of ca 1 – 2 %. The wear rate is non- carbon steel to this day. When the first ion-
linear (Fig. 61). This nonlinearity is critical for exchange membrane electrolyzers were intro-
determining the correct time to begin recoating duced in the late 1970s, the cathodes were also
so as to avoid unplanned stoppages. carbon steel. By the early 1980s the design
had evolved to stainless steel and nickel cath-
odes, and finally in the 1990s to exclusively
8.1.4. Anodes for Membrane Cells nickel cathodes. Depending on current density,
the hydrogen overpotential of carbon steel cath-
Structure. The variety of designs of mem- odes is about 300 mV. Active cathode coatings
brane cells has led to a range of anodes active can lower the overpotential by 200 – 280 mV,
area structures; the common principles are the thus providing significant energy savings. Ac-
need to support the membrane and gas release tive coatings have often been described in the
to the back of the anode surface. Therefore, thin literature and used in water electrolysis for over
flattened expanded, perforated metals or louver 40 years. With the development and evolution of
type structures with and without perforations are the ion-exchange membrane technology, active
used [171]. cathode coatings are coming into general use.
Chlorine 63

The patent literature covers many different pensive coating. Most diaphragm electrolyzers
types of coatings, and new ones are being pub- use heat-cured polmer – asbestos separators (di-
lished regularly. The two basic approaches to ac- aphragms) that are vacuum deposited after the
tivation are high-surface area coatings and cat- cathode coating is applied. This curing opera-
alytic coatings. Both bare nickel and carbon steel tion can destroy the activity of certain coatings.
show lower hydrogen overpotential once in op- All cathode coatings are susceptible to poi-
eration and their surfaces roughen. In fact by soning by impurities that make their way into the
grit blasting bare nickel cathodes and roughen- catholyte with the deionized water or are compo-
ing the surface, the long-term overpotential can nents of the piping, electrolyzer etc. These im-
be reduced by 30 – 40 mV. More common are purities tend to blind the activity of the coatings
porous nickel-type coatings that offer high sur- over a period of time that depends on their con-
face area and good chemical resistance. These centration. Porous nickel coatings in diaphragm
coatings consist of two or more components. electrolyzers are less susceptible to blinding by
At least one of the components is leached out impurities because spalling of the brittle coating
in caustic to leave the porous high surface area makes the coatings self-cleaning. The platinum
nickel [172]. These coatings are typicall nickel- group metal coatings are subject to damage from
zinc [173], nickel – aluminium – Raney nickel reverse currents during electrolzer outages. Pre-
[174], nickel – aluminium [175], or nickel – sul- cautions are needed to protect the coatings with
fur [176]. A variety of additives are recom- reducing agents [183] or by cathodic protection
mended for strength, life, and resistance to poi- [184].
soning by impurities. Rough coatings of nickel – Active cathode coating have become the stan-
nickel oxide mixtures [177] and nickel with em- dard throughout the chlor-alkali industry for new
bedded activating elements such as ruthenium construction with the ion-exchange membrane
[178] are also used. Sintered nickel coatings electrolyzer technology. In most of the older di-
are described in patents [179] as well as being aphragm electrolyzer plants, problems with ap-
available from Huntington Alloys. Nickel coat- plication of the cathode coatings and generally
ings containing platinum group metals, primar- lower power costs have obviated the use of active
ily platinum and/or ruthenium, have been sold cathode coatings. While there are more recent
by Dow [180], Johnson Matthey, and ICI [181]. developments applicable to the diaphragm tech-
The coatings used for diaphragm and mem- nology in the way of active cathode coatings,
brane electrolyzers differ because of the differ- many of these developments remain the propri-
ent substrates (carbon steel and nickel, respec- etary information of the technology and coating
tively) and the different operating conditions. suppliers.
The weak 11 % caustic in diaphragm cell liquor
is less corrosive than the strong 33 % caustic
of a membrane electrolyzer. The less expensive 9. Comparison of the Processes
and more fragile coatings like nickel – zinc can
be used in diaphragm electrolyzers. Membrane The advantages and disadvantages of the three
electrolyzer suppliers favor the platinum group chlor-alkali processes are summarized in Table
metal coatings. 18. The three chlor-alkali processes can be com-
The shape of the cathode structure is an im- pared in respect to the quality of the chlorine and
portant factor affecting the choice of cathode caustic produced, and the equipment and oper-
coating. The complex cathodes of diaphragm ating costs.
electrolyzers lend themselves to liquid sys- Today the membrane process is the state of
tems (e.g., electroplating or electroless baths) the art for producing chlorine and sodium hy-
that can coat the entire structure by immersion droxide or potassium hydroxide. All new plants
[182]. Membrane electrolyzers, which are pri- are using this technology. The production ca-
marily of a filter-press design, have flat cath- pacity of chlor-alkali plants using the membrane
odes that are easy to coat by spraying or paint- process reached about 21 % of total world pro-
ing. The lower operating current density of di- duction capacity in 1995 and is predicted to in-
aphragm electrolyzers means more cathode area crease to about 28 % by 2001 (Table 19) [132].
per unit of production; this requires a less ex-
64 Chlorine

Table 18. Advantages and disadvantages of the three chlor-alkali processes

Process Advantages Disadvantages

Diaphragm process use of well brine, low electrical energy consumption use of asbestos, high steam consumption for caustic
concentration in expensive multistage evaporators, low purity
caustic, low chlorine quality, cell sensitivity to pressure variations

Mercury process 50 % caustic direct from cell, high purity chlorine anduse of mercury, use of solid salt, expensive cell operation, costly
hydrogen, simple brine purification environmental protection, large floor space

Membrane process low total energy consumption, low capital investment,use of solid salt, high purity brine, high oxygen content in
inexpensive cell operation, high-purity caustic, chlorine, high cost of membranes
insensitivity to cell load variations and shutdowns,
further improvements expected

The diaphragm cell capacity remains constant caustic, is required for 20 – 30 % of the demand
and there is a decline in mercury cell capacity. in industrialized countries. Even the use of pu-
rification processes (see page 24) does not re-
Table 19. World chlorine market 1995 and 2001 (in %)
duce the NaCl content below 0.03 wt %. In ad-
1995 2001 dition to the NaCl and NaClO3 , the levels of Si,
Diaphragm 52 49 Ca, Mg, and sulfate impurities are higher than
Mercury cell 22 18 for the mercury and membrane processes.
Membrane 21 28
Others 5 5
Market, 106 t/a 41 43.3
9.2. Economics

The conditions for a conversion from the mer- The wide variation in the main cost factor, that
cury and the diaphragm process to the membrane for electrical energy, which varies from region
process are discussed below. to region by a factor of up to three, makes a
direct comparison of production costs problem-
atic. Further, the cost of electrical energy is in-
creasing in different regions at drastically dif-
9.1. Product Quality ferent rates, depending on the basic source of
energy and customs. Rapidly changing foreign
Table 20 shows typical composition values for
exchange rates also make international compar-
the chlorine and caustic produced by the di-
isons difficult.
aphragm, mercury, and membrane processes.
A detailed discussion of the capital invest-
Chlorine produced by the mercury process can
ment and operating costs for the three processes
be used directly for most uses. Chlorine pro-
for a 200 000 t/a-plant in 1991 is given in [185].
duced by the diaphragm or membrane process
A comparison of the investment costs does not
contains up to 2 % O2 , depending on the pH of
make sense today for the mercury process, be-
the anolyte. This oxygen can be removed by con-
cause no mercury cell plant and only a few di-
densation and evaporation of the chlorine.
aphragm cell plants were built since then. All
The sodium hydroxide solution from the mer-
new plants are using the membrane process.
cury process is the purest of the three; the
amounts of NaCl and NaClO3 are especially low.
However, the quality of caustic from the mem- 9.2.1. Equipment
brane process is almost as good. A main draw-
back of the diaphragm process is the high con- The expenses for the rectifier, chlorine and hy-
centration of NaCl and NaClO3 in the caustic drogen systems, HCl system, caustic storage,
solution. This sodium hydroxide solution can- utilities, and engineering and construction over-
not be used for some processes. A chloride- heads are approximately the same for the three
free grade, commonly referred to as rayon-grade processes.
Chlorine 65
Table 20. Product qualities: typical compositions of chlorine, caustic, and hydrogen

Product and contents Process


Diaphragm Mercury Membrane
Unpurified Purified

Chlorine gas (from cells),


vol%
Cl2 96.5 – 98 98 – 99 97 – 99.5
O2 0.5 – 2.0 0.1 – 0.3 0.5 – 2.0
CO2 0.1 – 0.3 0.2 – 0.5
H2 0.1 – 0.5 0.1 – 0.5 0.03 – 0.3
N2 1.0 – 3.0 0.2 – 0.5

NaOH solution, wt %
NaOH 50.0 50.0 50.0 50.0
NaCl 1.0 0.025 0.005 0.005
Na2 CO3 0.1 0.1 0.05 0.04
Na2 SO4 0.01 0.01 0.0005 0.0001
NaClO3 0.1 0.001 0.0005 0.001
SiO2 0.02 0.02 <0.001 0.002
CaO 0.001 0.001 0.001 0.0001
MgO 0.0015 0.0015 0.0002 0.0001
Al2 O3 0.0005 0.0005 0.0005 0.0001
Fe 0.0007 0.0007 0.0005 0.0004
Ni
Cu 0.0002 0.0002 0.00001 0.0001
Mn
Hg none* none* 0.00001 none*
NH3 0.001

Hydrogen gas, vol%


H2 >99.9 >99.9** >99.9
−6
* < 10 %.
** Hydrogen gas from the mercury process contains mercury: 1 µg/m3 – 10 mg/m3 , depending on the purification process. The hydrogen
gas from the other two processes is free of mercury.

Cells. The complex mercury cells are con- for either process is approximately the same, 4
siderably more expensive than the simpler di- – 7 % of the total.
aphragm and membrane cells. There is no de- Caustic Concentration. The elaborate multi-
velopment in mercury cell technology. Improve- stage evaporators required for the concentration
ments are being made in diaphragm cells (higher of the diaphragm-cell caustic and the separa-
current densities, longer service times), but the tion of NaCl and Na2 SO4 must be nickel plated
relative advantage of the membrane cells is ris- because of the corrosiveness of the cell liquor
ing fast with considerable increase in current containing NaCl and NaClO3 . These evapora-
density and improved membrane performance: tors cost 20 – 35 % of the total. The evaporators
Fewer cells are needed for a given production for the membrane process may be constructed of
capacity. stainless steel and are much smaller because the
Brine System. The brine system for the di- essentially salt-free cell liquor is more concen-
aphragm process is the simplest of the three – trated, costing 3 – 4 % of the total. The mercury
there is neither sulfate precipitation nor dechlori- process produces 50 % caustic directly, evapo-
nation – and makes up only 3 – 4 % of the capital ration is not required.
investment. The brine system is the most com- Facilities for Handling Salt. The mercury
plex for the membrane process, for fine purifi- and membrane plants require storage and han-
cation by ion exchange is necessary. However, dling facilities for solid salt. If a diaphragm plant
the two- or three-fold greater depletion of the uses well brine, only small facility is needed for
brine in the membrane process allows the brine the recycling of the salt from the caustic evapo-
system to be smaller than that for the mercury ration.
process. Therefore, the cost of the brine system
66 Chlorine

Mercury. In addition to the capital cost of The price of electrical energy varies widely
mercury itself, there is the expense of the equip- from region to region. The relatively broad range
ment to prevent emission of mercury into the of possible current densities combined with the
environment and to remove mercury from the steep increase in the cell voltage with current
products (see page 25). This equipment costs density for the diaphragm and membrane cells
10 – 15 % of the total capital investment. allows optimization of the current density with
The investment cost of a new (green-fields) respect to the local energy price. That is, if elec-
chlor-alkali project in the USA is estimated to trical energy is relatively expensive, a greater
be between 250 000 and $ 300 000 per tonne per number of cells, and thus a greater capital in-
day chlorine capacity in 1998 [186]. vestment, can be tolerated to reduce the specific
energy consumption and thus minimize total unit
production cost [187].
9.2.2. Operating Costs
The fixed costs for operators and other person- 9.2.3. Summary
nel, taxes, insurance, repairs, and maintenance
are about the same for all three processes. The 20 In spite of the advantages of the membrane tech-
% lower depreciation of the membrane process nology, about 75 % of all chlorine is produced
is offset by the additional expense for purchase in mercury and diaphragm cells, operating in ca.
and replacement of the membranes and for the 500 plants around the world. Diaphragm tech-
more elaborate brine purification. nology prevails in the United States (70 %), Rus-
Of the variable costs, the expense for salt, sia, and China, and mercury technology in West-
precipitants, and anode reactivation are roughly ern Europe (64 %). Continued production from
the same. The difference among the three pro- these plants is economical under special circum-
cesses shows up in the consumption of energy, stances.
as electricity and steam. If 1 t of steam is taken For mercury cell users, the question of today
to be equivalent to 400 kWh of electrical energy, is whether the old, depreciated plant is compet-
then the comparison in Table 21 can be made. itive with new membrane cell plants. The alter-
The differing total energy consumptions are il- natives are:
lustrated in Figure 62. – Further production in the mercury cell plant
– Conversion to the membrane process
– Phasing out the old plant
Candidates for further production are plants
of medium to large size, with low electricity
costs, with very high quality products, with
high emissions standards, with high mainte-
nance standards (low repair costs), or which
produce speciality products which cannot be
obtained in membrane cells (e.g., alkoxides or
dithionites).
All producers that do not fulfill one or more
of these conditions are candidates for conver-
sion. The more the existing infrastructure can
be used, the greater the benefits resulting from
conversion. The investment for the conversion
of a middle-sized plant (100 000 t/a) is between
$ 550 and $ 800 per tonne of chlorine capacity
per year [188]. This investment includes the cost
for the membrane cells, secondary brine purifi-
Figure 62. Relative consumption of energy (electricity
cation and additional changes of the infrastruc-
and steam) in the three chlor-alkali processes in producing ture. The costs are specific for each existing plant
50 wt % NaOH and depend on:
Chlorine 67

Table 21. Energy consumed to produce 1 t of chlorine plus 1.13 t of caustic soda (50 %) in the three chlor-alkali processes

Energy Process
Diaphragm Mercury Membrane

Electricity for 2300 – 2900 3100 – 2200 –


3400 2600
electrolysis, kWh
Steam equi- 800 – 1000 0 200 – 400
valent, kWh
Total, kWh 3100 – 3900 3100 – 2400 –
3400 3000
Relative 100 % 93 % 78 %
energy costs

– Chlorine quality (e.g., the oxygen content) a plant with a capacity of 1000 t/d to membrane
– The use of existing buildings. The materials technology is ca. 90 million dollars [186].
of the existing brine treatment area In the first few years after the introduction
– The possibility of using the dilute caustic of membrane technology, diaphragm cells in
(32 – 36 %) within the plant without concen- several plants were equipped with membranes
tration (retrofit) to reduce the cost of steam for cell
– Use of the existing electrical equipment, rec- liquor concentration, to give a small reduction
tifiers, busbars in electricity consumption and better quality
– Possible capacity enlargement because of the of caustic. This procedure is economic where
lower specific energy per tonne of chlorine steam is very expensive [187, 191].
– Dismantling and disposal of mercury-
contaminated parts of the old plant.
Normally the decision for a conversion is ini- 10. Other Production Processes
tiated by plans for an expansion of the produc-
tion capacity or by environmental legislation. 10.1. Electrolysis of Hydrochloric Acid
Each change in the plant structure or in the cost
structure may lead to reevaluation of the future of Electrolytic decomposition of aqueous hy-
the electrolysis plant. Therefore, each plant has drochloric acid is used to produce chlorine and
to be considered individually [189, 190]. For the hydrogen. The first pilot plant was set up by
European chlor-alkali industry a detailed analy- G. Messner in 1942 in Bitterfeld, Germany, and
sis of the impact of a conversion of all mercury since 1964 eight full-scale plants have been com-
cells to the membrane technology on the com- missioned in Europe and the United States, a
petitiveness of the industry is given in [110]. total capacity of 540 000 t/a [192]. Hydrogen
The situation is different for diaphragm cell chloride is a byproduct of many organic indus-
plants. These plants are still economic where trial processes. Electrolysis of hydrochloric acid
inexpensive brine (e.g., from solution mining) competes with chemical processes in which ei-
is available, energy costs are comparably low ther hydrogen chloride is used to produce chlori-
(e.g., from cogeneration of electricity and steam nated hydrocarbons directly, e.g., by oxychlori-
on site), and when the market price for caus- nation, or where chlorine is produced by chem-
tic is determined by the lower quality of di- ical reaction, e.g., in the KEL chlorine process
aphragm caustic. In countries like the United (see page 69). The advantages of the electrolytic
States it will be difficult to economically justify process are very pure products without further
conversion. In contrast to the mercury process, treatment, reliability (simple design), ease of op-
improvements to the cells are still being made, eration, flexibility (5 : 1 turndown ratio), and low
resulting in lower operating costs and savings in energy consumption even with small installa-
solid waste disposal. The investment to convert tions.
68 Chlorine

Principles. Hydrochloric acid (22 wt % HCl) in the graphite plates, which are cemented in
is fed into the cells in two separate circuits, a frames made of HCl-resistant plastics. At the
catholyte circuit and an anolyte circuit. During bottom of the frames, channels feed in the elec-
electrolysis the concentration is reduced to ca. trolyte. The gases rise up the plates and pass
17 %, and the temperature increases from 65 to through ducts into collection channels in the
80 ◦ C. A part of the depleted acid is separated upper part of the cell. Chlorine leaves the cell
from the catholyte stream, concentrated in the with the anolyte, and hydrogen leaves with the
absorption plant to ca. 30 %, and fed back into catholyte. The end plates of the electrolyzer are
the main stream. The electrolyzer is bipolar, with made of steel lined internally with rubber and
pairs of electrodes arranged like the leaves of are held together by spring-loaded tension rods.
a filter press. A diaphragm or membrane (e.g., The electric current is supplied via graphite ter-
Nafion 430) separates the anode compartment minals. The unit rests on insulated steel frames.
from the cathode compartment to prevent mix- The effective surface of the electrodes is 2.5 m2 ,
ing of the gaseous products. and the current loading can be up to 12 kA.
The reversible standard decomposition po- DeNora and General Electric are developing
tential of hydrochloric acid is 1.358 V, made up an electrolyzer with a solid polymer electrolyte
of the anode potential, the discharge of chloride (SPE) based on Nafion [195]. In addition to a
ions with formation of chlorine, and the cathode voltage savings of 20 %, it is hoped that com-
potential, the discharge of hydroxonium (H3 O+ ) pletely chloride-free hydrogen gas can be pro-
ions with formation of hydrogen. In practice duced.
(> 15 % HCl, 70 ◦ C), the decomposition poten-
tial is ≤ 1.16 V. Operation. A simplified flow diagram of the
The graphite electrode plates are not attacked process as operated by Bayer – Hoechst – Uhde
by 22 % hydrochloric acid. A poly(vinyl chlo- is shown in Figure 63.
ride) (PVC) fabric constitutes the diaphragm.
Chlorine dissolved in the anolyte diffuses
through the diaphragm and is reduced at the
cathode, causing a loss of 2 – 2.5 % of the theo-
retical current yield. The increase of cell volt-
age when current flows is mainly because of
the hydrogen overpotential at the graphite cath-
ode and the resistance of the electrolyte. De-
polarizing agents (polyvalent metal ions) in the
catholyte reduce the overpotential by ≤ 300 mV
at 4 kA/m2 [193].
The conductivity of hydrochloric acid is max-
imized at a concentration of 18.5 wt %. High
temperatures improve the conductivity, but to
avoid increased vapor pressure of HCl and ma- Figure 63. Simplified flow diagram of a hydrochloric acid
electrolysis
terial problems, the temperature is kept below a) Absorption column; b) Heat exchanger; c) Strong acid
85 ◦ C. Modern cells have a voltage of ca. 1.90 V tank; d) Catholyte collecting tank; e) Catholyte filter;
at 4.8 kA/m2 , corresponding to an energy con- f) Catholyte supply tank; g) Electrolyzer; h) Hydrogen –
sumption of 1400 – 1500 kWh per tonne of chlo- catholyte separator; i) Chlorine – anolyte separator;
k) Anolyte collecting tank; l) Anolyte filter; m) Anolyte
rine. supply tank; n) Weak acid line to absorber

Diaphragm Cells. Hydrochloric acid elec-


trolysis cells are manufactured by Hoechst – In the absorption column, the hydrogen chlo-
Uhde [194]. Each Hoechst – Uhde electrolyzer ride gas is absorbed adiabatically by depleted
consists of 30 – 36 individual cells that are hydrochloric acid from the catholyte. In the up-
formed from vertical graphite plates connected per section of the column, an absorber removes
in series, between which there are diaphragms. the remaining hydrogen chloride and the water
To improve gas release, vertical slits are milled vapor by absorption in a water stream, which
Chlorine 69

makes up the water balance of the process. The ric proportions. Experience has shown that there
30 wt % acid that is produced is then cooled, tends to be a surplus of either chlorine or sodium
purified if necessary by activated carbon, and hydroxide. Chlorine may, however, be produced
supplied to the anolyte and catholyte circulation competitively without the byproduct sodium hy-
systems. droxide by nonelectrolytic methods. The starting
The electrolyte is pumped through a filter and material is usually hydrogen chloride, which is
heat exchanger to a gravity feed tank for the catalytically oxidized to chlorine by oxygen, air,
electrolyzer unit. The gases produced are freed nitric acid, sulfur trioxide, or hydrogen peroxide.
from the electrolytes in separators, and the elec- Other processes start from ammonium chloride
trolytes flow back into their respective collecting or metal chlorides.
tanks to be resaturated. The working life of the
PVC diaphragms, 1 – 2 years, depends on the
impurities in the acid. The concentrated acid is, 10.2.1. Catalytic Oxidation of Hydrogen
therefore, purified carefully [196]. Chloride by Oxygen
The product gases are saturated with water
vapor and hydrogen chloride at the partial va- A catalyst is essential for the economic oxida-
por pressures of 20 % hydrochloric acid. Both tion of hydrogen chloride to chlorine by air or
product streams are cooled. Sodium hydroxide oxygen (Deacon Process), and the catalyst must
solution is used to wash the hydrogen, removing be active at low temperature and have adequate
chlorine and hydrogen chloride and producing a life. There are many patents claiming improved
99.9 % product. The chlorine, which is dried by catalysts and equipment. Most of the catalysts
sulfuric acid, contains ca. 0.5 % hydrogen and are oxides and/or chlorides of metals on vari-
ca. 0.05 % carbon dioxide. The hydrogen over- ous substrates. Only three processes have been
potential can be reduced by activation of the commercialized.
cathodes.
The KEL Chlorine Process. The process
Membrane Cells [197, 198]. Since 1992 developed by Kellogg [197] uses concentrated
Bayer has replaced the woven fabric cloth in the sulfuric acid (ca. 80 %) with ca. 1 % nitrosyl-
diaphragm cells by anion-exchange membranes sulfuric acid as the catalyst. From 1975 to 1988
of the sulfonate type. Only hydrated protons are Du Pont operated a full-scale plant in Corpus
able to pass from the anolyte to the catholyte, so Christi, Texas, recovering up to 600 t/d of chlo-
that the whole cell and the electrolyte systems rine. The plant was shut down due to a change
are simplified. Together with an optimized sur- in the structure of the plant and because of
face of the electrodes for better gas release, this material problems after more than 10 years of
leads to: operation. The raw material, from a fluorinated
– Lower cell voltage of 300 mV, correspond- hydrocarbon plant, consisted of waste gases that
ing to a power consumption of 1300 kWh per contained hydrogen chloride [200]. Figure 64
tonne of chlorine at 4.8 kA/m2 shows a simplified flow diagram.
– Longer life of the cell components Sulfuric acid catalyst is fed into the top of
– Higher product quality the stripper column. The hydrogen chloride gas
– Improved safety of operation reacts with the catalyst to form nitrosyl chloride:
– Simplified process
A similar electrolytic process for recovering HCl+NOHSO4 → NOCl+H2 SO4
chlorine from anhydrous HCl, also using mem-
brane cell technology, has been developed by
DuPont [199]. The oxygen, the ultimate oxidizing agent,
blows the remaining hydrogen chloride out of
the sulfuric acid, which becomes more concen-
10.2. Chemical Processes trated and also is cooled in a flash vaporizer. This
acid is then fed back into the process. Nitrosyl
The chlor-alkali process produces chlorine and chloride, hydrogen chloride, oxygen, and water
sodium hydroxide solution in fixed stoichiomet-
70 Chlorine

vapor flow as a gaseous stream into the oxidizer The cooled, dried chlorine gas still contains
and react there, increasing the temperature: ca. 2 % hydrogen chloride and up to 10 % oxy-
gen. Both are removed by liquefaction.
2 NOCl+O2 → 2 NO2 +Cl2 The net reaction is

NO2 +2 HCl → NO+Cl2 +H2 O 4 HCl+O2 → 2 Cl2 +2 H2 O

In the absorber – oxidizer, the rest of the hy-


drogen chloride is oxidized. Concentrated sulfu- The installation at Corpus Christi operated at
ric acid is fed in at the top, reacts with the oxides 1.4 MPa and 120 – 180 ◦ C. On account of the ag-
of nitrogen to form nitrosylsulfuric acid, absorbs gressive nature of the chemicals, expensive ma-
the water that has formed, and is conducted back terials, such as tantalum-plated equipment and
into the stripper: pipes, must be used. For outputs of 250 – 300 t
of chlorine per day, this process can be more
economical than the electrolysis of hydrochlo-
NO+NO2 +2 H2 SO4 → 2 NOHSO4 +H2 O
ric acid, depending on local conditions.
NOCl+H2 SO4 → NOHSO4 +HCl The Shell Chlorine Process. The catalyst
developed by Shell consists of a mixture of cop-
per(II) chloride and other metallic chlorides on
a silicate carrier [201]. The reaction of the sto-
ichiometric mixture of hydrogen chloride and
air takes place in a fluidized-bed reactor at ca.
365 ◦ C and 0.1 – 0.2 MPa. The yield is 75 %. The
water condenses out from the gas stream, and
the hydrogen chloride is removed by washing
with dilute hydrochloric acid. After the residual
gas has been dried with concentrated sulfuric
acid, the chlorine is selectively absorbed, e.g.,
by disulfur dichloride. After desorption and liq-
uefaction, the chlorine has a purity > 99.95 %.
A manufacturing unit was built by Shell in
the Netherlands, 41 000 t/a, and another in In-
dia, 27 000 t/a, but both have been closed down
owing to the prolonged surplus of chlorine on
the market.

The Mitsui MT-Chlorine Process. The cat-


alyst consists of chromium(III) oxide on a sil-
icate carrier [202]. In a fluidized-bed reactor,
hydrogen chloride is reacted with oxygen at a
temperature of 415 ◦ C to give chlorine gas with
a conversion rate of 73 – 77 %. The reactor is
made from nickel-lined low carbon steel. The
concentration of the purified product is > 99.5
% Cl2 .
A commercial plant for 30 000 t Cl2 /a is suc-
Figure 64. Flow diagram of the KEL chlorine process (sim- cessfully operating since 1988, with an expan-
plified) sion to 60 000 t Cl2 /a in 1990.
a) Stripper; b) Oxidizer; c) Absorber – oxidizer; d) Acid
chiller; e) Acid cooler; f) Vacuum flash evaporator
Chlorine 71

A two-stage cyclic fluidized bed process for vapor at high temperature and may also con-
converting HCl to chlorine is described in [203]. tain brine mist and traces of chlorinated hydro-
The catalytic oxidation process combines the carbons, and is normally at atmospheric pres-
exothermic oxidation of 60 – 70 % of the HCl at sure. Before the chlorine can be used, it must be
380 – 400 ◦ C in a fluidized bed of copper oxy- cooled, dried, purified, compressed, and where
chlorides impregnated on zeolite with the trans- necessary, liquefied. A simplified flow sheet is
fer of the reaction products to a second reactor shown in Figure 65.
operating at 180 – 200 ◦ C where the rest of HCl
is converted.
11.1. Cooling
10.2.2. Oxidation of Hydrogen Chloride by Table 22 shows the volume, water content, and
Nitric Acid heat content of 1 kg of chlorine gas at 101.3 kPa
The nitrosyl chloride route to chlorine is based as a function of temperature. To avoid solid chlo-
on the strongly oxidizing properties of nitric rine hydrate formation, the gas is not cooled be-
acid: low 10 ◦ C [206]. Cooling is accomplished in ei-
ther one stage with chilled water or in two stages
with chilled water only in the second stage.
6 HCl+2 HNO3 → 2 Cl2 +2 NOCl+4 H2 O
Table 22. Volume, moisture content, and enthalpy of 1 kg of
chlorine gas at 101.3 kPa as a function of temperature t
2 NOCl+2 H2 O+O2 → 2 HCl+2 HNO3
t, ◦ C Volume, m3 Water Heat
content, content,
The practical problems lie in the separation of Dry Saturated * g/kg ** kJ/kg *
the chlorine from the hydrogen chloride and ni-
trous gases. The dilute nitric acid must be recon- 0 0.312 0.314 1.54 3.81
20 0.335 0.342 5.95 24.45
centrated. Corrosion problems are severe. Sug- 40 0.357 0.385 69.50
gested improvements include (1) oxidation of 19.7
concentrated solutions of chlorides, e.g., LiCl, 60 0.380 0.473 61.5 188.41
70 0.392 0.565 112.0 325.73
by nitrates followed by separation of chlorine 80 0.404 0.756 222.0 623.83
from nitrosyl chloride by distillation at 135 ◦ C
* Chlorine gas saturated with water vapor at temperature t.
or (2) oxidation by a mixture of nitric and sulfu- ** Grams of H2 O per kg of Cl2 .
ric acids with separation of the product chlorine
and nitrogen dioxide by liquefaction and frac- The chlorine gas can be cooled indirectly in a
tional distillation [204]. tubular titanium heat exchanger so that the cool-
ing water is not contaminated and the pressure
drop is small. The resultant condensate is either
10.2.3. Production of Chlorine from fed back into the brine system of the mercury
Chlorides process or dechlorinated by evaporation in the
Alkali-metal chlorides, ammonium chloride, case of the diaphragm process.
and other metallic chlorides are reacted, usu- The chlorine gas can be cooled directly in
ally with nitric acid, to produce nitrate fertilizers packed towers. Water is sprayed into the top
[205]. Chlorine is not produced directly, but it and flows countercurrent to the chlorine. This
can be obtained from the intermediate products treatment thoroughly washes the chlorine; how-
nitrosyl chloride or hydrochloric acid. ever, dechlorination of the wastewater consumes
a large amount of energy. The cooling water
should be free of traces of ammonium salts to
11. Chlorine Purification and avoid the formation of nitrogen trichloride.
Closed-circuit direct cooling of chlorine
Liquefaction combines the advantages of the two methods.
Chlorine produced by the various processes, es-
pecially by electrolysis, is saturated with water
72 Chlorine

Figure 65. Simplified flow diagram of a chlorine processing plant


a) Chlorine gas cooler (primary); b) Chlorine demister; c) Blower or fan; d) Chlorine gas cooler/chiller (secondary);
e) Condensate collection tank; f) Drier, first stage; g) Drier, second stage; h) Sulfuric acid mist separator; i) Sulfuric acid
circulation pump; k) Cooler for circulating sulfuric acid; l) Sulfuric acid feed tank; m) Cooler for sulfuric acid feed

The chlorine-laden water from the cooling tower Activated carbon filters can adsorb organic
is cooled in titanium plate coolers and recycled. impurities and may be regenerated by heating to
The surplus condensate is treated like the con- 200 ◦ C.
densate from indirect cooling. Spray towers, as Gaseous impurities can be removed by ab-
well as packed towers, are used. Water carry- sorption of the chlorine in a suitable solvent,
over is removed by demisters, which reduce the such as carbon tetrachloride, water, or disulfur
amount of sulfuric acid used for drying. dichloride, followed by desorption. This can be
coupled with further processes, such as the re-
covery of chlorine from the waste gas remaining
11.2. Chlorine Purification after liquefaction [15, pp. 418 – 422].
A wash with concentrated hydrochloric acid
Water droplets and impurities such as brine mist removes the dangerously explosive nitrogen
are mechanically removed by special filter ele- trichloride [208]. Scrubbing with liquid chlorine
ments with glass wool fillings. The efficiency (see Fig. 67) mainly reduces the content of or-
varies with the gas throughput. A commonly ganic impurities and carbon dioxide, but it can
used device is the Brink demister [207]. Instead also lower the bromine content. When the chlo-
of glass wool, porous quartz granules can be rine is cooled down to near its dew point, liquid
used. chlorine scrubbing is often combined with com-
In electrostatic purification, the wet chlorine pression by turbo or reciprocating compressors.
gas is passed between wire electrodes in ver-
tical tubes. The electrodes are maintained at a
d.c. potential of 50 kV with a current density 11.3. Drying
of 0.2 mA/m2 . The particles and droplets in the
Drying of chlorine is carried out almost ex-
chlorine become charged and collect on the tube
clusively with concentrated sulfuric acid (96 –
walls. The resultant liquid is fed back into the
98 wt %) [209]. Depending on the desired fi-
brine system or chemically treated before dis-
nal concentration of the waste acid, drying can
posal.
be a two-, three-, or four-stage process. The
acid and chlorine flow countercurrently. The fi-
Chlorine 73

nal moisture content depends on the concen- sion from increasing the temperature enough to
tration and temperature of the acid in the fi- ignite material in contact with the chlorine.
nal stage (Fig. 66). An upper limit is 50 ppm Wet chlorine gas can be compressed 20 –
H2 O. Low-temperature liquefaction (− 70 ◦ C) 50 kPa by a single-stage blower or fan with a
demands lower moisture content, which can be rubber-lined steel casing and titanium impeller.
achieved with molecular sieves, whereby 2.5 It can also be compressed in liquid-ring com-
ppm is possible [210]. pressors, so that further treatment of the chlo-
rine can be accomplished in smaller equipment
[211]. Sulfuric acid ring compressors are used
for throughputs of 150 t of dry chlorine gas per
day per compressor and for pressures of 0.4 MPa
or, in two-stage compressors, 1.2 MPa. The heat
of compression is removed by cooling the cir-
culating liquid; cooling of the gas is not nec-
essary. Advantages are simplicity of construc-
tion, strength, and reliability, but efficiency is
low [212].
Reciprocating compressors were formerly lu-
bricated with sulfuric acid, but are now available
as dry-ring compressors (no lubrication). They
can compress up to 200 t per day. Multistage
compressors produce pressures up to 1.6 MPa.
The heat of compression of each stage must be
removed by heat exchangers or by injection of
liquid chlorine (see Fig. 67). Well-purified chlo-
rine gas is essential for trouble-free operation
[213].
Figure 66. Drying chlorine with sulfuric acid
Attainable moisture content as a function of concentration
and temperature of the acid

The packed towers usual in the first stages


are constructed of rubber-lined steel or glass-
fiber-reinforced poly(vinyl chloride). The heat
liberated on dilution of the circulating acid is
removed by titanium heat exchangers, and the
weak acid is dechlorinated chemically or by
blowing air. Often the acid is recirculated after
reconcentration to 96 % by heating under vac-
uum. Generally, columns with bubble cap plates Figure 67. Multistage reciprocating compressor for chlo-
or sieve trays are used at the final stage. The dry- rine liquefaction at 1 MPa with cooling water at 15 ◦ C with
ing is effective, but the pressure drop is great. liquid chlorine scrubbing
Occasionally, spray towers are used to dry chlo- a) Low-temperature cooling and scrubbing column;
b) Collection tank for impurities; c) Three-stage com-
rine. pressor; d) Intermediate cooler, stage 1; e) Intermediate
After drying, the chlorine gas is passed cooler, stage 2; f) Liquefier; g) Chlorine collection vessel;
through a demister or a packed bed to remove h) Chlorine storage tank; i) Chlorine storage tank on load
sulfuric acid mist. cells

Turbo compressors are most economical


11.4. Transfer and Compression when they operate with large amounts of chlo-
rine. Each unit compresses up to 1800 t/d.
In all operations involving compression, care In multiple-stage operation, pressures up to
must be exercised to prevent the heat of compres-
74 Chlorine

1.6 MPa are reached. Labyrinth seals are used of enough inert gas to keep the mixture below
on the high-speed shafts. Requirements for cool- the explosive limit. Multistage installations can
ing and gas purity are like those of recipro- liquefy over 99.8 % of the chlorine gas.
cating compressors. Screw compressors handle High-pressure (0.7 – 1.6 MPa) liquefaction
low rates of chlorine and give pressures up to with water cooling (Fig. 67) does not require a
0.6 MPa. Sundyne blowers are one-stage high- cooling plant. Therefore, it has the lowest energy
speed centrifugal compressors handling 80 – cost of all methods; however, the high construc-
250 t per day and giving pressures up to 0.3 tion cost must be set against this.
MPa. Liquid chlorine injection is used for cool- Medium-pressure (0.2 – 0.6 MPa) liquefac-
ing [214]. Membrane compressors are used for tion with cooling (− 10 to − 20 ◦ C) is especially
pressurizing storage tanks with chlorine gas to useful when only a part of the chlorine is to be
transfer liquid chlorine to other vessels [215]. liquefied and the remaining gas is to be reacted
Liquid chlorine is pumped with canned pumps at the liquefaction pressure, e.g., with ethylene
[216]. to form ethylene dichloride. The residual gas can
be fed into the compressor suction systems, pro-
vided that the increased inert gas content does
11.5. Liquefaction not interfere with the subsequent process. Oth-
erwise, the residual gas must be scrubbed free
The most suitable liquefaction conditions can of chlorine or liquefied in a second stage.
be selected within wide limits. Important fac- Figure 68 shows a two-stage liquefaction
tors are the composition of the chlorine gas, the by the Uhde system, which operates at 0.3 –
desired purity of the liquid chlorine, and the de- 0.4 MPa and − 20 ◦ C in the first stage and
sired yield. There are nomograms that give the − 60 ◦ C in the second stage, with a yield
relationship between the chlorine concentrations of 99 % [220]. The refrigerant is difluo-
of the incoming and residual gases, liquefaction romonochloromethane. The gaseous refrigerant
yields, pressures, and temperatures [217]. In- is compressed, liquefied by water cooling, and
creasing the liquefaction pressure increases the collected in a container. The liquid refrigerant is
energy cost of chlorine compression, although sprayed into the shell of the chlorine liquefier,
the necessary amount of cooling decreases, re- where it evaporates, absorbing heat and cool-
sulting in an overall reduction in energy require- ing the chlorine, which flows from the liquefiers
ment (Table 23) [218]. at − 15 ◦ C (first liquefier) or − 55 ◦ C (second
Table 23. Electrical energy requirement for compression and
liquefier) [221]. The residual gas from the first
liquefaction of 1 t of chlorine gas horizontal liquefier contains < 5 % hydrogen. It
Liquefaction pressure, MPa
is fed into the second liquefier, which is at an
0.1 0.3 0.8 1.6
angle of 60◦ and has a strong, low-volume con-
struction. There the gas mixture passes through
Energy for compression, the explosive concentration limits. In case of an
kWh/t 5 23 42 57
Energy for cooling, kWh/t 87 68 27 3 explosion, there is a comprehensive control sys-
Combined energy, kWh/t 92 91 69 60 tem to ensure safety:
Starting temperature, ◦ C −36 −8 25 53 The explosion pressure is vented by means of
Final temperature, ◦ C −42 −17 14 40 a bursting disk to a residual gas absorber. Simul-
taneously the residual gas from the first stage is
Any hydrogen is concentrated in the resid- passed directly into this absorber. The chlorine
ual gas. To keep the hydrogen concentration be- gas to the second stage is shut off, and an inert
low the 6 % explosive limit, conversion of gas to gas purge is introduced. Finally, the liquid chlo-
liquid should be limited to 90 – 95 % in a single- rine exit valve is closed to prevent back flow of
stage installation. Higher yields may be obtained the liquid chlorine into the second liquefier and
by condensing the chlorine from the residual from there into the absorber.
gas in a second stage, which is constructed to With normal-pressure (ca. 0.1 MPa) lique-
reduce the risk from explosion [219]. This is faction and low temperature (< − 40 ◦ C), cryo-
achieved by the use of sufficiently strong equip- genic storage of the liquid chlorine is possible.
ment to withstand explosions or by the addition
Chlorine 75

Figure 68. Flow diagram of a two-stage chlorine liquefaction plant at intermediate pressure – Uhde system
a) Chlorine gas compressor; b) Refrigerant collector, stage 1; c) Refrigerant condenser, stage 1; d) Chlorine liquefier, stage
1; e) Refrigerant separator, stage 1; f) Refrigerant compressor, stage 1; g) Liquid chlorine storage tank; h) Chlorine liquefier,
stage 2; i) Bursting disk; j) Refrigerant separator, stage 2; k) Refrigerant condenser, stage 2; l) Refrigerant collector, stage 2;
m) Refrigerant compressor, stage 2

This process is advantageous when large quanti- The snift gas is then cooled by cooling water
ties of chlorine must be liquefied as completely to 45 ◦ C and then further cooled to − 12.2 ◦ C
as possible. Attention must be paid to the in- by Freon. Gas is sent to the absorber, whereas
creased solubility of other gases at low tem- liquid is either returned to chlorine storage or is
peratures, especially carbon dioxide [206]. This used for reflux at the stripper.
carbon dioxide can be removed from the liquid The off-gas enters the bottom of the chlo-
chlorine by passage of hot chlorine gas [219]. rine absorber and passes upward through the two
An absorption – desorption process by Akzo packed sections of the tower while cold carbon
is based on carbon tetrachloride [222]. It requires tetrachloride flows downward. All of the chlo-
little energy and yields over 99.8 % of a pure liq- rine and the nitrogen trichloride is absorbed in
uid chlorine that is almost free of carbon diox- the carbon tetrachloride while the noncondens-
ide. A similar process by Diamond Shamrock able gases remain in the gasphase and are re-
has been described [224]. moved from the system.
The chlorine-rich carbon tetrachloride leaves
the bottom of the chlorine absorber at ca. 10 ◦ C
11.6. Chlorine Recovery and is forced by pressure difference to the chlo-
rine stripper. Chlorine stripper feed enters the
Chlorine can be recovered from the tail gas from middle of the column and flows downward
liquefaction with a chlorine recovery system. through two packed sections, releasing chlorine
Tail gas from liquefaction and chlorine from as it is heated. A thermosiphon reboiler is pro-
the plant evacuation system together with the vided at the base of the stripper. By heating
snift compressor and stripper recycle streams are the liquid above 65 ◦ C, the absorbed nitrogen
supplied to a snift compressor suction knock- trichloride decomposes to nitrogen and chlorine
out drum. The gas is compressed by the snift gas.
gas compressor to 7.0 kg/cm2 with a discharge Chlorine boiled off in the stripper passes up-
temperature of 85 ◦ C. ward through a packed top section of the col-
umn where it is scrubbed and purified by liq-
76 Chlorine

uid chlorine from the discharge knock-out drum. shell around a double-enveloped low-pressure
The stripper overhead stream, a mixture of chlo- storage tank can provide such a facility. To vent
rine and a small amount of inerts, is sent to the chlorine, there must be an absorption or lique-
chlorine liquefaction system or recycled to the faction system. In the course of all operations,
suction knock-out drum to maintain the stripper the design pressure should not be exceeded. The
reflux [222, 223]. dimensions of branches and the amount of pipe
work should be minimized. Bottom connections
from storage tanks are not recommended for
12. Chlorine Handling small chlorine users. Large branches should al-
ways be located in the gas space of a vessel.
Both the chlorine industry and governmental or- The pipework system should be provided with
ganizations are well aware of the risks of chlo- remotely operable valves to permit isolation in
rine. In the United States and Canada, The Chlo- case of emergency.
rine Institute [225] has established standards and Before being put into service, the whole stor-
recommendations for safe transport and han- age system must be degreased, cleaned, and
dling of chlorine since 1924. In Europe, Euro dried to achieve a dew point of − 40 ◦ C in the
Chlor, an association of major Western Euro- purge gas at the outlet of the system. No sub-
pean chlorine manufacturers, publishes recom- stance that could react with the chlorine can be
mendations, codes, and memorandums for chlo- allowed to enter the storage system. The filling
rine handling and transport concerning Euro- ratio in the tank should never exceed 95 % of the
pean conditions and regulations [226]. Both or- total volume of the vessel; for pressure storage
ganizations distribute manuals and pamphlets tanks, this corresponds to 1.25 kg of liquid chlo-
worldwide. Surveys of existing national and rine per liter of vessel capacity at 50 ◦ C (Fig.
international regulations for the handling and 69).
transport of hazardous chemicals are available
[2, 34, 227, 229].

12.1. Storage Systems


Chlorine is liquified and stored at ambient or
low temperature [230, 233]. In both cases the
pressure in the storage system corresponds to
the vapor pressure of liquefied chlorine at the
temperature in the stock tank. Pressure storage
is recommended for all usual customers [234,
235]. Euro Chlor recommends a maximum ca-
pacity of 300 – 400 t for individual tanks. For the
large storage capacities required by producers,
usually a low-pressure storage system, operat-
ing at a liquid chlorine temperature of ca. − 34 Figure 69. Proportion (%) of a one-liter vessel occupied
◦ by 1.25 kg of liquid chlorine as a function of temperature
C, is chosen. A low-pressure system needs a
cooling or recompression system, and, for this
reason, it is basically unsuitable for small chlo- Typical measuring and control equipment of
rine consumers [236, 237]. a pressure storage tank is shown in Figure 70.
A few major design aspects must be men- The ISO codes for process measurement control
tioned. Any risk of fire or explosion must be functions and instrumentation are explained in
eliminated. All tanks having an external connec- Table 24 [238]. The measuring equipment of a
tion below the liquid level should be placed in a low-pressure storage system needs supplemen-
liquor-tight embankment (bund). In the event of tary devices, for example, a temperature indica-
leakage the liquid should be collected in a small tor with an alarm and, in the case of a double-
area to reduce the rate of vaporization. The outer
Chlorine 77

shell vessel, a device to determine the quality 12.2. Transport


of the purging gas inside the double shell. The
vessel and an external envelope should be pro- Within a chemical plant and over distances of
tected against overpressure or underpressure. In several kilometers, chlorine can be transported
low-pressure systems, the chlorine is removed by pipelines, either as gas or liquid [24, 239].
by vertical submerged pumps, canned pumps be- Every precaution should be taken to avoid any
low the vessel, or ejector pumps operating with vaporization of chlorine in a liquid-phase sys-
a flow of liquid chlorine produced by external tem or any condensation in a gas-phase system.
pumps. Wherever liquid chlorine could be trapped be-
tween two closed valves or wherever the system
Table 24. ISO codes and miscellaneous symbols for process could be overpressurized by thermal expansion,
measurement control functions and instrumentation
an expansion chamber, a relief valve, or a rupture
Codes Function or Instrumentation disk should be provided [240, 241].
AA analysis alarm Commercial chlorine is transported as a liq-
CW cooling water uid, either in small containers (cylinders and
dPI difference pressure indicating drums) or in bulk (road and rail tankers, barges,
FA flowrate alarm
FI flowrate indicating and ISO containers). The design, construction,
FIA flowrate indicating alarm system of labeling, inspection, and commission-
FICA flowrate indicating controlling alarm ing are covered by national and international reg-
FRA flowrate recording alarm
HZ hand operated emergency acting
ulations [227]. Cylinders have a chlorine con-
H high tent up to 70 kg. A protective hood is provided
L low to cover the valve during transport. The ton con-
LA level alarm tainers (drums) have a capacity of 500 – 3000 kg
LIA level indicating alarm
LIC level indicating controlling of chlorine. Drums are equipped with two valves
M moisture analysis near the center of one end and connected with
PA pressure alarm internal eductor pipes.
PCZA pressure controlling emergency acting alarm
PI pressure indicating
The capacity of tank cars (rail tankers)
PIA pressure indicating alarm ranges from 15 to 90 t. Special angle valves are
PIAS pressure indicating alarm switching mounted on the manhole cover on top of the ves-
PIC pressure indicating controlling sel. In Europe, pneumatic valves are normally
PRC pressure recording controlling
PSA pressure switching alarm
used [230 – 232]. During loading and unload-
QRA quality recording alarm ing, these valves can be closed rapidly and re-
TA temperature alarm motely in case of an accident. They have an in-
TC temperature controlling ternal safety plug, providing a tight seal against
TI temperature indicating
TIA temperature indicating alarm the passage of gas or liquid chlorine in the event
TIC temperature indicating controlling of failure of the body of the valve.
TRA temperature recording alarm In North America, the eductor pipe inside the
WI weight indicating
WIA weight indicating alarm
vessel has an excess-flow valve at the top, im-
measuring device mediately below the manhole cover. This valve
remote control valve closes the eductor pipe when the rate of liquid
––– control line flow exceeds a set rate [2, 24]. North Ameri-
can tank cars have a spring-loaded safety relief
valve, which protects the vessel against over-
Periodic inspection and retesting of the whole
pressure in case of external heat. The tanks have
system, including a visual examination, a thick-
thermal insulation. In Europe thermal insulation
ness test of the wall of the vessel and pipes, and
and safety relief valves are not used or recom-
an examination of the welds and the surfaces
mended.
under any thermal insulation, is recommended.
Road tankers and ISO containers have a chlo-
Hydraulic retesting is accompanied by risk of
rine capacity of 15 – 20 t. The design of and the
corrosion and is, therefore, not favored.
equipment on chlorine pressure road tankers is
similar to these of rail tankers. In North Amer-
ica, large amounts of chlorine are transported by
78 Chlorine

Figure 70. Discharge of liquid chlorine by padding to pressure storage


a) Liquid-chlorine rail tanker; b) Flexible connection; c) Plug; d) Viewing glass; e) Remote-control tank valves; f) Protective
membrane; g) Storage vessel; h) Rupture disk; i) Relief safety valve; j) Buffer vessel for liquid chlorine

tank barges [24]. These barges usually are of the 12.3. Chlorine Discharge Systems
open-hopper type with several cylindrical unin-
sulated pressure vessels. The total capacity of All containers should be discharged in the same
barges ranges from 600 to 1200 t. The chlorine order as received. They must be placed where
is transported at low temperature. any external corrosion, risk of fire, explosion, or
damage is avoided [24]. At normal room tem-
Classification and Labeling. According to perature, the discharge rate of chlorine gas from
Directive EEC 67/548, Annex I, chlorine (In- a single 70-kg cylinder is ca. 5 kg/h and the rate
dex no. 017–001–00–7) is classified as toxic of a drum is ca. 50 kg/h. The flow of chlorine gas
and dangerous to the environment. The R and can be increased by a higher ambient tempera-
S phrases are: ture or by connecting two or more containers. A
system of two or more containers must be care-
R 23 toxic by inhalation
R 36/37/38 irritating to eyes, respiratory system and skin
fully operated and controlled to avoid overfilling
R 50 very toxic to aquatic organisms by transfer of chlorine from warm to cool con-
S 7/9 keep container tightly closed and in a tainers. Direct heating of containers is not rec-
well-ventilated place ommended [24]. The best way to determine the
flow rate and container content is to observe the
For transportation, chlorine is in class 2, no. weight of the container [228]. A flexible tube is
2TC (toxic, corrosive) in ADR, RID, and ADNR used to connect a mobile container with the fixed
in Europe, and in class 2.3 in the IMDG Code (p. piping system. Any reverse suction from the con-
2116) and ICAO Code. All vessels must be la- suming plant must be prevented by a baromet-
beled with the denomination 268/1017 Chlorine ric leg or other adequate precaution if the chlo-
and with labels for dangerous goods: cylinders rination process runs at atmospheric pressure.
and drums with labels 2.3 (toxic gas) and 8 (cor- Pressurized processes need a pressure control-
rosive); railroad tankers with labels 2.3, 8, and ling system with automatic isolation valves.
13 (shunt carefully); and for shipping, labels 2.3, Uninsulated tanks have a maximum gaseous
8, and MP (marine pollutant). discharge rate of ca. 2 t/h. The chlorine gas can
be used only for low-pressure chlorination pro-
cesses and at low rates. This method increases
Chlorine 79

Figure 71. Liquid chlorine vaporizer


a) Liquid-chlorine drum; b) Buffer vessel; c) Flexible coil; d) Chlorine vaporizer; e) Protective membrane; f) Relief safety
valve; g) Rupture disk; h) Barometric leg; i) Water pump; j) Water heater

the risk of concentrating nitrogen trichloride and 12.4. Chlorine Vaporization


other nonvolatile residues in the liquid phase
within the tank. In all other circumstances, the When large amounts of chlorine gas are required
liquid chlorine should be transferred into a fixed or when the chlorination process needs pressur-
storage vessel and then vaporized in a special ized gas, liquid chlorine must be vaporized and
installation. superheated to avoid liquefaction [242, 243]. It
Liquid chlorine is discharged by putting the is advisable to operate the vaporizer at a suf-
tank under pressure with dry inert gas or dry ficiently high temperature to accelerate the de-
chlorine gas. The inert transfer gas must have a composition of nitrogen trichloride. As a source
dew point below − 40 ◦ C at atmospheric pres- of heat, steam with a maximum allowed tem-
sure and must be clean and free of impurities perature of 120 ◦ C is used when the vaporizing
such as dust or oil. Before closing the valves, system is constructed of mild steel. Water above
the tanks must be vented to avoid the risk of high 60 ◦ C is also suitable, as shown in Figure 71. Di-
pressure in the container on account of the addi- rect electrical heating is not appropriate because
tional partial pressure of the inert gas. The use of there is always a risk of overheating the steel.
an inert gas requires the availability of a chlorine Coil-in-bath vaporizers use a coiled tube or
absorption or neutralization system. Discharge a spiral located in a vessel of hot water (Fig.
with pressurized chlorine gas requires a chlo- 71). Generally, they are used for small through-
rine vaporizer or a special chlorine compressor. puts; they are simple in design and construction.
Articulated arms, flexible hoses, and steel coils Double-envelope vaporizers have compact con-
are used for the flexible connections. Remote- struction and are easy to operate and to maintain.
control valves installed close to the ends of the Vertical tube vaporizers have a large surface area
flexible connections limit leakage in the event of and allow a high flow rate. Kettle vaporizers are
a failure. also constructed for large unit capacities [244].
Recommendations on technical equipment, Every effort must be taken to avoid the reverse
installation, taking into operation, checks, and suction of water or organic materials into the va-
handling are provided by The Chlorine Institute porizer. The recommended water and nitrogen
and Euro Chlor. trichloride content of introduced liquid chlorine
must not be exceeded. Vaporizers operating at
low temperature or with a constant liquid level
80 Chlorine

Figure 72. Absorption equipment for the treatment of gases containing chlorine
a) Buffer vessel; b) Vent fan; c) Packed tower; d) Circulating pump; e) Heat exchanger (cooler)

need to be purged to avoid dangerous concen- Where this is not possible, several methods can
tration of nitrogen trichloride [245, 246]. be used to decompose the hypochlorite: con-
trolled thermal decomposition, catalytic decom-
position [249], acidification, for example, with
12.5. Treatment of Gaseous Effluents sulfuric acid

Gaseous effluents containing chlorine arise from NaCl+NaOCl+H2 SO4 → Na2 SO4 +Cl2 +H2 O
various sources and must be treated in such a way
as to obtain a tolerable concentration of chlorine
when they are released into the air. The vent gas
may contain other substances, such as hydro-
gen, organic compounds, CO2 , etc., which must 12.6. Materials
be considered in design and operation of an ef-
fluent treatment installation [247, 248]. The choice of material [250] depends on the de-
Operation of the collection system below at- sign and operating conditions and must take into
mospheric pressure facilitates the purging of account all circumstances. A chlorine manufac-
chlorine vessels, pipes, etc. The risk of corrosion turer should be consulted to confirm the suitabil-
in dry chlorine installations by moisture from the ity of a material. Any use of silicone materials
treatment system must be excluded. The most in chlorine equipment should be avoided.
commonly used and recommended reagent is
caustic soda. The effluents are treated in an ab- Dry Chlorine Gas (water < 40 ppm by
sorption system, such as packed absorption tow- weight). Carbon steel is the material most used
ers, venturi scrubbers, etc. An example of a flow for dry chlorine gas. It is protected by a thin
sheet for a large plant is shown in Figure 72. layer of ferric chloride. For practical purposes
To avoid any formation of solid salts, the the recommended temperature of these mate-
recommended concentration of caustic soda is rials is ≤ 120 ◦ C. High-surface areas, such as
< 22 wt %. The operating temperature should steel wool, or the presence of rust and organic
not exceed 55 ◦ C; under normal conditions a substances increase the risk of ignition of steel.
temperature of ca. 45 ◦ C is usual. A cooling The resistance of stainless steels to chlorine
system may be necessary. In large chlorine ab- at high temperature increases with the content of
sorption units, the sodium hypochlorite solution nickel. For stainless steels containing less than
that is produced can be used in other processes. 10 wt % nickel, the upper temperature limit is
Chlorine 81

150 ◦ C. High-nickel alloys, such as Monel, In- are recommended [253]. In wet chlorine gas,
conel, or Hasteloy C, are suitable up to 350 – rubber or synthetic elastomers are acceptable.
500 ◦ C. Poor mechanical strength limits the use Even at temperatures up to 200 ◦ C, PTFE is re-
of nickel. Copper is used for flexible connections sistant against wet and dry chlorine gas and liq-
and coils, but it becomes brittle when stressed uid chlorine.
frequently. Materials resistant because of protection by a
Because titanium ignites spontaneously in chloride surface layer are not recommended for
dry chlorine, it must be avoided. Graphite, glass, protective membranes, rupture disks, and bel-
and glazed porcelain are used where there is lows. Suitable materials are tantalum, Hasteloy
a risk of moisture in the dry chlorine gas, and C, PTFE, PVDF, Monel, and nickel.
poly(vinyl chloride) (PVC) or chlorinated PVC
and polyester resins are suitable if the tempera-
ture limits of these materials are regarded. 12.7. Safety
Liquid Chlorine. Unalloyed carbon steel In hazard and risk assessment studies, the design
and cast steel are used with liquid chlorine. Low- of chlorine installations and equipment and the
temperature chlorine systems apply fine-grain operating and maintenance concepts are exam-
steels with a limited tensile strength to guarantee ined in detail to minimize risks [254]. However,
good conditions for welding. To avoid erosion of there remains a certain risk, and all efforts must
the protective layer, practice is to limit the ve- be taken to protect people and the environment
locity of the liquid to less than 2 m/s. Organic in the case of a chlorine emergency. The pen-
materials – rubber lining, ebonites, polyethy- etrating odor and the yellow-green color of a
lene, polypropylene, PVC, chlorinated PVC, cloud indicate chlorine in the air. If around-the-
polyester resins, and silicone – are dangerous clock surveillance by operators is not possible,
[251]. Zinc, tin, aluminum, and titanium are not automatic leak detectors are available. Safety in
acceptable. For certain equipment, copper, sil- handling chlorine depends largely on the educa-
ver, lead, and tantalum are appropriate. tion and training of employees. An emergency
plan should be brought to the attention of the
Wet Chlorine Gas. Wet chlorine gas rapidly personnel involved. Computer-assisted systems
attacks most common metallic materials with can be used in certain circumstances [255]. Peri-
the exception of tantalum and titanium. To as- odic exercises and safety drills should be carried
sure a protective oxide layer on the surface of out.
the titanium, sufficient water must be present in All people on a chlorine plant are advised to
the chlorine gas. If the system does not remain carry escape-type respirators. The use of filter
sufficiently wet, titanium ignites spontaneously masks is prohibited where there is a risk of a
[252]. high concentration of chlorine. Anyone who en-
Most organic materials are slowly attacked ters an area with high chlorine concentrations
by wet chlorine gas. Rubber-lined iron is suc- should be equipped with self-contained breath-
cessfully used up to 100 ◦ C. At low pressure ing apparatus and full protective clothing suit-
and temperature the use of plastic materials like able for dealing with liquid chlorine. Protective
PVC, chlorinated PVC, and reinforced polyester equipment, safety showers, eye-wash facilities,
resins is advantageous. Polytetrafluoroethy- and emergency kits [24] must be quickly acces-
lene (PTFE), poly(vinylidene fluoride) (PVDF), sible.
and fluorinated copolymers like tetrafluoroethy- A means of indicating the actual wind direc-
lene – hexafluoropropylene (FEP) are resistant tion should be located near the chlorine installa-
even at higher temperature. Ceramics have been tion.
progressively replaced by plastics. Impregnated Fixed or mobile water curtains can be used
graphite is suitable up to 80 ◦ C; the impregna- to divert the dispersion of a chlorine gas cloud
tion should be resistant to wet chlorine. [256]. However, the direct discharge of water
into liquid chlorine and on the area of a chlorine
Materials for Special Parts. After the ban leak must be avoided.
of asbestos as material for gaskets, substitutes
82 Chlorine

In most countries, chlorine manufacturers Sampling ISO 1552 [269, 270]


have organized groups of experts who are well Moisture ISO 2121 [271], ASTM E410
versed and drilled in handling chlorine and can [272]
be called at any time in case of chlorine emer- ISO 2202 [273]
Chlorine content ISO 2120 [274], ASTM E412
gency. [275]
The Chlorine Institute has released pamphlets Gaseous components DIN 38 408, part 4 [276]
and recommendations covering all aspects of NCl3 [277]
Mercury [278], ASTM E506 [279]
safety, e.g., first aid [257], emergency response
plans [258], protective equipment [259], preven-
tion of injuries to personnel [260], prevention of The residue is weighed, and the organic con-
chlorine releases [261], and estimating the area stituents are taken up in acetone, hexane, or di-
affected by a chlorine release [262]. Emergency ethyl ether and determined by gas chromatog-
kits have been developed for sealing leaks in raphy. The inorganic residue is analyzed. For
chlorine containers, drums, and tank cars [263]. quick analysis, liquid chlorine can be introduced
Euro Chlor offers a Chlorine Safety Manual directly onto a silica gel column of a gas chro-
[264], recommendations for emergency inter- matograph.
vention [265], and for safe design, construction,
operation of equipment [266]. Chlorine Gas. The chlorine gas can be ana-
Emergency plans have been established for lyzed for chlorine content, gaseous impurities,
accidents during transportation and use, e.g., hydrogen, organics, and moisture:
CHLOREP [267] in North America and TUIS
(Transport Unfall Informations System) in Ger- 1) Chlorine content. One method for process
many. monitoring and control of chlorine concentra-
tion is measurement of thermal conductivity.
2) Gaseous impurities. A known amount of chlo-
rine gas is passed through a solution of potas-
13. Quality Specifications and sium iodide or phenol to absorb the chlorine.
Analytical Methods The residual gases (O2 , N2 , H2 , CO, CO2 )
are collected in a gas burette, measured, and
13.1. Quality Specifications analyzed by gas chromatography or with an
Orsat apparatus.
Liquid chlorine of commercial quality must have 3) Hydrogen. A known amount of air is added to
a purity of at least 99.5 wt % [268]. The wa- the residual gas after removal of the chlorine
ter content is < 0.005 wt %, and solid residues to ensure excess oxygen, and the volume re-
are < 0.02 wt %. The impurities are mainly CO2 duction is measured after the hydrogen is con-
(≤ 0.5 wt %), N2 , and O2 (each 0.1 – 0.2 wt %). sumed on a heated platinum coil. The hydro-
There are traces of chlorinated hydrocarbons gen content can be continuously monitored by
(originating from rubberized or plastic piping) thermal conductivity measurement.
and inorganic salts such as ferric chloride. The 4) Organics. Organic components can be deter-
chlorine may also contain small amounts of mined most conveniently by gas chromatog-
bromine or iodine, depending on the purity of raphy.
the salt used in the electrolytic process. 5) Moisture. A known amount of chlorine gas
is passed through a drying tube filled with
a weighed amount of phosphorus pentoxide.
13.2. Analytical Methods The moisture content is determined from the
weight gain of the drying tube (ISO 2121).
Industrial liquid chlorine is mainly analyzed by Continuous determination can be carried out,
the methods in ISO regulations. The liquid chlo- e.g., by absorption with phosphorus pentox-
rine is evaporated at 20 ◦ C, and this gas is then ide and measuring the current required to elec-
analyzed. trolyze the absorbed water or by the electrical
conductivity after absorption in sulfuric acid.
Chlorine 83

Detection of Chlorine. Chlorine can be rec- vanometry, potentiometry, colorimetry, and UV


ognized by smell or color. Small amounts can spectroscopy [281].
be detected by the blue coloration of starch –
iodide paper, although other oxidizing agents
can produce the same effect. Another method for 14. Uses
chlorine detection depends on its ability to com-
bine with mercury. If the unknown gas mixture is The first industrial use of chlorine was to pro-
shaken with water and mercury, all the chlorine duce bleaching agents for textiles and paper and
disappears and the remaining water has a neutral for cleaning and disinfecting. These were liq-
reaction. However, if the chlorine contains some uid bleaches (solutions of sodium, potassium,
hydrogen chloride, the water becomes acidic and or calcium hypochlorite) or bleaching powder
reacts with silver nitrate solution to give a white (chlorinated lime). Chlorine was then regarded
precipitate (AgCl) that is soluble in aqueous am- merely as a useful chemical agent.
monia. Leaks in pipes or equipment are detected Since 1900s, chlorine has achieved con-
by testing with the vapor from aqueous ammo- stantly increasing importance as a raw material
nia: a thick white cloud of chloride forms. for synthetic organic chemistry. Chlorine is an
Quantitative Determination of Free Chlo- essential component of a multitude of end prod-
rine. The gas mixture can be shaken with a ucts, which are used as materials of construc-
potassium iodide solution, and the liberated io- tion, solvents, pesticides, etc. In addition, it is
dine can be then determined by titration. Chlo- contained in intermediates that are used to make
rine in alkaline solution can be reduced to chlorine-free end products. It is these areas of
chloride by potassium or sodium arsenite, and use that allow chlorine production to increase.
the arsenite can be then oxidized to arsenate. The direct use of elemental chlorine, e.g., for
The end point is detected by spot tests with sterilizing water, has declined in some areas,
starch – iodide paper. Excess arsenite is back- but not in others. For example, in Germany, it
titrated with acidified potassium bromate solu- is < 0.1 %, but in the United States it is 6 %
tion. Small amounts of chlorine, e.g., in drinking (1982). The WHO recommends chlorine as the
water, can be determined by photometric mea- most useful agent for water disinfection [282].
surement of the yellow color produced by the The percentage of world chlorine consump-
reaction with o-tolidine in hydrochloric acid so- tion of various product groups in 1997 was as
lution [280]. follows [283]:
To determine both chlorine and carbon diox-
Vinyl chloride 33
ide, the chlorine is absorbed by a solution that Misc. organic products 6
contains known amounts of acid and potassium Solvents 19
arsenite, and the chlorine is determined by back- Pulp and paper 5
Water treatment 6
titration of the arsenite. The carbon dioxide, Others (inorganic products, etc) 31
which is not absorbed by this solution, is then
absorbed by potassium hydroxide solution.
Detection Tubes. Commercial detection The number of possible reactions of chlo-
tubes (Drägerwerk, Lübeck; Auergesellschaft, rine, and therefore the number of intermediates
Berlin) are available for measuring chlorine and end products, is remarkably large. Some im-
in air. They have various ranges: 0.2 – 3, 2 – portant reactions are shown in Figure 73 along
30, and 50 – 500 ppm. The Chlorometer (Zeiss- with the areas of application of the end products
Ikon, Berlin) can determine the free chlorine [284]. Figure 74 also shows some of the applica-
content of water in a few minutes. tions of chlorine. About one-third of the finished
For the protection of the environment and products are chlorine-free, although made with
control of working conditions, traces of chlo- the help of chlorine [285].
rine as small as 0.01 – 10 ppm must be deter-
mined. There are many types of apparatus on Inorganic Products. The use of chlorine
the market for measuring workplace concentra- to produce pure hydrochloric acid, bleaching
tions or emissions. They depend on physico- agents, and other inorganic products is increas-
chemical methods, such as conductometry, gal- ing, although slowly enough that this forms a
84 Chlorine

Figure 73. Important reactions of chlorine and the uses of the end products

steadily decreasing proportion of the total con- States: 13 % in 1983, 9 % in 1994, 6 % in 2000,
sumption. The production of some metal chlo- 0 % in 2010 (forecast); the decline in Canada
rides is increasing, e.g., titanium tetrachloride, and in Scandinavia is even sharper (1983: 60 –
aluminum chloride, and silicon tetrachloride. 70 %; 1998: near 0 %) [286].

Bleaching. Countries with large paper and Organic Products. Organic intermediates
cellulose industries had a large chlorine con- and end products accounted for up to 85 % of the
sumption for bleaching purposes. Ecological chlorine use in developed countries. The prod-
concerns about AOX (absorbable organic halo- uct with the largest chlorine requirement is vinyl
gens) in the effluents of papermills have led to re- chloride, which accounted for 33 % of world
placement of elemental chlorine by other bleach- chlorine consumption in 1997. This proportion
ing agents, including ozone, hydrogen perox- will increase further and outweigh consumption
ide, and chlorine dioxide. As a consequence, the losses in other areas and is therefore the reason
use of chlorine is dropping sharply; e.g., United for increasing chlorine capacities worldwide.
Chlorine 85

Figure 74. The chlorine tree (courtesy of Euro Chlor)

Chlorine reacts with methane to give the The following products are obtained from
solvents methyl chloride, methylene chloride, propylene: propylene oxide, propylene glycol,
chloroform, and carbon tetrachloride. Butane carbon tetrachloride (via propylene dichloride),
and pentane are the starting materials for glycerine (via allyl chloride), and epoxy resins
hexachlorobutadiene and hexachloropentadi- and synthetic rubber (via epichlorhydrin).
ene, the basis of many pesticides. Chlorinated Aromatic derivatives are also important.
paraffins are starting materials for detergent raw The reaction of chlorine with benzene pro-
materials, plasticizers, and lubrication additives. duces the following compounds, among oth-
The chlorination of ethylene first gives 1,2- ers: monochlorobenzene, dichlorobenzene, and
dichloroethane, which is an important starting hexachlorobenzene, which are used as sol-
material for a variety of further products, in- vents or pesticides. A more important use of
cluding vinyl chloride, ethylene oxide, ethy- monochlorobenzene, however, is as an interme-
lene glycol, ethylenediamine, and chloral. Other diate in the manufacture of aniline, and dyes. A
products of the chlorination of ethylene are further important synthesis route is to start from
trichloroethylene, perchloroethylene, etc. toluene to produce toluenediamine and with the
86 Chlorine

help of phosgene to convert this to toluene di- – 85 % of pharmaceuticals are made using chlo-
isocyanate, which is used in plastics. rine
– 98 % of Western Europe’s drinking water is
Environmental Aspects. The production purified by chlorination
and use of chlorine, as well as certain chlori-
nated substances bear risks for human health and These figures demonstrate the economic impor-
the environment. International discussions are tance of the chlorine industry.
underway on legally-binding conventions and Figure 75 shows how world chlorine produc-
protocols thereof. Matters of concern include tion has developed from 1965 to 1995. The fig-
the use of mercury, asbestos, chlorine contain- ures are only estimates, as many important na-
ing materials which are persistant, toxic, and tions do not publish this information.
liable to bioaccumulate (e.g., polychlorinated
biphenyls, dioxins and furans, DDT, endrine,
chlordane), and chemicals having an ozone-
depletion potential (e.g., chlorofluorocarbons).
Regarding the principles of “Responsible Care”
and of “Sustainable Development” [287] the
chemical industry is managing this challenge:
Once a hazard is realised, the substance in ques-
tion is handled individually by investigating its
properties and its impact on the environment on
a scientific basis. Risk assessments and risk –
benefit studies are prepared, the results are pub-
lished as fact sheets, position papers, scientific
publications and contributions to the discussion
with governments, authorities, and the public.
Practical measures for reducing the risks are
continuous efforts to reduce or eliminate emis-
sions (mercury, asbestos), offer recycling con- Figure 75. World chlorine production (1965 – 1995)
cepts for products (PVC, chlorinated solvents),
improve production processes and transporta- Chlorine capacities 1995 (in 103 t/a) in the
tion safety, cease the production (CFCs, pen- main producer countries was as follows [290]:
tachlorophenol in most countries) or marketing
United States 11 860
of problematic compoundes, and developing Japan 4250
for alternative processes (bleaching in the pulp Former Soviet Union 3800
and paper industry). All these measures are in- China 3750
fluencing the use pattern of chlorine: the use Germany 3690
Brazil 1660
for bleaching, for C1 compounds, for solvents, Canada 1630
and for tetraethyllead will further decrease; the France 1580
importance of organic chemicals will rise and United Kingdom 1270
India 1230
PVC will expand especially in the construction Italy 1130
sector. A study conducted by ECOTEC [288]
illustrates the chlorine flow within the industrial
production system and to external consumers. World chlorine capacity was about
40 × 106 t/a in 1983, it rose to 48 × 106 t/a in
1998, and is forecast to reach 55 × 106 t/a by
15. Economic Aspects 2005 [283].
Figure 76 demonstrates the growth of chlo-
In Western Europe in 1995 [289]: rine capacity by economic regions [291]. Com-
– Almost 2 × 106 jobs were related to chlorine pared to 1983, the capacities in 1997 remained
– 55 % of European chemical turnover de- nearly constant in Western Europe, Eastern Eu-
pended on chlorine rope, and Africa. In North America, reductions
Chlorine 87

Figure 76. Chlorine production capacities by economic regions (1983 and 1997)

in Canada were outweighed by growth in the amount of caustic, leading to high caustic prices
United States. Huge capacities were erected in and low chlorine prices. This situation seems to
the Middle East and above all in Asia and Ocea- change in cycles every 6 to 8 years. For example,
nia (China, India), where capacities were tripled. from 1988 to 1991 the market value for chlo-
From the growth between 1997 and 2005, North rine fell from 150 $/t down to 0, but from March
America’s share will be 35 %, Asia’s 40 %, and 1992 to August 1993, the price rose to 230 $/t.
the Middle East’s 15 % [283]. After 2000 China The price for caustic was 320 $/t in 1991; it fell
will become the second biggest producer after to 50 $/t in March 1994 and climbed to 270 $/t
the United States, with an annual chlorine ca- in December 1994 (Fig. 77) [294].
pacity of about 5.5 × 106 t. Because of the fixed ratio of chlorine to caus-
All chlorine producers are listed in [292]. A tic in production and the different location of
detailed review of the chlor-alkali market for uses, worldwide trade flows have been created
1994 is given in [293]. for both products. Chlorine is traded as chlo-
The production costs of 1 electrochemical rinated derivatives, particularly as EDC, VCM,

unit (= 1 t chlorine + 1.13 t sodium hydroxide) and PVC, accounting for 85 %, and chlorinated
depends up to 60 % on the price of electricity. At solvents. The equivalent annual amount of chlo-
3.5 c/ /kWh, they are about 250 $/ecu (Fig. 77), so rine is approximately 4.5 × 106 t (1995 to 2000).
the total value of production is ca 24 × 109 $/a. The main exporting countries are the United
The chlor-alkali industry is one of the biggest States (to South East Asia, South America), Rus-
consumers of electrical energy, consuming ca. sia (to West Europe), West Europe (to South East
0.15 × 1012 kWh/a. Asia, China), Japan (to China), and Middle East
A prediction of the market prices for chlorine Asia (to Japan , China). [295].
and for caustic soda is very difficult. Firstly, most
chlorine is used captively to avoid transporta-
tion, or it is sold on the basis of long-term con- 16. Toxicology
tracts. In relation to production, the amounts on
the spot markets are small. Secondly, the produc- Chlorine gas is dangerous to health because it
tion of chlorine is strictly coupled to that of the is a powerful oxidant. In the physiological pH
caustic soda. A strong demand for chlorine cre- range, it is converted to hypochlorous acid, a cy-
ates a surplus of caustic, leading to high prices totoxic substance. The extent to which the cells
for the chlorine and a drop in prices for caustic; are damaged depends on the gas concentration,
weak demand for chlorine reduces the available the exposure time, the water content of the tis-
88 Chlorine

Figure 77. Production costs and product prices in the United states from 1987 to 1997 [294]
a) Product value for 1 ecu (1 t Cl2 + 1.13 t NaOH; b) Price for caustic soda; c) Production costs for 1 ecu at 3.5 c//kWh electricity
price; d) Price for chlorine into PVC

sue, and the health of the person exposed to the development of toxic tracheobronchitis. After a
gas. Besides getting into the eyes, larynx, and latent period of several hours with fewer symp-
trachea, chlorine also reaches the bronchi and toms, pulmonary edema may occur because of
the bronchioles. Because of its moderate solu- alveolar membrane destruction. This is indicated
bility in water, chlorine affects the alveoli only by increased shortness of breath, restlessness,
at high concentrations. In this respect, chlorine and cyanosis. Subsequently, a further complica-
differs from other gases with low water solubil- tion can occur after several days in the form of
ity and high lipid solubility, such as phosgene, pneumonia caused by superinfection of the in-
nitrogen monoxide, and nitrogen dioxide. Initial jured pulmonary tissue.
moderate bronchial irritation is followed by the A clear dose – effect relationship of chlo-
development of a toxic pulmonary edema be- rine gas at different concentrations in humans
cause of increased alveolar injury. has not yet been published. On the basis of
The olfactory threshold of chlorine gas is results obtained from animal experiments, the
0.2 – 3.5 mL/m3 . Prolonged exposure seems to LC50 for healthy humans is assumed to be
raise the olfactory threshold. Concentrations of 300 – 400 mL/m3 at a 30-min exposure [296].
3 – 5 mL/m3 are tolerated for up to 30 min with- No deaths have occurred in animal experiments
out any subjective feeling of malaise. At concen- at 30-min exposures for concentrations below
trations between 5 and 8 mL/m3 , mild irritation 50 mL/m3 . Death following acute intoxication
of the upper respiratory tract and the conjunc- is caused by a fulminant pulmonary edema.
tiva is observed. In addition, running of the eyes Many investigations deal with the toxic-
and coughing are observed at concentrations of ity of low chlorine concentrations. Investiga-
15 mL/m3 and higher. Above 30 mL/m3 , the fol- tions in humans indicate the possibility of re-
lowing symptoms are observed: nausea, vomit- versible damage to the lung function parameters
ing, oppressive feeling, shortness of breath, and at concentrations starting at 1.0 mL/m3 , whereas
fits of coughing, sometimes leading to bronchial 0.5 mL/m3 is a no-observable-effect-level [297,
spasms. Exposure to 40 – 60 mL/m3 leads to the 298]. In monkeys no effects have been observed
Chlorine 89

at 0.5 mL/m3 in a one-year study [299]. Long- 12. D. W. F. Hardie, W. W. Smith: Electrolytic
term investigations of workers exposed to chlo- Manufacture of Chemicals from Salt, The
rine, e.g., in chlor-alkali electrolysis plants or in Chlorine Institute, New York 1975.
pulp manufacture, however, do not indicate in- 13. A. Schmidt: Angewandte Elektrochemie,
creased rates of mortality or morbidity caused Verlag Chemie, Weinheim 1976.
by pulmonary diseases [300 – 303]. 14. Society of Chemical Industry: Diaphragm
No indications of carcinogenicity or muta- Cells for Chlorine Production, London 1977.
genicity of chlorine have been detected in an- 15. J. J. McKetta, W. A. Cunningham:
imal experiments or encountered in industrial Encyclopedia of Chemical Processing and
Design, vol. 7, Marcel Dekker, New
medicine [304]. In Germany, the MAK value
York-Basel 1977.
is 0.5 mL/m3 (1.5 mg/m3 ) with a 15-min ex-
16. R. Dandres: Le Chlore, 3rd ed., Institut
cursion factor of 1 [304]; in the USA, the national de recherche et de sécurité pour la
TLV is 0.5 mL/m3 (1.5 mg/m3 ) with a STEL of prévention des accidents du travail et des
1.0 mL/m3 (3 mg/m3 ). maladies professionelles, Paris 1978.
17. M. O. Coulter: Modern Chlor-Alkali
Technology, vol. 1, Ellis Horwood, London
17. Acknowledgement 1980.
18. C. Jackson: Modern Chlor Alkali Technology,
The entire topic was coordinated by Peter vol. 2, Ellis Horwood, Chichester, Great
Schmittinger Britain, 1983.
19. K. Wall, Modern Chlor-Alkali Technology,
vol. 3, Ellis Horwood, Chichester, Great
18. References Britain, 1986.
20. N. M. Prout, Modern Chlor-Alkali Technology,
General References vol. 4, Elsevier Appl. Science, London 1990.
21. T. C. Wellington, Modern Chlor-Alkali
1. K. J. Vetter: Elektrochemische Kinetik, Technology, vol. 5, Elsevier Appl. Science,
Springer Verlag, Berlin-Heidelberg 1961. Barking 1992.
2. J. S. Sconce: Chlorine, Its Manufacture, 22. M. G. Beal in R. W. Curry (ed.): Modern
Properties and Uses, Reinhold Publ. Co., New Chlor-Alkali Technology, vol. 6, SCJ, London
York 1962. 1995.
3. Kirk-Othmer, vol. 1, 1978, pp. 799 – 883. 23. S. Sealey: Modern Chlor-Alkali Technology,
4. P. Gallone, G. Milazzo: Electrochemistry, vol. 7, The Graham House, Cambridge 1998.
Elsevier, Amsterdam-London-New York 1963. 24. The Chlorine Institute: Pamphlet 1 Chlorine
5. F. Matthes, G. Wehner: Manual, Ed. 6, 1997.
Anorganisch-Technische Verfahren für die 25. Int. Soc. Electrochem.: Extended Abstracts,
Grundstoffindustrie, VEB Deutscher Verlag 23rd Meeting, Stockholm 1972.
für Grundstoffindustrie, Leipzig 1964. 26. The Electrochemical Society: Extended
6. Gmelin, system no. 6, Chlor (1927). Abstracts of Industrial Electrolytic Division,
7. Winnacker – Küchler, 3rd ed., vol. 1, The Chlorine Institute, Spring Meeting, San
pp. 228 – 239; Francisco, May 12nd – 17th, 1974.
Winnacker – Küchler, 4th ed., vol. 2, 27. Int. Soc. of Electrochemistry: Extended
pp. 379 – 480. Abstracts, 27th Meeting, Zürich 1976.
8. R. Powell: Chlorine and Caustic Soda 28. Technischer Arbeitskreis Elektrolyse:
Manufacture, Noyes Data Corp., Park Ridge, Hinweise für den Umgang mit Chlor, Verlag
N.J. ,1971. Chemie, Weinheim 1978.
9. R. G. Smith: Chlorine; An Annotated 29. The Electrochemical Society: Extended
Bibliography, The Chlorine Institute, New Abstracts, Spring Meeting Seattle, Princeton,
York 1971. N.J., May 1978.
10. A. Kuhn: Industrial Electrochemical 30. AIChE: Sympos. Ser. 75, no. 185, 1979.
Processes, Elsevier, Amsterdam 1971. 31. Oronzio de Nora Symposium: Chlorine
11. P. Gallone: Trattato di ingegneria
Technology, Venezia, May 1979.
elettrochimica, Tamburine Editore, Milano
32. AIChE: Sympos. Ser. 77, no. 204, 1981.
1973.
90 Chlorine

33. J. S. Robinson: Chlorine Production 57. Bayer, DE 3 216 418-A1, 1983 (R. Schäfer).
Processes, Noyes Data Corp., Park Ridge, 58. Bayer, DE 2 709 728-C2, 1981 (R. Schäfer).
N.J., ,1981. 59. K. Maycock et al., SCI Chlorine Symposium,
34. Joint Chlorine Institute: BITC Meeting, London, June 1997.
London, June 1982, The Chlorine Institute, 60. Hoechst, DE 2 905 125-A1, 1980 (M. Schott,
New York. J. Russow).
35. Elektrochemische Verfahrenstechnik, 61. Z. I. Lifatova et al., SU 865 800-A1, 1981.
Dechema Monographie vol. 97, Verlag 62. Bayer, DE 1 177 122-B, 1964 (E. Zirngiebl).
Chemie, Weinheim 1984. 63. Ebara Infilco, JP 54–81 174-A2, 1979 (M.
36. Technische Elektrolysen, Dechema Kanekawa, K. Akagi).
Monographie vol. 98, Verlag Chemie, 64. Processing, June 1978.
Weinheim 1985. 65. W. Glas, Chem. Ing. Tech. 43 (1973) 832.
66. F. Lüns, BBC-Nachrichten, 1/2 (1974) 36 – 42.
Specific References 67. H. A. Horst, Chem. Ing. Tech. 43 (1971) 164.
68. Siemens, DE 3 123 877-C1, 1983 (G.
37. C. W. Scheele, Sv. Akad. Handl. 35 (1774) 88. Achilles).
38. J. Eidem, L. Lunevall in [26] Hundt & Weber Schaltgeräte, EP 11 820-A1,
39. R. E. Means, T. R. Beck, Chem. Eng. (N.Y.) 91 1980 (E. Stratmann).
(1984) no. 22, 46 – 51. 69. F. Lappe, Chem. Ing. Tech. 42 (1970) 1228.
40. The Chlorine Institute: Properties of Chlorine 70. Fiat Final Report 732, 797, 816, 834.
in SI Units, Pamphlet 72, 2nd ed., July 1986. 71. W. Kramer in [36]
41. J. J. Martin, D. M. Longpre, J. Chem. Eng. 72. T. Sugino, K. Aoki, J. Electrochem. Soc. Jpn.
Data 29 (1984) 466 – 473. 27 (1959) no. 1 – 3, E 17.
42. L. Ziegler, Chem. Ing. Techn. 22 (1950) 229. 73. G. Hauck, W. Dürr, Chem. Ing. Tech. 39
43. E. M. Aleta, P. V. Roberts, J. Chem. Eng. Data (1967) 720.
31 (1986) 51 – 54. H. J. Antweiler, J. P. Schäfer, Chem. Ing. Tech.
44. I. A. A. Ketalaar, Electrochem. Technol. 5 43 (1971) 180.
(1967) 143. 74. D. S. Nielson, O. Platz, Reliabil. Eng. 4 (1983)
45. R. F. Schwab, N. H. Doyle, Electrochem. 1 – 18.
Technol. 5 (1967) 228. 75. PPG Ind., US 4 155 819-A, 1979 (W. W.
Y. E. Frolov, A. S. Maltsera, A. N. Baratov, A. Carlin).
I. Roslovskii, Khim. Promst (Moscow) 7 76. Diamond Shamrock, US 4 073 706-A, 1978
(1977) 530 –532. (Z. Nagy).
46. The Chlorine Institute: Explosive Properties of 77. H. Tsukuda et al., J. Electrochem. Soc. Jpn. 32
Gaseous Mixtures Containing Hydrogen and (1964) 1.
Chlorine, Pamphlet 121, 2nd ed., June 1992. 78. D. W. Wabner, P. Schmittinger, F. Hindelang,
47. IUPAC, White Book on Chlorine 68 (1996) 9. R. Huß, Chem. Ing. Tech. 49 (1977) 351.
48. V. M. Sedivy, Industrial Minerals, 343 (1996) 79. L. J. J. Janssen in [18], p. 271.
April, 73. 80. Society of Chemical Industry, Electrochemical
49. Hoechst, EP 35 695-B1, 1983. Technology Group: Intern. Meeting on
50. Solvay & Cie., EP 30 756-A1, 1981. Electrolytic Bubbles, 13 – 14 Sept. 1984,
51. A. I. Postoronko et al., SU 537 027-A, 1976. London.
Jamestown Chem.: Calcium Sulphate 81. BASF, US 3 502 561, 1970 (W. Rasche, E.
Inhibitor. Wygasch, G. Csizi).
52. S. L. Fuks, T. M. Orchinnikova, Zh. Prikl. 82. A. Zimmer, L. Franke, Chem. Tech. (Leipzig)
Khim (Leningrad) 56 (1983) no. 8, 36 (1984) no. 2, 64 – 69.
1757 – 1761. 83. H. Hund, Chem. Ing. Tech. 39 (1967) 702.
53. K. Harada, Cleveland Meeting: The 84. W. Vielstich, Chem. Ing. Tech. 34 (1962) 346.
Electrochem. Soc., May 1966. Bayer, DE 1 792 588-C3, 1977 (E. Zirngiebl).
54. Olin, US 4 277 447-A, 1981 (R. F. Chambers, Murgatroyd’s Salt & Chem., DE 1 915 765,
N. Rachima). 1969 (M. J. Lockett).
55. Toa Gosei Chem. Ind., JP 57–106 520-A2, 85. M. M. Jaksic, I. M. Csonka, Electrochem.
1982 (K. Suzuki, R. Nagai, H. Goto). Technol. 4 (1966) 49.
56. Hoechst, DE 3 037 818-A1, 1982 (S. 86. G. Barthel, Chem. Anlagen + Verfahren 8
Benninger). (1973) 59 – 60.
Chlorine 91

87. P. Theissing et al., Chem. Ing. Tech. 51 (1979) 110. A. A, Lindley, An Economic and
237. Environmental Analysis of the Chlor-Alkali
F. Hine: Topics in Pure and Applied Production Process, prepared for the European
Electrochemistry, SAEST, Karaikudi, India. Commision (DG III, Chemicals), Final Report,
88. Uhde: Alkalichloridelektrolyse nach dem June 30, 1997.
Quecksilberverfahren. 111. Euro Chlor, GEST 97/228, The Eurochlor
89. Olin, US 4 004 989-A, 1977 (R. W. Ralston). Program, Fourth Techn. Seminar, Proceedings,
90. G. Hauck, Chem. Ing. Tech. 34 (1962) 369. Feb. 1997.
91. Euro Chlor: Control of Worker Exposure to 112. ELTECH Systems Corporation Technical
Mercury in the Chlor-Alkali Industry, 4th ed., Service Department: Diaphragm Electrolytic
October 1997. Chlorine Cell, Technical Data Manual,
92. The Chlorine Institute: Pamphlet 125, ELTECH Systems Corporation, Chardon,
Guidelines: Medical Surveillance and Hygiene Ohio 1982.
Monitoring Practices for Control of Workers 113. ELTECH Systems Corporation Technical
Exposure to Mercury in the Chlor-Alkali Service Department: Diaphragm Chlorine
Industry, 3rd ed., December 1994. Cell, Operating Manual, ELTECH Systems
93. Environmental Protection Agency: Control Corporation, Chardon, Ohio 1983.
Techniques for Mercury Emissions from 114. ELTECH Systems Corporation, US 4 410 411,
Extraction and Chlor-Alkali Plants, Publ. No. 1983 (R. W. Fenn, III, E. J. Pless, R. L. Harris,
AP-118, Washington, DC. K. J. O’Leary).
94. A. Gellera, L. Cavalli, G. Nucci, Analyst 115. ELTECH Systems Corporation, US 3 674 676,
(London) 109 (1984) 1537 – 1539. 1972 (Edward I. Fogelman).
95. H. Klotz, Chem. Ing. Tech. 52 (1980) 444. 116. L. C. Curlin, T. F. Florkiewicz, R. C.
96. Akzo: Anode Current Control by Computer Matousek: “Polyramix, A Depositable
ACCS-System, 1972. Replacement for Asbestos Diaphragms”,
97. A. I. Vasilev et al., SU 1 024 528-A1, 1983. London International Chlorine Symposium,
98. Bayer, DE 1 767 840-B2, 1975 (R. Schlee, W. 1988.
Büsing). 117. R. N. Beaver: The Dow Diaphragm Cell, Dow
Hoechst, DE 3 244 033-A1, 1984 (K. Chemical, Freeport, Texas 1985.
Lehmann, H. Valentin, K. Gnann). 118. Dow Chemical Company, US 4 497 112.
99. Montedison, EP 93 452-A2, 1983 (F. Lo Vullo, 119. Oronzio De Nora Impianti Elettrochimici
E. Malvezzi, P. Balboni). S.p.A.: Design Features of Glanor Diaphragm
100. Solvay & Cie., EP 85 999-A1, 1983 (J. P. Electrolyzer, Oronzio DeNora Impianti
Detournay, J. Defourny). Elettrochimici S.p.A., Milano.
101. Heraeus Elektroden, EP 68 076-A2, 1983 (P. 120. PPG Industries: PPG’s Bipolar Diaphragm
Fabian). Electrolyzers, PPG Industries, Inc., Pittsburgh,
102. H. Becker, Dechema Monographie, vol. 61, Pa.
Verlag Chemie, Weinheim 1968, p. 59. 121. PPG Industries: Licensing From PPG
103. BITC Working Group Mercury Analysis, Chemicals, PPG Industries, Inc., Pittsburgh,
Anal. Chim. Acta 72 (1974) 37 – 48; Pa.
84 (1976) 231 – 257; 122. Occidental Chemical Corporation: Hooker
87 (1976) 273 – 281. Chlor-Alkali Systems, A Record of
104. AKZO: Das IMAC-TMR-Verfahren. Achievement in Diaphragm Cell Technology,
105. EEC Directive 82/176, 22. 3. 1982. Occidental Chemical Corporation, Niagara
106. Euro Chlor Publ. Env. Prot. 11, Mercury Falls, N.Y.
Housekeeping, 4th ed., September 1998. 123. Uhde GmbH: Alkaline Chloride Elektrolysis
107. Euro Chlor: Good Mercury Housekeeping by the Diaphragm Process; System Hooker,
Third Techn. Seminar, Proceedings, December Uhde GmbH, Dortmund.
1993. 124. ELTECH Systems Corporation: ELTECH
108. Parcom Decision 90/3, Reducing Atmospheric Modified Diaphragm Cells, ELTECH Systems
Emissions from Existing Chlor-Alkali Plants, Corporation, Chardon, Ohio 1985.
June 14, 1990. 125. ELTECH Systems Corporation, US 3 591 483,
109. Euro Chlor: Chlorine Industry Review US 3 707 454, 1971 (Richard E. Loftfield,
1996 – 1997, p. 10. Henry W. Laub).
92 Chlorine

126. ELTECH Systems Corporation, US 3 928 166, 145. Du Pont, US 4 021 327, 1977 (W. G. Grot).
1975 (Kevin J. O’Leary, Charles P. Tomba, 146. Asahi Glass, JP 59/40 231, 1984 (Y. Oda, T.
Robert W. Fenn, III). Morimoto, K. Suzuki).
127. ELTECH Systems Corporation: Caustic 147. The Dow Chemical Company: Membrane
Purification System 1985. Chlor-Alkali Process, Texas Applied Science
128. G. A. Carlson, E. E. Estep, “Porous Cathode and Technology Laboratories, Freeport, Texas,
Cell for Metals Removal from Aqueous 1993 (H. S. Burney).
Solutions”, paper presented at the 148. D. Bergner, Chem.-Ztg. 107 (1983) 281.
Electrochemical Society Meeting, May, 1972. 149. H. B. Beer, GB 1 147 442, 1965.
129. W. G. Grot, Chem. Ing. Tech. 44 (1972) 167. H. B. Beer, GB 1 195 871, 1967.
Diamond Shamrock, US 4 025 405, 1977 (R. 150. V. de Nora, J.-W. Kühn von Burgsdorff, Chem.
L. Dotson, K. J. O’Leary). Ing. Tech. 47 (1975) 125 – 128.
130. Asahi Glass, US 4 065 366, 1977 (Y. Oda, M. Electronor Corporation, US 3 616 445, 1967
Suhara, E. Endoh). (G. Bianchi, V. de Nora, P. Gallone, A.
131. H. Ukihashi, M. Yamabe, ACS Polymer Nidola).
Division Workshop on Perfluorinated Ionomer 151. S. Trasatti, G. Lodi in S. Trasatti (ed.):
Membranes, Lake Buena Vista, Florida, Feb. Electrodes of Conductive Metallic Oxides, part
23 – 26, 1982. A, Elsevier Scientific Publishing Company,
132. World Chlor-alkali Analysis, Vol. 1, CMAI, Amsterdam-Oxford-New York 1980,
Texas, USA May 1997. pp. 301 – 358.
133. T. A. Davis, J. D. Genders, D. Pletcher, Ion 152. S. Trasatti, W. E. O’Grady in H. Gerischer, C.
permeable membrane, Alresford 1997. W. Tobias (eds.): Advances in
134. Asahi Chemical Industry, US 4 178 218, 1979 Electrochemistry and Electrochemical Eng.,
(M. Seko). vol. 12, J. Wiley & Sons, New York-Chichester
135. K. Schneiders, J. Schulze: Improvement in the 1981, pp. 177 – 261.
Hoechst-Uhde Membrane Cell, Modern 153. GB 1 402 414, 1971 (N. W. J. Pumphrey, B.
Chlor-Alkali-Technology, Vol. 6, Hesketh).
1995. GB 1 484 015, 1973 (B. Hesketh, C. Pownall,
136. Akin, Optimierung einer N. W. J. Pumphrey).
Chloralkalielectrolyse-Membranzellen- 154. C. Conradty, DE-OS 1 813 944, 1968 (G.
Kaskade unter Berücksichtigung ihres Thiele, D. Zöllner, K. Koziol);
Energieverbrauchs, Dissertation, Dortmund C. Conradty, DE 2 255 690, 1972 (K. Koziol,
1984. K.-H. Sieberer, H.-C. Rathjen).
137. Tokuyama Soda, US 4 190 514, 1980 (S. 155. C. Modes, W. C. Heraeus: abstract and poster
Matsuura, T. Oku, M. Kuramatani, N. during the 13th Int. Congress and General
Kramoto, Y. Ozaki). Assembly, Int. Union of Crystallography,
138. PPG, US 4 323 595, 1982 (C. N. Welth, J. O. Hamburg, August 1984.
Snodgrass). 156. Yu. E. Roginskaya, O. V. Morozova,
139. H. S. Burney: Chlor-Alkali Process, in Ralph. Electrochimica Acta 40 (1995) 817 – 822.
E. White et al. (eds.), Modern Aspects of 157. D. V. Kokoulina, T. V. Ivanova, Y. I.
Electrochemistry, no. 24, Plenum Press, New Krasavitskaya, Z. I. Kudryavtseva et al.,
York, 1993. Elektrokhimiya 13 (1977) no. 10, 1511 – 1515.
140. Du Pont: Recommended Brine Impurity 158. J. E. Currey in Encycl. of Chem. Proc. and
Limits for Chloralkali Plants, Aug. 1991. Design, vol. 7, Marcel Dekker, New York
141. Bayer AG: The Importance of Electrolyte 1978, pp. 305 –450.
Quality to Membrane Chloralkali Cells, Uhde 159. D. Bergner, S. Kotowski, J. Appl. Electrochem.
Chlorine Symposium 1995 (F. R. Minz). 13 (1983) 341 – 350.
142. Du Pont: Effect of Impurities on Membrane 160. Y. M. Kolotyrkin, Denki Kagaku 47 (1979)
for Chloralkali Production, Wilmington, no. 7, 390 – 400.
Delaware Aug. 1989 (James T. Keating). 161. D. W. Wabner, P. Schmittinger, F. Hindelang,
143. Rolf Rodermund: Chlorat-Reduktion im R. Huß, Chem. Ing. Tech. 49 (1977) 351.
Solekreislauf einer Chlor-Alkali-Electrolyse, 162. B. Busse, R. Scannell, Electrochemische
Dissertation, Dortmund 1983. Verfahrenstechnik-Energietechnik,
144. Asahi Glass, US 4 255 523, 1981 (H. Ukihashi, Stoffgewinnung und Bioelektrochemie, Mono.
T. Asawa, T. Gunjima). GdCh. Tagung Monheim 1996.
Chlorine 93

163. F. Beck, Electrochimica Acta 34 (1989) 186. Th. F. Florkiewicz, The Future of Diaphragm
811 – 822. Cell Technology, Chlorine Institute Spring
164. Oronzio De Nora Impianti Elettrochimici Meeting, St. Louis, Missouri, March 1998.
S.p.A., GB 2 051 131, 1980 (A. Pellegri). 187. M. Esayian, J. H. Austin: “Membrane
165. A. Nidola in S. Trasatti (ed.): Electrode of Technology for Existing Chloralkali Plants”,
Conductive Metallic Oxides, part B, Scientific E. I. du Pont de Nemours Co., Wilmington,
Publishing Company, Del., 1984.
Amsterdam-Oxford-New York 1981, 188. VCI Verband der Chemischen Industrie (ed.):
pp. 627 – 659. Zukunft der Alkalichlorid-Electrolyse-Anlagen
166. C. Traini, Proc. of O. De Nora Symposium, nach dem Amalgamverfahren, Frankfurt a. M.
15 – 18 May 1979. 1998.
167. Diamond Shamrock Corp., US 3 674 676, 189. M. R. Marcan: “Mercury/Membrane Cells –
1970 (E. I. Fogelman). when to Change”, Modern Chlor-Alkali
168. V. de Nora, Chem. Ing. Tech. 47 (1975) no. 4, Technology, vol. 5, Elsevier Publ., London,
141. 1992, p. 177.
169. Heraeus Elektrochemie, EP 0 687 312, 1994 190. R. W. Curry: “Issues involved in membrane
(B. Busse, R. Scannell). conversion of a chlor-alkali business”, Modern
170. S. Kotowski, B. Busse, Chlor-Alkali Technology, vol. 5, Elsevier
“Anodenbeschichtungen in Membranzellen zur Publ., London 1992, p. 185.
Chlor- und Laugeproduktion”, paper presented 191. S. Higuchi, Soda to Enso 35 (1984) 383.
at the Annual DECHEMA Meeting 1984. 192. H. Isfort, W. Stockmans in [36]
171. Uhde GmbH, DE 3 219 704, 1982 (H. Schmitt, 193. K. Kerger, Chem. Ing. Tech. 43 (1971) 167.
H. Schurig, W. Strewe). 194. Uhde: Chlor und Wasserstoff aus Salzsäure
172. N. P. Fedot’ev, N. V. Berezina, E. G. Kruglova, durch Elektrolyse.
Zh. Prikl. Khim. (Leningrad) 21 (1948) 195. General Electric, DE 2 856 882-A1, 1979 (R.
317 – 328. M. Dempsey, A. B. Conti).
173. Toa Gosei Chemical Industry Co. Ltd., JP 196. H. Klotz, P. Orth in [36]
31–6611, 1956 (K. Sasaki). 197. The M. W. Kellogg Co.: Hydrogen Chloride to
174. E. I. DuPont de Nemours & Co., US 4 116 804, Chlorine, The Kel-Chlor Process, presentation
1978 (C. R. S. Needes). at the 20th Chlorine Plant Managers Seminar,
175. Diamond Shamrock Corp., US 4 024 044, New Orleans, Feb. 9 1977.
1977 (J. R. Brannan, I. Malkin). 198. Krupp Uhde: “Chlorine and Hydrogen from
176. F. Hine, M. Yasuda, M. Watanabe: “Studies of Acid by Electrolysis”, company brochure,
the Nickel – Sulfur Electrodeposited Cathode”, 1993.
Denki Kagaku 47 (1979) no. 7, 401 – 408. 199. Chemical Week 1996, 14. Aug., 26.
177. Asahi Chemical, EP-A 0 031 948, 1981 (M. M. Mortensen et al., Chem. Eng. Sci. 51
Yoshida, H. Shiroki). (1996) 10, 2031.
178. Chlorine Engineers Corp. Ltd., US 4 465 580, 200. The M. W. Kellogg Co., Hydrocarbon
1984 (K. Kasuya). Process. 60 (1981) no. 11, 143.
179. Tokuyama Soda, US 4 190 516, 1980 (Y. 201. Shell, BE 599 241-A, 1961 (W. S. Engel, F.
Kajimaya, T. Kojima, Y. Murakami, S. Wattimena).
Matsura). 202. Mitsui Chemicals Inc.: “The Mitsui MT-Chlor
180. Dow Chemical, EP-A 0 129 734, 1985 (N. R. Process, Catalytic Oxidation of Hydrogen
Beaver, L. E. Alexander, C. E. Byrd). Chlordie”, company brochure.
181. ICI, EP-A 0 129 374, 1984 (J. F. Cairns, D. A.
203. M. Mortensen et al., Chem. Engng. Scie. 51
Denton, P. A. Izard).
(1996) 2031.
182. Diamond Shamrock Corp., US 4 104 133,
204. Institut Français du Pétrole, FR 1 294 706-A,
1978 (J. R. Brannan, I. Malkin, C. M. Brown).
1962 (P. Bedagne et al.).
183. Kanegafuchi Kagaku Kogyo Kabushiki
205. C. P. van Dijk, Chem. Econ. Eng. Rev. 4 (1972)
Kaisha, EP-A 0 132 816, 1985 (Y. Samejima,
no. 12, 42.
M. Shiga, T. Kano, T. Kishi).
206. H. Schmidt, F. Holzinger, Chem. Ing. Tech. 35
184. Olin Corp., US 4 169 775, 1979 (H. C. Kuo).
(1963) 37.
185. Harrimann Chemsult: Chlor-Alkali Prospects
207. Monsanto Enviro-Chem. Systems: Fiber Bed
for Chlorine, Caustic Soda and Soda Ash in
Mist Eliminator, 1979.
the 1990’s, London 1991.
94 Chlorine

208. Dow Chemical, US 3 568 409-A, 1971 (B. 229. D. P. Meinhardt: “Chlorine”, Chem. Eng.
Mac Ferguson, J. F. Gilbert, D. N. Glew). (1981) Oct. 5, 125 – 133.
209. CJ 134 “The Drying and Liquefication of 230. N. C. Harris, J. P. Shaw, European Chlorine
Chlorine and the Phase Diagram Cl2 – H2 O”, Storage Practice, paper presented at the
April 1967. Chlorine Institute’s 23rd Plant Manager
210. Koninkl. Nederlandsche Zoutindustrie, US Seminar, 1980.
3 534 562-A, 1970 (D. Meyer, T. Thijssen). R. Papp, Chlorine ’74, Chemistry and
211. ICI, DE 1 946 096-C3, 1973 (K. J. Howliston). Industry, 15 March (1975) 243 – 246.
212. G. Alberti, DECHEMA Monogr. 66 (1971) 231. Euro Chlor Techn. Publ. 75/46, Pneumatically
1193 – 1211. Operated Valves for Use on Rail and Road
R. Gabbioneta: Garo. Tankers, and ISO-Containers for Liquid
Nash International: HC-7 Chlorine Chlorine, 8th ed., November 1996.
Compressor. 232. The Chlorine Institute, Pamphlet 93,
213. Sulzer Burckhardt, Prospectus No. 1393 a. Pneumatically Operated Valves for Use on
214. Sundstrand: Sundyne Zentrifugalverdichter, Tank Cars, Ed. 1, September 1996.
1980. 233. Winnacker-Küchler: 4th ed., pp. 428 – 429.
215. Euro Chlor Techn. Publ. 79/79 Chlorine 234. Euro Chlor Techn. Publ. 72/10, Pressure
Transfer Compressors, 3rd ed. February 1999. Storage of Liquid Chlorine, 8th ed., June 1993.
216. Euro Chlor Techn. Publ. 83/119 Canned 235. The Chlorine Institute, Pamphlet 5,
Pumps for Use with liquid Chlorine, 3 rd ed. Non-refrigerated Liquid Chlorine Storage, Ed.
September 1987. 5,
217. H. Abel, F. Özvegyi, Kälte-Klima-Rundsch. 7 October 1992.
(1969) 100. 236. Euro Chlor Techn. Publ. 73/17, Low Pressure
218. Inf. Chim. 98 (1971) 55. Storage of Liquid Chlorine, 5th ed., February
219. H. Hagemann, Chem. Ing. Tech. 39 (1967) 744. 1996.
220. R. F. Schwab, W. H. Doyle, Electrochem. 237. The Chlorine Institute, Pamphlet 78,
Technol. 5 (1967) 228. Refrigerated Liquid Chlorine Storage, Ed. 2,
221. B. Liggenstorfer: “Kälteanlage zum June 1994.
Verflüssigen von Chlor”, Tech. Rundsch. 238. ISO 3511/1.
Sulzer 63 (1981) no. 2, 53 – 56. 239. A. F. Timofeer et al.: “Steps to Prevent
222. AKZO: Chlorine Liquefaction by Absorption Failures of Chlorine Containers”, Khim.
and Distillation, 1974. Promst (Moscow) 10 (1978) no. 3, 202 – 205;
223. EC Position Paper PP VII: “Use of Sov. Chem. Ind. (Engl. Transl.) 10 (1978)
Carbontetrachloride in Chlorine Production”, no. 3, 222 – 225.
Feb. 1995. 240. Euro Chlor Techn. Publ. 79/81, Dry Liquid
224. T. A. Liederbach, Chem. Eng. Prog. 70 (1974) Chlorine Piping Systems Located inside
64. Producers’ or Consumers’ Plants, 7th ed.,
225. Euro Chlor, Ave E Van Nieuwenhuyse 4, Box September 1986.
2, B-1160 Brussels, Fax: +32 2 6767241, 241. The Chlorine Institute, Pamphlet 60, Chlorine
e-mail: eurochlor@cefic.be: Publications. Pipelines, Ed. 4, May 1997.
226. Technical and Safety Information, catalog 242. The Chlorine Institute, Pamphlet 9, Chlorine
1997/98, The Chlorine Institute, Inc., 2001 Vaporizing Equipment, Ed. 5, April 1997.
“L” Street, N.W. Suite 506, Washington D.C. 243. Euro Chlor, Techn. Publ. 75/47, Chlorine
20036, Fax (202) 223–7225. Vaporizers, 8th ed., January 1980.
227. R. Dandres: Le Chlore, 3rd ed., L’Institut 244. Armstrong Chemtec, Form 54 – 2/97,
National de Recherche et de Sécurite pour la Vaporizers.
Prevention des Accidents du Travail et de 245. Euro Chlor Techn. Publ. 76/55, Acceptable
Maladies Professionelles, Paris 1978. Levels of Nitrogen Trichloride in Liquid
H. Dorias: Gefährliche Güter, Springer Verlag,
Chlorine, 9th ed., September 1990.
Berlin-Heidelberg-New York-Tokyo 1984.
246. The Chlorine Institute, Pamphlet 21, Nitrogen
228. G. Payne: “Safe Handling and Storage of
Trichloride – A Collection of Reports and
Liquefied Gases, in Particular Chlorine and
Papers, Ed. 4, 1997.
Chlorofluorohydrocarbon Refrigerants”,
247. The Chlorine Institute, Pamphlet 89, Chlorine
Trans. Inst. Chem. Eng. 42 (1964) no. 4,
Scrubbing Systems, Ed. 1, December 1991.
92 – 99.
Chlorine 95

248. Euro Chlor Techn. Publ. 76/52, Equipment for 263. The Chlorine Institute, Instruction Booklets,
the Treatment of Gaseous Effluents Containing Chlorine Institute Emergency Kit “A” for
Chorine, 9th ed., October 1986. 100-lb. and 150-lb. Chlorine Cylinders, Kit
249. M. E. Fakley et al., The ICI “B” for Chlorine Ton Containers, Kit “C” for
HYDECAT-Process in [21], p. 23. Chlorine Tank Cars and Tank Tucks, Ed. 8,
250. Euro Chlor Techn. Publ. 79/82 Choice of 1996.
Materials of Construction for Use in Contact 264. Euro Chlor Chlorine Safety Manual 90/159,
with Chlorine, 7th ed., October 1995. Aspects of Transport Safety, November
251. W. A. Statesir: “Explosive Reactivity of 14 – 15, 1990.
Organics and Chlorine”, Chem. Eng. Prog. 69 265. Euro Chlor Techn. Publ. 92/176, Chlorine
(1973) no. 4, 52 – 54. Emergency Equipment, 1st ed. September
252. Euro Chlor Third Technical Seminar 1993.
Proceedings, Techn. Publ. 93/192 How to Use 266. Euro Chlor Techn. Publ. 93/191, 93/193,
Steel and Titanium Safety, December 1993. 93/194, “Safe Design”, “Safe Constructions”,
253. The Chlorine Institute, Pamphlet 95, Gaskets “Safe Operations” in Eurochlor Third
for Chlorine Service, Ed. 1, October 1993. Technical Seminar Proceedings, December
254. T. B. Meslin: “Assessment and Management of 2 – 3, 1993.
Risk in the Transport of Dangerous Materials: 267. The Chlorine Institute, Chlorep: How to Get
The Case of Chlorine Transport in France”, Help in Chlorine Emergencies.
Risk Anal. 1 (1981) no. 2, 137 – 142. 268. British Standard Institution: Specification for
G. B. Frame: “Determination of Risk from a Liquid Chlorine, BS 3947, 1965.
Chlorine Spill or Major Release”, Chem. Can. 269. ISO 1552: Liquid Chlorine for Industrial Use
32 (1980) no. 8, 27 – 29. — Method of Sampling.
255. C. L. Melancon, Dow Chemicals Emergency 270. The Chlorine Institute, Pamphlet 77, Sampling
Response System, paper presented at The Liquid Chlorine, Ed. 3,
Chlorine Institute’s 22nd Plant Manager’s April 1997.
Seminar 1979. 271. ISO 2121: Liquid Chlorine for Industrial Use
R. Sklarew et al., Emergency System for Toxic — Determination of the Water Content —
Chemical Releases, Pollution Engineering, Gravimetric Method.
July 1982. 272. ASTM – E410 Standard Method of Test for
256. J. M. Buchlin, Aerodynamic Behaviour of Moisture and Residue in Liquid Chlorine,
Liquid Spray-Design Method of Water Spray 1986.
Curtain, von Karman Institute, Rhode Saint 273. ISO 2202: Liquid Chlorine for Industrial Use
Genese, Belgium 1980, Test Report No. 171. — Determination of the Water Content —
257. The Chlorine Institute, Pamphlet 63, First Aid using an Electrolytic Analyzer.
and Management of Chlorine Exposures, Ed. 274. ISO 2120: Liquid Chlorine for Industrial Use
5, May 1993. — Determination of the Content of Chlorine
258. The Chlorine Institute, Pamphlet 64, by Volume in the Vaporized Product.
Emergency Response Plans for Chlorine 275. ASTM – E412 Standard Method of Assay of
Facilities, Ed. 4, November 1995. Liquid Chlorine, 1986.
259. The Chlorine Institute, Pamphlet 65, Personal 276. DIN 38 408 Teil 4: Gasförmige Bestandteile
Protective Equipment for Chlorine and (Gruppe G) Bestimmung von freiem und
Sodium Hydoxide, Ed. 3, November 1995. Gesamtchlor (G 4).
260. The Chlorine Institute, Pamphlet 85, 277. Anal. Chim. Acta, 156 (1984) 221 – 233.
Recommendations for Prevention of Personnel 278. Anal. Chim. Acta, 87 (1976) 273 – 281.
Injuries for Chlorine Producer and User 279. ASTM – E506 Standard Test Method for
Facilities, Ed. 3, November 1994. Mercury in Liquid Chlorine, 1984.
261. The Chlorine Institute, Pamphlet 86, 280. Deutsche Einheitsverfahren zur
Recommendations for Chlor-Alkali Wasseruntersuchung, Verlag Chemie,
Manufacturing Facilities for the Prevention of Weinheim 1971.
Chlorine Releases, Ed. 3, November 1994. 281. M. L. Langhorst, Am. Ind. Hyg. Assoc. J. 43
262. The Chlorine Institute, Pamphlet 74, (1982) 347 – 360.
Estimating the Area Affected by a Chlorine The Chlorine Institute: Atmospheric
Release, Ed. 3, in revision. Monitoring Equipment for Chlorine, 2nd ed.,
96 Chlorine

March 1982, CI Member Information Report Herstellung von Chlor, Natronlauge und
138. Chlorderivaten, 4th Technical Seminar,
W. König et al., Chem. Ind. (Feb. 1984) 69. Leipzig, 1997.
282. World Health Organisation, 296. P. Davies, I. Hymes: “Chlorine toxicity criteria
“Recommendations for Drinking Water for hazard assessment”, Chem. Eng., June
Quality”, 2nd ed., Vol.1, December 1993. 1985, 30 – 33.
W. König et al., Chem. Ind. 36 (Feb. 1984) 69. 297. H. H. Rotman, M. J. Fliegelman, T. Moore, R.
283. Mary. C. Blackburn, Chemical Market G. Smith, D. M. Anglen, C. J. Kowalski, J. G.
Associates, Inc.: “Chlorine and Caustic – Weg: “Effects of low concentrations of
which Way is up”, Krupp Uhde Symposium, chlorine on pulmonary function in humans”, J.
Dortmund, May 6 – 8, 1998. Appl. Physiol.: Respir. Environ. Excercise
284. W. König, K. Krinke, H. Kron et al., Chem. Physiol. 54 (1983) no. 4, 1120 – 1124.
Ind. 36 (Feb. 1984) 69. 298. R. P. F. Schins, H. Emmen, L. Hoogendijk, P.
285. Euro Chlor: Chlorine Industry Review J. A. Borm: “Nasal Inflammatory and
1994 – 1995. Respiratory Parameters in Human Volunteers
286. Chem. Engng. News, 1994, Nov. 21, 17. during and after Repeated Exposure to
287. Euro Chlor, Proceedings of the Technical Chlorine”, Eur. Respir. J. 16 (2000) 626 – 632.
Seminar Leipzig “Care in Practise”, Germany 299. “Chronic Inhalation Toxicity Study on
1997. Chlorine in Non-Human Primates”, CITT
288. Euro Chlor: Chlorine flow in Europe 1995, Activities Chemical Institute of Toxicology 4
Jan. 1997. (1984) no. 8, 1 – 3.
289. Euro Chlor: Chlorine Industry review 300. B. Grenquist-Norden: “Respiratory effects of
1994 – 1995. industrial chlorine and chlorine dioxide
290. Harriman Chemsult, Chlor-Alkali Prospects exposure”, Institute of Occupational Health,
for Chlorine, Caustic Soda and Soda Ash in University of Helsinki, Finland, 1983.
the 1990’s, Autumn 1991, London, England. 301. F. Schuckmann (Occupational Health
291. B. Hileman et al., Chem. & Eng. News, 21st Department of Hoechst AG): “Pulmonary
November 1994, p. 12. Function Study on Workers with Long-Term
292. The Chlorine Institute, Pamphlet 10, North Exposure to Chlorine”, in: Medichem
American Chlor-Alkali Industry Plants and proceedings, XI. International Congress,
Production Data Book, Calgary, Alberta, Canada, 26. – 29. 9. 1983,
April 1997. pp. 475 –483.
The Chlorine Institute, Pamphlet 16, 302. B. G. Ferris, Jr., S. Puleo, H. Y. Chen:
Chlor-Alkali Producers Outside North “Mortality and morbidity in a pulp and a paper
America July 1997. mill in the United States: ten-year follow-up”,
293. Chemical Economics Handbook – SRI Br. J. Ind. Med., 36 (1979) no. 2, 127 – 134.
International, CEH Marketing Report, 303. L. R. S. Patil, R. G. Smith, A. J. Vorwald, T. F.
“Chlorine/Sodium Hydroxide”, 1996, USA, Mooney: “The health of diaphragm cell
Zurich, Tokyo. workers exposed to chlorine”, Am. Ind. Hyg.
294. ANORGANICA Ltd, Monthly Newsletter Assoc. J. 31 (1970) 678 – 686.
“Soda – Chlorine” Issue Number 197, 29th 304. H. Greim (ed.): “Chlor”,
May 1997. Toxikologisch-arbeitsmedizinische
295. Euro Chlor, GEST 97/230, D. Hutchinson: Begründungen von MAK-Werten,
Derzeitige und zukünftige Tendenzen bei der WILEY-VCH, Weinheim 2004.

You might also like