Page 308 - Failure Analysis Case Studies II
P. 308

293










                                           I .25














                            Fig.  10.  A  layer of  magnetite sludge had formed in the water space at  the bottom of  the radiator. Pitting
                            corrosion had occurred where the sludge covered the steel. Dimensions in mm.


                       until the concentration of sodium nitrite in the water was 1300ppm (parts per million, or mgl-’).
                       The sodium nitrite was the active ingredient in the inhibitor package; the sodium borate was added
                       as a buffer to keep the pH of the water at about 9. However, after only 2 years the new radiators
                       started leaking. The offending radiators were removed, and one was cut open. As shown in Fig. 10,
                       a layer of black sludge had formed in the water space at the bottom of the radiator. The corrosion
                       had originated beneath this deposit, and had developed into deep pits. One of these had penetrated
                       right  through  to the  outside  surface. However, there  was no  evidence of  any  pitting  above  the
                       sludge.
                         The water had been analysed every 2 months to make sure that the inhibitor concentration was
                       up  to  the  recommended  level.  In  most  cases, it  was  found  that  the  Concentration  had  fallen
                       significantly since the previous service, and extra inhibitor had to be put in to make up the shortfall.
                       The average concentration reading was  E 500 ppm. However, on several occasions the nitrite level
                       was as low as 100 ppm. An analysis of the mains water gave: pH E 8; C1-  z 500 ppm; SO:-  E 100 ppm;
                       high hardness. However, samples of water taken from the system itself gave: pH z 10; C1- z 250 ppm
                       (max); SO:-  z 300 ppm (max); strong indications of nitrite-oxidizing bacteria.

                       6.2.  Failure analysis
                         The first thing that one notices about the system is the high concentration of C1-  and SO:-,  The
                       minimum concentration of sodium nitrite which must be added to compensate for these aggressive
                       ions can be estimated from Eqn (5). The ionic mass of CI-  is 35, and the molar mass of NaCl is 58.
                       The 250ppm of C1-  is therefore equivalent to (58/35) x 250=414ppm of NaC1. The ionic mass of
                       SO:-  is 96, and the molar mass of Na2S0, is 142. The 300ppm of SO:-  is therefore equivalent to
                       (142/96)  x 300 = 444 ppm of Na,SO,.  The combined mass of NaCl and Na2S0, is 858 ppm. Equation
                       (5) shows that this must be balanced by  at least 858ppm of NaNO,.  As seen earlier, the average
                       concentration of inhibitor was  E 500ppm, and the minimum concentration was  E 100ppm. These
                       levels are far below the minimum requirement, so it is hardly surprising that the radiators corroded
                       rapidly. The situation was still worse in practice because the sludge would have screened the metal
                       from  the  inhibitor.  The  routine  analyses  show  that  the  inhibitor  level  regularly  fell  from  the
                       recommended dose of 1300ppm to as little as l0Oppm after only 2 months. Presumably nitrite was
                       consumed by  the nitrite-oxidizing bacteria.  This would have been avoided if  a biocide had  been
                       added to the water when the system was first filled.
   303   304   305   306   307   308   309   310   311   312   313