ORP is flaky
Take a look at
this paper which was done roughly in the same timeframe as the others they reference. Basically, the paper is just concluding that the OTO test sucks, which we already knew, and that the Free Chlorine (FC) measurment even with DPD isn't useful when chlorine is combined with other substances such as CYA, which we also already know. ORP does not correlate better than FC when there is no CYA present and when CYA is present they didn't try calculating the theoretical HOCl concentration. Don't get sold into the ORP hype that is only coming from ORP manufacturers.
ORP is very flaky and all over the map between different sensors from different manufacturers. I describe how terrible this situation is in
this post. I tried to do a correlation analysis using data in the Aquarius paper and was having a heck of a time because the graphs within the paper itself are terribly inconsistent! For example, reading the ORP off of the "ORP millivolts versus Free Chlorine" on page 1 at 2 ppm FC and pH 9 gives around 625 mV while the "ORP millivolts versus pH value" graph on page 4 gives around 540 mV. I find that I can get a decent correlation if I ignore the graph on page 1 and use instead the two graphs on page 4 (the first having detailed data for a pH of 7.5) and I have added that to my PoolEquations spreadsheet. However, note that Aquarius has the steepest ORP mV change per doubling in HOCl at 46 mV per doubling compared to most other sensors which have around 20-28 mV per doubling (Sensorex is a weird exception with 84 mV per doubling).
See
this post for how ORP correlates with calculated HOCl in real pools. In this study of 620 samples from 194 pools, comparing ORP readings from a portable Oakton ORP measuring device vs. the built-in ORP controllers in many of the pools, 30 out of 130 (23% of those that had built-in ORP controllers) were off by more than 100 mV from each other measuring the same pool water at roughly the same time! To depend on ORP in any absolute sense is pretty much ridiculous. It's OK for process control when manually adjusting a setpoint to get to the FC level you want, but other than that it's not very useful. Oxidizers such as MPS will register high ORP, but are not very fast at killing pathogens while the presence of reducing agents will lower ORP even if they don't react with chlorine or react very slowly. ORP is measuring a thermodynamic value (even then, it doesn't match theory in the case of HOCl) and not reaction rates so it is easily fooled. The presence of hydrogen gas from SWG systems can affect the readings (though not actual kill times) as can sunlight.
The World Health Organization (WHO) and others have set 650 mV as the disinfection standard for water in spite of all the problems noted above. At a pH of 7.5 (and 77ºF temperature), 650 mV is achieved with no CYA in the water with 0.023 ppm FC for the Chemtrol sensor, 0.082 ppm FC for the Oakton sensor, 0.25 ppm FC for the Aquarius sensor, and 0.62 ppm FC for the Sensorex sensor. So which one do you choose? It's nuts, yet no manufacturer cops to this problem.
Low chlorine levels are difficult to maintain
Trying to maintain FC levels too much below 0.5 ppm is difficult in pools because a relatively small chlorine demand can readily consume that amount of chlorine locally thereby creating regions with no chlorine for longer than one would like. In higher bather load pools, it may be difficult to circulate the water well enough and fast enough to replenish a chlorine level of, say, 0.2 ppm FC (diffusion is slow in still water; it is physical circulation/movement of the water that distributes the chlorine). This is one of the big advantages of using CYA in the water since it is an HOCl buffer so can hold a chlorine reserve while maintaining a low HOCl level that is still fast at killing pathogens. The chlorine is released from CYA very quickly -- the half-life of one species is 0.25 seconds so half of the reserve can release in that time (normally far less is needed from the reserve). This is far faster than the many minutes it would take for circulation to get chlorine from the controller to any particular spot in the pool or from one part of the pool to another.
Appropriate active chlorine level for sanitation
As for what level of active chlorine (hypochlorous acid) is needed for decent sanitation, that is a legitimate debate. Kill times are usually quoted by using the product of chlorine concentration in ppm (C) with time in minutes (T) at a given kill rate (log reduction or % kill) for a given condition of pH and temperature. Most heterotrphic bacteria have a 99% (2-log reduction) kill at room temperature (77ºF) at a pH of 7.5 with a CT of 0.04. See the PDF file with a table in
this link at the CDC to see more difficult pathogens. With a minimum FC that is 7.5% of the CYA level, this is equivalent to an FC of around 0.06 ppm FC with no CYA and has an ORP (at 77ºF) of 682 mV with Chemtrol, 638 mV with Oakton, 557 mV with Aquarius, and 373 mV with Sensorex. Pathogens with a CT of 0.04 get 99% killed in 0.04/0.06 = 0.67 minutes or around 40 seconds. As for uncontrolled bacterial growth, one looks at the 50% kill rate in 15 minutes which is the fastest generation rate for bacteria. I write about how to calculate this in
this post. So the 0.06 ppm FC would kill pathogens faster than they can reproduce if they had a 2-log (99% kill) CT of 0.06*15*2/0.301 = 6 or lower. In the CDC table, pretty much the only pathogens not killed reasonably quickly are the bacteria
Burkholderia pseudomallei,
Vibrio cholerae (rugose strain; smooth strain is easy to kill),
Yersinia enterocolitica, and all of the protozoa though most especially
Cryptosporidium parvum (the more recent CT value for Crypto is 15,300 as this table is a bit out of date). In practice, it is only Crypto that is a problem in chlorinated pools. Even
Giardia lamblia with its CT of 15 is reasonably handled since protozoan oocysts do not reproduce in swimming pool water -- they only reproduce in your gut when the cysts are broken open. A 99.9% kill with a CT of 15 is equivalent to a 99% kill with a CT of 7.5 and that takes 7.5/.06 = 125 minutes or around 2 hours to achieve at the FC/CYA ratio minimum for manually dosed pools recommended on this forum. That won't prevent person-to-person transmission during a fecal accident, but it will prevent widespread outbreaks. This is one reason I think the standard for commercial/public pools could be higher with an FC/CYA ratio of around 0.2 (i.e. an FC that is 20% of the CYA level -- 6 ppm FC with 30 ppm for outdoor pools and 4 ppm FC with 20 ppm CYA for indoor pools) which will do a 99% kill of Giardia in just under 40 minutes.
There is a downside to not using any CYA at all and having higher FC levels such as found in most indoor pools. The rate of production of volatile and irritating nitrogen trichloride is theoretically much higher in such pools since it is roughly proportional to the active chlorine (hypochlorous acid) level as I describe in
this post.
Richard