574
A. Golomb
After three months, the plating shop foreman reported no detectable change in plating
quality. Nickel salt requirements for bath make-up were virtually eliminated, and boric
acid usage had dropped to below half the regular quantity. Deionized water usage was
minimal: the only water input exactly matched volumetric evaporative losses from the
bath. Perhaps even more significant in its wider implications, the system was, from a
pollution standpoint, essentially a 'closed loop'.
A gradual decline in flux was observed during the initial 3-month test period (Fig. 3).
This is attributed, at least in part, to precipitation of ferric hydroxide as a fine film over
the membrane surface. The presence of iron in the bath and rinse waters results from
local galvanic action on steel piece parts which occasionally drop inadvertently from the
racks during plating and rinsing. (That the iron should be present in the ferne state rather
than ferrous is predictable from the known presence of hydrogen peroxide added to the
plating bath as a depolarizing agent. Furthermore, oxygen levels close to saturation would
be anticipated, since air agitation of the bath and rinse tanks is used to promote efficient
mixing.) Experimentation indicated that membrane fouling could be kept well under
control by carefully adjusting the feed solution pH. Gradual addition of dilute sulphuric
acid to the feed holding tank maintained it a pH 4.0 ± 0.5. As can be seen from Fig. 3,
occasional acid flushing of the unit with a more strongly acidified feed at pH 2.0 for 30
minutes had a restorative effect on the flux rate. (Alternatively, flushing with a 0.3% w/w
aqueous solution of oxalic acid was found to be very effective, but can only be
conveniently accomplished during a lapse in plating production.) Typical results of the
RO field trial during the initial 3-month period (after 400 hrs of operation) are
summarized in Table II. These results clearly lend weight to the predictions based on the
earlier laboratory experimental work.
PROJECTED COST ESTIMATES
A summary of an order-of-cost analysis for 'closed loop' reverse osmosis recovery of
nickel plating chemicals is given in Table III. Three levels of operation are considered,
with dragout rates of 15, 50 and 500 gal/day respectively. The assumptions made are as
follows:(i) Relative to the field trial data, process flow rates are taken to be proportionately
higher for the larger scales of operation, where concentrations remain the same,
(ii) In estimating equipment size and cost, a flux rate of 10.7 gfd at 450 psi has been
used throughout,
(iii) Capital costs are based on straight line amortization over a five-year period at an
interest rate of 8.5 percent. Operation is assumed on 240 days per year, with three
8-hr. shifts per day.
(iv) As a first approximation, for the three levels considered, combined operating and
labour costs equivalent to 0, 5% and 25% respectively of 1 man per 24-hr day at
$8.00/hr have been applied. Membrane replacement costs are based on a membrane
life of 18 months.
Total operating and maintenance costs per 1000 gal of permeate are:
A. Golomb
After three months, the plating shop foreman reported no detectable change in plating
quality. Nickel salt requirements for bath make-up were virtually eliminated, and boric
acid usage had dropped to below half the regular quantity. Deionized water usage was
minimal: the only water input exactly matched volumetric evaporative losses from the
bath. Perhaps even more significant in its wider implications, the system was, from a
pollution standpoint, essentially a 'closed loop'.
A gradual decline in flux was observed during the initial 3-month test period (Fig. 3).
This is attributed, at least in part, to precipitation of ferric hydroxide as a fine film over
the membrane surface. The presence of iron in the bath and rinse waters results from
local galvanic action on steel piece parts which occasionally drop inadvertently from the
racks during plating and rinsing. (That the iron should be present in the ferne state rather
than ferrous is predictable from the known presence of hydrogen peroxide added to the
plating bath as a depolarizing agent. Furthermore, oxygen levels close to saturation would
be anticipated, since air agitation of the bath and rinse tanks is used to promote efficient
mixing.) Experimentation indicated that membrane fouling could be kept well under
control by carefully adjusting the feed solution pH. Gradual addition of dilute sulphuric
acid to the feed holding tank maintained it a pH 4.0 ± 0.5. As can be seen from Fig. 3,
occasional acid flushing of the unit with a more strongly acidified feed at pH 2.0 for 30
minutes had a restorative effect on the flux rate. (Alternatively, flushing with a 0.3% w/w
aqueous solution of oxalic acid was found to be very effective, but can only be
conveniently accomplished during a lapse in plating production.) Typical results of the
RO field trial during the initial 3-month period (after 400 hrs of operation) are
summarized in Table II. These results clearly lend weight to the predictions based on the
earlier laboratory experimental work.
PROJECTED COST ESTIMATES
A summary of an order-of-cost analysis for 'closed loop' reverse osmosis recovery of
nickel plating chemicals is given in Table III. Three levels of operation are considered,
with dragout rates of 15, 50 and 500 gal/day respectively. The assumptions made are as
follows:(i) Relative to the field trial data, process flow rates are taken to be proportionately
higher for the larger scales of operation, where concentrations remain the same,
(ii) In estimating equipment size and cost, a flux rate of 10.7 gfd at 450 psi has been
used throughout,
(iii) Capital costs are based on straight line amortization over a five-year period at an
interest rate of 8.5 percent. Operation is assumed on 240 days per year, with three
8-hr. shifts per day.
(iv) As a first approximation, for the three levels considered, combined operating and
labour costs equivalent to 0, 5% and 25% respectively of 1 man per 24-hr day at
$8.00/hr have been applied. Membrane replacement costs are based on a membrane
life of 18 months.
Total operating and maintenance costs per 1000 gal of permeate are: