Experimental and Computer-Simulated Rin,,, €4”4 - -

by Thomas H, Cook, Brian W. Cook and Marlon E. Hopper Black Hills State College, Spearfish, SD

INTRODUCTION water rinse is commonly used be- A tween metal finishing operations

to prevent contamination of a subse- quent processing bath. Contamination of a process bath must be avoided to prevent product rejects. Once a process bath becomes contaminated, it must be chemically or otherwise purified or dumped and freshly made up.

A few years ago it was common prac- tice to flow tap water into the rinse and allow the contaminated rinse to over- flow and enter a sewer or waterway. Recently, environmental regulations have generally stopped the use of flow- ing rinsewater. Thus, metal finishers have had to periodically haul away their stagnant rinsewater or continuously re- move contaminants from their flowing rinses. Hauling or treatment of rinse- water is expensive.

Presented in this article are rinsing principles which can effectively reduce or eliminate the high cost of dealing with rinsewater. This is the final rinse article in a series of three. The first showed how a 40% improvement in rinsing can be attained by double-dip- ping the steel work in a single clean rinse tank. The second article demon- strated how a four-fold improvement in draining “basket” steel parts can be accomplished by using a suitable wet- ting agent.*

EXPERIMENTAL VERTICAL SHEET APPARATUS

Shown in Fig. 1 is the experimental vertical sheet apparatus. The device consists of three main parts:

1 . The six-inch high, single vertical sheet with drip tips on the lower edge.

2. The four, three-sixteenths inch diameter, horizontal connecting rods.

3. The vertical hook for hanging. The apparatus has a total surface area

of 13.26 square feet, and was specific- ally designed to allow free liquid drain- age.

The apparatus was constructed by laying the long, narrow, steel sheet on

Fig. 1. Vertical sheet apparatus. “A” is top view and “B” is constructed m give the best possible draining and is high sheet with drip tips. The mtalwface area is 13.26 square fegt.

view. devi- is a six-inch

Fig. 2. Basket apparatus. “A” is top view and “B”is bottom view. This device is constructed to give very poor draining by being filled with flat6 steel work pieces. The device is six inches high and has a total surface area of 13.26 square feet.

NOVEMBER 1987 - - _ _ 59 - -

the floor, placing flexible cardboard on the steel and rolling the steel and card- b o d up together. Four, *-six- teenth inch diameter holes were then drilled through two mutually perpen- dicular diameters of the spirally bun- dled steelcardboard layers.

After drilling, the cardboard was re- moved and the three-sixteenth inch diameter rods were inserted into the four horizontal connecting rod holes. Assembled in this way, the concentric spiral of the vertical sheet has a one- fourth inch space between successive tumS.

EXPERIMENTAL BASKET APPARATUS

Shown in Fig. 2 is the experimental steel basket apparatus which is filled with flat-S steel parts. The device con- sists of a thin steel cylindrical shell, a perforated disc-shaped bottom and a vertical hook for hanging. The ap- paratus was laced together by passing a small, soft steel wire through suitably drilled holes in the shell and bottom.

The flat-S parts were made from one- half inch wide by one and one-half inch long, thin steel sheet and were bent into a flat-S shape (described later). A suf- ficient weight of flat-S steel parts were put into the basket to give a total surface area, including the basket, of 13.26 square feet. The device was constructed to give poor draining by trapping liq- uid.

EXPERIMENTAL RINSING PROCEDURE

Shown in Fig. 3 is the experimental rinsing procedure. Seven ten-quart plastic buckets were used in the rinsing experiments. The first five buckets had exactly the same mixture of zinc chlo- ride (to inhibit acid attack on the steel) and hydrochloric acid so that the total chloride concentration 'was close to 11.22% (weight/volume).

3

Fig.

This 'acid was made by dissolving 800 grams of zinc metal in 2,000 ml of concentrated (37%) hydrochloric acid, decanting the clear liquid, adding 2,000 ml more of concentrated hydro- chloric acid, and then adding distilled water to make 16,000 ml of solution. Small volumes of concentrated hydro- chloric acid were added to the acid baths after each experiment to maintain a chloride concentration of 11.22%. The last two buckets contained exactly 8,000 ml of distilled water.

The rinsing procedure consisted of immersing and withdrawing the experi- mental apparatus (vertical sheet or bas- ket) several times in each blank acid and then into the working acid so as to maintain the working acid at 11.22% chloride throughout the experiment. Next, the apparatus was withdrawn and drained over the working acid for 20 Seconds, then immersed into the rinse tank. The apparatus was held in the rinse tank for 10 seconds, withdrawn and drained over the rinse tank for 20 seconds (single-dip rinse), then im- mersed into the process tank. The ap- paratus was withdrawn and immersed into the process tank 15 times (to re- move residual chloride from the appa- ratus), drained for exactly 20 seconds over the process tank and then placed on a paper towel for about 20 seconds. The apparatus was then immersed into the first blank acid to begin the next cycle. Each experiment consisted of 12 complete cycles.

In performing these rinsing experi- ments, four selections (from a total of eight parameters) were made. These parameters were:

1. Type of apparatus (sheet or bas- ket).

2. Rinsewater dump and refill fre- quency (stagnant or dynamic).

3. Rinse dip method (single-dip or double-dip).

4. Use of draining aid (no wetting agent or with a wetting agent).

By making each of the four selec- tions, a total of 16 unique rinsing exper-

vertical sheet device and the baske vice had the same surface area. A cro section view of a flat-S steel part

paratus, draiaing for ve the rinse tank and

then transfemng the apparatus to the next (process) tank.

A doubledip rinse' consisted of im- mersing the apparatus in the rinsewater

ck into the same

ume; 3 ml of wetting agenV10,OOO ml

3. Experimental rinsing procedure. Procedure consisted of immersing, draining and transferring the apparatus from tank to tank. Draining times were 20 seconds and d/p time in the rinse tank was 10 seconds. The blank acids pfsvented dilution of the working acid from the previous cycle.

60 METAL FINISHING

*in ,- I I-....

I

I 1

I I. . . . ., . . ._ _._ ,

. _. -. . - ' j . :

of liquid) of the aid to all of the acid baths, the rinsewater and the process solution. The wetting agent required several minutes of vigorous mixing to dissolve in each of the solutions and produced a slightly sudsy solution.

ANALYTICAL TESTING OF THE SOLUTIONS

To test the chloride in each solution, a sample ranging between 1 .OO ml (for the acids) to 10.0 ml (for the most dilute solutions) was pipetted into a 200-ml erlenmeyer flask. One gram of reagent grade calcium carbonate was added to neutralize excess acid, and sufficient distilled water was added to bring the total volume to 100 ml. One ml of 5% sodium chromate was added as an indicator and the chloride was titrated with standardized 0.5 M silver nitrate solution.

The chloride concentrations of the acids were determined before and after each experiment. The chloride concen- trations of the rinsewater and process solution were determined after the fourth, eighth and twelfth complete cy- cles. All chloride concentrations were expressed as weightlvolume percent.

Corrections were made in the experi- mental percent chloride results for the titrant blank of 0.13 ml and for the reduced volumes in the rinse and proc- ess baths because of sample with- drawal. The chloride results of the rinse and process baths were also normalized to correspond to values resulting from a working acid bath of exactly 11.22% chloride. The steel surfaces of the ex- perimental sheet and basket devices be- came systematically rougher and slight- ly pitted with each successive experi- ment. These small, systematic changes were monitored by measuring the vol- ume of liquid on the device before and after each experiment.

SHEET EXPERIMENTAL RINSING RESULTS

The sheet experimental rinsing re- sults are shown in Table I. Values are percent chloride after four, eight and 12 complete cycles. These values were corrected and normalized as previously described. Asterisks are adjacent to the chloride concentrations of the process baths after 12 complete cycles so that these rinse results can be directly com- P d .

The top half of Table I gives results for stagnant rinses (rinse water un-

changed for all 12 cycles) and the bot- tom half of Table I gives results for dynamic rinses (rinse tank dumped and refilled with distilled water after each cycle).

As Table I shows, the two important improvements in rinsing sheet are dou- ble-dip rinsing (in a single rinse tank) and maintaining clean rinsewater (dy- namic rinse). Obviously, double-dip rinsing lengthens processing time, and maintaining clean rinsewater can cause pollution or costly rinsewater treat- ment.

Surprisingly, the wetting (draining) agent caused the rinsing to be worse in three-fourths of the experiments shown

waters. The wetting agent (present in all the baths) greatly reduced rusting of the steel sheet while draining over the near-neutral rinse and process tanks. Without the wetting agent, the sheet rusted so much that it had to be cleaned every cycle. The values in parentheses in Table I were generated using a computer program.

BASKET (FILLED WITH FLAT-S STEEL PARTS) EXPERIMENTAL RINSING RESULTS

The basket rinsing results are shown

in Table 11. These values are percent chloride after four, eight and 12 com- plete cycles. Values after the twelfth cycle for the process bath are marked by an asterisk and can be directly com- pared. The top half of Table 11 is for stagnant rinses and the bottom half is for dynamic rinses.

-agent and maintaining clean rinse- water. Double-dip rinsing a basket is

the sheet (0.325/0.007 = 46.4).

an the sheet

All real work of this six-inch hi& (0.0391/0.0101 = 3.87).

Rinse Process Rinse PrOCeSS 4) 0.131 (0.132) 0.003 (0.003) 4) 0.127 (O.li3) 0.001 (0.001) 8) 0.261 (0.263) 0.007 (0.007) 12) 0.389 (0.392) 0.013 (0.013)'

Single-Dip Wth Wetting Agent Double-Dip With Wetting Agent Rinse Process PrOCeSS

4) 0.125 (0.126) 0.001 (0.001) ) 0.001 (0.001) 8) 0.251 (0.251) 0.004 (0.004) 8) 0.257 (0.255) 0.004 (0.004) 12) 0.372 (0.374) 0.008 (0.008)' 12) 0.385 (0.380) 0.008 (0.00s)'

Dynamic Rinse

SingleOip, No Wetting Agent Double-Dip. No Wetting Agent Rinse PrOCeSS Rinse PrOCeSS

4) 0.0330 (0.0332) 0.0023 (0.0023) 4) 0.0373 (0.0375) O.OOO8 (0.oooS) 8) 0.0326 (0.0332) 0.0044 (0.0044) 8) 0.0378(0.0375) 0.0014 (0.0014) 12) 0.0331 (0.0332) 0.0066 (0.0066)' 12) 0.0372 (0.0375) 0.0026(0.0020)'

Single-Dip With Wetting Agent Double-Dip With Wetting Agent . Rinse Process Rinse Process

4) 0.0318(0.0311) 0.0033(0.0035) 4) 0.0320 (0.0320) O.oooS(O.OOO7) 8) 0.0310(0.0311) 0.0071 (0.0070) 8) 0.031 9 (0.0320) 0.0014 (0.001 4) 12) 0.0312(0.0311) 0.0101 (0.0103)' 12) 0.0318(0.0320) 0.0021 (0.0021)'

Note: Values are percent chloride for rinse and process baths after 4, 8 and 12 complete cycles. These values are corrected for volume decrease due to sample withdrawal for testing and normalized for 11.22% chloride in the "working" add. Values in parentheses are from the computer rinse program. Values marked by asterisks can be directly compared.

NOVEMBER 1987 61

I .

I .

Table II. Basket (Filled with Flat4 Steel Parts) Experimental Rinsing Results.

Stagnant Rinse

Single-Dip, No Wetting Agent Double-Dip, No Welting Agent Rinse Process Rinse Process

4) 0.915 (0.903) 0.056 (0.060) 4) 0.922 (0.903) 0.044 (0.047) 8) 1.739 (1.734) 0.188 (0.183) 8) 1.755 (1.734) 0.155 (0.160) 12) 2.496 (2.498) 0.374 (0.361)" 12) 2.544 (2.498) 0.325 (0.327)'

Single-Dip With Wetting Agent Double-Dip With Wetting Agent Rinse Process Rinse Process

4) 0.418 (0.416) 4) 0.385 (0.383) 0.010 8) 0.821 (0.818) 8) 0.752 (0.754) 0.031 12) 1.200 (1.204) 12) 1.117 (l.ll2) 0.065 (0.066)*

Dynamic Rinse

Double-Dip, No Wetting Agen; Rinse Process

4) 0.2285 (0.2283) 0.0282 (0.0282) 4) 0.2398 (0.2394) 0.0204 (0.0207) 8) 0.2279 (0.2283) 0.0550 (0.0542) 8) 0.2371 (0.2394) 0.04oO (0.0397) 12) 0.2285(0.2283) 0.0771 (0.0780)' 12) 0.2421 (0.2394) 0.0575 (0.0571)'

Single-Dip With Wetting Agent Double-Dip With Wetting Agent Rinse Process Rinse Process

4) 0.0979 (0.0960) 0.01 27 (0.0131) 4) 0.1014 (0.0970) 0,0046 (0.0047) 8) 0.0980 (0.0960) 0.0261 (0.0258) 8) 0.0951 (0.0970) 0.0094 (O.QO93) 12) 0.0927(0.0960) 0.0391 (0.0380)* 12) 0.0963 (0.0970) 0.0141 (0.0137)'

Note: Values are percent chloride for rinse and process baths after 4, 8 and 12 complete cycles. These values are corrected for volume decrease due to sample withdrawal for testing, and normalized for 11.22% chloride in the "working" acid. Values in parentheses are from the computer rinse program. Values marked by asterisks can be directly compared.

size would contaminate a process tank within the values established by Table I (low limits) and Table II (upper limits). Larger steel workpieces take longer to drain2 and would cause greater process bath contamination than shown in Tables I and LI unless the draining times were longer.

Horizontally bundled pipes drain very much like a basket filled with flat- s parts.' This poor draining results from each pipe being in contact with six surrounding pipes, giving 12 capil- lary contacts on the outside of each pipe within the bundle.

Because all real work is likely to be a geometrical mix of the sheet and bas- ket devices, double-dip rinsing and an effective wetting agent will likely be beneficial in most rinsing operations; however, as will be shown later, dou- bledip rinsing is not effective if the rinsewater is heavily contaminated. Also, some wetting agents are not com- patible with subsequent process opera- tions. For example, some organic wet- ting agents cause a black, nonadherent cathode deposit in electroplating.

THE RINSE MODEL ; Steel workpieces, previously cleaned in acid, which have just been withdrawn from rinsewater have four

types of liquid on them. Shown in Fig. 4 are the four types: adhered rinse- water, bridged rinsewater, cupped acid and acid layer.

Of these four types of liquid, three can be experimentally shown to be pre- sent or can be easily conceptualized. For example, apiece of steel withdrawn from water is coated with adhered rinsewater. By careful observation, bridging water can be seen to fill the capillary space between closely spaced steel pieces. Cupped acid is easily con- ceptualized because the higher density acid would tend to remain in the bottom of cup-shaped workpieces during rins- ing.

Only the acid layer cannot easily be shown to be present when an acid- treated workpiece is rinsed in water. Several years ago it was observed that the third (pracess) tank in a three-tank system reached 1.54% sulfate in eight months of steel processing. In this sys- tem, the first tank was sulfuric acid averaging 17% sulfate and the second tank was a flowing rinse averaging only 0.05% sulfate. The workpieces were 40 feet long, large diameter pipes, racked well apart from one another. Also, the draining between the tanks averaged 30 seconds.

By the theory of dilution, the process

tank could not have a higher sulfate concentration than that of the rinse, that is, 0.05%. The correct explanation is that the acid-layer on the pipes was

ned under 0.4% sul- . Dipping twice re-

shown by the data in Table I. In three-

, resulting in no acid

doubledip rinsing (as compared with single-dip rinsing) results in a 5.3-fold

ume quantity, although diffusion (mix- and rinse layers may

single-dip and dou- ble-dip rinsing, the same amount of acid adheres to the steel work upon withdrawal from the acid. The large, 5.3-fold increase in contamination of the process bath with singledip rinsing occurs because the acid layer is not washed off in the rinse tank. Thus, in single-dip rinsing (Fig. 5) the rinse chloride concentration is lower (0.03348%) than the rinse chloride con- centration found in double-dip rinsing (0.03394%) because in double-dip rins-

is washed off in the

In single-dip rinsing, the acid layer is removed in the process bath, whereas in doubledip rinsing the acid layer is removedin the rinse tank. Thus, single- dip rinsing allows some acid (the acid layer) to be directly transferred from the acid solutions to the process bath.

These rinse calculations are based on simple dilution theory, except for an experimental draining ratio which ranges between 0.9303 and 0.9952. This ratio was determined experimen- tally (three times) and the results are

62 METAL FINISHING

shown in Fig. 6. Thus, distilled water does not drain as well as solutions con- taining 1.272% or more chloride. -

The base-line draining ratio of 0.9303 is the ratio 174.9 mV188.0 ml. Distilled water has a draining ratio of 1.oooO and all solutions in this study above 1.272% chloride have a draining ratio of 0.9303. For solutions greater than 0% chloride and less than 1.272% chloride, the draining ratio can be cal- culated from the empirical formula: R = 1 - [0.0385849 log(4C + l)]; where R is the draining ratio, log is the loga- rithm base e (2.71828 . . .), and C is the percent chloride of the solution which is draining.

The reason distilled water does not drain as well as chloride (or perhaps zinc or free acid) containing solutions was not investigated. THE COMPUTER RINSE PROGRAM

For the rinse model, it is necessary to press the keys on a calculator 306 times to determine the chloride concen- trations of both rinses and both process Table 111. Computer Rinse Program

(Program Variables Are Defined in Table Iv)

10 INPUT TANK V0LUMES";T 20 INPUT "ADHERED LIQUID";A 30 INPUT "BRIDGED LIQUID";B 40 INPUT "ACID LAYER";L 50 INPUT "ACID C0NTAMINANT";CI 60 INPUT "CYCLES (WORK L0ADS)";D 70 INPUT "CVCLES/DUMP RINSE";DR 80 INPUTTYCLE INTERVALTO PRINT";

Qo 90 RT=T:PT=TA=A+B

100 FOR I = I TO D 110 IF I/DR=l/DR THEN GR=O:RT=T 120 C = CI:Z = A:GOSUB

130 C = GR/RT:GOSUB

140 GP= GP+ G:PT = PT + Y 150 GP= GP + L'CI:PT= PT+ L 160 C = GP/PT:Z = (A + L):GOSUB

170 IF RQQ = I/QQ THEN PRINT USING

210: GR =GR+G:RT= RT+Y

210:GR = GR-G:RT= RT-Y

210:GPz GP - G:PT= PT-Y

'### ##.#### ##.###x"; I,GR/RT,GP/PT

180 A$ = INKEY$:IF A$ = THEN 200 190 A$=INKEY$:IF A$=""THEN 190 200 NEXT I:END 210 IF C<= 1.272 THEN R = 1 -

(.0385849'LOG(4'C + 1)) ELSE R = .9303

220Y=R'Z 230 G=Y*C 240 RETURN

__

WITH ACID LAYER

NO ACID

iRlNSE

PROCESS.

. PROCESS. 8600ml

(O.~SI)(26.09)(0.0339~ - 8000 +(0.9951)(26.09)

- (0.9303)(26.09)(11.22$0 - 8000+(0.9303)(26.09)

-

Fig. 5. Single-dip rinsing with acid layer and double-dip rinsing with no acid layer. Single-dip rinsing allows acid to be transferred directly from acid bath to process tank. Double- dip rinsing removes acid layer in the rinse tank.

solutions in Fig. 5 . These time-con- suming calculations give the chloride concentrations in these two experi- ments for only the first complete cycle. Additional cycles require considerably more calculations for each cycle, and a single error would cause all subse- quent results to be incorrect.

The chloride concentrations of rinse and process baths after many cycles (e.g. 1 ,OOO cycles) must be calculated to understand rinsing in most practical applications. A suitably programmed personal computer is the ideal tool to perform the many calculations.

Table III gives the computer rinse program based on the rinse model. This program is in BASIC and was written for an IBM-PC. The program uses the

four types of liquid clinging to the steel workpieces and other data as inputs. The sum of the adhered and bridged liquids are taken as a single input be- cause these liquids are exchanged (acid for rinse) in the rinsewater. The sum of the acid layer and cupped acid are taken as a single input because this acid is not exchanged for rinsewater in the rinse tank.

The program computes the mass of chloride that is transferred from tank to tank and sums these values for the rinse and process solutions for each cycle. It also determines the volume of liquid in the rinse and process tanks for each cycle. By determining the chloride mass and the total bath volumes, the percent chloride (weightholume) is

NOVEMBER 1987 63

B 'sOl-----T P 1-R I

I 3 4

x CMR(0E

Fig. 6. Volume of liquid in a six-inch high basket filled with flat-S steel parts after 30 seconds draining vs per- cent chloride in liquid.

computed for rinse and process baths and printed for each complete cycle.

Definitions of the variables for computer rinse program are in Table IV. Details of program operation can be determined by plugging in numbers for the program variables and then cal- culating the chloride concentrations (in rinse and process baths) for a few cy- cles.

From the results of the 16 rinse ex- periments (Tables I and II) and from the total clinging liquid in these exper- iments, the separate adhered rinse, bridged rinse, cupped acid and acid layer volumes were determined. These separated volumes, in ml, are shown in Table V. By using these separate volumes in the computer program, the average error between computer-simu- lated results and experimental results was less than 2 2%. The computer re- sults are shown in parentheses in Tables I and II adjacent to the corresponding experimental results. The computer re- sults match the experimental results quite well, thus substantiating the validity of the rinse model and com- puter program.

The adhered rinse (16.0 ml) for the sheet device was determined from other experiments.' The adhered rinse for the basket device was estifnated at 20.0 ml on the basis that draining through flat4 parts is a longer, less direct pathway for adhered rinse than on vertical sheet.

The small variations in the separate volumes shown in Fig. 5 were a result of systematic changes in the draining devices and nonreproducible packing of the flat-S parts. The bridged rinse volume for the sheets, which changed from 7.8 to 12.8 ml, was a result of the acid enlarging the holes through which pass the horizontal connecting rods. This enlargement caused a larger gap between the rods and the holes,

~- I . - L ~ . -

1 _

Table IV. Deflnttlons of Variables In the Computer Rinse Program

T = initial volume in each tank (ml). A = volume of liquid adhered to the Adhered acid is washed off in the rinse t B = volume of bridged liquid which fills t workpiece~. Bridged liquid is carried from ta rinse tank and exchanged for bridged rinse L = volume of the acid layer which is strongly through the rinsewater. The acid layer is transferred tank. Cupped acid is included in this variable (ml). CI = weight/volume percent of contaminant in acid ( D = number of cydes to run the program or the nu all the tanks (no units). DR = number of cycles to run the program or the the rinse tank before dumping the rinse tank and QQ = cycle interval to print. Enterin would mean every tenth cycle wo be computed (no units). I = a sequential numbering which prints loads going through all tanks (no units). RT = volume of liquid in rinse tank during a specific PT = volume of liquid in process tank during a s GR = mass of pure acid in rinse tank = 1 gram). GP = mass of pure acid in process tank during a specific cycle (centigrams). C = "working variable" equal to wei line 210 (% chloride in this study). 2 = "working variable" equal to volu at line 210 (ml). Y = actual volume (determined by empirical drainin to the next (ml). G = mass of pure acid carried from onetank to the Lines 180 and 190 stattlstop the program by pressing any key. R = draining ratio. Calculated according to line 210 (no me).

Table V. Separate Values for Adhered Llquld, Bridged liquid, Cupped Acid, and Acid Laver

liquid carried from one tank

Sheet

Single- Double- Single- Double- Dip Rinse Dip Rinse Dip Rinse Dip Rinse

16.0 16.0 16.0 16.0 7.9 8.6 0.56 0.060

No Wetting Agent With Wetting Agent

Adhered rinse, ml Bridged rinse, ml 9.5 7.8-12.8 Layer acid, ml 0.33-0.34 0-0.028

Basket

Single- Double- Single- Double- Dip Rinse Dip Rinse Dip Rinse Dip Rinse

No Wetting Agent W~ Wetting Agent

20.0 20.0 20.0 20.0 54.2 55.0

Adhered rinse, ml Bridged rinse, mi 158.6-162.0 157.5-162.0

1.8 0.22 Layer CL Cup ackl, ml 1.8-2.3 0-0.009

Note: All values are in ml, and a range indicates systematic changes m the draining apparatus or nonreproducible packing of flat4 steel workpieces. These volumes for adnerd rinse and bridged rinse are for distilled water. If the rinsewater is between 0%-1.272% chloride, then adhered rinse and bridged rinse volumes equal Table V values times 'R" where R = 1 - (.0385849'LOG(4'C + 111; C = percent chloride in rinsewater; LOG is natural log base e. Adhered acid and bridged acid equals Table V values times 0.9303.

thus increasing the bridged rinse vol- me. The acid layer systematically in- :reased slightly due to minor acid pit- ing of the steel surface. The bridged inse variations for the basket apparatus between 158.6 and 162.0 ml) were a esult of nonreproducible packing of

the flat-S steel parts. The result of wetting agent on the

acid layer on the sheet was quite unex- pected. As shown in Table V, the wet- ting agent increased the acid layer on the sheet for both single- and double- dip rinses. For single-dip rinsing with-

METAL FINISHING 64

out wetting agent, the acid layer 0.34 ml, whereas with wetting agent the layer increased to OS6 ml. The cor- responding values for double-dip rins- ing were 0.0 ml without wetting agent and 0.060 ml with wetting agent. This increased acid layer caused by the wet- ting agent could explain the observation that, in the presence of a wetting agent, rusting of the steel was greatly inhib- ited. Since acid (low pH) inhibits rust, perhaps the thicker acid layer was the actual cause of rust inhibition. The thicker acid layer caused by the wetting

PRACTICAL APPLICATIONS OF THE COMPUTER RINSE P

A large nse cycles are involved in most-practical industrial ap- plications. Shown inFig. 7arethe com- puter simulated results for rinse and

sheet for many cycles (in a contami- nated, stagnant rinse) is that a wetting agent slightly improves the purity of the process solutions.

For many cycles, double-dip rinsing does not measurably improve rinsing. These results are directly opposite to those found for 12 cycles (in relatively clean rinse) where double-dip rinsing gave the major rinsing improvement as shown in Table I.

Figure 8 shows the computer simula- tion when the rinse tank is periodically dumped and refilled and is being used to rinse sheet. The frequency of dump- ing and refilling is: 100 cycles (left); 50 cycles (center); and 10 cycles' (right). Figure 8 shows that double-dip rinsing becomes important in cleaner rinsewater.

Figure 9 gives the computer rinse simulation for the flat-S filled basket being rinsed 1,OOO times in a stagnant rinse. As Fig. 9 shows, the use of a wetting agent is quite important in rins- ing a basket. With contaminated, stag- nant rinse, double-dip rinsing the bas- ket is apparently not important. Figure 9 also shows (as compared with Fig. 7) that the process solution is more heavily contaminated by rinsing bas- kets than by rinsing she&.

Figure 10 present5 the computer re- sults for periodic dumping and refilling the rinse tank which is being used to rinse the basket device. The frequency

-

9

18

7

6

5

4

3

2

I

0 0 m200300400

Fig. 8. Computer simulated results for periodically dumped rinse bt'vertical sheet for 1,OOO work loads (cycles). Rinse dumped after every 100 CJ&~ fh$tthird of figure). Rinse dumoed after evew 50 cvcles Icenhr thid of figure). Rinse dumped after every 10 cycles (right third i f fig&).

of dumping the rinse tank is: 100 cycles (left); 50 cycles (center); and 10 cycles (right). As Fig. 10 shows, a wetting agent is extremely important in rinsing a basket filled with flat-S steel parts. Once again, double-dip rinsing be- comes important if the rinsewater is quite clean.

From Figs. 7 and 9, it is clear that using a stagnant rinse will cause the process tank to become contaminated. Thus, rinsewater must be periodically withdrawn from the rinse tank and re- placed by clean rinsewater. Often, con- taminated rinsewater is chemically treated and disposed of in a sewer or waterway. Occasionally, contaminated rinsewater is hauled away, generally at considerable expense. Altematively , contaminated rinsewater can be evapo-

'

NOVEMBER 1987 . . ~ " --. __ - I

t e a small volume and -acid tank. Polluted used to make up for in a hot acid tank.

Contaminated rinsewater can be used

up fresh acid, one hot dipgalvanizing company has stopped using their water treatment facilities in two plants. The operators of these plants understood the

tanks were required in one plant. In using two rinse tanks, the one

closer to the acid (the dirtier rinse) is pumped into the acid tank to make up

65 _ _

P OR SINGLE DIP

OR SINGLE DIP

PROCESS TANK: NO WETTING AGENT &AFTER SUUGLE DIP RINSE OR AFTER DOUBLE DIP RINSE

Fig. 9. Computer simulated results for stagnant tfnsing of basket filled with flat-S steelparts for 1,OOO work loads (cycles).

I

1 ' 0 100 200 300 40(

mnm Lo*Ds Fig. 10. Computer simulated results brperiodiicay dumped tfnse for basket filled with flat6

steel parts for 1,OOO work loads (cycles). Rinse dumped after every 100 cycles (/en thirij of figure). Rinse dumped after every 50 cyc/es (center third of figure). Rinse dumped after every 10 cycles (right third of figum).

for evaporation and is used to prepare fresh acid baths. The rinse closer to the process tank (the cleaner rinse) is pumped into the dirtier rinse tank when- ever possible. Clean tap water is added only to the cleaner rinse tank, never to the dirty rinse Qr to the acid tank. By not using the two water treatment facilities, the company is saving ap- proximately $200,000 per year.

In laboratory experiments and in plant operations, hot rinses and near- neutral rinses are usually bad for the steel work, and can be expensive. Hot rinsewater can dry while the steel is draining, which exposes the metallurg- ically clean steel surface to air oxida-

\ tion. Hot, thick steel work, (e.g. steel pickled in hot sulfuric acid), must be cooled in room temperature rinse so

that it will not dry and rust during rinse draining. Table VI shows approximate, experimental times required to cool the steel work to prevent drying and rusting during draining.

Near-neutral rinses having a pH be- meen 5 and 9 promote rusting of the steel surface. Workpieces generally should not be stored for extended periods in near-neutral rinses. Rinses with a pH below 5 or above 9 tend to inhibit rusting even if the rinsewater dries during draining. For rinsing after an acid bath, a steel rinse tank must be lined (with rubber or acid-proof fiber- glass) to prevent corrosion by the low pH rinsewater.

In laboratory experiments, air bub- bling in the rinse bath or lateral move- ment of the workpieces improved rins-

Table W. ~pproximate DIP n m in Rinsewater to

tDip SecondDip

5 10 18

45

Imshmg me surrace of con- _ . . . -

A . *. ,. "

tinuous sheetln themse tank i s highly effective ih removing . the . . acid . . layer. . . -

up new acid. Currently, there are loca- tions in developing countries wliere in-

s directly compete for industrialization pro- lation increases, this

competition for clean water will inten- sify. MF

a m a s H. cook is a pmfessor of chemistry at Black Hills State College. He received a PhD in inorganic chemis- try from the University -of. Wyoming. He is an international consultant on hot dip galvanizing.

Brian W. Cook is a sophomore majoring in computer science at South Dakota School of Mines and Technol- ogy. He is the son of Laurie N. and Thomas H. Cook.

Marlon E. Hopper is a sophomore majoring in chemistry (math minor) at Black Hills State College.

METAL FINISHING 66