Deionized water is often used in precision cleaning. It's a tried-and-true process, but with a number of hidden pitfalls.

In general, there are three general grades of water: tap water, distilled water and deionized (DI) water. In terms of precision cleaning, neither tap water nor distilled water are sufficiently pure to handle the job as both are contaminated, to greater or lesser degrees, with minerals and organics. So you must jump to DI-water.

The quality of DI water usually is measured by the water's resistance to electric current (in OHM-cm). Quoting from Finishing.com: "Deionized water quality depends on a variety of factors, including raw water composition, ion exchange types and quantities, and the number of resin tanks in the system.

Two-bed deionizers use separate tanks, one containing cation resin, the other containing anion resin. A two-bed weak vase deionizer typically produces water with electrical resistance of about 50 kOHM-cm. A two-bed strong base deionizer typically produces water with electrical resistance of about 200 kOHM-cm."

Now, how pure is that?

• 50 kOHM is pretty standard and can be produced easily and inexpensively by many in-house deionizing systems.
• 1 megaohm of resistivity is the minimum required for true precision cleaning.
• In the really high-end world, such as the semiconductor industry, 20 megaohms is the norm. 20 megaohm water is so hungry for ions it will cut through steel.

Obviously, the cost, energy consumption, through-put and handling issues all jump exponentially as the purity of the water increases. The more pure the water, the more hungry it is for ions, and the more contamination it will attract unless the packaging and handling is tightly controlled.

Now, you didn't mention any specifics about application into which you were working. But unless you have an elaborate and energy-intensive DI generating system, your water will generally not be clean enough for highly aggressive cleaning of normal PCBs.

We have seen people try to use DI-water, for example, on the bench top. This is futile. If the water is pure enough to be an aggressive cleaner, it instantly will become contaminated as soon as the bottle or container is opened, and at that point you might as well be cleaning with distilled water and save some money.

People also try to use DI-water in ultrasonic tanks. We just visited a company with thirty 5-gallon ultrasonic cleaners, and they were changing the DI-water every 60 minutes in every machine. This was because they found that after an hour or so the cleaning had stopped no matter how long they ran the machines.

Again, the issue here is handling: an open-topped ultrasonic cleaner holding DI-water is always going to be recontaminated within minutes, and ordinary water won't clean (without additives).

The only viable option is a tightly-sealed, closed-loop system which purifies the water, performs the cleaning task, and then recycles the water. These tend to be expensive, energy-hungry, and relatively slow in through-put.

For defluxing cleaning applications, the recommended DI water quality should be 1 to 10 Microsiemens-cm. Any conductivity that is higher than 10 Microsiemens-cm is not recommended.

The resistivity is the reciprocal of the conductivity. That afore mentioned range corresponds to 1 to 0.1 Megohm-cm. Any resistivity value below 0.1 Megohm-cm is not recommended either.

The resistivity of DI water is higher as 10⁸ Ohm. With a bubble equipment (additional carbonic acid) you can achieve a value smaller as 10⁷ or 10⁶ Ohm.

The problem is, do you use normal DI water, you can produce electrostatic charge on the PCB or electronic device (ESDS).

This electrostatic charge can damage your electronic device shortly or later in the final product. You cannot control the damage or degradation of ESDS.

The maximum resistivity of treated water should be 18,3 M-Ohms.

Classification for Electronic assembly water resistivity does exist. Should be able to find this on the internet for confirmation. I believe you will find three classes and grades if memory serves me right.

1. 18.3- 1.0,
2. Greater than 1.0 - 500
3. Greater than 500

Most aqueous cleaners with closed looped system will operate 18.3 to about 1 or 2 M ohms at which time the resin beds will be exchanged.

I hope this information helps. I would confirm the numbers as they could have changed. this was data points in the mid-90's

Resisitivity of a rinse water is a function of surface area being cleaned.

We rinse to a resisitivity of 200 Mohm's b/c that is the recommendation from our equipment supplier.

We also, checkfor active material in areas of the PCBA that are specific to damage. Example, we will spot check that there is no ionic activity in on or near QFPs and other areas that may have residual flux left behind.

Ionic contamination testing is required as ionic residues remaining on both the PCB manufacturing process and the soldering process may affect the reliability of a finished assembly.

A Contaminometer measures ionic contamination by essentially immersing a sample in a test solution of propanol and de-ionized water to dissolve the contaminants.

The dissolved ionic substances cause a change in conductivity of the test solution; this change is precisely measured and converted into a contamination value expressed as ug/cm2 NaCl equivalence.

In practice, it's impossible to completely remove all contaminants, so maximum threshold levels such as 0.2ug/cm2 NaCl equivalence are commonly accepted for modern densely packed assemblies.

Using high resistivity water has been the most effective cleaning with both batch and in line cleaning. We use 18.2 meg ohm water typically but change the mixed beds on recirculation at 10 meg ohm.

We clean 6 days a week with two shifts and we clean finished assemblies, with no clean flux, bare boards prior to any assembly and components prior to assembly.

We have found under proper cleaning conditions (time and temperature) the use of 18.2 meg ohm DI water will not damage, BGA, QFP, DIP, flip chip, bare boards with OSP, Tin Lead, SAC LEAD FREE HASL, SAC assemblies, tin plated parts.

I have a lot of data on cleaning if you need additional information please contact me directly.

I was recently involved on a project for wafer cleaning and they had a spec minimum of 14 megs

There is not a published specification applied to removing flux from populated assemblies. As a practical suggestion, it is important to consider why DI water is used.

Solvency:

DI water, with its near absence of ions is hungry. It wants to absorb ionic content.

Absence of solids:

Tap (drinking) water normally contains minerals. These minerals can leave residues behind as the water evaporates (like spots on a car). DI water may be mechanically removed via airknifes or rapidly evaporated (a common technique in batch-format defluxing systems) without concern for solids left behind.

Contrast:

Some automatic batch-format defluxing systems are equipped with real-time cleanliness testing capabilities. In this case, one must use DI water in the rinse cycles.

Because DI water is non highly ionic (it does not conduct electricity well), it makes a very effective contrast agent when compared to defluxing chemicals. Most defluxing chemicals are highly ionic and are easily detected in DI water.

Because modern batch-format defluxing systems test the electrical conductivity of each rinse cycle, even small trace quantities of defluxing chemical can be detected, forcing additional rinse cycles until there is an absence of detected chemicals in the rinse water.

Recommended Quality of DI Water:

DI water is available at qualities between 1 M-Ohm and 18 M-Ohm. While higher qualities (above 10 M-Ohm are common in the semiconductor industry, a quality of about 5 M-Ohm is adequate for all defluxing applications.

The cost of DI water averages about 6 cents per gallon. Based on this low cost, I would always recommend the use of DI water.