[ID: PI7910-1155 / Seite 1]

Wissenschaft und Technik

Technik

Microbiological Study of a New

Design of PW/WFI
Pretreatment and Production System
Shlomo Sackstein
Biopuremax Ltd, Israel

The main challenge for the manufacture of Water for Injection (AC). Apart from the possible growth
(WFI) is meeting the maximum bacterial levels, since meeting the in the softeners and AC, all incoming
chemical criteria is usually easily attained. The microbiological bacteria from the municipality are
removed by the RO membrane and
criteria are very hard to meet as the regulatory maximum is less
concentrated in the reject compart-
than 10 cfu/100 ml. Internal action levels and/or alert levels must ment. This high concentration is
always be below 10 cfu/100 ml, so keeping reliably within specifi- conducive to biofilm growth with
cations is a very hard task. the associated reduction in RO per-
In the case study, the Electrolytic Scale Reduction (ESR) and Hydro formance.
Optic Dechlorination (HOD) are pretreating water feeding a hot In order to forestall out of specifi-
cation bacteriological results mainte-
water sanitizable Reverse Osmosis (RO) and hot water sanitizable
nance departments usually resort to
Continuous Electrodeionization (CEDI). This system, constantly an overkill approach and sanitize the
produced WFI quality water over a period of 3 years. This per- equipment on a weekly basis, with
formance is remarkable in achieving WFI standards, even though very little option to change as they
there was only a single membrane barrier. This is because of the are severely limited in the ability to
low level of bacteria in the pretreatment feeding the RO even predict the microbial growth over
different seasons with changing bac-
though there are very high levels in the feed water.
terial levels in the feed water.
A new system for pretreatment of
1. Introduction are used to design equipment which RO feed water in pharma plants will
can achieve the essential chemical be presented in this article. The sys-
Most modern and GMP Purified criteria. Given accurate and exten- tem actively reduces incoming bacte-
Water (PW) generation systems are sive feed water data, very reliable rial levels, freeing the RO from bacte-
based on Reverse Osmosis (RO) chemical results can be expected. rial ingress, while effectively disin-
membranes. The system has to Systems are designed to meet the fecting the feed water on its passage
meet both chemical and microbio- bacterial criteria in spite of highly through the pretreatment equipment
logical criteria to achieve operational unstable parameters in the feed into the membranes. This system op-
and regulatory goals. water. The final bacteriological per- erates without organic media and
The common chemical criteria to formance of a system is an amalgam without resins. There is no need for
be met in the RO pretreatment are: of: initial design, installation stand- chemical disinfection as the system
hardness reduction and chlorine re- ards, day to day operation and regu- is self-cleaning. This system operates
moval. The microbiological criteria lar maintenance. This also includes while consistently reducing the bac-
are a not to be exceeded total count feed water bacteriological levels and terial levels and preventing microbial
and absence of pathogens. microbial species, feed water TOC colonization of the equipment. The
The design of the systems, both levels and water temperature as system operates with no moving
pretreatment and production, must well as other parameters that are parts and without need of rinses or
meet the chemical criteria while tak- hard to define and measure. Because backwashes.
ing into account many common and of the wide variety of impacting pa-
well known design issues which de- rameters, it is very hard to predict in
pend on feed water parameters. The advance the propensity of bacteria to 2 . B ac k g r o un d
feed water parameters are easily as- grow on the organic media usually
certained by laboratory analysis. Well used in the pretreatment, e.g. soft- Water systems need to be operated
known and tested engineering rules ener resins and/or Active Carbon and maintained in a controlled man-

Pharm. Ind. 79, Nr. 10, 1–4 (2017)

© ECV • Editio Cantor Verlag, Aulendorf (Germany) Sackstein • New Design of PW/WFI 1
 [ID: PI7910-1155 / Seite 2]

Wissenschaft und Technik

Technik

ner. In addition, the source water has The regulatory authorities expect tor cylinder has the effect of raising
to meet potable water standards as the system owner to control all as- the pH, the H+ concentration on the
defined by the relevant organizations pects of the sampling of the system. central anode will have the effect of
in the US, EU, Japan or by the WHO. This includes both the chemical and lowering the pH. This effect will
As bacteria will proliferate if condi- microbial results [7]. In a media- cause some of the naturally occur-
tions are favorable, even low levels of based water system, the only way to ring chlorides in the feed water to
feed water microorganisms can lead achieve this control is constant dili- be activated as free chlorine.
to out of control downstream results gence and fast reactions in stamping The following formula denotes
[1]. If water microbial parameters ex- out microbial outbreaks. This is a this reaction:
ceed the potable water limits, after high, manpower intensive and time Cl- + H+ → HCL
entering the boundaries of the PW/ consuming effort. Microbial control HCl + HCl → Cl2 + H2
WFI generation, the operator will be in media based systems is limited This free chlorine by-product dis-
hard pressed to demonstrate control in efficacy and has the potential for infects the ESR and destroys biofilm
of microbial level in the system. long periods of system and plant as long as the system is in operation.
If the bacterial levels in the treat- downtime.
ment stages are increasing as the n 3.2 Removal of Oxidants in
water advances through the system, RO Feed Water by Way of UV
this demonstrates loss of control 3. System Based on Irradiation – Hydrodynamic
even if the product water is still Electrolytic Scale Reduction, Optical Dechlorination
within acceptable limits. H y d r o Op t i c D e c h l o r i n a t i o n , Hydrodynamic Optical Dechlorination
The US, Europe and WHO have Hot W ater Sa nitizable R O- (HOD)
standards of “zero” detectable “ob- E DI It is commonly known that UV
jectionable microorganisms” even radiation can destroy free chlorine
though they are sometimes detected n 3.1 Electrolytic Scale [9, 10]. This reaction is a fast and
in the municipal water [2, 3]. Carbon Reduction and Hydro Optic reliable process that removes the ox-
beds are the most prone to develop Dechlorination RO idizing form of chlorine and neutral-
high levels of microorganisms of all Pretreatment izes the possible damage that could
sorts. An FDA inspector observed An Electrolytic Scale Reduction occur to downstream RO and Con-
that “the carbon filter is probably the (ESR) is a unit that reduces scale so tinuous Electrodeionization (CEDI)
weakest element in any DI system from as to eliminate the potential of scale as the by-products of the reaction
the standpoint of microbial attack” [4]. precipitation on downstream RO are not oxidizing [9].
Bacteria can form mats and clumps membranes. The removal of scale The process for breakup of chlor-
on resin beads. The softener resin is precipitation in the membranes will ine in the common hypochlorous
organic and in conjunction with slow allow regular and continuous opera- acid form is given as follows:
flow and eddies, that are needed for tion of the RO with normal recovery 2HOCl → 2HCl + O2
proper ion exchange, proliferation of percentages. A typical disinfection UV will gen-
unwanted growth is almost inevita- The system core has arrays of erate at the end of the lamp life a
ble. Control of this biofilm is possible Stainless Steel (SS) reactors equipped dosage level [11] of 30,000 mJ/cm2.
but only by considerable investment with central electrodes. When an The dosage levels needed for removal
in routine maintenance. The needed electrical field is effected between of chlorine depend on the free chlor-
maintenance can range from back- the electrode and the cylindrical re- ine concentration. If the HOD is
wash and regeneration to sanitiza- actor a current flows through the specified to remove up to 1 ppm of
tions and rebeddings. water. Some of the water dissociates free chlorine, the dosage levels
Extensive guidance for microbial into OH- and H+ ions [8]. needed are in the range of
sampling can be found in the ISPE The OH- concentration on the in- 1,500,000–1,800,000 mJ/cm2 [12]. The
Good Practice Guide: Sampling for side of the reactor cylinder has the very high level of UV radiation will
Pharmaceutical Water, Steam and effect of raising the pH and causing effectively dechlorinate the water
Process Gases [5]. Areas that are wor- precipitation of hardness scale. Some and, as a welcome by-product, will
thy of attention were singled out. of the scale adheres to the cylinder also disinfect with a large over kill
These are: deionization steps, carbon and some will drop to the bottom of of more than 50 times.
filters, depth filters, RO, Ultrafiltration the reactor.
(UF) and the final deionization proc- The scale builds up in the reactor n 3.3 ESR-HOD-RO-EDI
ess. It was recommended to perform until removed manually or automati- Integrated System
species identification of the resident cally. All the above components can be in-
microflora and build up a picture of In the same way that the OH- con- tegrated to generate PW. The ESR
the local microbial population [6]. centration on the inside of the reac- will reduce the scale in the feed

Pharm. Ind. 79, Nr. 10, 1–4 (2017)

2 Sackstein • New Design of PW/WFI © ECV • Editio Cantor Verlag, Aulendorf (Germany)
 [ID: PI7910-1155 / Seite 3]

Figure 1: System Flow Diagram (Source: The figure was made by the author/Biopuremax Ltd).

water. It will also generate free chlor- culated from the municipal water The ESR and HOD technologies
ine as a by-product of the precipita- feed all the way to the PW/WFI stor- are no more susceptible to bioburden
tion process in the reactors. This free age tank without leaving any areas than SS pipes. All the surfaces are
chlorine is destroyed by the HOD that have not been heat sanitized. manufactured of SS or quartz and
that is installed in series to the ESR. are easily hot-water sanitizable. As
The dechlorinated and scale reduced n 3.4 Continuous Bioburden the ESR will also generate free chlor-
water can now be utilized as feed Reduction (CBR) ine, the system will actively reduce
water for a conventional RO and As ion exchange, RO and UF are all bioburden. The HOD destroys the
EDI system (Fig. 1). particularly susceptible to microbial free chlorine with high doses of UV
The RO and EDI will remove the contamination, it is essential to con- irradiation which also disinfects the
ionic components and soluble or- sider mechanisms for microbiologi- water. When the system is returned
ganic compounds so as to meet PW cal control [13]. to operation after a filter or UV lamp
monograph standards. Water for In- A well designed process should be replacement, a sanitizing cycle will
jection can be generated by adding a capable of controlling bioburden and remove contamination that might
second stage to the RO or by adding endotoxins from beginning to end. have been introduced by exposing
final Ultra Filtration (UF). This is true of water systems as well the system to the outside environ-
For startup or after maintenance, as of any other production system. ment.
a sanitizing cycle is needed. As in
traditional designs, chemicals can
be used for sanitization, but a hot
water sanitization can also be uti- n Table 1
lized as the ESR-HOD combination Site Data for Bioburden Criteria.
is fabricated of SS. There are many
advantages of the hot water wash Total Micro E. COLI Pseudomonas Coliforms Fungus
over the chemical sanitization. The CFU/100 ml CFU/100 ml CFU/100 ml CFU/100 ml CFU/100 ml
Average Average Average Average Average
hot water sanitization will penetrate
biofilm with ease in comparison to Product 0 0 0 0 0
Water
chemical disinfectants [11], with the
added advantage that with hot water WFI Criteria < 10 CFU/ < 1 CFU/ < 1 CFU/ < 1 CFU/ < 1 CFU/
100 ml 100 ml 100 ml 100 ml 100 ml
there is no need to validate chemical
removal before restarting the system. Number of 141 141 141 141 141
Samples
In addition, the hot water will be cir-

Pharm. Ind. 79, Nr. 10, 1–4 (2017)

© ECV • Editio Cantor Verlag, Aulendorf (Germany) Sackstein • New Design of PW/WFI 3
 [ID: PI7910-1155 / Seite 4]

Wissenschaft und Technik

Technik

n Table 2 n LITERATURE
Site Data for Chemical Criteria. [1] USP 38 General chapter <1231>.
[2] https://safewater.zendesk.com/hc/en-us/ar
Endotoxin TOC (ppb) Heavy Metals Nitrate < 0.1 mg/l ticles/214284228-5-What-are-EPA-s-drink
(EU/ml) Average (ppm) Average ing-water-regulations-for-pathogens-and-
indicators-
Average Average [3] Guidelines for Drinking Water Quality,
Product Water < 0.005 < 50 < 0.1 < 0.1 Third Edition, Volume 1, Geneva 2008,
Table 5.2, Table 7.7.
WFI Criteria < 0.25 < 500 < 0.1 < 0.2 [4] Theodore H. Meltzer, “High purity water
Number of 33 Online 33 32 preparation, for the semiconductor,
Samples pharmaceutical and power industries”,
page 347, 1993.
[5] ISPE Good Practice Guide: Sampling for
Pharmaceutical Water, Steam and Proc-
ess Gases, 2.3.5.6 Performance Qualifica-
4 . Ca s e S t ud y systems do not withstand high bio- tion Sampling, Table 2.2, ISPE 2016.
burden feed water and are liable to [6] PDA, Technical Report #69: “Bioburden
and Biofilm Management in Pharma-
An ESR-HOD-Single pass RO-EDI has grow pathogens and to contaminate ceutical Manufacturing Operations”,
been operating for over 3 years. Ta- downstream filters and RO. Paragraph 4.2.3.1, 5, 2015.
ble 1 summarizes the microbiological This new system stops prolifera- [7] Dr. Lynne A. Ensor, Director, Division of
Microbiology Assessment, US FDA, “Mi-
results within this time period. The tion of biofilm and pathogens and crobiological Quality Considerations in
results meet WFI criteria as they are meets the pharmaceutical market Non-Sterile Pharmaceutical Product
below detectable limits. demands of a well-designed pretreat- Manufacture”, PDA Microbiological
Seminar, Porto, 14 Feb. 2017.
Note: All microbial counts are ment capable of controlling biobur- [8] Nissan Cohen, Shlomo Sackstein “Chem-
averages and are all zero. Over the den from beginning to end. The ESR- ical and Media-Free Pretreatment for Bi-
full sampling period, there was not HOD pretreatment for pharmaceuti- opharma RO”, Pharmaceutical Engineer-
ing, Vol 34, No 4, July/August 2014.
one positive colony forming unit cal water is simple while providing [9] Barry Collins, Gary Zoccolante, “Dech-
(CFU). The results are shown at a effective results. lorination in Pharmaceutical Water Sys-
100 ml resolution. This system operates without tems”, Pharmaceutical Engineering,
Feb. 2007, Volume 4, Issue 3.
Table 2 summarizes the chemical chemicals, media, or resins, and [10] Uri Levy, Ph.D. and Ori Demb, “Queries
characteristics of the water in the eliminates the need for: regenera- Regarding Short-Wavelength Dechlorina-
same 3 year period. The endotoxin tion, complicated instrumentation tion”, Internal Documentation, Atlan-
tium, Oct. 12, 2010.
average is below detectible limits, and feedback loops. The system op- [11] Theodore H. Meltzer, “Pharmaceutical
the online TOC was always below erates without moving parts and Water Systems”, page 169, 113, 1997.
50 ppb, heavy metals and nitrate without need of rinses or back [12] Dr. Mark Wilf, “Alternative Dechlorina-
tion Methods in Reverse Osmosis Appli-
were below the WFI criteria. In short, washes. The CBR, demonstrated in cations”, Pharmaceutical Engineering,
the system met WFI criteria for a the case study, is exactly what is Sept./Oct. 2013, Volume 33, Number 5
period of 3 years without any Out needed to supply long term low lev- [13] Annex 2: “WHO Good Manufacturing
Practices: Water for Pharmaceutical use”,
of Specification (OOS) occurrences. els of bacteria in the system. Stand- WHO Technical Report Series No. 970,
ard SS fabrication materials facili- 2012.
tate hot water sanitization and serve
5 . Co n c l us i o n to “reset” the bacteriological levels
on a regular basis. Correspondence:
Pretreatment systems for RO have In the case study, the combination Shlomo Sackstein
remained almost unchanged for the constantly produced WFI quality Biopuremax Ltd
last 3 decades. The industry demands water, even though there was only a 4 Hasdnaot
a new system, in which water is not single membrane barrier. This is be- POB 12644
wasted and no waste stream is gen- cause of the low level of bacteria in Zip 4672831 (Israel)
erated as a by-product. The existing the pretreatment feeding the RO. e-mail: [email protected]

Pharm. Ind. 79, Nr. 10, 1–4 (2017)

4 Sackstein • New Design of PW/WFI © ECV • Editio Cantor Verlag, Aulendorf (Germany)