Analytical Advances: New Techniques for Detecting Leaks

While clinical trial outcomes and FDA's safety and efficacy criteria are more often in the public eye, GMP requirements for biopharmaceutical processing — including site inspections, facility audits, and records compliance — are equally important and just as strict.

To help ensure purity, GMP requirements dictate that the vessels in which recombinant proteins or small molecules are mass-produced, transported, and refined must be frequently checked for sterility. Unintended bacterial or other contamination in commercial scale vessels can result in lost product, lost monies, and a possible FDA enquiry. A worst-case scenario involves the contamination going unnoticed until it causes an adverse reaction in a patient. While these situations have occurred in the past, they are almost completely avoidable through accurate and consistent process monitoring.

At issue in almost every case of batch contamination is the vessel's integrity. Vessels often have many inlets along the walls to accommodate stirrers, process instrumentation, vents, and fill-and-drain connections. To help ensure that the bioreactors and fermentors are emptied of residue between batches, the vessels are usually drained, cleaned, then steamed before being sealed and charged for the next process run. During this cleaning process, nearly infinitesimal leaks can be created when valves are operated and clamp fittings are opened and closed. Age also plays havoc with vessels. And as valves and other seals wear over time, they can develop leaks too. While some vessel perfusions are visible and can be repaired, others may go unnoticed because they are too small, located along a seam, or hidden from sight. With leak paths unchecked and unrepaired, a product can be compromised, resulting in the loss of the campaign and up to hundreds of thousands of dollars in product and production hours.

Currently, vessel manufacturers do not have an absolute or uniform methodology for testing finished vessels' integrity and certifying them "leak tight." Similarly, biotech and pharmaceutical companies using vessels and piping for fermentation, compounding, transport, and other process operations don't follow a universally recognized integrity standard or leak testing method. Many leak rate standards are developed in house and vary from company to company and even from one manufacturing site to the next.

A variety of leak detection methods are commonplace, but industry experience shows that current leak detection methods do not adequately or consistently ensure that all leak paths will be identified before a batch run. However, a new leak detection method based on the use of inert helium tracer gas and a portable leak detection device produced by Varian, called HeliTest, offers an improved method.

Current Methods

Pressure Decay.

Industry has traditionally used pressure decay and visual methods (such as the use of soap and water solutions) to detect leaks in process vessels. In the pressure decay method, the empty vessel is charged with several pounds of pressure of clean, dry air. The vessel is then isolated, and technicians determine the amount of pressure lost over a given time period using pressure gauges. In this case, guidelines — usually established in house and with company-to-company variability — are used to determine how much pressure loss, representing the sum of the vessel's leaks, is acceptable over the test period. Anything below a determined value (based on manufacturing managers' experience) is deemed "acceptable" because these levels previously have not resulted in lost product due to bacterial or other contaminant infiltration. This testing methodology has several drawbacks, including susceptibility to temperature fluctuations resulting in skewed pressure reading, operator error, and gauge inaccuracy (gauges are typically low-accuracy Bourdon tube types). Using this method, a leaking tank is likely to be identified, but determining the source and location of the leak is nearly impossible.

Visual Detection. In visual leak detection, a soap and water solution is typically employed to reveal the source of a leak. The empty vessel is pressurized before testing. Then, technicians squirt the solution at suspected leak sites around the vessel. Leaks will be confirmed as the compressed air escaping from the vessel forms bubbles in the solution.

The drawbacks with this test method include a reliance on visual identification, the inability to find the smallest leaks, inconsistent testing methods by technicians, and inadequate application of the soap and water solution to tank penetrations in the horizontal plane (that is, the solution runs off the vessel before bubbles can form).

Ammonia. A less common method of leak detection is introducing ammonia into the process vessel before charging it for the process run. The operator, using test kits, can identify ammonia gas escaping through leak paths. In addition to the safety issues of working with ammonia, there are concerns about this methodology's ability to produce accurate and repeatable results because it depends on visual interpretation of color variation. And because the color change does not appear immediately, it is also difficult to find the precise location of the leak. Additionally, ammonia vapors may contaminate the vessel, necessitating an additional cleaning step.

The Helitest Method

The lightweight HeliTest can detect leaks that are not visible, it does not require the use of toxic chemicals, and it is portable and simple to operate. The HeliTest is based on a technology called selective ion pump detection (SIPD). It detects helium based on a patented technology using the physical properties of quartz membranes and a unique ionizing process of gases that penetrate the membrane.

Before testing for leaks, the vessel must be emptied and then filled with helium — the vessel is typically pressurized to approximately two atmospheres (absolute), with a 25% concentration of helium. To use the HeliTest (which can be worn on a neck strap) the operator aims the attached flexible wand sequentially at all potential leak sites using the "sniffer" technique. The wand is attached to an internal sampling pump that continuously draws in sample gas. Inside the HeliTest, this pump discharges the sample over a heated silica capillary impermeable to all gas species except helium. Then, the helium molecules passing through the capillary are ionized. The resulting ions are "counted" and generate an electric signal allowing a read-out of helium concentration on the display. The leak rate reading is proportional to the quantity of helium that becomes ionized as it escapes the vessel and is drawn into the sampling wand. The HeliTest provides both a visual and audio indication of a leak and its relative size.

Helium is inert, inexpensive, and occurs in low levels — 5 ppm — in the atmosphere, so it is an ideal tracer gas for leak detection. Because only helium can penetrate the silica capillary, the HeliTest only measures helium concentration; therefore, leak readings represent actual process vessel leaks. The HeliTest can measure helium concentrations as low as 2 ppm. Generally, readings of 20 ppm or higher represent leaks requiring immediate attention.

The HeliTest has been used in manufacturing and production settings for over three years with no reports of product loss attributable to leaks not found by this instrument. BPI