How our Sensors Work

The transducer - a quartz crystal microbalance

Brims Ness uses a quartz crystal microbalance sensor. A crystal, about the size of a thumbnail, is excited electronically, making it oscillate at 10 MHz, which is 10 million oscillations per second. When mass, or weight, is attached to the crystal, the oscillations slow down, the same principle that causes a thick guitar string to have a lower pitch than a thin string. This device is a transducer (Lat. "to lead across") because it converts the change in mass on the crystal surface to an electronic output.

The Receptors

Our chemical receptors are thin films that we coat on our crystals. Our targets are ions, which are charged particles. In table salt, which comprises sodium chloride, sodium, (Na⁺) is a cation and chloride (Cl^-) is an anion.

We have two types of receptors, exchangers and adsorbers.

Exchangers

As seen on the left, the blue ions are more preferred by the coating on the crystal than the green ions. If the blue ions are heavier than the green ions, the sensor will experience a frequency drop when the exchange takes place. The frequency drop is directly proportional to the mass increase.

Adsorbers (adsorb, to collect on a surface)

Some of our coatings attract ions to a surface of the crystal without any exchange. The frequency drop is directly proportional to the mass increase.

Reversibility

Our exchangers are reversible.High concentrations of the green ions will exchange with blue ions on the crystal. Also, in some cases, when a crystal is saturated with target ions on an adsorber sensor, we can flush the target from the crystal and reset to its starting position.

Interferences

An interference is an unintended frequency response. It may be caused by competing ions in the sample stream. In some cases, we remove the competing ions upstream in the sample flow using adsorbents.

In other cases we configure a series of exchangers so we can convert an "alphabet soup" of unknown ions into known quantities. Thus in the setup below, the red ion is the target ion. By situating two exchange columns upstream from a pair of crystals, we are able to determine the presence of a target ion in any alphabet soup of cations and anions.

Quantification

So far we've talked about how we identify the presence of an ion in a sample stream. We have two ways of determining the concentration of the target ion.

Calibration

A microprocessor in our sensor system calculates the concentration based on an algorithm developed in our labs at Brims Ness. That algorithm compares the frequency drop to a known concentration of the target ion. For example, the graph below shows the frequency response to a known concentration of phosphate. The graph shows a frequency response of a crystal over a 24 hour period. At the one hour mark, a 300 ppb (parts per billion) phosphate solution was introduced to a sample stream. (The sensor was configured as an adsorber). The slope, indicating the rate of change in the frequency is proportional to the concentration of the target in the sample stream.

Calculation

This method is appropriate for very low concentrations. We see above that we have a 2,000 Hz drop at 300 ppb phosphate. We get a clean, obvious signal at that concentration. If we run a 300 ppt concentration (parts-per-trillion), the slope is more gradual, for it takes longer for the ions to accumulate on the crystal. The concentration is 1,000 times more dilute. We expect that the response we see at 2 seconds above, will take about 2,000 seconds for the more dilute solution to get the same response.

The algorithm accounts for the more extended time period. We incorporate a metering pump to control the flow of the sample stream, so we know exactly what the denominator is in the parts per trillion calculation. We calculate directly the numerator that compares the frequency response to the mass of the accumulated target ions.

Other

Brims Ness has developed a new means of amplifying our signal at very low concentrations. We will introduce that invention in due course.

BNC patents US 5,990,684, US 6,232,783, others pending