Biomedical Engineering Reference
In-Depth Information
Sensors monitoring optical density are sometimes also considered as NIR
probes, since the wavelength range used in many instruments is on the order of
900-1,100 nm; however, these sensors determine the intensity of just one wave-
length peak and are not spectrometers (e.g., [
88
]).
3.3.6 Orphaned Sensors and Methods
Several types of sensors or analyzers have not yet been discussed here but are the
subject of and are referred to in other papers. Some of these seem to have been a
short-lived hype and have faded away in the recent scientific literature and/or have
not made their way to commercialization for process monitoring.
One example is biosensors. These cannot be used inside the sterile barrier (in
situ) because they do not withstand proper sterilization. However, their use in a
process analyzer operated in bypass mode (and with the sample not returned to the
process but discarded) is feasible [
115
]. In spite of extensive studies in the past,
many authors have come to the conclusion that implementation of biosensors is
strongly hindered by their limited stability. Improvements are reported for a few
enzymes, but the results are obviously not very satisfactory: Vojinovic et al. [
116
]
stated in this context that operational stability allowing ''up to 8 h continuous
lactate conversion'' with a lactate oxidase-based sensor and virtually no activity
loss had been achieved. This is definitely less than the usual duration of a (batch)
production process. In another work [
117
], they investigated immobilized glucose
(GO), alcohol (AO), lactate (LO), galactose (GalO), and
L
-amino acid oxidases
(LAAO) together with horseradish peroxidase (HPR) and found that shelf-life was
as high as 6 months for GO/HRP, AO/HRP, and LAAO/HRP. After 1,400 and
8,000 FIA injections, respectively, GalO and LAAO had lost half of their original
activity. Katrlik et al. [
118
] employed sensors made with Gluconobacter oxydans
for 1,3-propanediol measurements, but those analyses were made manually after
sampling.
Another example is electronic noses (arrays of gas sensor elements) and elec-
tronic tongues (arrays of potentiometric or voltammetric elements). Although such
arrays have been used to characterize biologically important aspects such as the
metabolic burden of recombinant organisms or the indirect estimation of product
concentration [
119
,
120
], the number of such publications has decreased signifi-
cantly in the last decade. The background is probably that (a) one does not know
(exactly) what the sensor array elements do or could sense, (b) the limited stability
of the (gas sensor) array elements, and (c) the evaluation of the high-dimensional
datasets (with probably more than 1,000 array elements = dimensions) requires
multivariate regression models (chemometrics) which are not easy to build and to
validate. However, recently, an electronic nose was used to distinguish various
physiological forms of microbes, e.g., vegetatively growing and sporulating/
sporulated Bacilli [
121
], and electronic tongues have been used to monitor
methane [
122
] or to discriminate between species that can often occur as unwanted
food contaminants [
123
].
Search WWH ::
Custom Search