TIBTEC-1074; No.

of Pages 9

Review

On-line monitoring of large

cultivations of microalgae and
cyanobacteria
Ivo Havlik1, Patrick Lindner1, Thomas Scheper1, and Kenneth F. Reardon2
1
Leibniz University Hannover, Institute of Technical Chemistry, Callinstrasse 5, 30167 Hannover, Germany
2
Colorado State University, Fort Collins, CO 80523-1370, USA

Large cultivations of microalgae will benefit from on-line processes based on each of these options are being
monitoring to achieve process control and improved commercialized.
productivity. This monitoring requires reliable sensors  Owing to the production economics, microalgae are
for on-line, in situ measurement of both physicochemical grown at scales much larger than is typical for other
and biological process variables. Although standard in- cultivation processes (e.g., ponds of 100 ha or larger [5,6]).
dustrial sensors can be used for many physicochemical  Mixing in these large systems is slow and thus
variables, monitoring methods for most biological quan- important process parameters vary spatially.
tities rely on sensors that are currently suitable only for  Because it is not possible to maintain an axenic culture
laboratory scale or off-line use. Here, we review these in such large systems, there is interest in monitoring the
methods and discuss new approaches that could be species composition of a culture to track the level of both
adapted. We suggest that these new methods should beneficial (e.g., synergistic bacteria) and harmful (e.g.,
be noninvasive and based on approaches that have rotifers) contaminants.
already been applied to other bioprocesses; examples  The product is often intracellular and produced in the
discussed here are in situ microscopy, flow cytometry stationary phase.
(FC), IR spectroscopy, and software sensors.
Ideally, measurement devices for microalgal cultiva-
Monitoring needs for cultivation of microalgae tions would be on-line, rapid, sterilizable, stable, and
The use of photosynthetic microorganisms, namely micro- selective, and these methods would need at most infre-
algae and cyanobacteria, as feedstocks for the production of quent calibration. These are the same objectives demanded
fuels and chemicals has received widespread attention in for monitoring bioreactors in other areas of biotechnology;
the past few years [1–4], and microalgae have been cultivat- therefore, reviews of those technologies are recommended.
ed commercially since the 1960s [5]. Considerable efforts For example, existing in situ sensors for bioreactors for
have been devoted to research, development, and commer- disposable sampling units have been recently reviewed [7].
cialization of many aspects of the production process, in- In this review, we discuss which cultivation parameters
cluding strain selection and engineering, cultivation are of interest and survey the currently available monitor-
systems, harvesting, and conversion of the lipid, starch, ing options as well as measurement methods that could be
or whole biomass to the desired products. Some of the basic developed or adapted for the monitoring of large microalgal
features of microalgal cultivations are summarized in Box 1. cultivations. We emphasize on-line measurement methods
However, monitoring of these microalgal and cyanobacterial because they can be automated to provide information for
cultivations has largely been overlooked. process control. The advantages and disadvantages of these
Although new developments continue to emerge, culti- methods, along with their application scope, robustness, and
vation monitoring is relatively well established in indus- suitability for on-line implementation are discussed.
trial, food, and pharmaceutical biotechnology. However,
the cultivation of microalgae and cyanobacteria (hence- Process variables in microalgal cultivations
forth referred to simply as algae) presents several new The parameters of interest for bioprocess monitoring can
challenges that will require both the careful evaluation of generally be categorized as physical, chemical, or biologi-
existing monitoring approaches and the development of cal, as well as by the phase in which they are measured
new methods. Specifically: (gas, liquid, and biomass) [8]. As illustrated in Figure I in
 Many microalgal strains can grow phototrophically, Box 1, the most important of these for microalgal cultiva-
heterotrophically, and mixotrophically, and cultivation tions are:
 Physical: light in the photosynthetically active radiance
Corresponding author: Reardon, K.F. ([email protected]). (PAR) range; temperature; mixing intensity.
Keywords: bioprocess monitoring; microalgal cultivation; in situ microscopy;
optical sensors.  Chemical: dissolved CO2 and oxygen; pH; nitrogen,
0167-7799/$ – see front matter
phosphorus, and other nutrient concentrations; extra-
ß 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tibtech.2013.04.005 cellular product concentrations.
Trends in Biotechnology xx (2013) 1–9 1
TIBTEC-1074; No. of Pages 9

Review Trends in Biotechnology xxx xxxx, Vol. xxx, No. x

Box 1. Microalgal cultivation

As photosynthetic organisms, microalgae require for their cultiva- open systems [9,12,86,87]. In both reactor types, CO2 sources may be
tion light, suitable temperature and pH, and a carbon and nutrient sparged into the liquid medium [88]. In most cases, the desired
supply [4]. The most important growth factor for phototrophic product is intracellular, and thus after the desired biomass or product
microalgae is light, which supplies the energy necessary for level is achieved the cells must be harvested and extracted [3].
microalgal biomass production. The energy supplied by light to
an individual cell depends on several factors: photon flux density,
cell density, optical path length, and mixing intensity. In full
sunlight (about 2000 mmol photons/m2/s; [16]), about 90% of
photons captured by photosynthetic pigments are dissipated as
heat and fluorescence. The second most important process Photon
parameter is the CO2 used as carbon source. About 1.7–3 g CO2/g flux CO2 O2
biomass is necessary for growth biomass [9,15]. The third most
important parameter in microalgal growth is temperature; the
allowable and optimal ranges depend upon the strain but are
typically 15–35 8C and 25–30 8C (freshwater algae), respectively.
Temperatures above 35 8C are lethal for many species [4,22].
Nutrients (primarily nitrogen and phosphorus) are necessary for pO2 Photosynthec efficiency
growth, and their depletion leads to accumulation of reserve pCO2
Chlorophyll, other pigments
substances as lipids, starch, and pigments [85]. These substances pH
T Biomass composion
are often the desired products of the cultivation (in some Cell morphology
[Cells]
instances, the goal is simply biomass production), therefore, [Nutrients] Populaon composion
management of nutrient levels is critical.
Many different microalgal cultivation systems have been described
in the literature [9,86]. In general, these are of two types: open TRENDS in Biotechnology
systems, such as ponds and raceways, and closed PBRs. PBRs allow
the cultivation to be performed with greater control and can achieve Figure I. Important process variables to monitor in the gas, liquid, and biomass
phases of microalgal cultivations.
higher productivities, but have much higher capital costs than the

 Biological: biomass concentration; biomass composition; inclined PBR, in which measurements of incident irradi-
other species; physiological state; photosynthetic effi- ance, temperature, pH, pO2, pCO2, CO2/gas flow rate, and
ciency (PE); solar-to-biomass conversion efficiency; cell CO2 concentration with an IR gas analyzer have been
morphology. reported [15].
Sensors used for monitoring these quantities are sum-
These variables influence the process performance and marized in Table 1 for physicochemical variables and in
can be impacted by design and initial process settings. Table 2 for biological variables.
However, they are most effectively adjusted via active pro-
cess control informed by on-line or off-line measurements.
For microalgal cultivation systems based upon photo- Physicochemical variables
trophic growth, the provision of light is perhaps the most Light intensity. The optimal light intensity in the PAR
important process consideration. The incident light inten- (400–700 nm) range for most microalgae is 10–250 mM
sity at the reactor surface and the light intensity profile photons/m2/s; as a reference point, direct sunlight yields
within the cultivation medium are dependent upon the up to 2000 mM/m2/s [16]. PAR constitutes energetically (in
environment and the design of the photobioreactor (PBR) W/m2) about 40% of the visible sunlight [17]. The decisive
or pond/raceway [9]. Furthermore, the light intensity pro- value affecting microalgal growth is the light intensity
file changes during a process because of the influence of the inside the culture, not the intensity impinging on the surface
optical properties of the microalgal suspension. Irradiance of the culture medium, because the latter is attenuated
is characterized by the photon flux density (PFD) imping- before reaching a microalgal cell by absorption and by
ing on the surface of the cultivation medium, the light shadowing by other cells. The PFD is usually measured
penetration depth, the light intensity profile, and the on the liquid (pond) or outer (PBR) surface, and light pene-
light–dark cycle frequency [10–12]. tration into the microalgal suspension is subsequently mod-
eled to obtain the desired irradiance profile within the
Available technologies for on-line monitoring bioreactor [18–20]. Measurements inside a PBR are rare
The process variables that can be monitored on-line with [21]. For open outdoor PBRs, a model based on the Lambert–
current technologies are primarily the physicochemical Beer law with an experimentally determined extinction
variables in the liquid and gas phases. Some biological coefficient has been used [10]. The PFD is usually measured
variables can be monitored on-line as well; these include for the PAR range by a planar (2p) or a spherical (4p)
biomass concentration and the concentrations of chloro- quantum sensor [22]. It is important to note that the light
phyll and other pigments. These measurements can be intensity (and the duration of that intensity) that reaches an
used to derive values of the growth rate, PE, and photo- algal cell is a result of a complex interplay among mixing,
synthetic quantum yield (PQY). cell concentration, and incident light intensity [9,23].
Owing to their larger size, measurements in ponds and
raceways are usually limited to incident irradiance, tem- Temperature. Microalgae have a temperature optimum
perature, pH, pO2, and nutrient concentrations [13,14]. that depends upon the species. Temperatures above the
The most instrumented outdoor PBR type is the thin-layer optimal range should be avoided [4,22] as they cause a
2
TIBTEC-1074; No. of Pages 9

Review Trends in Biotechnology xxx xxxx, Vol. xxx, No. x

Table 1. Physicochemical variables monitored and controlled on-line in microalgal cultivations

Monitored variable Sensor type Acceptable range What happens when Control options Refs
value out of range
Photon flux density  Quantum sensor: 10–250 mM/m2/s  Low: slow growth  PBR design [20–22,37,89,90]
(for a single cell) flat cosine, fiberoptic (optimal);  High: PE sinks  Culture density
spherical, PAR dosimeter, 0–2000 mM/m2/s at PFD >250  mixing
integration solarimeter (actual)  High: Photoinhibition
at >1500
Temperature  Thermoelement (Pt-100) 15–35 8C  Low: slow growth  Water spraying [14,16,20,21,
 High: culture death  Water bath 24,37,40,42]
 Heat exchanger
 Shading
pH pH glass electrode 7–10  Growth rate decrease  CO2 injection [13,14,16,17,
Optical pH sensor 21,57,89]
pO2 (liquid phase) DO electrode (Clark) <15–25 mg/l  High: growth rate  Aeration [20,21,30,31]
Optical pH sensor decrease
Oximeter
O2 (gas phase) Paramagnetic analyzer Depends on O2(l)  High: growth rate – [15,36]
Polarometric analyzer and mixing intensity decrease
pCO2 (liquid phase) pCO2 electrode >0.1 kPa  Low: growth rate  CO2 injection [26,30,31]
IR analyzer + flow meter decreases below 0.1 kPa
CO2 (gas phase) IR analyzer >0.15%  Low: growth rate  CO2 feed [27,30–32,91]
Mass spectrometer decreases
Inorganic nutrients UV spectroscopy Varies with nutrient  Low: growth limitation;  Nutrient addition [39,92]
Colorimetric assays lipid or starch
Ion-selective electrodes accumulation
(modified)
Mixing None Re <6000–10 000  Low: CO2 limitation/O2  Gas addition rate [18,86]
inhibition  Agitation intensity
 High: mechanical
damage to cells

large drop in photosynthetic quantum yield [17]. In larger the purge air or by using flue gas from a source such as a
cultivation systems, shading, immersion in a water tank, fossil fuel combustion process [15]. Carbon is the dominant
or water spraying (evaporative control) is used. Shading, nutrient in terms of amount: theoretically, at least 1.65–
however, leads to diminished yields [24]. In raceways and 1.85 g CO2/g biomass must be supplied [9,15]. The CO2
ponds, temperature is rarely controlled. Temperature is partial pressure in the liquid phase should be kept above
usually measured by a standard industry Pt-100 sensor. 0.1–0.2 kPa to avoid limitation [15]. CO2 in the liquid
phase can be estimated from the equilibrated gas-phase
pH. In photoautotrophic microalgal cultivations with light CO2 concentration using an IR analyzer [30]. Direct mea-
intensities above the compensation point (at which the rate surement of CO2 in the liquid phase using currently avail-
of photosynthesis is equal to the rate of respiration), the able pCO2 electrodes can be problematic because of the
culture pH will rise if not controlled. The pH value influ- frequent maintenance and recalibration that is required
ences the growth rate, both directly because each micro- for these sensors. CO2 control is usually a byproduct of pH
algal species has an optimal range pH for growth [22], and control where CO2 flow into a PBR is measured using a
indirectly, by affecting nutrient availability [25]. The ac- mass flow meter and dosed by pH controller [17,21,31]. CO2
ceptable pH range for most cultivated microalgal species is can also be measured in the outlet gas with an IR analyzer
7–9, with a typical optimum about 8.2–8.7. Some species [31,32] and CO2 dosed accordingly [30].
are adapted to more acidic or basic environments [26]. pH
is measured using glass electrodes or optical pH probes, Oxygen in liquid and gaseous phases. As the product of
and is normally controlled by injection of CO2 into the photosynthesis, oxygen exerts an inhibitory effect on the
culture or into the inlet gas. pH control using mineral acids culture growth rate starting at oxygen concentrations
may have advantages in some cases. More sophisticated above 200–500% of saturation value, that is, above 15–
control algorithms for pH have been investigated with the 25 mg/l O2 [9]. The dissolved oxygen concentration is rec-
aim to enhance the efficiency of CO2 utilization [27,28]. ognized as a reliable and sensitive indicator of the micro-
Improvement of pH control using a control algorithm algal culture state as related to growth and productivity.
reduces pH gradients in the culture, increasing the average Oxygen is usually continuously removed by sparging the
photosynthesis rate and biomass productivity [28]. culture with CO2-enriched air, thus combining oxygen
removal, CO2 supply, and pH control in one step. Oxygen
CO2 in liquid and gaseous phases. A steady supply of CO2 in the liquid phase is measured by a Clark electrode,
is necessary for photoautotrophic growth of microalgae. optical oxygen sensor [33], or oximeter, either directly in
Relying on ambient CO2 in open ponds leads to growth the cultivation medium [20,31,34,35], or in the degasser or
limitation due to low CO2 diffusion rates from the atmo- other bypass [21]. The sensor measurements are then used
sphere [29]. CO2 can be supplied by injecting pure CO2 into to adjust the air or other gas flow rate into the culture.
3
TIBTEC-1074; No. of Pages 9

Review Trends in Biotechnology xxx xxxx, Vol. xxx, No. x

Table 2. Monitoring methods of biological variables in microalgal cultivations with on-line use or on-line potential
Monitoring method Monitored variable Sensor type Off-line/On-line Comment Refs
(concentration)
OD, turbidity  Cell mass concentration  OD sensor Off-line and on-line Wavelength choice [31,34,40–43,47]
 Turbidity sensor depends upon pigments
Color analysis  Cell mass concentration  CCD camera Off-line Data analysis software is [20,35,48]
critical
ISM  Cell number concentration  Microscope + CCD On-line Image-analysis software [33,50,52–54]
 Cell mass concentration is critical
 Cell morphology
 Population composition
(contamination)
Absorbance  Pigments  Spectrophotometer Off-line Fatty acid levels may be [58,59]
spectrum  Fatty acids estimated by correlation
to pigment ratio
Fluorometry  Photosynthetic efficiency  Pulse amplitude Off-line and on-line PAM identifies stress [9,55,56,90,93,94];
 Quantum yield modulated leading to lipid production YSI, Qubit, bbe
 Lipids fluorometer onset Moldaenke
 Pigments
IR radiation  Lipid  ATR flow system Off-line and on-line Data analysis software is [64,67–70]
(MIR, NIR, FTIR)  Protein  Fiberoptic probe critical
 Carbohydrate content
FC  Lipid content  Flow cytometer Off-line and on-line Sample processing [74–76,79,95]
 Cell size necessary for lipids and
starch

Oxygen can also be measured in outlet gas with a para- Microalgal concentrations have also been estimated by
magnetic [36] or polarometric gas analyzer [15]. software sensors using measurements of oxygen [44], local
irradiance [45], color images of light distribution in a PBR
Nutrients. Microalgae require many inorganic ions for [46], and a hardware sensor using four monochromatic
growth, the most important of which are nitrate or ammo- photodiodes with artificial neural network (ANN) pattern
nium, phosphate, and iron. Nitrogen and phosphate are recognition [47]. The macroscopic color intensity of a micro-
also of interest for monitoring because their limitation is algal suspension has been correlated with biomass concen-
used to induce lipid accumulation [37]. Off-line measure- tration [48], although it should be noted that other factors
ment of inorganic ions by ion chromatography and colori- might alter the culture color.
metric assays is common, but there are also technologies
for on-line (based on UV spectroscopy) or at-line (automat- Cell count, cell morphology, and contamination. Off-line
ed assays) measurement that have been developed and counting of microalgal cells by analysis of microscopic
commercialized, initially for wastewater treatment pro- images was first reported in 1989 [49] and complete com-
cesses. Ion-selective electrodes (ISEs) are also available; mercial systems are now available that process culture
although these are generally not robust enough for long- samples. The microalgal cell count was found to correlate
term, on-line use, a modified ammonium ISE has been used with the macroscopic color intensity of the microalgal
for monitoring bacterial cultivation for 7 days [38]. An suspension in a cultivation flask [48]. Recently, an on-line,
‘electronic tongue’, an array of ISEs with data analysis flow-through in situ microscopy (ISM) system featuring
by artificial neural networks, has been developed for the real-time image analysis was developed for monitoring
measurement of ammonium, potassium, sodium, chloride, biomass-related parameters (cell number concentration,
and nitrate [39]. cell morphology, and cell size distribution) as well as
protein crystallization and the stability of enzyme carriers
Biological variables [33,50–53]. This system has recently been adapted to
Biomass concentration. Biomass can be estimated on-line microalgal cultivations by adding a flow-through cell that
using optical density (OD) measurements at various wave- can be cleaned of adhering microalgae by temporarily
lengths, calibrated by off-line dry weight or microscopic cell increasing the flow rate [54]. The flow cell is installed in
counts. Care must be taken in selecting the measurement a bypass loop. In addition, the image processing algorithms
wavelength and in determining the calibration because the are adjusted for recognition of microalgal cells, because
pigments and organelles of microalgae can affect light each cell type requires a specific algorithm.
absorbance significantly, and the cellular composition is The potential of this device is illustrated in Figure 1, in
likely to change with time. The use of a commercial tur- which a cell count computed on-line by ISM in a Chlamy-
bidity sensor [34,40] and a self-constructed near IR (NIR) domonas reinhardtii cultivation is compared with off-line
absorbance sensor [41,42] has been reported. OD measure- OD measurement. ISM can furnish automated information
ment can also be combined with fluorescence measure- about cell count and cell morphology in numerical form,
ments [31,43] or in combined sensors (e.g., the EXO and can also be used as simple operator support by contin-
sensor line from YSI Inc. for water quality monitoring). uously supplying raw images for visual evaluation.
4
TIBTEC-1074; No. of Pages 9

Review Trends in Biotechnology xxx xxxx, Vol. xxx, No. x

70 Methods with potential for on-line monitoring

On-line cell count (cells / image)

60 5 Although existing sensor systems can provide on-line mea-

surements for the physical parameters of interest in micro-
50 4
algal cultivations, monitoring systems for important
40 chemical and biological quantities are not nearly as well

OD550
3
30 developed. This is largely due to the high degree of com-
Key: Cell count (on-line) 2 plexity of cells and their chemical environment. A variety
20
12h asym. median (on-line)
1 of analytical methods are available for assay-type mea-
10 Opcal density (off-line)
surements of cells, but most of these require sample pro-
0 0 cessing and cannot readily be adapted to process
0 50 100 150 200 250 300 350
Culvaon me [h] monitoring. The most promising methods for translation
TRENDS in Biotechnology to on-line monitoring are those based on the measurement
of optical properties. These approaches include measure-
Figure 1. A comparison of off-line cell density measurements with data computed
from in situ microscopy (ISM) analysis during a cultivation of Chlamydomonas
ment of fluorescence and the use of visible and infrared
reinhardtii demonstrates the ability of ISM to provide accurate data without spectroscopy, and can be used to estimate concentrations of
sampling. The algae were grown in a 4.4-l photobioreactor sparged with 3% CO2 in pigments, lipids, and biomass [7,60,61]. It may also be
air at 1 vvm. The temperature was controlled at 26 8C. OD was measured in a
Uvicon spectrophotometer at 550 nm. Reproduced from [54].
possible to adapt FC for on-line monitoring to enable the
measurement of cell size and shape distributions. Al-
though on-line or at-line measurements of microbial culti-
vations have been demonstrated for each of these
Photosynthetic efficiencies and quantum yield. Chloro- approaches, considerable work remains to ascertain their
phyll fluorescence measurements using pulse amplitude applicability (accuracy, information value, and potential
modulated (PAM) fluorometry have become a standard interferences) in real microalgal cultivations.
noninvasive method for determining photosynthetic per-
formance and physiological stress leading to lipid accumu- 2D fluorometry
lation [55]. Fluorescence is measured in situ using a PAM The use of 2D process fluorometers enables the simulta-
fluorometer [21,31,41,43,56]. Although the true PE is de- neous measurement of several analytes by scanning
fined in terms of carbon fixed as carbohydrate, a function- through a range of excitation and emission wavelengths
ally important analog is the solar-to-biomass conversion and using statistical methods to extract information from
efficiency, which is defined as the part of light energy that the resulting data. This approach has enabled the mea-
is conserved in produced microalgal biomass. This quantity surement of proteins, vitamins, coenzymes, biomass, and
is calculated from on-line measurements of the photon flux glucose in cultivation media [7], including for on-line mon-
absorbed in the reactor volume, volumetric biomass pro- itoring and control [60,62,63]. Although 2D fluorometers
ductivity, and biomass combustion enthalpy [30,36,57]. are available that rapidly scan a full range of wavelengths,
PQY characterizes the efficiency of excitation capture by versions in which a set of specific excitation and emission
open photosystem II reaction centers, and is measured by wavelength pairs are measured would enable the construc-
using a PAM fluorometer [21,36,57]. There exist robust tion of simpler devices that still reveal important informa-
optical fluorescence sensors for measurement of the con- tion about the culture. Fluorescence spectra of cultivation
centration of photosynthetic pigments in water that could broths are often complex and contain overlapping peaks,
be used for biomass estimation. Unfortunately, these sen- thus, the integration of data processing using multivariate
sors are designed for measuring very low pigment concen- analyses such as principal component analysis (PCA) will
trations translating into upper limits of tens of milligrams be necessary. Microalgae contain pigments at levels that
biomass per liter. Examples include the EXO Total Algae vary with time and conditions during cultivation, there-
Sensor from YSI, the Algal Online Monitor from Qubit fore, the application of 2D fluorometry has the potential to
Systems, and the AlgaeTorch and FluoroProbe from bbe provide information of use for monitoring.
Moldaenke.
IR spectroscopy
Biomass composition. Although this parameter is one of IR spectroscopy is an optical technique capable of detect-
the most interesting for microalgal cultivations, there are ing a wide variety of organic compounds, and thus pos-
few reports of reliable on-line measurements. PAM fluo- sesses considerable potential for on-line monitoring of
rometry has been used for neutral lipid estimation by bioprocesses [61,64,65]. Organic molecules exhibit specific
correlating lipid yield and photosynthetic efficiency during signatures in the IR region of the electromagnetic spec-
physiological stress [55]. Chlorophyll a and lipid contents trum. The most useful parts of this region for monitoring
in microalgae were simultaneously estimated by three- are the mid-IR (MIR; 200–4000/cm) and the near-infrared
color analysis in culture samples [35]. The total fatty acid (NIR; 4000–13 000/cm). In addition, NIR und MIR spec-
content could be correlated with carotenoid-to-chlorophyll troscopic sensors can simultaneously detect several com-
ratio obtained by measuring absorbance spectra of micro- pounds [66]. As a result of the nature of the spectral signals
algal suspension [58,59]. In addition, the carotenoid and from these techniques, the application of multivariate
chlorophyll content of microalgae have been correlated analytical techniques such as PCA and partial least
with color coordinates measured in a culture sample using squares analysis (PLA) is necessary to obtain the desired
a colorimeter [20]. information.
5
TIBTEC-1074; No. of Pages 9

Review Trends in Biotechnology xxx xxxx, Vol. xxx, No. x

5 on-line implementation are difficult, and until recently,

only off-line applications have been reported in microalgal

Increased lipid/decreased carbohydrate

3
5 2 5
cultivations. The first commercial system capable of at-line
29 9
7
7
operation in a microalgal bioprocess including automatic
11 3 2
22 2 sampling and sample dilution was announced in January
15 11 15 7 0
0 29 9
53
2013 by FlowCam [77]. Other FC on-line monitoring sys-
25
15 18 18 11
9 tems such as the Cytosense Flow Cytometer [78] are
25
29 designed for monitoring low to very low cell counts in
PC2

25
18
22
22
marine or lake ecosystems. In yeast fermentation monitor-
ing, FC coupled with an autosampler and a flow injection
-5 Key: system for sample processing has been used in an at-line
Day 0 mode [79].
Low
Increased lipid and carbohydrate
Inter Computer-aided monitoring methods: software sensors
High
Process mathematical models can be used together with
-10
hardware sensors to estimate values of process variables
-10 -5 0 5 10
PC1 that cannot be measured directly. This approach – based
TRENDS in Biotechnology on control theory – is widely used in all engineering dis-
ciplines including bioprocess engineering. Such computer-
Figure 2. Fourier transform IR (FTIR) measurements can be used to distinguish
aided estimators have been termed ‘software sensors’,
algal cells with differing cellular compositions. Illustrated here is a principal
component analysis of off-line FTIR spectra (1800–950/cm) of Scenedesmus reflecting the fact that they can replace or support classical
subspicatus cells. Samples were obtained at different nitrogen levels in the hardware sensors. Another function of software sensors is
medium, with green circles denoting the high-N medium (initially 19.7 mg/l N as
NO3 ), blue squares the intermediate level (initially 3.0 mg/l N as NO3 ), and red
the reduction of noise and response time; the former is
triangles the low level (initially 0.8 mg/l N as NO3 ). The number next to each accomplished by digital signal filtering and the latter by
symbol indicates the day of growth. PC1 and PC2 refer to principal components relying on a mathematical process model or a correlation.
derived from the data analysis. Reproduced from [70].
Several estimator types have been established: data-
driven empirical estimators are represented by correla-
tions or ANNs, while process model-based estimators are
NIR spectra or absorbance at specific wavelengths have represented by mass and energy balance models, adaptive
been used for the analysis of processed and dried micro- observers, and statistical filtering techniques based on the
algal samples. For example, NIR and Fourier transform IR Kalman–Bucy filter [80]. All of these estimator types are
(FTIR) spectra of four microalgal species were collected widely used as software sensors in various bioprocesses for
and regression models constructed for prediction of lipid measuring biomass and product concentration. These are
concentration [67]. The quantitation of biomass composi- most commonly paired with gas analyzers, multi-wave-
tion as proteins, lipids, carbohydrates, or silicate by FTIR length fluorescence sensors, and dissolved oxygen sensors
in the 4000–700/cm range has been reported, using either [81]. Measurement methods such as MIR spectroscopy that
microalgal biomass or isolated components (Figure 2) produce complex signals must be accompanied by an eval-
[68–70]. uation algorithm to produce valid data from the primary
Recently, devices to interface spectrometers to biopro- measuring information. ANNs can be trained to map com-
cess units have been developed. Two classes of such devices bined signals of different physical sensors to any desired
are attenuated total reflection (ATR) flow cells, in which a process variable or its rate, provided the sensitivity of the
sample stream is directed through the measuring chamber process variable to those signals is nonzero, without know-
of an FTIR spectrometer [71], and autoclavable optical ing the underlying causal connection or having to formu-
fibers with ATR crystals at the tip [72]. At least two late a mathematical model. For complicated relations, for
companies, Mettler Toledo International Inc. and Remspec example, several sensors being combined to provide a
Corp., have commercialized ATR optical fiber probes. single output, partial models can be formulated as simple
correlations, deterministic mathematical models, obser-
FC vers, ANNs, or fuzzy-logic models, and combined into a
Multiparameter FC provides information on cell size and software sensor and process predictor as demonstrated for
cell components by measuring light scatter and, if desired, industrial-scale beer fermentation [82].
fluorescence of individual cells after fluorescent dye stain- At this time, only a few implementations of software
ing. FC can be used as a fast and reliable method for off-line sensors in microalgal processes have been attempted, most
monitoring of lipid content [73]. FC methods in microbial of which focus on the estimation of biomass concentration.
biofuels production for monitoring cellular lipid content, Using correlations, the biomass concentration in a micro-
enzyme activity, cell viability, and cell count have recently algal cultivation has been estimated by software sensors
been reviewed [74,75]. The total cellular lipid content has using measurements of local irradiance [45], color images of
been measured after staining with Nile Red and calibrat- light distribution in a PBR [46], or macroscopic color inten-
ing measured fluorescence to values determined by stan- sity in a cultivation flask [48]. A predictive pH controller
dard methods. Cell size and cell shape distribution of based on a CO2-uptake model and measurement of input
microalgae have been measured with FC by forward scat- and output CO2 balance has been constructed [27].
ter [76]. As a result of the necessary staining, at-line and Biomass concentration, specific growth rate, photosynthetic
6
TIBTEC-1074; No. of Pages 9

Review Trends in Biotechnology xxx xxxx, Vol. xxx, No. x

efficiency, and average light intensity in a PBR have been monitoring microalgal lipid production, sensors should
estimated with a growth model and dissolved oxygen mea- be combined that provide signals connected with cell lipid
surement [44]. Biomass concentration and the nitrogen content, both directly as IR and indirectly as fluorescence
concentration in the biomass and medium have been esti- and absorbance [58]. Lipid content is sometimes linked to
mated continuously in microalgal cultivations by a model- the color of a microalgal culture, therefore, color analysis in
based interval observer using intermittent off-line biomass conjunction with pattern recognition by ANN could be
measurements [83]. A software sensor using an ANN has used. Software sensors for evaluating single- and multi-
processed the signal of a flow-through hardware sensor sensor signals, combined correlations, mathematical mod-
containing four monochromatic photodiodes, and the els, and ANNs can also be constructed [82].
ANN has been trained to combine simultaneous measure- If noninvasive measurements are not suitable and some
ments on four different wavelengths into the resulting type of sampling is applied (whether automated or manu-
biomass concentration signal [47]. An on-line multivariate al), the field of acceptable methods widens, and the speed,
sensor using support vector regression and several prepro- stability, and simplicity of the measurement method be-
cessing algorithms to correlate a Raman spectrum signal come its prominent features. In this case, automated FC,
with the concentrations of biomass, glucose, and oil content automated cell counting, lipid determination using micro-
in a microalgal bioreactor has been constructed [84]. wave drying and NMR analysis, and ion-selective electro-
des could be adapted and used. However, those sensors are
Concluding remarks and future perspectives labor and capital intensive, and present some serious
Currently, on-line measurements in microalgal cultiva- drawbacks under harsh process conditions.
tions focus on physicochemical process variables. Methods There are unique challenges facing the application of on-
that can be implemented for on-line, direct measurement line measurement systems to microalgal cultivations, in-
of biological characteristics, including the concentrations cluding the significant changes in cell size, shape, and color
of substrates, products, or cells, are emerging slowly due to that can occur during a cultivation. In addition, few indus-
the complexity of the biological phase and its interactions trial-scale cultivations will be axenic, and the nature of the
with cultivation environment. The adaptation of successful non-target fraction of the culture may have a detrimental
off-line analytical methods to on-line application includes effect on some measurements. Finally, the placement and
not only adapting the sensor hardware but also the design density of sensors within a very large cultivation system
of data processing methods for effective, intelligent data are considerations that require more evaluation. Although
management. A large amount of raw data must be evalu- some recommendations have been made (e.g., placement
ated using chemometric models and multivariate analysis along the strongest mass gradient [9]) decisions will ulti-
to supply valid process data. All of the methods for on-line mately be made on the basis of the performance of each
monitoring of biological variables reviewed here require sensor system and the value of the information obtained.
more sophisticated evaluation methods implemented in Despite these challenges, the potential for on-line mon-
software to deliver meaningful results. Evaluation meth- itoring methods to improve the productivity and reliability
ods have thus become an integral part of bioprocess sen- of microalgal cultivations is high. This application area is
sors, which could be categorized as low-level software new, and exciting developments are expected in the next
sensors. Higher-level software sensors combining more few years.
easily measurable process variables can substitute for
missing sensors for measuring difficult variables that Acknowledgments
are not yet available. This work was supported by the Sustainable Bioenergy Development
Although the measurement of physicochemical vari- Center of Colorado State University.
ables can be accomplished by standard sensors, the goals
of miniaturization, reliability, robustness, and cost cutting Disclaimer statement
K.F.R. is a founder and the Chief Technology Officer of OptiEnz Sensors,
are motivation to consider optical sensors as alternatives
LLC. No product of OptiEnz Sensors is mentioned in this article. None of
in measuring dissolved oxygen, CO2, and pH [33]. The most the other authors has any conflict of interest to disclose.
promising methods for on-line, in-situ measurement of
biological variables, the physiological state, and product
References
concentrations seem to be computer-aided optical sensors 1 Malcata, F.X. (2011) Microalgae and biofuels: a promising partnership?
measuring OD, pigment fluorescence, culture color, IR Trends Biotechnol. 29, 542–549
spectra and ISM images. ISM can supply in almost real 2 Scott, S.A. et al. (2010) Biodiesel from algae: challenges and prospects.
time those culture parameters that are obtainable only Curr. Opin. Biotechnol. 21, 277–286
with difficulty by other on-line methods. As a side benefit, 3 Brennan, L. and Owende, P. (2010) Biofuels from microalgae – a review
of technologies for production, processing, and extractions of biofuels
microscopic images can be displayed for visual control of and co-products. Renew. Sust. Energ. Rev. 14, 557–577
the culture by the operator. Additional work is required to 4 Masojidek, J. and Torzillo, G. (2009) Mass cultivation of freshwater
develop an in situ cleaning procedure or self-cleaning microalgae. In Applications in Ecological Engineering (Jorgensen, S.E.,
surface materials for optical sensors because fouling by ed.), pp. 176–186, Elsevier
5 Borowitzka, M.A. (1999) Commercial production of microalgae: ponds,
adhering cells seriously impairs the vision field.
tanks, tubes and fermenters. J. Biotechnol. 70, 313–321
Software sensors based on the combination of single 6 Lee, Y.K. (2001) Microalgal mass culture systems and methods: their
sensor signals should be designed to provide estimates and limitation and potential. J. Appl. Phycol. 13, 307–315
predictions of the respective product concentration to fa- 7 Glindkamp, A. et al. (2009) Sensors in disposable bioreactors status and
cilitate forward process control. As an example, when trends. Adv. Biochem. Eng. 115, 145–169

7
TIBTEC-1074; No. of Pages 9

Review Trends in Biotechnology xxx xxxx, Vol. xxx, No. x

8 Scheper, T. and Reardon, K.F. (1992) Sensors in biotechnology. In 34 Bosma, R. et al. (2007) Prediction of volumetric productivity of an
Sensors: A Comprehensive Survey (Goepel, W. et al., eds), pp. 1023– outdoor photobioreactor. Biotechnol. Bioeng. 97, 1108–1120
1046, VCH Verlagsgesellschaft 35 Su, C.H. et al. (2008) Simultaneous estimation of chlorophyll a and lipid
9 Posten, C. (2009) Design principles of photo-bioreactors for cultivation contents in microalgae by three-color analysis. Biotechnol. Bioeng. 99,
of microalgae. Eng. Life Sci. 9, 165–177 1034–1039
10 Doucha, J. and Livansky, K. (2009) Outdoor open thin-layer microalgal 36 Cuaresma, M. et al. (2011) Horizontal or vertical photobioreactors?
photobioreactor: potential productivity. J. Appl. Phycol. 21, 111–117 How to improve microalgae photosynthetic efficiency. Bioresour.
11 Slegers, P.M. et al. (2011) Design scenarios for flat panel Technol. 102, 5129–5137
photobioreactors. Appl. Energy 88, 3342–3353 37 Rodolfi, L. et al. (2009) Microalgae for oil: strain selection, induction of
12 Grobbelaar, J.U. (2010) Microalgal biomass production: challenges and lipid synthesis and outdoor mass cultivation in a low-cost
realities. Photosynth. Res. 106, 135–144 photobioreactor. Biotechnol. Bioeng. 102, 100–112
13 Ketheesan, B. and Nirmalakhandan, N. (2011) Development of a new 38 Miller, A.Z. et al. (2010) Laboratory-induced endolithic growth in
airlift-driven raceway reactor for algal cultivation. Appl. Energy 88, calcarenites: biodeteriorating potential assessment. Microb. Ecol. 60,
3370–3376 55–68
14 Chinnasamy, S. et al. (2010) Biomass and bioenergy production 39 Gutierrez, M. et al. (2008) Nutrient solution monitoring in greenhouse
potential of microalgae consortium in open and closed bioreactors cultivation employing a potentiometric electronic tongue. J. Agric.
using untreated carpet industry effluent as growth medium. Food Chem. 56, 1810–1817
Bioresour. Technol. 101, 6751–6760 40 Quinn, J. et al. (2011) Microalgae bulk growth model with application
15 Doucha, J. et al. (2005) Utilization of flue gas for cultivation of to industrial scale systems. Bioresour. Technol. 102, 5083–5092
microalgae (Chlorella sp.) in an outdoor open thin-layer 41 Marxen, K. et al. (2005) A photobioreactor system for computer
photobioreactor. J. Appl. Phycol. 17, 403–412 controlled cultivation of microalgae. J. Appl. Phycol. 17, 535–549
16 Kromkamp, J.C. et al. (2009) Short-term variations in photosynthetic 42 Sandnes, J.M. et al. (2006) Real-time monitoring and automatic density
parameters of Nannochloropsis cultures grown in two types of outdoor control of large-scale microalgal cultures using near infrared (NIR)
mass cultivation systems. Aquat. Microb. Ecol. 56, 309–322 optical density sensors. J. Biotechnol. 122, 209–215
17 Sukenik, A. et al. (2009) Photosynthetic performance of outdoor 43 Nedbal, L. et al. (2008) A photobioreactor system for precision
Nannochloropsis mass cultures under a wide range of cultivation of photoautotrophic microorganisms and for high-
environmental conditions. Aquat. Microb. Ecol. 56, 297–308 content analysis of suspension dynamics. Biotechnol. Bioeng. 100,
18 Eriksen, N.T. (2008) The technology of microalgal culturing. 902–910
Biotechnol. Lett. 30, 1525–1536 44 Li, J. et al. (2003) Online estimation of stirred-tank microalgal
19 Garcia-Malea, M.C. et al. (2006) Continuous production of green cells of photobioreactor cultures based on dissolved oxygen measurement.
Haematococcus pluvialis: modeling of the irradiance effect. Enzyme Biochem. Eng. J. 14, 51–65
Microb. Technol. 38, 981–989 45 Su, W.W. et al. (2003) State and parameter estimation of microalgal
20 Lopez, M.C.G.M. et al. (2006) Comparative analysis of the outdoor photobioreactor cultures based on local irradiance measurement. J.
culture of Haematococcus pluvialis in tubular and bubble column Biotechnol. 105, 165–178
photobioreactors. J. Biotechnol. 123, 329–342 46 Jung, S.K. and Lee, S.B. (2006) In situ monitoring of cell concentration
21 Masojidek, J. et al. (2009) A two-stage solar photobioreactor for in a photobioreactor using image analysis: comparison of uniform light
cultivation of microalgae based on solar concentrators. J. Appl. distribution model and artificial neural networks. Biotechnol. Prog. 22,
Phycol. 21, 55–63 1443–1450
22 Barsanti, L. and Gualtieri, P. (2006) Algae. Anatomy, Biochemistry, 47 Briassoulis, D. et al. (2010) An experimental helical-tubular
and Biotechnology, CRC Press, Taylor & Francis Group photobioreactor for continuous production of Nannochloropsis sp.
23 Quinn, J.C. et al. (2012) Scale-up of flat plate photobioreactors Bioresour. Technol. 101, 6768–6777
considering diffuse and direct light characteristics. Biotechnol. 48 Cordoba-Matson, M.V. et al. (2010) Evaluation of Isochrysis galbana
Bioeng. 109, 363–370 (clone T-ISO) cell numbers by digital image analysis of color intensity.
24 Ugwu, C.U. and Aoyagi, H. (2008) Influence of shading inclined tubular J. Appl. Phycol. 22, 427–434
photobioreactor surfaces on biomass productivity of C. sorokiniana. 49 Brown, L.M. et al. (1989) Computer-based image analysis for the
Photosynthetica 46, 283–285 automated counting and morphological description of microalgae in
25 Suh, I.S. and Lee, C.G. (2003) Photobioreactor engineering: Design and culture. J. Appl. Phycol. 1, 211–225
performance. Biotechnol. Bioprocess Eng. 8, 313–321 50 Rehbock, C. et al. (2010) Development of a flow-through microscopic
26 Livansky, K. et al. (2006) CO2 partial pressure – pH relationships in the multitesting system for parallel monitoring of cell samples in
medium and relevance to CO2 mass balance in outdoor open thin-layer biotechnological cultivation processes. J. Biotechnol. 150, 87–93
Arthrospira (Spirulina) cultures. Arch. Hydrobiol. 165, 365–381 51 Prediger, A. et al. (2011) In situ microscopy for online monitoring of
27 Berenguel, M. et al. (2004) Model predictive control of pH in tubular enzyme carriers and two-phase processes. Chem. Eng. Technol. 34,
photobioreactors. J. Process Control 14, 377–387 837–840
28 Sanchez, J.L.G. et al. (2003) Minimization of carbon losses in pilot-scale 52 Bluma, A. et al. (2010) In-situ imaging sensors for bioprocess
outdoor photobioreactors by model-based predictive control. monitoring: state of the art. Anal. Bioanal. Chem. 398, 2429–2438
Biotechnol. Bioeng. 84, 533–543 53 Hopfner, T. et al. (2010) A review of non-invasive optical-based image
29 Grobbelaar, J.U. (2007) Photosynthetic characteristics of Spirulina analysis systems for continuous bioprocess monitoring. Bioprocess
platensis grown in commercial-scale open outdoor raceway ponds: Biosyst. Eng. 33, 247–256
what do the organisms tell us? J. Appl. Phycol. 19, 591–598 54 Havlik, I. et al. (2013) Monitoring of microalgal cultivations with on-
30 Doucha, J. and Livansky, K. (2006) Productivity, CO2/O2 exchange and line, flow-through microscopy. Algal Res. 2, http://dx.doi.org/10.1016/
hydraulics in outdoor open high density microalgal (Chlorella sp.) j.algal.2013.04.001
photobioreactors operated in a Middle and Southern European 55 White, S. et al. (2011) PAM fluorometry as a tool to assess microalgal
climate. J. Appl. Phycol. 18, 811–826 nutrient stress and monitor cellular neutral lipids. Bioresour. Technol.
31 Cerveny, J. et al. (2009) Photobioreactor for cultivation and real-time, 102, 1675–1682
in-situ measurement of O2 and CO2 exchange rates, growth dynamics, 56 Masojidek, J. et al. (2011) Productivity correlated to photobiochemical
and of chlorophyll fluorescence emission of photoautotrophic performance of Chlorella mass cultures grown outdoors in thin-layer
microorganisms. Eng. Life Sci. 9, 247–253 cascades. J. Ind. Microbiol. Biotechnol. 38, 307–317
32 Nedbal, L. et al. (2010) Experimental validation of a nonequilibrium 57 Hulatt, C.J. and Thomas, D.N. (2011) Productivity, carbon dioxide
model of CO2 fluxes between gas, liquid medium, and algae in a flat- uptake and net energy return of microalgal bubble column
panel photobioreactor. J. Ind. Microbiol. Biotechnol. 37, 1319–1326 photobioreactors. Bioresour. Technol. 102, 5775–5787
33 Lindner, P. et al. (2011) Disposable sensor systems. In Single-Use 58 Solovchenko, A.E. et al. (2009) Carotenoid-to-chlorophyll ratio as a
Technology in Biopharmaceutical Manufacture (Eibl, R. and Eibl, D., proxy for assay of total fatty acids and arachidonic acid content in the
eds), pp. 67–81, J. Wiley & Sons green microalga Parietochloris incisa. J. Appl. Phycol. 21, 361–366

8
TIBTEC-1074; No. of Pages 9

Review Trends in Biotechnology xxx xxxx, Vol. xxx, No. x

59 Solovchenko, A. et al. (2011) Stress-induced changes in optical 77 Fluid Imaging Technologies Inc. (2013) FlowCAM1 particle imaging
properties, pigment and fatty acid content of Nannochloropsis sp.: within processing line. Algae Industry Magazine (http://
implications for non-destructive assay of total fatty acids. Mar. www.algaeindustrymagazine.com/flowcam-particle-imaging-within-
Biotechnol. (NY) 13, 527–535 processing-line/)
60 Hantelmann, K. et al. (2006) Two-dimensional fluorescence 78 Pomati, F. et al. (2011) An automated platform for phytoplankton
spectroscopy: a novel approach for controlling fed-batch cultivations. ecology and aquatic ecosystem monitoring. Environ. Sci. Technol.
J. Biotechnol. 121, 410–417 45, 9658–9665
61 Roychoudhury, P. et al. (2007) Multiplexing fibre optic near infrared 79 Broger, T. et al. (2011) Real-time on-line flow cytometry for bioprocess
(NIR) spectroscopy as an emerging technology to monitor industrial monitoring. J. Biotechnol. 154, 240–247
bioprocesses. Anal. Chim. Acta 590, 110–117 80 Komives, C. and Parker, R.S. (2003) Bioreactor state estimation and
62 Podrazky, O. et al. (2003) Monitoring the growth and stress responses control. Curr. Opin. Biotechnol. 14, 468–474
of yeast cells by two-dimensional fluorescence spectroscopy: first 81 Odman, P. et al. (2009) On-line estimation of biomass, glucose and ethanol
results. Folia Microbiol. 48, 189–192 in Saccharomyces cerevisiae cultivations using in-situ multi-wavelength
63 Tartakovsky, B. et al. (1996) Application of scanning fluorometry for fluorescence and software sensors. J. Biotechnol. 144, 102–112
monitoring of a fermentation process. Biotechnol. Prog. 12, 126–131 82 Simutis, R. et al. (1993) Fuzzy-aided neural network for real-time state
64 Landgrebe, D. et al. (2010) On-line infrared spectroscopy for bioprocess estimation and process prediction in the alcohol formation step of
monitoring. Appl. Microbiol. Biotechnol. 88, 11–22 production-scale beer brewing. J. Biotechnol. 27, 203–215
65 Roychoudhury, P. et al. (2006) The potential of mid infrared 83 Goffaux, G. et al. (2009) Continuous – discrete interval observers for
spectroscopy (MIRS) for real time bioprocess monitoring. Anal. monitoring microalgae cultures. Biotechnol. Prog. 25, 667–675
Chim. Acta 571, 159–166 84 Nadadoor, V.R. et al. (2012) Online sensor for monitoring a microalgal
66 Riley, M.R. et al. (2001) Simultaneous measurement of 19 components bioreactor system using support vector regression. Chemometr. Intell.
in serum-containing animal cell culture media by Fourier transform Lab. 110, 38–48
near-infrared spectroscopy. Biotechnol. Prog. 17, 376–378 85 Li, Y.Q. et al. (2008) Effects of nitrogen sources on cell growth and lipid
67 Laurens, L.M.L. and Wolfrum, E.J. (2011) Feasibility of spectroscopic accumulation of green alga Neochloris oleoabundans. Appl. Microbiol.
characterization of algal lipids: chemometric correlation of NIR and Biotechnol. 81, 629–636
FTIR spectra with exogenous lipids in algal biomass. Bioenerg. Res. 4, 86 Kunjapur, A.M. and Eldridge, R.B. (2010) Photobioreactor design for
22–35 commercial biofuel production from microalgae. Ind. Eng. Chem. Res.
68 Jakob, T. et al. (2007) A complete energy balance from photons to new 49, 3516–3526
biomass reveals a light- and nutrient-dependent variability in the 87 Chen, C-Y. et al. (2011) Cultivation, photobioreactor design and
metabolic costs of carbon assimilation. J. Exp. Bot. 58, 2101–2112 harvesting of microalgae for biodiesel production: a critical review.
69 Pistorius, A.M.A. et al. (2009) Monitoring of biomass composition from Bioresour. Technol. 102, 71–81
microbiological sources by means of FT-IR spectroscopy. Biotechnol. 88 Kumar, K. et al. (2011) Development of suitable photobioreactors for
Bioeng. 103, 123–129 CO2 sequestration addressing global warming using green algae and
70 Dean, A.P. et al. (2010) Using FTIR spectroscopy for rapid cyanobacteria. Bioresour. Technol. 102, 4945–4953
determination of lipid accumulation in response to nitrogen 89 Pruvost, J. et al. (2011) Systematic investigation of biomass and lipid
limitation in freshwater microalgae. Bioresour. Technol. 101, 4499– productivity by microalgae in photobioreactors for biodiesel
4507 application. Bioresour. Technol. 102, 150–158
71 Schenk, J. et al. (2007) A simple method to monitor and control 90 Obata, M. et al. (2009) Using chlorophyll fluorescence to monitor yields
methanol feeding of Pichia pastoris fermentations using mid-IR of microalgal production. J. Appl. Phycol. 21, 315–319
spectroscopy. J. Biotechnol. 128, 344–353 91 Meiser, A. et al. (2004) Optimization of eicosapentaenoic acid
72 Holm-Nielsen, J.B. et al. (2008) On-line near infrared monitoring of production by Phaeodactylum tricornutum in the flat panel airlift
glycerol-boosted anaerobic digestion processes: evaluation of process (FPA) reactor. J. Appl. Phycol. 16, 215–225
analytical technologies. Biotechnol. Bioeng. 99, 302–313 92 Wu, Y-H. et al. (2012) Biomass production of a Scenedesmus sp. under
73 Brennan, L. et al. (2012) Enhancement of BODIPY505/515 lipid phosphorous-starvation cultivation condition. Bioresour. Technol. 112,
fluorescence method for applications in biofuel-directed microalgae 193–198
production. J. Microbiol. Methods 90, 137–143 93 Masojidek, J. and Prasil, O. (2010) The development of microalgal
74 Silva, T.L.d. et al. (2012) Applications and perspectives of multi- biotechnology in the Czech Republic. J. Ind. Microbiol. Biotechnol. 37,
parameter flow cytometry to microbial biofuels production processes. 1307–1317
Trends Biotechnol. 30, 225–232 94 Masojidek, J. et al. (2007) Photosynthesis in microalgae. In Handbook
75 Hyka, P. et al. (2013) Flow cytometry for the development of of Microalgal Culture: Biotechnology and Applied Phycology
biotechnological processes with microalgae. Biotechnol. Adv. 31, 2–16 (Richmond, A., ed.), pp. 20–39, Blackwell Science
76 Davis, R.W. et al. (2012) Multiplex fluorometric assessment of nutrient 95 Gouveia, L. et al. (2009) Neochloris oleabundans UTEX #1185: a
limitation as a strategy for enhanced lipid enrichment and harvesting suitable renewable lipid source for biofuel production. J. Ind.
of Neochloris oleoabundans. Biotechnol. Bioeng. 109, 2503–2512 Microbiol. Biotechnol. 36, 821–826