O₂-Sensitive Inks for Measuring Total (Aerobic) Viable Count Using Micro-Respirometry

Abstract

The popular method of micro-respirometry (μR) for measuring total viable (aerobic) count (TVC) utilises luminescence-based O₂ sensors that are difficult to fabricate and therefore expensive. A simple method is described for making inexpensive, ink-based potential substitutes that utilise the same O₂-sensitive dyes. The sensitivity of such inks is readily increased by using dyes with a long lifetime in the absence of O₂, τ_o, and/or an ink resin/polymer with a high O₂ permeability, P_m(O₂). Response modelling of the μR-based TVC system and subsequent testing using a range of O₂ sensors of different sensitivity show that there is little to be gained by making the O₂ sensor either very sensitive or insensitive, and that the best O₂ sensors are dyes such as Pt(II) tetraphenyltetrabenzoporphyrin (PtBP), with τ_o = ca. 40–50 μs. Further work shows that a simple-to-make PtBP ink can be used as a direct replacement for the expensive O₂ sensor used in commercial instruments for measuring TVC based on μR. In addition, the PtBP can be replaced by an even less expensive O₂-sensitive dye, Pt(II) meso-tetra(pentafluorophenyl)porphyrin (PtTFPP). The potential use of inexpensive O₂-sensitive inks as an alternative to any expensive commercial counterpart based on the same O₂-sensitive dye is discussed briefly.

1. Introduction

The measurement of the total viable count (TVC) of aerobes, measured in colony-forming units (CFUs)/mL, is an essential part of microbiology and plays an important role in food safety, environmental monitoring, and clinical analysis [1]. The most established method for measuring TVC is the aerobic plate-counting method (APC) [2,3,4], but it is time-consuming, laborious (as it involves multiple dilutions and counting), expensive (as it consumes a lot of plasticware) and slow, with a time to result of up to 1–3 days [5,6]. Consequently, there is a real demand for a more rapid, high-throughput, cost-efficient method for measuring TVC that is amenable to automation. One such method is micro-respirometry (μR) which is based on monitoring the consumption of O₂ as a function of incubation time, t, in a growth medium inoculated with the bacterium under test [7,8,9,10]. Recent studies include the use of μR for, (i) quality control in pharmaceuticals [11] and food samples [12], (ii) fundamental bacterial and animal cell respiration studies [13,14,15,16,17], and (iii) use with different growth media [18,19].

It is common practice to use μR to measure TVC, and in this method (μR-TVC), the level of dissolved O₂ is monitored using an O₂ sensor, which contains a phosphorescent dye, D, that is dynamically quenched by O₂, since its electronically excited state lifetime, τ, is related to the ambient level of O₂ dissolved in the growth medium (%O₂) via the Stern–Volmer equation,

τ_o/τ = 1 + K_sv%O₂

(1)

where τ and τ_o are the lifetimes of D in the presence and absence of O₂, respectively, and K_sv is a measure of the sensitivity of the O₂ sensor. Further details concerning the mechanism of O₂ sensing are given in S1 in the Electronic Supplementary Information (ESI) file.

In μR-TVC, usually only τ is reported as a function of incubation time, t, and not %O₂, although the calculation of the latter from the former, via Equation (1), is trivial. A typical plot of observed variation in τ, and %O₂, vs. incubation time, t, when 1 mL 10⁴ CFU/mL E. coli is used to inoculate 9 mL of growth medium, is illustrated in Figure 1. The typical ‘bioreactor’ used for this work is a Falcon™ tube with an O₂ sensor set in its base, as illustrated in Figure S1 in the ESI. From the data set illustrated in Figure 1, the lifetimes in the absence and presence of air are 41.6 and 20.4 μs; thus, from Equation (1), K_sv = 0.050%O₂⁻¹ and k_q = (K_sv/τ_o) = 0.0012%O₂⁻¹μs⁻¹. In μR-TVC, the threshold time (TT) at the lifetime midway point τ(TT) = (τ_o + τ(air)/2); so, 31 μs in Figure 1 is determined from a set of τ vs. t runs for inoculums of known TVC. The results of this work are then used to generate a straight-line calibration plot of log(CFU/mL) vs. TT, which can then be used to calculate the TVC of any further samples from their measured TT values. For example, a typical set of τ vs. t data generated using different inoculums of E. coli are illustrated in Figure S2a in the ESI, along with the subsequent calibration graph of log(CFU/mL) vs. TT generated from these data in Figure S2b.

One barrier to the widespread use of μR in the measurement of TVC is the preparation of the O₂ sensor ‘dot’ placed at the bottom of the incubation cell (see S1 in the Electronic Supplementary Information (ESI) file), which is ‘slow and difficult to control and standardise’ [20] and, therefore, costly. Consequently, most commercial O₂ sensors, such as those listed in Table 1, are expensive, which is a barrier to the widespread use of μR-TVC, in which the Falcon™ tube/O₂ sensor dot is used as a consumable. The structures and photophysical properties of the different dyes listed in Table 1 are given in S3, Table S1 in the ESI [21].

In most μR studies, the dye used in the O₂ sensors is a Pt-based porphyrin, often PtBP and PtTFPP. As you might expect, the formulation of commercial O₂ sensors is proprietary, but some idea of their complex nature is provided by Santovito et al., who used an O₂ sensor solution supplied by Agilent Ireland (Cork) to deposit a droplet at the bottom of their incubation cell, a Falcon™ tube, which dries to form a water insoluble sensor ‘dot’ [22]. These researchers noted that the commercial O₂ sensor solution comprised polymeric microparticles, impregnated with PtBP and suspended in ethanol containing 5 wt.% of a hydrogel, stored in the dark at 4 °C [22].

Table 1. Commercial phosphorescent sensors and their properties in water at 20 °C.

Company	Product	Dye	Cost ($) **	τo */μs	τ(air) */μs	Ksv/%O2−1	Ref.
Oculer	TVC Assay	PtBP	6.5	44.1	22.5	0.046	[23]
PyroScience	OXSP5	PtBP	33	52.5	20.8	0.073	[24]
PreSens	PSt3	PtTFPP	33	59.0	17.5	0.115	[24]
Ocean Insight	FOSPOR	PtTFPP	36	22.4	9.1	0.070	[24]
OxySense	O2xyDot	[Ru(dpp)3](ClO4)2	4.3	6.5	0.9	0.30	[25]

*: Lifetime τ_o and τ(air) determined are O₂-free (produced by adding sodium sulfite) and air-saturated water, respectively. PtBP: Platinum(II) tetraphenyltetrabenzoporphyrin, PtTFPP: Platinum(II) meso-tetra(pentafluorophenyl)porphyrin, [Ru(dpp)₃](ClO₄)₂: Tris(4,7-diphenyl-1,10-phenanthroline) ruthenium(II) perchlorate. **: cost per indicator.

Table 1. Commercial phosphorescent sensors and their properties in water at 20 °C.

Company	Product	Dye	Cost ($) **	τo */μs	τ(air) */μs	Ksv/%O2−1	Ref.
Oculer	TVC Assay	PtBP	6.5	44.1	22.5	0.046	[23]
PyroScience	OXSP5	PtBP	33	52.5	20.8	0.073	[24]
PreSens	PSt3	PtTFPP	33	59.0	17.5	0.115	[24]
Ocean Insight	FOSPOR	PtTFPP	36	22.4	9.1	0.070	[24]
OxySense	O2xyDot	[Ru(dpp)3](ClO4)2	4.3	6.5	0.9	0.30	[25]

*: Lifetime τ_o and τ(air) determined are O₂-free (produced by adding sodium sulfite) and air-saturated water, respectively. PtBP: Platinum(II) tetraphenyltetrabenzoporphyrin, PtTFPP: Platinum(II) meso-tetra(pentafluorophenyl)porphyrin, [Ru(dpp)₃](ClO₄)₂: Tris(4,7-diphenyl-1,10-phenanthroline) ruthenium(II) perchlorate. **: cost per indicator.

In this paper, we examine how the TT value and therefore the sensitivity of the μR-TVC method is affected by the sensitivity (i.e., K_sv) of the O₂ sensor. In addition, we demonstrate how a simple, easy-to-make, inexpensive ink-based O₂ sensor can replace an expensive commercial O₂ sensor used in μR-TVC.

2. Materials and Methods

2.1. Materials

Unless stated otherwise, all chemicals and solvents were purchased from Merck (Dublin, Ireland) in the highest purity available. Polyvinyl butyral (PVB), molecular weight (MW) ≈ 40,000–70,000, polystyrene (PS), MW ≈ 192,000, beads/pellets, and polyvinyl chloride (PVC), MW ≈ 233,000, were purchased from Sigma-Aldrich (Dorset, UK). All gases used in this work, namely, 1%, 5% and 21% O₂/Ar mixes, 100% O₂ and Ar, were purchased from BOC (Surrey, UK). Pt(II) and Pd(II) tetraphenyltetrabenzoporphyrin (PtBP and PdBP), and Pt(II) meso-tetra(pentafluorophenyl)porphyrin (PtTFPP) were purchased from Frontier Specialty Chemicals (Logan, UT, USA). KWIK STIK stock cultures of Escherichia coli ATCC 8739 (E. coli) were obtained from Microbiologics (St Cloud, MN, USA). Luria-Bertani (LB) broth and LB agar were purchased from Sigma-Aldrich (Dorset, UK).

2.2. Methods

2.2.1. Preparation of O₂-Sensitive Inks

The Pt(II) and Pd(II) porphyrin-based inks were prepared using the PtBP, PtTFPP and PdBP dyes listed in Table S1 in the ESI and a polymer, such as PVB, PS or PVC, as outlined in

Table 2. Preparation details for the O₂-sensitive porphyrin-based inks.

Ink (Dye/Polymer)	Preparation
Dye */PVB	A 0.1% w/v dye solution was prepared by adding 3 mg dye to 3 mL dichloromethane (DCM). The ink was then prepared by adding 0.1 g of the 0.1% w/v dye solution to 1 g of a 10% w/v PVB/ethanol (EtOH) solution.
Dye */PS	As above, but using a 5% w/v PS/ethyl acetate solution.
Dye */PVC	As above, but using a 2% w/v PVC/DCM solution.

Dye *: Individual inks were made for each of the following dyes: PtBP, PdBP and PtTFPP.

.

The dye–ion pair complex [Ru(dpp)₃²⁺(Ph₄B₂)₂] was prepared using a previously reported method [26], and a 0.1% w/v [Ru(dpp)₃²⁺(Ph₄B₂)₂] dye/PVB ink was then prepared using 3 mg of the dye–ion pair complex in 3 mL acetone, followed by adding 0.1 g of the 0.1% w/v dye solution to 1 g of a 10% w/v PVB/ethanol (EtOH) solution.

2.2.2. Preparation of E. coli Stock and Calibration Dispersion of Bacteria

Details of the preparation of the plate agar growth medium and primary culture plates are given in S4 of the ESI. To make the overnight stock of a liquid culture of E. coli, a single colony of the bacterium under test was taken from the primary culture plate using a sterile inoculating loop and suspended in 10 mL LB broth placed in a 15 mL Falcon™ tube, which was then incubated overnight at 30 °C. The loading of the resulting overnight culture was then confirmed using a conventional aerobic plate-counting (APC) method [2,3] and was typically ca. 10⁸ CFU/mL. Serial one-in-ten dilutions of this overnight bacterial culture were then performed in LB broth to produce suspensions of different known concentrations, spanning a range 10⁸–10¹ CFU/mL, for calibrating the μR-TVC system. All μR-TVC runs were carried out in triplicate, and average values were used.

2.2.3. μR-TVC Using O₂-Sensitive Inks

The preparation of the LB broth used in this work is described in S4 of the ESI. In a typical experiment, serial dilutions of E. coli cultures varying in concentration from 10⁸ to 10¹ CFU/mL were prepared and 1 mL of each dispensed into 15 mL Falcon™ tubes containing 9 mL of sterile LB broth and a PtBP/PVB O₂ sensor ‘dot’ set in its base, to give a final volume of 10 mL. Each inoculated tube was then incubated at 37 °C, and an in-house-constructed, phase-correlated detection-based instrument for measuring luminescence lifetime, τ, was used to record the subsequent variation in τ of the O₂ sensor as a function of incubation time, t, as illustrated in Figure 1. Section S5 in the ESI describes in more detail the phase-correlated detection-based instrument and method used to measure τ for the different O₂ sensor inks.

2.2.4. Other Methods

UV–Vis spectra were recorded using a Cary 60 UV–Vis spectrophotometer (Agilent, Dublin, Ireland). Different gas blends of O₂ with Ar were generated as required using a Cole–Parmer rotameter-based gas blender (Cole-Parmer, Cambridge, UK). An Anéolia Legend O₂/CO₂ gas analyser (Anéolia, Moissy Cramayel, France) was used to verify the %O₂ in all blended gas mixtures. A commercial instrument for measuring the TVC of aerobes based on micro-respirometry, the Oculer Rapid 930 (Tipperary, Ireland), was used to compare the performance of the PtBP/PVB ink with that of their proprietary O₂ sensor.

3. Results and Discussion

3.1. Modelling the Effect of O₂ Sensor Sensitivity on μR-TVC

As noted earlier and illustrated in Figure S2 of the ESI, in μR-TVC, a straight-line calibration graph of log(CFU/mL) vs. threshold time, TT, must first be generated prior to use for measuring the TVC of samples with an unknown bacterial load. Although the value of TT will depend upon several experimental parameters, such as incubation temperature and the nature of the growth medium, it will also depend upon the sensitivity of the O₂-sensor, K_sv, and to date, this dependency has not been studied. The Stern–Volmer constant is directly related to the product of the lifetime of the excited state of the O₂-sensitive dye in the absence of O₂, τ_o, and the bimolecular rate constant, k_q, for the quenching reaction between the luminescent excited state of the dye and ambient O₂,

K_sv = τ_o·k_q

(2)

where k_q depends on the O₂ permeability of the medium (usually a polymer) which surrounds and encapsulates the dye molecules. Thus, the sensitivity of the O₂ sensor used in μR-TVC can be varied systematically by either (i) fixing k_q (by using the same encapsulating polymer) and varying τ_o (by using different dyes, such as those listed in Table 1), or (ii) fixing τ_o (by using the same O₂-sensitive dye, such as PtBP) and varying k_q (by using different encapsulating polymers, which will have different values of O₂ permeability, P_m(O₂), where k_q is ∝ P_m(O₂)). The effects of these two very different ways to vary K_sv on a typical τ vs. incubation time, t, profile, and so its TT value, can be readily calculated using a model based on a smoothed version of the typical %O₂ vs. t profile illustrated in Figure 1 for an inoculum of 10⁴ E. coli, and the Stern–Volmer equation, Equation (1). This model-smoothed O₂ vs. t profile, based on the data in Figure 1, is illustrated in Figure S7 in the ESI.

Using this simple model, Figure 2 illustrates the predicted variation in sensor lifetime, τ, as a function of t, for a series of theoretical O₂ sensors with different values of τ_o (varied over the range 15–200 μs), but with a fixed value of k_q (=0.0012%O₂⁻¹μs⁻¹). Using the data in these profiles, it is possible to calculate the model-predicted variation in TT (a measure of the sensitivity of the μR-TVC method, as a function of τ_o, a plot of which is illustrated in the insert diagram in Figure 2). Both plots in Figure 2 show that although TT increases (sensitivity decreases) with increasing τ_o, this effect is slight, i.e., <10 min, which translates to a <5% change in TT for a 13-fold change in K_sv, since τ_o varied from 15 to 200 μs. Further work shows a similar weak effect on TT (and so sensitivity of the μR-TVC method) for a series of simulated O₂ sensors, with a fixed value of τ_o (=45 μs), and k_q varied over the range (0.4 to 5.6) × 10⁻³%O₂⁻¹μs⁻¹, see S7 in the ESI.

In practice, a <5% reduction in TT, brought about by decreasing the sensitivity of the O₂ sensor by using a dye with a shorter τ_o and/or an encapsulating polymer with a lower O₂ permeability (a measure of which is k_q), will have little impact on the sensitivity of the μR-TVC method, as it will not make it significantly faster and will most likely introduce unwanted technical problems, such as a reduction in experimental reproducibility, when using very high or low sensitivity sensors. Therefore, the practice of μR in general and μR-TVC is dominated by O₂ sensors which use the dyes PtBP and PtTFPP, which have O₂ sensitivities that are neither very low nor very high.

Support for the above model predictions was provided by a study of different O₂-sensitive inks, in which three of the inks employed different O₂-sensitive lumophores (PtBP, PdBP and Ru(dpp)₃²⁺), with different τ_o values, encapsulated in the same polymer, PVB, and three used the same lumophore, PtBP, encapsulated in three very different polymers (PVB, PS and PVC). Each ink was used in the same μR-TVC experiment as described in Figure 1, involving an inoculum of 1 mL of 10⁴ CFU/mL of E. coli in 9 mL of growth medium, followed by incubation at 37 °C, and monitoring τ vs. t. The results of this work, namely, τ(air), τ_o, K_sv, and TT are listed in

Table 3. Different O₂ sensors and their properties and recorded TT values in growth media at 37 °C.

Dye	Polymer	τo/μs	τ(air)/μs	KSV/%O2−1	TT */h
PtBP	PVB	41.6	20.4	0.050	3.80
	PS	46.2	11.0	0.152	3.80
	PVC	45.7	30.2	0.024	3.85
PdBP	PVB	290.0	72.0	0.144	3.60
[Ru(dpp)32+(Ph4B2)2]	PVB	6.0	5.6	0.003	3.15

*: for an E. coli inoculum of 10⁴ CFU/mL.

and show that in all cases, despite a significant variation in K_sv, TT varies very little, as predicted by the model-based simulations reported earlier.

For example, for the three O₂-sensitive inks in which the dye was fixed (PtBP) and the polymer (and thus k_q) varied, the measured values of TT differed only by only ca. 1%, despite K_sv exhibiting a 6-fold change. Similarly, for the three O₂-sensitive inks in which the polymer, PVB, (and so k_q) was fixed and the dye was varied, the measured values of TT differed by only 17%, for a 36-fold change in K_sv. The short TT value (3.15 h) for the [Ru(dpp)₃²⁺(Ph₄B₂)₂]/PVB ink appears to suggest it would be a better (faster) sensor than the PtBP/PVB ink (3.80 h), but a comparison of the lifetimes of the former with those of the latter, in the absence and presence of air (i.e., 6.0 and 5.6 μs, cf. 46.2 and 20.4 μs) shows a much lower dynamic range, which would make, in practice, the accurate measurement of TT more difficult. Following on from this work, the PtBP/PVB ink was used in all subsequent studies as it yielded a lifetime range (approx. 41.6–20.4 μs) and sensitivity (K_sv = 0.050%O₂⁻¹) that were similar to that of a commercial O₂ sensor (from Oculer Ltd., Tipperary, Ireland) used in their μR-TVC instrument (τ_o and τ(air) = 43.0 and 20.8 μs, respectively; K_sv = 0.051%O₂⁻¹); they also used the same dye (PtBP) and so had the same emission λ(max) (770 nm). Thus, the PtBP/PVB ink O₂ sensor generated in this work should be an inexpensive alternative to the Oculer proprietary sensor (see Figure S1 in the ESI) and therefore compatible for use with the Oculer Rapid 930 and its τ vs. t recording technology.

3.2. Characterisation of the PtBP/PVB O₂ Sensor in Growth Medium

The sensitivity of the PtBP/PVB O₂-sensitive sensor was assessed by measuring the variation in its lifetime, τ, upon exposure to the LB growth medium used in μR, saturated with different blends of O₂/Ar gas of known %O₂ at 20 °C. Note that room temperature (RT), 20 °C, was chosen for this work, rather than the usual 37 °C employed in μR work; this was for the purpose of simplicity, given that warming up saturating gases to a fixed temperature much above RT is not simple and provides little additional information. The resulting plot of τ vs. %O₂ is illustrated in Figure 3.

The Stern–Volmer plot of the data in Figure 3 is illustrated in the insert plot, from which a value for K_sv = 0.041 ± 0.001%O₂⁻¹ was calculated, which is consistent with that (0.050%O₂⁻¹) reported in Table 3 for the same PtBP/PVB sensor at 37 °C, given P_m(O₂) (and so k_q) increases with increasing T.

In other work, the 90% response and recovery times of the PtBP/PVB sensor to aerobic and anaerobic conditions were measured at 20 °C. In this work, the lifetime of the sensor was monitored with respect to t as it was subjected to a continuous cycle of air and argon purging. Figure 4 illustrates the results generated, from which 90% response and recovery times of 70 and 80 s, respectively, were calculated. The latter times are much shorter that the transition period over which the %O₂ changes sharply from 21 to 0%, ca. 45 min, see Figure 1; thus, it would appear that the response of the O₂ sensor is sufficient for use in μR-TVC.

3.3. μR-TVC Using a PtBP/PVB Ink and the Oculer Rapid 930

In a typical experiment, serially diluted samples of the E. coli stock solution of a known concentration, spanning the range 10⁸–10¹ CFU/mL, were prepared and used (1 mL) to inoculate the sterile growth medium (9 mL) in the 15 mL Falcon™ tube ‘bioreactor’, with a PtBP/PVB sensor ‘dot’ set in its base (see Figure S1 in the ESI). Each inoculated tube was then placed in the Oculer Rapid 930, where it was incubated at 37 °C. The recorded variations in the lifetime of the O₂ sensor ‘dot’, τ, as a function of incubation time, t, are illustrated in Figure 5a.

From the results illustrated in Figure 5a, τ_o and τ(air) = 42.0 and 18.6 μs, so that τ_TT = 30 μs, and this value was used to determine the value for TT for each inoculum. All such μR experiments were repeated in triplicate, and the average values of log(CFU/mL) and TT were then plotted against each other, as illustrated in Figure 5b. As in all μR-TVC studies, this straight-line calibration graph was then used to determine the bacterial load/TVC of E. coli in any subsequent test sample from its measured value for TT.

The above results show that the inexpensive, easy-to-make PtBP/PVB ink ‘dot’ sensor can act as a ready substitute for the more expensive (see Table 1) PtBP-based proprietary Oculer O₂ sensor. Confirmation of the latter was achieved by using the same experimental setup to generate the equivalent τ vs. t profiles and log(CFU/mL) vs. TT plot, as illustrated in Figure 5a and Figure 5b, respectively, but this time using the proprietary Oculer O₂ sensor ‘dot’ set in the base of a Falcon™ tube. The results of this latter set of experiments are illustrated in S2 in the ESI and are near-identical to those illustrated above. In particular, the log(CFU/mL) vs. TT calibration graph has an almost identical gradient (−1.08) and intercept (8.24), see Figure S2b in the ESI, compared to that recorded using the system with a PtBP/PVB ink sensor (−1.06 and 8.15), see Figure 5b.

The reproducibility of the PtBP/PVB ink method for producing an O₂ sensor dot for μR was tested by depositing 10 sensor ‘dots’ into 10 Falcon™ tubes and then using each of these in a series of μR-TVC runs, as described in Figure 1, with the Oculer Rapid 930 at 37 °C and an inoculum of 10⁴ CFU/mL of E. coli. The results of this work are illustrated in S8 in the ESI. Each of the 10 Falcon™ tubes produced near identical ‘S’-shaped plots of τ vs. t, from which the following average values for τ(air), τ_o and TT, 19.9 ± 0.4 μs, 41.6 ± 0.1 μs and 3.83 ± 0.02 h, respectively, were calculated. The reproducibility of these key operational parameters is striking and underlined by the fact that the coefficient of variation of τ(air), ca. 2.0%, is similar to that of 1.7% reported by others for the same parameter based on 20 identical commercial O₂ sensors used for μR, which employed the same dye, PtBP [22].

Finally, the stability of the PtBP/PVB ink dot when used in the Oculer Rapid 930 at 30 °C was tested in the absence and presence of O₂ over 30 days. In this work, a Falcon™ tube containing the PtBP/PVB sensor dot was filled with 10 mL of water and 250 mg of sodium sulfite (and efficient O₂ scavenger), alongside another with just 10 mL of water. These two filled tubes were then sealed, and the lifetimes of the O₂ sensors set in their bases were monitored over 30 days. The results are illustrated in S9 of the ESI and reveal that the measured parameters, τ_o and τ(air), exhibit negligible drift (<1%) over the 30-day monitoring period. These results show the PtBP/PVB ink dot when used in the Oculer Rapid 930 is very stable in aqueous solution, in the absence and presence of O₂, even over a long period of time (30 d). This latter feature is very desirable since, along with reproducibility, the stability of O₂ sensor is paramount in μR-TVC, since for some aerobes, the measured value of TT can be many days, depending upon the growth kinetics of the bacterial species under test and the incubation temperature.

3.4. Micro-Respirometry Using a PtTFPP/PVB Ink

The PtBP/PVB ink is attractive as it can be used as a direct substitute for the proprietary, expensive commercial PtBP-based O₂ sensors that use the same dye. In this work, it has been used to replace the O₂-sensitive dot in a Falcon™ tube that is sold by the company for use with the Oculer Rapid 930 for measuring the TVC of aerobes. It should also be possible to use this ink to replace any commercial O₂ sensor that uses the same dye, see Table 3. However, the PtBP dye is quite expensive, at ca. USD 779 per 100 mg [27], and although the chemical cost of a PtBP/PVB O₂ sensor ‘dot’ (in terms of chemicals) is only ca. USD 0.011 each, see S10 in the ESI, it would be good to reduce this further, especially as the O₂ sensor dots are used as a consumable in μR. With this in mind, PtTFPP represents an attractive alternative to PtBP, as it is a much less expensive O₂-sensitive dye, at ca. USD 132 USD per 100 mg [28], with a lifetime (τ_o = 40.1 μs) similar to that of PtBP (τ_o = 41.6 μs). Thus, a PtTFPP/PVB ink was used as an alternative to the PtBP/PVB ink to make the O₂ sensor used in a series of μR-TVC runs that were otherwise identical to those reported in Figure 5, and the results of this work are illustrated in Figure 6. In this work, τ for the PtTFPP/PVB ink sensor was measured using a custom-built lifetime probe with a 530 nm phase-modulated excitation LED, as described in S5 of the ESI.

As in Figure 5, the data in the τ vs. t profiles, generated using the PtTFPP/PVB ink and illustrated in Figure 6a, were used to produce the straight-line log(CFU/mL) vs. TT calibration graph shown in Figure 6b, with m and c values that are very similar to those determined for the PtBP/PVB ink sensor, see Figure 5. The PtTFPP/PVB ink has the advantage of a much lower chemical cost (7× lower) than the PtBP/PVB ink, both of which are considerably less expensive than the retail cost of any commercial O₂ sensor; Table 1 presents a comparison of the calculated costs per sensor ink dot in S10 in the ESI.

4. Conclusions

The popular method of μR for measuring TVC utilises luminescence-based O₂ sensors, which are expensive as they are difficult to fabricate. The latter can be readily replaced using much less expensive, easy-to-make, ink-based sensors that utilise the same O₂-sensitive dyes. The sensitivity of such inks can be readily increased by using dyes with long lifetimes in the absence of O₂, τ_o, and/or an ink resin/polymer with a high O₂ permeability. Response modelling of the μR-based TVC system and subsequent testing using a range of O₂ sensors of different sensitivity show that there is little to be gained by making the O₂ sensor either very sensitive or insensitive and that the best O₂ sensors for this work use a dye, such as PtBP, with a τ_o value that falls in the range of ca. 40–50 μs, as this is relatively easy to measure. Further work shows that the PtBP/PVB ink can act as a substitute for a much more expensive O₂ sensor used in a commercial system for measuring TVC based on μR, and that the PtBP can itself be replaced by the much less expensive O₂-sensitive dye, PtTFPP. It is likely that most of the inexpensive O₂-sensitive inks reported here can serve as substitutes for an expensive commercial counterpart based on the same O₂-sensitive dye, such as those sold by Oculer, PreSens and Ocean Insights, cf. Table 1 and Table 3. Finally, it is worth noting that O₂-μR can only be used for studying aerobes, or, like E. coli, facultative anaerobes (under an aerobic atmosphere), although anaerobes are often just as important, if not more so, in the same fields, such as food safety, water quality assessment, and clinical practice. Thus, this group is currently engaged in exploring ways to measure the respiration of anaerobes using non-O₂-based sensors.