SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results and Discussion
  6. Summary and Suggestions for Future Study
  7. Acknowledgments
  8. References

A 2010 study exposed Staphylococcus aureus to ultraviolet (UV) radiation and thermal heating from pulsed xenon flash lamps. The results suggested that disinfection could be caused not only by photochemical changes from UV radiation, but also by photophysical stress damage caused by the disturbance from incoming pulses. The study called for more research in this area. The recent advances in light-emitting diode (LED) technology include the development of LEDs that emit in narrow bands in the ultraviolet-C (UV-C) range (100–280 nm), which is highly effective for UV disinfection of organisms. Further, LEDs would use less power, and allow more flexibility than other sources of UV energy in that the user may select various pulse repetition frequencies (PRFs), pulse irradiances, pulse widths, duty cycles and types of waveform output (e.g. square waves, sine waves, triangular waves, etc.). Our study exposed Escherichia coli samples to square pulses of 272 nm radiation at various PRFs and duty cycles. A statistically significant correlation was found between E. coli’s disinfection sensitivity and these parameters. Although our sample size was small, these results show promise and are worthy of further investigation. Comparisons are also made with pulsed disinfection by LEDs emitting at 365 nm, and pulsed disinfection by xenon flash lamps.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results and Discussion
  6. Summary and Suggestions for Future Study
  7. Acknowledgments
  8. References

Ultraviolet (UV) radiation was first demonstrated as a germicidal agent in 1877 (1–3). Ultraviolet germicidal irradiation (UVGI) damages the DNA of viruses and bacteria, rendering it incapable of replicating. The Illuminating Engineering Society of North America (IESNA) (4) and the Deutsches Institut für Normung (DIN) (5) have adopted standardized germicidal efficiency curves that peak in the UV-C range near 265 nm (UV-C: 180–280 nm). These curves also include some of the UV-B wavelengths (UV-B: 280–315 nm) and none of the UV-A wavelengths (UV-A: 315–400 nm). Research has been performed on the germicidal quality of UV-A and visible radiation, and suggests that the longer UV wavelengths have more of a “threshold” type of damage mechanism than the shorter UV wavelengths, which was noted decades ago by Hollaender (6) and this wavelength dependence was investigated in greater detail in 2008 by Vermeulen et al. (7) Various types of UVGI lamps have been used for air (8), water (9,10) and surface disinfection (11,12), and an overview of the current UV radiation technologies for air and surface disinfection was recently published (13).

Pulsed xenon flash lamps have been used for water and surface disinfection, particularly in the food industry (14–16). These flash lamps provide the benefit of delivering a broadband, powerful germicidal exposure to the sample in a short period of time. A 2005 study of Escherichia coli by Wang et al. (17) indicated that the shape of the action spectrum for pulsed UV radiation peaks near 270 nm, and is similar to that for continuous-wave (CW) UV radiation. A 2010 study of pulsed disinfection of Staphylococcus aureus by Krishnamurthy and Irudayaraj indicated that incoming pulses not only could render the irradiated cells unable to replicate, but could also cause photophysical cell damage from the repeated disturbance from incident pulses. Examples of such damage included cytoplasm shrinkage, cell wall and/or membrane damage, and even destruction of the organisms (e.g. cell-wall rupture and cellular content leakage) (18). Disinfection with visible and/or UV-A radiation appears to be mainly driven by these effects, and is qualitatively consistent with the “threshold” patterns observed by Hollaender and Vermeulen for disinfection with longer wavelengths (6,7). These effects have also been described as “pulsed light disintegration” (13).

Research has been encouraged into improving the effectiveness of pulsed disinfection by optimizing the pulse width and pulse repetition frequency (PRF). Xenon flash lamp systems are somewhat limited in this area. Systems can be made with adjustable PRF settings, although the ability to adjust the pulse duty cycle is constrained by concerns with overheating the flash lamp. The recent developments in light-emitting diode (LED) technology offer a promising avenue for investigation in this area. LEDs that emit in the UV-C range require less power, and offer a greater range of PRFs, pulse lengths and duty cycles. Further, technology improvements will continue to make LEDs more cost-effective.

Research on disinfection with UV-emitting LEDs has already begun. In 2007, Hamamoto et al. (19) performed a study of CW UV water disinfection with LEDs that emitted at 365 nm. As reported in their study, after 315 J cm−2 of 365 nm exposure, E. coli populations were reduced by 5.7 log units, indicating a sensitivity of 0.018 log kills per J cm−2 of 365 nm exposure. In 2010, Li et al. (20) performed a study of using pulsed LEDs emitting at the same wavelength, disinfecting samples of Candida albicans and E. coli. The pulses were emitted at various PRFs and duty cycles, and their results indicated that the germicidal effectiveness did vary with these parameters. One of their measurements exposed samples at 50% duty cycle and PRFs of 0.1, 1, 10, 100 and 1000 Hz, and CW exposure. The PRF that yielded the highest sensitivity for disinfection of both pathogens was 100 Hz. Another of their measurements exposed samples at a fixed PRF of 100 Hz, and varying duty cycles at 25%, 50% and 75%, and CW exposure. The duty cycle that yielded the highest sensitivity in that test was the 25% duty cycle. The enhancement in sensitivity was an approximately five-fold increase in log kills per J cm−2radiant exposure compared with continuous irradiation.

Our study exposed samples of E. coli to pulsed UV-C radiation from an LED, which emitted at 272 nm. Samples were exposed to pulsed exposures at PRFs of 0.25, 0.5, 1, 5, 10, 25, 50 and 100 Hz, and duty cycles at 10%, 25%, 50%, 75% and 90%, and also CW exposure. Sensitivity to disinfection was determined, and estimates of the time-effectiveness and cost-effectiveness for disinfection were calculated.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results and Discussion
  6. Summary and Suggestions for Future Study
  7. Acknowledgments
  8. References

Experimental setup.  A schematic diagram of the experimental setup is shown in Fig. 1. An Agilent 33220A function/arbitrary waveform generator provided the current to the LED, and the waveform generator was configured to produce square pulses at 6.8 V. A 105 ω resistor was placed in series with the LED in order to regulate the current. Thirty samples were exposed to pulses emitted at various PRFs, and these were: 1, 5, 10, 25, 50 and 100 Hz. Further, these pulses were emitted at different duty cycles: 10%, 25%, 50%, 75% and 90%. Three additional samples received pulses emitted at 0.5 Hz, at 25%, 50% and 75% duty cycle. Two samples received CW exposure.

image

Figure 1.  Schematic diagram of experimental setup.

Download figure to PowerPoint

The illumination source was a UVTOP270 LED from Sensor Electronic Technology, Inc. The diode emitted a band of wavelengths centered at 272 nm, with a full width half-maximum of 10 nm. The LED’s spectral emission was measured with an Ocean Optics HR2000CG-UV-NIR spectrometer, and is displayed in Fig. 2. The spectral emission did appear to significantly change as the temporal output of the LED was adjusted. Two sample waveforms of the pulse emissions were measured by a Tektronix TDS 3032B oscilloscope with a Thorlabs DET2-SI detector, and are shown in Fig. 3a–b. The LED had a flat window exit optic, which diverged at approximately 60° half angle. The sample was placed at 2.5 inches below the LED, which ensured uniform illumination. At this location, the incident CW irradiance was 5.5 μW cm−2. The power consumption from the LED was 204 mW while operating.

image

Figure 2.  Spectral distribution of the LED.

Download figure to PowerPoint

image

Figure 3.  (a,b) Two sample waveforms of pulses emitted by the LED. Figure (a) illustrates pulses emitted at 100 Hz and a 10% duty cycle, and figure (b) illustrates pulses emitted at 50 Hz and a 50% duty cycle.

Download figure to PowerPoint

Radiometric measurements of the pulses were performed with an International Light IL1700 radiometer system, using a SED033 detector that was specially calibrated with a QNDS2 neutral filter and “wide eye” diffuser, for pulsed measurements in the UV-C range. The radiant exposures received by the samples were as follows: the 10%, 25% and 50% duty cycle samples received 250 μJ cm−2, 500 μJ cm−2 and 1.0 mJ cm−2 respectively; the 75%, 90% and 100% duty cycle samples received 1.3 mJ cm−2 each. The difference in radiant exposures provided to the samples was necessitated by the much-longer time periods required to accumulate the radiant exposure at the lower duty cycle emission settings.

E. coli sample preparation.  Overnight liquid cultures of E. coli DH1 were prepared in tryptic soy broth (TSB) and used to seed tryptic soy agar (TSA) plates. E. coli was serially diluted 1:10 in phosphate-buffered saline from 10−1 to 10−6 dilution. Then 10 μL spots from each dilution were plated onto the TSA plate, representing final dilutions of 10−3 to 10−8. After the agar plates were placed underneath the LED, the top cover of the agar plates was removed during the exposure to UV emissions (control samples received no exposure). All plates were incubated overnight at 37°C and colonies counted to determine the killing effect of the UV-C radiation. Correlation of the radiant exposure measured in mJ cm−2 required for various levels of log inactivation was determined. The initial number of colony forming units per milliliter (cfu mL−1) is abbreviated as N0, the remaining cfu mL−1 is abbreviated as N. The sensitivity was then calculated by the following equation: [(-log N/N0)/mJ cm−2].

Statistical analysis.  A two-way (two factor) analysis of variance (ANOVA) was used to analyze the sensitivity for disinfection measurements. The ANOVA test compared the levels of each factor (PRF and duty cycle) and determined whether any of the levels had statistically different log kills per radiant exposure averages compared with the other levels within that factor. To determine whether a factor had levels with statistically significant differences in average log kills per J cm−2, the P-value for each factor, calculated from the ANOVA test, was analyzed. If the P-value for duty cycle or pulses per second was less than the predetermined alpha (α) of 0.05, then it was concluded that at least two of the levels had statistically significant averages, and a Tukey post hoc test was used to determine exactly which pairs of levels were significantly different for a specific factor. Using an α cutoff of 0.05 meant that we were allowing at most a 5% chance of finding a statistically significant factor or significant difference between levels of a factor by random chance. In essence, the calculated P-value was the probability that the same results and differences would have been found assuming that the null hypothesis (all levels of a specific factor have equal means for kills per J cm−2) was true. Additionally, two-way ANOVA test calculated how much of the observed variation in the kills per J cm−2 calculations was accounted for by each factor in the study.

Surface plots were also included in the statistical analysis to illustrate combinations of levels for the two factors that had higher averages of log kills per J cm−2. With duty cycle on the y-axis and PRF on the x-axis, bands were created within the surface plots to display average log kills per J cm−2 for combinations of the two factors.

The same two-way ANOVA method was used to analyze how PRF and duty cycle affected the time-effectiveness and energy-effectiveness calculations. However, because the energy-effectiveness and time-effectiveness calculations were calculated in a similar fashion, the statistical results were exactly the same for the two measures.

Results and Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results and Discussion
  6. Summary and Suggestions for Future Study
  7. Acknowledgments
  8. References

Variations in disinfection sensitivity with duty cycle and PRF

Figure 4 illustrates the varying sensitivity of E. coli for disinfection, as a function of duty cycle and PRF. Statistically significant correlations were found between sensitivity and duty cycle (P-value < 0.001), and between sensitivity and PRF (P-value = 0.017). The Tukey post hoc test revealed that the disinfection sensitivities could be separated into three statistically distinct groups according to duty cycle (10%, 25% and 50–90% duty cycle). This test also indicated that the variations in sensitivity with PRF were less distinct, as only the sensitivities determined for samples illuminated at 1 vs 100 Hz PRF were statistically distinct.

image

Figure 4.  Measured sensitivities of Escherichia coli for pulsed 272 nm disinfection are shown, varying with duty cycle and pulse repetition frequencies (PRFs). As shown in the figure, there appears to be an inverse relationship with sensitivity vs duty cycle, and (to a less-consistent extent) sensitivity vs PRF. The sensitivity measured for CW irradiation and pulsed xenon irradiation is also provided.

Download figure to PowerPoint

ANOVA results indicated that 86% of the variations in log kills per J cm−2 was accounted for by the changes in PRF and duty cycle. In an ANOVA setting, accounting for 86% of the variation is very high and indicates that, from the measurements taken, these two factors largely explained any differences observed in the variation of log kills per J cm−2.

The observed patterns in the data are summarized in Table 1, including a comparison with the pulsed UV disinfection study with 365 nm LEDs by Li et al. The two CW samples are also provided for comparison in Fig. 4, although there was not a sufficient number of CW samples to include in the statistical analysis.

Table 1. Variations in disinfection sensitivity with duty cycle and repetition rate, and comparison to pulsed disinfection with 365 nm LED.
 272 nm365 nm, Li et al.
E. coli sensitivity to pulsed vs CW emissionsPulsed sensitivity up to 3.8× higher than CW sensitivityPulsed sensitivity up to 5.2× higher than CW sensitivity
Pulsed emissions: effect of varying PRFIn general, sensitivity decreased as PRF increasedSensitivity increased with PRF from 0.1 to 100 Hz, then decreased at PRF = 1 kHz
Pulsed emissions: effect of varying duty cycleSensitivity decreased as duty cycle increasedSensitivity decreased as duty cycle increased
Pulsed parameters for highest sensitivity1 Hz 10% duty cycle100 Hz 25% duty cycle
E. coli sensitivity to 272 vs 365 nm radiation E. coli was consistently more sensitive to 272 nm radiation. Results ranged from 35× to 142× more disinfection sensitivity, depending on the PRF and duty cycle settings in each study

It is not known why the observed changes in sensitivity with PRF were not always consistent with the sensitivity changes observed in the study by Li et al. (20). This may be because the damage mechanisms from pulsed UV-C and UV-A radiation were likely affecting different areas of the cell structure (18). It is also worth noting that the culture methodology used for this test differed from that used by Li et al. (20). In the Li et al.’s study, sensitivity was assessed for biofilms of E. coli K12 prepared by adding 1 mL of an E. coli K12 suspension (1.0 × 107 CFU mL−1) in Luria-Bertani broth to sample wells and aerobically incubated for 2 days, whereas the current study assessed the sensitivity of E. coli cultured on agar plates. Considering that less active cells/biofilms are more resistant to nonspecific damage due to a reduced rate of replication and enzymatic activity, as well as up-regulation of defense mechanisms, one could argue that the difference in sensitivity was simply related to differences in gene expression. Future studies should investigate this question by testing for the biomarkers involved in cell damage from pulsed UV-C exposures.

Exposure for the most time-effective disinfection

For the user community, it is necessary to also consider which emission characteristics would provide the most rapid disinfection. For example, even though the most sensitivity in our study was measured for 1 Hz and 10% duty cycle, it is also true that a sample that is exposed to 1 Hz and 50% duty cycle will receive its radiant exposure five times more quickly, and may be disinfected faster. Using the sensitivity data from Fig. 4, we can calculate the rate of disinfection that one might expect over longer periods of time. The results of these calculations are shown in Fig. 5. For comparative purposes, the time-effectiveness of pulsed xenon systems for E. coli disinfection were also calculated, from results reported by MacGregor et al. (14) and Rowan et al. (21). Figure 5 shows that the four samples of CW and pulsed xenon disinfection showed the fastest disinfection rates, although more samples would be needed to increase the statistical validity of that observation. Also, time-effectiveness for LED disinfection could be improved with different designs of LEDs (e.g. LEDs with less-divergent, more tightly focused beams illuminating the sample). Statistically, ANOVA results indicated that duty cycle was significantly correlated with time-effectiveness (P-value < 0.001), and the correlation with PRF was not significant (P-value = 0.182). The Tukey post hoc test revealed that only the samples illuminated at 90% duty cycle led to time-effectiveness values that were statistically distinct from the rest.

image

Figure 5.  Predicted emission parameters that would provide the most time-effective disinfection are shown, along with calculations of the time-effectiveness from two studies with pulsed xenon flash lamps.

Download figure to PowerPoint

Exposure for the most energy-efficient disinfection

Another key area of interest for the user community is energy efficiency. As of this writing, no LED-based UV-C disinfection system is known to exist. However, we can use the data from this study to estimate the potential energy savings associated with the use of UV-C LEDs for disinfection. When the test LED was illuminating the sample, its power consumption was 204 mW, and the irradiance incident on the samples was 5.5 μW cm−2. Using these data, and the sensitivity data from Fig. 4, we can calculate the most energy-efficient parameters for disinfecting samples. The results are summarized in Fig. 6, and for comparative purposes, the energy efficiency of pulsed xenon systems used in studies by MacGregor et al. (14) and Rowan et al. (21) for E. coli disinfection was also calculated. As shown in the figure, one would expect the LED pulsing at 10% duty cycle to provide the most energy-efficient disinfection. Statistically, ANOVA results indicated that duty cycle was significantly correlated with energy-effectiveness (P-value < 0.001), and the correlation with PRF was not significant (P-value = 0.182). The Tukey post hoc test revealed that only the samples illuminated at 90% duty cycle led to energy- effectiveness values that were statistically distinct from the rest.

image

Figure 6.  Predicted emission parameters that would provide the most cost-effective disinfection, along with the calculations of the cost-effectiveness from two studies with pulsed xenon flash lamps.

Download figure to PowerPoint

This is only a theoretical analysis of one model of LED, however. The energy efficiency could be improved further by using LEDs designed to focus the emissions more directly to the sample, as suggested previously. Also, these estimates may not apply to UV disinfection of air and water, as the disinfection efficiency would be decreased because pathogens would continuously flow past the disinfecting source (16). In order to determine the most cost-effective method for disinfection, a side-by-side comparison of the available UVGI systems would need to be performed, including not only their electricity costs, but also the overall cost to build the systems, and their periodic maintenance costs as well.

Summary and Suggestions for Future Study

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results and Discussion
  6. Summary and Suggestions for Future Study
  7. Acknowledgments
  8. References

LEDs provide a flexible, highly portable, energy-efficient method for disinfection, and pulsed emissions appear to provide a promising method for enhancing this technology. More research is encouraged in the following areas:

  •  The duty cycles and PRFs that provide the optimum disinfection sensitivity should be investigated further. This study provides a starting point for such an effort, but a more thorough set of samples and cross checks are necessary to clarify the dependence of disinfection sensitivity with duty cycle and PRF.
  •  The wavelength dependence of the different damage mechanisms for disinfection should be investigated further, to include the biomarkers associated with cell damage, and the effects induced by UV-A and visible radiation that may occur below the lethal threshold level. This could help lead to more-efficient disinfection systems that use combinations of multiple wave length sources, possibly emitting at different temporal rates (e.g. pulsed vs CW).
  •  Time-effectiveness and cost-effectiveness comparisons of UVGI systems using the various technologies (i.e. LEDs, low- and medium-pressure mercury lamps, and pulsed xenon flash lamps) are encouraged. The suitability of these systems for air, surface and water disinfection should continue to be investigated.
  •  The IES and DIN action spectra for disinfection do not include wavelengths higher than the UV-B range (4,5). It may be beneficial for a new standardized action spectrum to be developed, which includes UV-C, UV-B, UV-A and visible wavelengths.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results and Discussion
  6. Summary and Suggestions for Future Study
  7. Acknowledgments
  8. References

Acknowledgements— The views expressed in this article are the views of the authors and do not reflect the official policy or position of the Department of the Army, the Department of Defense or the U.S. government. Use of trademarked name does not imply endorsement by the U.S. Army but is intended only to assist in identification of a specific product.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results and Discussion
  6. Summary and Suggestions for Future Study
  7. Acknowledgments
  8. References
  • 1
    Downes, A. and T. P. Blunt (1877) The influence of light upon the development of bacteria. Nature16(402), 218.
  • 2
    Downes, A. and T. P. Blunt (1877) Researches on the effect of light upon bacteria and other organisms. Proc. Roy. Soc. Lond.26, 488500.
  • 3
    Downes, A. and T. P. Blunt (1878) On the influence of light upon protoplasm. Proc. Roy. Soc. Lond.28, 199212.
  • 4
    Illuminating Engineering Society of North America (IESNA) (2000) Nonvisual effects of radiant energy. In IESNA Lighting Handbook, 9th Edition (Edited by M.S. Rea), pp. 151155. IESNA, New York, NY.
  • 5
    Deutsches Institut für Normung (DIN) (2000) Strahlungsphysik im optischen Bereich und Lichttechnik. DIN Document 5031-10:2000-03, Beuth Verlag GmbH, Berlin, Germany.
  • 6
    Hollaender, A. (1943) Effect of long ultraviolet and short visible radiation (3500 to 4900 Å) on Escherichia Coli. J. Bacteriol.46(6), 531541.
  • 7
    Vermeulen, N., W. J. Keeler, K. Nandakumar and K. T. Leung (2008) The bactericidal effect of ultraviolet and visible light on Escherichia coli. Biotechnol. Bioeng.99(3), 550556.
  • 8
    Commission International de l’Eclairage (CIE) (2003) Ultraviolet Air Disinfection. CIE Document 155:2003, CIE, Vienna.
  • 9
    Hill Jr., W. F., F. E. Hamblet, W. H. Benton and E. W. Akin (1970) Ultraviolet devitalization of eight selected enteric viruses in estuarine water. Appl. Microbiol.19(5), 805812.
  • 10
    Hoyer, O. (1998) Testing performance and monitoring of UV systems for drinking water disinfection. Water Supply16(1/2), 419442.
  • 11
    Rastogi, V. K., L. Wallace and L. S. Smith (2007) Disinfection of Acinetobacter baumannii-contaminated surfaces relevant to medical treatment facilities with ultraviolet C light. Milit. Med.172(11), 11661169.
  • 12
    Stibich, M., J. Stachowiak, B. Tanner, M. Berkheiser, L. Moore, I. Raad and R. F. Chemaly (2011) Evaluation of a pulsed-xenon ultraviolet room disinfection device for impact on hospital operations and microbial reduction. Infect. Control Hosp. Epidemiol.32(3), 13.
  • 13
    Kowalski, W. (2009) The Ultraviolet Germicidal Irradiation Handbook —UVGI for Air and Surface Disinfection. Springer-Verlag, New York.
  • 14
    MacGregor, S. J., N. J. Rowan, L. McIlvaney, J. G. Anderson, R. A. Fouracre and O. Farish (1998) Light inactivation of food-related pathogenic bacteria using a pulsed power source. Lett. Appl. Microbiol.27, 6770.
  • 15
    Fine, F. and P. Gervais (2003) Efficiency of pulsed UV light for microbial decontamination of food powders. J. Food Protect.67(4), 787792.
  • 16
    Pataro, G., A. Muñoz, I. Palgan, F. Noci, G. Ferrari and J. G. Lyng (2011) Bacterical inactivation in fruit juices using a continuous flow Pulsed Light (PL) system. Food Res. Int.44, 16421648.
  • 17
    Wang, T., S. J. MacGregor, J. G. Anderson and G. A. Woolsey (2005) Pulsed ultra-violet inactivation spectrum of Escherichia coli. Water Res.39, 29212925.
  • 18
    Krishnamurthy, K. and J. Irudayaraj (2010) Microscopic and spectroscopic evaluation of inactivation of Staphylococcus aureus by pulsed UV light and infrared heating. Food Bioprocess Technol.3, 93104.
  • 19
    Hamamoto, A., M. Mori, A. Takahashi, M. Nakano, N. Wakikawa, M. Akutagawa, T. Ikehara, Y. Nakaya and Y. Kinouchi (2007) New water disinfection system using UVA light-emitting diodes. J. Appl. Microbiol.193, 22912298.
  • 20
    Li, J., K. Hirota, H. Yumoto, T. Matsuo, Y. Miyake and T. Ichikawa (2010) Enhanced germicidal effects of pulsed UV-LED irradiation on biofilms. J. Appl. Microbiol.109, 21832190.
  • 21
    Rowan, N. J., S. J. MacGregor, J. G. Anderson, R. A. Fouracre, L. McIlvaney and O. Farish (1999) Pulsed-light inactivation of food-related microorganisms. Appl. Environ. Microbiol.65(3), 13121315.